You are on page 1of 117





20 N. Wacker Drive
Glenview, Illinois 60025 Chicago, Illinois 60606
ACKNOWLEDGEMENT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
F O R E W O R D. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 TYPES OF PRECAST SEGMENTAL CONSTRUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.4 ALTERNATE DESIGN PROPOSALS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.5 APPLICABILITY OF PRECAST SEGMENTAL CONSTRUCTION . . . . . . . . . . . . . . . . . . . . . . 5
1.6.1 Lievre River Bridge, Quebec . . . . . . . . . . . . . . . . . . . . . . . b .................... 6
1.6.2 Bear River Bridge, Digby, Nova Scotia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6.3 JFK Memorial Causeway, Corpus Christi, Texas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6.4 Muscatatuck River Bridge, North Vernon, Indiana . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6.5 Vail Pass Bridges, Colorado. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6.6 Kishwaukee River Bridge, Winnebago County, Illinois . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6.7 Sugar Creek Bridge, Parke County, Indiana. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6.8 Turkey Run Bridge, Parke County, Indiana. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6.9 Pennsylvania State University Test Track Bridge. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6.10 Other Precast Segmental Bridges in Planning, Design and Construction. . . . . . . . . . . . 11
CHAPTER 2. CONSIDERATIONS FOR SEGMENT DESIGN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1 GENERAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 PRINCIPAL DIMENSIONS OF SEGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 DETAIL DIMENSIONS OF SEGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 PIER AND ABUTMENT SEGMENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.5 POST-TENSIONING TENDONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5.2 Permanent Post-Tensioning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5.3 Temporary Post-Tensioning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5.4 Layout of Post-Tensioning Tendons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6 MILD STEEL REINFORCEMENT CAGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.7 SHEAR KEYS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.8 EPOXYJOINTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25


3.1 GENERAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2 DEVELOPMENT OF PRELIMINARY BRIDGE DETAILS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2.1 Selection of Span Arrangement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2.2 Abutment Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.3 PierDetails ......................................................... 29
3.2.4 Horizontal and Vertical Curvature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2.5 BearingDetails.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3 LONGITUDINAL ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.1 Erection Moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.3.2 Creep Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Creep Effects Resulting from Change of Statical System. . . . . . . . . . . . . 31 The Effect of Creep on Moments due to Support Settlements . . . . . . . . 33 The Effect of Creep in Reducing Restraint Forces due to Shrinkage. . . . 34 Determination of the Creep Factor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Example Creep Factor Calculations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Influence of Creep on Super-structure Moments. . . . . . . . . . . . . . . . . . . 37
3.3.3 Analysis for Superimposed Dead Load and Live Load. . . . . . . . . . . . . . . . . . . . . . . . . 41
3.3.4 Analysis for the Effects of Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.3.5 Shear Lag . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Computer Analysis of Shear Lag in Single-Cell Box Girder Bridges. . . . . 44 Consideration of Shear Lag in Bridge Designs. . . . . . . . . . . . . . . . . . . . . 49
3.3.6 Ultimate Strength Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.4 TRANSVERSE ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.4.2 Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.4.3 Symmetrical Box Girder Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.4.4 Antisymmetrical Loading . . . . . . . . . . . . . . . . . . . . . . . . . ! .................... 53
3.4.5 Evaluation of the Contributions of Transverse Bending, Longitudinal Bending
and Torsion to Resistance of Antisymmetrical Loading . . . . . . . . . . . . . . . . . . . . . . . 54
3.5 ANALYSIS AND TRANSVERSE POST-TENSIONING OF DECK SLABS . . . . . . . . . . . . . . . . 59
3.5.1 Live Load Plus Impact Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.5.2 Transverse Post-Tensioning of Deck Slabs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.6 ANALYSIS AND CORRECTION OF DEFORMATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.6.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.6.2 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Phase A - Free Cantilever. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Intermediate Phases B, B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 . Phase C - Final Continuous System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.6.3 Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Correction of Deformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Correction of Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Correction of Superimposed Curvature. . . . . . . . . . . . . . . . . . . . . . . . . . 67 . $ Example Alignment Calculations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Notes on Alignment Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.7 COMPUTER PROGRAMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.7.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.7.2 Sources of Computer Programs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69


4.1 FABRICATIONXIF PRECAST SEGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.1 .l General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.1.2 Methods of Casting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 The Long-Line Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 The Short-Line Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.1.3 Formwork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.1.4 Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.1.5 JointSurfaces ....................................................... 75
4.1.6 BearingAreas ....................................................... 76
4.2 HANDLING AND TRANSPORTATION OF PRECAST SEGMENTS. . . . . . . . . . . . . . . . . . . . . 76
4.3 METHODS OF ERECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.3.1 Cranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.3.2 WinchandBeam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.3.3 LaunchingGantry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.3.4 Progressive Placing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.3.5 Erection Tolerances. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.3.6 Design of Piers and Stability During Construction. . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Single Slender Piers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Moment Resisting Piers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

CHAPTER 5. DESIGN EXAMPLE, NORTH VERNON BRIDGE, INDIANA. . . . . . . . . . . . . . . . . . . . . 85

5.1 GENERAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

5.2 STRUCTURE DIMENSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

5.3 ORDER OF ERECTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

5.4 POST-TENSIONING DETAILS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

5.5 DESIGN REQUIREMENTS AND LOADING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.6 DESIGN PROCEDURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
5.6.1 Step 1. Free Cantilever Plus Initial Cantilever Group 1 Post Tensioning . . . . . . . . . . . 89
5.6.2 Step 2. Completion of Tail Span Plus Continuity Group 2 Post-Tensioning . . . . . . . . 91
5.6.3 Step 3. Completion of Center Span. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.6.4 Step 4. Addition of Superimposed Dead Loads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.6.5 Step 5. Application of Live Load and Temperature Load . . . . . . . . . . . . . . . . . . . . . . 95
5.6.6 Step 6. Influence of Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Step 6a. Box Girder Dead Load Moment Redistribution Due to Creep. . 97 Step 6b. Post-Tensioning Moment Redistribution Due to Creep . . . . . . . 98 Step 6c. Effect of Prestress Losses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
5.6.7 Step 7. Final Stress Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
5.6.8 Step 8. Calculation of Transverse Moments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

, APPENDIX....................................................................... 107
SEGMENTAL BOX GIRDER BRIDGES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
AND CANADA WITH CROSS SECTIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
A.3 NOTATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
A.4 REFERENCES.............................................................. 118

Extreme care has been taken to have data and

information in the Precast Segmental Box Girder
Bridge Manual as accurate as possible. However,
as the Post-Tensioning Institute and Prestressed
Concrete Institute do not actually make designs
or prepare engineering plans, they cannot accept
responsibility for any errors or oversights in the
use of Manual material in bridge project designs
or in the preparation of engineering plans.

The majority of the technical material in this manual was developed under a contract
with the consulting firm of Bouvy, Van Der Vlugt & Van Der Niet/Segmental Technology
and Services (BVN/STS). H. H. Janssen prepared most of the BVN/STS material. The creep
and shrinkage data in Chapter 3 reflects the procedures in Comite Europeen du Beton/
Federation lnternationale de la Precontrainte bulletin dinformation No. 111 published in
October, 1975. Portions of Chapters 1, 2 .and 4 were adapted from the article by Jean
Muller of Enterprises Campenon Bernard, Ten Years Experience in Precast Segmental
Construction which was initially published in the January-February 1975 Journal of the
Prestressed Concrete Institute. Some of the material in Chapter 4 was taken from Recom-
mended Practice for Segmental Construction in Prestressed Concrete developed by the PCI
Committee on Segmental Construction and first published in the March-April 1975 Journal
of the Prestressed Concrete institute. The computer analysis for the effects of shear lag
presented in Chapter 3 was conducted by Professor Alex Scordelis of the University of
California at Berkeley. The prepublication drafts of the manual were reviewed by com-
mittees of the Prestressed Concrete Institute and the Post-Tensioning Institute. General
editorial work in development of the manual was by Clifford L. Freyermuth, Post-Tension-
ing Institute.


In the period since the conclusion of World War II, prestressed concrete in various forms
I has emerged as a major factor in long span bridge construction. A number of prestressed
concrete box girder bridges with spans ranging to 700 ft. (210 m) have either been
completed or are underway in the U.S. and Canada. A prestressed concrete deck has been
selected for the Dames Point Bridge in Jacksonville, Florida. This cable-stayed bridge will
have spans of 650 ft., 1300 ft., and 650 ft. (200-400-200 m) for a total length of 2600 ft.
(800 m). The Pasco-Kennewick cable-stayed bridge in the State of Washington utilizing pre-
cast segmental construction will be completed in 1978 and has a main span of 981 ft. (299
I ml.
In the late 1940s, and in the 1950s, many innovative construction methods were devel-
oped in Europe for replacement of war damaged bridges. These construction methods pri-
marily related to the use of prestressed concrete. In particular, the cast-in-place cantilever
method of segmental bridge construction developed by the firm of Dyckerhoff & Widmann
in Germany opened the way to construction of concrete bridge spans in excess of 700 ft.
(210 m).
Beginning in the mid 1960s, the Freyssinet Organization developed technology in France
for the use of precast segmental box girder bridges. This technology subsequently spread to
countries throughout the world, including, in recent years, Canada and the United States.
As a contribution to the continuing evolution of prestressed concrete bridge construction,
the Prestressed Concrete Institute and the Post-Tensioning Institute are pleased to present
this joint publication on precast segmental box girder bridges.

CHAPTER 1 Zealand adopted the method. Many other
DEVELOPMENT OF PRECAST countries are today using the precast segmental
SEGMENTAL BRIDGE CONSTRUCTION techniques for various applications. The first
known application of precast segmental box girder
1.1 Introduction bridge construction in North America was a high-
way bridge over the Lievre River in Quebec. The
The earliest known application of precast seg-
Lievre River Bridge was built in 1967 and has a
mental bridge construction was a single span
main span of 260 ft. (79 m) with end spans of 130
county bridge in New York State built in 1952.
ft. (40 m). The Bear River bridge near Digby, Nova
The bridge girders were divided longitudinally into
Scotia, shown in Fig. 1.4 contains six interior spans
three precast segments which were cast end to end.
of 265 ft. (81 m) and end spans of 203 ft. (62 m).
After curing, the segments were transported to the
The Bear River bridge was opened to traffic in
job site where they were reassembled and post-
December 1972.
tensioned with cold joints.
The first U.S. precast segmental box girder
The development of long span prestressed con-
bridge was built near Corpus Christi, Texas and was
crete bridge construction techniques in Europe is
opened to traffic in 1973. The Corpus Christi
outlined in the Foreword. Of particular signifi-
bridge, shown in Fig. 1.5, has a central span of 200
cance was the development of cast-in-place canti-
ft. (61 m) and end spans of 100 ft. (30.5 m). Sub-
lever segmental construction in Germany by the
sequent to the Corpus Christi bridge, precast
firm of Dyckerhoff & Widmann, Inc. The technolo-
segmental bridges have been completed in Indiana
gy of cast-in-place segmental construction was
and Colorado, and a bridge of this type is now
adapted and extended for use with precast seg-
under construction in Illinois. A simple span pre-
ments in the Choisy-le-Roi Bridge over the Seine
cast segmental bridge has been constructed at the
River south of Paris in 1962. The Choisy-le-Roi
Pennsylvania State University test track as a re-
Bridge, designed and built by Enterprises
search project sponsored by the Federal Highway
Campenon Bernard, is shown in Fig. 1.1. Several
Administration and the Pennsylvania Department
other structures of the same type were built in due
of Transportation. Numerous precast segmental
course. At the same time, the techniques of pre-
bridges have been designed for other locations in
casting segments and placing them in the structure
the U.S. and Canada, and it is expected that this
were continually refined.
technique will be widely used in the years ahead.
A major innovation for construction of precast
segmental bridges was the launching gantry which
was used for the first time on the Oleron Viaduct,
shown in Fig. 1.2, which was built between 1964 1.2 Types of Precast Segmental Construction
and 1966. The Oleron Viaduct launching gantri/ Two main types of precast segmental bridge
is shown in Fig. 1.3. The launching gantry makes it construction have developed which may be differ-
possible to move segments over the completed part entiated by the use of either cast-in-place concrete
of the structure and place them in cantilever over or epoxy joints.
successive piers. Use of a launching gantry per- A number of precast segmental bridges have
mitted completion of the Oleron Viaduct at an been built using cast-in-place joints 3 to 4 in. (76
average of 900 linear feet (270 m) of finished deck to 102 mm) wide between segments. This procedure
per month. While the launahing gantry is a very eliminates the need for match-casting and reduces
useful means of erection in many cases, erection the dimensional precision required in casting the
can also be accomplished by use of cranes and segments, but it has major disadvantages including
other means as described in Section 4.3. the requirement of falsework to support the seg-
Experience with major precast segmental bridges ments while the cast-in-place joint cures, and sub-
in Europe allowed the refinement of the construc- stantial reduction in construction speed. On
tion process. Improvements were made in precast- balance, the use of cast-in-place joints is not gen-
ing methods and in the design of erection equip- erally attractive and for this reason this type of
ment to permit use of larger segments and longer joint will not be considered further in this manual.
spans, and which could accommodate horizontal The prevailing system of precast segmental
curvature in the roadway alignment. bridge construction uses an epoxy resin jointing
The technique of precast segmental construc- material. The thickness of the epoxy joint is on the
tion not only gained rapid acceptance in France order of l/32 in. (0.8 mm). The use of an epoxy
but spread to other countries. For example, the joint requires a perfect fit between the ends ot
Netherlands, Switzerland and later Brazil and New adjacent segments. This is achieved by casting each


Fig. 1.1 - Choisy-le-Roi Bridge over the Seine River, Francet7)

Fig. 1.2 - Oleron Viaduct, Francef7)

segment against the end face of the preceding one The primary disadvantages of precast segmental
(matchcasting) and then erecting the segments construction relate to the need for a somewhat
in the same order in which they were cast. This higher level of technology in design, and the neces-
manual will consider only design and construction sity for a high degree of dimensional control during
techniques for bridges using match-cast segments manufacture and erection of the segments. At the
and an epoxy resin jointing material. moment, the temperature and other weather
restrictions of epoxy jointing materials is also a
limiting factor. The large number of successful
1.3 Advantages of Precast Segmental projects in Europe and other parts of the world,
Bridge Construction and the growing number of completed projects in
The advantages of the use of precast segmental North America suggest that these obstacles will not
construction techniques to the bridge engineer are inhibit rapid growth in the use of precast segmental
as follows: bridge construction.

1. The economy of precast prestressed concrete

construction is extended to a span range of 100 to 1.4 Alternate Design Proposals
400 ft. (30 to 120 m), and even longer spans may
Up to the present time, precast segmental bridge
be economical in circumstances where use of heavy
projects in North America have been primarily
erection equipment is feasible.
selected as the result of competitive bidding against
2. The precast segments may be fabricated while other superstructure types. Given the economic
the substructure is being built, and rapid erection conditions of the forseeable future, it is felt appro-
of the superstructure can be achieved. priate that alternate proposals for any type of
3. The method makes use of repetitive industria- superstructure should be permitted at either the
lized manufacturing techniques with theinherent owners or the contractors option on all major
potential for achieving high quality and high bridge projects. Such a procedure would enhance
strength concrete. competition, minimize construction costs, and
encourage the innovation necessary to assure
4. The need for falsework is eliminated and all
progress in the development of bridge construc-
erection may be accomplished from the top of the
tion techniques. To the fullest extent practicable,
completed portions of the bridge. These aspects
the contract documents should permit reasonable
may be particularly important for high-level cross-
flexibility in span arrangements and other details
ings, in cases where it is necessary to minimize
necessary to assure economical application of alter-
interference with the bridge environment, or where
native construction techniques. As one example of
heavy traffic must be maintained under the bridge
this point, the optimum ratios of end spans to
during construction.
intermediate span for three-span continuous rein-
5. The structure geometry may be adapted to any forced concrete or structural steel bridges are
horizontal or vertical curvature or any required usually not economical for segmental construc-
roadway superelevation. tion. For economy of a three-span precast segmen-
6. The effects of concrete shrinkage and creep may tal bridge erected in cantilever, the end spans
be substantially reduced both during erection and should be approximately 50 percent of the length
in the completed structure because the segments of interior spans. Of course, in long viaducts a
will normally have matured to full design strength portion of the end span can be built on falsework
before erection. without significantly affecting the overall structure
7. Except for temperature and weather limitations economy. However, generally it is not equitable to
related to mixing and placing epoxy, precast seg- select strucJIural parameters to maximize economy
mental construction is relatively insensitive to wea- of one type of construction, and then require that
ther conditions (see the weather restrictions on use any alternate design conform precisely to those
of epoxy in Appendix Section A.l). parameters (presuming some flexibility is per-
mitted by the factors controlling the structure
8. The esthetic potential of concrete construction. geometry).
9. Enhanced durability of bridge decks through It must also be recognized that use of alternate
precompression of the concrete and elimination of designs may entail some disadvantages. In particu-
cracking, and through use of high quality concrete lar, additional engineering costs may be involved.
produced under conditions that permit a high level Value engineering incentive clauses providing for
of quality control. alternative designs normally consider the additional
Fig. 1.3 - Oleron Gantryt7)

Fig. 1.4 - Bear River Bridge near Digby, Nova Scotia

engineering costs to both the contractor and the
owner in establishing the net cost reduction result-
ing from the alternative design proposal. For a
various types of segmental concrete superstructures,
these additional costs may be minimized by
advance recognition of the available construction
options in the contract documents. This may be
accomplished by using general rather than specific
details in the contract plans in such a way that the
specific details on the fabrication drawings or con-
struction plans for the options exercised by the
contractor can be checked against the contract
drawings. As examples of this procedure, contract
drawings may use Pe (force x eccentricity) diag-
grams or envelopes for the post-tensioning require-
ments rather than a specific number, size and loca-
tion of tendons; and envelopes which indicate max-
imum and minimum construction and service load Fig. 1.6 - Rhone - Alpes Motorway Overpasses, Switzer-
stresses along the structure.
Other factors contributing to selection of precast
segment4 construction are described in Section
1.5 Applicability of Precast Segmental 1.3.
Construction In recent years, the advantages of precast seg-
The USR of precast segmental bridge construc- mental construction have been extended to shorter
tion found initial acceptance for the span range of span freeway overpasses in several European
160 to 350 ft. (50 to 110 m). When thecantilever projects. The most notable application in this
method of &on is used, this span range is still category is the Rhone-Alpes Motorway which in-
considered to be the basic area of application. volved construction of 150 overpasses over a 5-year

Fig. 1.5 - Corpus Christi Bridge, Texas

anchored along the deck surface as illustrated in
Fig. 1.8. The total construction time for a single
overpass (foundations, piers, and superstructure)
using this technique is less than 2 weeks.
A procedure for precast segmental construction
developed primarily for the span range of 100 to
160 ft. (30 to 50 m) is the concept of progressive
placing discussed in Section 4.3.4. With this proce-
dure, segments are placed continuously from one
end of the deck to the other in successive canti-
levers on the same side of the various piers rather
than in balanced cantilever at each pier.

1.6 Applicatiohs of Precast Segmental

Construction in North America

1.6.1 Lievre River Bridge, Quebec

The Lievre River Bridge in Quebec, shown in
Fig. 1.9, was the first North American bridge of
precast segmental box girder construction. The
bridge, which was completed in 1967, utilizes a
two-cell box section and has spans of 130 ft. -
260 ft. - 130 ft. (40-79-40 m). The 92 ton
(84 t) pier segments of the Lievre River Bridge
were cast-in-place on the piers and the remainder
Fig. 1.7 - Rhone - Alpes Motorway, Overpasses, Switzer-
of the superstructure was match-cast using a cast-
ing bed set up on the river bank. Typical segments
of the bridge were 9 ft. 6 in. (2.9 m) long and
weighed from 38 to 52 tons (35 to 47 t). The
casting of segments extended from January
through June. During the winter months, the cast-
ing operation was protected by an enclosure of
plastic sheeting supported on reusable steel trusses.
The enclosure was assembled in sections 20 ft.
(6.1 m) long and was lifted by crane to the
required location as work advanced. Under normal
weather conditions, the erection pace for the
bridge was two segments per day. Erection began
in August and the bridge was completed the same
Fig. 1.8 - Rhone - Alpes Motorway Overpasses, Switzer-

period. The bridges are three-span structures with

main spans ranging from 60 to 100 ft. (18 to 30
m). The construction procedure for the Rhone-
Alpes bridges is shown in Fig. 1.6. Significant fea-
tures of these bridges includethe complete elimina-
tion of the normal closure joint, and the use of
conventional post-tensioning tendon profiles
instead of the cantilever type tendon arrangement.
Stability during construction is provided by tem-
porary supports close to the piers asshown in Fig.
1.7, and by temporary post-tensioning bars Fig. 1.9 - Lievre River Bridge, Quebec
1.6.2 Bear River Bridge, Digby, Nova Scotia erection. Eight cast-in-place closure segments 4 ft.,
A precast segmental superstructure was selected (1.2 m) long were used at the center of the spans
for the Bear River Bridge near Digby, Nova Scotia, to join the abutting precast cantilever sections into
when alternate bids found precast segmental a fully continuous structure. Casting of the super-
construction at $3.36 million, compared to the low structure units began in mid March and was com-
bid for a steel structure of about $3.60 million. pleted by the end of August. Erection started the
Another motivation for selection of the precast first of July and was completed at the end of Octo-
superstructure was the fact that Nova Scotia does ber, 1972. Grouting of tendons and placement of
not have steel fabricating facilities that would have curbs, sidewalks and guardrails required about 1%
accommodated the Bear River Bridge. This meant months following erection of the last segment.
that the money for superstructure labor and
materials would largely have been spent outside
the Province. On the other hand, selection of the
precast segmental superstructure resulted in use of
predominantly local labor sources and local
materials. The combination of direct cost savings
and use of local labor and materials led to the
selection of the precast segmental superstructure
even though there had only been one prior use of
this type of construction in Canada.
A construction view of the Bear River Bridge
is presented in Fig. 1 .lO. The bridge has six interior
spans at 265 ft. (80.8 m) each, and symmetrical
end spans of 203 ft. 9 in. (62.1 m) for a total
length of 1997 ft. 6 in. (608.8 m). The precast Fig. 1.10 - Bear River Bridge, Nova Scotia
sections are 37 ft. 6 in. (11.4 m) wide and 11 ft. 10
in. (3.6 m) deep. Most sections were 14 ft. 2 in.
1.6.3 J F K Memorial Causeway,
(4.3 m) long and weighed about 90 tons (82 t).
Corpus Christi, Texas
The top slab of the box is post-tensioned trans-
versely to achieve a thickness of 10 in. (254 mm) The JFK Memorial Causeway is shown shortly
at the centerline of the,section. after it was opened to traffic in the summer of
1973 in Fig. 1.5. The precast segmental box girder
The geometry of the bridge included a variety
portion of the bridge, the first of its kind in the
of circular, spiral, and parabolic curves as well as
United States, is shown in Fig. 1.11 as it appeared
tangent sections. In plan, the east end of the bridge
in late February, 1973. Erection of the 100 ft.
has two sharp horizontal curves connected to each
(30.5 m) end span and 100 ft. cantilever are com-
other and to the west end tangent by two spiral
plete on one side and about one third complete on
curves. In elevation, the bridge is on a 2044 ft.
the other side.
(623 m) vertical curve with tangents of 5.5 and 6.0
Precast segmental construction was selected for
percent. There is approximately 28 ft. to 30 ft.
the JFK Memorial Causeway following a compre-
(8.5 to 9.1 m) difference in elevation between the
hensive model test program at the University of
roadway surface at the abutments and at the center
Texas at Austin. Fig. 1.12 shows a general view of
of the bridge. Two sets of short line forms were
the model bridge during testing. Results and con-
used to cast the segments to meet the exacting
clusions from this test program indicated that this
geometry requirements. To attest to the accuracy
type of construction is safe and dependable. ) *
with which the segments matched the planned
Specific conclusions resulting from the tests are as
geometry, two to four segments were erected each
working day, and only nominal elevation adjust-
ments were required in the abutting cantilevers 1. The segmental bridge model safely carried
where the cast-in-place closures were completed at the ultimate design loads for all critical moment
the centers of the spans. and shear loading configurations on which its
The bridge required 145 precast segments. Two design had been based, as specified by the 1969
segments were constructed each working day, one Bureau of Public Roads Ultimate Strength Design
in each short line form. The segments were cast Criteria.
directly against the face of the matching segment *Numbers in parenthesis refer to references listed in Appendix
in the bridge which assured a perfect fit during Section A.4.
Fig. 1 .I 1 - Corpus Christi Bridge, Texas

Fig. 1.12 - Corpus Christi Bridge model test

2. The deflection under design live load in four 10. Bolts used for the temporary connection of
lanes (only three lanes required by live load the pier segments to the main piers yielded locally
reduction factors) was approximately L/3200 in under the most critical unbalanced loading,
the main span. This is much less than L/800 which although the calculated direct tensile stress was less
is generally considered as acceptable. than the actual yield strength. The bolts used in
3. Positive tendons in the main span were the model were also below the yield strength later
designed as for an ideal three-span continuous specified for the bolts in the prototype. Yielding
beam. Since the completed bridge was supported was apparently caused by the large gap between
on neoprene pads which have no vertical restraint the pier segments and the pier, with consequent
against uplift, the outer ends were able to rise off local bending, and was accentuated by the stress
their supports at loads greater than the design ulti- concentrations in the threads.
mate load, so that the structure did not act contin- 11. The theoretical calculations were generally in
uously at the ultimate conditions under main span good agreement with the experimental results al-
positive moment loading. Even so, there was suffi- though there were some appreciable deviations
cient reserve strength in the main span to carry between the experimental and theoretical values of
design ultimate load. strain in the top slab in some stages of cantilever
4. Under tests to failure with very high construction.
combined moment and shear loading, flexural
cracks appeared near the epoxy joints in the top
slab near the main pier. However, they joined the 1.6.4 Muscatatuck River Bridge,
diagonal tension cracks and did not extend along North Vernon, Indiana
the joints. There was no sign of any direct shear The second application of precast segmental
failure at the joints. In tests of the full bridge construction in the U.S. was the widening of a 45
model, approximately 75 percent of the theoretical year old open-spandrel arch bridge on U.S. 50 over
ultimate shear load was applied in the maximum the Vernon Fork of the Muscatatuck River in
shear loading test prior to failure of the bridge Jennings County, Indiana. The 22 ft. (6.7 m)
during that test by flexure. No sign of shear dis- wide precast segmental box sections were erected
tress was evident. Subsequent tests of a three- just 1 ft. (0.3 m) away from the deck of the
segment model under severe shear loading as a existing arch bridge. The two decks are joined
cantilever section indicated that full shear strength by a longitudinal neoprene joint to provide a new
of the unit was developed. Hence, the epoxy joint 44 ft. (13.4 m) wide driving surface. A view of the
technique used did not reduce the design shear Muscatatuck River Bridge is presented in Fig. 1.13.
strength. Complete design calculations for the Muscatatuck
5. During erection of the first few segments, River Bridge are presented in Chapter 5.
tensile stresses occurred in the bottom slab as pre-
dicted in the design. These stresses resulted from
1.6.5 Vail Pass Bridges, Colorado
the large amount of prestress in the top slab at this
stage of erection. Temporary prestress devices Construction was completed on a series of four
successfully controlled the effects of these stresses. precast segmental bridges on Interstate 70 west of
6. Theoretical calculation of the load factor for Denver over Vail Pass in 1976. The lengths of the
live and impact loads required to form the first bridges ranged from 390 to 830 ft. (119 to 253 m),
plastic hinge agreed very well tiith the experimental and the main span lengths were either 200 ft.
results. These tests proved the accuracy and appli- (61 m) or 210 ft (66 m). A single-cell box girder
cability of the ultimate load calculation procedure. section was used for the 42 ft. (12.8 m) wide seg-
7. Near failure, major cracks concentrated near ments. A construction view of one of the Vail Pass
the epoxy joints which had no continuous conven- Bridges is presented in Fig. 1.14. Alignment prob-
tional reinforcement. However, throughout the lems were encountered at the closure of the first
loading sequence, cracks were generally well dis- Vail Pass Bridge which required removal of a
tributed because of the effective grouting and the portion of the precast parapet, and use of an
strength of the epoxy joints. asphaltic wearing surface of varying thickness
8. Transverse moment capacity of the bridge (maximum thickness of the asphaltic surfacing was
cross section was very adequate, as shown by the about 11 in. (279 mm) at one point along one
punching shear load test results. gutter line). The cause of the misalignment has not
9. There was no adverse effect of the epoxy been specifically determined at this time.
joints on the slab punching shear strengths. Three later segments of l-70 over Vail Pass in-
eluded alternates for bridges with structural steel, recreational purposes, the State of Illinois assumed
precast segmental or cast-in-place segmental super- special obligations for preservation of adjacent
structures. In two of these cases, structural steel landscape. A number of types of bridges were
bridges were low, and in the third case, the low bid evaluated by the State of Illinois Department of
was for cast-in-place segmental construction. Transportation. These included an orthotropic
steel box girder, a tied arch, and a segmental con-
crete box girder. After an elaborate cost study, it
was determined that segmental concrete construc-
tion offered not only an economical solution,
but also it would most nearly fulfill environmental
and esthetic considerations.

Fig. 1.13 - Muscatatuck River Bridge, North Vernon,


Fig. 1.15 - Kishwaukee River Bridge, Illinois

1.6.7 Sugar Creek Bridge,

Parke County, Indiana
The Sugar Creek Bridge in Parke County,
Indiana, was completed in 1977. This three-span
bridge has end spans of 90 ft. (27 m), a central
span of 180 ft. (55 m), and utilizes a single-cell
box section 30 ft. (9.1 m) wide.
Fig. 1.14 - Vail Pass Bridge, Colorado

1.6.8 Turkey Run Bridge,

1.6.6 Kishwaukee River Bridge, Parke County, Indiana
Winnebago County, Illinois Also completed in 1977, the Turkey Run Bridge
A model of the Kishwaukee River Bridge is in Parke County, Indiana, has two 180 ft. (55 m)
shown in Fig. 1.15. The twin structures have end spans, and utilizes two parallel boxes, each 22 ft.
spans of 170 ft. (52 m) and three interior spans of (6.7 m) wide. This bridge was constructed with the
250 ft. (76 m) each. The total length of each struc- aid of temporary erection bents to reduce the
ture will be about 1,170 ft. (360 m). The bridge required depth of the box section.
will span a wooded river gorge about 100 ft. (30
m) above the normal water level in the river. Bids
were received by the Illinois Department of Trans- 1.6.9 Pennsylvania State University Test Track
portation on the Kishwaukee River Bridge on Bridge
September 2, 1976. The construction special pro- A curved [radius of curvature 546 ft. (166 m)]
visions permitted submission of alternate design precast segmental box girder bridge with a single
proposals for the crossing, but the low bid was span of 121 ft. (37 m) is now under test at the
submitted for precast segmental construction. Pennsylvania State University Pavement Durability
The bridge is in an environmentally sensitive Track. The segments of this bridge were assembled
area. Since the Kishwaukee River is used for on the ground adjacent to the bridge site, and the
entire superstructure was erected in one piece by 1.6.10 Other Precast Segmental Bridges in
cranes. Among the objectives of this research Planning, Design and Construction
project, sponsored by the Pennsylvania Depart- The structures completed or under contract
ment of Transportation and the Federal Highway listed above represent only the beginning of the
Administration, was the evaluation of details for applications of precast segmental bridge construc-
precast segmental bridges, and investigation of the tion in North America. Additional structures
applicability of precast segmental construction for known to be in various stages of planning, design
use as site-assembled grade separation bridges. Such and construction are listed in Table 1.1. Segment
bridges might be used in situations where transport cross sections for precast segmental bridges com-
length or haul weight restrictions do not permit the pleted to date, and for many of the bridges listed
use of precast l-girders. in Table 1.1 are presented in Appendix Section


T .ength, ft. T
Name Location Total Individual spans Box Girder Width

Fredericton Bridge Fredericton, 2540 394 max. 3 boxes

New Brunswick each box 27 wide

Kishwaukee River Highway 412 1090 170, 2 boxes

Winnebago County, 3@250, each box 42 wide
Illinois 170 2 - 2-lane roadways

Illinois River Highway 408 3300 390,550,390 2 boxes

between Pike & Scott 9@ 230 each box 42 wide
Counties, Illinois 2 - 2-lane roadways

lslington Ave. Extension Toronto, Ontario 600 2 boxes

each box 46 wide

St. Joseph River Benton Harbor, Michigan 408 98. 212,98 1 box 48-6% wide

l-205 Columbia River Bridge Oregon - Washington 10,000 620 max. multiple boxes
143total width

Pike County Kentucky 372 93.5, 185, 93.5 1 box 28 wide

Cline Avenue East Chicago, 6000 300 max. multiple boxes

Lake County, Indiana 110 total width
Kentucky River Frankfort, Kentucky 780 323 82 total width

Long Key Key West, Florida 12,144 118 1 box 40 wide

1 ft. = 0.3048 m


2.1 General
Much of the economy of precast segmental
bridges results from the standardization and indus-
trialization of the process of manufacturing the
segments. When design details permit repetition of
daily actions, one segment per day can be manufac-
tured from each form by a comparatively small
crew. To achieve this rate of production, it is im-
portant to avoid changes in the forms, to standard-

1 I1 I
Fig. 2.1 - Segment dimensions
ize the cage of mild steel reinforcement, and to use
a repetitive layout of the post-tensioning tendons.
. It is always necessary to thicken the bottom slab of w w

io-i -.o-
the segments near the pier. However, even this
minor variation in the details of the segments may
disturb somewhat the normal schedule of segment

2.2 Principal Dimensions of Segments Fig. 2.2 - Superstructure with parallel segments and
cast-in-place joint
The principal segment dimensions are top slab
width W, construction depth D, width of
bottom slab B, web spacing Is, and segment only 25 percent of the deflection permitted in
length L. These dimensions are shown for a typ- steel structures in the U.S. Span/depth ratios for
ical segment in Fig. 2.1. end spans are usually somewhat lower than for in-
In the most simple case, the segment width W terior spans. The shallower depth structures require
is selected as equal to the width of the bridge. more high strength post-tensioning materials. Var-
When the bridge width exceeds about 40 ft. (12 iable depth structures become appropriate for
m), or when it is necessary to minimize segment spans in excess of 250 to 300 ft. (75 to 90 m). In
weight or size, the structure width can be divided this case, the span/depth ratios have normally been
into a multiple of the segment width as shown in selected as 18 to 20 at support and 40 to 50 at
Fig. 2.2. In this case, the transverse connection of midspan.
the top slabs may be accomplished by transverse When webs are vertical, the bottom slab width
post-tensioning which extends through all the B follows from the width W and the struc-
boxes and the cast-in-place joint(s). turally acceptable length of the cantilever as dis-
As an alternative to use of multiple boxes for cussed below. Sloping webs present no problem
structures wider than about 40 ft. (12 m), single when the box girder depth is constant, but do
boxes with multiple webs have been used for require significant form adjustments for produc-
widths up to about 70 ft. (21, m). For intermediate tion of variable depth segments due to the varia-
widths, single box sections may be used with inte- tion in bottom slab width. A narrow bottom slab
gral transverse floor beams under the roadway is desirable to reduce segment weight since the
slab (e.g., St-Andre de Cubzac Viaducts) or boxed bottom slab area is usually a factor for structural
cantilevers (e.g., Chillon Viaduct). These alterna- consideration only in the negative moment area
tives are illustrated in Fig. 2.3, which in addition, adjacent to piers.
shows the evolution of segment size and weight The segment length L has a pronounced effect
for a number of European bridges. on the economy of a bridge. The selection of the
The construction depth D is determined by segment length determines the total number of
the spans. Most European bridges have been built segments that must be produced and erected. Since
with span/depth ratios of 18 to 20. However, ratios the majority of the cost involved in production and
of 20 to 30 are considered feasible and structurally erection is fixed per unit and only a small share of
satisfactory. Deflection tests on the model of the the cost is variable, economy is achieved by using
Corpus Christi Bridge with a span/depth ratio of the smallest number of segments consistent with
25 resulted in a deflection of only L/3200 which is transportation requirements and the capacity of

8.20 FT.

SEUDRE 3.30M 75
79 M 10.80 FT.
259 FT.

BLOIS 3.50M 75
91M 11.50 FT.
299 FT.

CHILLON 3.20M 80
104M 10.50 FT.
341 FT.


312 FT.

B3 SOUTH 2.50M-3.40M 50
50M 8.20 FT.-l 1.20 FT.
164 FT.

SAINT-CLOUD 41e 0 2.25M 130
106M n 7.40 FT.
348 FT. 1
+ 1 t

Fig. 2.3 - Segment details of various European bridgest7)

erection equipment. Since the cost of handling and 3. Local bending stresses due to wheel loads
erection increases with L, it is necessary to make applied directly over epoxy joints.
a study of the total in-place economy of various 4. Local anchorage bearing and splitting stresses
segment lengths to determine the most economical
for transverse post-tensioning (when used)
value. When segments must be transported over
require a minimum thickness of about 8% in.
highways, the weight and size limitations usually
(216 mm) for tendon forces ranging from 100
determine the value of L. to 120 kips (445 to 534 kN).
The spacing of webs s is normally determined
purely on structural criteria. In principle, any web In addition to the above structural considera-
spacing can be utilized if all pertinent structural tions, the top slab thickness must be adequate to
aspects are thoroughly investigated using, if necess- accommodate four layers of transverse and longitu-
more sophisticated structural dinal mild steel reinforcement, transverse and
ary, analysis
techniques. The need for such analysis is greatly longitudinal tendons, and minimum concrete cover
reduced when the web spacing is selected in such a of 2 in. (51 mm) on top and 1 in. (25 mm) on the
way that the ordinary beam theory can be applied bottom.
for longitudinal moments. The beam theory may The dimensions of haunches b, c and d
be used when the depth of the section is equal to in Fig. 2.1 are determined by the transverse bend-
or greater than l/30 of the span, and when the ing moments and by the space required for the
width W divided by the number of webs is not anchorages of the longitudinal post-tensioning
more than 7% percent of the span length. For sec- tendons (see Figs. 2.10 and 2.12). It is normally
tions such as shown in Figs. 2.1 and 2.2 the slab necessary to accommodate at least two layers of
cantilever C is about one-fourth WI. For box longitudinal tendons. A concrete depth of 14 in.
sections with more than two webs the slab canti- (356 mm) is required at anchorages of longitudi-
lever dimension should be selected to provide rea- nal strand tendons. A depth of 10 in. (254 mm)
sonable balance between cantilever and interior may suffice for bar tendons. Although it is
transverse moments. Use of these criteria for deter- essential to provide adequate space in the top slab
mining the number and spacing of webs also results and haunch thicknesses for the above considera-
in reasonable requirements for the depth of the top tions, it should also be kept in mind that the top
slab and the amount of transverse top slab rein- slab is the heaviest part of the box girder, and from
forcement. this standpoint it is desirable to keep those dimen-
Segment dimensions used on U.S. and Canadian sions as small as practical.
precast segmental bridges now completed or in The web thickness e is generally 14 in. (356
advanced stages of design are presented in Appen- mm) or more to provide room for the anchorage
dix Section A.2. hardware of 12-strand tendons which are a fre-
quently used tendon size. Minimum anchorage space
requirements for bar tendons is about 10 in. (254
mm). The 14 in. (356 mm) width may also be de-
sirable or necessary to accommodate the bursting
2.3 Detail Dimensions of Segments and splitting force from anchorages for 12.strand
The concrete dimensions of top slab, webs, bot- tendons. This thickness may be reduced when ten-
tom slab and haunches are determined by structur- dons are anchored in ribs or anchor blocks. Thick-
al considerations and by numerous practical factors nesses as small as 8 in. (203 mm) have been used
related to production of the segments. with strand tendons when webs were vertically pre-
The top slab thickness (a in Fig. 2.1) usually stressed. When shear forces near supports are re-
ranges from 7 to 10 in. (175 to 250 mm). It is duced by upward shear from the post-tensioning
necessary to consider the following structural tendons and segment depth is within the limits
described in Section 2.2., the shear stress require-
factors in selecting the top slab thickness:
ments for highway bridges are generally met when
1. Bending moments in the transverse direction the total width of webs amounts to 7 or 8 percent
caused by slab dead load, permanent loads of the bridge width. The principal tensile stresses
and live load. resulting from combination of vertical shear stresses
2. Compression zone requirements for longitu- and compressive stresses reach a maximum value at
dinal bending moments normally need be con- the intersection of the top slab and the web. Efforts
sidered in determining top slab thickness only should be made to keep these principal stresses
in structures with spans of 350 ft. (110 m) within allowable limits [see AASHTO Bridge
or more. Specifications,@ @@ion 1.6.6. (B)] , and to avoid
the use of additional reinforcement for this limits. The bottom slab thickening for this purpose
purpose. This requires the widening of the webs should be reduced to the minimum thickness re-
f as shown in Fig. 2.1. quired in the shortest distance possible to facili-
The web is a stiff element in the box section and tate manufacturing of the segments.
provides substantial moment restraint to the top The dimensions of the bottom slab haunches
slab, and consequently transverse moments at the (h and i in Fig. 2.1) have a major structural
junction of the web and top slab are high. In- task in the longitudinal negative moment area of
creased concrete thickness, obtained by widening transferring the change of force in the bottom slab
to the web f as shown in Fig. 2.1, reduces the to the webs. This function is illustrated in Fig. 2.4.
amount of reinforcement required. Particular The force differential AF is transferred by longitu-
attention should be given to lapping of reinforce- dinal shear, and is the highest in the negative
ment in this area to avoid discontinuity in areas of moment area. The bottom slab haunches also assist
high moment. in transmitting transverse bending moments
A different situation exists in positive and nega- between the bottom slab and the webs, and reduce
tive moment areas relative to the required bottom the amount of reinforcement required for this
slab thickness g. The structural significance of purpose.
the bottom slab in the positive moment area re-
lates only to the bottom slab contribution to the 2.4 Pier and Abutment Segments
section properties. As a result, the bottom slab Pier and abutment superstructure segments
thickness is usually reduced in positive moment differ from typical interior superstructure segments
areas to the minimum required to carry the slab in that they normally require a diaphragm to assist
dead load, and the space required for reinforce- the webs in distributing the high shear forces to
ment and concrete cover. Space for one layer of the bearings. As illustrated in Fig. 2.5, vertical and
tendons, mild steel reinforcement, and concrete transverse post-tensioning can be used to transfer
cover require a minimum bottom slab thickness of
the shear from the webs through the diaphragm to
about 7 in. (178 mm). In the negative moment
the bearings. The amount of post-tensioning
area, the bottom slab thickness is controlled by
utilized for this purpose is a function of the shear
high compressive stresses. Thickening of the bot-
forces in the webs. In addition to the post-tension-
tom slab near piers is nearly always required to
ing tendons, the pier and abutment segment dia-
keep the compressive stresses within the allowable
phragms are normally heavily reinforced with non-
prestressed reinforcement. The tendons extending
across the diaphragm in Fig. 2.5 must be tied into
F = total compressive the diaphragm with bonded reinforcement to resist
force in half of
bottom slab of
tendon splitting stresses at the corners of the open-
single box girder at ings. Precise analysis of diaphragm stresses requires
Section 1.
use of finite element or other similar analytical
techniques. However, an approximate analysis
based on force resolution is usually sufficient. As
shown in Fig. 2.5, it is essential that an opening be
maintained in both pier and abutment segment
F + AF = corresponding
compressive force at diaphragms sufficiently large to permit movement
Section 2. of men and equipment.



PLAN AF = shear force at the

connection of web
and bottom slab.



Fig 2.4 - Longitudinal shear transfer by bottom slab to
web haunches Fig. 2.5 - Pier and abutment segments

dons with little interference with the erection
process. When tendons are anchored at the face of
a segment, a scaffold is normally used as shown in
Fig. 2.6 to facilitate installation and stressing of
tendons. With interior ribs, or web stiffeners,
these operations are accomplished from inside the
box. However, segments with interior ribs are
more difficult to manufacture, and selection of
segment details in a particular case requires con-
sideration of all aspects of manufacture, erec-
tion, and installation, stressing and grouting of
the tendons.
Continuity tendons are normally placed and
Fig. 2.6 - Use of scaffold for stressing of tendons stressed after the erection process and after the
closing of the castlin-place joints. Details for an-
erection. The anchorages for permanent longi- chorage of continuity tendons in the top slab over
tudinal tendons to be stressed during erection may the webs are presented in Fig. 2.8. This anchorage
be located either in the webs at the face of the detail has the disadvantage of allowing dirt, water
segment, or in special web stiffeners cast into the and extraneous material to enter the tendon ducts.
segment for the purpose of providing a location This may cause blockages and other problems.
for anchorage of permanent and temporary ten- Details for anchorage of continuity tendons in
dons that does not interfere with the erection the bottom slab are shown in Fig. 2.9. Continuity
process. Fig. 2.6 shows stressing of tendons with tendons may also be anchored in web stiffeners as
anchorages located in the web faces. Fig. 2.7 illustrated in Fig. 2.7. The stressing pockets for
shows details of a segment with an interior stif- anchors in the top slab should be kept as small
fening rib which provides a location for installa- as possible to minimize conflicts with mild steel
tion, stressing and anchorage of longitudinal ten- reinforcement or transverse post-tensioning ten-


(I) J O I N T
(2) W E B K E Y

Fig. 2.7 - Details of segment with web stiffener() HORIZONTAL SECTION C-C
- A N C H O R A G E

-Spiral roinforcemont when required

..~ . .
by porr-r.n.,onrnB.

NOTE: Block-out dimensions

reinforcement dotoils
with the port-tanrioni
system used.


Fig. 2.8 - Top slab anchorage block-out

NOTE: Specific reinforcement details and dimensions

of concrete build-out vary with different
post-tensioning systems

a 4 - ##6 Hairpins I
Length 2-6 (0.8 m) - Typical for 12.strand tendon. Reinforcement requirement for other tendon sizes will vary.

#4 = 13 mm dia.
SECTION #6 = 19 mm dia.
4'.4" (VARIES)
(1.3 m)




- 2 - #4 Hairpins @ 5 in. (127 mm) ctrs.

PLAN (approx. HS show)

Fig. 2.9 - Bottom slab anchorage build-out

dons. For larger strand tendons, used for longitudi- feasible provided adequate measures are taken to
nal post-tensioning, mild steel reinforcement is overcome non-alignment of ducts at the joint
normally required to assist in distribution of the caused by casting tolerances. Transverse tendons
prestressing force into the segment. Anchorage and may be installed in flat bundles of three or four
tendon coupler blockout details to be used with strands to maximize the tendon eccentricity.
bar tendons on the Kishwaukee River Bridge in In segments at and adjacent to piers, there are a
Illinois are shown in Fig. 2.10. large number of longitudinal and transverse ten-
Vertical post-tensioning is occasionally used to dons, and careful detailing and placement are re-
accommodate high shear stresses, and for connec- quired to assure that sufficient space is provided
tion of the superstructure to piers or abutments for proper placement and vibration of the con-
so that moments can be transmitted. Connections crete. For this reason, it is usually recommended
between the superstructure and the substructure that the transverse tendons be placed on top of
are made by vertical tendons which pass through the longitudinal tendons (also see discussion in
the pier segments and are anchored in the pier. In Section 3.5.2 relative to bar tendon details used
some cases, coupling of the vertical tendons is for Kishwaukee River Bridge).
necessary, particularly when access to the anchor-
ages at the surface of the piers is difficult. Most
tendons used to connect the superstructure to the 2.5.3 Temporary Post-Tensioning
substructure are relatively short, so it becomes Most segmental structures with epoxy joints
important that allowance be made for the anchor are erected as cantilevers. Permanent cantilever
seating loss. Vertical post-tensioning in webs, some- post-tensioning is applied after a segment has been
times called prestressed stirrups, may be used to erected at each end of the cantilever. As a result,
help offset high principal stresses(3). during the placing of the first segment at one end,
The stress in short vertical tendons may be sig- the element has to be attached to the cantilever
nificantly affected by anchorage seating losses. by means of temporary post-tensioning. The tem-
Lift-off tests are recommended to ensure that the porary post-tensioning also provides compression
correct stress has been applied to prestressed stir- of not less than 50 psi (0.35 MPa) in the joints to
rups. The maximum ultimate strength of these be sure that the joints are properly closed and that
individual tendons has to be limited to about 200 the excess epoxy is squeezed out. It is recom-
kips (890 kN) in order that they can be incorpor- mended that uniformly distributed compressive
ated within the normal web thickness. These ten- stress be applied across the joints to avoid small
dons normally have an active stressing anchor and a differences in the thickness of the epoxy joint
blind or passive dead-end anchor which is em- which could affect the structure geometry. The
bedded in the concrete. It is strongly recom- temporary post-tensioning usually consists of bars
mended that web tendons be installed vertically to because of the short length of the tendons (about
avoid passing through the joints. two times the length of the segments). In the bar
Except for smaller segments, transverse post- tendon details used for the Kishwaukee River
tensioning of top slabs is recommended to mini- Bridge (Fig. 2.101, the permanent longitudinal
mize the top slab thickness and to provide assur- post-tensioning also serves to provide the tem-
ance against the development of longitudinal porary compression during erection. This facili-
cracking in the top slab. The transverse tendons in tated the construction process through elimina-
bridges only one segment wide can be stressed at tion of temporary stressing operations.
any time after the segments have been removed Temporary tendons, when required, may be lo-
from the forms. cated inside or outside the segments. It is often
Transverse post-tensioning may also be used to simplest to place the bars in the top and bottom
connect the top slabs of superstructures containing slabs of the segments. The anchors may be placed
more than one segment in the transverse direction, in recesses at the joints. Alternatively, the connec-
as illustrated by Fig. 2.2. These tendons run tion may be made by use of temporary steel
through the longitudinal cast-in-place joint be- attachments such as illustrated in Fig. 2.11. Be-
tween the segments. Placing, stressing and grouting cause the temporary bars are reused it is recom-
of these tendons is done after erection and obvi- mended that prestressing force be limited to about
ously requires careful control of the deflections 55 percent of the ultimate strength of the bars.
of adjacent catilevers. To facilitate placing the ten- The holes and the recesses for temporary tendons
dons, the width of the longitudinal joint must not and anchorages should be grouted after the perma-
be less than 2 ft. (0.6 m). Narrower joints are nent post-tensioning has been stressed.

PLAN Blockout to be filled

with mnrrrt+
-_. . . . --. .-.-..
;,;c,A, errin 9of
Bloc Lout heopeng Cut Bar.raftar rtressing
rfor g r o u t i n g \-- , __j /-as requared to fit blockout.

. :. .
. -.

Sepment Jt. /


Segment Jt./

Blockout in Blockout to be filled
with concrete l ftor
stressing of tendon

Ie Tendon Lfo? Longit: duct


Fig. 2.10 - Stressing and coupler blockout details - Kishwaukee River Bridge
anchored at the same location at the segment
joints. In developing the tendon layout to comply
with the above requirements, the number of ten-
dons required is the design consideration of most
Some practical suggestions relative to location
and detailing of tendon layouts aoae as follows:
1. Tendon spacing must be sufficient to permit
placement and vibration of concrete without
development of voids or honeycomb. A clear
distance of 1% in. (38 mm) is required be-
tween tendons during grouting to minimize
the possibility of grout transmission between
Fig. 2.11 - Temporary steel fittings attached to deck for
anchoring temporary prestressing bars adjacent ducts at the joints between seg-
ments. A typical layout of ducts meeting
In place of permanent vertical post-tensioning those requirements is presented in Fig. 2.12.
between pier segments and piers, post-tensioning 2. The bending radius of the tendons is deter-
may be employed temporarily to provide a mo- mined largely by the duct material. A semi-
ment connection during cantilever erection only. rigid duct of corrugated metal is preferable,
After erection has been completed and the con- and the minimum bending radius of such ducts
tinuity tendons have been placed and stressed, is about 15 ft. (4.6 m). Pre-bending requires
the temporary vertical post-tensioning at piers may an additional operation and complicates
be removed. This permits use of sliding bearings placement of the ducts. Sharp bends are un-
at piers in the finished structure to accommodate desirable from the standpoint of installing
volume changes due to temperature, shrinkage, and tendons, friction losses, and the high concen-
creep. trated forces resulting on the concrete.
3. A free passage of 5 in. (127 mm) minimum
width should be provided between tendons
located over the segment webs for proper
2.5.4 Layout of Post-Tensioning Tendons placement and vibration of concrete.
Unlike design of conventionally reinforced con- 4. Crossing of longitudinal tendons in the nar-
crete structural elements where a quantity of rein- row part of the web should be avoided.
forcement may be the final result of design calcu- 5. Tendon eccentricities should be made as large
lations, a practical tendon layout always requires as possible. Cantilever tendons can be spread
an iterative design process in which the designer laterally into the top slab and a second layer
and the detailer continuously exchange informa- of tendons can be accommodated in the top
tion. In the preliminary design stage, concrete slab haunches as shown in Fig. 2.12. Tendons
sections are assumed and bending moments and anchored in the first few segments remain
shear forces are calculated. Subsequently, an ini- within the web reinforcement because of
tial number and eccentricity of tendons required to bending radius limitations. This results in
counteract the bending stresses is determined along some loss of eccentricity. Midspan conti-
with the number and slope of tendons counter- nuity tendons are placed in the bottom slab.
acting shear forces. The preliminary design is com- 6. Cantilever strand tendons are anchored in
pleted by determination of the required mild the webs and top slab haunches, or on web
steel reinforcement. The preliminary design re- stiffeners. Cantilever bar tendons may be
sults must then be evaluated by the detailer on the anchored in the slab as shown in Fig. 2.10.
drawing board to see whether or not the prelimi- Shear tendons are anchored in webs. Con-
nary design assumptions can be achieved in prac- tinuity tendons are anchored as described in
tice. This is usually not the case on the first try, Section 2.5.2. The anchorage of continuity
and further iterations are then made. Detailing of tendons in the top slab combined with an-
post-tensioning tendons requires consideration of chorage of cantilever tendons in the webs
minimum radius of curvature, spacing require- provides a connection between the two
ments and avoidance of conflicts with mild steel overlapping tendon systems through con-
reinforcement. Further, because of formwork crete compression. In a layout where ten-
limitations, tendons are always located and dons are anchored in top and bottom slabs
the tendon anchors which may become an
important factor near the supports. Pre-
stressed stirrups may also be used to accom-
modate shear forces near supports.
8. Tendon lengths should be made as short as
possible. However, use of very short tendons
requires careful consideration of diffusion
of the prestress into the section and the pre-
stress losses due to seating of the anchorage.
Fig. 2.12 - Tendon spacing and 12-strand tendon anchor- From the structural viewpoint, the tendon
age details in top slab haunches layout may be in accordance with the bend-
ing moment diagram. However, the erection
procedure and the available anchorage loca-
tions usually require substantial adjustments
to the tendon layout resulting solely from
structural moment requirements.

2.6 Mild Reinforcement Cage

The amount of longitudinal and transverse rein-
forcement required is determined by the design
calculations or from the nominal minimum amounts
required to provide toughness during curing, hand-
ling and erection of the segments.
During production of the segments, the rein-
forcement is assembled and wire tied outside the
form to make a solid cage that can be lifted into
Fig. 2.13 - Tendon layout influence on the mode of shear
the form without damage. Spot welding of crossing
transfer between top and bottom slab tendons
bars in forming the reinforcement cage requires
control of the carbon content of the bars to as-
only, the connection between tendon systems sure weldability without producing brittleness.
is by shear in the webs. The shear transmis- Spot welding of reinforcement should be permitted
sion was accommodated in the bar tendon only when authorized by the Engineer. Tendon
details used for the Kishwaukee River Bridge ducts frequently pass through layers of reinforce-
by extending all longitudinal tendons one seg- ment. Details should be developed to accommo-
ment length beyond the point required by date the tendon trajectory without cutting the
design moments. The two means of providing reinforcement. Fig. 2.14 shows a possible solu-
a connection of the two tendon systems are tion to the case where tendons are located in the
illustrated in Fig. 2.13. Both systems have top slab and anchored in the web. The top slab
been used successfully, but the designer and web haunches permit use of two types of
should keep in mind the difference by which hairpin bars, a and b, which permit the tendons
forces are transmitted between the two sys- to pass easily.
tems of tendons.
7. The slope of continuity tendon anchorages
with respect to the top slab should be about
25 degrees as shown in Fig. 2.8. This shortens
the block-out to acceptable limits (the block-
outs interrupt the transverse reinforcement)
and also reduces the tendency of the anchor
to break out vertically. The 25 degree slope is
also appropriate for cantilever tendons an-
chored in webs. The vertical component of
the tendon is then about 40 percent of the
tendon force. This provides a substantial re- Fig. 2.14 - Reinforcement details to permit anchorage of
duction in the shear forces in the webs above top slab tendons in web
Shear Keys
Shear keys in the webs serve the dual purpose of
transferring shear during erection and providing a b
guide to assure the correct vertical position of the
segment. Horizontal alignment is obtained by use
of a guide in the top slab. The erection shear re- ~
sults from the weight of one or more segments Fig. 2.16 - Reinforcement requirements near web shear
(depending on the erection speed) or the upward keys
force resulting from inclined post-tensioning ten-
dons. Stability during erection is obtained through In conjunction with the loading cases in Figs.
the combined action of the shear keys and the tem- 2.15(a) and 2.15(b), reinforcement should be
porary (or permanent) post-tensioning in the top provided in webs to contain potential crack devel-
and bottom slab. As indicated in Section 2.5.3, opment in both the upward and downward direc-
the temporary post-tensioning is proportioned to tions as shown in Fig. 2.16.
provide a uniform compression of not less than Recent European bridges have utilized multiple
50 psi (0.35 MPa) across the entire joint. The shear keys in the webs such as shown in Fig. 2.17.
forces R, acting on the shear key and the joint The multiple key eliminates the need to reinforce
due to segment weight and temporary post-tension- the shear key and the adjacent web area, and it has
ing are illustrated in Fig. 2.15(a), and due to the further significant advantage of relieving the
segment weight and final cantilever post-tension- epoxy of any shear transmission function. The
ing in Fig. 2.15(b). The use of single web shear large number of interlocking keys [(l) in Fig.
keys such as shown in Fig. 2.15 requires careful 2.171 in the webs carry all the shear across the
attention to reinforcement details in the shear joint without any assistance from the epoxy.
keys and in the web area adjacent to the keys. Note also the keys across the top slab [(2) in Fig.
2.171 which assist in obtaining segment alignment
during erection and which may also provide shear
transfer due to concentrated loads on the deck.
The use of the multiple key web design in Fig.
2.17 is associated with a web stiffener (3) which
contains tendon duct and anchorages for perma-
nent (4) and temporary post-tensioning (6). The
top slab has vertical holes (5) adjacent to the stif-
fener which permit an attachment for handling
the segment. The use of multiple web keys requires
w = Segment weight
F1 = Temporary prestress a substantial web area free of anchorage pockets,
in top slab tendon holes, and other interruptions which would
Fz = Temporary prestress
in bottom slab
RI = Force on joint
R2 = Shear key force

Fig. 2.15 (a) - Forces on shear key due to temporary post-

tensioning and segment weight

(1) Castellated web key.

(2) Slab key for alignment.
(3) Web stiffener.
F = Force in permanent (4) Tendon duct and anchorage for final assembly
tendon (5) Insert for handling and temporary assembly
Fi2 = Shear key force
RI = Force in joint
(6) Tendon ducts for temporary assembly
Fig. 2.17 - Precast segment with multiple keys and web
Fig. 2.15 (b) - Forces on web shear keys stiffener()
reduce the available shear area of the keys. This
leads to use of web stiffener details such as shown
in Fig. 2.7, which involve additional effort during
production of the segments.

2.8 Epoxy Joints

As indicated in Section 2.7, the function of the
epoxy joint is, to an extent, dependent on the
design of the shear keys. However, in all cases, the
epoxy will serve the following purposes:
1. During placement of segments, the epoxy
acts as a lubricant which, in conjunction with
the keys in the web and top slab, assists in
guiding the segment into proper alignment.
2. The epoxy layer acts as a stress distribution
material during erection and during post-
tensioning. This is illustrated by the fact that
the thin layer of epoxy cannot be pressed out
of the joint entirely. In addition, any small
cavities and pores in the faces of the seg- Fig. 2.18 - Application of epoxy resin by gloved hand
ments are filled.
3. Epoxy can restore the tensile and shear
case, little difference between a precast structure
strength of the concrete across the joint.
with joints and a monolithic cast-in-place structure.
4. Epoxy is required to serve as a joint sealant Application of epoxy to the joint surfaces is
to prevent water from entering into tendon
accomplished by hand immediately prior to ap-
ducts, and also to prevent grout leaks at
plication of the temporary post-tensioning, as
illustrated in Fig. 2.18. Prior to application of the
Concern is occasionally expressed about the lack epoxy, the joint surfaces are either sand blasted or
of reinforcement extending through joints of pre- wire brushed to remove any surface laitance. This
cast segmental bridges. Actually, there is a great is usually done while the segments are stockpiled
deal of grouted high strength post-tensioning rein- awaiting erection.
forcement continuous through all joints. This rein- Recommended specifications and tests for
forcement exerts a very large compressive force epoxies to be used in joints of segmental bridges
across the joint which ensures that the joint will are presented in the Tentative Design and Con-
be under compression (or perhaps very low tensile struction Specifications for Precast Segmental
stresses at the bottom slab) under service loads. Box Girder Bridges developed by the Prestressed
The safety of the structure in both shear and flex- Concrete Institutes Bridge Committee. These
ure at ultimate load is, of course, determined on specifications are presented in Appendix Section
the basis of a cracked section, and there is, in this A.l.
CHAPTER 3 the superstructure and substructure design, and
ANALYSIS OF PRECAST should be considered in selecting preliminary
Selection of the span arrangement and other
3.1 General considerations preliminary to the analysis phase
are considered in the following sections.
The material presented in this chapter deals
primarily with those aspects of precast segmental
bridge design that differ from or require more de-
tailed consideration than conventional types of 3.2.1 Selection of Span Arrangement
continuous prestressed concrete structures. Back- In selecting the span arrangement for a precast
ground information on the fundamentals of analy- segmental bridge, it is necessary to consider the
sis of continuous prestressed concrete structures method of construction. When cantilever con-
may be obtained from References 2, 4, 5 and 18, struction is used, the segments are erected in bal-
Appendix Section A.4. anced cantilever starting from a pier and placing
In general, analysis and design of precast seg- segments on either side in a symmetrical opera-
mental box girder bridges should conform to the tion. This method of erection results in typical
latest edition of the Specifications for Highway superstructure components consisting of one-
Bridges published by the American Association of half of the main span length cantilevered from the
State Highway and Transportation Officials6, or piers as shown in Fig. 3.1 (a). If the end span is
to other applicable specifications for railway or selected as 65 to 70 percent of the interior span as
rapid transit structures. Additional specifications in Fig. 3.1 (a), the small section of the superstruc-
developed by the Prestressed Concrete Institute for ture adjacent to the abutment will require use of
consideration by the American Association of falsework or some other erection procedure.
State Highway and Transportation Officials to To provide a transition between span lengths Ll
provide specific coverage of precast segmental and L2, for example at the transition between
bridges are presented in Appendix Section A.1. approaches and main spans in a viaduct, an inter-
In order to provide background on those aspects mediate span of average length will optimize the
of precast segmental bridge design that may re- use of the cantilever concept, as illustrated in
quire special consideration, the discussions in the Fig. 3.1 (b).
following sections on the influence of creep,
shear lag, temperature effects, and transverse
analysis are presented in much more detail than
may be necessary for routine designs. As suggested
by the specifications in Appendix A.l., elastic
analysis using beam theory may be used in the (4
design of precast segmental bridges of normal
proportions. Consideration is given to shear lag in
the immediate vicinity of the supports when seg-
ments are wider and/or shallower than normal
0.66-07OL I L 1 065-070L
(see Section 2.2).
Notation is generally explained as it is used in
the text. In addition, notation is presented in
Appendix Section A.3.

3.2 Development of Preliminary Bridge Details

As in any bridge design, it is necessary to assume
cross section dimensions and span lengths of a
precast segmental bridge before an analysis can be
made. The selection of the superstructure cross
section, normal span/depth ratios, and other
pertinent aspects of superstructure design are
discussed in Chapter 2. The method of erection, Fig. 3.1 - Span arrangements for precast segmental
as discussed in Section 4.3, also has an affect on bridges()

With end span length on the order of 65 to 70
percent of the interior spans, a special segment
may be used at the abutment and one or two seg-
ments may be temporarily cantilevered out to
reach .the first balanced cantilever as shown in Fig.
3.3. (b).
When end spans are only 50 percent of the
length of interior spans, as in Fig. 3.3 (c), an up-
lift reaction has to be transferred to the abutment
during construction and in the completed struc-
ture. Abutment details that may be used to accom-
+SDmm plish this are shown in Fig. 3.3 (d). Here, the webs

of the main box girder deck are cantilevered under

the expansion joint into slots in the main abut-
ment wall. Neoprene bearings are placed above

i I

1 1

Fig. 3.2 - Effect of hinge location on deflection

Continuous bridges over 2000 ft. (610 m)

long have been built without permanent hinges
or expansion joints in the superstructure. It is de-
sirable to keep the number of joints to a minimum
to reduce maintenance costs and improve riding
quality. This may be accomplished by use of piers
which permit longitudinal volume changes of the
superstructure (for example the Chillon Viaduct
shown in Fig. 4.18), or by the use of bearing
details that will accommodate substantial move-
ment. In very long structures, intermediate expan-
sion joints become necessary. Location of these
joints near the dead load contraflexure point, as
shown in Fig. 3.1 (c), will be helpful in reducing PRESTRESSING TENDONS

deflection of the joint. Fig. 3.2 shows a compari-

son of deflections and angle changes due to live
load in a 259 ft. (79 m) span with hinges located at
mid-span and near the point of contraflexure.

3.2.2 Abutment Detaild7 LONGITUDINAL SECTION

When geometric restraints will not permit opti-
mum pier locations or span arrangements, abut-
ment details may be developed to facilitate the
construction procedure. Fig. 3.3 (a) shows a deck
section cantilevered over a front abutment wall to
achieve a longer than normal end span. A conven-
tional bearing is provided at the front abutment
wall in Fig. 3.3 (a) and a rear prestressed tie is
used to counteract uplift and to permit cantilever
construction to proceed out to the first joint
Jl where a connection is made with the cantilever SECTION C-C
construction starting from the first intermediate
pier. Fig. 3.3 - Alternatives for construction of end spans()

the webs to transmit the uplift force and, at the To accommodate large movements and heavy
same time, to allow the deck to expand freely. loads, the use of more expensive pot type bearings
using neoprene to absorb rotation and a teflon
layer to permit volume changes may be appro-
3.2.3 Pier Details priate. Design information on these bearings is
available from suppliers.
Pier details should be developed with consider-
Heavy pier reactions during erection, or tem-
ation given to the need to provide stability to the
porary prestressing of the pier segment to the pier,
cantilevers during construction. Some details that
may require use of temporary bearing pads of steel
have been used to accomplish this are discussed
or concrete. Details of this type are shown in Sec-
and illustrated in Section 4.3.6.
tion 4.3.6 (see Figs. 4.20 and 4.21). The use of
four bearings at piers as shown in Fig. 4.21 sub-
stantially reduces the positive longitudinal live
3.2.4 Horizontal and Vertical Curvature
load moments in the superstructure, as illustrated
As noted in Chapter 1 and elsewhere, precast in Fig. 3.5.
segmental construction is readily adapted to nearly
any horizontal and vertical alignment by adjusting
the segment dimensions during casting. The Bear 3.3 Longitudinal Analysis
River Bridge, shown in Figs. 1.4 and 1 .lO, and the
Saint-Cloud Viaduct in France, shown in Fig. 3.4, 3.3.1 Erection Moments
are examples of bridges on curved alignment.
During erection, the moments over the piers
increase with the addition of each pair of segments,
as illustrated in Fig. 3.6. The additional moment
3.2.5 Bearing Details caused by adding segments No. 8 at each end of
Most European bridges have utilized laminated the cantilever is shown by the shaded area in Fig.
neoprene bearings. However, the European specifi- 3.6. These moments are resisted by post-tension-
cations for design of neoprene bearings are con- ing tendons in the top slab which may be anchored
siderably less restrictive than U.S. specifications. at the face of the segments or in build-outs inside

Fig. 3.4 - Saint-Cloud Bridge, Paris, France()

LIVE LOAD = 4K/ft (58.4 kN/m)

260 260 260



1 ft. kip = 1.356 kN-m


Fig. 3.5 - Comparison of superstructure live load moments with simple and double pier supports()

the box section. The use of build-outs makes it cast segmental bridges during erection are modi-
possible to place the segments and stress the ten- fied by thechange in statical system due to coupling
dons in two separate operations, but tends to com- cantilevers and the post-tensioning used to connect
plicate the process of manufacturing the segments. the cantilevers into a continuous structure. Subse-
The amount of post-tensioning required to main- quent to casting the closure joint and stressing
tain zero tensile stress in the top slab under the of the continuity tendons, the influence of con-
erection moments (including weight of any erec- crete creep modifies both the cantilever and con-
tion equipment) is readily calculated from the tinuity moments as will be illustrated in the fol-
simple formula: lowing sections.
Creep deformation of concrete is that part of
ME p We)
-=-+- the inelastic deformation not caused by shrinkage.
Zt A Zt Creep deformations occur as a result of the inelas-
tic response of concrete to long term loadings such
where M, = erection moment, in. lb. as dead load, post-tensioning forces, and perma-
Z, = section modulus with respect to nent displacements. Restraint of creep deforma-
top fibers, in.3 tions causes redistribution of moments. This hap-
P = post-tensioning force, lb.
pens, for example, when statical systems are
A = cross sectional area of pier seg-
changed by connecting a cantilever structure into
ment, ins2 a continuous structure. The effect of permanent
e = eccentricity of post-tensioning
deformations by external causes is reduced by
force, in.
creep. This occurs in the case of support settle-
The concrete area in the bottom slab at the pier ments.
must be sufficient to maintain compressive stresses The relationship between creep deformations
to the value allowed by the specifications. The and elastic deformations is linear. The ratio is
stress f,, is calculated as: called the creep factor 6. The following relation-
ship can be expressed for $ :
f,, = !$+;-F

where z,-, = section modulus with respect to bot-

tom fibers, in.3
where e,, = creep strain
Ee = elastic strain
u = stress
3.2.2 Creep Analysis E = elastic modulus of concrete at age of
The moments existing in the cantilevers of pre- 28 days
(27 120

~-9~-7[+ [-"I 3~i-p -[--i-r-F[-q-Tp 1 6 j 7 j e j 9 j 10 1 n j

Fig. 3.6 - Dead load moment development during cantilever erectiont2)

for various age concretes by simply subtracting 1

from the ordinate. A more detailed procedure for
evaluation of 8 is presented in Section
The following sections illustrate the effect of
concrete creep on the magnitude of moment
redistribution and reduction of the effects of
deformations due to shrinkage and support settle-
ments in precast segmental bridges. Creep Effects Resulting From Change of

Statical System Due to Closure of Central
Fig. 3.8 (a) shows a double cantilever with an
open joint at B. The elastic deflection is 6 and the
angle of rotation at the ends of the cantilevers is
Q as shown in Fig. 3.8 (b). If the joint remains
open, the deflection at time t will have increased to
S(l + #,) and the angle of rotation to a(1 + #,),
where #, is the creep factor at time t. For a uni-
formly distributed load q applied when the con-
0 37I42?;a42% 3 1 5 6 5 40 ml crete is 28 days old, and a length of cantilever Q:
Duration of Loading
Fig. 3.7 - Concrete strains vs. age and duration of
loading(7 where I = moment of inertia of the cantilever
The relationship between total concrete strain and E = elastic modulus of concrete at 28 days
the reference strain of a 28day old concrete sub- If the joint at B is closed after application of the
jected to short term load is illustrated in Fig. 3.7. load, the increase in angle of rotation a#, is re-
The value of @I can be estimated from this figure strained. As a result, the moment M, develops as

A graph of (l-e-@t) vs. values of 4 is presented in
Fig. 3.9.
Using the relationships for (II and p:

Substituting in the above, noting that 2!? = L

(l-e+) = qLZ(l-e+t)
M, t qQ2
6 24

By evaluating the equation for M, for a large

value of @t it is found that M, = qL2/24 which is
the same moment that would have been obtained
if the joint at B had been closed before the load q
was applied. This illustrates the fact that moment
redistributions due to creep following a change
in the statical system tend to approach the
moment distribution that relates to the statical
system obtained after the change.


Fig. 3.8 - Deformation of cantilevers before and after


shown in Fig. 3.8 (cl. The moment M,, if acting

in the cantilever, causes rotation at 6 defined aso.
The magnitude of fl may be calculated as:

The restraint moment M, produces both elastic

and creep deformations. During a time interval
dt, the creep factor increases by d$,. As a result,
QI increases by (ud&, and p increases by pd@,
(creep) and dp (elastic). From these relations and
the fact that there is no net increase in disconti-
nuity after the joint is closed we may write the
general compatibility of angular deformation
expression :
Fig. 3.9 - Variation in creep factors for both creep and

(cu-PI Referring to Fig. 3.10, the general relationship may

Integrating this expression: be stated:

-jr = In(cY-p)+C M,, = (1-e-G) (M,,-M,)

where M,, = creep m o m e n t resulting from
Evaluating the constant of integration:
change of statical system
When $t = 0, p = 0 --f C = -lna M , = moment due to loads before change
of statical system
-P = (I-em@t) M,, = moment due to same loads applied
a on changed statical system

Fig. 3.10 - Moment curves for cantilever system (I), fixed-end system (II), and cantilever system with later construction to
form fixed-end system (I I I) The Effect of Creep on Moments due to Solving this equation as in Section
Support Settlements X, = P (l-e+t)
Fig. 3.11 (a) illustrates a beam fixed at end A
and supported at end B. In Fig. 3.11 (b), the beam
is assumed to settle suddenly at B a distance 6.
The effect of this settlement is an additional mo-
ment at A which can be calculated as:

M = -PR
where P =-
In Fig. 3.11 (c), the support has been removed at
B and the beam is loaded with a load equal to P.
The deflection resulting from the load P in the
time interval dt increases by 6d@,. In Fig. 3.11
(d), the support is again applied at B and the in-
crease of the deflection 6d@, resulting from the
load P is presumed to be eliminated by upward
displacement caused by an increase in the support
reaction in an amount of X,. The level of support
B does not change between Figs. 3.11 (b) and
3.11 (d). The increase in the support reaction
X, induces both elastic (by dX,) and creep (by
X,d@,) deformations. Since there is no further
deflection after Fig. 3.11 (b), the elastic and creep
deformations due to the reaction X, may be
equated to the creep deformation due to P. This (4 1%
gives the following expression: Fig. 3.11 - The effect of creep on moments due to support
(dX, + Xtd&) = Pd& settlements

The support reactions at B vary as follows: $J = 1.0, the value of e-d = 0.368, the final super-
Immediately after settlement the support at B structure moments due to the 1 in. settlement at
carries: R - P support 2 are as shown in Fig. 3.12 (c).
After the creep process, the support carries:
R- P+ P (l-et) = R- Pe$t
In a similar manner, the moments at A due to the The Effect of Creep in Reducing Re-
settlement vary: straint Forces due to Shrinkage
Immediately after settlement: M = -PP
For this analysis, it is assumed that the shrink-
After the creep process: age at infinity, +,, develops with time at the same
M = -PQ + P (l-e-#t)Il rate as the creep factor. This assumption leads to
the equation:
M = -pQe-@t
The ultimate effect of creep on the reaction at
B and moment at A resulting from a support
settlement can be evaluated from the above form-
ulas by considering the value of e-4 for various where E,ht = shrinkage strain at time t
values of 4 as follows:
Esh = shrinkage strain at infinity
6 1 23 4 5 Development of the restraining force due to
e-4 0.368 0.135 0.05 0.018 0.007 0 shrinkage will be illustrated for the beam AB
It can be seen from the above that the effect of a shown in F/g. 3.13 (a) which is fixed against hori-
support settlement is reduced to zero by a large zontal movement at both ends. Due to shrinkage,
the beam shortens by:
value of $J~. As in the case of change in the statical
system, the creep redistributions have the tendency A sht = Esht Q

to approach the distribution belonging to the If the restraint to horizontal movement in the joint
system obtained after the change. at B is temporarily released, the beam would
To illustrate the application of the above, Fig. shorten due to shrinkage. Applying an axial force
3.12 (a) shows a three-span superstructure sub- S, to the beam at B as shown in Fig. 3.13 (b),
jected to a settlement of 1 in. at support 2. Fig. the beam elongates according to:
3.12 (b) shows the moment diagram resulting from
the 1 in. settlement at support 2. For a value of AS,=s,B

For the same time interval, the force S, induces


95,-O I 190-0 95,-O

elastic (dS,QIEA) and creep (S,Qd#,IEA) elonga-
tions which are equal to the shrinkage during the

\E,= 320 x lo6 K-W

1 ft. = 0.3048 m
1 k-ft. = 1.356 kN-m
1 k-ft.2 = 0.413 kN-mZ
Fig. 3.12 - Superstructure moments due to support
settlement Fig. 3.13 - Restraint force resulting from shrinkage

time interval. This leads to the following expres- where 4 ct ,t,) = magnitude of the creep factor
sion: at time t for a concrete specimen
E,,, ad@,/@ = S,Qd@,fEA+dS,Q/EA loaded at time t,.
Q dm = magnitude of delayed elasticity
-dS,/EA at infinity
(Es,, /G3,/EA) = factor variable from zero to

unity indicating the variation of

Integrating this expression as in Section @d with time
gives: @f- = magnitude of flow at infinity
Or(t) -Of&) = factor variable from zero to
s, = Esh EA unity indicating the variation of
@ & with time
to = theoretical age of concrete at
The quantity E,~ EA is the force required if all loading (days)
the shrinkage were taken elastically. t = theoretical time after casting
Setting this quantity equal to S, the above equa- (days)
tion becomes: The numerical value of delayed elasticity after an
infinite time has been determined as #d, = 0.4.
s = s (1-e) The recoverable nature of this part of the creep
f 0 factor will have consequences only for temporary
loads acting on a structure, such as those applied
where 4 = 4.. during construction by launching girders or other
temporary erection equipment. For dead load,
A graph of the values of (l-e-@)/@ is presented in post-tensioning forces, and other permanent
Fig. 3.9. This graph illustrates the reduction of loads #d is added to the value of @f.
shrinkage restraint forces by creep. The value of The variation of @d with time is shown in Fig.
(l-e-@)/@ for 4 = 2.0 is about 0.43. This indicates 3.14, where the factor 0, is given as the ordinate,
that shrinkage restraint forces would be reduced 57 and the duration of the loading (t-t,) is the ab-
percent for 4 = 2.0. In general, the creep reduction scissa.
of the effects of a slow process, like shrinkage, The fact that fld depends only on the duration
can be evaluated by division of the results obtained of the loading explains the elastic tIatUre of @d.
from a fast process like a sudden support settle- With time, the full deformation due to loading or
ment or a sudden change in statical system by the unloading will develop. By comparison of Figs.
creep factor 4. 3.14 and 3.16, fld develops somewhat faster than
pf: 30 percent of @d takes place in one day, 50
percent after 30 days, and 90 percent within a
year. Determination of the Creep Factor) The magnitude of the flow, @f,, at infinity de-
pends on the relative humidity of the ambient
The creep factor, $J, was defined in Section
medium and the composition of the concrete.
3.3.2 as the ratio of creep strain to elastic strain.
For the precise determination of its value, @I
must be considered to be the sum of recoverable
creep, @d, and irrecoverable creep, I#J~:
4 = 4d + Gf
Recoverable creep and irrecoverable creep are
referred to below as delayed plasticity and fld
flow, respectively.
Both C#I~ and @r are time dependent, but accord-
ing to different relations. These relations are in- 03

troduced into the expression for $ in the following

1 I I I I I
manner: 1 5 10 50 100 500 low sow 10.000
t-t,, days

Fig. 3.14 - Variation of the delayed elasticity with

4 = @d,Pd(t--to) + @Jr, p(t) - h,)] time()
Table 3.1 Variation of PC1 and h with humidity of ambient medium and composition of the concrete()

Concrete Composition
Relative Humidity of Stiff Concrete Plastic Concrete Soft Concrete Thickness
Ambient Medium Slump l/2-314 Slump 1-2 Slump 3-6 Factor
(13-19mm) (25-51mm) (76-152mm)
P Cl P Cl P cl x

In water- 100% 0.60 0.80 1 .oo 30

In Damp Atmosphere;
Over Water-go% 0.975 1.30 1.625 5
Outdoors-70% 1.50 2.00 2.50 1.5
Dry Atmosphere:
Interior of Building-40% 2.25 3.00 3.75 1

These factors are represented by PC, in Table 3.1 The variation of #f with time is shown in Fig.
@f- also depends on the theoretical thickness 3.16. The ordinate shows the factor of develop-
hth of the structural element in combination with ment of and the abscissa the time t, in days. In
the relative humidity of the atmosphere. These contrast to delayed elasticity, ed, the time scale
factors are represented by pc2. The value of #f, in Fig. 3.16 begins at the time the concrete is cast.
at infinity is the product of &, and Bc2 : Therefore, the influence of the age at loading, t,,
is obtained from the expression [flf(t)-~rct,b]. The
Of, =I&, x PC2
dependence of the rate of development of @Jr on
The theoretical thickness, hth, is evaluated from: the thickness of the member and the relative hu-
midity of the environment is indicated in Fig.
hth = CC 3.16 by the different curves for various theoretical
where X = theoretical thickness factor, taken As suggested by Fig. 3.7, loading of concrete
from Table 3.1 at an early age greatly increases the final flow fac-
A, = area of concrete section, cm2 tor, #f. In addition to age at loading, an adjustment
I-c = perimeter of concrete section in contact in creep effect calculations may be necessary when
with the atmosphere, cm a rapid hardening cement is used, or when the
After evaluating hth as above, the value of PC2 process of cement hydration is hampered because
can be taken from Fig. 3.15 and the value of of low temperatures. Such corrections may be
@f, can be calculated. made by calculating a theoretical age for the con-
crete by use of the formula:

a ; [T(,., + lo] At
where t = theoretical age
a = 1 .O for ASTM cement Types I and I I
Q = 2.0 for ASTM cement Type II I
QI = 3.0 for cement having highly accel-

When concrete cures at 20 C (68O F) and normal

55 1 0 20 40 60 60 2160 hardening cement is used, theoretical time and real
h,h, cm 1 cm = 0.39 in. time are equivalent. Theoretical time and real time
are also equivalent when loading takes place im-
Fig. 3.15 - Effect of member thickness on flow() mediately after the curing process is over. This is

36 Example Creep Factor Calculations
To provide a numerical example of creep factor
calculations, a three-span example bridge will be
assumed which has 44 segments produced at a
rate of one segment per day over a period of nine
weeks. The average concrete thickness is 0.32 m
(12.6 in.). Slump of the concrete was 1% in. /38
mm). A three-week erection period starts four
weeks after production of the last segment. The
structure is made continuous by casting a midspan
splice one week after completion of segment
0.25 erection, and the bridge is erected over water.
The creep factor to be used for the moment
redistribution calculations is obtained as follows:
0 10 loo 1000 8000
Time t, days 1 cm = 0.39 in.

Fig. 3.16 - Variation of flow with time)

where: #.&, = 0.4
fld(t--to) is obtained frOmI Fig. 3.14 at age
of seven days. The delayed elasticity that
occurs during the week after erection
normally the case for precast segmental bridges.
while the structure is not continuous
If the age of loading has been assumed as 7
amounts to &(t-t,) = 0.38. Only the re-
days in the creep calculation, an equivalent age
mainder (l-Pd(t--t,j) = 1.0 - 0.38 = 0.62
can be obtained by:
contributes to the moment redistribution.
- curing 7 days at 20 C and use of Type I or The value of #r, is calculated from:

1 (20+10)7 #f, = ljcl x&2

Type II cement since: = 7 &, is taken from Table 3.1 The value is
30 1.3.
Theoretical thickness hth = h2Ac/~ =
- curing 4 days at 16O C (61 F) and use of Type
5(0.32) = 1.60 m (5.25 ft.)
2(16+10)4 The corresponding value of Bc2 = 1.12
I I I cement since: = 7
is taken from Fig. 3.15.
The values for Sfttj and flfct,, are taken
- curing 3 days at 13X0 C (56O F) and use of ce- from Fig. 3.16. The value of Prctj at t =
ment having highly accelerated strength infinity equals 1. The average age of the
3 (13.5 + 10) 3 concrete at loading, based on the indi-
gain since: = 7 cated time schedule is 9/2 + 4 + 3 +l =
30 12% weeks, and from Fig. 3.16, the cor-
responding value for Pfct,) = 0.3.
Alternatively, use of normal cement and curing of
4 days at 16 C and 3 days at 13.5 C gives a theo- Therefore:
4 = 0.4 (0.62) + 1.3 x 1.12 (l-0.3)
retical age of only:
6 = 0.25 + 1.02 = 1.27
(16 + 10) 4 + (13.5 + 10) 3 Moment redistribution calculations will
= 5.5 days, and load-
30 be carried out for:
4 Low = 0.85 x 1.27 = 1.07
ing should be postponed for 1.5 days.
4 High = 1.15 x 1.27 = 1.46
Due to the importance of the creep factor in
design calculations for precast segmental bridges
and the inherent uncertainty in determination of
the creep factor, it is recommended that calcula- Influence of Creep on Superstructure
tions be made using values of the creep factor in- Moments
creased and decreased 15 percent from the theor- The theoretical considerations of the influence
etical value. of creep in redistribution of moments presented in

Section are applied to actual bridge exam- Moment Calculations

ples for a variety of loading conditions in this

Section. The effects of dead load, cantilever pre-
stress, continuity prestress, and other loadings that
may cause moment redistribution are treated
separately. The general procedure is as follows (the
step numbers below do not necessarily relate to
the diagram numbers shown in the various exam-
-3- Casting of midspan
ples) : splice completed at
joints B and F.
Step 1. Bending moments are determined during Bending moments
the erection phase. at end of Step 3
are as Step 2
Step 2. Bending moments are determined in the -& Elastic distribution
continuous condition (the elastic moment of 3 (mid-span
splice completed at
distribution that would have occurred Dj. Bending mo-
if the structure had been erected in one ments continuous
single step). MB = MF = +2985
Step 3. The difference between the moments of k-ft.
MC = ME = -7548
Step 2 and Step 1 is calculated. This dif- K-ft.
M ,, = +4702 k-ft.
ference is always a moment diagram con-
-5- Difference between
sisting of straight lines, since it is merely I diagrams 4 and 2:
-_ -~~
the result of changed fixities (boundary M = +12250 -
7548 = +4702 k-ft.
conditions). -6- Creep moments
Step 4. The diagram obtained in Step 3 is multi- obtained by
plied by the factor (l-e-@) and the 6- II of diagram 5 by
2351 (1 - e+j = 0.5
creep moments are obtained. (for example
Step 5. Bending moments of Steps 1 and 4 are only)
0800 -7- Dead load mo-
added in order to find the moment dis- ments at infin-
tribution at infinity. 7 -==zx-- ity obtained from
2180 I -
1351 diagrams 2 and 6:
It should be noted that at any time between -lk-ft.= 1.356 kN-m

erection and infinity, the bending moments in the Fig. 3.17 (b) - Effect of creep on dead load moments -
structure will be between the values calculated in Example 1

Steps 1 and 5.
Comparing the examples in Figs. 3.17,3.18, and
3.19 it is seen that the final dead load bending Construction Procedure
moments in the structure depend on the order in Structure weighs 5k/lin.
which the joints are closed in the structure. In ft. (73 kN/mj
a Step 1. - Erect cantilev-
these same figures, it is seen that the magnitude of ers over sup-
the moment redistribution due to creep also de- ports C and E.

pends on the construction sequence and the num-

ber of spans in the structure.
ing segments
between AB
and FG on
Construction Procedure close joints B
100 ft. = 30.5 m and F, and re-
Structure weighs 5 k/ move false-
lin. ft. (73 kN/mj work.
Step 1. Erect cantilevers
Fig. 3.18 (a) - Effect of creep on dead load moments -
over supports C
Example 2

F-G on false-
k. P P work,close j o i n t s
B and F and re- Figs. 3.20 and 3.21 illustrate that the effect of
move falsework creep on the moments resulting from continuity
i P P
Step 3. Concrete splice
at D post-tensioning depends on the construction se-
Fig. 3.17 (a) - Effect of creep on dead load moments - quence and the order in which the tendons are
Example 1 stressed.

Moment Calculations Moment Calculations

Bending moments Bending moment

at C and E; result at support C, re-
of Step 1. sult of Step 1
Mc = 12250 k-ft. MC = 12250 k-ft.
- 2 - Bending moments -2- Bending moments
not effected by at joint B, support
Step 2. C, result of Step 2.
- 3 - Construction com- I I
pleted. Bending MB = +1575
moments resulting MC = 12250 k-ft.
from Step 3: - 3 - Bending moments
at joint B, supports
MB = +1471 k-ft. C and E, joint D,
MC = -12597 k-ft. result of Step 3
MD = -347 k-ft.
Elastic distribu- I MB = +1575,
tion of bending MC, ME = -12250
moments in con- Mg=O
tinuous structure: - 4 - Bending moments
at joint B,supports
MB = +2985 k-ft.
C,E,G, joints D
MC = -7548 k-ft.
MD = +4702 k-ft. and F, result of
Step 4 MB =
-5- Difference be-
+1575, MC, ME,
tween diagrams 4
MG = -12250,
and 3 Mg,MF=O
M = 4702 + 347 =
- 5 - Bending moments
5049 k-ft.
resulting from Step
-6- Creep moments 5, end of erection.
obtained by
multiplication of M =1575-
I I diagram 5 by 0.3 (38) = +1564
6 k-ft.
I (l-e+1 = 0.5 (in
212s example only) MC = -12,250 -
M,, = 0.5 x 5049 38 = -12,288 k-ft.
= 2525 k-ft. MD = +47 k-ft.
Dead load mo- ME = -12,250 +
ments at infinity 131 = -12,119

obtained from dia-
grams 3 and 6 MF = -178 k-ft.
M = -12250-
M =+1471+ 48% = -12;736
O.! (2525) = 2228 k-ft.
1 k-ft. = 1.356 kNm k-ft. hlkf = +1429 k-ft.
MC = -12597 +
Elastic bending
2525 = -10072
moment distribu-
MD = -347 + 2525 tion continuous
= +2178 k-ft. bridge.
Difference be-
tween diagrams 5
Fig. 3.18 (b) - Effect of creep on dead load moments - and 6.
Example 2 Creep bending
moments, ob-
tained by multi-
plication of dia-
gram 7 with factor
Construction Procedure (1 -e@) (here
chosen to be 0.5)
Structure weighs 5kllin.
Dead load bending
ft. (73 kN/m)
moments at infin-
Step 1 - Erect cantilever ity obtained by
over support C. addition of dia-
Step 2 - Erect tailspan grams 5 and 8.
segment be-
AGC D E F G HI tween A and B Fig. 3.19 (b) - Effect of creep on dead load moments -
on falsework. Example 3
, 100li P, ll
4.0 P, n
1.0 P, ,g - concrete joint
at B Construction Procedure
r.,. - remove false-
work Prestressing force
. . ,.;.. .:. Step 3 - Erect cantilever F,=F2=1000k=F
over support E M = Fe assumed not to
: : 1 - concrete joint rr*&mn II. rEWDOY F,- IL*DOW F, g vary with time
at D Eccentricity e = 3-0
1 Step 4 - Erect cantilever f (0.9 m)
h P P P 100
over support G. - Step 1 - Both halves of
: -concrete joint the structure
A1 at F.
100 ft. = 30.5 m erected
Step 5 - Erect tailspan
segment be-
tween H and I is stressed
1 k=4.448kN
on falsework
Step 2 - Midspan joint at
-concrete joint
C is concreted
at H. - midspan conti-
- remove false-
nuity prestress
is stressed.

Fig. 3.19 (a) - Effect of creep on dead load moments - Fig. 3.20 (a) - Effect of creep on moments due to conti-
Example 3 nuity post-tensioning - Example 1

Moment Calculations Moment Calculations

-2- Elastic moment

distribution if
tendons Fg were
stressed in contin-
3 uous system.
-3- Difference be- I - 3 - Difference be-
3- I tween diagrams 1
.484 F. tween 1 and 2.
and 2.
A- Creep Moments
-4- Creep bending obtained by multi-
moments obtained 4- 1 plication of 3 by
4 by multiplication I O.Z.?h f 1 -a-@) taken
of 3 b factor
I O.emf. as 0.5.
(1-e 4), taken as
0.75s F. -5- Final moments by
continuity pre-
s- I(1 I II,
stress Fg, obtained
I1 +
0.242P. 11 O.lo* P.
~,,551. -5- $gaE by addition of 1
and 4.
- 6 - Bending moments
-6- Elastic bending resulting from
moment distribu- Step 2. Tendons
tion by stressing Ft are stressed in
of tendons F2. the continuous
These tendons are system and are
6 I I i II stressed in contin- therefore not
uous system and subject to creep
therefore not moment redistri-
subject to creep bution.
moment redistri- Final moments
bution. due to all continu-
r= -7- Total bending mo- ity prestress ob-
merits by prestress tained by addition
of 5 and 6.
Ft and F2, ob-
tained by addition
of 3 and 6. Fig. 3.21 (b) - Effect of creep on moments due to conti-
nuity post-tensioning - Example 2
Fig. 3.20 (b) - Effect of creep on moments due to conti-
nuity post-tensioning - Example 1

Construction procedure
1 1 Prestressing force = F
d r I I d
(for simplicity assumed
10 ,70 30
100 I.0 100 constant over length and
Construction Procedure time)
100 ft. = 30.5 m Eccentricity = e
Prestressing force F, =
F2 = F
Eccentricity = e
M = Fe assumed not to
A a C D E F G vary with time
Step 1 - Erect cantilever
over supports C
and E
I 1
P P - concrete mid- -2- Elastic distribu-
span joint at D ,o.m P. 0.335F. tion of bending
- stress contin- 2-F /i moments by pre-
k n i I
uity tendon Fg. stress if stressed in
o.*wJ 5%
Step 2 - Erect segments continuous bridge.
in tailspan be-
tween A and B
(F and G)
- stress continuity
tendons F t .
-4- Creep bending
moments due to
Fig. 3.21 (a) - Effect of creep on moments due to conti- cantilever prestress
nuity post-tensioning - Example 2 I obtained by multi-
I plication of dia-
l I gram 3 with factor
(l-e+), taken as

Fig. 3.22 shows the influence of creep on the

moments due to the cantilever post-tensioning. In
this case, the effect of creep is independent of the
construction sequence since the stressing of the Fig. 3.22 - Effect of creep on moments due to cantilever
tendons does not change the statical system. post-tensioning

3.3.3 Analysis for Superimposed Dead Load The longitudinal effects of temperature cause
and Live Load the total structure length to increase or decrease,
The main loadings on a precast segmental box and where there is a temperature difference be-
girder bridge, the dead load of the box girder tween the top slab and the remainder of the box
superstructure and the prestressing force exerted section, longitudinal bending moments and shears
by the post-tensioning tendons, were discussed in result. The change in overall length of structure
Section 3.3.2 with major emphasis given to mo- may be accommodated by expansion joints, ex-
ment redistribution resulting from creep. After pansion bearing details, and/or flexure of piers.
the structure has been erected and completely The effects of a temperature differential between
post-tensioned, the response of the superstructure top and bottom slabs is illustrated for simple span
to additional superimposed dead load and to live and continuous bridges.
load is considered in the same manner as for any For consideration of longitudinal temperature
continuous bridge. The response of the structure differential effects on a simply supported box
to these loads is elastic. The superimposed dead girder bridge, Fig. 3.23 (a) shows a structure where
load is subject to additional creep deformation, but the top slab temperature is increased At degrees
this deformation does not cause significant re- with respect to the bottom of the section. The
distribution of moments. normal expansion of the top slab is restrained
Consideration of the effects of live load on the by the webs and the remainder of the box sec-
transverse design moments and the use of trans- tion. For purposes of analysis, the deformation
verse post-tensioning in deck slabs is considered in of the box section may be considered to be pre-
Sections 3.4 and 3.5, respectively. vented by exerting external forces P at the centroid
of the top slab level as shown in Fig. 3.23 (a).
Concrete stresses in the top slab will be:

3.3.4 Analysis for the Effects of Temperature f, = EaAt

The effects of temperature on a precast seg- where E = modulus of plasticity of concrete
cr = linear coefficient of thermal expan-
mental bridge superstructure are similar to the
temperature effects on any bridge superstructure sion
in the longitudinal direction. For illustrative pur- Under the loading condition in Fig. 3.23(a) the
poses, calculations evaluating longitudinal tem- stresses in the webs and bottom slab remain zero.
perature effects are presented below. It is noted, If the area of the top slab is A, the required force
however, that the Standard Specifications for High- P will be:
way Bridges of the American Association of State
Highway Officials6) permit stress increases of 25 P=f,A
to 40 percent for loading combinations that in- In Fig. 3.23(b), external equilibrium is restored
clude temperature and shrinkage effects. Since the by removing forces P by superimposing forces P
shrinkage effects are substantially reduced due to which are equal in magnitude but are in opposite
the maturity of the concrete before a continuity directions (P = P). The force P may be considered
connection is made, the permissible stress increase to act at the centroid of the full cross section as
is usually substantially more than the actual tem- shown in Fig. 3.23 (c) by introducing the moment:
perature and shrinkage effects on a precast seg-
mental box girder superstructure. Furthermore, the M = P (c, - e)
longitudinal thermal stresses are primarily of con- The concrete stresses resulting from the equiva-
cern relative to the possibility of crack develop- lent thermal force and moment are shown in Fig.
ment at service load (which is accepted as a matter 3.23 (d):
of course in reinforced concrete structures), and
the longitudinal temperature stresses would have
minimal, if any, effect on the strength of the f,, = -E&At
fc2 = +EaAt ;
The effects of temperature are generally believed
to be more significant in the transverse direction
where temperature stresses may act in combina- fc3 (top fiber) = +EtitA (c,-e) F
tion with the effect of transverse post-tensioning
of deck slabs. These effects are considered in Sec- fc3 (bottom fiber) = - EaAtA (c,-e) p
tions 3.4.7 and 3.5, respectively.


f, = E&t
where E = modulus of elasticity of concrete
(Y = linear coefficient of
thermal expansion



Fig. 3.23 - Analysis for temperature differential between top and bottom slabs

where: 2. The restraint moments M,, shown in Fig. 3.25
+ = tension (cl, required to rejoin the ends of the girders
- = compression over the supports are calculated.
6 = total area of section 3. The total temperature effects on the continuous
I = moment of inertia of section structure are obtained by adding the moments
Applying these equations to the cross section and stresses resulting from the calculations in 1
and section properties in Fig. 3.24 for a top slab and 2 above.
temperature increase of 18 F (10 C), with OL =
5.5 x 1 O6 in./in./OF (9.9 x lo6 m/m/C), and E =
4 x lo6 psi (27.6 x lo3 MPa) [SO00 psi (34.5
MPa) concrete], the stresses become:
f Cl = - 4000 x 5.5 x 1 O6 x 18 = -0.396 ksi I!,

(-2.73 MPa)
f c 2 = +0.396 x 1929.6/3614.4 = + 0.211 ksi
(+ 1.46 MPa)
fc3t= +0.396 x 1929.6 (18.5 - 4) x 18.5/
1142 x lo3 = + 0.180 ksi (+ 1.24 MPa)

f c3b = -0.396 x 1929.6 (18.5 - 4) (48 - Fig. 3.25 - Procedure for analysis of a three span structure
for temperature differential stresses
18.5)/1142 x lo3 =-0.286 ksi
(-1.97 MPa)
Total top fiber stress: -0.396 + 0.211 + 0.180 = The calculation procedure for continuous super-
-0.005 ksi (-0.035 MPa) structures described above in general terms is ap-
Total bottom fiber stress: 0.211 - 0.286 = -0.075 plied in the following to the continuous bridge
ksi (-0.518 MPa) with five equal spans shown in Fig. 3.26 (a).
Proceeding with the first step in the analysis,
the superstructure is considered to be cut over
each support, and a constant equivalent thermal
moment, M, is applied over the full length of all
girders as shown in Fig. 3.26 (b). M causes equal
1 ft. = 0.3048 m
rotations at each girder and over the supports. In
1 in. = 25.4 mm order to rotate the girders back to the same slopes
at the supports, bending moments MI and M2
Fig. 3.24 - Superstructure cross section assumed for tem-
perature differential analysis must be applied resulting in the moment diagram
shown in Fig. 3.26 (c). The total slope at support
2 resulting from the constant temperature moment
From these calculations it is seen that a tempera-
M acting on simple spans l-2 and 2-3 may be cal-
ture increase in the top slab with respect to the re-
culated using moment-area or slope-deflection
mainder of the cross section causes very small
compressive stresses when the superstructure is techniques as:
simply supported. MP MQ MP
In the case of continuous superstructures, re- slope= -+-=-
2EI 2EI El
sistance to the rotation at the supports resulting
from temperature differentials between top and
By the same procedure, the slope due to MI and
bottom slabs generates additional moments and
M2 at support 2 is:
flexural stresses. For the three span structure
shown in Fig. 3.25 (a), the procedure for calcula-
M,n M,Q2M,11 M,a M,Q
tion of temperature moments and stresses is as - -
follows: 3EI +3EI+-=
6EI 3EI + 6EI
1. The continuous superstructure is considered to
be cut over the supports into three simply sup- Setting the slope due to the temperature moment
ported spans as illustrated in Fig. 3.25 lb). equal to the slope resulting from M, and M2
The temperature stresses and rotations at sup- provides the following:
ports can then be calculated for equivalent
thermal force and moment as for simple span 2M,Q M211 MI1
bridges as described above. zl+zEi-=El

shrinkage permitted by the specifications. Further,
A A A A I (a) the stress is less than 50 percent of the modulus
1 2 3 4 5 6 of rupture of the concrete so temperature stresses
would not be expected to cause cracking in the
The moments M, and M2 cause a change in sup-
, port reactions. For the above example the change
in reactions at supports 1, 2, and 3 will be respec-
I (b) tively +24M/1911, -3OM/19Q, and +6M/19P.
For spans 1, 2, and 3, respectively, and for II =
80 ft. (24.4 m) and M = P(c, - e) = E c@t (c, -
e) = 4 x lo6 x 144/1000x 5.5 x 10m6 x 18 x 13.4 x
(1.54 - 0.33) = 924 ft. kips (1253 kN-m). The
changes in support reactions are: +14.6 kips,
1 2 3 4 5 6 -18.2 kips, and +3.6 kips (+65.0, -80.9, +16.0
kN). The weight of the girder is 3.75 kips/ft.
(4 (54.7 kN/m) which provides dead load reactions
at supports 1, 2, and 3 of 119 kips, 339 kips, and
292 kips (525, 1503, 1294 kN). Therefore, the
change in dead load reactions due to the tempera-
Fig. 3.26 - Moments in a five-span continuous superstruc-
ture due to temperature differentials. ture differential is, for this structure, on the order
of 12 percent for the exterior support and 1.2 to
5.4 percent for interior supports.

A similar equation is developed for support 3:

M,Q + -=
5M$? -
6EI 6EI El
Solving these two equations simultaneously for MI
and M2 gives: (4
M, =gM

The total bending moment diagram is, therefore,
the sum of the diagrams in Figs. 3.27 (a) and 3.27
(b), as shown in Fig. 3.27 (c). The stresses due to
this moment diagram and the axial forces due to
+o 196
the temperature differential are calculated as fol-
lows for span 3-4: Fig. 3.27 - Moments and stresses in a five-span continuous
superstructure due to a temperature differential of 18OF
f,, = - 0.396 ksi (-2.73 MPa) (IOOC) between top and bottom slabs
f,, = +0.211 ksi (+1.46 MPa)
feat= +1/19 x 0.180 = +O.OlO ksi (+0.07
f c3b =- l/19 x 0.286 = -0.015 ksi (-0.10 3.3.5 Shear Lag
MPa) Computer Analysis of Shear Lag in Single-
The combined stresses for span 3-4 are shown in
Cell Box Girder Bridges
Fig. 3.27 (d). The compressive stress of 0.07 ksi
(0.52 MPa) calculated for the simple span case, Computer analyses of four single celled box
becomes a tensile stress of 0.195 ksi (1.35 MPa) girder bridges shown in Fig. 3.28 were performed
in the continuous case. While this is a significant to provide data on the magnitude of shear lag
stress, the magnitude is much less than the 25 to effects. The computer model assumed rigid dia-
40 percent stress increase for temperature and phragms at the pier and at abutments. The cross

A 150
:: 150
E 150
0 100 ::

Fig. 3.28 - Superstructure details assumed for computer

analysis of shear lag

sectional dimensions and thicknesses of these four

bridges were intentionally chosen to exaggerate
the shear lag effects. The analyses were performed
using a computer program, MUPDlt8), which is
based on the folded plate method using elasticity
theory. Longitudinal force distributions obtained
from these computer analyses were plotted at vari- Fig. 3.29 - Loading cases for computer analysis of shear
ous sections and compared with forces calculated lag
by elementary beam theory. The ratios between
the peak forces found from the MUPDI computer peak forces occur. Results are given at several sec-
analyses and the forces at the same points found tions along the span which are deemed important.
by elementary beam theory give a measure of the These include sections at midspan; maximum +
effects which are commonly lumped under the M; maximum - M (center support); several sec-
designation shear lag. The forces may be ex- tions close to the center support; and sections
pressed in terms of stresses by dividing by the slab where concentrated live loads act.
or web thicknesses. A careful study of Tables 3.2 and 3.3 reveal a
The analyses were performed for four different number of important facts. In the following, the
loading conditions shown in Fig. 3.29: 1) dead ratio of the longitudinal force N, obtained from
load; 2) prestress; 3) live loadplus impact for max- the MUPDI analysis to that obtained from ele-
imum negative moment; 4) live load plus impact mentary beam theory will be called force ratio
for maximum positive moment. Loadings 5 and 6 for brevity.
in Fig. 3.29 were obtained by superposition of
results for both sides of the bridge in load cases 1. Comparing force ratios of structure A with those
3 and 4, respectively. The combination of four of structure B, they are seen to be very similar.
bridges and four loading conditions required six- The same is true comparing results for structure
teen separate analyses. C with those of structure D. This indicates the
Since the major interest in this investigation was force ratios are essentially independent of varia-
the ratio of the peak longitudinal forces from the tion in depth for a given span (within the span
MUPDI analysis to the forces at the same points depth ratio range between 20 and 30).
found by elementary beam theory, these are sum-
marized in Tables 3.2 and 3.3. Results are given at 2. Comparing force ratios of structures A and
four points on the cross-section a, b, c, d where the B with those of structures C and D, it can be

xb=Bottom Slab

Table 3.2 Summary of results for longitudinal force ratios for structures A and B

Structure A Structure B
L = 150 ft. D = 7.5 ft. L = 150 ft. D = 5.0 ft.
Dist. X Ratio of N, Ratio of N,
from MUPDI 3/Beam Analysis MUPDI 3/Beam Analysis
Load Support
Case Ft. Remark a b C d a b C d

1 56.25 MAX+M 1.04 1.05 1.04 1.06 1.04 1.07 1.04 1.06
75 MIDSPAN 1.04 1.07 1.05 1.07 '1.04 1.07 1.05 1.07
Dead 144 1.07 1.08 1.07 1.08 1.08 1.11 1.09 1.10
Load 146 1.11 1.12 1.11 1.10 1.13 1.16 1.13 1.15
148 1.33 1.32 1.37 1.34 1.35 1.37 1.40 1.39
150 MAX-M 1.41 1.44 1.51 1.50 1.44 1.46 1.52 1.50
2 75 MIDSPAN 0.84 1.02 0.78 1.02 0.92 1.02 0.91 1.02
135 1.02 2.14 1.03 2.11 1.01 0.61 1.02 0.59
Pre- 144 1.07 1.51 1.03 1.55 1.06 3.46 1.06 3.63
Stress 146 1.05 1.49 1.06 1.54 1.05 3.00 1.06 3.00
148 1.07 1.42 1.09 1.45 1.06 2.36 1.08 2.39
150 MAX-M 1.09 1.37 1.12 1.41 1.07 1.96 1.10 2.07
3 75 MIDSPAN 1.20 1.17 1.17 1.17 1.12 1.13 1.11 1.13
87 PT.LOAD 1.33 1.38 1.47 1.40 1.33 1.33 1.36 1.35
2 Lanes 144 1.25 1.28 1.26 1.28 1.23 1.25 1.26 1.25
LL+l 146 1.35 1.32 1.32 1.28 1.32 1.33 1.32 1.31
for 148 1.58 1.49 1.63 1.63 1.63 1.58 1.70 1.61
-M 150 MAX-M 1.75 1.69 1.91 1.80 1.81 1.66 1.91 1.74
4 60 PT.LOAD 1.12 1.21 1.21 1.22 1.15 1.18 1.20 1.18
75 MIDSPAN 1.07 1.07 1.08 1.09 1.04 1.07 1.05 1.07
2 Lanes 144 1.14 1.17 1.08 1.20 1.18 1.26 1.22 1.21
LL+1 146 1.25 1.31 1.23 1.31 1.33 1.32 1.35 1.29
for 148 1.56 1.60 1.60 1.26 1.64 1.60 1.09 1.19
+M 150 MAX-M 1.70 1.65 1.76 1.81 1.75 1.68 1.78 1.77
5 75 MIDSPAN 1.10 1.06 1.09 1.07 1.03 1.07 1.04 1.07
87 PT.LOAD 1.11 1.13 1.20 1.13 1.13 1.13 1.14 1.13
4 Lanes 144 1.09 1.07 1.07 1.08 1.10 1.09 1.10 1.10
LL+I 146 1.15 1.11 1.13 1.08 1.14 1.16 1.14 1.14
for 148 1.29 1.32 1.31 1.35 1.35 1.36 1.39 1.37
-M 150 MAX-M 1.40 1.43 1.50 1.48 1.45 1.44 1.52 1.47
6 60 PT.LOAD 1.00 1.07 1.07 1.08 1.04 1.07 1.08 1.07
75 MIDSPAN 1.03 1.02 1.04 1.03 1.00 1.03 1.00 1.04
4 Lanes 144 1 .oo 1 .oo 0.96 1.03 1.05 1.13 1.08 1.08
LL+I 146 1.06 1.12 1.08 1.13 1.17 1.16 1.18 1.13
for 148 1.28 1.37 1.30 1.37 1.36 1.38 1.39 1.42
+M 150 MAX-M 1.35 1.41 1.41 1.50 1.41 1.46 1.43 1.50

1 ft.= 0.3048m

~b=Bot+om Slab

Table 3.3 Summary of results for longitudinal force ratios for structures C and D

Structure C Structure D
L=300ft.D=15ft. L=3OOft.D=lOft.

! Dist. X Ratio of N, Ratio of N,

from MUPDI B/Beam Analysis MUPDI 3/Beam Analysis
Load Support
Case Ft. Remark a b C d a b C d

1 112.5 MAX+M 0.99 1.01 1.00 1.02 0.99 1.01 0.99 1.01
150 MIDSPAN 0.99 1.02 0.99 1.02 0.99 1.01 1 .oo 1.01
Dead 294 1.02 1.02 1.02 1.02 1.04 1.04 1.04 1.04
Load 296 1.08 1.12 1.11 1.13 1.10 1.13 1.12 1.13
298 MAX-M 1.13 1.19 1.17 1.22 1.14 1.20 1.18 1.21
300 1.13 1.20 1.18 1.23 1.13 1.20 1.17 1.21
2 150 MIDSPAN 0.85 1.00 0.77 1.01 0.96 1.00 0.95 1.00
270 0.99 1.09 0.99 1.11 0.99 1.33 0.99 1.27
Pre- 294 1 .oo 1.08 1.00 1.09 1.00 1.22 1.00 1.23
Stress 296 1.00 1.10 1.01 1.10 1 .oo 1.17 1.01 1.18
298 1.02 1.11 1.03 1.10 1.01 1.12 1.03 1.13
300 MAX-M 1.02 1.11 1.04 1.09 1.01 1.10 1.03 1.09
3 150 MIDSPAN 1 .oo 1.04 1 .oo 1.03 1.00 1.02 1.03 1.00
174 PT.LOAD 1.08 1.10 1.09 1.12 1.05 1.11 1.09 1.09
2 Lanes 294 1.12 1.11 1.14 1.11 1.09 1.12 1.10 1.11
LL+I 296 1.22 1.22 1.24 1.25 1.18 1.17 1.19 1.19
for 298 1.29 1.30 1.34 1.32 1.26 1.22 1.29 1.25
-M 300 MAX-M 1.17 1.31 1.38 1.37 1.25 1.21 1.29 1.23
4 120 PT.LOAD 1.04 1.05 1.11 1.04 1.02 1.04 1.05 1.04
150 MIDSPAN 1 .oo 1 .oo 1 .oo 1.02 1.00 1.02 1.00 1.00
2 Lanes 294 1.08 1.10 1.10 1.08 1.11 1.09 1.09 1.09
LL+I 296 1.15 1.24 1.18 1.27 1.20 1.17 1.20 1.18
for 298 1.21 1.32 1.30 1.36 1.27 1.21 1.27 1.23
+M 300 MAX-M 1.29 1.29 1.39 1.33 1.26 1.20 1.32 1.26
5 150 MIDSPAN 1.00 1.02 0.98 1.02 1.00 1.01 1.01 1.00
174 PT.LOAD 1.00 1.02 1.00 1.04 1.00 1.04 1.01 1.03
4 Lanes 294 1.04 1.02 1.05 1.02 1.02 1.05 1.04 1.04
LL+I 296 1.09 1.12 1.12 1.13 1.08 1.10 1.09 1.11
for 298 1.14 1.19 1.18 1.19 1.13 1.13 1.15 1.16
-M 300 MAX-M 1.07 1.21 1.20 1.23 , 1.11 1.12 1.14 1.14
6 120 PT.LOAD 1.00 1.01 1.04 1.00 1 .oo 1.01 1.01 1.00
150 MIDSPAN 1.00 1.00 0.99 1.01 0.99 1.02 0.99 1.00
4 Lanes 294 1 .oo 1.03 1.03 1.00 1.05 1.03 1.03 1.03
LL+I 296 1.04 1.14 1.07 1.15 1.10 1.10 1.10 1.10
for 298 1.07 1.20 1.15 1.23 1.14 1.13 1.15 1.15
+M 300 MAX-M 1.14 1.19 1.22 1.20 1.13 1.13 1.17 1.16

1 ft. = 0.3048m

seen that the latter are considerably lower indi-
cating that an increase in span results in a de-
crease in force ratio. This is logical since it is
generally recognized that shear lag is inverse-
ly proportional to the span length to plate
width ratio.

3. For a given structure, considering the dominant

forces for any of the loadings, the force ratios
are highest at the center support and drop off
rapidly a few feet away. (Note that nearly all
force ratios are less than 1 .lO at 6 ft (1.8 m) from
the center support.) The dead load longitudinal
force variation across the section of structure f3
at 6 ft. away from the support and at the sup-
port is shown in Figs. 3.30 and 3.31, respec- Fig. 3.31 - Longitudinal force variation in structure 13 at
tively; and similar drawings are presented for center support

structure D in Figs. 3.32 and 3.33. The force

ratios in the midspan positive moment regions
are much smaller. The force ratios are primarily
a function of shear lag, which in turn is a func-
tion of the magnitude of the shear, which is
greatest at the center support. The forces can
be expressed in terms of stresses at the various
points by dividing by the web or slab thickness,

4. For the important dead load case 1, the force

ratios at the center support ranged from 1.41 to
1.52 for structures A and 8, and from 1.13 to
1.23 for structures C and D; while at midspan,
they ranged from 1.04 to 1.07 for structures A 635 t
and 8 and from 0.99 to 1.02 for structures C VERTICAL
and D. NOT TO
5. The force ratios for the dominant stresses under
1 ff = 0.3046 ml
the prestress load case 2 were generally much Force I k,pr
1 k = 4.446 kN
smaller than the force ratios for dead load. For
structures A and B, some high values of force Fig. 3.32 - Longitudinal force variation in structure D
six feet from center support

ratio resulted at points b and d due to the rela-

tively small absolute value of the force at those
points calculated by beam analysis. When com-
pared to small initial values of N, from beam
analysis, the values of N, from MUPDI gave
large force ratios, even though the numerical
force increase was not large. For dominant
forces in the top slab, the force ratios for the
prestress load case ranged from 1.01 to 1.12 for
structures A and B, and from 0.99 to 1.09 for
structures C and D.

6. As seen in the key to load cases shown in Fig.

Fig. 3.30 - Longitudinal force variation structure 6 six feet 3.29, loadings 3 and 4 represent 2 lanes of live
from center support loading, plus impact, placed on one half of the

eters chosen for the computer analyses were in-
tentionally selected to provide an upper bound to
the magnitude of the shear lag effect that could
be expected in a bridge. The shear lag effect from
the prestressing counteracts the shear lag due to
dead load and live load. In this regard, the model

3 985
used in the above computer analysis, which con-
siders the bridge post-tensioned by continuous ten-

1 dons from end to end of the bridge, probably un-

derestimates the actual magnitude of shear lag due
to prestressing in a segmental structure. The use of
) 5.140~

partial length tendons concentrated directly over

the webs in the negative moment area would result
in a higher stress concentration at these points
DIMENSION counteracting shear lag effects even more than
SCALE FOR results from the continuous tendon assumption
used in the MUPDI analysis.
An important finding from the computer analy-
sis was the very limited length of structure in
which significant shear lag effects were found to
Fig. 3.33 - Longitudinal force variation in structure D
at center support
occur. As illustrated in Section, the maxi-
mum effects are 10 percent only 6 ft. (1.8 m)
transverse cross-section. Thus, force ratios for from the center of the support. In most designs,
these loadings reflect not only the effect of this would mean that shear lag effects are only
shear lag, but also of eccentric loading. As significant within the pier section. The computer
mentioned earlier, live load forces are much analyses also show the most significant effects on
smaller than dead load forces. For load cases 3 the short span (150 ft.) (45.7 m) structures with
and 4, force ratios at the center support ranged higher width to span ratios.
from 1.65 to 1.91 for structures A and B and It is felt that the above discussion of the magni-
from 1 .17 to 1.39 for structures C and D. At tude and length of structure affected in conjunc-
the sections where the concentrated live loads tion with the specification requirement of zero
acted (near midspan) the force ratios ranged tensile stress in the top slab under full service
from 1.12 to 1.47 for structures A and B and load, which in itself provides a tensile stress resid-
from 1.02 to 1.12 for structures C and D. ual capacity in the concrete in excess of 500 psi
(3.45 MPa) between service load and the initia-
7. As seen in the key to load cases shown in Fig. tion of cracking, provide sufficient justification for
3.29, loadings 5 and 6 represent 4 lanes of live disregarding explicit consideration of the shear
loading, plus impact, placed symmetrically on lag effects in most practical bridge design projects.
the transverse cross-section. Thus, force ratios For shorter span structures (150 ft.) (45.7 m)
for these loadings are due only to the effect of with wide (40 ft.) (12.2 m) single cell segments,
shear lag. Force ratios at the center support shear lag might be considered in the pier segment
ranged from 1.35 to 1.52 for structures A and by providing some nominal residual compressive
B, and from 1.07 to 1.23 for structures C and D, stress under peak negative moments, or by use of
which are very similar to the force ratios for the computer programs such as MUPDI 3 to evaluate
dead load case. At the sections where the the magnitude of the shear lag effect.
concentrated live loads acted (near midspan) the
force ratios ranged from 1 .OO to 1.20 for struc-
tures A and B and from 1.00 to 1.04 for struc-
tures C and D.
3.3.6 Ultimate Strength Analysis
Precast segmental bridges erected in cantilever
will normally have excess ultimate strength ca- Consideration of Shear Lag in Bridge pacity under full loading conditions because the
Designs negative moment tendons are proportioned to
As noted in Section, the section param- maintain zero tensile stress in the top slab in any
Moments in ft.-kips
1 ft.-k = 1.356 KN-m

Fig. 3.34 - Ultimate moment curves vs. capacity for a three-span segment of a precast segmental bridgeC2)

condition of erection or service loading. Therefore,

the combination of negative moment tendons and
positive moment continuity tendons will usually
provide more than adequate longitudinal moment
capacity to meet the load factor requirements
under loading conditions which produce maximum
moments in the continuous structure.
Under partial loadings which produce maximum
positive moments in one span, a check should be
made to assure that the structure has the negative
moment capacity required in the adjacent un-
loaded spans to withstand any moment reversal
that might occur. Additional tendons may be re-
quired in the top slab at midspan to assure continu-
ity between the top slab negative moment tendons.
This check is important to avoid the possibility
that a negative moment hinge might form in an
unloaded span before the sections in the loaded
span have reached their ultimate capacity.
Fig. 3.34 shows ultimate moment curves for a
three-span segment of a precast segmental bridge.
Curve (a) shows the required moment capacity
under full loading of all spans. Curve (b) shows the
required moment capacity under loading of the
central span only. Note that negative moment ca-
pacity is required at the center of the unloaded
spans under the partial loading. The ultimate
moment capacity of the structure is indicated be- Fig. 3.35 - Initial loading and reaction assumption for
tween the shaded areas. transverse analysis



t 7

Xm,+R2) X(R1+R2) X(R, -R2) h(R2-R2)

=P =P

Fig. 3.36 - Non-symmetrical loading and reaction assumption for transverse analysis

3.4 Transverse Analysis 0 t
3.4.1 General
Transverse moments, shear and axial forces in
box girders are analyzed taking into consideration A
the longitudinal geometry, torsional properties,

and transverse geometry of the box girder. Inter-
mediate diaphrams are generally not required, and
the design method presented in the following sec- -47
tions does not include consideration of them.

3.4.2 Principles
In Fig. 3.35 (a), loading 2P per unit length is
assumed to be constant over the length of a simply
supported box girder with section ABCD. Con-
sider the corners of the box girder supported as
shown in Fig. 3.35 (b). The analysis reduces then
to simple case of a frame. This analysis is carried
out and transverse moments, shear and axial
forces are calculated. Also the support reactions
R,, Ra, Ra, and R4 are evaluated.
Non-symmetrical loading as indicated in Fig.
3.36 (a) would cause bearing forces or support
reactions as shown : fb)

R, > R, and R4 = -Rg Fig. 3.37 - Transverse analysis for symmetrical loading

The fact that previously assumed supports are not

present must be accounted for by subsequent
loading of the box girder by forces opposite to the box. The direction of T, is the same as that of
R R,, Ra, and R4. These forces are shown in the load P. The directions of T, and T3 are as
F/g: 3.36 (b). For a subsequent analysis of the box shown, since they must be at right angles to the lon-
girder by forces RI , R2, RB, and R,,, these loads gitudinal shear forces in top slab and bottom slab
are rearranged in symmetrical and antisymmetrical caused by the rate of change of longitudinal
components as shown in Fig. 3.36 (c). bending moments. Over a length L the rate of
change of the shear forces in top slab, web, and
bottom slab is T, , T2, and T3, respectively.
Obviously T2 equals the vertical load P on L.
3.4.3 Symmetrical Box Girder Loading
However, in the horizontal direction equilibrium
Symmetrical loading of the box girder as shown can only be obtained by addition of transverse
in Fig. 3.37 (a) causes longitudinal bending and axial forces in top slab and bottom slab as shown.
shear that has been accounted for in the calcula- These axial forces are equal to rates of change of
tion of longitudinal prestressing. Transverse mo- shear forces T, and TB, being T, and Ts as is
ments are, because of the placement of the load shown in Fig. 3.37 (b). T, and TB are obtained
at the webs, secondary in nature and usually neg- from the rates of change of the shear stress which
ligible. Not negligible, however, are the transverse may be calculated as illustrated in Fig. 3.38. The
axial forces which are: tension in webs, tension shear stress diagram over the bottom slab, maxi-
in bottom slab, and compression in the top slab. mum value 7, is shown in Fig. 3.38 (b). The value
Top and bottom slab axial forces are a conse- of T may be calculated as:
quence of the rate of change of longitudinal shear
as is shown in the following. The box girder shown Pbdz Pbz
in Fig. 3.37 (b) is cut through the longitudinal r=dl=-I
centerline. Support and loading P are indicated.
Shear forces T,, TZ, and T3 occur in top slab, where I is the moment of intertia of the half sec-
web and bottom slab, respectively, in a section of tion shown. From the distribution of the shear




Fig. 3.38 - Transverse analysis for symmetrical loading

stress over the top and bottom slab as shown in

Fig. 3.38 (b):
Fig. 3.39 - Antisymmetrical box girder loading effects
Tn =T n - dTb
1 3

Substituting the value of 7 from above: Since the box girder is relatively stiff in the
transverse direction, the response of the structure
Tr, = T3f - Pb2dz
21 to upward and downward forces -P and +P is to
balance transversely. This results in transverse
The transverse axial force diagram caused by cen- moments M, and horizontal and vertical shear
tral loading of 2P is as indicated in Fig. 3.38 (4. forces S,., and S, as shown in Fig. 3.39 (b). There
The shortening or elongation of the individual are also horizontal and vertical displacements h
members due to axial loads sets up transverse mo- and v.
ments which can usually be neglected. These displacements h and v cannot occur with-
out the resistance of the top slab and bottom slab
(h) and webs (v) in the longitudinal direction.
Deflection v of web AD will cause longitudinal
3.4.4 Antisymmetrical Loading
bending stresses, compression -T at D and tension
Antisymmetrical loading of the box girder as +T at A. Because of compatability of strains,
shown in Fig. 3.39 (a) affects the structure in the equal stresses +T occur in the top slab CD due to
following ways: horizontal displacement h as shown in Fig. 3.39
1. In the transverse direction, transverse bending (c). This illustrates that, as a result of transverse
and torsional shear are induced. deformations, bending moments and shear forces
2. In the longitudinal direction, moments and are set up in the longitudinal direction of the box
shear forces are set up acting in the planes girder. The longitudinal forces act in the planes of
of the bottom slab and top slab. the slabs and webs and, as a result, part of the ex-

3.4.5 Evaluation of the Contributions of Trans-

,rJ verse Bending, Longitudinal Bending and

Torsion to Resistance of Antisymmetrical
The top half of a box girder section with unit
length L is shown in Fig. 3.41 (a) with the hori-
zontal forces acting on it. Horizontal equilibrium
leads to the expression:
2S,, +T,,=t,,
The left half of the box section with unit length
L illustrated in Fig. 3.41 (b) shows the vertical
forces acting on it. Vertical equilibrium leads to:
2S, + T, + t, = .P

1 H A complete box section with length L is shown in

Fig. 3.42 (a) with the forces acting on it. Moment
Fig. 3.40 - Horizontal forces and shear forces acting on equilibrium of the forces in Fig. 3.42 (a) leads to
box girder the expression:
ternal load P, say T,, is carried by the webs direct-
ly to the supports. At the same time, shear forces
T,., are acting in the top and bottom slab. The ratio
of T, and T,, follows directly from the geometry
of the box section as a consequence of equal
stresses T at the corners.
After having determined the basic consequences
of transverse deformations, the box girder may
be cut at the horizontal neutral axis. Fig. 3.40 (a)
shows the top half to the box girder and the hori-
zontal forces, discussed up to this point, acting on
it. The lack of horizontal equilibrium is restored
by the torsional shear forces. A torsional moment,
uniformly applied over the length of the box
girder, by loads +P and -P per unit length, changes
at the rate of MIt per unit length; where M, =
PH. Assuming the concrete thickness d to be small
with respect to box girder dimensions V and H,
the shear forces t, are constant per unit length of
web or slab. Torsional shear forces, therefore, are
in the webs t, = t,,V, and in top and bottom
slab t,, = t,H as indicated in Fig. 3.40 (b). The
T t
value of the various torsional shear forces may be

calculated as follows:
to zTd=-
t, = t,V
fh = t, H
where t, = torsional unit shear force
= torsional shear stress
M: = torsional moment per unit length of
box girder
t, and t,, = rate of change of torsional shear
force in the web and slab, respec- Fig. 3.41 - Equilibrium of horizontal and vertical forces
under antisymmetrical loading

The relations of S, and S,, and t, and tIh fol-
low from the geometry of the box girder:

--Z S;=M,

fh t,
-z-z to

The above equations permit solution for all un-


3.4.6 Example Transverse Analysis Calculations

The box girder section shown in Fig. 3.43 has
a simply supported span of 40.00 m (131.2 ft.)
length. The moment of inertia of the full section is
2.76m4 (319.5 ft.4). A linear load of 10 t/m
(6.8 k/ft.) is present over the full length of the box
girder. Web and slab thicknesses are 0.3 m (1.0 ft).
Consider the box supported at four corners as
shown in Fig. 3.44.

IO t/m

Fig. 3.42 - Box section equilibrium under antisymmetrical

I 15
ti t 0,3
Fig. 3.42 (b) shows a box section with unit length 0
L indicating displacements h and v and the forces zoo
resisting these displacements. This leads to: ? 0 . 3
1 I -d

1ml = 3.28 ft.

1 t/m = 0 678 km

INote: 10 tonnes 0
force in the c.g.1.
mctnc wonl =
170 =p 9.8 x ,w Nmtcml
f!q 1 an tbs SI sysom. Thir
looom analyar II I the
_ 570 c.g.*. syrtsml

d = web thickness Fig. 3.43 - Example transverse analysis box girder section

d, = slab thickness
P = rotation of corner
L = unit length
L = span length V tc

In the longitudinal direction:

EdV3 Ed, H3
v=crT,L4 12 and h = OLT,, L4 -
/ I 12

The rate of change of longitudinal shear forces

T,, and T, are considered external uniform dis-
tributed load. The coefficient (Y equals384
L for the Fig. 3.44 - Support assumptions for example transverse
deflection of a simply supported beam. analysis
1.72 1 I

- 3 . 4 2
4 +

-b. L - ' 1.15

(a) Moments, t-m/m (b) Axial Forces, t/m

1 t/m = 0.678 klft.

1 t-m/m = 2.2 k-ft./ft. . b

892 (cl Support Forces, t/m 1.08
Fig. 3.45 - Moments, axial forces, and support forces for example transverse analysis

FE l-T-[ + p-j +
1m2 1 ID8 I5 15 L LB2

Fig. 3.46 - Adjustment of support forces for example transverse analysis

Bending moments, axial forces and support From moment equilibrium:
forces are obtained from a conventional moment - 3 . 9 2 x 5.7 + T, x 5.7 - T,, x 1.7 + t, x 5.7 +
distribution calculation. The resulting diagrams are fh x 1.7 = 0
presented in Figs. 3.45 (a), 3.45 (b), and 3.45 (c). I
The supports not actually present are taken into making use of fy, = -
account by the loads of Fig. 3.46 (a), which in th 5.7
turn are subdivided into the loads of Figs. 3.46 fh = 3.35 t,

(b), 3.46 (cl, and 3.46 (d).

The central loading of Fig. 3.46 (b) causes and as a consequence of equal longitudinal stresses
1 at the corner of the box girder:
longitudinal bending only. Transverse moments are
negligible. However, axial forces are developed
0.85T, 2.85Th
which are shown in Fig. 3.47. The transverse axial
force is evaluated as: 1 .73 = 5.73
Th= 11.24T,
5 x 0.3 x 2.B52 x 0.85
substituting in the moment equilibrium equation:
l/2 = 3.75 t/m
( l/2 x 2.76 > -22.34 - 13.41 T, + 11.4 t, = 0
(2.54 k/ft.) Solving the above equations:
The loading of Fig. 3.46 (c) causes transverse and = 0.026 t/m (18 IbJft.1
longitudinal bending and torsion. In accordance = 0.30 t/m (203 IbJft.1
with Section 3.4.5: = 0.95 t/m (644 Ib./ft.)
t, = 1.99 t/m (1350 Ib./ft.)
th = 6.67 t/m (4520 Ib./ft.)
Corner moments M, = 0.95 x 2.85 = 2.71 t-m/m
(5.96 k-ft./ft.)
Resulting bending moment and axial force dia-
grams are presented in Figs. 3.48 (a) and 3.48 (b).
Axial forces are obtained from:
1 for displacements h.and v:
2 x 2.71
top slab: = 3.18 t/m (2160 Ib./ft.)
check: 3.18 = t,, - Th = 6.67 - 0.3 = 3 18

5 ~~4Oyz$d] web. 2 x 2.71

v= 384 5.7
= 0.95 t/m (644 Ib./ft.) at top

Substituting these values of h and v in the above 3.92 - 0.95 = 2.97 t/m (2010 Ib./ft.)
equation: at bottom

check: 2.97 - 0.95 = 2.02 = t, + T,
S x 1.72
= 1.99 + 0.03
o.33 =

404 -
12 x T, 12 x T,
o.3x5.73x1 .7 +0.3x5.7x1.73 1
S = 177.17[$ +G]

= 5.45 T, + 61.3 T,
1 t/m = 0.678 k/ft.
From vertical equilibrium:
Fig. 3.47 - Transverse axial forces for example transverse
3.92 = t, + T, + 2 S, analysis, t/m


(a) Moment, t-m/m

(a) Moments, t-m/m


1 tlm = 0.678 k/ft.

1 t-m/m = 2 2 k-ft./ft. (b) Axial

Fig. 3.48 - Bending moment and axial force diagrams re-

sulting from example transverse analysis loading case of
Fig. 3.46 (c)

(b) Awal Forces. t/m

Fig. 3.50 - Final moment and axial force diagrams result-

ing from example transverse analysis

1. Transverse bending moments are influenced

1 t/m = 0.678 k/ft. 6.48 considerably by torsion to the extent that maxi-
1 t-m/m = 2.2 k-ft./ft. (a) Moment, t-m/m mum moments occur at places other than expected
in the case of a regular frame [compare Figs.
3.45 (a) and 3.50 (a).]
2. Transverse axial tensile forces cannot be neglect-
ed since they increase the required amount of rein-
forcement. These forces are particularly signifi-
- OJ7 an cant in the bottom slab.
(b) Axial Force, t/m 3. Axial compressive forces reduce the required
Fig. 3.49 - Bending moment and axial force diagrams re- amount of reinforcement. This is particularly
sulting from example transverse analysis loading case of significant in the webs at the connection with the
Fig. 3.46 (d)
top slab. At these points, compressive forces are
high and occur simultaneously with high moments.
A solution for loading case d in Fig. 3.46 is ob-
tained in a similar manner. Moment and axial force 4. Corner moments as given in Figs. 3.48 (a) and
diagrams are presented in Figs. 3.49 (a) and 3.49 3.49 (a) caused by loading indicated in Figs.
(b), respectively. 3.46 (c) and 3.46 (d) may be approximately
The final results of the investigation shown in calculated as Pe/8 where P is the vertical or hori-
Figs. 3.50 (a) and 3.50 (b) are obtained by addi- zontal force, and e is the width and depth of the
box respectively.
tion of the results given in Figs. 3.45, 3.47, 3.48,
and 3.49. 5. When span/depth ratios are constant, longitud-
Conclusions from the example transverse analy- inal bending has very little influence on trans-
sis calculations are as follows: verse moment distribution.


3.4.7 Transverse Temperature Effects 3.5 Analysis and Transverse Post-TensioMg

Tensile stresses in the box girder cross section of Deck Slabs
may be generated by the following temperature
effects: 3.5.1 Live Load Plus impact Analysis
Analysis for live load plus impact moments
1. At sections near the supports, the relatively and shears in deck slabs of precast segmental
thin top slab may cool much more rapidly bridges requires consideration of the effect of con-
than the thicker bottom slab. This will centrated loads on variable depth plates which
cause tensile stresses around the exterior of are integral parts of a tubular frame. Design of such
the cross section. slabs is accomplished by use of charts of influence
surfaces for variable depth plates.gv lo)
2. With strong and prolonged sun radiation on For cantilever slab moments, the use of the in-
the bridge surface, the air in the interior of a fluence charts simply requires computing the sum-
hollow box girder may become heated to mation of the ordinates of the wheel loads plotted
over 100 F (38 C). When the outer air on the influence surface and multiplying by the
cools during the night, the temperature dif- magnitude of the wheel load to obtain the mo-
ference between the interior and outer air ment per unit length for the point under considera-
produces transverse flexural moments in the tion. For interior span positive moments, the
webs and slabs which cause tensile stresses influence surfaces are used to determine fixed end
around the exterior of the cross section. moments for various positions of the load. The
Fig. 3.51 shows the moments and stresses fixed end moments are then used in a frame
in a single cell box girder at midspan and at analysis to determine the effects of live load on
the support for a temperature difference of the frame.
27O F (15 C) between the air inside and More extensive discussion of calculation of live
outside the box.(12) load moments using influence surfaces and a de-
sign example for a transversely post-tensioned deck
3. Thick concrete elements exposed to intense are presented in Post-Tensioned Box Girder
sun radiation are subject to substantial ten- Bridge Manual published by the Post-Tensioning
sile stresses when the exterior surfaces cool I nstitute.(lg)
due to the lag in response of the interior con- The analysis of two or more box girders con-
crete to the temperature change. nected by a common slab requires consideration
of the flexural and torsional restraints at supports
The significant tensile stresses shown for the as well as the flexural and torsional response of
bridge section in Fig. 3.51 illustrate the desir- the box girders and the connection slabs. This
ability of avoiding the use of thick concrete webs analysis may be accomplished by an extension of
and slabs which are highly rigid with respect to the analytical procedures described in Section
transverse flexure. The flexural stiffness is, of 3.4. A detailed procedure to accomplish this analy-
course, a function of both the thickness and length sis has been published. Alternatively, the analy-
of the structural element. This factor becomes sis of single or multiple cell box girder sections
more significant when the transverse temperature may be made by use of one of the available com-
stresses are combined with the transverse tensile puter programs.
stresses in webs that result from the transverse
post-tensioning of deck slabs as discussed in Sec-
tion 3.5. The -joint between the web and bottom 3.5.2 Transverse Post-Tensioning of Deck Slabs
slab near supports is a point where the combined Transverse post-tensioning of deck slabs offers
tensile stresses may become high, and, at this point, the following advantages in comparison with non-
it is particularly important that any cracks which prestressed transverse reinforcement:
may result from the various effects be anticipated 1 The deck slab thickness is reduced with re-
in the design. These tensile stresses and potential su lting reductions in concrete quantities
cracks may be accommodated by use of a conser- and dead load moments and shears.
vative design of nonprestressed shear reinforce-
ment, or by the use of prestressed stirrups. The 2 Longer slab spans may be achieved which
latter option has the advantage of providing a permits reduction in the number of webs
much higher degree of assurance against cracking required in wide structures. This reduces
in the webs. forming costs and concrete quantities.

Mid span Support
cross-section cross-section

- 7.78

II1 2.82 - 41.51

1 ft. = 0.3048 m
1 in. = 25.4 mm
1 ft.-k/ft. = 0.45 t-m/m
Transverse bending moment in ft-kips/ft
1 psi = 0.0069 MPa
for temperature difference T, - Ti = 27F (15C)

Corresponding Point 1 1 2 2
Edge Stresses Span 5564 *220 +485 + 68
(psi) Support ?483 f 84 +446 +446

Fig. 3.51 - Transverse moments and stresses due to a temperature difference of 27OF between the outer and inner surfaces of a
box girder()

3 A high level of assurance is provided against t o see if the combined prestressed and nonpre-
the development of longitudinal cracking in stressed reinforcement in the transverse direction
the deck slab. This provides a more durable is sufficient to meet the load factor requirements.
deck with minimal potential maintenance If not, the amount of either the prestressed or
costs. nonprestressed reinforcement should be increased
as required.
4 In the area of top slab anchorages, such as
Tendon profiles for transverse deck slab rein-
illustrated in Fig. 2.8, transverse compres-
forcement may vary depending on the type of ten-
sion is helpful in counteracting tensile stresses
don material and on other design and construction
in the slab which result from concentrated
requirements. Tendon geometry used for the Kish-
anchorage forces.
waukee River Bridge is shown in Figs. 3.52 and
5 For wide segments, the use of transverse 3.53. Fig. 3.52 illustrates the use of bar tendons,
post-tensioning in the deck slabs usually and Fig. 3.53 the geometry proposed in the design
results in reduced overall structure cost. drawings. The placement of the bar tendons in
the center of the slab was selected in this case to
Transversely post-tensioned deck slabs also nor- provide a means of support for the longitudinal
mally have transverse and longitudinal nonpre- tendons. While this increased the required amount
stressed reinforcement in the top and bottom of of transverse post-tensioning by about 30 percent,
the slab. This contributes to the flexural capacity this increase in cost was offset by reduction in
of the slab in ultimate strength calculations and labor requirements for placement of the longitud-
provides the necessary flexural capacity to permit inal tendons. The tendon profile shown in Fig.
removal of the section from the forms and handl- 3.53 was selected to more closely approximate
ing prior to stressing of the transverse tendons. the moment diagram.
The transverse post-tensioning is proportioned to One additional factor that must be considered
limit the tensile stresses in the deck slabs to the when transverse post-tensioning of the deck slab
design values. Subsequently, the slab is checked is used is the effect of the transverse elastic short-

E Sor Qirder

Fig. 3.52 - Transverse and longitudinal post-tensioning, Kishwaukee Bridge, Illinois

1 ft. = 0.3048 m
1 in. = 25.4 mm

Fig. 3.53 - Transverse tendon geometry from design drawings, Kishwaukee River Bridge, Illinois

ening of the deck slab in generating Wditional 3.6 Analysis and Correction of Deformations
transverse moments and stresses. The lateral bend-
ing of the webs sets up fixed end moments that
3.6.1 General
must be distributed throughout the transverse
frame. An analysis of this effect on a cross sec- The development of segmental construction has
tion of a post-tensioned box girder bridge cast-in- made it economical to build slender concrete
place on falsework is shown in Fig. 3.54. bridges with long spans. As a result, the magnitude
For wide sections, such as this, relatively high of the deformations and deflections may be in-
tensile stresses are generated by the slab short- creased to such an extent that they require more
ening. Even in narrower sections that might be attention and usually need adjustment during con-
expected in a precast segmental bridge, this effect struction. The amount of deformation is further
may be substantial and should be considered in increased by erection of a structure in free can-
the design. These stresses become highest near piers tilever. The deformations require correction of the
where th& transverse frame elements are thickest. geometry of a structure during segment fabrica-
A design check should be made to assure that tion which can only be based on an effective pre-
stresses resulting from transverse post-tensioning diction of the deformations.
of the deck slab, in conjunction with the transverse Erection of a typical span in a multispan bridge
temperature stresses discussed in Section 3.4.7, usually starts at a pier by placing segments alter-
dre not sufficiently high to cause cracking at nately on both sides in free cantilever until mid-
the bottom exterior corner as illustrated in Fig. span is reached. The newly erected cantilever is
3 . 55. () The magnitude of these stresses and the then connected to the completed part of the struc-
potential for crack development are minimized ture by casting the midspan splice. This procedure
by use bf the thinnest possible concrete sections is repeated for each additional span, however, with
consistent with strength requirements and with different resulting deformations since these depend
segment design recommendations presented in on the statical system in which the addition takes
Chapter 2. place. Obviously, this statical system changes

13.12 13.12' 13.12' 13.12' 13.12'
V, = 67.197 lb//f V, = 67,197 lb//f

.-. -. ~j+&~--

- - - -
2' I
t I I
Midspan cross-section

1 ft. = 0.3048 m Corresponding Pomt 1 1" 2' 2"

1 I".= 25.4 mm
edge stresses Midspan t295 ?130 t309 t 31
1 ft.-k/ft = 0.45 t-m/m
1 ps, = 0.0069 MPa (psi1 Support +488 t 54 ?766 +340

Fig. 3.54 - Transverse bending moments due to normal force component of post-tensioning in deck slab2)

Bridge axis+ ing (cantilever, continuity, and losses), and dead

load. As mentioned above, total deformations
are obtained by summing up the contributions of
each intermediate phase of construction. Also,
the changes occurring after completion of the
structure are added. The various phases are:
Phase A: condition of free cantilever
Phase B,B: intermediate phases (connection of
a new cantilever to completed
Fig. 3.55 - Potential cracking due to temperature stresses structure)
and elastic shortening of slab due to transverse post-tension- completed structure
ing(12) Phase C:
Deformations are either hand or computer calcu-
lated. In the latter case, the influence of time de-
pendent properties such as modulus of elasticity
numerous times in the construction process. The
of concrete, influence of creep, shrinkage, and
analysis of deformations therefore implies the sum-
relaxation losses on tendon forces, and differences
mation of deformations in all successive inter-
in the creep factors of individual segments can be
mediate phases. This is a tedious and complex,
integrally entered into the calculations. In the case
but, nonetheless unavoidable, aspect of the design
of hand calculations, this is not feasible and sim-
calculations. plifications are needed.
The following sections are based on the assump-
tion of hand calculated deformations. It is com-
3.6.2 Analysis mon practice to consider deformations due to
Important contributions to deformations, elastic bending moments only, since those by axial and
as well as creep, are made by self weight, prestress- shear forces are usually negligible.

63 Phase A - Free Cantilever This time is also needed for the determination of
Loading conditions are: the contribution of creep to the deformations.
Steel relaxation varies significantly for different
1. Elastic deflection due to self weight.
post-tensioning materials (wire, strand or bar),
2. Elastic deflection due to initial cantilever
and low relaxation materials are available (relax-
ation losses for low relaxation strand are in the
3. Creep deformation of 1 and 2 for the dura-
range of 25 percent of the values in Fig. 3.56 (a)).
tion of this phase.
For this reason, use of relaxation curves for the
The deflected shape of the completed cantilever is specific material to be used is recommended.
easily calculated. Elastic deflections due to self Although creep starts from erection of the first
weight and prestress are calculated assuming a segment onwards, without the use of a computer
Youngs modulus of elasticity: it is not practical to calculate total creep deforma-
tion as the sum of the effects of each successive
E = 33 w32 fi
step. A reasonable approximation is obtained when
where f, = cylinder strength of concrete in the completed cantilever is considered to creep
psi at the time of erection during a time interval which starts when the can-
w = unit weight of concrete in lb. per cu. ft. tilever is halfway complete and ends when a con-
E may be assumed constant for precast segments nection with the completed structure is made.
after the segment age reaches 28 days. This time interval is different for each canti-
The prestressing force used for the calculation is lever arm as illustrated by Fig. 3.57.
the total of initial tendon forces reduced only Creep deformations are obtained by multiplica-
by friction losses and part of the steel relaxation tion of the elastic deformations by a creep factor.
loss. The relaxation loss is evaluated from a relaxa- The creep factor used here is
tion-time curve based on test results by the steel
supplier, or from typical relaxation curves such as 4 '2'1 = @d, fld(tZpt, ) + @f,
[~f,,-~ftJ for part a
given in Figs. 3.56 (a) and 3.56 (b), and an esti-
mate of the time the cantilever is in phase A. 4 t4t2 = #d, (3d(t4-2) + @fee[oft4 -aft,] for part b

I /iiiitiiiiii Iiiiiiiiiiii i iiiiiiiiiii i iiiiiwi

I ,I,
f,, = 700

30 50

Fig. 3.56 (a) - Relaxation loss curves for stress-relieved strand



Fig. 3.56 (b) - Relaxation loss curves for 150 K 1% in. diameter bars (1030 MPa - 32 mm@)

rckx.ed at t2 -closed at t4

h The required calculations are simplified if car-

A ried out for a simply supported span. The effect
ri time scale t
of fixity may be treated separately and may then
be added to the simple span calculations. Fig.
to t1 t2 t3 t4 3.58 illustrates this procedure. Span BC is assumed
part a of cantilever I creeps during interval t2-t,
part b during t4-t2
to be part of a structure with a number of equal
spans. After application of continuity prestress,
Fig. 3.57 - Effect of construction sequence on creep time
interval as a free cantilever
this span is loaded with the concentrated load V,
the weight of the midspan splice, and the moment-
The sequence of erection and the time schedule area of the continuity prestress.
must be known or assumed prior to the start of Both these loads cause secondary moments,
the deformation calculations. Fig. 3.58 (b), which affect the deformations of
span BC and all preceding spans. The total elastic
deformation is obtained by summation of the three Intermediate Phases B,B bending moment diagrams shown in Fig. 3.58 (c)
of the simply supported span BC. Creep deforma-
Deformations in this phase are those from: tions are found by multiplying the elastic values by
1. The weight of cast-in-place splices. a creep factor. The creep influence is limited to
2. Continuity prestress in the span considered. that part of the creep taking place in the period
3. Continuity prestress in the adjacent spans. between closing of the splices in spans BC and CD
4. Creep deformation resulting from 2 and 3 respectively. The remainder of the creep deforma-
I- above. tion is assumed to occur in the final continuous

2. Elastic and creep deformation by prestress
3. Creep deformations by self weight, cantilever
prestress and continuity prestress.
Determination of the elastic deformation by
lb) superimposed dead load (weight of topping, curbs,
railings, etc.) needs no further comment. The
creep deformation is obtained by multiplication
of the elastic value by @ct, -t,), with t, being the
time of application of the dead load.
For the amount of deformation by prestress
losses, a simplification is made. The total amount
of the losses caused by creep, shrinkage and re-
laxation is reduced by the part of the relaxation
(dJ loss deducted in phase A. All other losses are con-
I sidered to take place in the final system. This
negative prestressing force F again causes elastic
and creep deformation and is written, therefore, in
a simplified form as:

where F i = initial prestressing force

Ff = final prestressing force
A B c D t, = time of completion of structure

Fig. 3.58 - Superstructure deformations in phase 8 The determination of the creep deformations
by self weight and prestressing in the completed
structure is based on the solution presented in
Section 3.3. Evaluation of the creep deformations
in this phase can be restricted to those occurring
system. With reference to Fig. 3.57, the creep fac- in the final system. The creep effects of the inter-
tor to be used is: mediate phases B,B are then neglected; the
error is small, since the most important contri-
# (t4t2) = @d, fld(t,yt2) + @f, [oft,- fq
bution, the creep of the forward cantilever arm,
has been taken into account. After a few spans
The total deformation is shown in Fig. 3.58 (d)
have been completed, the statical system during
indicating a rotation over pier C, bringing down the construction closely resembles the final com-
forward cantilever arm. Also, this rotation in- pleted structure.
creases by creep while the structure is in phase B.
Addition of a new span, Fig. 3.58 (e), again causes
secondary moments which will affect span BC as
well, Fig. 3.58 (f), and so will the connection of 3.6.3 Alignment
each successive span. For this reason, it is easier
The need for correction of deformations should
to calculate the deformation due to secondary
be investigated for all precast segmental bridges.
moments after summation of all contributing
The use of match-cast joints makes alignment cor-
moment diagrams. The rotation of each forward
rections during construction awkward and un-
arm, however, must be determined just before
desirable. In the casting yard, corrections are al-
closure of the next span.
ways minor and are easily accommodated by the
casting equipment. Adjustments of alignment can
be made during construction by use of stainless
steel shims in the joints. The following procedure Phase C - Final Continuous System
of alignment correction for a bridge with several
Deformations in this phase consist of: equal spans illustrates the principles. Corrections
1. Elastic and creep deformation by superim- consist of those resulting from deformation, ro-
posed dead load. tations, and superimposed curvatures.

66 Correction of Deformations
The correction curve of each cantilever arm
equals the deformation curve but with opposite
sign. Typical deflection curves are shown in Fig.
3.59. The theoretical curve is approached by ,.
straight lines one or more segments long. The
difference between a curve and approximating
straight lines obtained in this way is not visible
provided the angular changes are kept below 0.001 Fig. 3.60 - Cantilever rotations due to continuity post-
radians as shown.
As indicated in Section, the deforma-
tion of the forward cantilever arm will be differ-
ent from the backward arm, because of different
creep behavior and the rotation caused by con-
tinuity prestress. Although it is possible to make
additional corrections during casting for forward
and backward cantilever arms, it proves simpler
to make such corrections by counter rotations.

Fig. 3.61 - Correction for rotations due to continuity


correction /

Fig. 3.59 - Typical deflection curves Correction of Rotation

Fig. 3.62 - Correction for roadway curvature
Due to continuity prestressing in the end span,
the forward cantilever arm rotates over an angle CY
as shown in Fig. 3.60. A similar rotation p occurs
in the subsequent spans. Starting erection of the
first cantilever with a counter rotation of ~1 - %(3
used are sufficiently adaptable, any shape of
would bring the forward cantilever arm to a slope
bridge, including vertical curves, horizontal curves,
of 43 after stressing of the end span continuity
superelevation, etc., can be achieved by superim-
tendons. The subsequent span then automatically
posing the difference between the desired curva-
starts with a counter curve of %p as well, and this
ture and the straight axis (the shaded area of Fig.
situation repeats itself until completion of the
3.62) on the corrections previously described.
structure as shown in Fig. 3.61.
The continuity prestress obviously affects not
only the forward cantilever arm but also the re- Example Alignment Calculations
mainder of the completed part of the structure.
However, the resulting up and downward curves Part of a bridge is shown in Fig. 3.63. The de-
from this source are usually part of the deforma- flection X of span LM is the value calculated for
tion corrections made in the form. This also ap- the sum of elastic and creep deformations caused
plies to the angle changes occurring at the splices. by the continuity prestress of all adjacent spans.
The camber Y of span MO and the rotation of
the forward cantilever arm OP are those calculated
for elastic and creep deformations caused by
continuity prestress of span MO only. It is clear Correction of Superimposed Curvature that corrections for span MO will be based on a
The desired alignment of most bridges differs reduced camber Y - X. After erection, the de-
from a straight line. Provided the casting forms flected shape of the cantilever arms NOP (support

\ Fig. 3.66 - Summation of geometry corrections
Fig. 3.63 - Example alignment calculations

The discontinuity at 0 does not exist since MO

is the final bridge axis after step 5, whereas OP
shows the situation after step 4 only. Notes on Alignment Calculations

1. With the procedure illustrated in Figs. 3.63
Fig. 3.64 - Deflected shape of cantilever after erection through 3.66, only the deformations during con-
struction are covered. After completion, addi-
tional deformations will occur. These can be
treated, if found to be of considerable magni-
tude, similar to corrections of superimposed
curvature as described in Section

2. The corrections described are based on deforma-

tion calculations. It is essential to check the re-
012 sults of such calculations by field measurements.
Such comparative measurements should always
Fig. 3.65 - Correction of deflected cantilever geometry to take place in the morning at the same hour in
obtain a straight axis order to minimize the considerable effect of
movements due to temperature variations.

= 0) will be as indicated in Fig. 3.64. The correc-

tion to obtain a straight axis, shown shaded in Fig. 3.7 Computer Programs
3.65, is arrived at by:
1. Drawing the deformation line a due to contin- 3.7.1 General
uity prestress in all spans (Y-X in Fig. 3.63).
2. Reducing curve a by the free cantilever de- In some cases, hand calculations may be suffi-
flection, resulting in curve b. ciently accurate for the final design of a precast
3. Rotation of the axis by an angle %p. segmental bridge. However, for more complex
superstructures, the use of a computer program to
Verification of this result is illustrated in Fig. 3.66: assist in the analysis becomes most helpful. Fur-
1. The correction is introduced with opposite ther, the calculation of deflections becomes
sign in curve d. very cumbersome by hand unless substantial ap-
2. The free cantilever deflection is superimposed proximations are introduced, and a computer pro-
in curve e. gram is an invaluable aid in providing a more pre-
3. The rotation of %(3 is added in curve f. In this cise estimate of time-related deflections. The
situation the midspan splice is cast. sources listed below have programs developed or
4. Continuity prestress is added resulting in adapted specifically for use in design of precast
curve g. segmental bridges. Additional programs undoubt-
5. Deflection by continuity prestress of all ad- edly exist that could be used more or less directly
jacent spans results in the final geometry h. to analyze precast segmental bridges.

3.7.2 Sources of Computer Programs
Detailed information on computer program
services may be obtained from the following:

Dyckerhoff & Widmann, Inc.

529 Fifth Avenue
New York, New York 10017
(2 12) 953-0700
Engineering Computer Corporation
P.O. Box 22526
Sacramento, California 95831
i. (916) 922-9316
Europe Etudes BC Program
The Preston Corporation
2426 Cee Gee
San Antonio, Texas 78217
(512) 828-6264
Center for Highway Research
The University of Texas at Austin
Austin, Texas 72717
Segmental Technology and Services
P.O. Box 50825
Indianapolis, Indiana 46520
(3 17) 849-9686
University of California at Berkeley
Department of Civil Engineering
Berkeley, California 94720

CHAPTER 4 of the casting of the joints during erection and less
on the accuracy of the segments. Curvature and
twisting of the structure may be obtained within
the joint.
The principle of the match-cast joint is that the
4.1 Fabrication of Precast Segments 4,
connecting surfaces fit each other very accurately,
so that only a thin layer of filling material is
4.1 .l General Considerations needed in the joint. Each segment is cast against
During design of a segmental structure, consider- its neighbor. The sharpness of line of the assembled
ation should be given to the formwork necessary construction depends mainly on the accuracy of
to achieve economy and to obtain efficiency in the manufacture of the segments.
production. It is generally preferable to use as few
units as possible, consistent with economic ship-
ping and erection. 4.1.2 Methods of Casting
In the case of girder segments, economy and Segments to be erected with wide joints may be
speed of production may be increased by: cast separately. Match-cast joint members are cast
1. Keeping the length of the segments equal and by the long-line or short-line method.
keeping them straight, even for curved struc-
2. Proportioning the segments or parts of them, The Long-Line Method
such as keys and web stiffeners, in such a way
Principle-All of the segments are cast, in their
that easy stripping of the forms is possible. correct relative position, on a long line. One or
3. Maintaining a constant web thickness in the more formwork units move along this line. The
longitudinal direction. formwork units are guided by a pre-adjusted soffit.
4. Maintaining a constant thickness of the top An example of this method is shown in Figs. 4.1
flange in the longitudinal direction. through 4.3
5. Keeping the dimensions of the connection Advantages-A long line is easy to set up and to
between webs and the top flange constant. maintain control over the production of the
6. Bevelling corners to facilitate casting. segments. After stripping the forms it is not
7. Avoiding interruptions of the surfaces of webs necessary to take away the segments immediately.
and flanges caused by protruding parts for Disadvantages-Substantial space may be re-
anchorages, inserts, etc. quired for the long line. The minimum length is
8. Using a repetitive pattern, if practical, for normally slightly more than half the length of the
tendon and anchorage locations. longest span of the structure. It must be con-
9. Minimizing the number of diaphragms and structed on a firm foundation which will not
stiffeners. settle or deflect under the weight of the segments.
10. Avoiding dowels which have to pass through In case the structure is curved, the long line must
the forms. be designed to accomodate the curvature. Because
11. Minimizing the number of blockouts. the forms are mobile, equipment for casting,
Variation of the cross section of girder segments is curing, etc., has to move from place to place.
generally limited to changing the depth and width
of the webs and the thickness of the bottom
flange. Curves in the vertical and horizontal direc- The Short-Line Method
tion and twisting of the structure are easily ac-
Principle-The segments are cast at the same
place in stationary forms and against a neighboring
Segmental construction is distinguished by the
element. After casting, the neighboring element is
type of joint between elements. The following
taken away and the last element is shifted to the
types have been used:
place of the neighboring element, clearing the
1. Wide (broad) joints (this type of joint is not space to cast the next element. A horizontal cast-
considered in the design procedures presented ing operation is illustrated in Figs. 4.4 through 4.6.
in this manual). Segments intended to be used horizontally may
2: Match-cast joints. also be cast vertically. A photograph of a short-
The precision of line of segments assembled line form is presented in Fig. 4.7.
with wide joints depends mainly on the accuracy Advantages-The space needed for the short-

R n


\ Soffit
Fig. 4.1 - Cross section of formwork using long-line method(14)

/Outside Fotmwork
Inside Formwork


1 I
Fig. 4.2 - Start of casting (long-line method)(14)


Fig. 4.3 - After casting several segments (long-line method)(4)



=+#t+# f# - C A R R I A G E

Fig. 4.4 - Formwork for short-line method(4)

@ido Formwork /Bulkhead

Fig. 4.5 - Just before separation of segments (short-line method)14)

Fig. 4.6 -Just before casting next segment (short-line method)(4)

Fly. 4 7 Shott line form used for Purls Belt Brdges)

line method is small in comparison to the long- be flexible in order to accomr%date slight dif-
line method, approxim&tefy ttiree times the length ferences of dimensions with the previously cast
of a segment. The entire process is centralized. segment. They must be designed in such a way
Horizontal and vertieaf c-urves and twisting of that the.. necessary adjustments ior the desired
the structure are obtained by adjusting the po- camber, curvature and twisting can be achieved
sitionof the neighboring segment. accurately and easily.
Special consideration must be given to those
Disadvantages-To obtain the desired structural
parts of the forms that have to change in dimen-
configuration, the neighboring segments must be
sions. To facilitate alignment or adjustment, spec-
accurately positioned.
ial equipment such as wedges, screws, or hydraulic
jacks should be provided. Anchorages of the
tendons and inserts must be designed in such a
4.1.3 Formwork way that their position is rigid during casting.
Formwork must be designed to safely support Fittings must not interfere with stripping of the
all loads that might be applied without. undesired forms. If accelerated steam curing using tempera-
deformations or settlements. Soil stabilizat$n OP ture in excess of approximately 160 F (71 C) is
the foundation may be required, or the forr&ork~ foreseen, the .influences of the deformations of the
may be designed so thaf adjustments can be made + . forms, caused- .: by 1 heating and cooling, must be
to compensate for settlement. -* consid&e$ig prder to avoid development of cracks
Since prodUction of, segments is based onreusing in the -concrete. External vibrato& must be
the forms as much aspossible,.the formwork must_ attached at locations that will achieve maximum
be sturdy and s$%ai attention must be giLen to consolidation and permit easy exchange during
construction details. Forms must also be easy to the casting operations. Internal vibration may
handle. Paste leakage through formwork joints also be required.
must be prevented. This can normally be achieved Holes for prestressing tendons may be formed by:
by using a flexible sealing material. Special atten- 1. Rubber hoses which are pulled out after harden-
tion must be given to the junction of tendon ing of the concrete.
sheathing with the forms. The forms may need to 2. Sheathing which remains after hardening of the
concrete. Flexible sheathing made out of spirally characteristics of concrete required by the design
wound metal is usually stiffened from the inside may vary somewhat depending on whether the
by means of dummy cables, rubber or plastic segments are cast in the field or in a plant. The
hoses, etc., during the casting operation. results will be affected by curing temperature
3. Rigid sheathing with smooth or corrugated walls and type of curing. Liquid or steam curing or
may be ,used that will not deform significantly electric heat curing may be used.
under the pressure of wet vibrated concrete and A sufficient number of trial mixes must be made
for which there is no danger of perforation. to assure uniformity of strength and modulus of
4. Movable mandrels. elasticity at all significant load stages. Careful
Holes must be accurately positioned, particu- selection of aggregates, cement, gdmixtures and
larly when a large number of holes is required. water will improve strength and modulus of elas-
Horizontal and vertical tolerance for tendon ticity and will also reduce shrinkage and creep.
holes within the segment should not exceed +% Soft aggregates and poor sands must be avoided.
in. (13 mm) from the theoretical location. Ten- Creep and shrinkage data for the aggregates and/or
don ducts shall be match-cast in alignment at concrete mixes should be available or should be
segment faces. determined by tests.
Formwork that produces typical box girder Corrosive admixtures such as calcium chloride
segments within the following tolerances is con- may not be used. Water-reducing admixtures and
sidered good workmanship. also air-entraining admixtures which improve
concrete resistance to environmental effects such
Width of web. . . . . . . . . . . . . . +3/8 in. (10 mm)
as deicing salts and freeze and thaw actions are
Depth of bottom slab. . . ~~t00in.(13mmt00)
highly desirable. However, their use must be
Depth of top slab . . . . . . . . . . . &l/4 in. (6 mm)
rigidly controlled in order not to increase undesir-
Overall depth of segment. . . . . . *l/4 in. (6 mm)
able variations in strength and modulus of elastic-
Overall width of segment. . . . . . *l/4 in. (6 mm)
ity of concrete. The cement, fine aggregate, coarse
Length of match-cast segment. . +_ l/4 in. (6 mm)
aggregate, water and admixture should be com-
Diaphragm dimensions . . . . . . ?1/2 in. (13 mm)
bined to produce a homogeneous concrete mix-
Grade of roadway and soffit . . . .+1/8 in (3 mm)
ture of a quality that will conform to the mini-
Depending upon the detail at bridge piers, the
mum field test and structural design requirements.
tolerances for the soffit of a pier segment may need
Care is necessary in proportioning concrete mixes
to be limited to *l/16 in. (1.6 mm). The tolerance
to ensure that they meet specified criteria. Reliable
of a segment should be determined immediately
data on the potential of the mix in terms of
after removing the forms. If specified tolerances
strength gain and creep and shrinkage performance
are exceeded, acceptance or rejection should be
should be developed for the basis of improved
based on the effect of the over-tolerance on final
design parameters. Proper vibration should be used
alignment and on whether the effect can be cor-
to afford use of lowest slump concrete and to
rected in later segments. In match-cast construc-
allow for the optimum consolidation of the con-
tion, a perfect fit is established between segments.
Limits for smoothness and out-of-squareness of
the joint should be established.

4.1.5 Joint Surfaces

4.1.4 Concrete Requirements concerning surface quality must
Uniform quality of concrete is essential for seg- be stricter for match-cast joints than for wide
mental construction. Procedures for obtaining joints filled with mortar or concrete. Surfaces
high quality concrete are covered in PCI and PCA should be oriented perpendicular to the pain
publications.6*6 Both normal weight and post-tensioning tendons to minimize shearing
structural lightweight concrete can be made con- forces and dislocation in the plane of the joint
sistent and uniform with proper mix proportioning during post-tensioning. Inclination with respect
and production controls. Ideal concrete for seg- to a plane perpendicular to the longitudinal axis is
mental construction will have as near as practical permitted for joints with assured friction resis-
zero slump and 28-day strength greater than the tance. The inclination should generally not exceed
strength specified by structural design. It is rec- 20 degrees. Larger inclination, but not more than
ommended that stati&caI,methods be used to eval- approximately 30 degrees may be permitted if the
uate concrete mixes. * inclined surface area is located close to the neutral
The methods and procedures used to obtain the axis and does not exceed 25 percent of the total

joints surface area. 4.3 Methods of Erection7*20)
For match-cast joints, the surface, including
formed keys, should be even and smooth, to avoid 4.3.1 Cranes
point contact and surface crushing or chipping
Mobile cranes moving on land or floating on
off of edges during post-tensioning. Holes or
barges are commonly used where access is available
sheathing for tendons must be located very pre-
as illustrated in Fig. 4.8. Occasionally, a portal
cisely when producing segments joined by post-
crane straddling the deck has been used with
tensioning. Care is required to prevent leakage or
tracks installed on temporary trestles on either
penetration of joint-filling materials into the
side of the bridge. The capacities of cranes readily
duct, blocking passage of the tendons.
available in the United States and Canada makes
this method of erection more attractive than it is
in Europe.
4.1.6 Bearing Areas
Bearing areas at reactions should be even, with- 4.3.2 Winch and Beam
out ridges, grooves, honeycomb, etc., to assure In this method, illustrated in Fig. 4.9, a lifting
uniform distribution of bearing forces. It may be device attached to an already completed part of
desirable to place bearing elements like pads or the deck raises the segments which have been
steel plates in the forms before casting. Other- brought to the bridge site by land carrier or barge.
wise, cement mortar or epoxy may be required The segments are lifted into place by winches
on contact surfaces. carried at deck level on a short cantilever mechan-
ism anchored on the bridge. In the first applica-
tions of this type of erection in Europe, the seg-
ment over the pier had to be placed independently
4.2 Handling and Transportation of Precast (either cast-in-place or handled by a separate
Segments 4, mobile crane). Recently, this drawback has been
overcome. Now the precast pier segment may be
Segments should be handled carefully in a
placed on the pier with the same basic equipment
manner that limits stresses to values compatible
cantilevered temporarily from a tower attached
with the strength and age of the concrete. It should
to the pier.
be verified that the segment weights are less than
the capacity of the lifting equipment. Highway and
site transportation may produce dynamic stresses
4.3.3 Launching Gantry
which may be considered by use of an impact
coefficient. Special care of cantilever projections is In this method, a special machine travels along
often needed to prevent cracking. Location of lift- the completed spans and maintains the work flow
ing hooks and inserts should be determined care- at the deck level. The crane gantry, which was
fully to avoid excessive stresses in the segment first used for the Oleron Viaduct, has contributed
during handling, and they should have a safety significantly to the development of precast seg-
factor of 1.75 to 2.00 when all loads and stresses mental construction. The principle behind seg-
have been considered. Storage of units at the site mental erection using the crane gantry system is
should be arranged to minimize damage, deflec- shown in Fig. 4.10. An essential component in
tion, twist, and discoloration of the units. Stock- the system is a truss girder which has a length some-
piling should be limited to avoid excessive direct what greater than the maximum bridge span. The
or eccentric forces. Special precautions may be system consists essentially of:
required to avoid settlement of foundations made 1. A main truss where the bottom chords act as
to support the stored segments. Inserts, anchorages rolling tracks.
and other imbedded items may need to be pro- 2. Three-leg frames which may or may not be fixed
tected from corrosion and from penetration of to the main truss. The rear and center frames
water or snow during cold weather. In cases where allow the segments to pass through them longi-
extensive transportation of segments is required, tudinally.
it is recommended that a segment should not be 3. A trolley which can travel along the girder and
erected before it is certain that the subsequent is capable of longitudinal, transverse, and
segment has been safely transported. vertical movement as well as horizontal rotations.
Fig. 4.8 - Segment erection by crane, Corpus Christi Bridge, Texas
3. Finally, the segment placing trolley is used as a
launching cradle with the help of an auxiliary
tower bearing on the newly placed pier seg-
ment. The gantry is then transferred to its initial
position one span further thus allowing the seg-
ment placing cycle to repeat itself [Fig. 4.10
(c)l .
For structures combining vertical and horizon-
tal curvatures, including variable superelevation,
the launching gantry can be designed to follow
the geometry of the bridge while maintaining
operational stability and segment placing capa-
bility. In the last few years, several important
Fig. 4.9 - Winch and beam erection, St. Andre de Cubzac technical improvements have been made in gantry
Bridge, France() design. These advancements are exemplified
starting at the Chillon Viaduct in Switzerland,
and later at the Saint-Cloud Bridge where 143-
ton (130 t) segments were easily placed in a 337
ft. (102 m) span with a 1090 ft. (332 m) radius
of curvature (see Figs. 3.4 and 4.11). It should be
noted that, on certain structures, a somewhat
different approach is used in designing the launch-
ing gantry system (see Fig. 4.12). The total length
of the truss girder is now slightly greater than
twice the maximum span length. In this system,
all three gantry supports rest directly over a pier.
Although the investment cost is higher in this
system than in the original concept, this type of
f ~~
54.00 gantry has several advantages:
t 1. The completed deck carries no gantry reactions.
2. Stability against unsymmetrical loading due to
unbalanced cantilever erection may be pro-
vided by the gantry.
3. The pier segment may be placed and adjusted
during the normal placing cycle for the preced-
i 106.00 ing cantilever spans.
4. Construction time may be further reduced if
Fig. 4.10 - Operational stages of a launching gantry (first two placing trolleys are used.
In this advanced system, segments may be
moved in place over the completed bridge or
beneath the bridge. This procedure was used on the
large Rio-Niteroi Bridge where all segments were
To complete a full construction cycle for a typical
floated on pontoons and lifted into place by four
span, the gantry assumes three successive positions:
540 ft. (164 m) long launching gantries weighing
1 . For placing typical segments in cantilever, the 400 tons (363 t) each (see Fig. 4.13). A similar
center leg rests directly over a pier while the approach was also used for the 83 South Via-
rear leg is seated towards the end of the pre- ducts near Paris.
viously completed deck cantilever [Fig. 4.10
2. For placing the segment over the adjacent pier,
the girder is moved along the completed deck 4.3.4 Progressive Placing
until the ceder leg reaches the end of the can- The latest development of precast segmental
tilever. The front leg rests on a temporary cor- construction embodies the concept of progressive
bel fixed to the pier while the pier segment is placing. This approach actually comes directly
placed and adjusted into position [Fig. 4.10 (b)l . from cantilever design. Here, segments are placed
Fig. 4.11 - Launching gantry, St. Cloud Bridge, Parist7)

at each pier. 7
When the-deck reaches one pier, permanent
bearings are installed and construction proceeds
to the next span. Some noteworthy advantages
of the method are:

1. The operations are continuous and are per-

formed at the deck.level.
2. The method seems to be of interest primarily in
the 100 to 160 ft. (30 to 50 m) span range
where conventional cantilever construction is
not always economical,
3. During construction, the piers are not subjected
to significant unbalanced moments although
the vertical reaction is substantially increased.

One disadvantage of the method is that con-

struction of the first sqt2must be carried
. out . with,
a specialsystem.
It should also be noted that the stresses inthe
/ deck are completely reversed duringcanstruetion
F i g . 4 . 1 2 r .Operationat stag& o f a l a u n c h i n g gantri~
, (second ty&)t and after completion,_Conseq~uently, special sta-
% */ j: 1 t
bilization devices must be used temporarily to
keep. the concrete stresses within safe limits and
continuously from one end of tfie deck to the., to minimize the amount of temporary prestress.
other insuccessive Cantilevers.ontlie same side of A tower and .guy cable system has been used effea-4
the various piers rather than in balan&dcantilever tively to control theundesirabfe temporary stresses.
Fig. 4.13 - Launching gantry for Rio-Niteroi Bridge, Brazi117

Figs. 4.14 and 4.15 show schematically the princi-

ple of progressive segment placing together with
some of the construction details.

Fig. 4.15 - Construction sequence (isometric view) using

progressive segment placing17)

4.3.5 Erection Toleranced 4,

Maximum differential between outside faces of
adjacent units in the erected position should not
exceed l/4 in. (6 mm). The most important item
of tolerance or acceptance is the final geometry of
the erected superstructure. The evaluation of the
deck surface of each segment used in the cantilever
portions of the bridge superstructure, measured
after closure sections are in place, must not vary
from the theoretical profile grade elevation by
more than that specified for the project. The
gradient of the deck surface of each segment
should not vary from the theoretical profile grad-
ient by more than 0.3 percent. More liberal tol-
Fig. 4.14 - Construction sequence (elevation) using pro- erances may be acceptable if the design incorpo-
gressive segment placingt7) rates a wearing surface.
4.3.6 Design of Piers and Stability During construction has been provided by the equipment
Construction used for placing the segments.
With double piers, two parallel walls make up Single Slender Piers the pier structure, which usually rests on a single
foundation. Such a configuration was successfully
If the piers in the finished structure are designed
used for a number of European bridges, including
solely to transfer the deck loads to the foundations
the Chillon Viaduct illustrated in Fig. 4.18. Stabil-
(including horizontal loads), there is the likelihood
ity during construction is excellent and requires
that the piers will be unable to resist the unsym-
little temporary equipment, except for some
metrical moments due to the cantilever construc-
bracing between the slender walls to prevent elas-
tion (i.e., with one segment plus the equipment
tic instability.
load). Thus, temporary shoring is often required
(see Figs. 4.16 and 4.17) at considerable cost.
More recently, the stability of the cantilever under

SEGMENT WEIGHTS: 60 TO 40 t Moment Resisting Piers
- ?.tAX. STATICAL REACTION IN SUPPORT: ,060 , Moment resisting piers are designed to with-
stand the unbalanced moments during construc-
tion while temporary vertical prestress rods make a
rigid connection between the deck and the pier
cap. The Corpus Christi Bridge shown in Fig. 4.19
PROVISIONAL SUP utilized moment resisting piers.
When the ratio between span lengths and pier
height allows it, the rigid connection and the
corresponding frame action may be maintained
1 t = 1.1 ton
1 m = 3.28 ft. permanently between the deck and piers. This
frame action is also achieved by use of twin neo-
Fig. 4.16 - Stability during construction() prene bearings which allow for deck expansion.

Fig. 4.17 - Temporary erection shoring at pier, Pierre-Benite Bridge, France()

Fig. 4.18 - Twin piers, Ghillon Viaduct,,Switrer21and(j

-- --
Fig. 4.19,- Moment resisting piers, Corpus Christi Bridge, TexaO
Fig. 4.21 - Twin neoprene bearings in final structure(2)

This type of pier detail is shown in Fig. 4.20

where the elastomeric bearings are indicated as
Fig. 4.20 - Piers with twin neoprene bearings during con- (1 ), the vertical erection post-tensioning between
struction(21) pier and super-structure is shown as (21, and the
temporary concrete bearing pads are shown as (3).
After completion of erection and continuity post-
Flat jacks are usually placed between the pier top tensioning, the vertical post-tensioning at the pier
and the deck soffits to permit the removal of and the temporary concrete bearing pads are re-
temporary bearings and installation of the per- moved, leaving the neoprene bearings in place as
manent ones. shown in Fig. 4.21.


5.1 General
1 m = 3.28 ft.
The North Vernon Bridge over the Muscatatuck
Fig. 5.1 - Span arrangement
River in Indiana was built parallel to an existing
reinforced concrete arch bridge with the purpose
of doubling the capacity of the existing roadway.
The spans were therefore fixed to meet those of
the arch, as indicated in Fig. 5.1. Cost estimates
for widening the bridge with another arch proved
too expensive and led to consideration of both
steel and concrete alternatives. The presence of a
precast concrete plant in the vicinity of the bridge
site, and the feasibility of segment erection by
mobile crane made it possible that even this small
structure with a total deck area of only 8855 sq. I
ft. (823 m*) could be built competitively using
precast segmental construction.
a-AREA A (M2)
C, (Ml 1.002 1.066 1.122 1.151
c-MOMENT OF INERTIAI (MO) 4.0239 4.423 4.750 4.908
d-MOMENT OF RESIST 4.0151 4.1482 4.2328 4.268
5.2 Structure Dimensions 2,
lM3) 2.3115 2.6378 2.9306 3.0824
e-KERNEL BEAM Kt IM) 0.555 0.608 0.705 0.74
The total bridge length of 381 ft. (116.04 m) Kb (Ml 0.956 0.938 0.925
f-THICKNESS BOTTOM D 041 0.252 0.302 0.33
is made up of 2 end cross girders of 5 ft. 3 in. (1.6
m), 44 segments of 8 ft. 0 in. (2.44 m) length, 2
Fig. 5.3 - Cross section dimensions and segment properties
pier segments of 9 ft. 0 in. (2.74 m) length, and a
cast-in-place splice of 5 ft. 3 in. (1.6 m). The span
and segment dimensions are shown in Fig. 5.2. for all segments except for the two segments lo-
In consideration of the length of the main span, cated on either side of the two pier segments. In
the depth of the box girder was selected as 9 ft. these segments, the bottom slab thickness was
0 in. (2.745 m). The resulting span/depth ratio of increased from 8 in. (0.20 m) to 13 in. (0.33 m)
21 .l is well within the economical limits. The box in order to reduce the compressive stress in the
girder dimensions and section properties are pre- bottom fibers resulting from the negative support
sented in Fig. 5.3. These dimensions are constant moments.

I2 3 4 5 6 7 6 9 IO II 12 13 1st l5 I6 I7 I6 I9 20 21 22 23 24 25 26

2xJ3 9x 2.44 9 x 2.44

29.8 I 29.01
1 m = 3.28 ft.
Fig. 5.2 - Segment dimensions and joint numbers

5.3 Order of Erection 5.4 Post-Tensioning Details
The erection sequence for the structure is i n Except as noted below, the post-tensioning is
three steps as indicated in Fig. 5.4. carried out by tendons consisting of twelve % in.
Step 1: The segmental cantilevers are erected diameter 270 k strands (13 mm $I, 1862 MPa)
from each pier. with an ultimate force of 495 kips (2202 kN).
Step 2: The precast end cross girders a r e All tendons are stressed initially to 70 percent of
erected. their ultimate force. The effective force level in
Step 3: The midspan splice is cast-in-place. the example design calculations at time of pre-
stressing is reduced to allow for anchor seating and
friction losses. The final tendon forces after losses
are 60 percent of ultimate or lower. ,

a P B
The post-tensioning tendons are arranged in
groups as follows:
Group 1: Cantilever post-tensioning consists of

26 tendons, 13 in each web (See Fig.
Group 2: Tail span continuity post-tensioning
consists of 2 tendons, one in each web
S T E P 2
(See Fig. 5.6).
Group 3a: Center span continuity post-tensioning
consists of 8 tendons, 4 in each web,
> S T E P 3
located in the bottom slab at midspan
and anchored in the top slab (See Fig.
Fig. 5.4 - Erection sequence 5.7).


Fig. 5.5 - Cantilever tendon layout

Fig. 5.6 - Tail span continuity tendons

201 202 203 20 4 3 b

b .\
-1 . . \
Y- .w-- - - - - - <- - tJ- -- \ I-1 ---l-----------.~--J. 205
. .
5 \ . ., . .
t . . .
- - - t=-- L- -a.-,- ---r-- _.-_ __.

Fig. 5.7 - Center span continuity tendons: Bottom slab (Group 3a)
Top slab (Group 3b)



G R O U P 3 a
Fig. 5.8 - Location of tendons (eccentricities)

5. Final tendon forces are 60 percent of ultimate

or lower.
The design is carried out for loading by:
1. Dead load during construction
2. Initial prestress
3. Superimposed permanent loads
4. Live loads
5. Temperature differential
fig. 5.9 - Loading for erection moment stability 6. Creep under box girder dead load
7. Creep under post-tensioning
Group 3b: Center span continuity post-tensioning 8. Loss of prestress
consists of four 6-strand tendons lo-
cated in the top slab. These tendons are
anchored in the pier segments (See 5.6 Design Procedure
Fig. 5.7).
The design of the North Vernon Bridge is pre-
The precise location of the tendons in the section
sented in accordance with the following steps:
is indicated in Fig. 5.8.
Step 1: Free cantilever plus initial cantilever
Group 1 post-tensioning. Stress control in
5.5 Design Requirements and Loading* all phases of erection.
Step 2: Completion of end span plus initial con-
The design is carried out by elastic methods
tinuity Group 2 post-tensioning. Stress
to meet the following criteria:
1. Concrete bending stresses within allowable limits Step 3: Concreting of midspan splice plus initial
for 5500 psi (38 MPa) concrete. continuity Group 3 post-tensioning. Stress
2. No tension allowed for combinations of all control.
The design requirements presented here are those selected for the
3. Cracking safety under 110 percent of dead load North Vernon Bridge and are generally somewhat more conserva-
and 125 percent of live load.* tive than required by current American Association of State High-
way Officials Bridge Specifications(6). and the PCI Tentative De-
4. Ultimate load capacity of 115 percent of dead sign and Construction Specifications for Precast Segmental Bridges
load and 225 percent of live load.* presented in Appendix Section A.I.

Step 4: Addition of permanent loads. Stress con- In all cases, provision must be made to accom-
trol. modate additional temporary erection loads on
Step 5: Addition of variable loads. Stress control. the structure, and stress and stability checks must
Step 6: Influence of time. be made for the structure under these loadings.
Step 6a: Dead load moment redistribution due to Such erection loads can be intentional (for exam-
concrete creep. Stress control. ple, movement of a launching girder over the struc-
Step 6b: Post-tensioning moment redistribution ture), or unintentional (storage of post-tensioning
due to concrete creep. Stress control. tendons or a large group of visitors on the struc-
Step 6c: Prestress losses. Stress control. ture). Consideration of erection loads has been
Step 7: Final stress control omitted in the presentation of this design exam-
Step 8: Transverse section analysis. ple for simplicity.
Other calculations required to complete the
design are made by procedures common to con-
ventional post-tensional box girder bridges or con-
ventional reinforced concrete design and are not
presented here. These calculations relate to the All of the following design example diagrams and
following: dimensions are in c.g.s. metric units

1. Calculation of end cross girder and pier segment Dimensions = meters (3.28083 ft)
reinforcement. Forces = metric tonnes (2204.62 lb)
2. Support forces and bearing requirements. Bending moments = tonnes x meters (7232.98 ft. I b.)
3. Road joint movements. Stress = tonnes/sq. meter (1.422 psi)
4. Principal shear stresses at service load. The relationship to SI metric units is:
5. Ultimate moments, safety to failure. Force: 1 t = 9.8 kN = 2204.62 lb.
6. Ultimate shear, safety to failure. (1 lb. = 4.448 Newtons)
7. Substructure loading during erection. Moment: 1 t-m = 9.8 kN-m = 7232.98 ft.-lb.
8. Temporary prestressing of segments during (1 ft.-lb. = 1.356 kN-m)
erection. Stress: 1 t/m2 = 9.8 kPa = 1.422 Ib./in.2
9. Reinforcement of keys. (1 Ib./in.2 = 6.895 kilopascals)

5.6.1 Step 1. Free Cantilever Plus Initial Can- force diagram by the section modu-
tilever Group 1 Post-Tensioning lus of the bottom fiber and dividing
by the section area F/A x Zb. This
In Step 1, stresses are calculated for loading due
is the bottom fiber moment due to
to the dead load of the free cantilever box girder
the axial compression from post-
section and the Group 1 cantilever post-tensioning.
tensioning (t-m).
The post-tensioning is shown in Fig. 5.5 and con-
sists of 13 tendons in each web. A check is made
for unbalance during erection. The calculations are
made as follows: (See Fig. 5.2)

1. Calculate the effect on the supporting struc-

ture caused by unbalance of segment x+n+l
(See Fig. 5.9) Check stability of the assembly.
The stability is in this case assured by placing
two supports on a wide pier.
2. Calculate concrete stresses in each joint due
to dead load of segments x+n+ 1.
3. Calculate forces in the tendons present in
the segments x to x+n. Consider friction
losses and, if judged necessary, steel relaxa-
tion. Subsequently calculate concrete stresses
in each joint due to post-tensioning.
4. Comply with stress limitations for all values
of n in each joint. 7

At completion of erection of one cantilever,

the bending moments are as shown in Fig. 5.10
(only half of the cantilever is shown, the other
half is identical):
Diagram 5 Box girder dead load bending mo-
ments (t-m). 9
Diagram 6 Tendon force diagram for Group 1
post-tensioning (t).
Diagram 7 Eccentricity of Group 1 tendons,
Diagram 8 Bending moment diagram due to
Group 1 post-tensioning, diagram
6 x 7 (t-m).
Diagram 9 IO
Bending moment diagram obtained
by multiplication of the tendon
force diagram hy the section modu-
lus (moment of resistance in Fig.
5.3) and dividing by the section
area F/A x Z,. This is the top fiber
moment due to the axial compres-
sion from post-tensioning (t-m).
Diagram 10 Bending moment diagram obtained Fig. 5.10 - Step 1. Free cantilever plus initial cantilever
by multiplication of the tendon group 1 post-tensioning

Diagram 11 Check top fiber moments (indirect-
(Fig. 5.11) ly checking stresses). Moment dia-
gram (a) is obtained by adding dia-
grams 8 and 9 from Fig. 5.10.
This is the top fiber moment due to
the combination of bending and
axial force resulting from post-
tensioning. Diagram c + a is diagram
a reduced by the dead load moment
diagram (diagram 5 in Fig. 5.10).
The top fiber compressive stress
control limits are indicated by the
f, x Z, diagram. Allowable com-
pressive stress at this stage is 2150
t/m2. In this case 2150 x 1.422 =
3057 psi or approximately 0.55 x
5500 = 3025 psi.
Diagram 12 Check bottom fiber moments (indi-
rectly checking stresses). Moment
diagram b is obtained by adding
diagrams 8 and 10 from Fig. 5.10.
This is the bottom fiber moment 12

due to the combination of bending

and axial force from post-tension-
ing. Diagram c + b is the addition
of diagram 5, box girder dead
load moment from Fig. 5.10, to
diagram b. The bottom fiber com-
pressive stress control moment dia-
gram based on f, = 0.55 x 5500 =
3025 psi, or 2150 t/m2, is also indi-
cated in Fig. 5.11 [2150 x 3.0824 Fig. 5.11 - Step 1. Check top fiber and bottom fiber
= 6627 t-m (47,932 ft. kips)] . moments

5.6.2 Step 2. Completion of Tail Span Plus Add diagrams 3 and 5 to obtain
Continuity Group 2 Post-Tensioning 6b. This is the combined axial
(expressed as a moment) and
The completion of the tail span is achieved by
bending moment effect of the post-
addition of the end span cross girders and installa-
tensioning on the bottom fiber
tion of the Group 2 post-tensioning shown in Fig.
5.6. This post-tensioning consists of one tendon in
each web. For analytical purposes, the changes
with respect to Step 1 are:
1. End cross girder is added 2 3 4 5 6 7 8 8 1 0

2. End support is added

3. Continuity Group 2 post-tensioning is installed
4. Two supports at piers are replaced by one
support at the center of the pier
With reference to Fig. 5.12, the calculations to
account for the above changes proceed as follows:
Diagram 1 Determine box girder dead load
bending moment diagram due to in-
troduction of end support and end 3

cross girder.
Diagram 2 Determine force diagram of Group
2 post-tensioning tendons and the
tendon eccentricities.
Diagram3 Determine the bending moment dia-
gram due to Group 2 tendons.
The structure is simply supported
and the bending moment equals the
force multiplied by the eccentricity.
Diagram 4 The tendon force diagram multi- 5
plied by the top section modulus,
Z,, and divided by the section
area, A, expresses the axial com-
pression due to post-tensioning in 6a
terms of a top fiber moment (t-m).
Diagram 5 The tendon force diagram multi-
plied by the bottom section mod-
ulus, Zb, and divided by the section
area, A, expresses the axial com-
pression due to post-tensioning in
terms of a bottom fiber moment
(t-m). -172
Diagram 6 Add diagrams 3 and 4 to obtain 6 b
diagram 6a. This is the combined
axial (expressed as a moment)
and bending moment effect of the
post-tensioning on the top fiber Fig. 5.12 - Completion of tail span plus continuity group
(t-m). 2 post-tensioning

See diagrams of Fig. 5.13 with numbers corres-
ponding to those below, and diagrams from previ-
ous figures as noted.
Diagram 7 Add bending moment diagrams due
to box girder dead load from Steps
1 and 2 (diagram 5 from Fig. 5.10
plus diagram 1 from Fig. 5.12)

Diagram 8 Add diagram 11 a of Step 1 to

diagram 6a of Step 2. +tSS

Diagram 9 Check top fiber moments com-

pared to allowable by adding dia-
gram 7 and the results of calcula-
tion 8 above. As can be seen, there
is a large margin between the maxi-
mum permissible moment of 9176
t-m and the moment in the struc-

Diagram 10 Add diagram 12b of Step 1 to dia-

gram 6b of Step 2.
Diagram 11 Check bottom fiber moments com-
pared to allowable by adding dia-
gram 7 to the results of calculation -I?59
10 above. Again the structure mo-
ment is much less than the permis- Fig. 5.13 - Step 2 continued. Check top fiber and bottom
sible value of 6627 t-m. fiber moments

5.6.3 Step 3. Completion of Center Span account for the effect of the axial
At this stage, the cast-in-place midspan splice is force.
completed and continuity post-tensioning in Diagram 5 Add diagrams 2e and 3 to obtain
Groups 3a and 3b is placed and stressed. Group 3a diagram 5a which is the total
post-tensioning consists of four tendons in each effect of the post-tensioning with
web which are located in the bottom slab at mid- respect to the top fiber, expressed
span. Group 3b post-tensioning consists of four 6- as a moment.
strand top slab tendons. Both Group 3a and Group Add diagrams 2e and 4 to obtain
3b post-tensioning are shown in Fig. 5.7. diagram 5b which is the total
The calculation procedure illustrated in Fig. effect of the post-tensioning with
5.14 for this step is as follows: respect to the bottom fiber, ex-
pressed as a moment.
Diagram 1 Calculate the bending moment dia-
gram due to the additional weight 456!69

of the midspan cast-in-place seg-

The tendon force diagram and
eccentricities shown in Fig. 5.14
are for all tendons in Groups 3a
and 3b.
Diagram 2 Determine bending moment dia-
grams due to post-tensioning Groups
3a and 3b. The post-tensioning is
stressed in the continuous system,
and the resulting moment diagrams
are obtained as follows:
2a. Assume hinges at supports on
piers, calculate post-tensioning force
2b. Calculate the bending moments
due to post-tensioning Groups 3a
and 3b for the hinged span CE
(moment = force x eccentricity). d

2c. Calculate angle of rotation at

C and E by the moment diagram
obtained in 2b. e
2d. Calculate the secondary moment
required to rotate the joint closed
at C and E.
2e. The addition of diagrams 2b
and 2d is the bending moment
diagram resulting from post-ten-
sioning in the continuous system.
Diagram 3 Multiply the tendon force diagram
of post-tensioning Groups 3a and
3b by Z, and divide by the section
area, A. This provides an equivalent
top fiber moment diagram to ac-
count for the effect of the axial
Diagram 4 Multiply the tendon force diagram
5 b
of post-tensioning Groups 3a and
3b by Zb and divide by the section
area, A. This provides an equivalent
bottom fiber moment diagram to Fig. 5.14 - Step 3. Completion of center span

Top and bottom fiber stresses are checked in 5.6.4 Step 4. Addition of Superimposed Dead
Fig. 5.15 in terms of moments as follows: Loads
At this stage the effect of permanent superim-
Diagram 6 Check top fiber moment by addi- imposed dead loads due to addition of curbs,
tion of diagrams 1 and 5a of Step railings and toppings is considered. Permanent
3 to diagram 9 of Step 2. The mo- superimposed loads are treated separately from live
ments (and stresses) are satisfac- loads because permanent loads cause creep defor-
tory in all locations. mations of the structure. The amount of the super-
imposed dead load is 1.525 t/m (1.03 kip/ft.).
Diagram 7 Check bottom fiber moment by ad-
With reference to Fig. 5.16, the calculation pro-
dition of diagrams 1 and 5b of Step
cedure is as follows:
3 to diagram 11 of Step 2. Again
the moments are well within the Diagram 1 Calculate bending moments due to
allowable values throughout the superimposed loads.
length of the structure. Diagram 2 Check top fiber moments by add-
ing diagram 1 above to diagram 6
of Step 3. All top fiber moments
are within the allowable.
Diagram 3 Check bottom fiber moments by
adding diagram 1 above to diagram
7 of Step 3. All bottom fiber mo-
ments are within the allowable.

Fig. 5.15 - Step 3 continued. Check top fiber and bottom

fiber moments

Fig. 5.16 - Step 4. Addition of superimposed dead loads.

Check top fiber and bottom fiber moments

5.6.5 Step. 5. Application of Live Load and Diagram 4 Check top fiber moments with
Temperature Load (Fig. 5.19) respect to allowable by combining
diagram 2 of Step 4 (Fig. 5.16)
The live load on the structure is HS20-44.
with 3a and 3b of Step 5. All
The temperature loading consists of a 10 C
(18O F) temperature rise of the top slab with moments within the allowable.
respect to the webs and the bottom slab for maxi- Diagram 5 Check bottom fiber moments with
mum temperature effects and a 5O C (go F) tem- respect to allowable by combining
perature decrease of the top slab with respect to diagram 3 of Step 4 (Fig. 5.16)
the webs and bottom slab for minimum tempera- with diagrams 3a and 3b of Step 5.
ture effects. With the area of the top slab 1.988 All moments within the allowable.
m* (21.39 ft.*) and modulus of elasticity 3.5 x
lo6 t/m* (5 x lo6 psi), the force developed by a
1OoC temperature differential with a thermal co-
efficient, CY, of 0.00001 m/m/C (5.56 x 1O-6
in./in./OF) is 695.8 t (1534 kips). The eccentricity
of this force with respect to the neutral axis is
0.926m (3.04 ft.). The temperature differential
analysis procedure is presented in Section 3.3.4.
The temperature stresses calculated are converted
to equivalent bending moments. As illustrated by
Figs. 5.17, 5.18 and 5.19, the calculation proce-
dure is as follows:
Diagram 1 Calculate live load positive mo-
(Fig. 5.17) ments (diagram la) and negative
moments (diagram 1 b).
Diagram 2 Calculate maximum bottom fiber
(Fig. 5.17) temperature moments (2a) and
minimum bottom fiber temperature 2b

moments (2b), and maximum and +,*I

minimum top fiber temperature +4.

moments (Figs. 2c and 2d, respec- 2 c

Diagram 3 Combine diagrams 1 and 2 to pro-
(Fig. 5.18) vide:
3a. Maximum live load moment
plus maximum temperature mo-
ment bottom fiber.
3b. Minimum live load moment and
minimum temperature moment top Fig. 5.17 - Step 5. Application of live load and tempera-
fiber. ture load

3 b

Fig. 5.18 - Step 5 continued. Maximum live load plus Fig. 5.19 - Step 5 continued. Check top fiber and bottom
maximum temperature moments on bottom fiber and top fiber moments

5.6.6 Step 6. Influence of Time Diagram 5 Diagram 3 is multiolied bv the IOW
With passage of time, the moments in the struc- value of the creep factor Ci - e*2 )
ture are modified due to creep effects on box gird- to provide a low estimate of the
er dead load moments and post-tensioning mo- box girder dead load m o m e n t
ments, and by the effect of prestress losses. These redistribution.
three effects will be considered separately in the
following calculations. Step 6a will consider the
redistribution of box girder dead load moments 60
due to creep, Step 6b will cover the redistribution
of the post-tensioning moments due to creep, and
Step 6c will consider the effect of prestress losses.
In all of these calculations, high and low values
of the creep factors will be assumed as follows:
f#l, = 1.41 #* = 1.05
(1 -e-@I) =0.76 (1 -ee-@2) =0.65 Step 6a. Box Girder Dead Load Moment

Redistribution Due to Creep
With reference to Fig. 5.20, the calculation
procedure is as follows:
Diagram 1 Calculate box girder dead load mo-
ments (not including superimposed
dead load) in the continuous struc-
tu re.
Diagram 2 Calculate the box girder dead load
moments at completion of erec-
tion. Add diagram 5, Step 1 to
diagram 1, Step 2, and diagram 1, 4
Step 3.
Diagram 3 The difference between diagrams 1
and 2 above is as shown. 5
Diagram 4 Diagram 3 is multiplied by the high
value of the creep factor (1 - e-@l)
to provide a high estimate of the
box girder dead load creep moment Fig. 5.20 - Step 6. Influence of time. Step 6a. Box girder
redistribution. dead load moment redistribution due to creep

97 Step 6b. Post-Tensioning Moment Redis- Diagram 12 Combine diagrams 5 and 10 to ob-
tribution Due to Creep tain a low value of the total redis-
Diagram 1 The effects of cantilever post- tribution of post-tensioning mo-
(Fig. 5.21) tensioning (Group 1) on the contin- ments due to creep.
uous structure.
Diagram 2 The cantilever post-tensioning mo-
ments at the end of erection (dia- 6 b
gram 8 from Step 1, Fig. 5.10).
Diagram 3 The difference between diagrams 1
and 2 above.
Diagram 4 Multiply diagram 3 by the high
value of the creep factor (1 - e*l ),
giving a high estimate of the canti-
lever post-tensioning (Group 1)
moment redistribution due to creep.
Diagram 5 Multiply diagram 3 by the low
value of the creep factor (1 - e-#2 )
giving a low estimate of the canti-
lever post-tensioning (Group 1)
moment redistribution due to creep.
Diagram 6 Determine the effect of Group 2
continuity post-tensioning on the
continuous structure. 3
Diagram 7 The effect of Group 2 post-tension-
ing during construction (diagram 3, 4
Fig. 5.12).
Diagram 8 The difference between diagrams 6 5
and 7, above.
Diagram 9 Multiply diagram 8 by the high
value of the creep factor ( 1 -
e-@l 1, giving a high estimate of
the Group 2 post-tensioning mo- 7
t 21
ment redistribution due to creep.
Diagram 10 Multiply diagram 8 by the low value 6

of the creep factor (1 - e+2 1, 9

giving a low estimate of the Group IO

2 post-tensioning moment redistri-
bution due to creep. I I
Diagram 11 Combine diagrams 4 and 9 to ob-
tain a high value of the total redis-
tribution of post-tensioning mo- Fig. 5.21 - Step 6b. Post-tensioning moment redistribu-
ments due to creep. tion due to creep

98 Step 6c. Effect of Prestress Losses Diagram 8 Group 2 continuity post-tensioning
(Fig. 5.23) bending moments in continuous
Prestress losses due to friction, elastic shorten-
ing, shrinkage and creep have been calculated as system multiplied by loss per-
14 percent of initial forces or 18.610 t/m* (26,460 centage.
Diagram 9 Group 1 cantilever post-tensioning
bending moments in continuous
Diagram 2 Group 1 post-tensioning tendon
system multiplied by loss per-
(Fig. 5.22) force diagram multiplied by pre-
stress loss percentage. Group 3 continuity post-tensioning
Diagram 10
Diagram 3 Group 2 post-tensioning tendon bending moments in continuous
force diagram multiplied by pre-
system multiplied by loss per-
stress loss percentage. centage.
Diagram 4 Group 3 post-tensioning tendon
Diagram 11 Diagrams 8, 9 and 10 added to-
force diagram multiplied by pre- gether.
stress loss percentage. Diagram 11 added to diagram 6 to
Diagram 12
Diagram 5 Diagrams 2,3 and 4 added together. obtain total equivalent top fiber
Diagram 6 Diagram 5 multipled by 2, and bending moments due to losses.
divided by the section area, A. This Diagram 13 Diagram 11 added to diagram 7 to
is the prestress force loss effect on obtain total equivalent bottom fiber
the top fiber expressed as a mo- bending moments due to losses.
Diagram 7 Diagram 5 multiplied by 2, and di-
vided by the section area, A. This
is the prestress force loss effect on 162021222324252627

the bottom fiber expressed as a

moment. e -3

g a

0 a

II -a



Fig. 5.23 - Step 6c continued. Equivalent top fiber and

Fig. 5.22 - Step 6c. Effect of prestress losses bottom fiber bending moments due to prestress losses

5.6.7 Step 7. Final Stress Control

Diagram 1 Calculation of total time-related

(Fig. 5.24) (maximum and minimum) effects
from Steps 6a, 6b and 6c for top

Diagram 2 Calculation of total time-related

(Fig. 5.24) (maximum and minimum) effects
from Steps 6a, 6b and 6c for bot-
tom fibers.

Diagram 3 Final stress control for the top

(Fig. 5.24 fiber is evaluated by combining
diagram 1 above with diagram 4
from Step 5 (Fig. 5.19).

Diagram 4 Final stress control for the bottom

(Fig. 5.24 fiber is evaluated by combining
diagram 2 above with diagram 5
from Step 5 (Fig. 5.19).
I ,. !.
11, / 1

Fig. 5.24 - Step 7. Final stress control. Top fiber and

bottom fiber time-related bending moments


5.6.8 Step 8. Calculation of Transverse Moments

Transverse moments in the North Vernon Bridge
were calculated by use of a computer program
based on folded plate theory. The calculation pro-
cedure divides the box section into longitudinal
strips which may or may not be of constant
thickness. This makes it possible to include con-
sideration of the areas where slabs or webs are
thickened. The length of the strips is taken as the
span length for a simply supported box girder,
or as the distance between points of zero moment i P
in the case of a continuous box girder.
The results of the analysis are given at the con-
nections of the longitudinal strips. The results
provide bending moments and axial forces due to
box girder dead load, superimposed dead loads,
linear loads (curbs), and live loads. The live load
may be either uniformly distributed, or one or
more vehicles. In either case, live load moments
are obtained from influence lines calculated for
each section. The uniformly distributed live load or
a design vehicle is placed on the influence lines in
such a position as to give maximum positive or
negative moments. Because of the effect of load
distribution, influence lines for uniformly dis-
tributed live loads differ from influence lines for
+ve +I.150
vehicles. +QO88
+MI +Q-
Influence diagrams and moment and axial force -2.477 -2177

diagrams for the North Vernon Bridge are pre-

sented in Figs. 5.25 to 5.31. Size and location of AXIAL FORCE

transverse reinforcement are shown in Fig. 5.32.

Fig. 5.25 - Transverse moments and axial forces due to

box girder dead load

,325 t/m
2 3

4 IO 4

-QLlOl I
0 0



Fig. 5.27 - Transverse moments and axial forces due to

Fig. 5.26 - Transverse moments and axial forces due to
linear loads (one side only)
superimposed dead load

1 / I9 p g! /
-0.024 t on24
1-o.co2 1 SECTION 5;


Fig. 5.28 - Influence lines for vehicle Fig. 5.29 - influence lines for vehicle, continued

Fig. 5.30 - Transverse moments and axial forces due to Fig. 5.31 - Transverse moments and axial forces due to
uniformly distributed live load vehicular loading



, II
319 @ 6 / i
3194 @ 6

, / -- -- J
1 ft. = 0.3048 m
T- ---fi l\,.
f 7+ I 1 in. = 25.4 m m
, 4 1 3 Cal IO1

bar size is indicated by the first digit of bar numbers
Fig. 5.32 - Transverse reinforcement details
A. APPENDIX Elastic analysis and beam theory may be used in
the design of precast segmental box girder struc-
A.1 Tentative Design and Construction Specifica- tures. For box girders of unusual proportions,
tions for Precast Segmental Box Girder Bridges. methods of analysis which consider shear lag*
%The PCI Bridge Committee prepared tentative shall be used to determine stresses in the cross
design and construction specifications and accom- section due to longitudinal bending.
panying commentary in 1975 in the form of a pro- (B) Design of Superstructure
posed addition to the AASHTO Standard Specifi- (1) Flexure
cations for Highway Bridges. They were presented
The transverse design of precast segments for
to the AASHTO Committee on Bridges and Struc-
flexure shall consider the segment as a rigid box
tures for evaluation, and then were published by
frame. Top slabs shall be analyzed as variable depth
the Prestressed Concrete Institute (PC1 JOURNAL,
sections considering the fillets between the top
July-August 1975) to develop comments and dis-
and webs. Wheel loads shall be positioned to pro-
vide maximum moments, and elastic analysis shall
The PCI Bridge Committee evaluated the com-
be used to determine the effective longitudinal
ments received relative to the 1975 tentative speci-
distribution of wheel loads for each load location
fications as well as new information on design and
(see Article 1.2.8). Transverse post-tensioning of
construction of precast segmental box girder
top slabs is generally recommended.
birdges, and prepared the following version of the
In the analysis of precast segmental box girder
design and construction specifications for con-
bridges, no tension shall be permitted at the top of
sideration by the AASHTO Subcommittee on
any joint between segments during any stage of
Bridges at its 1977 Regional Meetings. The speci-
erection or service loading. The allowable stresses
fication proposals as presented in this section rep-
at the bottom of the joint shall be as specified in
resent the recommendations of the PCI Bridge
Article 1.6.6 (B) (2).
Committee, and may be modified prior to final
adoption as AASHTO Standard Specifications for (2) Shear
Highway Bridges. (a) Reinforced keys shall be provided in segment
The specification proposals are presented here in webs to transfer erection shear. Possible reverse
a format utilizing section numbers compatible with shearing stresses in the shear keys shall be
the 1973 AASHTO Standard Specifications for investigated, particularly in segments near a
Highway Bridges. Specifically, new sections of the pier. At time of erection, the shear stress car-
1973 AASHTO Specifications are proposed as ried by the shear key shall not exceed 2c
(b) Design of web reinforcement for precast
segmental box girder bridges shall be in ac-
cordance with the provisions of Article 1.6.13.
1.6.25 Precast Segmental Box Girders
2.4.33 (L) Precast Segment Manufacture and (3) Torsion
Erection In the design of the cross section, consideration
2.4.33 (M) Epoxy Bonding Agents for Pre- shall be given to the increase in web shear resulting
cast Segmental Box Girders from eccentric loading or geometry of structure.
2.4.33 (N) Inspection of Precast Segmental (4) Deflections
Box Girder Jointing Procedures
Deflection calculations shall consider dead load,
2.4.33 (0) Epoxy Bonding Agent Tests
live load, prestressing, erection loads, concrete
creep and shrinkage, and steel relaxation.
Deflections shall be calculated prior to manu-
1.6.25 Precast Segmental Box Girders facture of segments, based on the anticipated pro-
(A) General duction and erection schedules. Calculated deflec-
Except as otherwise noted in this section, the tions shall be used as a guide against which erected
provisions of Section 6 - Prestressed Concrete deflection measurements are checked.
shall apply to the analysis and design of precast
segmental box girder bridges. Deck slabs without
transverse post-tensioning shall be designed under
the applicable provisions of Section 5 - Concrete Defined as non-uniform distribution of bending stress over the
Design. cross section.

(5) Details (C) Design of Substructure
(a) Epoxy bonding agents for match-cast joints In addition to the usual substructure design con-
shall be thermosetting 100 percent solid siderations, unbalanced cantilever moments due to
compositions that do not contain solvent or segment weights and erection loads shall be ac-
any non-reactive organic ingredient except for commodated in pier design or with auxiliary struts.
pigments required for coloring. Epoxy bond- Erection equipment which can eliminate these un-
ing agents shall be of two components, a balanced moments may be used.
resin and a hardener. The two components
shall be distinctly pigmented, so that mixing
produces a third color similar to the concrete
in the segments to be joined, and shall be
1.6.25 Precast Segmental Box Girders
packaged in pre-proportioned, labeled, ready-
to-use containers. (A) General
Epoxy bonding agents shall be formulated to Material strengths and allowable stresses need be
provide application temperature ranges which no different from other prestressed concrete
will permit erection of match-cast segments at bridges; therefore, current limits in Standard Speci-
substrate temperatures from 40F (5C) to fications for Highway Bridges should apply. How-
115F (46C). If two surfaces to be bonded ever, higher strength concrete has advantages and
have different substrate temperatures, the should be used when available. Higher strength
adhesive applicable at the lower temperature concrete has more durability, not only because of
shall be used. the mix design but also because of the greater
If a project would require or benefit from quality control required to produce it.
erection at concrete substrate temperatures Precast segmental box girders may be designed
lower than 4OF, the temperature of the by beam theory with consideration of shear lag.
concrete to a depth of approximately 3 in. Shear lag need only be investigated for segments
(76 mm) should be elevated to at least 40F to wider than 40 ft. (12m) used on 150 ft. (46m)
insure effective wetting of the surface by the spans or less, because of the shallow depth.
epoxy compound and adequate curing of the
(B) Design of Superstructure
epoxy compound in a reasonable length of
time. An artificial environment will have to be Influence surfaces for design of constant and
provided to accomplish this elevation in variable depth deck slabs have been published
temperature and should be created by an (see References 5 and 6, page 109).
enclosure heated by circulating warm air or
by radiant heaters. In any event, localized The following limitations are recommended:
heating shall be avoided and the heat shall be 1. When beam theory is used, single cell boxes
provided in a manner that prevents surface should be no more than 40 ft. (12m) wide, includ-
temperatures greater than 11OF (43C) during ing cantilevers. For bridges wider than 40 ft.,
the epoxy hardening period. Direct flame jett- multiple box cross sections or multiple cell boxes
ing of concrete surfaces shall be prohibited. are usually used. Single cell boxes of width greater
Epoxy bonding agents shall be insensitive to than 40 ft. can be used if carefully analyzed for
damp conditions during application and, after shear lag to determine the portion of cross section
curing, shall exhibit high bonding strength capable of handling longitudinal moment.
to cured concrete, good water resistivity, low 2. For maximum economy, the span-to-depth
creep characteristics and tensile strength ratio for constant depth structures should be 18
greater than the concrete. In addition, the to 20. However, span-to-depth ratios of 20 to 30
epoxy bonding agents shall function as a have been used when required for clearances or
lubricant during the joining of the match-cast esthetics. The shallower depths require the use of
segments being joined, and as a durable, more high strength post-tensioning steel which may
watertight bond at the joint. See Article 2.4.33 cause congested cross sections. Variable depth
(M) for epoxy bonding agent specifications. structures usually have span-to-depth ratios of 18
to 20 at the supports and 40 to 50 at midspan.
(b) Articles 1.6.24 (C) and 1.6.24 (F) relating to
flange thickness and diaphragms shall not 3. Width-to-depth ratios should also be consid-
apply to precast segmental box girders. ered. A shallow box girder that is too wide begins

to behave as a slab. No criteria have been estab- 4. Maisel, V. I., and Roll, F., Methods of Analysis
lished, but when the width-todepth ratio is greater and Design of Concrete Boxbeams with Side
than six, considering the total width of the section Cantilevers, Technical Report No. 42.494,
including slab cantilevers, it is recommended that Cement and Concrete Association, 52 Grosvenor
the designers use multiple cell boxes or carefully Gardens, London, SWlW OAQ, November, 1974.
analyze the cross section. 5. Pucher, Adolph, Influence Surfaces of Elastic
4. Proper fillets should be used in the cross sec- Plates, 4th Edition, 1973 (English), Springer
tion to allow stress transfer around the box per- - Verlag New York, Inc.
imeter and to provide ample room for the large
6. Homberg, Helmut, Double Webbed Slabs,
number of tendons.
(Dalles Nervurees Platten Mit Zwei Stegen),
5. Diaphragms should be considered. These are 1974 (English), Springer - Verlag New York,
usually required only at piers, abutments, and ex- Inc.
pansion joints.
6. The thickened bottom slab in pier segments,
when required for stresses, should taper down or 2.4.33 Prestressed Concrete
step down to the minimum midspan segment bot-
tom slab thickness in as short a distance as is prac- (L) Precast Segment Manufacture and Erection
tical. (1) Manufacture of segments
7. Web thicknesses should be chosen for pro- Each segment shall be match-cast with its ad-
duction ease. If post-tensioning anchorages are jacent segments to ensure proper fit during erec-
located in the webs, web thickness may be gov- tion. As the segments are match-cast they must be
erned by the anchorage requirements. precisely aligned to achieve the final structure
8. Permanent access holes into the box section geometry. During the alignment, adjustments to
should be limited in size to the minimum func- compensate for deflections are made.
tional dimension and should be located near points All tendon ducts are placed during production.
of minimum stress. The conduit to enclose grouted, post-tensioned
tendons shall be mortar tight, made of galvanized,
(C) Design of Substructure
ferrous metal, and may be either rigid with a
Unbalanced cantilever moments occur during smooth inner wall, capable of being curved to the
erection only and are usually greater in magnitude proper configuration, or a flexible, interlocking
than service load moments. Wind loads in combi- type. Couplers for either type shall also provide a
nation with erection loads could develop critical mortar tight connection. Rigid conduit may be
stresses and, thus, wind loads should be consid- fabricated with either welded or interlocking
ered in accordance with Article 1.2.22. seams. Galvanizing of welded seams for rigid con-
duit or of conduit couplers will not be required.
During placing and finishing of concrete in a seg-
Selected References ment, inflatable hoses capable of exerting suffi-
cient pressure on the inside walls shall be placed
The following selected references provide some internally in all conduits and shall extend a mini-
useful guidelines in the design and construction of mum of 2 ft. (0.6m) into the conduit in the pre-
precast prestressed segmental box girder bridges: viously cast segment. Either type of conduit shall
1. PCI Committee on Segmental Construction, be capable of withstanding all forces due to con-
Recommended Practice for Segmental Con- struction operations without damage. Other types
struction in Prestressed Concrete, PCI JOUR- of conduit and/or internal protection systems are
NAL, V. 20, No. 2, March-April 1975, pp. 22- permitted subject to the approval of the Engineer.
41. (2) Erection of Segments
2. Muller, Jean, Ten Years of Experience in Pre- Segments are usually erected by the cantilever
cast Segmental Construction, PCI JOURNAL, method from each pier without falsework, al-
V. 20, No. 1, January-February 1975, pp. 28-61. though temporary supports may be used. With
3. Swann, R. A., A Feature Survey of Concrete the approval of the Engineer, other systems of
Box Spine-Beam Bridges, Cement and Concrete erection may be considered.
Association, 52 Grosvenor Gardens, London Match-cast segments shall be erected using
SW1 W OAQ, 1972. epoxied joints. Pressure shall be provided on the

joint by means of post-tensioning. The pressure surface dry (no visible water).
shall be as uniform as possible with a minimum of Instructions furnished by the supplier for the
30 psi (0.21 MPa) at any point. safe storage, mixing and handling of the epoxy
Deflections of cantilevers shall be measured as bonding agent shall be followed. The epoxy shall
erection progresses and compared with computed be thoroughly mixed until it is of uniform color.
deflections. Any deviation from the required align- Use of a proper sized mechanical mixer operating
ment shall be corrected by either modifying the at no more than 600 RPM will be required. Con-
segment geometry during the casting operation or tents of damaged or previously opened containers
by inserting stainless steel screen wire shims in the shall not be used. Mixing shall not start until the
epoxy joints during erection. The maximum thick- segment is prepared for installation. Application of
ness of shims at any joint shall be l/16 in. (1.6mm). the mixed epoxy bonding agent shall be according
Provision shall be made to permit alignment ad- to the manufacturers instructions using trowel,
justments of a completed cantilevered portion of rubber glove or brush on one or both surfaces to
the box girder before the midspan splice connect- be joined. The coating shall be smooth and uni-
ing adjacent cantilevers is constructed. form and shall cover the entire surface with a mini-
(3) Grouting mum thickness of l/16 in. (1.6mm) applied on
both surfaces or l/8 in. (3.2mm) if applied on one
Grouting of the ducts shall be done in accord-
surface. Epoxy should not be placed within 3/8 in.
ance with Article 2.4.33 (I). Under normal condi-
(9.5mm) of prestressing ducts to minimize flow
tions, grouting shall be accomplished within 20
into the ducts. A discernible bead line must be ob-
calendar days following installation of tendons.
served on all exposed contact areas after tempo-
For delays beyond 20 days, tendons shall be pro-
rary post-tensioning. Erection operations shall be
tected with a water soluble oil or approved equal
coordinated and conducted so as to complete the
protective agent.
operations of applying the epoxy bonding agent
Protection of the tendon ducts against splitting
to the segments, erection, assembling, and tem-
from freezing of water in ducts must be provided
porary post-tensioning of the newly joined segment
until cement grout can be used. Use of some other
within 70 percent of the open time period of the
type grout should be considered when erecting in
bonding agent.
these low temperatures.
The epoxy material shall be applied to all sur-
(M) Epoxy Bonding Agents for Precast Segmental faces to be joined within the first half of the gel
Box Girders time, as shown on the containers. The segments
All epoxy bonding agents shall meet the require- shall be joined within 45 minutes after applica-
ments of Article 1.6.25 (B) (5) (a). Two-part tion of the first epoxy material placed and a mini-
epoxy bonding agents shall be supplied to the erec- mum average temporary prestress of 50 psi (0.35
tion site in sealed containers, pre-proportioned MPa) over the cross section should be applied with-
in the proper reacting ratio, ready for combining in 70 percent of the open time of the epoxy mater-
and through mixing in accordance with the manu- ial. At no point of the cross section shall the tem-
facturers instructions. All containers shall be porary prestress be less than 30 psi (0.21 MPa).
properly labeled to designate the resin component The joint shall be checked immediately after
and the hardener component as well as the tem- erection to verify uniform joint width and proper
perature range for its application. The substrate fit. Excess epoxy from the joint shall be removed
temperature range of 40F to 115F (5C to 46C) where accessible. All tendon ducts shall be swabbed
may be divided into either two or three applica- immediately after stressing, while the epoxy is
tion ranges for bonding agents. Such ranges shall still in the non-gelled condition, to remove or
overlap each other by at least 6F (3C). smooth out any epoxy in the conduit and to seal
Surfaces to which the epoxy material is to be any pockets or air bubble holes that have formed
applied shall be free from oil, laitance, form re- at the joint.
lease agent, or any other material that would pre- If the jointing is not completed within 70 per-
vent the material from bonding to the concrete cent of the open time, the operation shall be ter-
surface. All laitance and other contaminants shall minated and the epoxy bonding agent shall be
be removed by light sandblasting or by high pres- completely removed from the surfaces. The sur-
sure water blasting with a minimum pressure of faces must be prepared again and fresh epoxy shall
5000 psi (35 MPa). Wet surfaces should be dried be applied to the surface before resuming jointing
before applying epoxy bonding agents. The sur- operations.
face should be at least the equivalent of saturated As general instructions cannot cover all situa-

tions, specific recommendations and instructions Specification: 30 minutes minimum on one
shall be obtained in each case from the Engineer quart (0.95Q) and one gallon (3.79Q) quantities
in charge. at the maximum temperature of the designated
Epoxy bonding agents shall be tested to deter- application temperature range. (Note: gel time is
mine their workability, gel time, open time, bond not to be confused with open time specified in
and compression strength, shear, and working tem- Test 3).
perature range. See Article 2.4.33 (0) for test Test 3 - Open Time of Bonding Agent
methods and recommended specification limits.
This test measures workability of the epoxy
The frequency of the tests shall be stated in the
bonding agent for the erection and post-tensioning
Special Provisions of the Contract.
operations. As tested here, open time is defined as
The Contractor shall furnish the Engineer sam-
the minimum allowable period of elapsed time
ples of the material for testing, and a certification
from the application of the mixed epoxy bonding
from a reputable independent laboratory indicating
agent to the precast segments until the two seg-
that the material has passed the required tests.
ments have been assembled together and tempo-
(N) Inspection of Precast Segmental Box Girder rarily post-tensioned.
Jointing Procedures Testing Method: Open time is determined using
In addition to the material acceptance tests, test specimens as detailed in the Tensile Bending
which should be initially performed by a neutral Test (Test 4). The epoxy bonding agent, at the
testing laboratory and then checked by the owners highest specified application temperature, is mixed
organization, the owners inspector should make together and applied as instructed in Test 4 to the
regular checks of the epoxy jointing procedures. concrete prisms which shall also be at the highest
Data such as weather, ambient temperature, con- specified application temperature. The adhesive
crete surface temperature, adhesive batch number, coated prisms shall be maintained for 60 minutes
and the jointing time should be noted. The inspec- at the highest specified application temperature
tor should frequently sample and record data such with the adhesive coated surface or surfaces ex-
as the observed gel time of the epoxy bonding posed and uncovered before joining together.
agent, the surface conditions of the segments The assembled prisms are then cured and tested as
being joined, the adequacy of coverage of the ad- instructed in Test 4.
hesive, the amount of material being squeezed Specification: The epoxy bonding agent is ac-
from the joints, and the approximate open time of ceptable for the specified application temperature
the epoxy. An approximate determination of the only when essentially total fracturing of concrete
open time can be noted from behavior of lap joint paste and aggregate occurs with no evidence of
samples spread on small cement-asbestos boards. adhesive failure.
(0) Epoxy Bonding Agent Tests Construction situations may sometimes require
application of the epoxy bonding agent to the
Test 1 - Sag Flow of Mixed Epoxy Bonding Agent precast section prior to erecting, positioning and
This test measures the application workability of assembling. This operation may require epoxy
the bonding agent. bonding agents having prolonged open time. In
Testing Method: ASTM D 2730 for the desig- general, where the erection conditions are such
nated temperature range. that the sections to be bonded are prepositioned
Specification: Mixed epoxy bonding agent must prior to epoxy application, the epoxy bonding
not sag flow at l/8 in. (3.2mm) minimum thick- agent shall have a minimum open time of 60 min-
ness at the designated minimum and maximum utes within the temperature range specified for
application temperature range for the class of its application.
bonding agents used.
Test 4 - Three Point Tensile Bending Test
Test 2 - Gel Time of Mixed Epoxy Bonding Agent This test, performed on a pair of concrete
Gel time is determined on samples mixed as prisms bonded together with epoxy bonding agent,
specified in the testing,method. It provides a guide determines the bonding strength between the
for the period of time the mixed bonding agent bonding agent and concrete. The bonded concrete
remains workable in the mixing container and dur- prisms are compared to a reference test beam of
ing which it must be applied to the match-cast jo concrete 6x6~18 in. (150x150x460mm).
joint surfaces. Testing Method: 6x6x9 in. (150xl50x230mm)
Testing Method: ASTM D 2471 (except that concrete prisms of 6000 psi (41 MPa) compressive
one quart and one gallon quantities shall be tested). strength at 28 days shall be sandblasted on one 6x6

in. side to remove mold release agent, laitance, tion temperature range.
etc., and submerged in clean water at the lower Testing Method: ASTM D 648.
temperature of the specified application tempera- Specification: A minimum deflection tempera-
ture range for 72 hours. Immediately on removing ture of 122F (5OC) at fiber stress loading of 264
the concrete prisms from the water, the sand- psi (1.8 MPa) is required on test specimens cured
blasted surfaces shall be air dried for one hour at 7 days at 77F (25C).
the same temperature and 50 percent RH and each Test 7 - Compression and Shear Strength of
shall be coated with approximately a l/16 in. (1.6
Cured Epoxy Bonding Agent
mm) layer of the mixed bonding agent. The adhe-
sive coated faces of two prisms shall then be placed This test is a measure of the compressive strength
together and held with a clamping force normal to and shear strength of the epoxy bonding agent
the bonded interface of 50 psi (0.35 MPa). The as- compared to the concrete to which it bonds. The
sembly shall then be wrapped in a damp cloth slant cylinder specimen with the epoxy bonding
which is kept wet during the curing period of agent is compared to a reference test cylinder of
24 hours at the lower temperature of the specified concrete only.
application temperature range. Testing Method: A test specimen of concrete
After 24 hours curing at the lower temperature is prepared in a standard 6x12 in. (15Ox300mm)
of the application temperature range specified for cylinder mold to have a height at midpoint of 6 in.
the epoxy bonding agent, the bonded specimen and an upper surface with a 30-degree slope from
shall be unwrapped, removed from the clamping the vertical. The upper and lower portions of the
assembly and immediately tested. The test shall specimen with the slant surfaces may be formed
be conducted using the standard ASTM C78 test through the use of an elliptical insert or by sawing
for flexural strength with third point loading and a full sized 6x12 in. cylinder. If desired, 3x6 in.
the standard MR unit. At the same time the two (75xl50mm) or 4x8 in. (lOOx200mm) specimens
prisms are preapred and cured, a companion test may be used. After the specimens have been moist
beam shall be prepared of the same concrete, cured for 14 days, the slant surfaces shall be pre-
cured for the same period and tested following pared by light sandblasting, stoning or acid etching,
ASTM C78. then washing and drying the surfaces, and finally
Specification: The epoxy bonding agent is ac- coating one of the surfaces with a 10 mil (0.25mm)
ceptable if the load on the prisms at failure is thickness of the epoxy bonding agent under test.
greater than 90% of the load on the reference test The specimens shall then be pressed together
beam at failure. and held in position for 24 hours. The assembly
shall then be wrapped in a damp cloth which shall
Test 5 - Compression Strength of Cured Epoxy be kept wet during an additional curing period of
Bonding Agent 24 hours at the minimum temperature of the
This test measures the compressive strength of designated application temperature range. The
the epoxy bonding agent. specimen shall then be tested at 77F (25C) follow-
Testing Method: ASTM D 695. ing ASTM C 39 procedures. At the same time as
Specification: Compressive strength at 77F the slant cylinder spcimens are made and cured, a
(25C) shall be 2000 psi (14 MPa) minimum after companion standard test cylinder of the same con-
24 hours cure at the minimum temperature of the crete shall be made, cured for the same period,
designated application temperature range and 6000 and tested following ASTM C 39.
psi (41 MPa) at 48 hours. Specification: The epoxy bonding agent is ac-
Test 6 - Temperature Deflection of Epoxy Bond- ceptable for the designated application tempera-
ture range if the load on the slant cylinder speci-
ing Agent
ment is greater than 90 percent of the load on the
This test determines the temperature at which companion cylinder. The bond strength on the
an arbitrary deflection occurs under arbitrary slant surface (shear), determined by dividing the
testing conditions in the cured epoxy bonding specimen test load by the area of the elliptical slant
agent. It is a screening test to establish perform- surface, shall be at least 3000 psi (21 MPa) at 48
ance of the bonding agent throughout the erec- hours.

A.2 Summary of Precast Segmental Concrete
Bridges in the United States and Canada With
Cross Sections
Note: for metric dimensions
1 ft. = 0.3048 m
1 in. = 25.4 mm

Fig. A.2.3 Corpus Christi, Texas

Spans: 100 feet - 200 feet - 100 feet
Bridge Length: 400 feet
Two Segments Wide
Segment Length: 10 feet

l- 1

Fig. A.2.1 Lievre River Bridge, Quebec

Spans: 130 feet - 260 feet - 120 feet
Bridge Length: 520 feet
Segment Length: 9 feet 6 inches
,.W 20.0

Fig. A.2.4 Vail Pass, Colorado

End Spans: 160 feet
c ROADWAY Main Spans: 210 feet
t- Segment Length: 7 feet 4 inches

c 4
1'.ll'%" 3,-o" l'e 2'-(1 6'-6"
-. I- -I " .

Fig. A.2.2 Bear River Bridge, Nova Scotia Fig. A.2.5 North Vernon, Indiana
End Spans: 203 feet 9 inches Over Muscatatuck River
Interior Spans: 265 feet Spans: 95 feet - 190 feet - 95 feet
Bridge Length: 1997.50 feet Bridge Length: 380 feet
Segment Length: 14 feet 2 inches Segment Length: 8 feet

Fig. A.26 Kishwaukee River Bridge, Illinois , -9 lo.0

End Spans: 170 feet

interior Spans: 250 feet Fig. A.2.9 Pike County, Kentucky
Northbound : 1090 feet Bridge Length: 372 feet
Southbound: 1090 feet Spans: 93.5 feet - 185 feet - 93.5 feet
Segment Length: 7 feet O-5/8 inches Segment Length: 7 feet 10 inches

Fig. A.2.7 Parke County, Indiana

Bridge Length: 276 feet Fig. A.2.10 Lake Oahe Crossing Missouri River,
Spans: 90 feet - 180 feet - 90 feet North Dakota
Segment Length: 8 feet Bridge Length: 3020 feet
Spans: 179 feet - 10 @ 265 feet - 179 feet
Segment Length: 8 feet 4 inches


Fig. A.2.8 Turkey Run, Indiana Fig. A.2.11 Scottdale Bridge, Michigan
Bridge Length: 322 feet Bridge Length: 407 feet
Spans: 180 feet - 180 feet Spans: 97 feet - 206.5 feet - 97 feet
Segment Length: 8 feet Segment Length: 8 feet

Fig. A.2.15 Akron Bridge, Ohio
Westbound: 3660 feet
Eastbound: 3646 feet
Spans: Variable 100 to 290 feet
Segment Length: 6,7 and 8 feet
Fig. A.2.12 Illinois River, Illinois
Eastbound: 3329.5 feet
Westbound: 3203.5 feet
Approach Spans: 175 feet - 230 feet
Main Spans: 390 feet - 550 feet - 390 feet
Segment Length: 10 feet

Fig. A.2.16 Lake County Ramps, Indiana

Bridge Length: 6240 feet
(Ramps plus Mainline)
Spans: Variable 100 to 315 feet
Fig. A.2.13 Zilwaukee, Michigan
Segment Length: 7 feet 9 inches
Bridge Over Saginaw River
Bridge Length: 8000 feet
Two segments Wide
Spans: Variable 155-392 feet
Segment Length: 8 feet to 12 feet

l-3-1,4 ,4.11,16

Fig. A.2.14 St. Louis Missouri

Bridge Length: 405 feet
Spans: 100 feet - 200 feet - 100 feet
Segment Length: 9 feet 4 inches

A.3 Notation c, = distance from centroid to top fiber
A = cross sectional area of segment c,, = distance from centroid to bottom fiber
d = web thickness
A = area of top slab
A, = area of concrete section d, = slab thickness
= d = thickness of top and bottom slab (Fig.
E 28-day modulus of elasticity of concrete
Fi = initial prestressing force 3.38)
e = eccentricity of post-tensioning force
Ff = final prestressing force
= base of natural logrithms = 2.718. . .
H = horizontal distance center to center of
;I, = 28-day compressive strength of concrete
I = moment of inertia test cylinders
L = span length f, = concrete stress
L = unit length along span f cb = bottom fiber compressive stress
h = horizontal displacement
M, = transverse moments
Mcr = creep moment resulting from change of h th = theoretical thickness of structural element
statical system with respect to relative humidity
II = cantilever length (Fig. 3.8)
ME = erection moment
= uniformly distributed load (Fig. 3.8)
M, = moment due to loads before change of q
t = theoretical age
statical system (Fig. 3.10)
t = theoretical time after casting (days)
MI, = moment due to same loads, considered to
produce M,, applied to changed statical t, = theoretical age of concrete at time of
system (Fig. 3.10) loading (days)
M, = moment at time t t, = rate of change of torsional shear force in
M, = torsional moment per unit length of box top and bottom slabs
girder t, = rate of change of torsional shear force in
N, = ratio of longitudinal forces obtained from webs
computer analysis to forces obtained from t, = time of completion of the structure
elementary beam theory t, = time of application of the dead load
= V = vertical displacement
P post-tensioning force
P = load causing deflection 6 (Fig. 3.11) w = unit weight of concrete
= Z = vertical dimension from centerline of box
P loading per unit length (Fig. 3.35)
R = reaction before settlement (Fig. 3.11) section to centerline of slab
R = support reactions (Fig. 3.36) AS, = member elongation due to shrinkage re-
Sh = horizontal shear force in transverse analysis straint force, St
s, = E,h EA, elastic shrinkage restraint force A f,ht 8, member shortening due to shrink-
sht =

St = shrinkage restraint force adjusted for the age at time t

effect of creep At = temperature differential
s, = vertical shear force in transverse analysis At = number of days at ambient temperature T
Tw = ambient temperature during At days a = elastic angle change at end of cantilever
T, = shear force in top slab (Fig. 3.8)
T2 = shear force in web a = a factor used in determining theoretical
T3 = shear force in bottom slab age related to the type of cement used
T, = rate of change of shear force in top slab a = coefficient of linear thermal expansion
T2 = rate of change of shear force in web a = rotation of forward cantilever arm adja-
T, = rate of change of shear force in bottom cent to end span (Fig. 3.60)
slab = angle change due to restraint moment
T,, = shear forces in top and bottom slab (Fig. 3.8)
T, = portion of external load carried to sup- = rotation of corner of box section
ports by webs = rotation of forward cantilever arm adja-
v = vertical distance from center of top slab cent to interior span (Fig. 3.60)
to center of bottom slab = factor reflecting the influence of the rela-
B Cl
x, = increase in support reaction due to elas- tive humidity of the ambient medium and
tic and creep deformation (Fig. 3.11) the composition of the concrete on Qr
2, = top section modulus P c2
= factor reflecting the influence of the rela-
2, = bottom section modulus tive humidity of the ambient medium and
b = horizontal dimension from centerline of the theoretical thickness of the concrete
box section to centerline of web hth On @r

= factor variable from zero to unity indi- a = stress
cating the variation of #d with time 7 = maximum shear stress in bottom slab
rClfW - Brct,, = factor variable from zero to unity [Fig. 3.38(a)]
indicating the variation of $f with 7 = torsional shear stress
time 4 = ecr/ee, creep factor, also = @d + @f
6 = elastic deflection (Fig. 3.8) @t = E,,/E~ at time t
EC, = creep strain @d = creep due to delayed elasticity or re-
E, = elastic strain coverable creep on removal of load
Esh = shrinkage strain at infinity Of = creep due to flow, not recoverable
Esht = shrinkage strain at time t ~kt,) = magnitude of the creep factor at time t for
x = factor used in determining the theore- a concrete specimen loaded at time t,
tical thickness hth (Table 3.1) @d, = magnitude of delayed elasticity at
P = perimeter of concrete section in contact infinity
with the atmosphere @f, = magnitude of flow at infinity

A. 4 References 12. Leonhardt, F., and Lipproth, W. Conclusions Drawn

from Distress of Prestressed Concrete Bridges Beton-
1. Kashima, S., and Breen, J. E., Construction and
und Stahlbetonbau, No. 10, Vol. 65, pp. 231-244,
Load Tests of a Segmental Precast Box Girder Bridge
Berlin, October, 1970 (in German).
Model Research Report 121-5 (s), Center for High-
13. Brown, R. C., Jr., Burns, N. H., and Breen, J. E.,
way Research, The University of Texas at Austin,
February, 1975. Computer Analysis of Segmentally Erected Pre-
2. cast Prestressed Box Girder Bridges, Research Re-
Post-Tensioning Institute, Post-Tensioning Manual,
port 121-4, Center for Highway Research, The Uni-
Glenview, Illinois, 1976.
versity of Texas at Austin, November, 1974.
3. Ruesch, H., and Kupfer, H., Bemessung von Spann-
14. PCI Committee on Segmental Construction, Rec-
betonbauteilen, Chapter M, Beton-Kalender, Wil-
ommended Practice for Segmental Construction in
helm Ernst and Son, Berlin, 1977 (in German).
Prestressed Concrete, Journal of the Prestressed
4. Lin, T. Y., Design of Prestressed Concrete Structures,
Concrete Institute, Vol. 20, No. 2, March - April,
Second Edition, John Wiley & Sons, Inc., New York,
5. Leonhardt, F., Prestressed Concrete Design and Con- 15. Prestressed Concrete Institute, Manual for Quality
struction, Second Edition, Wilhelm Ernst & Sons, Control for Plants and Production of Precast Pre-
Berlin, Munich, 1964. stressed Concrete Products, Prestressed Concrete In-
6. American Association of State Highway and Trans- stitute, Chicago, 1977.
portation Officials, Standard Specification for High- 16. Portland Cement Association, Design and Control of
way Bridges, Twelfth Edition, 1977, American Asso- Concrete Mixtures, Eleventh Edition, Portland Ce-
ciation of State Highway and Transportation Officials, ment Association, Skokie Illinois, 1968.
Washington, D.C. 17. Comite Europeen du Beton /FIP, Bulletin dinfor-
7. Muller, Jean, Ten Years of Experience in Precast mation No. 111 October, 1975.
Segmental Construction, Journal of the Prestressed 18. Freyermuth, Clifford L., Design of Continuous
Concrete Institute, Vol. 20, No. 1. January - Feb- Highway Bridges with Precast, Prestressed Concrete
ruary, 1975. Girders, Journal of the Prestressed Concrete Insti-
8. Scordelis, A. C., Analysis of Continuous Box Girder tute, Vol. 14, No. 2, March - April, 1969, pp. 14-39.
Bridges, SESM 67-25, Department of Civil Engineer-
ing, University of California, Berkeley, November, 19. Post-Tensioning Institute, Post-Tensioned Box Girder
1967. Bridge Manual, Post-Tensioning Institute, Glenview,
9. Homberg, Helmut, Fahrbahnplatten. mit Verand- Illinois, 1978.
lither Dicke Springer-Verlag, New York, 1968. 20. Libby, James R., Long Span Precast, Prestressed
10. Homberg, Helmut, and Ropers, Walter, Fahrbahn- Girder Bridges, Journal of the Prestressed Concrete
platten mit Verandlicher Dicke, Springer-Verlag, Institute, Vol. 16, No. 4, July - August, 1971, pp.
New York, 1965. 80-98.
11. Muller, Jean, Concrete Bridges Built in Cantilever, 21. Freyssinet International, Precast Segmental Can-
Societe des lngenieurs Civils de France, British Sec- tilever Bridge Construction, Technical Brochure,
tion, 1963. May, 1973.