Professional Documents
Culture Documents
Integral Abutment and Jointless Bridges
Integral Abutment and Jointless Bridges
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
TABLE OF CONTENTS
Session I: Current Practices with Design Guidelines and Foundation Design
Integral Abutment and Jointless Bridges
V. Mistry
12
30
41
50
61
73
84
97
113
iii
125
136
148
163
Integral Abutment Pile Behavior and Design Field Data and FEM Studies
C. Bonczar, S. Brea, S. Civjan, J. DeJong, D. Crovo
174
185
199
211
222
233
244
iv
257
270
281
292
302
312
323
Author Index
337
vi
SESSION I:
CURRENT PRACTICES WITH DESIGN
GUIDELINES AND FOUNDATION
DESIGN
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.
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 costeffective 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
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.
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.
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
10
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.
11
1
2
12
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.
13
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.
14
Table 1: Number of IAJB Designed and Built Since 1995 and In-Service
DESIGNED
BUILT
IN SERVICE
since 1995
since 1995
(TOTAL)
Integral Abutment
~ 7000
~ 8900
~ 13000
Full Integral
~ 5700
~ 6400
~ 9000
Semi Integral
~ 1600
~ 1600
~ 4000
Deck Extension
~ 1100
~ 1100
~ 3900
21%
21%
20%
PERCENT OF STATES
18%
18%
15%
10%
10%
10%
8%
8%
5% 5%
5% 5%
501 - 1000
Over 1000
3% 3%
0%
None
1 - 10
11 - 20
21 - 50
51 - 100
101 - 500
Figure 2: Percent of States that reported designing and building Integral Abutments within
the range specified (since 1995).
15
PRESTRESSED
CONCRETE
GIRDERS
MAXIMUM SPAN
Full Integral
Semi Integral
Deck extensions
Integral Piers
TOTAL LENGTH
Full Integral
Semi Integral
Deck extensions
Integral Piers
MAXIMUM SKEW
Full Integral
Semi Integral
Deck extensions
Integral Piers
MAXIMUM
CURVATURE
Full Integral
Semi Integral
Deck extensions
Integral Piers
RANGE
STEEL GIRDERS
60 200
90 200
90 200
120 200
150 1175
90 3280
200 750
300 400
15 70
20 45
20 45
15 80
0 10
0 10
0 10
3 - No Limit
16
MAXIMUM SPAN
Full Integral
Semi Integral
Deck extensions
Integral Piers
TOTAL LENGTH
Full Integral
Semi Integral
Deck extensions
Integral Piers
MAXIMUM SKEW
Full Integral
Semi Integral
Deck extensions
Integral Piers
MAXIMUM
CURVATURE
Full Integral
Semi Integral
Deck extensions
Integral Piers
RANGE
65 - 300
65 - 200
80 - 200
100 - 300
150 - 650
90 - 500
200 - 450
150 - 1000
15 - 70
30 - 40
20 - 45
15-No Limit
0 - 10
0 - 10
0 - 10
0 - No Limit
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 nonuniformity 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.
17
40%
DESIGN PROCEDURES
36%
36%
DETAILS
PERCENT OF STATES
30%
28%
24%
23%
21%
20%
18%
15%
13%
13%
10%
10%
8%
0%
Primary Load
Considerations
Secondary Load
Considerations
Substructure/foundation
Abutment/Backfill
Approach Slab
Jointless retrofit
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
18
100%
92%
90%
90%
80%
77%
79%
PERCENT OF STATES
70%
60%
54%
50%
40%
30%
20%
10%
0%
Design integral (fully and
semi) abutments whenever
possible.
Figure 4: Percent of States that Answered <YES> with Regard to Their FUTURE Plans for
Jointless Bridge Construction.
19
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/ft2 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.
72%
70%
60%
59%
PERCENT OF STATES
50%
44%
41%
40%
33%
28%
30%
21%
20%
15%
10%
0%
Passive Earth
Pressure
Temperature.
Creep
Shrinkage
Settlement
Additional forces
due to skew
Additional forces
due to curvature
Other. Describe
below in
comments
section.
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.
20
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 fullintegral 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.
21
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.
50%
46%
40%
PERCENT OF STATES
33%
30%
20%
13%
10%
8%
0%
Strong Axis Parallel to CL of
Bearings
Designer's Option
ORIENTATION OF PILES
22
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, FoundationQuestion 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.
33%
30%
PERCENT OF STATES
26%
20%
18%
10%
8%
0%
Combination
Active Pressure
Passive Pressure
23
Other
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).
51%
50%
41%
PERCENT OF STATES
40%
30%
26%
20%
10%
0%
Axial Forces (No Bending)
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
24
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.
70%
69%
60%
PERCENT OF STATES
50%
40%
30%
20%
13%
13%
10%
8%
10%
8%
10%
0%
Require compacted
Require
backfill
uncompacted backfill
Use Expanded
Polystyrene (EPS)
Use other
compressible
material behind
abutment.
Require additional
inspection during
construction
Other. Describe in
comments section
below.
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).
25
26
50%
46%
PERCENT OF STATES
40%
30%
28%
26%
20%
15%
10%
10%
8%
3%
0%
Cracking of integral
abut. backwall
Detailing
Detrimental rotation
of integral abutment
backwall
27
Setlement of
approach slabs
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 nonuniformity 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.
28
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.
29
30
31
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.
Approach Slab
Sub-Base
Sleeper Slab
Expanded
Polystyrene
Fill
HP14x89
Ground Level
Deck
Min. 18
18
17
Prestressed
Bulb Tee
Girder
Anchor Plate
and Beam Seat
Retaining Wall
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 inservice behavior of integral abutment bridges as well as the behavior of the piles
32
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.
33
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, soilstructure 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.
R/C Bridge
Approach Slab
8 Slab
Prestressed Conc.
Bulb-T Beam
Convergence Meter
Strain Gage
Drain
Ground Level
1-3
3-3
14 CFT pile
Reference Pile
Tiltmeter
Strain Gages
10
90
Ground
Level
5 spaces at 4=20
1-3
Parallel to
Roadway
Strain Gage
Centerline of
Abutment
Pile
Plan
Elevation
34
ABUTMENT BEHAVIOR
Deck
Coldest Day
Girder
Construction Temperature
Hottest Day
04/26/04
03/27/04
02/26/04
01/27/04
12/28/03
11/28/03
10/29/03
09/29/03
08/30/03
07/31/03
Construction Day
07/01/03
120
110
100
90
80
70
60
50
40
30
20
10
0
-10
06/01/03
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.
35
1.5
Rotation (degrees)
Coldest Day
Hottest Day
1.0
Construction Day
0.5
Bent 1
Average Bent 1
0.0
-0.5
Average Bent 6
-1.0
04/26/04
03/27/04
02/26/04
01/27/04
12/28/03
10/29/03
09/29/03
08/30/03
07/31/03
07/01/03
06/01/03
11/28/03
Bent 6
-1.5
0.7
Coldest Day
Hottest Day
0.6
Construction Day
0.5
0.4
0.3
0.2
Inward
0.1
Bent 1 Center
0.0
36
04/26/04
03/27/04
02/26/04
11/28/03
10/29/03
09/29/03
08/30/03
07/31/03
07/01/03
-0.2
01/27/04
Bent 6 Center
Bent 6 SE
Outward
12/28/03
-0.1
06/01/03
0.8
(1)
T
L
0.7
Hottest Day
0.6
Construction Day
0.5
Calculated
0.4
0.3
0.2
Inward
0.1
0.0
Bent 1 Center
04/26/04
12/28/03
11/28/03
10/29/03
09/29/03
08/30/03
07/31/03
07/01/03
06/01/03
-0.2
03/27/04
Coldest Day
Outward
02/26/04
-0.1
01/27/04
0.8
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.
37
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.
38
39
REFERENCES
1.
2.
3.
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.
40
41
42
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.
43
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
1
2
3
4
5
6
7
Description
Totally deteriorated or in failed condition
Used for conditions between 1 and 3
Serious deterioration or not functioning as originally designed
Used for conditions between 3 and 4
Minor deterioration and is functioning as originally designed
Used for conditions between 5 and 7
New condition
44
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.
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.
46
47
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.
2.
3.
4.
No curved girders.
5.
Steel H-piles or cast-in-place concrete piles are used; however cast-inplace 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
48
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 96035, 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, 3rd Edition". Structures Design and Construction Division, New York State
Department of Transportation, April 2002. (Revised January 2004)
49
ABSTRACT
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
50
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
30o 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 30mile 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.
51
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 35o skew angle.
Table 1. Geometric Guidelines
Connecticut
Span Length
Skew
Steel (ft) Concrete (ft) Angle (Degrees)
----20
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
200
330
300
No Limit
330
330
590
600
No Limit
590
30
30
--No Limit
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.
52
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.
PU
MY
MX
+
+
1 .0
0.85 AS FY i M UY M UX
(1)
53
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
Connecticut
Maine
Massachusetts
New Hampshire
Rhode Island
Vermont
13
U-back
10
--None
13
U-back
--U-back
----10
Earth Pressure
Full Passive
Full Passive
Modified pressure
coefficient:
K=0.43+5.7[1-e-190(dt/H)]
----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 Uwingwalls 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
54
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.
SeismicForce(kN )
1 2
H K (kN / m)
2
(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
55
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
56
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 16 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
57
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
58
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
CONCLUSION
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.
59
REFERENCES
1.
2.
60
INTRODUCTION
61
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
300 ft for 0 skew
450 ft
Steel Bridges
150 ft for 30 skew 30 max skew
500 ft for 0 skew
750 ft
30 max skew
750 ft
Concrete Bridges
250 ft for 30 skew 45 max skew
Total Movement at Abutment
1 in
2 in
45 max skew
2 in
62
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 Hpiles 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].
64
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 semiintegral 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
65
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
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.
66
Superstructure Rotation
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.
Figure 9. Original Approach Slab
The front edge of the approach slab seat is
Connection Detail
heavily chamfered to prevent spalling of
concrete.
67
68
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.
69
CONCLUSIONS
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.
2.
3.
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
70
SESSION II:
CASE STUDIES
71
72
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.
73
74
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
75
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
76
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. TWalls 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
77
78
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
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
79
Transverse joints
Longitudinal joints
Finished deck
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
80
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
Owner:
Contractor:
Designer:
Burke, Jr., M. P., Integral Bridge Design is on the Rise, Modern Steel Construction, JulyAugust, 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 02033, Structures Design and Construction Division, New York State Department of
Transportation, 2002.
81
82
83
84
INTRODUCTION
85
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.
Figure 3 Elevation
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)
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
87
88
89
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
Figure 6 Saw Cut Detail
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
Figure 7 Deck / Approach Panel Section at Abutment
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
90
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
91
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
Figure 9 Approach Panel
bridge wing walls. Our chosen approach
Longitudinal Joint Detail
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.
92
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.
93
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
94
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.
95
ACKNOWLEDGEMENTS
96
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, fivespan, 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 4foot 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.
97
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'-71/2"). 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.
98
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 semiintegral 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
99
Figure 2
Semi-Integral Abutment with Approach Slab
West Virginia Division of Highways
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
Lost River Bridge No. 1
US 220 Ramp Connector
Bridge
Edgewood Drive Bridge
Dumpling Run Bridge
Elizabeth Bridge
Sauerkraut Bridge
Length
666.1'
615'
Span Arrangement
173.9' 262.5' 229.7'
136'-171.5'-171.5'-136'
586'
543'
536'
524.9'
111.25'-111.25'-140'-111.25'-111.25'
167'-209'-167'
163'-210'-163'
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.
100
Alter
nate
Bridge
Type
Span
Arrangement
Girder
Spacing
Steel plate
girder
Steel plate
girder
165'-200'-165'
= 530'
165'-200'-165'
= 530'
9'-0"
9'-0"
101
Girder
Web
Size
72" x
1/2"
72" x
1/2"
No. of
Girders
Cost
(Millions)
Comment
$4.96
$4.99
Use 50-ksi
steel.
Use 50-ksi
steel in
positive
moment
regions and
70-ksi steel
in negative
moment
regions.
Alter
nate
Bridge
Type
Span
Arrangement
Post165'-200'-165'
tensioned = 570'
concrete Ibeam
Continuou
s
prestresse
d concrete
I-beam
and two
singlespan
prestresse
d concrete
I-beams
80'-150'-150'150'-80' =
610'
Girder
Spacing
10'-0"
9'-0"
Girder
No. of
Web
Girders
Size
102"8
deep
Type J
beam
Cost
(Millions)
Comment
$5.12
Five
concrete Igirder
would be
spliced
together by
posttensioning.
60"9
deep and
96"deep
Modifie
d Type
IV
Beams
$5.24
150'
interior
spans
would be
designed as
simple
spans for
dead load
and
continuous
spans for
live load.
102
Bridge Type
Span Lengths
Girder Spacing
Girder Size
No. of Girders
Bearings
Steel Type
Skew
Abutment Type
Pier Type
Alternate 1
Welded steel plate
girders
165'-1200'-165' = 530'
8'-0"
57" web with varying
flange thickness
10
High-load, multirotational bearings
AASHTO M270 grade
50W and Grade HPS
70W
38
Semi-integral
Four-column bent with
concrete cap
Alternate 2
Welded steel plate girder
196'-6"-230'-0"-196'-6" =
623'
8'-0"
66" web with varying
flange thickness
10 total (five each from
eastbound and westbound
structures)
High-load, multirotational bearings
AASHTO M270 Grade
50W and Grade HPS
70W
0 (Radial)
Integral
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.
103
104
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.
105
Figure 3
Alternate 1
106
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.
107
Figure 4
Alternate 2
108
Selection of Alternate
109
Figure 5
Coal River Integral Abutment Detail
110
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
111
was done to isolate the bridge movement and to protect the adjacent
roadway barrier and crash impact attenuator.
Figure 7
Approach Slab and Pavement Joint Details
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
112
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
113
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 SuperdeckTM. The deck crosssection 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%.
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].
114
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.
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
115
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.
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 MicroMeasurements 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
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
116
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.
117
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:
118
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 HS25 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 106 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 106 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
119
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].
120
121
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.
122
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 CycleControl 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.
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.
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.
123
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,46th
International SAMPE Symposium and Exhibition, Long Beach, California, 2001, pp
1321- 1330.
124
ABSTRACT
125
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:
126
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
Figure 1 Continuous for Live Load Detail
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.
127
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
128
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
Figure 3 Steel Continuous for Live Load
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
129
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
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
131
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
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
133
134
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.
135
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
136
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]:
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.
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.
Superstructure Type
and Pile Shape
Maximum
Integral Abutments
End Span
Concrete Beam
120 feet
HP 10x57
Steel Girder
HP 10x57
105 feet
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
138
All abutment piles are a minimum length of 2.5 times the prebore depth,
from bottom of footing to bearing end.
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.
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.
139
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)
640 FEET
CONCRETE
790 FEET
140
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
Span
Length
122-130
Skew
Bridge
Length
252
I-80/2nd Ave.
I-80/14th St.
244
117-127
4o-01-38
I-80/Two-Track
Concrete I-Beam
295
0o
I-80/Single Track
Concrete I-Beam
196
52-57-5777-52
47-87-62
13o-30-39
Ramp A/UPRR
Concrete I-Beam
185
52-67-66
7o-14-45
Ramp D/UPRR
Concrete I-Beam
185
52-67-66
18o-15-31
0o
141
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.
In 2001, two integral abutment bridges were designed for the I-235
construction project in Polk County, Iowa. The 9th street over I-235 bridge is a 3span, continuous steel plate girder bridge with span lengths of 83-104-122. I235 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:
4 EI
w
L pile
w =
142
WL2girder
24 EI
(1)
M = P / 2
The resulting extreme fiber bending stress
fb =
Mw + M
S
(3)
(2)
Cm f b
1.0
fa
(1 ' ) Fb
Fe
(4)
(5)
IOWA US 20 PROJECT
Bridge Name
Ramp C
US 20 WB
US 20 EB
Ramp D
Alignment
Bridge Length
Tangent
300
Tangent
271
Tangent & curved
345
Tangent & curved
290
Span Length
100-100-100
97-97-77
115-115-115
90-110-90
Skew
0o
0o
0o
0o
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.
143
T-REX PROJECT
144
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.
145
Structure
No.
17
Girder Type
18
Buried Slab on
P/S Box Girder
P/S I-Girder
27
P/S Box-Girder
34
Buried Slab on
P/S I-Girder
P/S
I-Girder,
Ballasted LRT
Bridge
Steel
Plate
Girder
P/S
I-Girder,
Direct Fixation
LRT Bridge
Buried Slab on
P/S I-Girder
P/S
I-Girder,
Ballasted LRT
Bridge
P/S I-Girder
38
48 & 49
72
73
74
77
83
P/S
I-Girder,
Direct Fixation
LRT Bridge
Foundation
Bridge
System
Length
W12x72 on 2
53
Dia. caisson
2 Dia.
36 to 45
Caisson
2 Dia.
39 to 42
Caisson Wall
W10x68 on 2 59 to 81
Dia. caisson
2 Dia.
218
Caisson
2-6 Dia.
Caisson
3 Dia.
Caisson
2 Dia.
Caisson
W14x89 on
2-6 Dia.
caisson
2 Dia.
Caisson
W14x109 on
3-6 Dia.
caisson
Span
Length
53
Skew
36 to 45
19o-4617
Varies
39 to 42
Varies
59 to 81
Varies
109-109
Varies
330
165-165
420
155-165100
42o-4800
0o
61 to 72 61 to 72
Varies
112
112
53o-3633
76
76
390
125-140125
45o-0450
0o
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.
146
REFERENCES
1.
2.
3.
4.
147
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
148
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 slabon-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
149
connected together with cast-in-place concrete over the girders and between the
slab elements (Figure 1& 2, System B).
150
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 castin 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.
151
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.
152
153
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 castin-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.
154
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
155
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 selfweight 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.
156
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.
157
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.
158
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.
159
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-9601, 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/CSAS6-00,
7. Computers and Structures Inc., SAP2000 Version 7.44
8. MIDASoft Inc. MIDAS/Civil Version 5.80
160
SESSION III:
MAINTENANCE AND
REHABILITATION
161
162
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
163
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.
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
164
be seen in Figure 3. Wing walls are perpendicular to the abutment and are
separated from it by an expansion joint.
165
166
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.
167
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.
2-D FEM
3-D FEM
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.
168
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.
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
169
40
N02
N03
N04
S02
S03
S04
20
Temperature Change (oF)
Temperature Change ( F)
2-D FEM
40
20
0
-20
-40
-60
0
-20
-40
-60
-80
-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
-80
-0.8
Displacement at CL of Girder(in)
-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)
a) North Abutment
b) South Abutment
N02
N03
N04
40
Temperature Change ( F)
20
0
a) North Abutment
-20
-40
-60
-80
-0.8
-0.7
-0.6
-0.4
-0.3
-0.2
-0.1
0
Displacement at JM Location (in)
2-D FEM
40
-0.5
S02
0.1
S03
0.2
0.3
0.4
S04
20
0
b) South Abutment
-20
-40
-60
-80
-0.8
-0.7
-0.6
-0.5
-0.4
-0.3
-0.2
-0.1
0.1
0.2
0.3
0.4
170
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.
5.38' Data
7.21' Data
8.13' Data
9.05' Data
5.38' Data
7.21' Data
8.13' Data
9.05' Data
5.38' Dense
7.21' Dense
8.13' Dense
9.05' Dense
5.38' Dense
7.21' Dense
8.13' Dense
9.05' Dense
60
40
40
20
20
Temperature Change ( F)
Temperature Change ( F)
60
0
-20
-40
-60
0
-20
-40
-60
-80
-80
-45
-40
-35
-30
-25
-20
-15
-10
Average Pressure in 2002 (psi)
-5
-45
-40
-35
-30
-25
-20
-15
-10
Average Pressure in 2004 (psi)
171
-5
9.05' Data
8.13' Data
7.21' Data
8.13' Data
5.38' Data
55
gage malfunction
45
35
25
15
5
-5
Jan-02
Mar-02
9.05' Data
May-02
Jun-02
Aug-02
Oct-02
45
Dec-02
8.13' Data
7.21' Data
25
15
5
5.38' Data
9.05' Data
65
55
55
45
35
25
15
5
May-02
Jun-02
Aug-02
Oct-02
Mar-04
Apr-04
Jun-04
Aug-04
Oct-04
Dec-04
65
Mar-02
gage malfunction
35
-5
Jan-04
5.38' Data
55
-5
Jan-02
7.21' Data
65
65
Dec-02
8.13' Data
7.21' Data
5.38' Data
gage malfunction
45
35
25
15
5
-5
Jan-04
Mar-04
Apr-04
Jun-04
Aug-04
Oct-04
Dec-04
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
172
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.
173
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
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.
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/ft3 (1.48 to 1.61 Mg/m3) 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.
175
110
140
30
45
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
H)
(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.
176
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 oC and -19.8oC (+15.5 oF to -35.6oF) 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
177
by pea stone properties when the bridge contracts and by backfill properties when
the bridge expands.
-35.6 Data
10
-35.6 Data
10
-27.6 Data
-27.6 Data
Distance From Bottom of Abutment (ft)
-17.7 Data
-9.1 Data
15.5 Data
-72 Up Bound
-36 Up Bound
-5
-10 Up Bound
36 Up Bound
15 Up Bound
-10
-T
Bridge Contraction
-15
-0.8
-0.6
+T
Bridge Expansion
-0.4
-0.2
0
Lateral Displacement (in)
0.2
-9.1 Data
15.5 Data
-72 Dense
-36 Dense
-5
-10 Dense
15 Dense
36 Dense
-72 Loose
-10
-T
Bridge Contraction
-15
-0.8
15 Low Bound
0.4
-17.7 Data
5
-36 Loose
-10 Loose
15 Loose
-0.6
-0.4
-0.2
0
Lateral Displacement (in)
36 Low Bound
+T
Bridge Expansion
0.2
0.4
36 Loose
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.
N03
N04
Upper Bound
Lower Bound
S02
60
40
40
20
Temperature Change ( F)
Temperature Change ( F)
N02
60
0
-20
-40
-60
-80
-0.25
S03
S04
Upper Bound
Lower Bound
20
0
-20
-40
-60
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
-80
-0.25
0.20
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
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
178
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.
Upper Bound - Dense Backfill
2.0
1.5
N02
N03
N04
1.0
0.5
0.0
-0.5
-1.0
Yield DisplacementLower Bound - Dense
Backfill
-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
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
179
+T
500
Calculated Yield Moment =
709 k-in
-500
+T
500
0
-500
-1000
-1000
-1500
-250
1000
Moment at Top of Pile (k-in.)
1000
1500
1500
-200
-150
-100
-50
50
100
150
200
-1500
-250
250
-200
-150
-100
-50
50
100
150
200
250
Temperature Change ( F)
Temperature Change ( F)
Figure 4. Pile Moment Variation with Temperature for Different Backfill and Pea Stone
Conditions
180
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.
181
5.38' Lower-Dense
5.38' Upper-Dense
5.38' Upper-Loose
7.21' Lower-Dense
7.21' Upper-Dense
7.21' Upper-Loose
9.05' Lower-Dense
9.05' Upper-Dense
9.05' Upper-Loose
60
40
20
0
-20
-40
-60
-80
-45
-40
-35
-30
-25
-20
-15
-10
Average Pressure in 2002 (psi)
-5
60
5.38' Data
7.21' Data
8.13' Data
9.05' Data
5.38' Dense
7.21' Dense
8.13' Dense
9.05' Dense
60
0
-20
9.05' Data
8.31' Dense
9.05' Dense
20
0
-40
-60
-60
-80
-80
-45
-40
-35
-30
-25
-20
-15
-10
Average Pressure in 2002 (psi)
-5
-45
60
-40
-35
-30
-25
-20
-15
-10
Average Pressure in 2004 (psi)
-5
5.38' Data
7.21' Data
8.13' Data
9.05' Data
5.38' Dense
7.21' Dense
8.13' Dense
9.05' Dense
60
40
40
20
20
8.13' Data
7.21' Dense
-20
-40
Temperature Change ( F)
7.21' Data
5.38' Dense
40
20
Temperature Change ( F)
40
5.38' Data
0
-20
-40
-60
5.38' Data
7.21' Data
8.13' Data
9.05' Data
5.38' Dense
7.21' Dense
8.13' Dense
9.05' Dense
0
-20
-40
-60
-80
-80
-45
-40
-35
-30
-25
-20
-15
-10
Average Pressure in 2002 (psi)
-5
-45
-40
-35
-30
-25
-20
-15
-10
Average Pressure in 2004 (psi)
-5
Figure 5. Effects of Soil and Pea Stone Properties on Earth Pressures Behind Abutments
182
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).
FEM w/out self wt.
Le fix-fix
1200
T(-)
900
Moment at Pile Top (kip-in)
FEM
600
300
0
Calculated yield moment
= 709 k-in
-300
-600
T(+)
-900
-1200
-4
-3
-2
-1
Figure 6. Pile Moment Versus Temperature Induced Displacements at Pile Top Evaluation
of Equivalent Cantilever Method
FEM w/out self wt.
FEM
Le fix-fix
1200
900
T(-)
600
300
0
Calculated yield moment
= 709 k-in
-300
-600
-900
T(+)
-1200
-4
-3
-2
-1
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
183
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.
184
185
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.
186
187
OBJECTIVE
188
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
189
M s = s E B AB e2' +
2
(1)
' t
e2 + = distance between mid-depth of slab and centroid of composite
2
section
(2)
M r = M p M d (1 e ) M s
2
2
where
Mp = moment caused by prestressing force about centroid of composite
member,
Md = 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.
190
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
(3)
191
1
1
(4)
E p A p
E s As
1 +
1 +
E
A
E
A
d
d
d d
where
Ep = modulus of elasticity of precast panels
Ap = area of precast panels
Es = modulus of elasticity of steel reinforcement in CIP deck
As = area of steel reinforcement in CIP deck
Ed = modulus of elasticity of CIP deck
Ad = area of CIP deck
The restraint moment at the center pier of a two span symmetric bridge is
calculated as
1 e 2
3
3
M r = M p (M d ) precast 1 e 1 (M d )CIP 1 e 2 M s
2
2
[(
)]
(5)
where
2I d
Ld
=
2 I d 3I m
+
Ld
Lm
Id = moment of inertia of diaphragm region (area between support points at
center pier)
Ld = length of diaphragm region
Im = moment of inertia of main spans
Lm = length of main spans
Mp = moment caused by prestressing force about the centroid of composite
member
Ms = differential shrinkage moment adjusted for restraint of precast panels and
steel reinforcement
(M d ) precast = mid-span moment due to dead load of precast panels
192
2 = creep coefficient for creep effects initiating when CIP topping is cast
c =
c (t 0 )
E c (t 0 )
[1 + (t , t 0 )]
(6)
193
shrinkage. Thus, the total strain in concrete due to applied stress and shrinkage is
given by
c (t ) = c (t 0 )
1 + (t , t 0 ) c (t ) 1 + (t , )
+
d c ( ) + cs (t , t 0 )
0
E c (t 0 )
E c ( )
(7)
where, cs(t,t0) is the free shrinkage strain occurring between ages t0 and t,
c (t ) = c (t 0 )
1 + (t , t 0 )
1 + (t , t 0 )
+ c (t )
+ cs (t , t 0 )
E c (t 0 )
E c (t 0 )
(8)
(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 Ec
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,
Ie
o
1
=
2
E e Ae I e Be Be
Be N
Ae M
(10)
where Ae , Be and I e 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 Ee = Ec / (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
194
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)
N
N
N
N N
+
+
+
(11)
=
M M creep M shrinkage M relaxation M temp. gradient
NUMERICAL ILLUSTRATION
195
Description
Girder
Concrete strength at transfer of prestress
(3days)
Concrete strength fc
Slab
Concrete strength fc
prestressed and non-prestressed steel
Modulus of elasticity
initial prestressing force
Diaphragms
size
spacing
bituminous overlay, thickness
Properties
3480 psi
8860 psi
3350 psi
29000 ksi
920 kips
1x 4
at 1/3 points of the
span
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
st
rd
supported condition at 61 day to a continuous system at 63 day.
196
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
197
ACKNOWLEDGEMENT
198
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 15o 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
200
AASHTO 1998. LRFD Bridge Design Specifications. 2od Edition, Washington, D.C.
AASHTO 1996. Standard Specifications for Highway Bridges, Washington, D.C.
Wassermann, E.P. 2001. Design of Integral Abutments for Jointless Bridges, Structure
Magazine, May 2001, pp. 24-33
Roman, E.K. 2004. Evaluation of Integral Abutments, Masters Thesis, Stevens Institute of
Technology, Hoboken, N.J.
Khodair, Y.A. 2004. Numerical and Experimental Analysis of Integral Abutment Bridge,
Ph.D. Dissertation, Stevens Institute of Technology, Hoboken, N.J.
201
202
203
204
205
206
207
208
209
Figure 10. Soil Pressure Along the Abutment Compared with Coulomb.
210
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,
Henderson, Las Vegas, NV 89074
3
PE, Senior Project Manager, PBS&J, 2270 Corporate Circle, Suite 100,
Henderson, Las Vegas, NV 89074
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-byside 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.
211
Since seismic forces were not significant for this project, seismic analysis
is not addressed in this paper.
INTRODUCTION
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.
212
Load Case
Component (Box girder
1220x686 mm) (DC1)
Barrier Rail (DW1)
Load (KN/m)
94.10
Description
Acting on Simple Girder
4.67
43.09
12.29
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
Analysis
Factored
1 Lane
Load
Presence
factor
1.20
446
535
2 Lanes
1.00
892
892
3 Lanes
0.85
1338
1137
Load Case
Reactions (KN)
Load Case
25% Truck
25%Tandem
5% (Truck + Lane)
5%(Tandem + Lane)
Load (KN)
160
111
50
40
213
(1)
THERMAL FORCES: The design mean temperature was 15oC, with rise
20oC and fall 25oC.
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
Wind
Pressure
Superstructure
Wind
Pressure
Substructure
Wind
Pressure
Vehicle
On
Load
4.4 KN/m
Description
Acting on Simple Girder
On
1.3 KN/m2
On
1.5 KN/m
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.
214
215
(a) Stage 1
(b) Stage 2
216
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,
pu = 1.5 DL = 228.14 KN / Girder / Side
(2)
217
(3)
(5)
where,
a
M n = As f y (d )
2
(6)
By using fc=28 MPa, fy=414 MPa, and 4#22 bars at bottom of abutment,
then Mn=2170 KN-m. So,
M r = 0.9 M n = 1953KN m > M u
(7)
inspection,
Mr >
reinforcing
wu L p
2
= 298 KN
(8)
(9)
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.
Pu = 228 KN
(9)
The distributed loads are coming from abutment, approach slab, stone
veneer and live loads:
wu = 147.73KN / m
(10)
(11)
(12)
(13)
(13)
219
DC
DW
EH
ES
Truck
Lane
IM_Truck
CE
BR
WS
WL
TU1(+)
TU2(-)
SUM(+)
SUM(-)
Service Loads
Shear Moment
(KN)
(KN-m)
836
0
534
178
0
335
0
61
142
278
85
98
47
92
0
0
31
556
0
0
0
0
0
235
0
-294
1675
1831
1370
-294
Strength I
Load Factor
Shear (KN)
Max Min Max
Min
1.25 0.9 1045
753
1.5 0.65 801
347
1.5
0.9
0
0
1.5 0.75
0
0
1.75
249
0
1.75
148
0
1.75
82
0
1.75
0
0
1.75
54
0
0
0
0
0
0
0
1.2
0.5
0
0
1.2
0.5
0
0
2380
1100
1846
1100
Moment (KN-m)
Max
Min
0
0
266
115
502
301
92
46
486
0
171
0
161
0
0
0
973
0
0
0
0
0
281
117
-353
-147
4154
579
-6
315
LL Case
1
2
3
Lane
1
2
3
Lane Factor
1.2
1
0.85
Final factor
1.2
2
2.55
Govern factor
2.55
(14)
M u total
= 346 KN m / m
La
(15)
By using #19 bars @ 200mm spacing, the abutment resistant moment capacity is
calculated as:
220
M R = M n = 360 KN m / m
(16)
DOWEL DESIGN
Vn = 182.5KN > Vu
(18)
CONCLUSIONS
221
222
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
223
1 4 ft
P u llin g S la b
(1 4 ft x 1 0 ft x 2 .5 ft)
C o n crete B a lla st B lo ck s
(6 ft x 2 .5 ft x 2 ft)
J a ck
L o a d C ell
L o a d C ell
3 ft
2 ft
1 ft
1 ft
R ea c tio n B ea m
P u ll P a d
A b u tm en t
3 8 ft
H P 10x42
224
14 ft
Concrete Ballast Blocks
(2 ft x 2.5 ft x 6 ft)
Abutment Slab
(14 ft x 10 ft x 2.5 ft)
HP12x53
W24x117
Load Cell
Jack
Load Cell
3 ft
2 ft
1 ft
36 ft
14 in Prestressed Pile
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
225
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.
226
70
13th Test
60
9th Test
10th Test
50
40
14th Test
30
15th Test
20
10
0
0.00
0.50
1.00
1.50
2.00
2.50
The behavior of steel H-pile integral abutments is controlled by the pileabutment 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.
227
70
60
3rd T est
50
1st T est
40
30
20
10
0
0
0.2
0.4
0.6
0.8
1.2
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
228
60
50
40
Cycle # 1
Cycle # 25
30
Cycle # 75
20
Cycle # 50
10
0
0.00
0.20
0.40
0.60
0.80
1.00
1.20
60
50
Cycle #4
40
Cycle # 50
30
20
Cycle # 75
10
0
0
0.2
0.4
0.6
0.8
229
1.2
1.4
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.
120
100
80
60
40
1 kip =4.45 kN
1 in =25.4 mm
20
0
0
10
230
12
120
100
80
60
40
20
0
0
DISCUSSION
231
ACKNOWLEDGMENT
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.
232
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
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
2
233
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 soilstructure interaction problem that requires a model for the complete bridge
including concrete super and substructures, piles as well as soil. In addition, the
234
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
235
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. FBMultiPier can perform static, dynamic or pushover analysis.
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.
236
MODEL IMPLEMENTATION
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.
237
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.
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
238
(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 (=30o) and dense sands (=40o). 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 = 80oF.
Figure 4: Site View of Abutment and Detail A Nonlinear Springs on Pile Nodes
239
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.
240
LATERAL DISPLACEMENT
ALONG THE ABUTMENT
(FB-MultiPier)
0
-5
DEPTH (ft)
-10
-15
-20
-25
-30
-35
-40
-0.46
-0.41
-0.36
-0.31
-0.26
-0.21
-0.16
-0.11
-0.06
-0.01
DISPLACEMENT (in)
LL
LD
DD
DL
DEPTH (ft)
-10
-15
-20
-25
-30
-35
-40
-0.46
-0.41
-0.36
-0.31
-0.26
-0.21
-0.16
-0.11
-0.06
-0.01
DISPLACEMENT (in)
LL
LD
DD
DL
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
241
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.
MOMENT ALONG THE PILE DEPTH
(FB-MultiPier)
POINTS AT THE PILE HEADS -FE
0
-5
LD(FE)
DEPTH (ft)
-10
LL(FE)
DD(FE)
DL(FE)
-15
-20
-25
-30
-35
-40
-95
-85
-75
-65
-55
-45
-35
-25
-15
-5
15
25
35
45
MOMENT (k.ft)
LL
LD
DD
DL
LL(FE)
LD(FE)
DL(FE)
DD(FE)
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,
242
Lowell for providing us the information required to model the study bridge and
being able to compare the results.
REFERENCES
1.
2.
3.
4.
Kunin, J., and Alampalli, S. 2000. Integral Abutment Bridges: Current Practice in United
States and Canada, J. of Performance of Constructed Facilities:104-111.
Faraji, S., Ting, J.M. and D.S. Crovo. 2001. Nonlinear Analysis of Integral Bridges: FiniteElement Model, J. of Geotechnical and Geoenvironmental Engineering: 454-461.
FB-MultiPier Users Guide, version 4: Bridge Software Institute. University of Florida,
Gainesville, Florida.
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.
243
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
244
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 highperformance 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 longterm performance with respect to integral abutments have continued.
ON-GRADE EXPANSION JOINT
ON-GRADE PAVEMENT
APPROACH SLAB
CONTINUOUS DECK
FINISH GRADE
INTEGRAL ABUTMENT
WINGWALL
FLEXIBLE PILE
GROUP
BEDROCK
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 ( f c' =10 ksi, or 69 MPa) 72-in. (1.83-m)
deep bulb-tee girders support the 8.25-in. (21-cm) thick reinforced concrete ( f c'
=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.
245
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
Abutment
(SOLID65)
(b)
Depth
(ft=0.3m)
50
300
P (kip=4.448 kN) on 1-ft spacing
(a)
250
KzAT
200
150
25
100
|
|
|
50
0
0
Winkler
Soil Springs
(COMBIN39)
(c)
0.2
0.4
0.6
0.8
Lateral Deflection (in.=2.54 cm)
0.1
0.2
0.3
Normal Deflection (in.=2.54 cm)
0.4
0.5
Pile
(BEAM24)
|
|
|
0.4
0.3
0.2
0.1
0
0
Figure 2. (a) Winkler Model of Horizontal Soil; Support to Substructure; (b) Lateral ForceTranslation 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].
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
247
(a)
SOLID DECK
ELEMENTS
RIGID LINK
ELEMENTS
REPRESENTED
BT-72 GIRDER
GIRDER
ELEMENTS
Figure 3. Finite Element Model (a) Full Model Indicating Variation in Reinforcement; (b)
Schematic Showing the Modeling Approach for the Prestressed Concrete Composite
Superstructure.
248
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
249
Location, in.
(a)
(b)
(c)
(d)
(e)
72
72
72
72
60
60
60
60
48
48
48
48
36
36
36
36
24
24
24
24
12
12
12
12
50
0
100 2
75
o
T(y), F
0
5 -100
T 10
-4
-75
-50
o
T(y), F
0
-5
T actual
.
o+ y
T F E
-4
-3
T 10
-2
-4
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.
yt
(3-a)
(3-b)
250
Figure 4(c) and (e) present the gradients derived for simulating the AASHTOproposed design gradients, respectively for increase and decrease.
RESULTS
251
-T,=0
+T,=0
-T,=27
+T,=27
-T,=55
+T,=55 Abutment 1
7
Abutment 2
Elevation (ft)
3.5
-3.5
-3.5
-7
-7
-10.5
-10.5
-14
-14
-17.5
-17.5
Elevation (ft)
-0.5
0
0.5
x (in.)
-21
-1
3.5
3.5
-3.5
-3.5
-7
-7
-10.5
-10.5
-14
-14
-17.5
-17.5
-21
-1
Abutment 2
-11
-4
-15
-8
-19
-12
-23
-16
3.5
-21
-1
Abutment 1
-0.5
0
0.5
z (in.)
-21
-1
-0.5
-0.5
0 0.5
x (in.)
0 0.5
z (in.)
-27
-20
-250 -150 -50 50 150 250 -150 -50 50 150 250 350
Mz (kip-ft)
Mz (kip-ft)
-11
-4
-15
-8
-19
-12
-23
-16
-27
-20
-350 -250 -150 -50 50 150
-150 -50 50 150 250
M (kip-ft)
M (kip-ft)
y
Figure 5. Average Abutment/Pile Deflection and Pile Moment Distribution in Both Abutments
Under Positive and Negative Thermal Gradients With Varying Bridge Skew
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
252
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.
253
254
SESSION IV:
CONSTRUCTION PRACTICES
255
256
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
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
257
Description of Bridge
The project involves a replacement bridge situated on Route 60 over the
Jackson River in Alleghany County, Virginia. The integral backwall (semiintegral) 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/m2 ($2.00/ft2) in 1997.
258
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.
259
RESULTS
260
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 oC (90 oF).
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 oC (50 to 80 oF), following a prolonged
period of cool weather. At the same time, pressures registered by sensors 2 and 3
(behind the EPS) were much lower.
261
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 wellgraded 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
263
264
(1)
265
266
1.
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.
4.
5.
6.
7.
This study was supported by the Federal Highway Administration. The author
expresses his gratitude to Art Wagner and Linda DeGrasse of the Virginia
267
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.
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. 248257.
5.
6.
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.
268
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.
269
270
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.
SUPERSTRUCTURE
EMBEDMENT
TYPICAL DEPTH
OF BACKFILL
ROADWAY
SURFACE
CONSTRUCTION
JOINT
TYPICAL
STRINGER
INTEGRAL
ABUTMENT
PILE EMBEDMENT
PRE-DRILLED
SHAFT
LOOSE STONE
DRIVEN
HP-PILE
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.
271
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
AXIAL LOAD P
SUPERSTRUCTURE
ROTATIONS
DIRECTIONAL
MOVEMENT
INTEGRAL
ABUTMENT
AND PILE
NEUTRAL
POINT
OF MAX
MOMENT
PILE HEAD
ROTATIONS
GENERAL LOCATION
OF NEXT LARGEST
MOMENT IN PILE
POINT OF 0
DEFLECTION
EQUIVALENT
FIXED/FIXED COLUMN
TRANSLATED
PILE HEAD
DEFLECTION
INCREASING
DEPTH
CURVATURE REVERSES AT
LOCATION OF PILE FIXITY
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 2nd 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.
272
LOCATION OF
HINGE IN PILE
PILE
POTENTIAL LOCATION
OF HINGE IN PILE
HINGE
LOCATION
INSET
SOIL SURROUNDING PILE MUST
RESIST GROSS BUCKLING
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
273
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:
Pile Head Shear Le ( P ) = 3
12 EI
P
where:
6 EI
M
(1)
(2)
Le(P) = equivalent length of column for lateral load at the pile head
Le(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
274
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 nonlinear 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
2nd 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:
b M n = b M p = ZFy
(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].
275
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.
SOIL PRESSURE NORMAL TO
ABUTMENT CAUSES LATERAL
COMPONENT OF MOVEMENT
SQUARE ABUTMENT
PILES ORIENTED FOR
WEAK AXIS ALONG
ABUTMENT CENTERLINE
SKEWED ABUTMENT
SUPERSTRUCTURE
EXPANSION
SUPERSTRUCTURE
EXPANSION
C mx f by
fa
C mx f bx
1 .0
+
+
Fa
(4)
fa
f
Fbx 1 a Fby
1
F'
F ' ex
ey
f by
fa
f
+ bx +
1 .0
0.472 Fy Fbx Fby
276
(5)
P
+
0.85 As Fcr
MC
P
M
+
1.0
0.85 As Fy M p
M u 1
A
F
s e
1.0
(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:
277
My
Pu
M
+
+ x 1.0
0.85 As Fy M uy M ux
(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]:
- p
- p
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
bf - t w
b = Flange projection =
2
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;
278
(9)
279
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 28December 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.
280
ABSTRACT
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:
281
expansion joints
primary direction of
thermally induced displacement
superstructure
bearings
abutments
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.
282
approach slabs
(optional)
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
283
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.
284
summer
position
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-
285
long-term position
of abutment
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-
286
287
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
288
geosynthetic
tensile
reinforcement
(a)
resilient-EPS geofoam
compressible inclusion
(b)
EPS-block
geofoam
lightweight fill
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.
289
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).
SUMMARY AND CONCLUSIONS
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
290
291
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
292
SITE CONDITIONS
Material
Unit Weight
(pcf)
Coarse
to
medium 145.7 155.2
SANDSTONE
Fine to shaley SANDSTONE
162.5 164.4
UCS
(psi)
4790 7240
Youngs
Modulus
(ksi)
3,800 4,200
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
293
E (ksi)
400
600
800
1000
E (ksi)
1200
100
0.0
1400
48.0
400
500
4.0
52.0
E AVG = 592 ksi
56.0
58.0
60.0
62.0
300
2.0
50.0
54.0
200
6.0
8.0
E AVG = 378 ksi
10.0
12.0
14.0
16.0
E AVG = 378 ksi
64.0
18.0
66.0
20.0
Figure 1. Elastic Moduli Profiles from Pressuremeter Testing at Pier 1 and Pier 2.
294
600
An interim p-y curve methodology uses the ultimate resistance of the rock
(pur) and the initial modulus of rock (Eir) 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 (Ecore) [8, 9, 10], and
between the modulus reduction ratio (Emass/Ecore) 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
295
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.
100 %
4.00E-04
90 %
Percent Greater Than or Equal To
3.50E-04
3.00E-04
2.50E-04
2.00E-04
1.50E-04
1.00E-04
Mean -/+ 1 Standard Deviation
5.00E-05
80 %
70 %
60 %
50 %
40 %
30 %
20 %
10 %
0.00E+00
3000
4000
5000
6000
7000
8000
9000
0%
3000
10000
4000
5000
6000
7000
8000
9000
10000
(a)
(b)
100 %
2.00E-07
00
0
00
0
00
0
16
00
14
00
00
0
00
0
12
00
10
00
80
0
00
0
60
0
00
0
40
0
00
0
20
0
0.00E+00
60 %
50 %
40 %
30 %
20 %
10 %
0%
90
00
00
10
00
00
0
4.00E-07
70 %
70
00
00
80
00
00
6.00E-07
80 %
50
00
00
60
00
00
8.00E-07
90 %
30
00
00
40
00
00
1.00E-06
0
10
00
00
20
00
00
1.20E-06
(c)
(d)
Figure 2. Values Obtained from the One Standard Deviation Method (i.e., (a) and (b)) and
the 90th Percentile Method (i.e., (c) and (d)).
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,
296
homogeneous and isotropic solid, the relationship of the elastic modulus over the
initial part of the p-y curve is given by [3, 4]:
Eir = ki Emi
(1)
where ki is a dimensionless constant derived from experiment and Emi 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 pur 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]:
1.4 x
pur = r (UCS )B1 +
for 0 x 3B
B
(2)
(3)
After computing the pur and Emi 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]:
p = E mi ( y ) for y <ya
p
p = ult
2
ym
(4)
(5)
p = pult
(6)
where ya 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 ym = km B where
km 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.
900
800
700
P (kips/in)
600
500
400
300
200
100
0
0.00
0.02
Defined by Eqn (4)
0.04
0.06
0.08
Y (in)
EL. +654.0 ft
EL. +652.0 ft
EL. +650.0 ft
EL. +646.0 ft
EL. +644.0 ft
EL. +642.0 ft
298
EL. +648.0 ft
0.10
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].
Deflection (in)
-0.2
0
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0
Pier 2 - Rock
3
Pier 2
20
Pier 1
30
40
6
9
12
Pier 1 - Soil
50
60
Pier 1 - Rock
Bottom of Socket - Pier 2
Bottom of Socket - Pier 1
10
15
18
Figure 5 presents the predicted moments for each pier structure for the
corresponding deflections noted in Figure 4.
299
Moment (kip-ft)
20
6000.0
8000.0
0
3
Pier 2
30
9
Pier 1
40
12
Pier 1 - Soil
50
Pier 1 - Rock
-2000.0
0.0
2000.0
4000.0
0
Pier 2 - Rock
Finished Grade Pier 1 and Pier 2
10
15
18
300
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
301
ABSTRACT
302
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
303
304
(1)
(3)
(4)
(5)
(6)
t avg =
i Ti Ai Ei
i =1
(7)
Ai Ei
i =1
Ad E d d + Ag E g g
Ad E d + Ag E g
305
(8)
T =
t avg
(9)
1.0Tmin shade + 9 F
(10)
Tmax eff
(11)
1.04Tmin shade + 3F
(12)
Tmax eff
(13)
Tmax eff
Tmin shade
Tmax shade
T solar
T1
306
Bridges experience length changes because of temperature changes and timedependent 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 slabto-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 bridgeend 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.
307
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.
3.
4.
308
a 98 percent confidence level that the movement will be less than that
calculated.
5.
309
increment data provide a reasonably accurate temperature range, on a sitespecific basis, for calculation of bridge expansion and contraction.
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.
2.
Max. Meas.
(in)
-26/+102
Bridge
Calculated (in)
w/Site T
w/NCHRP
276
w/Climatic
Atlas
w/ASHRAE
1.93
1.74
1.20
1.65
1.51
-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
-11/+93
2.91
2.79
2.38
2.96
2.92
1.93
2.53
11.17
2.91
1.74
2.43
13.30
2.94
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
1.10
1.20
NA
NA
1.85
2.05
1.35
1.45
Case No.
310
Bridge Contraction
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Emerson, M., Bridge Temperatures Estimated From the Shade Temperature, TRRL
Laboratory Report 696, Transport and Road Research Laboratory, Crowthorne,
Berkshire, England, 1976, 54 pp.
British Standards Institution, Steel, Concrete, and Composite Bridges, Part I, British
Standard NS 5400, Crowthorne, Berkshire, England, 1978.
Imbsen, R.A., et al., Thermal Effects in Concrete Bridge Superstructures, NCHRP Report
276, Transportation Research Board, Washington, DC, September 1985, 99pp.
AASHTO, AASHTO Guide Specifications for Thermal Effects in Concrete Bridge
Superstructures, Washington, DC, 1989.
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. 21172134.
U.S. Department of Commerce, Climatic Atlas of the United States, Ashville, NC, June
1968.
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.
Climatic Atlas of the United States, Harvard University Press, 1954.
ASHRAE, Fundamentals Handbook, New York, NY, 1993.
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.
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.
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.
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.
AASHTO, AASHTO LRFD Bridge Design Specifications, Third Edition, Washington,
DC 2004.
Oesterle, R.G., et al., Jointless and Integral Abutment Bridges, Analytical Research and
Proposed Design Procedures, Draft Final Report, FHWA, March 2002.
American Concrete Institute, Committee 209, Prediction of Creep, Shrinkage, and
Temperature Effects in Concrete Structures, ACI 209R-92, Detroit, MI, 1992, 47 pp.
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.
Bazant, Z.P., and Panula, L., Creep and Shrinkage Characterization for Analyzing
Prestressed Concrete Structures, PCI Journal, May/June 1980, pp. 86-121.
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. 149155.
Expansion Joints in Buildings, Technical Report No. 65, National Academy of Sciences,
1974.
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.
311
ABSTRACT
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
312
Figure 1 shows the passive soil pressure response, Pp, due to thermal
expansion and soil/abutment interface friction, Faf, assuming no rotation of the
superstructure. For rotational equilibrium:
Faf (L cos ) = Pp (L sin )
(1)
(2)
where tan = friction coefficient for interface of formed concrete and soil.
Substituting 2 into 1:
tan =
or
sin
= tan
cos
(3)
313
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, Fa, to wall/soil interface friction, Faf., in Fig. 1, rotational
equilibrium is found by:
(Fa + Faf) (L cos ) = Pp (L sin )
(4)
(5)
n=
cos
(6)
314
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-inplace 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.
315
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.
316
317
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.
SUMMARY AND CONCLUSIONS
319
Pp
Lc
os
Faf
Ls
in
Faf
Pp
W
Figure 1. Soil pressure load, Pp, and soil abutment interface friction, Faf.
1.368
1.2
tan - tan
for = 20
Fa
Pp
0.828
0.8
0.475
0.4
0.213
0
0
10
20
30
40
50
60
Skew Angle,
Figure 2. Relationship between force required for abutment lateral resistance, Fa, and
passive pressure response, Pp, to restrain lateral movement.
1.0
0.94
0.87
0.77
0.64
0.50
0.5
0.34
10
20
30
40
50
60
70
Skew Angle,
Figure 3. Relationship between end normal movement, ln, and end thermal expansion, l.
320
1.2
1.0
L/W = 3.15, S
w/
abut
= 0.19
L/W = 4.73, S
w/ S
abut
= 0.29
L/W = 9.45, S
w/
abut
= 0.58
0.8
0.6
t1
0.4
0.2
20
25
30
35
40
45
50
55
60
65
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.
L/W = 3.15, S
w/
abut
= 0.19
L/W = 4.73, S
w/
abut
= 0.29
L/W = 9.45, S
w/
abut
= 0.58
2500
2000
1500
3000
1000
20
25
30
35
40
45
50
55
60
500
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.
321
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
Chandra, V., et al., Draft Report on Precast Prestressed Concrete Integral Bridges,
State-of-the-Art, Precast/Prestressed Concrete Institute, 1995, 113 pp.
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.
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.
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.
Oesterle, R.G., et al., Jointless and Integral Abutment Bridges, Analytical Research
and Proposed Design Procedures, Draft Final Report, FHWA, March 2002.
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.
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.
Poulos, H.G., and Davis, E.H., Pile Foundation Analysis and Design, John Wiley and
Sons, 1980, 397 pp.
322
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.
323
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
324
325
Backfill Properties
600
500
400
300
200
100
0
0
Figure 4 Results of Haavistonjoki backfill large-scale cyclic tri-axial test. Resilient modulus
at various stress levels.
326
1500
1200
900
600
300
0
0
0,25
0,5
0,75
1,25
Axial strain, %
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/m3 and
friction angle = 42. Earth pressure cells F and G (label 2.2) at Haavistonjoki bridge pier
T3.
Shear strain in the backfill near the abutment can be approximated by the
equation (Bolton 1993, Koskinen 1997):
327
= 2 * y / H,
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 Gmax to G = (0,1 0,2) x Gmax 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.
328
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.
329
10.2.04
11.2.04
12.2.04
13.2.04
14.2.04
15.2.04
16.2.04
180
160
-2
H
I
J
-4
120
-6
K
L
M
100
-8
80
-10
60
-12
Temperature [C]
140
N
O
P
Q
V
40
-14
20
-16
Bridge
temperature
-18
-20
-20
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].
330
Date
5.2.04
9.2.04
13.2.04
17.2.04
21.2.04
25.2.04
29.2.04
2
-5
-10
-15
-1
-20
-2
-25
-3
-30
-4
-35
-5
F
Displacement at support T4
331
Displacement [mm]
1.2.04
0
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 Ecrushed_rock = 80 000 kPa
and Ecrushed_rock = 250 000 kPa. Other soil parameters used are presented in
Table1.
Table 1 Properties of soil.
Soil
material
Crushed
rock
Crushed
rock,
reduced
Clayey
Silt,
reduced
Moraine
[kN/m3]
sat
[kN/m3]
E50ref = Eoedref
or E
[kN/m2]
[]
[]
Rinter
22
23,1
80 000
0,3
45
0,5
20
22,0
29 700
0,3
42
0,19
17
18,1
9 300
0,3
33
0,19
23
23,0
80 000
0,3
45
0,7
332
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 E50ref = 80 000 kN/m2.
With crushed rock E50ref = 250 000 kN/m2 the increase was 170 kPa - 80 kPa = 90
kPa, which corresponds to measured values.
333
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. Ecrushed_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.
334
Figure 16. Horizontal displacement according to subgrade coefficient for large diameter
piles. pm = ultimate limit of horizontal resistance, ym = corresponding displacement
(Tiehallinto 1999)
SUMMARY
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.
3.
4.
335
5.
6.
7.
8.
Tiehallinto (Finnish Road Administration): Steel Pipe Piles. Code of practice. 1999. (in
Finnish)
9.
10. Tiehallinto (Finnish Road Administration): The supplement bridge design instructions.
2002. (in Finnish)
336
AUTHOR INDEX
Alampalli, S.
Aluri, S.
Arockiasamy, M.
Ashraf, S.
41
113
185
73
Basu, P.
Bergmann, M.
Bermudez, R.
Bonczar, C.
Bowser, J.
Brea, S.
Burdette, E.
244
73
199
163, 174
125
163, 174
222
Camp, J.
Chang, J.
Chovichien, V.
Christou, P.
Civjan, S.
Conboy, D.
Crellin, B.
Crovo, D.
125
211
30
233
163, 174
50
163
163, 174
Deatherage, J.
DeJong, J.
Deng, Y.
Dunker, K.
222
163, 174
211
136
Farre, J.
Frosch, R.
211
30
Jayakumaran, S. 73
Kerokoski, O.
323
Knickerbocker, D. 244
Kutschke, W.
292
GangaRao, H.
113
Goodpasture, D. 222
Grajales, B.
292
Hassiotis, S.
Hoit, M.
Hoppe, E.
Horvath, J.
Howard, S.
Huckabee, P.
Huh, B.
Husain, I.
199
233
257
281
222
270
148
148
Ingram, E.
222
337
Laaksonen, A.
Liu, D.
Lopez, J.
Lotfi, H.
Low, J.
323
136
199
312
148
Maberry, S.
Magliola, R.
Maruri, R.
McCormick, M.
McVay, M.
Michael, K.
Mistry, V.
125
136
12
148
233
97
3
Norrish, C.
73
Oesterle, R.
302, 312
Penafiel, P.
Perkun, J.
Peterson, B.
Petro, S.
211
97
84
12
Shekar, V.
Sivakumar, M.
Stoothoff, E.
113
185
50
Volz, J.
302
Wasserman, E.
Weakley, K.
Wenning, M.
Wetmore, J.
White, H.
244
61
30
84
41
Yannotti, A.
41