Integral Abutment and Jointless Bridges

THE 2005 FHWA CONFERENCE
Integral Abutment and Jointless Bridges

(IAJB 2005)

March 16 18, 2005

Baltimore, Maryland

Organized by:
Constructed Facilities Center
College of Engineering and Mineral Resources
West Virginia University

Conference Sponsors:
Federal Highway Administration USDOT
West Virginia Department of Highways - WVDOT

ii
iii

TABLE OF CONTENTS

Session I: Current Practices with Design Guidelines and Foundation Design

V. Mistry 3

Integral Abutments and Jointless Bridges (IAJB)
2004 Survey Summary
R. Maruri, S.Petro 12

The In-Service Behavior of Integral Abutment Bridges:
Abutment-Pile Response
R. Frosch, M. Wenning, V. Chovichien 30

New York State Department of Transportation's Experience with
Integral Abutment Bridges
A. Yannotti, S. Alampalli, H. White 41

Integral Abutment Design and Construction: The New England Experience
D. Conboy, E. Stoothoff 50

VDOT Integral Bridge Design Guidelines
K. Weakley 61

Session II: Case Studies

Case Study: A Jointless Structure to Replace the Belt Parkway Bridge
Over Ocean Parkway
S. Jayakumaran, M. Bergmann, S. Ashraf, C. Norrish 73

Case Study Jointless Bridge Beltrami County State Aid Highway 33
Over Mississippi River in Ten Lake Township, Minnesota
J. Wetmore, B. Peterson 84

Design and Construction of Dual 630-foot, Jointless, Three-span
Continuous Multi-girder Bridges in St. Albans, West Virginia,
United States, Carrying U.S. Route 60 over the Coal River
J. Perkun, K. Michael 97

Integral Abutment Bridges with FRP Decks Case Studies
V. Shekar, S. Aluri, H. GangaRao 113

iv

New Mexicos Practice and Experience in Using Continuous Spans
for Jointless Bridges
S. Maberry, J. Camp, J. Bowser 125

Integral Abutment Bridges Iowa and Colorado Experience
D. Liu, R. Magliola, K. Dunker 136

Moose Creek Bridge Case Study of a prefabricated Integral Abutment
Bridge in Canada
I. Husain, B. Huh, J. Low, M. McCormick 148

Session III: Maintenance and Rehabilitation

Field Data and FEM Modeling of the Orange-Wendell Bridge
C. Bonczar, S. Brea, S. Civjan, J. DeJong, B. Crellin, D. Crovo 163

Integral Abutment Pile Behavior and Design Field Data and FEM Studies
C. Bonczar, S. Brea, S. Civjan, J. DeJong, D. Crovo 174

Effects of Restraint Moments in Integral Abutment Bridges
M. Arockiasamy, M. Sivakumar 185

Full-Scale Testing of an Integral Abutment Bridge
S. Hassiotis, J. Lopez, R. Bermudez 199

Analysis and Design of Integral Abutment by LRFD Method
Y. Deng, J. Farre, J. Chang, P. Penafiel 211

Behavior of Pile Supported Integral Abutments
E. Burdette, S. Howard, E. Ingram, J. Deatherage, D. Goodpasture 222

Soil Structure Analysis of Integral Abutment Bridges
P. Christou, M. Hoit, M. McVay 233

Behavior of Two-Span Integral Bridges Unsymmetrical About the Pier Line
D. Knickerbocker, P. Basu, E. Wasserman 244

Session IV: Construction Practices

Field Study of Integral Backwall with Elastic Inclusion
E. Hoppe 257

v
Plastic Design of Steel HP-Piles for Integral Abutment Bridges
P. Huckabee 270

Integral-Abutment Bridges: Geotechnical Problems and Solutions Using
Geosynthetics and Ground Improvement
J. Horvath 281

P-y Curves from Pressuremeter Testing at Kings Creek Bridge,
WV Route 2, Hancock County, West Virginia
W. Kutschke, B. Grajales 292

Effective Temperature and Longitudinal Movement in
R. Oesterle, J. Volz 302

Transverse Movement in Skewed Integral Abutment Bridges
R. Oesterle, H. Lotfi 312

Soil-Structure Interaction of Jointless Bridges
O. Kerokoski, A. Laaksonen 323

Author Index 337
vi

1
SESSION I:
CURRENT PRACTICES WITH DESIGN
GUIDELINES AND FOUNDATION
DESIGN

2

3


Vasant C. Mistry; Federal Highway Administration; Washington, D. C.

ABSTRACT

The most frequently encountered corrosion problem involves leaking
expansion joints and seals that permit salt-laden run-off water from the roadway
surface to attack the girder ends, bearings and supporting reinforced concrete
substructures. Because neither the materials used nor the pains taken to mitigate
joint leakage can fully resolve these problems, other options such as, the
construction of jointless bridges, the use of integral or semi-integral abutments,
and moving the joints beyond the bridges should be sought.

Since 1987, numerous States have adopted integral abutment bridges as
structures of choice when condition allow. At least 40 States are now building
some form of jointless bridges. While superstructures with deck-end joints still
predominate, the trend appears to be moving toward integral.

This paper presents some of the important features of integral abutment
and jointless bridge design and some guidelines to achieve improved design. The
intent of this paper is to enhance the awareness among the engineering
community to use Integral Abutment and Jointless Bridges wherever possible.

WHY JOINTLESS BRIDGES?

One of the most important aspects of design, which can affect structure life
and maintenance costs, is the reduction or elimination of roadway expansion
joints and associated expansion bearings. Unfortunately, this is too often
overlooked or avoided. Joints and bearings are expensive to buy, install, maintain
and repair and more costly to replace. The most frequently encountered corrosion
problem involves leaking expansion joints and seals that permit salt-laden run-off
water from the roadway surface to attack the girder ends, bearings and supporting
reinforced concrete substructures. Elastomeric glands get filled with dirt, rocks
and trash, and ultimately fail to function. Many of our most costly maintenance
problems originated with leaky joints.

Bridge deck joints are subjected to continual wear and heavy impact from
repeated live loads as well as continual stages of movement from expansion and
contraction caused by temperature changes, and or creep and shrinkage or long
term movement effects such as settlement and soil pressure. Joints are sometimes
subjected to impact loadings that can exceed their design capacity. Retaining
hardware for joints are damaged and loosened by snowplows and the relentless
pounding of heavy traffic. Broken hardware can become a hazard to motorists,

4
and liability to owners. Deck joints are routinely one of the last items installed on
a bridge and are sometimes not given the necessary attention it deserves to ensure
the desired performance. While usually not a significant item based on cost,
bridge deck joints can have a significant impact on a bridge performance. A wide
variety of joints have been developed over the years to accommodate a wide
range of movements, and promises of long lasting, durable, effective joints have
led States to try many of them. Some joint types perform better than others but
all joints can cause maintenance problems.

Bearings also are expensive to buy and install and more costly to replace.
Over time steel bearings tip over and seize up due to loss of lubrication or buildup
of corrosion. Elastomeric bearings can split and rupture due to unanticipated
movements or ratchet out of position.

Because of the underlying problems of installing, maintaining and repairing
deck joints and bearings, many States have been eliminating joints and associated
bearings where possible and are finding out that jointless bridges can perform
well without the continual maintenance issues inherent in joints. When deck
joints are not provided, the thermal movements induced in bridge superstructures
by temperature changes, creep and shrinkage must be accommodated by other
means. Typically, provisions are made for movement at the ends of the bridge by
one of two methods: integral or semi-integral abutments, along with a joint in the
pavement or at the end of a reinforced concrete approach slab. Specific
guidelines for designing and detailing jointless bridges have not yet been
developed by AASHTO so the States have been relying on established experience

A 1985 FHWA report on tolerable movement of highway bridges examined
580 abutments in 314 bridges in the United States and Canada. Over 75 percent
of these abutments experienced movement, contrary to their designers intent,
typically much greater movement vertically than horizontally. The following
paragraph is from the report.

The magnitude of the vertical movements tended to be substantially greater
than the horizontal movements. This can be explained, in part, by the fact that in
many instances the abutments moved inward until they became jammed against
the beams or girders, which acted as struts, thus preventing further horizontal
movements. For those sill type abutments that had no backwalls, the horizontal
movements were often substantially larger, with abutments moving inward until
the beams were, in effect, extruded out behind the abutments.

The use of expansion joints and bearings to accommodate for thermal
movements does not avoid maintenance problems; rather, the provision to these
items can often facilitate such problems.

5
In this 40-year national experience, many savings have been realized in initial
construction costs by eliminating joints and bearings and in long-term
maintenance expenses from the elimination of joint replacement and the repair of
both super and substructures. Designers should always consider the possibilities
of minimum or no joint construction to provide the most durable and cost-
effective structure. Steel superstructure bridges up to 400 ft. long and concrete
superstructure bridges up to 800 ft.
long have been build with no joints,
even at the abutments.

The impact on the total project
cost and quality is best illustrated by
the figure shown on the right. As is
seen, the decisions made at the
design stage account for over 80
percent of the influence on both cost
(first and life-cycle) and quality
(service life performance) of the
structure. Decisions made in the initial stages of design establish a program that
is difficult and costly to change once detailed design or construction begins.

The following quote is very appropriate for bridge engineering:

Quality is never an accident. It is always the result of high intention, sincere
effort, intelligent direction, and skillful execution. It represents the wise choice of
many alternatives.

This is especially true when the Engineer begins the task of planning, designing
and detailing a bridge structure. The variables are many, each of which has a
different, first and life cycle, cost factor. The question to be asked continuously
through the entire process is what value is added if minimum cost is not selected?
Another question to be asked is what futures should be incorporated in the
structure to reduce the first and life cycle cost and enhance the quality? Most of
the variables are controlled by the designer. These decisions influence the cost
and quality of the project; for better or for worse!

WHAT IS AN INTEGRAL ABUTMENT BRIDGE?

Integral abutment bridges are
designed without any expansion joints
in the bridge deck as shown by the
figures on the right. They are generally
designed with the stiffness and
flexibilities spread throughout the
structure/soil system so that all supports
accommodate the thermal and braking

6
loads. They are single or multiple
span bridges having their
superstructure cast integrally with their
substructure. Generally, these bridges
include capped pile stub abutments.
Piers for integral abutment bridges
may be constructed either integrally
with or independently of the
superstructure. Semi-integral bridges
are defined as single or multiple span
continuous bridges with rigid, non-
integral foundations and movement
systems primarily composed of
integral end diaphragms, compressible backfill, and movable bearings in a
horizontal joint at the superstructure-abutment interface.

WHY INTEGRAL ABUTMENTS?

As stated earlier, integral abutment and jointless bridges cost less to construct
and require less maintenance then equivalent bridges with expansion joints. In
addition to reducing first costs and future maintenance costs, integral abutments
also provide for additional efficiencies in the overall structure design. Integral
abutment bridges have numerous attributes and few limitations. Some of the
more important attributes are summarized below.

Simple Design- Where abutments and piers of a continuous bridges are each
supported by a single row of piles attached to the superstructures, or where self-
supporting piers are separated from the superstructure by movable bearings, an
integral bridge may, for analysis and design purposes, be considered a continuous
frame with a single horizontal member and two or more vertical members.

Jointless construction - Jointless construction is the primary attribute of the
integral abutment bridges. The advantages of jointless construction are numerous
as has been stated earlier.

Resistance to pressure - The jointless construction of integral bridges distributes
longitudinal pavement pressures over a total superstructure area substantially
greater than that of the approach pavement cross-section.

Rapid construction - Only one row of vertical piles is used, meaning fewer piles.
The back wall can be cast simultaneously. Fewer parts are required. Expansion
joints and bearings are not needed. The normal delays and the costs associated
with bearings and joints installation, adjustment, and anchorages are eliminated.

7
Ease in constructing embankments - Most of the embankment is done by large
earth moving and compaction equipment requiring only little use of hand operated
compaction equipment.

No cofferdams - Integral abutments are generally built with capped pile piers or
drilled shaft piers that do not require cofferdams.

Vertical piles (no battered piles) - At abutment a single row of vertical piles is
used.

Simple forms - Since pier and abutment pile caps are usually of simple rectangle
shape they require simple forms.

Few construction joints are required in the integral abutment bridges, which
results in rapid construction.

Reduced removal of existing elements - Integral abutment bridges can be built
around the existing foundations without requiring the complete removal of
existing substructures.

Simple beam seats - Preparation of load surface for beam seat can be simplified
or eliminated in integral bridge construction.

Greater end span ratio ranges - Integral abutment bridges are more resistant to
uplift. The integral abutment weight acts as a counterweight. Thus, a smaller end
span to interior span ratio can be used without providing for expensive hold-
downs to expansion bearings.

Simplified widening and replacement - Integral bridges with straight capped-
pile substructures are convenient to widen and easy to replace. Their piling can
be recapped and reused, or if necessary, they can be withdrawn or left in place.
There are no expansion joints to match and no difficult temperature setting to
make.

The integral abutment bridge acts as a whole unit.

Lower construction costs and future maintenance costs.

Improved ride quality - Smooth jointless construction improves vehicular riding
quality and diminishes vehicular impact stress levels.

Design efficiency - Design efficiencies are achieved in substructure design.
Longitudinal and transverse loads acting upon the superstructure may be
distributed over more number of supports.

8
For example, the longitudinal load distribution for the bent supporting a two
span bridge is reduced 67 percent when abutments are made integral instead of
expansion. Depending upon the type of bearings planned for expansion
abutments, transverse loadings on the same bent can be reduced by 67 percent as
well.

Added redundancy and capacity for catastrophic events - Integral abutments
provide added redundancy and capacity for catastrophic events. Joints introduce a
potential collapse mechanism into the overall bridge structure. Integral abutments
eliminate the most common cause of damage to bridges in seismic events, loss of
girder support. Integral abutments have consistently performed well in actual
seismic events and significantly reduced or avoided problems such as back wall
and bearing damage, associated with seat type jointed abutments. Jointless design
is preferable for highly seismic regions.

Improve Load distribution - Loads are given broader distribution through the
continuous and full-depth end diaphragm.

Enhance protection for weathering steel girders

Tolerance problems are reduced or eliminated - The close tolerances required
with expansion bearings and joints are eliminated or reduced with the use of
integral abutments.

RECOMMENDED BEST PRACTICES

The following best practices are believed to contain the key elements to
ensure quality improvements in designing and constructing Integral Abutment and
Jointless Bridges.

Develop design criteria or office practices for designing integral abutment
and jointless bridges.

In extending the remaining service lives of existing bridges, develop
criteria for evaluating and retrofitting bridges with joints to integral or
semi-integral structures.

Establish an annual workshop between joint specialists of various State to
exchange information in the areas of design, construction and maintenance
of joints and jointless bridges since there is continuing innovation and
changing technology. This will help leverage the expertise of limited
manpower in all the States and allow more effective communication of
What works and what does not.

The decision to install an approach slab should be made by the Bridges
and Structures Office, with consultation from the Geotechnical group.

9
The decision should be based upon long-term performance and life cycle
costs, rather than just first costs to the project.

Standardize practice of using sleeper slabs at the end of all approach slabs.
An irregular crack and pavement settlement typically develops at the
interface of the approach slab and the approach pavement. Develop a
method to control and seal this cracking, and if not already provided,
develop a method to channel the water coming through this crack away
from the pavement without allowing material to be washed away.

RECOMMENDED DESIGN DETAILS FOR INTEGRAL ABUTMENTS

Use embankment and stub-type abutments.
Use single row of flexible piles and orient piles for weak axis bending.
Use steel piles for maximum ductility and durability.
Embed piles at least two pile sizes into the pile cap to achieve pile fixedly
to abutment.
Provide abutment stem wide enough to allow for some misalignment of
piles.
Provide an earth bench near superstructure to minimize abutment depth
and wingwall lengths.
Provide minimum penetration of abutment into embankment.
Make wingwalls as small as practicable to minimize the amount of
structure and earth that have to move with the abutment during thermal
expansion of the deck.
For shallow superstructures, use cantilevered turn-back wingwalls
(parallel to center line of roadway) instead of transverse wingwalls.
Provide loose backfill beneath cantilevered wingwalls.
Provide well-drained granular backfill to accommodate the imposed
expansion and contraction.
Provide under-drains under and around abutment and around wingwalls.
Encase stringers completely by end-diaphragm concrete.
Paint ends of girders.
Caulk interface between beam and backwall.
Provide holes in steel beam-ends to thread through longitudinal abutment
reinforcement.
Provide temporary support bolts anchored into the pile cap to support
beams in lieu of cast bridge seats.
Tie approach slabs to abutments with hinge type reinforcing.
Use generous shrinkage reinforcement in the deck slab above the
abutment.
Pile length should not be less than 10 ft. to provide sufficient flexibility.
Provide prebored holes to a depth of 10 feet for piles if necessary for
dense and/or cohesive soils to allow for flexing as the superstructure
translates.

10
Provide pavement joints to allow bridge cyclic movements and pavement
growth.
Focus on entire bridge and not just its abutments.
Provide symmetry on integral bridges to minimize potential longitudinal
forces on piers and to equalize longitudinal pressure on abutments.
Provide two layers of polyethylene sheets or a fabric under the approach
slab to minimize friction against horizontal movement.
Limit use of integral abutment to bridges with skew less than 30 degree to
minimize the magnitude and lateral eccentricity of potential longitudinal
forces.

SUMMARY

There are many advantages to jointless bridges as many are performing well
in service. There are long-term benefits to adopting integral bridge design
concepts and therefore there should be greater use of integral bridge construction.
Due to limited funding sources for bridge maintenance, it is desirable to establish
strategies for eliminating joints as much as possible and converting/retrofitting
bridges with troublesome joints to jointless design.

The National Bridge Inventory database notes that eighty percent of the
bridges in the United States have a total length of 180-ft. or less. These bridges
are well within the limit of total length for integral abutment and jointless bridges.
Where jointless bridges are not feasible, installation of bridge deck joints should
be done with greater care and closer tolerances than normal bridge construction to
achieve good performance. Since 1987, numerous States have adopted integral
abutment bridges as structures of choice when conditions allow. At least 40
States are now building integral and/or semi-integral abutment type of bridges.
Preference range from Washington State and Nebraska, where 80-90 percent of
structures are semi-integral; to California and Ohio, which prefer integral, but use
mix, depending upon the application; to Tennessee, which builds a mix of both
integral and semi-integral, but builds integral wherever possible.

While superstructures with deck-end joints still predominate, the trend appears
to be moving toward integral. Although no general agreement regarding a
maximum safe-length for integral abutment and jointless bridges exists among the
state DOTs, the study has shown that design practices followed by the most DOTs
are conservative and longer jointless bridges could be constructed.

There are several activities underway that will affect the way States are
designing jointless bridges in the future. These include a joint AASHTO/NCHRP
task force responsible for initiating and drafting AASHTO design guide
specifications and synthesis report on current practices for integral and
semi-integral abutment bridges, FHWA-sponsored research study on Jointless

11
Bridges, update of LRFD specs to address jointless bridge design issues, and
future workshops. An excellent reference document on current issues regarding
jointless bridges is the FHWA Region 3 Workshop manual on Integral Abutment
Bridges, November 1996.

Continuity and elimination of joints, besides providing a more maintenance
free durable structure, can lead the way to more innovative and aesthetically
pleasing solutions to bridge design. As bridge designers we should never take the
easy way out, but consider the needs of our customer, the motoring public first.
Providing a joint free and maintenance free bridge should be our ultimate goal.
The best joint is no joint.

REFERENCES

1. Wasserman, Edward P. And Walker, John H., Integral Abutments for Steel Bridges, October
1996.
2. Burk, Martin P., Jr., An Introduction to the Design and Construction of Integral Bridges,
FHWA, West Virginia DOT and West Virginia University, Workshop on Integral
Bridges, November 13-15, 1996.

12
Integral Abutments and Jointless Bridges
(IAJB) 2004 Survey Summary

Rodolfo F. Maruri, P.E.
1
and Samer H. Petro, P.E.
2

ABSTRACT

Integral Abutment and Jointless Bridges (IAJB) have been used for decades
and the criteria for using them and detailing has varied from state to state. The
main advantage of IAJB is the elimination of joints, which after they start leaking,
account for 70% of the deterioration that occurs at the end of girders, piers and
abutment seats. FHWA promotes the usage of Integral Abutment and Jointless
bridges, where appropriate, as one method of building bridges that will last 75-
100 years with minimal maintenance.

In 1995 and 1996, Federal Highway Administration (FHWA) in conjunction
with the Constructed Facilities Center (CFC) at West Virginia University (WVU)
conducted a survey and workshop about Integral Abutment Bridges [1]. In 2004,
another survey [2] was developed by FHWA and the CFC at WVU, using similar
questions as the 1995 survey and incorporating additional questions, to obtain a
status of usage and design for Integral Abutments and Jointless bridges. The
survey was distributed by AASHTO subcommittee on Bridge and Structures to all
50 states Department of Transportation (DOT), District of Columbia DOT, Puerto
Rico Highway and Transportation Authority and Federal Lands Highways
Division (referred to as states in the paper).

This paper summarizes the responses received to date from the states. The
survey was divided into different topic areas which included General Issues,
Design and Details, Foundation, Abutment/Backfill, Approach Slabs, Retrofit
(Jointed to Jointless), and Other Issues. Integral Abutments, as defined in the
survey and in this paper, refers to the monolithic construction of the abutment
with the deck in order to eliminate the joints at the end of the bridge. This
includes the use of Full, Semi Integral Abutments and Deck Extensions. Jointless
bridges refers to the elimination of joints at the piers through the usage of integral
pier caps, continuous spans and continuous for live load construction.

The purpose of the survey was to obtain a snapshot about the usage of integral
abutments and jointless bridges from the states, their policy, their design criteria
and other issues. The results of the survey are presented in this paper and will be
used to disseminate information between states and help FHWA encourage the
usage of IAJB.

1
Rodolfo F. Maruri, P.E., Federal Highway Administration, Richmond, Virginia.
2
Samer H. Petro, P.E., Gannett Fleming, Inc., Morgantown, West Virginia.

13
INTRODUCTION
Integral abutments and jointless bridges (IAJB), when properly designed and
constructed, perform better than bridges with expansion joints because they
minimize maintenance, extend service life of bridge components including
bearings, abutment and pier seats, paint system and superstructure. Although
integral abutments have been designed and constructed successfully for decades,
the design and analysis of these structures have relied primarily on a fragmented
body of technical references, and design and construction details have varied from
state to state.

The Federal Highway Administration (FHWA) in conjunction with the
Constructed Facilities Center (CFC) of West Virginia University (WVU)
conducted a survey of integral abutment design and construction as part of a
three-day workshop on integral abutment and jointless bridges scheduled for
March 16-18, 2005 in Baltimore, Maryland. The IAJB 2004 survey was sent to
all 50 States Department of Transportation (DOT), DC DOT, Puerto Rico
Highway and Transportation Authority and the Federal Lands Highway Division
(referred to as states in the paper). The survey was conducted with the intention
that bridge designers and owners will use the information to promote usage and
design practices for integral abutment and jointless bridges.

The IAJB 2004 survey include questions about the number of integral
abutments designed, built and in service, the criteria used for design and
construction, including maximum span lengths, total length, skews and curvature
and problems experienced with integral abutment bridges. In addition, the survey
questioned the states about their design considerations such as thermal movement,
passive earth pressure, approach slabs, foundation and pile design and retrofitting
of non-integral abutments to integral abutments.

For consistency in the terminology used, the survey defined and provided a
sketch of full integral abutments, semi integral abutments, and deck extensions.
Full Integral Abutment was described as a capped pile stub type abutment with
or without a hinge between superstructure and foundation cap, semi-integral
abutment was described as a rigid, non-integral foundation with movement
system primarily composed of internal end diaphragms and movable bearings in a
horizontal joint at the superstructure-abutment interface and a deck extensions
was described as extension the deck over the top of the backwall and place joint
behind the abutment backwall to prevent deterioration of the end of superstructure
beams [2].

Of the fifty-three (53) states surveyed, thirty-nine (39) states (74%) responded
(Figure 1). In addition, other states indicated that they will submit the responses
to the survey in the future.

14

Figure 1: States that Responded to the IAJB 2004 Survey

According to the responses, there are approximately 13000 integral abutment
(IA) bridges, of which approximately 9000 are full integral abutment bridges,
approximately 4000 are semi-integral abutment bridges and approximately 3900
deck extension bridges in-service (Table 1). The increase in the number of
integral abutments from the numbers reported in the 1995 survey [1] can be
attributed to the acceptability of the benefits of integral abutments, familiarity
with design and construction issues and a larger sample of responding states (39
respondents in 2004 versus 18 in 1995). The numbers reported are approximate
since the National Bridge Inventory (NBI) data, which is kept by all the states
with information about their bridges, does not differentiate between the different
types of abutments and most states do not have other methods for maintaining an
inventory bridges and/or integral abutments.

The following sections in the paper, discuss the survey answers, analysis and
compilation of answers and conclusions based on responses provided.

GENERAL ISSUES
This section of the survey questioned the states about their use of integral
abutments since the last workshop [1], the number of integral abutment bridges in
service, the states policy for design of jointless bridge construction and the
criteria used for integral abutments and jointless bridges. A breakdown of
integral bridges designed and built since 1995 and the total number of in-service
integral abutment bridges is shown in Table 1.

15
Table 1: Number of IAJB Designed and Built Since 1995 and In-Service

DESIGNED
since 1995
BUILT
since 1995
IN SERVICE
(TOTAL)
Integral Abutment ~ 7000 ~ 8900 ~ 13000
Full Integral ~ 5700 ~ 6400 ~ 9000
Semi Integral ~ 1600 ~ 1600 ~ 4000
Deck Extension ~ 1100 ~ 1100 ~ 3900

The survey responses indicate an increase in the number of integral abutments
of over 200% in the last 10 years. As in 1995, Tennessee continues to have over
2000 integral abutments bridges, but Missouri reports having 4000 integral
abutment bridges, which represent the largest amount of integral bridges. An
increase in the number of integral abutments, since 1995, is most evident in the
northern states where Illinois, Iowa, Kansas and Washington all reported having
over 1000 in-service. In addition, Michigan, Minnesota, New Hampshire, North
Dakota, South Dakota, Oregon, Wyoming and Wisconsin, reported having
between 100-500 integral abutment bridges in-service. Unlike the northern states,
the southern states like Florida, Alabama and Texas do not use integral abutment
and reported having one or less bridges with integral abutments.

As illustrated in Figure 2, fifty-one percent (51%) of the responding states
indicated that they designed and built over 50 integral abutment (IA) bridges since
1995, of which 21% built 101-500 integral abutment bridges, 5% built 501-1000
integral abutment bridges and 5% built over 1000 integral abutment bridges.
3%
8%
10%
23%
18%
5% 5%
3%
18%
10%
8%
21%
5% 5%
15%
21%
0%
10%
20%
None 1 - 10 11 - 20 21 - 50 51 - 100 101 - 500 501 - 1000 Over 1000
NUMBER OF BRIDGES DESIGNED USING INTEGRAL ABUTMENTS (RANGE)
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S
Integral Abutment (Designed)
Integral Abutment (Built)

Figure 2: Percent of States that reported designing and building Integral Abutments within
the range specified (since 1995).

16
Analysis of responses shows a similar distribution with the in-service integral
abutments bridges as in Figure 2. For in-service integral abutment bridges, fifty-
nine percent (59%) of the responding states indicated having over 50 integral
abutment bridges in-service, of which 31% have 101-500 integral abutment
bridges, 3% have 501-1000 integral abutment bridges, and 15% have over 1000
integral abutment bridges. The 2004 survey revealed that approximately 85% of
the states (about 33 states) constructed integral abutment bridges equally with
either simple spans or multiple spans.

The responses to General Question 4, regarding the states criteria for using
integral abutments, show that a majority of the states do not limit the maximum
span within the bridge, but do limit the total length of the bridge and the skew of
the bridge. Table 2 summarizes the criteria range provided by the states for
prestressed concrete girder and steel bridges for maximum span, total length of
bridge, maximum skew of bridge and maximum curvature.

Table 2: Range of Design Criteria Used For Selection of Integral Abutments.

PRESTRESSED
CONCRETE
GIRDERS
RANGE

STEEL GIRDERS RANGE
MAXIMUM SPAN

MAXIMUM SPAN
Full Integral 60 200

Full Integral 65 - 300
Semi Integral 90 200

Semi Integral 65 - 200
Deck extensions 90 200

Deck extensions 80 - 200
Integral Piers 120 200

Integral Piers 100 - 300
TOTAL LENGTH

TOTAL LENGTH
Full Integral 150 1175

Semi Integral 90 3280



Integral Piers 150 - 1000
MAXIMUM SKEW

MAXIMUM SKEW
Full Integral 15 70

Semi Integral 20 45



Integral Piers 15-No Limit
MAXIMUM
CURVATURE

MAXIMUM
CURVATURE

Full Integral 0 10

Semi Integral 0 10


Integral Piers 3 - No Limit

Integral Piers 0 - No Limit

17
Based on the experience of many states, their comments and their established
design criteria, it can be concluded that a majority of new bridges could be built
using integral abutments. Colorado, Iowa and Tennessee indicated that they built
the majority of their new bridges using integral abutments.

The utilization of integral abutments with curved bridges is not widely
accepted based on survey responses. Four states reported that they allow the use
of curved girder bridges with integral abutments and three (3) more allow the
construction of curved bridges with straight girders and integral abutments. An
alternative mentioned to account the forces in curved bridges and/or long bridges
is the use of integral abutments with an expansion joint elsewhere on the bridge.

The IAJB 2004 survey shows that although progress has been made in the
construction of integral abutments since 1995, there is still a lot of variability in
the usage and criteria used for selection of integral abutments. The non-
uniformity of selection criteria for integral abutments indicates that this is an area
where standardization is warranted.

DESIGN AND DETAILS

This section of the IAJB 2004 survey questioned the states regarding their
changes to the design procedures or details, future plans for jointless bridges,
policy regarding the use of integral abutments, forces and loads used to design
integral abutments and other design issues. The following presents the questions
asked in the survey and their respective responses.

Question 1 in this section asked whether the design procedures or details
changed since August of 1995 with regard to loads, substructure design, backfill,
approach slabs and jointless retrofit of bridges. The survey results indicated that
less than 25% of the states that responded changed their design procedures
regarding primary and secondary loads, substructure design, abutment/backfill
and approach slab since 1995. However, this percentage increased with regard to
changing the details associated with the same issues of integral abutments. The
largest change occurred in the details for foundation/substructures and approach
slabs (38% and 36% respectively). Figure 3 illustrates the percentage of states
that responded <YES> in each of the issues noted.

Some of the changes reported in the survey include Iowa specifying a 10-foot
prebored hole filled with bentonite for each pile (8-foot prebored hole used prior
to 2002), Virginias accountability of lateral forces on skewed integral abutment
bridges, and Connecticuts incorporation of approach slabs in all bridges to
minimize bump/settlement problems at the bridge/approach fill interface. In
addition to these detailing changes, several states noted that they have changed
their design to incorporate the Load Factor Design (LFD) specifications and/or
the Load and Resistance Factor Design (LRFD) specifications.

18

24%
23%
21%
15%
18%
8%
10%
13%
36%
28%
36%
13%
0%
10%
20%
30%
40%
Primary Load
Considerations
Secondary Load
Considerations
Substructure/foundation Abutment/Backfill Approach Slab Jointless retrofit
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S

DESIGN PROCEDURES
DETAILS

Figure 3: Percent of States that Reported Changing their Design Procedures and Details
since 1995.

Design and Details Question 2 and 3 asked about the states future plans for
jointless bridge construction, including the future use of integral abutments,
continuous spans, retrofit of existing bridges, and policy about elimination of
joints. The survey revealed that over ninety percent (90%) of the states have a
policy to eliminate as many joints as possible and construct jointless simple and
continuous span bridges whenever possible. However, only 77% indicated that
they will design integral (fully and semi) abutments whenever possible and 79%
noted that they will design bridges as jointless whenever they meet the design
criteria for jointless bridges (Figure 4). The difference in the percentages between
eliminating as many joints as possible (92%) and using integral abutments
whenever possible (77%) can be attributed to states that do not extensively use
chemicals for deicing of bridges in the winter and therefore do not have a policy
of incorporating integral abutments in their bridge design (Figure 5).

Noteworthy comments included Oregons comment about problems with
multiple-span jointless bridges; Arizonas comment about having problems with
integral abutment approach slabs which is the reason Arizona does not use
integral abutments anymore; Vermont noting that they do not use integral
abutment extensively because of scourability issues; and Washington State noting
that they preferred using semi-integral type abutments because they are more
economical since it avoids the transfer of seismic forces into the substructure

19
79%
92%
54%
90%
77%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Design integral (fully and
semi) abutments whenever
possible.
Design bridges as jointless
whenever they meet the
design criteria for jointless
bridges.
Eliminate as many deck joints
as possible.
Retrofit existing bridges and
eliminate deck joints where
possible
Design (jointless) simple and
continuous span bridges
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S

Figure 4: Percent of States that Answered <YES> with Regard to Their FUTURE Plans for
Jointless Bridge Construction.

Figure 5: Future Plans for IAJB Design and Construction

20
Design and Details Question 4 dealt with the forces, including passive and
active earth pressure, temperature, creep, shrinkage, settlement, additional loads
due to skew layout, additional forces due to curvature and other forces that states
account for in the design of integral abutments. The survey revealed that 72% of
the states account for temperature related forces (Figure 6). In addition, states
also noted that they account for temperature (temperature gradient, thermal
expansion and contraction in longitudinal and transverse direction) in their design
(Design and Details Question 5), but the procedure for accounting for the
thermal expansion and contraction varied widely.
The survey results also indicate that 59% of the states surveyed accounted for
passive earth pressures, but only 21% of the states allow for curved bridges with
integral abutments and account for the additional forces due to the curvature of
the bridge (Figure 6).
Noteworthy comments about design of integral abutments include Illinois
practice to designed only for vertical loads, North Dakotas practice to use 1000
lb/ft
2
to account for various loads (passive pressure, thermal, creep and shrinkage
loads) and Iowas use of the a simple, fixed-head pile model which does not
consider passive or active pressure and is based on research conducted by
Greimann and Abendroth at Iowa State University during the 1980s.

59%
33%
44%
41%
21%
15%
72%
28%
0%
10%
20%
30%
40%
50%
60%
70%
Passive Earth
Pressure
Temperature. Creep Shrinkage Settlement Additional forces
due to skew
Additional forces
due to curvature
Other. Describe
below in
comments
section.
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S

Figure 6: Percent of States That Account for the Forces Listed in the Design of Integral
Abutments.

Thirty-three percent (33%) of the responding states account for creep effects
when designing integral abutment bridges (Design and Details Question 4 and
6), while Georgia, Illinois, Iowa and other states indicated that they do not
account for creep movements.

21
As expected, the majority of the states responded that they use computer
software to design integral abutment bridges (78%). However, the program
and/or method used varied widely. Several states, including California, Illinois
and North Dakota indicated that they use hand calculations and charts, while other
states noted that they have developed their own in-house spreadsheet, using Excel
and MathCAD, to design integral abutment bridges. Structural programs and
finite element software like STAAD, STRUDL, and RISA are used by
Pennsylvania, Rhode Island and North Carolina to design integral abutments
while Tennessee, New Hampshire, Virginia and New Jersey use COM624P
and/or L-pile for pile design.

FOUNDATIONS

The monolithic construction of the deck with integral abutment (backwall)
requires special design for the backwall and supporting piles of integral abutments
and jointless bridges. The design of the foundation for integral abutments needs
to account for the expansion and contraction of the bridge due to thermal
movement. The resulting soil pressures due to thermal expansion and restraining
effects due to jointless construction of the bridge have been recognized as the
controlling load for design of integral abutments and piles. Designing and
detailing of integral abutments to handle these forces is critical for the proper
performance of integral abutments.

The 2004 IAJB survey questions where chosen to obtain an understanding
about how states are designing foundations for integral abutments, including
criteria used to select foundation type, type of pile, orientation of pile, pile design
considerations, pressure used in the design of integral abutments and special
details utilized to reduce the pressures at the integral abutment.

The survey responses (Foundation, Questions 1 and 2) indicate that full-
integral abutment with steel bearing piles is the most commonly type of integral
abutments (~ 70%). However, several states noted that they are currently
designing and/or creating standards for semi-integral abutments. The comments
provided indicated that semi-integral abutments are commonly used with the
uncharacteristic designs that incorporate larger skews, higher abutment walls and
unique soil conditions.

Washington State noted they preferred using semi-integral abutments because
they are more economical since they avoid transferring seismic forces into the
substructure. New Hampshire indicated that they use deck extensions extensively
since the foundation design is not an issue. The use of deck extensions is
predominant in the northeast region (New Hampshire, New York, Connecticut
and Maine) as is evident in the large number of in-service deck extensions in this
region.

22
Nevada and Hawaii indicated that in addition to steel bearing piles (H piles
and pipe piles), friction piles and spread footings, they are using drilled shafts for
foundations of integral abutments. Noteworthy, even though steel bearing piles
were the most common type of pile used for integral abutments, there was no
consensus on the typical orientation of the pile (Foundation, Question 4). Thirty
three percent (33%) of the responding states orient the piles with the strong axis
parallel to the centerline of bearing, 46% orient the piles with the weak axis
parallel to the centerline of bearing, 8% (3 states) leave it to the discretion of the
Engineer and the remaining 13% did not provide a comment or noted that the
question was not applicable because of their use of symmetric piles (Figure 7).
The non-uniformity of pile orientation seems to indicate that this is an area where
further standardization is warranted.

33%
46%
8%
13%
0%
10%
20%
30%
40%
50%
Strong Axis Parallel to CL of
Bearings
Weak Axis Parallel to CL of
Bearings
Designer's Option Not Applicable (Symmetric Pile)
or No Answer Provided
ORIENTATION OF PILES
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S

Figure 7: Typical Pile Orientation Use for FULL Integral Abutments

Responses to Foundation Question 3 indicate that a number of states have
developed office practices that allow designers to detail integral abutments
without doing complicated analysis. These states use the office practices in
conjunction with geotechnical recommendations based on soil parameters to
decide the type of foundation used. Based on the comments provided, there is no
evidence of problems relating to the type of foundation used for integral
abutments.

The use of MSE wall has increased dramatically over the years and as such
the use of integral abutments where the MSE walls serves as a component of the
integral abutments has increased correspondingly. Foundations-Question 5,
questioned about the states policy regarding the combination of integral

23
abutments and MSE walls. Based on the survey responses, the preferred detail is
to offset the MSE wall from the integral abutment and footing between two (2)
feet to five (5) feet. According to comments, the offset provides space for
construction of MSE wall and offsetting of MSE straps around abutment piles. In
addition to offsetting of the integral abutment behind the MSE wall, several states
noted that they have special requirements for the placement of piles in the MSE
backfill including the use of sleeves filled with sand. The detailing of MSE
abutments with integral abutments is inconsistent based on the responses received
and is another area where guidelines based on all available research would be
beneficial to states that are currently using this type of construction and/or plan to
use it.

The soil pressure used for the design of integral abutments and its piles has
been the subject of controversy and much research. The survey, Foundation-
Question 6, shows that there is still no consistent design method used with regard
to soil pressures. The majority of the respondents indicated that they use passive
pressure (33%) and/or a combination of passive and active pressures (18%).
Active pressures, however, is used by a minority of respondents (8%) and other
combination of pressure and/or methods was used by 26% of the states
responding (Figure 8). The survey was not specific enough to make any
conclusions about the variability of pressures used in the design of integral
abutments.

18%
8%
26%
33%
0%
10%
20%
30%
Combination Active Pressure Passive Pressure Other
PRESSURE USED FOR DESIGN
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S

Figure 8: Typical Soil Pressures Used for Design of Substructure

24
The limits or capacities used for piles provide another opportunity for
standardization. Based on the comments provided (Foundation-Questions 7, 8
and 9), states use AASHTO in combination with statewide practices that limit
lateral deflection of pile, computer programs and other methods to determine the
capacity of piles. The pile capacity is based on the axial capacity of the pile
(using 0.25*fy as stipulated in AASHTO Standard Specification, section 4.5.7.3
or other) [41% of states], or a combination of axial/bending capacity based on
beam-column analysis and frame analysis [51% of states]. In addition to
accounting for bending due to expansion/contraction of superstructure, 26% of the
states also account for the bending due to superstructure rotation in the horizontal
plane (skew bridges) (Figure 9).

41%
51%
26%
0%
10%
20%
30%
40%
50%
Axial Forces (No Bending) Bending Forces Due to
Expansion/Contraction of Superstructure
Bending Forces Due to Superstructure
Rotation in the Horizontal Plane (skew
bridges)
TYPE OF FORCE USED TO DETERMINE CAPACITY OF PILE
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S

Figures 9: Forces Used to Determine Capacity of Piles for Integral Abutments.

ABUTMENT/BACKFILL

The handling of the backfill behind the integral abutments can have a
significant effect on the performance of integral abutments and as a result has
been discussed and researched over the past decades. A review of the answers
and comments provided in the IAJB 2004 survey (Abutment/Backfill Question
1 and 2), show that most states require the fill behind the integral abutment to be
compacted (69%) as compared to 15% for uncompacted fills. Interestingly, in
addition to using compacted fills there are a number of states that require the use
of expanded polystyrene (EPS), other compressible materials behind abutment,
lightweight fills and additional inspection during construction in order to reduce

25
and/or control the earth pressures exerted on integral abutments during expansion
cycles (Figure 10).

The other survey questions in this section inquire about whether the states
specify the minimum length of approach fill required behind the integral abutment
(Abutment/Backfill - Question 3) and whether the states limit the maximum
height of integral abutments (Abutment/Backfill Question 4). In both questions,
the analysis of the responses indicated that the states specified the length of
approach fills and/or limited the height of the integral abutments 31% of the time,
but the majority of the states did not limit or specified these parameters. The
comments for Abutment/Backfill Question 4 inferred that the limit for the
height applied only to the full integral abutment. Washington State indicated that
they have used a 30-foot high semi-integral abutment.

13%
10% 10%
8%
8%
13%
69%
0%
10%
20%
30%
40%
50%
60%
70%
Require compacted
backfill
Require
uncompacted backfill
Use Expanded
Polystyrene (EPS)
Use other
compressible
material behind
abutment.
Use Lightweight fill Require additional
inspection during
construction
Other. Describe in
comments section
below.
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S

Figure 10: Percent that Responded <YES> to Listed Requirement for Approach Backfill.

APPROACH SLABS
Some of the most common problems associated with integral abutments are
the settlement and the cracking of approach slabs. Fortunately, these problems do
not cause a significant disruption of traffic or a decrease of the service life of the
bridge. The questions in this section were designed to find out the design
procedure and details used for approach slabs (Approach Slabs Questions 1 and
2), the problems experienced with approach slabs (Approach Slabs Question 3)
and the criteria about the usage of approach slabs behind integral abutments
(Approach Slabs Question 4).

26

The comments provided in Approach Slabs - Questions 1 and 2, indicate that
there is no consistency in the detailing of approach slabs. Thirty-one percent
(31%) of respondents indicated that they use a sleeper pad at the end of approach
slab, 26% indicated that they float the slab on the approach fills and 30%
indicated that they do both.

Many states indicated that they have or are using corbels on the abutment
backwall for the support of the approach slab, while other states indicated that
they use reinforcing projecting from the abutment backwall to tie the approach
slab to the abutment backwall, and other states are using a combination. Based on
the responses received, it is evident that the detailing of approach slabs, including
the connection to the abutment backwall and the interface between the approach
slab and approach fills is an area where standardization and guidelines would be
beneficial.

Review of comments provided for Approach Slabs Question 3, indicates
that approach slab settlement, cracking, and bump at the interface between the
approach slab and approach fill are the major problems with approach slabs. The
answers and comments provided with this question are consistent with the
answers provided in Other Issues Question 1 (Figure 11).

In order to mitigate some of the problems with approach slabs, several states
are using buried approach slabs and/or select fills under the approach slab while
other states have filled voids under approach slab with grout, resurfaced approach
slabs with asphalt, and/or used an overlay. Surprisingly, a state noted that the
reasons they do not use integral abutment bridges anymore are because of the
bump formed at the end of the approach slab and settlement problems under
approach slabs due to poor drainage.

RETROFIT

Retrofit of jointed decks to jointless decks has been increasing as the
condition of the decks deteriorates and a complete deck replacement is needed.
Forty-nine percent (49%) of the respondents indicated that they have a policy to
retrofit existing bridges whenever possible. Virginia has used a poor-man
continuity detail, obtained from Utah DOT, for complete and partial re-decking
projects with great success. For re-decking projects, Virginia has found that the
main cost of using a continuous deck with simple spans is the need to retrofit the
existing bearings. As the age of the bridges in the United States continues to
increase and concrete decks need to be replaced to improve rideability and the
condition rating of the decks, consideration should be given to retrofitting existing
simple span to continuous (jointless) deck. Based on the responses in the survey,
this is an area where further guidelines would be beneficial to those states that
have not yet established a retrofit policy.

27
Conversion of non-integral abutment bridges to integral or semi-integral
abutment bridges is not widely used and only 39% of the responding states
indicated that they have a policy to investigate its feasibility. New Mexico notes
that this is very common retrofit and that the selection of bearings is critical;
Missouri notes that they consider retrofitting only on short spans and small skew;
Virginia notes that they try to incorporate this retrofit with major superstructure
replacement projects; and South Dakota notes that they incorporate this type of
retrofit with a major renovation such as a deck replacement or if there are severe
problems at the abutment.
OTHER ISSUES
The last section of the survey inquired about the problems states are having
with integral abutments (Other Issues Question 1), special details/design
procedure for bearing in integral abutments (Other Issues Question 2) and list of
recent research in the area of integral/jointless bridges (Other Issues Question
3). Details and design procedure for bearings and recent research in the area of
integral/jointless bridges will be provided during the IAJB 2005 Conference. A
summary of the problems reported with Other Issues Question 1 is shown in
Figure 11.

The compilation of Other Issues Question 2 revealed that 36% of the states
use the same procedure for design of integral abutment bearings as for regular
abutment bearings, 13% design the bearings based on thermal movements and
38% use other methods to design and detail the bearings of integral abutments.

15%
26%
10%
8%
3%
46%
28%
0%
10%
20%
30%
40%
50%
Cracking of integral
abut. backwall
Cracking of deck at
integral abutment
Cracking of approach
slabs
Cracking of wingwall Detailing Detrimental rotation
of integral abutment
backwall
Setlement of
approach slabs
P
E
R
C
E
N
T

O
F

S
T
A
T
E
S

Figure 11: Problems Experienced with Integral Abutments.

28

CONCLUSIONS / RECOMMENDATIONS

The IAJB 2004 survey revealed that design practices and details vary greatly
from state to state. For example, maximum span limits, curvature and skew
effects, thermal movement limits and creep effects are just a few criteria that
differ considerably between states according to the surveys received. Therefore,
the papers authors believe that standardization and/or guidelines are warranted.

The cost associated with proper maintenance of joints, subsequent
deterioration of bridge components when joints do not perform satisfactorily and
FHWAs goal to built bridges with 75-100 years service life with minimal
maintenance, are some of the reasons that the authors consider that integral
abutment and jointless bridges should be the construction of choice, whenever
feasible. Based on the compilation of the surveys, the following conclusions and
recommendations are made:

1. Develop guidelines that incorporate the research done in the area of
integral abutments. These guidelines should include problems
experienced by other states, as well as design guidelines and examples.
Examples of areas identified in the 2004 IAJB survey where non-
uniformity in design and detailing are apparent include,
a. Criteria used for selection of integral abutments.
b. Forces and pressures used to design integral abutment and integral
abutment piles.
c. Orientation of integral abutment piles.
d. Design of integral abutments with curved bridges.
e. Detailing of approach slab at bridge interface and approach fill
interface.
2. Promote the issuance of a national policy for the use of integral abutments,
especially in states that indicated that they do not have a policy regarding
the future use of integral abutments.
3. Develop guidelines and/or additional information to increase the use of
continuous jointless and/or continuous decks with simple span
superstructures, whenever appropriate.
4. Develop guidance and/or additional information on the use of deck
extensions to eliminate joints at abutments. These guidelines should
incorporate criteria on when to use them, problems experienced by other
states, and design guidelines.
5. Develop guidelines and/or additional information for detailing of integral
abutments around MSE walls.

29
6. Develop guidelines and/or additional information for handling of backfill
behind integral abutments.
7. Develop guidelines and/or additional information for detailing of approach
slabs to minimize cracking and mitigate problems at the approach slab and
roadway fill interface.
8. Develop guidelines and/or additional information for the retrofitting of
existing bridges to eliminate joints at piers and abutments.
9. Create a new survey incorporating comments given in the IAJB 2004
Survey and clarifying areas which were unclear according to submitters.
10. Re-issuance as a Technical Advisory by Federal Highway Administration
and/or policy guidelines for Integral Abutments.

These recommendations and conclusions presented within this paper are the
opinions of the authors, based on the IAJB 2004 survey, and are not necessarily
endorsed by Federal Highway Administration. However, a goal of the conference
is for the CFC at WVU to provide recommendations to Federal Highway
Administration for its consideration and issuance as a Technical Advisory and/or
policy guideline.

ACKNOWLEDGEMENT

The authors would like to thank all the individuals who completed the survey
and returned it to the CFC at WVU. Their work is the basis for the survey
summary and this paper. In addition, we would like to acknowledge the support
of Malcolm T. Kerley, chairman of AASHTO sub-committee on Structures and
Bridges, for sending the IAJB 2004 survey to all the states Bridge Engineers.

REFERENCES

1. GangaRao, H., Thippeswamy, H., Dickson, B. Franco, J., 1996. Survey and Design of
Integral Abutment Bridges, Constructed Facilities Center at West Virginia University,
Morgantown, West Virginia.

2. Maruri, R., Petro, S., GangaRao, H., 2004. IAJB 2004 Survey, Federal Highway
Administration and Constructed Facilities Center at West Virginia University, Morgantown,
West Virginia.

30
The In-Service Behavior of Integral Abutment Bridges:
Abutment-Pile Response

Robert J. Frosch, Purdue University
Michael Wenning, American Consulting, Inc.
Voraniti Chovichien, Thai Engineering Consultants Co, Inc.

ABSTRACT

Integral bridges have been used for many years across many regions of the
country. However, empirical guidelines have often limited their use. While
removal of limits imposed by these guidelines may be warranted, there are many
questions regarding the behavior of these structures that remain unanswered. In
particular, the interaction of the abutment, pile and soil remains uncertain. In
Indiana, the decision to explore extension of the limits has resulted in a study to
ascertain the in-service behavior of integral abutment bridges. Through several
field instrumentations, new light is being shed on the behavior and performance
of these bridges. The behavior of integral abutment bridges is concentrated in the
response of the abutment-pile-soil system. Therefore, this response is the focus of
this paper.

INTRODUCTION

The Indiana Department of Transportation (INDOT) has used empirical limits
for integral bridge construction. These limits are similar to those used by many
departments; namely bridges less than 300 ft in total length with skews no greater
than 30 degrees [1]. Like most departments, INDOT has observed many
advantages to this type of construction and wishes to increase these limits.
Among the leading advantages are simpler construction and reduced maintenance.

Many unanswered questions, however, have kept INDOT from opening the
limits on these designs. Among them are:

How far can empirical details be stretched without additional analysis?
Should H-piles be oriented in their strong or weak axis for best results?
What methods of design should be proscribed when bridges fall outside
the empirical limits?
What components of the bridge require additional design?
How much does the mass of the bridge buffer the daily temperature
changes?
Does the bridge skew enter into the design and if so how?

31
In May 1998, the Indiana Department of Transportation made a decision that
is likely to change the course of bridge design in the state. They authorized the
design of a jointless bridge 302 m (990 ft) long, over 3 times longer than any
previously built in Indiana. This decision began a series of events that resulted in
a research study at Purdue University [2,3]. While the overall goal of the Purdue
research study was to provide answers to the various questions, the first phase of
research was directed towards providing an understanding of the in-service
behavior of these bridges.

SR 249 OVER US 12 THE FIRST BRIDGE

The SR 249 Bridge is designed to carry traffic over US 12 and nine railroad
tracks into the Port of Indiana at the northernmost part of the state. The ten-span
bridge has a 13-degree skew and a total length of 990 ft (302 m). Individual
spans ranged from 87 to 115 ft (26.4 to 35.0 m). The intent of this project was to
build a bridge using standard construction materials and details and monitor it to
evaluate the assumptions that were made in design.

Due to the size of the bridge, alternate plans were produced for both a steel
girder and prestressed bulb-tee option. The continuous, composite prestressed
bulb-tee option received the low bid and was constructed. The four girders were
5 ft deep, made of semi-lightweight concrete (130 pcf) and spaced at 10 ft-4 in.
(3.15 m) on the 38ft-4 in. (11.7 m) wide bridge.

The girders sat on elastomeric bearing pads that were designed to deflect with
the anticipated temperature change. Per INDOT detailing standards, the
diaphragm encapsulating the ends of the beams over the piers were cast the full
width of the bridge seat, or 50 inches in this case. The bottom of the diaphragm
rested on a layer of polystyrene that was intended to allow the superstructure to
expand or contract without locking up at the piers. A keyway is provided to
restrain the superstructure from excessive movement.

The design of the end bent presumed that the bridge length would expand and
contract with annual temperature variations. For this region, AASHTO
temperature ranges of 45 and 60 degrees from construction temperatures were
used for the concrete and steel options, respectively. This amounted to 1.6 in. and
2.3 in. of anticipated movement in each direction.

Once the movement was computed, a model was set up to calculate the
resisting force of the soil. Soil borings at the site revealed that the bridge would
be founded on seams of peat and marl that extended about 45 ft below ground
line. The geotechnical report advised against adding any weight to the existing
spill slopes due to large anticipated settlements. As a result, the approaches for
the bridge had to be constructed of expanded polystyrene fill. This created a
situation that pulled the project out of its standard construction methods

32
Sub-Base
Prestressed
Bulb Tee
Girder
HP14x89
Expanded
Polystyrene
Fill
Approach Slab
Deck
17
Retaining Wall
18
Min. 18
Anchor Plate
and Beam Seat
Sleeper Slab
Ground Level
Sub-Base
Prestressed
Bulb Tee
Girder
HP14x89
Expanded
Polystyrene
Fill
Approach Slab
Deck
17
Retaining Wall
18
Min. 18
Anchor Plate
and Beam Seat
Sleeper Slab
Ground Level
category. The piles supporting the end bents would have to be designed as free
standing for the first 19.68 ft. The computer program COM624P was used to
model the spring coefficients for the various soil layers. The anticipated
deflection at the pile head was inputted to obtain the maximum moment and point
of fixity for the pile. The point of fixity was assumed to be the second location
where the deflection diagram crossed the zero point. The portion of the pile
extending into the cap was covered with polystyrene to obtain a pinned
connection. Piles were then designed as columns with a height from the point of
fixity to the bottom of the end bent.

Steel encased concrete (shell) piles, 14 in. diameter, would have been the first
choice of support for these soil conditions. However, when an analysis was
performed, the thickness of the piles would have been excessive. Larger
diameters were investigated but the same thickness was always required. Since
the stiffness of the pile increases the force required to move it the predetermined
distance, the moment increased linearly with the pile section modulus. It was
determined that the shape of the pile would have to change to obtain a better ratio.

H piles were then investigated in both strong and weak axis orientation.
Ultimately a strong axis orientation was used to avoid the possibility of local
flange buckling. A schematic of the end bent detail is provided in Figure 1.

Figure 1: SR 249 End Bent Detail

The bridge was instrumented with a combination of strain, tilt, crack and
temperature meters. In addition to this bridge, INDOT has continued to build and
instrument others to measure the response of other types of bridges, piles, skews
and soil conditions.

FIELD STUDIES

Overall, four bridges in Indiana have been instrumented to observe the in-
service behavior of integral abutment bridges as well as the behavior of the piles

33
supporting these structures. These bridges range in length from 150 to 990 ft
providing a spectrum of behavioral data [2,3].

While the SR 249 Bridge provided excellent information regarding the
behavior of a relatively long integral structure, this structure was not typical in
regards to the design of the end bent. Therefore, several other bridges including
the SR18 over Mississinewa River Bridge (Figure 2) were also selected for
instrumentation. There are several reasons that the SR18 Bridge in particular was
selected.

1. The bridge was designed and constructed according to typical integral
abutment details.
2. The bridge exceeded the length limitation of INDOT and could provide
much needed data regarding bridge length.
3. The skew of the structure was small. Therefore the research could focus
on the effects of bridge length.

Figure 2: SR18 over Mississinewa River Bridge

To better understand the soil-pile-abutment-system, the five-span, continuous
prestressed, concrete bulb-tee integral bridge was instrumented. This bridge is
located east of the city of Marion in Grant County, Indiana on the westbound
lanes of State Road 18 crossing the Mississinewa River. The total bridge length is
367 ft with a skew angle of 8.

To evaluate the abutment movement, tiltmeters and convergence meters were
provided on the end bents (Bents 1 and 6). The convergence meter was oriented
horizontally and operated perpendicular to the abutment. This meter measured
the relative displacement between the end bent and the reference pile to determine
the longitudinal abutment movement. The locations of the convergence meters,
tiltmeters and pile strain gages are shown in Figure 3.

34
Tiltmeter
10
8 Slab
R/C Bridge
Approach Slab
Prestressed Conc.
Bulb-T Beam
End Bent Backfill
3-3
Strain Gages
1-3
14 CFT pile
Convergence Meter
Reference Pile
3
Drain
Ground Level
Strain Gage
Tiltmeter
10
8 Slab
R/C Bridge
Approach Slab
Prestressed Conc.
Bulb-T Beam
End Bent Backfill
3-3
Strain Gages
1-3
14 CFT pile
Convergence Meter
Reference Pile
3
Drain
Ground Level
Strain Gage
Centerline of
Abutment
1
-
3
Parallel to
Roadway
8
5

s
p
a
c
e
s

a
t

4
=
2
0
Abutment
Pile
Plan
Elevation
N
90
Ground
Level
Strain Gage
Centerline of
Abutment
1
-
3
Parallel to
Roadway
8
5

s
p
a
c
e
s

a
t

4
=
2
0
Abutment
Pile
Plan
Elevation
N
90
Ground
Level
Centerline of
Abutment
1
-
3
Parallel to
Roadway
8
5

s
p
a
c
e
s

a
t

4
=
2
0
Abutment
Pile
Plan
Elevation
N
90
Ground
Level
Strain Gage
Strain gages were installed on piles, not only at ground level but also along
the length of Pile 6 of the western bent (Figure 4) to evaluate the in-service, soil-
structure response and to determine the response of the entire pile rather than only
at the base of the abutment. All strain gages except the ones at ground level were
installed prior to pile driving to provide the strain profile along the length of the
pile enabling investigation of overall pile behavior. The strain gages at ground
level were installed after driving. These gages on Pile 6 allow calculation of pile
bending down the length of the pile and estimate of the deflected shape. Strain
gages on the south face were installed to provide redundancy, locate the neutral
axis, and evaluate out-of-plane movement of the pile.

Figure 3: End Bent Instrumentation (SR18)

Figure 4: Strain Gages along the Pile Length (SR18)

35
-10
0
10
20
30
40
50
60
70
80
90
100
110
120
0
6
/
0
1
/
0
3
0
7
/
0
1
/
0
3
0
7
/
3
1
/
0
3
0
8
/
3
0
/
0
3
0
9
/
2
9
/
0
3
1
0
/
2
9
/
0
3
1
1
/
2
8
/
0
3
1
2
/
2
8
/
0
3
0
1
/
2
7
/
0
4
0
2
/
2
6
/
0
4
0
3
/
2
7
/
0
4
0
4
/
2
6
/
0
4
A
i
r

T
e
m
p
e
r
a
t
u
r
e

(
F
)
Construction Day
Hottest Day
Coldest Day
Construction Temperature
Deck
Girder
-10
0
10
20
30
40
50
60
70
80
90
100
110
120
0
6
/
0
1
/
0
3
0
7
/
0
1
/
0
3
0
7
/
3
1
/
0
3
0
8
/
3
0
/
0
3
0
9
/
2
9
/
0
3
1
0
/
2
9
/
0
3
1
1
/
2
8
/
0
3
1
2
/
2
8
/
0
3
0
1
/
2
7
/
0
4
0
2
/
2
6
/
0
4
0
3
/
2
7
/
0
4
0
4
/
2
6
/
0
4
A
i
r

T
e
m
p
e
r
a
t
u
r
e

(
F
)
Construction Day
Hottest Day
Coldest Day
Construction Temperature
Deck
Girder
ABUTMENT BEHAVIOR

The temperature on the SR18 Bridge was measured by temperature gages
located on a girder and in the deck as shown in Figure 5. The temperature
measured by both gages was almost identical; however, the response of the deck
was slightly slower than that of the girder.

Figure 5: Air Temperature

The rotation of the abutment was measured by tiltmeters located on the east
and west faces of the end bents (Bents 1 and 6). The rotations of the abutments
were filtered by taking the average of the data recorded between the time interval
four hours before and four hours after the desired measurement time. The filtered
rotations of both bents are plotted in Figure 6. The results indicate that both bents
translated and hardly rotated. The date of deck casting is noted as the
construction day. This day is significant in that it signifies the time at which the
structure became integrally connected.

36
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0
6
/
0
1
/
0
3
0
7
/
0
1
/
0
3
0
7
/
3
1
/
0
3
0
8
/
3
0
/
0
3
0
9
/
2
9
/
0
3
1
0
/
2
9
/
0
3
1
1
/
2
8
/
0
3
1
2
/
2
8
/
0
3
0
1
/
2
7
/
0
4
0
2
/
2
6
/
0
4
0
3
/
2
7
/
0
4
0
4
/
2
6
/
0
4
L
o
n
g
i
t
u
d
i
n
a
l

M
o
v
e
m
e
n
t

(
i
n
.
)
Bent 6 Center
Bent 1 Center
Bent 6 SE
Inward
Outward
Construction Day
Hottest Day
Coldest Day
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0
6
/
0
1
/
0
3
0
7
/
0
1
/
0
3
0
7
/
3
1
/
0
3
0
8
/
3
0
/
0
3
0
9
/
2
9
/
0
3
1
0
/
2
9
/
0
3
1
1
/
2
8
/
0
3
1
2
/
2
8
/
0
3
0
1
/
2
7
/
0
4
0
2
/
2
6
/
0
4
0
3
/
2
7
/
0
4
0
4
/
2
6
/
0
4
L
o
n
g
i
t
u
d
i
n
a
l

M
o
v
e
m
e
n
t

(
i
n
.
)
Bent 6 Center
Bent 1 Center
Bent 6 SE
Inward
Outward
Construction Day
Hottest Day
Coldest Day
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0
6
/
0
1
/
0
3
0
7
/
0
1
/
0
3
0
7
/
3
1
/
0
3
0
8
/
3
0
/
0
3
0
9
/
2
9
/
0
3
1
0
/
2
9
/
0
3
1
1
/
2
8
/
0
3
1
2
/
2
8
/
0
3
0
1
/
2
7
/
0
4
0
2
/
2
6
/
0
4
0
3
/
2
7
/
0
4
0
4
/
2
6
/
0
4
R
o
t
a
t
i
o
n

(
d
e
g
r
e
e
s
)
Average Bent 1
Average Bent 6
Construction Day
Hottest Day
Coldest Day
Bent 1
Bent 6
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0
6
/
0
1
/
0
3
0
7
/
0
1
/
0
3
0
7
/
3
1
/
0
3
0
8
/
3
0
/
0
3
0
9
/
2
9
/
0
3
1
0
/
2
9
/
0
3
1
1
/
2
8
/
0
3
1
2
/
2
8
/
0
3
0
1
/
2
7
/
0
4
0
2
/
2
6
/
0
4
0
3
/
2
7
/
0
4
0
4
/
2
6
/
0
4
R
o
t
a
t
i
o
n

(
d
e
g
r
e
e
s
)
Average Bent 1
Average Bent 6
Construction Day
Hottest Day
Coldest Day
Bent 1
Bent 6

Figure 6: Abutment Rotation
The movements of the abutments were measured by convergence meters, and
the displacements are plotted in Figure 7. The data was zeroed immediately prior
to casting. These results indicate that the abutment movement corresponds well
with temperature. For instance, as the temperature decreases (contraction phase),
both abutments move toward each other as anticipated. Furthermore, the
displacements were essentially identical indicating symmetrical behavior.

Figure 7: Abutment Displacement

37
The measured movement of Bent 1 was compared to the thermal movement
calculated according to Equation 1 as shown in Figure 8. It can be seen that the
calculated abutment movements are greater than the measured values. This
difference is most likely due to backfill restraint, pile resistance and friction from
the approach slab.

L = (T)L (1)

where:
= thermal coefficient of concrete, taken as 6.0x10
-6
/F;
T = temperature change
L = half of the total span length, 367 ft/2 = 183.5 ft.

-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0
6
/
0
1
/
0
3
0
7
/
0
1
/
0
3
0
7
/
3
1
/
0
3
0
8
/
3
0
/
0
3
0
9
/
2
9
/
0
3
1
0
/
2
9
/
0
3
1
1
/
2
8
/
0
3
1
2
/
2
8
/
0
3
0
1
/
2
7
/
0
4
0
2
/
2
6
/
0
4
0
3
/
2
7
/
0
4
0
4
/
2
6
/
0
4
L
o
n
g
i
t
u
d
i
n
a
l

M
o
v
e
m
e
n
t

(
i
n
.
)
Bent 1 Center
Inward
Outward
Construction Day
Hottest Day
Coldest Day
Calculated
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0
6
/
0
1
/
0
3
0
7
/
0
1
/
0
3
0
7
/
3
1
/
0
3
0
8
/
3
0
/
0
3
0
9
/
2
9
/
0
3
1
0
/
2
9
/
0
3
1
1
/
2
8
/
0
3
1
2
/
2
8
/
0
3
0
1
/
2
7
/
0
4
0
2
/
2
6
/
0
4
0
3
/
2
7
/
0
4
0
4
/
2
6
/
0
4
L
o
n
g
i
t
u
d
i
n
a
l

M
o
v
e
m
e
n
t

(
i
n
.
)
Bent 1 Center
Inward
Outward
Construction Day
Hottest Day
Coldest Day
Calculated

Figure 8: Calculated vs. Measured Movements

PILE BEHAVIOR

Stresses and strains along the pile length over various temperature change
ranges, T, were determined by grouping the strain according to the temperature
range (Figure 9). The average strains of each temperature range were calculated.
The increment of the temperature change range is 10 F 5% except for T equal
to 0 F. At T = 0 F, the range considered was from -1 to 1 F. It is noted that
the construction temperature was considered as 60 F, and all temperature
changes are referenced from this temperature.

38

Figure 9: Pile Stresses with Depth

Deflections along the pile depth were computed by integrating the moment of
the area under the curvature diagram considering the deflection measured at the
pile top as measured by the convergence meter located at the center of the eastern
bent. The deflected shape of Pile 6 over various temperature change ranges was
estimated as shown in Figure 10. The estimated deflected shapes correspond very
well to the temperature change, T. Double curvature bending occurs with the
inflection point located between a depth of 4 and 8 ft. It should be noted that the
deflection at the bottom of the pile was not directly measured. This value was
assumed for calculation of the displacement shape and was considered reasonable.

Figure 10: Pile Displacement with Depth

39

SUMMARY AND CONCLUSIONS

Based on the field study, the following conclusions have been reached.

1. The abutment responds to temperature changes, and its movement can be
estimated conservatively using the theoretical thermal
expansion/contraction of the superstructure, L = (T)L. The actual
displacement is expected to be slightly less due to backfill restraint, pile
resistance and approach slab friction.

2. The abutment primarily translates or slides longitudinally in response to
thermal expansion and contraction of the bridge. Only minor rotations of
the abutment occur and for analysis purposes can be ignored.

3. Piles integrally connected with the abutment bend in double curvature.
Lateral displacements in the soil correspond directly with temperature
changes. Measures to eliminate the integral abutment-pile connection can
be used such as in the SR249 structure to provide for a pinned connection.
This connection eliminated the double curvature response.

4. For satisfactory bridge performance, the structure must be detailed and
constructed properly.
a. Piles must be constructed and oriented as designed.
b. Intermediate piers should be designed to accommodate lateral
displacement or the connection must be detailed to minimize
lateral force transfer. If the piers are not designed for the lateral
displacement, locking of the superstructure to the intermediate
piers must be prevented through isolation.

The overall goal of the research program was to provide answers to questions
that have limited the design of integral abutment bridges. While the scope of this
paper is limited to the measured response, two additional phases of research have
been conducted including an analytical and laboratory experimental study. The
comprehensive view provided through these three phases of research has provided
answers to most of the questions originally posed. An ongoing study is
completing the investigation into bridge skew. While all three phases of research
have been essential to providing answers, the measured in-service response was
indispensable for not only the calibration of analytical models, but for a true
understanding of behavior.

40
REFERENCES

1. Indiana Department of Transportation. 1992. Bridge Design Memorandum #233
Revised, Inter-Department Communication, Indianapolis, IN, 5 pp.
2. Durbin, K.O. 2001. Investigation of the Behavior of an Integral Abutment Bridge,
Masters Thesis, Purdue University, West Lafayette, IN, 2001, 138 pp.
3. Chovichien, V. 2004. The Behavior and Design of Piles for Integral Abutment Bridges,
Doctoral Dissertation, Purdue University, West Lafayette, IN, 489 pp.

ACKNOWLEDGEMENTS

The authors would like to gratefully acknowledge the Indiana Department of
Transportation. Funding for the research program conducted by Purdue
University was provided through the Joint Transportation Research Program
(JTRP) through Project No. SPR-2393. Thanks are also extended to Katrinna
Durbin and David Fedroff for their contributions to this research study.

41
New York State Department of Transportation's
Experience with Integral Abutment Bridges

Arthur P. Yannotti, P.E. New York State, Department of Transportation
Sreenivas Alampalli, PhD, P.E. New York State Department of Transportation
Harry L. White 2nd, P.E. New York State Department of Transportation

ABSTRACT

The New York State Department of Transportation (NYSDOT) has been
using integral abutment bridges since the late 1970's. Since that time, the design
methodology and details have been modified several times to improve
performance. Semi-Integral abutments were introduced in 1998. Approximately
450 integral and semi-integral abutment bridges have been constructed in New
York and thus far, their in-service performance has been excellent. They are the
preferred abutment type for NYSDOT. This paper examines the evolution of the
design and construction practices and explains the reasons for the modifications.

HISTORY

Traditionally, bridges in New York were constructed with a joint in the
deck slab at abutments to accommodate thermal movements. By the 1970's,
multiple span bridges were commonly designed as continuous spans, thus
eliminating the deck joints over the piers. Deck joints were of a number of types,
including various types of neoprene seals, finger joints and open trough systems.
However, all types were prone to leak and allow water containing road salt to
drain onto the underlying superstructure beams, bearings, abutment backwalls and
bridge seats.

In order to eliminate the "leaking joint" problem, the first integral
abutment bridges constructed by NYSDOT were built in the late 1970's. These
were typically single span steel bridges with a span length under 100 feet. A
single row of steel H piles with the strong axis parallel to the girders was used so
that bending occurred about the weak axis. The steel piles were extended into a
short abutment stem. The steel girders were erected on the steel piles and
attached to them by welding to a cap plate that was welded to the top of the piles.
The girders and piles were then encased in concrete as the deck slab was placed
creating an integral type abutment. Design assumptions included equal
distribution of the vertical load to the piles. Bending moments in the piles were
ignored, and the abutments and wingwalls were designed for full passive earth
pressure. A section of an early typical integral abutment with a steel
superstructure is shown in Figure 1.

42

Figure 1. Early Steel Superstructure Integral Abutment

Superstructures with adjacent prestressed concrete box beams are
commonly used in New York. Early attempts were made in the 1980's to adapt
integral abutments to this type of superstructure. The principal differences were
that there was no positive attachment of the beams to the piles and a sloping
backwall was used to allow the approach slab to be poured continuously with the
deck. Design assumptions were the same as for bridges with steel superstructures.
A section of an early typical integral abutment with adjacent prestressed concrete
beam superstructure is shown in Figure 2.

Integral abutment bridges were occasionally used in the 1970's and 1980's
on an experimental basis. Their performance proved to be very good and other
benefits emerged. The elimination of the footing and the relatively simple
concrete forming, compared to conventional abutments, made them advantageous
because of their lower cost and more rapid construction than conventional
abutments.

43

Figure 2. Early Prestressed Concrete Superstructure Integral Abutment

EVALUATION

By 1996 there were more than 155 integral structures in-service. For
spans longer than 160 feet, steel H-piles with bending about the weak axis were
specified. Maximum allowable bridge length was 500 feet. Most of these bridges
have their principal features in common, although details have varied greatly, and
a study by the NYSDOT Structures Design and Construction [1] concluded that
with few exceptions, these structures have performed well. Typical design details
[2] for integral bridges of that era are shown in Figures 1 and 2.

In 1996, New York initiated two studies to evaluate the in-service
performance of integral abutment and jointless bridges [3-5]. The first study [3]
evaluated 84 integral bridges and 105 jointless bridge decks in-service at the time
the study was initiated, through field inspection of the relevant bridge
components. The second study [4, 5] investigated the validity of some of the
design assumptions regarding soil pressure on abutments and load distribution
among piles by surveying other transportation agencies nationwide and in
Canada. This was conducted to document current design and construction
practices related to integral and jointless bridges and to make further
improvements to New York guidelines.

44
The main objective of the field inspection and analysis of inspection data
[3] was to evaluate general performance of integral abutment and jointless
bridges, and to determine the problem details needing improvement. As part of
the field evaluation, engineers inspected several visible components, influenced
by the abutment details used. These included the abutment itself, approach slab,
the first 5 feet of the deck and the wearing surface near the abutments. Condition
ratings (see Table 1), as per the NYSDOT Bridge Inspection Manual [6], were
utilized to rate these on a 1 (total deteriorated) to 7 (new condition) scale.
Settlement of approach slabs was also recorded during the field inspections. The
attributes of the bridge (such as age, superstructure type, skew, bridge length, etc.)
were obtained from bridge inventory data. A statistical analysis was performed,
as needed, to investigate the effects of various attributes in a rational format to
supplement visual observations. Results indicated that these bridges have been
functioning as designed and showed superior performance when compared to
conventional bridges of similar age and exposure.

Table 1. New York State Department of Transportation Condition Ratings [6]
Rating Description
1 Totally deteriorated or in failed condition
2 Used for conditions between 1 and 3
3 Serious deterioration or not functioning as originally designed
5 Minor deterioration and is functioning as originally designed
7 New condition

Thirty-nine transportation agencies responded to the study which surveyed
several transportation agencies in the United States and Canada [4, 5]. Of the
responders, integral abutment bridges are being built by at least 30 agencies.
Bridge owners rated their in-service performance as "good" or "excellent". They
experienced only such minor problems as cracking and settlement of bridge
approaches, but indicated that their in-service performance is at least equal or
better than conventional bridges. The design practices and assumptions relating
to thermal limits, soil pressures and piles varied considerably and were based
largely on past experience.

MODIFICATIONS

Several modifications were made to integral abutment details as a result of
this study. The most important had to do with the way the abutment backwall,
approach slab and deck slab were placed for bridges with superstructures
consisting of adjacent prestressed concrete box beams. Adjacent prestressed
concrete box beams are a very common type of superstructure in New York, so
this was an issue of considerable importance. Although it is more difficult to
achieve a positive connection with the substructure than is the case with steel
girders, satisfactory results had been obtained by extending prestressing strands

45
and reinforcing bars from the ends of the beams into the abutment backwall. In
order to facilitate construction, the deck slab, approach slab and abutment
backwall had been placed in a single pour (See Figure 2). This detail required the
abutment backwall to be sloped at an approximate 45 angle against the approach
fill so that the approach slab could be placed in the initial pour without backfilling
against the abutment backwall.

The study showed that this older detail was unsatisfactory because
significant transverse cracking in the deck and transverse and longitudinal
cracking in the approach slab near the abutment typically occurred. It was
determined that this was caused by the inability of the approach slab to
accommodate any settlement of the approach fill. A revised detail was developed
(See Figure 3) that allowed rotation of the approach slab at the abutment when
settlement occurred. This was accomplished by using a vertical backwall, placing
the deck slab and approach slabs in separate pours with a formed joint and
eliminating the horizontal reinforcing steel across the deck slab transition to
approach slab joint. Reinforcing bars are now placed at 45 into both the deck
slab and approach slab. This detail minimizes the moment capacity at the joint
and allows the approach slab to rotate if there is settlement of the approach fill.

Figure 3. Current Prestressed Concrete Superstructure Integral Abutment

Integral abutments with steel superstructures typically used a detail where
reinforcement was run horizontally between the deck slab and approach slab

46
beneath the saw cut joint to tie them together. The study showed there was a
similar cracking pattern to that observed with prestressed concrete
superstructures. A detail at the approach slab/ deck slab/ backwall interface
similar to the prestressed concrete superstructure detail was adopted. It has
proven satisfactory in reducing cracking at these locations.

Because of the generally very good performance of integral abutments,
other changes in design criteria were made at this time. The allowable exposed
height of the abutment stems was increased and the allowable bridge length was
also increased. In addition, integral abutments with flared and U-shaped
wingwalls were also allowed.

FURTHER MODIFICATIONS

Since the 1996 study, integral abutment construction and performance has
been continually monitored. One of the difficulties sometimes encountered in
integral abutment construction has been the close tolerances required for driving
piles. The practice, in steel superstructure integral abutments, of welding the
girders to the extended piles required that the piles be driven within one inch from
their plan locations. In difficult pile driving conditions, this was not easy to
achieve, even with the use of a pile driving template restraining the piles. Based
on the performance of the adjacent prestressed concrete details and similar details
used in other states, a detail was developed that would not require such precision
in final pile locations. In this case, piles are driven and a partial depth concrete
cap is placed. The steel girders are then erected on the concrete cap. Four
leveling bolts are used to adjust the girder elevation. The girder ends are then
encased in the abutment stem. Separate concrete placement is then made for the
deck and approach slabs (see Figure 4). This revised detail has proven to be more
easily constructed with no detectable difference in performance from the welded
pile cap.

An additional abutment type was also developed for the situation where
full integral abutments cannot be used. Semi-integral abutments offer some
significant advantages over typical integral abutments and jointless deck slabs.

Typical integral abutments require a minimum length of piling to ensure
that the piles have sufficient fixity to resist the horizontal displacements of the
superstructure. This can be a barrier to the use of integral abutments where the
ledge rock is near the surface. Since semi-integral abutments use conventional
bearings to accommodate superstructure movements (see Figure 5), the abutment
stem can be supported by conventional means such as spread footings or shorter
piles. The girders are integrally cast in a concrete backwall that is independent of
the rest of the abutment. The encased girders eliminate the need for a deck joint.
The approach slab is connected to the superstructure with typical integral
abutment details.

47

Figure 4. Current Steel Superstructure Integral Abutment

Figure 5. Semi-Integral Abutment

48
CURRENT INTEGRAL ABUTMENT CRITERIA

Integral abutments are now the first choice when selecting an abutment
type. As of 2004, there are 447 bridges with integral abutments in New York.
Of this number, 290 have concrete superstructures and 147 have steel
superstructures. The longest bridge in service with integral abutments is a four
span bridge with a total length of 350 feet. Criteria for integral abutments are
contained in the New York State Bridge Manual [7]. NYSDOT currently limits
the use of integral abutment to the following criteria:

1. Bridge length to less than 650 feet. There is no limitation on
individual span length.

2. Skew is limited to a maximum of 45.

3. Abutment reveal is limited to five feet. Abutment reveal is the
dimension from the bottom of the girder to the finished grade
under the bridge at the abutment stem.

4. No curved girders.

5. Maximum grade of the bridge is limited to 5%.

Steel H-piles or cast-in-place concrete piles are used; however cast-in-
place piles may only be used when the total bridge length is less than 160 feet.
Piles must be driven a minimum of 20 feet and are placed in pre-augured 10 feet
deep holes if the bridge length exceeds 100 feet. Steel H-piles are orientated with
the strong axis parallel to the girders so that bending occurs about the weak axis
of the pile. Wingwalls are separated from abutment stems when their length
exceeds thirteen feet to minimize the bending moment caused by passive earth
pressures. Piles are still designed to carry vertical loads equally and there is no
explicit requirement to consider bending moment in piles.

FUTURE MODIFICATIONS

Modifications that are currently under construction include increasing the
allowable abutment reveal dimension farther. This is based on successful practice
by other agencies. Definite requirements to analyze pile bending at longer bridge
lengths are also needed. In addition, changing the orientation of H-piles so that
bending occurs about the strong axis is also being considered for longer span
bridges. It is expected that the use of integral abutments will continue to increase
as their behavior becomes better understood.

49
SUMMARY

The New York State Department of Transportation (NYSDOT) has been using
integral abutment bridges since the late 1970's and thus far has constructed
approximately 450 integral and semi-integral abutment bridges. Due to their
excellent in-service performance, they are the preferred abutment type for
NYSDOT. This paper gives the evolution of the design and construction
practices.

REFERENCES

[1] "Integral Bridge Abutments: Report on Condition and Performance." Structures Design and
Construction Division, New York State Department of Transportation, December 1995.

[2] "Bridge Detail (BD) Sheets I1A thru IA3, Integral Abutments." Engineering Instruction 96-
035, Structures Design and Construction Division, New York State Department of
Transportation, July 2, 1996.

[3] Alampalli, S., and Yannotti, A. "Long-Term performance of Integral Bridges and Jointless
Decks." Transportation Research Record 1624, National Research Council, Washington,
D.C., pp. 1-7, 1998.

[4] Kunin, J. and Alampalli, S., "Integral Abutment Bridges: Current Practice in the United States
and Canada." Special Report 132, Transportation Research and Development Bureau,
New York State Department of Transportation, Albany, NY, June 1999.

[5] Alampalli, S., and Kunin, J., "Integral Abutment Bridges: Current Practice in the United States
and Canada." Journal of Performance of Constructed Facilities, ASCE, Vol. 14, No. 3,
pp. 104-111, August 2000.

[6] Bridge Inspection Manual 1997". Structures Design and Construction Division, New York
State Department of Transportation, December 1997. (Revised 1999)

[7] "Bridge Manual, 3
rd
Edition". Structures Design and Construction Division, New York State
Department of Transportation, April 2002. (Revised January 2004)

50
Integral Abutment Design and Construction:
The New England Experience

Darren W. Conboy, P.E., Associate Vice President, Edwards and Kelcey
Erik J. Stoothoff, Structural Engineer, Edwards and Kelcey

ABSTRACT

The use of integral abutments has been expanding in New England.
Historically, each of the New England states have used a variety of jointless
bridges, such as rigid frames and 3-span structures with cantilevered end spans.
Over the past 20 years the construction of modern integral abutment bridges has
increased throughout the region. The purpose of this paper will be to compare the
design and construction practices used in each state. The limitations on use,
design criteria and methodology and construction details of each state will be
summarized and compared. In addition, issues that were encountered during
construction and current performance and conditions will be presented.
Information on research and bridge instrumentation projects in the region will
also be summarized.

All of the transportation agencies in New England have had favorable
experiences with the performance of integral abutment bridges. As has been
reported in other publications, there is a wide range of design and construction
practices in use around the country. This is also true in New England. This paper
will serve to confirm the understanding that regardless of the specific design
philosophy and construction details, integral abutments perform well and should
be used wherever possible.

INTRODUCTION

In the last twenty years, transportation agencies in each of the six New
England states have constructed over 130 new integral abutment bridges. The
past ten years has seen a marked increase in the rate of integral abutment
construction. Each of the six state transportation agencies now indicate that
integral abutments shall be considered as the first choice for all new or
replacement structures. The use of integral abutments is often constrained by
factors, such as soil conditions, depth of structure limitations or vertical clearance
restrictions, therefore the percentage of new bridges actually utilizing them varies
from state to state.

The Maine Department of Transportation (MDOT) has been the leader,
having constructed 60 integral abutment bridges since 1983. Forty-five of the

51
bridges have been built since 1997. Superstructure types have been equally split
between steel and concrete. A two-span steel structure with an overall length of
295 feet and a 30 skew angle, which is currently under design, will be longest
built by the Department. Approximately 75% of their bridges are now being
designed with integral abutments. They have found that for stream crossings, in
particular, integral abutments provide the most cost-effective solution by
eliminating the need for cofferdams. In 1993, MDOT was the first state agency in
New England to incorporate guidelines and design procedures for integral
abutments into their Bridge Design Manual [1].

The Massachusetts Highway Department (MassHighway) has built 31 integral
abutment bridges since 1991, including 27 in the last ten years. Fifteen bridges
with steel superstructures have been built with an average length of 156 feet. The
longest steel bridge, which has an overall length of 354 feet, is constructed on a
30
o
skew angle. That bridge, which was constructed in 1991, has performed well
and is in good to excellent condition. Sixteen concrete bridges have been built
with an average length of 77 feet and a maximum length of 280 feet.
MassHighway developed detailed design guidelines and standard details, which
were added to their Bridge Manual [2] in 1999.

The other New England state agencies have been somewhat less active;
however, each is committed to using integral abutments where appropriate. The
Vermont Agency of Transportation (VTrans), which has built five bridges and has
another ten under design, has established an Integral Abutment Committee. The
Committee, supported by a research contract with Wiss, Janney, Elstner
Associates, Inc. (WJE), has completed a detailed report [3], which includes design
criteria and a suggested general design procedure. They also intend to fully
instrument and monitor two bridges to verify their design and behavioral
assumptions.

The Connecticut Department of Transportation (CDOT) (with approximately
20 bridges constructed), the Rhode Island Department of Transportation (RIDOT)
(five bridges) and the New Hampshire Department of Transportation (NHDOT)
(four bridges) have not developed their own comprehensive design procedures or
standard details. They each allow bridge engineers to utilize established
procedures from other states or organizations, such as the American Iron and
Steel Institute.

The Maine Turnpike Authority has actually constructed some of the longest
integral abutment bridges in the region. As part of their recently completed 30-
mile mainline widening program, the Authority utilized integral abutments on six
bridges. They constructed four concrete bridges with overall lengths between 244
and 356 feet and two steel bridges with lengths of 258 and 273 feet.

52
GEOMETRIC GUIDELINES

The New England states generally follow the recommendations of the 1980
FHWA Technical Advisory [4] for bridge length as summarized in Table 1.
MDOT is currently reviewing their length limitations and anticipate they will be
increased. The limitations on skew angle typically serve as a reminder to
designers that, for bridges constructed on skews above those limits, a more
rigorous three-dimensional analysis will be required. A 1993 study by Caswell
[5] for MDOT determined for skews less than 22 the friction force between the
concrete abutment and soil is greater than the transverse component of the passive
earth pressure, therefore the transverse forces can be ignored. VTrans has
constructed one bridge with a 35
o
skew angle.

Table 1. Geometric Guidelines
Span Length Skew

Steel (ft) Concrete (ft) Angle (Degrees)
Connecticut --- --- 20
Maine 200 330 30
Massachusetts 330 590 30
New Hampshire 300 600 ---
Rhode Island No Limit No Limit No Limit
Vermont 330 590 20

DESIGN PROCEDURES

This section will focus on the design procedures of the three New England
states, Maine, Massachusetts and Vermont that have published design guidelines.

Pile Design

Maine and Massachusetts, the two states with the most integral abutment
experience, have come to similar but different conclusions regarding design
procedures and construction details for piles. Both prefer to orient their piles for
bending about the weak axis. MDOT defines weak axis orientation as piles
aligned with their weak axis perpendicular to the centerlines of the superstructure
beam members. This also serves to facilitate their standard construction detail of
welding the bottom flange of steel beams to the tops of the piles. MassHighway
requires the weak axis of the pile to be parallel to the centerline of the abutment.

Maine requires that the allowable stress design (ASD) method be utilized for
substructure elements, while Massachusetts utilizes the load factor design (LFD)
method for substructure elements.

53
Section 3.9.7.2 of the MassHighway Bridge Manual [2] requires that pile
capacity must meet the combined axial load and bending criteria in AASHTO
Article 10.54.2 [6]. They have modified the standard interaction formula
(AASHTO Eqn. 10-156) by multiplying the standard maximum moment strength
in the weak axis direction by a new Coefficient of Inelastic Rotational Capacity
(
i
) as shown in Equation (1). The new coefficient provides a mechanism to
account for the significant additional rotational capacity that compact HP-pile
sections, bending about their weak axis, have beyond the elastic rotation at which
a plastic hinge first forms. AASHTO Article 10.48.1 states that the additional
rotational capacity of a compact section is three times the elastic rotation.
Theoretically, the pile can continue to rotate and translate due to thermal
movement up to three times the elastic rotation present at formation of the plastic
hinge without an increase in stress or buckling of the compression flange. This
additional capacity was accounted for by adding
i
, which conservatively has a
maximum value of 1.75 rather than the theoretical maximum of 3.0.

0 . 1
85 . 0
+ +
UX
X
UY i
Y
Y S
U
M
M
M
M
F A
P
(1)

P
u
= Applied axial load determined from analysis.
A
s
= Cross sectional area of pile.
M
y
; M
x
= Applied moment determined from analysis and P- moment.
M
uy
; M
ux
= Maximum moment strength based on slenderness criteria.
i
= Coefficient of inelastic rotational capacity:
For Compact Sections:
i
= 1.75 (to account for inelastic rotational
capacity of weak axis bending only).
For Non-Compact Sections:
i
= 1.0 (to account for lack of inelastic
rotational capacity).

MassHighways pile selection analysis incorporates all applicable forces,
including thermal expansion of the superstructure, superstructure deflection,
bridge skew effects and seismic effects. However, MassHighway does follow the
recommendations of Caltrans [7] that seismic loads should not control the number
or size of piles. MDOT allows designers to bring the stress level in the piles due
to combined axial loads and bending moments to the full yield strength of the
steel. In cases where the bending stresses reach yield, MDOT instructs designers
to orient the piles in the strong direction in lieu of developing a plastic hinge
under weak axis bending. MDOT uses an additional design check [1] of the piles
that the live load girder end rotation stresses do not exceed 0.55F
y
.

The VTrans design procedure [3] allows either LFD or the Load and
Resistance Factor Design (LRFD) method. The designer may choose to orient the
piles in either the weak or strong direction, with orientation in the direction of
bridge movement. VTrans prefers to use compact pile sections, but does not
provide any guidance for increasing the capacity of compact piles due to inelastic

54
rotation in the weak direction. Non-compact piles may be specified but only for
skews less than 20.

Connecticut [8] requires only vertical loads for the design of piles. Rhode
Island designs their integral abutment piles using the LRFD method neglecting
skew effects. Both Connecticut and Rhode Island design their piles for weak
direction bending, and size the piles such that the loads do not allow a plastic
hinge to develop.

Abutment Design

Similar to pile design, the New England states have differing design
procedures and construction details for the abutment stems and wingwalls. Table
2 summarizes the major components of abutment stem and wingwall design.

Table 2. Abutment Stem and Wingwall Details

Abutment
Height (ft.)
Wingwall
Orientation
Wingwall
Length (ft.)
Earth Pressure
Connecticut 8 U-back --- Full Passive
Maine 12 Parallel 10 Full Passive
Massachusetts 13 U-back 10
Modified pressure
coefficient:
K=0.43+5.7[1-e
-190(d
t
/H)
]
New Hampshire --- U-back --- ---
Rhode Island None --- --- ---
Vermont 13 U-back 10 Full Passive

WINGWALLS

MassHighway, MDOT and VTrans all design wingwalls for different forces
based upon the different configurations. MDOT designs wingwalls as an
extension of the abutment stem for full passive earth pressure. VTrans prefers U-
wingwalls but will allow flared or parallel walls to be used if the situation
warrants. They require the U-wingwalls be designed for active earth pressure.
Flared wingwalls are designed for a two-component earth pressure diagram
comprised of passive earth pressure acting perpendicular to the centerline of the
abutment stem and active earth pressure acting perpendicular to the centerline of
the bridge. Parallel wingwalls are designed for full passive earth pressure.

MassHighway, which also uses U-wingwalls as their standard construction
detail, has the most extensive wingwall design procedures. In addition to
designing the walls for active horizontal earth pressure, MassHighway requires
that the capacity of the wall be checked for vertical passive earth pressure acting

55
on the base of the wall due to abutment rotation as depicted in Figure 1 on the
following page.

MassHighway [2] also states that a primary function of integral wingwalls is
to assist in the resistance of transverse seismic forces. Their design policy calls
for the required length of the wingwall to be as determined in Equation 2.

Required Wingwall Length:
) / (
2
1
) (
) (
2
m kN K H
kN ce SeismicFor
m L
= (2)

ABUTMENT STEMS

Both MassHighway and VTrans design their abutment stems as a continuous
beam between the ends of the superstructure beam members, and have a
construction detail where the horizontal reinforcing bars are passed through holes
cut in the webs of the beams. While the two states have similar reinforcing
details, VTrans designs this reinforcing for full passive earth pressure. Based on
studies by the University of Massachusetts, MassHighway has implemented a
modified horizontal earth pressure coefficient to more accurately model the earth
force applied to the abutment stem under thermal movement. This modified
horizontal earth pressure coefficient is shown in Table 2. MDOT does not believe
in penetrating the webs of the main beams to pass reinforcing steel. MDOT does
Figure 1 Force Diagram from MassHighway Bridge Manual

56
concede that shear cracking of the abutment stem may occur, but is preferable to
reducing the shear capacity of the superstructure beam members.

CONSTRUCTION DETAILS

There are several construction details that are unique to individual New
England states. The following sections discuss the construction details set forth
by MDOT, VTrans and MassHighway in their design guidelines and construction
standards, as well as highlighting some of the construction details used in the
other states for integral abutment bridges constructed through their limited
programs.

Figure 2 shown below is the typical integral abutment section for steel bridges
found in the MDOT Bridge Manual [1] and Figure 3 is the typical section from
the VTrans Report on Integral Bridge Abutment Design [3].

Piles

The standard pile-to-abutment stem connection detail among New England
states is a 2-foot to 3-foot embedment of the pile in the abutment stem. The
exception is in Maine where the standard connection detail for steel bridges is to
weld the top of the piles to the bottom flange of the beams.

Massachusetts and Vermont both pre-drill an oversized hole for the first 8 feet
of the pile installation. According to MassHighway, this allows the pile to be
more flexible since the pile head is surrounded by loose soil. It also assists with
maintaining pile location tolerances. The two states use different material to fill
the oversized hole; MassHighway uses a crushed stone or pea stone, while
VTrans specifies compacted sand. Officials from MDOT have not seen a need to
pre-drill the piles, however if the first pile displays significant travel during the
installation, MDOT will recommend pre-excavating to assist the pile driving
effort.

Figure 2 Typical MDOT Steel Bridge Detail Figure 3 Typical VTrans Steel Bridge Detail

57
Abutments

There are four methodologies used in New England to set the superstructure.
As already discussed, MDOT sets a steel superstructure by attaching the beams to
the piles. When constructing a precast concrete superstructure, MDOT uses a 1-
6 neoprene pad to set the beams. MassHighway and NHDOT cast concrete
pedestals on top of the first abutment pour, along with an erection pad to provide
proper beam seat elevations, and are held in place by two anchorbolts at each
beam seat. Vermont and Connecticut both use anchor bolts with leveling plates
and nuts to provide proper beam seat elevations. Maine uses a drilled-in
reinforcing bar when constructing precast concrete bridges as hold-down
reinforcement.

All of the New England states use approach slabs at the ends of their integral
abutment bridges, with the exception of Maine, which only uses approach slabs
for steel bridges that exceed 80 feet in total length and 140 feet for concrete
bridges. Massachusetts has a standard approach slab detail that is supported on a
shelf constructed into the abutment backwall where it is free to slide. The
approach slab is buried and slopes down and away from the bridge. A concrete
key provides horizontal restraint at the end of the approach slab. The roadway
joint is located at the interface between the approach slab and abutment backwall.

When MDOT does use approach slabs, they may either be at grade or buried.
Only when buried are approach slabs connected to the abutment. Roadway joints
are used only when the total bridge lengths exceed 140 feet and 230 feet for steel
and concrete superstructures, respectively. These roadway joints are located at
the interface between the approach slab and abutment backwall.

VTrans, CDOT and NHDOT utilize at-grade approach slabs supported on
shelves constructed into the abutment backwall, similar to MassHighway,
however these states call for a pinned connection between the approach slab and
abutment. VTrans and CDOT have roadway joints at the interface between the
approach slab and abutment. NHDOT utilizes a sleeper slab at the end of the
approach slab and includes an asphaltic roadway joint at the end of the approach
slab over the sleeper slab.

RESEARCH AND STRUCTURE INSTRUMENTATION

MassHighway has sponsored research on the various aspects of integral
abutment behavior. The University of Massachusetts completed two research
projects in 1998. The first project [9] was an analytical study on the behavior of
integral abutment bridges under thermal expansion and seismic loads. The study
included investigating the use of both linear and non-linear soil resistance behind
the abutment walls. The effect of skew was also investigated. The study resulted
in recommendations for analysis and design guidelines.

58
The second study [10] investigated passive earth pressures behind integral
bridge abutments. Passive loading tests were conducted on a rigid concrete
retaining wall to study the effect of wingwall orientation on lateral earth pressure
development. The study showed that the lateral earth pressure was greater for the
U-wall orientation preferred by MassHighway than for flared or parallel
wingwalls. The study led to the development of the lateral earth pressure formula
shown in Table 2, which was incorporated into the MassHighway design
guidelines.

In addition, MassHighway installed an extensive instrumentation system on a
three-span integral abutment bridge two years ago. The results of that program,
which are also being monitored by the University of Massachusetts, are being
reported in separate paper [11] at this conference.

MDOT is currently supporting research on the behavior of pile-supported
integral abutments at bridge sites with shallow bedrock. Shallow bedrock, which
is a common occurrence in Maine, would normally preclude the use of integral
abutments. The goal of the research is to provide a better understanding of the
behavior of integral abutments with short piles, which would allow their use to be
expanded to these soil conditions. The first phase of the study [12], performed by
the University of Maine, consisted of development of an analytical model to
investigate soil/structure interaction and the structural response of bridges
supported on piles less than 13 feet long. The second phase of the study, now
underway, includes instrumentation of actual bridge and further analytical
modeling.

VTrans is also planning to instrument two bridges. They are currently in the
process of selecting the appropriate locations. The intent is for one bridge to be
oriented in a north-south direction and the second in an east-west direction. WJE
will assist with the development of the actual instrumentation program.

OBSERVATIONS

Some issues that were identified by state bridge officials include:

Construction

Pile Driving Tolerance: Typically, piles are required to be driven to
within 3 inches of the location specified on the plans. When there has
been a problem in achieving this level of accuracy, pre-drilling, pre-
excavating or the use of a template has successfully addressed the
problem.
Stage Construction: MDOT has experienced a fit-up problem when
stage constructing a steel beam, integral abutment bridge. This, however,

59
can be attributed to their detail of welding the beam to the pile. In this
case, all of the piles and the entire abutment cap were constructed in the
first stage. The second stage beams did not fit properly on the top of piles
due to movement that had occurred in the first stage structure.
Abutment Stem Cracking: Some early structures experienced cracking in
the abutment stems, which was addressed by revising the reinforcing steel
layout.

Performance

Approach Roadway Pavement Cracking: Pavement cracking has been
observed, particularly for bridges with no roadway joints or on greater
skew angles. This is expected and not considered a problem.
Splitting or Cracking in Asphaltic Roadway Joints: The Maine Turnpike
Authority has noted these issues on two bridges. They have been
attributed to settlement of the approach roadway fill and rotation of the
approach slabs rather than a specific problem with the integral abutment
details. Other states reporting similar problems feel improper joint
installation is the issue.

CONCLUSION

Despite several differences in design philosophies and construction details, all
of the states have come to the same conclusion; that integral abutment bridges
shall be the first design alternate considered when a new or replacement bridge is
proposed. State bridge officials, inspectors and construction and maintenance
personnel indicate that, in general, integral abutment bridges have been easier to
construct and have performed as well or better than bridges with traditional
bearings and roadway joints. Overall, all of the New England states have
expressed a very positive experience so far with all aspects of integral abutment
bridges. All New England states have made a commitment to promoting the
increased use of integral abutments in new bridges.

ACKNOWLEGMENTS

Much of the information contained in this paper was obtained from discussions
and correspondence from the following agency staff members: CDOT Bryan
Reed and James McCann; MDOT Eric Calderwood and Leanne Timberlake;
Maine Turnpike Authority Robert Driscoll (HNTB Corporation); MassHighway
Daniel Crovo; RIDOT Rahmat Noorparvar; and VTrans George Colgrove
and members of the Integral Abutment Committee.

60
REFERENCES

1. Maine Department of Transportation, 2003 Bridge Design Guide, Chapter 5
2. Massachusetts Highway Department, 1995 Bridge Manual Part I & Part II, Metric
Edition, Part I-Section 3.9, Part II-Section 12
3. Vermont Agency of Transportation Integral Abutment Committee, 2004. Report on
Integral Abutment Bridge Design
4. Federal Highway Administration, January 1980, Integral, No-Joint Structures and
Required Provisions for Movement, FHWA Technical Advisory T 5140.13
5. Caswell, E.B., May 1993, Review Current Policy on the Use of Jointless Bridges,
Volume 1, Maine Department of Transportation
6. American Association of State Highway and Transportation Officials, 2002 Standard
Specifications for Highway Bridges, 17
th
Edition, Section 10.54.2
7. California Department of Transportation, 1996, Memos to Designers 5-1
8. Connecticut Department of Transportation, 2003 Bridge Design Manual, Section 5.
9. Ting, J.M. and S. Faraji, March 1998, Streamlined Analysis and Design of Integral
Abutment Bridges, University of Massachusetts Transportation Center
10. Thomson, T.A. and A.J. Lutenegger, July 1998, Passive Earth Pressures in Integral
Bridge Abutments, University of Massachusetts Transportation Center
11. Bonczar, C., S. Civjan, S. Brena, B. Crellin, and D. Crovo, 2005, Field Data and FEM
Modeling of the Orange-Wendell Structural and Geotechnical Considerations,
presented at The 2005-FHWA Conference, Integral Abutment and Jointless Bridges,
March 16-18, 2005
12. DeLano, J.G., T.C. Sandford and W.G. Davids, May 2004, Behavior of Pile-Supported
Integral Abutments at Bridge Sites with Shallow Bedrock, The University of Maine

61
VDOT Integral Bridge Design Guidelines

Keith Weakley, P.E.

Assistant District Bridge Engineer
Virginia Department of Transportation-Staunton District
811 Commerce Road
Staunton, VA 24402

Tel.: 540-332-9107
Fax: 540-332-7886
E-mail: Keith.Weakley@vdot.virginia.gov

ABSTRACT

The Virginia Department of Transportation (VDOT) developed a set of
guidelines to address the design and detailing of jointless bridges. These
guidelines address structural aspects and constructibility issues. Structural details
are provided for incorporation into contract plans, and a commentary is provided
to aid the designer in the selection and design of several types of jointless bridges.

INTRODUCTION

Jointless bridge construction may be defined as the practice of constructing
bridges without deck joints [1]. It is becoming one of the bridge engineers
responses to joint-related damage caused by the use of deicing chemicals. The
Virginia Deptartment of Transportation (VDOT) has been promoting the use of
jointless bridges for many years. Over the past several years, the VDOT Jointless
Bridge Committee has been revising the existing design and detailing standards to
incorporate new technology and lessons learned from construction of these types
of bridges in the past two decades. In addition to these supplements and
revisions, an effort was made to help the design engineer select the most
appropriate type of a jointless bridge and provide practical guidance on the design
of the structural elements.

DESIGN AND SELECTION

Overview

VDOT currently promotes three types of jointless construction: full integral,
semi-integral and deck extension. The full integral bridge is characterized by
beams or girders cast into a concrete end diaphragm which is connected to a
concrete pile cap supported by a single row of piles (Figure 1). The semi-integral

62
bridge (or integral backwall) is similar to the full integral bridge except that the
concrete end diaphragm is not rigidly connected to the substructure (Figure 2).
The deck extension is simply the extension of the end of the deck slab over the
traditional backwall and into the adjoining approach pavement. The main
beams/girders are not cast into a concrete end diaphragm (Figure 3). In each case,
the bridge superstructure interacts with the approach fills.

Figure 1.
Full Integral
Figure 2.
Semi-Integral
Figure 3.
Deck Extension

VDOT has developed a set of criteria for the use of jointless bridges, as shown in
Table 1. Should a designer wish to exceed these limits (length, skew and other
criteria), a design exception is required from the State Structure and Bridge
Engineer.

Table 1. VDOT Length and Skew Limits for Jointless Bridges
Full Integral Semi-Integral Deck Extensions
300 ft for 0 skew 450 ft 450 ft
Steel Bridges
150 ft for 30 skew 30 max skew 30 max skew
500 ft for 0 skew 750 ft 750 ft
Concrete Bridges
250 ft for 30 skew 45 max skew 45 max skew
Total Movement at Abutment 1 in 2 in 2 in

In addition, VDOT has developed a selection algorithm to aid the designer in
determining the most appropriate type of a bridge. It starts with a full integral as
a top choice. Various criteria that may exclude it are then analyzed. If a full
integral bridge is deemed infeasible, the designer is directed to consider a semi-
integral option. Similar checks are performed, and if the semi-integral bridge is
deemed infeasible, the designer is steered to a deck extension. The last choice is a
conventional design, with a joint between the slab end and backwall.

63
Full Integral Bridge

The full integral bridge is the top design choice of VDOT. According to
research performed at the Virginia Polytechnic Institute and State University
(Virginia Tech), steel H-piles are most suitable to support fully integral
abutments. The inherent flexibility of steel H-piles allows them to endure constant
flexure induced by the cyclic thermal strains of the superstructure [2]. The
VDOT approach has been to use a lighter pile section (HP 10 x 42) and orient H-
piles in a weak axis-bending configuration. VDOT has opted to significantly
reduce pile stresses by detailing a hinge between the pile cap (footing) and the
concrete end diaphragm (integral abutment). The hinge serves as a moment-relief
device, thereby resulting in the transfer of only the lateral load to foundation piles.
The original hinge detail was essentially a construction key cast along the
centerline of the integral abutment, as shown in Figure 5. Full-scale testing at
Virginia Tech determined that the key did not function as intended [2].

Figure 5. Old Hinge Detail Figure 6. Revised Hinge Detail

The research study resulted in a more rotationally flexible hinge detail, as
shown in Figure 6. It consists of strips of high durometer neoprene along either
side of the dowels along the centerline of the integral abutment, through which
vertical loads are transmitted to the footing. The dowel line serves to transmit
shear loads.

When this type of a bridge is constructed, consideration should be given to the
temporary support of beams. When steel beams or girders are used, anchor bolts
are set into the footing along the centerline, much like the dowels. Leveling nuts
and plates are used to support the bottom flange of the beam and adjust the
elevation of the beam, as shown in Figure 7.
Construction plans typically
include a table of elevations for the
tops of the plates on which the
beams will rest. This approach
works well with steel beams and
girders, however, it is not suitable
for heavy prestressed concrete
beams. In these cases, the method
of temporary support is left up to
the contractor, subject to a review by VDOT.

Figure 7. Steel Beam Support Detail

64
The design of each structural element of a full integral bridge is relatively
simple. The integral backwall is designed for the passive earth pressures. For the
purposes of design, it is treated as a continuous beam on its side with the main
bridge beams acting as supports. The pile cap is designed to handle vertical loads
from the superstructure and passive pressures that develop during thermal
expansion. In addition to these forces, it may also be necessary to design piles to
resist lateral loads. If a bridge is skewed, lateral loads also develop to restrain the
rotation induced by the eccentricity of the passive pressure forces generated
during thermal expansion.

Semi-Integral Bridge

The second preferred type of a jointless bridge is the semi-integral design.
Though not as economical as the full integral, it is more adaptable to varying field
conditions. In some cases, it is possible to retrofit existing bridges to a semi-
integral configuration. The semi-integral design is usually considered when the
minimum pile length of 25 feet cannot be provided for the full integral option.

The semi-integral bridge has some of the same features as the full integral, as
well as some that are unique. The integral backwall overlaps the abutment and
overhangs it slightly. A small gap is provided between the integral backwall and
the substructure, to allow it to move freely in the longitudinal direction. A drip
bead is made on the extension to prevent weeping along the overlap of the integral
backwall and the abutment. As with the skewed full integral bridges, lateral loads
develop in skewed semi-integral bridges to resist the rotation developed by the
thermal expansion. This condition has been addressed at VDOT for more than a
decade and was recently corroborated by a field study [3]. In the semi-integral
bridge, unlike in the full integral, forces must the transferred to the substructure
in a way such that the longitudinal movement of the
superstructure is not compromised.
VDOT addresses this situation by
providing a rub plate at the acute
corners of the bridge (Figure 8).
The rub plate serves as a bearing
between the vertical edge of the
superstructure and the wing haunch.
It allows for the transfer of the large
horizontal thrust from the
superstructure to the abutment,
while allowing longitudinal sliding.
Typically, these rub plate assemblies have been made of stainless steel, with shear
studs cast into the concrete. The wing haunch is designed as a vertical cantilever,
which typically results in a more heavily reinforced section.

Figure 8. Rub Plate Detail

65
Deck Extension

The third type of a jointless bridge is the deck extension. Deck extensions are
used when neither the full integral nor the semi-integral option meet the
conditions of the particular structure, and for retrofitting existing structures.
Usually, this is the case if the bridge length and/or skew exceeds the guidelines
set forth in the VDOT Structure & Bridge Office Practice (Table 1). Deck
extensions are essentially a cross between the conventional bridge and the semi-
integral design. Though the substructure does have a backwall, and the girders
are not embedded in a concrete end diaphragm, the deck slab overlaps the
backwall, resulting in the elimination of the deck joint. The deck slab is extended
from the end diaphragm across the top of the shortened backwall. The extension
is lengthened by an additional 3 inches to provide for a drip bead, as with the
semi-integral option. A -inch layer of expanded polystyrene is used between the
abutment backwall and the deck slab to seal the joint and provide a gap for
deflection and longitudinal movement. VDOT does not use materials with a high
stiffness to reduce the amount of loading the cantilever slab may experience.
Although the exposure to passive earth pressures is lessened because of the
shallow depth, the superstructure still undergoes rotation in skewed structures,
much like the semi-integral design. The load generated in the acute corners of the
bridge is accounted for by using rub plates (as in the semi-integral), though they
tend to be much smaller. This type of detail is amenable to the use of the
Massachusetts-type approach slab. Since the approach slab is buried below the
plane of the deck extension, there is no conflict with the longitudinal movements
of the superstructure (Figure 3).

Backfill

The interaction with the approach embankment has become an issue of
particular concern. Since the typical soils found in roadway fills can exert
relatively high passive pressures, the resulting loads necessitate larger and thereby
more expensive members. This problem has led VDOT to adopt the use of
elasticized expanded polystyrene (EPS) as a backwall-backfill interface material.
The EPS effectively acts as a cushion between the backwall and the roadway fill.
VDOT practice is to use a layer of this material between the bottom of the integral
backwall and the approach slab seat.

Field studies conducted by the Virginia Transportation Research Council
(VTRC) [4] show that the use of elasticized EPS material has consistently and
substantially reduced passive earth pressures behind the integral backwall. As a
result of this research, VDOT has been able to use select backfill materials,
thereby reducing settlements behind the integral backwall. Consequently, the
need for approach slabs has been eliminated in many cases.

Current VDOT practice is to specify the thickness of EPS on the plans,
showing the placement details and including a pay item for that thickness on a per

66
square yard basis. In the case of a single-span integral, VDOT specifies the EPS
placement on the up grade end only (e.g., the abutment of higher elevation) to
establish a dominant passive pressure at the lower abutment. This creates less
stress on the bearings and anchor bolts, as the thermal expansion will push
against the soil backfill. Since both bearings are detailed as expansion bearings,
the net effect of the elastic inclusion is not diminished.

It is also important to ensure that the loading produced on the integral
backwall by the backfill material during construction is equalized (whether or not
the EPS is used). Failure to do so could result in the drift of the superstructure.
For this reason VDOT specifications limit the differential fill height during
backfill operations to a maximum of 6 inches, regardless of the number of bridge
spans.

CONSTRUCTION AND PERFORMANCE ISSUES

During the construction and monitoring of jointless bridges, issues have
surfaced that were subsequently addressed in design details and specifications.
These observations have provided feedback to streamline the design process.

Backfill

Since the integral backwall is constantly moving, problems arose with the
excessive settlements of the approach backfill. The expanding superstructure
pushes against the adjoining fill. When a bridge contracts, material falls into a
void that was created during expansion. It results in amplified settlement directly
adjacent to the end of the bridge, necessitating frequent maintenance of the
approach pavement. The latest design methodology is to stipulate the elastic
inclusion as an interface between the backwall and the backfill.

Longitudinal Superstructure Movement

The use of wings that are in line with the superstructure (U-back wings) is
typical in many areas, especially in the case of confined right-of-way, typical of
many urban settings. This type of configuration requires a barrier on the top of the
wingwall in line with the bridge deck parapet. Initially VDOT addressed the use
of U-back wings by allowing a gap between the integral superstructure and the
static substructure. Even though the gap is sized for the predicted movement of
the superstructure plus a factor of safety, the possibility of a conflict exists. On
several occasions, failure of this detail has caused problems and/or damage to the
structure. As a result, VDOT recommends eliminating all possible obstacles to
the longitudinal expansion and contraction of the integral superstructure. In the
case of U-back layout, the wings should be located outside the edge of the
superstructure. This detail should allow for adequate room to accommodate
guardrail posts.

67
Superstructure Rotation

Though VDOT currently addresses the issue of rotational tendencies in
skewed integrals, this approach has been reinforced by the behavior of a bridge
that was built in 1993 and has since been continuously monitored. The skew was
only 5 degrees and no rub plates were used on the structure. Sufficient earth
pressure developed to cause a noticeable horizontal rotation in the superstructure
of 323 ft long and 85 ft wide [5]. This behavior has led to the use of rub plates on
all skewed bridges, regardless of size, skew, or span configuration.

Approach Slabs

In the past practice, the type of a superstructure has had little impact on the
approach slab support. With the introduction of integral bridges, this has become
an area of greater concern. Currently, this is addressed by connecting one end of
the approach slab directly to the integral backwall and by supporting the other end
on a sleeper pad. In the past, the connection with the integral backwall was
accomplished with two straight horizontal reinforcing bars, one in the plane of the
top mat of the approach slab reinforcement and one in the plane of the bottom
mat. These bars were embedded in the approach slab and the integral backwall
(Figure 9). Over time, transverse cracks appeared in the approach slab,
approximately 20 inches from the contact with the integral backwall. It was
determined that these cracks coincided with the ends of the reinforcing bars.
When the sleeper pad settles, moments are produced at the end of the approach
slab, resulting in the tension cracks forming in the top surface.
To alleviate this problem, VDOT
has opted for a bar configuration that
facilitates flexible rotation of the approach
slab at the backwall connection. The two
straight bars were eliminated, and an
inclined bar was passed through the point
of rotation at the edge of the approach slab
seat (Figure 10). This design provides a
more positive connection of the approach
slab to the integral backwall while
allowing the inevitable rotation to occur.
The front edge of the approach slab seat is
heavily chamfered to prevent spalling of
concrete.

Figure 9. Original Approach Slab
Connection Detail

68
With the implementation of
structural backfill, approach slabs
can be eliminated in many cases.
However, VDOT details still provide
the approach slab seat, should the
need arise to install one at a later
date. When the approach slab is not
used, the pavement structures rests
on this seat, allowing it to bridge
over the EPS material. VDOT has
also installed several buried
approach slabs on an experimental
basis. They have not been in service
for a sufficient amount of time to verify their performance; however, they offer
potential for eliminating costly approach slab repairs. This experimental type of
an approach slab is similar to the Massachusetts design for conventional bridges,
as characterized by its location below the pavement structure. In addition, the
VDOT version provides mechanisms for the translation of the integral
backwall and the rotation of
the approach slab itself, as
shown in Figure 11. This
allows for thermal movements
of the bridge and approach
settlements to occur without
unduly stressing either the
approach slab or the integral
backwall. Mitigation of
settlements can be
accomplished with simple
pavement overlay operations,
eliminating the need for mud
jacking or similarly intensive remedial measures.

Staged Construction

Staged construction calls for special considerations at integral bridges.
Problems typically arise during the second stage. When casting the second stage
of an integral bridge, deflections caused by the non-composite dead loads produce
rotations at the beam ends. To accommodate these rotations and to prevent torsion
of the integral backwall, VDOT recommends a closure pour between the stage 1
and stage 2 integral backwalls. This pour may also extend into the deck itself, but
it is dependent on the differential deflections in the superstructure between the
two deck pours. VDOT is currently developing pertinent standards for staged
construction.

Figure 10. Revised Approach Slab Connection
Detail

Figure 11. Experimental Approach Slab Detail

69
Experimental Abutments

There are always cases that fall outside the recommended guidelines for
jointless bridges. It is not feasible to have design details that work in every
situation. However, VDOT has developed a jointless detail that can be used in
virtually any situation. Though experimental at this stage, the alternate abutment
detail allows the implementation of jointless concepts in situations that may far
exceed the recommended limits of movement, length, skew, curvature, etc. This
detail consists of an abutment that is a hybrid between a traditional abutment with
a backwall and a semi-integral design, as shown in Figure 12. The superstructure
girders are embedded in the concrete end diaphragm much like a semi-integral,
although the end diaphragm is not in contact with the backfill material. The gap
is bridged using a tooth joint, or an expansion dam of some type.

The movement is accommodated in the area between the end diaphragm and
the abutment backwall. Drainage is also collected in this area, which serves as a
flume to transfer the runoff to the slope protection on either side of the abutment.
Unpainted weathering steel can be used throughout the bridge with this design, as
there are no open joints in the proximity of the girders or bearings. The additional
cost associated with the initial abutment construction is offset by the initial and
long term savings associated with the elimination of paint. This type of abutment
also eliminates problems associated with the approach slab attachment, since a
conventional detail can be employed.

Figure 12. Alternate Jointless Detail

70
CONCLUSIONS

VDOT is continuously striving to improve design guidelines and construction
details for jointless bridges in an effort to ensure the optimum performance.
Incremental progress is achieved through methodical research and practical
feedback obtained from construction monitoring. Limits on the use integral
bridges are being gradually relaxed, as new information becomes available.
Bridges constructed to date in Virginia have performed remarkably well.
Consistent and measured application of the jointless bridge technology has been a
goal of the Virginia Department of Transportation.

ACKNOWLEDGMENTS

The author thanks the VDOT Jointless Bridge Committee for their hard work
and patience in developing design standards and details and the Virginia
Transportation Research Council for their technical assistance and research.

REFERENCES

1. Virginia Department of Transportation, Structure & Bridge Division. 2003 Volume V,
Part 2 Structure & Bridge Design Office Practice. Richmond.

2. Arsoy, S. December 2000. Experimental and Analytical Investigations of Piles and
Abutments of Integral Bridges. Doctoral Dissertation. Department of Civil and
Environmental Engineering, Virginia Polytechnic Institute & State University,
Blacksburg.

3. Steinberg, E., Sargand, S. and Bettinger, C. 2004. Forces in Wingwalls of Skewed
Semi-Integral Bridges. ASCE Journal of Bridge Engineering, November/December
2004.

4. Hoppe, E.J. 2004. Field Study of Integral Backwall with Elastic Inclusion. Proceedings
of the Integral Abutment and Jointless Bridge Conference. Baltimore, Md., 16-18 March
2005.

5. Hoppe, E.J., and Gomez, J.P. 1996. Field Study of an Integral Backwall Bridge.
Virginia Transportation Research Council, Charlottesville

71
SESSION II:
CASE STUDIES

72

73
Case Study: A Jointless Structure to Replace the Belt
Parkway Bridge Over Ocean Parkway

S. Jayakumaran, PhD., P.E., Project Structural Engineer, Gannett Fleming
Michael Bergmann, P.E., Structures Design Supervisor, New York State
Department of Transportation
Syed Ashraf, P.E., Project Geotechnical Engineer, Gannett Fleming
Charles Norrish, P.E., Project Manager, Gannett Fleming

ABSTRACT

This paper presents the design of a 65.8 m (216 ft) long three span
continuous, jointless bridge using the semi integral abutment concept to relieve
the substructures of most of the seismic forces.

As part of their capital program, the New York City Department of
Transportation (NYCDOT) proposed to replace the rapidly deteriorating Belt
Parkway Bridge over Ocean Parkway in Brooklyn, New York in 1999. On the
basis of a competitive selection process, the Granite Halmar (contractor)/ Gannett
Fleming (designer) team was selected for this replacement project, under a
Design-Build contract.

The bridge carries the Belt Parkway, a regional corridor traveled by
166,000 vehicles daily, over Ocean Parkway, a scenic landmark in south
Brooklyn. The existing bridge was a two span steel multi girder structure with a
wall type center pier and full height reinforced concrete abutments. The proposed
bridge is a three (3) span structure, longer and wider to span over the proposed
service roads on Ocean Parkway underneath and to accommodate full width
shoulders on the bridge. The three spans are 18.9 m (62 ft.), 32.8 m (108 ft.) and
14.1 m (46 ft.) respectively for a total length of 65.8 m (216 ft.). The abutments
consist of a single row of pipe piles supporting a concrete cap beam, and the piers
are multi column bents with concrete capbeams supported by minipiles.

The design incorporated the semi integral abutment concept, with the
backwall cast integrally with the deck and the deck joint moved over to the
approach slab beyond the abutment. In addition, the bridge was made continuous
for live load at the piers by introducing girder splices over the piers. The structure
was fixed at the west pier and allowed to expand at all other supports. Under a
regular design, the continuity in the superstructure along with the fixity only at
one pier would have resulted in extensive seismic forces on that one pier.
However, under this proposed semi integral abutment concept with the backwalls
cast integrally with the deck, the passive pressure generated by the embankment
behind the backwalls was utilized in absorbing the seismic forces, thereby
reducing the resulting seismic forces at the fixed west pier.

74
In addition to eliminating deck joints, the use of stainless steel
reinforcement, and the use of precast (Inverset) deck elements considerably
enhanced the service life of the bridge, beyond the 50-year life required by
NYCDOT.

Due to staging of construction, the superstructure of the first half the
bridge was erected in ten days and the second half of the bridge was also
completed in record time. The bridge replacement was completed by November
2004, a month ahead of the already aggressive schedule.

INTRODUCTION

Belt Parkway, winding through the southern coast of Brooklyn, is a
regional corridor in New York City traveled by 166,000 vehicles daily. The
roadway crosses numerous features, including Ocean Parkway, a scenic landmark
in South Brooklyn designed by Frederick Olmsted. Belt Parkway carries three
lanes of traffic in either direction. The bridge carrying Belt Parkway over Ocean
Parkway was identified for replacement in 1999. The level of deterioration was
clearly evident; the bridge was supported by timber shoring at its abutments and
the Belt Parkway was plated over at the abutment locations.

In addition to the deteriorating bridge, the adjoining four loop ramps at the
Belt Parkway/Ocean Parkway Interchange were found to be geometrically
substandard. Traffic studies revealed that the intersection would reach its
capacity by year 2005. The New York City Department of Transportation
(NYCDOT) soon realized that the construction of a new bridge could not be
accomplished without improving the loop ramps. NYCDOT then collaborated
with the New York State Department of Transportation (NYSDOT) and decided
to expand the scope of work to include not only a new bridge, but also a new
interchange configuration.

NYCDOT and NYSDOT soon recognized that the extensive scope of
work that this project would entail, coupled with the existing condition of the
bridge, demanded an expeditious solution. It became clear that Design-Build was
the optimum procurement method. Design-Build was estimated to save three to
four years over a conventional Design-Bid-Build procurement. Design-Build
would also foster the most cost-efficient and innovative solution possible since it
challenges the consultant and contractor to think together and creatively. Since
the project involved a corridor in the National Highway System, it was eligible for
federal funding. At the time the project was let out, the Federal Highway
Administration (FHWA) allowed the use of Design-Build in New York State
under its experimental SEP-14 Program for innovative delivery methods.
NYCDOT secured the necessary SEP-14 approval and an 80 percent federal
funding was thereafter secured.

75
In order to expedite the delivery of this project the NYCDOT included an
$85,000 per day liquidated damages clause in the contract. The contractors were
expected to pay $85,000 a day for each day that that the construction runs over the
scheduled critical duration. The critical duration was established as the
duration of construction activities that impacts traffic on Belt Parkway.

Through a competitive procurement process, the Granite Halmar
(contractor)/ Gannett Fleming (consultant) team was selected for the
implementation of the project. Given the very stringent liquidated damages
clause, the team's concept for the expedited replacement of bridge was to use the
Inverset Deck System. In addition to the Inverset Deck, the team planned to use
other prefabricated elements including T-Walls for the abutments and wing walls,
precast cap beams and precast parapets. The use of prefabricated elements lends
speed and efficiency to erection of the bridge at the site, there is less activity at
the site resulting in less cost as well as less adverse impact to the community and
the overall quality of the prefabricated elements is superior to field built units.

Precast Inverset deck units are two girder systems with a precast concrete
deck. The girders are usually spaced between six to eight feet with a 1-1/2 feet to
two feet overhang on both sides. Based on the overall width of the bridge, the
girder spacing had to be selected so that the units would be of manageable weight.
In addition, the width selection of the units had to be chosen so that the joints
align with the proper lane configuration, as well as the staging line at the center of
the bridge required for maintenance, protection of traffic and the two-stage bridge
construction.

In order to enhance the life of the bridge the team introduced the concept
of a Jointless Bridge. The concept also included semi integral abutments to
eliminate joints at the abutments. Integral abutment bridge designs have been on
the rise since the 1990s [1]. By 1996, around 155 structures have been built in
New York State with the Integral Abutment concept [2].

This paper will now focus on the innovative ideas that were introduced to
achieve a Jointless Bridge while using precast elements, which contributed to
the overall success of the project.

EXISTING BRIDGE

The existing bridge built in 1941, was a two (2) span structure with a wall
type pier and full height abutments. The superstructure consisted of simple spans
with steel multi girder deck systems. The span lengths were 21.82 meters each
for a full length of 43.64 meters. The width of the structures was 23.92 meters,
which was only adequate to accommodate three (3) lanes in each direction
without any shoulders. The abutments, pier and the wingwalls had stone
cladding. Due to the limited span lengths, Ocean Parkway underneath at the

76
bridge was necking down to the mainline roadway only. The service roads of
Ocean Parkway on either side of the mainline were interrupted at the bridge.

The bridge exhibited heavy deterioration with severe corrosion, deck
deterioration and bearing failure. Timber blocking was installed at the underside
of the bridge, to prevent deck spalls falling on the roadway below. At the east
abutment, temporary timber shoring was provided to support some of the stringers
and the deck was plated over on top.

PROJECT SCOPE

The design had to follow the standards established by NYSDOT for the
Belt Parkway corridor. The vertical clearance over Ocean Parkway had to be
improved from an existing of 4.30 meters to a proposed of 4.55 meters. The sight
distance also had to be improved from an existing of 75 meters to a proposed
sight distance of 130.8 meters. This project ultimately included reconfiguring six
(6) exit and seven (7) entrance ramps to the Belt Parkway within the 1,386 meter
project limits and the replacement of the existing two-span bridge with a longer
(65.736 meters), and wider (40.714 meters) structure. The project also included
the rehabilitation of approximately 250 meters of Ocean Parkway and the
incorporation of new lighting, drainage and various other safety improvements.
The objective was to bring this segment of the Belt Parkway up to current design
standards and provide a minimum 50-year useful service life. In order to achieve
the required fifty (50) year life for the bridge the NYCDOT specified the use of
stainless steel reinforcement for the deck.

Another scope item identified was service roads along Ocean Parkway.
Within the project limits there was no separation of the main line Ocean Parkway
from its service roads. It was decided that the new design would introduce malls
that would create this separation. The western mall was to carry both a bicycle
lane and a pedestrian path, while the eastern mall was to be landscaped. The
space required for these improvements dictated that the new bridge length exceed
the old length by 22.06 meters.

PROPOSED BRIDGE

As described previously in order to accommodate six lanes with standard
shoulders, the new bridge had to be made wider. Also to accommodate the
service roads and malls on Ocean Parkway, the proposed bridge had to be made
longer. Location of the malls underneath provided the area for the location of
intermediate piers for the new bridge. Based on this, the new bridge was
configured to be a three (3) span structure, with the second span going over the
Ocean Parkway mainline and the flanking spans going over the service roads.
The three (3) spans of the new bridge were 18.874, 32.804 and 14.058 meters.

77
The width of the structure was 40.714 meters. The north half of the bridge was
made wider so that during the construction of the south half of the bridge, six
lanes of traffic can be accommodated on the completed north half.

The abutments were full height abutments and the piers were multi
column bents. At the abutments, the bridge was supported by a row of tapertube
piles with a cast in place cap beam to receive the bridge bearings. T-Walls were
constructed around the capbeam to retain the embankments at the abutments. T-
Walls were selected for the abutments and wing walls to enable rapid
construction, versatility for implementing staged construction, capacity to
incorporate architectural features, and flexibility in the sub-structure to adjust to
the heterogeneous geotechnical conditions of the site. The T-Wall units were
erected in approximately five (5) days per abutment/wing wall. The abutment
piles had to be accurately located to avoid the T-Wall stems.

Pile foundations were selected for the replacement bridge to address the
compressible organic layer within the existing soil profile. The basic foundation
design was for 16-inch diameter tapertube piles for the abutments. For the piers,
tapertube piles were installed outside the fascia of the existing bridge where there
was sufficient working room. Given the constraints of the Maintenance and
Protection of Traffic for the contract, mini piles were used within the footprint of
the existing bridge installed from Ocean Parkway beneath the existing structure.

The superstructure was designed as continuous for live load, having one
pier "fixed". In order to mitigate the very high seismic forces at this pier, semi
integral abutments with monolithic back walls were used. This configuration
enabled the back walls to develop the passive resistance of the embankment
behind the back walls, mitigating the seismic forces, thereby reducing the seismic
load on the "fixed" pier.

Battered piles were needed at the "fixed" pier to resist the seismic forces.
The west pier was chosen as the location for the fixed pier because it was most
suitable location for battered piles, with the most clearance between the proposed
and existing piles.

A general elevation of the bridge showing the concepts is shown in Figure 1.

JOINTLESS BRIDGE

The Design-Build team incorporated a Jointless Bridge concept to
enhance the overall quality and durability of the structure. The longitudinal joints
between the Inverset units were cast in place with reinforcement to provide
continuity across the joints. Stringers were field spliced at the piers and a
transverse closure pour completed the continuity of the deck. The semi integral
abutment approach eliminated the need for a joint assembly at the abutment.

78
Several details needed to be investigated and developed for the seamless
implementation of the Jointless Bridge concept, especially with the use of
precast Inverset Deck Units. These details are discussed below.

Elimination of Joints at the Piers

In a usual cast in place deck bridge, elimination of joints at the pier is
easily achieved by splicing the steel stringers prior to the placement of the deck.
With Inverset units, the deck elements are precast and field splicing the steel
stringers will not be feasible if the deck extends the full span length. Therefore,
we developed a system of stopping the concrete deck short of the full length with
the stringers extending the full span length beyond the precast deck to facilitate
the splicing. After splicing of the stringers, additional reinforcement was placed
and a closure pour placed to complete the continuity. With this system the
Inverset units needed to be supported at both ends. Therefore at the piers, girders
from the adjacent spans were supported by a bearing each. Once the Inverset
units were supported the stringers were field spliced using steel plates. The
splices however, posed several challenges. The challenges and the methods of
resolutions are discussed below.

Closure Pours The length of the closure pours had to be minimized to
maximize the use of precast elements and expedite construction. This was
achieved by the use of larger diameter M24 (1 in. diameter) bolts for the splices.
The overall closure pour length was 2.00 meters. Deck reinforcement was
extended from each precast Inverset unit. Additional lap bars were placed and the
closure pours completed.

Interference with the Bearings At the bottom flanges, the bolting pattern
was changed to avoid the bearings. Splice bolts were placed on either side of the
bearings. Due to this interference and the resulting bolt spacing, the splice length
at the bottom was longer than the top splice.

Interference with the Stiffeners In a usual splice detail the top and
bottom flanges are sandwiched by splice plates on both sides of the flanges. With
Inverset units, the stringers are sent to the field complete with stiffeners and
connection plates. The splice plates to be placed below the flanges were
interfering with the stiffeners. The bearing stiffeners, as per NYSDOT standards
needed to be mill to bear or welded to the flanges. It would have been almost
impossible to fit those splice plates in the field between the stiffeners and the
flanges within the allowable tolerances. To overcome this problem, splice plates
were placed only on one side of the flanges and the splice plates on the inside of
the flanges were eliminated.

Different Beam Depths The depths of the beams on the adjacent spans
were different, especially at the West Pier, due to the significant difference in the
spans; the beam depths (W 920 vs. W840) varied as much as 48 mm. In order to

79
account for this difference in depth the splice plate was located on the inside of
the bottom flange of the deeper girder and on the outside of the bottom flange of
the shallower girder. Also additional filler plates were included to make up the
difference in depths.The splice plate details are shown in Figure 2.

Semi Integral Abutments

The use of semi integral abutment concept facilitated the elimination of
the deck joints at the abutments. With this configuration, the backwall is cast
together with the deck which spans over the abutment capbeam. The approach
slab sits on a ledge at the deck / backwall interface beyond the abutments
capbeam. Hence the usual deck joint at the front of the backwall is eliminated.
Usually this joint at the front of the backwall leaks, leading to severe deterioration
in the bearings and the bridge seat.

Backwall cast with the deck

Another advantage of this semi integral abutment configuration is that the passive
pressure behind the backwall is utilized in restraining the bridge against
longitudinal movements, especially in a seismic event. This semi integral
abutment concept is a standard detail in the NYSDOT bridge design guidelines
[3] and is accepted practice.

Longitudinal Closure Pours

Cast in place longitudinal closure pours were placed between the
Inverset units. Hoop reinforcement was placed to extend on either side of the
units, and additional longitudinal rebars were placed through the loops and the
closure pours placed in the field. This cast in place closure pours allowed the
minor elevation differences between adjacent Inverset units to be eliminated.
Also a smooth riding surface is attained for the bridge deck.

80

Transverse joints Longitudinal joints Finished deck

Check for fit in the field

With the extremely tight construction schedule, there would be no time
available to address misfits during the erection of the superstructure in the field.
To mitigate this potential problem, the Design-Build team elected to construct a
full-scale mock-up of the Belt Parkway Bridge at the Fort Miller fabrication yard.
Bridge components were set to proposed grades, and all mating pieces were field
drilled and assembled. This virtually eliminated any fit problems during
superstructure erection.

Entire Bridge Assembled at Fort Millers Yard to check for fit.

CONSTRUCTION STAGING

The bridge was constructed in two stages. During the first stage the
westbound half of the bridge was demolished and replaced with the new section.
A temporary Mabey Bridge was installed on the south side of the existing bridge
to maintain the three lanes of eastbound traffic during this phase. Once the
westbound section was completed, all six lanes were accommodated on that new
bridge and the remainder (eastbound half) of the bridge was demolished and
replaced. The new westbound half had to be six meters wider than the required

81
width, in order to accommodate six lanes of traffic during the second phase of the
project.

Stage I of the bridge replacement was completed around July 2004 and the
entire bridge replacement was completed in November 2004. This is about a
month ahead of the already aggressive schedule set out by the contractor.

CONCLUSIONS AND RECOMMENDATIONS

Currently the bridge is open to traffic in both directions, and it is
performing very well. This case study shows that Jointless Bridges are very
effective and will result in extensive cost savings in reduced maintenance costs
and enhance life. Construction of similar bridges, especially in congested areas
with very high maintenance costs shall be encouraged.

It is recommended that this Jointless Bridge along with other bridges in
the City be monitored for its performance over the coming years, and future
bridge designs be modified to accommodate the innovations used in this bridge to
achieve a jointless structure.

ACKNOWLEDGEMENTS

Owner: New York City Department of Transportation
Chris Sklavounakis, Director, Emergency/ Design Build Contracts

Contractor: Granite Halmar Construction Company, Inc.
Paul Atkins, Area Manager

Designer: Gannett Fleming Engineers and Architects, P.C.
Charles Norrish, Vice President

Resident Engineering Inspection:
HAKS Engineers / KS Engineers JV
Jerry Liebowitz, Vice President, HAKS Engineers and Surveyors

REFERENCES

1. Burke, Jr., M. P., Integral Bridge Design is on the Rise, Modern Steel Construction, July-
August, 1990.

2. Distribution of Pile Loads and Earth Pressures for Design of Integral Abutment Bridges,
Study Proposal for Research Project 226-1, New York State Department of
Transportation, September 1996.

3. Bridge Detail Sheets BD-IA1 thru BD-IA5, Integral Abutments, Engineering Bulletin 02-
033, Structures Design and Construction Division, New York State Department of
Transportation, 2002.

82

83

84
Case Study Jointless Bridge
Beltrami County State Aid Highway 33 Over
Mississippi River in Ten Lake Township, Minnesota

Jeffrey Wetmore, P.E. Senior Project Manager, Edwards and Kelcey, Inc. and
Bruce Peterson, P.E. Project Manager, Edwards and Kelcey, Inc.

ABSTRACT

In Climbing Mount Improbable Richard Dawkins makes the observation that
in nature, evolution makes use of existing structures and functions, while
engineers seek a start with a clean piece of paper. The re-design Minnesota Bridge
04519 is an example of the natural, evolutionary process applied to an
engineering problem.

Expansion joint devices and bearings are among the most maintenance
intensive and troublesome components of a bridge. Leaking joints lead to
deterioration of other components, including reinforced concrete, beams, bearings
and paint. Owners with limited maintenance budgets and staffs, seek to reduce the
maintenance associated with these components.

After completion of the final design and preparation of construction plans for
Bridge Number 04519, the Beltrami County Engineer requested a change, based
on his experience in another state. He asked to eliminate the expansion joints and
use integral abutments on the bridge. He expected these changes would reduce
maintenance. The case study of Minnesota Bridge Number 04519 illustrates a
seldom-used approach, taking advantage of the existing bridge elements in
developing a jointless bridge.

Working with the County Engineer and consulting with the Minnesota
Department of Transportations State Aid Bridge Engineer and the State Bridge
Engineer, the designer proposed moving the expansion devices away from the
abutments by tying the approach panels to the deck slab and providing a joint
between the concrete approach panels and the bituminous roadway.

The bridge has been in service approximately eight years. An inspection of the
structure and examination of maintenance records, for the bridge and approach
roadways, provide information about serviceability of the bridge and the degree of
success at accomplishing the goal of reducing maintenance costs. While, there
may be bridges in other states that share similar details, the authors are not aware
of any similar structures in Minnesota.

85
Figure 1 Location Map
Figure 2 General Plan
INTRODUCTION

In most structural applications stiffness is more critical for serviceability than
strength. We applied this principle in the redesign of Bridge 04519 for Beltrami
County. Since the concept of making the bridge deck jointless was introduced
after design and detailing had been completed, we sought to utilize the design
results and details that had been developed for a conventional bridge, rather than
starting with a clean sheet of paper, or utilizing typical details for integral
abutments. This approach allowed us to explore ideas that had not previously
been implemented on jointless bridges in Minnesota.

Beltrami County is a rural
county in northern Minnesota.
County State Aid Highway 33 is
a low volume road, with 20 year
projected average daily traffic
(ADT) of less than 1000. The
seasonal temperature range is
large, even by Minnesota
standards with low temperatures
of more than 40 degrees
Fahrenheit below zero. (See
Figure 1. Location Map)

Bridge 04519 was designed
to replace a structurally
deficient and functionally
obsolete bridge. The existing
bridge and approaches were
narrow, the profile did not meet
sight distance requirements and

86
Figure 5 Abutment Section
Figure 3 Elevation
the bridge was load posted. The site had numerous constraints, including wetlands
that would be impacted by adding the required shoulders and improving the
profile to provide
sight distance. The
area also was
expected to contain
cultural resources
Native American
burials.

To satisfy the
constraints cost
effectively, the bridge
was proposed as a
three-span prestressed
concrete beam bridge with pile bent
piers and semi-tall parapet abutments.
The beams utilized were 36 inches deep,
spanning 53 feet from bearing line to
bearing line, for an overall length of
167-8. (See Figure 2. General Plan,
Figure 3. Elevation, and Figure 4.
Typical Section)

The semi-tall parapet abutments
provide an optimal solution for the span
length to abutment height trade off.
Since both the cost of abutments and the
costs of superstructure supporting
elements are square functions of the
height of abutment and span length
respectively, a minimum solution to this
problem is the least sum of the squares.
The optimum solution does not occur
when either square is at its maximum
value and the other is at its minimum.
Thus semi-tall abutments are usually the preferred solution compared to stub
Figure 4 Typical Section

87
abutments or tall abutments. (See Figure 5. Abutment Section)

A three span alternative minimized structure costs while maintaining an
adequate waterway. It was necessary to minimize the depth of structure to permit
recreational boat traffic under the bridge while to minimizing the roadway grades.
The semi-tall abutments, and wing walls parallel to the roadway, minimized the
impacts to the adjacent wetlands, and the excavation in areas that could contain
cultural resources.

EVALUATION OF MODIFICATIONS

For the reasons stated above, Bridge Number 04519 had been designed as a
conventional Minnesota State Aid highway bridge - using parapet abutments and
expansion joints at both ends of the bridge. Given this conventional design,
conversion to a jointless bridge with integral abutments would have entailed a
significant redesign effort. Several options were available, including revising the
bridge length to accommodate the shorter abutment height common with integral
abutments. The incorporation of a conventional integral abutment would have
required redesign and re-detailing of all abutment drawings. Changing to the more
flexible wing walls parallel to the abutment stem would have required re-grading
the approach embankments, resulting in additional wetland impacts, and a new
permit application, with higher costs of mitigation due to the greater impacts to
wetlands.

Another alternative would have been to make the stiff parapet abutment
integral with the superstructure. This approach is common in Tennessee.
Tennessee DOT uses jointless bridges with longer lengths, larger skew angles and
taller abutments than other DOTs do. Given the more extreme temperatures that
Minnesota bridges experience compared to bridges in Tennessee and the limited
experience that Minnesota has with integral abutment bridges, we considered this
approach too risky.

Instead, we explored alternatives to an integral abutment that would eliminate
expansion joints at the abutments, while minimizing the redesign and detailing
costs and time of such a significant design change. Our alternative would use, to
the extent possible, the designs, details and plans that had been developed. This
included using the superstructure, bearing, abutment, pier and approach panel
details, while moving the expansion devices off of the abutments where leaking
joints can cause damage.

Relocating the expansion joints involved several tradeoffs that we evaluated
qualitatively. We proposed moving the expansion devices by connecting the
approach panels to the deck slab and providing a joint between the approach
panels and the bituminous roadway. Concrete approach panels are a typical
feature of bridges in Minnesota. The approach panel bridges the area of fill behind

88
the abutment, providing a smoother transition and reducing roadway settlement
from the fill behind the abutments. Because approach panels were used,
connecting the approach panel to the bridge deck did not result in additional
materials or work in the construction of the bridge. We were able to take
advantage of an element that was already used, which only needed a slight change
to modify its function.

The designers had seen a case study of a bridge in Pennsylvania,
demonstrating this concept of moving the expansion joints off the bridge to the
end of the approach panels. This solution did not represent an untried departure
from bridge engineering practices, but only an extension of its application to
Minnesota.

One of the tradeoffs identified from relocating the expansion joints was that
damage to the bridge, beams, bearings and abutments would be exchanged for
damage to the pavement abutting the approach panel. The concrete approach
panel has much higher strength and stiffness than bituminous pavement. In the
conventional applications there is relatively little movement of the concrete panel
relative to the bituminous pavement, so the concrete panel inflicts little damage
on the pavement. By relocating the joint, we were knowingly creating a situation
where there could be movements of the approach panel and forces imposed on the
bituminous pavement.

With our goal of utilizing the previous design and details, but relocating
expansion joints to reduce damage to the bridge, we performed the redesign and
detailing. We performed approximate analyses to consider shrinkage and
differential thermal movements of the bridge superstructure, approach panels and
pavement. Key considerations included determining the locations of tensile strains
from the volume changes and from external forces. We planned to modify details
that created fixity and restraints to minimize these conditions. The examination of
details was conducted with an emphasis on stiffness and ductility, rather than
strength and allowable stress. We followed Hardy Cross advice, Strength is
essential, but otherwise unimportant.

Starting at the center of expansion, (span 2) and moving to the extreme limits
of the structure (approach panel / bituminous roadway joint) we examined the
changes required to eliminate the deck expansion joints at the abutments, and to
accommodate the resulting differential movements and forces.

In span 2 and all of the superstructure, we expected the tension and
compression forces across the entire superstructure would be slightly greater than
a conventional structure due to the added restraint in the deck from the connection
to the bridge approach panels, which are not free to move. In a conventional
bridge, all longitudinal restraint is a result of forces transmitted through the
bearings to the beams. Both fixed bearings and expansion bearings transmit forces
to the beams. Based on the design procedures used for elastomeric bearings and

89
the temperature dependent properties of the elastomer, we calculate that the
restraint forces can equal up to 15 to 20 percent of the vertical dead load on the
bearings. Typical design procedures do not account for these tension or
compression loads, yet beams and superstructures perform adequately. Other
internal loads include shrinkage of the concrete in the beams, and shortening of
the beams due to creep under the prestress loads. Shrinkage of the concrete deck
will tend to impose compression stresses in the top flange of the beam and tension
stresses in the bottom fibers of the beam since this load is eccentric to the beam.
Overall, the beam will be put into compression, and the deck will be in tension in
self-equilibrating forces. Additional thermal forces include temperature gradients
from conditions where heating or cooling of the deck occurs more rapidly than in
the beams. There are also gradients from heating of a fascia beam that depend on
the orientation of the bridge fascia to the sun. Several research efforts have
examined the strains and stresses in beams and decks that are possible with the
restraint caused by fixity at the abutments. The calculated stresses can be high,
but field observation of integral abutment bridges has not shown detrimental
affects on bridges with similar lengths, and geometry to the subject bridge. Based
on this analysis, we believed our design was within the limits where experience
has shown adequate serviceability could be attained without detailed analysis.
Therefore we determined that the deck and beams required no changes to the
original design or details.

The piers also required no changes to the original design. The added
longitudinal restraint provided by the approach panels resulted in both the
external forces and the internal forces on the piers being lower in the new design.
External forces applied to the superstructure in the longitudinal direction are
distributed to the piers, to the expansion bearings at the abutments and to the
approach panels. The coefficient of friction between the concrete approach panels
and the soil provides a stiffer load path than the piers or abutment bearings,
meaning that most of the longitudinal external loads would be redistributed from
the piers to the approach panels by the design change. The restraint provided by
the approach panels would also reduce the deflections and internal forces on the
piers. Restraint at the ends of the deck reduces the movement at all points within
the deck relative to the unrestrained condition.

The approach panels provide restraint in the transverse direction. In addition
to the panel to soil friction, the approach panels are confined between the wing
walls of both abutments. The deck acts as a horizontal diaphragm, transmitting
transverse loads to the approach panels. Since this behavior results in the pier
design being more conservative, it was unnecessary to perform a detailed analysis
to determine the amount of load that was redistributed from the piers to the
approach panels.

90
Figure 7 Deck / Approach Panel Section at Abutment
Figure 6 Saw Cut Detail
The elimination of the expansion joint at the abutments changes the support
fixity condition of the beams. Beams are typically analyzed as simply supported,
free to translate and rotate.
Eliminating the joint
provides partial restraint
against both rotation and
translation. The rotational
restraint creates a negative
moment in the deck at the
abutment support, and also
reduces the positive moment
in mid-span. Bridges do not
meet the idealized condition
of zero horizontal restraint at
any bearing. At locations
where the deck is continuous
over a support, such as at piers (using Minnesota standard detailing practices)
beams also do not meet the idealized conditions assumed for design. The
continuous deck provides restraint. Typically additional steel is placed in the
negative moment region to control cracking. A V groove is placed in the bottom
of the deck and a saw cut is made in the top of the deck so that a single crack is
formed in a controlled location. (See Figure 6. Saw Cut Detail) Based on the
standard practices, we knew that the beams and the deck performed adequately
with the negative moments, typical in prestressed beam bridges.

Integral abutment details create a negative moment condition at the abutment
support. The guidelines and typical details that several states have published
indicate that with attention to detailing the crack control and serviceability is
adequate
without
detailed
analysis of the
magnitude of
the moments
at the
abutment. The
details we
examined for
integral
abutments
most likely
provide a
greater
amount of rotational restraint than the detail we proposed. Integral abutment
details usually include a heavy abutment diaphragm cast integrally against the
beam ends and connected to the abutment seat with reinforcing bars. Section

91
Figure 8 Approach Panel Transverse Section
properties of the abutment diaphragm are larger than the section properties for the
approach panel. The elastomeric bearings under the beams on our proposed detail
will allow some rotation of the beam end by permitting translation of the bottom
flange of the beam. (There is no translation of the beam relative to the abutment
on integral abutment detail bridges.) The condition at the abutment will not be
significantly different from the condition at a pier where the deck is continuous.
Based in this qualitative analysis we believed the tension in the top fibers of the
deck would be in the same approximate range as they are over a pier where the
deck is continuous, and that serviceability would be adequate without special
considerations. We included considerations such as additional reinforcement over
the support, and the provision of a contraction joint that could be sealed. (See
Figure 7, Deck / Approach Panel Section at Abutment)

We evaluated several other conditions in the deck near the abutment. The
section properties change at the termination of the beams. Both the area and the
flexural capacity of the super-structure change. With the large change in flexural
capacity, we expected strain concentrations in the deck would occur that would
have the potential to increase the cracking at the location at the end of the beams.
Comparing the conditions at the abutment to the conditions at a pier, we
concluded that conditions for strain concentration and deck cracking were not
significantly different at the abutment than they are over a pier. Since the deck
depth transitioned 12 inches at the beam end to match the approach slab
thickness we think there is a less severe transition in section at the abutments
than at the piers.

We next evaluated the elastomeric expansion bearings at the abutment. The
thickness of the elastomer was sized to provide a shear strain of 50 percent to
accommodate a full temperature range for concrete bridges, and allowances for
creep and shrinkage of the prestressed beams. We expect some movement at the
abutment bearings since there has not been an attempt to completely restrain to
the deck at the abutments. But we expect the total movement of the bearings will
be less than for a conventional bridge designed with an expansion joint at the
abutment.

Another condition we evaluated at the deck to approach panel interface was
the vertical stiffness of the beams and bearings compared to the stiffness of the
abutment back wall. The elastomeric bearings and the soil have a much lower

92
Figure 9 Approach Panel
Longitudinal Joint Detail
stiffness than concrete. Each passage of a live load would result in a vertical
deflection of the bridge deck and approach panel. The abutment back wall would
offer a much stiffer support than the condition on either side of the back wall.
With short transitions between areas of differing stiffness, we believed that the
strains from differential movements would be relieved by cracking of the concrete
approach panel, a condition we wanted to avoid. Our solution was to soften the
abutment back wall and lengthen the distance between differential vertical
movement points, by placing a one inch layer of joint filler on top of the abutment
back wall. (See Figure 7. Deck / Approach Panel Section at Abutment)

The durability of the joint filler was not a concern. If the filler deteriorated the
deck would maintain a gradual transition between the stiffness of the bearings
supporting the bridge deck and the soil supporting the approach panel. If the soil
immediately behind the abutment back wall
settled slightly or was displaced, the
transition distance between points of support
would increase slightly. The 12-inch thick
approach panel can easily span a sufficient
distance to eliminate reasons to be
concerned about settlement, loss of soil or
loss of joint filler. (See Figure 8. Approach
Panel Transverse Section)

We examined two options for the
interface between the approach panel and
bridge wing walls. Our chosen approach
was to form a vertical joint between the
approach panel and the wing wall, with the barrier mounted on the wing wall. The
other option was to extend the approach panel over the wing wall and mount the
barrier on the approach panel. With the second option the approach panel would
have been supported on elastomeric bearing material with an expansion capacity
and vertical stiffness similar to the abutment bearings, for reasons discussed
above.

The durability of longitudinal joint between the approach panel and the wing
wall was another area of concern. Separating the two concrete pours using joint
filler would allow differential movement, but we expected the movements to
degrade the joint filler in a few years. We determined that even if the joint filler
degraded, there would not be an adverse effect. The joint would fill with fine
particles but would not create additional restraint. Since the joint is located at the
gutter line, water and chloride runoff concerned us. Minimizing the infiltration of
water into the wing wall back fill is a standard practice, even though the backfill
material is free draining granular material. We adapted a standard joint detail that
we thought had the greatest tolerance for longitudinal movement. The detail
includes a backer rod and silicon sealant (See Figure 9. Approach Panel
Longitudinal Detail). Periodic maintenance can be performed on this type of joint.

93
The silicon can tolerate high shear strains. We also modified to the profile of the
approach panel to include a small lip and included an under-drain system to
intercept water before it infiltrated the backfill to reduce the exposure of the
concrete to chlorides.

The last area of concern, where the differential movements are the largest, is
the transverse joint between the approach panel and the bituminous pavement.
Movements of unrestrained concrete elements, subject to a temperature range of
150 degrees Fahrenheit can be easily calculated and provide an upper bound of
movement of about one inch.

Our analysis of the situation and several factors regarding the construction of
the bridge and roadway pavement lead us to conclude that the differential
movements are likely less than calculated and that there are mitigating factors and
tradeoffs that allow a jointless bridge to function adequately.

The approach panel is restrained against movement by the friction between
the soil and the concrete. From the vertical loads involved and the coefficient of
friction between the soil and concrete it is clear the panels are not fully restrained.
Some amount of movement between the approach panel and the pavement must
be accommodated. This problem is not unique to this style of jointless bridge.
Bridges with integral abutments have the same problem. An integral abutment
does not totally restrain the superstructure. Design assumptions and detailing
criteria make the integral abutments as flexible as possible, allowing movement
between the abutment and the pavement. Bridge 04519 was only 167 feet long,
much shorter than the guidelines several states use for allowing integral
abutments, so other bridges and pavements performed adequately with
movements larger than we were planning.

The tradeoff we accepted is that if the movement of the approach panel caused
damage to the pavement, periodic repair of the pavement was cheaper and easier
than bridge and joint repair. Pavement repair would be routine maintenance,
performed by county maintenance crews. No special plans, materials, equipment
or skills would be required, only those the county maintenance forces possess.

INSPECTION RESULTS

We obtained copies of the Mn/DOT Pontis Bridge Inspection Reports and the
Mn/DOT Structure Inventory Reports. These reports were current through
November 23, 2003. We also conducted our own on-site inspection to confirm the
written data, and to observe and document the performance of key bridge and
roadway elements.

94
Figure 10 - Deck Cracking
Figure 11 Approach Panel/Pavement Joint

The Structure Inventory Report
listed condition codes for all
elements as 8 and 9, except deck
geometry with an appraisal of 7. All
structural appraisals were 8s.

The Bridge Inspection Report
provides more detailed condition
information than the Structure
Inventory Report. The first noted
deficiency was for deck cracking
radiating from edge of the top flange
of the beam to the corner of the
overhang. (See Figure 10. Deck Cracking)

The second deficiency noted, in the
2001 inspection, was that the bituminous
pavement adjacent to the approach panel
required patching. The 2002 inspection
noted that patching had been done. A
later inspection noted the condition
unchanged (OK). (See Figure 11.
Approach Panel / Pavement Joint) Other
notes addressed items that are unrelated
to the bridge being jointless, including
rip rap damage and graffiti on
abutments, beams and railings.

Our field inspection of the bridge,
conducted November 25, 2004,
confirmed the conditions noted in the
annual inspections. We documented the
general layout of the bridge, conditions
and structural details with photographs
at that time.

ADDITIONAL COMMENTS

The only unexpected behavior of the bridge is the cracking of the deck,
evident in each corner. The cracks formed from the corner of the top flange of the
precast beam to the notch where the deck changed width to fit between the wing
walls. The designers had not contemplated any transverse restraint at this location,
and had not considered that the notch would behave as a reentrant corner. It is
well understood that slabs that are restrained at reentrant corners are subject to

95
cracking. When detailing slabs, it is desirable to eliminate details containing
reentrant corners. It is normal practice to place additional reinforcement
perpendicular to the direction of possible cracking at locations where reentrant
corners cannot be eliminated.

We recommend additional of reinforcement for control of cracking at these
locations. The cracks are tight. There is efflorescence on the underside of the deck
at the crack locations, however, the reinforcing steel is epoxy coated, so
deterioration of the reinforcement is not expected. Annual inspections will
monitor this condition. The locations are readily accessible for both inspection
and repair, if ever required.

Further analysis may result in a better understanding of the restraint or volume
change effects that produced cracks at these four locations. Matching the
approach panel with the deck, to cover the wing walls and placing the barriers on
the approach panel could eliminate the cracking at this location. This may lead to
other difficulties such as attaching the approach guardrail to the barrier. If a
separate end post is used for the guardrail connection, the crack locations may
move, but may not be eliminated.

Small economies of construction could be gained if a design was started from
scratch by recognizing the load paths through the deck to the approach panels,
reducing the horizontal forces used to design the substructures. Based on
subsequent review of the options for detailing, the forces the bridge elements are
subject to and the various load paths that can be designed, it is clear that the
horizontal forces that were applied to the substructure elements are greater than
required. Re-analysis could result in a slight reduction of the number of piles
required for the piers and for the abutments, since the longitudinal and transverse
forces can be redirected to a load path through the approach panels.

It may be possible to eliminate the end diaphragms at the abutments and
instead use intermediate diaphragms steel cross frames to simplify forming of
the deck. The stiffening function of the end diaphragm is not critical with the
jointless bridges, since the deck is continuous into the approach panel and there is
not a free, unsupported edge of slab, as there is in a conventional deck.

CONCLUSIONS

Eight years of service is too soon to judge the results for a bridge that is
expected to last 75 years or more, but the initial results indicate that the bridge is
performing as intended, and the tradeoffs made appear justified.

96
ACKNOWLEDGEMENTS

The authors would like to acknowledge Thomas Martin of the Minnesota
Department of Transportation, Bridge Management Unit for his assistance in
providing Pontis Bridge Inspection Reports and Structure Inventory Reports. We
also would like to acknowledge Dave Graber, Senior Engineering Technician
from Beltrami County, for his assistance discussing the annual bridge inspections
and his recollections of the construction inspections he performed in 1996.

We would also like to acknowledge retired Beltrami County Engineer, Tom
Kozojed, for allowing us to try an unconventional approach to bridge design and
detailing. Donald Flemming, retired Minnesota State Bridge Engineer, was also
instrumental in permitting the project to go forward, as was Kevin Western,
former Minnesota State Aide Bridge Engineer. When new approaches are taken
in solving traditional problems many individuals have the opportunity to stifle
innovation by withholding support from the new ideas. Fredrich Hayek can be
paraphrased in this regard. Progress only occurs when the few act in a way
discouraged by the majority. We want to acknowledge the leadership by these
individuals who allowed our unproven approach to be tried.

97
Design and Construction of Dual 630-foot, Jointless,
Three-span Continuous Multi-girder Bridges in
St. Albans, West Virginia, United States,
Carrying U.S. Route 60 over the Coal River

John Perkun, P.E., SAI Consulting Engineers, Inc.
and Keith Michael, P.E., SAI Consulting Engineers, Inc.

ABSTRACT

The aging bridge that carries U.S. 60 over the Coal River in St. Albans, West
Virginia was programmed for replacement by the West Virginia Division of
Highways (WVDOH) in the late 1990s. In 2000, the WVDOH retained SAI
Consulting Engineers, Inc. to design a replacement for the existing 570-foot, five-
span, riveted steel girder bridge. The existing bridge has four travel lanes and two
3-foot sidewalks. The existing superstructure consisted of three main girders with
floorbeams and stringers, was on a 38-degree skew, and utilized sliding plate
expansion dams at the two abutments. The replacement bridge would be wider
than the existing bridge and would consist of four 12-foot travel lanes, a raised 4-
foot center median, two 6-foot shoulder/bicycle lanes, two concrete parapets, and
two 5-foot sidewalks (78'-7-1/2").

In accordance with standard WVDOH design policy, one of the primary goals
of the project was to minimize or eliminate bridge joints where possible. During
the preliminary span arrangement study phase, two types of replacement
structures were studied and are shown below.

1. One bridge that would carry the entire proposed 78'-7-1/2" cross-section.
This alternative would consist of a 530-foot-long, three-span, continuous
steel bridge with skewed (38-degree) abutments and piers.
2. Two separate, parallel 39'-3-3/4"-wide bridges with perpendicular
abutments and piers. Each bridge would be a 630-foot, three-span,
continuous steel bridge. The use of perpendicular abutments and piers
required these substructure units to be longitudinally offset by 30.5 feet.

The foundations chosen for the piers (single-shaft, hammerhead type) were
deep-drilled shafts bearing on bedrock. The drilled shafts were as deep as 60 feet
and were particularly adaptable to the perpendicular substructure configuration
chosen. The integral abutments were founded on a single row of steel H-piles
with the weak axis of the pile oriented in the direction of the bridge movement.

Special details were generated due to the jointless design, primarily to take
care of bridge movements at the abutments and approach slabs.

98
The bridge is currently under construction and is due for completion by
December 2005.

INTRODUCTION AND DESCRIPTION OF PROJECT

One of the primary problems facing bridge engineers and bridge owners over
the years has been the degradation of bridge structures due to failure of joints
placed in the bridge to handle the normal movements. Once the integrity of the
joint is compromised, many problems can arise including water infiltration to
vulnerable parts of the bridge structure, failure of the joint to provide the
movement required, and damage to the deck or riding surface itself. This is
especially true in climates subject to freeze-thaw cycles where water infiltration
through a failed joint can cause dramatic problems to steel and concrete bridge
elements alike.

One solution to this joint failure problem has been to design bridges with no
joints thus averting, to a great degree, problems as described in the previous
paragraph. This, of course, provides challenges relative to how the movements of
the bridge can be effectively handled and to provide longevity for the bridge
structure. Since the displacements of these jointless bridges cannot occur on the
bridge superstructure, provisions for movement must be made at the ends of the
bridges through the use of integral and semi-integral abutments.

One such challenge was put forth by the West Virginia Division of Highways
for the replacement of an aging bridge in St. Albans, West Virginia. The existing
structure is a five-span continuous girder-floorbeam-stringer structure (90'-120'-
150'-120'-90'). It is supported by concrete abutments supported on timber piles
and four multi-column piers with solid web walls. Two of the piers are supported
on timber piles and two are supported on spread footings founded on rock below
the riverbed. The overall bridge length is 570 feet.

The structure carries a four-lane divided highway with sidewalks on each side
of the bridge. The out-to-out bridge width is 60'-0" with a 48'-0" roadway width.
Each traffic lane is 12'-0" wide. There are no shoulders. The sidewalks are 3'-0"
wide with a 1'-0"-wide concrete parapet barrier.

The replacement bridge would be wider than the existing bridge and would
consist of four 12-foot lanes, a raised 4-foot center median, two 67-foot
shoulder/bicycle lanes, two concrete parapets, and two 5-foot sidewalks (78'-7-
1/2"). In accordance with West Virginia Division of Highways policies, bridges
are to be designed with a minimum number or no joints at all. The challenge was
in placeto design one of the State of West Virginias longest jointless bridges.

99
WEST VIRGINIA DIVISION OF HIGHWAYS DESIGN PHILOSOPHY FOR
JOINTLESS BRIDGES AND INTEGRAL ABUTMENTS

In the 1990s, the West Virginia Division of Highways adopted standards for
design and construction of jointless bridges. These standards were contained in
West Virginia Structural Directive 26 (SD-26). This standard provided design
guidelines along with construction details for design of the bridge abutments.
This was the design standard in place during the design phase of this project. The
content of SD-26 has recently been incorporated into the newly published West
Virginia Division of Highways Bridge Design Manual, Section 3.9 Jointless
Bridge Abutments.

The Standard for Jointless Bridge Abutments states, Fully integral and semi-
integral abutments shall be used whenever possible to eliminate deck expansion
joints. The standard also establishes an upper limit threshold for anticipated
bridge movements at 2 inches for the use of integral abutments. It also provides
guidelines for the use of approach slabs whenever anticipated movements exceed
one-half inch. See Figures 1 and 2 for West Virginia Division of Highways
details.

Figure 1
Semi-Integral Abutment with Approach Slab
West Virginia Division of Highways

100

West Virginia has a growing list of jointless bridges that employ integral and
semi-integral abutments. See Table 1.

Table 1. West Virginia Jointless Bridges
Bridge Name Length Span Arrangement
Lost River Bridge No. 1 666.1' 173.9' 262.5' 229.7'
US 220 Ramp Connector
Bridge
615' 136'-171.5'-171.5'-136'
Edgewood Drive Bridge 586' 111.25'-111.25'-140'-111.25'-111.25'
Dumpling Run Bridge 543' 167'-209'-167'
Elizabeth Bridge 536' 163'-210'-163'
Sauerkraut Bridge 524.9' 154.2'-216.5'-154.2'

In the case of the Coal River Bridge, the total projected movement of the
bridge was 3 inches and clearly indicated the use of approach slabs but seemed to
rule out the use of a jointless design. However, further studies would lead to the
incorporation of integral abutments into the design.

Figure 2
Semi-Integral Abutment with Approach Slab
West Virginia Division of Highways

101
SPAN ARRANGEMENT PHASE

The span arrangement study investigated three-span and five-span
replacement structures. The proposed typical section consisted of four 12'-0"
travel lanes, two 6'-0" shoulders, two 5'-0" sidewalks, and a 4'-0" median. A
concrete barrier with an aluminum pedestrian railing will be constructed on the
outside edge of the sidewalk. A concrete safety-shape parapet will be on the
inside edge of the sidewalk to separate the sidewalk from the shoulder.

The bridge will be designed for HL93 live load utilizing Load and Resistance
Factor Design (LRFD) in accordance with 1998 AASHTO specifications and the
latest interim revisions, WVDOH LRFD Policy Statement dated June 23, 1998
and West Virginia Structural Directives. Structural steel will be AASHTO M270,
Grade 50 (Weathering Steel), and AASHTO M270, Grade HPS70 (Weathering
Steel). Pile foundations will be designed in accordance with 1996 AASHTO
Standard Specifications for Highway Bridges, 16
th
edition (LFD/ASD) with the
latest interims. Drilled shaft foundations will be designed in accordance with
Publication No. FHWA-IF-99-025: Drilled Shafts: Construction Procedures and
Design Methods Utilizing LRFD.

The deck will be an 8-inch monolithic deck using Class H concrete. Stay-in-
place formwork will be specified and used for the girder design. The empirical
deck will be used. The eastbound and westbound lanes of the bridge have cross-
slopes of 0.02 ft./ft. The sidewalks on each side of the bridge will have cross-
slopes of 0.02 ft./ft. The bridge deck and superstructure design will accommodate
future deck replacement while maintaining two lanes of traffic, one lane for each
travel direction. The proposed bridge will provide a hydraulic opening that will
accommodate the 100-year flood with several feet of freeboard.

The following alternates were initially considered:

Table 2. Alternatives Considered
Alter
nate
Bridge
Type
Span
Arrangement
Girder
Spacing
Girder
Web
Size
No. of
Girders
Cost
(Millions)
Comment
1 Steel plate
girder
165'-200'-165'
= 530'
9'-0" 72" x
1/2"
9 $4.96 Use 50-ksi
steel.
2 Steel plate
girder
165'-200'-165'
= 530'
9'-0" 72" x
1/2"
9 $4.99 Use 50-ksi
steel in
positive
moment
regions and
70-ksi steel
in negative
moment
regions.

102
Alter
nate
Bridge
Type
Span
Arrangement
Girder
Spacing
Girder
Web
Size
No. of
Girders
Cost
(Millions)
Comment
3 Post-
tensioned
concrete I-
beam
165'-200'-165'
= 570'
10'-0" 102"-
deep
Type J
beam
8 $5.12 Five
concrete I-
girder
would be
spliced
together by
post-
tensioning.
4 Continuou
s
prestresse
d concrete
I-beam
and two
single-
span
prestresse
d concrete
I-beams
80'-150'-150'-
150'-80' =
610'
9'-0" 60"-
deep and
96"-
deep
Modifie
d Type
IV
Beams
9 $5.24 150'
interior
spans
would be
designed as
simple
spans for
dead load
and
continuous
spans for
live load.

Because the concrete options were more expensive, it was decided to
eliminate them from further consideration. The steel options would be the only
options advanced for further studies.

After submission of the span arrangement report, the West Virginia Division
of Highways decided to revise Alternate 2. The revision involved lengthening of
the bridge and providing a span arrangement of 195'-240'-195'. The reason for
this was three-fold:

1. Placing the abutments for the new bridge outside of the existing abutments
eliminates any interference with the existing piling. This is particularly
important for the pre-drilled and placed steel piling necessary for the
integral abutment design.
2. The 240-foot main span provided a larger hydraulic opening in the Coal
River.
3. This span arrangement allowed for the design of two separate side-by-side
bridges.

The decision was made to advance Alternates 1 and 2 for more detailed
studies.

103
PRELIMINARY DESIGN PHASE

The following is a summary of the two alternates advanced for further studies:

Table 3. Two Alternates Advanced for Further Study
Alternate 1 Alternate 2
Bridge Type Welded steel plate
girders
Welded steel plate girder
Span Lengths 165'-1200'-165' = 530' 196'-6"-230'-0"-196'-6" =
623'
Girder Spacing 8'-0" 8'-0"
Girder Size 57" web with varying
flange thickness
66" web with varying
flange thickness
No. of Girders 10 10 total (five each from
eastbound and westbound
structures)
Bearings High-load, multi-
rotational bearings
High-load, multi-
rotational bearings
Steel Type AASHTO M270 grade
50W and Grade HPS
70W
AASHTO M270 Grade
50W and Grade HPS
70W
Skew 38 0 (Radial)
Abutment Type Semi-integral Integral
Pier Type Four-column bent with
concrete cap
Single-column with
hammerhead cap

With the two final arrangements chosen, more detailed studies were
undertaken. The following issues were studied in more detail:

1. Thermal movements and jointless bridge behavior exceeding the West
Virginia Division of Highways 2.5 maximum movement for integral
abutments.
2. Geotechnical and preliminary foundation recommendations
THERMAL MOVEMENTS AND JOINTLESS BRIDGE BEHAVIOR

Alternate 1 was analyzed with the assumption that semi-integral abutments
would be utilized and that the bridge superstructure would be fixed at the piers.

The semi-integral abutment configuration would provide for the connection of
the approach slab to the semi-integral abutment cap. A high-load, multi-rotational
bearing would be placed under the steel girder to allow for movements.

Details would have to be generated where off-structure elements of the bridge
interface with the moving elements of the semi-integral abutment structure.
These details will be discussed in the Final Design section of this paper.

104
The pronounced skew of the Alternate 1 bridge and the unpredictability of
how the skewed approach slab would behave under movement were concerns to
the design team. In theory, the skewed configuration should be capable of
handling the predicted 3-inch bridge movement but, in practice, it is usually
prudent to design these types of bridge elements with little or no skew.

Alternate 2 was analyzed with the assumption that integral abutments would
be utilized and that the bridge superstructure would be fixed at the piers.

The integral abutment configuration would provide for a connection of the
steel girder directly to the concrete abutment and the approach slab. A single row
of predrilled and driven H-piles, with the weak axis of the H-pile oriented in the
major bridge movement direction, would allow for projected movements of the
bridge.

Details would have to be generated where off-structure elements of the bridge
interface with the moving elements of the integral abutment structure. These
details will be discussed in the Final Design section of this paper.

The idea of designing two separate bridges and providing radial abutments
and piers tended to simplify the bridge design. The elimination of the skew
abutment/approach slab configuration in favor of parallel offset abutments and
approach slabs instilled greater confidence in the design team. It was thought that
the bridge movements would be more predictable in this configuration. It also
tended to simplify detailing of various bridge elements with the elimination of the
skew.

GEOTECHNICAL AND FOUNDATION STUDIES

A geotechnical study was undertaken in order to determine the required
substructures for Alternates 1 and 2. Eighteen borings were drilled. Due to the
considerable depth to sound rock, deep foundations are recommended.

Alternate 1

The abutments for the skewed option will be semi-integral abutments
supported by piles due to the bridge skew of 38
o
. In accordance with SD26-1,
integral abutments are limited to 30
o
skew. The wingwalls at each end of the
bridge will be 24 feet in length and placed on pile-supported footings. The piles
should be provided with driving tips and driven to refusal to Elevation 513' at
Abutment 1 and to Elevation 516' at Abutment 2. Rock slope protection will be
placed at each end of the bridge in front of the abutments.

105
Alternate 2

For the radial option, integral abutments supported by one row of piles are
recommended. The estimated pile tip elevation for Abutment 1 is Elevation 513'
and for Abutment 2 is Elevation 516'. The piles will be embedded 2 feet into the
abutment cap. The eastbound and westbound abutments will be offset from each
other. A soldier pile wall with concrete lagging will be constructed between the
two abutments. The wall will also serve as part of the shoring system required for
half-width construction. Details for the wall will allow for separate movement of
each of the integral abutments (eastbound and westbound). The longitudinal edge
of the approach slab will be supported by the wall.

Each wingwall will be 24 feet long and will be supported on H-pile footings.
A strip seal will be placed between the abutment and wingwall in order to allow
for the longitudinal thermal movement. Rock slope protection will be placed at
each end of the bridge in front of the abutments.

PIERS 1 AND 2

Alternate 1

For the skewed option (Alternate 1), each pier location will have one
continuous pier supporting the eastbound and westbound roadways. A drilled
shaft foundation and a footing supported by piles or multiple caissons were
considered for the piers. The drilled shaft foundation consists of four drilled
shafts, which will be drilled through riverbed and claystone to the siltstone. The
use of a web wall to connect the drilled shafts was considered, but would require a
cofferdam to facilitate placement below the water level. This would be very
expensive and, therefore, the use of a web wall was eliminated. Rock sockets will
be drilled into the siltstone layer.

The footings supported by caissons/piles will consist of a four-column bent
supporting the pier cap. The top of the footing will be placed at the scour depth,
15 feet below the riverbed. The caissons will be drilled through the riverbed to
the siltstone layer and a rock socket will then be drilled into the siltstone layer.
According to the geotechnical studies, the claystone layer is unsuitable for
supporting the pile foundations and is susceptible to scour. The piles would be
predrilled and driven through the claystone layer to refusal in the siltstone to an
approximate elevation of 514'.

Because of the depth to competent rock, spread footings are not considered a
viable option for the pier foundations. The claystone strata are considered
susceptible to scour; therefore, the pier footing should extend to the siltstone
layer. The siltstone is approximately 40 feet below the riverbed and thus would
require a very large cofferdam and a large amount of claystone excavation. This
option would be extremely expensive.

106

Figure 3
Alternate 1

107
Alternate 2

Piers for the radial option will be hammerhead piers supported by a single,
large-diameter drilled shaft. The drilled shaft will be drilled through the riverbed
and claystone to the siltstone. A rock socket will be provided into the siltstone.
Since the superstructure consists of two separate structures, one for the eastbound
and one for the westbound, a total of four piers will be constructed. Due to the
possibility of scour, the drilled shafts will be designed to be stable and structurally
sufficient without soil support down to the total scour depth.

Piers supported on a footing with piles or multiple caissons were not
considered for the radial option. Due to the scour depth, the top of footing would
be 15' below the riverbed. In order to construct the footing, it would be necessary
to build a cofferdam and excavate claystone. Due to the cost of the cofferdam and
the claystone excavation, this option was ruled out.

108

Figure 4
Alternate 2

109

Selection of Alternate

The West Virginia Division of Highways selected Alternate 2 to advance to
final design. The primary reasons for this decision are:

1. The parallel offset abutment configuration provided for more predictable
behavior than the skewed configuration.
2. Separate parallel bridge structures could more easily be phased into the
construction sequencing of the bridge.
3. The pier construction was greatly simplified with the use of a single-
column, drilled-shaft configuration. Disturbance to the riverbed would be
significantly reduced.
4. Interference with existing deep foundations was eliminated by lengthening
the bridge to avoid placing the new offset abutments on the existing
abutments and piling.

FINAL DESIGN

Alternate 2 as stated previously was advanced to final design. In order to
accommodate the projected 3-inch movement and provide for a jointless bridge,
special details are required and are discussed below.

Final Superstructure Design and Details

In order to provide for an integral abutment, the steel girders and the concrete
abutment must be joined together. The West Virginia Division of Highways
Bridge Design Manual provides guidance on accomplishing this.

Figure 5 shows the final integral abutment detail for the Amandaville of Coal
River Bridge. The key features of this detail are:

110

Figure 5
Coal River Integral Abutment Detail

1. A single row of pre-drilled and driven HP 12x84 piles was chosen to
handle the longitudinal movement of the bridge. The piles are placed so
that the weak axis of the pile is in the direction of the major movement of
the bridge to facilitate the movements anticipated.
2. The approach slab is placed on a haunch on the backside of the abutment
and joined to it with a series of hooked reinforcement bars. This allows
the approach slab to move in unison with the abutment.
3. The steel girders were placed on a 1-1/8-inch-thick elastomeric bearing
pad.
4. After the steel girders were set, a 3/4-inch erection stud was drilled and
placed to anchor the girder into position prior to final concrete placement.

The drilled shaft foundations at the four pier locations required shafts that
were 7'-0" in diameter and were an average of 50 feet deep. Each shaft had a
6'-6"-diameter rock socket that was 11'-0" deep. The pier column and
hammerhead cap project out of the water an average of 22'. The steel girders are

111
fixed at the piers and the bearings specified were 4-5/8"-thick laminated
elastomeric bearing pads.

Some interesting details were required off the bridge structure proper that
interfere with the integral abutments and approach slabs. Where the approach
slab interfaces with the roadway pavement, shoulder and concrete safety barrier,
several special details had to be developed.

Figure 6
Concrete Barrier Sliding Plate Expansion Dam

1. A strip seal expansion dam with a 3" opening was placed on the
sidewalks. The calculated movements of the bridge were 1.3 at 120
degrees and 2.4 at 30-degrees, so a 3 opening was provided. A steel
sliding plate was placed over the strip seal to prevent a pedestrian tripping
hazard. The design team was concerned that degradation of the sidewalk
pavement at the end of the approach slab would lead to a constant
maintenance problem and a tripping hazard. This led to the inclusion of
the strip seal expansion dam.
2. The strip seal expansion dam was carried into the adjacent concrete safety
barrier. A 3" opening was provided. A steel sliding hood plate in the
shape of the concrete safety barrier was placed over the strip seal. This

112
was done to isolate the bridge movement and to protect the adjacent
roadway barrier and crash impact attenuator.

3. In accordance with WVDOH, a joint was placed where the approach slab
interfaces with the existing bituminous concrete pavement. In time, it is
expected that the movement of the bridge will begin to degrade the
bituminous concrete pavement.

Another interesting special detail involves allowing for movement in the
integral abutment stem where it intersects the longitudinal closure wall; a 3-1/2"
opening was specified. The opening is between the rear face of the abutment
stem and the beginning of the longitudinal closure wall. This was done at both
abutment locations.

CONCLUSION

In conclusion, the challenge that was presented by the WVDOH to design a
630-foot-long jointless bridge was met successfully. Guidelines provided by the
WVDOH and also their experience and willingness to push the envelope were
crucial in putting together a successful project.

The project went under construction in January 2004, and in the first week of
January 2005 the westbound lanes of the new bridge was opened to traffic.
Currently, the eastbound lanes are under construction and the project is on
schedule for a December 2005 opening.
Figure 7
Approach Slab and Pavement Joint Details

113
Integral Abutment Bridges with FRP Decks Case
Studies

Vimala Shekar
Srinivas Aluri
Dr. Hota V.S. GangaRao

Constructed Facilities Center, West Virginia University
Morgantown, WV 26506-6103

ABSTRACT

To allow free expansion and contraction between superstructure and
abutments, the traditional construction method has incorporated joints and
bearings. But during in-service life of bridges, these joints and bearings become
potential places for accumulation of debris and deicing chemicals, thereby
weakening concrete and corroding steel stringers leading to high life cycle cost
including maintenance cost. As a way to reduce initial and maintenance cost,
engineers recommend building bridges without joints. Hence, transportation
departments of various states in U.S. have been building integral abutment
bridges since 1960s. Over the years, these jointless bridges have proven to be
successful and have shown good performance. The Constructed Facilities Center
at West Virginia University (CFC-WVU) has expertise in designing integral
abutment bridges with traditional concrete decks and also fiber reinforced
polymer (FRP) composite decks. In this paper, the in-service performance
evaluations of two integral abutments bridges (i.e., Market Street and Laurel Lick)
with FRP decks have been highlighted. In addition, the behavior of an integral
abutment bridge with FRP composite deck is correlated with the behavior of a
jointless concrete deck bridge.

INTRODUCTION

Over the years, bridge maintenance cost has been a growing problem in
highway bridges. Hence, engineers have recommended constructing bridges
without joints, reducing the initial and maintenance costs. Construction of
jointless bridges is simpler and faster than construction of bridges with joints,
because they require fewer parts, less material and are less labor intensive [1].
The transportation departments of various states in U.S. have been building
integral abutment bridges since 1960s and are also using jointless bridges for
replacement of deteriorated structures. These jointless bridges have proven to be
more efficient [2].

The Constructed Facilities Center at West Virginia University (CFC-WVU)
has done several research and development studies on the behavior of jointless

114
bridges and has developed specifications including standard design procedures for
integral abutment bridges. CFC-WVU not only has expertise in designing
jointless bridges with traditional concrete decks, but also has built integral
abutment bridges with lightweight modular decks using fiber reinforced polymer
(FRP) composite materials [3]. CFC-WVU has been in the forefront of
developing new and cost-effective glass FRP structural shapes for bridge
applications. In the state of West Virginia alone, there are twenty-three (23)
bridges that have been rehabilitated using FRP composite materials and two of
them have integral abutments. In this paper, the in-service performance
evaluations of two integral abutments bridges (i.e, Market Street and Laurel Lick)
with FRP decks are summarized, as part of the on-going monitoring work of
CFC-WVU with the West Virginia Department Transportation Division of
Highways (WVDOT-DOH). In addition, a behavioral comparison of an integral
abutment bridge with FRP composite deck and concrete deck is discussed.

DESCRIPTION OF BRIDGES

The two integral abutment bridges discussed in this paper were built using
FRP composite bridge decks, designed by CFC-WVU and manufactured by
Creative Pultrusion, Inc. under the trade name of Superdeck
TM
. The deck cross-
section consists of double-trapezoid and hexagonal shapes as shown in Figure 1.
The FRP bridge deck component is made of E-glass fiber with vinylester resin.
The fiber architecture of the deck consists of several fabric layers, mats and
rovings with a fiber volume fraction of about 40 ~ 45%.

Figure 1 Cross-Section of SuperDeck
TM

MARKET STREET BRIDGE

The Market Street Bridge, originally was a two span concrete filled steel grid
deck stiffened with steel stringers supported on steel girders with sidewalks on
each side [4]. The overall length of this three-lane bridge was 165 long and 59
wide. It had two full heights cut stone abutments and one solid concrete pier. The
bridge had an average daily traffic of 6900 and was designed to resist HS-15
loading [5]. The deck and girders were in poor conditions and the pier was
showing some signs of distress, however the abutments remained in good
condition [4].

115

The new bridge (opened to traffic in 2001) is a single span jointless bridge
with a center-to-center bearing length of 177 and deck width of 56. The bridge
is constructed with a fiber reinforced polymer (FRP) deck supported by seven (7)
steel girders having a center-to-center spacing of 8 feet and 6 inches. The deck
was connected to girders by means of shear studs and grouted with concrete. A
polymer concrete overlay of 3/8 thick was placed on the FRP deck. The bridge
has an average daily traffic of 6900-10,000 and is designed for HS-25 type of
loading. The steel girders are embedded in concrete pile caps. A detail of the
integral abutment superstructure is shown in Figure 2, and was designed by Alpha
Associates, Morgantown, WV.

Figure 2. Details of Intergral Abutment in Market Street Bridge

Laurel Lick Bridge

The Laurel Lick Bridge is a single span jointless bridge located in Lewis
County, West Virginia. The center-to-center bearing length of the bridge is 19
with a deck width of 16. The bridge was constructed with a fiber reinforced
polymer (FRP) deck supported by six (6) WF glass fiber reinforced polymer
(GFRP) stringers having a center-to-center spacing of 2 6 and was opened to
traffic in May 1997. The FRP deck panels were joined in the field using shear
keys (full depth hexagonal component) to provide mechanical interlocking in
addition to adhesive bonding. The deck was connected to the stringers by means
of mechanical fasteners (blind bolt) and adhesive bonding (Pliogrip) as shown in

116
Figure 3. A polymer concrete overlay of 12.7 mm (0.5) thick was placed on the
FRP deck. The bridge has an average daily traffic of 100 and is designed for
AASHTO HS-25 loading.

Figure 3. Connection of FRP Deck to Stringer on Laurel Lick Bridge
INSTRUMENTATION PLAN

In order to measure the deck and stringer strains under a static load, the
bridges were instrumented with several electrical resistance strain gages. A ruler
and theodolite were used to measure deflections on the Laurel Lick Bridge. Due
to accessibility problems, deflections were not measured on the Market Street
Bridge. To evaluate the dynamic response of the bridge, piezoelectric
accelerometers were mounted on the bridges. Data from strain gages and
accelerometers were acquired using a System 5100 scanner from Vishay Micro-
Measurements controlled through a laptop computer. The scanner acquires data
from fifteen strain sensors and five LVDTs/accelerometers with 16-bit A/D
resolution. The maximum sampling rate of 50 scans per second (50 Hz) per
channel was used to acquire all the dynamic test data. Typically, the first natural
frequency of these bridges (an important consideration in DLA and damping
calculations) is less than 10 Hz. Therefore, a 50 Hz sampling rate is considered
adequate to prevent any aliasing.

TEST RESULTS AND DISCUSSIONS

Static Response of Two Bridges

Three static load tests were performed on the Market Street Bridge and five
load tests were performed on the Laurel Lick Bridge. The third load test data of
the Market Street Bridge was discarded due to corruption of the data. The first

117
and second tests of the Market Street Bridge were conducted on May 2003 and
February 2004, respectively. Each of the two load tests on Market Street bridge
included three load positions: 1) Two trucks positioned in two lanes, one truck in
the left lane and another in the center lane (Load Case 1); 2) Two trucks
positioned in two lanes one truck in the center lane and another in the right lane
(Load Case 2); and 3) Two trucks positioned back to back in the center lane (Load
Case 3). In the first two load cases, the center of gravity of the truck axles was
positioned at the center of bridge span to induce maximum bending moment in
bridge superstructure. A typical load pattern for the Market Street Bridge is
shown in Figure 4.

Figure 4. Two Trucks positioned in Two Lanes (Load Case 1), One Each on Lane 1 and
Lane 2 on Market Street Bridge

In the case of the Laurel Lick Bridge, a total of five load tests were
conducted. The first load test was conducted on September 1998, the second load
test on March 2003 and the third load test on January 2001. The fourth and fifth
loads tests were conducted on April 2003 and May 2004, respectively. Each of the
first two load tests (i.e., in 1998 and 1999) included four load positions: 1) The
center of the rear axle of the loaded truck was positioned at the center of bridge
(Load Case 1); 2) The center of the rear axle of the loaded truck was positioned
towards the edge of bridge at mid-span (Load Case 2); 3) The center of the rear
axle of the loaded truck was positioned on the deck module joint (Load Case 3)
and 4) The rear axle of the loaded truck was positioned at the bridge end i.e., on
the abutment (Load Case 4). Since the first two load tests results revealed that
first load position (i.e., Load Case 1) was most critical among the four load
positions, the remaining three load tests (i.e., the tests conducted during 2001,
2003 and 2004) were performed only for Load Case 1 position. A typical load
pattern on the Laurel Lick Bridge is shown in Figure 5.

118

Figure 5. Center of Rear Axles of Truck Positioned in Center of Bridge Load Case 1 on
Laurel Lick Bridge

During the static load tests, strains and deflections on the stringers and deck
were recorded before and during the load application on both bridges. The strain
and deflection data were used to evaluate:

Degree of structural composite action between the deck and supporting
stringers.
Transverse load distribution factors for computing the design moments for
the stringers.
Deck stresses and deflections.

DEGREE OF STRUCTURAL COMPOSITENESS

The degree of structural compositeness is defined as a ratio of in-plane
deformation or strain to bending at the deck-stringer [6]. In the field, strains
measured on the bottom of the FRP deck and at the bottom of the top and bottom
flanges of stringers are used to compute the degree of structural compositeness
between the FRP deck and stringers. For the Market Street Bridge, the deck is
connected to the stringers through shear studs and the average degree of
compositeness was found to be about 100%. The composite moment of inertia of
the FRP deck with steel plate girders for the Market Street Bridge is only 7%
higher than the steel plate girder moment of inertia. This is because the FRP
modular deck panels were pultruded with hollow-core trapezoidal sections and do
not contribute significantly to the composite moment of inertia. But, in a
conventional pre-cracked concrete deck, the composite moment of inertia is about
250% of the stringer moment of inertia and the concrete slab and steel stringer are
assumed to act as a unit with the top of the slab in compression and top of the
steel plate girder in compression, with no slippage in between the slab and
stringer. Similarly, in the case of Laurel Lick Bridge, the FRP deck is connected

119
to the GFRP stringers by mechanical fasteners and adhesive bonding, and the
composite action for this bridge was also close to 100%.

TRANSVERSE LOAD DISTRIBUTION FACTORS (TLDF)

TLDFs on bridges are computed by dividing maximum measured strain in
the stringers (at mid-span) by the summation of peak strains in all of the stringers.
On the Market Street Bridge, since the center of the bridge is crowned, the bridge
deck behaves as a hinge at the center width of the deck. Therefore the TLDF is
calculated assuming that out of seven (7) stringers, about 50% of stringers (i.e.,
four stringers) are always effective in contributing towards the TLDF. Hence for
the Market Street Bridge the maximum transverse load distribution factor was
found to be 0.78 for Load Cases 1 and 2. The TLDF of 0.78 on the Market Street
Bridge translates to S/5.4, which compares favorably with the current AASHTO
(1998) equation of S/5.5 (for two or more traffic lanes in concrete deck bridges
stiffened with steel stringers). Similarly for the Laurel Lick Bridge, the maximum
TLDF was found to be 0.22 for Load Case 1, which translates to S/5.7, which
compares well with the AASHTO (1998) equation of S/6.0 (for single lane glue
laminated panels supported with glued laminated stringers).

DECK STRESSES AND DEFLECTIONS

Strains obtained from static load tests are used to evaluate the performance of
the Market Street and Laurel Lick bridges. For the Market Street Bridge, the
maximum measured global strain (for a truckload equivalent to the AASHTO HS-
25 loading) was 59 microstrains, which translates to a global stringer stress of
about 1.7 ksi, while the allowable stress in steel stringers is 27.5 ksi. In the case of
the Laurel Lick Bridge, the maximum strain in the GFRP stringer was 429
microstrain (for a truckload equivalent to the AASHTO HS-25 loading), which
translates to a global stringer stress of 1.9 ksi. (assuming an E value of the FRP
stringer as 4.59 x 10
6
psi). The stress for the stringer is within the allowable limits
of GFRP stringers (2 ksi).
The maximum measured bottom deck strain in the Market Street Bridge (for a
truckload equivalent to the HS-25 loading) was 52 microstrains, which translates
to a deck stress of about 158 psi. The deck stress is well within the allowable
limits of 2440 psi. (Note: The allowable strain of the FRP composite deck is
assumed to about 20% of the ultimate strain i.e, 4000 microstrains and a modulus
of elasticity of 3.05 x 10
6
psi). In the case of the Laurel Lick Bridge, the
maximum deck strain was 44 microstrains, which is translated to a deck stress of
134 psi that is well within the allowable deck stress.

In both bridges (Market Street and Laurel Lick), although induced deck
stresses are significantly low compared to the allowable limits, designs of
composite bridges are mostly driven by deflection criteria. Typically the

120
allowable deck deflection relative to the stringers on composite bridges is S/500,
S being center-to-center of stringers [7]. Due to accessibility problems, the deck
deflections on the Market Street Bridge were not measured.

The maximum measured global deflection for a truck load equivalent to the
AASHTO HS-25 loading on the Laurel Lick bridge was found to be 0.223 in,
which is well within the AASHTO limit of S/500 (i.e., 0.44 in).

Dynamic Response of the Two Bridges

FRP bridge decks are lightweight (~ 25 - 11 lb/sq.ft) and have lower stiffness
than steel. Therefore bridges with FRP decks could have higher amplitudes of
vibration due to moving traffic than bridges with conventional decks, potentially
causing cracks on the wearing surface. Further vibrations could be perceived by
users that a bridge structure may be unsafe in spite of the actual structural
integrity.

Field dynamic tests were conducted on Laurel Lick and Market Street bridges
to evaluate the Dynamic Load Allowance (DLA) factors [8]. Field test results
from the Market Street Bridge indicate that the DLA factors are well within the
1998 AASHTO LRFD [9] Bridge Design Specification of 33%. Results from the
Laurel Lick Bridge show that the DLA could be as high as 93% [8], however it
should be noted that the deck and stringer strains are well within the ultimate
failure strains.

Deck accelerations were also evaluated for the Market Street Bridge and
checked against the pedestrian bridge vibration serviceability criterion from 1983
Ontario Highway Bridge Design Code (OHBDC) [10]. For several test cases the
deck accelerations were found to be over the 1983 OHBDC limit states. In
addition, during the testing, it was found that the deck vibration due to moving
trucks was clearly perceptible to pedestrians. However, further research is
underway to mitigate these high deck vibrations using different damping schemes.

INTEGRAL ABUTMENT BRIDGES FRP VS CONCRETE DECKS

In this section, various issues related to integral abutment bridges with FRP
composite decks are correlated with concrete bridge decks. The issues include: 1)
TLDF, 2) effective span length, 3) deck cracking, 4) temperature gradient, 5)
dynamic response and 6) installation time.

On the Market Street Bridge, the TLDF for the interior girder was found to be
S/5.4, which is close to the TLDF of the McKineleyville Bridge (S/6.3) - an
integral abutment bridge with concrete deck [11].

121
Based on the maximum global stringer strain of 59 microstrains on the Market
Street Bridge, effective span length of the girder was found to be about 0.87 times
the span of girder. For a typical single span integral abutment bridge with
concrete deck the effective span length was equal to span of the girder.

With respect to bridge deck cracking, the most notable cracks in the
McKinleyville Bridge were observed near the abutment in the longitudinal
direction. These cracks were located over the stringers, which are attributed to
existence of transverse cracks in the bridges, thus relieving the stress in one
section of concrete deck while increasing the stress in other sections. Moreover, in
a typical concrete integral abutment bridge, the area near the abutment acts like a
continuous beam, with the stringers acting as supports. As cracking further
increases, the section properties of the beam decreases to a point where it cannot
handle the tremendous forces generated by a vehicle when it enters the bridge
[11]. However, in the case of both FRP deck integral abutment bridges, i.e.,
Market Street and Laurel Lick bridges, there was no sign of cracking near the
abutments, as the deck was reinforced with glass fabrics on the abutment and the
FRP is able to resist as much tension as it resists compression.

The temperature gradient on the McKinleyville Bridge was found to be about
10
o
F between the top and bottom of deck. By applying a temperature of 30
o
F at
the deck top and 5
o
F at the deck bottom, the thermal stress was found to be about
162 psi [11], while in the case of the Market Street Bridge, for a temperature
gradient of 26
o
F, the temperature induced stress was about 90 psi. Normally a
temperature gradient of 36
o
F

is not a critical scenario in the field. The critical
temperature has been measured as high as 140
o
F at top and 30
o
F at the bottom of
an FRP deck in which thermal stress will be as high as 1300 psi [12]. Hence a
designer should properly account for thermally induced stress when designing
bridges with FRP decks. It should be noted that though the overall (thermal and
live load stresses) induced stresses are significantly below the allowable limit, the
design of composite bridges is mostly driven by deflection criteria.

The installation time for FRP bridge decks is very short compared to
conventional decks. The erection time of an FRP composite deck is about an
eighth to a tenth of a conventional concrete deck. The construction crew needed
for FRP decks is also smaller to that of conventional decks. It was estimated that a
typical construction crew (without any special training) of six people need one
day to install the FRP deck modules for a bridge of about 500 ft
2,
such as that of
Wickwire Run Bridge [13]. However, on a large job as that of the Market Street
Bridge, 10,800 ft
2
, the deck was installed by a crew of five people working 6
hours a day for 6 days and additional 200 hours were needed to provide glass
fabric reinforcement over the field joints [14].

122

CONCLUSIONS

From the static test results on the Market Street Bridge and the Laurel Lick
Bridge, the following conclusions were drawn:

1) The degree of compositeness between the FRP deck and supporting
stringers in both bridges (Market Street and Laurel Lick) is about
100%. In the case of the Market Street Bridge, the composite moment
of inertia is only 7% higher than the steel stringer moment of inertia.
This is because the FRP modular deck panels were pultruded with
hollow-core trapezoidal sections and therefore do not contribute
significantly to composite moment of inertia.

2) The TLDF for the Market Street Bridge was found to be S/5.4, which
is close to the AASHTO equations of S/5.5 (for two or more traffic
lanes with a concrete deck bridge), which is also close to a typical
integral abutment bridge with a concrete deck. In the case of the
Laurel Lick Bridge, the TLDF was S/5.7, which is close to the
AASHTO equations S/6.0 (for single lane glue laminated bridge
decks).

3) The maximum induced static stresses in FRP decks for both bridges
were well within the allowable stresses of the composite material.
Also, the stringer deflection in the Laurel Lick Bridge was well within
the design limits.

4) Since there was no significant change in the stresses in the deck and
the stringers between each of the load tests, we can conclude that the
overall structural performance of both jointless bridges was good.

5) Dynamic Load Allowance factors for the Market Street Bridge were
within the 1998 AASHTO LRFD Bridge Specifications limits.

6) Unlike jointless bridges with concrete decks, there was no sign of
cracking on the abutment for both integral abutment bridges with FRP
decks.

7) The erection time of an FRP composite deck is about an eighth to a
tenth of a conventional concrete deck.

123

ACKNOWLEDGMENTS

The authors wish to acknowledge the financial support provided by the West
Virginia Department of Transportation (WVDOT) and the Federal Highway
Administration (FHWA) - US Department of Transportation for this study. The
efforts of WVDOT personnel in planning, traffic control and providing trucks for
field tests is greatly appreciated.

REFERENCES

1. Burke, M.P., Jr., 1987, "Bridge Approach Pavements, Integral Bridges, and Cycle-
Control Joints," Transportation Research Record 1113, Washington, D. C., pp. 54-70.

2. Franco J.M., 1999, "Design and Field Testing of Jointless Bridges," Masters Thesis,
West Virginia University, Morgantown, WV 26505.

3. Shekar V., Samer H.Petro., GangaRao H.V.S. 2003, Fiber Reinforced Polymer
Composite Bridges in West Virginia, Journal of the transportation research board,
Eighth International Conference on Low-Volume Roads, Vol 2. No. 1819, pp 378-384.

4. Whipp R., 2001, Constructing the Market Street Bridge, Polymer Composites II 2001.
Applications of Composites in Infrastructure Renewal and Economic Development,
Morgantown, WV.

5. Dave Sada., 2003, Personnel Communication.

6. Lopez, R. A., 1995, Analysis and Design of Orthotropic Plates Stiffened by Laminated
Beams for Bridge Superstructures, Dissertation, Department of Civil and Environmental
Engineering, West Virginia University, Morgantown, WV.

7. GangaRao, H.V.S., Shekar, V., (2002), Specifications for FRP Highway Bridge
Applications: Acceptance Test Specifications for FRP Decks and Superstructures, CFC
Report to USDOT-FHWA, West Virginia University.

8. Aluri S., Chandrashekar Jinka, Hota V.S. GangaRao, 2005, Dynamic Response of Three
Fiber Reinforced Polymer Composite Bridges, Accepted for publication in Journal of
Bridge Engineering,

9. AASHTO, 1998, LRFD Bridge Design Specifications, The American Association of
State Highway Transportation Officials, Washington, D. C.

10. Ontario Ministry of Transportation and Communications (OMTC) (1983). 1983 Ontario
Highway Bridge Design Code. Second Edition.

11. Franco M.J., 1999, Design and Field Testing of Jointless Bridges, Masters Thesis,
Constructed Facilities Center, West Virginia University, Morgantown, WV.

12. Krit L., 2004, Theoretical and Experimental Analysis of FRP Bridge Decks under
Thermal Loads , Ph.D Dissertation, College of Engineering and Mineral Resources,
West Virginia University, Morgantown, WV.

124

13. Roberto Lopez-Anido, Dustin L. Troutman and John P. Busel, 1998, Fabrication and
Installation of Modular FRP Composite Bridge Deck, Proceedings of International
Composites Expo 98, Nashville, Tennessee, pp. 4-A/1 to 4-A/6

14. Hota V.S. GangaRao, Krit Laosiriphong. Design and Construction of Market Street
Bridge, Society For The Advancement of Material and Process Engineering,46
th
International SAMPE Symposium and Exhibition, Long Beach, California, 2001, pp
1321- 1330.

125
New Mexicos Practice and Experience in Using
Continuous Spans for Jointless Bridges

Steven Maberry; Jimmy D. Camp, State Bridge Engineer; Joan Bowser: NMDOT

ABSTRACT

This paper presents a cursory review of jointless bridge development in New
Mexico. After briefly outlining some key settings in New Mexico, it classifies
New Mexico jointless bridges into five recognizable types. Then the paper
provides a brief review that uses a few project examples to illustrate problems
faced in New Mexico jointless bridges and how some major details developed in
different bridge types, to address these issues. The subject then turns to consider
opportunities for quality improvement. The final paper account is a brief
discussion of methods that might be used to address these problem opportunities.

NEW MEXICO BRIDGE CONSTRUCTION ENVIRONMENT

The New Mexico Department of Transportation (NMDOT) faces the same
jointed bridge problems that other transportation departments face. The problems
include all the oft-expressed issues of water-borne damage. Concrete on pier caps
and on girder beam-ends suffer from joint water intrusion. Bearings deteriorate
and seize. Not so often expressed, but noted in New Mexico and no doubt
experienced by others, is the infrequent, but very dangerous road hazard created
when a fabricated bridge joint separates from its concrete embedment and
becomes perilous road junk protruding out of the deck or a projectile hurdling
toward high-speed traffic.

New Mexico also exhibits a few problems that are not, perhaps, shared to the
same extent by some states. Aggregate Alkali Silica Reactivity (ASR) is some of
the worst in the nation. New Mexico has instituted ASR controlling policies.
These policies include testing and establishing approved aggregate and water
sources. ASR-controlling additives like Class F fly ash, blended cement, ground
granulated blast furnace slag, silica fume and lithium nitrate have also helped.
Also, some admixtures are prohibited or discouraged. Older bridges, however,
suffer from uncontrolled ASR.

Further, New Mexico topography ranges between 2,817 and 13,161 feet above
sea level. The state contains mesas, mountain ranges, canyons, valleys and
arroyos. The climate is mild arid to semiarid. This climate, coupled with high-
altitudes, leads to large daily and annual temperature swings. The average daily

126
temperature range, low to high, is 25 to 35 F [1]. Such temperature swings are
hard on concrete. Not only do they cause repeated temperature-induced
movement, but they also lead to frequent winter freeze-thaw cycles. Movement
coupled with delayed maintenanceleads to joint failure. Subsequently, failed
joints introduce water onto substructures. Frequent freeze-thaw cycles then
exacerbate damage, and ASR amplifies deterioration, especially on older bridges
where ASR was less well controlled.

One of New Mexicos greatest assets, its dramatic geology, presents further
complications. Subterranean geological structures that support bridges and bridge
approaches vary widely: from caliches to weak clay, from buried gravel beds to
basalt, from rounded, weak sands to angular, well-graded grains. Such wide
variations prevent establishing a standard foundation design approach. Therefore,
New Mexico must retain options for designing with diverse foundation types.

THE JOINTLESS SOLUTION IN NEW MEXICO

In contrast to other states jointless bridge designs, New Mexico has pursued
retaining pier and abutment bridge bearings. Not pursuing integral abutments with
jointless bridge decks may seem odd. However, with its varied geology, New
Mexico frequently needs more foundation optionsincluding spread footings,
pile groups too stiff for flexible integral abutments and drilled shafts. Therefore,
New Mexico has pursued semi-integral abutment types and retained bearings over
pier caps.

Jointless Bridge Types
One can identify five bridge types that frequently either occurred in the past or
made some debuts in New Mexicos bridge inventory types. These five types are:

Pre-stressed concrete that is simple for dead construction loads and
continuous for live loads (predominant type for new construction).
Pre-stressed concrete that is simple span for the girder beams and
continuous for the deck (predominant type for rehabilitation).
Steel bridge continuous for dead and for live load.
Steel bridge that is simple for dead construction loads and continuous for
live loads.
Pre-stressed concrete that is continuous span for dead and live load
achieved inadvertently through incorrect construction placement.

127
PRE-STRESSED CONCRETE,
SIMPLE FOR DEAD,
CONTINUOUS FOR LIVE
This bridge type uses pre-cast,
pre-stressed concrete beams. They
are first set on bearings to span from
pier to pier, simply supported. Then,
using the detail sketched in Figure 1,
the deck and diaphragms are poured
simultaneously.

As deck concrete loads the
beams during construction, the
beams rotate on their bearings as
simply supported. Once the
diaphragm cures, though, it provides
a compressive base at the bottom of the (already dead-load rotated) beams. The
cured deck and beam composite provides paths for negative moment tension
stress through deck reinforcing steel and girder harped strands. When there is a
large deck to pour, the pour might be staged. In staged construction, the mid-span
deck sections are poured first so that the beam still rotates under the dead load.
Then, the deck above the pier support and diaphragm for that pier are poured
simultaneously.

New Mexico now builds most new bridges as this bridge type. It results in a
jointless bridge, but typically the pre-stressed beams are still placed on bearings.
Abutments are semi-integral and piers do not absorb most of the bridge
movementthe bearings do.

PRE-STRESSED CONCRETE, SIMPLE SPAN GIRDERS, CONTINUOUS
DECK
New Mexico bridge rehabilitations have evolved to where the design
maintains the girder beams as simple spans, but the decks become continuous.
Early rehabilitations did not successfully use this simply supported beams
approach and resulted in unexpected consequences. NMDOT Bridge Section has
developed details that now address some of those unexpected consequences,
resulting in more satisfactory diminished maintenance and lengthened bridge life.

In this type, the original bridge deckjointed over every pier capis
removed, either over each pier or in its entirety. Design for the new deck achieves
a continuous deck, but leaves the girder beams still acting predominately as
simple spans. To address bridge behavior changes and possible bearing
degradation, bearings are usually changed to allow for added thermal movement
from the conversion and to correct existing deterioration.

Figure 1 Continuous for Live Load Detail

128

Figure 2 illustrates details responsible for keeping the span girders simple
under a continuous deck. The expanded polystyrene embedded in the diaphragms
over each pier forms a break that interrupts continuity between girder beams and
allows the beam ends to rotate, just as it did prior to making the bridge deck
continuous. With this design, the deck remains subject to tension from negative
moments and cracks readily over the piers. The rigid foam encased and protected
by the steel cap shown here, spreads cracks over a greater distance. The foam
allows unrestricted beam-end rotation and the steel protects the foam during
construction. Without this detail, cracks in the deck tend to concentrate into a
single, wider crack over the pier-cap centerline. With it, cracks tend to spread
over the longitudinal length bounded by the steel capstypically resulting in
three narrower cracks instead of a single, wide crack. In some cases, multiple
cracks appear rather than three distinct ones. With the tension cracking relieved in
more than one crack, the width of each crack is substantially reduceda more
desirable outcome. These caps also relieve the very ends of the beam top flanges
from fully participating in resisting the negative moment (and consequential
tension).

Another detail to note in Figure 2 is the saw cut in the bridge deckthis can
be a single cut, as shown here, or multiple cuts. These expansion control cuts are
then filled with High Molecular Weight Methacrylate (HMWM) or other sealing
compounds (e.g. low-viscosity, low surface tension polymers). New Mexico is
funding research that might establish the time frame, in our climate and
conditions, when crack formation stabilizes and thus suggest when to seal these
cracks.

STEEL BRIDGE CONTINUOUS FOR DEAD AND LIVE LOADS
Although pre-stressed concrete beams dominate bridge design in New
Mexico, there are some standard continuous steel bridge examples in New
Mexico. The beam girders on these bridges are spliced to be continuous from
Figure 2 Continuous Decks Only

129
initial placement; hence, the steel girders support both deck and the imposed live
loads as continuous beams.

STEEL BRIDGE, SIMPLE FOR DEAD, CONTINUOUS FOR LIVE LOAD
Recently, New Mexico completed bridges that forego the material and labor
intensive bolted splices in continuous steel girders. Instead, the steel beam-ends
rest on each pier cap (much like pre-stressed concrete) and merely include a
continuity strap. Figure 3 illustrates the details for this jointless bridge type. This
approach to steel bridge design has been outlined in recent reports [2]. Like
concrete beams, the steel beam carries construction deck dead loads as simple
span. Once the diaphragm between bridge-ends cures, it provides a compressive
base to transfer loads across the diaphragm. The composite action and the
continuity strap transfer negative-moment tension to the deck reinforcement and
the top beam flange. The end result is an approach that behaves similar to simple
for dead and continuous for live load pre-stressed concrete beams. There is also
construction simplification through eliminating the more complicated field joint.
To this particular jointless approach, maintaining one web thickness throughout
and avoiding flange width or thickness transitions add beam fabrication
simplification. While using slightly more
material and having slightly more weight,
these consistencies in the web and flange,
coupled with no mid-span joints (using only
the continuity strap shown in Figure 3)
produce a bridge superstructure with fewer
complicationstranslating to fewer
mistakes, less costly fabrication and a robust
design.

CONCRETE CONTINUOUS FOR DEAD
AND LIVE LOAD
In New Mexicos inventory of jointless
bridges, there are, unfortunately, some
bridge examples that have the distinction of
being pre-stressed concrete continuous for
both construction dead load and for live
load. These bridges were not designed this way. Instead, they were the result of
construction sequence field changes.

The notion incorporated in pre-stressed concrete beams, continuous for live
load only (not dead load and live load), is that the deck and diaphragms be poured
simultaneously (as previously described). This way, when the two opposing girder
beams experience loading from the deck concrete pour, their ends freely rotate in
opposite directions as they deflect at the mid-span in response to the added load
(in contrast, abutments do not have opposing beam rotations, and the diaphragm
Figure 3 Steel Continuous for Live Load

130
construction staging is not so critical). Then, when both the deck and the
diaphragm have set, the beam and deck respond in concert to additional loading
from live load.

At times, the builder, unaware of the reasons for simultaneous deck and
diaphragm pouring, completed the diaphragms prior to loading the beams with
deck concrete. The girder beam-ends, then, became set in the cured diaphragms,
unable to rotate with the subsequent deck dead load. NMDOT is not without some
contributory responsibility in this. When asked if this construction procedure
would be allowed, NMDOT representatives in the field have consented to the
procedure change. The query never reached the design engineer. The result is a
bridge inadvertently made continuous for both dead load and live load.
HISTORY OF JOINTLESS BRIDGE PROBLEM RESOLUTIONS

A few of the unexpected results experienced in New Mexicos jointless
bridges have been hinted at in this paper so far. While eliminating joints in the
bridge deck has proven to be a boon to bridge life and maintainability, there have
been problemsmany of them unexpected. This section extracts a select few
from New Mexicos experience and provides their resolution.

Interstate-25 over Rio Grande South of Albuquerque
Many of the more interesting problems were concentrated in rehabilitations.
An early rehabilitation was the conversion of a multiple simple-spans bridge with
expansion joints over each pile bent. The original design used simple spans, deck
and pre-stressed concrete beams between each pile bent. The conversion reduced
the number of bridge joints, but did not eliminate them entirely. For this bridge, it
was thought that the total span was too great to convert entirely to a jointless
bridgeespecially in light of plans to retain the original bearings, which were not
designed for the potential movement in a fully jointless total span.

In converting from a jointed, standard simple span, the design recognized
potential problems associated with shear changes induced by full continuous
bridge conversion. Therefore, instead of converting both beams and deck to
continuous, the bridge deck was made continuous, but the girder beams remained
simple. At each joint to be eliminated, the deck was demolished back several feet
and separated from the diaphragm underneath. The tops of existing girder beams
were covered with foam board for three feet to either side of the pier cap
centerline. This foam served to break the deck-to-beam bond in this highest
negative moment region and to absorb some differential shear. The deck was then
re-poured in the space resulting from the demolition around the joint, eliminating
the joint.

In the immediate months after completion, two realities surfaced. The first
was the pile bent stiffness. Sixteen round piles filled with concrete support each

131
pile bent. With five beam lines, that is just over three piles per beam. The second
reality was that the existing bearings were not replaced or fully refurbished. The
bridge deck, now continuous for four spans, instead of a joint on every span,
proceeded to move. The old bearings failed to adequately accommodate for this
four-fold movement, and, instead, transferred much of this movement into the pier
cap. The substructure stiffness resisted any movement, and the concrete pier caps
suffered the consequences. Damage ensued to the pier caps and large chunks split
and cracked. Pier cap repairs and bearing modifications have since addressed this
problem. The bridge functions now with the reduced maintenance originally
planned.

Lessons learned: When converting from jointed to jointless, bearings require
thorough consideration. They may need complete replacement to allow the new,
greater bridge movement. Bearings may also need replacement just to restore
original function after years of deterioration. Further, had the substructure been
less stiff, it could have deflected and accommodated some of the movement.

The Approach Slab Saga
In jointless bridge designs first years, New Mexico included an approach
slab. These early approach slabs were unsupported at the off-bridge ends. In time,
sub-grade pumpingwhere water mixes with the soil and traffic movement
causes the slab to act as a pump and move the soil/water out from under the
slabcaused degradation of slab support and subsequent slab settling and
damage.

The first reaction was to remove the approach slab in its entirety and extend
the pavement to the bridge end with a joint installed at the transition from
pavement to bridge. The consequences of this action were about water (again) and
differential settlement. First, when the joint inevitably degraded, runoff dumped
water directly onto the abutment and bridge end. Second, even very minor
settlement of the approach grade resulted in a notable, and uncomfortable, bump
over the lip formed by the different heights between approach pavement and
bridge end.

New Mexicos present response to the approach problem has been a return to
using approach slabs. However, a sleeper now supports the slab at the off-bridge
endresulting in bridging the voids when pumping removes soil from beneath a
slab.

Lessons learned: Although sub-grade underneath an approach slab is prone to
settlement and erosion, an approach slab accomplishes two things. It provides
vehicles with a smoother transition from pavement to bridge when the approaches
and bridge, as they often do, experience differential settlement. It also moves the
end expansion jointand all its associated movement and maintenance
problemsaway from the abutment and bridge end. The sleeper support has

132
reduced the approach slab settlement and associated problems, but poor
compaction or deep sub-grade consolidation still often lead to sleeper settlement.

Interstate-10 Over Rio Grande in Las Cruces
In 2002, New Mexico undertook a major rehabilitation on the Interstate-10
over Rio Grande Bridge. The project included converting the regularly jointed
bridge into a jointless bridge deck with the pre-stressed concrete beams remaining
simple spansas NMDOT had learned to approach existing bridge
rehabilitations. The original bridge construction was completed in 1971. The
bridge was actually two bridges, one east bound, and one west bound. Each bridge
consisted of twelve spans with seven beam lines per bridge.

Early into the project, it became clear that manyalmost allof the beam
ends required repair. NMDOT had anticipated some beam-end repairs, but not the
number involved. With eleven piers, seven beam lines, and two separate bridges,
this translated into over 300 beam-end repairs. Such a large number of individual
beam end repairs threatened to exceed the estimated cost and available funds.
The designers set to seeking a solution to this cost overrun. Inspired by the
now typical approach to new bridge construction, they noted that encasing beam-
ends within a diaphragm over the piers, similar to that used in new construction,
would eliminate any need to repair beam-endsencasement would become the
all-inclusive end repairs needed. The construction bid included a unit amount for
substructure concreteas well as a per beam-end repair amountand it was clear
that encasing the beam-ends would result in lowering costs to within available
funds. All that was needed was an approach that would break the way such a
diaphragm contributed to converting the concrete beams into continuous beams.
The solution was the detail previously shown in Figure 2. This approach, arising
from the beam-end repair problem, has now become the most frequent way
NMDOT addresses bridge deck rehabilitations that convert existing simple span
bridges into jointless bridges.

Lessons learned: Bridge beam-end repairs can be a labor-intensive, costly
item that demolishes original cost estimates. Using an interrupted diaphragm
approach like that shown in Figure 2 provides an affordable, universal beam-end
repair coupled to a good jointless conversion.

JOINTLESS BRIDGE IMPROVEMENT OPPORTUNITIES

A jointless approach to bridge design remains attractive. It is unlikely that
anyone will develop an ideal bridge expansion joint. Eliminating the joint and
allowing the bridge deck to roof the substructure support elements remains a
superior approach. Several areas, however, remain opportunities for improvement
around the jointless philosophy. Several of these are briefly discussed here.

133

Negative Moment Cracking
Continuous decks, or continuous deck/beam composites, naturally introduce
negative moment over supports. Since concrete responds to any substantial
tension by developing a crack, concrete decks over piers typically develop distinct
cracking in this negative moment region. Neither the steel cap/foam board shear
bond disconnect nor the saw cut with filler actually solves cracking. At this time,
pre-stressing the deck with post-tensioning is the only known sure cure. However,
post-tensioning is rarely selected for this alone due to disproportionate expense.
Further, the ideal 5500-psi minimum strength for post tensioning cast-in-place
concrete has been difficult to achieve in New Mexico. Opportunities for negative
moment region cracking prevention still exist.

Construction Sequencing
The conditions that resulted in the inadvertent concrete beams continuous for
both dead and live load have occurred more than once. In 1992, New Mexico
replaced a bridge carrying U.S. 70 over the Rio Grande in Las Cruces. It was
inadvertently made into this type of bridge when the diaphragms were poured and
cured before the deck was added. Subsequently, diaphragm cracking and spalling
ensued as the beams rotated within the green-cured diaphragms when the deck
dead load was applied. This problem was repeated in 2004, when NMDOT
replaced Interstate-10 over University Road and railroad track in Las Cruces. In
this case, the builder asked permission to pour the diaphragms first, and
NMDOTs district field representative assentedalbeit with the advisory that any
adverse consequences would be the contractors responsibility. There were
consequences.
Despite the fact that the contractor subsequently made good on the
consequences, the district had, twelve years earlier, experienced the same issue
and failed to recognize the penalty involved in pouring diaphragms out of
sequence.

We recognize a need for delineating this more mysterious designer choice
on the plansperhaps with a note emphasizing not to violate this sequence
without specific consent of the bridge designer. It is also obvious that we might
benefit from some training program that passes these lessons onto field project
managers so that institutional memory is not lost with retirement, promotions or
turnover.

Semi-Integral Abutment Behavior
Though we have found it expedient to concentrate on using semi-integral
abutments, they do not always behave as predicted. Alternatively, they may
behave as expected, but there are known adverse outcomes.

134
First concerns revolve about expected abutment movement in contrast to
actual abutment movement. Just as jointed bridges often move in unexpected
ways at their joints, so too, we discover that many of the semi-integral abutments
may not actually move as anticipated. Recently, several commonly applied details
have surfaced that are not actually allowing full movement. For example, it
appears that possible use of bituminous materials specified to serve as a spacer
between concrete pourswhich are expected to allow bridge movementmay
not be as compliant as assumed.

Construction details in the field also thwart design intent. Materials meant to
break cold bonds between separate concrete pours and allow sliding movements
may not be applied full-width or full-lengthresulting in seizing or friction
resistance to movement. Bolts intended to allow sliding movement at bearings or
guardrail transitions may get improperly installed, upsetting intended free
movement. To respond to this, at least at the bearings, NMDOT now uses details
at bridge supports where anchor bolts are no longer used, but more can be done.

Fully integral abutments might address some of these problemsa flexible
abutment offers fewer opportunities for other bridge end details to hang up and
drag a stiffly emplaced abutment. However, New Mexicos varying geology
encourages retaining variety in foundation options. We need drilled shafts and
spread-footingsand these foundations are not conducive to flexible, fully
integral abutments. We have also found that Mechanical Stabilized Earth (MSE)
walls do not behave well when driven piles, instead of spread footings, are in the
abutment foundations. It is difficult, if not impossible, to achieve full compaction
in the MSE wall around driven piles. Spread footings work better in MSE wall
performance.

Approach Slab
Even when these roving bridge-ends do move properly, there are
complications. Approach slabs continue to settle. The differential movement
between approach slabs and wing-walls stresses the seals along the sides
resulting in early failure and introducing water beneath the approach slabs.
Subsequently, the live loads actuate the approach slab deflection and pump the
sub-grade soil out from underneath. The sleeper support reduces settlement but
does not eliminate it.

We are experimenting with new approach slab details to prevent compromised
seals from allowing water to flow underneath approach slabs. We are also
considering adding injection ports to the wing-wall sides to provide readily
available access for grout or epoxy injections underneath approach slabs.
Sometimes, we have considered the approach slabs to be sacrificial and easier to
replace or repair than the bridge end and abutmentsthe districts, which are
responsible for maintenance, however, do not seem to completely endorse this
attitude toward approach slabs.

135

METHODS FOR IMPROVEMENT

In routine state bridge operations, it is easy to overlook reviewing habits and
practice in customary design approaches. Outcomes research would be a valuable
tool. Whether it is a formal part of standard bridge management or an activity that
individual design unit supervisors undertake, someone needs to review the
performance of similar bridge designs two years or more after their completion. It
is not adequate to review the bridges immediately after construction. Nor is it
reasonable to expect the biannual bridge inspections to recognize design details
that affect bridge quality. Instead, design engineers need to investigate the
outcome of their own details after the bridges have been in service for some years.

Abutments need examination. New Mexico might want to consider an
increased use of integral abutments. We should carefully consider abutment
movements and our assumptions about specified materials. Assuming that
previous designs have functioned satisfactorily due to the absence of adverse
feedback from the field is not adequate. Several materials that we have considered
flexible do not actually perform as expected. More care should be applied to
insuring true bond breaks by adequate detailing and by attention to construction
quality in the field.

Approaches need examination for better materials and better details. Field
elevation measurements, post construction and later, would establish the true
degree of settlement. We also need to consider alternate approach slab design
details that tolerate settlement without depending on sealant materials to prevent
water ingress.

Communicating the more subtle reasons for our design approaches to both our
own field personnel and to contractor personnel needs improvement. Better notes
and better detailing may be adequate. Training and improved paths for
communication feedback would also serve us well.

REFERENCES

1. Western Regional Climate Center. Climate of New Mexico. Reno Nevada: WRCC Desert
Research Institute Website. http://www.wrcc.dri.edu/narratives/NEWMEXICO.htm (Jan
2005)

2. Lampe, N., and A. Azizinamini. Steel Bridge System, Simple for Dead Load and Continuous
for Live Load. Proc. Conference of High Performance Steel Bridge, Nov 2000, Baltimore,
MD.

136
Integral Abutment Bridges Iowa and Colorado
Experience

David Liu and Robert A. Magliola, Parsons, Chicago, IL
Kenneth F. Dunker, Iowa Department of Transportation, Ames, IA

ABSTRACT

Integral abutment or jointless bridges have many advantages over full height
abutment or stub abutment bridges. They eliminate or reduce expansion joints in
bridge superstructures. They also simplify design, detailing, and construction. For
the last several years, Parsons has designed more than a dozen integral abutment
bridges in Iowa and Colorado. In this paper, the Iowa and Colorado design criteria
for integral abutment bridges are reviewed. Case studies of integral abutment
bridges for several projects are presented. The girder types used in these projects
are welded steel plate girders, prestressed concrete I-girders, prestressed box
girders, and buried slab on prestressed concrete I-girders or box girders. A variety
of foundation systems, such as end bearing H-piles, friction bearing H-piles,
drilled shafts, a combination of H-piles (or W sections) and drilled shafts, or
caisson walls are used in these bridges.

INTRODUCTION

By eliminating or reducing expansion joints and expansion bearings, integral
abutment bridges reduce the cost for construction and maintenance. This type of
bridge can increase design efficiency, add redundancy and capacity for
catastrophic events, enhance load distribution for girders or beams at bridge ends,
provide better protection for weathering steel girders, speed up construction,
reduce tolerance problems, and provide greater end span ratio ranges. With these
benefits, integral abutment bridges have been growing rapidly for the last decade.

Integral abutment bridges have their limitations. The maximum bridge length
is limited by allowable thermal movement of about 2 inches [1]. Like other states,
the Iowa DOT and the Colorado DOT have their own bridge length limits ranging
from 300 feet to 790 feet depending on the type of bridges and skew angles [1],
[2]. Integral abutments apply to straight girders only. That means up to 25% of
new steel bridges each year can not use integral abutments.

A typical section through an integral abutment for a steel girder bridge is
shown in Figure 1. Steel S sections or plain elastomeric pads are provided under
steel girders and concrete beams as temporary bearing points prior to casting the
abutment backwall. H-piles are placed in prebored holes of about 10 feet deep.

137
Deeper prebored holes could increase the bridge length limits [2]. When driving
problems may occur for H-piles, H-piles can be used in the top portion of the
foundation, and the bottom of these H-piles can be embedded in cast-in-place
drilled shafts. This hybrid foundation system can provide the flexibility that is
needed for integral abutments and at the same time it can avoid the driving
problem for H-piles.

There are several methods for the analysis of integral abutment piles.
Wasserman presents a design procedure as follows [3]:

Calculate the thermal movement demand;
Calculate the plastic-moment capacity of the embedded pile:
x y p
Z F M = .
Run COM624P program with fixed head condition and see if the plastic
rotation of the pile is developed;
Calculate the column capacities from AASHTO Article 10.54, and
develop the resulting interaction diagram;
Calculate the adequacy of the backwall to resist passive pressure due to
thermal expansion by assuming a uniformly increasing load applied to a
simple beam.

The research on integral abutment pile design, conducted by Iowa State
University under Iowa DOT Project HR-273 recommends two alternates to
determine the pile loads. Alternate one accounts for the elastic stresses produced
by the horizontal displacement of the abutment and the stresses induced by the
vertical load. Alternate one is recommended for concrete, timber and some steel
piles that have a limited amount of ductility [4].

Alternate two assumes that the inelastic stresses in the pile due to the
longitudinal movement of the superstructure have no significant effect on pile
capacity. However, the secondary P- effect is considered in this alternate. In
other words, alternate two neglects the stresses produced by the horizontal
displacement but considers the stresses caused by the vertical load on the
displaced pile. Alternate two is recommended for steel piles in which their
moment-rotation capacity exceeds the moment-rotation demand at the plastic
hinge location [4].

Other rational approaches to analyze the pile loads are the use of the LPILE
program and a finite element model. The LPILE program can be used to find the
maximum moment in the pile with different boundary conditions at the top of the
pile. One of the boundary conditions is the fixed condition with anticipated
thermal movement. The other is the pinned condition with the associated moment.
The finite element program can be used to model girders, piers, abutment
diaphragms and piles under thermal loading and lateral soil pressure.

In this paper, the Iowa and Colorado design criteria for integral abutment
bridges are reviewed. The integral abutment design and its details for several

138
projects are presented. The different types of piles, such as end bearing H-piles,
friction bearing H-piles or a combination of H-piles (or W sections) and drilled
shafts used in these projects are discussed.

IOWA DESIGN CRITERIA [2]
Integral and stub abutments are the two commonly used types of abutments in
Iowa. In order to minimize construction and maintenance costs, Iowa prefers
integral abutments for typical bridges. The bridge length limits adopted in 2002
for use of standard integral abutments are given in Table 1.

Table 1. Bridge Length Limits for Use of Integral Abutments (Iowa)
Superstructure Type
and Pile Shape
Length and Skew Limits for Standard
Integral Abutments
Maximum
End Span
Concrete Beam
HP 10x57

575 feet at 0-degree skew to 425 feet at
45-degree skew, with linear interpolation
of length for intermediate skew
120 feet

Steel Girder
HP 10x57

400 feet at 0-degree skew to 300 feet at
45-degree skew, with linear interpolation
of length for intermediate skew
105 feet

The following are some of the conditions that the bridge length limits in the
Table 1 are based on:

Integral abutments are placed at both ends of the bridge. If a working
integral abutment is feasible at only one end of a bridge, the maximum
length limit for the bridge shall be one-half the limit in the table, with no
change in maximum end span length.
Each abutment pile is loaded to 6 ksi or less. If HP 10x57 piles are loaded
to 9 ksi, the maximum end span length for a steel girder or concrete beam
bridge shall be reduced by 15 feet.
All abutment piles for bridges longer than 130 feet are placed in prebored
holes 10-feet deep and filled with bentonite slurry. Bentonite slurry is

139
assumed to provide no bearing capacity or lateral support for the piles.
Prebored holes may be increased in depth to 15 feet to reduce or eliminate
downdrag forces.
All abutment piles are a minimum length of 2.5 times the prebore depth,
from bottom of footing to bearing end.
In cases where a MSE retaining wall is used near an integral abutment,
each pile shall be sleeved with a corrugated metal pipe (CMP) to control
compaction near the pile as the embankment and MSE wall are built. At
the top, the CMP sleeve shall be blocked temporarily with framing lumber
so that the pile remains at the center of the sleeve. The CMP sleeve should
be at least as long as the required prebore and shall be filled with sand to
the elevation of the bottom of prebore and then with bentonite to the top of
the CMP sleeve.

A typical integral abutment partial plan and section for a steel girder bridge is
presented in Figure 1. For skews less than 30 degrees pile webs are oriented
parallel with the abutment. For concrete beam and steel girder bridges with skews
greater than 30 degrees, piles shall be oriented for weak axis bending with pile
webs perpendicular to centerline of roadway.

Figure 1. Integral Abutment Partial Plan and Section (Iowa)

For bridge lengths up to and including 130 feet prebored holes are not
required, and piles need not be loaded with impact. Impact should be considered
for piles with prebored holes.

140
COLORADO DESIGN CRITERIA [1]

The maximum structure lengths for integral abutment bridges are shown in
Table 2. These lengths are based on the center of motion located at the middle of
the bridge and a temperature range of motion of 2 inches. The temperature range
assumed is 80 degree F for concrete decked steel structures and 70 degree F for
concrete structures.

Table 2. Bridge Length Limits for Use of Integral Abutments (Colorado)
TYPE OF GIRDER MAXIMUM STRUCTURE LENGTH
STEEL 640 FEET
CONCRETE 790 FEET

A typical integral abutment section for a steel girder bridge is presented in
Figure 2. If caissons or spread foundations are used in lieu of the piles shown in
Figure 2, sliding sheet metal with elastomeric pads may be used on top of
caissons or spread foundations when a pinned connection does not provide
enough flexibility.

For pretensioned or post-tensioned concrete bridges a provision for creep,
shrinkage, and elastic shortening should be provided. If this shortening plus
temperature fall motion exceeds 1 inch, temporary sliding elements between the
upper and lower abutment may be used.

141

Figure 2. Typical Abutment Section (Colorado)
IOWA I-35/I-80 PROJECT

From 1998 to 1999, Parsonss Chicago office performed the final design of
six bridges for the I-35/I-80 construction project in Polk County, Iowa. The
integral abutments were adopted for two continuous welded plate girder bridges
and four pretensioned prestressed concrete I-beam bridges. All abutments are
founded on H-piles, and the abutments for two steel bridges were constructed in
three stages with the use of the mechanical splicers. Since the skews in these
bridges are moderate, no special conditions were considered in the pile design.

Table 3. Bridge Geometrics for I-35/I-80 Project (Iowa)
Bridge Name Type of Girder Bridge
Length
Span
Length
Skew
I-80/2
nd
Ave. Steel Plate Girder 252 122-130 0
o

I-80/14
th
St. Steel Plate Girder 244 117-127 4
o
-01-38
I-80/Two-Track Concrete I-Beam 295 52-57-57-
77-52
0
o

I-80/Single Track Concrete I-Beam 196 47-87-62 13
o
-30-39
Ramp A/UPRR Concrete I-Beam 185 52-67-66 7
o
-14-45
Ramp D/UPRR Concrete I-Beam 185 52-67-66 18
o
-15-31

Table 3 presents the geometrics of these bridges. A typical integral abutment
partial plan and section for concrete I-beam bridges are shown in Figure 3. The

142
end bearing H-piles were used in this project. The typical integral abutment
partial plan and section for steel girder bridges are similar to Figure 1.

Figure 3. Typical Abutment Section for Concrete I-Beam Bridges (Iowa)

IOWA I-235 PROJECT

In 2001, two integral abutment bridges were designed for the I-235
construction project in Polk County, Iowa. The 9
th
street over I-235 bridge is a 3-
span, continuous steel plate girder bridge with span lengths of 83-104-122. I-
235 W. B. over Easton Blvd. bridge is a single span, 172 feet long, steel plate
girder bridge with a skew of 45 degrees.

Since one of these two bridges has a skew angle of 45 degrees and the other
has a bridge length of 309 feet which exceeds the maximum length of 300 feet
allowed at that time, the pile design for both bridges was based on the research
report conducted by Iowa State University under Iowa DOT Project HR-273. As
mentioned in the introduction, alternate two is more applicable for H-piles. The
pile service load design procedure for alternate two is outlined as follows:

Find the applied bending stress:
The moment due to the end span DL+LL+I rotation is

w
pile
L
EI
M
4
=
EI
WL
girder
w
24
2
= (1)

The moment due to the thermal movement is

143
2 / = P M (2)
The resulting extreme fiber bending stress

S
M M
f
w
b
+
= (3)

Check the stability Equation:

0 . 1
) 1 (
'
+
b
e
a
b m
a
a
F
F
f
f C
F
f
(4)

Check the yield equation:

0 . 1
472 . 0
+
b
b
y
a
F
f
F
f
(5)

Determine plastic hinge rotation demand:
According to the Iowa Bridge Design Manual, most H-piles satisfy the
ductility check except HP 12x53 and HP 14x73 shapes at 50 ksi [2].
IOWA US 20 PROJECT

In an ongoing project for US 20 across the Mississippi River in Dubuque
County, Iowa, integral abutments are used for four approach bridges. They are 3-
span, continuous steel plate girder bridges. These integral abutments are placed on
one side only, and fixed bearings are used for two adjacent piers. Since the bridge
lengths exceed the allowable lengths shown in Table 1, approval from IDOT was
requested.

Table 4. Bridge Geometrics for US 20 Project (Iowa)
Bridge Name Alignment Bridge Length Span Length Skew
Ramp C Tangent 300 100-100-100 0
o

US 20 WB Tangent 271 97-97-77 0
o

US 20 EB Tangent & curved 345 115-115-115 0
o

Ramp D Tangent & curved 290 90-110-90 0
o

Due to the site restraint, M.S.E. walls are placed in front of abutments. Each
H-pile is sleeved with a corrugated metal pipe (CMP) to control compaction near
the pile as the embankment and MSE wall are built. Because of the existence of
compressible soil layer, a down drag force is considered in pile design. A
maximum prebore depth of 15 feet is used to reduce down drag forces.

144
The geometrics of these four bridges is presented in Table 4. A typical integral
abutment section is shown in Figure 4. The corrugated metal pipes are extended to
the bottom of the prebored holes. Figure 5 presents the blocking detail for the
CMP. The blocking is placed at the top and bottom of the CMP. The bottom
blocking shall be treated material and shall remain in place. The top blocking
shall be removed prior to backfilling inside the CMP.

Figure 4. Typical Abutment Section for US 20 Bridges (Iowa)

Figure 5. Corrugated Metal Pipe Blocking Detail (Iowa)

T-REX PROJECT

In a recent Southeast Corridor Multi-Modal project (T-REX) in Denver,
Colorado, integral abutments with different foundation systems were used in most
bridges. Table 5 presents the geometrics of some of these bridges. The integral

145
abutments used in Structure No. 34 are supported by W10x68 piles in the top
portion, and the bottoms of these W-piles are embedded in cast-in-place drilled
shafts. W-piles provide the flexibility that is needed for the integral abutments.
The use of cast-in-place drilled shafts can avoid the driving problems H-piles may
experience. The details for this type of integral abutment are shown in Figure 6.

The LPILE program was used to analyze the piles and drilled shafts. The fixed
condition at the top of W10x68 piles was assumed. Both W10x68 piles and drilled
shafts were designed for axial compression and bending.

Since most of these bridges are short simple bridges and the thermal
movement is very small, caissons and caisson walls were used in this project. For
longer span bridges, mainly LRT bridges, the expansion devices that consist of
sliding sheet metal with elastomeric pads were used to reduce the demand for
foundation flexibility.

146
Table 5. Bridge Geometrics for T-REX Project (Colorado)
Structure
No.
Girder Type Foundation
System
Bridge
Length
Span
Length
Skew
17 Buried Slab on
P/S Box Girder
W12x72 on 2
Dia. caisson
53 53 19
o
-46-
17
18 P/S I-Girder 2 Dia.
Caisson
36 to 45 36 to 45 Varies
27 P/S Box-Girder 2 Dia.
Caisson Wall
34 Buried Slab on
P/S I-Girder
W10x68 on 2
Dia. caisson
38 P/S I-Girder,
Ballasted LRT
Bridge
2 Dia.
Caisson
218 109-109 Varies
48 & 49 Steel Plate
Girder
2-6 Dia.
Caisson
330 165-165 42
o
-48-
00
72 P/S I-Girder,
Direct Fixation
LRT Bridge
3 Dia.
Caisson
420 155-165-
100
0
o

73 Buried Slab on
P/S I-Girder
2 Dia.
Caisson
74 P/S I-Girder,
Ballasted LRT
Bridge
W14x89 on
2-6 Dia.
caisson
112 112 53
o
-36-
33
77 P/S I-Girder 2 Dia.
Caisson
76 76 45
o
-04-
50
83 P/S I-Girder,
Direct Fixation
LRT Bridge
W14x109 on
3-6 Dia.
caisson
390 125-140-
125
0
o

CONCLUSION

With the advantages of jointless, simple design, easy detailing, and quick
construction, integral abutment bridges are the preferred structure type in many
projects. Depending on bridge lengths and site conditions, the foundation systems
that could be used for integral abutments are end bearing H-piles, friction bearing
H-piles, drilled shafts, a combination of H-piles (or W sections) and drilled shafts,
and caisson walls. Since integral abutment bridges are limited to straight girders,
future research is needed to provide guidance on integral abutments for curved
bridges. The effects of approach slab and compressive soil on integral abutment
pile design should be investigated as well. A unified design guideline for integral
abutment bridges should be developed.

147

REFERENCES

1. Colorado Department Transportation. 2002. Bridge Design Manual, August 1, 2002.
2. Iowa Department Transportation Office of Bridges and Structures. 2004. Bridge
Design Manual, July 14, 2004.
3. Wasserman, E.P. 2001. Design of Integral Abutments for Jointless Bridges, 2001
Lecture Series: Bridges, ASCE Illinois Section, Structural Group.
4. Greimann, L.F., R.E. Abendroth, D.E. Johnson, and P.B. Ebner. 1987. Pile Design and
Tests for Integral Abutment Bridges, Final Report and Addendum, Iowa DOT Project
HR-273, College of Engineering, Iowa State University, Ames.

Figure 6. Typical Integral Abutment Section for T-REX Project (Colorado)

148
Moose Creek Bridge Case Study of a prefabricated
Integral Abutment Bridge in Canada

Iqbal Husain, P.Eng., Ministry of Transportation, Ontario
Ben Huh, P.Eng., McCormick Rankin Corporation, Mississauga, Ontario
John Low, P.Eng., Stantec Consulting Ltd., Hamilton, Ontario
Mike McCormick, P.Eng., Ministry of Transportation, Ontario

ABSTRACT

This paper describes the design and construction issues of a single span
prefabricated integral abutment bridge. The superstructure consists of precast
prestressed girder/full depth deck elements and the substructure consists of
precast abutment stem and wingwall units. The units were fabricated at the
precast plant and transported to the site where they were assembled using closure
strips.

Integral abutment bridges are the most common type of bridges now used in
Ontario. Moose Creek Bridge is the first integral abutment type bridge built in
Ontario using the T-shaped prefabricated girder deck system. The abutments were
made monolithic with the deck by casting concrete closure segments.

Use of prefabricated bridge systems to construct bridges reduces construction
time considerably and enhances the quality due to fabrication in a controlled
environment. Prefabricated bridge systems also provide a number of other
significant advantages such as reduced traffic impact, improved construction zone
safety and less disruption to the environment. Ministry of Transportation
investigated the suitability of the T-shaped girders with closure concrete pours to
connect the T-beams. Scaled models of the system were tested in the Ministrys
research lab and were found to behave adequately under static and cyclic load
tests. It was decided to test the concept by constructing a prototype bridge and to
evaluate construction issues and performance under site conditions. Moose Creek
Bridge was selected due to its size, simplicity and construction schedule.

INTRODUCTION

Moose Creek Bridge is the first fully integral abutment bridge in Canada
using the precast elements and segments for both the superstructure and
abutments. First integral and semi-integral abutment bridges in the Province of
Ontario were designed and built in 1960's. It is only in 1990's that the Province
increased its effort in the design and construction of such bridges and issued two
reports [1,2] to establish the general guidelines for the planning, design and
construction of such structures in the Province. The performance of some of these
bridges is monitored visually and any deficiencies are recorded. Ministry of

149
Transportation (MTO) published a report on findings of this monitoring in 1999
(3). More than 150 new structures have been built and appear to be performing
well without any major cause for concern. The simplicity, economy, durability
and performance of integral abutment bridges in recent years has considerably
influenced the type of structures being used in the province of Ontario, so much
so that almost 60% of bridges being built are of integral abutment type.

It is, therefore, not surprising that when MTO decided to increase its effort in
the use of prefabricated elements and systems for the design and construction of
the bridges it chose an integral abutment structure first. The Moose Creek Bridge
is located on Highway 101 in Northern Ontario approximately 450 km from
Toronto. Owing to the advanced stage of deterioration it was planned to replace
the existing bridge with a cast-in-place concrete slab on standard precast
pretensioned concrete girder deck and cast in place integral abutment. The
conventional construction method for thin-slab bridge decks typically involves
on-site casting of the full-depth reinforced concrete deck on top of the naked steel
or precast prestressed concrete girders. This method generally requires at least a
month for casting and curing concrete, and the quality of cast-in-place concrete is
sometimes difficult to ensure on site, especially in remote areas of Ontario.
During the detail design stage it was considered necessary to explore a
prefabricated design and construction for this bridge to overcome some
scheduling problems and the desire to demonstrate the feasibility and
effectiveness of prefabrication technology in Ontario.

A single 22m (72) span Integral Abutment design using T-shaped precast
prestressed girder deck elements with minimal infill closure strips was considered
for the superstructure. The substructure also consisted of precast elements
supported on single row of steel H-piles made monolithic with the deck by casting
a closure strip at top end the abutment.

RESEARCH AND LABORATORY TESTING OF SCALE MODELS AT
MTO

For a prefabricated design it is important that the joints between the elements
and systems can transfer the loads and forces adequately as in a monolithic
structure and provide the desired durability. To this end the Bridge Office of the
Ministry of Transportation carried out in-house research and tested scaled models
of potential superstructure systems to validate the design assumptions. Two
potential systems were envisaged:

Prefabricated Bridge System A consisted of T-shaped precast composite slab-
on-girder (steel or prestressed concrete) elements transported to the site and
assembled together with cast-in-place concrete closure strips (Figure 1 & 2,
System A). Prefabricated Bridge System B consisted of full-depth precast
concrete deck slab elements laid across new or existing girders on site and

150
connected together with cast-in-place concrete over the girders and between the
slab elements (Figure 1& 2, System B).

Figure 1. Prefabricated Bridge Systems

Prefabricated T-Girders with closure strips
System A

Prefabricated Slab Panels with closure strips
System B

Figure 2. Test Models

151

Laboratory tests were performed to study the structural behaviour of the
overall bridge system under service loads, the long-term load effect on the
longitudinal cold joints, and the ultimate load-carrying capacity of the concrete
deck after being subjected to repetitive load cycles (Figure 3). Seven million load
cycles were applied at two locations on the deck, and the results of this cyclic load
test showed little evidence of distress and no major impact on structural behaviour
or serviceability. Other than hairline cracks along the cold joints, no major signs
of distress were observed. The ultimate limit test, performed after the cyclic load
test, demonstrated ample strength in the bridge deck to resist design wheel loads.
The ministry recently published the findings of a series of laboratory test on these
models to study the overall structural behaviour of these bridge systems under
service, ultimate and cyclic loadings and the performance of longitudinal cold
joints [4]. Encouraged by the excellent structural behaviour and integrity of the
joints MTO decided to proceed with a full-scale prototype application of this
concept in the Ontarios first prefabricated bridge.

Figure 3. The Experimental Set-Up and The Loading System

DESIGN AND CONSTRUCTION

The Moose Creek Bridge [5] utilizes a 22m (72) long a slab on girder type
superstructure to span the creek (Figure 4). The total bridge width of 14.64m (48)
is comprised of six precast pre-tensioned concrete CPCI 1200 girders cast
integrally with full depth concrete deck at 2.45m (8) centre-to-centre spacing,
and two cast-in-place safety barriers as shown in Figures 4 and 5. Full depth cast-
in place concrete closure strips were detailed to complete the connection between
the T-shaped girder deck sections (See Figure 5). The deck surface is protected by
waterproofing membrane and a bituminous pavement overlay is provided as
riding surface. The superstructure is supported by full height, precast concrete
abutment units founded on steel H-piles.

152

Figure 4. Moose Creek Bridge as Built

Figure 5. Typical Section

The bridge was designed in accordance with the first edition of Canadian
Highway Bridge Design Code (CCAN/CSA-S6-00) [6]. It was designed using the
simplified methods of analysis in the code by idealizing the deck configuration as
a two dimensional frame for the preliminary design purpose [7]. It was difficult to
use the simplified method of analysis accurately for the live load distribution at
the service stage because of integral abutment configuration of the bridge. The
code is not quite specific about this situation. Thus, the finite element method of
analysis was selected as the preferred method for predicting the complex loading
behaviour, considering the three-dimensional nature of the structure.

153
Figure 6. 3D Analysis Model for Composite Action

The three-dimensional finite element program MIDAS/CIVIL [8], utilizing
quadrilateral plate elements subjected to in-plane and bending effects, was used to
idealize the deck and abutment stems. The plate element has four corner nodes
that have three translational and three rotational degrees of freedom. Three-
dimensional beam elements were used for the CPCI 1200 girder sections and steel
H-piles as displayed in Figure 6.

Precast Substructure System

Over the years cast-in-place (CIP) concrete substructures have proven to be
very versatile. The familiarity with CIP concrete structures is universal in both
design and construction as they are generally more forgiving in case field
adjustments are needed. They can adapt to varying site conditions and
construction tolerances are easily achieved. However, CIP substructures are time
consuming to construct and the quality of workmanship is not always guaranteed.
Prefabricated substructure design provides an opportunity to apply advanced
technologies and use new materials under controlled environment to enhance and
expedite bridge construction. Specifically, the prefabricated substructure
consisting of abutment units offers rapid construction, durable performance and
an attractive appearance.

The performance of the abutment-pile connection is an important aspect in the
overall behavior of integral abutment bridges. A single row of piles was used to
support the abutments. Predrilled oversize holes filled with loose dry sand were
provided around each pile to provide flexibility and allow the piles to move
laterally. The connection between the abutment wall and pile top was analyzed for
both pinned and fixed conditions for the pile design. When the connection is
considered as pinned between the pile top and abutment, it may allow free
rotation of the pile top about an axis perpendicular to the bridge longitudinal axis.
If the connection is designed as fixed, plastic bending moments may develop at
the pile top due to thermal movements and effect of vehicular traffic and may
result in the formation of a plastic hinge [1].

154

Figure 7. Plan and Elevation View of Abutment Units

The abutment wall was divided into three equal units for each abutment due to
their size and weight restriction on shipping and erection as shown in Figure 7.
The design of the precast abutment unit is no different from the design of a cast-
in-place abutment. The difference is in the details of connections. A typical unit
was 4.5m (14.8) wide x 1.0m (3.3) deep x 3.3m (10.8) high in order to meet the
weight and size restriction for transportation in Ontario. Each abutment wall unit
had two openings for connecting to pile top to provide stability during
construction as shown in Figure 8. The top of piles was embedded at least 600mm
(2) into the abutment walls and was adequately reinforced to transfer the bending
and bursting forces. The reinforcing steel was carefully detailed in the precast
section to allow for maximum construction tolerances of 150mm (6) in each
direction for the pile placement. In order to facilitate the concrete placement over
the piles, the top of opening in the precast section was tapered. Each abutment
consisted of three wall units and two wingwall units. The exterior abutment units
were precast with the lower wingwall portions to which the wingwall units were
secured with the help of mechanical anchorage. Once abutment wall units were
erected over the top of the piles, they were held in place until the placement of
reinforcing steel and concrete in the vertical closure strips, the pile tops and the
bearing seat pedestals (Figure 9). After concrete had reached its specified
strength, T-shaped girder elements were placed in their final locations in a
specified sequence as shown in Figure 11. The continuity between the abutment
and the superstructure was achieved by placing a cast-in-place closure segment
joining the abutment and the deck together.

155

Figure 8. Typical Abutment Stem Units with Dimensions

Figure 9. Erection of Abutment Wall Units

Precast T-Shaped Girder Details

Concrete deck girder elements can eliminate the requirement for formwork
and placing of steel in the field and considerably reduce the need for cast-in-place
concrete, usually associated with conventional slab on I-girder bridges, resulting
in reduced construction time. The primary consideration when designing a
structure with precast components is the weight and size of the components. The
shipping cost is very high when transporting oversize components over long
distance or to remote areas is required. A deck element is a precast, prestressed
concrete I-girder with an integral deck in a T-shape that is cast and prestressed
with the girder shown in Figure 10. All girders were prestressed with twenty-eight
12.7mm (0.5) diameter, 1860MPa (270ksi) low-relaxation steel strands as shown

156
in Figure 10. These T-shaped girders were manufactured in precast concrete
plants under closely controlled and monitored conditions, transported to the
construction site and erected such that flanges of adjacent units abut each other.
Sections that are not too long or too heavy for transportation by truck can be used
to construct bridges using these elements.

Figure 10. Typical T-Shaped Girder Unit with Reinforcing Details

Variable deck thicknesses were used in the T-shaped girders to control the
camber. The camber is a by-product of prestressing forces applied internally or
externally to bridge girders. Camber of a girder at any age is algebraic sum of the
upward deflection due to prestressing and the downward deflection due to self-
weight and other applied loads. In traditional slab-on-girder type structure, the
thickness of cast-in-place bridge decks (or haunch) can be adjusted to compensate
for camber and to maintain the required deck profile on bridges. When a full deck
slab is precast as a part of the girder there is no opportunity for any screed
elevation adjustment on the deck. As a result, the girder elements were cast
integrally with 240mm to 290mm thick variable deck in longitudinal direction to
compensate for the differential camber for these girders.

157

Figure 11. Erection of Girder Units

Connection details between deck girder elements are also very important
aspect for constructability, performance and serviceability of the structure. The
detail of infill strip that was designed for ease of construction and provide leak
proof joint performance over the service life is shown in Figure 12. The 70mm
(2.75) thick self-form section did not only eliminated formwork during concrete
placement on site but also ensured structurally integrated deck with cast-in-place
concrete, as illustrated in Figure 13. The 40mm (1.6) x 75mm (3) shear key was
provided in the precast section to ensure structural integrity in the cast-in-place
concrete strip and the precast deck slabs. A customized wood form was used for
fabricating these T-shaped girders. The deck portion on top of the girders had a
crossfall of 2%, same as conventional straight bridges. Only one wood form was
used for all six T-shaped girders needing only a slight modification, due to the
end details of deck sections. Well-assembled wooden form could be used to cast
up to 10 or 15 precast components. The standardization of cross-section and shape
may enable the fabricator to design steel forms that would allow customization
with minor adjustments resulting in savings.

Figure 12. Typical End Details of Deck Sections

158

Figure 13. Ready for concreting closure strip joints

Cast-In-Place Concrete Joints

The bridge deck elements and the abutment units were designed assuming a
continuous frame action at the joints linking the bridge deck to the abutment. A
monolithic connection between precast deck elements and abutment units was
achieved by cast-in-place concrete closure strip placement, providing enough
continuity of reinforcing steel essential to achieve full moment connection. The
continuity between T-shaped deck elements was achieved by connecting them
together with cast-in-place concrete closure strips as shown in Figure 14.

Formwork at the underside of the deck for the closure strip was considered
undesirable as it would be difficult to access from underneath the deck and would
require time for curing of concrete. To eliminate formwork completely a nib
detail, as shown in Figures 13 and 14, was configured. The nib section of the deck
needed to be a minimum thickness to accommodate the reinforcing steel and the
minimum concrete over requirement. A 25mm (1) diameter form backer and
sealant were used between ends of prefabricated decks and tip of the nib to
provide flexibility in erection. To maintain the required development of the
reinforcing steel through the closure strip, the transverse reinforcing steel for the
deck was bent in U-shaped loop to achieve a full strength connection as was
already established through model tests described earlier. The placing of loop
reinforcing steel was alternated with the reinforcing steels in adjacent deck
elements. In order to ensure the precision required in placement of this
reinforcement work it was necessary to use templates during the casting of these
elements.

159

Figure 14. Typical CIP Deck Joint Details

Figure 15. Before CIP concrete (left), After CIP concrete (right)

The barrier walls and approach slab were only cast-in-place concrete
components for this bridge other than in-fill strips and joints between the deck
and abutment. In future, the exterior T-girder sections could be prefabricated
integrally with barrier wall or the precast barrier walls could be assembled on top
of the prefabricated deck on site.

In spite of the obvious benefits and advantages of prefabricated precast
elements and systems, their use also raises concerns about certain design,
construction and performance issues that are perceived to influence the structural
integrity and long term durability of the bridge system and its components. These
issues include long term performance of the connections between adjacent
element and units, the handling of camber and cross slope for long span
structures, distribution and continuity for live load, lateral load resistance, skew
effects, future maintenance and other factors that influence constructability and
performance.

160
CONCLUSIONS

The development and use of prefabricated wall units and deck elements for
integral abutments bridges is a major step in using prefabricated bridges in the
province of Ontario as a large proportion of bridges built in Ontario are integral
abutment.

The successful completion of the bridge without any significant design,
fabrication, transportation or erection problems is quite encouraging and it is
expected that the techniques developed here would be used in more future
projects.

To take full advantage of this system it would be necessary to extend it to two
span and multi-span structures that would require the development of the details
for continuity at the pier location and provision of reinforcing steel for negative
bending moment in the deck.

The proposed construction method would significantly reduce on site
construction time that in turn would result in safety, increased durability, effective
traffic management and economy in user and traffic management cost.

A standardization of the cross-section with different depth of girders and span
lengths is expected to follow from this experience. It would also have an impact
on cost reduction, which was higher than expected for this project.

It is suggested that a dialogue should be initiated between the designer,
fabricator and the contractor at planning stage and maintained through all stages
of the project to ensure the use of cost effective design, fabrication ad
construction techniques.

REFERENCES

1. Husain I., Bagnariol D., "Integral Abutment bridges", Structural office report No. SO-96-
01, Structural Office, Ministry of Transportation, Ontario, 1996
2. Husain I., Bagnariol D., Semi-integral Abutment Bridges, Bridge Office Report No.
BO-99-03, Bridge Office, Ministry of Transportation, Ontario, 1999
3. Husain I., Bagnariol D., Performance of Integral Abutment Bridges, Bridge Office
Report No. BO-99-04, Bridge Office, Ministry of Transportation, Ontario, 1999
4. Au A., Lam C, Tharmabala B. Prefabricated Bridge Technology in MTO Proceedings
Concrete Bridge Conference, Charlotte, North Carolina, USA, 2004
5. Huh B. and Low J, 2004, Moose Creek Bridge, the First Field Application of Fully
Prefabricated Bridges in Ontario, 2004 PCI National Bridge Conference
6. CSA International, 2000, Canadian Highway Bridge Design Code (CHBDC) CAN/CSA-
S6-00,
7. Computers and Structures Inc., SAP2000 Version 7.44
8. MIDASoft Inc. MIDAS/Civil Version 5.80

161
SESSION III:
MAINTENANCE AND
REHABILITATION

162

163
Field Data and FEM Modeling of the Orange-Wendell
Bridge

Christine Bonczar
a
, Sergio F Brea
b
, Scott Civjan
c
, Jason DeJong
b
, Benjamin
Crellin
a
, Daniel Crovo
d
a
Graduate Student, University of Massachusetts at Amherst

b
Assistant Professor, University of Massachusetts at Amherst
c
Associate Professor, University of Massachusetts at Amherst
Department of Civil and Environmental Engineering, 130 Natural Resources Rd.,
Amherst, MA 01003
d
Assistant Bridge Engineer, Massachusetts Highway Department, 10 Park Plaza,
Boston, MA 02116

ABSTRACT

The response of a three-span (270 ft. total length) integral abutment bridge
located in Orange, Massachusetts was evaluated through field monitoring and
extensive two and three dimensional finite element modeling. This research
project is being performed at the University of Massachusetts Amherst and is
funded by the Executive Office of Transportation and the Federal Highway
Administration. Design assumptions and integral abutment bridge behavior are
discussed by comparing measured field data with results from analytical models.
The Orange-Wendell Bridge is instrumented with 85 gages for assessing bridge
behavior. Thermal seasonal effects on the bridge were of primary interest. Over
three years of data have been collected including abutment pressures, rotations
and deflections as well as temperatures, pile strain and inclinometer readings.
Only abutment deflection, rotation, backfill pressure and their interactions are
discussed in this paper. It is shown that measured abutment rotations can
constitute a significant component of the total longitudinal displacement of the
superstructure. Behavioral differences in the soil-abutment interaction at the
North and South abutments have been observed. The effects of rapid temperature
changes occurring primarily in the spring on backfill pressures behind the
abutments are also discussed.
INTRODUCTION

The Massachusetts Highway Department, through the Massachusetts
Transportation Research Program, commissioned a field study to validate and
economize the design of integral abutment bridges. This paper presents data from
field monitoring and Finite Element Modeling (FEM) of a three-span integral
abutment bridge located in central Massachusetts. Of primary concern were
seasonal temperature effects on the structure. Measurement of global movements
and abutment soil pressures are presented. Parametric FEM results are also
presented to provide an understanding of bridge behavior and sensitivity to

164
variations in soil properties. Additional information on soil-structure interaction
and the influence of soil properties are reported in a companion paper [1].
BRIDGE DESCRIPTION

The integral abutment bridge described in this study spans over Millers River on
Wendell Depot road in Orange, MA. The three-span continuous structure was
designed for two lanes of traffic using an HS20-44 design truck. The bridge has a
total length of 270 ft (82.3 m) and a 32 ft (9.8 m) width. Exterior spans are 80 ft
(24.4 m) long and the interior span is 110 ft (33.5 m) long. The span is
perpendicular to the abutments. The north end is exposed to direct sunlight, while
the south span is predominantly shaded. Details of the structure are shown in
Figures 1 through 3.

Figure 1. Orange-Wendell Bridge Elevation

The superstructure consists of an 8 in. (200 mm) concrete deck supported on four
48 in. (1220 mm) deep steel plate girders. The girders are evenly spaced every
8.67 ft (2.64 m) across the bridge starting at 3.00 ft (915 mm) from the deck edge
on each side. Concrete guardrails are provided along both sides of the bridge
deck. A 5.33 ft (1.63 m) sidewalk is provided along the east side of the bridge
deck. The steel girders are embedded into the abutment walls at both ends of the
bridge, and supported on elastomeric bearing pads on the two interior concrete
bents. They are braced laterally with cross frames at approximately 20 ft (6.1 m)
spacing throughout the length of the bridge. The bents consist of a 48 in. (1220
mm) wide by 42 in. (1070 mm) deep bent cap supported on three 42 in. (1070
mm) diameter concrete column/pier shafts that extend approximately 45 ft (13.7
m) into the riverbed.

Each abutment wall is supported on 8 HP 10x57 steel piles equally spaced every
4.1 ft (1.25 m). The pile tops are embedded approximately 2 ft (0.6 m) into the
bottom of the abutment and driven approximately 60 ft (18.3 m) into the ground.
The top 10 ft (3.0 m) of the piles were driven into an augered hole that was
backfilled with sub rounded pea stone after pile driving. The piles are oriented
with their weak axis parallel to the transverse bridge axis to offer the least
resistance to bending. This was done to provide maximum deformation capacity
to accommodate thermal displacements of the bridge. Approach slab details can

165
be seen in Figure 3. Wing walls are perpendicular to the abutment and are
separated from it by an expansion joint.

Figure 2. Section at Interior Column Bent

Figure 3. Abutment and Approach Slab Details

166

1 Inclinometer housing along pile
2 Tilt meter
3 Joint meter
4 Strain gages in Pile
5 Earth pressure cells
6 Strain gages in girders
7 Concrete block protective housing
8 Reference pile
9 Connection box

BRIDGE INSTRUMENTATION AND DATA RECORDING

The bridge is monitored using 85 instruments that capture critical parameters of bridge
behavior during seasonal thermal fluctuations. Six different types of gages were used in
this bridge: earth pressure cells, joint meters, tilt meters, temperature gages, strain gages
and thermistors. Instrumentation is concentrated near the abutment components (walls or
piles), as shown in Figure 4.

Figure 4. Abutment Construction and Instrumentation Details

Longitudinal (parallel to the roadway) and transverse displacements of the bridge
are measured using joint meters on the east and west sides of each abutment.
These instruments measure the displacement of the abutment walls relative to
reference piles located on each side of the bridge. Reference piles are offset
laterally from the abutment and are assumed to be minimally affected by
abutment movements. One tilt meter is located at each end of both abutment walls
to monitor rotation near the base of the walls.

Earth pressure cells are located behind the abutment walls at different heights to
measure backfill pressure distribution behind the walls. The vertical spacing
between the pressure cells is 11 in. (280 mm). In the south abutment, a vertical
array of five earth pressure cells is aligned below each of the two interior girders
of the bridge. The earth pressure cells in the north abutment were aligned at the
centerline of the abutment and in vertical arrays between the three piles on the
west side of the bridge. Five pressure cells were used for the centerline location
and three cells were placed in each space between the piles on the west side of the
abutment.

167
Vibrating wire strain gages were placed on the two exterior HP piles supporting
each abutment (Fig. 4), attached to each flange. These instruments are
concentrated in the top portion of the piles and are located at depths equal to 0.5,
2.5, 4.5, 6.5, and 8.5 ft (150, 760, 1070, 1370, and 1610 mm) measured from the
bottom of the abutments. Strain gages were also placed on the top and bottom
flanges of the exterior girders at 7 in. (177 mm) from the face of the abutment
walls.

Temperature gages are located on the underside of the bridge deck at 10 ft (3.0 m)
from the face of the abutments. These instruments provide readings considered to
be representative of the ambient temperatures occurring at each edge of the bridge
deck. In addition, each instrument has an embedded thermistor that monitors
temperature at specific gage locations.

All instruments are connected to a CR10X Campbell Scientific data-logging unit
through the use of six 16-channel multiplexer cards. The system is programmed to
acquire and store data daily from all the instruments every 6 hours starting at
12:30 a.m. The readings are downloaded from the data logger periodically by
modem and processed on a personal computer using a spreadsheet. Data
acquisition of the bridge initiated on 2 January 2002 and is reported through
December 2004.
FEM MODELING

Detailed two (2-D FEM) and three dimensional (3-D FEM) finite element
models were constructed to evaluate the performance of the Orange-Wendell
Bridge analytically using GT-STRUDL [2]. The 3-D model was used to
accurately represent the thermal response of the bridge and compare it with
measured parameters. The 2-D model was developed to provide modeling
recommendations that could be used efficiently in routine design. The 2-D model
was calibrated to match the 3-D model response and the measured field data.

In the 3-D model, physical characteristics of the bridge that required particular
attention in the model construction were the connection between the deck and
girder elements, the connection between deck and abutment elements and the
connection between girders and bent cap supports. Connections between these
components of the bridge were established using either rigid links (deck-abutment
and girder-bent cap) or element offsets (girder-deck). Soil-structure interaction
effects were approximately captured using non-linear spring elements. Geometric
non-linear effects were considered for all members, and material non-linearity
was included as a plastic hinge fiber element at critical pile members.

The simplified 2-D model of the bridge was constructed using beam elements.
Composite properties of the entire superstructure width (deck and 4 girders) were
used for the deck-girder members. Properties were calculated at four different
sections along the bridge because of changes in the thickness of the girder flanges.

168

Deck-girder members were attached to the abutment at the elastic neutral axis of
the transformed deck-girder sections. Interior supports at bent caps were modeled
using pin supports, and abutment piles were assumed fixed at 15 ft (4.6 m) below
the abutment. Results presented herein are for the 2-D model, though similar
results were obtained from the 3-D models. A schematic of the 2-D model can be
seen in Figure 5.

Figure 5. FEM Modeling of the Orange-Wendell Bridge

Two node, non-linear springs were used for soil-springs behind the abutment
and along the length of the piles. National Cooperative Highway Research
Program [3] design curves were used to model passive and active pressure effects
of soil behind the abutment, but comparisons were also made with the
Massachusetts Highway Department Bridge Design Manual [4]. Stiffness values
for the passive mode of soil springs in the MassHighway guidelines are lower
than those included in the NCHRP 343 Report. Because of uncertainty in field
values of soil density, spring properties were calculated for dense and loose soils
for comparison purposes. Symmetric soil springs based on the Hyperbolic
Tangent Method [5] were used for the pile soil interface. Soil around the top 10
feet of the piles was assumed to be loose soil, as the augered hole filled with pea
stone would provide minimal soil restraint. A detailed discussion on assumptions
used to model the top 10 ft of pea stone at the piles is presented in the
accompanying paper [1]. Soil properties assumed for the model are listed in Table
1.

2-D FEM

3-D FEM

169

Table 1. Assumed Soil Properties for FEM Analysis

Soil Type
' (lb/ft
3
) (deg.)
k
l
(lb/in
3
)
Loose 110 30 45
Medium-Dense 125 37 97
Dense 140 45 220

RECORDED READINGS

The construction of the bridge was finished in late 2000. The average ambient
temperature was approximately 70.5 F (21.4C) when initial readings were taken
after casting of the deck and approximately 66.4F (19.1C) after abutment
backfill and bridge completion. In the discussions that follow the temperature
change in the bridge was considered to vary from reference average initial
ambient temperature equal to 66.4F (19.1C). The recorded ambient temperatures
have ranged from 10F (-23C) to 100F (37.8C) during the monitoring period.
Measured field data displacements and rotations were compared with FEM
models that did not include the self-weight of the bridge because the initial
reading in the field was taken after construction concluded. On the other hand,
self-weight was included in the analytical models for comparisons with measured
backfill pressures and pile strains because field zero readings were obtained prior
to deck and backfill placement.

Displacement and Rotations of Bridge Abutments

Comparisons between measured displacement data and FEM results can be seen
in Figures 6 and 7. Bridge expansion (positive) and contraction (negative) were
referenced to several key points. Rotation and displacement data measured 1 ft
(0.3 m) above the abutment base (joint meter location) were used to estimate
displacements at other locations. It was found that expansion and contraction of
the superstructure calculated using current design guidelines [4] matched
relatively well with measured displacements at the centerline of the girder (Figure
6). Resulting abutment rotations caused displacements at the girder centerline to
be significantly higher than displacements at the joint meter location (Figures 6
and 7). Yearly bridge expansion data for both locations has changed from 2002
through 2004. Higher bridge expansion at the south abutment in 2004, for
example, can be clearly observed in Figures 6b and 7b at both displacement

170
reference points. This is attributed to changes in abutment rotation and
displacement from year to year, likely caused by loosening of the pea stone
around the top 10 feet of the piles as discussed by Bonczar et al. [1].

Although the model predicts symmetric response of the two abutments
significantly more contraction was measured at the north abutment and slightly
higher expansion at the south abutment. The observed differences can be due to
unequal backfill conditions, possibly compounded by thermal differences due to
unequal solar heating at north and south abutments. In general north abutment
motions match more closely with FEM results.

-80
-60
-40
-20
0
20
40
-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4
Displacement at CL of Girder(in)
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
2-D FEM N02 N03 N04

a) North Abutment b) South Abutment
Figure 6. Measured and Calculated Abutment Displacement at Centerline of Girder

-80
-60
-40
-20
0
20
40
-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4
Displacement at JM Location (in)
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
2-D FEM N02 N03 N04

-80
-60
-40
-20
0
20
40
-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4
Displacement at JM Location (in)
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
2-D FEM S02 S03 S04

Figure 7. Measured and Calculated Displacement at Joint Meter Location

-80
-60
-40
-20
0
20
40
-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0.4
Displacement at CL of Girder(in)
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
2-D FEM S02 S03 S04
a) North Abutment
b) South Abutment

171
Abutment Soil Pressures

Backfill pressures behind the south abutment measured during 2002 and 2004
are compared in Figure 8 with 2-D FEM results obtained using soil spring
properties for dense soil behind the abutments. Measured and analytical pressures
were compared at four depths by averaging pressure cell readings at each of the
levels indicated in the figures. Data points correspond to average measured
readings over 3-day periods throughout each year. At depths up to approximately
9 ft (2.7 m), field data are in reasonable agreement with the FEM model that
assumed dense backfill conditions behind the abutment, particularly for 2002.
However, measured pressures near the bottom of the abutment are significantly
higher than calculated with the 2-D FEM for both years, and the difference is
more pronounced in 2004. This may be related to initial density and stiffness of
the pea stone that encases the top 10 ft of the pile and the changes in properties
from abutment movements in the first four years. Methods of defining spring
properties may also have influenced pressures at specific points. More refined
analytical models were developed to explain this difference as discussed in the
companion paper [1].

Peak backfill pressures measured behind the abutments over 3-day periods for
2002 and 2004 are shown in Figure 9. At the two deepest locations reported, the
highest yearly pressures occurred during the spring season for both years. The
pressures measured at the two shallower locations generated in the spring months
were comparable to those generated in the summer. Although the bridge
experienced its highest expansion during the summer months, short-term
temperature differentials were much higher during the spring season. These rapid
temperature changes in the spring contributed to the larger observed pressures.
Although pressures near the bottom of the abutments were higher during 2004 the
same trends in seasonal behavior were observed. The complex seasonal behavior
of backfill soil partially explains the scatter observed in the pressure cell readings
and the difference between field data and FEM results using simplified soil-spring
models.

-80
-60
-40
-20
0
20
40
60
-45 -40 -35 -30 -25 -20 -15 -10 -5 0 5
Average Pressure in 2002 (psi)
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
5.38' Data 7.21' Data 8.13' Data 9.05' Data
5.38' Dense 7.21' Dense 8.13' Dense 9.05' Dense

a) Data from 2002 b) Data from 2004
Figure 8. Average Soil Pressure behind South Abutment
-80
-60
-40
-20
0
20
40
60
-45 -40 -35 -30 -25 -20 -15 -10 -5 0 5
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)

172

-5
5
15
25
35
45
55
65
Jan-02 Mar-02 May-02 Jun-02 Aug-02 Oct-02 Dec-02
A
v
e
r
a
g
e

N
o
r
t
h

P
r
e
s
s
u
r
e

(
p
s
i
)
gage malfunction

-5
5
15
25
35
45
55
65
Jan-04 Mar-04 Apr-04 Jun-04 Aug-04 Oct-04 Dec-04
A
v
e
r
a
g
e

N
o
r
t
h

P
r
e
s
s
u
r
e

(
p
s
i
)
8.13' Data 7.21' Data 5.38' Data
gage malfunction

a) North abutment 2002 b) North abutment 2004
-5
5
15
25
35
45
55
65
Jan-02 Mar-02 May-02 Jun-02 Aug-02 Oct-02 Dec-02
A
v
e
r
a
g
e

S
o
u
t
h

P
r
e
s
s
u
r
e

(
p
s
i
)
-5
5
15
25
35
45
55
65
Jan-04 Mar-04 Apr-04 Jun-04 Aug-04 Oct-04 Dec-04
A
v
e
r
a
g
e

S
o
u
t
h

P
r
e
s
s
u
r
e

(
p
s
i
)
gage malfunction

c) South abutment 2002 d) South abutment - 2004
Figure 9. Seasonal Fluctuations of Peak Soil Pressures

CONCLUSIONS

Three years of field data have been collected at the Orange-Wendell Bridge
and compared with extensive FEM of the structure. Maximum bridge thermal
displacements at the girder centerline were predicted reasonably well using
common design recommendations [4]. Displacements measured at girder
centerlines and joint meter locations vary significantly due to abutment rotations.
Additionally, these differences have varied from year to year. Differences in the
response of the north and south abutments have also been observed. Soil pressures
behind the abutments match reasonably well with FEM predictions at shallow
depths. However there was significant scatter in readings at pressure cell
locations. These variations were partly due to complex seasonal variation of
backfill pressures.

ACKNOWLEDGEMENTS

The research reported was prepared in cooperation with the Executive Office of
Transportation, Massachusetts Highway Department and the United States

173
Department of Transportation, Federal Highway Administration, through the
Massachusetts Transportation Research Program. The contents of this paper
reflect the views of the authors, who are responsible for the facts and the accuracy
of the data presented herein. The contents do not necessarily reflect the official
view or policies of the Massachusetts Highway Department or the Federal
Highway Administration.

REFERENCES

1. Bonczar, C. Brea, S. F, Civjan, S. A., DeJong, J., and Crovo, D. Integral Abutment Pile
Behavior and Design Field Data and FEM Studies. Proceedings: Integral Abutment
and Jointless Bridges Conference, Federal Highway Administration, Baltimore, MD,
March 2005.

2. GTSTRUDL, Structural Design & Analysis Software, Version 7.0, Georgia Tech
Research Corporation, Atlanta, Georgia.

3. National Cooperative Highway Research Program (NCHRP 1991). Manuals for the
Design of Bridge Foundations, R.M. Barker, J.M. Duncan, K.B. Rojiani, P.S.K. Ooi,
C.K. Tan, and S.G. Kim, eds., Report No. 343, Transportation Research Board,
Washington, D.C.

4. MassHighway Bridge Manual (1999). Design guidelines and Standard Details for
Integral Abutment Bridges, Massachusetts Highway Department, Boston, MA.

5. Ting, J.M. and Faraji, S. (1998). Streamlined Analysis and Design of Integral Abutment
Bridges. Report No. UTMC 97-12, University of Massachusetts Transportation Center,
Amherst, Massachusetts.

174
Integral Abutment Pile Behavior and Design Field Data
and FEM Studies

Christine Bonczar
a
, Sergio F Brea
b
, Scott Civjan
c
, Jason DeJong
b
, Daniel Crovo
d

a
Graduate Student, University of Massachusetts at Amherst

b
Assistant Professor, University of Massachusetts at Amherst
c
Associate Professor, University of Massachusetts at Amherst
Department of Civil and Environmental Engineering, 130 Natural Resources Rd.,
Amherst, MA 01003
d
Assistant Bridge Engineer, Massachusetts Highway Department, 10 Park Plaza,
Boston, MA 02116
ABSTRACT

Pile behavior in integral abutment bridges can be very complex. Research being
conducted at the University of Massachusetts Amherst and funded by the
Massachusetts Executive Office of Transportation and the Federal Highway
Administration is addressing several key issues. The Orange-Wendell Bridge is
used as a basis for the study. The Orange-Wendell Bridge is a three-span (270 ft.
total length) integral abutment bridge in central Massachusetts. Over 3 years of
instrumented data and one year of bi-weekly manual pile inclinometer readings
have been collected. Additional data from the time of construction is also
included. Recorded ambient temperatures have ranged from approximately 10F
(-23.3C) to 100F (37.8C). In addition, extensive non-linear finite element
modeling (FEM) has been performed to model bridge behavior. A parametric
analysis was conducted on both 3-D and 2-D non-linear FEM of the entire bridge
structure. Pile elements included non-linear geometric effects and non-linear
material properties. Non-linear spring curves representing soil effects were
included in the models at pile, pier and abutment locations. Additional models
considered only equivalent length cantilever pile sections as are often used in
design. Field data and modeling results specific to the following will be
addressed: pile deformations, moment at the top of abutment piles, properties of
abutment and pile backfill materials, effects of pile yielding and pile design
assumptions.

INTRODUCTION

The Massachusetts Highway Department, through the Massachusetts
Transportation Research Program commissioned a field study to validate and
economize the design of integral abutment bridges. This paper compares data
from field monitoring with finite element models (FEM) of a three-span integral
abutment bridge located in central Massachusetts, with particular emphasis on
pile behavior and effects of abutment backfill properties. Of primary concern
were seasonal temperature effects on the bridge. Additionally results from FEM

175
parametric studies conducted to evaluate the effects of soil and backfill properties
on the bridge response to temperature fluctuations are presented. A companion
paper [1] presents details of the bridge, instrumentation configuration, monitoring
procedure, representative data obtained during 2002 and 2004 and a description of
the basic FEM constructed for the bridge. Ambient temperatures have ranged
from 10F (-23.3C) to 100F (37.8C) over the 3 years of readings obtained
(January 2002 through December 2004). The bridge was completed in late 2000.
BRIDGE AND FEM DESCRIPTION

The integral abutment bridge described in this study crosses over Millers River on
Wendell Depot road in Orange, MA. Details of the three-span continuous bridge
as well as the basic FEM model of the bridge are described in a companion paper
by Bonczar et al. [1]. Measured pressures behind the abutment were significantly
higher at deep locations near the abutment base from those calculated using a 2-D
FEM model of the bridge. Significant seasonal variations of peak pressures were
also observed. A detailed parametric analysis on the effects of backfill and soil
properties on bridge behavior is reported in this paper. The goal was to provide
information for designers on important parameters to consider when modeling
integral abutment bridges. Modeling assumptions specific to the soil and backfill
spring modeling are defined below.

Pea Stone Modeling

As described in the companion paper [1], the top 10 feet (3 m) of all abutment
piles were driven through a 24-in. (610-mm) pre-drilled hole that was filled with
pea stone after the driving operation. This method was intended to allow
additional movement of the pile in the expected plastic hinge region. Pea stone
characteristics are not easily available, and as will be discussed later, its properties
can change significantly because of seasonal compaction. GEI Consultants, a
consultant during bridge construction, conducted confined compression tests of
the pea stone, and provided lower and upper bounds for its density equal to 92.5
to 100 lb/ft
3
(1.48 to 1.61 Mg/m
3
) and friction angles between 27 and 40 degrees
(Table 1). The tests consisted of cyclic one dimensional confined stiffness testing
of 3/8 in, sub rounded pea stone and indicated significant stiffness variation upon
reloading of the samples and previous strain history. It is noted that sub-rounded
pea stone had the least variation of all materials tested. Because test specimens
were confined, these values are not representative of the Orange-Wendell Bridge
conditions, where the pea stone is surrounded by soil.

176
Table 1. Assumed Pea Stone and Soil Backfill Properties
Pea Stone Properties
' (lb/ft
3
) (deg.)
k
1
(lb/in
3
)
Lower Bound 92.5 27 10
Upper Bound 100 40 45
Loose 110 30 45
Backfill Properties
' (lb/ft
3
) (deg.)
k
1
(lb/in
3
)
Loose 110 30 45
Dense 140 45 220

Actual conditions for the placed pea stone include soil surrounding the drilled
hole that will deform upon loading, resulting in a non-rigid compliant boundary
with a lower stiffness and a condition that may enable additional accommodation
of particles. This is expected to result in a very loose soil condition at the top of
the piles. Values assumed for the parametric study of pea stone properties are
listed in Table 1. Also listed are loose soil properties that were assumed at the pea
stone location for preliminary modeling reported by Bonczar et al. [1]. It is noted
that loose soil properties are similar to the upper bound of pea stone properties.
Therefore, upper bound pea stone properties used for the top 10 feet of all piles
resulted in similar abutment rotation, displacement and pile deformations with
temperature change as those obtained by modeling pea stone with loose soil
properties as discussed in [1].

Abutment Backfill Modeling

To evaluate the influence of soil stiffness on bridge response, soil spring
properties representing abutment backfill were calculated for dense or loose soil,
as indicated in Table 1. Soil properties measured during geotechnical borings at
the Orange-Wendell Bridge are closer to the dense soil properties listed in Table
1. This assumption also matches field data as discussed in Bonczar et al. [1].

Two node non-linear springs were used for the soil-spring elements behind the
abutment. Design curves contained in NCHRP Report 343 [2] were used to
calculate the spring stiffness for passive and active conditions of the soil behind
the abutment. The coefficient of passive earth pressure for the soil springs was
calculated using [3]:

] 1 [ 7 . 5 43 . 0
) ( 190 H
T
e K

+ = (1)
Where:
H = height of soil behind abutment

T
= maximum thermal movement

Equation 1 was used in the development of the soil force-deflection curves used
to model springs behind the abutment in the finite element models.

177

COMPARISON OF MEASURED AND CALCULATED RESULTS

The parametric FEM studies were evaluated by comparing the calculated response
of the bridge with measured field data. As mentioned previously, FEM analyses
of the bridge were initially conducted with loose and dense backfill properties.
Efforts then concentrated on more accurate modeling of soil properties along the
pile length with particular attention to the top 10 ft (3 m) where the piles are
surrounded by pea stone. Although the laboratory tests conducted on pea stone
did not reflect confinement conditions in the field, they gave insight on the
significant variation of stiffness with cycling so it was expected that compaction
properties would change in different seasons and years. Therefore, these refined
FEM analyses were evaluated for the upper and lower bound pea stone properties
and dense or loose backfill properties defined in Table 1.
Assessment of Pile Deformations

Pile deformations in the field have been measured manually using an inclinometer
from July 2003 to July 2004, though readings were not always possible due to
freezing of water seeping into the inclinometer casing. Comparisons of summer
2003 through winter 2004 field data (averaged for all 4 inclinometers) for
temperature changes between +8.6
o
C and -19.8
o
C (+15.5
o
F to -35.6
o
F) with
FEM results for similar temperature changes are shown in Figure 1. The influence
of upper and lower bound properties of pea stone (Table 1) on calculated pile
deformations for dense soil conditions are shown in Figure 1a. It can be seen that
lower bound pea stone properties provide better agreement with inclinometer data
from 2003/2004, primarily in deflected shape of piles and location of point of
inflection. The deformations are concentrated in the top 10 ft of the piles
coinciding with the depth of the section filled with pea stone. Additionally, for the
same thermal displacement of the superstructure (top of the abutment), using
lower bound pea stone properties results in higher displacements at the bottom of
the abutment than using upper bound properties of pea stone, especially during
bridge contraction. Clearly there is less restraint to pile deformation when lower
bound properties are used, which consequently affects abutment rotation as well.

Figure 1b shows the FEM pile deformation results when lower bound properties
were used for pea stone, and backfill soil properties were varied from dense to
loose. When the bridge contracts no difference in pile deformation with varying
soil properties was observed because the active pressures behind the abutment are
very similar for the two density conditions. However, there is significant variation
in pile deformations and abutment rotations for bridge expansion. It can be seen
that field data matches FEM results better when assumptions of dense backfill
properties are made, as would be expected from materials and procedures used in
construction. It was found that calculated pile deformations were influenced most

178
by pea stone properties when the bridge contracts and by backfill properties when
the bridge expands.

-15
-10
-5
0
5
10
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4
Lateral Displacement (in)
D
i
s
t
a
n
c
e

F
r
o
m

B
o
t
t
o
m

o
f

A
b
u
t
m
e
n
t

(
f
t
)
-35.6 Data
-27.6 Data
-17.7 Data
-9.1 Data
15.5 Data
-72 Up Bound
-36 Up Bound
-10 Up Bound
15 Up Bound
36 Up Bound
-72 Low Bound
-36 Low Bound
-10 Low Bound
15 Low Bound
36 Low Bound
+T
Bridge Expansion
-T
Bridge Contraction
Upper Bound -
Depth to 1st pt. of inflection
Lower Bound and Inclinometer Data-

a) Varying Pea Stone Properties-Dense Backfill b) Varying Abutment Backfill-Lower
Bound Pea Stone

Figure 1. Abutment and Pile Deformation

Figure 2 shows the yearly variation of north and south abutment rotation
compared with calculated abutment rotation using upper or lower bound pea stone
properties and dense backfill. Abutment rotation decreased during contraction
from year to year. This is an indication that the pea stone properties may have
initially been at a compaction state better captured by the upper bound values
used in the analyses, but became relatively looser over four years of bridge
service. At the south abutment rotations appear to remain relatively unchanged
from year to year. Since bridge expansion values are not significantly different
from year to year it is assumed that backfill properties have not varied
significantly.

-80
-60
-40
-20
0
20
40
60
-0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20
Abutment Rotation (deg)
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
N02 N03 N04 Upper Bound Lower Bound

a) North Data 2002/2003/2004 b) South Data 2002/2003/2004

Figure 2. Variation of Abutment Rotation during Monitoring Period

The relationship between pile top displacement and abutment rotation is presented
in Figure 3 for dense backfill and two different properties of pea stone. The
calculated curves are compared with data from the north abutment from 2002
through 2004. Calculated displacements necessary to cause yielding in the piles
are shown as horizontal lines for dense backfill and the two bounds of pea stone
-15
-10
-5
0
5
10
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4
Lateral Displacement (in)
D
i
s
t
a
n
c
e

F
r
o
m

B
o
t
t
o
m

o
f

A
b
u
t
m
e
n
t

(
f
t
)
-35.6 Data
-27.6 Data
-17.7 Data
-9.1 Data
15.5 Data
-72 Dense
-36 Dense
-10 Dense
15 Dense
36 Dense
-72 Loose
-36 Loose
-10 Loose
15 Loose
36 Loose
+T
Bridge Expansion
-T
Bridge Contraction
Dense & Loose -
-80
-60
-40
-20
0
20
40
60
-0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10 0.15 0.20
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
S02 S03 S04 Upper Bound Lower Bound

179
indicated in Table 1. It can be observed that measured data points lie mostly
within the area defined by yield displacements when lower bound properties of
pea stone are used, but exceed the yield displacement for upper bound properties.
It should be noted, however, that upper bound properties were more likely in 2002
when measured displacements were smaller. Therefore it does not appear that
significant yielding of piles has occurred. It is also observed that calculated
displacement-rotation curves exhibit a significant change in slope after their
intersection with the yield displacement lines for bridge contraction, but this
change is less apparent during expansion due to restraint by the backfill.

-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
-0.40 -0.30 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60
D
i
s
p
l
a
c
e
m
e
n
t

a
t

T
o
p

o
f

P
i
l
e

(
i
n
)
Upper Bound - Dense Backfill Lower Bound - Dense Backfill N02 N03 N04
Yield Displacement-
Upper Bound - Dense
Backfill
Yield Displacement-
Lower Bound - Dense
Backfill

Figure 3. Relationship between Pile Displacement and Abutment Rotation
Maximum Moment in Piles

The calculated moment at the pile top was minimally affected by the pea stone
properties for bridge expansion or contraction. Figure 4a shows the moment at the
pile top as a function of temperature change using upper and lower bound pea
stone properties in the FEM analysis, both assuming dense soil backfill. The
calculated yield moment in the piles occurs at DT = 70F (39C) and DT = -80F
(-44C) for the assumed upper bound pea stone properties, and at DT = 90F
(50C) and DT = -95F (-53C) for the lower bound assumption. Because of the
higher restraint provided by pea stone using upper bound properties, piles reach
the yield moment at a lower temperature change compared with using lower
bound pea stone properties. The implication of this finding is that if the pea stone
loosens over years of bridge service, a larger temperature change will be required
to generate yielding in the piles. Similar data is presented for dense and loose
backfill properties in Figure 4b combined with lower bound pea stone properties.
The calculated yield moment occurs at DT = 90F (50C) for dense backfill, at DT
= 45F (25C) for loose backfill, and at DT =-95F (53C) for dense or loose
backfill. In this case loose backfill conditions generate higher pile moments than
dense backfill for the same temperature change, but only when the bridge
expands. The maximum temperature changes at the Orange-Wendell Bridge have
been 33.6F (18.7C) and -71.4 F (39.7C) referenced to a construction

180
temperature of approximately 66F (19C), so no yielding of piles is expected
even when self weight moments are included.

Piles were instrumented with 40 strain gages at different depths, but not all
instruments remained functional after pile driving or years in service. To date
only incipient pile yielding has been inferred from strain gage data, which
supports the observations on required temperature changes necessary to produce
yielding.
-1500
-1000
-500
0
500
1000
1500
-250 -200 -150 -100 -50 0 50 100 150 200 250
Temperature Change (
o
F)
M
o
m
e
n
t

a
t

T
o
p

o
f

P
i
l
e

(
k
-
i
n
)
Upper Bound-Dense Backfill Lower Bound-Dense Backfill
Calculated Yield Moment =
709 k-in
+T T

-1500
-1000
-500
0
500
1000
1500
-250 -200 -150 -100 -50 0 50 100 150 200 250
Temperature Change (
o
F)
M
o
m
e
n
t

a
t

T
o
p

o
f

P
i
l
e

(
k
-
i
n
.
)

Lower Bound-Dense Backfill Lower Bound-Loose Backfill
Calculated Yield Moment =
709 k-in
+T T

a) Effect of Pea Stone Properties on Pile Moment b) Effect of Backfill Properties on Pile
Moment
Figure 4. Pile Moment Variation with Temperature for Different Backfill and Pea Stone
Conditions

Soil Pressure Behind Abutment

Soil pressures generated behind the abutments were also investigated
analytically using assumptions of lower and upper bound pea stone properties.
During bridge expansion, it was noted that the lower bound pea stone properties
resulted in higher pressures behind the abutment than those generated using upper
bound properties in the FEM models. Because greater deformation is induced near
the base of the abutments when loose material surrounds the top of piles, backfill
pressures are higher than those corresponding to piles with higher restraint. On
the other hand, for a given restraint condition at the top of piles, if backfill is
modeled using loose soil properties, the pressures generated behind the abutment
would be lower than those for dense soil properties because of changes in soil
spring constants. These results are shown in Figure 5a, which depicts pressures at
three different depths below the abutment top for three model parameters (upper
bound pea stone/dense backfill, lower bound pea stone/dense backfill, and upper
bound pea stone/loose backfill). Additionally, FEM results assuming lower bound
pea stone properties with dense backfill are overlaid on field data for 2002 and
2004 in Figures 5b through 5e for the north and south abutments, respectively.
Earth pressure data was averaged at each level for each respective abutment.
South and north field data remain constant from year to year except near the
bottom of the abutments where earth pressure increases with respect to
temperature change. This increase in pressure may be justified by the pea stone

181
properties becoming relatively looser from 2002 to 2004 as well as seasonal
fluctuations and local pressure concentrations as discussed in Bonczar et al. [1].

Pile Design Issues

Piles are often designed using an equivalent cantilever model [3]. In this
method the complex soil-pile interaction is simplified using a model consisting of
a fixed-end pile element with its length calculated to generate a similar maximum
moment in the pile to that obtained through more intensive calculations. The
method used by the Massachusetts Highway Department, where the top of the pile
element is fixed against rotation but free to translate, was evaluated in this
research. An imposed displacement equal to the design displacement value is
applied to the top of the pile and the induced moment is used for design. For the
Orange-Wendell Bridge, Bonczar et al. [1] have reported significant differences in
displacement at the centroid of the deck/girder elements (design value) and
displacement at the top of the pile due to abutment rotation. The amount of
rotation also affects the pile moment. The equivalent cantilever length obtained
for the soil types used at the Orange-Wendell Bridge for an average design
displacement of 0.9 in. (22.9 mm) using the Massachusetts Highway Department
Design recommendations [4] was 9.25 feet (2.8 m).

Figure 6 shows a preliminary evaluation of the equivalent cantilever method
for pile design. The design curves indicated on the plot for the equivalent length
of 9.25 ft. (2.8 m) correspond to elements with fixed rotation at both ends of the
equivalent cantilever. Application of an axial load (dead load of the structure)
results in a slight increase in moment due to second order effects.

Three plots from the 2-D FEM modeling are also presented in Figure 6. These
are referenced to displacement realized at the bottom of the abutment. FEM
results indicate a slightly higher moment than the equivalent cantilever results for
a given design displacement. This is due to abutment rotation reducing FEM
displacements at the top of piles from design displacement while the cantilever
method ignores these rotations. The effects of including dead load are more
pronounced for the FEM models because the self-weight produces a moment at
the girder-abutment connection that adds to second order effects.

182
-80
-60
-40
-20
0
20
40
60
-45 -40 -35 -30 -25 -20 -15 -10 -5 0 5
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
5.38' Lower-Dense 7.21' Lower-Dense 9.05' Lower-Dense
5.38' Upper-Dense 7.21' Upper-Dense 9.05' Upper-Dense
5.38' Upper-Loose 7.21' Upper-Loose 9.05' Upper-Loose

a) FEM Earth Pressure Results

-80
-60
-40
-20
0
20
40
60
-45 -40 -35 -30 -25 -20 -15 -10 -5 0 5
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)

b) North Abutment 2002 c) North Abutment 2004

-80
-60
-40
-20
0
20
40
60
-45 -40 -35 -30 -25 -20 -15 -10 -5 0 5
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)

d) South Abutment 2002 e) South Abutment 2004

Figure 5. Effects of Soil and Pea Stone Properties on Earth Pressures Behind Abutments

Figure 7 compares results from the equivalent cantilever method and
displacements at the girder centerline calculated using 2-D FEM analyses. Note
that the equivalent cantilever results are identical to those presented in Figure 6 as
this model neglects any abutment rotation. The cantilever model matches the
FEM results that include self-weight for bridge expansion very well, and is
conservative when bridge contraction is considered. Results indicate that piles of
the Orange-Wendell Bridge would be below yield when the design expansion
displacement is reached, and would be near their plastic moment capacity for the
design contraction displacement. These extreme displacements have not been
-80
-60
-40
-20
0
20
40
60
-45 -40 -35 -30 -25 -20 -15 -10 -5 0 5
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)
-80
-60
-40
-20
0
20
40
60
-45 -40 -35 -30 -25 -20 -15 -10 -5 0 5
T
e
m
p
e
r
a
t
u
r
e

C
h
a
n
g
e

(
o
F
)

183
measured at the site yet, so pile yielding is expected. It is seen that the equivalent
cantilever method included in the Massachusetts Highway Department Design
Manual [4] provides an excellent correlation to pile moments for bridge
expansion and is conservative for bridge contraction (Figure 7).

-1200
-900
-600
-300
0
300
600
900
1200
-4 -3 -2 -1 0 1 2 3 4
Displacement at Pile Top (in)
M
o
m
e
n
t

a
t

P
i
l
e

T
o
p

(
k
i
p
-
i
n
)

FEM w/out self wt. Le fix-fix w/out axial load FEM
Le fix-fix FEM no rotn. w/out self wt.
T(+)
T(-)
Calculated yield moment
= 709 k-in

Figure 6. Pile Moment Versus Temperature Induced Displacements at Pile Top Evaluation
of Equivalent Cantilever Method

-1200
-900
-600
-300
0
300
600
900
1200
-4 -3 -2 -1 0 1 2 3 4
Displacement at Centerline of Girder (in)
M
o
m
e
n
t

a
t

P
i
l
e

T
o
p

(
k
i
p
-
i
n
)

FEM w/out self wt. Le fix-fix w/out axial load FEM Le fix-fix
T(+)
T(-)
Calculated yield moment
= 709 k-in

Figure 7. Pile Moment Versus Temperature Induced Displacements at Centerline of Girder
Evaluation of Equivalent Cantilever Method

CONCLUSIONS

Over three years of field data acquired automatically and one year of manual
inclinometer readings have been collected at the Orange-Wendell Bridge.
Extensive FEM analyses have also been performed for the structure. This paper
presented the effects of pea gravel and backfill soil properties on the bridge
behavior.

Pile deformations are influenced predominantly by pea stone properties during
bridge contraction and backfill properties during bridge expansion. It was found
that models assuming dense backfill and very loose pea stone properties

184
adequately modeled the soil properties in 2004 while 2002 data indicated the pea
stone may have originally been near the assumed upper bound properties. Pile
moments are generally below the yield moment of the piles, indicating that little
or no yielding should be expected in the bridge. This finding correlates with
measured field data. Maximum moments in the piles were also affected by pea
stone and backfill properties. Moments are higher for upper bound pea stone and
loose backfill. Pressures behind the bottom of the abutment are increased when
looser properties are assumed for pea stone, and are increased significantly as the
backfill properties become denser. There is significant scatter in field data for
abutment backfill pressures. The equivalent cantilever method currently used in
the Massachusetts Highway Department Bridge Design Manual [4] provided a
reasonable approximation to moments in piles for the Orange-Wendell Bridge.

ACKNOWLEDGEMENTS

The research reported was prepared in cooperation with the Executive Office of
Transportation, Massachusetts Highway Department and the United States
Department of Transportation, Federal Highway Administration, through the
Massachusetts Transportation Research Program. The contents of this paper
reflect the views of the authors, who are responsible for the facts and the accuracy
of the data presented herein. The contents do not necessarily reflect the official
view or policies of the Massachusetts Highway Department or the Federal
Highway Administration.

REFERENCES
1. Bonczar, C., Brea, S. F., Civjan, S. A., DeJong, J., Crellin, B. and Crovo, D. Field Data
and FEM Modeling of the Orange-Wendell Bridge. Proceedings: Integral Abutment and
Jointless Bridges Conference, Federal Highway Administration, Baltimore, MD, March
2005.

2. National Cooperative Highway Research Program (NCHRP 1991). Manuals for the
Design of Bridge Foundations, R. M. Barker, J. M. Duncan, K. B. Rojiani, P. S. K. Ooi,
C. K. Tan, and S. G. Kim editors, Report No. 343, Transportation Research Board,
Washington, D.C.

3. Abendroth, R.E. and Greimann, L.F. (1989). Rational Design Approach for Integral
Abutment Bridge Piles. Journal of the Transportation Research Board, Transportation
Research Record 1223, pp. 12-23.

4. MassHighway Bridge Manual (1999). Design Guidelines and Standard Details for
Integral Abutment Bridges, Massachusetts Highway Department, Boston.

185
Effects of Restraint Moments in Integral Abutment
Bridges
By M. Arockiasamy
1
and M. Sivakumar
2
1
ACI member M. Arockiasamy, is Professor and Director of Center for
Infrastructure and Constructed Facilities in the Department of Civil Engineering
at Florida Atlantic University, Boca Raton, FL. His research interest includes
concrete durability, bridges, advanced polymer matrix composite materials, and
infrastructure systems.
2
ACI student member M. Sivakumar, has earned his Ph.D., at Florida Atlantic
University, and is a Design Engineer, PTE Strand Co Inc., Miami, FL. His
research interest includes creep, shrinkage and thermal induced stresses in
concrete structures.

ABSTRACT

The integral abutment bridge concept has received considerable interest
among bridge engineers owing to their enormous benefits due to elimination of
expensive joints and reduced installation and maintenance costs. The
superstructure of integral abutment bridges is made continuous through a
composite cast-in-place concrete deck slab over prestressed concrete or steel
girders and continuity diaphragms. The girders are often rigidly connected with
the abutments. Restraint moments develop in the superstructure due to the
continuity and time-dependent creep, shrinkage and thermal effects. This makes
the design of integral abutment bridges different from other conventional bridges.
There is a need to develop an analytical procedure to determine the secondary
effects induced due to creep, shrinkage and temperature.

The objective of this paper is to present the state-of-the-art on restraint
moments in continuous precast prestressed girder bridges and its relevance to
integral abutment bridges. Approximate and more rigorous methods are available
to determine the effects of creep and shrinkage. The PCA method reports the
influence of creep of the precast girders, and of the differential shrinkage between
the precast girders and the situ-cast deck slab on continuity behavior. The CTL
approach was based on a series of computer simulations to study the effects of
variation in time-dependent material behavior and variation in bridge design
parameters on the resultant service moments in the bridge girders. A modified
restraint-moment calculation method (P-method) was developed for full-span
prestressed concrete form panels that accounts for the length and stiffness of the
diaphragm, the different initiation time for creep and the restraint of cast-in-place
(CIP) concrete shrinkage. A flexibility based analytical tool was presented to
predict time-dependent restraint moments and the effectiveness of the continuity
connection under service live loads. Non-linear time-dependent analysis was
reported to examine the design considerations for integral abutment bridges.

186
BACKGROUND

A deck joint is one source of problems for the bridge structure causing debris
to accumulate and solidify over a period of time preventing free expansion of the
deck, which could create excessive stresses at the joints. The joints thus need
constant periodical maintenance. The cost associated with the long-term
maintenance of jointed bridges can be reduced by eliminating the joints in the
bridge and creating a continuous jointless structure. The integral abutment bridge
concept has received considerable interest among bridge engineers owing to their
enormous benefits due to elimination of expensive joints, the speed and simplicity
in the construction, reduced installation and maintenance costs. The
superstructure of integral abutment bridges does not have joints either at abutment
or at intermediate supports and is cast integrally with joints either at abutment or
at intermediate supports and is cast integrally with abutments that are generally
supported by a single row of piles. The superstructure of integral abutment
bridges is made continuous through a composite cast-in-place concrete deck slab
over prestressed concrete or steel girders and continuity diaphragms. The girders
are often rigidly connected with the abutments. Restraint moments develop in the
superstructure due to the continuity and time-dependent creep, shrinkage and
thermal effects. Unlike jointed bridges, the design and construction of jointless
integral abutment bridges require some additional considerations, for example, the
time-dependent restraint moment at the integral abutment and intermediate piers.
This makes the design of integral abutment bridges different from other
conventional bridges.

The continuity of the girders in the integral abutment bridge can be achieved
by providing continuous reinforcement in the deck over the piers and a concrete
diaphragm between the ends of the girders at interior supports and through stub
type abutment at the ends. Many different methods are available to achieve
continuity of the superstructure (Ma et al., 1998) and several options available for
superstructure - abutment continuity (Burke, 1993). The girders act as simple span
members for its own self weight, before the continuity connection is cast. Once
the continuity diaphragm and deck and the abutments are cast, the composite
girder/deck section will carry live loads and superimposed dead loads as a
continuous composite structure.

A multi-span bridge with a number of precast prestressed concrete girders
with simple supports over the piers undergoes an increase in camber over time
due to creep under a sustained prestressing force and other effects. No restraint
moments are induced over the supports. However, when the girder ends are
restrained by a pier diaphragm or abutment as in the case of jointless integral
abutment bridges, then the bridge becomes a statically indeterminate structural
system, allowing time-dependent secondary forces to develop due to creep,
shrinkage and temperature effects. Time-dependent effects induce restraint
moments in the continuity connections. These restraint moments can produce
stresses that are higher than those due to gravity loading alone.

187

The continuity connection is subject to negative and positive moments and
reinforcement is provided in the deck over the support to resist the negative
moment. Positive moment reinforcement extends from the ends of the girder into
the diaphragm (Fig 1).

A good understanding of the behavior of integral abutment bridges under the
time-dependent restraint moment is essential in order to appropriately design the
continuity connection reinforcement.

Figure 1 a. Typical continuity connection details in precast prestressed concrete girder
integral abutment bridge

Figure 1 b. Typical continuity connection details at the integral abutment

188

OBJECTIVE

The objective of this paper is to present the state-of-the-art on restraint
moments in continuous precast prestressed girder bridges and its relevance to the
integral abutment bridges.

EFFECT OF CREEP AND SHRINKAGE ON CONTINUITY

Creep and shrinkage effects are important parameters in integral abutment
bridges. Creep significantly affects the structural behavior at service loads. Creep
causes losses of prestress and increase in deformations and deflections, which
may impair the serviceability of the bridge.

The precast prestressed girders are sufficiently old before they are installed in
place in the field. Most of the concrete drying shrinkage in the precast member
will have occurred before the cast-in-place (CIP) concrete is cast. The shrinkage
of the new concrete in the bridge deck exceeds the remaining shrinkage in the
precast girder. This differential shrinkage creates shear forces between relatively
old precast prestressed concrete girder and new cast-in-place concrete deck slab,
thus producing axial forces and bending moments.

The girder concrete creeps under the prestressing force causing an increase in
upward camber in the girder (Fig 2 a). Under the combined effects of creep and
shrinkage of concrete and relaxation of prestressing steel, the camber in the
prestressed concrete girders gradually changes with time depending on the type of
structure and construction methods. The upward deflection is counteracted due to
differential shrinkage and to some extent by creep due to the dead load of the
girder and dead load of the deck slab and other additional superimposed dead
loads at later stages of construction (Fig 2 b).

Figure 2 a. Creep deformation in simple span composite construction due to prestressing
force

189

Figure. 2 b. Deformation in simple span composite construction due to differential
shrinkage, sustained dead load of the girder and deck slab

In a composite simple span bridge, the time-dependent effects result in
changes of stresses in the composite section in addition to the loss of prestress and
changes in camber. For a simple span construction, where member ends are free
to rotate, the restraining force due to the differential shrinkage between the
precast girder and the cast-in-deck will cause the composite section to deform.
In a jointless multi-span bridge, the structure becomes continuous after the
deck is cured. The rotations at the girder ends are restrained and any time-
dependent deformations that occur thereafter will induce forces and restraint
moments in the girders over the supports (Fig.3). Positive restraint moments will
develop over the supports due to the creep in concrete resulting from the
prestressing force and the continuity of the multi-span integral abutment bridge
may be reduced by the effects of creep in the girders.

Axial shortening and rotation at the girder ends will induce stresses in the
continuous deck. The result is a downward deflection of the composite deck-
girder system and a negative restraint moment that helps to maintain continuity by
offsetting positive restraint moment caused by creep effects.

RESTRAINT MOMENTS

Approximate and more rigorous methods are available to determine the effects
of creep and shrinkage. A brief overview of the published literature is presented
on the continuity of precast, prestressed concrete girders with cast-in-place decks
and diaphragms.

PCA Approach

The influence of creep and shrinkage of the concrete on the long-term
behavior of composite continuous construction was investigated in the research
and development laboratories of Portland Cement Association (PCA). The study
was based on an extensive laboratory testing of two-span continuous precast
prestressed I shaped girders with a situ-cast deck slab (Mattock, 1961). The
experimental study was restricted to two half-scale continuous bridge girders over
a period of approximately two years from the time of construction. One girder
incorporated a positive moment connection at the interior support while the other
did not. The uniform differential shrinkage moment, M
s
in a composite concrete
section at any time is given by

190
|
.
|
\
|
+ =
2
'
2
t
e A E M
B B s s
(1)
where
s
= differential shrinkage strain,
E
B
= elastic modulus for deck slab concrete
A
B
= cross-sectional area of deck slab
|
.
|
\
|
+
2
'
2
t
e = distance between mid-depth of slab and centroid of composite
section

Figure 3 a. Creep deformation in continuous composite jointless construction due to
prestressing force

Figure 3 b. Deformation in continuous composite jointless construction due to differential
shrinkage, sustained dead load of the girder and deck slab

The restraint moment, M
r
at the center support of a two span continuous
bridge is calculated as
( )
|
|
.
|
\
|
|
.
|
\
|
=
e
M e M M M
s d p r
1
2
3
1
2
3
(2)
where
M
p
= moment caused by prestressing force about centroid of composite
member,
M
d
= mid-span moment due to dead load,
e = base of Naperian logarithm (2.7183)
= creep co-efficient = ratio of creep strain to elastic strain at time of
investigation.

191

The total restraint moment at the pier is equal to the sum of the three
components of shrinkage, dead load creep and creep due to prestress. It was
concluded that the deformations due to creep and differential shrinkage did not
influence the ultimate load carrying capacity of the continuous girders and the
influence of creep and shrinkage is restricted to deformations and the possibility
of cracking at the service load level.

CTL Approach

Another approach for the analysis of precast prestressed girders made
continuous was developed by Oesterle, Glikin and Larson (1989) at Construction
Technology Laboratories Inc., (CTL) as a part of National Cooperative Highway
Research Program (NCHRP). This approach takes into account the stiffness and
the length of connection at the continuous supports. The computer program
BRIDGERM was developed to calculate the time-dependent restraint moments in
precast prestressed multi-span bridges. The analysis in BRIDGERM was carried-
out by superimposing incremental restraint moments over a series of time-
intervals. The total restraint moment at the pier is equal to the sum of the three
components of shrinkage, dead load creep and creep due to prestress.

The restraining moment at a series of discrete time, T is given by
( )( )
|
|
.
|
\
|
+ + =
T
ST PT D T
T
T
e
M e M M M
1
' 1 ' ' (3)
where,
M
D
= restraint moment due to dead load
M
PT
= restraint moment due to creep under prestress
T
= creep coefficient at time T

It was reported that continuity performance is highly dependent on the age of
the girder when the diaphragm and deck are cast. With continuity established at
an early girder age, positive restraint moment develops, dependent on the amount
of positive reinforcement provided in the girder connection at the supports.

McDonagh and Hinkley (2003) developed RMCalc, using the same algorithm
as BRIDGERM to compute the restraint moments. This software program
(RMCalc), in conjunction with ACI 209R-92, provides a simple means of
computing the restraint moments.

P-Method

Peterman and Ramirez (1998) proposed a modified restraint-moment
calculation method for full-span prestressed concrete form panels that accounts
for the length and stiffness of the diaphragm, the different initiation time for creep

192
and the restraint of cast-in-place (CIP) concrete shrinkage. The differential
shrinkage moment is estimated by multiplying using the following factors to
account for restraint resulting from the precast members and the reinforcement in
the CIP slab:
|
|
|
|
.
|
\
|
+
|
|
|
|
|
.
|
\
|
+
d d
s s
d d
p p
A E
A E
A E
A E
1
1
1
1
(4)
where
E
p
= modulus of elasticity of precast panels
A
p
= area of precast panels
E
s
= modulus of elasticity of steel reinforcement in CIP deck
A
s
= area of steel reinforcement in CIP deck
E
d
= modulus of elasticity of CIP deck
A
d
= area of CIP deck

The restraint moment at the center pier of a two span symmetric bridge is
calculated as
( ) ( ) | | ( ) ( )
( )
|
|
.
|
\
|
|
.
|
\
|
=

2
2
2 1
1
2
3
1 1
2
3

e
M e M e M M M
s CIP d precast d p r

where
=
m
m
d
d
d
d
L
I
L
I
L
I
3 2
2
+

I
d
= moment of inertia of diaphragm region (area between support points at
center pier)
L
d
= length of diaphragm region
I
m
= moment of inertia of main spans
L
m
= length of main spans
M
p
= moment caused by prestressing force about the centroid of composite
member
M
s
= differential shrinkage moment adjusted for restraint of precast panels and
steel reinforcement
( )
precast d
M = mid-span moment due to dead load of precast panels
( )
CIP d
M = mid-span moment due to dead load of CIP topping
1
= creep coefficient for creep effects initiating when prestress force is
transferred to the precast panels
(5)

193
2
= creep coefficient for creep effects initiating when CIP topping is cast
( ) | |
1
1

e = change in expression ( )
1
1

e occurring from time CIP topping
is cast to time corresponding to restraint moment calculation

It was suggested that the calculation of restraint moment at interior piers due
to time dependent effects should include provisions for possible cracking of the
CIP concrete and the restraint of CIP concrete shrinkage due to the precast panels
and steel reinforcement. The design moments at service loads should include the
calculated moments due to the restraint of the time-dependent effects.

Non-linear Continuity Analysis

Mirmiran et al., (2001) carried out an analytical study to determine the
performance of continuity connections for precast, prestressed concrete girders
with cast-in-place decks. A flexibility based analytical tool was presented to
predict time-dependent restraint moments and the effectiveness of the continuity
connection under service live loads. It is recommended that a minimum amount of
positive moment reinforcement should be provided to avoid a significant loss of
continuity and to control cracking of the diaphragm under service loads.

Non-linear Time-dependent Analysis

An analytical study was carried-out to examine the design considerations for
integral abutment bridges (Arockiasamy and Sivakumar, 2003). The analysis
includes nonlinearity due to cracking of the concrete, as well as the time
dependent deformations of composite superstructure due to creep, shrinkage and
temperature to predict instantaneous linear and non-linear time dependent long-
term behavior due to creep and shrinkage effects. The effect of a fully continuous
deck slab is incorporated in the model. The redundant forces in integral abutment
bridges are derived considering the time-dependent effects of creep and
shrinkage.

Under sustained stress, the strain increases with time due to creep and the total
strain
c
including the instantaneous and creep strains at any time t is

( )
( )
( ) | |
0
0
0
, 1 t t
t E
t
c
c
c
+ = (6)

where ( )
0
, t t is the ratio of creep strain to instantaneous strain and a function
of the age at loading, t
0
and the age t for which the strain is calculated,
c
(t
0
) is the
sustained concrete stress and E
c
(t
0
) is the modulus of elasticity of concrete at age
t
0
. This linear relationship, which is true within the range of stresses under
sustained loads, allows superposition of the strain due to changes in stresses and

194
shrinkage. Thus, the total strain in concrete due to applied stress and shrinkage is
given by

( ) ( )
( )
( )
( )
( )
( )
( ) ( )
0
0
0
0
0
,
, 1 , 1
t t d
E
t
t E
t t
t t
cs c
t
c c
c c
c

+
+
+
+
=
(7)

where,
cs
(t,t
0
) is the free shrinkage strain occurring between ages t
0
and t,
( ) t
c
is the gradual increment in concrete stress introduced during the period t
0
to t. A reduced creep coefficient can be used to calculate the creep strain, if the
stress is applied gradually. With this simplification, the integral equation can be
eliminated and Eqn. (7) can be modified as

( ) ( )
( )
( )
( )
( )
( )
( )
0
0
0
0
0
0
,
, 1 , 1
t t
t E
t t
t
t E
t t
t t
cs
c
c
c
c c
+
+
+
+
= (8)

where is the aging coefficient.

For any applied moment M
i
and axial force N
i
, the instantaneous strain
distribution
oi
and curvature
i
are obtained as

( )
)
`
=
)
`
i
i
c i
oi
M
N
A B
B I
B AI E
2
1
(9)

where A, B and I are respectively the area, first and second moment of area of
the transformed cross section about a reference point O .The transformed
sectional property is calculated based on the modulus of elasticity of concrete E
c
at the time of application of the loads. The increments of top fiber strain
o
and
curvature produced by the axial force N and moment M, gradually applied
about the top reference level, may be obtained as,

( ) )
`
=
)
`
M
N
A B
B I
B I A E e e
e e
e e e e
o
2
1
(10)

where
e
A ,
e
B and
e
I are respectively the area, first and second moment of
area of the age-adjusted transformed cross section about a reference point O .The
transformed sectional property is calculated based on the age-adjusted modulus of
elasticity of concrete
e
E = E
c
/ (1+) at the time of evaluation of strain and
curvature. Since the sectional properties vary with time due to the change in the
modulus of elasticity of concrete with respect to time, the center of gravity of the
cross-section will also vary. In order to avoid numerical calculations of the

195
geometric centroid at each time step, the reference point O may conveniently be
chosen at the top fiber of the deck slab.

The restraining forces are calculated as the sum of the four terms given by
eqn. (11)
gradient temp relaxation shrinkage creep
M
N
M
N
M
N
M
N
M
N
.
)
`
+
)
`
+
)
`
+
)
`
=
)
`
(11)

NUMERICAL ILLUSTRATION

A numerical illustration is presented in the following, the details of which are
reported in the reference (Sivakumar, 2004). The analysis takes into account
different construction stages (Table 1). A typical two-span continuous integral
abutment bridge with equal spans of 115 ft. (35 m) with prestressed concrete
composite deck was analyzed for illustration. The AASHTO LRFD (1994) Bridge
Specifications for live load were used in the simulations. Additional
superimposed dead load due to the self-weight of the slab, diaphragm, overlay,
barriers etc., was determined as 1.62 kip/ft. (23.68 kN/m). Fig. 4 shows the
prestressed concrete girder composite cross-section used in the analysis. The
properties of concrete, prestressing steel and the details of overlay and diaphragm
are given in Table 2.

Table.1 Assumed erection schedule

Figure 4. Prestressed concrete girder composite cross-section

196
Table 2 Material properties
Description Properties
Girder
Concrete strength at transfer of prestress
(3days)
3480 psi
Concrete strength fc 8860 psi
Slab
Concrete strength fc
3350 psi
prestressed and non-prestressed steel
Modulus of elasticity
29000 ksi
initial prestressing force 920 kips
Diaphragms
size
spacing

1x 4
at 1/3 points of the
span
bituminous overlay, thickness 3

Time-dependent change in the bending moment along the span for various
stages of construction is shown in Fig. 5. It can be seen that the moment at the
mid-span due to the time-dependent effects, tends to approach the value of a
simply supported girder. The deck slab and the abutment attain sufficient strength
during the period 61 to 63 days and thus contribute to the composite action of the
system. The bending moment variation along the span changes from the simply
supported condition at 61
st
day to a continuous system at 63
rd
day.

Figure 5 Time-dependent change in bending moments along the span

197

Figure 6 Time-dependent change in restraint moments at the interior support based on
BRIDGERM (Oesterle, 1989)

The restraint moments shown in Fig 6 are based on the computer program
BRIDGERM, which follows the PCA method with certain modifications. The
analysis uses a simplified model that considers the finite length of the support
regions. For each time-step, the three components of the restraint moments
(differential shrinkage, creep due to dead loads, and creep due to prestress) are
calculated using the rate of creep method. The time-dependent material properties
for concrete are based on ACI-209 recommendations.

DISCUSSION

Examination of the computed time-dependent restraint moment values in the
present study shows redistribution over a period of time (Fig. 5). The
instantaneous moments along the span due to the transfer of prestress at 3 days
marginally increase up to 60
th
day due to the creep effect resulting from the
prestress. The additional superimposed dead load due to the casting of the deck
slab at the 61
st
day induces additional instantaneous moments. With the curing of
the deck slab, the bridge system becomes continuous on the 63
rd

day resulting in
increase in the restraint moments with a corresponding decrease in the midspan
positive moments. The restraint moments at the supports can be seen to decrease
(from 63
rd

day to 180 days) due to the effect of differential shrinkage and the
creep resulting from the dead load of girder, deck slab and superimposed dead
loads. During the same period the positive moment at midspan shows a
corresponding increase which can be attributed to the continuity and the time-
dependent effects. However, the application of live load on the 181
st
day results in
increases in the restraint moments as well as the midspan positive moments.

198
ACKNOWLEDGEMENT

The authors wish to express sincere thanks to Florida Department of
Transportation (FDOT) for the financial support of the study presented in this
paper (research project: Design Consideration for Integral Abutment Bridges in
Florida, Contract No. BC342), Principal investigator: Dr M. Arockiasamy, Project
Manager: Marc Ansley). They wish to express their appreciation to Dr P.
Scarlatos, Professor and Interim-Chair, Department of Civil Engineering, and Dr
Karl K. Stevens, Dean, College of Engineering, Florida Atlantic University for
their continued interest and encouragement.

REFERENCES

ACI Committee 209 (1992) report on Prediction of Creep, Shrinkage, and Temperature Effects
in Concrete Structures, American Concrete Institute, Detroit, MI.
AASHTO LRFD Bridge Design Specification, (1998) American Association of State Highway
and Transportation Officials, Washington, DC.
Arockiasamy, M., and Sivakumar, M., (2003), Design Considerations for Integral Abutment
Bridges in Florida, Final Report, FDOT Contract No. BC-342, Florida Atlantic
University, Boca Raton, FL.
Burke, M.P.Jr.,(1993) The Design of Integral Concrete Bridges, Concrete International, June, pp
37-42.
Ma, Z., Huo, X., Tadros, M.K., Baishya, M., (1998), Restraint Moments in Precast/Prestressed
Concrete Continuous Bridges, PCI Journal, Nov.-Dec., pp 40-57.
Mattock, A. H., (1961), Precast-Prestressed Concrete Bridges 5. Creep and Shrinkage Studies,
Journal of the PCA research and Development Laboratories, May, pp 32-64.
Mc Donagh, M.D., and Hinkley,K. B., (2003) Resolving Restraint Moments: Designing for
Continuity in Prestressed Concrete Girder Bridges, PCI journal, July-August, pp 104-
119.
Mirmiran A., Kulkarni S., Castrdale R., Miller R., and Hastak M., (2001), Nonlinear Continuity
Analysis of Precast, Prestressed Concrete Girders with Cast-in-place Decks and
Diaphragms, PCI journal, Sep-Oct , pp 60-80.
Oesterle, R. G., Glikin, J.D., and Larson, S.C., (1989), Design of Precast Prestressed Bridge
Girders Made Continuous, NCHRP Report 322, November, pp 1-97.
Peterman R. J., and Ramirez J. A., (1998), Restraint Moments in Bridges with Full-Span
Prestressed Concrete Form Panels, PCI journal, January-February, pp54-73.
Sivakumar, M., (2004), Creep and Shrinkage Effects on Integral Abutment Bridges, Ph.D.
Thesis, Florida Atlantic University, Boca Raton, FL.

199
Full-Scale Testing of an Integral Abutment Bridge

By Sophia Hassiotis
*
Jose A. Lopez** and Ricardo Bermudez***
*Associate Professor, Department of Civil, Environmental and Ocean
Engineering, Stevens Institute of Technology, Hoboken, N.J., USA
**NJDOT Division of Design Services, Trenton, N.J. 08625
***Sensing Systems Corp., New Bedford, Massachusetts 02745

ABSTRACT

An integral-abutment bridge is designed to transfer the temperature and
traffic-induced horizontal loading to its foundation. The mechanism eliminates
bearings, which have been a source of expensive rehabilitation. Although integral
abutments have been used successfully by many states, a nationally accepted
design methodology does not exist for their design and construction. Instead, each
highway department depends on the experience of its engineers to push the design
envelope.

The New Jersey Department of Transportation is in the process of revising its
design specifications on integral bridges, and to this end, it is funding an
extensive testing program to monitor the Scotch Road Integral Abutment Bridge.
It is our intention in this paper to share some of the data that we have been
gathering for the past two years. In addition we are discussing some of the most
relevant conclusions.

Specifically, we are gathering data every two hours using 1) thermocouples to
monitor temperature on the superstructure, 2) stain gages to measure bending
moments on the piles supporting the abutment, and axial stresses developed in the
girders and 3) pressure gages to measure the soil pressure variation behind the
abutment and in the MSE wall. We are also monitoring the displacement, rotation
and internal strains of the abutment. We have been witnessing an excellent
correlation between the temperature and displacement, however we have been
measuring a steady build up of soil pressure behind the abutment.

EXPERIMENTAL SET-UP

Scotch Road is an urban, 4-lane highway crossing Interstate 95 in the vicinity
of Trenton, New Jersey. The existing highway bridge on Scotch Road was
replaced with a new, wider structure. The new bridge is a 2-span continuous
HPS485W (HPS70W) steel- girder structure supported on a conventional pier
with fixed bearings and integral abutments.

The integral abutment bridge was designed in accordance with several design
codes, including the AASHTO [1] for structural components, following the LRFD

200
method, and AASHTO [2] for the design of the pile foundations and MSE walls,
following the Service Load Design Method. The design method prescribed by
Wasserman [3] was closely followed and slightly modified to reflect NJDOT
integral abutment design requirements.

The bridge is 90.9 m (298 ft.) long, built of steel plate girders spaced at 3.35
m (11 ft.) on center across a width of 31.8 m (104.3 ft.) A multi-column bent
supported on spread footings comprises the pier. The structure has a skew of
about 15
o
measured from the centerline of the bearing to the centerline of bridge.

The abutments average 3.34 m (11 ft.) high and 900 mm (3 ft.) thick. Each of
the abutments is supported on a single row of 19 HP14x102 (HP360x152) piles,
oriented for weak-axis bending. The piles are approximately 11.75 m (38.5 ft.)
long. A 600 mm (2 ft.) diameter corrugated steel sleeve was placed around each
pile and subsequently backfilled with granular material to increase the flexibility
of the system. Compacted crushed stone was used as a backfilling material
between the piles and the steel mesh tying the components of the MSE wall. The
soil behind the abutment and below the approach slab is well-compacted I-9
porous fill.

The bridge was instrumented during construction and has been monitored
continually for the past two years. Figures 1 and 2 show the location and type of
measurements. The following types of measuring devices were used: (1) strain
gauges along the depth of the piles (as well as inside the abutment mass), (2) soil
pressure cells for measuring the horizontal soil pressure behind the abutment, on
the galvanized sleeves surrounding the piles, and at two elevations on the MSE
wall, (3) inclinometers for measuring the rotations at the connection between the
abutment and the stringers, (4) round displacement transducers connected to four
strain gages for measuring the longitudinal displacement at the relief slab and (5)
thermocouples to monitor the temperature of the concrete slab and the steel
girders.

DATA GATHERING AND ANALYSIS

An extended literature review of integral abutment bridges can be found in
Roman [4]. Numerical evaluation of the Scotch Road Bridge can be found in
Khodair [5]. In this paper we intend to present the latest set of data and
summarize some of the conclusions that can be reached.

Figure 3 is a plot of the temperature variation with time. During the summer,
the top of the deck reached up to 20
o
F higher than ambient. In the mean time, the
highest and lowest temperatures in the girders were measured at the bottom and
were up to 10
o
F higher than ambient. During the winter, the deck and the top and
bottom of the girders were approximately at ambient temperatures.

201
Figure 4 is a plot of the longitudinal displacement of the bridge. Analysis
showed that the bridge displacement is a linear function of the temperature as
expected and used in the design. Apparently, the stiffness of the integral abutment
does not constrain the thermal extension of the bridge during working
temperatures.
Figure 5 is a plot of the rotation of the bridge. Rotation is not taken into
consideration in the design of the substructure, but it can be considerable
especially in the summer months that see a rotation almost double the one
measured in the winter. This is due to the fact that the temperature variation along
the depth of the girder, which is the main cause of the rotation, is negligible
during the winter.

Figure 6 is a plot of the axial stresses developed at the top of the piles due to
the displacement and rotation of the abutment. In general we found that the
expected values are below the measured values. This is due to the fact that the
abutment is flexible and does not transmit all of the displacement to the piles.

Figure 7 shows the variation of the pressure behind the abutment. Figure 8
shows the variation of pressure for the second stage of construction. Figure 9 is a
3D presentation of the data where pressure is plotted against displacement and
time. Figure 10 shows the variation of pressure along the abutment compared to
Coulomb active, at-rest and passive pressures. In general, we see a build-up of
pressure behind the abutment. A steady build-up can be explained by
plastification and flow of granular material during cyclic loading, which is known
in the literature as strain ratcheting. In addition, we see jumps in pressure after
the winter months. We can explain these reading by two possible mechanisms:
1) during the spring months, as the temperatures start to increase, the bridge is
pushing into frozen or semi-frozen ground and is meeting with high resistance and
2) during the low temperatures, soil moves in behind the bridge to fill the gap left
by the shrinking structure. Further research is needed to isolate which of the
above-mentioned reactions is responsible for the pressure build-up and to
conclude on the possible ramifications on the integral abutment design.

ACKNOWLEDGEMENTS

The research is sponsored by the New Jersey Department of Transportation.

REFERENCES

1. AASHTO 1998. LRFD Bridge Design Specifications. 2
od
Edition, Washington, D.C.
2. AASHTO 1996. Standard Specifications for Highway Bridges, Washington, D.C.
3. Wassermann, E.P. 2001. Design of Integral Abutments for Jointless Bridges, Structure
Magazine, May 2001, pp. 24-33
4. Roman, E.K. 2004. Evaluation of Integral Abutments, Masters Thesis, Stevens Institute of
Technology, Hoboken, N.J.
5. Khodair, Y.A. 2004. Numerical and Experimental Analysis of Integral Abutment Bridge,
Ph.D. Dissertation, Stevens Institute of Technology, Hoboken, N.J.

202

Figure 1. Plan View of Instrumentation

203

Figure 2. Side-View of Instrumentation

Figure 3. Temperature on Top and Bottom of Stringer 2.

204

Figure 4. Longitudinal Displacement at the Relief Slab.

205

Figure 5. Rotation of the Stringer at the Connection with Abutment.

206

Figure 6. Axial Stress at Pile Top.

207
Figure 7. Soil Pressure along the Abutment Wall. Stage I Construction.

208

Figure 8. Soil Pressure along the Abutment Wall. Stage II Construction.

209

Figure 9. Soil Pressure as a function of Displacement and Time.

210

Figure 10. Soil Pressure Along the Abutment Compared with Coulomb.

211
Analysis and Design of Integral Abutment by LRFD
Method

Yong Deng
1
, Joe Farre
2
, Jaime Chang
3
and Percy Penafiel
4

1
Ph.D, PE, Senior Bridge Engineer, PBS&J, 2270 Corporate Circle, Suite 100,
Henderson, Las Vegas, NV 89074
2
PE, Senior Program Manager, PBS&J, 2270 Corporate Circle, Suite 100,
3
PE, Senior Project Manager, PBS&J, 2270 Corporate Circle, Suite 100,
4
EI, PBS&J, 2270 Corporate Circle, Suite 100, Henderson, Las Vegas, NV 89074

ABSTRACT

There are no standard methods and specifications for the analysis and
design of integral abutment. This paper was prepared in accordance with the
analysis and design method developed for Long Lake Outlet Bridge that is part of
a Federal Highway Administration (FHWA) project in Wyoming. This paper
provides an analysis method for analyzing integral abutments. The subject bridge
length (End-to-End) is 19.25 meters and consists of typical pre-tensioned side-by-
side 686 mm x 1220 mm pre-cast box girder with integral abutments founded on
micropiles. The abutments were designed in accordance with AASHTO LRFD
Specifications and micropiles were designed by the ASD method. Micropile
design is beyond this scope of this paper and therefore is not included in this
paper.

Construction procedures were considered in the analysis and design of the
bridge. First, the dead loads from the typical pre-tensioned girders and
superimposed dead loads, girder and deck slab were applied as vertical loads with
separate element actions. Considering that some strength has already been
obtained when the barrier rail is constructed, for conservative purpose, the barrier
rail load was applied with full deck slab strength and composite section to obtain
the negative moment at integral abutments. Live loads were applied to the
composite section to obtain negative moments, lateral and vertical loads.

In order to consider longitudinal loads such as braking load, shrinkage
loads and temperature loads, the finite element analysis method was applied. Soil
lateral pressure and surcharge forces were taken into consideration and applied at
the abutments. Since soil-structure interactions are complex, the L-pile program
was used to determine the micropile contributions to the structure-soil interaction.
The STAAD program was introduced to analyze the general structure behavior.

212
Since seismic forces were not significant for this project, seismic analysis
is not addressed in this paper.

INTRODUCTION

This bridge is a side by side 686mm x 1220mm pre-cast, pre-stressed concrete
box girder superstructure, with a simple span of 19.14 m. Integral abutments were
used in this project. The abutments were integrated with the pre-cast pre-stressed
box girder by using dowels and anchored bars to take shear load, moment and
vertical load.

This bridge was located in Wyoming. It has 18.50 meters span (bearing
centerline to centerline). The girder lengths were 19.25 meters. The bridge out-to-
out width is 12.00 meters and curb-to-curb width is 11.00 meters. The bridge was
aligned with 0
o
skew. A typical section is shown in Fig.1. Nine 1220x 686 mm
girders were put side by side to take design loads. Overhang is 0.51 m with
thickness varying from 200 mm to 250 mm. A typical 150 mm concrete deck slab
was poured after tightening of girders. The barrier rail was typical WYDOT TL-3
barrier rail and was designed to resist TL-3 collision load.

Due to the traffic requirement, construction was staged. Stage 1
construction was carried out by building the north side. Then the traffic was
switched to the new bridge. For stage II, after demolishing the old bridge, the
south side bridge was built.

The bridge integral configuration is shown in Fig.2. When installing the
1220 x 686 mm pre-cast pre-stressed box girders and 150 mm deck slab, the box
girders act as simple span bridge. Then the #32 dowel were built to resist
longitudinal shear forces, and the back wall reinforcement to resist moment.
When considering the structural resistance horizontal forces and moment, the
abutment-pile system is introduced.

SUPERSTRUCTURE DESIGN LOADS

DEAD LOAD: The Dead Loads include component (DC1), Barrier Rail
(DW1), Top Slab (DW2) and future wearing surface (DW3). The results are listed
in Table 1.

213
Table 1: Dead Loads
Load Case Load (KN/m) Description
Component (Box girder
1220x686 mm) (DC1)
94.10 Acting on Simple Girder
Barrier Rail (DW1) 4.67 Acting on Simple Girder
Top Slab (DW2) 43.09 Acting On Simple Girder
Future Wearing Surface
(DW3)
12.29 Acting On Finished Structure

LIVE LOAD: The Design Live load is HS-20. The Load Combination is
based on AASHTO LRFD Table 3.6.2.1-1 [1]. The live load is a moving load. In
order to produce the maximum reactions to the abutment, influence lines were
used to determine the controlling forces. Table 2 shows the analysis results, which
include a 33% impact force per AASHTO LRFD [1]. The truckload plus lane load
will govern the design as shown in the table.

Table 2: Live Loads
Reactions (KN)
Load Case
Load
Presence
factor
Analysis Factored
1 Lane 1.20 446 535
2 Lanes 1.00 892 892
3 Lanes 0.85 1338 1137

BRAKING FORCE: According to AASHTO LRFD 3.6.4 [1], there are a
few combinations for braking force. They are the 25% truck load, 25% tandem
load, 5% (truck + lane) load or 5%(tandem +lane load multiplying by the lane
number and present factor, acting 1.83 meters above surface). From this table, it is
obvious that truckload governs the design as shown.

Table 3: Breaking Force (BR)
Load Case Load (KN)
25% Truck 160
25%Tandem 111
5% (Truck + Lane) 50
5%(Tandem + Lane) 40

CENTRIFUGAL FORCE: This bridge is located on a curve with radius of
400 meters. According to AASHTO LRFD 3.6.3 [1], the centrifugal force will be:

KN W C CF 200 = = (1)

214

THERMAL FORCES: The design mean temperature was 15
o
C, with rise
20
o
C and fall 25
o
C.

SEISMIC FORCES: Seismic acceleration is 0.075g. According to
AASHTO LRFD 3.10 [1], no special analysis will be required.

WIND LOAD: Based on AASHTO LRFD 3.8 [1], the design wind
velocity is 160 km/h. The wind load results are listed in Table 4.

Table 4: Wind Load
Load Case Load Description
Wind Pressure On
Superstructure
4.4 KN/m Acting on Simple Girder
Wind Pressure On
Substructure
1.3 KN/m
2
Acting on Simple Girder
Wind Pressure On
Vehicle
1.5 KN/m Acting On Simple Girder

STRUCTURAL ANALYSIS MODEL

The integral abutment is shown in Fig.1 and Fig.2. The height of the abutment
varies from 3.274 m to 3.690 m to match roadway superelevation. The thickness
of the abutment is 750 mm typically. The bottom elevation of the abutment is 0.6
m below finish grade, seated on the 244mm micro-pile. By using L-pile program
analysis, 9 micropiles were used on each abutment (micro-piles were designed
separately and are beyond the scope of this paper). The structure was modeled as
two stages. Stage 1 was a simple beam, which was during construction. Stage 2
was a frame including the pile effects, which was the finished bridge. The
calculation models are shown in Fig.3.

The effective pile length can be estimated from CALTRANS BRIDGE
DESIGN AIDS [2] as Lp=9d (2196mm). For stage 1, half loads were transferred
to each abutment by bearing pads, and the Finite Element Analysis computer
program (STAAD) was used to analyze the load cases under stage 2.

215

Figure 1: Typical Integral Abutment Elevation

Figure 2: Integral Abutment Typical Section

216

Figure 3: Calculation Model for Integral Abutment Bridges

INTEGRAL ABUTMENT DESIGN

The typical cross-section of the integral abutment is shown in Figure 2.
The abutment was seated on 9 micro-piles with 8 equal spacings of 1300 mm. The
abutment model can be assumed as a continuous beam as shown in Figure 4.

(a) Stage 1
(b) Stage 2

217

Figure 4: A Calculation Model of Integral Abutment

Top and Bottom Reinforcement Design

The top and bottom reinforcement design was the vertical load-based
flexural design. There were two load cases: Case A, construction stage, which is
based on strength III limit state; Case B, final stage, which is based on strength I
limit state.

Case A

Case A includes the reactions of girder and deck slab, and abutment
related weight. According to AASHTO LRFD [1], a load factor of 1.5 is used for
strength III limit state. The reaction from the girder and deck slab is:

Maximum DL = 152.08 KN/girder/side, therefore,

Side Girder KN DL p
u
/ / 14 . 228 5 . 1 = = (2)

The abutment weight is assumed evenly distributed between the
micropiles for each element:

m KN w
u
/ 8 . 105 ) 02 . 8 50 . 62 ( 5 . 1 = + = (3)

(a) Original Structure
(b) Finite Element Model

218
In order to calculate the max positive moment, assume a simple span
girder between piles, and multiply by 0.8 to account for continuity. The positive
moment will be:
m KN
p L w Lp P
M
u u
u
=
=
20 . 77
)
8
3 . 1 8 . 105
4
3 . 1 14 . 228
( 8 . 0
)
8 4
( 8 . 0
2
2
(4)

The required reinforcement at the bottom of the pile cap is:

n r
M M = 9 . 0 (5)

where,
)
2
(
a
d f As M
y n
= (6)

By using fc=28 MPa, f
y
=414 MPa, and 4#22 bars at bottom of abutment,
then M
n
=2170 KN-m. So,

u n r
M m KN M M > = = 1953 9 . 0 (7)

Therefore, Case A doesnt govern design.

By inspection,
u r
M M
3
4
> , therefore the minimum reinforcing
requirements of AASHTO LRFD 5.7.3.3.2 [1] are satisfied.

The depth of compression block, or steel reinforcement, is small relative
to the section effective depth. This means the maximum reinforcement
requirements of AASHTO LRFD 5.7.3.3.1 [1] are satisfied.

For shear design, the ultimate shear force is:

KN
L w
P V
p u
u u
298
2
=
+ = (8)

The shear resistance, V
r
is:

KN V
n r
2104 V = = (AASHTO LRFD 5.8.2.4) (9)

Shear reinforcement is not required from Case A calculations. For
constructability, use #16@305mm spacing.

219

CASE B

Case B is the final stage of structure and includes all factored loads. The
maximum factored girder reactions due to the girder dead loads and deck act as
point load through bearing pads transferring to abutment.

KN P
u
228 = (9)

The distributed loads are coming from abutment, approach slab, stone
veneer and live loads:

m KN w
u
/ 73 . 147 = (10)

Similarly, the maximum moment is calculated by assuming simple
reaction at mid-span between piles, and then multiplying by 80% to account for
continuity:

KN M
u
8 . 142 = (11)

The resistant moment is calculated as:

u r
M m KN M > =1953 (12)

The factored shear force is:

KN V
u
549 = (13)

and shear resistant is as calculated previously:

KN V
r
2104 = (13)

The minimum and maximum reinforcement by checking is good.
Therefore, for this case, the abutment design is not governed by load, but
geometry and constructability.

VERTICAL REINFORCEMENT DESIGN

In order to design the reinforcement at the back-wall, the finite element
program STAAD was used to calculate the forces for the structure under stage 2.

220
The negative moments at abutment are summarized in Table 5. The live
loads multipliers are shown in Table 6. For vertical loads, it is easier to check by
hand by using the moment distribution method. The longitudinal loads are
distributed to each abutment by their stiffness. The following will focus on the
temperature loads.

Table 5 Negative Moment Results and Combination (Strength I)
Service Loads Strength I
Load Factor Shear (KN) Moment (KN-m) Shear
(KN)
Moment
(KN-m) Max Min Max Min Max Min
DC 836 0 1.25 0.9 1045 753 0 0
DW 534 178 1.5 0.65 801 347 266 115
EH 0 335 1.5 0.9 0 0 502 301
ES 0 61 1.5 0.75 0 0 92 46
Truck 142 278 1.75 249 0 486 0
Lane 85 98 1.75 148 0 171 0
IM
_Truck
47 92 1.75 82 0 161 0
CE 0 0 1.75 0 0 0 0
BR 31 556 1.75 54 0 973 0
WS 0 0 0 0 0 0 0
WL 0 0 0 0 0 0 0
TU1(+) 0 235 1.2 0.5 0 0 281 117
TU2(-) 0 -294 1.2 0.5 0 0 -353 -147
SUM(+) 1675 1831 2380 1100 4154 579
SUM(-) 1370 -294 1846 1100 -6 315

Table 6: Live Loads Multiplier
LL Case Lane Lane Factor Final factor Govern factor
1 1 1.2 1.2
2 2 1 2
3 3 0.85 2.55
2.55

The abutment length is:

m La 12 = (14)

The factored moment is:

m m KN
La
M
M
total u
u
/ 346 = =
(15)

By using #19 bars @ 200mm spacing, the abutment resistant moment capacity is
calculated as:

221

m m KN M M
n R
/ 360 = = (16)

Vertical reinforcement at front face of the abutment is usually governed by
temperature requirements. Using #16 at 150mm meets the AASHTO requirement.

DOWEL DESIGN

Dowels were designed based on the horizontal forces, usually longitudinal
forces, such as soil pressure, braking forces, temperature forces and live load
introduced horizontal reactions at dowel level. From the STAAD program
analysis and load combinations, the governing ultimate force is:

KN V
u
5 . 145 = (17)
Using a #32 dowel, the resistant force is calculated from friction formula as:

u n
V KN V > = 5 . 182 (18)

CONCLUSIONS

This paper has presented a general design method of integral abutment
based on the AASHTO LRFD method. The design considered all AASHTO
LRFD required loads. The paper divided the design into several simple models.
For the abutment top and bottom reinforcement design, vertical loads were
considered. For the vertical reinforcement and dowel designs, horizontal forces
were mainly considered. The front face vertical and horizontal reinforcement
were governed by temperature requirement. Future research should focus on the
pile-soil interactions for series pile diameters, especially micropiles.

REFERENCES

1. AASHTO LRFD Bridge Design Specifications, Third Edition, 2004.
2. CALTRANS, BRIDGE DESIGN AIDS, Department of Transportation, State of
California.
3. Richard M. Barker and Jay A. Puckett, Design of Highway Bridges, John Wiley &
Sons, Inc., 1997.
4. FHWA, Micropile Design and Construction Guidelines, NTIS, Virginia, June 2000.

222
Behavior of Pile Supported Integral
Abutments

Edwin G. Burdette
1
, Samuel C. Howard
2
, Earl E. Ingram
3
,
J. Harold Deatherage
4
, and David W. Goodpasture
5

1
Professor of Civil Engineering, The University of Tennessee, Knoxville, TN
37996-2010, Phone 865-974-7704, email: eburdett@utk.edu
2
Graduate Research Associate, The University of Tennessee, Knoxville, TN
3
President, Ingram Concrete, New Braunfels, TX
4
Professor of Civil Engineering, The University of Tennessee, Knoxville, TN
5
Professor Emeritus of Civil Engineering, The University of Tennessee,
Knoxville, TN

ABSTRACT

Two sets of field tests on integral abutments were preformed at The
University of Tennessee in research sponsored by the Tennessee Department of
Transportation (TDOT). Tests on concrete abutments supported by steel H-piles
were followed by tests on prestressed concrete piles. The purpose of the tests was
to evaluate the design criteria used by TDOT to design integral abutment bridges.
The nations leader in the utilization of integral bridges, TDOT continues to
extend the length limits of such bridges. The research described here provided
realistic field data upon which to evaluate design criteria.

Piles were driven into residual red clay soil. Concrete abutments were built
on top of the piles to simulate the behavior of actual integral abutments. Lateral
load was applied to the piles to induce horizontal displacements consistent with
those that occur due to temperature change. Six abutments with steel piles and
four with concrete piles were tested, first to the displacement limits corresponding
to TDOT criteria, and then to displacements well beyond the current design limits.
Tests to failure of two abutments with steel piles and one with concrete piles are
particularly interesting.

The test results showed that current TDOT design criteria for maximum
horizontal displacement are conservative. The test data indicated that current
criteria could be extended to accommodate somewhat larger displacements and
correspondingly longer lengths for jointless bridges.

223
INTRODUCTION

The only good joint is no joint. This philosophy, articulated almost forty
years ago by Henry Derthick, who was then the Engineer of Structures for the
Tennessee Department of Transportation (TDOT), has led the state of Tennessee
and others to adopt designs that minimize joints in bridges. As a result the use of
integral abutments has become commonplace, with Tennessee being the national
leader in the implementation of designs using integral abutments. The rationale
for integral abutments is presented by Wasserman and Walker [1].

The use of integral abutments to achieve jointless bridges is applicable to
bridges of short to moderate length, the limit of length being determined by the
maximum lateral displacement that an abutment can withstand without damage
that would threaten the serviceability of the bridge. The design criteria used by
TDOT limit the horizontal movement of a pile at the ground surface to 1.0 in.
(25.4 mm) in each direction. These criteria are explained more fully in Reference
1. The limits of bridge length consistent with these criteria are 500 ft. (160 m)
and 800 ft. (240 m) for steel and concrete bridges, respectively. These lengths,
however, have been exceeded. A steel bridge 575 ft. (175 m) long and a concrete
bridge 1,175 ft. (360 m) long are currently in service, the latter being the longest
jointless bridge known to the authors.

The desire by TDOT to extend the length boundaries of jointless bridges led
to a major research project sponsored by TDOT and carried out by the Civil and
Environmental Engineering Department at The University of Tennessee. This
project was carried out in two phases. Phase I tested concrete abutments
supported by steel H-piles; Phase II tested concrete abutments supported by
prestressed concrete piles.

OBJECTIVE

The objective of the research was to investigate the behavior of integral
abutments supported by steel and concrete piles and to assess the adequacy of
TDOTs design criteria. The specific objective of this paper is to describe briefly
the results of load tests performed on both steel and concrete piles. Emphasis in
the paper is the description of the overall load-deflection behavior of the pile-
abutment system including a discussion of the factors that limit horizontal
displacement of both types of piles. Appropriate limits of horizontal displacement
are discussed.

TEST SET-UP

An extensive testing program, which consisted of a series of field tests on
steel H-piles and prestressed concrete piles supporting integral abutments, was

224
sponsored by TDOT and performed by personnel from The University of
Tennessee, Knoxville. Six integral abutments with steel HP 10x42 piles and four
abutments on prestressed concrete piles were tested during the period from 1997
through 2003. The steel piles were driven into undisturbed clay soil 38 ft. (11.6
m) and the concrete piles were driven into the ground 36 ft. (11 m). Bending was
about the strong axis. Each 14-in. (0.356 m) square concrete pile was prestressed
with 6, in. (12.7 mm) diameter, 270k tendons. The axial load on the pile at the
beginning of the test varied from 65 kips (289 kN) for the first three steel piles to
approximately 90 kips (400 kN) for the later steel and all prestressed piles. Each
pile was embedded 12 inches (0.305 m) {24-in. (0.610 m) for 1 steel pile} into a
30 in. (0.762 m) {5 steel piles} or 36 in. (0.914 m) {1 steel pile and 4 prestressed
piles} wide simulated abutment. A slab was cast on top of the abutment to
provide ballast, to provide a way to apply lateral load to the system and to provide
a way to restrain the system as lateral load was applied. The test setup is
illustrated in Figures 1 and 2.

Abutment
Concrete Ballast Blocks
(6 ft x 2.5 ft x 2 ft)
1 ft
HP10x42
1 ft
2 ft
3 ft
14 ft
Pull Pad
Pulling Slab
(14 ft x 10 ft x 2.5 ft)
Load Cell
Reaction Beam
38 ft
Jack
Load Cell

Figure 1. Steel H-Pile Test Setup (Reaction Beam)

225
2 ft
Concrete Ballast Blocks
(2 ft x 2.5 ft x 6 ft)
Load Cell
Jack
W24x117
HP12x53
Load Cell
Abutment Slab
(14 ft x 10 ft x 2.5 ft)
14 in Prestressed Pile
36 ft
3 ft
14 ft
1 ft

Figure 2. Prestressed Concrete Pile Test Setup (Hold-down Beam)

The concrete used in the abutments was TDOTs Class A mix which calls for
a minimum compressive strength of 3,000 psi (20.7 MPa) at 28 days. The
average 28-day cylinder strength of the concrete in the abutments and pull slabs
was 3950 psi (27.2 MPa). The 28-day compressive strength of the concrete in the
piles was 5,700 psi (39.3 MPa).

Lateral load was applied to the simulated abutment to produce a horizontal
deflection at the top of the pile. The reaction beam or hold-down beams shown
in Figures 1 and 2 restrained rotation of the pile, respectively, to simulate the
restraint provided by the bridge deck system in an integral abutment bridge. The
abutment rotational restraint induced reverse curvature in the pile in a way similar
to that in an actual bridge with an integral abutment. With the steel piles and the
reaction beam, as horizontal load was applied, the axial load on the pile caused by
the weight of the loading slab and abutment was reduced by the amount of the
upward reaction. The loading was applied in most of the tests at a rate of 1 inch
(25.4 mm) in 4 to 6 hours to simulate the expansion and contraction in a bridge
due to temperature change.

An Optim Megadac model 3108 data acquisition system with pc computer
control and data collection was used for the projects. Included in the
instrumentation for the steel pile studies were two Interface 100-kip capacity load
cells, two or three RDP Electrosence type LDC300A Linear Variable Differential
Transformers (LVDTs), Hitec HBW-35-240-6 welded strain gages and custom
Sensotec pressure sensors. Strain gages and pressure sensors were spaced at
eighteen or nine inch centers.

The instrumentation for the tests on prestressed concrete piles was relatively
simple. The primary variables measured in the tests were two forces, the pulling

226
force and the hold-down force, and two deflections, the longitudinal movement of
the pile at the ground surface and one foot above the ground surface at the bottom
of the abutment. Vertical deflections of each of the four corners of the abutment
slab were also measured; and strain gages, both mechanical and electrical foil,
measured the strains on both the pulling face and opposite face of the pile in the
one foot space between the ground and the bottom of the abutment.

TEST PROGRAM

The test regiment for the piles consisted of at least three tests to 0.5 in. (12.7
mm) followed by three or more tests to 1.0 in. (25.4 mm) lateral displacement in
one direction. Then the apparatus was relocated and the regimen repeated in the
opposite direction. After the last 1.0 in. (25.4 mm) test was completed, more
stringent tests were then conducted. On Concrete Pile No. 2, the lateral
displacement was increased to approximately 10 inches (254 mm); and repeated
cyclical loading to displacements of greater than 1 inch (25.4 mm) was performed
on Concrete Pile No. 4 and Steel Pile No. 6. The latter two piles were subjected
to a hundred cycles of loading.

BEHAVIOR OF PILES AND ABUTMENTS

Typical Load Deflection Behavior

As the horizontal deflection of a pile at the ground surface approached 0.5 in.
(12.7 mm), the prestressed pile cracked. While this crack reduced the stiffness of
the pile cross-section significantly, the effect on the overall load-deflection
behavior was much less pronounced. Cracking in the abutments was so minor as
to be almost insignificant. The element that limited the amount of horizontal
displacement for a prestressed pile was the pile itself. Figure 3 shows typical
load-deflection behavior of an integral abutment with prestressed concrete piles.

227
Figure 3. Typical Prestressed Concrete Load Deflection Curves

The behavior of steel H-pile integral abutments is controlled by the pile-
abutment interface. As a steel H-pile is deflected laterally, the abutment is
subjected to stresses that tend to cause it to crack. While the TDOT criteria were
found to work adequately in previous testing, the cracking of the abutment was
the limiting factor in determining the actual upper limit of horizontal movement at
the ground surface that could be tolerated with steel H-piles [2, 3, and 4]. Since
the earlier abutment tests were completed in 1999, TDOT has adopted a design
width of integral abutment of 36 in. (0.914 m), twenty percent greater than the
earlier widths, a change which allows a larger displacement before significant
cracking occurs. Figure 4 illustrates typical load-deflection behavior of a steel
H-pile supported integral abutment with the 36 in. (0.914 m) width.
0
10
20
30
40
50
60
70
0.00 0.50 1.00 1.50 2.00 2.50
Horizontal Deflection (in.)
L
a
t
e
r
a
l

L
o
a
d

(
k
i
p
s
)
15th Test
14th Test
13th Test
10th Test
9th Test

228
0
10
20
30
40
50
60
70
0 0.2 0.4 0.6 0.8 1 1.2
Hori zontal Defl ecti on (i n.)
L
a
t
e
r
a
l

L
o
a
d

(
k
i
p
s
) 1st Test
3rd Test

Figure 4. Typical Steel H-Pile Load - Deflection Curves

Cyclical Tests

When the basic test regimen was completed on Concrete Pile No. 4 and Steel
H-Pile No. 6, a series of cyclic tests to horizontal displacements of approximately
1.10 in. was performed. Slightly more than 100 cycles of loading were applied to
each pile. These tests were performed at a significantly faster rate of loading than
that used in the basic test regimen. Load-displacement curves illustrating this
cyclic loading are shown in Figures 5 and 6. These curves illustrate the fact that,
while there is some softening of the load-deflection behavior with repeated
cycles of loading, the pulling load at maximum deflection is reduced only a small
amount. The resistance of the pile to horizontal movement remains relatively
large, even after repeated cycles of loading to displacements, which exceed the
one-inch limit in TDOTs design criteria. The cyclic tests demonstrate the ability
of integral abutments to sustain repeated loads causing large horizontal
displacements with no loss of function

229
0
10
20
30
40
50
60
0.00 0.20 0.40 0.60 0.80 1.00 1.20
L
a
t
e
r
a
l

L
o
a
d

(
k
i
p
s
)
Cycle # 1
Cycle # 75
Cycle # 50
Cycle # 25

Figure 5. Cyclical Behavior of Prestressed Concrete Pile Integral Abutment

0
10
20
30
40
50
60
0 0.2 0.4 0.6 0.8 1 1.2 1.4
L
a
t
e
r
a
l

L
o
a
d

(
k
i
p
s
)
Cycle #4
Cycle # 50
Cycle # 75

Figure 6. Cyclical Behavior of Steel H-Pile Integral Abutment

230
Failure Tests

Failure tests were conducted on Concrete Pile No. 2 and Steel H-Pile No. 6.
Load was applied at a faster rate than that used in the normal test regimen,
approximately 2 in. (50.8 mm) per hour. The load was increased until the limit of
travel in the pulling ram was reached; the load was held and the ram reset; and
loading was continued. Loading continued until the pile or abutment was
essentially destroyed. Failure of the concrete pile occurred when the concrete had
ruptured in bending at a section at the pile abutment interface (tension on the side
of the pile in the direction of pull) and at a section approximately 4 ft. (1.22 m)
below the ground surface (compression on the side of the pile in the direction of
pull). In the test to failure of the steel pile, the concrete abutment limited the
amount of horizontal displacement. The load-deflection for the tests to failure of
Concrete Pile No. 2 and Steel H-Pile No. 6 are shown in Figures 7 and 8,
respectively.

0
20
40
60
80
100
120
0 2 4 6 8 10 12
L
a
t
e
r
a
l

L
o
a
d

(
k
i
p
s
)
1kip =4.45 kN
1in =25.4 mm

Figure 7. Prestressed Concrete Pile Integral Abutment Test to Failure

231
0
20
40
60
80
100
120
0 1 2 3 4 5 6
L
a
t
e
r
a
l

L
o
a
d

(
k
i
p
s
)

Figure 8. Steel H-Pile Integral Abutment Test to Failure

DISCUSSION

Reference 2 gives a thorough discussion of the behavior of the pile-abutment
interface for the tests performed on steel piles with abutments that were 30 in.
(0.762 m) wide. While some cracking in the abutment occurred at horizontal
displacements of 1 in. (25.4 mm), the conclusion was drawn that displacements as
large as 1.5 in. (38 mm) were acceptable in that the integrity of the abutment was
not compromised at this displacement.

In the tests of steel piles with 36 in. (0.914 m) wide abutments, the abutment
concrete was essentially pristine for displacements up to 1 in. (25.4 mm). There
was some indication of the onset of minor flaking of concrete on the bottom
surface of the pulling slab originating at the tips of the flanges of the pile at 1 in.
(25.4 mm) displacement. As this displacement was increased to 2 in. (50.8 mm)
and beyond, a thin layer of concrete in a localized area flaked off, with no
apparent effect on the load-deflection behavior of the system. Even at
displacements of 3 in. to 4 in. (76 mm to 102 mm), the pile-abutment interface
continued to function adequately.

The abutments in the concrete pile tests remained essentially undamaged
throughout the entire range of testing. In the tests of abutments on prestressed
concrete piles, the amount of horizontal displacement was limited by the piles. In
the tests of abutments on steel piles, the abutments themselves limited the amount
of horizontal displacement that the system could tolerate without loss of integrity.

232
As the horizontal displacement of a concrete pile at the ground surface
approached 1 in. (25.4 mm), the pile cracked just below the pile-abutment
interface. While this cracking reduced the stiffness at that section significantly,
the overall load-displacement behavior was only slightly altered. When the
pulling load was removed, the force in the prestressing steel closed the crack.

CONCLUSIONS

As noted in the Introduction, the length limit of jointless bridges is determined
by the longitudinal displacement that an abutment can withstand without
sustaining damage that threatens the serviceability of the bridge. It turned out in
the tests that the maximum longitudinal displacement that an abutment can
withstand is limited by different elements depending on the type of pile. With
steel piles the displacement was limited by cracking in the abutment, while in the
concrete piles, the piles themselves limited the longitudinal displacement. For
both types of piles, however, the tests showed that TDOTs design limitation of
1.0 in. (25.4 mm) is conservative. While no specific recommendation of
maximum allowable displacement was formulated through the research, a value
of displacement of as much as 1.5 in. (38.1 mm) would appear to be reasonable
for the prestressed concrete piles and somewhat greater than that for steel H-piles.

ACKNOWLEDGMENT

The financial support and personal encouragement by the Tennessee
Department of Transportation and the Federal Highway Administration is much
appreciated. The conclusions reported herein are those of the authors and do not
necessarily reflect the views of either sponsoring agency.

REFERENCES

1. Waserman, Edward P. and John Houston Walker, Integral Abutments for Steel Bridges,
Highway Structures Design Handbook, Vol. 11, Chapter 5, American Iron and Steel Institute,
Oct. 1996.

2. Burdette, Edwin G., Earl E. Ingram, David W. Goodpasture, and J. Harold Deatherage,
Behavior of Concrete Integral Abutments, Concrete International, Vol. 24, No. 7, July
2002.

3. Burdette, Edwin G., J. Harold Deatherage, and David W. Goodpasture, Final Report,
Behavior of Laterally Loaded Piles Supporting Bridge Abutments, Center for
Transportation Research, The University of Tennessee, December 1999.

4. Ingram, Earl E., Edwin G. Burdette, David W. Goodpasture, and J. Harold Deatherage,
Evaluation of Applicability of Typical Column Design Equations to Steel H-Pile Supporting
Integral Abutments, AISC Engineering Journal, Vol. 40, No. 1, First Quarter 2003.

233
Soil Structure Analysis of Integral Abutment Bridges

Petros M. Christou
1
, Marc I. Hoit
2
,

& Mike C. McVay
3
University of Florida, Department of Civil and Coastal Engineering

ABSTRACT

Integral Abutment Jointless Bridges (IAJB) are defined as simple or multiple
span bridges in which the bridge deck is cast monolithically with the abutment
walls. Different aspects such as good response under seismic loading, low initial
costs, elimination of bearings and less maintenance make this kind of bridge very
attractive for potential owners. The main issue related to the analysis of this type
of structures is dealing with the soil-structure interaction of the abutment walls
and the supporting piles. Other concerns include the transfer of stresses between
the different parts of the structure under the application of dead and live loads
under service or extreme events. For instance, lateral displacements due to
thermal stresses may be considerable as a result of the monolithic nature of the
structure. Also dynamic loads (seismic) or deformations may force a number of
the structural components (piles, piers, etc) to have inelastic response. In addition,
with a fixed pile head condition, lateral deformations/loads may introduce
significant axial loads/deformations within the piles/shafts. Since the analysis of
this kind of a structure is complex, the analytical model must be able to account
for the nonlinear soil behavior as well as be able to model the nonlinear (i.e.
inelastic) structural response for the collective bridge. This paper describes an
alternate model that is proposed for the analysis of IAJB using the commercially
available FB-MultiPier
software which provides the ability to model a bridge

system (abutment to abutment) along with the foundation and the soil. The soil
response is characterized by nonlinear springs (p-y for the lateral, t-z for the
vertical and - for the torsional soil response). The material nonlinearity of the
piles, intermediate piers and pile bents is handled with the implementation of a
discrete element model incorporating nonlinear stress-strain response of the steel
and concrete.

GENERAL

Bridge structures react to the temperature changes and they contract or expand
depending on the thermal stresses that are developed internally. It is common
practice to design a bridge to allow the superstructure (bridge deck) to move in
response to the internal stresses which avoids excessive forces in the substructure
and the supporting foundation. Different components such as expansion joints,
extension bearings and roller supports [1] are utilized to allow for superstructure
1
Petros M. Christou, PhD, Bridge Software Institute, University of Florida, Gainesville, FL 32611
2
Marc I. Hoit, Professor of Civil and Coastal Engineering, University of Florida, Gainesville, FL 32611
3
Mike C. McVay, Professor of Civil and Coastal Engineering, University of Florida, Gainesville, FL 32611

234
movement relative to the substructure. The alternative to this kind of bridge is the
integral abutment concept, which eliminates the joints.

IAJB are single or multiple span bridges which have the bridge deck
(superstructure) cast monolithically with the abutments (substructure), forming a
connection that is capable of transferring moments (moment connection) in
addition to axial and shear forces. The major characteristic of the IAJB is that the
bridge deck is cast without any expansion joints. Traditionally the foundation of
the abutments consists of a single row of flexible piles that are placed in pre
augered holes up to a certain depth. The holes are then backfilled with granular
material. The pile heads are either fixed at the point of connection with the
abutment walls (i.e. transfer moments) or they are pinned (no moment transfer).
The intermediate piers can be either cast independently from the bridge deck and
essentially act as roller supports or cast monolithically forming a moment
connection.

As the temperature changes, the length of the bridge deck also changes. The
lateral thermal change causes lateral deflections to the abutments, the foundation
piles and the adjacent soil. In order to accommodate the lateral movement, the
abutments and the piles/shafts are forced to displace and rotate. The soil moves
either outwards (deck expansion) or inwards (deck contraction) and it is either in
the passive state (deck expansion) or an active state (deck contraction). It is
evident that the longer the bridge deck and the larger the temperature variations
the larger will be the change in length of the bridge deck. This in fact is one of the
main concerns when dealing with IAJB because the magnitude of the strains in
the piles and the abutments can become excessive. As a result there is a legitimate
argument regarding the maximum allowable length of the IAJB. The vertical
loading on the bridge is mainly due to the dead weight of the structure as well as
any superimposed load that acts on the bridge at a given time. This loading is
transferred to the ground through the piles under the abutments as well as the
internal piers.

The main issue related to the analysis of this type of structures is dealing with
the soil-structure interaction of both the abutment walls as well as the supporting
piles. The behavior of the structural components including the piles can either be
linear or nonlinear depending on the amount of the applied forces. The behavior
of the soil on the other hand, is nonlinear. This complication generates a
nontrivial problem where the different element responses are interdependent and
any attempt to analyze the different parts of the bridge independently will involve
considerable assumptions and approximations. The movement of the structural
components is opposed by the soil behind the abutment walls, the soil next to the
piles and the internal pier foundation. The magnitude of the forces that are
developed in the soil is directly related to the magnitude and the nature of the
displacement of the structural elements. Consequently this is a challenging soil-
structure interaction problem that requires a model for the complete bridge
including concrete super and substructures, piles as well as soil. In addition, the

235
lateral soil behavior/resistance is nonlinear and is a function of displacements.
Consequently, a significant challenge is to determine the interactions between the
bridge, the abutments and the foundation and calculate the forces that will be used
for the design of the bridge.

Different methods and innovative procedures have been used to tackle the
analysis of the IAJB. The approaches generally involve either two-dimensional or
three-dimensional models. The most common method is to model the
superstructure separately from the foundation. For the superstructure, an
equivalent stiffness of the pile-soil system is employed based on assumed
displacements. If the computed displacements dont agree with the assumed
values then new substructure stiffness is found from the computed values. This is
an iterative procedure that performs successive superstructure solutions and
foundation stiffness corrections until compatibility is obtained. These methods
tend to be simple and can be very efficient depending on the experience of the
engineer to obtain an educated first guess. However they tend to oversimplify
the soil/ foundation part of the system and consequently the accuracy of the
results is dependent on the quality of the simplifying assumption. Other methods
include more sophisticated finite element models which account for all the
elements in the structure including the foundation and the surrounding soil [2].
These methods can potentially be very accurate depending on the capabilities of
the software that will be used and they can account for the potential nonlinearities.
However in order to reach an acceptable level of accuracy the model must be
discretized using a collection of different elements which may not be a trivial task
at times. The process of obtaining a model which will produce desirable results
can become time consuming and is dependent on the ability of the engineer to
obtain realistic responses.

This paper describes an alternate model that is proposed for the analysis of
IAJB using the commercially available FB-MultiPier
software which provides

the ability of modeling a bridge system (abutment to abutment) along with the
foundation and the soil. The model is simple to create yet it addresses all the
issues regarding the interaction of the superstructure with the substructure and the
foundation. The model allows for nonlinear behavior of the piles and the
intermediate piers and also includes nonlinear soil springs to describe the
response of the soil behind the abutment walls as well as the response of the soil
around the piles.

FB-MultiPier

The commercially available software package FB-MultiPier
[3] was used for

this study. FB-MultiPier
is capable of modeling a system which consists of

foundations (soil, piles, pile caps etc) and piers (pile bents) which are connected
with a bridge superstructure (Figure 1). The piles can be selected from a library of
preexisting sections or they can by customized by the user. The soil is described

236
by different layer strata and it is characterized by p-y curves for the lateral
response, t-z curves for the vertical response and - curves for the torsional
response of the soil. The pile cap is modeled with plate elements. The pier
structures (pier columns, pier caps, bent caps) are modeled with three-dimensional
nonlinear discrete elements. The model for the bridge superstructure consists of a
single line of three-dimensional discrete elements, which are connected to the pier
(pile bent) caps through bearing elements that are placed at the locations of the
girders on the cap. The user has the flexibility of assigning any type of bearing
connection (integral, pinned, roller) or any custom bearing response, which is
characterized by a graph. All of the structural elements that are involved in the
model, except the pile cap, can have either linear or nonlinear material response.
P effects are also included. The pile cap can only have linear behavior. FB-
MultiPier
can perform static, dynamic or pushover analysis.

Figure 1: Three Dimensional Proposed Model Implemented in FB-MultiPier

THE BRIDGE

The bridge that was selected for this study (Figure 2) is located over Nashua
River in Fitchburg, Massachusetts and it is the same bridge that is found in Faraji
et al [2]. The choice of the particular bridge was based on published results that
were used for comparison with the findings of this study. The length of the bridge
is 150 ft and its width is 54 ft. It consists of three continuous spans, which are
symmetric about the center of the bridge. The length of the two end spans is 45 ft
and the length of the intermediate span is 60 ft. The bridge deck is supported by
two intermediate piers, which are monolithically joined at the two end abutment
walls. The abutment walls are supported by a single row of foundation piles.

237
MODEL IMPLEMENTATION

The three-dimensional discrete element used for the analysis is described in
the FB-MultiPier
Users Guide [3]. The element is very similar to a conventional

three-dimensional frame element used in many structural analysis programs. Note
that the discrete element has two additional nodes at its quarter points in order to
model the nonlinear response of the element and P effects.

The composite bridge deck consists of a concrete slab that rests on seven
girders (Figure 3) spanning along the length of the bridge and also a series of
eight beams, which are used to tie the girders in the transverse direction. The
concrete deck is 8.5 in thick. The girders are W36x135 and they are spaced at 9 ft
on center. The transverse beams consist of channels and angles and they are lined
every 22.5 ft in the outer spans and every 20 ft in the center span. The
superstructure is simplified and modeled as a single line of three-dimensional
discrete elements. The superstructure is framed on the middle piers and the end
abutments through rigid links, which have the ability to transfer, shear, moments
and axial forces to the pier and pile bent caps. Also the links are used to offset the
superstructure from the caps and simulate the physical distance from bridge deck
to the pile caps. In order to have a valid model, the properties of the
superstructure elements must represent the whole bridge. Of interest are the area
(A), the plastic centroid (PC) and the moments of inertia (I
xx
, I
yy
) of the deck. The
eccentricity of the deck is taken as the distance from the pier cap to the plastic
centroid of the section.

Figure 2: Study Bridge over Nashua River in Fitchburg, Massachusetts

The intermediate piers (Figure 3) consist of three concrete columns, a pier cap
and two rows of bearings. The columns are 30 inches in diameter and they are
placed every 21.5 ft. The dimensions of the pier cap are 3.5 ft wide, 3 ft deep and
56.5 ft long. The columns and the pier cap are modeled with discrete elements.
in

238
The columns extended a distance of 14 ft into the ground. The columns as well as
the pier cap were assumed to have nonlinear behavior. Above the pier cap there
are two rows of bearing elements that are placed at the geometric locations where
the girders connect to the pier cap. The bearing elements were loaded with the
girder dead load and are connected together with rigid links to simulate the
rigidity of the superimposed deck.

The end abutments (Figure 4) consist of the abutment wall, a single line of
seven foundation piles and a line of bearing elements. The dimensions of the wall
are 30 in thick and 8 ft high. The piles are of type HP 12x74 and they are spaced
every 9 ft on center. The pile heads are fixed at the abutment walls and therefore
they form a moment connection. In the absence of a wall (plate or shell) element
in FB-MultiPier
the abutment wall was modeled using discrete elements (we will
refer to them as wall elements). FB-MultiPier
is able to model multiple sections

along the pile length. The wall elements actually form the upper part of the piles
and they have the same cross section properties as the actual wall. In order to
assure that these elements form a rigid block and provide the adequate rigidity of
the wall in the transverse direction, FB-MultiPier
provides the option to use

additional elements (Extra Members). The Extra Members are rigid discrete
elements (user defined) that can be added to the structure to connect elements and
force them to respond in a rigid manner. During the course of this study it was
found that the results were not affected by the presence of these elements and
therefore they were dropped to simplify the model. Above the abutments there is a
line of bearing elements located at the geometric centers of the superimposed
girders. As with the intermediate piers, the load is applied to the cap at these
bearings.

Figure 3: Cross Section of Composite Bridge Deck

The response of the soil to the foundation piles and the internal piers is
modeled with nonlinear springs (Figure 4) that are applied to each pile node
(Figure 5). The lateral behavior of the springs is characterized by nonlinear curves
given by American Petroleum Institute (API) based on soil type. For this study,
sand was considered and curves given by ONeill et al [4] were employed. The
latter require the angle of internal friction (), unit weight and sub grade modulus

239
(K) of the soil, sand. To characterize the different lateral springs along the pile,
based on soil properties, the soil input is (FB-MultiPier
) based on layers (Figure

6). The program will automatically generate the characteristic springs on the
nodes along the piles. For this study, the soil was modeled in three layers
extending to a depth below the pile tips. The first layer extended downward a
depth of 8 ft and represents the soil behind the abutment walls. The second layer
was 6 ft thick and represents the soil that was pre augered during the pile
installation. The third layer extended below the pile tip, i.e. 32 ft below the
ground surface. Two different levels of compaction for the sand layers were
studied. Loose sands (=30
o
) and dense sands (=40
o
). The two types of soil
were implemented behind the abutment wall and adjacent to the piles.

The loading considered for this study consisted of the bridge dead load and the
thermal loads. The latter loads on the abutments as well as the piles were expected
to control the design of IAJB. The simple unit weight of the materials was used to
calculate the self-weight of the bridge. Even though FB-MultiPier
provides the
option to calculate and apply the self-weight automatically, it was decided to
calculate the weight of the structure and apply it directly on the bridge bearings.
The thermal load was calculated from a temperature change of T = 80
o
F.

Figure 4: Site View of Abutment and Detail A Nonlinear Springs on Pile Nodes

240

Figure 5: Thin Element Mode of the Study Bridge

Figure 6: Soil Layer Data Input in FB-MultiPier

RESULTS

Of interest from the analysis were the displacements of the abutment walls as
well as the induced moments in the piles within the structure. Figure 7 shows the
lateral displacements along the depth of the abutment wall and the foundation
piles as obtained from the proposed model. For comparison, the displacements
reported by Faraji et al [2] based on a finite element analysis are given in Figure
8. Evidently, the FB-MultiPier
model yielded a displacement at the top of the

abutment wall of about 0.36 in to 0.44 in depending on the compaction of the soil
behind the abutment wall. The finite element model from Faraji et al [2] gave a
displacement of 0.43 in to 0.46 in for the same conditions.

241
LATERAL DISPLACEMENT
ALONG THE ABUTMENT
(FB-MultiPier)
-40
-35
-30
-25
-20
-15
-10
-5
0
-0.46 -0.41 -0.36 -0.31 -0.26 -0.21 -0.16 -0.11 -0.06 -0.01
DISPLACEMENT (in)
D
E
P
T
H

(
f
t
)
LL LD DD DL

Figure 7: Plot of Lateral Displacement Along the Abutment (FB-MultiPier
)

LATERAL DISPLACEMENT
ALONG THE ABUTMENT
(Faraji et al)
-40
-35
-30
-25
-20
-15
-10
-5
0
-0.46 -0.41 -0.36 -0.31 -0.26 -0.21 -0.16 -0.11 -0.06 -0.01
DISPLACEMENT (in)
D
E
P
T
H

(
f
t
)
LL LD DD DL

Figure 8: Plot of Lateral Displacement Along the Abutment (Faraji et al [2])

Presented in Figure 9 are the moments along the length of the HP-piles from
FB-MultiPier
analysis along with Faraji et al [2] results at the pile heads for the
different soil backfilling conditions. Evidently, the results reveal a similar pattern
for the pile displacements and moments for both methods of analyses. The
differences, which are small for both displacements and moments, were attributed
mainly to the soil models and properties used in both analyses. For instance, for
the FB-MultiPier
analysis, the lateral soil springs were calculated based on

ONeils p-y representation whereas the Faraji et al [2] finite element model
employed lateral springs calculated separately for that back analysis. It should be

242
noted also that the FB-MultiPier
soil model includes the presence of vertical (t-z

curves) and torsional (- curves) springs which are absent in the finite element
model Faraji et al [2]. In other smaller issues, FB-MultiPier models the composite
bridge deck as a series of discrete elements whereas Faraji et al [2] modeled the
composite deck with the more accurate plate elements. Finally the FB-MultiPier
employs q-z curves for pile tip resistance, whereas the Faraji et al [2] finite
element model assumed the tip of the piles were fixed. Judging from the results
the difference in the modeling of the tip bearing resistance is probably
insignificant. However, the presence of the vertical springs does have an impact to
the results since they influence the movement of the overall bridge.

Based on the results of the two methods and the similarities in the response of
the soil behind the abutment walls we can draw some conclusions as to the
validity of the simplified model that is presented in this work. Obviously there are
some issues that need to be addressed and require further study concerning the
more detailed simulation of the composite deck. The overall performance of the
model, together with the ease of implementation, suggest that this is an efficient
way for the analysis of the IAJB and a reliable alternative to the more complicated
and time consuming modeling methods.

Figure 9: Plot of the Moment Along the HP Pile (FB-MultiPier
)

ACKNOWLEDGEMENT

The authors would like to thank Mr. Henry Pate from the Tennessee
Department of Transportation and Dr Edwin Burdette from the University of
Tennessee, Knoxville for the data provided during the course of this work. The
authors also like to thank Dr John M. Ting of the University of Massachusetts,
MOMENT ALONG THE PILE DEPTH
(FB-MultiPier)
POINTS AT THE PILE HEADS -FE
LL(FE) LD(FE) DL(FE) DD(FE)
-40
-35
-30
-25
-20
-15
-10
-5
0
-95 -85 -75 -65 -55 -45 -35 -25 -15 -5 5 15 25 35 45
MOMENT (k.ft)
D
E
P
T
H

(
f
t
)
LL LD DD DL LL(FE) LD(FE) DL(FE) DD(FE)

243
Lowell for providing us the information required to model the study bridge and
being able to compare the results.

REFERENCES

1. Kunin, J., and Alampalli, S. 2000. Integral Abutment Bridges: Current Practice in United
States and Canada, J. of Performance of Constructed Facilities:104-111.
2. Faraji, S., Ting, J.M. and D.S. Crovo. 2001. Nonlinear Analysis of Integral Bridges: Finite-
Element Model, J. of Geotechnical and Geoenvironmental Engineering: 454-461.
3. FB-MultiPier
Users Guide, version 4: Bridge Software Institute. University of Florida,

Gainesville, Florida.
4. ONeill, M.W. and Murchison, J.M. 1983. An Evaluation of p-y Relationships in Sands,
University of Houston Res. Rept. GT-GT-DF02-83 to the American Petroleum Institute,
Houston Texas.

244
Behavior of Two-Span Integral Bridges Unsymmetrical
About the Pier Line

David Knickerbocker, Hardesty and Hanover, LLP
Prodyot K. Basu, Vanderbilt University
Edward P. Wasserman, Tennessee Department of Transportation

ABSTRACT
The lack of symmetry in jointless bridges may result from unequal span
lengths, and/or unequal dimensions of the end abutments caused by, say, unequal
pile lengths and height of backwall. Finite element modeling is used in the study
and the results are validated with experimental data from two integral bridges
located in Middle Tennessee. These two-span HPC bridges consist of CIP
reinforced concrete deck slab over pretensioned concrete girders. Parametric
studies are undertaken by finite element modeling in parallel with field
observations, and interesting conclusions are drawn with respect to volumetric
loads.
INTRODUCTION
For bridge owners and designers, the mechanism of the integral abutment
along with minimization of intermediate expansion joints is proving to be an
increasingly attractive alternative to expansion joints and bearings.
Accommodation of movements by these mechanical means works well in theory.
However, tight tolerance requirements and difficulty of maintaining systems
whose performance is highly sensitive to corrosive agents, impact and
introduction of foreign objects, all common elements in the bridge environment,
has presented an obstacle to cost-effective maintenance [1, 2]. In the integral
bridge, expansion joints are replaced by an alternative mechanism in the
substructure. Abutments are supported by a single row of vertical piles; the
relative flexibility of this configuration allows for the requisite thermal
movements of the jointless superstructure. As the abutments and piles are
designed to undergo displacement against the backfill, integral bridge behavior is
based on soil-structure interaction. Intermediate piers are usually designed to
allow movement in the superstructure, providing further stability. At each end of
the bridge deck, an approach slab is used to convey deck movements well beyond
the bridge to the highway pavement interface, where a roadway expansion joint is
usually provided.

If the conventional jointed bridge works well in theory and less so in practice,
its jointless counterpart has proven practical and economical in service [3], and
current efforts toward development of rational analysis techniques will promote
more widespread adoption of the technology [4, 5, 6]. While omission of bridge
expansion joints and bearings relieves construction and maintenance costs,
analysis and design of integral bridges are more complex than for conventional

245
bridges. The nonlinear soil/structure interaction between the backfill material and
the abutments and piles acts to complicate the analysis. The nonlinear resistance
of soil to substructure deformation is not easily characterized without
sophisticated numerical modeling.
Under the FHWAs programs implemented to promote the use of high-
performance concrete in highway bridges, the authors were involved in design,
construction, and instrumentation of a high-performance concrete bridge built in
1999 [7]. The Porter Road Bridge see Figure 1 comprises two equal 159-ft
(48.5 m) continuous spans with integral abutments on a 27 skew, overpassing TN
S.R. 840 outside of Nashville, TN. The instrumentation program and
corresponding recorded behavioral data was found to be valuable in the
characterization of integral bridges, and studies of live load behavior and long-
term performance with respect to integral abutments have continued.

In addition to collection and processing of performance data, a finite element
model was constructed for in-depth analysis. This paper focuses on the results of
the analysis, with a focus on the behavior of the integral abutment details,
particularly degree of skew and depth of abutment and supporting piles, under
effects inducing expansion and contraction of the superstructure. Positive and
negative thermal gradients, and differential slab shrinkage, are imposed on the
model to investigate the effects on integral behavior.
Integral Abutment Bridge Study Specimen
The 32-ft (9.75-m) wide, two-lane Porter Road Bridge has two spans of 159 ft
(48.5 m) each with a 27 skew. Note in Figure 1 that the two abutments differ
significantly in size, as do the lengths of the respective piles. Four lines of
prestressed high-performance concrete (
'
c
f =10 ksi, or 69 MPa) 72-in. (1.83-m)
deep bulb-tee girders support the 8.25-in. (21-cm) thick reinforced concrete (
'
c
f
=7 ksi, or 34.5 MPa) deck in each span. For development of continuity in the
superstructure, the pier diaphragm wall and abutment backwalls are cast
monolithically with the deck, and the precast girders are made continuous with
the end-walls and deck via embedded reinforcement. The pier is designed to
deflect with the movements of the superstructure, and is detailed to behave as a
roller support, to allow the superstructure to translate and rotate relative to it.
Figure 1. Schematic Jointless Porter Road Bridge, with Integral Abutments.
WINGWALL
FLEXIBLE PILE
GROUP
INTEGRAL ABUTMENT
CONTINUOUS DECK
ON-GRADE PAVEMENT
FINISH GRADE
BEDROCK
ON-GRADE EXPANSION JOINT
APPROACH SLAB

246
The flexible abutment piles are oriented to rotate about their strong axis,
perpendicular to the abutment beam.

During manufacturing, two girders were outfitted with embedded vibrating
wire strain and temperature sensors in three vertical sections: at the two ends and
at the center. More sensors were placed in the deck above these locations prior to
casting for realization of fully composite behavior. Digital tiltmeters were also
used at the connections, for assessment of continuity, and measurement of
midspan deflection was performed using a catenary as well as by surveying
methods.
FINITE ELEMENT MODEL
The detailed finite element model was created using the general-purpose FEM
software package ANSYS V. 7.0 [8], which is capable of handling the required
nonlinear behavior of the soil reaction, as well as that of various types of
structural concrete components. The model consists of over 5,500 beam-column
and solid concrete elements for the bridge structural elements, and over 1,000
nonlinear springs for simulation of nonlinear soil behavior. To appreciate the
effect of skew, details of the Porter Road Bridge were modeled first with no skew,
and subsequently the geometry was skewed to varying degrees using a computer
program. Results are reported for a normal case of zero skew (=0), the as-built
case of =27, and an extreme skew case with =55.
Backfill
Resistance to translation of the integral abutment system was modeled using
equivalent Winkler springs, as in Figure 2(a). Passive resistance against pile
translation was modeled using ANSYS nonlinear spring elements COMBIN39.
The unidirectional springs were placed to resist forward, backward and sideways
translation of the piles. Soil reaction against concrete surfaces of abutments,
backwalls and wingwalls was modeled using the same element type, with
adjustment to soil stiffness elements made according to the adjacent area of
substructure to be resisted.

For sandy soil adjacent to piles, P- curves, approximating the nonlinear
response to lateral pile deflection, for granular soil are characterized by the
American Petroleum Institute [9] as

(
= y
Ap
kX
Ap P
u
u
tanh (1)
at a specific depth, X, where p
u
= ultimate bearing capacity (force/length), k =
initial modulus of subgrade reaction (force/volume), and y = lateral deflection.
The values for k and p
u
taken as the lesser of p
us
(shallow) and p
ud
(deep) are
functions of the angle of internal friction f, the effective soil weight g (weight
density units), depth X, and effective pile diameter D. For HP12x53 piles, the
tributary width in both directions is 12 in. (30.5 cm), and equivalent nonlinear
springs were placed at a spacing of 12 in. Figure 2(b) reports the results of the
above expression applied on the tributary area at discrete depths of 1 to 50 ft.

247
These nonlinear backfill reaction curves were integrated into the model in the
longitudinal and transverse directions, and made to act in compression as well as
virtual tension to account for compression on two sides of the pile in each
direction.
Resistance offered by the well-graded, consolidated fill behind the abutments,
backwalls, and wingwalls, was estimated using nonlinear P- curves provided by
NCHRP [10] for densely consolidated soil. Figure 2(c) presents this data for unit
tributary area and depth. The stiffness trend of each soil spring element was
then a multiple of the given curve, according to its tributary area and depth, in
inches.
Structural Components
The steel piles were modeled using the thin-walled plastic beam element
BEAM24, which allows for the exact cross-section to be defined, and is capable
of nonlinear plastic behavior.

The 8-node concrete element SOLID65 was used to simulate the abutments,
wingwalls, backwalls, diaphragm, pier, and deck. This element is specifically
0 0.1 0.2 0.3 0.4
0
0.1
0.2
0.3
0.4
0.5
Normal Deflection (in.=2.54 cm)
F
o
r
c
e

p
e
r

A
r
e
a
-
D
e
p
t
h
,

P
/
(
w
*
h
*
d
)

[
l
b
/
i
n
3
]
(c)
0 0.2 0.4 0.6 0.8 1
0
50
100
150
200
250
300
Lateral Deflection (in.=2.54 cm)
P

(
k
i
p
=
4
.
4
4
8

k
N
)

o
n

1
-
f
t

s
p
a
c
i
n
g
Depth
(ft=0.3m)
50
25
1
|
|
|
|
|
|
(b) A Ab bu ut tm me en nt t
( (S SO OL LI ID D6 65 5) )
P Pi il le e
( (B BE EA AM M2 24 4) )
W Wi in nk kl le er r
S So oi il l S Sp pr ri in ng gs s
( (C CO OM MB BI IN N3 39 9) )

z
KzA
T

(a)
Figure 2. (a) Winkler Model of Horizontal Soil; Support to Substructure; (b) Lateral Force-
Translation Curves for Granular Pile Backfill for Tributary Area at Varying Depth [9]; (c)
Force per Area-Depth vs. Translation for Well Graded, Dense Backfill Adjacent to Abutments
[10].

248
designed to handle reinforced or plain concrete behavior, and includes capability
to simulate cracking and crushing. Reinforcement may be modeled as smeared
throughout the element, requiring input of a reinforcement volume ratio. In the
model shown in Figure 3(a), difference in color indicates variation in the
reinforcement ratio; the amount of steel in the deck increases toward the central
pier to account for negative moment induced there.
The of the superstructure, and thus movement of the supporting elements, are
significant. These changes may be due to change in temperature, as well as time-
dependent effects of creep and shrinkage. Creep coefficients of high performance
concrete tend to be less than for traditional concrete mixes [11], while
autogeneous shrinkage is more pronounced, and thermal strain behavior follows
closely with that of traditional concrete.

The majority of creep and shrinkage deformations in the precast girders were
complete by the time the girders were made continuous at the time of casting the
deck slab. Since the girders bear the weight of the slab, but ensuing creep is
undertaken by the composite section, creep due to deck casting was not
considered to have a significant effect. Differential deck slab shrinkage,
inherently shortening the superstructure, directly affects the integral abutment
system; temperature change acts similarly on the integral substructure. Here, the
investigation for its response to positive and negative temperature change is
presented, with omission of results on shrinkage strains, since they generally
produce results similar to that of decrease in temperature.
THERMAL LOADING
Due to the relatively low thermal conductivity of concrete, the distribution of
temperature along the depth of a concrete superstructure is markedly nonlinear. A
great deal of precast girders were simulated using the 2-node elastic beam element
BEAM4, located along the centroid of the precast members. The 156-ft (47.5-m)
girders are discretized into 2-ft (0.6-m) increments. BEAM4 elements were also
Figure 3. Finite Element Model (a) Full Model Indicating Variation in Reinforcement; (b)
Schematic Showing the Modeling Approach for the Prestressed Concrete Composite
Superstructure.
REPRESENTED
BT-72 GIRDER
GIRDER
ELEMENTS
RIGID LINK
ELEMENTS
SOLID DECK
ELEMENTS
(b) (a)

249
used, with very high stiffness as rigid links, to accomplish the composite
connection (also at 2-ft increments) between solid deck elements and line beam
elements located along the girder centroid, as shown in Figure 3(b). These rigid
link elements are also used at the ends of the girders to connect the beams to the
top and bottom of the endwall. The roller support at the pier/diaphragm
interface is modeled via embedded anchor bolts through the centerline of the pier
and diaphragm, again using the element BEAM4, along with spring elements at
the interface to simulate the presence of relatively low resistance to rotation.
Finally, BEAM4 elements are used to model the steel cross-bracing members
between girders at third-points of each span.
Simulated Loads
In jointless bridge analysis, effects that primarily influence changes in size
and/or shape study into this phenomenon has been performed, resulting in design
guidelines [12] and an approximated stepwise linear temperature gradient for use
in design [13].

Considering the cross section of the tributary composite girder in Figure 4(a),
application of the AASHTO design increased temperature profile [Figure 4(b)]
would induce axial and bending strains according to

=
t
b
y
y
dy y b y T
A
) ( ) (
1

(2-a)

=
t
b
y
y
ydy y b y T
I
) ( ) (
1
(2-b)
in which cross-sectional area of the transformed section A = 1380 in.
2
(8903
cm
2
), moment of inertia I = 1,080,000 in.
4
(0.450 m
4
), y-axis ordinate at top of
section y
t
= 28 in. (71 cm), and at section bottom y
b
= -53.75 in. (-136.5 cm),
coefficient of thermal expansion , and T(y) and b(y) are the thermal gradient and
width function, respectively, along the y-axis. The strain resulting from the
AASHTO design gradient is presented in Figure 4(c). Design thermal gradient
and resulting strains are presented for temperature decrease in Figure 4(d-e).

250
The challenge in applying thermal loading to the FE model is that thermal
gradients can be induced in the current setup only in a two-increment stepwise
linear variation with depth: linearly through the deck, and again linearly through
the precast girder. Simulation of a given thermal gradient therefore involves
choosing temperature values T
1
at the bottom of the girder, T
2
at the girder/deck
interface, and T
3
at the top of the deck, such that differences in the effects are
minimized. Most importantly, movements induced by the thermal gradient should
be the same. This was accomplished by setting the results of Equations 2 equal
for the actual thermal gradient T
actual
(y) and for simulated approximation
T
FE
(y, T
1
, T
2
, T
3
) for the FE model, as follows:

=
t
b
t
b
y
y
FE
y
y
actual
dy y b T T T y T dy y b y T ) ( ) , , , ( ) ( ) (
3 2 1
(3-a)

=
t
b
t
b
y
y
FE
y
y
actual
ydy y b T T T y T ydy y b y T ) ( ) , , , ( ) ( ) (
3 2 1
(3-b)

Solving Equations 3 for, say T
2
and T
3
in terms of T
1
leaves one unknown
value, for which an optimal solution can be found using the relationship for beam
bending stress
) ( ) ( y T y y + =
(4)

and solving for the remaining term T
1
by minimizing the difference between
the stress profiles, as in
0
) , , , ( ) (
1
3 2 1
1
=
dT
T T T y T y T d
T
FE actual

(5)

50 75 100
0
12
24
36
48
60
72
L
o
c
a
t
i
o
n
,

i
n
.
T(y),
o
F
2 3 4 5
0
12
24
36
48
60
72
T
10
-4
T
actual
o
+
.
y
T
FE
-100 -75 -50
0
12
24
36
48
60
72
T(y),
o
F
-5 -4 -3 -2
0
12
24
36
48
60
72
T
10
-4
(c)
(a)
(d) (b) (e)
Figure 4. (a) Transformed Composite Girder Section; (b) Design Positive Thermal Gradient
(AASHTO, 1998) with Base Temperature 50F; (c) Thermal Strain Imposed by Theoretical and
Equivalent FEM Thermal Gradient, and Resulting Deformation Strain; (d) Design Negative
Thermal Gradient [14] with Base Temperature -50F; (e) Associated Strains.

251
Figure 4(c) and (e) present the gradients derived for simulating the AASHTO-
proposed design gradients, respectively for increase and decrease.
RESULTS
Results of simulated design thermal effects, presented above, are summarized
in Figure 5 in the form of abutment/pile deflection, and pile moment distribution
with depth. In all cases, the trends shown are averages for the group of (10) pile
lines. Comprehensive results of this study are reported elsewhere [14].

The trends are defined by their respective loading, either positive (+T) or
negative (-T) AASHTO design thermal gradient for concrete, and by their angle
of skew , in degrees. Abutment 1 is the eastern abutment, found at the left side
of the image of Figure 1. Note again that it is characterized by a significantly
deeper abutment beam than its counterpart, and average pile length is less than for
the Abutment 2 piles. Deflections defined in the x-direction are perpendicular to
the abutment, and those in the z-direction are parallel to it. Therefore the skew is
the angle between the axis of the bridge superstructure and this fixed coordinate
system. The pile sections are oriented with their principal axes aligned with this
z-axis (along the face of the abutment) as is the standard bridge construction
practice in Tennessee. Thus moments M
z
are about the piles principal axis, and
those denoted M
y
are about the weak axis. Finally, elevations reported in the
graphs are reported with a zero baseline at the top of the abutments.

252
DISCUSSION AND CONCLUSIONS
The deflection results show the effect of the difference in abutment beam
heights. Under thermal expansion, Abutment 2 translates about 3 times more than
Abutment 1, due to the significantly greater surface area bearing against the soil.
In thermal contraction, net movement is found again to tend toward the smaller
abutment side, with the larger abutment exhibiting most of the movement. In this
case the smaller system is actually stiffer, since the piles are engaged with the soil
very near the superstructure, and passive pressure is not a factor in this scenario.
These results reinforced field observations: a change in temperature from the
bridges setting temperature was observed to result in rotation of the intermediate
pier toward Abutment 2, whether that temperature change was an increase or
decrease.

Off-axis translations and forces are increased with increase in skew. The
magnitude of such increase for the 55 skew case is on the order of the in-line
deformations and forces. With the construction practices used in Tennessee,
forces acting at a skew to the strong axis of the bridge introduce flexibility to the
Figure 5. Average Abutment/Pile Deflection and Pile Moment Distribution in Both Abutments
Under Positive and Negative Thermal Gradients With Varying Bridge Skew
-1 -0.5 0 0.5 1
-21
-17.5
-14
-10.5
-7
-3.5
0
3.5
7
x
(in.)
E
l
e
v
a
t
i
o
n

(
f
t
)
Abutment 1
-T,=0
+T,=0
-T,=27
+T,=27
-T,=55
+T,=55
-1 -0.5 0 0.5 1
-21
-17.5
-14
-10.5
-7
-3.5
0
3.5
7
Abutment 2
x
(in.)
-250-150 -50 50 150 250
-27
-23
-19
-15
-11
Abutment 1
M
z
(kip-ft)
-150 -50 50 150 250 350
-20
-16
-12
-8
-4
Abutment 2
M
z
(kip-ft)
-1 -0.5 0 0.5 1
-21
-17.5
-14
-10.5
-7
-3.5
0
3.5
7
z
(in.)
E
l
e
v
a
t
i
o
n

(
f
t
)
-1 -0.5 0 0.5 1
-21
-17.5
-14
-10.5
-7
-3.5
0
3.5
7
z
(in.)
-350-250-150 -50 50 150
-27
-23
-19
-15
-11
M
y
(kip-ft)
-150 -50 50 150 250
-20
-16
-12
-8
-4
M
y
(kip-ft)

253
integral system; this acts to relieve stresses associated with bridge expansion and
contraction. Another behavioral aspect associated with the skewed bridge case is
tensile stresses at the obtuse corner of the concrete deck. Some cracking was
noted in the actual bridge at this location, and the results of the simulations were
no different.

From the driven pile records for the Porter Road Bridge, it was found that pile
lengths vary significantly from one pile to the next, with a range of from 5 to 60
ft. The actual pile lengths are reflected in the model, however the effects are not
readily apparent from the reported results since only average values are used here.
The significance is not that the pile lengths vary, which is to be expected, but that
some piles only reach 5 ft beyond the bottom of the abutment. Short piles may
represent a savings and possibly the illusion of sound construction practice to
the contractor, and a bridge construction should be carefully monitored to assure
compliance with specifications.

It has been shown here how the factors including bridge skew, depth of
abutment beam and depth of piles influence the balance of integral bridge
deformation and loading under volumetric changes. Another principle factor at
play is stiffness of the abutment and pile backfill. It is of interest to provide
guidelines for design of integral bridges given geometric attributes and stiffness
of backfill; furthermore it is key to the success of integral bridges that
construction is held to the specifications designed for. When these challenges are
met, the public will get the full practical benefit of the integral abutment bridge.
REFERENCES
1. Burke, M. P. J. (1989). Bridge Deck Joints. NCHRP Synthesis Report No. 141,
Transportation Research Board, Washington, DC.
2. Wasserman, E. P. (2001). "Design of Integral Abutments for Jointless Bridges."
Structure, 24-33.
3. Burke, M. P. J. (1990). "Integral Bridges." Transportation Research Record, 1275, 53-61.
4. Wolde-Tinsae, A. M., and Greimann, L. F. (1988). "General Design Details for Integral
Abutment Bridges." Civil Engineering Practice, Journal of the Boston Society of Civil
Engineers/ASCE, 3(2), 7-20.
5. Soltani, A. A., and Kukreti, A. R. (1992). "Performance Evaluation of Integral Abutment
Bridges." Transportation Research Record, 1371, 17-25.
6. Kunin, J., and Alampalli, S. (2000). "Integral Abutment Bridges: Current Practice in
United States and Canada." ASCE Journal of Performance of Constructed Facilities,
14(3), 104-111.
7. Basu, P.K., and Knickerbocker, D. (2002). High Performance Concrete Bridges, Final
report of TN DOT Project No. TNSPR-RES1162, Nashville, TN.
8. SAS. (2002). ANSYS 6.1 Users Manual. Swanson Analysis Systems. Houston, PA.
9. American Petroleum Institute. (1993). Recommended Practice for Planning, Designing,
and Constructing Fixed Offshore Platforms Working Stress Design, 20th edition, API
RP2A-WSD, Washington, D.C.
10. National Cooperative Highway Research Program (NCHRP). (1991). Manuals for the
Design of Bridge Foundations. R. M. Barker, J. M. Duncan, K. B. Rojiani, P. S. K. Ooi,
C. K. Tan, and S. G. Kim. Eds. Rep. 343, Transportation Research Board, Washington,
DC.

254
11. Roller, J. J., H. G. Russell, R. N. Bruce, and B. T. Martin. Long-Term Performance of
Prestressed, Pretensioned High Strength Concrete Bridge Girders. PCI Journal, Vol. 40,
No. 6, November-December 1995, pp. 48-59
12. Imbsen, R. A., Vandershaf, D. E., Shamber, R. A., and Nutt, R. V. (1985). Thermal
Effects in Concrete Superstructures. NCHRP Report No. 276, Transportation Research
Board, Washington, DC.
13. AASHTO. (1998). LRFD Bridge Design Specifications, Customary Units - Second
Edition, American Association of State Highway and Transportation Officials,
Washington D. C.
14. Basu, P.K., and Knickerbocker, D. (2004). Behavior of Jointless High Performance
Concrete Bridges, Final report of TN DOT Project No. TNSPR-RES 162, Nashville, TN.

255
SESSION IV:
CONSTRUCTION PRACTICES

256

257
Field Study of Integral Backwall with Elastic Inclusion

Edward J. Hoppe, Ph.D., P.E.
Senior Research Scientist
Virginia Transportation Research Council
530 Edgemont Road
Charlottesville, VA 22903

Tel.: 434-293-1960
Fax: 434-293-1990
E-mail: Edward.Hoppe@vdot.virginia.gov

ABSTRACT

An integral bridge 100 m (331 ft) long was constructed with a layer of
elasticized expanded polystyrene (EPS) 0.25 m (10 in) thick attached to the
backwall. The bridge has been monitored for a period of 5 years following
construction. Significantly attenuated lateral earth pressures have been recorded
at the backwall. The settlement of the approach fill has been acceptable. Field
data indicate that the elasticized EPS layer has been functioning effectively in
allowing the superstructure to interact with the adjoining backfill material.

INTRODUCTION

Jointless construction is considered an effective design option to reduce bridge
maintenance costs. It is also regarded as being well suited to resist seismic loads
(1). Although these attributes make the integral bridge an increasingly popular
choice, there are soil-structure interaction issues unique to this type of design that
remain unresolved. Of particular concern is the excessive settlement of approach
embankments, resulting from thermally induced cyclic movements of the
superstructure. In many cases, rectifying this condition can be fairly expensive
because the integral bridge approach slab (if provided) cannot be overlaid with
pavement.

To address this problem, the Virginia Department of Transportation (VDOT)
conducted a study designed to test the feasibility of using elastic inclusion at the
integral backwall. The design was completed in mid-1997, and the bridge was
opened to traffic in October 1999. Field monitoring has been conducted
continuously for 5 years.

PURPOSE AND SCOPE

The objective of this study is to test the concept of an elastic inclusion serving
as an interface between a structure and a stiff backfill. The elastic inclusion is

258
intended to accommodate thermally induced lateral movement of the
superstructure with the main purpose of reducing the approach fill settlement. In
addition, the role of the elastic inclusion is to reduce earth pressures acting on the
structure. Long-term field monitoring includes automated measurements of earth
pressures and the resulting strains in steel girders. Thickness of the EPS layer is
also measured periodically to determine if the material remains elastic over time.

METHODOLOGY

Description of Bridge

The project involves a replacement bridge situated on Route 60 over the
Jackson River in Alleghany County, Virginia. The integral backwall (semi-
integral) bridge is 100 m (331 ft) long and 16.6 m (54.5 feet) wide (overall), with
three-span continuous steel plate girders, and no skew. Fixed bearings are
provided over two piers, and expansion bearings are installed at abutments. There
are no approach slabs constructed at the bridge. The average daily traffic (ADT)
is 12,771 vehicles with 7% trucks (2002 traffic data).

The experimental detail involves a layer of elasticized EPS 0.25 m (10 in)
thick placed on the back of integral backwall to absorb a limited range of elastic
movement without impacting the adjoining embankment fill. The EPS material
terminates at 0.76 m (2.5 ft) below grade. This distance was selected arbitrarily to
allow space for placement of one earth pressure sensor for measuring backfill
stress directly on the backwall. Two other pressure sensors were installed at the
backwall, behind the EPS layer. A separation geotextile was placed on the EPS
layer to prevent damage from the adjoining granular backfill material. Select
backfill is classified as VDOT Type I-21B, consisting of a dense-graded
aggregate with 100% of particles passing the 50 mm (2 in) sieve and 4% to 7% of
particles passing the No. 200 sieve. It was compacted in lifts of approximately
0.20 m (8 in) each with a hand-operated compactor in the proximity to the
backwall. The underlying soil is approximately 10 m (33 ft) of clay and silty sand
fill with N-values ranging between 5 and 13, underlain by 3 m (10 ft) of clayey
sand natural soil with an N of 50, and limestone bedrock. Abutments are
supported on steel piles (HP 10x42) driven to bedrock.

The elastic inclusion is composed of a layer of glued polystyrene porous
drainage material 0.10 m (4 in) thick laminated with a layer of elasticized
polystyrene block 0.15 mm (6 in) thick. According to the manufacturer, the
stress-strain behavior of both layers is essentially identical. The material cost was
quoted at $21.53/m
2
($2.00/ft
2
) in 1997.

259

Instrumentation

Electronic instrumentation consists of earth pressure cells installed at the
backwall, strain gages attached to girder flanges, and linear displacement
transducers and tiltmeters placed on the backwall. Earth pressure cells and strain
gages are of the vibrating-wire type. All sensors are interfaced with Campbell
Scientific CR-10X dataloggers, sampling every hour. Figure 1 shows the location
of earth pressure cells at the integral backwall. These cells are positioned at 0.63
m (2.08 ft), 1.12 m (3.67 ft), and 1.60 m (5.25 ft) below the grade level and at a
6.3-m (20.6-ft) horizontal distance from the wingwall face. The uppermost cell
(sensor 1) is in direct contact with the backfill material. An EPS layer covers the
remaining cells (sensors 2 and 3). All pressure cells are recessed in the backwall,
with the sensing surface flush with the backwall surface.

A simple telltale gage was installed to measure the thickness of elastic
inclusion in service. This gage consists of a 200 by 200 by 2 mm (8 by 8 by 0.09
in) aluminum plate attached to the face of EPS, with a connecting stainless steel
threaded rod of 5 mm (3/16 in) diameter, protruding through the opening in the
backwall, as shown in Figure 1. The gage is located at approximately the same
depth as earth pressure sensor 3, but with a small horizontal offset of 0.45 m (18
in). Periodic measurements of the length of the protruding rod (distance x)
indicate the magnitude of EPS compression. These manual measurements are
typically conducted during the hottest and coldest times of the year to reveal the
full range of EPS working strains and to detect creep. All ambient air
temperatures are recorded under the deck, in the proximity to the backwall. These
temperatures are typically more moderate (cooler in the summer and warmer in
the winter by about 8 to 10 degrees) as compared to topside readings.

260

Figure 1. Integral Backwall Instrumentation.

RESULTS

The backfill behind the instrumented abutment was compacted on 8/11/1999.
Immediately after compaction, the EPS strain was 11%, resulting from the lateral
pressure exerted by the backfill material, coupled with compaction-induced
stresses. Approximately 2 weeks later, on 8/25, the strain relaxed to 8%, with the
ambient air temperature about 3 degrees lower. Subsequently, the observed
pattern was that of EPS strain increasing and decreasing with the rising and
falling of the air temperature.

Figure 2 shows the strain-temperature-time behavior of the elastic inclusion
from November 1999 to January 2005. The graph shows data points reflecting
measurements made during the highest, lowest and mid-range of recorded
ambient air temperatures. Bars represent the EPS strain, and triangles mark the
corresponding air temperature under the bridge. The range of working strains
from 4% to 13% corresponds to 23 mm (0.9 in) of backwall movement. Numbers
on the time axis indicate days elapsed from the time of backfill compaction to the
end of the following year. The latest reading was taken at 1,990 days, on
1/19/2005.

261

Figure 2. Strain-Temperature-Time Behavior of Elasticized EPS.

The largest recorded earth pressures at sensors 2 and 3 were 19.7 kPa (411
psf) and 22.5 kPa (470 psf), respectively. These magnitudes were reached
repeatedly during prolonged periods of summer hot weather, with air
temperatures hovering around 30
o
C (90
o
F).

The largest sensor 1 pressure reading of 417.8 kPa (8,723 psf) was recorded
on 4/18/2004. Figure 3 shows earth pressure data for the period 4/13/2004 to
4/20/2004. As seen from the air temperature record, this was a period of a rapid
warm-up from approximately 10 to 26
o
C

(50 to 80
o
F), following a prolonged
period of cool weather. At the same time, pressures registered by sensors 2 and 3
(behind the EPS) were much lower.

262

Figure 3. Earth Pressures from 4/13/2004 to 4/20/2004.

In the following week the air temperature stabilized in a relatively narrow
range of approximately 18 to 26
o
C (64 to 80
o
F). The maximum pressure at
sensor 1 was only a fraction of the previous spike, and pressures at other sensors
reached similar levels at the corresponding air temperatures. It appears that after
the initial backfill resistance was broken by the expanding superstructure, the
subsequent peaks were greatly reduced. This pattern of earth pressure behavior
was observed at other times. The next largest earth pressure reading at sensor 1
was 339 kPa (7,077 psf), recorded on 4/16/2002.

Approach elevations were taken periodically between 11/1/1999 and
12/16/2004. Since the bridge was opened to traffic in 1999, the approach
pavement was patched only once in the immediate vicinity of the bridge, because
of excessive differential settlement. The remedial work, consisting of placing
approximately 50 mm (2 in) of plant mix over a strip of roadway 0.6 m (2 ft)
wide, was performed on 9/5/2002. Embankment elevation records point to the
approach zone most prone to settlement (although still manageable) extending to

263
about 1.5 m (5 ft) beyond the backwall, where approximately 11 mm (0.43 in) of
settlement occurred between 12/19/2002 (following asphalt patching) and
12/16/2004. No excessive differential settlements were observed in the remaining
segment of the bridge approach, with the maximum of 14 mm (0.6 in) recorded at
12.6 m (41 ft) beyond the backwall during the 5-year monitoring period.

Laboratory direct shear tests conducted at the interface of granular backfill
and a concrete specimen yielded a residual friction angle of 31 degrees. The
concrete specimen was coated with the same type of waterproofing compound as
was the integral backwall (SurePoxy LMLV). A residual friction angle of 35
degrees was measured for the uncoated concrete-backfill interface. Tests were
performed using the Large Direct Shear Box (2) with internal dimensions of 635
by 406 by 25 mm (25 by 16 by 1 in).

DISCUSSION

The magnitude of the actual passive earth pressure acting on the integral
backwall in service has been a subject of a debate. It is important to recognize
that depending on the theory used, estimates can vary widely. Wasserman (3)
advocates the use of the Rankine theory to calculate passive pressure in a
conservative way. Duncan and Mokwa (4) propose the Log Spiral theory and
claim that passive pressures can induce large loads in integral bridges.
Thippeswamy et al. (5) recommend neglecting earth pressure loads altogether in
the analysis and design of jointless bridges. More field studies are needed to
resolve this issue. Passive loads can be of concern at relatively tall integral
abutments. The Canadian Foundation Engineering Manual (6) states that a
compacted backfill requires very little movement to generate fully passive
conditions.

In addition to reducing passive earth pressures, the purpose of an elastic
inclusion is to absorb cyclic backwall movements without disturbing the adjacent
backfill material, which results in amplified settlement. Approach settlement
behind a bridge abutment can be minimized by a thorough compaction of a well-
graded granular backfill; however, the resulting stiffness of such material can
result in a generation of substantial passive pressures at the integral structure.
Consequently, the design has either to accommodate these elevated stresses or
incorporate a low-stiffness flexible layer at the backwall-backfill interface. For
optimum performance, it is essential that the backwall inclusion remains elastic in
response to diurnal and seasonal movements of the superstructure.

Carder and Card (8) identified a number of potentially applicable materials for
use as compressible layers. Selected candidates include polystyrene,
polyethylene foam, geocomposites and rubbers with the suitability to the task to
be verified through further research. Some of these materials were subsequently
subjected to extensive laboratory testing by Carder et al. (9). Unfortunately, no

264
data were reported on the elasticized expanded polystyrene. The authors
concluded that it has not been investigated further as it is expensive to produce
and for this reason is not likely to be a cost effective solution for use on
construction sites. This opinion was presumably based on the analysis of the
U.K. construction environment only.

The EPS is a unique lightweight material, with a density of only about 1% that of
a traditional earth fill. It is typically manufactured in rectangular blocks. In the
U.S. practice, the term geofoam is used as a synonym for the EPS block geofoam.
A colloquial term Styrofoam is technically incorrect, since it is the registered
trademark of a particular product line of Dow Chemical.

The report by Stark et al. (10) is intended to synthesize the state of the practice
of geofoam use in roadway embankments (although it contains no references to
the elasticized EPS). It identifies the need for long-term stress-strain-time-
temperature testing of geofoam and recommends creep test durations of at least 20
months to extrapolate performance for a 50-year design life.

The EPS has been routinely used as a lightweight fill in embankment
construction when low bearing capacity soils are encountered. In Norway, more
than 100 projects involving the use of EPS have been successfully completed
since 1972 (11). Some of these projects involve placement of EPS block at bridge
approach embankments to mitigate differential settlement.

Regular EPS material exhibits a linear-elastic stress-strain relationship up to
about 1% strain (12). EPS that has been strained beyond the yield point and then
unloaded is considered elasticized. EPS that is elasticized by temporary loading
to between 60% and 70% strain subsequently exhibits linear-elastic behavior up
to approximately 10% strain and linear (proportional) stress-strain behavior up to
about 30% strain. Elasticized EPS has greater stiffness in the direction
perpendicular to the compressed axis, which is advantageous for the support of
the overlying pavement section. These properties make the elasticized EPS
potentially suitable for high-strain applications.

The results of laboratory direct shear tests on the backfill-backwall interface
indicate the residual interface friction angle () of approximately 31 degrees.
Assuming that tan =
2
3
tan , it leads to an estimated angle of internal friction
() of 42 degrees for the backfill material. The relationship between the
coefficient of passive earth pressure (K
p
), and , as stated by Powrie (13), is as
follows:

K
p
= {[1 + sin
cos( + )] / [1 sin
]} x
( + ) tan
(1)

where sin = sin / sin
.

265
The equation leads to an estimated K
p
of 11.9. For comparison, passive earth
pressure coefficients were also computed by Rankine, Coulomb, and Log Spiral
theories, with the resulting values of 5.0, 34.8, and 16.7, respectively. The wide
range of K
p
calculated by the different methods illustrates the difficulty of
providing the correct value of earth pressure for design.

The pressure spike of 417.8 kPa (8,723 psf), registered at sensor 1,
corresponds to the passive earth pressure coefficient of 27.3, based on the
estimated bulk unit weight of the backfill material of 24.1 kN/m
3
(153 pcf). This
translates into a substantial, although transient, passive load acting on the
abutment. The pattern of observed backfill-structure interaction is that of a
significant earth pressure buildup during a period of a sudden warm-up,
corresponding to the superstructure expanding relatively quickly against the
adjoining backfill. It is indicative of a stress-strain behavior of a dense granular
material, reaching a high peak strength (and a correspondingly high internal
friction angle) before decreasing to a residual strength. The magnitude of 417.8
kPa (8,723 psf) and the corresponding K
p
of 27.3 should be kept in perspective.
This event occurred only once in the 5-year monitoring period. A typical pattern
of recorded earth pressures implies that the Log Spiral theory offers a reasonable
approximation of K
p
; however, the occurrence of high transient earth pressures
cannot be ignored and should be accounted for in the bridge design process when
no elastic inclusion is present.

In contrast, earth pressures are more attenuated and more uniform at the
backwall section covered with the elasticized EPS. Maximum ratios of horizontal
to vertical earth pressures at sensors 2 and 3 are calculated at 0.7 and 0.6,
respectively. Thus, the presence of the elastic inclusion results in a substantial
decrease in mobilized earth pressures. With the backwall height of 1.8 m (6 ft),
the observed movement range of 23 mm (0.9 in) corresponds to a displacement-
height ratio of 0.0125. It is indicative of fully passive earth pressure conditions in
dense granular materials (14).

Field measurements of the EPS thickness show that in the 5 years of
monitoring the material remained elastic in the working strain range of 4% to
13%, with no evidence of appreciable creep. It is important to recognize that this
type of performance cannot be attained with a regular (non-elasticized) EPS
geofoam.

The lateral extent of observed settlements indicates that a relatively short
approach slab of about 1.5 m (5 ft) would adequately serve the purpose of
providing a grade transition. A shorter approach slab would be easier for the
superstructure to push and pull during cyclic movements, if the embankment
settlement and the corresponding slab rotation are relatively small. Large
anticipated settlements require the use of long approach slabs.

266
The use of approach slabs at integral bridges can be a particularly troublesome
maintenance issue if excessive settlement occurs. The solution is to provide a
better quality backfill or to eliminate the slab and rely on periodic resurfacing of
the roadway. Recent U.K. practice points to a diminished use of approach slabs,
and the recently constructed bridges appear to be performing well (15).

It is very likely that the relatively small embankment settlements were
attributable to the combined action of the high-quality backfill material and the
elastic inclusion. Placement of the select backfill alone could have generated
excessive lateral stresses attributable to potentially high backfill stiffness. The
observed performance of the elasticized EPS also suggests that it can be used to
protect the integral and non-integral bridge abutments from lateral pressures
exerted by the compaction equipment. Current VDOT specifications stipulate
that only lightweight compactors be used in the direct proximity of a wall, but this
practice often results in excessive approach settlements because of inadequate
compaction (bump at the end of the bridge). The presence of an elastic inclusion
would likely help to dissipate lateral stresses induced by a heavier (intermediate)
equipment while allowing for a more thorough backfill compaction. Matsuda et
al. (16) propose that the EPS material be used to reduce earth pressure during
construction and to absorb dynamic loads attributable to earthquakes.

Where approach slabs are used, the EPS layer can be extended up to the
underside of the slab. If no approach slab is constructed, the EPS should
terminate below the pavement section, typically at the approach slab seat
elevation (0.46 m [18 in] for VDOT design). The German code of practice (17)
with regard to lightweight embankments constructed with regular EPS requires
that the thickness of roadway material in contact with the upper surface of the
EPS block should not be less than 0.30 m (1 ft), to allow for adequate compaction.

CONCLUSIONS

The elastic inclusion consisting of a layer of elasticized EPS 0.25 m (10
in) thick has performed effectively during the 5 years of field monitoring. The
material remained elastic in the observed working strain range of 4% to 13%.
Field tests indicate reduced earth pressures and approach settlements at the
integral bridge.

The function of the elastic inclusion is to interface a stiff backfill mass
with a structure without generating excessive stresses and settlements. The
presence of a well-compacted select backfill material at the bridge approach is
essential for good results (low maintenance). The use of inferior backfill
negates advantages provided by the elastic inclusion.

Approach elevation data indicate a zone of increased settlement in a close
proximity to the backwall. It appears that relatively short approach slabs, if

267
required, are satisfactory to provide a grade transition. Shorter approach slabs
are easier for the superstructure to push and pull during cyclic movements,
and will exert less stress on the backwall if they settle. The alternative
solution is to resurface the approach roadway occasionally. The bridge under
study performed well without approach slabs.

In view of a relatively limited experience with the long-term behavior of
the elasticized EPS, it may be prudent to restrict its design strain range to 10%
until more field data become available. Long-term creep of the material is of
primary concern.

RECOMMENDATIONS

1. Elasticized EPS geofoam layer should be installed on integral bridge
backwalls and wingwalls, where elevated earth pressures and approach
settlements are a concern.

2. Select backfill material should be placed against a wall covered with the
elasticized EPS. The lateral extent of the select backfill should be based
on the estimated passive failure surface.

3. Medium-heavy roller compactor should be used to compact backfill
material adjoining the wall covered with the elasticized EPS. The elastic
inclusion should be designed to absorb additional strains caused by the
compactive effort.

4. Elasticized EPS should be designed not to exceed 10% strain. Greater
working strains should be explored through instrumentation and long-term
monitoring of select structures.

5. Short approach slabs of approximately 1.5 m (5 ft) should be used where
relatively small embankment settlements are anticipated.

6. The option of having no approach slab should be examined where
practical.

7. Other elastic inclusion candidate materials (alternative products with
comparable properties) should be evaluated through a series of laboratory
and field tests.

ACKNOWLEDGMENTS

This study was supported by the Federal Highway Administration. The author
expresses his gratitude to Art Wagner and Linda DeGrasse of the Virginia

268
Transportation Research Council for their helpful field support and data
processing. Thanks are also expressed to the VDOT Staunton District bridge
designers and VDOT Jointless Bridge Committee members for their valuable
guidance and inspiration throughout this project.

REFERENCES

1. Soltani, A., and Kukreti, A. (1982). Performance evaluation of integral abutment
bridges. Report No. 179. Transportation Research Board, Washington, D.C.

2. Shallenberger, W.C., and Filz, G.M. (1996). Interface strength determination using a
large displacement shear box. Proceedings of the Second International Congress on
Environmental Geotechnics. Osaka, Japan, 5-8 November 1996.

3. Wasserman, E.P. (2001). Design of integral abutments for jointless bridges. Structure,
May 2001, pp. 24-33.

4. Duncan, J.M., and Mokwa, R.L. (2001). Passive Earth Pressures: Theories and Tests.
Journal of Geotechnical and Geoenvironmental Engineering, Vol. 127, No. 3, pp. 248-
257.

5. Thippeswamy, H.K., GangaRao, H.V.S., and Franco, J.M. (2002). Performance
Evaluation of Jointless Bridges. Journal of Bridge Engineering, Sept.-Oct. 2002, pp.
276-289.

6. Canadian Geotechnical Society (1992). Canadian Foundation Engineering Manual (3
rd

edition). Vancouver, Canada.

7. Hoppe, E.J., and Gomez, J.P. (1996). Field Study of an Integral Backwall Bridge.
Virginia Transportation Research Council, Charlottesville.

8. Carder, D.R., and Card, G.B. (1997). Innovative structural backfills to integral bridge
abutments. Report No. 290. Transport Research Laboratory, Crowthorne, Berkshire,
U.K.

9. Carder, D.R., Barker, K.J., and Darley, P. (2002). Suitability testing of materials to
absorb lateral stresses behind integral bridge abutments. Report No. 552. Transport
Research Laboratory, Crowthorne, Berkshire, U.K.

10. Stark, T.D., Arrelano, D., Horvath, J.S., and Leshchinsky, D. (2004). Geofoam
Applications in the Design and Construction of Highway Embankments. NCHRP Web
Document 65 (Project 24-11). Transportation Research Board, Washington, D.C.

11. Frydenlund, T.E. (1991). Expanded Polystyrene: A Lighter Way Across Soft Ground.
Veglaboratoriet, Oslo, Norway.

12. Horvath, J.S. (1995). Geofoam Geosynthetic. Horvath Engineering, P.C. Scarsdale,
N.Y.

13. Powrie, W. (1997). Soil Mechanics: Concepts and Applications. E & FN Spon, London,
U.K.

269
14. Cernica, J.N. (1995). Geotechnical Engineering: Soil Mechanics. John Wiley and Sons,
New York.

15. Lock, R.J. (2002). Integral Bridge Abutments. Master of Engineering Project Report.
University of Cambridge, Department of Engineering. Cambridge, U.K.

16. Matsuda, T., Ugai, K., and Gose, S. (1996). Application of EPS to backfill of abutment
for earth pressure reduction and impact absorption. Proceedings of International
Symposium on EPS Construction Methods. Tokyo, Japan, 29-30 October 1996.

17. Bundesanstalt fr Straenwesen BASt (1995). Merkblatt fr die Verwendung von
EPS-Hartschaumstoffen beim Bau van Straendmmen. Kln, Germany.

270
Plastic Design of Steel HP-Piles for Integral Abutment
Bridges

Preston A. Huckabee, MSCE, P.E., Gill Engineering Associates, Inc., M.ASCE

ABSTRACT

Integral abutment bridges (IAB) are slab or slab on stringer bridges
incorporating abutments monolithic with the superstructure and generally founded
on a single row of HP-piles to minimize the resistance to superstructure thermal
movement. The piles are embedded into the abutment concrete sufficiently to
rigidly connect at the pile head. The monolithic nature of the integral abutment to
bridge superstructure connection, forces the abutments to move with the
superstructure as it undergoes thermal movements. This movement of the
superstructure induces stresses within the piles that can force them to deform
inelastically. The rigid pile to abutment connection, and the fixity at the pile base
due to soil embedment, causes the piles to behave as fixed-fixed columns
translated at the top through a distance, .

In order to ensure adequate strength as piles undergo inelastic or plastic
deformations, the Massachusetts Highway Department (MHD), Bridge Section,
through its Bridge Quality Partnership, a public/private effort to develop bridge
design standards, has developed a methodology for sizing pile sections, based on
work by Greimann, et al, at Iowa State University [1]. This design methodology
utilizes a ductility based approach, in conjunction with the AASHTO Load Factor
Design (LFD) column strength equation to size pile sections for use as
foundations in integral bridges. This design methodology was released as part of
the December, 1999 revisions to the Bridge Design Guidelines of the
Massachusetts Highway Department Bridge Manual.

DISCUSSION OF PHYSICAL REQUIREMENTS

Integral abutment bridges are monolithic in nature in that the bridge
superstructure is embedded into the abutments. While this produces a very robust
3-sided frame with regard to the superstructure and abutments, it also demands a
foundation that is relatively compliant to permit superstructure expansion and
contraction without producing detrimental impact to the substructure. Typical
practice is to construct the abutment on a single row of HP-piles, usually 1 pile
per stringer. The piles are driven into a predrilled shaft several diameters deep of
loose stone, then into underlying in-situ soil (see Figure 1). This is intended to
permit abutment movements while minimizing resistance from the piles as the
loose material at the top reduces resistance to movement.

271
Integral abutment piles are embedded into the abutment that is rigidly attached
to the superstructure. As the superstructure moves the piles must go along for
the ride without impeding this movement and while remaining adequate to resist
the gravity loads applied (Figure 2). In general, IAB piles are required to exceed
yield in order to permit thermal expansion and contraction to occur in the
superstructure.

ROADWAY
SURFACE
TYPICAL
STRINGER
PRE-DRILLED
SHAFT
SUPERSTRUCTURE
EMBEDMENT
PILE EMBEDMENT
LOOSE STONE
DRIVEN
HP-PILE
T
Y
P
I
C
A
L

D
E
P
T
H
O
F

B
A
C
K
F
I
L
L
CONSTRUCTION
JOINT
INTEGRAL
ABUTMENT

Figure 1. Typical Integral Abutment

The movement of the pile top due to thermal expansion and contraction
induces rotations within the pile just under the abutment. The deflection curve of
a pile fixed at the top, and translated through a distance , produces reverse
curvature in the pile with the maximum moment occurring under the abutment, at
the top point of fixity. The next larger moment occurs at some depth below grade
as the deflection curve reverses. Finally, the pile reaches effective fixity in the
soil at the point of zero deflection. This may be idealized as a fixed/fixed column
translated through a displacement , with the equivalent length of beam found by
using classical equations [2] (see Figure 2).

In a typical 3-sided frame there is the potential for an unstable condition
occurring if a sufficient number of plastic hinges form in the columns (see Figure
3). However, in an integral bridge, plastic hinges may be allowed to form in the
columns without loss of stability provided the presence of the soil surrounding the
piles and the backfill behind the abutments. The soil around the piles must
prevent them from gross Euler buckling while they deflect due to thermal
movements. The soil behind the abutments must prevent the structure from
racking as it moves through its thermal cycles.

272

DISCUSSION OF DESIGN ASPECTS

There are 2 criteria that must be met to find a particular pile acceptable,
geotechnical and structural. The geotechnical aspect involves adequate length of
pile embedment, and/or adequate end bearing to transfer vertical loads to the soil.
The structural aspect involves the adequacy of the pile section to resist vertical
loads without gross Euler or local buckling, while undergoing bending due to
thermal movements and superstructure rotations at the pile head.

THERMAL MOVEMENT RANGE
DIRECTIONAL
MOVEMENT
SUPERSTRUCTURE
ROTATIONS
PILE HEAD
ROTATIONS
NEUTRAL
INTEGRAL
ABUTMENT
AND PILE
EQUIVALENT
FIXED/FIXED COLUMN
TRANSLATED
PILE HEAD
ROTATIONS
AXIAL LOAD P
CURVATURE REVERSES AT
LOCATION OF PILE FIXITY
PILE FIXED AT TOP
POINT
OF MAX
MOMENT
POINT OF 0
DEFLECTION
GENERAL LOCATION
OF NEXT LARGEST
MOMENT IN PILE
PILE HEAD
DEFLECTION
INCREASING
DEPTH

Figure 2. Integral Abutment Pile, Typical Pile Deflection Curve
and Equivalent Fixed/Fixed Column

There are three levels of rotation associated with local buckling involved in
the design of a steel section, full plastic rotation, inelastic buckling rotation, and
elastic buckling rotation. AISC 2
nd
LRFD Code Commentary B-5 for Local
Buckling [3] states that compact sections are capable of between 3 to 5 times the
rotation capacity at which a fully plastic stress distribution forms on a section
assuming linear elastic behavior up to full plastic; non-compact sections reach a
state of inelastic flange local buckling rotation, at which the flanges exceed yield
but buckle prior to a fully plastic section forming; and slender sections reach a
state of elastic flange local buckling prior to reaching yield. The choice of a
compact or non-compact pile section will mitigate flange local buckling. It was
decided that the use of slender sections as IAB piles would be avoided.

The geotechnical aspects are beyond the scope of this paper, except insofar as
they specifically affect the structural adequacy of the pile through soil structure
interaction. The soil must be adequate in stiffness to support the pile to prevent
Euler buckling, as determined by in-situ testing, and to prevent the structure from
racking, generally accomplished by densified backfill behind the abutments.

273
The structural aspects may be simplified conceptually as the combined effects
of axial load on the pile (gravity loads) and bending (thermal movements) and
whether or not the pile has adequate strength and stability to resist these. In order
to determine the acceptability of a particular pile, a methodology must be applied
to compare bridge movements and axial loads to the capacity of the pile section.

BACKFILL PRESSURE
ON ABUTMENT
RESISTS RACKING
LOCATION OF
HINGE IN PILE
POTENTIAL LOCATION
OF HINGE IN PILE
HINGE
LOCATION
SUPERSTRUCTURE
RACKING
PILE
SOIL SURROUNDING PILE MUST
RESIST GROSS BUCKLING
INSET

Figure 3. Three Sided Frame Integral Abutment
Inset Shows Frame Instability

DISCUSSION OF SOIL-STRUCTURE INTERACTION

There are three geotechnical aspects concerning the structural design of IAB
piles; adequate support for the pile to resist Euler Buckling, adequate backfill
behind the abutments to prevent the structure from racking and determination of
an equivalent pile length to simplify the analysis of the pile as a column. The
MHD Geotechnical Section performed a study [2] to determine the impacts for
each of the above criteria, as well as other aspects of design, for integral abutment
bridges.

The first soil-structure aspect is support of the pile to resist Euler Buckling.
For most soils except extremely soft soils such as peat, Euler buckling will not
occur in piles [2,4,5]. An excerpt from NCHRP, Synthesis of Highway Practice
42, Design of Pile foundations by Vesi [4], states very simply for Pile Buckling:

Experience shows that buckling of fully embedded piles is
extremely rare, even in soft soils, as long as they are capable of
supporting a pile in friction.

Thus only pile strength was viewed as being the controlling design criteria for
IAB piles. However, in the Geotechnical analysis for the bridge design under
consideration, MHD policy [6] requires a determination be made as to the

274
adequacy of the soil underlying the abutment to provide support to the piles to
resist buckling. If the soil is determined to be too soft, then the piles are to be
designed as free standing columns in accordance with the AASHTO [7] standards
and no ductility is permitted to be utilized.

The second soil-structure aspect of adequate backfill behind the abutments to
prevent structure racking was studied by the University of Massachusetts at
Amherst [8] using full-scale tests to find soil passive pressures when loaded by
integral abutments. The study consisted of construction of a full-scale abutment
wall loaded against backfill soils of various densities to determine induced
passive pressures.

The MHD Geotechnical Section [2] determined from the UMass study that a
good balance of induced passive pressure, which must be accounted for in the
design of the superstructure to substructure connection, and resistance to racking
could be achieved by standard MHD backfilling methods. Further, modeling at
the University of Massachusetts at Lowell found that densified backfill reduces
pile stresses when the bridge expands as the backfill resists this movement and
reduces rotations in the pile head from the abutments movement, thus reducing
the ductility demand on the pile in this direction [9].

The third soil-structure aspect is the determination of pile length to fixity in
soil. The MHD Geotechnical Section [2] provided a method for the simplified
analysis of the pile as an equivalent length column. A parametric study was
performed using the computer program COM624P to analyze various pile
sections and layered soil combinations. The piles were analyzed assuming fixed
heads and were moved various distances at the pile top ranging from to 2.
From this study lateral forces at the pile head, bending moments and deflections
were found.

These results were then used to find equivalent lengths of fixed/fixed columns
using the classical equations:

(1)

(2)

where: L
e(P)
= equivalent length of column for lateral load at the pile head
L
e(M)
= equivalent length of column for moment at the pile head
EI = flexural rigidity of the pile
D = thermal expansion
P = lateral load at top of pile
M = moment at top of pile
3
) (
12
Shear Head Pile
P
EI
L
P e
=
M
EI
L
M e
=
6
Moment Head Pile
) (

275

Multivariable equations were then derived combining the pile head deflection
, the pile stiffness EI/d and the equivalent length of pile Le, which was found as
the average for each pile analysis for Pile Head Shear, equation (1) and for Pile
Head Moment, equation (2). The use of equivalent column lengths simplifies
incorporation into finite element modeling or hand analysis by removing the non-
linear behavior from the analysis. This was found to be acceptable given the
inherent uncertainties associated with determining soil properties.

STRONG VS. WEAK AXIS PILE BENDING AND EFFECT OF SKEW

The MHD design standard requires that IAB piles be oriented for weak axis
bending, i.e. oriented along the centerline of the integral abutment. This is done
to minimize pile resistance to abutment movement, and to ensure that the piles
will not suffer flange local buckling regardless of unbraced length.

As was previously stated, compact and non-compact sections will yield prior
to local buckling occurring, however, this can be dependant upon unbraced length
depending upon the orientation of the member in bending. Compact section
criteria are valid for both strong and weak axis bending according to AISC LRFD
2
nd
Edition [3]. AISC Beam Section: Design Strength of Beams, Flexural design
Strength for Cb=1.0 States:

The flexural design strength of compact (flange and web local
buckling <p) I-shaped and C-shaped rolled beams (as defined in
Section B5 of the LRFD Specification) bent about the major or
minor axis is:

(3)

In minor axis flexure this is true for all unbraced lengths, but for
bending about the major axis the distance
b
between points braced
against lateral movement of the compression flange or between
points braced to prevent twist of the cross section shall not exceed
p

Given the approximate nature of determining actual unbraced lengths in soil,
it was decided that by orientating the piles for weak axis bending, flange local
buckling in the pile could be avoided. However, as the skew of the bridge
increases, both strong and weak axis bending occurs in IAB piles.

The MHD Bridge Manual [6] limits integral bridge skews to 30. For single
span bridges with skews up to 20 it is permitted to analyze the piles in the weak
axis as if they had 0 skew. This is a result of research that showed adequate
results of analyzing moderately skewed bridges as if they were square [9].
y p b n b
ZF M M
= =

276

For multiple span bridges and for single span bridges with skews above 20,
MHD policy requires that bridges be analyzed in 3-dimensions with the actual
skew of the piles. This is done to capture the lateral movement of the bridge as it
expands and contracts due to the resistance of backfill on the abutments and the
skew of the piles, see Figure 4. The resulting bending moments in both pile axes,
axial loads and P moments are all combined in the IAB pile interaction equation.
Further, MHD policy limits the use of non-compact pile sections for bridge skews
up to 20 only; compact pile sections are required when bridge skew exceeds 20.
SQUARE ABUTMENT
PILES ORIENTED FOR
WEAK AXIS ALONG
ABUTMENT CENTERLINE
SUPERSTRUCTURE
EXPANSION
SOIL PRESSURE NORMAL TO
ABUTMENT NO LATERAL COMPONENT
SUPERSTRUCTURE
EXPANSION
SKEWED ABUTMENT
PILES ORIENTED FOR
WEAK AXIS ALONG
ABUTMENT CENTERLINE
SOIL PRESSURE NORMAL TO
ABUTMENT CAUSES LATERAL
COMPONENT OF MOVEMENT

Figure 4. Square and Skewed Integral Abutments

BASIS FOR MHD INTEGRAL PILE DESIGN POLICY

Briefly, the design methodology outlined in the Greimann, et al/Iowa DOT
report Pile Design and Tests for Integral Abutment Bridges [1] is to permit the
pile to deflect under thermal movements. An analysis is then made to determine
the equivalent length of column and if the section is compact or non-compact,
then 1 of 2 approaches is applied to determine suitability.

If the pile section is non-compact, insufficient ductility for a plastic hinge to
form, as defined in the AISC ASD Manual Section B5 [10], the combined
bending moment due to thermal movements plus superstructure rotation and the
applied axial load is compared to the allowable stresses within the pile section
found using AASHTO Table 10.32.1A. The applied and allowable stresses are
compared using the AASHTO ASD interaction formulas for beam columns for
stability and for strength, AASHTO Section 10.36 Combined Stresses [7]. If the
pile section meets the criteria the design is adequate.

(4)

(5)
0 . 1
'
1
'
1
: 42) - (10 Eqn AASHTO Stability
|
|
.
|
\
|
+
|
|
.
|
\
|
+
by
ey
a
by mx
bx
ex
a
bx mx
a
a
F
F
f
f C
F
F
f
f C
F
f
0 . 1
472 . 0
: 43) - (10 Eqn AASHTO Strength + +
by
by
bx
bx
y
a
F
f
F
f
F
f

277

If the pile section is compact, sufficient ductility for plastic hinging, as
defined in the AISC ASD Manual Section B5 [10], the bending moment is
calculated only by the P moment, equal to the axial load multiplied by the
largest thermal movement. Once again, the applied axial and bending stresses are
compared to the allowable stresses using the AASHTO ASD interaction formulas
for beam columns for stability and for strength, equations (4) and (5). If the pile
section meets the criteria the design is adequate.

MHD INTEGRAL PILE DESIGN POLICY

The MHD Bridge Manual [6] requires bridge foundations to be designed by
the AASHTO Load Factor Design (LFD) requirements [7]. It was felt that a
ductility based approach to the design of IAB piles similar in nature to the
Response Modification Methodology employed in AASHTO Division IA for
Seismic Design [7] would provide satisfactory results for IAB pile design.
Therefore, because the pile was intended to be permitted to achieve yield, and in
keeping with the MHD design standards for other foundation types, the Load
Factor method was chosen as the appropriate design approach. Also, it was
decided that this method would be easily incorporated when MHD moves over to
the AASHTO LRFD Specifications.

Using Greimann et al [1] as a basis, a ductility approach for the design of
Integral Abutment Piles was developed for the MHD Bridge Manual by
modifying the AASHTO LFD [7] requirements for elements subjected to
combined axial load and bending Section 10.54.2.1:

(6)

(7)

Because of the support for the pile provided by the soil to resist Euler
buckling, the MHD design policy requires only a strength check for the pile
adequacy. Therefore, the AASHTO strength equation (7) for combined axial load
and bending was chosen for this policy. The MHD Bridge Manual [6] Integral
Abutment Design Section 3.9.7.2 contains the modified AASHTO design
equation for IAB piles and is as follows:

3.9.7.2 Determine the structural adequacy of the preliminary pile section
selected from the Geotechnical Report using the following strength criteria:
0 . 1
1
85 . 0
: 155) - (10 Eqn AASHTO Stability
|
|
.
|
\
|
+
e s
u
cr s
F A
P
M
MC
F A
P
0 . 1
85 . 0
: 156) - (10 Eqn AASHTO Strength +
p y s
M
M
F A
P

278

(8)

Where:
Pu = Applied axial load determined from analysis;
As = Cross sectional area;
My; Mx = Applied moment determined from analysis & P-D moment
Muy ; Mux = Maximum moment strength based on the slenderness
criteria;

i
= Coefficient of inelastic rotational capacity defined herein;
Fy = Yield Stress, preferably 250 MPa (36 ksi), but not to exceed 450
MPa (65 ksi) (AISC [3], page 4-10);

For compact sections the maximum moment shall be:
Muy = ZyFy and Mux = ZxFy
i
= 1.75 (to account for inelastic rotational capacity for weak axis
bending only)

Where:
Zy;Zx Plastic section modulus for respective axis;
Z< 1.5 S (AISC [3], Section F1.1)
S= Section modulus for respective axis;

For non-compact sections the maximum capacity shall be (AASHTO
10.48.3)[7]:

(9)

i = 1 (to account for lack of inelastic rotational capacity)

Where:
Mp= Plastic moment capacity for respective axis;
Mp=ZyFy of ZxFy;
Zy;Zx Plastic section modulus for respective axis;
Z< 1.5 S (AISC [3], Section F1.1)
S= Section modulus for respective axis;
My; Mx Yield moment capacity for respective axis;
My = SyFy and Mx = SxFy;
= Flange slenderness for Section
= b/tf
b = Flange projection =
tf = Flange thickness
p = Flange slenderness for compact section;
p = 10.8 (AASHTO Table 10.48.1.2A for Grade 250 (36) steel)
r = Flange slenderness for non-compact section;
0 . 1
85 . 0
+
+
ux
x
uy
y
y s
u
M
M
M
M
F A
P
|
|
.
|
\
|

=
|
|
.
|
\
|

=
p - r
p -
Mx) - (Mp - Mp Mux and
p - r
p -
My) - (Mp - Mp Muy
2
t - b
w f

279
r = 11.6 (AASHTO Table 10.48.2.1A for Grade 250 (36) steel)

If the analysis results indicate that the piles are inadequate, the Design
Engineer shall increase the pile size and/or add additional piles and reanalyze
until and adequate pile size and spacing is determined.

The ratio of bending about the weak axis to the weak axis capacity, My/Muy,
was modified by adding in the term
i
in the denominator to account for the
ductility of the pile section. To be somewhat conservative in this decision, the
full minimum ductility capacity of compact sections equal to 3 for plastic hinging
was not chosen for this factor, but rather a reduced value of 1.75. Also, this factor
is applied only to the weak axis component of pile bending. If both axes of the
pile are in bending, as in a skewed bridge, the strong axis term is not modified by
a ductility factor owing to the difficulties associated with determining unbraced
length in soil.

Equation (9) is recognition that compact sections may exceed yield but not
reach full plastic hinging in bending. The equation is a linear interpolation
between bending moment at yield to bending moment at full plastic, by the ratio
of flange slenderness of the section under consideration to the slenderness for full
plastic hinging and the slenderness required for yield. The equation is an
adaptation of one contained in AISC [3] on page 4-11 but is based on the cited
AASHTO [7] criteria.

CONCLUSION

The methodology developed by the Massachusetts Highway Department for
the design of integral abutment bridge foundation piles utilizes a ductility based
approach to permit and incorporate yielding in the piles as the bridge cycles
through expansion and contraction. By limiting the selection of piles to compact
or non-compact sections, and employing the modified LFD column strength
interaction equation, robust pile foundations may be designed using plastic
design, thus ensuring an economical use of the material in a structure type that
will provide a long service life.

REFERENCES
1. Greimann, L. F., R. E. Abendroth, D. E. Johnson, and P.B. Ebner. December 1987. Pile Design
and Tests for Integral Abutment Bridges, Final Report, College of Engineering, Iowa
State University, and Iowa Department of Transportation, pp 85-108.
2. Ernst, H., P. Connors and V. McGrath. April 1998. A Preliminary Report on the Geotechnical
Aspects of Integral Abutment Bridges, Geotechnical Section Internal Report,
Commonwealth of Massachusetts, Massachusetts Highway Department.
3. American Institute of Steel Construction, Manual of Steel Construction Load and Resistance
Factor Design, Volume I, 2nd ed., 1994.
4. Vesi, A. S. 1977. Design of Pile Foundations, NCHRP Synthesis of Highway Practice.,
42,Transporation Research Board, pp40.

280
5. Cummings, A. E.. 1938. The Stability of Foundation Piles Against Buckling Under Axial
Load, Division of Engineering and Industrial Research, National Research Council,
Highway Research Board, Proceedings of the 18th Annual Meeting, November 28-
December 2, 1938.
6. Commonwealth of Massachusetts, Massachusetts Highway Department Bridge Manual, Part I,
1995 with December 1999 Revisions.
7. AASHTO Standard Specifications for Highway Bridges, 17th ed., 2002
8. Lutenegger, A. J., T. A. Thomson, 1998. Passive Earth Pressures in Integral Bridge
Abutments, Report No UMTC-97-16, University of Massachusetts Transportation
Center, University of Massachusetts, Amherst.
9. Ting, J. M., S. Faraji, 1998. Streamlined Analysis and Design of Integral Abutment Bridges,
Report No UMTC-97-13, University of Massachusetts Transportation Center, University
of Massachusetts, Amherst.
10. American Institute of Steel Construction, Manual of Steel Construction, Allowable Stress
Design, 9th ed., 1989.

281
Integral-Abutment Bridges: Geotechnical Problems and
Solutions Using Geosynthetics and Ground Improvement

John S. Horvath, Ph.D., P.E.
Professor; Manhattan College; Civil and Environmental Engineering Department;
Bronx, NY 10471-4098; john.horvath@manhattan.edu

ABSTRACT
The integral-abutment bridge (IAB) concept was developed at least as far
back as the 1930s to solve long-term structural problems that can occur with
conventional bridge designs. Unfortunately, the IAB concept as executed
historically turns out to have its own inherent post-construction flaws. However
they are fundamentally of a geotechnical, not structural, nature. As a result, bridge
engineers, who are more familiar with dealing with structural issues, have been
slow to recognize the true source of IAB problems and develop appropriate
permanent solutions for them. Thus IABs represent an interesting case study in
soil-structure interaction that requires the coordinated attention of both
structural/bridge and geotechnical engineers working as a multidisciplinary team
if the concept is to be improved for better long-term performance. This paper is
intended to be an contribution toward that goal and illustrates the potential use of
modern geotechnolgies for IAB problem solving.

THE EVOLUTION OF INTEGRAL-ABUTMENT BRIDGES

The conventional design concept used for most bridges with a short to
medium span length consists of a superstructure resting on abutments as shown in
Figure 1. There may be one or more intermediate piers but their absence or
presence is not relevant to the present discussion and does not affect any of the
conclusions and recommendations made in this paper.
Because of natural, seasonal variations in atmospheric air temperature, the
bridge superstructure will change in temperature and concomitantly change
dimensions, primarily in the longitudinal direction as also shown in Figure 1.
Typical ranges of longitudinal displacement for relatively modest span lengths are
of the order of several tens of millimetres (one inch). However, the abutments
supporting the superstructure are for all practical purposes insensitive to air
temperature so remain spatially fixed year 'round. The relative displacement
between the moving superstructure and fixed abutments is accommodated by a
synergistic combination of expansion joints and bearings as shown in Figure 1.
Thus the key elements of conventional bridge design can be summarized as
follows:

282
Seasonal thermal displacements of the superstructure are natural and
unavoidable. The magnitude of these displacements turns out to be relatively
insensitive to the specific materials (steel versus portland-cement concrete
(PCC)) used for the bridge components.

The abutments are for all intents and purposes rigid structural elements that
are fixed in space and time (foundation settlements are not considered here as
they do not materially enter into the present discussion). This makes the
ground retained by the abutments also fixed in space and time.

There are explicit measures taken to isolate the moving superstructure from
the fixed abutments + soil and vice versa, at least in terms of longitudinal
displacements.

Although the design concept shown in Figure 1 has been used for a long time
and works well enough in practice, the expansion joint/bearing detail is often a
source of significant post-construction structural maintenance and expense during
the life of a bridge. Therefore, the IAB concept was developed to eliminate the
troublesome and costly expansion joint/bearing detail. This is accomplished by
physically and structurally connecting the superstructure and abutments as shown
conceptually in Figure 2.
Figure 1. Basic Elements of a Conventional Bridge
abutments
superstructure
expansion joints
primary direction of
thermally induced displacement
bearings

283
IABs have been used for road bridges since at least the early 1930s in the
U.S.A. [1]. Over the years and in different countries they have variously and
synonymously been called frame bridges, integral bridges, integral bridge
abutments, jointless bridges, rigid-frame bridges and U-frame bridges. There is
also a design variant called the semi-integral-abutment bridge. Relevant to this
paper is the fact that collectively such bridges have seen extensive and growing
use worldwide in recent years because of their economy of construction in a wide
range of conditions and using a variety of structural materials. Thus they comprise
an important aspect of modern transportation-engineering practice as evidenced
by the specialty conference for which this paper was prepared.

PROBLEMS WITH THE IAB CONCEPT

Overview

Although the IAB concept has proven to be conceptually successful in eliminating
expansion joint/bearing problems as well as economical in initial construction for
a wide range of span lengths, it has not turned out to be problem- and
maintenance-free in actual service. This is because the IAB concept suffers from
an inherent, fundamental flaw. Specifically, the IAB concept fails to explicitly
and proactively address how the relative displacement between the moving
superstructure and fixed ground is to be accommodated. This derives from the
fact that the IAB concept fails to recognize that it does not, and cannot,
fundamentally alter nature and the laws of physics and the resulting tendency of a
bridge superstructure to undergo seasonal temperature and length changes in its
longitudinal direction. All that has changed between conventional versus IAB
bridge designs (i.e. Figure 1 versus Figure 2) are the details of how this thermally
induced displacement occurs, and the nature of the resulting problems and
maintenance issues it generates. Thus IABs as currently designed still have
maintenance costs as did their jointed predecessors, which inflates the true life-
cycle cost of an IAB.
Figure 2. Basic Elements of an Integral-Abutment Bridge
approach slabs
(optional)

284

Causes

The fundamental cause of in-service problems for IABs as they are currently
designed is illustrated in Figure 3. As the bridge superstructure goes through its
seasonal length changes, it causes the structurally connected abutments to move
inward and away from the soil they retain in the winter, and outward and into the
retained soil during the summer. The specific mode of abutment movement is
primarily rigid-body rotation about the bottoms of the abutments although there is
a component of rigid-body translation (pure horizontal displacement) of the
abutments as well. Because rotation is dominant, the magnitude of the range of
horizontal displacements is thus greatest at the top of each abutment.

At the end of each annual thermal cycle, there is often a net displacement of
each abutment inward toward each other and thus away from the retained soil as
shown in Figure 3. The primary reason for this is that the inward winter
displacement is typically of sufficient magnitude to cause an active earth pressure
'soil wedge' to develop adjacent to each abutment and follow the abutment
inward, with the soil slumping downward somewhat in the process. Because of
the fundamentally inelastic nature of soil behavior, this inward/downward soil
displacement is not fully recovered during the outward summer cycle. It is
relevant to note that this net inward/downward soil displacement will occur no
matter what type of soil is used and how well it was compacted during original
construction. It is also of interest to note that this tendency to develop a net
inward displacement of the abutments is exacerbated when the bridge
superstructure is composed primarily of PCC due to the inherent post-construction
shrinkage of PCC.

Consequences

There are two significant consequences of the annual thermal cycle of IABs.
The first was recognized at least as far back as the 1960s [2,3,4] and is the
relatively large lateral earth pressures that develop on the abutments during the
annual summer expansion of the superstructure. These pressures can approach the
theoretical passive state, especially along the upper portion of the abutments
where horizontal displacements are largest. Passive earth pressures are typically
an order of magnitude greater than the at-rest pressures for which a bridge
abutment should typically be designed. This tenfold increase in lateral earth
pressures far exceeds any normal margin of structural safety built into the design
and thus can result in structural distress and even failure of an abutment.

285
Recent research indicates that this long recognized seasonal increase in lateral
earth pressures may be a more significant and potentially problematic issue than
initially thought. This is because the summer-seasonal increase in pressures is not
necessarily constant over time but can increase over time. The reason is that not
only is one seasonal cycle of inward-outward-inward displacement nonlinear, but
each succeeding season is nonlinear with respect to the preceding one. This means
each winter the abutment moves inward slightly more than it did the preceding
winter and each summer it moves outward slightly less than it did the preceding
summer. As a result of this net soil displacement inward toward the abutments
and the fact that the bridge superstructure still expands each summer the same
amount as the preceding year, the summer lateral earth pressures increase over
time as the soil immediately adjacent to each abutment becomes increasingly
wedged in. This overall behavior is a geo-phenomenon called ratcheting. The soil
mechanics behavior causing ratcheting is quite complex but is well and
thoroughly described in the literature [5].

Because ratcheting causes each summer's lateral earth pressures to be
somewhat greater in magnitude than those from the preceding year, it means
structural failure of the abutments may take years, even decades, to develop, a
happenstance observed in practice for other types of earth-retaining structures
where thermally induced ratcheting occurs [5,6,7]. Given the relatively long
design life of most IABs (typically 100 years or more), ratcheting represents a
potentially serious long-term source of problems, primarily structural distress and
failure of the abutments.

The second significant consequence of the annual thermal cycle of IABs is
also related to the net inward displacement of the abutments and has become fully
appreciated only in recent years. This is the subsidence pattern that develops
adjacent to each abutment as shown in Figure 4. This is the result of the above-
Figure 3. Thermally Induced IAB Abutment Displacement
summer
position
winter
position
final position at end of
annual temperature cycle
initial position at start of
annual temperature cycle (shaded area)
superstructure

286
described phenomenon of accumulated, irreversible soil-wedge slumping behind
each abutment. The consequences of this subsidence depend on whether or not an
approach slab was constructed as part of the bridge. If there is no slab, there will
be a difference in road-surface elevation occurring over a short distance creating
the classical 'bump-at-the-end-of-the-bridge' condition. If there is a slab, initially
it will span over the void created underneath it by the subsided soil. However,
with time and traffic the slab can fail structurally in flexure.
Subsidence behind IAB abutments has received much more interest in recent
years compared to the traditional concern over increased lateral earth pressures.
This is because experience indicates subsidence develops and becomes
problematic relatively soon (a few years at most) after an IAB is placed in service
[8,9,10,11,12] whereas the ratcheting buildup of lateral earth pressures might not
create problems for decades as noted above. For example, [11] noted that a survey
of 140 IABs with approach slabs in the State of South Dakota, U.S.A. found a
void under virtually every slab. The void depths ranged from 13 to 360 mm (0.5
to 14 in), and the voids extended as much as 3 m (10 ft) behind the abutment.

PROBLEM SOLUTIONS

Overview of Past Efforts

Although most recent research into defining IAB problems has focused on the
newly identified issue of subsidence behind abutments, some recent work related
to developing actual solutions for IAB problems has focused on the traditional
issue of lateral earth pressures alone [1,11,13]. Specifically, various types of
relatively compressible materials (generically referred to herein as compressible
inclusions) such as either resilient or normal expanded polystyrene (EPS)
geofoam and tire shreds have been placed behind IAB abutments. Conceptually, a
compressible inclusion is intended to serve as a sacrificial cushion between a
relatively rigid earth retaining structure and the adjacent ground with the overall
goal of reducing lateral earth pressures. A recent overview of the compressible-
Figure. 4. Ground-Surface Subsidence behind IAB Abutments
subsidence zone (and void development if approach slab used)
new ground surface
due to long-term
abutment displacement
long-term position
of abutment

287
inclusion concept with an emphasis on earth retaining structures used for
transportation facilities can be found in [14].

While these research efforts are a step in the right direction, available
information suggests they are significantly incomplete. Research to date indicates
that although the use of a compressible inclusion can be highly effective in
reducing the summer increase in lateral earth pressures it is totally ineffective for
controlling subsidence behind the abutments. In fact, experience indicates the
presence of a compressible inclusion may even exacerbate the subsidence
problem even as it addresses the summer lateral earth pressure problem. This is
because the highly compressible nature of a compressible inclusion that is so
desirable under summer expansion of an IAB becomes a detriment when winter
contraction occurs. As the superstructure contracts and pulls each abutment away
from the retained soil, the relatively weak compressible inclusion between
abutment and soil is unable to restrain the soil from slumping and displacing
inward toward the abutment. This actually results in subsidence behind the
abutment that is larger in magnitude than if no compressible inclusion were
present. This has been observed for at least one IAB that was studied thoroughly
where a compressible inclusion consisting of recycled tire fragments was used
with an approach slab [11]. It has also been observed in large-scale, 1-g physical-
model testing where a compressible inclusion composed of resilient-EPS geofoam
was used with a coarse-grain-soil backfill [12].

Proposed Improved Solutions: Basic Concepts

Because of the current extensive use of IABs, there is a critical need to
develop solutions to correct the behavioral deficiencies inherent in all IABs as
they are typically designed and constructed at the present time. As noted
previously, past efforts to develop improved designs based on the use of a
compressible inclusion alone have not been a total success because they did not
address both problems, i.e. the seasonal buildup of lateral earth pressures on the
abutments and ground-surface subsidence adjacent to the abutments.

The key to developing improved solutions that address both problems is a
thorough understanding and appreciation of the fundamental physical processes
that affect both conventional bridges as well as IABs. The following key concepts
and considerations were identified by the author and used to develop the
improved solutions presented subsequently:

Expansion and contraction of a bridge superstructure due to seasonal
temperature changes is inevitable and unavoidable as it is a natural
phenomenon that simply cannot be changed in any significant way. This
displacement will occur regardless of the specific structural concept used for
the bridge design. This means the tendency for differential horizontal
displacement between an IAB and the ground surface adjacent to its
abutments is unavoidable and must be addressed explicitly.

288
The ground adjacent to IAB abutments must be made inherently self-stable on
a permanent, year-round basis to prevent development of subsidence during
the seasonal winter contraction of the IAB. In essence, the ground itself must
provide the non-yielding, seasonally constant retention function formerly
provided by the abutments of conventional bridge design (Figure 1).

There must be a design detail involving a structural element or material
between the self-stable ground and moving IAB abutments to reliably and
predictably accommodate the relative movement between them. This detail
conceptually replaces the expansion joint/bearing detail of conventional
bridge design (Figure 1). Simply leaving a void between ground and abutment
as has been done occasionally in the past is not considered an acceptable
design detail. Experience indicates that a void is difficult to construct
routinely and reliably in practice [15], and it cannot be depended on to remain
for the long service life of a bridge.

Proposed Improved Solutions: Concept Details

A detailed numerical study was conducted by the author to both define the key
behavioral aspects of IABs as well as investigate potential solutions using
geosynthetics [16]. That study drew heavily on the knowledge gained during the
1990s about geofoams in general and EPS geofoam in particular [17]. Although
the revised designs developed and presented in [16] will increase the construction
cost of IABs, the anticipated superior post-construction, in-service performance of
such IABs should more than make up for the increase by reducing future
maintenance and repair costs. Similarly, implementing these revised designs
retroactively on existing IABs should be cost effective by reducing their future
maintenance and repair costs.

Two different design concepts were developed to accommodate different site
conditions. Both are shown schematically in Figure 5. The one likely to be more
cost effective in most applications is shown in Figure 5(a) and is appropriate for
sites where compression and/or stability of the native soils underlying the
approach embankment to the bridge is not an issue. The concept utilizes
geosynthetic tensile reinforcement (likely geogrids or geotextiles) to create a
mechanically stabilized earth (MSE) mass within the retained soil adjacent to
each abutment. This reinforced soil mass would be inherently self-stable for the
design life of the bridge. In addition, a relatively thin (typically of the order of
150 mm (6 in thick) layer of resilient-EPS-geofoam would be used as a
compressible inclusion in a 'chimney' orientation between the abutment and MSE
mass. This durable inclusion is highly compressible and thus functions as the
desired expansion joint between the abutment and MSE mass. Note that the
compressible inclusion also thermally insulates the retained soil (against winter
freezing) and the geosynthetic tensile reinforcement (from summer heat which
can increase geosynthetic creep), and can be designed to also serve as a drain for
ground water. Functionally, this compressible inclusion allows the reinforcement

289
within the soil to strain in tension (which prevents the soil from displacing inward
and downward toward the abutment each winter) as well as allows the abutments
to move seasonally in either direction with minimal restraint. Thus summer
increases in lateral earth pressures are reduced to relatively small magnitudes.
Overall, lateral earth pressures acting on the abutments are significantly reduced
from current design levels, which would achieve a cost savings in the structural
design of the abutment.

The other design alternative is shown in Figure 5(b). A self-stable wedge of
some kind of geofoam (most likely EPS blocks [18] but alternatively foamed
PCC) or geocomb blocks would be used as a solid lightweight-fill material in lieu
of the MSE mass. A relatively thin layer of highly compressible resilient-EPS
geofoam is again used multifunctionally as a compressible inclusion/thermal
insulation/chimney drain. This alternative is expected to be the one of choice for
sites where the soils underlying the approach embankment are soft and
compressible. Use of a lightweight fill material would minimize settlements and
enhance stability of the ground adjacent to the bridge as well as greatly reduce the
loads acting on the abutment and the deep foundations that would likely be
supporting it in such soil conditions. Solid lightweight-fill materials such as
various types of geofoam are particularly attractive here as they are inherently self
stable even when constructed with vertical side slopes. Although geofoam
materials are inherently more expensive than soil on a strictly volumetric
comparison the resulting overall savings would likely more than compensate for
the use of a geofoam material in lieu of soil. The benefits of accelerated
construction by using geofoam materials should also be considered.
Figure 5. Proposed New IAB Design Alternatives
resilient-EPS geofoam
compressible inclusion
(a)
(b)
geosynthetic
tensile
reinforcement
EPS-block
geofoam
lightweight fill

290

Proposed Improved Solutions: Additional Comments

For the sake of completeness, it should be noted that in cases where the
ground underlying the approach embankment is weak and compressible there are
other potential alternatives to using a solid lightweight-fill material as shown in
Figure 5(b) that might be cost-effective on a project-specific basis [19]. This
includes using a granular type of lightweight fill material (expanded-shale
aggregate, tire shreds, etc.) in combination with geosynthetic tensile
reinforcement to stiffen and retain the material as an equivalent MSE mass as
shown in Figure 5(a). Alternatively, some type of ground improvement might be
performed to strengthen and stiffen the native soils in situ prior to constructing the
approach embankment. After performing the ground improvement, the approach
embankment could be constructed using normal soil and an MSE mass adjacent to
the abutments as shown in Figure 5(a).

IABs are an interesting example of how new problems were inadvertently
created in the process of solving old problems. IAB problems are fundamentally
geotechnical in nature and can manifest themselves both structurally and
geotechnically any time in the life of an IAB. The primary cause of both short and
long-term IAB problems is the irreversible net inward and downward
displacement of the soil retained by IAB abutments. This will occur regardless of
the type of soil used and how well it was compacted during construction. The
resulting problems consist of irreversible subsidence behind the abutments and the
ratcheting buildup of 'summer' lateral earth pressures on abutments. Either or both
of these outcomes can result in serviceability or collapse failures of the bridge
components and thus are serious.
Research also indicates that relatively simple and cost-effective design
solutions to eliminate these problems can be achieved using a variety of modern
geosynthetics and/or ground-improvement technologies in an innovative,
synergistic fashion. Because the problems encountered with IABs as currently
designed turn out to be a complex problem in soil-structure interaction, any
successful solution must address the need to both:

Support the ground adjacent to the abutments on a permanent, year-round
basis and

Provide for a compressible inclusion (essentially an expansion joint) between
abutment and adjacent ground to serve as an engineered, in-ground
replacement to the expansion joint/bearing detail of conventional bridges.
Several specific suggestions for improved IAB designs were presented in this
paper. An important benefit is that these solutions can be implemented as part of

291
the rehabilitation of existing IABs in addition to being applicable for new
construction.

REFERENCES

1. Carder, D. R. and G. B. Card 1997. "Innovative Structural Backfills to Integral Bridge
Abutments", Proj. Rpt. 290, Trans. Res. Lab., U.K.
2. Broms, B. B. and I. Ingleson 1971. "Earth Pressure Against the Abutments of a Rigid Frame
Bridge", Gotechnique, Inst. of Civil Engr., U.K., 21 (1), 15-28.
3. Card, G. B. and D. R. Carder 1993. "A Literature Review of the Geotechnical Aspects of the
Design of Integral Bridge Abutments", Proj. Rpt. 52, Trans. Res. Lab., U.K.
4. Sandford, T. C. and M. Elgaaly 1993. "Skew Effects on Backfill Pressures at Frame Bridge
Abutments", Trans. Res. Rec. 1415, Natl. Acad. Press, 1-11.
5. England, G. L. 1994. "The Performance and Behaviour of Biological Filter Walls as Affected
by Cyclic Temperature Changes", Serviceability of Earth Retaining Structures, R. J. Finno
(ed.), ASCE, 57-76.
6. England, G. L. and T. Dunstan 1994. "Shakedown Solutions for Soil Containing Structures as
Influenced by Cyclic Temperatures - Integral Bridge and Biological Filter", Proc. of the Third
Intl. Conf. on Struc. Engr, Singapore.
7. England, G. L., T. Dunstan, C. M. Tsang, N. Mihajlovic and J. B. Bazaz 1995. "Ratcheting
Flow of Granular Materials", Static and Dynamic Prop. of Gravelly Soils, M. D. Evans and R.
J. Fragaszy (eds.), ASCE, 64-76.
8. Briaud, J.-L., R. W. James and S. B. Hoffman 1997. "Settlement of Bridge Approaches (the
Bump at the End of the Bridge)", Syn. of Hwy. Prac. 234, Natl. Acad. Press.
9. Ng, C. W. W., S. Springman and A. R. M. Norrish 1998. "Centrifuge Modeling of Spread-
Base Integral Bridge Abutments", Jour. of Geotech. and Geoenv. Engr., ASCE, 124 (5), 376-
388.
10. Reid, R. A., S. P. Soupir and V. R. Schaefer 1998. "Use of Fabric Reinforced Soil Walls for
Integral Abutment Bridge End Treatment", Proc. Sixth Intl. Conf. on Geosyn., R. K. Rowe
(ed.), Ind. Fabrics Assoc. Intl., 573-576.
11. Reid, R. A., S. P. Soupir and V. R. Schaefer 1998. "Mitigation of Void Development under
Bridge Approach Slabs Using Rubber Tire Chips", Recycled Matls. in Geotech. App., C.
Viupulanandan and D. J. Elton (eds.), ASCE, 37-50.
12. Reeves, J. N. and G. M. Filz 2000. "Earth Force Reduction by a Synthetic Compressible
Inclusion", report submitted to GeoTech Systems Corp. and Virginia's Ctr. for Innovative
Tech. by Virginia Tech, Dept. of Civil Engr.
13. Humphrey, D. N., N. Whetten, J. Weaver, K. Recker and T. A. Cosgrove 1998. "Tire Shreds
as Lightweight Fill for Embankments and Retaining Walls", Recycled Matls. in Geotech.
App., C. Viupulanandan and D. J. Elton (eds.), ASCE, 51-65.
14. Horvath, J. S. 2004. "Controlled Yielding Using Geofoam Compressible Inclusions: New
Frontier in Earth Retaining Structures", Geotech. Engr. for Trans. Proj., ASCE, 1925-1934.
15. Edgar, T. V., J. A. Puckett and R. B. D'Spain 1989. "Effects of Geotextiles on Lateral
Pressure and Deformation in Highway Embankments", Geotex. and Geomem., Elsevier, 8 (4),
275-292.
16. Horvath, J. S. 2000. "Integral-Abutment Bridges: Problems and Innovative Solutions Using
EPS Geofoam and Other Geosynthetics", Res. Rpt. CE/GE-00-2, Manhattan Coll., Civil Engr.
Dept.
17. Horvath, J. S. 1995. Geofoam Geosynthetic, Horvath Engr., P.C.
18. Stark, T. D., D. Arellano, J. S. Horvath and D. Leshchinsky 2004. "Geofoam Applications in
the Design and Construction of Highway Embankments", NCHRP Web Document 65.
19. PIARC Technical Committee C12 on Earthworks, Drainage, Subgrade 1997. "Matriaux
Lgers pour Remblai/Lightweight Filling Materials", Doc. 12.02.B, PIARC - World Road
Association, La Defense, France.

292
P-y Curves from Pressuremeter Testing at Kings Creek
Bridge, WV Route 2, Hancock County, West Virginia

Walter G. Kutschke, P.E., of URS Corporation, Pittsburgh, PA and
Braulio Grajales, P.E., of URS Corporation, Tampa, FL

ABSTRACT

This paper details the site characterization and lateral load analyses for the
replacement of West Virginia State Route 2 over Kings Creek, Hancock County,
West Virginia. Pressuremeter testing was used to determine the in-situ modulus of
elasticity of the bearing sandstone stratum for the development of site specific p-y
data for lateral load analyses of the plumb, deep foundation system. This paper
will review the available information regarding laterally loaded large diameter
bored piles socketed into weak rock and will focus on the development of p-y
curves from pressuremeter data. The paper also presents two statistical procedures
to analyze subsurface data that exhibits significant scatter.

INTRODUCTION

This $3.6 million bridge replacement project, located in the northern
panhandle of West Virginia, involves the replacement of West Virginia State
Route 2 over Kings Creek along a different alignment. The existing 309 ft, three
span structure provides grade separation over the Kings Creek valley floor,
located approximately 65 ft beneath the structure. The superstructure consists of
a non-redundant steel plate girder floorbeam system that is structurally deficient
and functionally obsolete.

The replacement structure consists of a 324-ft three span, prestressed concrete
I-beam bridge that utilizes integral abutment construction. Abutments are founded
on a single row of steel H-piles driven onto the underlying sandstone rock. Pier
construction consists of a hammerhead cap with a single 84-in diameter pier
column. The pier foundation design utilizes a single shaft that extends from a 78-
in diameter rock socket, to an 84-in diameter drilled shaft and into the pier
column eliminating the need for a pile cap. Pier 1 is the tallest pier with a height
differential between bearing pad to rock socket tip of 97 ft. Given the relatively
tall pier heights supported by a single rock socket, there is significant potential for
the development of large moment and shear force in the plumb foundation
system. As such, lateral load analyses for the pier foundation system became a
critical aspect of the foundation design. This paper will focus on the development
of site-specific p-y data for lateral load analyses of the plumb, drilled shaft
foundation system.

293
SITE CONDITIONS

The project site is situated in the Appalachian Plateaus Physiographic
Province, which is only structurally a plateau since stream action has eroded the
surface over millions of years resulting in a region of high relief. Small, narrow
valleys (or hollows) twist through the resulting mountains. The older surface is
evident in the pattern of hilltops all tending to reach the same elevation. Rocks
exposed along the project corridor are horizontally bedded, sedimentary rocks of
the Conemaugh and Allegheny Group, Pennsylvanian Period. Rock strata
generally consist of cyclic sequences of sandstone, siltstone, shale, limestone, and
coal.

Site characterization consisted of detailed line mapping, core borings, in-situ
pressuremeter testing and laboratory testing. Test borings advanced for the pier
foundations encountered gray, medium to coarse-grained sandstone to dark gray,
shaley carbonaceous sandstone. Further examination of the rock cores indicates
very intense to thickly bedded planes (Relative Dip (RD) = 0), average hardness,
moderate to slight weathering and moderate fracturing with RD = 0 to 70.
Percent rock core recoveries are generally 100 and Rock Quality Designations
(RQD) vary from 22 to 100 percent, but generally greater than 70 percent. The
Rock Mass Rating for this material ranges from 47 to 67, indicating a Class III or
Fair rating.

Laboratory tests on representative rock cores included unconfined
compressive strength (UCS) (ASTM D2938) and Youngs modulus (ASTM
D3148) testing. Table 1 provides a summary of the test data.

Table 1. Unconfined Compressive Strength Test Data
Unit Weight UCS Youngs
Modulus
Material
(pcf) (psi) (ksi)
Coarse to medium
SANDSTONE
145.7 155.2 4790 7240 3,800 4,200
Fine to shaley SANDSTONE 162.5 164.4 6230 8480 - - -

It is important to note that the Youngs modulus data obtained from the intact
rock cores does not represent the Youngs modulus of the rock mass; the
laboratory test does not account for jointing and fracturing of the rock mass [1, 2].

PRESSUREMETER TESTING

In addition to the core borings and laboratory-testing program, site
characterization for the proposed pier foundations consisted of in-situ
pressuremeter testing in the sandstone stratum. The pressuremeter, or rock

294
dilatometer, is a cylindrical radially expandable borehole probe that determines
deformations by measuring the total volume change of the probe. The dilatometer
is inserted in a N-sized borehole at the lowest test elevation. The membrane is
inflated to a predefined pressure increment and held at this pressure for one
minute; after which, the injected volume and pressure are recorded. The operation
is repeated until the maximum pressure or maximum injected volume is reached.
The volume change of the probe is measured by monitoring the displacement of a
piston within the probe to eliminate the parasitic expansion of the tubing and
pumping system. The probe is then deflated and raised to the next test elevation.
The use of pressuremeter testing in weak rocks is relatively well known [3, 4, 5,
6, 7].

Reduction of pressuremeter data involves plotting a pressure-volume curve
and selecting an elastic moduli value over which the pressure-volume curve is
linear. Figure 1 presents the elastic moduli profiles at pier 1 and 2. Depth to rock
beneath finished grade at Pier 1 is 48.5ft and at Pier 2 is 0.1ft. The elastic
moduli data exhibit significant scatter and no apparent trend with depth can be
discerned; the elastic moduli profiles identify a lower value of 136 ksi and an
upper value of 1,200 ksi.

Pier 1 Test Boring B-3
48.0
50.0
52.0
54.0
56.0
58.0
60.0
62.0
64.0
66.0
400 600 800 1000 1200 1400
E (ksi)
D
e
p
t
h

B
e
n
e
a
t
h

G
r
o
u
n
d

S
u
r
f
a
c
e

(
f
t
)
E
AVG
= 592 ksi
E
AVG
= 1,200 ksi

Pier 2 Test Boring B-7
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
20.0
100 200 300 400 500 600
E (ksi)
D
e
p
t
h

B
e
n
e
a
t
h

G
r
o
u
n
d

S
u
r
f
a
c
e

(
f
t
)
E
AVG
= 554 ksi
E
AVG
= 378 ksi
E
AVG
= 136 ksi
E
AVG
= 378 ksi

Figure 1. Elastic Moduli Profiles from Pressuremeter Testing at Pier 1 and Pier 2.

295

ANALYSIS OF SUBSURFACE DATA

An interim p-y curve methodology uses the ultimate resistance of the rock
(p
ur
) and the initial modulus of rock (E
ir
) to predict a p-y curve [3]. The ultimate
resistance of the rock can be estimated from the UCS of the intact rock. The initial
modulus of the rock can be taken from the initial slope of a pressuremeter curve
to provide a site specific, in-situ value. Alternatively, the initial modulus of the
rock can also be obtained from published correlations between UCS of the intact
rock specimens and elastic modulus of the intact rock (E
core
) [8, 9, 10], and
between the modulus reduction ratio (E
mass
/E
core
) and the average RQD value [1,
10].

Review of Figure 1 indicates that the pressuremeter data contains significant
scatter. As such, site characterization involved a statistical analysis of the UCS
and elastic modulus data. The evaluation consisted of applying two different
methods of analysis, namely the one-standard deviation method, and the 90th
percentile method [11]. These methods are described below:

One Standard-Deviation Method

The one-standard deviation method generates a range of values that are
delimited by a lower bound equal to the mean minus one standard deviation, and
an upper bound equal to the mean plus one standard deviation, and reduces the
initial data by discarding the values outside the generated range (i.e., smaller than
the mean minus one standard deviation, or greater than the mean plus one
standard deviation). The design value is the mean of the modified or reduced data.
Figures 2(a) and 2(b) present the results from the one-standard deviation method
for the UCS and pressuremeter test data, respectively.

The one-standard deviation method results in a mean UCS equal to 6,481 psi
with a standard deviation of 1,102 psi and a mean elastic modulus equal to
505,392 psi with a standard deviation of 386,482 psi. Three UCS values were
discarded since two of them fell below the lower boundary (i.e. 5,379 psi) and one
above the upper boundary (i.e. 7,583 psi). The modified mean UCS is equal to
6,647 psi, which is greater than the initial mean UCS value. Similarly, three
elastic modulus values are discarded since one of them fell below the lower
boundary (i.e. 118,909 psi) and two of them above the upper boundary (i.e.
891,874 psi). The modified mean elastic modulus value is equal to 381,076 psi,
which is less than the initial mean elastic modulus value.

296
90
th
Percentile Method

The 90th percentile method selects a value which 90 percent of the data are
greater than or equal to. This method arranges the data in ascending order and
then counts back 90 percent of the values to obtain the design value. An
approximate design value can also be obtained graphically.

Figures 2(c) and 2(d) present the results obtained from the 90th percentile
method. The core data and pressuremeter test results used in the analysis were the
same as those used in the one-standard deviation analysis. Figures 2(c) and 2(d)
indicate a UCS design value equal to 5,010 psi and an elastic modulus design
value equal to 130,029 psi which 90 percent of the UCS and elastic modulus data
meet or exceed these values, respectively.

0.00E+00
5.00E-05
1.00E-04
1.50E-04
2.00E-04
2.50E-04
3.00E-04
3.50E-04
4.00E-04
3000 4000 5000 6000 7000 8000 9000 10000
Unconfined Compressive Strength (psi)
P
r
o
b
a
b
i
l
i
t
y

D
i
s
t
r
i
b
u
t
i
o
n

f
(
x
)
Mean -/+ 1 Standard Deviation

(a)
0 %
10 %
20 %
30 %
40 %
50 %
60 %
70 %
80 %
90 %
100 %
3000 4000 5000 6000 7000 8000 9000 10000
Unconfined Compressive Strength (psi)
P
e
r
c
e
n
t

G
r
e
a
t
e
r

T
h
a
n

o
r

E
q
u
a
l

T
o

(b)
0.00E+00
2.00E-07
4.00E-07
6.00E-07
8.00E-07
1.00E-06
1.20E-06
0
2
0
0
0
0
0
4
0
0
0
0
0
6
0
0
0
0
0
8
0
0
0
0
0
1
0
0
0
0
0
0
1
2
0
0
0
0
0
1
4
0
0
0
0
0
1
6
0
0
0
0
0
Elastic Modulus (psi)
P
r
o
b
a
b
i
l
i
t
y

D
i
s
t
r
i
b
u
t
i
o
n

f
(
x
)
Mean -/+ 1 Standard Deviation

(c)
0 %
10 %
20 %
30 %
40 %
50 %
60 %
70 %
80 %
90 %
100 %
0
1
0
0
0
0
0
2
0
0
0
0
0
3
0
0
0
0
0
4
0
0
0
0
0
5
0
0
0
0
0
6
0
0
0
0
0
7
0
0
0
0
0
8
0
0
0
0
0
9
0
0
0
0
0
1
0
0
0
0
0
0
Elastic Moduli of Rock from Pressuremeter Test (psi)
P
e
r
c
e
n
t

G
r
e
a
t
e
r

T
h
a
n

o
r

E
q
u
a
l

T
o

(d)
Figure 2. Values Obtained from the One Standard Deviation Method (i.e., (a) and (b)) and
the 90
th
Percentile Method (i.e., (c) and (d)).

P-Y CURVE ANALYSIS

A first step in developing site-specific p-y curve data is to estimate the initial
elastic modulus of the rock. To represent a strip from a beam resting on an elastic,

297
homogeneous and isotropic solid, the relationship of the elastic modulus over the
initial part of the p-y curve is given by [3, 4]:

E
ir
= k
i
E
mi
(1)

where k
i
is a dimensionless constant derived from experiment and E
mi
is the initial
modulus of the rock. Three approaches for estimating the initial modulus of the
rock were considered. The first one is estimating the initial elastic modulus of the
rock mass by averaging measured elastic modulus values of intact specimens
from UCS tests and correcting them using a modulus reduction ratio as a function
of the average RQD [1, 2]. This methodology results in an estimated rock mass
modulus of 800,000 psi. However, this value should be used cautiously due to the
limited number of tests performed. The second approach is to use the modified
mean elastic modulus value (i.e. 381,076 psi) estimated from the one-standard
deviation analysis. The third approach is to use the elastic modulus value (i.e.
130,029 psi) estimated from the 90th percentile method. The initial elastic
modulus of the rock mass values obtained from the three different approaches is
also spread considerably. For this reason, a design value of 130,029 psi was
chosen as the recommended value for computing the initial portion of the p-y
curve in the sandstone material. The probability of the representative elastic
modulus of the rock mass being less than 130,029 psi is 16.6%.

The ultimate resistance is rarely, if ever, developed in practice, but prediction
of the ultimate resistance for p-y curves is necessary to reflect the non-linear
behavior. The ultimate resistance p
ur
is a function of the UCS of the intact rock,
the diameter of the shaft (B), the depth below the rock surface (x) and a strength
reduction factor (
r
) and is given by [3, 4]:

( )
|
.
|
\
|
+ =
B
x
B UCS p
r ur
4 . 1
1 for 0 x 3B
(2)

( )B UCS p
r ur
2 . 5 = for x > 3B (3)

As is readily evident, the UCS value is directly proportional to the ultimate
resistance; therefore careful selection of this value is critical for proper site
characterization. Two approaches for calculating a representative UCS value were
used. The first approach is to use the modified mean UCS value (i.e. 6,647 psi)
estimated from the one-standard deviation analysis and the second approach is to
use the UCS value (i.e. 5,010 psi) estimated from the 90th percentile method.
UCS data also exhibits important variation and no apparent trend with depth can
be discerned, as in the case for the elastic moduli. The UCS values obtained from
the two different approaches are separated by more than 1,600 psi. Consequently,
a UCS design value of 5,010 psi was chosen for the recommended value for
computing the final portion of the p-y curves in the sandstone material. The
probability of the representative UCS of the rock being less than 5,010 psi is
9.2%.

298

After computing the p
ur
and E
mi
values, a family of p-y curves can now be
developed. The straight-line portion, curved portion and horizontal portion are
respectively given by [3, 4]:

( ) y E p
mi
= for y <y
a

(4)
2
2
|
|
.
|
\
|
=
m
ult
y
y p
p

(5)
ult
p p = (6)

where y
a
defines the limit of the linear portion of the p-y curve and is found by
solving for the intersection of Equation (4) and Equation (5) and y
m
= k
m
B where
k
m
is a constant ranging from 5E-4 for vuggy limestone to 5E-5 for sandstone
containing very closely spaced discontinuities; these values are based on limited
test data [3,4].

Figure 3 presents the computed p-y curve data for Pier 1.

0
100
200
300
400
500
600
700
800
900
0.00 0.02 0.04 0.06 0.08 0.10
Y (in)
P

(
k
i
p
s
/
i
n
)
EL. +654.0 ft EL. +652.0 ft EL. +650.0 ft EL. +648.0 ft
EL. +646.0 ft EL. +644.0 ft EL. +642.0 ft
Defined by Eqn (4)
Defined by Eqn (5)
Defined by Eqn (6)

Figure 3. p-y Curve Data for Pier 1.

299
LATERAL LOAD ANALYSIS

There exist several analysis and design methods specifically for rock-socketed
shafts under lateral loading [3, 12, 13]. However, their application in practice
remains very limited since it has been customary practice to adopt the techniques
developed for laterally loaded, plumb foundation elements in soil to solve the
problem of rock-socketed drilled shafts under lateral loading [14, 15]. This
practice often results in erroneous design and excessive shaft length [3].

The structural design for the piers was performed in accordance with
AASHTO LRFD Design Methodology [2] and utilized the computer programs
COM624P [16] and GT-STRUDL [17]. COM624P analyses were preformed to
obtain the theoretical points of fixity in the transverse and longitudinal directions
for both piers. A two-dimensional GT-STRUDL model of the entire bridge was
used to determine the distribution of longitudinal loads from the superstructure to
the piers. Two cases were analyzed. Case 1 included spring constants at the
abutments to approximate the resistance of the integral abutment piles and the
passive soil force. Case 2 assumed the spring constants at the abutments to be
negligible. Case 2 results (which assume that the piers take all loads from the
superstructure) were used to design the piers. The temperature loading was equal
for both piers. The externally applied wind and braking forces were
approximately 20 percent greater for Pier 2 (compared to Pier 1), which reflect the
difference in the stiffness of the piers. Figure 4 presents the predicted lateral
deflection for each pier structure for SER-I Load combination [2].

0
10
20
30
40
50
60
-0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Deflection (in)
P
i
e
r

1

E
m
b
e
d
m
e
n
t

(
f
t
)
0
3
6
9
12
15
18
P
i
e
r

2

E
m
b
e
d
m
e
n
t

(
f
t
)
Pier 1 - Soil
Pier 1 - Rock
Pier 1
Pier 2
Pier 2 - Rock
Finished Grade -
Pier 1 and Pier 2
Bottom of Socket - Pier 2

Figure 4. Estimated Lateral Deflection for SER-I Load Combination.

Figure 5 presents the predicted moments for each pier structure for the
corresponding deflections noted in Figure 4.

300

0
10
20
30
40
50
60
-2000.0 0.0 2000.0 4000.0 6000.0 8000.0
Moment (kip-ft)
P
i
e
r

1

E
m
b
e
d
m
e
n
t

(
f
t
)
0
3
6
9
12
15
18
P
i
e
r

2

E
m
b
e
d
m
e
n
t

(
f
t
)
Pier 1 - Soil
Pier 1 - Rock
Pier 1
Pier 2
Pier 2 - Rock
Finished Grade -
Pier 1 and Pier 2

Figure 5. Predicted moments for deflections noted in Figure 4.


A prudent subsurface investigation for the characterization of a rock mass is
necessary because secondary structure (e.g., joint and/or fracture planes, etc.) can
dictate the behavior of a pile under lateral load. Site characterization involved
geological mapping, core borings, laboratory testing and in-situ pressuremeter
testing. Methods presented herein described two different statistical analyses for
the variable UCS and pressuremeter data.

The available subsurface data allowed for the development of site-specific p-y
curve data for lateral load analyses in weak rock. Lateral load analyses indicated
that for the drilled shafts to develop fixity (i.e., required embedment to preclude
the shafts from acting as fence post members), pier 1 required an embedment of
approximately 40-ft (developed prior to the rock socket) while pier 2 required an
embedment of approximately 6-ft into the sandstone stratum.

A field-load test with instrumented piles is desirable for improving the lateral
load design for this project; especially when only one drilled shaft is to be used
for each pier foundation, and to add to the experimental database of laterally
loaded drilled shafts embedded in soft rock.

301

REFERENCES

1. FHWA HI-88-042, Drilled Shafts, Federal Highway Administration, July 1988.
2. AASHTO LRFD Bridge Design Specifications, 1998 with 1999 to 2003 Interim
Revisions, Washington, DC.
3. Reese, L. C. 1997. Analysis of Laterally Loaded Piles in Weak Rock, Journal of
Geotechnical Engineering, ASCE, 123(11):1010-1017.
4. Wyllie, D. C. 1992. Foundations in Rock, E & FN Spon, London.
5. Robertson, P.K., Hughes, J.M.O., Campanella, R.G. and Sy, A. 1984. Design of
Laterally Loaded Displacement Piles Using a Driven Pressuremeter, Laterally Load
Deep Foundations: Analysis and Performance, ASTM STP835, J. A. Langer, E.T.
Mosley, and C.D. Thompson, Eds., American Society for Testing and Materials, pp.
229-238.
6. Anderson, J.B., Townsend, F.C., and Grajales, B., 2003. Case History Evaluation of
Laterally Loaded Piles, Journal of Geotechnical Engineering, ASCE 129(3):187-
196.
7. Dittes, M., and Labuz, J. F. 2002. Field and Laboratory Testing of St. Peter
Sandstone, Journal of Geotechnical Engineering, ASCE, 128(5):372-380.
8. Peck, R. B.1976. Rock Foundations for Structures, Vol. 2, ASCE Specialty
Conference on Rock Engineering for Foundations and Slopes. American Society of
Civil Engineers, Boulder, CO, pp. 1-21.
9. Deere, D. V. 1968. Geological Considerations, Chapter 1, Rock Mechanics in
Engineering Practice, K. G. Stagg and O. C. Zienkiewicz, eds. John Wiley and Sons,
Inc., NY, pp. 1-20.
10. Bieniawski, Z. T. 1984. Rock Mechanics Design in Mining and Tunneling. A. A.
Balkema, Rotterdam/Boston, pp. 272.
11. State of Florida, Department of Transportation, Soils and Foundation Handbook,
Ed. 2000. State Materials Office, Gainesville, Florida.
12. Carter, J. P., and Kulhawy, F. H. 1992. Analysis of Laterally Loaded Shafts in
Rock, Journal of Geotechnical Engineering, ASCE, 118(6):839-855.
13. Zhang, L., Ernst., and Einstein, H. H. 2000. Nonlinear Analysis of Laterally Loaded
Rock-Socketed Shafts, Journal of Geotechnical Engineering, ASCE, 126(11): 955-
968.
14. Amir, J. M. 1986. Piling in Rock, Balkema, Rotterdam. The Netherlands.
15. Gabr, M. A. 1993. Discussion on Analysis of Laterally Loaded Shafts in Rock.
Journal of Geotechnical Engineering, ASCE, 119(12):2015-2018.
16. FHWA SA-91-048, COM624P Laterally Loaded Pile Analysis Program for The
Microcomputer, Federal Highway Administration, June 1993.
17. GT-Strudl, Version 25.0, Computer Aided Structural Engineering Center, School of
Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta,
Georgia.

302
Effective Temperature and Longitudinal Movement in

Ralph G. Oesterle
1
, and Jeffery S. Volz
2

1
Construction Technology Laboratories, Inc., 5400 Old Orchard Road, Skokie, IL
60077-1030; PH 847-972-3216; ROesterle@CTLGroup.com
2
The Pennsylvania State University; Civil Eng. Dept.; jsv116@psu.edu
ABSTRACT
Jointless bridges promote reduced maintenance costs, improved riding quality,
lower impact loads, reduced snowplow damage to decks and approaches, and
improved seismic resistance. In spite of many of these recognized benefits, the
behavior of such structures is not yet fully understood, and nationally adopted
design criteria are still lacking. This paper presents results from an experimental
and analytical research program, funded by the Federal Highway Administration,
on the behavior of jointless and integral abutment bridges. The experimental work
included testing and monitoring of bridge models and a bridge structure in the
field, tests of bridge components, and a field survey of fifteen jointless bridges.
Experimental results have resolved many questions regarding environmental
effects and long-term and time dependent loading in combination with live and
dead load. The analytical work evaluated the response of jointless bridges with
respect to various design parameters. The research indicated that analysis
procedures can be used to adequately quantify the structural response if accurate
material and environmental parameters are known. Simplified design procedures
are recommended based on this research.

A study of the effects of longitudinal bridge movement on jointless integral
abutment bridges was a major focus of the research. A bridge will expand and
contract from seasonal and diurnal variations in temperature and will contract
with concrete creep and shrinkage strains. Piers and abutments must be designed
to accommodate this movement, and the superstructure must be capable of
carrying the forces induced by the stiffness of the piers and abutments. An
important first step to understand the effect of longitudinal movement is to
determine expected movement. Factors involved include an effective temperature
range with seasonal and diurnal components. Diurnal components include daily
shade temperature change and a solar effect. Other factors include the coefficient
of thermal expansion, creep, shrinkage and restraint from piers and abutments.
The overall variability of these factors causes uncertainty in the determination of
bridge movements. Therefore the research program included studies to define
appropriate temperature ranges for bridge design, dependent on location and type
of bridge and to determine expected abutment and pier support movements and
the potential variability of those movements.
EFFECTIVE TEMPERATURE RANGE
The effective temperature is the temperature that governs the overall
longitudinal movement of the bridge superstructure. Determination of the
effective temperature is a complex problem influenced by shade temperature,
solar radiation, wind speed, material properties, surface characteristics and section

303
geometry. Many of these factors are highly variable and not necessarily related. A
variety of researchers have attempted to relate one of these environmental factors,
namely shade temperature, to effective temperature with limited success [1].
Emerson developed a relationship between instantaneous shade temperature
and effective temperature by first developing a relationship between mean shade
temperature and effective bridge temperature [1]. She found that a 48-hour mean
shade temperature correlated well with concrete bridge effective temperatures,
and a 24-hour mean shade temperature correlated well with composite steel
bridge effective temperatures. The use of mean shade temperatures minimized the
random nature of most of the variables affecting bridge temperature. Emerson
then related the mean shade temperatures to instantaneous shade temperatures
based on meteorological data recorded throughout the United Kingdom.
Emerson's approach was subsequently incorporated into British Standard BS 5400
[2] to predict minimum and maximum effective bridge temperatures based on
geographical distribution of minimum and maximum shade temperatures.
Emerson's approach was applied to effective temperatures in NCHRP Report
276 [3] with extrapolation to cover the larger range of minimum and maximum
shade temperatures experienced within the United States. The design guidelines in
NCHRP Report 276 are the basis for the temperature data in AASHTO Guide
Specifications for Thermal Effects in Concrete Bridge Superstructures [4].
However, isotherms in NCHRP Report 276 represent normal daily minimum and
maximum values that do not appear to represent appropriate extreme shade
temperatures for calculation of maximum bridge movement. As an example,
Girton measured effective maximum temperature ranges of 115F and 117F over
a 2-year period for a concrete bridge and a composite steel bridge, respectively,
located in Iowa [5]. Using NCHRP Report 276, an effective design temperature of
approximately 70F was determined for the concrete bridge and 90F was
determined for the composite steel bridge.

Therefore, a study was completed as part of the FHWA Jointless Bridge
Project to determine more appropriate minimum and maximum shade
temperatures depending on the location within the United States. A variety of
climatological references were evaluated, including:
1. Climatic Atlas of the United States, U.S. Department of Commerce [6].
2. "Daily Normals of Temperature, Precipitation, and Heating and Cooling
Degree Days, 1961-1990," Climatography of the United States [7].
3. Climatic Atlas of the United States, Harvard University Press [8].
4. ASHRAE Fundamentals Handbook, "Chapter 24, Weather Data," [9].
The first two references contain normal daily minimum and maximum shade
temperatures as well as the most extreme temperatures ever recorded. The normal
daily minimum and maximum values were the basis of the data in NCHRP Report
276 [3] and are not considered sufficiently extreme. The absolute extreme values
are considered too excessive for design purposes. However, Harvard University's
Climatic Atlas of the United States [8] contains isotherms of the lowest and
highest temperatures experienced during a normal year, and the ASHRAE
Fundamentals Handbook [9] contains tabularized data of outdoor air temperatures
based on 99 percent confidence intervals. During a normal summer, there would

304
be approximately 30 hours at or above the ASHRAE maximum temperature, and,
during a normal winter, there would be approximately 22 hours at or below the
ASHRAE minimum temperature. The last two references were evaluated further
to determine if they yielded reasonable extreme values for minimum and
maximum shade temperatures based on geographical location.

Table 1 is a comparison between the maximum measured bridge movement
and the calculated bridge movement based on the various minimum and
maximum shade temperature sources and four bridges with available measured
movement and temperature data [10]-[12]. Both the Climatic Atlas [8] and the
ASHRAE [9] data are compared to site-measured minimum and maximum shade
temperatures and the NCHRP Report 276 isotherms. In all cases, the minimum
and maximum shade temperatures are converted to effective bridge temperatures
through the application of Emerson's relationships, including the extrapolations in
NCHRP Report 276. The calculated bridge movement is then equal to the product
of the effective temperature range, length and coefficient of linear expansion.

The comparison in Table 1 reveals that isotherms in NCHRP Report 276
significantly under predict the anticipated bridge movements, while site-measured
minimum and maximum shade temperatures provide the closest agreement with
measured bridge movement. Both the Climatic Atlas and the ASHRAE data result
in calculated bridge movements between these two extremes. It is important to
note that the measured movements may not represent extremes for these bridges,
since they were only measured over a 1- to 3-year interval.

It is apparent that both the Climatic Atlas and the ASHRAE data provide more
appropriate minimum and maximum shade temperatures. However, the ASHRAE
data are based on more recent weather data. In addition, with an extensive listing
of 755 individual stations throughout the United States, the data are easier to
apply, particularly in the mountainous western region, than isothermal maps. As a
result, the ASHRAE weather data are recommended for use to determine
minimum and maximum shade temperatures based on bridge location. From these
values, minimum and maximum effective bridge temperatures can be determined
through the application of Emerson's relationships, including the extrapolations in
NCHRP Report 276. In reviewing Emerson's work, however, it was observed that
a nonlinear relationship to determine 24-and 48-hour mean shade temperatures
from maximum and minimum shade temperatures was applied to the data above
82F and below 10F. This nonlinear relationship did not appear justified because
of this lack of available data in the United Kingdom. Therefore, for the FHWA
Jointless Bridge Project, temperature data from 10 cities throughout the United
States, representing a range of climatic conditions, were examined. Regression
analyses for these data relating 24- and 48-hour mean shade temperatures to
maximum and minimum shade temperatures revealed the following:
1. Only a small difference existed between the 24- and 48-hour relationships,
indicating that either value could be applied.
2. A linear relationship appeared justified at all locations throughout the
range of data, with a correlation of 0.97 with all cities combined.
Based on this temperature study, the following relationships are recommended
to determine the 24-hour mean shade temperatures corresponding to the minimum
and maximum shade temperatures taken from the ASHRAE tables:

305
T
24 min
= 1.035 T
min shade
+ 7.43F (1)
T
24 max
= 0.977 T
max shade
- 9.05F (2)
where: T
24 min
= minimum 24-hour mean shade temperature
T
24 max =
maximum 24-hour mean shade temperature

The next step, similar to Emerson's work, is to determine relationships
between the minimum and maximum effective bridge temperatures from T
24 min

and T
24 max,
respectively. For the Jointless Bridge Project, these relationships
were determined based on measurements of two full-scale test bridges constructed
for the project [13]. For the minimum effective bridge temperatures, the effective
bridge temperatures at 6:00 a.m. were measured and were related to a 24-hour
mean shade temperature. An essentially uniform thermal distribution existed
within a bridge girder at this time. For maximum effective bridge temperatures,
the thermocouple located at midheight of the girders was measured at the time of
maximum thermal gradient, approximately 1:00 p.m., and this value was related
to a 24-hour mean shade temperature. This thermocouple reading represents a
reasonable approximation of the maximum effective bridge temperature,
accounting for all of the diurnal effects except direct solar radiation. The data
indicated that a linear relationship existed for both the minimum and maximum
effective bridge temperatures. For concrete, the relationships were as follows:

T
min eff
= 0.963 T
24 min
+ 1.49F (3)
T
max eff
= 0.988 T
24 max
+ 5.45F (4)
For composite steel, the relationships were as follows:
T
min eff
= 1.004 T
24 min
4.16F (5)
T
max eff
= 1.113 T
24 max
+ 7.34F (6)
where: T
min eff
= minimum effective bridge temperature
T
max eff
= maximum effective bridge temperature, excluding direct solar
effects
The final component to be obtained is the direct solar effect. Comparisons of
various proposed thermal gradients with data accumulated from the test girders
indicate that the temperature gradients in AASHTO LRFD [14] provide
reasonable upperbounds to solar radiation differentials. An additional uniform
temperature increment, T , caused by direct solar radiation can be calculated
from these gradients through application of the following equations:

1
1
=
=
=
m
i
i i
i i i i
m
i
t
E A
E A T
avg
(7)
=
g g d d
g g g d d d
E A E A
E A E A

+
+
(8)

306
T =
avg t
(9)

where:
t
avg
= average thermal strain in cross section
i
= coefficient of thermal expansion in subsection i
d
= coefficient of thermal expansion of deck concrete
g
= coefficient of thermal expansion of girder material
= effective coefficient of thermal expansion of a composite cross
section
E
i
= modulus of elasticity in subsections
E
d
= modulus of elasticity of deck
E
g
= modulus of elasticity of girders
T
i
= diurnal increment of temperature in subsection i
A
i
= cross-sectional area of subsection i
A
d
= cross-sectional area of the deck
A
g
= cross-sectional area of the girder

This solar increment, T , is a function of solar zone, girder type, and cross
section. A parametric study of prestressed and composite steel girder cross
sections with a range of deck thickness and girder spacing was therefore carried
out in the Jointless Bridge Project. For the range of girder/deck parameters
studied, the T ranged from 0.18 to 0.28 T
1
for concrete and 0.14 to 0.19 T
1
for
composite steel, where T
1
is the solar incremental temperature at the top surface
of the deck from AASHTO LRFD [14]. Proposed design values for T , based
on a 95 percent probability that they represent upperbound values, are 0.26 T
1
for
concrete and 0.18 T
1
for composite steel.

Combining the above relationships for minimum and maximum shade
temperature versus 24-hour mean shade temperature, 24-hour mean shade
temperature versus effective bridge temperature and direct solar effects, the
equations developed to predict minimum and maximum effective bridge
temperatures for concrete are as follows:
T
min eff
= 1.0T
min shade
+ 9 F (10)
T
max eff
= 0.97T
max shade
-3 F+ T
solar
(11)
For composite steel, the relationships are as follows:
T
min eff
= 1.04T
min shade
+ 3F (12)

T
max eff =
1.09T
min shade
-3 F+ T
solar
(13)
where: T
min eff
= minimum effective bridge temperature
T
max eff
= maximum effective bridge temperature
Tmin shade = minimum shade temperature from the ASHRAE weather
data based on bridge location
Tmax shade = maximum shade temperature from the ASHRAE weather
data based on bridge location
T
solar =
uniform temperature change from direct solar radiation
based on girder type and bridge location equal to 0.26T
1
for
concrete bridges and 0.18T
1
for composite steel bridges
T
1
= solar incremental temperature from AASHTO LRFD.

307
Application of the above relationships to the bridges studied in Table 1
indicates that this approach provides closer agreement with the measured bridge
movement as shown in Table 2.
MONTE CARLO STUDY FOR EFFECTS OF PARAMETER
VARIABILITY
Bridges experience length changes because of temperature changes and time-
dependent volume changes associated with concrete creep and shrinkage. In
integral abutments bridges, it is important to estimate the maximum expansion
and maximum contraction at each end of a bridge to determine the longitudinal
displacement expected for the abutment pile and the available pile strength
associated with the longitudinal displacement. It is also important to determine the
expected movement at each pier and the joint width needed for the approach slab-
to-pavement joint. Another important movement to be estimated is the maximum
total thermal movement at each end that results from the total effective
temperature range. A conclusion from the abutment soil interaction study in the
Jointless Bridge Project [15] is that the starting point to determine maximum
passive pressure should conservatively be at the point of maximum contraction.
The maximum passive pressure is related to the end movement, with re-expansion
for the full effective temperature range.
The calculation of length change for a prestressed concrete bridge can be
accomplished through use of typical design values for the coefficient of thermal
expansion combined with creep and shrinkage strains determined from ACI
209R-92 recommendations [16]. However, overall variability in these factors
provides uncertainty in the calculated end movements. Although a coefficient of
thermal expansion for concrete is typically assumed to be 5.5 to 6.0 millionths/F,
it is known that this value can range from approximately 3.0 to 7.0 millionths/F
[17]. Also, the variability of creep, shrinkage and modulus of elasticity of
concrete is known to be significant [18]. In addition, restraint to length changes
from abutments and piers, combined with the variability of the restraint caused
primarily by the variability of the soil, leads to uneven movement at each end of a
bridge and uncertainty in the magnitude of the movement at each end. Finally, the
effective bridge temperature and the age of the girders at completion of the
superstructure is typically unknown. Therefore, the relative magnitude of
expansion and contraction and the starting point for creep and shrinkage
calculations are uncertain.

Monte Carlo studies were carried out in the Jointless Bridge Project to
investigate the effects of the variability of parameters on the variability of
calculated movements. Two four-span bridge models, one with prestressed concrete
and one with steel composite beams, were selected for the Monte Carlo studies.
Bridge movements were calculated for a large number of runs (typically 2000 runs)
using the statistical variation of material parameters affecting the movement. Within
each run, a computer program randomly selects values, including coefficient of
thermal expansion, temperature at construction, creep and shrinkage parameters of
concrete, modulus of elasticity of concrete and soil stiffness, based on statistical
distributions of the values of these parameters. The variation in calculated bridge-
end abutment movements are then used to determine a 98 percent confidence
interval on maximum calculated movements. The prestressed concrete bridge model
was modified to simulate the conditions of a cast-in-place concrete bridge. A
modification to the prestressed bridge model incorporated the Emanual and Hulsey
method [19] of calculating the coefficient of thermal expansion of concrete.

308

Table 3 presents the values of the magnification factors,, for various
conditions based on the results of the Monte Carlo studies. These factors are
intended to modify the calculated values to account for uncertainty in the
calculations. Case 1 includes the magnification factors for maximum expected
movement from the assumed "as constructed" condition. A primary factor
affecting the magnitude of these magnification factors is the uncertainty of the
construction temperature. The values for total movement account for
uncertainty in the calculation of the overall change in length of the bridge.
However, because of uncertainty in the stiffness of the abutments and piers, the
s for the calculation of movement at each end are somewhat larger.
Case 2 addresses re-expansion from full contraction. Case 3 is a modification of
Case 1(a), using a more certain coefficient of thermal expansion estimate based
on Emanual and Hulsey methodology for calculating the coefficient of thermal
expansion [19].

The following procedure is presented for determining the maximum end
movements of a prestressed concrete integral abutment bridge. (It is assumed that
the bridge has unknown construction timing and no specific data on material
properties are available.):

1. Determine the effective bridge temperature range from Equations (10) and
(11). A mean temperature during the construction season for most
locations can be determined from reference [20].

2. Calculate end movements using:
a. Typical design values of 6.0 millionths/F for coefficient of thermal
expansion, creep and shrinkage values from ACI 209R-92, and modulus
of elasticity of concrete of 57,000
'
c
f , lbf/inch
2
.

b. Use a procedure presented by Zederbaum [21] to determine a point of
zero movement or point of fixity within the bridge based on the stiffness
of the piers and abutments.

3. For maximum expansion shortly after construction (Case 1(a) bridge
expansion, Table 3), use the temperature differential between the
maximum effective bridge temperature and the mean construction
temperature for thermal expansion and a time span equal to one-fourth of
the construction season for creep and shrinkage contraction with the
girders assumed to be 90 days old at the time of casting of the deck. Based
on Monte Carlo studies, the calculated end movements should be
increased by a factor of 1.60 to account for uncertainties with a 98
percent confidence level that the movement will be less than that
calculated.

4. For maximum contraction after several years of service (Case 1(a)
bridge contraction, Table 3), use the temperature differential between the
minimum effective bridge temperature and the mean construction
temperature for thermal contraction, and ultimate creep and shrinkage
values with the girders assumed to be 10 days old at the time of casting of
the deck. Based on the Monte Carlo studies, the calculated end movement
should be increased by a factor of 1.35 to account for uncertainties with

309
a 98 percent confidence level that the movement will be less than that
calculated.

5. For maximum thermal re-expansion from a starting point of full
contraction (Case 2, Table 3), the full range of effective bridge
temperatures should be used without any creep or shrinkage movement.
The resulting calculated end movements should be multiplied by 1.20 to
account for uncertainties in the calculation.

Similar procedures can be used to determine the maximum expansion and
contraction end movements for the cast-in-place concrete bridges. However,
shortening caused by creep is not a factor. Calculated end movements for the cast-
in-place concrete bridges should be increased by the factors for Case 1(b) (1.60
for maximum expansion and 1.40 for maximum contraction).

The procedures used to estimate the maximum end movements of composite
steel bridges are also similar to the procedures outlined above for the prestressed
concrete bridges, except that a modulus of elasticity of 29x10
6
lbf/inch
2
and a
coefficient of thermal expansion of 6.5 millionths/F should be used for the steel
girders. The results of the Monte Carlo study for composite steel bridges indicated
that calculated end movements should be increased by the factors for Case 1(c)
(1.70 for maximum expansion and 1.50 for maximum contraction).

The factor for Case 2, maximum thermal re-expansion from a starting point
of full contraction, is the same for all three types of bridges.

Results of the analyses for the coefficient of thermal expansion of concrete
using the Emanual and Hulsey model [19] indicated that calculated end
movements should be increased by factors of 2.05 for maximum expansion and
1.45 for maximum contraction. The values calculated in this simulation are
greater than the values calculated for the Case 1(a) condition because of the
difference in calculating design thermal movements rather than the variability of
computer-predicted movements. The coefficient of thermal expansion design
value of 6.0 millionths/F recommended by AASHTO and used in the Case 1(a)
study is conservatively high when compared to the average of the database
developed for the Monte Carlo study. The average for the coefficient of thermal
expansion for concrete in the database for the Monte Carlo study is
4.9 millionths/F. Therefore, the conventional design calculated value for
movement using 6.0 millionths/F already includes some margin of safety. The
absolute values of bridge movements calculated by the simulation using the
Emanual and Hulsey model are lower than those calculated for the baseline
condition. This is expected since the Emanual and Hulsey database had a lower
coefficient of variation. Therefore, the use of the Emanual and Hulsey model to
estimate the value of the coefficient of thermal expansion of concrete is
recommended if sufficient information regarding concrete composition is
available to the designer.
A summary of findings and conclusions is as follows:
1. Equations (10) through (13) were developed to determine effective annual
temperature ranges for concrete and composite steel bridges. These
equations, combined with the ASHRAE weather data and the solar

310
increment data provide a reasonably accurate temperature range, on a site-
specific basis, for calculation of bridge expansion and contraction.
2. Magnification factors presented in Table 3 were determined from Monte
Carlo analyses to account for uncertainty in calculations of bridge
movement in order to estimate maximum expected bridge expansion,
contraction and re-expansion from full contraction.

It should be noted that the procedures developed for estimation of bridge
effective temperatures and movement ranges not only apply to jointless bridges,
but can also be used to estimate design movements for expansion joints in jointed
bridges.
Table 1. Comparison Between Measured and Predicted Bridge Movement
Calculated (in)
Bridge
Max./Min.
Shade
Temp. F
Max. Meas.
(in)
w/Site T
w/NCHRP
276
w/Climatic
Atlas
w/ASHRAE

Boone River [9]
=325 ft.
-26/+102 1.93 1.74 1.20 1.65 1.51
Maple River [9]
=320 ft.
-20/+113 2.53 2.41 1.76 2.36 2.18
Kinsport [10]
=2700 ft.
+1/+93 11.17 12.82 9.78 12.96 12.49
Cass County [11]
=450 ft.
-11/+93 2.91 2.79 2.38 2.96 2.92

Table 2. Comparison Between Measured and Modified Predicted Bridge Movement
Boone River Maple River Kingsport Cass County
Max. Meas. (in) 1.93 2.53 11.17 2.91
Calc. w/Eq (10)-(13)(in) 1.74 2.43 13.30 2.94

Table 3. Values of magnification factor
Bridge Expansion Bridge Contraction
Case No. Design Condition
Total End Total End
1(a) Prestressed Bridge 1.50 1.60 1.30 1.35
1(b) Cast-in-Place Concrete 1.50 1.60 1.30 1.40
1(c) Composite Steel 1.50 1.70 1.45 1.50
2
Re-Expansion After Full
Contraction
1.10 1.20 NA NA
3
Emanual and Hulsey
Modification to Case 1(a)
1.85 2.05 1.35 1.45

311
REFERENCES
1. Emerson, M., Bridge Temperatures Estimated From the Shade Temperature, TRRL
Laboratory Report 696, Transport and Road Research Laboratory, Crowthorne,
Berkshire, England, 1976, 54 pp.
2. British Standards Institution, Steel, Concrete, and Composite Bridges, Part I, British
Standard NS 5400, Crowthorne, Berkshire, England, 1978.
3. Imbsen, R.A., et al., Thermal Effects in Concrete Bridge Superstructures, NCHRP Report
276, Transportation Research Board, Washington, DC, September 1985, 99pp.
4. AASHTO, AASHTO Guide Specifications for Thermal Effects in Concrete Bridge
Superstructures, Washington, DC, 1989.
5. Girton, D.D., et al., Validation of Design Recommendations for Integral-Abutment
Piles, Journal of Structural Engineering, ASCE, Vol. 117, No. 7, July 1991, pp. 2117-
2134.
6. U.S. Department of Commerce, Climatic Atlas of the United States, Ashville, NC, June
1968.
7. Daily Normals of Temperature, Precipitation, and Heating and Cooling Degree Days,
1961-1990, Climatography of the United States, No. 84, National Climatic Data Center,
Ashville, NC.
8. Climatic Atlas of the United States, Harvard University Press, 1954.
9. ASHRAE, Fundamentals Handbook, New York, NY, 1993.
10. Girton, D.D., et al., Validation of Design Recommendations for Integral-Abutment Piles,
Iowa DOT Research Project HR-292, Iowa State University, College of Engineering,
September 1989.
11. Tennessee DOT Research Project No. 77-27-2, Thermal Movements of Continuous
Concrete and Steel Structures, Final Report, University of Tennessee, Civil Engineering
Department, January 1982.
12. Jorgenson, J.L., Behavior of Abutment Piles in an Integral Abutment Bridge, Report No.
FHWA-ND-1-75-B, Engineering Experiment Station, North Dakota State University,
Fargo, ND, November 1981.
13. Tabatabai, H.; Oesterle, R.G.; and Lawson, T.J., Jointless and Integral Abutment Bridges,
Experimental Research and Field Studies, Draft Final Report, FHWA, August 2001, 972
pp.
14. AASHTO, AASHTO LRFD Bridge Design Specifications, Third Edition, Washington,
DC 2004.
15. Oesterle, R.G., et al., Jointless and Integral Abutment Bridges, Analytical Research and
Proposed Design Procedures, Draft Final Report, FHWA, March 2002.
16. American Concrete Institute, Committee 209, Prediction of Creep, Shrinkage, and
Temperature Effects in Concrete Structures, ACI 209R-92, Detroit, MI, 1992, 47 pp.
17. Kosmatka, S.H., and Panarese, W.C., Chapter 13: Volume Changes in Concrete,
Design and Control of Concrete Mixtures, Thirteenth Edition, Portland Cement
Association, Skokie, IL 1988, pp. 151-162.
18. Bazant, Z.P., and Panula, L., Creep and Shrinkage Characterization for Analyzing
Prestressed Concrete Structures, PCI Journal, May/June 1980, pp. 86-121.
19. Emanuel, J.H., and Hulsey, J.L., Prediction of the Thermal Coefficient of Expnasion of
Concrete, Journal of American Concrete Institute, Vol. 74, No. 4, April 1977, pp. 149-
155.
20. Expansion Joints in Buildings, Technical Report No. 65, National Academy of Sciences,
1974.
21. Zederbaum, J., Factors Influencing the Longitudinal Movement of Concrete Bridge
Systems With Special Reference to Deck Contraction, Concrete Bridge Design, ACI
Publication SP-23, American Concrete Institute, Detroit, MI, 1969, pp. 75-95.

312
Transverse Movement in Skewed Integral Abutment
Bridges

Ralph G. Oesterle
1
, and Hamid R. Lotfi
1

1
Construction Technology Laboratories, Inc., 5400 Old Orchard Road, Skokie, IL
60077; PH 847-965-7500; ROesterle@CTLGroup.com ; Hlotfi@CTLGroup.com
ABSTRACT
Jointless bridges promote reduced maintenance costs, improved riding quality,
lower impact loads, reduced snowplow damage to decks and approaches and
improved seismic resistance. In spite of many of these recognized benefits, the
behavior of such structures is not yet fully understood, and nationally adopted
design criteria are still lacking. This paper presents results from an experimental
and analytical research program, funded by the Federal Highway Administration,
on the behavior of jointless and integral abutment bridges. The experimental work
included testing and monitoring of bridge models and a bridge structure in the
field, tests of bridge components and a field survey of fifteen jointless bridges.
Experimental results have resolved many questions regarding environmental
effects, and long-term and time dependent loading in combination with live and
dead load. The analytical work evaluated the response of jointless bridges with
respect to various design parameters. The research indicated that analysis
procedures can be used to adequately quantify the structural response if accurate
material and environmental parameters are known. Simplified design procedures
are recommended based on this research.

The analytical phase of the research program included a study of the effects of
skew angle on the response of jointless integral abutment bridges to restrained
longitudinal expansion. With skewed bridges, the soil passive pressure developed
in response to thermal elongation has a component in the transverse direction.
Therefore, skewed bridges respond to temperature change with both longitudinal
and transverse movements. Analyses were carried out to demonstrate the
relationships between skew angle and transverse forces on skewed abutments.
These studies were conducted to provide procedures to either determine forces
required to resist in-plane rotation of the superstructure associated with skew, or
estimate expected movement of skewed abutments not specifically designed to
restrain the movement. The analytical procedures were used to compare with
results from a field study of a skewed bridge that was monitored as part of the
experimental phase of the research program. These procedures were then used to
perform a sensitivity study to demonstrate the relationship between transverse
movement and longitudinal expansion for various skew angles and ratios of
bridge length to width.
BACKGROUND
A skewed bridge has abutments at an angle, , with respect to the bridge deck
as shown in Fig. 1. With skewed integral abutments bridges, the soil passive
pressure developed in response to thermal elongation has a component in the
transverse direction as illustrated in Fig. 1. Within certain limits of the skew
angle, soil friction on the abutment will resist the transverse component of passive

313
pressure. However, if the soil friction is insufficient, then, depending on the
transverse stiffness of the abutment, either significant transverse forces or
significant transverse movements could be generated.

Because of potential problems and uncertainty related to the response of
skewed integral abutments, many state DOTs limit the skew angle. A typical limit
for maximum skew angle for integral abutment bridges used by many states is
30. However, maximum skew angle limits in various states range from 0 to no
limit [1]. Therefore studies were carried out in the FHWA Jointless Bridge Project
to:

1. Develop a relationship between skew angle and abutment soil friction for
limiting skew.
2. Develop a relationship for the magnitude of forces required to restrain
transverse movement in integral abutment bridges with large skew angles.
3. Develop a relationship between skew angle and expected transverse
movement for a typical integral stub abutment with no special design
features to restrain this movement.
4. Compare analytical results with field data for a skewed bridge that was
monitored as part of the experimental portion of this project.
5. Perform a sensitivity study to demonstrate the relationship between
transverse movement and longitudinal expansions for various skew angles
and ratios of bridge length to width.
This work was accomplished by developing equilibrium and compatibility
equations for end abutment forces and, for the case of a typical stub abutment,
solving these relationships for various skew angles and bridge length-to-width
ratios.
ANALYSES FOR TRANSVERSE RESPONSE TO THERMAL
EXPANSION
Skew Angle Limit for Limiting Transverse Effects
Figure 1 shows the passive soil pressure response, P
p
, due to thermal
expansion and soil/abutment interface friction, F
af
, assuming no rotation of the
superstructure. For rotational equilibrium:
F
af
(L cos ) = P
p
(L sin ) (1)
and from interface behavior:
F
af
= P
p
tan (2)
where tan = friction coefficient for interface of formed concrete and soil.
Substituting 2 into 1:
tan =
sin
cos
= tan
= (3)
or

314
Therefore, the bridge superstructure can be held in rotational equilibrium
until the skew angle exceeds the angle of interface friction. Integral abutments are
typically backfilled with granular material. NCHRP Report 343 lists a friction
angle of 22 to 26 for formed concrete against clean gravel, gravel sand
mixtures, and well-graded rock fill with spalls [2]. Based on these data, the angle
of = 20 represents a reasonably conservative skew angle limit below which
special considerations for transverse forces or transverse movement are not
needed.
Forces Required to Resist Transverse Movement
With larger skew angles, the integral abutment can either be designed to resist
the transverse force generated by the soil passive pressure in an attempt to guide
the abutment movement to be predominantly longitudinal or the abutment can be
detailed to accommodate the transverse movement. Adding lateral resistance of
the abutment, F
a
, to wall/soil interface friction, F
af
., in Fig. 1, rotational
equilibrium is found by:
(F
a
+ F
af
) (L cos ) = P
p
(L sin ) (4)
substituting from (2) into (4):
F
a
= P
p
(tan - tan ) (5)
F
a
is the summation of abutment lateral resistance from pile and passive
pressure on the substructure surface perpendicular to the abutment. Figure 2
shows the relationship between F
a
and P
p
, assuming the interface friction angle,,
to be 20. As shown in Fig. 2, the force required to resist transverse movement is
a significant portion of the soil passive pressure, P
p
. It should be noted that P
p
is
not necessarily full passive pressure, but can be determined for the end movement
using relationships calculated by Clough and Duncan [3]. The end movement to
consider in calculating passive pressure is the end movement normal to the
abutment,

n
. As illustrated in Fig. 3, this end movement is:
n
=

cos (6)
where

is the maximum expected end movement for thermal re-expansion
from the starting point of full contraction for the full range of effective bridge
temperatures as discussed in a companion paper in this conference proceedings
[4].
From Fig. 3, it can be seen that

n
is reduced with respect to

as the skew
angle, , increases. This relationship helps offset the increase in F
a
/P
p
with
increasing . However, F
a
will still be a sizeable portion of P
p
.

For relatively short bridges and/or bridges in locations with small effective
temperature ranges, it may be feasible to design the abutment substructure to
resist F
a
. It should be understood though that for whatever means used to develop
F
a
(battered pile and/or lateral passive soil resistance), lateral movements are
required to develop the resistance, F
a
. Therefore, details anticipating some
transverse movement should be used. The expected movements are a function of
the relative stiffnesses of response for P
p
and F
a
. It should also be noted that
adding battered piles to an integral abutment for lateral loading will also increase
the stiffness in the longitudinal direction, which induces more demand on the
superstructure and connections between the girders and abutments.

315

Expected Transverse Movement With Typical Integral Abutment
METHOD OF ANALYSES
To investigate the relationship between skew angle and expected transverse
movement for a typical integral stub abutment, a set of relationships were derived
based on equilibrium and compatibility of end abutment forces in the plane of the
bridge superstructure. For this analysis, the superstructure is assumed to act as a
rigid body with rotation, , about the center of the deck (for a longitudinally
symmetrical bridge). The rotation occurs to accommodate the thermal end
movement, . Forces considered in response to this movement include soil
pressure on the abutment and wingwalls, wall-soil interface friction on the
abutment and pile forces normal to and in line with the abutment and wingwalls.
Details of the forces, stiffness and equations of compatibility and equilibrium are
provided in the report on the analytical work for the Jointless Bridge Project [5].

A spreadsheet program was used to solve for rotational equilibrium in the
plane of the deck. For a given end thermal movement, , the equilibrium
position can be found using an iterative analysis by progressively increasing the
rotation angle, , until the sum of the in-plane moments is zero.

RESULTS OF ANALYSES FOR INSTRUMENTED BRIDGE
As part of the experimental program for the Jointless Bridge Project [6], a
heavily skewed bridge in Tennessee was instrumented and monitored for 1 year.
This bridge carries U.S. Interstate 40 (I-40) over Ramp 2B in Knox County,
Tennessee. It has a three-span steel-plate girder superstructure with an overall
length of 415.92 ft and integral abutments. This structure is sharply skewed with a
skew angle of 59.09. The three span lengths are 139.83 ft, 208 ft, and 68.08 ft.
The superstructure consists of 12 steel-plate girders spaced at 135 inches
composite with a 9.25-inch thick concrete slab. The deck slab consists of stay-in-
place prestressed deck panels and a cast-in-place concrete topping. The bridge
was instrumented to monitor the longitudinal and transverse movements of the
east abutment as well as obtain an indication of restraint to the longitudinal
expansion.

The east abutment wall is supported on 18 HP10x42 piles. Relatively long
wingwalls turned parallel to the longitudinal axis of the bridge are used. These
wingwalls are each supported on three piles. The piles are oriented for strong axis
bending with movement normal to the abutment or wingwalls. Approximately
every other pile is battered at 2:12 in the strong axis direction. The height of the
abutment, is 13.0 ft.

The east abutment was analyzed using the relationships for rotational
equilibrium. Lateral load versus deflection relationships were developed for the
piles using COM624P [7] for weak and strong axis bending in stiff clay (site
condition for the instrumented bridge). The analyses for the effect of the battered
piles [8] indicated that battering at 2:12 does not significantly influence the lateral
stiffness. The Clough and Duncan relationship [3] for loose sand was used for the
passive pressure relationship.

316
Polynominals were fitted to the pile lateral load versus the deflection data
generated by COM624P [7] and a polynomial was fitted to the Clough and
Duncan curve for soil pressure [3]. These polynominals were used with the
equations in the spreadsheet program to solve for the that results in rotational
moment equilibrium.

Based on the experimental data, an end movement of = 0.781 inch was
used in the analysis. The distance to the center of rotation was estimated based on
the ratio of the measured east end movement of 0.781 inches to the overall
calculated bridge expansion of 2.26 inches. Based on this ratio of longitudinal
movement, the center of rotation was estimated to be 143.8 ft from the east end.

The best measurement for superstructure rotation was a deformation gage for
transverse movement located on the center girder at the first interior pier from the
east end. This gage provided consistent data throughout monitoring and
theoretically is not affected by transverse thermal expansion. Based on a range of
movement of 0.203 inch corresponding to a longitudinal end movement of =
0.781 inch, a superstructure rotation angle of = 0.000224 was determined. Using
the spreadsheet to determine rotational equilibrium, an angle of = 0.000226 was
calculated. The calculated value indicated very good agreement with the
measured data. Based on the analysis, the transverse movement at the acute corner
of the east end is 0.688 inch and at the obtuse corner it is 0.092 inch.

In addition to monitoring transverse movement, selected bridge girders were
instrumented with strain gages to obtain an indication of restraint to the
longitudinal expansion. The gages were read at 6:00 a.m. every morning in order
to eliminate strains associated with diurnal temperature gradients. The resulting
changes in strain were used to calculate the longitudinal restraint of movement
from the coldest day to the warmest day. The resulting estimated forces in the
combined deck and steel girders are 355 kips, 153 kips and 43 kips (all
compression) for the north, center, and south instrumented girders, respectively
(An explanation for the variation in compression in these three girders is provided
in the following section of this paper).

The analysis using the spreadsheet for the abutment forces and movement
determined a total longitudinal restraint force of 1755 kips of compression for a
longitudinal end movement of 0.781 inch. Distributing this force to the 12 bridge
girders, the average restraint force is 146 kips. The average of the compressive
restraint forces estimated from the strain gages in the three instrumented girders is
184 kips. These data indicate that the calculated restraint force is the right order of
magnitude and in good agreement with the measured data. In addition, the
measured and calculated data indicate that, with the large skew angle of 59, the
restraint to longitudinal expansion does not have a major influence on the stresses
in the deck and girders. The largest change in strain of 222 millionths relates to a
stress change of 6,660 psi. This change occurs in the positive moment region for
dead and live loads at a location 20 ft from the east abutment. Therefore, at this
location, the restraint compression has a beneficial effect. At the location of the
first interior pier, the moment accompanying the restraint changes sign and the
girder cross sectional area is increased significantly. Therefore, this restraint force
has a negligible effect in the negative moment region.

317

SENSITIVITY ANALYSES FOR THE EFFECTS OF SKEW ANGLE ON
TRANSVERSE MOVEMENT AND LONGITUDINAL RESTRAINT
To demonstrate the effects of skew angle on expected transverse movement
and longitudinal restraint forces, further analyses were carried out using the
spreadsheet program. Variables included skew angle and the length-to-width ratio
for the bridge. The abutment for the instrumented bridge is a relatively typical
type of stub abutment used in Tennessee and was used as the baseline abutment
for the analyses.

The instrumented bridge is relatively wide compared to the length. The ratio
of length to width (L/W) for this bridge is 3.15. To demonstrate the sensitivity to
the bridge L/W ratio, the analyses were repeated for abutments reduced to 2/3W
and 1/3W of the instrumented bridge. This corresponds to L/W ratios of 4.73 and
9.45. The abutment piles were reduced to 12 piles and 9 piles for L/W ratios of
4.73 and 9.45, respectively. The length of the wingwalls at each skew angle were
kept constant.

The length of the wingwalls required for skewed abutments to support the
sloping soil is a function of the skew angle as follows:

w
=
sh
abut
cos
(7)
where = skew angle
h
abut
= height of the abutment
s = slope of the embankment away from the abutment

Three piles were used to support wingwalls with skew angles greater than 40.
Two piles were used to support wing walls with smaller skew angles.

The results of these analyses for the ratio of transverse movement to
longitudinal movement,
t1
/, for a of 1 inch, are shown in Fig. 4. The
transverse movement,
t1
, is the transverse movement of the acute corner of the
bridge deck. This is the corner that experiences the greatest transverse movement
because of the skew angle.

The results in Fig. 4 demonstrate the increase in the transverse movement with
increasing skew angle. The data in Fig. 4 also demonstrate the increase in
transverse movement with decreasing L/W ratio. It should be pointed out that the
change in L/W was accomplished in the analyses by decreasing the width and
keeping the length constant. Also, the length of the wingwall at each skew angle
was constant. Therefore, the results in Fig. 4 predominantly demonstrate the
effects of increasing the ratio of the length of the wingwalls to the length of the
abutment wall. The data in Fig. 4 show that increasing the wingwall length
relative to the abutment wall length (which includes increasing the number of
wingwall piles relative to the number of abutment wall piles) can significantly
decrease the transverse movement. However, the wingwalls and abutment have to
be designed to transmit the wingwall forces into the superstructure.

318
In the instrumented bridge, the estimated restraint forces in the girders
determined from strain gages was 355 kips, 153 kips, and 43 kips for the north,
center and south instrumented girders, respectively. The bridge girder response to
the wingwall moments is a possible explanation for the increase in compression in
the north girder and the decrease in compression in the south girder as compared
to the center girder. The moments induced on the north wingwall will induce
tension (decreased compression) in the north exterior girder and will increase
compression on the first north interior girder. The north instrumented girder was
the first north interior girder. Likewise, the moments induced on the south
wingwall will induce increased compression on the south exterior girder and
tension (decreased compression) on the first south interior girder. The south
instrumented girder was the first south interior girder.

Figure 5 shows the resulting total longitudinal restraint force for these
analyses and demonstrates the decrease in longitudinal restraint with increasing
skew angle. For the full-width bridge with L/W = 3.15, the longitudinal restraint
at a skew angle of = 60 is approximately 60 percent of the longitudinal
restraint at = 25. For the larger L/W = 9.45, the ratio of longitudinal restraint at
= 60 is approximately 70 percent of the restraint at = 25. This demonstrates
the increase in restraint resulting from the increase in resistance to lateral moment
because of the larger ratio of wingwall length to abutment length.

Since the baseline abutment used in these analyses is a relatively typical stub
abutment (but also relatively deep, with an abutment height of 13.0 ft and with
strong axis pile bending for movement normal to the abutment versus weak axis
bending for movement parallel to the abutment), the data in Fig. 4 represent a
reasonably large estimate for the transverse movement of skewed abutments.
Although there is significant uncertainty for actual soil and pile stiffness, the
maximum expected end movement, , discussed in the companion paper [4],
includes a multiplier to account for uncertainty. Therefore, it is suggested that the
data in Fig. 4 can be used by designers to determine an approximate estimate for
expected transverse movement in skewed integral abutments resulting from the
restraint of longitudinal thermal expansion. In addition, the relationships between
longitudinal restraint force and skew angle shown in Fig. 5 can be used to
estimate the relative decrease in restraint forces in a skewed bridge. Also, the
transverse movements can be used to estimate the transverse forces on the
wingwall resulting from passive soil load and pile and to estimate longitudinal
and transverse movement for the abutment pile for biaxial bending considerations.
All of the other components of movement and forces can be determined from
t1

and using equations presented in the full analytical report [5]. For more
accurate estimates, the equations in the full report can be used to analyze specific
skewed abutments (and/or be modified for other pile orientations).

It should be noted that the transverse movement,
t1
, discussed above is the
movement of the acute corner of the bridge deck related to rigid body rotation of
the superstructure resulting from the abutment passive restraint of longitudinal
thermal expansion, . There will also be transverse thermal expansion of the
abutment. This transverse thermal expansion will be limited compared to the
longitudinal expansion, because the temperature change is moderated by the fact
that the abutment wall is exposed to the ambient air temperature on one side only
and the abutment is not appreciably exposed to solar radiation. Therefore,
depending on the width of the bridge and the skew angle, the transverse thermal
expansion may or may not add significant additional transverse movement to the
abutment wingwalls. This additional transverse movement would add to
t1
at the

319
acute corner and subtract somewhat from transverse movement at the obtuse
corner,
t2
. In order to estimate this transverse thermal expansion of the abutment,
it is suggested that the temperature range be determined from the difference
between the maximum 24-hour mean temperature, T
24
max
, and the mean
temperature during construction as described in the companion paper [4]. This
temperature range can be used with half the bridge width, W/2, to estimate an
additional transverse movement at the acute corner to be added to the
t1
and
subtracted from
t2
determined from the rigid body rotation.

To detail the abutments for the transverse movement of the corners, all
interfaces of the integral abutment with other components, such as approach
pavement, barrier walls, pavement for slope protection and drainage components,
should be detailed to accommodate this movement. In addition, the foundation
and pier structure stiffness will likely be significant for movement parallel to the
pier cap. Therefore, it is recommended that the connection between the bottoms of
the girders and diaphragms and the pier caps be flexible in this direction. This
approach, however, may not be appropriate for seismic design. In this case, design
of the diaphragms should consider the interior pier restraint of the rigid body
rotations that result from passive abutment restraint of longitudinal thermal
expansion.


A summary of findings and conclusions is as follows:

1. Integral abutments for skewed jointless bridges experience a component of
soil passive pressure in the transverse direction in response to thermal
elongation.
2. A skew angle of 20 represents a reasonable upper limit for skewed
integral bridges below which special considerations for transverse forces
or transverse movements are not needed.
3. With larger skew angles, the integral abutment can either be designed to
resist the transverse forces generated by the soil passive pressure or the
abutment can be detailed to accommodate the transverse movement.
4. With increasing skew angle, the transverse force required to resist
transverse movement becomes a sizeable portion of the passive pressure
resisting the movement normal to the abutment,
n
. It may only be
feasible to design sharply skewed abutments to resist transverse movement
for relatively short bridges and/or bridges in locations with small effective
annual temperature ranges.
5. Relationships were developed for the in-plane rotational equilibrium of
skewed integral abutment bridges. The results of the analysis for
transverse movement and longitudinal restraint forces based on these
relationships are in good agreement with data from the bridge that was
instrumented as part of the experimental program.
6. Results of analyses in Figs. 4 and 5 can be used to determine an
approximate estimate for expected transverse movement and relative
reduction in longitudinal restraint forces in a typical integral stub
abutment. For more accurate estimates, the equations in the full analytical
report [5] can be used to analyze specific skewed abutments (and/or
modified for other pile orientations) to determine transverse movement
related to rigid body rotation of the superstructure. Consideration should
also be given to transverse thermal expansion of the abutment.

320

L
F
af
P
p
P
p
F
af
W
L
s
in

L

c
o
s

Figure 1. Soil pressure load, P
p
, and soil abutment interface friction, F
af
.
Fa
0 10 20 30 40 50 60
0.4
0.8
1.2
0
Pp
0.213
0.475
0.828
1.368
tan - tan
for = 20
Skew Angle,

Figure 2. Relationship between force required for abutment lateral resistance, F
a
, and
passive pressure response, P
p
, to restrain lateral movement.
10 20 30 40 50 60 70
n
0
0.5
1.0 0.94
0.87
0.77
0.64
0.50
0.34
Skew Angle,
=
n

Figure 3. Relationship between end normal movement, l
n
, and end thermal expansion, l.

321
L/W = 3.15,
Skew Angle (degrees)
65 60 55 50 45 40 35 30 25 20
1.2
1.0
0.8
0.6
0.4
0.2
0

t1
S /
w
S
abut
= 0.19
L/W = 4.73, S /
w
S
abut
= 0.29
L/W = 9.45, S /
w
S
abut
= 0.58

Figure 4. Relationship between transverse movement at the acute corner,
t1
, and thermal
expansion, l, for an expansion of 1 inch with constant length bridge, L = 415.92 ft, and
varying L/W.

20 25 30 35 40 45 50 55 60
Skew Angle (degrees)
500
1000
1500
2000
2500
3000
R
e
s
u
l
t
a
n
t

F
o
r
c
e

(
k
i
p
s
)
L/W = 3.15, S /
w
S
abut
= 0.19
L/W = 4.73, S /
w
S
abut
= 0.29
L/W = 9.45, S /
w
S
abut
= 0.58

Figure 5. Relationship between resultant longitudinal restraint force and skew angle for
thermal expansion, l, of 1 inch with constant length bridge, L = 415.92 ft, and varying L/W.

322

REFERENCES
1. Chandra, V., et al., Draft Report on Precast Prestressed Concrete Integral Bridges,
State-of-the-Art, Precast/Prestressed Concrete Institute, 1995, 113 pp.
2. Barker, R.M.; Duncan, J.M.; Rojiani, K.B.; Ooi, P.S.K.; Tan, C.K.; and Kim, S.G.
Manual for the Design of Bridge Foundations, NCHRP Report 343, Transportation
Research Board, Washington, DC, December 1991, 308 pp.
3. Clough, G.W., and Duncan, J.M., Earth Pressures, Foundation Engineering
Handbook, Second Edition, Edited by H.Y. Fung, Van Nostrand Reinhold, New
York, NY, 1991, pp. 223-235.
4. Oesterle, R.G., and Volz, S.V., Effective Temperature and Longitudinal Movement
in Integral Abutment Bridges, Proceeding of the FHWA Conference on Integral
Abutment and Jointless Bridges, (IAJB2005), Baltimore, March 16-18, 2005.
5. Oesterle, R.G., et al., Jointless and Integral Abutment Bridges, Analytical Research
and Proposed Design Procedures, Draft Final Report, FHWA, March 2002.
6. Tabatabai, H.; Oesterle, R.G.; and Lawson, T.J., Jointless and Integral Abutment
Bridges, Experimental Research and Field Studies, Draft Final Report, FHWA,
August 2001, 972 pp.
7. Shih-Tower, W., and Reese, L.C., COM624P Laterally Loaded Pile Analysis
Program for Microcomputer, Version 2.0, Report No. FHWA-SA-91-048, FHWA,
Office of Technology Applications, Washington, DC, 1991.
8. Poulos, H.G., and Davis, E.H., Pile Foundation Analysis and Design, John Wiley and
Sons, 1980, 397 pp.

323
Soil-Structure Interaction of Jointless Bridges

Olli Kerokoski, Lic.Tech. M.Sc. (Civ.Eng.), Researcher olli.kerokoski@tut.fi
Anssi Laaksonen, M.Sc. (Civ.Eng.), Researcher anssi.laaksonen@tut.fi
Laboratory of Foundation and Earth Structures, Tampere University of
Technology, Finland

ABSTRACT

In the research project presented in this paper the main subjects of interest are:
earth pressures after cyclic abutment displacements, behavior of pavement near
the abutments and bridge construction details. The instrumentation of
Haavistonjoki Bridge at Tampere-Jyvskyl highway has been completed and is
discussed in this paper. The results show, for example, that the behavior of large
steel pipe piles under an integral abutment can be sufficiently predicted by
structural calculations. The measured earth pressures on the bridge abutments
were quite high because the backfill was well compacted.

INTRODUCTION

Traditionally, long bridges have been built with bearings and expansion joints.
Because of several problems involved in the traditional practice, the jointless
practice has been adopted. Jointless or integral abutment bridges have their
superstructure cast integrally with their substructure. The superstructure is
permitted to expand and contract without joints.

In Finland more than 1000 jointless bridges have been built during recent
decades. The maximum length of a symmetrical road bridge has been 70 m. The
passive earth pressure on the integral abutment is estimated to be mobilized after
a quite small abutment displacement, for example 0,002 times the height of the
abutment in dense sand.

Practice in Some Other Countries in Europe

In North Europe concrete slab bridges and concrete slab beam bridges are
commonly used. In North Europe and in Great Britain the allowable expansion
length is approximately the same. The effect of temperature fluctuation and
corresponding cyclic abutment movement on earth pressures is considered in
Swedish and in British standards. In Germany there are no specific standards for
the integral bridges, but the research and development has been going on during
recent years and several integral bridges have been constructed.

324
General Principals and Limitations

Integral abutment bridges fit best for places, where there is enough space to
build a sufficiently long deck with low-gradient slope, and where the embankment
is stable and its settlements are small.

As the wall supporting the earth moves horizontally towards the earth, the
earth pressure is increasing non-linearly from the at-rest pressure to the ultimate
pressure called passive earth pressure. Coulomb earth pressure theory considers
the friction between the wall and the soil, and defines the passive earth pressure
precisely enough, if and only if the wall friction angle is less than 0,4 times the
internal friction angle of the ground.

The horizontal displacements presented in research reports and design
manuals required to activate passive earth pressure vary considerably. Generally it
is estimated that the relative displacement magnitude of 0,0050,050 times H is
adequate, where H is the height of the wall. According to a group of German
researchers the displacement equal to 0,025 x H is enough to mobilize the passive
earth pressure. The displacement equal to 0,0025 x H is enough to mobilize half
of the passive earth pressure, and that is also the recommended maximum
displacement. The German recommendation for the maximum top displacement
of a wall rotating around its base is 4 times the maximum of horizontally moving
wall. In many research reports the earth pressure distribution on the rotating wall
is stated to be totally different from the horizontally moving wall case.

Various means to minimize the earth pressure at the bridge abutment and the
irreversible strains inside the approach embankment have been found. An elastic
vertical board behind the abutment and reinforcement of backfill material are
acting satisfactorily together.

Frozen soil is plastic and the mechanical properties of soil depend, e.g., on the
loading rate, the temperature and the stress state. The horizontal displacement of
bridge piers according to temperature change is very slow compared to general
strain rates measured in tri-axial and compression tests. Therefore there is no
physical basis for the Finnish practice to use rigid horizontal support at the
surface of the groundwater level. Regardless, frozen soil is much harder and
stiffer than unfrozen soil.

The information on long and skewed integral bridges is limited concerning
circumstances and practice in North European countries. However, the earth
pressures at the obtuse corner were found to be much higher than at the acute
corner.

In addition to the axial loading, the cyclic abutment movement and rotation
and the embankment strains exert forces to the abutment piles. These forces
reduce considerably the vertical load capacity of slender piles. As the integral

325
abutment and the embankment move simultaneously, the horizontal resistance of
the soil against the piles is found to be smaller. Therefore it is easier for the pile
top to move horizontally inside the soil compared to an individual pile.

TEST SITE AND TEST PROGRAM

The Haavistonjoki Bridge in Finland is a 50 m long 3-span slab bridge. On the
bridge, four piers and two abutments support the reinforced concrete deck
according to Figures 1 3. Eight 711 mm steel pipe piles support the piers and
the abutments. The bridge was constructed and instrumented in the summer of
2003. The test program included almost 200 gauges, from which a data logger
automatically collects the data to a computer in every 15 minutes for many years.

Figure 1 Haavistonjoki Bridge immediately after accomplishment of construction.

326

Figure 2 Haavistonjoki Bridge on Highway number 9 between Tampere and Jyvskyl.
Side view. The transition slab is situated inside of the embankment.

Figure 3 Haavistonjoki Bridge. Cross-section.

Backfill Properties

The backfill material was well-compacted # 060 mm crushed rock. The
material was tested at the geotechnical laboratory of Tampere University of
Technology, see Figures 4 6.
0
100
200
300
400
500
600
0 100 200 300 400 500 600 700 800
Sum of principal stresses, kPa
R
e
s
i
l
i
e
n
t

-
m
o
d
u
l
u
s
,

M
P
a

Figure 4 Results of Haavistonjoki backfill large-scale cyclic tri-axial test. Resilient modulus
at various stress levels.

327

0
300
600
900
1200
1500
0 0,25 0,5 0,75 1 1,25
Axial strain, %
A
x
i
a
l

c
o
m
p
r
e
s
s
i
o
n

s
t
r
e
s
s
,

k
P
a

Figure 5 Results of Haavistonjoki backfill large-scale static tri-axial test. Cell pressures
were 15, 30, 60, 100 and 150 kPa.

The replacement fill area and some site investigation results are presented in
Figure 6.

Figure 6 Basic soil material data. Replacement fill area: unit weight = 20 kN/m
3
and
friction angle = 42. Earth pressure cells F and G (label 2.2) at Haavistonjoki bridge pier
T3.

ESTIMATED MAXIMUM SHEAR STRAIN IN BACKFILL

Shear strain in the backfill near the abutment can be approximated by the
equation (Bolton 1993, Koskinen 1997):

328

= 2 * y / H,

where is shear strain, y is bridge horizontal displacement and H is abutment
height.

The horizontal bridge abutment displacement y
max
from the temperature
change T = 30 C in Haavistonjoki Bridge is following:

y
max
= * T * L/2 = 10 * 10
-6
1/C * 30 C * 50 000 mm / 2 = 7,5 mm,

where L is bridge length and is coefficient of thermal expansion.

The corresponding shear strain is

= 2 * y / H = 2 * 7,5 mm / 2500 mm = 6 * 10
-3

Shear modulus at a certain shear strain level may be estimated with the
diagram in Figure 7. In Haavistonjoki Bridge, the shear modulus of the backfill is
decreasing from G
max
to G = (0,1 0,2) x G
max
due to the large deformations
according to Figure 7. As a result, the plastic strains in a backfill material will be
quite large.

Figure 7 Typical variation in shear modulus according to shear strain level (Tiehallinto
2001).

TEST RESULTS

The main results have been presented in Figures 8 to 10. The focus is on the
period between 11.2. 15.2.2005 as the maximum changes take place.

In Figure 8 the measured earth pressures and measuring points are presented.
Earth pressures are found to be smaller near the wing walls.

329
Figure 9 shows the earth pressure changes and bridge temperatures during
11.2. 15.2.2005. The earth pressures follow the temperature changes, increasing
as temperature rises.

Figure 8 Measured earth pressures on bridge abutment T4 between 11.2.... 15.2.2004 at
earth pressure cells HQ [kPa].

330
-20
0
20
40
60
80
100
120
140
160
180
10.2.04 11.2.04 12.2.04 13.2.04 14.2.04 15.2.04 16.2.04
E
a
r
t
h

p
r
e
s
s
u
r
e

[
k
N
/
m

2
]
-20
-18
-16
-14
-12
-10
-8
-6
-4
-2
0
T
e
m
p
e
r
a
t
u
r
e

[
C
]
H
I
J
K
L
M
N
O
P
Q
V
W
Bridge
temperature

Figure 9 Measured changes in average bridge temperature and earth pressures on bridge
abutment between 11.2.... 15.2.2004 at earth pressure cells HQ [kPa].

Figure 10 presents the measured horizontal abutment displacements at support T4
under the transition slab and the corresponding earth pressures at pier T3. The
earth pressure increase varies between 16 20 kPa in cells F and G.

331
-35
-30
-25
-20
-15
-10
-5
0
1.2.04 5.2.04 9.2.04 13.2.04 17.2.04 21.2.04 25.2.04 29.2.04
Date
E
a
r
t
h

p
r
e
s
s
u
r
e

[
k
N
/
m
2
]
-5
-4
-3
-2
-1
0
1
2
D
i
s
p
l
a
c
e
m
e
n
t

[
m
m
]
F G Displacement at support T4

Figure 10 Measured horizontal abutment displacements at support T4 under the transition
slab and corresponding earth pressures at Haavistonjoki bridge pier T3 in February 2004.

COMPARISON WITH A FEM-MODEL

In a 2D-FEM-model of Haavistonjoki Bridge columns and piles have been
modeled by calculating the real bending stiffness and sectional area and dividing
the values by the bridge half width.

The steel pipe piles are supposed to behave like a composite structure because
of reinforced concrete inside the piles.

The coefficient of elasticity, E, in the soil in the vicinity of the piles is reduced
by a factor (3 times pile width/ bridge half width) = 3 * 0,711m / 5,75m = 0,371,
see Figure 11.

332

Figure 11 Plaxis 2D-FEM-model of Haavistonjoki Bridge.

Plaxis 2D FEM code and Hardening Soil earth model were used to calculate
the Haavistonjoki Bridge deformations and stresses. The results of two separate
crushed rock elastic modules were compared, namely E
crushed_rock
= 80 000 kPa
and E
crushed_rock
= 250 000 kPa. Other soil parameters used are presented in
Table1.

Table 1 Properties of soil.
Soil
material

[kN/m
3
]
sat

[kN/m
3
]
E
50
ref
= E
oed
ref
or E
[kN/m
2
]
[] [] R
inter

Crushed
rock
22 23,1 80 000 0,3 45 8 0,5
Crushed
rock,
reduced
20 22,0 29 700 0,3 42 5 0,19
Clayey
Silt,
reduced
17 18,1 9 300 0,3 33 0 0,19
Moraine 23 23,0 80 000 0,3 45 4 0,7

Extra soil data:
Cohesion c = 1 kPa 0 kPa,
In crushed rock the soil dilatation in embankment and surrounding the
piles ended when the unit weight reduced to 19 kN/m
3
,
Tension cut-off was on,
Unloading-reloading modulus E
ur
ref
= 3 x E
50
ref
,
E
50
= Confining stress dependent stiffness modulus
,

E
50
= E
50
ref
x (-
3
/ p
ref
)
m
, if cohesion is c = 0,
Amount of stress dependency is given by power m = 0,5,
Reference pressure p
ref
= 30 kPa = minor principal stress
3
.

The calculations were done in nine construction stages according to real
construction stages. The results have been presented in Figures 12 15.

333

Because of the large deformations in the soft soil below the backfill, the
average pressure at the bottom of abutment was about 80 kPa before it moved
according to a prescribed displacement. So the increase in earth pressure was only
about 120 kPa - 80 kPa = 40 kPa with crushed rock with E
50
ref
= 80 000 kN/m
2
.
With crushed rock E
50
ref
= 250 000 kN/m
2
the increase was 170 kPa - 80 kPa = 90
kPa, which corresponds to measured values.

Figure 12 Calculated earth pressures [kN/m
2
] on Haavistonjoki Bridge abutment after
forced displacement of 7 mm at deck level. The transition slab at level +19,2 m is clearly
visible. The average total earth pressure under the transition slab is 60 kPa and at the
bottom of the abutment 120 kPa. E
crushed_rock
= 80 000 kPa.

Figure 13 Deformed mesh from Haavistonjoki Bridge 2D-FEM-model. Displacements are
scaled up 200 times. Extreme displacement is 7,7 mm. Abutment rotation is (7,44mm -
5,91mm) /2 400mm = 1/1570. E
crushed_rock
= 80 000 kPa.

334

Figure 14 Relative shear stresses at Haavistonjoki Bridge abutment. E
crushed_rock
= 80 000
kPa.

Figure 15 Total shear stresses at the 5 m long transition slab top and bottom surface after 7
mm abutment displacement. Dotted lines indicate limit shear resistance. Maximum shear
stress at the slab support is 114 kPa and at the end 27 kPa. E
crushed_rock
= 80 000 kPa.

The Haavistonjoki Bridge was also analyzed with a 3D FEM-model and soil
displacement-force- springs. In Figure 16 the Finnish practice for pile horizontal
displacement resistance is introduced.

335

Figure 16. Horizontal displacement according to subgrade coefficient for large diameter
piles. p
m
= ultimate limit of horizontal resistance, y
m
= corresponding displacement
(Tiehallinto 1999)

SUMMARY

A transition slab is a stiff constructional member in the longitudinal direction
of bridge structure compared to soil stiffness. The majority of high soil stresses
and deformations occur several meters apart from the abutment. Earth pressures
against the abutment are quite small under the transition slab. Therefore also the
abutment rotation is smaller than rotation calculated with a 3D soil spring model
without a transition slab.

The transition slab serves as a 1-span beam or slab carrying the overburden
and dead load of the transition slab.

The abutment displacement influences directly on the soil deformations and
stresses around the upper part of abutment supporting piles. There isnt a
stationary and independent pile supporting soil layer around the piles, because the
abutment movement modifies the surrounding soil.

The 2D-model of a 3D-reality satisfactorily represents the real structure with
the assumptions used.

The jointless bridge research work will continue for some years at Tampere
University of Technology.

REFERENCES

1. Bolton M. D.: What are partial factors for? Limit State Design in Geotechnical
Engineering, International Symposium, Vol. 3. Danish Geotechnical Society,
Copenhagen, p. 565 583. 1993.

2. Kerokoski Olli: Soil-Structure Interaction of Jointless Bridges. Literary research.
Tampere University of Technology. 2005. (in Finnish)

3. Kerokoski Olli: Soil-Structure Interaction of Jointless Bridges. Calculation. Tampere
University of Technology. 2005. (in Finnish)

4. Kerokoski Olli: Soil-Structure Interaction of Jointless Railway Bridges. Literary research.
2005. (in Finnish)

336

5. Koskinen, Mauri: Soil-Structure Interaction of Jointless Bridges on Piles. Doctoral thesis.
Tampere University of Technology, Publications 200. 1997.

6. Laaksonen Anssi. Soil-Structure Interaction of Jointless Bridges. Haavistonjoki Bridge
instrumentation. Master thesis. Tampere University of Technology. 2004. (in Finnish)

7. Plaxis Manual. 2D - Version 8. A.A. Balkema publishers. Netherlands 2002.

8. Tiehallinto (Finnish Road Administration): Steel Pipe Piles. Code of practice. 1999. (in
Finnish)

9. http://alk.tiehallinto.fi/sillat/julkaisut/steelpipepiles1999.pdf Third edition. 2000. (in
English)

10. Tiehallinto (Finnish Road Administration): The supplement bridge design instructions.
2002. (in Finnish)

337
AUTHOR INDEX

Alampalli, S. 41
Aluri, S. 113
Arockiasamy, M. 185
Ashraf, S. 73

Basu, P. 244
Bergmann, M. 73
Bermudez, R. 199
Bonczar, C. 163, 174
Bowser, J. 125
Brea, S. 163, 174
Burdette, E. 222

Camp, J. 125
Chang, J. 211
Chovichien, V. 30
Christou, P. 233
Civjan, S. 163, 174
Conboy, D. 50
Crellin, B. 163
Crovo, D. 163, 174

Deatherage, J. 222
DeJong, J. 163, 174
Deng, Y. 211
Dunker, K. 136

Farre, J. 211
Frosch, R. 30

GangaRao, H. 113
Goodpasture, D. 222
Grajales, B. 292

Hassiotis, S. 199
Hoit, M. 233
Hoppe, E. 257
Horvath, J. 281
Howard, S. 222
Huckabee, P. 270
Huh, B. 148
Husain, I. 148

Ingram, E. 222
Jayakumaran, S. 73

Kerokoski, O. 323
Knickerbocker, D. 244
Kutschke, W. 292

Laaksonen, A. 323
Liu, D. 136
Lopez, J. 199
Lotfi, H. 312
Low, J. 148

Maberry, S. 125
Magliola, R. 136
Maruri, R. 12
McCormick, M. 148
McVay, M. 233
Michael, K. 97
Mistry, V. 3

Norrish, C. 73

Oesterle, R. 302, 312

Penafiel, P. 211
Perkun, J. 97
Peterson, B. 84
Petro, S. 12

Shekar, V. 113
Sivakumar, M. 185
Stoothoff, E. 50

Volz, J. 302

Wasserman, E. 244
Weakley, K. 61
Wenning, M. 30
Wetmore, J. 84
White, H. 41

Yannotti, A. 41

Integral Abutment and Jointless Bridges

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Integral Abutment and Jointless Bridges

Uploaded by

Copyright:

Available Formats

THE 2005 FHWA CONFERENCE

Integral Abutment and Jointless Bridges

software which provides the ability to model a bridge

software which provides

[3] was used for

is capable of modeling a system which consists of

can perform static, dynamic or pushover analysis.

Users Guide [3]. The element is very similar to a conventional

is able to model multiple sections

provides the option to use

) based on layers (Figure

model yielded a displacement at the top of the

analysis, the lateral soil springs were calculated based on

soil model includes the presence of vertical (t-z

Users Guide, version 4: Bridge Software Institute. University of Florida,

You might also like