OPTIMIZATION OF SPAN-TO-DEPTH RATIOS IN HIGH-STRENGTH CONCRETE GIRDER BRIDGES

by

Sandy Shuk-Yan Poon

A thesis submitted in conformity with the requirements for the degree of Master of Applied Science Graduate Department of Civil Engineering University of Toronto

Copyright by Sandy Shuk-Yan Poon (2009)

Optimization of Span-to-Depth Ratios in High-Strength Concrete Girder Bridges Sandy Shuk-Yan Poon Master of Applied Science Graduate Department of Civil Engineering University of Toronto 2009

ABSTRACT
Span-to-depth ratio is an important bridge design parameter that affects structural behaviour, construction costs and aesthetics. A study of 86 constant-depth girders indicates that conventional ratios have not changed significantly since 1958. These conventional ratios are now questionable, because recently developed high-strength concrete has enhanced mechanical properties that allow for slenderer sections. Based on material consumption, cost, and aesthetics comparisons, the thesis determines optimal ratios of an 8-span highway viaduct constructed with high-strength concrete. Three bridge types are investigated: cast-in-place on falsework box-girder and solid slabs, and precast segmental span-byspan box-girder. Results demonstrate that total construction cost is relatively insensitive to span-todepth ratio over the following ranges of ratios: 10-35, 30-45, and 15-25 for the three bridge types respectively. This finding leads to greater freedom for aesthetic expressions because, compared to conventional values (i.e. 18-23, 22-39, and 16-19), higher ranges of ratios can now be selected without significant cost premiums.

ii

ACKNOWLEDGEMENTS
I would like to express my deepest gratitude to my supervisor, Professor Paul Gauvreau, whose encouragement, guidance, and support enabled me to complete this thesis. I am also indebted to my research colleagues for their insightful advice and assistance throughout my graduate studies: Cathy Chen, Billy Cheung, Davis Doan, Negar Elhami Khorasani, Eileen Li, Kris Mermigas, Jason Salonga, Jimmy Susetyo, Brent Visscher, and Ivan Wu. Lastly, I would like to thank my family for their support and encouragement over these past two years.

iii

TABLE OF CONTENTS
Abstract .............................................................................................................................................. ii Acknowledgements........................................................................................................................... iii Table of Contents ............................................................................................................................. iv List of Figures ................................................................................................................................. viii List of Tables .................................................................................................................................... xi List of Symbols ............................................................................................................................... xiii 1 Introduction ............................................................................................................................... 1 1.1 1.2 1.3 2 The Significance of Optimizing Span-to-Depth Ratio ........................................................ 1 Objectives and Scope .......................................................................................................... 5 Thesis Structure................................................................................................................... 6

Typical Span-to-Depth Ratios of Existing Bridges................................................................. 7 2.1 2.2 2.3 2.4 Cast-in-Place Box-Girder .................................................................................................... 7 Cast-in-Place Slab ............................................................................................................. 12 Precast Segmental Box-Girder .......................................................................................... 16 Concluding Remarks ......................................................................................................... 18

Analysis Overview ................................................................................................................... 19 3.1 3.2 Analysis Model ................................................................................................................. 19 Materials............................................................................................................................ 21 Prestressing Tendons ................................................................................................. 21 Concrete Covers ........................................................................................................ 22

3.2.1 3.2.2 3.3

Loads ................................................................................................................................. 22 Load Combinations and Load Factors ...................................................................... 22 Live Loads................................................................................................................. 23

3.3.1 3.3.2 3.4

Design Requirements ........................................................................................................ 24 Ultimate Limit States Design Requirements ............................................................. 24 Flexural Strength ................................................................................................... 24 Shear Strength ....................................................................................................... 25 Serviceability Limit States Design Requirements ..................................................... 26 Stress ..................................................................................................................... 26 Vibration ............................................................................................................... 26 Deflection .............................................................................................................. 27 iv

3.4.1 3.4.1.1 3.4.1.2 3.4.2 3.4.2.1 3.4.2.2 3.4.2.3

3.5 4

Other Preliminary Analysis Assumptions ......................................................................... 27

Analysis of Cast-in-Place on Falsework Bridges .................................................................. 28 4.1 4.2 Cast-in-Place on Falsework Construction ......................................................................... 28 Cast-in-Place on Falsework Box-Girder ........................................................................... 28 Model ........................................................................................................................ 28 Cross-Section ........................................................................................................ 29 Prestressing Tendon Layout .................................................................................. 30 Analysis Results ........................................................................................................ 31 Structural Behaviour and Dimensioning ............................................................... 31 Vibration Limits .................................................................................................... 32 Deflections ............................................................................................................ 33 Material Consumption........................................................................................... 34 Limiting Factors of Span-to-Depth Ratios ............................................................ 36

4.2.1 4.2.1.1 4.2.1.2 4.2.2 4.2.2.1 4.2.2.2 4.2.2.3 4.2.2.4 4.2.2.5 4.3

Cast-in-Place on Falsework Solid Slab ............................................................................. 38 Model ........................................................................................................................ 39 Cross-Section ........................................................................................................ 39 Prestressing Tendon Layout .................................................................................. 39 Strip Method versus Beam Model ............................................................................. 40 Analysis Results ........................................................................................................ 42 Structural Behaviour and Dimensioning ............................................................... 42 Maximum Reinforcement Criterion ...................................................................... 43 Vibration Limits .................................................................................................... 44 Deflections ............................................................................................................ 45 Material Consumption........................................................................................... 46 Limiting Factors of Span-to-Depth Ratios ............................................................ 47

4.3.1 4.3.1.1 4.3.1.2 4.3.2 4.3.3 4.3.3.1 4.3.3.2 4.3.3.3 4.3.3.4 4.3.3.5 4.3.3.6 5

Analysis of Precast Segmental Span-by-Span Box-Girder .................................................. 48 5.1 5.2 Precast Segmental Span-by-Span Construction ................................................................ 48 Model ................................................................................................................................ 49 Cross-Section ............................................................................................................ 50 Elevation and Prestressing Tendon Layout ............................................................... 50

5.2.1 5.2.2 5.3

Longitudinal Bending Moments ....................................................................................... 51 Construction Moments .............................................................................................. 51 Moments due to Thermal Gradient ........................................................................... 54 v

5.3.1 5.3.2

5.4

Loss of Prestress................................................................................................................ 57 Friction Losses .......................................................................................................... 57 Creep and Shrinkage Losses ..................................................................................... 58 Losses due to Relaxation of Prestressing Steel ......................................................... 59 Total Prestress Losses ............................................................................................... 59

5.4.1 5.4.2 5.4.3 5.4.4 5.5 5.6

Behaviour of Unbonded Tendons at Ultimate Limit States .............................................. 60 Analysis Results ................................................................................................................ 61 Structural Behaviour and Dimensioning ................................................................... 61 Vibration Limits ........................................................................................................ 61 Deflections ................................................................................................................ 62 Material Consumption............................................................................................... 63 Limiting Factors of Span-to-Depth Ratios ................................................................ 64

5.6.1 5.6.2 5.6.3 5.6.4 5.6.5 6

Cost Comparisons ................................................................................................................... 65 6.1 Material Costs ................................................................................................................... 65 Material Unit Prices .................................................................................................. 65 Concrete Material Unit Price ................................................................................ 65 Cast-in-Place versus Precast Concrete .................................................................. 66 Falsework versus Erection Truss........................................................................... 67 Formwork .............................................................................................................. 67 Prestressing Tendons............................................................................................. 67 Material Cost Comparisons ....................................................................................... 67 Concrete Cost Comparison ................................................................................... 68 Prestressing Cost Comparison............................................................................... 69 Reinforcing Steel Cost Comparison ...................................................................... 70 Total Superstructure Cost ...................................................................................... 73

6.1.1 6.1.1.1 6.1.1.2 6.1.1.3 6.1.1.4 6.1.1.5 6.1.2 6.1.2.1 6.1.2.2 6.1.2.3 6.1.2.4 6.2

Overall Construction Costs ............................................................................................... 76 Construction Cost Breakdown .................................................................................. 76 Total Construction Cost Comparison ........................................................................ 77

6.2.1 6.2.2 6.3 6.4

Other Cost Factors ............................................................................................................ 78 Sensitivity Analysis........................................................................................................... 79 Sensitivity with Respect to Changes in Material Unit Prices .................................... 79 Sensitivity with Respect to Changes in Construction Cost Breakdown .................... 83

6.4.1 6.4.2 6.5

Concluding Remarks ......................................................................................................... 85 vi

Aesthetics Comparisons .......................................................................................................... 86 7.1 Visual Impact of Span-to-Depth Ratio .............................................................................. 86 Effects of Viewing Points ......................................................................................... 92 Other Factors that Affect Visual Slenderness ........................................................... 94

7.1.1 7.1.2 7.2 7.3 8

Evolution of the Visually Optimal Span-to-Depth Ratio .................................................. 97 Concluding Remarks ....................................................................................................... 102

Conclusions ............................................................................................................................ 103 8.1 8.2 8.3 8.4 8.5 8.6 Conventional Span-to-Depth Ratios ............................................................................... 103 Maximum Span-to-Depth Ratios .................................................................................... 103 Material Consumption Comparisons............................................................................... 104 Total Construction Cost Comparisons ............................................................................ 104 Aesthetic Comparisons.................................................................................................... 105 Optimal Span-to-Depth Ratios ........................................................................................ 105

Reference........................................................................................................................................ 107 Appendix A: Chapter 2 Supplementary Information ................................................................ 111 A.1 Cast-in-Place on Falsework Box-Girder ............................................................................ 112 A.2 Cast-in-Place on Falsework Solid Slab .............................................................................. 116 A.3 Precast Segmental Span-by-Span Box-Girder.................................................................... 118 Appendix B: Supporting Calculations ........................................................................................ 119 B.1 Flexural Strength for Bonded Tendons at ULS .................................................................. 119 B.2 Shear Strength at ULS ....................................................................................................... 120 B.3 Thermal Gradient Moments................................................................................................ 122 B.4 External Tendon Force ....................................................................................................... 124 B.5 Total Construction Cost ...................................................................................................... 125 Appendix C: Summary of Results ............................................................................................... 126 C.1 Cast-in-Place on Falsework Box-Girder............................................................................. 127 C.2 Cast-in-Place on Falsework Solid Slab .............................................................................. 128 C.3 Precast Segmental Span-by-Span Box-Girder .................................................................... 129 C.4 Sensitivity with Respect to Changes in Construction Cost Breakdown ............................. 130

vii

LIST OF FIGURES
Figure 1-1. Recommended ratios for cast-in-place box-girder ......................................................... 2 Figure 1-2. Recommended ratios for cast-in-place slab .................................................................... 2 Figure 1-3. Recommended ratios for precast segmental box-girder ................................................. 2 Figure 2-1. Span-to-depth ratios of cast-in-place box-girders ........................................................ 10 Figure 2-2. Span-to-depth ratios of cast-in-place box-girders ........................................................ 11 Figure 2-3. Span-to-depth ratios of cast-in-place slabs ................................................................... 13 Figure 2-4. Span-to-depth ratios of cast-in-place slabs ................................................................... 14 Figure 2-5. Span-to-depth ratios of precast segmental box-girders ................................................ 17 Figure 2-6. Span-to-depth ratios of precast segmental box-girders ................................................ 17 Figure 2-7. Span-to-depth ratios for all bridge types ...................................................................... 18 Figure 3-1. Typical plan and elevation ........................................................................................... 19 Figure 3-2. Typical deck arrangement ............................................................................................ 19 Figure 3-3. Summary of analysis cases ........................................................................................... 20 Figure 3-4. Live loads: CL-625 truck load (top); CL-625 lane load (bottom) ................................ 23 Figure 3-5. Flexural resistance: a) cross-section, b) concrete stains, c) equivalent concrete stresses, d) concrete forces .............................................................................................................................. 24 Figure 3-6. Construction cost economy from increasing the number of stirrup spacing ................ 25 Figure 3-7. Deflection limits for highway bridge superstructure vibration (CHBDC 2006) .......... 26 Figure 4-1. Moment comparison of bridges with constant and reduced end span length ............... 29 Figure 4-2. Typical cross-section for cast-in-place on falsework box-girder ................................. 29 Figure 4-3. Typical reinforcing steel layout .................................................................................... 30 Figure 4-4. Typical tendon profile .................................................................................................. 30 Figure 4-5. Changes in sectional modulus and cross-sectional depth ............................................. 32 Figure 4-6. Deflection for superstructure vibration limitation ........................................................ 33 Figure 4-7. Deflections: a) dead load, b) long-term, c) short-term ................................................. 33 Figure 4-8. Material consumptions for cast-in-place on falsework box-girder ............................... 35 Figure 4-9. Tendon arrangement that limits further increase in span-to-depth ratio....................... 36 Figure 4-10. Interior box cavity limitation ...................................................................................... 37 Figure 4-11. Height of access diminishes as span-to-depth ratio increases .................................... 37 Figure 4-12. Concrete reduction due to increase in L/h ratio for solid slab and box-girder ........... 38 Figure 4-13. Voided slab ................................................................................................................. 38 Figure 4-14. Typical cross-section for cast-in-place on falsework solid slab ................................. 39 viii

Figure 4-15. Typical reinforcing steel layout .................................................................................. 39 Figure 4-16. Transverse distribution of longitudinal bending moment in slabs.............................. 41 Figure 4-17. Maximum reinforcement criterion: a) concrete stains, b) equivalent concrete stresses, c) concrete forces .............................................................................................................................. 44 Figure 4-18. Deflection for superstructure vibration limitation ...................................................... 44 Figure 4-19. Deflections: a) dead load, b) long-term, c) short-term ............................................... 45 Figure 4-20. Material consumptions for cast-in-place on falsework solid slab .............................. 47 Figure 5-1. Precast segmental span-by-span construction method ................................................. 49 Figure 5-2. Span-by-span erection girder: a) overhead truss, b) underslung girder ........................ 49 Figure 5-3. Typical cross-section for precast segmental span-by-span box-girder ......................... 50 Figure 5-4. Typical reinforcing steel layout .................................................................................... 50 Figure 5-5. Typical tendon profile .................................................................................................. 51 Figure 5-6. Construction moments for segmental span-by-span method........................................ 52 Figure 5-7. Redistribution of dead load moments due to creep ...................................................... 54 Figure 5-8. Redistribution of dead load and prestress moments due to creep................................. 54 Figure 5-9. Thermal gradient effects ............................................................................................... 55 Figure 5-10. Moments due to thermal gradient ............................................................................... 56 Figure 5-11. Intentional angle changes ........................................................................................... 57 Figure 5-12. Long-term loss of prestress due to relaxation (Menn 1990) ....................................... 59 Figure 5-13. Compatibility conditions for bonded and unbonded tendons ..................................... 60 Figure 5-14. Deflection for superstructure vibration limitation ...................................................... 62 Figure 5-15. Deflections: a) dead load, b) long-term, c) short-term ............................................... 62 Figure 5-16. Material consumptions for precast span-by-span box-girder ..................................... 63 Figure 5-17. Access limited by height of interior box cavity.......................................................... 64 Figure 5-18. Access limited by height of interior box cavity and external tendons ........................ 64 Figure 6-1. Concrete material unit price ......................................................................................... 66 Figure 6-2. Concrete material cost comparison .............................................................................. 68 Figure 6-3. Total concrete cost comparison .................................................................................... 68 Figure 6-4. Prestressing tendon cost comparison ............................................................................ 69 Figure 6-5. Cost comparison of stirrups and minimum reinforcing steel ....................................... 71 Figure 6-6. Cost distribution of stirrups and minimum reinforcing steel ........................................ 72 Figure 6-7. Total reinforcing steel cost comparison ....................................................................... 73 Figure 6-8. Total superstructure material cost comparison ............................................................. 74 Figure 6-9. Total superstructure cost comparison (including cost of concrete placement)............. 75 ix

Figure 6-10. Total construction cost comparison ............................................................................ 78 Figure 6-11. Total construction cost comparison (+50% concrete unit price) ................................ 80 Figure 6-12. Total construction cost comparison (-50% concrete unit price) ................................. 80 Figure 6-13. Total construction cost comparison (+50% prestressing tendon unit price)............... 81 Figure 6-14. Total construction cost comparison (-50% prestressing tendon unit price) ............... 81 Figure 6-15. Total construction cost comparison (+50% reinforcing steel unit price) ................... 82 Figure 6-16. Total construction cost comparison (-50% reinforcing steel unit price) .................... 82 Figure 6-17. Total construction costs under changes in construction cost breakdown ................... 84 Figure 7-1. Cast-in-place on falsework box-girder with L=50m .................................................... 87 Figure 7-2. Cast-in-place on falsework solid slab with L=30m ...................................................... 88 Figure 7-3. Precast segmental span-by-span box-girder with L=50m ............................................ 89 Figure 7-4. Visual effects of increasing span-to-depth ratios from 10 to 35................................... 90 Figure 7-5. Effect of increasing span length (box-girder with h=2.5m) ......................................... 91 Figure 7-6. Viewed from 300m ....................................................................................................... 92 Figure 7-7. Viewed from 150m ....................................................................................................... 92 Figure 7-8. Viewed from 75m ......................................................................................................... 92 Figure 7-9. Effects of pier width-to-height ratio and span-to-depth ratio ....................................... 93 Figure 7-10. Effect of span-to-depth ratio as viewing angle becomes less oblique ........................ 94 Figure 7-11. Effect of bridge height on perceived superstructure slenderness ............................... 95 Figure 7-12. Effect of pier configuration on perceived superstructure slenderness ........................ 95 Figure 7-13. Effect of deck cantilever length on perceived superstructure slenderness ................. 96 Figure 7-14. Glenfinnan Viaduct, 1901 (Cortright 1997) ............................................................... 97 Figure 7-15. Slender bridges by Maillart ........................................................................................ 98 Figure 7-16. Waterloo Bridge over the Thames (Darger 2002) ...................................................... 99 Figure 7-17. Changis-sur-Marne Bridge, 1948 (Mossot 2007) ....................................................... 99 Figure 7-18. Sketches to evaluate aesthetic impact of span-to-depth ratios (O'Connor 1991) ....... 99 Figure 7-19. Neckar Valley Viaduct, 1977 (Leonhardt 1982) ...................................................... 100 Figure 7-20. Kocher Valley Viaduct, 1979 (Leonhardt 1982) ...................................................... 100 Figure 7-21. Pregorda Bridge, 1974 (Menn) ................................................................................. 101 Figure 7-22. Felsenau Bridge, 1974 (Menn) ................................................................................. 101 Figure C-1. Cast-in-place on falsework box-girder with L=50m.................................................. 130 Figure C-2. Cast-in-place on falsework solid slab with L=25m ................................................... 130 Figure C-3. Precast segmental span-by-span box-girder with L=40m.......................................... 130

LIST OF TABLES
Table 1-1. Description of recommended ratios ................................................................................. 3 Table 2-1. Summary of cast-in-place box-girders............................................................................. 7 Table 2-2. Summary of cast-in-place slabs (continued) .................................................................. 13 Table 2-3. Summary of precast segmental box-girders................................................................... 16 Table 3-1. Material properties ......................................................................................................... 21 Table 3-2. Material resistance factors (CSA 2006) ......................................................................... 21 Table 3-3. Prestressing tendon properties (CSA 1982) ................................................................... 21 Table 3-4. Corrugated metal duct properties (DSI 2008)................................................................ 21 Table 3-5. Concrete cover requirements (CSA 2006) ..................................................................... 22 Table 3-6. Load combination .......................................................................................................... 22 Table 3-7. Load factors ................................................................................................................... 22 Table 3-8. DLA factor (CSA 2006) ................................................................................................ 23 Table 3-9. Multi-lane loading modification factor (CSA 2006) ..................................................... 23 Table 4-1. Summary of structural response and dimensioning of cast-in-place on falsework boxgirder ................................................................................................................................................. 31 Table 4-2. Summary of material consumptions for cast-in-place on falsework box-girder ............ 34 Table 4-3. Results from beam model and strip method .................................................................. 42 Table 4-4. Summary of structural response and dimensioning of cast-in-place on falsework solid slab .................................................................................................................................................... 43 Table 4-5. Concrete strengths required to satisfy maximum reinforcement criterion ..................... 44 Table 4-6. Summary of material consumption for cast-in-place on falsework solid slab ............... 46 Table 5-1. Prestress losses due to friction ....................................................................................... 58 Table 5-2. Prestress losses due to anchorage set ............................................................................ 58 Table 5-3. Prestress losses due to creep and shrinkage ................................................................... 59 Table 5-4. Effective prestress after all losses .................................................................................. 59 Table 5-5. Prestress at ULS............................................................................................................. 61 Table 5-6. Summary of structural response of precast span-by-span box-girder ............................ 61 Table 5-7. Summary of material consumption for precast span-by-span box-girder ...................... 63 Table 6-1. Material unit prices ........................................................................................................ 65 Table 6-2. Concrete material unit price........................................................................................... 66 Table 6-3. Comparison of changes in cross-sectional depth and prestressing demand................... 70 Table 6-4. Total superstructure cost variations ............................................................................... 76 xi

Table 6-5. Construction cost breakdown (Menn 1990)................................................................... 77 Table 6-6. Material unit price changes ............................................................................................ 79 Table 6-7. Summary of material unit price sensitivity analysis ...................................................... 83 Table 6-8. Summary of cost study .................................................................................................. 85 Table C-1. Summary of results of cast-in-place on falsework box-girder analysis ...................... 127 Table C-2. Summary of results of cast-in-place on falsework solid slab analysis ........................ 128 Table C-3. Summary of results of precast segmental span-by-span box-girder analysis.............. 129

xii

LIST OF SYMBOLS
A Ac Ap As Av C c e(x) Ec Ep Es fc fcr fpu fpy fr fy h I Ic L L/h lp mop mp Mr Mr MSLS msp MULS n Gross cross-sectional area Area of concrete Area of prestressing steel Area of reinforcing steel Stirrup area Compressive force Depth of compression region Eccentricity of tendon at location x Concrete elastic modulus Prestressing tendon elastic modulus Reinforcing steel elastic modulus Concrete compressive strength Concrete tensile strength Prestressing tendon ultimate strength Prestressing tendon yield stress Free stress due to temperature gradient Reinforcing steel yield stress Girder depth Moment of inertia Moment of inertia of gross uncracked concrete section Span length Span-to-depth ratio Arc length of tendon between anchors Moment when qp is applied to prestressing band Moment when qp is applied to slab Flexural resistance Restraint moment SLS moment demand Self-equilibrating moment in strip method ULS moment demand Distance from base of cross-section to neutral axis (Section 5.3.2); Ep/Ec (Section 5.4.2) xiii

P P0 Pr qp S s T z (x) c D p x P p,rel 0 c cs (t) cu f (y) p0 p (t) c p s

Prestressing force Jacking force Axial restraint force Prestressing deviation force Sectional modulus Stirrup spacing Tensile force Moment lever arm Sum of angle changes of tendon between stressing locations and point x Thermal coefficient of concrete Dead load factor Prestress load factor Intentional angle change of tendon Deflection Loss of prestress force Unintentional angle change of tendon Prestress loss due to relaxation of steel Final strain Concrete strain Time-varying shrinkage strain Ultimate strain for concrete Free strain due to temperature gradient Thermal differential Coefficient of friction Jacking stress Effective prestress after all losses Creep coefficient Concrete resistance factor Prestressing tendon resistance factor Reinforcing steel resistance factor Final curvature of bending

xiv

1
1.1

INTRODUCTION
The Significance of Optimizing Span-to-Depth Ratio
Span-to-depth ratio, also known as slenderness ratio (L/h), is an important bridge design

parameter that relates a bridges span length to its girder depth. In the industry, this ratio is usually used to establish the superstructure depth and is chosen during the conceptual design phase before detailed calculations are performed. Selecting the ratio at an early stage of the design process permits approximate dimensional proportioning which is needed for preliminary analysis to evaluate the feasibility, cost-efficiency, and aesthetic merits of the design in comparison with alternative design concepts (ACI-ASCE 1988). The ratio is commonly chosen based on experience and typical values used in previously constructed bridges with satisfactory performance in order to ensure that the design does not deviate drastically from past successful practice. The ratio can also be determined by optimizing the combination of span length and superstructure depth to create a cost-efficient and aesthetically-pleasing structure, but this generally involves an iterative process. Therefore, instead of optimizing the span-to-depth ratio for every design concept, it is more common to select ratios from a range of conventional values. The choice of slenderness ratio is particularly critical in the design of girder-type bridges, because it directly affects the cost of materials and construction of the superstructure. For instance, using a high ratio (i.e. slender girder) reduces the concrete volume, increases the prestressing requirement, and simplifies the construction due to a lighter superstructure. Moreover, slenderness ratio has significant aesthetic impact, because the overall appearance of a girder-type bridge is highly dependent on the proportion of the superstructure (Leonhardt 1982). As stated previously, despite the significance of span-to-depth ratio, the industry has generally relied on the same proven range of ratios over the past decades. Figures 1-1 to1-3 show the recommended ranges of slenderness ratios outlined in different publications for three types of prestressed concrete constant-depth girders: cast-in-place box-girder, cast-in-place slab, and precast segmental box-girder. A brief description of the recommendations from each publication is given in Table 1-1.

2
40
ACI-ASCE 1988

30
AASHTO 1994

Span-todepth ratio

Duan et al. 1999 Hewson 2003

20

Leonhardt 1979 Menn 1990

10

Leonhardt 1979

Cohn & Lounis 1994 Multiple-cell box-girder

0 1975 1980 1985 1990 Year 1995

Incremental launching method

2000

2005

2010

Figure 1-1. Recommended ratios for cast-in-place box-girder

50 40 30
Menn 1990 ACI-ASCE 1988 AASHTO 1994 Cohn & Lounis 1994

Span-todepth ratio

20 10 0 1975

Cohn & Lounis 1994 Leonhardt 1979

Hewson 2003

Voided slab Solid slab

1980

1985

1990

Year

1995

2000

2005

2010

Figure 1-2. Recommended ratios for cast-in-place slab

25
ACI-ASCE 1988

20
AASHTO-PCI-ASBI 1997 Gauvreau 2006

Span-todepth ratio

15
Duan et al. 1999

10 5 0 1975 1980 1985 1990 Year 1995 2000 2005 2010

Figure 1-3. Recommended ratios for precast segmental box-girder

3
Table 1-1. Description of recommended ratios Author Leonhardt Year 1979 Description Fritz Leonhardt, a Professor of Civil Engineering at the University of Stuttgart, suggests ratios based on values from previously constructed prestressed concrete bridges with good performance. For cast-in-place single-cell box-girder, a ratio of 21 is recommended. The suggested ratio is lowered to around 12 to 16 when incremental launching method is used due to the large negative construction moments associated with this construction method. For castin-place slab, he suggests values from 18 to 36, with the higher values used for longer spans and for bridges with lighter traffic. The American Concrete Institute-American Society of Civil Engineers (ACI-ASCE) Committee 343 on Concrete Bridge Design defines span-to-depth ratio recommendations for common bridge types based on typical values. These recommendations are intended to provide general guidelines for preliminary design. For cast-in-place, post-tensioned multiple-cell box-girder, ACI-ASCE recommends ratios from 25 to 33. The recommended ratio for precast multiple-cell continuous box-girder is around 22. These ratios are higher than the ones for single-cell boxgirder, because a multiple-cell box section has more webs to accommodate tendons compared to a single-cell section with similar width. The recommended range of ratios is between 24 and 40 for cast-in-place, post-tensioned slab. Christian Menn is a Professor of Structural Engineering at the Institute of Structural Engineering in Zurich. His suggestions are based on existing bridges with satisfactory performance in terms of structural behaviour, aesthetics, and economics. He recommends ratios between 17 and 22 for cast-in-place box-girders, because girders with ratios below 17 would appear too heavy. On the other hand, girders with ratios above 22 have substantial cost increase due to the significantly higher longitudinal prestressing demand. Menn also suggests a maximum practical limit of 25 for solid slab and a maximum cost-effective slab depth of 0.8m. The American Association of State Highway and Transportation Officials (AASHTO) defines optional criteria for span-to-depth ratios in Cl.2.5.2.6.3 of the LRFD Bridge Design Specifications. These values are based on traditional maximum ratios of constant-depth continuous highway bridges with adequate vibration and deflection response. To ensure proper vibration and deflection behaviours, the maximum ratios are determined to be 25 for cast-in-place box-girder and 37 for cast-in-place slab. M.Z. Cohn is a Professor of Civil Engineering at the University of Waterloo and the span-todepth ratios suggested in this paper are part of the results of a Ph.D. thesis prepared by Z. Lounis. These ratios are established from a systematic, multi-level optimization approach that determines the ideal cross-sectional dimensions, span layouts and superstructure system based on cost, material consumption, and aesthetics. For cast-in-place single-cell box-girder, the optimum ratio is found to range from 12 to 20. The ratio increases with span length and decreases with bridge width (e.g. a ratio of 12 corresponds to a span of 20m and a width of 16m while a ratio of 20 corresponds to a span of 50m and a width of 8m). This range of ratios is slightly lower relative to the ones from other publications, because this study investigates a simply-supported system while the ratios from other publications are mostly based on continuous systems. A simply-supported girder tends to be deeper since it experiences greater moments at midspan compared to a continuous structure. Cohn & Lounis also suggest the range of optimum ratios for voided and solid slabs are 22 to 29 and 28 to 33 respectively. The American Segmental Bridge Institute (ASBI) has established various standard precast sections for segmental construction to enhance uniformity and simplicity for forming and production methods. Using these standard sections generally lead to practical and costeffective solutions. The ranges of span-to-depth ratios obtained from these standard sections are 17 to 19 for span-by-span method and 17 to 20 for balanced cantilever method. Lian Duan is a Senior Bridge Engineer with the California Department of Transportation and a Professor of Structural Engineering at Taiyuan University of Technology in China. A span-todepth ratio of 25 is recommended for cast-in-place multiple-cell box-girder based on typical values from existing bridges. A range of ratios from 12.5 to 20 is recommended for precast segmental box-girder. This range is based on frequently used standard precast sections from Federal Highway Administration (FHWA).

ACI-ASCE

1988

Menn

1990

AASHTO

1994

Cohn & Lounis

1994

AASHTOPCI-ASBI

1997

Duan et al.

1999

4
Table 1-1. Description of recommended ratios (continued) Author Hewson Year 2003 Description Nigel Hewson is a recognized expert in the design and construction of prestressed bridges and is an Associate Lecturer at the University of Surrey on this subject. He suggested a span-todepth ratio of 20 for cast-in-place single-cell box-girder and a maximum ratio of 20 for cast-inplace voided slab. Both of these recommendations are based on typical values. A span-to-depth ratio of 17 is recommended for precast segmental span-by-span constructed box-girder. This value corresponds to the lower limit of span length used for this construction method (30m) and the minimum height requirement of a box section to provide sufficient access space within the box (1.8m). The recommended ratio is lower than the one for cast-inplace box-girder, because a larger depth is needed to compensate for the reduced tendon eccentricity due to the use of external unbonded tendons.

Gauvreau

2006

As shown in the previous graphs, there has been no significant increase in the recommended span-to-depth ratio since 1979 despite the advancement in material strengths and construction technologies. Recent developments have resulted in high-strength materials which theoretically should lead to more slender structural components and longer span lengths. In particular, highstrength concrete with compressive strength of 40 to 140 MPa has been achieved by lowering the water-to-cement ratio and incorporating chemical admixtures (Kosmatka et al. 2002). Because of their enhanced mechanical properties like higher ultimate strengths and modulus of elasticity, highstrength concrete structures can resist the same level of loads using slenderer sections, resulting in lightweight structures. The reduction in self-weight is especially critical in long-span bridges, because the dead load consumes approximately 75% of the load-bearing capacity in long-span bridges constructed with normal-strength concrete (TRB 1990). High-strength concrete lowers the dead load contribution by using thinner sections and improves the load-bearing capacity by increasing strength, thus slenderer bridges with longer spans can be attained. High-strength concrete has been applied to various types of structures. For instance, concrete with compressive strength of 60 MPa is commonly used for large bridges in Europe (Muller 1999) while the building industry has been using concrete with strengths of over 100 MPa for years (Hassanain 2002). However, most short- and medium-span bridges are being constructed with concrete strengths of less than 50 MPa, because high-strength concrete is more expensive, especially if the designer still uses the typical span-to-depth ratios as defined decades ago based on normal-strength concrete (Hassanain 2002). For instance, the unit price of concrete rises by about 68% when the compressive strength changes from 30 MPa to 60 MPa (Dufferin Concrete 2009). This indicates a substantial material cost increase if the same guidelines for superstructure proportioning of normal-strength concrete bridges are applied to high-strength concrete bridges, causing the application of high-strength concrete in bridges to be economically unfeasible. Therefore, with the advent of high-strength materials, recommended span-to-depth ratios need to be

5 updated to match the improvement in material strength and stiffness and to provide an economic incentive for the application of these materials in bridges.

1.2

Objectives and Scope

The purpose of this thesis is to determine the ideal range of span-to-depth ratios for post-

tensioned girder bridges constructed with current high-strength materials based on aesthetic comparisons and optimization parameters such as material consumption and total construction cost. The three bridge types considered in this study are cast-in-place on falsework box-girder and solid slab, and precast segmental span-by-span box-girder. The objectives of this study are summarized as follows: Provide a study on the evolution of span-to-depth ratios in concrete girder bridges constructed over the past 50 years and establish a range of conventional ratios. Determine the amount of prestressing and the concrete strength needed to satisfy safety and serviceability requirements as a function of span-to-depth ratio for the three types of bridge considered. Compare the material consumptions and total construction costs for bridges with different slenderness ratios and determine the most cost-effective ratios. Investigate the sensitivity of the construction cost results with respect to changes in material unit cost and construction cost breakdown. Examine the visual impact of different span-to-depth ratios and especially evaluate the aesthetic influence of using the cost-effective ratios instead of conventional ones. Update the recommendations for span-to-depth ratios based on economic and aesthetic considerations. The results of this research are expected either to confirm that the conventional ratios are already optimal for new high-strength materials or to demonstrate that more slender sections can be attained. The study focuses on the superstructure only while the prestressing and concrete strength demands for the substructure are not explicitly accounted for. Also, only bridges with typical span lengths are analyzed in this study: 35m to 75m for cast-in-place box-girder, 20m to 35m for cast-inplace solid slab, and 30m to 50m for precast segmental box-girder.

1.3

Thesis Structure
The thesis is organized in eight chapters: Chapter 1 provides the background and motivation of optimizing span-to-depth ratio. Chapter 2 examines the span-to-depth ratios of existing bridges and discusses their changes over

the past 50 years. This information along with the span-to-depth ratio recommendations described in Chapter 1 leads to values for conventional slenderness ratios. These conventional ratios serve as a basis for cost and aesthetic comparisons in the later chapters. Chapter 3 outlines the general analysis model and method used for all three bridge types. It also provides a breakdown on all the analysis cases that need to be considered and discusses specific design criteria that must be satisfied. The specific analysis models and analysis results for cast-in-place on falsework box-girder and solid slab, and precast segmental span-by-span box-girder are described in Chapter 4 and 5 respectively. Analysis results include structural responses, material consumptions, and factors that limit further increase in slenderness ratio. The construction method and design issues unique to each bridge type are also discussed. Chapter 6 compares the material costs and total construction costs for bridges with varying span-to-depth ratios for the three bridge types. Optimal ratios with the lowest costs are determined and in particular, cost savings associated with using the optimal ratios instead of conventional ones are examined. Furthermore, a sensitivity analysis is performed to demonstrate the effects of changing unit costs and total construction cost breakdown on the analysis results. Chapter 7 explores the aesthetic impacts of varying span-to-depth ratios and discusses the public perception on visually optimal ratios. Chapter 8 provides a conclusion for this study by summarizing the optimal span-to-depth ratios for the three bridge types as well as their improvement over conventional ratios in terms of material consumptions, construction costs, and aesthetics. These optimal ratios lead to updated span-to-depth ratio recommendations for bridges constructed with current high-strength materials.

TYPICAL SPAN-TO-DEPTH RATIOS OF EXISTING BRIDGES

This chapter describes a study of 86 existing constant-depth girder bridges and presents a

compilation of their span-to-depth ratios. Specifically, the study determines the range of ratios typically used in the industry and examines its variations over the past 50 years. Three bridge types are considered: cast-in-place box-girder, cast-in-place slab, and precast box-girder. A majority of these bridges has span-to-depth ratios within the suggested ranges discussed in Chapter 1, indicating that a representative sample of bridges has been used.

2.1

Cast-in-Place Box-Girder
First, the study investigates 44 constant-depth cast-in-place box-girders. Table 2-1 provides the

basic information as well as a cross-sectional drawing for each bridge. Additional information, including the span arrangement, girder dimensions, designer and references, is given in Appendix A.1. Figure 2-1 shows the span-to-depth ratios with respect to the span lengths and compares these ratios with the recommended values described in Section 1.1. Figure 2-2 plots the ratios with respect to the completion years in order to illustrate the trend in slenderness ratio over time.
Table 2-1. Summary of cast-in-place box-girders
Bridge no. 1 Name Location Span-todepth ratio 17.7 Construction method N/A Cross-section

Grenz Bridge at Basel

Switzerland

Sart Canal-Bridge Weyermannshaus Bridge Eastbound Walnut Viaduct Taiwan High Speed Rail (1) & (2) Pregorda Bridge Almese Viaduct & Condove Viaduct

Belgium

12.0

Incremental launching N/A

N/A

Switzerland

18.9

U.S.A.

23.0

CIP on falsework

5&6

Taiwan

11.4

Span-by-span Span-by-span on falsework Balanced cantilever Balanced cantilever

Switzerland

22.2

8&9

Italy

18.2

10

Gravio Viaduct

Italy

18.2

Legend: N/A = no data CIP = cast-in-place

Split cross-section:

8
Table 2-1. Summary of cast-in-place box-girders (continued)
Bridge no. 11 Name Location Span-todepth ratio 18.2 Construction method Balanced cantilever Span-by-span on falsework N/A CIP on falsework Balanced cantilever Balanced cantilever Balanced cantilever Cross-section

Borgone Viaduct

Italy

12 13 14

Quadinei Bridge Altstetter Viaduct Reuss Bridge

Switzerland Switzerland Switzerland

20.0 21.6 17.7

15

Cerchiara Viaduct

Italy

18.5

16

Castello Viaduct

Italy

18.5

17

Costacole Viaduct

Italy

18.5

18

Ferroviario Overpass at Bolzano

Italy

28.1

N/A

19

Krebsbachtal Bridge

Germany

12.9

Incremental launching Incremental launching

20

Shatt Al Arab Bridge

Iraq

12.8

21

Ancona Viaduct

Italy

20.7

Segmental

22

Felsenau Bridge (approaches)

Switzerland

16.0

Span-by-span on falsework

23

La Molletta Viaduct

Italy

20.8

Segmental

24

Fosso Capaldo Viaduct Sihlhochstrasse Bridge Grosotto Viaduct, Grosio Viaduct, Tiolo Viadut Denny Creek Viaduct Woronora River Bridge

Italy

20.8

Segmental

25

Switzerland

29.5

N/A

26 to 28

Italy

20.0

Balanced cantilever

29 30

U.S.A. Australia

20.9 14.7

N/A Incremental launching N/A

Table 2-1. Summary of cast-in-place box-girders (continued)

Bridge no. Name Location Span-todepth ratio Construction method Balanced cantilever Cross-section

31

Valentino Viaduct

Italy

20.0

32

Giaglione Viaduct

Italy

20.0

Balanced cantilever

33

Venaus Viaduct

Italy

20.0

Balanced cantilever

34

Passeggeri Viaduct

Italy

20.0

Segmental

35

Brunetta Viaduct

Italy

20.0

Segmental

36

Pietrastretta Viaduct

Italy

20.0

Segmental

37

Deveys Viaduct

Italy

20.0

Segmental

38

Gruyre Lake Viaduct Interstate 895 Bridge over James River (approaches) Ltten Bridge

Switzerland

15.1

Span-by-span

39

U.S.A.

21.3

Balanced cantilever

N/A

40

Switzerland

18.1

N/A

41

Savona Mollere Viaduct

Italy

22.5

Segmental

42

Ruina Viaduct

Italy

19.3

N/A

43

Weinland Bridge

Switzerland

22.6

Span-by-span on falsework

44

Kocher Valley Bridge

Germany

21.2

Balanced cantilever

10

Bridge No. 35 30 25

1 5 20 30 40 42 | |||| | || | | ||| || | | || | | | | | |
1974

43 |

44 |

1958 1974 1992 1967 1992 1980 1992 1972 1978 1975 2001 2002 1975 2000 1994 1984

25 (Lian et al. , AASHTO, and minimum value of ACI-ASCE) 1971 20 (Hewson)

20 Span-todepth ratio 15 10 5 0 20

Incremental launching *Shaded region = Menn's range (17 to 22)

40

60

80 100 Span length (m)

120

140

160

Figure 2-1. Span-to-depth ratios of cast-in-place box-girders

Figure 2-1 demonstrates that all 44 cast-in-place box-girders have span lengths between 35.4m and 138m as well as span-to-depth ratios that range from 11.4 to 29.5. The frequency plot on the top shows that 42 out of 44 bridges (95%) investigated have span lengths from 35m to 75m which is the typical range for constant-depth box-girders as suggested by Hewson (2003). Above the frequency plot are bridge numbers that relate each data point to its corresponding bridge in Table 2-1. The frequency plot on the right shows a large concentration of bridges that have span-to-depth ratios between 17.7 and 22.6. In fact, 33 out of 44 bridges (75%) have ratios within the range of values recommended by Menn (17 to 22) and Hewson (20) which are based on existing bridges with satisfactory performance, indicating that the study sample is representative of typical bridges. Most bridges have ratios below 25 which is the traditional maximum value that ensures adequate vibration and deflection responses in cast-in-place box-girders according to American design standards (AASHTO 1994). Only 2 bridges (i.e. bridge no. 18 and 25), one with a multiplecell box-girder and the other with twin parallel box-girders, have ratios above 25. These two bridges, however, are within the range of ratios recommended by ACI-ASCE (1988) for post-tensioned castin-place multiple-cell box-girders (25 to 33). Higher ratios are expected for these types of crosssections, because additional webs can help accommodate the large amount of prestressing tendons associated with slender girders without sacrificing the efficiency of the tendon layout (i.e. lowering the tendon eccentricity by placing tendons in vertical layers within the webs). Also, the decrease in spacing between webs causes considerable reduction in transverse bending for wide cross-sections, thus lowering the transverse prestressing requirement.

11 Furthermore, 6 bridges have ratios of less than 15. Out of these 6 bridges, 4 are constructed with incremental launching which generally requires a deeper cross-section to resist the large negative moments during construction. They have ratios between 12 and 16 which is the typical range for incrementally launched single-cell box-girders recommended by Leonhardt (1979). The other two are railway bridges which also need a larger depth due to the greater live loads and more stringent serviceability requirements. For instance, ACI-ASCE (1988) suggested a typical span-todepth ratio of approximately 16 for cast-in-place multiple-cell box-girders that carry railroads. Figure 2-2 describes the variation in span-to-depth ratios over time. Only 37 out of the 44 bridges investigated are included in this graph due to the lack of data on completion year. These 37 bridges were completed between 1958 and 2002, and no significant variation in span-to-depth ratios is observed within this time span. The slenderness ratios commonly used by the industry have not increased over time despite the improvements in material strengths and advancements in construction technologies. As stated previously, 75% of these bridges follow the same guidelines recommended by Menn in 1986 and Hewson in 2003.
35 30 25 20 Span-todepth ratio 15 10 5 0 1955 1965 1975 Year 1985 1995 2005
Incremental launching *Shaded region = Menn's range (17 to 22) 25 (Lian et al. , AASHTO, and minimum value of ACI-ASCE) 20 (Hewson)

Figure 2-2. Span-to-depth ratios of cast-in-place box-girders

12

2.2

Cast-in-Place Slab
In addition to cast-in-place box-girders, 28 cast-in-place constant-depth slab bridges are also

investigated. Table 2-2 provides basic information as well as a cross-sectional drawing for each bridge and the detailed bridge information is given in Appendix A.2. Figure 2-3 and 2-4 relate the span-to-depth ratios to the span lengths and to the completion dates.
Table 2-2. Summary of cast-in-place slabs
Bridge no. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Name Khandeshwar Bridge Spadina Ave. Bridge #16, Hwy 401 Spadina Ave. Bridge #18A, Hwy 401 Spadina Ave. Bridge #18B, Hwy 401 Spadina Ave. Bridge #19, Hwy 401 Spadina Ave. Bridge #21A, Hwy 401 Spadina Ave. Bridge #21B, Hwy 401 Sindelfingen Footbridge L 333 Overpass at Bassum Waiblingen Footbridge Mako Bridge Kittelbaches Bridge San Francisco Airport Viaduct St. Vincent Street Overpass Bridge across Jan-Wellen-Platz Spadina Ave. Bridge #14, Hwy 401 Spadina Ave. Bridge #15, Hwy 401 Spadina Ave. Bridge #12, Hwy 401 Saale Bridge at Rudolphstein Bridge #20 at Hwy 401/427 Interchange Spadina Ave. Bridge #5, Hwy 401 Spadina Ave. Bridge #11 Hwy 401 Location India Canada Canada Canada Canada Canada Canada Germany Germany Germany Senegal Germany U.S.A. Canada Germany Canada Span-todepth ratio 20.3 22.2 22.2 22.2 22.2 22.2 22.2 55.7 21.9 42.5 23.9 19.3 19.2 35.6 25 38.6 Construction method N/A CIP on falsework CIP on falsework CIP on falsework CIP on falsework CIP on falsework CIP on falsework N/A N/A N/A Incremental launching N/A N/A CIP on falsework N/A CIP on falsework N/A N/A N/A N/A N/A N/A Cross-section N/A

61

Canada

38.6

CIP on falsework

62 63 64

Canada Germany Canada

38.6 20 31.7

CIP on falsework N/A CIP on falsework N/A

65

Canada

35.0

CIP on falsework

66

Canada

30

CIP on falsework

13
Table 2-2. Summary of cast-in-place slabs (continued)
Bridge no. 67 Name Spadina Ave. Bridge #22, Hwy 401 Spadina Ave. Bridge #23, Hwy 401 Hundschipfen Bridge McCowan Road Underpass Spadina Ave. Bridge #24, Hwy 401 Spadina Ave. Bridge #4, Hwy 401 Location Span-todepth ratio 30.5 Construction method CIP on falsework Cross-section

Canada

68 69 70

Canada Switzerland Canada

31.2 40.5 37.9

CIP on falsework N/A CIP on falsework N/A

71

Canada

33.4

CIP on falsework

72

Canada

34.7

CIP on falsework

Bridge No. 60

45 |

52 | | |
1986

56 58 ||| ||

60 63 | ||

68 || | |

70 72 | |

50
1996 2000 1963 1963 1975 2000 1993 1967 1992 1961 1994 1963 1963 37 (AASHTO) 1963 30.5 (Cohn & Lounis for solid slab)

40 Span-todepth ratio 30

1963 1964

1963

25 (Menn, Cohn & Lounis for voided slab) 20 (Hewson for voided slab)

20

10

Voided slab Solid slab

*Shaded region = Leonhardt's range (18 to 36)

0 0 10 20 30 Span length (m) 40 50 60

Figure 2-3. Span-to-depth ratios of cast-in-place slabs

As shown in Figure 2-3, the 28 bridges have span lengths between 13.2m and 47.5m and spanto-depth ratios from 19.2 to 55.7. The sample consists of 14 solid slabs and 14 voided slabs. Out of the 14 solid slabs, 7 that were mostly built in the 1960s have span lengths greater than 20m which is the current maximum economic span length for this type of slab (Gauvreau 2006). All voided slabs, except for bridge no. 72, have spans of less than the maximum typical span length of 46m as suggested by ACI-ASCE (1988). Also, most of the bridges (79%) have span-to-depth ratios that are within Leonhardts recommended range of 18 to 36 and are below AASHTOs maximum value for

14 adequate deflection and vibration behaviour. Since the majority of the sample has span-to-depth ratios within Leonhardts suggested range and spans similar to conventional values, the sample is fairly representative of typical slab bridges.
60 50

40 Span-todepth ratio 30

37 (AASHTO) 30.5 (Cohn & Lounis for solid slab) 25 (Menn, Cohn & Lounis for voided slab)

20

20 (Hewson for voided slab) Voided slab Solid slab *Shaded region = Leonhardt's range (18 to 36)

10

0 1960 1970 1980 1990 Year 2000 2010

Figure 2-4. Span-to-depth ratios of cast-in-place slabs

As shown in Figures 2-3 and 2-4, most of the bridges have span-to-depth ratios that cluster around two ranges: 13 bridges (46%) have ratios between 19 and 25 while 13 bridges have ratios between 30 and 40.5. The first range is composed of 7 solid slabs built in the 1960s and 6 voided slabs built after 1970. The latter range consists of 6 solid slabs and 7 voided slabs which are all constructed in the 1960s except for bridge no. 69. The remaining two bridges (bridge no. 52 and 54) with higher ratios of 42.5 and 55.7 are both pedestrian bridges which can be more slender due to the lower live load requirements. Therefore, there is a noticeable variation in the typical range of spanto-depth ratios depending on the construction year and function of the bridge. Figure 2-4 clearly illustrates the changes in typical span-to-depth ratios with respect to construction year. Out of the 19 bridges completed prior to 1975, 12 (63%) have slenderness ratios greater than 30 and 12 (63%) are solid slabs. Newer bridges are mainly voided slabs with lower slenderness ratios at around 20 due to the stricter code requirements in recent years. For instance, the Ontario Ministry of Transportation (MTO) sets the minimum non-prestressed reinforcement clear cover to be 7020 for the top surface of voided slabs in the MTO Structural Manual (2003) while the value is only 5020 in the Ontario Highway Bridge Design Code (OHBDC 1983). Likewise, the MTO Structural Manual limits the maximum span-to-depth ratio to 28 for all posttensioned slabs while no such provisions existed prior to 1975 (Scollard and Bartlett 2004). As shown in Figure 2-4, the first generation post-tensioned voided slabs constructed in the 1960s have

15 span-to-depth ratios of over 30 which required large amount of longitudinal prestressing. This resulted in the formation of longitudinal cracks above the voids due to the large concentrated posttensioning forces near the abutments, which created transverse splitting stresses, and due to the restraint of transverse concrete shrinkage imposed by the steel void forms. To solve this cracking problem, MTO recommended the addition of transverse prestressing to prevent shrinkage cracking and a decrease in span-to-depth ratio in 1975 in order to reduce the required prestressing force, which eventually led to the current maximum span-to-depth ratio limit of 28 (Scollard and Barlett 2004). As a result, in this study, 6 out of the 9 bridges built after 1975 have span-to-depth ratios below 28; the remaining 3 bridges are pedestrian bridges or European bridges. As stated before, the typical span-to-depth ratios for slab bridges vary considerably with time and bridge function. The impact of slab type (i.e. solid or voided), on the other hand, is not as significant. According to the literature discussed in Chapter 1, the conventional span-to-depth ratios for solid slab are expected to be higher than the ones for voided slab, because voided slabs are commonly used to reduce self-weight for longer spans that require slabs thicker than 800mm (Menn 1990). In fact, the study by Cohn and Lounis (1994) suggested that the optimum depth for voided slab is 12% to 20% thicker than the one for solid slab, resulting in optimum ratios of approximately 30.5 for solid slab and 25 for voided slab. The sample in this study indicates a small difference in span-to-depth ratios between the two slab types. Voided slabs have ratios that range from 19 to 35 while the range is from 22 to 39 for solid slabs (excluding the pedestrian bridges). These results are reasonable, because a voided slab is theoretically an intermediate cross-section between a solid slab and a box-girder and its range of ratios is expected to be in between the ones from solid slab and box-girder (i.e. 17 to 22 as determined in Section 2.1). The typical ratios of the voided slab might be closer to those of the solid slab or of the box-girder depending on its component dimensions. For instance, if the void diameter is less than 60% of the total slab depth, the longitudinal behaviour would resemble a solid slab (O'Brien and Keogh 1999). In this sample, the mean ratio for solid slab is 30 which is only slightly higher than the 27 for voided slab if all the bridges from 1960 to 2000 are considered. If only recently constructed bridges are considered (i.e. built after 1990), the conventional ratio for voided slab would decrease to around 20 which is the same as Hewsons suggested value. There is no data for recently constructed solid slabs, but the conventional ratio for solid slab is expected to follow the same trend when the entire sample is considered and be only slightly higher than the value for voided slab.

16

2.3

Precast Segmental Box-Girder

In this section, 14 precast segmental box-girders are examined. Table 2-3 provides basic

information as well as a cross-sectional drawing for each bridge and the detailed bridge information is given in Appendix A.3. Figures 2-5 and 2-6 relate the span-to-depth ratios to the span lengths and the completion dates.
Table 2-3. Summary of precast segmental box-girders
Bridge no. 73 Name Bukit Panjang LRT System 801 Location Singapore Span-todepth ratio 15.7 Construction method Segmental span-byspan Segmental with launching girder Cross-section

74

Wiscasset Bridge

U.S.A.

17.5

N/A

75

Chiovano Viaduct

Italy

17

Segmental

76

Collecastino Viaduct

Italy

17

Segmental

77

Fiumetto Viaduct

Italy

17

Segmental

78

San Leonardo Viaduct

Italy

17

Segmental

79

Petto Viaduct

Italy

17

Segmental

80

Cadramazzo Viaduct

Italy

16.7

Balanced cantilever

81

Fella IX Viaduct

Italy

16.7

Balanced cantilever

82

Malborghetto Viaduct

Italy

16.1

Balanced cantilever

83

Val Freghizia Viaduct

Italy

16.8

Balanced cantilever

84

Fella IV Viaduct

Italy

17.6

Balanced cantilever

85

Ngong Shuen Chau Viaduct Sutong Bridge Approach (Nantong side)

China

18.8

Balanced cantilever

86

China

18.8

Balanced cantilever

17

25

Bridge No. 73 75 | | ||

83 | |

84 |

86 |

20
1981 1992 1986 1998 1988 1986 1985

2007

2007 17 (Gauvreau)

15 Span-todepth ratio

10

5
*Shaded region = range from Duan et al. (12.5 to 20)

0 0 10 20 30 40 50 60 70 80 90 Span length (m)

Figure 2-5. Span-to-depth ratios of precast segmental box-girders

25

20
17 (Gauvreau)

15 Span-todepth ratio

10

5
*Shaded region = range from Duan et al. (12.5 to 20)

0 1980 1985 1990 1995 Year 2000 2005 2010

Figure 2-6. Span-to-depth ratios of precast segmental box-girders

According to Figure 2-5, all 14 bridges have slenderness ratios between 15.7 and 18.8 which are within the range of frequently used ratios suggested by Duan et al. (1999). Also, 13 bridges have spans between 30m and 60m which is a feasible and cost-effective span range for precast segmental constructed constant-depth girders recommended by ASBI (1997). Since most of the bridges in this study sample have span-to-depth ratios and span lengths within standard ranges, these bridges are assumed to be representative of typical precast segmental box-girders.

18 Moreover, out of the 14 bridges, 10 bridges (71%) have ratios within 5% from 17 which is the recommended value for precast segmental span-by-span construction from Gauvreau (2006). The lowest ratio is 15.7 for bridge no. 70 which is a railway bridge that requires a deeper girder to satisfy the more stringent serviceability requirements. Also, Figure 2-6 indicates that the typical span-to-depth ratios did not vary significantly from 1981 to 2007.

2.4

Concluding Remarks
This study examines the span-to-depth ratios of 86 constant-depth girder bridges in order to

determine the range of ratios typically used by the industry over the past 50 years. The slenderness ratios with respect to span lengths for all of these bridges are illustrated in Figure 2-7. The average span-to-depth ratio and the typical span lengths for each bridge type are also indicated on the graph.
70 60 50 Span-todepth 40 ratio 30 20 10 0 0 20 40 60 80 Span length (m) 100 120 140
Average ratios: 30 (CIP solid slab) 27 (CIP voided slab) 20 (CIP box-girder) 17 (precast segmental box-girder) Typical span lengths: 35 to 75m (CIP box-girder) 30 to 60m (precast segmental box-girder) 15 to 48m (CIP slab) CIP box-girder CIP voided slab CIP solid slab Precast segmental box-girder

Figure 2-7. Span-to-depth ratios for all bridge types

The primary findings of this investigation are summarized in Table 2-4. Average ratios within the typical ranges are considered as the conventional ratios and are used as a basis of comparison in this thesis.
Table 2-4. Summary of conventional span-to-depth ratios Bridge type Cast-in-place box-girder Cast-in-place voided slab Cast-in-place solid slab Precast segmental boxgirder Range of spanto-depth ratios 17.7 to 22.6 19 to 35 22 to 39 15.7 to 18.8 Number of bridges within this range 33 out of 44 (75%) 13 out of 14 (92%) 12 out of 14 (86%) 14 out of 14 (100%) Average ratio 20 27 30 17 Notes Range varies little between 1958 and 2002 Conventional ratio is closer to 20 for bridges completed after 1990 Used mainly from 1961 to 1975 Range varies little between 1981 and 2007

ANALYSIS OVERVIEW
This purpose of this analysis is to compute the amount of prestressing and the concrete strength

needed to satisfy design requirements for bridges with varying span lengths and span-to-depth ratios. These material consumption results are then used to compute construction cost as a function of span-to-depth ratio. By examining the variations in construction cost and aesthetic impacts, the study determines the most cost-optimal ratios for different bridge types. The three post-tensioned bridge types considered are cast-in-place on falsework box-girder, cast-in-place on falsework solid slab, and precast segmental span-by-span box-girder. This chapter describes the analysis model, material properties, applied loads, ultimate and serviceability limit states design requirements, as well as some preliminary analysis assumptions. The analysis is performed using the program SAP2000 and spreadsheet calculations.

3.1

Analysis Model
The analysis model is an 8-span highway viaduct with a straight profile. Typical plan and

elevation are shown in Figures 3-1.

Abutment Pier 1 Pier 2 Pier 3 Pier 4 Pier 5 Pier 6 Pier 7
Abutment

Bridge width b
Lend

Design lane 1 Design lane 2 C Bridge L Design lane 3 Design lane 4 L L L

PLAN

Lend

Depth h

ELEVATION

Figure 3-1. Typical plan and elevation

Two types of constant-depth cross-sections are investigated: single-cell box-girder and solid slab. Both types have 0.5m wide barriers, 90mm thick wearing surface and 3.5m wide design lanes as shown in Figure 3-2. Proportions of other cross-sectional components vary for different construction methods and are discussed in greater details in Sections 4.2.1.1, 4.3.1.1, and 5.2.1.
Barrier Design Lane 1 Design Lane 2 C L Design Lane 3 90mm thick wearing surface Design Lane 4

500

3500

3500

3500

3500

500

Figure 3-2. Typical deck arrangement

19

20 In the analysis, the span length and span-to-depth ratio are varied in the model to generate the analysis cases illustrated in Figure 3-3. It should be noted that for cast-in-place on falsework boxgirders, cases with spans of 75m are included in the study mainly for comparison purposes. In the industry, however, such long spans are generally constructed with cantilever method in regions where high labour costs deter the extensive use of falsework.
L = 35m L/h = 10, 15, 20, 25 L = 50m Cast-in-place on falsework box-girder L/h = 10, 15, 20, 25, 30 L = 60m L/h = 10, 15, 20, 25, 30, 35 L = 75m L/h = 10, 15, 20, 25, 30, 35 L =20m L/h = 30, 35, 40, 45 L = 25m L/h = 30, 35, 40, 45 8-span highway viaduct Cast-in-place on falsework solid slab L = 30m L/h = 30, 35, 40, 45, 50 L = 35m L/h = 30, 35, 40, 45, 50 L = 30m L/h = 15, 20, 25 Precast segmental, span-byspan box girder L = 40m L/h = 15, 20, 25, 30 L = 45m L/h = 15, 20, 25, 30 Legend: L = span length L/h = span-to-depth ratio

Figure 3-3. Summary of analysis cases

21

3.2

Materials
Tables 3-1 and 3-2 summarize the material properties and resistance factors used in the analysis.

To illustrate the effects of high-strength materials on span-to-depth ratios, a concrete compressive strength of 50 MPa is used in the analysis, because this value is the minimum strength requirement of high performance concrete as defined by MTO (OPSS 2007).
Table 3-1. Material properties Material Property 1 Concrete Compressive strength Tensile strength Elastic modulus Value fc = 50 MPa fcr = 2.8 MPa = 0.4 fc (CSA 2006) Ec = 28100 MPa
= 3000 fc + 6900 c 2300
1.5

(CSA 2006)

Prestressing tendons Ultimate strength Yield stress Effective prestress after all losses Elastic modulus Reinforcing steel bars Yield stress Elastic modulus
1 2

fpu = 1860 MPa fpy = 0.9 fpu = 1670 MPa p = 0.6fpu = 1120 MPa Ep = 200000 MPa fy = 400 MPa Es = 200000MPa

For solid slab analysis, higher concrete strengths (i.e. up to 80 MPa) are used for slender cases. For prestressing tendons and reinforcing steel bars, a bilinear stress-strain relationship is used: fp = Ep p fpy fs = Es s fy

Table 3-2. Material resistance factors (CSA 2006) Material Concrete Prestressing strands Reinforcing steel bars and wires Resistance factor c = 0.75 p = 0.95 s = 0.90

3.2.1

Prestressing Tendons

All analysis cases utilize size 15 seven-wire low-relaxation strands as prestressing tendons and standard size corrugated metal ducts with properties summarized in Tables 3-3 and 3-4.
Table 3-3. Prestressing tendon properties (CSA 1982) Property Nominal diameter Nominal Area Mass Value 15.24 mm 2 140 mm 1.109 kg/m

Table 3-4. Corrugated metal duct properties (DSI 2008) # strands 12 15 19 27 37 Outer diameter (mm) 94 97 106.3 121.4 138.4 Minimum block-out diameter (mm) 254 279 305 343 407 Transition length (mm) 508 575 640 702 890

22 The minimum clear distance between adjacent ducts is 40mm according to the Canadian Highway Bridge Design Code (CHBDC) (CSA 2006). A horizontal spacing of about the duct diameter is used between tendons to provide sufficient space for concrete placement and vibration. 3.2.2 Concrete Covers

Concrete cover requirements and tolerances from CHBDC are used (Table 3-5). The bridge is assumed to be exposed to the most severe environmental category (i.e. de-icing chemicals, spray or surface runoff containing de-icing chemicals, marine spray).
Table 3-5. Concrete cover requirements (CSA 2006) Component Top surface of bottom slab Top surface of top slab Soffit of top and bottom slabs Vertical surfaces Reinforcement or steel ducts Reinforcing steel Post-tensioning ducts Post-tensioning ducts Reinforcing steel Post-tensioning steel Reinforcing steel Post-tensioning ducts Concrete covers (mm) Cast-in-place Precast 40 10 40 10 60 10 60 10 130 15 50 10 70 10 70 10 90 10 120 10 45 10 65 10 60 10 80 10

3.3

Loads
Dead loads, live loads (truck and lane), and prestress loads based on CHBDC are used in the

analysis. Thermal gradient effects are considered only in the precast segmental box-girder analysis because they are more critical for precast girders which do not have continuous bonded steel. 3.3.1 Load Combinations and Load Factors

The load combinations and load factors used in the analysis under ultimate and serviceability limit states are summarized in Table 3-6 and Table 3-7 (CSA 2006). Fatigue limit state is not considered in the analysis, because the concrete is kept uncracked at service and the steel does not experience large stress cycles.
Table 3-6. Load combination Loads Ultimate limit states ULS1 ULS2 Serviceability limit states SLS1 SLS2 SLS3 Permanent loads D P D D 1.00 0 1.00 P P 1.00 0 1.00 Transitory loads L K 1.70 0 0.90 0.90 0 0 0 0.80 0 0 Table 3-7. Load factors Dead load Precast concrete Cast-in-place concrete Barriers Wearing surfaces Prestress load Secondary prestress effects Load factor D 1.10 1.20 1.20 1.50 Load factor P 0.95

Legend: D = dead load L = live load P = secondary prestress effects K = effects of strains due to temperature differential

23 As shown in Table 3-6, load combinations ULS2 and SLS3 only consider permanent loads. Without live loads, longitudinal tendons push the bridge upward while there might not be enough dead load to weigh the bridge down. Therefore, this load combination must be checked to ensure ULS and SLS requirements are satisfied when the bridge is hogging. This is particularly critical for slender bridges since they have large prestressing forces and small dead loads. On the other hand, load combination SLS2 considers pure live loads when the bridge is loaded with only one CHBDC CL-625 truck; this combination is used for superstructure vibration check. 3.3.2 Live Loads Axle no. 1
Axle loads, kN 50 2 125 3 125 4 175 5 150

The live loads considered in the analysis comprise of CL-625 truck and CL-625 lane loads as
3.6 m 1.2 m 3 6.6 m 18 m CL-625 Truck Load 6.6 m

illustrated in Figure 3-4.

Axle no. Axle loads, kN 1 50 2 4 175 5 150 125 125

3.6 m

1.2 m

6.6 m 18 m CL-625 Truck Load Uniformly distributed load 9kN/m

6.6 m

Axle loads, kN

40

100

100

140

120

3.6 m

1.2 m

Axle loads, kN

40

100

100

6.6 m Uniformly distributed load 9kN/m 18 m 140 CL-625 Lane Load

6.6 m

120

Figure 3-4. Live loads: CL-625 truck load (top); CL-625 lane load (bottom)
3.6 m 1.2 m 6.6 m The CL-625 truck load is increased by a 6.6 m dynamic load allowance (DLA) factor for SLS1 and 18 m ULS load combinations as shown in Table 3-8 (CSA 2006, Cl. 3.8.4.5.3). This DLA factor accounts CL-625 Lane Load

for the load increase due to impact from truck vibrations and it depends on the number of axles that are loaded to produce the maximum force effect. The more the axles, the lower is the DLA factor, because the probability of all of the axles being in phase is low. Furthermore, the model is under multi-lane loading and the traffic load moments are modified according to the number of design lanes loaded to produce maximum force effects as shown in Table 3-9. These modification factors account for the probability of simultaneously loading more than one lane.
Table 3-8. DLA factor (CSA 2006) # of axles loaded 1 2 3 or more DLA factor 1.40 1.30 1.25 Table 3-9. Multi-lane loading modification factor (CSA 2006) # of loaded design lanes 1 2 3 4 5 6 or more Modification factor 1.00 0.90 0.80 0.70 0.60 0.55

24

3.4

Design Requirements
All analysis cases are designed to satisfy CHBDC ULS and SLS requirements by increasing the

amount of prestressing tendons and stirrups. Also, for the solid slab analysis, concrete strengths are increased (up to 80 MPa) in order to reduce the prestressing demands in slender cases since the dimensions of such cases cannot accommodate the large amount of tendons required to satisfy ULS. The ULS check consists of flexural and shear strength requirements while the SLS check includes stress, vibration and deflection limitations. In the analysis, such design checks only consider longitudinal behaviour and ignore transverse behaviour based on the assumption that transverse reinforcement demands remain constant for analysis cases with the same span lengths and bridge type. This assumption is valid, because transverse behaviour depends highly on live loads which are the same for cases with the same span lengths. Although dead loads vary for these analysis cases, they contribute very little to the transverse calculations due to the small influence lengths. This assumption does not affect the results of this comparative study in which comparisons are made between bridges with the same span length but different span-to-depth ratios, and the relative values of material consumptions and construction costs are more important than the actual values. 3.4.1 Ultimate Limit States Design Requirements

This section describes the ultimate limit states design requirements which include flexural and shear strength checks. 3.4.1.1 Flexural Strength In ULS, the flexural resistance (Mr) is a pure couple between compression in the concrete and tension in the longitudinal prestressing tendons. The analysis assumes that other non-prestressed reinforcements do not contribute to flexural strength and that cracked concrete has no tensile strength. Also, an equivalent rectangular concrete stress distribution is used as shown in Figure 3-5.

c = 3.5x10-3 c

1fc a= 1c z fpy C Mr = Tz T = pApfpy Forces

Strain

Stress

Figure 3-5. Flexural resistance: a) cross-section, b) concrete stains, c) equivalent concrete stresses, d) concrete forces

Therefore, Mr = Tz = pfpyApz for cases with bonded tendons in which the changes in tendon strain are assumed to be equal to the changes in strain of the surrounding concrete. A sample calculation is shown in Appendix B.1. For cases with unbonded tendons (i.e. span-by-span boxgirders), this assumption is no longer valid and a different approach is used (refer to Section 5.5).

25 Furthermore, since the analysis model is a statically indeterminate system, negative bending moments exist around the supports. In such regions, the bottom slab thickness of the box-girder is proportioned such that the compressive depth c is within the bottom flange. Other flexural requirements that are considered include minimum and maximum reinforcement requirements. The first requirement states that the flexural resistance must be greater than 1.2 times the cracking moment (Mcr) or 1.33 times the moment demand at ULS, whichever value is smaller (CSA 2006, Cl. 8.8.4.3). The second requirement states that c/d should be less than 0.5 (refer to Figure 3-5) (CSA 2006, Cl. 8.8.4.5). This requirement ensures that the steel has yielded when the concrete crushes at a strain of 0.0035 at the extreme compression fibre and that significant plastic deformation has developed prior to failure. 3.4.1.2 Shear Strength The CHBDC sectional design model is used to compute shear strengths and a sample shear calculation is shown in Appendix B.2. Shear resistance (Vr) comprises of three components: concrete, prestressing tendons, and stirrups. Concrete shear resistance depends on the cross-section while the prestressing shear resistance depends on the vertical component of prestressing force which is determined by flexural requirements. Therefore, the only independent variable that can increase shear strength is the amount of stirrups. In the analysis, the stirrups are at least 20M and the minimum spacing of stirrups is 300mm as required by CHBDC to support longitudinal tendon ducts and this spacing is reduced if it does not provide sufficient shear resistance (CSA 2006, Cl. 8.14.6). Each analysis case uses at most two different stirrup spacing since further discretization yields little changes in the final results as shown in Figure 3-6.
$2,500 $2,000 Cost per deck area ($/m2) $1,500 Steel cost $1,000 $500 $0 1 2 3 Number of stirrup spacing 4 $277 $260 $258 $257 Total construction cost $2,290 $2,273 $2,271 $2,270

Figure 3-6. Construction cost economy from increasing the number of stirrup spacing

Figure 3-6 compares the cost results of for an analysis case when the number of stirrup spacing increases from one to four. The graph is obtained by computing the material consumption and total construction cost for a span-by-span constructed, precast segmental box-girder with a span length of 40m and span-to-depth ratio of 20. As the number of stirrup spacing increases, the amount of

26 stirrups needed decreases, because the stirrup layout is further refined to match the shear demand. This trend is illustrated in the graph, but the graph also indicates that the savings in steel diminishes as the number of stirrup spacing increases beyond two. For instance, the steel cost decreases by 6% when two spacing instead of one spacing are used. On the other hand, the steel cost decreases by only 1% when the number of spacing increases from two to four. Therefore, only two different stirrup spacing are used in the analysis. 3.4.2 Serviceability Limit States Design Requirements

This section describes the serviceability limit states design requirements which comprise of stress, vibration and deflection limitations. 3.4.2.1 Stress For SLS stress checks, CHBDC poses crack width limitations if the stress exceeds concrete cracking stress. This study, however, uses a more conservative approach in which cracking is not permitted during service in order to minimize durability issues. The concrete stress needs to be less than the tensile strength of concrete in order to avoid cracking:
M SLS S

A < fcr

[3-1]

where S = sectional modulus P = prestressing force A = gross cross-sectional area

3.4.2.2 Vibration Vibration limitations are checked according to CHBDC (CSA 2006, Cl.3.4.4) which states that the deflections under load combination SLS2 must be less than the value described in Figure 3-7. All analysis cases are assumed to experience frequent pedestrian use which requires the most stringent vibration criterion.

Figure 3-7. Deflection limits for highway bridge superstructure vibration (CHBDC 2006)

27 3.4.2.3 Deflection The CHBDC does not have actual deflection restrictions and the limits used in American codes are optional, because deflections in highway bridges generally do not pose a severe serviceability problem as in railway bridges. However, excessive deflections affect rider confidence and cause durability issues such as ponding (MacGregor and Bartlett 2000). The analysis considers both long-term and short-term deflections due to permanent loads. Longterm deflection long-term should not exceed 1/750 of the span length (Menn 1990). Any deflection that exceeds the limit needs to be balanced by camber imposed during construction such that longterm deflection reduces to zero (Chen and Duan 1999). Long-term deflection accounts for the minimum upward deflection due to prestress after all losses and the maximum downward deflection due to creep:
long term = elastic + creep elastic = dead load + prestress creep = elastic where = creep coefficient 2.0 for CIP = 1.5 for precast [3-2]
after all losses

On the other hand, short-term deflection accounts for the maximum upward deflection due to prestress before losses and the minimum downward deflection due to instantaneous dead load:
short term = prestress
before losses

+ instantaneous

dead load

[3-3]

This check is particularly critical for cases with slender cross-sections in which large prestress forces can cause the bridge to hog when there is no live load. Excessive camber poses a problem and cracking might occur at regions that are originally designed to resist compression. Live load deflections are considered under the vibration requirement. The actual live load deflections have less impact on rider comfort compared to the acceleration of motion that riders feel on the bridge.

3.5

Other Preliminary Analysis Assumptions

Tendon eccentricity in duct is neglected. The centre of gravity of tendons is assumed to be at the centroid of the duct. The increase in concrete to accommodate intermediate tendon anchors and deviators is assumed to be negligible compared to the concrete volume of the entire superstructure. Every span is assumed to have the same number of prestressing tendons. The amount of prestressing does not vary to suit the demand for each span. Substructure design is not considered in the analysis.

Other analysis assumptions include:

ANALYSIS OF CAST-IN-PLACE ON FALSEWORK BRIDGES

This chapter discusses the analysis of cast-in-place on falsework box-girder and solid slab. First,

Section 4.1 examines this construction method. The subsequent sections describe the analysis models for the two bridge types by considering the cross-sections as well as the prestressing tendon layouts. Lastly, this chapter summarizes the analysis results including structural behaviours, vibration limits, deflections, material consumptions, and the factors that restrict further increase of slenderness ratios.

4.1

Cast-in-Place on Falsework Construction

In cast-in-place on falsework construction, falsework for the entire bridge is assembled first.

The falsework needs to support concrete formwork as well as the full dead load of the bridge during construction. After concrete is placed, longitudinal internal tendons are installed into ducts and grout is placed inside the ducts such that tendons become bonded to the concrete. When the concrete gains sufficient strength, tendons are stressed and as a result, the girder hogs and is released from the formwork. The formwork and falsework are then removed. This is a labourintensive and slow method because it requires not only falsework erection but also on-site placement of reinforcing steel and concrete. Falsework is also expensive and it can disrupt traffic below the bridge. Moreover, cast-in-place concrete is subjected to on-site temperature and humidity changes, so greater effort is needed to ensure good quality compared to precast concrete which is manufactured in a controlled environment. Despite these disadvantages, the cast-in-place on falsework method is still used today, especially in regions where labour is inexpensive, primarily due to its simplicity (Gauvreau 2006). This flexible construction method can also be applied to bridges with tight curves and complex geometry (Hewson 2003). Furthermore, it provides opportunities for aesthetic expression since it can readily accommodate different geometries.

4.2
4.2.1

Cast-in-Place on Falsework Box-Girder

Model

Analysis is performed on 21 cases with interior span lengths of 35m, 50m, 60m, and 75m and span-to-depth ratios of 10, 15, 20, 25, 30, and 35. This set of span lengths is chosen because cast-inplace on falsework box-girders are economical for spans up to about 80m according to Menn (1990) and bridges with longer span lengths need to be haunched in order to reduce dead loads. The end spans are made 10m shorter than interior spans to balance moments along the entire bridge and to simplify the treatment of prestressing in the study. If the same span length is used throughout the 28

29 bridge, the end spans have greater moments than interior spans. A comparison of the moments obtained from the two span arrangements is shown in Figure 4-1.
-300 -200 -100 0 Moment 0 (kNm) 100 200 300 Constant span length Reduced span length at end spans 50 100 150 200 250 300 350 400 Distance (m)

Figure 4-1. Moment comparison of bridges with constant span length and reduced end span length

Typical span-to-depth ratios for this type of bridge range from 12 to 35 (Menn 1990). Therefore, a similar range is used in the analysis. According to Menn (1990), the most economical ratio is 15, but if both economic and aesthetic impacts are considered, the most optimal range becomes 17 to 22. 4.2.1.1 Cross-Section A typical cross-section used in the analysis is shown in Figure 4-2. Dimensions for the box components are based on values suggested by Gauvreau (2006). The deck is 15m wide and supports four 3.5m wide design lanes and two 0.5m wide barriers. A minimum thickness of 225mm is used for the top slab. This thickness is assumed to be sufficient to accommodate transverse tendons and to resist punching shears from wheel loads. The top slab thickness increases to 375mm near the webs such that the deck slab cantilever has enough strength and stiffness to resist transverse bending. The intersection of the top slab and web occurs at the quarter point of the deck slab in order to reduce transverse bending in the web. The web width is 450mm in order to accommodate reinforcing bars, internal bonded tendons, concrete clear covers and the spacing required to facilitate concrete placement and vibration. This width remains constant along the entire girder. On the other hand, the bottom slab thickness varies from a minimum of 200mm at midspan to the depth of the compressive stress zone created by negative moments at the supports.

Figure 4-2. Typical cross-section for cast-in-place on falsework box-girder

30 The reinforcement arrangement illustrated in Figure 4-3 is used and is assumed to be adequate in resisting transverse bending moments for all analysis cases. The stirrup spacing varies based on shear requirements; the remaining reinforcing steel is defined as minimum reinforcing steel in this study and the layout is the same for all analysis cases.

Figure 4-3. Typical reinforcing steel layout

4.2.1.2 Prestressing Tendon Layout The parabolic tendon layout used for this type of bridge is shown in Figure 4-4. The tendon is at the highest possible location at the supports and the lowest possible location at midspan as allowed by clear cover requirements of the webs. At the abutments, however, the tendon is located at the centroidal axis of the cross-section in order to eliminate unbalanced prestress moments. Ideally, the tendon layout should be parabolic between adjacent supports to provide maximum upward deviation forces, but this layout results in abrupt corners over the supports. For practical purposes, the tendon profile needs to be concave downward at the supports and the inflection point occurs at one-fifteenth of the span length (L/15).

Figure 4-4. Typical tendon profile

In this study, the tendon is assumed to extend from one end of the bridge to another and be stressed in one operation in order to simulate non-segmental construction. However, this layout is generally not feasible in actual construction, because the installation is difficult and the prestress loss due to friction is excessive for long bridges. In a real situation, segmented construction with staged prestressing is required. This results in a more practical tendon layout in which adjacent tendons would overlap at intermediate anchors. For the purpose of estimating loss of prestress, the more practical layout is used and the tendon is assumed to extend over one span only. The

31 overlapping of tendons between adjacent spans is not considered. Based on this layout, the effective prestress after all losses is estimated to be 60% fpu (1120MPa). 4.2.2 Analysis Results

This section summarizes the analysis results which include the structural response under ULS and SLS, the material consumptions, and the factors that limit further increase of slenderness ratios. 4.2.2.1 Structural Behaviour and Dimensioning Table 4-1 describes the ULS strength and SLS stress at the most critical location, the dimensioning of shear reinforcement, and the factor that determines the amount of prestress required for each analysis case. The sizing of prestressing tendons is described in Section 4.2.2.4.
Table 4-1. Summary of structural response and dimensioning of cast-in-place on falsework box-girder L L/h (m) 10 15 20 25 10 15 20 25 30 10 15 20 25 30 35 10 15 20 25 30 35 Ultimate limit states __Flexural strength__ __Shear strength__ MULS Mr MULS/Mr Av smin % of girder 2 (kNm) (kNm) (mm ) (mm) @ smin 37000 38200 97% 1200 35100 35600 99% 2000 33600 34800 97% 2000 243 9.0% 32100 32400 99% 2000 200 22% 75100 67400 63300 61200 60300 115000 99100 90200 85100 81300 73500 190000 161000 155000 142000 149000 143000 78400 68600 70700 69900 73600 118000 99700 96500 92800 88900 83300 199000 164000 173000 147000 166000 146000 96% 98% 90% 88% 82% 98% 99% 94% 92% 91% 88% 100% 98% 90% 97% 90% 98% 1200 1200 1200 2000 2000 1200 1200 1200 1200 1200 1200 1200 1200 1200 1200 1200 2000 224 147 102 133 106 219 138 99 76 63 52 182 112 79 59 47 60 5.4% 6.2% 17% 16% 18% 5.0% 8.0% 23% 29% 33% 30% 7.2% 13% 24% 29% 37% 28% Serviceability limit states __Stresses__ SLS (MPa) SLS/fcr 0.80 0.86 1.07 1.66 1.12 1.91 1.71 2.10 2.47 1.16 1.83 2.01 2.46 2.33 2.45 1.46 1.84 2.10 1.97 2.17 2.09 29% 31% 38% 59% 40% 68% 61% 75% 88% 41% 65% 72% 88% 83% 87% 52% 66% 75% 70% 78% 75% Governing factor for prestress requirement ULS flexural ULS flexural ULS flexural ULS flexural ULS flexural ULS flexural SLS stress SLS stress SLS stress ULS flexural ULS flexural SLS stress SLS stress SLS stress SLS stress ULS flexural ULS flexural SLS stress SLS stress SLS stress SLS stress

35

50

60

75

First, under ULS, the table describes the relationship between the flexural strength demand (MULS) and resistance (Mr). Table 4-1 also summarizes the stirrup requirements for satisfying shear demand by listing the stirrup area (Av), minimum stirrup spacing (smin) as well as the percentage of girder that needs stirrups to be placed at a spacing of smin. The remainder of the girder requires stirrups spaced at every 300mm which is the minimum spacing prescribed by CHBDC (CSA 2006,

32 Cl.8.14.3). All analysis cases, except for the cases with span of 35m and ratios of 10 and 15, need stirrup spacing of less than 300mm near the supports where shear forces are greatest. Cases with larger span lengths and span-to-depth ratios require small stirrup spacing of less than 100mm which should be increased by using a bigger reinforcement bar (i.e. larger than 20M) for construction purposes. However, the spacing described in Table 4-1 is used in this study, because the required volume of stirrup varies little with respect to stirrup spacing and it is of greater concern than the spacing in cost comparisons. Table 4-1 also compares SLS stresses with the factored cracking stress. SLS stresses (SLS) cannot exceed the concrete tensile stress (fcr = 2.8 MPa) to avoid cracking during service. Lastly, the table shows the factor that governs the amount of prestress needed for each analysis case. The governing factor is either ULS flexural strength or SLS stress; ULS shear strength does not govern since stirrups are added to resist any extra shear that is not balanced by the tendons and concrete. The ULS flexural resistance is proportional to the cross-sectional depth while the SLS stress depends on the sectional modulus which is related to the second power of the depth. For cases with low span-to-depth ratios, the sectional moduli are large and the SLS stresses are relatively small, thus ULS flexural strength governs the prestress requirement. However, as span-to-depth ratio increases, the sectional modulus decreases at a faster rate than the cross-sectional depth (Figure 4-5). As a result, SLS stress becomes more critical for slender girders and thus, it governs the prestress requirement for these cases.
35 30 25 Sectional modulus S (m3) or depth h (m) 20 15 10 h 5 0 5 10 15 20 L/h 25 30 35 40 Sbottom fibre Stop fibre

Figure 4-5. Changes in sectional modulus and cross-sectional depth

4.2.2.2 Vibration Limits Figure 4-6 describes the vibration deflection limits for bridges with frequent pedestrian use and the truck load deflections under SLS2. The truck deflections are acceptable for all analysis cases and vibration limits do not govern the prestressing requirement. The truck deflections are at least 58% less than the vibration limits.

33
80 L=75m Limit 60 Truck load deflection (mm) L=60m Limit 40 L=50m Limit 20 L=35m Limit 0 5 10 15 20 L/h 25 L=35m 30 35 40 L=50m L=75m L=60m

Figure 4-6. Deflection for superstructure vibration limitation

4.2.2.3 Deflections The following graphs summarize the dead load, long-term and short-term deflections in terms of span length over deflection (L/). As expected, deflection increases with increasing span lengths and span-to-depth ratios. The maximum camber required is 0.5m for the case with a span of 75m and span-to-depth ratio of 35 such that the long-term deflection essentially becomes zero.
12000 L=35m

Dead load deflection (L/) [down]

8000

4000

L=50m L=60m L=75m

0 5 6000 Long-term deflection (creep + elastic) (L/) [down] L=35m 10 15 20 L/h 25 30 35 40

4000

2000

L=50m L=60m L=75m 5 10 L=35m L=50m L=60m L=75m 15 20 25 30 35 40

0 L/h

15000 Short-term deflection 10000 (dead load+ prestress before loss) 5000 (L/) [down] 0 5

10

15

20

L/h

25

30

35

40

Figure 4-7. Deflections: a) dead load, b) long-term, c) short-term

34 4.2.2.4 Material Consumption Table 4-2 and Figure 4-8 summarize the material consumption for each analysis case as well as their variations from the baseline case (i.e. L/h=20, which is the conventional ratio defined in Chapter 2). These results are also illustrated in cross-section drawings in Appendix C.1.
Table 4-2. Summary of material consumptions for cast-in-place on falsework box-girder L (m) L/h 10 15 20 25 10 15 20 25 30 10 15 20 25 30 35 10 15 20 25 30 35 Volume 3 (m ) 2700 2410 2260 2174 4450 3860 3570 3420 3350 5890 5000 4600 4380 4270 4190 8290 6970 6340 6030 5850 5910 Concrete % change from baseline case +19% +6.5% 0% -3.9% +25% +8.1% 0% -4.0% -6.0% +28% +8.9% 0% -4.7% -7.1% -8.8% +31% +9.9% 0% -4.9% -7.7% -6.8% Prestressing tendon Number of % change from tendons baseline case 52 -52% 76 -30% 108 0% 136 +26% 76 100 144 184 240 92 120 160 192 248 300 136 184 240 300 384 528 -47% -31% 0% +28% +67% -43% -25% 0% +20% +55% +88% -43% -23% 0% +25% +60% +120% Reinforcing steel Mass % change from (ton) baseline case 216 +16% 193 +3.8% 186 0% 187 +0.5% 360 314 304 306 315 476 410 404 405 415 416 674 585 571 579 618 621 +18% +3.2% 0% +0.7% +0.4% +18% +1.5% 0% +0.2% +2.7% +3.0% +18% +2.5% 0% +1.4% +8.2% +8.8%

35

50

60

75

As shown in the graphs, when the girder becomes more slender, concrete volume decreases gradually at a decreasing rate. This trend exists because concrete volume depends highly on the cross-sectional depth which also decreases at a declining rate as span-to-depth ratio increases (Figure 4-8 a & b). Also, at higher span-to-depth ratios, the concrete volume varies less since the reduction in web concrete is counterbalanced by the increase in bottom slab thickness. For instance, for the cases with span length of 50m, concrete volume decreases by 880m3 as span-to-depth ratio increases from 10 to 20, but concrete volume only decreases by 220m3 as span-to-depth ratio increases from 20 to 30. In the latter case, the concrete reduction in the webs is 284m3, but the total concrete reduction is only 220m3 due to increase in bottom slab thickness. On the contrary, the amount of prestress increases as span-to-depth ratio increases, because slender bridges have lower flexural resistances and require larger prestress forces (Figure 4-8 c).

35 The number of tendons increases at an increasing rate because for slender cases with large prestress demand, the tendons need to be placed in more than one layer within the webs, thus lowering the resistance moment lever arm. The prestress layout becomes more inefficient as slenderness increases, resulting in an even greater demand for tendons to provide the same level of prestress. Reinforcing steel mass is attributed to the longitudinal and transverse steel reinforcing bars as well as the stirrups. The amount of steel reinforcing bars needed is proportional to the crosssectional area because they are placed according to Figure 4-3. Consequently, the graph for reinforcing steel mass follows the trend of concrete volume for cases with low span-to-depth ratios (Figure 4-8 d). These cases have deep girders that can resist shear using mostly stirrups installed with the minimum spacing needed to support the tendons (i.e. 20M bars spaced at 300 mm), so the required amount of reinforcing steel depends on the concrete volume. For the more slender girders, more stirrups are needed to resist shear and thus the reinforcing steel graph no longer follows the same trend as the concrete volume graph at higher span-to-depth ratios.
a) Concrete volume
10000 8000 6.0 6000 Concrete volume (m3) 4000 2000 0 5 10 15 20 L/h 25 30 35 40 L=75m L=60m L=50m L=35m Depth 4.0 h (m) 2.0 L=35m 0.0 5 10 15 20 L/h 25 30 35 40 L=50m L=75m L=60m

b)

Cross-sectional depth
8.0

c)

Number of prestressing strands

600 500 400 L=75m

d)

Reinforcing steel mass

800 L=75m 600

Prestress strands 300 200 100 0 5 10 15 20 25 L/h 30

Reinforcing L=60m steel mass 400 (ton) L=50m L=35m 200

L=60m L=50m L=35m

0 35 40 5 10 15 20 L/h 25 30 35 40

Figure 4-8. Material consumptions for cast-in-place on falsework box-girder

36 4.2.2.5 Limiting Factors of Span-to-Depth Ratios The upper bound of span-to-depth ratio for cast-in-place on falsework box-girders is restricted by the number of prestressing tendons that can fit inside the webs and by the minimum height of the interior box cavity. First, the maximum feasible span-to-depth ratio for cases with span lengths of 75m is limited to 35 due to the tendon arrangement within the webs. This case requires sixteen 37strand ducts arranged in four layers in the webs (current tendon arrangement in Figure 4-9). For cases having ratios beyond 35, more tendons are needed due to the reduced moment lever arm and section modulus, but they cannot be accommodated efficiently within the 450mm thick webs according to the current tendon arrangement. This inefficient arrangement of tendons decreases the eccentricity of tendons, thus further increasing the prestressing demand. Figure 4-9 compares the current tendon arrangement to a more efficient arrangement in which three ducts are placed in one layer in each web, thus increasing the eccentricity (i.e. h1 > h2). However, since the web thickness is kept constant for all cases in this study, the more efficient tendon arrangement is not used. The current arrangement cannot efficiently accommodate tendons needed for the cases with span lengths of 75m and slenderness ratios beyond 35.

More efficient tendon arrangement

Current tendon arrangement

h1

h2

Figure 4-9. Tendon arrangement that limits further increase in span-to-depth ratio

If the more efficient tendon arrangement from Figure 4-9 is used, the prestressing demand is expected to decrease. For example, for the case with span length of 75m and ratio of 35, using the more efficient arrangement increases the eccentricity by 89mm and decreases the prestressing demand by 3.0%. This decrease in tendon is relatively minor, because the advantage from additional tendon eccentricity is offset by the increase in dead load (i.e. concrete volume increases by 4.3% due to the thicker webs). With thicker webs, the maximum feasible span-to-depth ratio for cases with span lengths of 75m becomes 40 at which point the limiting factor is the minimum height requirement of the interior box cavity. Therefore, the results of this analysis depend on the assumed web thickness. Using thinner webs requires more prestressing tendons and reduces the maximum feasible span-to-depth ratio while larger web thickness reduces prestressing demand and expands the feasible range of ratios.

37 Another factor that restricts span-to-depth ratios is the minimum height requirement for the interior box cavity. As shown in Figure 4-10, the interior box cavity needs to be at least 1.0m such that workers have enough space to strip forms, stress tendons and perform maintenance and repairs.

height of access

Figure 4-10. Interior box cavity limitation

Although this height limit is not required by CHBDC, it is often used as good construction practice. This height requirement restricts the ratio for cases with span lengths of 35m, 50m, and 60m to 25, 30, and 35 respectively. As span-to-depth ratio increases, more tendons are used and the bottom slab becomes thicker in order to accommodate the compressive stress zone in negative moment regions. As the bottom slab thickness increases, the height of interior box cavity shortens. If this minimum height restriction does not exist, the box would essentially turn into a solid slab for the more slender cases (Figure 4-11).

Figure 4-11. Height of access diminishes as span-to-depth ratio increases

38

4.3

Cast-in-Place on Falsework Solid Slab

As shown in the previous analysis, box-girders are not feasible for slender cases (i.e. span-to-

depth ratio greater than 30) due to practical construction considerations. Any further reduction in cross-sectional depth essentially turns the box into a solid slab. This finding leads to the analysis of cast-in-place on falsework solid slabs described in this section. From the aspect of construction cost optimization, investigating span-to-depth ratios for solid slabs is valuable, because concrete savings as a result of reduced depth in solid slabs would be more than those in box-girders. As indicated by the shaded areas in Figure 4-12, reducing depth in solid slab eliminates a strip of concrete as wide as the soffit while reducing depth in box-girder only removes the web concrete.

Figure 4-12. Concrete reduction due to increase in span-to-depth ratio for solid slab and box-girder

Cast-in-place on falsework solid slabs are economical for short-span bridges due to simple formwork, prestress layout and concreting operations (Hewson 2003). The straightforward construction does not require a high level of technology or an extensive amount of labour. However, solid slabs are inefficient in terms of structural behaviour, because they result in relatively large dead loads and need more prestressing to get sufficient flexural stiffness and resistance compared to box-girders (Menn 1990). Due to its excessive dead load, solid slabs are generally used for shorter spans of less than 20m (Gauvreau 2006). To reduce the dead load for long span cases, voided slabs (Figure 4-13) are used instead of solid slabs. In voided slabs, stay-in-place forms are used to create the hollow cores. These forms must be anchored against uplift during concreting and they need vents and drainage openings, thus complicating the construction process. The voids also pose durability issues since inspection is not possible inside the voids. Due to the construction complications and durability concerns, voided slabs are not considered in the analysis despite the savings in concrete and reinforcing steel.

Figure 4-13. Voided slab

39 4.3.1 Model

The analysis considers 18 cases with span lengths of 20m, 25m, 30m, and 35m and span-todepth ratios of 30, 35, 40, 45, and 50. According to Gauvreau (2006), the maximum cost-effective slab depth is 0.8m and the maximum practical span-to-depth ratio is 25, resulting in a maximum span of 20m. This analysis uses a higher range of spans and ratios in order to demonstrate the impacts of slenderness. Bridges with spans shorter than 20m cannot achieve the proposed span-todepth ratios because the girders would be too slender to accommodate sufficient reinforcement (e.g. a 15m long span with a ratio of 35 only has a depth of 0.43m). Also, this set of ratios is chosen since the commonly used value is 30 based on the review of existing bridges in Chapter 2. Furthermore, span-to-depth ratios below 30 are not investigated, because a deep cross-section with a depth greater than approximately 0.8m has large dead loads and is not economical (Menn 1990). 4.3.1.1 Cross-Section A typical cross-section and reinforcement layout are shown in Figures 4-14 and 4-15. Compared to the box cross-section considered previously, the solid slab model has a wider deck (i.e. 22m) that supports six design lanes and shorter deck cantilevers (i.e. 3.75m) in order to emphasize slab behaviour. Also, the deck cantilevers are tapered to reduce dead load which is much higher in solid slabs than in box-girders. Longitudinal internal bonded tendons are grouped in bands over support lines based on recent practices in North America (Park and Gamble 2000). To simplify analysis, a solid wall pier is assumed and thus, only one band of tendons that spread over the entire spine of the cross-section is used. To carry loads into the longitudinal tendon band, transverse tendons are uniformly distributed along the length of the bridge.

Figure 4-14. Typical cross-section for cast-in-place on falsework solid slab

Figure 4-15. Typical reinforcing steel layout

4.3.1.2 Prestressing Tendon Layout The model has the same tendon layout as the one from the box-girder model (Figure 4-4).

40 4.3.2 Strip Method versus Beam Model

Designing reinforcement based on elastic sectional forces only is complicated in slabs, because bending moments exist in two orthogonal directions. To simplify the design, two methods are considered: strip method and beam model. Results from these two methods are compared in this section for three analysis cases: 1) L=25m, L/h=30; 2) L=30m, L/h=40; 3) L=35m, L/h=50. The strip method, proposed by Hillerborg (1996), generates lower bound solutions of the theory of plasticity. In contrast to yield line theory, the strip method solution provides adequate flexural safety at ULS. This method states that if the moments can be distributed such that equilibrium equations are satisfied and the reinforcements are designed for these moments, then the slab is safe at ULS. The solution only needs to fulfill equilibrium equations and not necessarily the compatibility criteria. More than one solution is possible since the slab is statically indeterminate. For design purposes, the solution that yields reinforcement economy and favorable behaviour under service conditions is used. In the simple strip method, load is assumed to be carried by strips that run in the longitudinal and transverse reinforcement directions. These strips are treated as beams and the moment in each strip can be solved using simple statics. To further simplify analysis, torsional moments are assumed to be zero in these strips. This moment distribution is preferred because torsional moments require more reinforcement. One way to apply the simple strip method in prestressed slabs is to divide the slab into strips that contain one tendon each. Since all the tendons are equal, strip width is varied in order to balance the load. Menn (1990) proposed a straightforward way to redistribute moments in slabs. In order to minimize cracking, the prestressing deviation forces (qp) are chosen to be 60-80% of the total dead load. To simplify analysis and construction, parallel prestressing tendons are arranged into narrow bands. These bands of tendons are idealized as beams in the longitudinal direction and slabs can redistribute moments onto the bands to reduce peak stresses. To account for this moment redistribution onto these reinforcement bands, a self-equilibrating moment (msp) is used:
msp = mp m0p where mp = moment where q p is applied to the original slab -m0p = moment when-qp is applied to the idealized slab in which the prestressing bands act as individual beams [4-1]

This self-equilibrating moment varies across the width of the cross-section since additional positive moments are concentrated at the reinforcement bands. Reinforcements are designed for the final moments after redistribution. Design moments in the analysis are computed based on Menns prestressing concept. First, the structure is divided into two 3.75m wide edge strips and one 14.5m wide prestress band as shown in

41 Figure 4-16 b. Longitudinal tendons are distributed evenly in the prestress band. The two edge strips are narrow relative to the prestress strip in order to ensure the deck cantilever load would travel into the prestress band due to the short load path. To compute m0p, - qp
b b0

is applied to the

prestress band where qp is 80% of the total dead load, b is the total slab width and b0 is the width of prestress band. Likewise, mp is obtained by applying qp to the entire slab. The redundant moment (msp) is the difference between mp and m0p. The redundant moment transfers forces from the edge strips to the prestress band. Since this moment is self-equilibrating, the total moment remains the Support Support Support same after redistribution as shown in Figure 4-16 g. Reinforcement is then designed such that the
Support A following is satisfied the factored moment (Figure 4-16 h):
A A
A A A

Support

Support Support

Support

Support
Support
Support

DL mDL + LL mLL + p msp < Mr where =load factor Mr =moment resistance

Support
Support
Support

[4-2]

The transverse distribution of longitudinal bending moment across the slab and the final design A moment are illustrated in Figure 4-16.
a) Elevation: Support Support
Edge strip Edge strip Edge strip
A b = 22m b = 22m b = 22m A A Prestress band b0 = 14.5m Prestress band b0 = 14.5m b band b = 14.5m = 22m Prestress 0 A A b = 22m Prestress band b0 = 14.5m A A b = 22m Prestress band b0 = 14.5m A
b = 22m Prestress band b0 = 14.5m

Edge strip Edge strip Edge strip

Support Support

Edge strip

Edge strip

Edge strip

Edge strip
Edge strip
Edge strip

b)

Section A-A:
Edge strip
Edge strip

2.00 2.00 2.00

c)

Dead load and live load moments mDL + mLL

Edge strip Edge strip

Prestress band b0 = 14.5m b = 22m (assumeb = 22mm ): m =

DL LL

2.00

Prestress band b0 = 14.5m Prestress band b0 = 14.5m

d)

-m0P due to application of deviation forces qp on prestress band: mP due to application of deviation forces qp on the entire slab width:

2.00 1.21 1.21 1.21 Edge strip Edge strip 2.00 1.21 1.21
2.00 0.80 0.80 0.80

e)

1.21

f)

Self-equilibrating moment mSP = -m0P + mP:

1.21

g)

Unfactored design moment = mDL + mLL + msp:

1.20 1.20 1.20
2.41 2.41 2.41

0.41 0.41 0.41 1.21 1.21 0.41

0.80 2.00 2.00 0.80 0.80 0.80 0.80 0.80 0.80 0.80m +m mDL+mLL DL 0.80= 2.00 LL m2.00 LL = DL+m = 2.00 0.80 0.80mDL+mLL 2.14 2.14 0.80= 2.00 2.14
m +m 0.80 DL LL
mDL+mLL 2.14= 2.00 0.80 0.80m +m DL LL 2.14= 2.00
2.14 mDL+mLL = +m m 2.00

h)

Design moment = DLmDL + LLmLL + pmsp: 1.20

2.41

0.41
3.29 0.41 3.29 3.29

1.20
2.41
2.41

Figure 4-16. Transverse distribution of longitudinal bending moment in slabs (values are divided by mDL)= 2.00 2.14 0.41
1.20
1.20
2.41

3.29

1.20 1.20

3.29 0.41 0.41 3.29

3.29

2.41

42 The three analysis cases are also investigated under the assumption that the slab behaves like a beam, so there is no moment redistribution. The longitudinal tendon requirements for ULS flexure determined from the two methods are summarized in Table 4-3.
Table 4-3. Results from beam model and strip method L=25m, L/h=30 L=30m, L/h=40 L=35m, L/h=50 Beam Strip % diff. Beam Strip % diff. Beam Strip % diff. 230 176 23% 360 351 2.5% 500 486 2.8% $ 2170 $ 2110 2.8% $ 2480 $ 2540 2.4% $ 2580 $ 2570 0.4%

Tendon requirement Total construction cost per deck area

The table shows that the prestress requirements and total construction costs obtained from the two methods are similar. Strip method does not change the results significantly because the analysis model is predominantly a simple, one-way system, so the moment distribution is similar to that of a beam. Therefore, the simpler beam assumption is used for the analysis. 4.3.3 Analysis Results

4.3.3.1 Structural Behaviour and Dimensioning Table 4-4 describes the ULS strengths and SLS stress at the most critical location and the dimensioning of shear stirrups for each analysis case. The sizing of prestressing tendon is described in Section 4.3.3.5. For all cases, the ULS flexural resistance (Mr) is at least 27% greater than the demand (MULS), so flexural strength does not govern the prestress requirement. However, to satisfy the maximum reinforcement criterion, larger concrete compressive strengths (fc) are needed. This criterion is discussed in greater details in Section 4.3.3.2. On the other hand, shear strength requirements can be fulfilled with the listed stirrup areas (Av) distributed at a minimum spacing of 300mm as described in CHBDC. Only minimum stirrups are needed because the cross-sectional widths and concrete strengths are large, thus resulting in high concrete shear resistances. Lastly, SLS stresses govern the prestressing requirement for all cases. Critical stress occurs at the top fibre of the support which is subjected to large negative moments.

43
Table 4-4. Summary of structural response and dimensioning of cast-in-place on falsework solid slab L (m) L/h MULS (kNm) 20 30 35 40 30 35 40 30 35 40 45 30 35 40 45 17500 15900 14800 28600 25500 23400 43900 39300 36500 32800 68900 61200 55900 51800 Ultimate limit states __Flexural strength__ Mr (kNm) MULS/Mr fc (MPa) 20000 19100 18900 34000 31600 30000 49600 45800 44300 42900 81100 74300 68100 64100 86% 83% 78% 84% 81% 79% 89% 86% 82% 76% 85% 82% 82% 81% 50 50 75 50 50 60 50 50 60 75 50 50 60 80 _Shear strength_ 2 Av (mm ) 600 600 600 600 1000 1000 600 1000 1000 1000 600 1000 1000 1000 Serviceability limit states __Stresses__ SLS (MPa) SLS/fcr 2.50 2.63 3.04 2.54 2.42 2.64 2.67 2.75 3.03 3.03 2.79 2.54 2.92 3.48 89% 93% 88% 90% 85% 85% 94% 97% 98% 87% 98% 90% 94% 97%

25

30

35

4.3.3.2 Maximum Reinforcement Criterion According to the maximum reinforcement criterion in CHBDC (CSA 2006, Cl.8.8.4.5), the flexural resistance at ULS should be developed with a c/d of less than 0.5 at regions where the moment capacity is close to the demand (c is the height of compressive stress region while d is the distance between the extreme compressive fibre and the tendon as shown in Figure 4-17). This ensures that the tendons would already be yielding and significant plastic deformations have occurred when concrete crushes such that the structure would fail in a ductile manner. This limitation is not satisfied for the more slender analysis cases. Slender bridges have short moment lever arms and thus large prestress requirements. A large tensile force in the tendon needs to be balanced by an equally large concrete compressive force, meaning that c is large. Slender bridges have small d due to physical limitation and large c due to moment requirements, resulting in large c/d ratios. For cases that do not satisfy this criterion, concrete strengths are increased such that prestress requirement and c decrease. Concrete strengths that are needed to satisfy the maximum reinforcement limit are summarized in Table 4-5. These higher strength concretes are more expensive and the additional expenses will be accounted for in the cost comparative studies in Chapter 6. For the cases that are crossed out in the table, concrete with strengths greater than 80 MPa are needed to fulfil the criterion. Such high strengths are not widely used in todays bridge industry and hence, the maximum span-to-depth ratio is limited by the maximum reinforcement criterion and the maximum practical concrete strength.

44
Table 4-5. Concrete strengths required to satisfy maximum reinforcement criterion

cu = -3.5x10 -3 c
p

0.85fc d C

fc required to satisfy c/d<0.5 limit (MPa) L/h 30 35 40 45 50 20m 50 50 75 150 25m 50 50 60 90 30m 50 50 60 75 135 35m 50 50 60 80 115

z fp T Figure 4-17. Maximum reinforcement criterion: a) concrete stains, b) equivalent concrete stresses, c) concrete forces

4.3.3.3 Vibration Limits Figure 4-18 describes the vibration deflection limits for bridges with frequent pedestrian use and the truck load deflections under SLS2. The truck deflections are acceptable for all analysis cases and vibration limits do not govern the prestressing requirement. The truck deflections are at least 52% less than the vibration limits. The deflection increases as the girder becomes more slender until the span-to-depth ratio reaches 35 for span lengths of 20m and 25m, and 40 for span lengths of 30m and 35m. These cases use larger concrete strengths as described previously (Section 4.3.3.2), resulting in greater stiffness as well as lower deflections and flexural frequencies (indicated by triangular markers on Figure 4-18).
50

40 Truck load 30 deflection (mm) 20

L=35m limit L=30m limit L=25m limit

10

L=20m limit

L=25m L=20m 30 35 40 45

L=35m L=30m

0 25 L/h 50

Figure 4-18. Deflection for superstructure vibration limitation

45 4.3.3.4 Deflections The following graphs summarize the dead load, long-term and short-term deflections in terms of span length over deflection (L/). The maximum camber required is 0.075m such that the long-term deflection essentially becomes zero for the case with a span of 35m and slenderness ratio of 40. As explained previously, cases with increased concrete strengths have lower deflections; these cases are indicated by triangular markers.
12000 Dead load deflection (L/) [down] L=20m 8000 4000 0 25 12000 Long-term deflection (creep + elastic) (L/) [down] 8000 4000 0 25 30000 Short-term deflection 20000 (dead load + prestress before loss) 10000 (L/) [up] 0 25 30 35 L/h 40 30 35 L/h 40 45 50 L=20m L=25m L=30m L=35m 30 35 L/h 40 45 50 L=25m L=30m L=35m

L=20m L=25m L=30m L=35m

45

50

Figure 4-19. Deflections: a) dead load, b) long-term, c) short-term

46 4.3.3.5 Material Consumption Table 4-6 and Figure 4-20 summarize the material consumption for each analysis case and its variation from the baseline case (i.e. L/h=30, which is the conventional ratio defined in Chapter 2). The cases with higher concrete strengths are italicized in the table and are indicated by triangular markers in the graphs. These results are also illustrated in cross-section drawings in Appendix C.2.
Table 4-6. Summary of material consumption for cast-in-place on falsework solid slab L (m) 20 L/h 30 35 40 30 35 40 30 35 40 45 30 35 40 45 Concrete % change from baseline case 0% -10% -18% 0% -11% -19% 0% -11% -20% -26% 0% -12% -20% -27% Prestressing tendon Number of % change from tendons baseline case 210 0% 264 +26% 322 +53% 260 312 364 300 350 414 480 414 480 528 576 0% +20% +40% 0% +17% +38% +60% 0% +16% +28% +39% Reinforcing steel Mass % change from (ton) baseline case 85.7 0% 86.9 +1.4% 88.0 +2.7% 111 113 114 124 130 131 132 148 156 156 156 0% +1.7% +2.5% 0% +4.8% +5.4% +5.8% 0% +5.0% +5.2% +5.3%

Volume 3 (m ) 2150 1930 1770 3210 2860 2600 4270 3790 3430 3160 5710 5050 4550 4170

25

30

35

Contrary to the non-linear decline in box-girders, concrete volume decreases at a relatively steady rate as slenderness increases for solid slabs (Figure 4-20 a). In box-girders, concrete volume diminishes at a decreasing rate since web concrete reduction is offset by bottom slab thickening. In contrast, increasing span-to-depth ratio in solid slabs simply removes an entire strip of slab, so the concrete volume reduction is directly proportional to the decrease in girder depth (Figure 4-20 b). The amount of prestressing is expected to increase at an increasing rate as slenderness rises based on observations described in Section 4.2.2.4 for box-girder. However, Figure 4-20 c shows that the prestress increases linearly with slenderness since larger concrete strengths are used and fewer tendons are needed for the slender cases. Also, solid slabs can accommodate a large number of tendons at the same elevation, so the tendons do not need to be placed in layers which reduce the prestressing efficiency. As shown in Figure 4-20 d, the amount of reinforcing steel remains relatively constant as spanto-depth ratio varies. For the cases with spans of 20m, the stirrup requirement is the same for all three cases. For the cases with the other span lengths, as span-to-depth increases from 30 to 35, the required stirrup area increases from 600mm2 to 1000mm2 but spacing remains at 300mm. This

47 increase in stirrup area has little effect because stirrups account for less than 10% of the total amount of reinforcement steel. As a result, the amount of reinforcement steel does not deviate more than 5% from the baseline case.
a) Concrete volume
6000

b)

Cross-sectional depth
1.2

4000 Concrete volume (m^3) 2000

L=35m L=30m L=25m L=20m

0.8 Depth (m) 0.4 L=25m L=20m

L=35m L=30m

0 25 30 35 L/h 40 45 50

0 25 30 35 L/h 40 45 50

c)

Number of prestressing strands

800

d)

Reinforcing steel mass

200

600 Prestress strands

L=35m L=30m

150 Reinf. steel 100 mass (ton) 50 L=25m L=20m

L=35m L=30m

400

L=25m L=20m

200

0 25 30 35 L/h 40 45 50

0 25 30 35 L/h 40 45 50

Figure 4-20. Material consumptions for cast-in-place on falsework solid slab

4.3.3.6 Limiting Factors of Span-to-Depth Ratios Maximum span-to-depth ratio in solid slabs is limited by the maximum reinforcement criterion which states that the height of the compressive stress region should be less than half the distance from the extreme compressive fibre to the centroid of tensile force (i.e. c/d < 0.5). This criterion is not satisfied for cases with ratios as low as 35 if a concrete strength of 50 MPa is used. In order to fulfil the criterion, cases with higher ratios use larger concrete strengths which reduce prestress demands (i.e. results in a smaller compressive stress region) and improve ductility. The maximum concrete strength used in the analysis is 80 MPa which limits the span-to-depth ratio to 40 for cases with spans of 20m and 25m, and to 45 for cases with spans of 30m and 35m.

ANALYSIS OF PRECAST SEGMENTAL SPAN-BY-SPAN BOXGIRDER

This chapter examines the optimization of span-to-depth ratio for precast segmental span-by-

span constructed box-girders. Section 5.1 provides a general description of the construction method; Section 5.2 discusses the analysis model; Section 5.3 examines the construction moments and moments due to thermal gradients; Section 5.4 explores the loss of prestress; Section 5.5 discusses the behaviour of unbonded tendons at ultimate limit states; and Section 5.6 summarizes the analysis results.

5.1

Precast Segmental Span-by-Span Construction

Precast segmental span-by-span construction method is ideal for long bridges with many short-

to-medium spans, because it has a high speed of construction when repetitive spans are used. For instance, a typical 40m span can be erected every 2 to 3 days using this method (Hewson 2003). Also, this method can be used for bridges that require access beneath the superstructure since the erection equipment is mainly above ground and the disruption to traffic below the bridge is minimized (Sauvageot 1999). The construction sequence is illustrated in Figure 5-1. In this method, an erection girder, either an overhead truss or underslung girder, is first supported on a previously completed deck or adjacent piers. Then precast segments are transported from the completed span and a crane located at the edge of the completed deck stacks the segments loosely on the erection girder. If an overhead truss is used, the segments are hanged with cables (Figure 5-2 a). If an underslung girder is used, segments are supported under the deck cantilevers or soffit (Figure 5-2 b). After all the segments of an entire span are placed, they are connected together with temporary prestress. A cast-in-place closure joint next to the pier segment stitches the new span with the previously completed span (Hewson 2003). Then permanent longitudinal external tendons, which overlap the previous span, are installed and the new span becomes continuous with the completed structure. The erection girder then launches forward to erect the next span. Since precast segments are used and tendons are only stressed once for the entire span, this method increases the ease and speed of construction. Efficiency in construction is further enhanced with the use of external tendons which increases the speed of precasting (Sauvageot 1999).

48

Direction of Construction Precast segments are placed onto erection girder Direction of Construction Erection girder
Erection girder

49

Direction of Precast segments are placed onto erection girder Construction Precast segments are placed onto erection girder

Erection girder

Prestressing tendons are installed

Prestressing tendons are installed

Prestressing tendons are installed

Erection girder launches forward

Erection girder launches forward

Erection girder launches forward

Figure 5-1. Precast segmental span-by-span construction method

Figure 5-2. Span-by-span erection girder: a) overhead truss, b) underslung girder (NRS 2008, OSHA 2006)

5.2

Model
Analysis is performed on 11 cases with span lengths of 30m, 40m and 50m and span-to-depth

ratios of 15, 20, 25, and 30. This set of span lengths is chosen based on lengths of erection girders that are typically used in the industry today. The maximum length is 50m because longer erection equipment would become too heavy and this method would no longer be economical (Sauvageot 1999). The range of slenderness ratios chosen for analysis is based on the conventional optimal ratio of 17 and the minimum cross-sectional depth of 1.8m (Gauvreau 2006). This depth is larger than the minimum depth used in cast-in-place on falsework box-girder analysis (i.e. 1.4m) since the precast bridge is longitudinally post-tensioned with external unbonded tendons which shorten the moment lever arm compared to internal tendons.

50 5.2.1 Cross-Section

A typical cross-section and reinforcement layout are shown in Figures 5-3 and 5-4. They are modified from AASHTO-PCI-ASBI standard sections (1997). Except for box height and bottom slab thickness, all other dimensions remain the same for every analysis case. The box webs are more slender compared to those for cast-in-place on falsework because they do not need to accommodate internal tendons. As a result, web thickness depends only on shear and transverse bending requirements. The bottom slab thickness depends on the thickness of the compressive stress zone at negative moment regions. The deck is 15m wide and can support four lanes of traffic which is the same as the cast-in-place on falsework case.

Figure 5-3. Typical cross-section for precast segmental span-by-span box-girder

Figure 5-4. Typical reinforcing steel layout

5.2.2

Elevation and Prestressing Tendon Layout

In precast segmental span-by-span construction, the girder is composed of interior segments, pier segments, and closure joints. The maximum length of interior segment is chosen to be 3m based on limitations of transporting segments from casting yard to construction site. The pier segments are 1m long and they support the anchorages of all longitudinal tendons. The 150mm long

51 cast-in-place closure joint fills the gap between the pier segments and interior segments and it allows for adjustments during placement of segments. Elevation and typical tendon layout of an end span is illustrated below.

Figure 5-5. Typical tendon profile

As shown in the diagram, longitudinal tendons are continuous between adjacent piers and they connect the precast segments of each span. Since the tendons are externally unbonded, a straight tendon profile is used instead of a parabolic profile. Each tendon begins at the centroidal axis of the girder for end spans or directly below the top slab for the interior spans. The tendon then continues within the box cavity towards the bottom slab. It is deviated at 1/3 of the span and assumes a horizontal profile. When the tendon reaches 2/3 of the span, it deviates again and continues towards the top slab of the pier segment. The tendon is then anchored at the exterior side of the pier segment. This anchorage location allows tendons from adjacent spans to overlap at the pier segments, thus providing continuity throughout the superstructure.

5.3

Longitudinal Bending Moments

This section discusses two types of longitudinal bending moments that are particularly critical

in precast segmental bridges: construction moments and thermal gradient moments. 5.3.1 Construction Moments

Because a segmental method is used, the structural system changes during construction and the final moments are affected by the stress history. Therefore, the moments at each stage of construction is computed to obtain the final moments at completion. The following illustration describes the dead load and prestressing moments (MDL and MP,tot) for a case with span of 40m and ratio of 20 at various construction stages. The first eight graphs show the moments caused by the additional new span at each construction stage. The last graph shows the final moments at the instant of completion which are obtained by summing the construction moments at all construction stages.

Figure 5-6. Construction moments for segmental span-by-span method

Stage 1: -40000 -20000 0 Moments (kNm) 0 20000 40000 Stage 3: MDL MP,tot 40 Distance (m) Moments (kNm) Stage 2: -40000 -20000 0 0 20000 40000 Stage 4: MDL 40

MP,tot 80 Distance (m)

-40000 -20000 0 Moments (kNm) 0 20000 40000 Stage 5: -40000 -20000 0 Moments (kNm) 0 40 80 Distance (m) 40 80

MP,tot 120

-40000 -20000 0 Moments 0 (kNm) 20000 40000 Distance (m) 40 80 120

MP,tot 160

MDL

MDL

MP,tot 120 160 200 Distance (m) MDL

20000
40000

52

Figure 5-6. Construction moments for segmental span-by-span method (continued)

Stage 6: -40000 -20000 0 Moments (kNm) 0 20000 MDL 40000 Stage 7: -40000 -20000 0 Moments (kNm) 0 20000 40000 Stage 8: -40000 -20000 0 Moments (kNm) 0 20000 40000 Final moments at the instant of completion: The final moments are the sum of all the construction moments. -40000 -20000 0 Moments (kNm) 0 20000 40000 MDL 40 80 120 160 200 240 280 MDL 40 80 120 160 200 240 280 MDL 40 80 120 160 200 240 40 80 120 160 200

MP,tot 240 Distance (m)

MP,tot 280 Distance (m)

MP,tot 320 Distance (m)

MP,tot 320 Distance (m)

53

54 In the span-by-span method, the dead load moments at the instant of completion are biased towards positive moment due to stress history compared to the same bridge built on falsework. However, the instantaneous span-by-span moments (MSBS) would redistribute over time and shift towards falsework moments (MFAL) because of creep. This redistribution of moments is approximated by the following formula (Menn 1990):
M = MSBS + (MFAL + MSBS ) [5-1]

The factor is chosen to be 0.7 for precast concrete. If cast-in-place concrete is used, would be 0.8 since it is more susceptible to effects of creep than precast concrete. The following graph shows the redistribution of dead load moments for a case with L=40m and L/h=20.
-40000 -20000 Dead load moment (kN) 0 0 20000 40000 Falsework Span-by-span M at infinity Moment at infinity 40 80 120 160 200 240 280 320 Distance (m)

Figure 5-7. Redistribution of dead load moments due to creep

When considering dead load moment only, significant redistribution of moments causes the moment demand at midspan to decrease and demand at supports to increase. However, prestress moments should also be included in the redistribution as shown in the following graph.
-40000 -20000 Dead load + 0 prestressing 0 moments (kN) 20000 40000

40

80

120

160

200

240

280

320 Distance (m)

Falsework

Span-by-span

M at infinity Moment at infinity

Figure 5-8. Redistribution of dead load and prestress moments due to creep

As illustrated in the graph, redistribution of the sum of prestress and dead load moments is negligible. Therefore, moment redistribution is not considered in the analysis and the instantaneous moments are used as moment demands for ULS and SLS checks. 5.3.2 Moments due to Thermal Gradient

For heavy bridges exposed to solar radiation, thermal gradients exist across the cross-section depths, meaning that the top deck is warm while the bottom flange is cold. As a result, the girder deflects upward under no external loads.

55 In a simply-supported girder, thermal gradient only causes deflection and not moments. On the other hand, it imposes a moment on continuous girders since there are internal restraints. These thermal gradient moments are directly proportional to the flexural stiffness. Therefore, thermal gradient is a particularly important issue for span-by-span constructed bridges in which crosssections are generally stiffer. Thermal gradient causes positive moments, which can crack the bottom flange, so additional bottom tendons are required in order to prevent such cracking. Also, thermal gradient moments are considered only in serviceability design checks, because the cracked flexural stiffness is small in ultimate limit states and thus thermal gradient moments are negligible. In the analysis, thermal gradient effects are considered for serviceability limit states and are obtained using the method described as follows. First, the continuous girder is made staticallydeterminate by removing internal restraints. A thermal differential (y) of 10C is then applied between the 225mm thick top deck and the rest of the box (CSA 2006, Cl.3.9.4.4). The free strain due to this temperature change is obtained by the following formula:
f = c (y) where c = thermal coefficient of concrete = 10 106 / [CSA 2006, Cl.8.4.1.3] [5-2]

This free strain profile follows the shape of the temperature change and is non-linear if unrestrained expansion is allowed at all elevations. However, the final strain profile (y) needs to be linear since plane section remains plane. The difference between the free strain and final strain profiles gives the restraint strain which is needed to restore compatibility. The thermal gradient and strain profiles are illustrated in the following diagram.

Figure 5-9. Thermal gradient effects

These self-equilibrating, restraint stresses due to non-linear thermal profile on a staticallydeterminate structure causes the primary temperature stresses fr(y):
fr y = Ec y f = Ec y c y [5-3]

56 The axial restraint force Pr and restraint moment Mr can then be computed from fr (y):
h h

Pr = Mr =

fr y b y dy = Ec
h

y c y b y dy
h 0

[5-4] [5-5]

fr y b y y n dy = Ec

y c y b y y n dy

However, both Pr and Mr are equal to zero because there are no external forces and internal redundancies have been removed to make the structure statically-determinate. By setting Equations 5-4 and 5-5 to zero, the final curvature of bending () and final strain (0) at elevation y=0 are obtained as follows. A complete derivation is shown in Appendix B.3.
c I c 0 = A =
h

y b y y n dy
0 h

[5-6] [5-7]

y b y dy n
0

According to the final strain profile, the final strain at elevation y can be obtained by:
y = 0 + y [5-8]

As a result, the primary thermal stresses become:

fr y = Ec 0 + y c (y) [5-9]

In a statically-determinate structure, these primary stresses only cause the girder to hog, but there is no net moment. However, in a continuous structure, the curvature created by primary stresses is incompatible and are restrained by piers. In order for the structure to maintain compatibility while a thermal deformation is imposed, positive restraint moments are needed at each end of the span:
M = -Ec I [5-10]

The final moment Mfinal is obtained using moment distribution (Figure 5-10). This positive restraint moment causes tensile stresses at the bottom fibre near the piers where the post-tensioning are arranged to resist negative moments. If the positive restraint moments are large enough to crack the bottom flange, additional bottom tendons would be required near the piers such that the resultant moments become negative.
0 0 Moments due to 2000 thermal gradient (kNm) 4000 6000 40 80 120 Distance (m) 160 200 240 280 320

Figure 5-10. Moments due to thermal gradient

57 Furthermore, secondary thermal stresses fs are induced by Mfinal:

fs y = Mfinal (y n) I [5-11]

The total thermal stresses that need to be considered in design checks are computed with the following equation:
ftemp y = fr y + fs y [5-12] Mfinal (y n) = Ec 0 + y c (y) + I

5.4

Loss of Prestress
Since external, unbonded tendons are used, the loss of prestress due to friction is lower and the

effective prestress after all losses is expected to be greater than 60%fpu (1120 MPa) which is the estimated value used for the cast-in-place on falsework analyses. Therefore, an explicit calculation of prestress losses is performed. With a more detailed calculation, the increase in prestress is less than 6% compared to the assumed value of prestress after all losses (1120 MPa = 60% fpu). This increase in stress is not significant enough to reduce the prestressing requirement. 5.4.1 Friction Losses

Friction losses are proportional to the deviation angles of tendons. Thicker sections have larger deviation angles and thus higher friction losses. The decrease in prestressing force due to friction losses P(x) are computed with the following formulae (Menn 1990).
P x = P0 1 e x if x 0.2 P0 x if x < 0.2 [5-13]

where x=distance from stressing location P0 =jacking force p0 =jacking stress=80%fpu =1490MPa =coefficient of friction=0.25 for external ducts x =sum of angle changes between stressing location and point x =x +x x =intentional angle changes refer to Figure 5-11 =1 +2 ++n =unintentional angle change due to construction tolerances and displacement of tendon during concreting =0 for external tendons

Figure 5-11. Intentional angle changes

58 The friction losses are summarized in the Table 5-1. The maximum loss is only 5% of fpu.
Table 5-1. Prestress losses due to friction Friction losses (% of fPu) L/h 15 20 25 30m 4.33% 2.70% 40m 4.91% 3.30% 2.37% 50m 5.09% 3.56% 2.62% L/h 15 20 25 Table 5-2. Prestress losses due to anchorage set Anchorage set losses (% of fPu) 30m 6.11% 4.68% 40m 5.63% 4.62% 3.76% 50m 5.12% 4.29% 3.68%

The loss of prestress due to anchorage set is a function of friction losses. The anchorage set is assumed to be 6mm and the prestress loss due to anchorage set is computed using Equation 5-14. (Collins and Mitchell 1997). The prestress losses due to anchorage set are summarized in Table 5-2.
set Ap Ep P x = 2p p where p= friction loss per unit length [kN/m] set =anchorage set = 6mm [5-14]

5.4.2

Creep and Shrinkage Losses

Creep and shrinkage losses for girders with external, unbonded tendons are computed using Equation 5-15 which is modified from Menns formula for bonded tendons using new compatibility conditions (Menn 1990).
n Ac t P t = 1 lp c0 x dx + cs t Ec Ac 1+ Ac l p Ic e2 x dx [5-15]

1 + n 1 + t

where t = time measured from initial loading e(x) = distance from centroid of gross uncracked concrete section to centroid of tendon Ec = modulus of elasticity of concrete at time of initial loading Ep = modulus of elasticity of prestressing tendon n = Ep /Ec Ac = area of gross uncracked concrete section Ap = area of tendon Ic = moment of inertia of gross uncracked concrete section t = time-varying creep coefficient = 1.5 for precast concrete = 2.0 for cast-in-place concrete = aging coefficient = 0.8 (Menn 1990) cs t = time-varying shrinkage strain under relative humidity of 60% Toronto value from CHBDC c0 x = concrete stress at tendon level due to initial load lp = arc length of tendon between anchors Integrations are made over length of tendon from anchors to anchors

The creep and shrinkage losses are summarized in Table 5-3.

59
Table 5-3. Prestress losses due to creep and shrinkage Creep and shrinkage losses (% of fPu) L/h 15 20 25 30m 7.85% 6.30% 40m 6.24% 9.71% 5.91% 50m 2.10% 6.11% 4.35%

5.4.3

Losses due to Relaxation of Prestressing Steel

Prestress loss due to relaxation of steel (p,rel) is estimated to be 59.3 MPa (3.19% fpu) for all cases according to the following relationship suggested by Menn (1990).
10 8 6 4 2 0 0.5 0.55 0.6 p0/fpu 0.65 0.7 0.75

p,rel /p0 [%]

Figure 5-12. Long-term loss of prestress due to relaxation (Menn 1990)

5.4.4

Total Prestress Losses

The effective prestress after all losses for every case is summarized in the following table. It is the average value along the length of the tendon.
Table 5-4. Effective prestress after all losses Effective prestress after all losses (% of fPu) 30m 40m 50m 61% 64% 62% 60% 64% 66% 63% 66%

L/h 15 20 25

As shown in the table, compared to the estimated value of effective prestress (60% fPu), the gain in stress is less than 6% after an explicit calculation. This gain in stress has an insignificant impact on prestress requirements, so an effective prestress after all losses of 60% fpu is used in serviceability checks.

60

5.5

Behaviour of Unbonded Tendons at Ultimate Limit States

The ULS flexural strength calculations for span-by-span method differ from those for previous

analyses. Previously, strain in bonded tendons can be directly solved with compatibility equations at the specific section of interest. In contrast, span-by-span method uses unbonded tendons, the tendon strain does not equal to the concrete strain at the same elevation and plane, and thus the tendon stress cannot be determined locally. Furthermore, since strains in unbonded tendons are averaged out between the anchorages, girders with unbonded tendons often have lower flexural resistances compared to girders with bonded tendons. Figure 5-13 compares the behaviours of the two types of tendons.

Figure 5-13. Compatibility conditions for bonded and unbonded tendons (Collins and Mitchell 1997)

To solve this problem, the unbonded tendon stress at ULS can be assumed to equal to the effective prestress after all losses (60%fpu) instead of the yielding stress (90%fpu), because girders with unbonded tendons often have lower flexural resistances than those with bonded tendons (Menn 1990). This conservative assumption causes the ULS be a governing factor for the prestress requirement and results in extra tendons that are unnecessary for satisfying ULS limits. In order to compute the actual prestress force in unbonded tendons at ULS, an iterative process is used. This process computes the global deformation of the entire tendon instead of just the strains

61 at one plane. The iterative method is explained in greater details in Appendix B.4. The prestress at ULS obtained from explicit calculations are summarized in the following table:
Table 5-5. Prestress at ULS Prestress at ULS (% of fPu) L/h 15 20 25 30m 73% 75% 40m 75% 76% 82% 50m 69% 75% 75%

These prestress at ULS are larger than the effective prestress after all losses, meaning that fewer tendons are needed compared to the conservative method proposed previously.

5.6

Analysis Results
This section summarizes the analysis results including structural behaviours at ULS and SLS,

material consumptions to satisfy design requirements as well as the factors that limit the slenderness of span-by-span precast box-girders. 5.6.1 Structural Behaviour and Dimensioning

The following table describes the ULS flexural strength and SLS stress at the most critical location as well as the stirrup spacing requirement to satisfy ULS shear. The sizing of prestressing tendons is discussed in Section 5.6.4. ULS flexural strength demand governs the prestress requirement for all cases because unbonded tendons result in a lower tendon stress that is averaged over the entire tendon length. Moreover, thermal gradient moments are not large enough to cause overall positive moments at the piers, thus bottom tendons are not needed at these regions.
Table 5-6. Summary of structural response of precast span-by-span box-girder L (m) 30 L/h 15 20 15 20 25 15 20 25 Ultimate limit states __Flexural strength__ __Shear strength__ 2 MULS (kNm) Mr (kNm) MULS/Mr Av (mm ) smin (mm) % of girder @ smin 38700 42900 90% 1200 207 15% 38700 43300 89% 2000 247 4.5% 79400 65500 65400 101000 97800 97900 83600 68600 71000 102000 99400 99700 95% 95% 92% 99% 98% 98% 1200 1200 1200 1200 1200 1200 141 165 119 189 149 116 21% 27% 32% 23% 29% 33% Serviceability limit states __Stresses__ SLS (MPa) SLS/fcr 0.58 21% -0.22 1.81 1.99 1.95 -0.10 1.49 1.96 67% 71% 70% 53% 70%

40

50

5.6.2

Vibration Limits

As shown in the following graph, vibration limits are satisfied for all analysis cases and do not affect tendon requirements. The truck deflections are at least 40% lower than the vibration deflection limits.

62
40

30 Truck load deflection (mm)

L=50m Limit L=40m Limit

20 L=50m L=40m 10 L=30m Limit 0 10 15 20 L/h 25 30 L=30m

Figure 5-14. Deflection for superstructure vibration limitation

5.6.3

Deflections

Deflection is not a limiting factor since the cross-sections are stiff and the resulting deflections are insignificant. The maximum camber required is 93mm.
20000 Dead load deflection (L/) [down] 15000 10000 5000 0 10 15 20 L/h 15000 Long-term deflection (creep + elastic) (L/) [down] 25 30 L=30m L=40m L=50m

L=30m L=40m L=50m

10000

5000

0 10 60000 Short-term deflection (dead load+ 40000 prestress before loss) 20000 (L/) [up] 0 10 15 20 L/h 25 30 15 20 L/h 25 30

L=40m L=30m L=50m

Figure 5-15. Deflections: a) dead load, b) long-term, c) short-term

63 5.6.4 Material Consumption

The following table and graphs summarize the material consumption for each analysis case as well as their variations from the baseline case (i.e. L/h=15). These results are also illustrated in cross-section drawings in Appendix C.3. Like the cast-in-place box-girder case, as slenderness ratio increases, concrete volume less compared to the increase in tendons. For instance, when the concrete volume decreases by only 4% from the baseline case, the number of prestressing tendons increases by 52% for the case with span of 30m and ratio of 20. On the other hand, reinforcing steel mass remains similar for all the cases with the same span length.
Table 5-7. Summary of material consumption for precast span-by-span box-girder L (m) 30 L/h 15 20 15 20 25 15 20 25 Volume 3 (m ) 1920 1840 2780 2660 2600 3640 3400 3350 Concrete % change from baseline case 0% -4.0% 0% -4.4% -6.6% 0% -6.7% -8.1% Prestressing tendon Number of % change from tendons baseline case 138 0% 210 +52% 198 224 296 204 256 333 b)
L=50m 3,000 Concrete volume 2,000 (m3) 1,000 0 10 15 20 L/h 25 30 L=40m L=30m 300 Prestress strands 200 100 0 10 15 20 L/h 25 30 L=30m

Reinforcing steel Mass % change from (ton) baseline case 159 0% 149 -5.8% 236 227 231 307 294 299 0% -3.9% -2.2% 0% -4.2% -2.5%

40

0% +13% +49% 0% +25% +63%

50

a)

Concrete volume
4,000

Number of prestressing strands

400 L=50m L=40m

c)

Reinforcing steel mass

400

300 Reinforcing steel mass 200 (ton) L=30m 100 0 10 15 20 L/h 25

L=50m L=40m

30

Figure 5-16. Material consumptions for precast span-by-span box-girder

64 5.6.5 Limiting Factors of Span-to-Depth Ratios

For all three span lengths analyzed (i.e. 30m, 40m, and 50m), the maximum slenderness ratio is limited by the minimum height of access inside the box. As stated previously in Section 4.2.2.5, a minimum access height of 1.0m is used to allow for sufficient space for construction, inspection and maintenance. The maximum slenderness ratios are restricted to 20 and 25 for the cases with spans of 30m and 40m respectively. For these two extreme cases, the access height is determined by the height of the interior box cavity as demonstrated in Figure 5-17 which shows the cross-section for the case with span of 40m and ratio of 25.

Figure 5-17. Access limited by height of interior box cavity

For the case with a span of 50m, the maximum span-to-depth ratio is 25 as shown in Figure 518. Compared to the previous two cases, the access height is further reduced due to the use of external tendons. If the slenderness ratio is increased to 30, the height of the interior box cavity and the access height would reduce to 1.24m and 0.83m respectively.

Figure 5-18. Access limited by height of interior box cavity and external tendons

This minimum access height limit is not a rigid design requirement and is used merely to ensure ease of construction. Higher span-to-depth ratios can be achieved for all three span lengths if only ULS and SLS requirements are considered. However, as shown in Figures 5-17 and 5-18, the addition prestress associated with higher slenderness ratios can only be accommodated with a second layer of tendons. Such tendon layout is less efficient, because it reduces the tendon eccentricity and as a result, more prestressing is needed. Due to the inefficient tendon configuration, these slender cases are expected to be less cost-effective. Nonetheless, these cases are not considered in this study since the minimum access requirement is not satisfied.

COST COMPARISONS
This cost study investigates the changes in superstructure and total construction costs, which are

based on the previously described material consumption results, when span-to-depth ratio varies. Comparison of these costs reveals the optimal span-to-depth ratio for each bridge type. The cost study also demonstrates the cost benefits of using the optimal ratios instead of conventional ratios. Furthermore, a sensitivity analysis is performed to examine the impacts of changing material unit prices and construction cost breakdown.

6.1

Material Costs
This section describes the material unit prices and provides a material cost comparison of all the

analysis cases. 6.1.1 Material Unit Prices

The unit costs for concrete, prestressing tendons and reinforcing steel are listed in Table 6-1. They are based on values obtained from cast-in-place post-tensioned highway bridges in Ontario (SNC-Lavalin 2008). All unit prices include installation costs (i.e. concrete placement and vibration, grouting of tendons).
Table 6-1. Material unit prices Material Concrete (fc = 50 MPa) Material only Formwork, falsework, and labour Longitudinal prestressing tendons (including anchorages) Reinforcing steel Unit 3 per m 3 per m 3 per m per kg per kg Unit price $ 1500 $ 250 $ 1250 $ 8.5 $ 5.0

6.1.1.1 Concrete Material Unit Price Concrete unit price varies with compressive strength, because higher strengths require more cementing material and admixtures. Material unit prices for concrete with strengths from 10 MPa to 60 MPa are summarized in Table 6-2. For strengths greater than 60 MPa, values are extrapolated from Figure 6-1 as indicated by hollow points. These prices account for the material only and do not include grouting, pumping, formwork, falsework, and labour (Dufferin Concrete 2009).

65

66
Table 6-2. Concrete material unit price Concrete compressive strengths (MPa) 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 Unit price (per m ) $ 149 $ 154 $ 158 $ 164 $ 171 $ 180 $ 192 $ 220 $ 250 $ 262 $ 287 $ 320 $ 353 $ 389 $ 428
3

$500 $400 $300 Extrapolated values

Concrete material price ($/m3) $200 $100 $0 0 20

40 60 80 Concrete strength (MPa)

100

Figure 6-1. Concrete material unit price

6.1.1.2 Cast-in-Place versus Precast Concrete The final results of this cost study are insensitive to the cost of casting and erection equipment since the study does not compare the costs between different construction methods and cost comparisons are mainly performed for bridges of the same type with varying spans and slenderness ratios. Therefore, in the analysis, the unit price of precast concrete is assumed to be the same as the one for cast-in-place concrete. In a real construction setting, the casting operation would cost relatively the same for different span lengths and construction methods, but the initial cost of precast is generally higher than castin-place concrete, because it needs to be custom-built and stored in a manufacturing plant. Yet, precast segments are more economically competitive when they are mass-produced. So, the unit price of precast concrete depends on the length of the bridge. Precast concrete would likely cost more if a short, 2-span bridge is analyzed instead of a long, 8-span viaduct. However, precast concrete increases the speed of construction and reduces the risk of downtime created by having additional on-site operations (e.g. in cast-in-place construction, workers need to build formworks, install rebars, pour and vibrate concrete on-site). Precasting also eliminates other construction problems related to cast-in-place such as poor quality control. These advantages of precast concrete cannot be easily quantified, so they are not considered in this cost study. As stated previously, the unit cost of precast concrete depends on the total length of the bridge due to the relatively high initial cost associated with precasting. This dependency becomes less significant when precast concrete are more widespread and the level of standardization increases. The study anticipates that precast segments will be readily available like standard I-sections, thus the impact of initial costs is reduced. As a result, the unit price of precast concrete is assumed to be the same as cast-in-place concrete in the cost analysis.

67 6.1.1.3 Falsework versus Erection Truss Span-by-span method has a cost reduction compared to cast-in-place on falsework since it does not require falsework. Some cost gain is expected because a multiple-span viaduct allows for the reuse of erection truss for each span while falsework needs to be assembled for the entire structure. This cost advantage only exists for long bridges with repetitive spans. For shorter bridges with fewer spans, the use of falsework might be more economical since the initial cost of erection truss is relative high. Another factor that affects the cost of span-by-span bridges is the span length, because the cost of erection equipment increases as span length increases. The cost of erection truss becomes prohibitive when the span length is greater than 50m due to the heavy dead load (Hewson 2003). On the other hand, the unit price for falsework remains relatively constant as span length increases. Therefore, a more extensive study is needed to investigate the relationship between cost, span length and number of spans for cast-in-place on falsework and span-by-span methods, but this is beyond the scope of this research. 6.1.1.4 Formwork The unit price of formwork increases as the cross-section becomes more complex. For instance, solid slabs require less complicated formworks compared to box-girders which have more surfaces. Likewise, formworks for bridges with external tendons are simpler because the cross-section has no internal tendon ducts. Variation in formwork price with respect to cross-sectional complexity is not considered in the cost analysis and a constant concrete unit price is used for all cross-sectional types. 6.1.1.5 Prestressing Tendons Both internal bonded and external unbonded tendons have the same unit price of $8.5/kg. Although external unbonded tendons are more expensive in terms of material costs, they require less labour due to the simple and rapid assembly. As a result, the cost is assumed to be the same as internal bonded tendons. 6.1.2 Material Cost Comparisons

This section compares the material costs per deck area for all the analysis cases. Only the superstructure material costs are compared because they are directly related to the span-to-depth ratios whereas other components of construction costs (i.e. mobilization, substructure, and accessories) are assumed to be independent of superstructure slenderness. The comparisons are illustrated using a series of graphs in which the maximum cost variation for each span length is indicated as a percentage next to the label for each line.

68 6.1.2.1 Concrete Cost Comparison Figure 6-2 summarizes the pure material cost of concrete of all the analysis cases evaluated. In general concrete cost decreases with increasing slenderness ratio. For solid slabs, however, cost increases drastically for the very slender cases which require concrete strengths higher than 50 MPa.
$300

$250

L=35m, 29% L=30m, 43% L=75m, 42% L=60m, 40% L=20m,42%

$200 Cost of concrete (material only) per deck area ($/m2) $150

L=50m, 33% L=35m, 24% L=25m, 12%

L=50m, 8.8% L=40m, 7.0% L=30m, 4.2%

$100

$50

L=20m CIP on falsework solid slab L=35m CIP on falsework box-girder L=30m Precast span-by-span box-girder 5 10 15 20 25 30 L/h 35 40 45 50 55

$0

Figure 6-2. Concrete material cost comparison

$1,600 L=75m, 42% $1,400 L=60m, 40% $1,200 L=50m, 33% L=35m, 24% $1,000 Cost of concrete per deck area ($/m2) $800 L=35m, 22% L=30m, 24% L=25m, 20% L=20m, 11%

L=50m, 8.8% L=40m, 7.0% L=30m, 4.2%

$600

$400 L=20m CIP on falsework solid slab L=35m CIP on falsework box-girder L=30m Precast span-by-span box-girder 5 10 15 20 25 30 L/h 35 40 45 50 55

$200

$0

Figure 6-3. Total concrete cost comparison

69 The cost effect of increasing concrete strength diminishes when the costs of formwork and falsework or erection truss are included as shown in Figure 6-3. In general, concrete cost decreases as span-to-depth ratio increases due to the reduction in superstructure depth. However, decrease in concrete cost slows down at higher span-to-depth ratios, because the cost reduction is offset by the increase in bottom slab thickness to accommodate compressive stress regions in box-girders or by the increase in concrete strengths to satisfy maximum reinforcement criterion in solid slabs. Concrete cost also varies substantially between different cross-section types even when the same span length and span-to-depth ratio are used. For instance, for cases with a span of 35m, the concrete costs are $1500/m2 for a solid slab with span-to-depth ratio of 30 and $836/m2 for a castin-place box-girder with ratio of 25. Concrete cost for the thinner solid slab (depth = 1.2m) is 79% more than the one for the deeper box-girder (depth = 1.4m) even though the cross-sectional depths vary by only 17%. This comparison indicates a significant reduction in concrete consumption, which translates into dead load reduction, occurs when a box section is used instead of a solid slab. 6.1.2.2 Prestressing Cost Comparison Unlike concrete, the cost of prestressing tendons varies substantially as span-to-depth ratio rises as shown in Figure 6-4.
$500 $450 $400 $350 Cost of prestress $300 steel per deck 2) area ($/m $250 $200 $150 $100 $50 $0 5 10 15 20 25 30 L/h 35 L=75m, 284% L=60m, 223% L=50m, 209% L=35m, 161% L=20m CIP on falsework solid slab CIP on falsework box-girder L=35m L=30m Precast span-by-span box-girder 40 45 50 55 L=50m, 62% L=40m, 49% L=35m, 39%

L=30m, 60%

L=25m, 40% L=20m, 53%

L=30m, 53%

Figure 6-4. Prestressing tendon cost comparison

The maximum cost differences amongst cases with the same span length are 284% for cast-inplace on falsework box-girder, 60% for cast-in-place on falsework solid slab, and 62% for precast

70 span-by-span box-girder. The increases in prestressing demand are significant because although the dead load (i.e. concrete volume) decreases as girder becomes more slender, this load reduction cannot compensate for the relatively large decline in moment resistance which is attributed to the decrease in flexural stiffness of the cross-section and in efficiency of tendon layout (tendon eccentricity is lower for slender sections). This problem is illustrated in Table 6-3 which shows the maximum change in concrete volume, girder depth, moment of inertia (proportional to flexural stiffness), and prestress demand between the case with the lowest span-to-depth ratio and the case with the highest ratio for a particular span length for each bridge type. For all three bridge types, the decrease in dead load is small relative to the decrease in girder depth and modulus of elasticity. As a result, more prestressing steels are needed to offset the reduced load capacity.
Table 6-3. Comparison of changes in cross-sectional depth and prestressing demand Factor Concrete volume Girder depth Moment of inertia Prestress demand CIP on falsework box-girder (L=75m) -29% -71% -94% +284% CIP on falsework solid slab (L=30m) -26% -33% -70% +60% Precast span-by-span box-girder (L=50m) -8.1% -40% -69% +62%

A comparison of prestressing steel costs between bridge types confirms that solid slab is the least efficient in resisting loads. For example, for a span length of 35m, the prestressing tendons required for a solid slab with a span-to-depth ratio of 30 is $275/m2 which is 2.4 times greater than the $117/m2 needed for a cast-in-place box-girder with span-to-depth ratio of 25. Solid slabs require more prestressing per unit deck area because they are heavier than box-girders with the same span lengths and depths. Precast span-by-span box-girder is more efficient than solid slabs but less efficient than cast-in-place box-girder, because it uses external tendons which reduce tendon eccentricities and thus lowers the prestress moments. For instance, for the cases with span length of 50m and span-to-depth ratio of 25, the prestressing tendons required in a precast box-girder ($215/m2) cost 30% more than the tendons needed in a cast-in-place box-girder ($165/m2) although the precast case has 7.6% less dead load. 6.1.2.3 Reinforcing Steel Cost Comparison Cost of non-prestressed reinforcing steel depends on the shear stirrup requirement and concrete volume. Figure 6-5 shows the cost of stirrups in solid lines and the cost of minimum reinforcing steels in dashed lines. In this study, the minimum reinforcing steel is defined as all the nonprestressed steels other than stirrups and is illustrated in Figures 4-3, 4-15 and 5-4.

71
$400

$350

L=75m, 42% L=60m, 40% L=50m, 33%

$300

CIP on falsework solid slab L=20m L=35m CIP on falsework box-girder L=30m Precast span-by-span box-girder Stirrups L=50m Minimum reinforcing steel L=30m

Cost of $250 reinforcing steel per deck area $200 ($/m2) $150

L=35m, 24%

L=50m, 8.8% L=40m, 7.0% L=30m, 4.2% L=20m, 22% L=25m, 23% L=30m, 35% L=35m, 37% L=75m, 110% L=60m, 84% L=50m, 76% L=35m, 40% L=50m, 23% L=40m, 18% L=30m, 18% L=35m, 257% L=30m, 413% L=25m, 180% L=20m, 423%

$100

$50

$0 5 10 15 20 25 30 L/h 35 40 45 50 55

Figure 6-5. Cost comparison of stirrups and minimum reinforcing steel

As shown in Figure 6-5, cost of stirrups generally increases with span-to-depth ratios since more stirrups are needed for slender cases which have lower shear resistances. The increase in stirrup cost for cast-in-place box-girders slows down for ratios beyond 30, because 25M bars, which result in a more efficient stirrup layout for these cases, is used instead of 20M bars. Also, solid slabs require fewer stirrups compared to box-girders because they have longer effective widths and higher concrete strengths for slender cases which result in greater concrete shear resistances. The large number of prestressing tendons, which are needed for adequate flexural resistances, further helps in resisting shears. For instance, for the cases with a span length of 35m, the stirrups required for a solid slab with a span-to-depth ratio of 30 cost $15/m2 while they cost 3.2 times more for a box-girder with a ratio of 25 (i.e. $48/m2). The dashed lines on Figure 6-5 show the costs of minimum reinforcing steel required for stability of the steel bar cage and for crack control. The minimum requirement of reinforcing steel is proportional to the concrete volume, so it decreases as span-to-depth ratio increases. The total cost of reinforcing steel is the sum of the costs of stirrups and minimum reinforcing steel and the contributions of these two cost components vary over different span-to-depth ratios. Figure 6-6 illustrates the contribution of each component by comparing the percentage distributions and the actual costs for the longest span case for each bridge type. As shown in Figure 6-6 a, the impact of stirrups on the total reinforcing steel cost increases as the girder becomes more slender

72 regardless of bridge type. This trend is a result of the reduction in minimum reinforcement due to decreasing concrete volume and the rise in shear demand due to increasing girder slenderness. As shown in Figure 6-6 b, for both cast-in-place and precast box-girders, the total reinforcing steel cost decreases as span-to-depth ratio increases until 20. For these cases, decrease in minimum reinforcing steel is significant due to the large reduction in concrete volume (shown in Figures 4-8 a and 5-16 a) while the increase in stirrups is relatively small. As a result, the total reinforcing steel cost exhibits an overall decreasing trend. When span-to-depth ratio increases beyond 20, the reduction in reinforcing steel diminishes and becomes insignificant compared to the increase in stirrups, resulting in an overall increase in total reinforcing steel cost. For solid slabs, the total reinforcing steel cost does not increase significantly even for span-todepth ratios above 20. Although the stirrups cost increases by more than two times over the entire range of ratios, this increase is offset by the equally large decrease in minimum reinforcement cost. The decrease in minimum reinforcement is relatively large compare to the one for box-girders with the same slenderness ratios, because solid slabs experience constant volume reduction as span-todepth ratio increases whereas the volume diminishes at a decreasing rate for box-girders (discussed previously in Section 4.3.3.5 and illustrated in Figure 4-20 a).
a) Percentage distribution of reinforcing steel cost
100% 80% 60% 40% 20% 0% 10 15 20 25 30 35 30 35 40 45 15 20 25 Minimum reinforcing steel Stirrups

b)

Distribution of reinforcing steel cost

$400 $300 $200 $100 $0 10 15 20 25 30 35 30 35 40 45 L=75m CIP on falsework box-girder L=35m CIP on falsework solid slab 20 25 L=50m Precast span-byspan box-girder 15 Minimum reinforcing steel Stirrups

Figure 6-6. Cost distribution of stirrups and minimum reinforcing steel

73 The total cost of reinforcing steel for all analysis cases is shown in Figure 6-7. The maximum variations in total reinforcing steel cost over the analysis range of ratios are only 18% for cast-inplace on falsework box-girder, 5.8% for cast-in-place on falsework solid slab, and 6.2% for precast span-by-span box-girder.
$400 L=75m, 18% L=60m, 18% L=50m, 18% $300 L=35m, 17% L=50m, 4.3% L=40m, 4.0% L=30m, 6.2%

$350

$250 Cost of reinforcing steel per $200 deck area ($/m2) $150

L=35m, 5.4% L=30m, 5.8% L=25m, 2.5% L=20m, 2.7%

$100 CIP on falsework solid slab L=20m L=35m falsework box-girder CIP on L=30m span-by-span box-girder Precast 5 10 15 20 25 30 L/h 35 40 45 50 55

$50

$0

Figure 6-7. Total reinforcing steel cost comparison

6.1.2.4 Total Superstructure Cost Figure 6-8 describes the total superstructure material cost, which consists of costs of concrete (material only), prestressing tendon, and reinforcing steel. The contribution of these three components are summarized in Figure 6-8 a. For deep box-girders, the superstructure material cost is mainly governed by concrete and reinforcing steel costs which decrease as span-to-depth ratio increases due to the reduction in concrete volume. The effect of increasing prestressing tendon cost is relatively minimal and is overshadowed by the reduction in concrete and reinforcement costs. Therefore, the total material cost exhibits an overall decreasing trend for low span-to-depth ratios. For the slender box-girders, the costs of prestressing tendon and reinforcing steel rise as span-todepth ratio increases. This cost increase is greater than the concrete cost reduction since the decreasing in concrete volume slows down at higher span-to-depth ratios. As a result, the total material cost experiences a net increase at higher ratios. For slender solid slabs, increase in material cost is also attributed to the increase in concrete cost from the use of higher strength concrete.

74
a) Percentage distribution of superstructure material cost
100% 80% 60% 40% 20% 0% 10 15 20 25 30 35 30 35 40 45 L=75m CIP on falsework box-girder L=35m CIP on falsework solid slab 20 25 L=50m Precast span-byspan box-girder 15 Reinforcing steel Concrete (material only) Prestressing tendon

b)

Total superstructure material cost

$2,000

$1,500

Total material cost per deck area ($/m2)

$1,000 L=75m, 42% L=60m, 22% L=50m, 19% L=35m, 6.3% L=50m, 12% L=40m, 9.2% L=30m, 6.7% L=35m, 27% L=30m, 28% L=25m, 11% L=20m, 28% CIP on falsework solid slab L=20m L=35m CIP on falsework box-girder L=30m Precast span-by-span box-girder 5 10 15 20 25 30 L/h 35 40 45 50 55

$500

$0

Figure 6-8. Total superstructure material cost comparison

Figure 6-9 describes the total superstructure cost which includes the falsework and casting costs of concrete in addition to the total material cost. As stated in Section 6.1.1, the cost of concrete casting and falsework is a function of concrete volume and is not related to concrete strength, and thus it generally decreases as span-to-depth ratio increases. This decreasing casting and falsework cost amplifies the cost reduction for deep girders and dampens the cost increase for slender girders. The dampening effect is especially obvious for solid slabs in which the concrete volume reduction associated with rising span-to-depth ratio is significant.

75
a) Percentage distribution of superstructure cost
100% 80% 60% 40% 20% 0% 10 15 20 25 30 35 30 35 40 45 L=75m CIP on falsework box-girder L=35m CIP on falsework solid slab 20 25 L=50m Precast span-byspan box-girder 15 Prestressing tendon Reinforcing steel Concrete (material and placement)

b)

Total superstructure cost

$2,000 L=75m, 18%

L=60m, 20% L=50m, 17% $1,500 L=35m, 14% L=50m, 3.3% L=40m, 2.6% Superstructure cost per $1,000 deck area ($/m2) L=30m, 0.3%

L=35m, 12% L=30m, 12% L=25m, 9.6% L=20m, 3.6%

$500 CIP on falsework solid slab L=20m CIP on falsework box-girder L=35m $0 5 10 15 20 25 30 L/h 35 L=30m Precast span-by-span box-girder 40 45 50 55

Figure 6-9. Total superstructure cost comparison (including cost of concrete placement)

The optimal span-to-depth ratio in terms of superstructure costs is determined by balancing the concrete cost reduction with the prestressing cost increase as span-to-depth ratio rises. In terms of superstructure material costs only, the cost optimal ratios are 15, 30, and 15 for cast-in-place on falsework box-girder, solid slab and precast span-by-span box-girder respectively. These optimal ratios increase to 25, 40, and 20 if the total superstructure costs are considered. This improvement in slenderness ratio demonstrates the construction economy related to slender and lighter superstructures. Although these optimal ratios based on total superstructure costs are higher than the conventional ones, the changes in superstructure cost between these two sets of ratios are actually

76 minor (less than 11%) as shown in Table 6-4. Using the more cost-efficient ratios instead of conventional ratios yields only insignificant total superstructure cost savings compared to the changes in individual material costs. The table also shows the maximum cost variations within the entire range of slenderness ratios analyzed in parentheses. Even if the least cost-efficient ratio is used, the total superstructure cost increases by only 20%.
Table 6-4. Total superstructure cost variations Analysis range of ratios Conventional ratio Optimal ratio Cost component Concrete Prestressing tendon Reinforcing steel Total superstructure CIP on falsework box-girder 10 - 35 20 25 -5.1% (42%) +28% (284%) -1.3% (18%) -0.6% (20%) CIP on falsework solid slab 30 - 50 30 40 -19% (24%) +53% (60%) -5.3% (5.8%) -11% (12%) Precast span-by-span box-girder 15 - 30 17 20 -4.1% (8.8%) +26% (62%) -3.6% (6.2%) -1.8% (3.3%)

Furthermore, Figure 6-9 shows that instead of refining cross-sectional component dimensions, using a more efficient structural system results in greater cost savings. For instance, for a span length of 35m, a cast-in-place solid slab with a slenderness ratio of 30 costs 59% more than a castin-place box-girder with a ratio of 25. Therefore, the superstructure cost savings associated with optimizing the span-to-depth ratios are negligible compared to the savings from choosing the proper bridge type for a given span length and girder depth.

6.2

Overall Construction Costs

This section discusses the impact of optimizing span-to-depth ratio on the total construction cost

which includes the costs of superstructure, substructure, mobilization and accessories such as bearings. First, the percentage breakdown of these cost components with respect to the total construction cost is described in Section 6.2.1. Section 6.2.2 then describes the total construction costs for all analysis cases computed using this cost breakdown. The analysis results indicate that varying the span-to-depth ratio has minor influence on the total construction cost, because only the superstructure cost depends on span-to-depth ratio whereas the cost of substructure and accessories are related more to the span length. 6.2.1 Construction Cost Breakdown

The construction cost breakdown used in the analysis is adopted from Menns cost study on 19 concrete highway bridges constructed between 1958 and 1985 in Switzerland (Table 6-5). Mobilization comprises of all the work that needs to be completed prior to the onset of construction. Such tasks include preparing site facilities, procuring equipments and establishing access to the construction site. Some examples of accessories are bearings, expansion joints, and drainage

77 systems. These two factors are independent of span-to-depth ratio, but they are related to the bridge site conditions, number of spans, and bridge length. Likewise, the substructure cost depends more on geotechnical conditions and bridge height. Therefore, the costs for these three items are computed for the baseline cases only and are assumed to be constant for all other analysis cases with the same span length.
Table 6-5. Construction cost breakdown (Menn 1990) Item Mobilization Structure Substructure Foundations Piers and abutments Total substructure Superstructure Total superstructure Total structure Accessories Total construction cost Cost (% of total construction cost) 8.0%

18.0% 5.5% 23.5% 23.5% 54.5% 78.0% 78.0% 14.0% 100.0%

6.2.2

Total Construction Cost Comparison

The total construction costs for all analysis cases are illustrated in Figure 6-10 and a sample calculation is shown in Appendix B.5. The trend and the most cost-optimal ratios are the same as the ones for superstructure cost, because the total construction cost is simply the sum of superstructure cost and the costs of substructure, mobilization and accessories which are constant amongst analysis cases with the same span length. However, factoring in this constant cost further diminishes the economic incentive of optimizing span-to-depth ratios. In fact, the total construction cost premiums of using the optimal ratios instead of conventional ratios are only 0.4% for cast-inplace box-girder, 5.8% for cast-in-place solid slab, and 1.0% for precast span-by-span box-girder. Even within the entire feasible range of span-to-depth ratios, maximum variations in total construction cost are only 11%, 6.2% and 1.8% for the three bridge types respectively. These cost reductions related to optimizing span-to-depth ratios for a given span are insignificant compared to the cost reduction from choosing a suitable bridge type and span arrangement. For instance, for a span length of 35m, using a cast-in-place box-girder with slenderness ratio of 25 instead of a cast-in-place solid slab with a ratio of 30 reduces the total construction cost by 37%. Cost improvement can also be attained by using shorter span lengths. For instance, using a span length of 50m instead of 75m reduces the total construction cost by 20% in cast-in-place box-girders.

78
$4,000

$3,500

L=75m, 10% L=60m, 11% L=50m, 9.0%

L=35m, 6.1% L=30m, 6.2% L=50m, 1.8% L=40m, 1.4% L=30m, 0.1% L=25m, 5.0% L=20m, 1.9%

$3,000 Total construction cost per deck area ($/m2)

$2,500

L=35m, 7.7%

$2,000

$1,500

$1,000 CIP on falsework solid slab L=20m L=35m falsework box-girder CIP on L=30m Precast span-by-span box-girder 5 10 15 20 25 30 L/h 35 40 45 50 55

$500

$0

Figure 6-10. Total construction cost comparison

6.3

Other Cost Factors

In addition to construction costs, bridge designers are also concerned with other cost factors

such as operation, maintenance, rehabilitation, and demolition costs. These factors are excluded in this cost study, but changing span-to-depth ratio is expected to have minor effects on these costs. These factors depend more on the overall design concept (e.g. structural system, cross-section type, span arrangement) than on the dimensioning of structural components (Menn 1990).

79

6.4

Sensitivity Analysis
The purpose of the sensitivity study is to investigate how sensitive are the analysis results to

changes in material unit prices and in breakdown of construction costs. The material unit prices vary in different locations due to differences in labour costs and technology levels. The construction cost breakdown also changes depending on the site conditions. 6.4.1 Sensitivity with Respect to Changes in Material Unit Prices

The first aspect to investigate is the impact of changes in material unit prices. A new set of superstructure costs are calculated based on the following variations in material unit prices:
Table 6-6. Material unit price changes Material Concrete Prestressing tendon Reinforcing steel Unit 3 per m per kg per kg Original price $ 1,500 $ 8.50 $ 5.00 +50% $ 2,250 $ 12.75 $ 7.50 -50% $ 750 $ 4.25 $ 2.50

These changes in unit prices are drastic and are chosen for only illustrative purposes such that a clear trend would be observed. A more realistic maximum price change would be around 20%. The effects of changing unit prices on total construction costs are illustrated in the following graphs. The maximum percentage differences in cost for each span length over the specified range of span-to-depth ratios are also indicated on the graphs. The dashed lines represent the costs under original unit prices while the solid lines represent the costs when the unit prices vary by 50%. The impacts of altering unit prices are illustrated in Figures 6-11 and 6-12 for concrete, Figures 6-13 and 6-14 for prestressing tendons, and Figures 6-15 and 6-16 for reinforcing steel. The graphs show that modifying concrete unit price has the greatest influence on the total construction cost, because concrete cost constitutes a large portion of the superstructure cost. For example, for a cast-in-place on falsework box-girder with span length of 75m and span-to-depth ratio of 10, increasing the concrete unit cost by 50% raises the total construction cost by 35%. On the other hand, increasing the prestressing cost by 50% only increases the total construction cost by 4.6%.

80
$5,000 L=75m, 12% $4,000 L=60m, 13% L=50m, 11% Total construction cost per deck area ($/m2) L=35m, 9.1% $3,000 L=25m, 6.4% L=50m, 2.2% L=40m, 1.7% L=30m, 0.5% L=20m, 3.8% L=35m, 8.7% L=30m, 8.0%

$2,000

$1,000 CIP on falsework solid slab L=35m L=75m CIP on falsework box-girder L=50m Precast span-by-span box-girder 0 5 10 15 20 25 L/h 30 35 40 45 50 55

$0

Figure 6-11. Total construction cost comparison (+50% concrete unit price)
$5,000

$4,000

Total construction cost per deck area ($/m2)

$3,000

$2,000

L=75m, 12% L=60m, 7.8% L=50m, 6.6% L=35m, 5.4% L=50m, 3.1% L=40m, 2.8% L=30m, 1.5%

L=35m, 3.8% L=30m, 3.1% L=25m, 1.9% L=20m, 2.8%

$1,000 L=25m CIP on falsework solid slab L=75m CIP on falsework box-girder L=50m Precast span-by-span box-girder $0 0 5 10 15 20 25 L/h 30 35 40 45 50 55

Figure 6-12. Total construction cost comparison (-50% concrete unit price)

81
$5,000

$4,000 L=75m, 11% Total construction cost per deck area ($/m2) $3,000 L=60m, 8.8% L=50m, 7.4% L=35m, 6.0% $2,000 L=35m, 5.5% L=30m, 4.9% L=50m, 2.7% L=40m, 2.3% L=30m, 1.3% L=25m, 3.5% L=20m, 2.4%

$1,000 L=25m CIP on falsework solid slab L=75m CIP on falsework box-girder L=50m Precast span-by-span box-girder $0 0 5 10 15 20 25 L/h 30 35 40 45 50 55

Figure 6-13. Total construction cost comparison (+50% prestressing tendon unit price)
$5,000

$4,000

Total construction cost per deck area ($/m2)

$3,000

L=75m, 13% L=35m, 9.3% L=60m, 13% L=50m, 12% L=35m, 9.7% L=30m, 8.3% L=50m, 2.5% L=40m, 1.9% L=30m, 1.1% L=25m, 6.8% L=20m, 3.9%

$2,000

$1,000 L=35m CIP on falsework solid slab L=75m falsework box-girder CIP on L=50m span-by-span box-girder Precast $0 0 5 10 15 20 25 L/h 30 35 40 45 50 55

Figure 6-14. Total construction cost comparison (-50% prestressing tendon unit price)

82
$5,000

$4,000 L=75m, 10% Total construction cost per deck area ($/m2) L=60m, 11% $3,000 L=50m, 9.1% L=35m, 7.7% L=50m, 1.7% L=40m, 1.5% L=30m, 0.2% L=35m, 6.7% L=30m, 6.1% L=25m, 4.7% L=20m, 2.9%

$2,000

$1,000 L=35m CIP on falsework solid slab L=75m CIP on falsework box-girder Precast span-by-span box-girder L=50m $0 0 5 10 15 20 25 L/h 30 35 40 45 50 55

Figure 6-15. Total construction cost comparison (+50% reinforcing steel unit price)
$5,000

$4,000

Total construction cost per deck area ($/m2)

$3,000

L=35m, 7.4% L=75m, 10% L=60m, 11% L=50m, 9.0% L=30m, 6.9% L=25m, 5.3% L=50m, 1.8% L=40m, 1.5% L=30m, 0.5% L=20m, 3.2%

$2,000

L=35m, 7.6%

$1,000 L=35m CIP on falsework solid slab L=75m CIP on falsework box-girder L=50m Precast span-by-span box-girder $0 0 5 10 15 20 25 L/h 30 35 40 45 50 55

Figure 6-16. Total construction cost comparison (-50% reinforcing steel unit price)

As shown in the graphs, the optimal span-to-depth ratios remain at 25, 40, and 20 for cast-inplace box-girder, cast-in-place solid slab, and precast box-girder regardless of unit price changes. Table 6-7 summarizes the maximum cost improvements of using these optimal span-to-depth ratios

83 instead of conventional ratios for the specified variations in unit price. Despite such drastic unit price changes, the cost improvements are only 0.6% 1.1% for cast-in-place box-girder, 5.8% 2.3% for cast-in-place solid-slabs, and 1.0% 0.5% for precast span-by-span box-girder. Table 6-7 also lists the maximum percentage changes in cost over the entire range of span-todepth ratio in parentheses. These values represent the maximum cost increase when the least costefficient span-to-depth ratios are used instead of the optimal ratios. For instance, for a given span length, using a span-to-depth ratio of 10 instead of the optimal ratio results in 11% increase in total construction cost for cast-in-place box-girder. This increase in cost changes to 12% if the concrete unit price is increased by 50%. For cases with altered unit prices, the changes in cost with respect to span-to-depth ratios deviate less than 3% from the ones for cases with original unit prices. Since the optimal ratios remain the same and the changes in cost improvements are negligible (less than 3%), results from the cost study are insensitive to modifications in unit prices.
Table 6-7. Summary of material unit price sensitivity analysis Unit cost change CIP on falsework box-girder CIP on falsework solid slab Precast span-by-span box-girder Range of L/h ratio 10 - 35 30 - 50 15 - 30 Optimal ratio 25 40 20 Conventional ratio 20 30 17 Original unit prices 0.4% (11%) 5.8% (6.2%) 1.0% (1.8%) +50% concrete 0.9% (12%) 7.9% (8.7%) 1.3% (2.2%) -50% concrete 1.5% (12%) 3.6% (3.8%) 0.5% (3.1%) +50% prestressing tendon 0.9% (11%) 5.2% (5.5%) 0.8% (2.7%) -50% prestressing tendon 1.0% (13%) 8.1% (9.3%) 1.5% (2.5%) +50% reinforcing steel 0.3% (11%) 6.3% (6.7%) 1.0% (1.7%) -50% reinforcing steel 0.4% (11%) 6.9% (7.4%) 0.9% (1.8%)

6.4.2

Sensitivity with Respect to Changes in Construction Cost Breakdown

The purpose of this section of the sensitivity study is to investigate the impact of altering the construction cost breakdown. Hitherto, the cost results are computed according to Menns breakdown of construction costs as described in Section 6.2.1. This breakdown changes under situations such as complications in geotechnical and hydraulic conditions, changes in bridge height which affect pier costs, and variations in mobilization cost for different construction sites. Previous cost calculations assume that the superstructure costs constitute 54.5% of the total construction cost for the baseline cases. Costs for the remaining items (i.e. substructure, mobilization and accessories) are computed based on the superstructure costs of the baseline cases and are set to be constant for the other cases with the same span lengths. Therefore, altering only the proportion of superstructure cost is sufficient in demonstrating the influence of different construction cost breakdowns. The total construction costs obtained from varying the superstructure component from 20% to 80% for cast-in-place on falsework box-girder with a span length of 50m are plotted in Figure 6-17.

84 As shown in the graph, changing the superstructure percentage basically shifts the curves vertically while the cost difference between cases with different span-to-depth ratios remains the same for each curve. As a result, the optimal ratio remains at 25 regardless of changes in the cost breakdown. The graph also shows the percentage cost improvements from conventional ratio to optimal ratio on the left (maximum cost variations over the entire range of span-to-depth ratios are shown in parentheses). Obviously, the cost improvement increases as superstructure percentage increases because the effect of reducing superstructure cost on the total construction cost is greater if superstructure cost forms a large portion of the total cost. Therefore, using optimal span-to-depth ratios poses more economic incentive if the superstructure percentage is higher. However, in spite of the changes in the cost breakdown, the saving from using the optimal ratio instead of conventional ratio is still less than 0.5% while the maximum cost variation within the entire range of ratio is less than 13%. The same pattern occurs in the other bridge types and span lengths as shown in Appendix C.4. Therefore, the values for optimal span-to-depth ratio as well as the cost variations between cases with optimal and conventional ratios are insensitive to changes in cost breakdown.
$8,000 Cost variation 0.1% (3.4%) $6,000 Total construction cost per deck area ($/m2) $4,000 0.3% (6.9%) 0.3% (8.6%) 0.4% (9.0%) 0.4% (10%) 0.5% (12%) 0.5% (13%) 40% 50% 54.5% (Menn) 60% 70% 80% Superstructure as % of total construction cost 20%

0.2% (5.2%)

30%

$2,000

$0 0 5 10 15 L/h 20 25 30 35

Figure 6-17. Total construction costs under changes in construction cost breakdown

85

6.5

Concluding Remarks
The results of this cost study are summarized in Table 6-8 which shows the percentage changes

in cost when optimal ratios instead of the conventional ones are used. Cost variations over the analysis range of ratios are also included within the parentheses. For all three bridge types, the costeffective ratios are higher than the conventional ratios, but the actual cost saving associated with using optimal ratios is less than 5.8%. The cost study also determines that the maximum cost variability is less than 11%, meaning that using the least cost-efficient ratio within the analysis range would only incur a relatively small additional cost. The range of cost-optimal ratios is thus expanded from the typical ranges defined in Chapter 2 to the analysis ranges of ratios indicated in Table 6-8.
Table 6-8. Summary of cost study Analysis range of ratios Typical range of ratios Conventional ratio Cost-optimal ratio Cost component Concrete Prestressing tendon Reinforcing steel Total superstructure Total construction cost CIP on falsework box-girder 10 - 35 17.7 - 22.6 20 25 -5.1% (42%) +28% (284%) -1.3% (18%) -0.6% (20%) -0.4% (11%) CIP on falsework solid slab 30 - 50 22 - 39 30 40 -19% (24%) +53% (60%) -5.3% (5.8%) -11% (12%) -5.8% (6.2%) Precast span-by-span box-girder 15 - 30 15.7 - 18.8 17 20 -4.1% (8.8%) +26% (62%) -3.6% (6.2%) -1.8% (3.3%) -1.0% (1.8%)

The results in Table 6-8 are found to be insensitive to changes in material unit price and construction cost breakdown. The optimal ratios, however, are determined based on parameters defined specifically for this study such as cross-section dimensions and span arrangements. If these parameters are altered, the optimums would likely be different. For example, the optimal ratios might increase if thicker webs are used for box-girders, because the prestressing requirement is reduced for slender cases due to the more efficient tendon layout as discussed in Section 4.2.2.5. Therefore, the actual values of optimal ratios determined in this study are expected to change in a real situation. Yet, the general finding regarding the variability in cost is still valid for a broad range of situations, because the study demonstrates that cost savings from the use of optimal ratios are minor compared to other construction cost components which are independent of span-to-depth ratio (e.g. costs of mobilization, substructure, and accessories). So, even if optimal ratio changes, the variability in cost is expected to remain relatively insignificant over the analysis range of ratios.

AESTHETICS COMPARISONS
The selection of slenderness ratio has significant impact on the overall appearance of girder

bridges. In particular, the ratio is an especially important visual criterion for highway overpasses, which are mostly observed from the highways passing beneath them, because the superstructure is the prime object of scrutiny from this view point (Elliott 1991). The previous cost study demonstrates that total construction cost varies by less than 11% over the range of span-to-depth ratios investigated. This finding provides the designer with more freedom for aesthetic expressions since he can choose from a wide range of slenderness ratios without much economic restrictions. This chapter examines these aesthetic opportunities by comparing the visual impacts of different span-to-depth ratios. This chapter also discusses some visually superior slenderness ratios by exploring existing bridges, which have been considered by the general public as aesthetically pleasing, and by examining some past studies on bridge aesthetics.

7.1

Visual Impact of Span-to-Depth Ratio

This section demonstrates the visual impact of altering slenderness ratios by comparing

drawings of bridges with various ratios as shown in Figures 7-1 to 7-3. These figures compare bridges with: a) conventional ratio obtained from Chapter 2; b) optimal ratio in terms of cost efficiency determined in Chapter 6; c) least cost-efficient ratio for each of the three bridge types considered. The total construction cost and its percentage variation from the cost of the optimal ratio are indicated in the parentheses. Each set of drawing includes a 3-D rendering from the vantage point of a driver, who is passing under the overpass at 150m away from the bridge, as well as 2-D elevation and cross-sectional views. The 2-D drawings are included to illustrate that although changing the slenderness ratio results in clear visual differences on paper, such differences might not be as apparent from the drivers perspective in a real situation. The drawings do not include barriers; the effect of barriers is discussed in Section 7.1.2. Also, all the 3-D drawings are obtained under the same lighting condition.

86

87

a)

Conventional L/h = 20 ($2450/m2, 0.4%)

b)

Optimal L/h = 25 ($2460/m2)

c)

Maximum cost L/h = 10 ($2670/m2, 9.0%)

Figure 7-1. Cast-in-place on falsework box-girder with L=50m

88

a) & c) Conventional and maximum cost L/h = 30 ($3000/m2, 5.8%)

b)

Optimal L/h = 40 ($2830/m2)

Figure 7-2. Cast-in-place on falsework solid slab with L=30m

89

a)

Conventional L/h = 17 ($2370/m2, 1.0%)

b)

Optimal L/h = 20 ($2350/m2)

c)

Maximum cost L/h = 25 ($2390/m2, 1.8%)

Figure 7-3. Precast segmental span-by-span box-girder with L=50m

As shown in Figures 7-1 and 7-3, using cost-effective span-to-depth ratios noticeably improves visual slenderness compared to the conventional ratios for cast-in-place on falsework and precast span-by-span box-girders. The visual difference is even more apparent when the optimal case is compared to the case with maximum construction cost. This finding indicates that varying the spanto-depth ratio can have a significant visual impact without substantial cost premiums (less than 11% variation in cost for all the analysis cases considered).

90 However, the visual impact of changing span-to-depth ratio is less obvious for the solid slab case, because visual difference diminishes as slenderness ratio increases beyond 25 as shown in Figure 7-4.
a) L/h=10 b) L/h=15

c)

L/h=20

d)

L/h=25

e)

L/h=30

f)

L/h=35

Figure 7-4. Visual effects of increasing span-to-depth ratios from 10 to 35

91 As stated previously, increasing the span-to-depth ratio by reducing the girder depth can enhance visual slenderness without incurring significant additional costs. Increasing the ratio by extending the span length has similar visual effect as shown in Figure 7-5. Yet, the cost premium associated with increasing span length is much more severe. For instance, for the cast-in-place boxgirder case depicted in the figure, increasing the span length from 50m to 75m adds 23% to the total construction cost.
a) L/h=20, L=50m ($2450/m2) b) L/h=30, L=75m ($3020/m2, 23%)

Figure 7-5. Effect of increasing span length (box-girder with h=2.5m)

92 7.1.1 Effects of Viewing Points

The visual impact of altering span-to-depth ratio becomes more or less noticeable depending on the location of the observer. First, as the observer approaches the bridge, the effect of changing slenderness ratio is more obvious. Figures 7-6 to 7-8 demonstrate the influence of viewing distance by comparing box-girders with span length of 50m and slenderness ratios of 10 and 20 when viewed from distances of 300m, 150m, and 75m.
a) L/h=10 b) L/h=20

Figure 7-6. Viewed from 300m a) L/h=10 b) L/h=20

Figure 7-7. Viewed from 150m a) L/h=10 b) L/h=20

Figure 7-8. Viewed from 75m

93 In addition to viewing distance, the viewing angle also influences the impact of span-to-depth ratio. Figure 7-9 shows that the pier width-to-height ratio has greater visual impact than the longitudinal span-to-depth ratio when the bridge is viewed from beneath along the length of the bridge. This viewing angle is of particular importance if pedestrians can walk below the bridge.
a) L/h=10 with wide wall piers b) L/h=20 with wide wall piers

c)

L/h=10 with narrow piers

d)

L/h=20 with narrow piers

Figure 7-9. Effects of pier width-to-height ratio and span-to-depth ratio

94 As the viewing angle becomes less oblique, the impact of pier dimensions diminishes while the span-to-depth ratio has growing influence on the perceived slenderness of the structure Figure 7-10.
a) L/h=10 b) L/h=20

Figure 7-10. Effect of span-to-depth ratio as viewing angle becomes less oblique

7.1.2

Other Factors that Affect Visual Slenderness

In addition to span-to-depth ratio, other factors also affect the perceived thickness of the superstructure. These factors include the bridge height, pier configuration, length of deck slab cantilevers in a box-girder, and railing type. In this section, each factor is investigated using a boxgirder model with a span length of 50m and slenderness ratio of 20.

95 First, Figure 7-11 compares a tall bridge with a low one. Although the span-to-depth ratios are the same for both bridges, the tall one appears to be more slender due to the larger opening under the bridge. This large opening contrasts with the slender superstructure, thus reducing the perceived girder depth. The low bridge, on the other hand, appears heavy because the girder depth is similar to the bridge height. In fact, to achieve sufficient slenderness and transparency, Menn (1990) suggested that the bridge height needs to be at least four times greater than the girder depth. This height suggestion ensures that there is a large contrast between the girder depth and the depth of the opening under the bridge, thus lowering the perceived thickness of the superstructure.
a) Bridge height/girder depth = 5 b) Bridge height/girder depth = 1.75

Figure 7-11. Effect of bridge height on perceived superstructure slenderness

The pier configuration also affects the visual impact of span-to-depth ratio (Figure 7-12). More pronounced piers clearly separate the individual span lengths. As a result, the perceived superstructure thickness still depends on the span-to-depth ratio. Conversely, narrow piers that are tucked in underneath the superstructure draw less attention and thus accentuate the continuity of the entire girder. The perceived slenderness is then related to the ratio between the visually uninterrupted length of the superstructure and the girder depth.
a) Pronounced piers b) Less obtrusive piers

Figure 7-12. Effect of pier configuration on perceived superstructure slenderness

96 Thirdly, the perceived slenderness of the superstructure is influenced by the length of deck slab cantilever in box-girders. Figure 7-13 compares a bridge with short cantilevers to a bridge with longer cantilevers. The latter bridge appears to be more slender because the long cantilevers cast shadows onto the girder webs and these shadows conceal a portion of the girder depth. The continuous shadow line along the girder also emphasizes the overall superstructure length. Hence, slenderness depends more on the bridge length than the span length in this case.
a) Short deck cantilevers b) Long deck cantilevers

Figure 7-13. Effect of deck cantilever length on perceived superstructure slenderness

Lastly, the railing type contributes to the visual slenderness of the superstructure. Figure 7-14 shows two bridges with open-type metal railings and concrete barrier walls. Any railing increases the perceived depth of the girder. However, the first case appears more slender because metal railings are more transparent compared to concrete barriers. Concrete barriers add an extra depth of solid concrete which represents a major visual bulk. These concrete barriers can be as tall as 1.37m for highway bridges according to CHBDC (CSA 2006) and this height increases with increasing traffic volume and speed (Dorton 1991).
a) Open-type metal railings b) Concrete barrier walls

Figure 7-14. Effect of railing type on perceived superstructure slenderness

97

7.2

Evolution of the Visually Optimal Span-to-Depth Ratio

As demonstrated previously, varying the span-to-depth ratio clearly has an impact on aesthetics.

Yet, visually optimal span-to-depth ratios cannot be easily defined because aesthetics is not a quantifiable attribute. The ratio that yields the best-looking bridge changes over time and depends on the background of the observer. This section explores the visually optimal slenderness ratio by examining the works and design philosophies of various prominent bridge designers. First, the aesthetically optimal slenderness ratio has changed throughout the history of concrete bridges. This section traces the development of slenderness in concrete arch and girder bridges. Arch bridges are considered, because early concrete bridges were mostly arches and the concept of arch slenderness, which is generally associated with the arch thickness and deck depth with respect to span length, is analogous to the influence of span-to-depth ratios on the slenderness of girder bridges. Most importantly, although the representation of slenderness is different for the two bridge types, both arch and girder bridges demonstrate evident improvement in slenderness due to economic or aesthetic reasons. In the 19th century, concrete bridges were generally deep and heavy because concrete was regarded as artificial masonry and arch was the primary structural form for concrete bridges at that time. Also, the society preferred the massiveness, rigidity, and embellishments associated with traditional masonry arches. This trend can be seen in the Glenfinnan Railway Viaduct 1901 which is one of the first major concrete bridges (Figure 7-14).

Figure 7-14. Glenfinnan Viaduct, 1901 (Cortright 1997)

Improvement in the slenderness of arches was evident by the late 19th century due to the development of reinforced concrete which allowed arches to be thinner and flatter. Robert Maillart utilized the new material and created the Stauffacher Bridge 1899 which is a slender three-hinged arch concealed with stone-cladding and ornaments such that it resembles traditional masonry arches (Figure 7-15 a). In his subsequent bridges, Maillart abandoned the traditional architectural forms and began his pursuit for slenderness and simplicity (Figure 7-15). First, he eliminated the stone-

98 cladding in the Zuoz Bridge 1901 to reveal the slender arch. He further enhanced slenderness by removing the spandrel walls in the Tavanasa Bridge 1906.
a) Stauffacher Bridge, 1899 (Billington 1997) b) Zuoz Bridge, 1901 (Billington 1990)

c)

Tavanasa Bridge, 1906 (Billington 1990)

d)

Salginatobel Bridge, 1930 (Billington 1990)

Figure 7-15. Slender bridges by Maillart

One of the major breakthroughs in bridge aesthetics is Maillarts Salginatobel Bridge 1930 which accentuates the simple three-hinged arch form devoid of any structurally unnecessary components (Figure 7-15 d). The bridges exceptional slenderness contrasts with the traditional masonry-like structures that were popular at Maillarts time. The design was chosen mostly for its economic efficiency instead of its aesthetic value. In fact, its aesthetic merit was not widely recognized outside of Switzerland until decades after the bridge was constructed. The Salginatobel Bridge is now praised as a masterpiece of structural art by scholars like Billington (1990) and Bill (1955). In the 1930s, the development of reinforced concrete also allowed for the construction of concrete girder bridges which could reach spans of over 70m. However, long-span reinforced concrete girders commonly had problems with deformations and cracking. To minimize these problems, the girders needed to be very deep, with span-to-depth ratios of less than 10 (Menn 1990). Constant-depth girders with such ratios would be heavy, so long-span reinforced concrete bridges were usually haunched. Two examples of these variable-depth girders are the Villeneuve-St. Georges Bridge 1939 and the Waterloo Bridge 1939 which are both haunched box-girders (Figure 7-16). Their span-to-depth ratios at the supports are 9.7 and 10 respectively while the midspan ratios are 31 and 32 (Menn 1990).

99

Figure 7-16. Waterloo Bridge over the Thames, 1939 (Darger 2002) Figure 7-17. Changis-sur-Marne Bridge, 1948 (Mossot 2007)

Significant improvement in slenderness for girder bridges only began with the introduction of prestressing technology to bridge construction in the 1940s. Prestressing enhances structural behaviour by reducing tensile stresses in the concrete and increasing the load-carrying capacity with the use of high-strength tendons. As a result, longer and more slender girder bridges were feasible. Some notable examples of slender prestressed bridges include Eugene Freyssinets series of post-tensioned bridges along the Marne River 1948 (Figure 7-17). These bridges achieved visual slenderness with the use of prestressing technology, which allowed for longer spans and thinner decks, and had a midspan span-to-depth ratio of up to 40. This is a large improvement in slenderness compared to the previous examples of reinforced concrete girders which had a midspan ratio close to 30. This improvement in slenderness through post-tensioning technology was mainly driven by material economy, structural efficiency and construction speed. The aesthetic value associated with slenderness was not fully appreciated by society until the 1960s which marked the beginning of an era that praised minimalism and simplicity. The publics perception on bridge slenderness was demonstrated in a bridge aesthetics survey conducted in 1969 by A.G.D. Crouch under the supervision of Colin OConnor at the University of Queensland (Crouch 1974, O'Connor 1991). Crouch was an Engineer with the Canberra Department of Works while OConnor was a Professor of Civil Engineering and has published a number of books and papers regarding bridge design. The bridge aesthetics survey investigated the relative visual merits of various substructure and superstructure proportion parameters (e.g. span-to-pier thickness ratio, span-to-column height ratio, etc.) through the use of simple sketches of bridges with varying proportions. The survey sample included 170 civil engineers, architects, and people with
Figure 7-18. Sketches to evaluate aesthetic impact of span-to-depth ratios (O'Connor 1991)

100 no education in structural design (a control group). The civil and architecture groups consisted of students, university staffs as well as people in practice. One particular superstructure proportion parameter that was investigated was the span-to-depth ratio (Figure 7-18). The response corresponded to a preferred ratio of 34.3 for engineers, 20.5 for architects and 24.4 for the control group. The difference among the three groups demonstrated that the visually optimal ratio depends on the background of the observer. Engineers preferred a higher ratio because they valued material efficiency. The other groups, in contrast, preferred a deeper girder because they saw depth as a sign of strength and visual elegance as a matter of good proportions. The sketches used in this survey might be too simple to confidently yield specific values for visually-optimal span-to-depth ratios. Nonetheless, the survey indicated that people in general preferred a more slender structure with a minimum span-to-depth ratio of 20 (i.e. the top two sketches in Figure 7-18). Since there was greater public resonance for visual slenderness, higher span-to-depth ratios gained unprecedented popularity in bridge designs in the 60s and 70s. Bridge designers, therefore, consciously pursued slenderness based on its aesthetic merits in addition to its material economy. In particular, two renowned bridge engineers designed a number of girder bridges with exceptional slenderness at this time: Leonhardt and Menn. Leonhardt believed that the slender bridge looks better than the clumsy one. A slender look is therefore a design feature well worth striving for (Leonhardt 1982). He also claimed that heavy bridges appear depressing whereas lighter bridges are more elegant. His design philosophy was demonstrated in the Neckar Viaduct 1977 and the Kocher Valley Viaduct 1979 (Figures 7-19 and 7-20). The former haunched girder has a slenderness ratio of 56 at midspan and 25 at the supports while the latter constant-depth girder has a ratio of 21.2. Both the haunched and constant-depth girders demonstrate significant improvement in slenderness compared to bridges from the 30s which used span-to-depth ratios of 30 at midspan and 10 at the supports for haunched girders and a ratio of 10 for constant-depth girders.

Figure 7-19. Neckar Valley Viaduct, 1977 (Leonhardt 1982)

Figure 7-20. Kocher Valley Viaduct, 1979 (Leonhardt 1982)

101 Menn, on the other hand, believed that visual elegance was related to the efficient use of material which could be demonstrated by slenderness and lightness of a structure (Menn 1990). Two of his designs that incorporated this concept of visual elegance are the Pregorda Bridge 1974 and the Felsenau Bridge 1974 (Figures 7-21 and 7-22). Like Leonhardts bridges, these two bridges are considerably more slender than the girder bridges constructed in the previous decades. The Pregorda Bridge is a constant-depth girder with a slenderness ratio of 22.2 while the main spans of the Felsenau Bridge are haunched and has ratios of 48 at midspan and 18 at the supports.

Figure 7-21. Pregorda Bridge, 1974 (Menn)

Figure 7-22. Felsenau Bridge, 1974 (Menn)

Other contemporary bridge experts are also strong advocates of bridge slenderness and they have expressed the importance of slenderness in bridge aesthetics in a number of publications. One of these bridge professionals is Edward Wasserman who is the Civil Engineering Director of the Structures Division for the Tennessee Department of Transportation and has been involved in the design of over 2200 bridges. He claimed that a span-to-depth ratio between 25 and 30 would produce a well-proportioned superstructure that appears to float gracefully whereas a lower ratio would result in a bridge that appears to loom heavily upon the landscape (Wasserman 1991). Likewise, Arthur Elliott, Bridge Engineer responsible for all bridge planning and design with the California Department of Transportation from 1953 to 1973, stated that a blocky, heavy, and poorly proportioned bridge is simply not beautiful (Elliott 1991). Frederick Gottemoeller, who has produced a number of publications regarding bridge aesthetics and has developed the aesthetic design guidelines for Maryland and Ohio, also recognizes the visual benefits of slenderness. He claimed that better-looking bridges are characterized by their simplicity, thinness, and continuity (Gottemoeller 2004). As shown in this historical study, the public perception of bridge aesthetics changes over time. In the early 1900s, people preferred heavier bridges whereas slender bridges were appreciated in the 60s and 70s. Although the visually optimal span-to-depth ratio cannot be easily determined, contemporary bridge designers generally favour a higher ratio and regard slenderness as a key element in a good-looking bridge.

102

7.3

Concluding Remarks
This chapter demonstrates that span-to-depth ratio has direct impact on perceived superstructure

slenderness by comparing 3-D drawings of bridges with varying ratios. Bridges with the most costeffective ratio are visually more slender than those with conventional ratios. The visual difference between the cost-effective ratio and conventional ratio is particularly noticeable for the cast-in-place on falsework and precast span-by-span box-girders. For cast-in-place solid slab, the difference is less obvious because the conventional ratio is already high (i.e. 30) and increasing the ratio beyond 25 is found to have negligible visual impact. This chapter further shows that the effect of varying span-to-depth ratio reduces as the observer moves away from the bridge or as the viewing angle becomes more oblique. Factors, such as low and protruding piers, short deck cantilever lengths, and thick concrete railing, also reduce the perceived slenderness. Lastly, a historical study indicates that the visually optimal slenderness ratio evolves over time and contemporary bridge engineers appreciate the aesthetic merit of superstructure slenderness.

CONCLUSIONS
Girder-type bridges have commonly been designed using conventional span-to-depth ratios

which have not changed significantly despite recent development in material strengths and construction technologies. This study determines the optimum slenderness ratios for three types of girder bridges constructed with high-strength concrete: cast-in-place on falsework box-girder and solid slab, and precast segmental span-by-span box-girder. The ratios are optimized based on material consumption and total construction cost criteria. Aesthetic comparisons are also performed to determine the visual impact of these optimum ratios. The primary results of this thesis are summarized as follows.

8.1

Conventional Span-to-Depth Ratios

A study of 86 constant-depth girder bridges reveals that the typical ranges of span-to-depth

ratios are 17.7 to 22.6 for cast-in-place box-girder, 19 to 35 for cast-in-place voided slab, 22 to 39 for cast-in-place solid slab, and 15.7 to 18.8 for precast segmental box-girder. The study demonstrates that the ratios for cast-in-place and precast segmental box-girders have not varied significantly from 1958 to 2007. The study also indicates that cast-in-place slabs constructed after 1975 are mostly voided slabs with slenderness ratios below 25 due to the more stringent code requirements in recent years.

8.2

Maximum Span-to-Depth Ratios

The maximum span-to-depth ratio, which satisfies safety, serviceability, and constructability

requirements, varies with bridge type and span length. For cast-in-place box-girder with spans of 35m, 50m, and 60m, the maximum ratios are 25, 30, and 35 respectively. These values are restricted by the interior box cavity height requirement which is necessary to provide sufficient space for workers. The maximum ratio is also 35 for the case with a span of 75m; it is limited by the number of tendons that can fit inside the webs of the box section. The maximum ratios for solid slab are 40 for spans of 20m and 25m, and 45 for spans of 30m and 35m. These ratios are governed by the maximum reinforcement criterion which ensures adequate ductile behaviour. This limitation is especially critical for solid slabs, because slabs have heavier dead loads compared to box-girders with the same ratio, so more reinforcements are needed. For precast segmental box-girder, the maximum ratios are 20 for a span of 30m, and 25 for spans of 40m and 50m. Like cast-in-place box-girders, these ratios are limited by the minimum height requirement for the interior box cavity. Even though the governing factor of maximum slenderness ratio is the same for both bridge types, the ratios for precast box-girders are lower, because external tendons are used, further reducing the height of the cavity within box-girders. 103

104

8.3

Material Consumption Comparisons

As slenderness increases, the prestressing and concrete strength demands increase while

concrete volume decreases. For both cast-in-place and precast segmental box-girders, the decrease in volume is small compared to the increase in prestressing since only a small amount of concrete at the webs is eliminated while the moment resistance is significantly lowered as the ratio increases. Reduction in concrete volume further diminishes for higher ratios because the bottom slab thickness needs to be increased to accommodate the larger compressive force in a slender girder. This increase in slab thickness counteracts the decrease in web volume, thus the reduction in concrete volume becomes less for higher ratios. Moreover, a concrete compressive strength of 50 MPa is sufficient to satisfy design requirements in every analysis case for both of these bridge types. Decrease in concrete volume in a solid slab is more significant since a large strip of concrete as wide as the soffit is removed as the ratio increases. Therefore, this volume reduction is proportional to the decrease in girder depth. On the other hand, prestressing demand increases with slenderness. This increase in prestressing is less than the one for cast-in-place box-girders, because solid slabs can accommodate many tendons at the same elevation while tendons in box-girders need to be placed in multiple layers within the webs which reduces the prestressing efficiency. Also, higher strength concretes (i.e. fc = 50 to 80 MPa) are used for slender slabs to satisfy ductility requirement by lowering the prestressing demand. As a result, prestressing consumption in solid slabs increases with span-to-depth ratio at a slower rate relative to box-girders. Material consumptions are compared on the basis of material costs. Considering only the pure material costs without the cost of formworks, falsework or precasting operations, the most efficient ratios are15 for both cast-in-place and precast box-girders and 30 for solid slabs. If the costs of concrete fabrication and placement are included, the most cost-efficient ratios for the three bridge types increase to 25, 40, and 20. This increase in cost-optimal ratios indicates that the construction economy related to a slenderer and lighter structure is a crucial aspect in the optimization of spanto-depth ratio. More importantly, over the entire range of ratios investigated (i.e. 10 to 35 for castin-place box-girder, 30 to 45 for cast-in-place solid slab, and 15 to 25 for precast segmental boxgirder), the maximum variations in pure material cost are 42%, 28%, and 12% while the maximum variations in total superstructure cost are only 20%, 12%, and 3.3% for each bridge type.

8.4

Total Construction Cost Comparisons

The total construction costs are computed assuming the superstructure accounts for 54.5% of

total construction cost. With the additional costs of mobilization, substructure, and accessories, the most cost-efficient ratios remain the same as the ones based on superstructure costs only. However,

105 the economy of using these optimal ratios diminishes when the total construction costs are considered; the maximum savings within the analysis range of ratios are reduced to less than 11%, 6.2%, and 1.8% for cast-in-place box-girder, cast-in-place solid slab, and precast segmental boxgirder respectively. This finding indicates that optimizing one particular structural component (i.e. superstructure span-to-depth ratio) does not result in significant economy. Greater cost savings emerge from the selection of an appropriate bridge type. For instance, within the same range of span lengths and span-to-depth ratios, a cast-in-place on falsework box-girder is more economical than a solid slab. Moreover, these findings are not sensitive to changes in material unit prices and in total construction cost breakdown; the cost-optimal ratios remain the same regardless of these changes. The variability in total construction cost over the analysis range of ratios is less than 13% when the material unit price is altered by 50% or when the superstructure cost contribution in the total cost breakdown rises from 54.5% to 80%.

8.5

Aesthetic Comparisons
Since the total construction cost does not vary significantly over the entire range of span-to-

depth ratios investigated, the designer has more freedom to select the slenderness ratio without much economic constraints. Varying the ratio changes the superstructure slenderness which is generally the most important visual component of a girder-type highway overpass. A historical study indicates that although the public perception on slenderness has evolved over time, contemporary bridge engineers, especially the ones from the 1960s and 1970s, recognize and appreciate the aesthetic merit of slenderness associated with high span-to-depth ratios. The visual impact of span-to-depth ratios is examined by comparing 3-D renderings of bridges with different ratios. This comparison determines that using cost-optimal ratios instead of conventional ones would result in considerable enhancement in the superstructure slenderness for cast-in-place and precast box-girders. Yet, the aesthetic impact of using the optimal ratio in solid slabs is negligible because increasing the ratio beyond 25 is found to have no apparent visual difference. Furthermore, the visual effects of varying span-to-depth ratio reduce as the observer moves away from the bridge or as the viewing angle becomes more oblique. In addition to lowering the slenderness ratio, other factors that reduce perceived slenderness include a low bridge height, protruding piers, short deck cantilever lengths, as well as solid concrete railing.

8.6

Optimal Span-to-Depth Ratios

Based on construction economy and aesthetics considerations, the optimal span-to-depth ratios

are established to be 25, 40, and 20 for cast-in-place on falsework box-girder, cast-in-place on falsework solid slab, and precast segmental span-by-span box-girder. However, these optimums are

106 expected to change in a real situation because they are determined based on specific parameters defined for this study such as cross-section dimensions and span arrangements. More importantly, the study demonstrates that, within the analysis range of ratios, the total construction cost is relatively insensitive to changes in the ratio. This finding is valid even if the optimum ratio changes, because the study shows that cost savings from optimizing span-to-depth ratio are generally minor compared to other construction cost components. The results of this study indicate that, compared to conventional ratios defined for normal-strength concrete bridges, a greater range of values can be used without significant cost premiums for high-strength concrete bridges (i.e. fc = 50 to 80 MPa): 10 to 35 for cast-in-place on falsework box-girder; 30 to 45 for cast-in-place on falsework solid slab; 15 to 25 for precast segmental span-by-span box-girder.

REFERENCE
AASHTO-PCI-ASBI. (1997)."Segmental Box Girder Standards for Span-by-Span and Balanced Cantilever Construction." American Segmental Bridge Institute, <www.asbi-assoc.org> (Feb. 1, 2009). AASHTO. (1994). AASHTO LRFD Bridge Design Specifications, American Association of State Highway and Transportation Officials, Washington, D.C. ACI-ASCE Committee 343. (1988). "Analysis and design of reinforced concrete bridge structures." American Concrete Institute, Detroit. Belli, A. (2003). "Interstate 895 Bridge, James River, Richmond, Virginia, USA." Structural Engineering International:Journal of the International Association for Bridge and Structural Engineering (IABSE), 13(3), 177-179. Bennett, M. V., and Taylor, A. J. (2002). "Woronora River Bridge, Sydney." Structural Engineering International: Journal of the International Association for Bridge and Structural Engineering (IABSE), 12(1), 28-31. Bill, M. (1955). Robert Maillart, Girsberger Zrich. Billington, D. (1990). Robert Maillart and the Art of Reinforced Concrete, The MIT Press, Cambridge. Billington, D. (1997). Robert Maillart: Builder, Designer, and Artist, Cambridge University Press, Cambridge. Cao, S., Wu, W., Tu, J., and Wu, E. (2006). "Design & Construction of Ngong Shuen Chau Viaduct." International Conference on Bridge Engineering-Challenges in the 21st Century, Hong Kong. Chen, W.-F., and Duan, L. (1999). Bridge Engineering Handbook, CRC Press, Boca Raton. Collins, M. P., and Mitchell, D. (1997). Prestressed Concrete Structures, Response Publications, Toronto. Cortright, R. (1997)."Glenfinnan Viaduct." Nicolas Janberg's Structurae, <en.structurae.de/photos/index.cfm?JS=54616> (June 1, 2009). Cremer, J.-M., Counasse, C., and Delforno, J.-Y. (2003). "The Sart Canal-Bridge, houdeng Aimeries, Belgium." Structural Engineering International: Journal of the International Association for Bridge and Structural Engineering (IABSE), 13(1), 19-22. Crouch, A. G. D. (1974). "Bridge Aesthetics - A Sociological Approach." Civil Engineering Transactions, CE16(2), 138 - 142. CSA (1982). "Steel for Prestressed Concrete Tendons." CSA G279-M1982, Canadian Standards Association, Rexdale. CSA. (2006). Canadian Highway Bridge Design Code, C. S. Association, ed., Canadian Standards Association, Mississauga, Ontario.

107

108 Darger, N. J. (2002)."Waterloo Bridge (1945)." Nicolas Janberg's Structurae, <en.structurae.de/photos/index.cfm?JS=3502> (June 1, 2009). Dorton, R. A. (1991). "Aesthetic Considerations for Bridge Overpass Design." Bridge aesthetics around the world, Transportation Research Board, National Research Council, Washington, D.C., 10-17. DSI. (2008). "DYWIDAG Post-Tensioning Systems." D.-S. International, ed., DYWIDAG-Systems International. Duan, L., Chen, K., and Tan, A. (1999). "Design Practice in Europe." Bridge Engineering Handbook, W.-F. Chen and L. Duan, eds., CRC Press Boca Raton, 10-11 to 10-37. Dufferin Concrete. (2009)."2009 Price List." Dufferin Concrete, <www.dufferinconcrete.ca> (June 1, 2009). Elliott, A. L. (1991). "Creating a Beautiful Bridge." Bridge aesthetics around the world Transportation Research Board, National Research Council, Washington, D.C., 215-229. Gauvreau, P. (2006). "Bridges." Post-Tensioning Manual, Post-Tensioning Institute, Phoenix, 195240. Gottemoeller, F. (2004). Bridgescape : The Art of Designing Bridges, Wiley, Hoboken. Hassanain, M. (2002). "Optimal Design of High-Performance Concrete Adjacent Precast Box Girder Bridges." 6th International Conference on Short & Medium Span Bridges:Developments in Short & Medium Span Bridge Engineering, CSCE, Vancouver. Hewson, N. R. (2003). Prestressed Concrete Bridges : Design and Construction, Thomas Telford, London. Hillerborg, A. (1996). Strip Method Design Handbook, E & FN Spon, London. Holgate, A. (1996). The Art of Structural Engineering: The Work of Jorg Schlaich and His Team, Ed. Axel Menges, Stuttgart. Holowka, M. (1979). "Deck Slab Investigation of a Voided Posttensioned Concrete Bridge Deck Bridge 20." Ontario Ministry of Transportation and Communications, St. Catharines. Janberg, N. (2009)."Structurae." <en.structurae.de> (June 1, 2009). Kosmatka, S. H., Kerkhoff, B., Panarese, W. C., MacLeod, N. F., and McGrath, R. J. (2002). Design and Control of Concrete Mixtures, Cement Association of Canada, Ottawa. Leonhardt, F. (1979). Grundlagen des Massivbrckenbaues, Springer-Verlag, Berlin/Heidelberg. Leonhardt, F. (1982). Brcken : sthetik und Gestaltung = Bridges: Aesthetics and Design, Deutsche Verlags-Anstalt, Stuttgart. Linse, A., and Wssner, K. (1978). "Kochertalbrcke-Entwrfe einer Grossbrcke. " Bauingenieur, 53(12), 453-463.

109 Liu, X. P., Qin, Z. P., Leung, Y. W., and Yue, L. (2007). "Construction of the precast segmental approach structures for Sutong Bridge." 5th International Conference on Current and Future Trends in Bridge Design, Construction and Maintenance, Thomas Telford Ltd, Beijing. MacGregor, J., and Bartlett, F. M. (2000). Reinforced Concrete: Mechanics and Design, Prentice Hall Canada, Scarborough. Meades, P., and Green, R. (1974). "Model Studies for Voided Posttensioned Concrete Slab Bridges." Ministry of Transportation and Communications, Ontario, Toronto. Menn, C. Christian Menn, Consulting Engineer: Projects." <www.christian-menn.ch> (July 1, 2009). Menn, C. (1990). Prestressed Concrete Bridges P. Gauvreau, ed., P. Gauvreau, translator, Birkhuser Verlag, Basel. Mossot, J. (2007)."Changis-sur-Marne Bridge " Nicolas Janberg's Structurae, (June 1, 2009). MTO (2003). "Structural Manual." Ontario Ministry of Transportation, Bridge Office, St. Catharines. Muller, J. (1999). "Design Practice in Europe." Bridge Engineering Handbook, W.-F. Chen and L. Duan, eds., CRC Press Boca Raton, 64-61 to 64-37. NRS. (2008)."Gantries-Guangzhou LRT Project." NRS Bridge Construction Equipment, <http://www.nrsas.com/v2009/project/pro_gantries.php> (June 1, 2009). O'Brien, E. J., and Keogh, D. L. (1999). Bridge Deck Analysis, E & FN Spon, London. O'Connor, C. (1991). "Empirical Assessment of Bridge Aesthetics: An Australian View." Bridge Aesthetics Around the World, Transportation Research Board, National Research Council, Washington, D.C., 230-240. OHBDC. (1983). Ontario Highway Bridge Design Code, Highway Engineering Division, Toronto. OPSS. (2007). Material Specification for Concrete - Materials and Production, Ontario Provincial Standard Specification. OSHA. (2006)."Overhead Launching Gantry Crane Safety and Health Information Bulletin." Occupational Safety & Health Administration, <www.osha.gov/dts/shib/shib050106.html> (June 1, 2009). Park, R., and Gamble, W. L. (2000). Reinforced Concrete Slabs, Wiley, New York. PEER. (2005)."Neckar Valley Viaduct, Weitingen, Germany (1978)." Pacific Earthquake Engineering Research Center, University of California, <http://nisee.berkeley.edu/elibrary/getpkg?id=GoddenG43-50> (July 1, 2009). Sauvageot, G. (1999). "Segmental Concrete Bridges." Bridge Engineering Handbook, W.-F. Chen and L. Duan, eds., CRC Press, Boca Raton.

110 Scollard, C. R., and Bartlett, F. M. (2004). "Rehabilitation Criteria for Post-Tensioned Voided-Slab Bridges." Canadian Journal of Civil Engineering, 31(6), 977-987. SNC-Lavalin. (2008). "Highway 400 Major MacKenzie Drive to Highway 11 Heritage Bridge Replacement Design Concept Study." Report prepared by SNC-Lavalin, submitted to Ontario Ministry of Transportation. TRB. (1990). Truck Weight Limits: Issues and Options, Transportation Research Board Washington, D.C. Vogel, T. (1997). Christian Menn-Brckenbauer, Birkhuser Verlag, Basel, 57. Wasserman, E. P. (1991). "Aesthetics for Short- and Medium-Span Bridges." Bridge aesthetics around the world Transportation Research Board, National Research Council, Washington, D.C., 58-66. Weinlandbrcke (1958). "Die Weinlandbrcke in der Umfahrungsstrasse von Andelfingen." Direktion der ffentlichen Bauten des Kantons Zrich (Administration of Public Buildings of the Canton of Zrich).

APPENDIX A: CHAPTER 2 SUPPLEMENTARY INFORMATION

111

A.1 Cast-in-Place on Falsework Box-Girder

Table A-1. Cross-sectional drawings for cast-in-place on falsework box-girders in Chapter 2
Bridge no. 1 Cross-section Bridge no. 13 Cross-section Bridge no. 23 & 24 Cross-section

14

25

15

26

5&6

16

27

17

28

18

29

19

31

10

20

32

11

21

33

12

22

34

112

Table A-1. Cross-sectional drawings for cast-in-place on falsework box-girders in Chapter 2 (continued)
Bridge no. 35 Cross-section Bridge no. 38 Cross-section Bridge no. 42 Cross-section

36

40

43

37

41

44

Table A-2. Cast-in-place on falsework box-girders in Chapter 2

Bridge no. 1 2 3 Name Grenz Bridge at Basel Sart Canal-Bridge Weyermannshaus Bridge Eastbound Walnut Viaduct Taiwan High Speed Rail (1) Taiwan High Speed Rail (2) Pregorda Bridge Almese Viaduct Condove Viaduct Location Switzerland Belgium Switzerland CIP on falsework CIP span-byspan CIP span-byspan Span-by-span on falsework Balanced cantilever Balanced cantilever 2*(21+ 3*40 + 20.7) = 323.4 2*(20.65+17*40+21.15) = 1443.6 24.4+35.4+38.1+ 38.1+34.4+37.16+ 38.1+38.6+38.6+ 34.0+24.7 = 382 3*25+25.5+40+25+2*30 +3*25 = 300.5 25+35+26.5+25+ 40+24.5+26.5+4* (30) = 322.5 2002 Incremental launching 13 spans Completion year Construction method Span arrangement (m) Span length L (m) 35.4 36 37.75 Depth h (m) 2 3 2 Span-todepth ratio L/h 17.7 12 18.88 Reference Drawing from Gauvreau (Cremer et al. 2003) Drawing from Gauvreau Designer: T.Y. Lin International for Department of Highways, Colorado Drawing from Gauvreau Drawing from Gauvreau Drawing from Gauvreau Designer: Christian Menn Drawing from Gauvreau (Vogel 1997) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data)

U.S.A.

1986

38.6

1.68

23.0

5 6

Taiwan Taiwan

2000 2000

40 40

3.5 3.5

11.4 11.4

7 8 9

Switzerland Italy Italy

1974 1990 1992

40 40 40

1.8 2.2 2.2

22.2 18.2 18.2

113

Table A-2. Cast-in-place on falsework box-girders in Chapter 2 (continued)

Bridge no. 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Name Gravio Viaduct Borgone Viaduct Quadinei Bridge Altstetter Viaduct Reuss Bridge Cerchiara Viaduct Castello Viaduct Costacole Viaduct Ferroviario Overpass at Bolzano Krebsbachtal Bridge Shatt Al Arab Bridge Ancona Viaduct Felsenau Bridge (approaches) La Molletta Viaduct Fosso Capaldo Viaduct Sihlhochstrasse Bridge Grosotto Viaduct Grosio Viaduct Tiolo Viaduct Denny Creek Location Italy Italy Switzerland Switzerland Switzerland Italy Italy Italy Italy Germany Iraq Italy Switzerland Italy Italy Switzerland Italy Italy Italy U.S.A. 1980 1992 1994 Balanced cantilever Balanced cantilever Balanced cantilever 45+15*55+45= 915 45+24*55+45 = 1410 45+9*55+45 = 585 Spans = 57.32 1975 1988 1988 1972 1992 1992 1992 1974 1975 1978 Incremental launching Incremental launching Segmental Span-by-span on falsework Segmental Segmental CIP on falsework Balanced cantilever Balanced cantilever Balanced cantilever 28.1+34+37.5+40*3+42. 5+40*2+37.5+32.5 = 412 40.30+28*42.50+34= 1264.3 40.30+6*42.50+ 36 = 331.3 34+5*42.50+34 = 280.50 Spans = 36.7 to 45 Total length = 134 Completion year 1992 1992 1967 Construction method Balanced cantilever Balanced cantilever Span-by-span on falsework Span arrangement (m) 2*(20.65+12*40+21.15) = 1043.6 2*(19.40+26*40+19.15) = 2157.1 Span length L (m) 40 40 40 41 42.5 42.5 42.5 42.5 45 45 Spans = 38.25 to 46.90 Total length = 761 Spans = 21.76 to 47.50 Total length = 2015.9 48*5 = 240 (40+2*50+40) +(40+2*50+40) = 360 (40+11*50+40) +(40+11*50+40) = 1260 46.9 47.5 48.0 50 50 54.5 55 55 55 57.32 Depth h (m) 2.2 2.2 2 1.9 2.4 2.3 2.3 2.3 1.6 3.49 3.65 2.3 3.0 2.4 2.4 1.85 2.75 2.75 2.75 2.744 Span-todepth ratio L/h 18.2 18.2 20 21.6 17.7 18.5 18.5 18.5 28.125 12.9 12.8 20.7 16.0 20.8 20.8 29.5 20 20 20 20.9 Reference DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) (Menn 1990) Drawing from Gauvreau Drawing from Gauvreau Designer: Christian Menn DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) Drawing from Gauvreau (Menn 1990) Drawing from Gauvreau DEAL Job Report (Segmental Bridge Data) Designer: Christian Menn (Menn 1990) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) Drawing from Gauvreau DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) Drawing from Gauvreau

114

Table A-2. Cast-in-place on falsework box-girders in Chapter 2 (continued)

Bridge no. 30 31 32 33 34 35 Name Woronora River Bridge Valentino Viaduct Giaglione Viaduct Venaus Viaduct Passeggeri Viaduct Brunetta Viaduct Pietrastretta Viaduct Deveys Viaduct Gruyre Lake Viaduct Interstate 895 Bridge over James River (approaches) Ltten Bridge Savona Mollere Viaduct Ruina Viaduct Weinland Bridge Kocher Valley Bridge Location Australia Italy Italy Italy Italy Italy Completion year 2001 1990 1992 1992 1992 1992 Construction method Incremental launching Balanced cantilever Balanced cantilever Balanced cantilever Segmental Segmental Span arrangement (m) Total length = 521 30.80+6*60+31 = 421.8 2*(30.9+9*60+31.05) = 1143.90 (30.7+9*60+ 30.9) + (30.95 +7*60+ 30.7) = 1083.25 (30.9+5*60+31.05)+(30. 9+4*60+31.05) = 663.90 (30.85+11*60+31.30)+(3 1.05+ 9*60+59.05) = 1352.25 (30.70+4*60+ 30.80)+(30.70+3*60+30. 90) = 543.10 (30.70+3*60+ 30.70)+(30.70+4*60+31. 70) = 542.80 Total length = 2043 Span length L (m) 58.7 60 60 60 60 60 Depth h (m) 4 3 3 3 3 3 Span-todepth ratio L/h 14.7 20 20 20 20 20 Reference (Bennett and Taylor 2002) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data)

36

Italy

1992

Segmental

60

20

DEAL Job Report (Segmental Bridge Data)

37

Italy

1992

Segmental CIP span-byspan Balanced cantilever

60

20

DEAL Job Report (Segmental Bridge Data)

38

Switzerland

1978

60.48

15.12

Drawing from Gauvreau

39 40 41 42 43 44

U.S.A. Switzerland Italy Italy Switzerland Germany

2002

64 65

3 3.6 3 3.8 3.9 6.5

21.3 18.1 22.5 19.3 22.6 21.2

(Belli 2003) Drawing from Gauvreau DEAL Job Report (Segmental Bridge Data) Drawing from Gauvreau (Menn 1990, Weinlandbrcke 1958) Designer: Fritz Leonhardt (Linse and Wssner 1978, PEER 2005) Drawing from Gauvreau

1994 1984 1958 1971

Segmental

35.3+67.5+36.5+67.5+35 .3 = 242.1 Spans = 44.5 to 73.5 Total length = 785 57+76+88+66 = 287 81+7*138+81=1128

67.5 73.5 88 138

Span-by-span on falsework Balanced cantilever

115

A.2 Cast-in-Place on Falsework Solid Slab

Table A-3. Cross-sectional drawings for cast-in-place on falsework solid slabs in Chapter 2
Bridge no. 46 47 Cross-section Bridge no. 51 58 Cross-section Bridge no. 64 65 Cross-section Bridge no. 70 71 Cross-section

48

60

66

72

49

61

67

50

62

68

Table A-4. Cast-in-place on falsework solid slabs in Chapter 2

Bridge no. 45 46 47 & 48 49 50 & 51 52 Name Location Completion year 2000 1967 1967 1967 1967 1986 Construction method N/A CIP on falsework CIP on falsework CIP on falsework CIP on falsework N/A 12.192+15.24+12.192 = 39.624 12.192+15.24+12.192 = 39.624 12.192+15.24+12.192 = 39.624 12.192+15.24+12.192 = 39.624 Span arrangement (m) Span length L (m) 13.2 15.24 15.24 15.24 15.24 16.7 Depth h (m) 0.65 0.686 0.686 0.686 0.686 0.3 Span-todepth ratio L/h 20.3 22.2 22.2 22.2 22.2 55.7 Reference Designer: Shirish Patel & Associates Consultants Private Limited (Janberg 2009) Department of Highways Ontario Department of Highways Ontario Department of Highways Ontario Department of Highways Ontario Designers: Schlaich, Bergermann und Partner (Holgate 1996, Janberg 2009) Designer: Haas Consult and Ingenieurbro Perlebery (Janberg 2009)

Khandeshwar Bridge Spadina Ave. Bridge #16, Hwy 401 Spadina Ave. Bridge #18A & B, Hwy 401 Spadina Ave. Bridge #19, Hwy 401 Spadina Ave. Bridge #21A & B, Hwy 401 Sindelfingen Footbridge

India Canada Canada Canada Canada Germany

53

L 333 Overpass at Bassum

Germany

1993

N/A

13.5+16.87+13.5

16.87

0.77

21.9

116

Table A-4. Cast-in-place on falsework solid slabs in Chapter 2 (continued)

Bridge no. 54 55 56 57 58 59 60 & 61 62 63 64 65 Name Location Completion year 1996 1975 1992 Construction method N/A Incremental launching N/A N/A 1964 1961 1963 1963 1994 CIP on falsework N/A CIP on falsework CIP on falsework N/A CIP on falsework 1963 CIP on falsework Spans = 18.3 to 22.25, Total length =1630 19.5+2*24.4+19.5 = 87.8 Total length = 536 18.2+25.7+2*23.4+29. +18.2 = 138 18.2+26.0+24.0+23.4+29. 4+18.2 = 139 23.4+8*31.2+23.4 16.4+23.2+21.9+31.4+24. 4= 117 15.24+2*32+23.17+32+1 5.2 = 149.655 19.96+34.08+29.56+27.8 8+36.57+34.13+25.50+26 .43+38.1+2*35.05+36.27 +36.2+36.5+25.603 = 477 18.2+38.7+23.1+20.2+34. 4+18.2 = 153 27.4+36.5+39.6+27.432 = 131 40.465 36+45.1+36=117 22.86+45.72+22.86 = 91.44 28.042+47.549+24.984 = 100.575 16+3*21+16+3*21+16 Span arrangement (m) Span length L (m) 17 21 21.65 22.25 24.4 25 29.4 29.4 31.2 31.4 32 Depth h (m) 0.4 0.88 1.12 1.16 0.686 1 0.762 0.762 1.56 0.99 0.914 Span-todepth ratio L/h 42.5 23.9 19.3 19.2 35.6 25 38.6 38.6 20 31.7 35 Reference Designer: Fischer + Friedrich Beratende Ingenieure (Janberg 2009) Designer: SFEDTP (Janberg 2009) Designer: Metz-Herder-Wendt und Billig, Schler-Plan Ingenieurgesellschaft mbH (Janberg 2009) Drawing from Gauvreau Department of Highways Ontario Drawing from Gauvreau Designer: Fritz Leonhardt (Leonhardt 1982, PEER 2005) Department of Highways Ontario Drawing from Gauvreau Department of Highways Ontario Drawing from Gauvreau (Janberg 2009) (Holowka 1979) Department of Highways Ontario Drawing from Gauvreau Department of Highways Ontario Drawing from Gauvreau Department of Highways Ontario Drawing from Gauvreau Department of Highways Ontario Drawing from Gauvreau Designer: Wolfgang Linder (Janberg 2009) (Meades and Green 1974) Department of Highways Ontario Drawing from Gauvreau Department of Highways Ontario Drawing from Gauvreau

Waiblingen Footbridge Mako Bridge Kittelbaches Bridge San Francisco Airport Viaduct St. Vincent Street Overpass Bridge across Jan-WellenPlatz Spadina Ave. Bridge #14 & #15, Hwy 401 Spadina Ave. Bridge #12, Hwy 401 Saale Bridge at Rudolphstein Bridge #20 at Hwy 401/427 Interchange Spadina Ave. Bridge #5, Hwy 401 Spadina Ave. Bridge #11 Hwy 401 Spadina Ave. Bridge #22, Hwy 401 Spadina Ave. Bridge #23, Hwy 401 Hundschipfen Bridge McCowan Road Underpass Spadina Ave. Bridge #24, Hwy 401 Spadina Ave. Bridge #4, Hwy 401

Germany Senegal Germany U.S.A. Canada Germany Canada Canada Germany Canada Canada

66

Canada

1963

CIP on falsework

38.1

1.27

30

67 68 69 70 71 72

Canada Canada Switzerland Canada Canada Canada

1963 1963 2000

CIP on falsework CIP on falsework N/A CIP on falsework

38.7 39.62 40.46 45.1 45.72 47.54

1.27 1.27 1 1.52 1.37 1.372

30.5 31.2 40.5 29.7 33.4 34.7

1963 1963

CIP on falsework CIP on falsework

117

A.3 Precast Segmental Span-by-Span Box-Girder

Table A-5. Cross-sectional drawings for precast segmental span-by-span box-girders in Chapter 2
Bridge no. 73 Cross-section Bridge no. 78 Cross-section Bridge no. 83 Cross-section

75

79

84

76

80 & 81

85

77

82

86

Table A-6. Precast segmental span-by-span box-girders in Chapter 2

Bridge no. 73 74 75 76 77 78 79 80 81 82 83 84 85 86 Name Location Completion year Construction method Segmental spanby-span Segmental with launching girder Segmental Segmental Segmental Segmental Segmental Balanced cantilever Balanced cantilever Balanced cantilever Balanced cantilever Balanced cantilever Balanced cantilever Balanced cantilever Span arrangement (m) Span length L (m) 34 36.85 39 39 39 39 39 40 40 45 47 60 60 75 Depth h (m) Span-todepth ratio L/h 15.7 17.5 17 17 17 17 17 16.67 16.7 16.1 16.8 17.6 18.75 18.75 Reference

Bukit Panjang LRT System 801 Wiscasset Bridge Chiovano Viaduct Collecastino Viaduct Fiumetto Viaduct San Leonardo Viaduct Petto Viaduct Cadramazzo Viaduct Fella IX Viaduct Malborghetto Viaduct Val Freghizia Viaduct Fella IV Viaduct Ngong Shuen Chau Viaduct Sutong Bridge Approach (Nantong side)

Singapore U.S.A. Italy Italy Italy Italy Italy Italy Italy Italy Italy Italy China China

1998 1981 1993 1991 1992 1992 1992 1985 1986 1986 1988 1985 2007 2007

Total length =7715.24 27.5+21*36.85+27.5 = 829 39+18*39+39 = 780 39+25*39+39 = 1053 39+11*39+39 = 507 39+5*39+39 = 273 39+11*39+39 = 507 2*(35+13*40+35) = 1180 2*(41+18*40+40.80) = 1604 2*(40+16*45+40) = 1604 (36+15*47+36) *2 = 1554 2*(45+6*60+45) = 960 45+4*60+45 50+19*75

2.16 2.1 2.3 2.3 2.3 2.3 2.3 2.4 2.4 2.8 2.8 3.4 3.2 4

DEAL Job Report (Segmental Bridge Data) (Janberg 2009) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) DEAL Job Report (Segmental Bridge Data) (Cao et al. 2006) (Liu et al. 2007)

118

APPENDIX B: SUPPORTING CALCULATIONS

B.1 Flexural Strength for Bonded Tendons at ULS
c = 3.5x10-3 c 1fc a= 1c z fpy Strain Stress Forces C Mr = Tz T = pApfpy

This calculation is performed for a cast-in-place on falsework box-girder analysis case with span length L of 50m and span-to-depth ratio L/h of 20.
MULS = moment demand = 53 000 kNm

Concrete properties
fc = 50 MPa fcr = 2.8 MPa c = 0.75

Prestressing tendon properties

fpy = 1670 MPa p = 0.95 Number of strands = 136 2 Ap = area of prestressing tendons = (136 strands)(140 mm ) = 0.019 m

Cross-sectional properties
b = width of compressive component = 15 m d = distance from extreme compression fibre to centroid = 2.19 m y = distance from base to centroid = 1.69 m 4 Ig = moment of inertia = 6.06 m

Flexural strength requirements

1 = 0.85 0.0015f c = 0.78 [CHBDC, Cl. 8.8.3] 1 = 0.97 0.0025f c = 0.85 T = p fpy A p = 0.95 1670 MPa 0.019m2 = 30 200 kN a= = T c 1 f c b

30 200 kN 0.75 0.78 50 MPa 15 m = 0.688 m z= d a 2

= 2.19 m = 2.16 m

0.0688 m 2

119

120
Mr = T z = 30 200 kN 2.16 m = 65 100 kNm Mr = 65 100 kNm > MULS = 53 000 kNm number of tendons is sufficient

Minimum reinforcement requirement

Mr > 1.2 Mcr = 1.33 MULS 1.2 Mcr 1.33 MULS [CHBDC, Cl.8.8.4.3]

fcr Ig = 12 000 kNm y = 70 500 kNm requirement is satisfied

Mr = 53 000 kNm > 1.2 Mcr = 12 000 kNm

Maximum reinforcement requirement

c d

< 0.5

[Cl.8.8.4.5]

a 0.0688 m = = 0.081m 1 0.85 c 0.081 m = = 0.037 < 0.5 d 2.91 m c=

B.2

Shear Strength at ULS

This calculation is performed for a cast-in-place on falsework box-girder analysis case with span length L of 50m and span-to-depth ratio L/h of 20. The following calculation aims to find the minimum stirrup spacing s needed to satisfy shear requirements.
VULS = shear demand = 11 300 kN MULS = moment demand = 5330 kNm Vp = component of prestressing force in the direction of shear = -3000 kN Mp = prestressing moment = 8300 kNm

Concrete properties
fc = 50 MPa fcr = 2.8 MPa c = 0.75 ag = aggregate size = 10 mm

Prestressing tendon properties

fpy = 1670 MPa p = 0.95 Ep = 200 000 MPa Number of strands = 136 2 2 Ap = area of prestressing tendons = (136 strands)(140 mm ) = 0.019 m 2 Aps = area of prestressing tendons in the flexural tension side = 0.568 m

Non-prestressed reinforcement properties

fy = 400 MPa s = 0.90

121
Es = 200 000 MPa 2 Av = area of stirrup = 0.0012 m (4-20M stirrups) 2 As = area of longitudinal non-prestressed reinforcement = 0.01 m

Cross-sectional properties
b = width of compressive component = 15 m d = distance from extreme compression fibre to centroid = 1.31 m h = height of section = 2.5 m 0.72 h dv = max = 1.8 m 0.9 d bv = minimum web width within dv = 0.938 m sz = crack spacing parameter = dv = 1.8 m 35sz sze = = 2.52 m 15 + a g fpo = 0.7fpu for bonded tendons = 1300 MPa p, for unbonded tendons

Mf + Vf Vp Aps fp0 d x = v = 0.003 = 0, 2 Es As + Ep Aps = 29 + 7000x = 0.4 1 + 1500x 0.88 +

s ze 2500

0 < x < 3.0 103

= 55 = 0.148

1300 1000 + sze

Vc = concrete shear resistance = 2.5c fcr bv dv = 3550 kN Vr = Vc + Vs + Vp Vs = shear resistance provided by stirrups = Vr Vc Vp = 10 800 kN s fy A v dv cot Vs = s Av Vs = = 0.0238 m s s fy A v dv cot Av s= = 420 mm > 300 = 300 0.0238 m

122

B.3 Thermal Gradient Moments

This section describes the derivations of moments and stresses due to thermal gradients.

Final curvature of bending :

= y + 0 y

The primary stresses:

fr y = Ec y f = Ec y c y = Ec 0 + y c (y) where c = thermal coefficient of concrete = 10 106 /

[CHBDC, Cl.8.4.1.3]

Axial restraint force Pr and restraint moment Mr:

h h

Pr = Mr =

fr y b y dy = Ec
h

y c y b y dy
h 0

fr y b y y n dy = Ec

y c y b y y n dy

Note the following relationships:

yb y dy = nA y 2 b y dy = I + n2 A b y dy = A

Now, equate both Pr and Mr to zero because there are no external forces and internal redundancies have been removed to make the structure statically-determinate. First, Pr is equated to zero to obtain the final strain at elevation y=0 (0):
Pr = 0
h 0

fr y b y dy = 0
h 0

Ec
h 0 h 0

y c y b y dy = 0

0 + y c y b y dy = 0
h h

0 b y dy +

yb y dy
0

c yb y dy = 0

0 A + nA

c yb y dy = 0

123
c A
h

0 =

y b y dy n
0

Similarly, Mr is equated to zero to obtain the final curvature of bending :

Mr = 0
h 0

fr y b y y n dy = 0
h 0

Ec
h 0 h 0

y c y b y y n dy = 0

0 + y c y b y y n dy = 0
h h h

0 yb y dy

0 b y ndy
h 0

c y b y ydy +
h 0 h 0

nc y b y dy

+
h

y 2 b y dy

yb y ndy = 0 nc y b y dy + I + n2 A n2 A = 0

0 nA 0 nA
h 0

c y b y ydy +
h 0

nc y b y dy c I
h 0

c y b y ydy + I = 0

y b y y n dy

To remove this incompatible rotation, restraint moment M is needed at each end of the span:
M = Ec I

This restraint moment is distributed to obtain the final moments caused by thermal gradient Mfinal. This final moment causes secondary stresses fs:
fs y = Mfinal (y n) I

The total thermal stresses that need to be considered in design checks are computed with the following equation:
ftemp y = fr y + fs y = Ec 0 + y c (y) + Mfinal (y n) I

124

B.4 External Tendon Force

This section describes an iterative process to calculate the prestressing force P in unbonded tendons to equilibrate external load Q at ULS.
Q

x P P

This method requires the knowledge of prestressing steel area and effective prestress after all losses. First, the force in the unbonded tendons is assumed to be Pi. Pi must be less than the Q yielding force Py but greater than the minimum force Pmin needed to equilibrate the external load Q. Then the change in length due to force lPF is computed using:
lPF = (Pi P ) l A P EP P0

where

P = effective prestress force after all losses = p,AP = 60%fpu AP lP0 = length of tendon when prestress force is P

The change in length due to deformation lPD, which is the actual length of tendon when loaded by c Q and Pi, is also computed:
C C

cp= p

lPD =

cP x dx
cpp T

T=Apfpy when p+p>py

where cP = concrete strain at PT level due to Q and Pi

cp the two changes If

cp in length are not equal, Pi is varied and the calculations are performed again. =p cp, average This process is iterated until lPF equals lPD, and Pi at this point is theactual force in the tendons at ULS. This iterative process is illustrated in the following figure. = = total deformation/undeformed length of tendon
cp, average p

cp

120 100 p

cp, average=p
PULS

Py

p,total = p+ p = cp+p 80 tal tendon strain varies along the tendon length Deformation (mm) 60
40 20 0 40000 -20

p,total =p+ p = cp, average+p Total tendon strain remains constant along the tendon length. Tendon stress is averaged out between adjacent anchorages. PTherefore, the flexural resistance is lower at critical locaions t min compared to the bonded case.

llPF PF
lPD lPD

42000

44000 P (kN)

46000

48000

50000

125

B.5 Total Construction Cost

The following calculation is performed for a cast-in-place on falsework box-girder with span length L of 35m and total bridge length of 260m. Material unit prices Concrete = $ 1500/m3 Prestressing tendons = $ 8.5/kg Reinforcing steel = $ 5.0/kg Other construction costs are more dependent on the span length and are assumed to be the same for all cases. They are computed based on the total superstructure cost of the baseline case.
Item Concrete 3 ($ 1500/m ) Prestressing tendon ($ 8.5/kg) Concrete volume Total concrete cost Tendon length per strand Number of strands Total tendon length Tendon mass Total prestressing tendon cost Mass of reinforcing steel per concrete volume Reinforcing steel mass Total reinforcing steel cost Unit m
3

10 2700 $ 4,050,000 359 52 18700 20600 $ 175,000

3

L/h 20 (baseline) 2410 2260 $ 3,610,000 $ 3,390,000 15 359 76 27300 30100 $ 255,000 80.0 193000 $ 963,000 $ 4,830,000 $ 3,910,000 358 108 38700 42600 $ 362,000 82.0 186000 $ 928,000 $ 4,680,000 $ 3,910,000

25 2170 $ 3,260,000 358 136 48700 53600 $ 456,000 86.1 187000 $ 936,000 $ 4,650,000 $ 3,910,000

m m kg

Reinforcing steel ($ 5.0/kg)

kg/m kg

80.0 216000 $ 1,080,000 $ 5,310,000 $ 3,910,000

Total superstructure cost (54.5% of total construction cost) Other construction cost (e.g. substructure, mobilization, accessories) (45.5% of total construction cost) Total construction cost Total construction cost per deck area

per m

$ 9,240,000 $ 2,360

$ 8,740,000 $ 2,240

$ 8,590,000 $ 2,200

$ 8,560,000 $ 2,200

APPENDIX C: SUMMARY OF RESULTS

126

C.1 Cast-in-Place on Falsework Box-Girder

Table C-1. Summary of results of cast-in-place on falsework box-girder analysis
Span-todepth ratio 10 Span length (m) 35 $ 2370 (107%) 4 15-strand ducts 50 $ 2670 (109%) 4 19-strand ducts 60 $ 2890 (112%) 4 27-strand ducts 75 $ 3280 (111%) 4 37-strand ducts

15

$ 2250 (102%) 4 19-strand ducts

$ 2490 (102%) 4 27-strand ducts

$ 2670 (103%) 4 31-strand ducts

$ 3030 (103%) 8 27-strand ducts

20

$ 2210 (100%) 4 27-strand ducts

$ 2450 (100%) 8 19-strand ducts

$ 2590 (100%) 8 27-strand ducts

$ 2950 (100%) 8 37-strand ducts

25

$ 2200 (99.4%) 8 19-strand ducts

$ 2450 (100%) 8 27-strand ducts

$ 2580 (99.5%) 8 27-strand ducts

$ 2950 (100%) 12 27-strand ducts

30

$ 2490 (102%) 8 37-strand ducts

$ 2600 (100%) 8 37-strand ducts

$ 3020 (103%) 12 37-strand ducts

35

$ 2660 (103%) 12 27-strand ducts

$ 3220 (109%) 16 37-strand ducts

127

C.2 Cast-in-Place on Falsework Solid Slab

Table C-2. Summary of results of cast-in-place on falsework solid slab analysis
Span-todepth ratio 30 Span length (m) 20 $ 2270 (100%) 10 27-strand ducts 25 $ 2650 (100%) 10 27-strand ducts 30 $ 3000 (100%) 12 27-strand ducts 35 $ 3450 (100%) 18 27-strand ducts

35

$ 2240 (98.1%) 12 27-strand ducts

$ 2560 (96.6%) 12 27-strand ducts

$ 2890 (96.3%) 14 27-strand ducts

$ 3320 (96.3%) 20 27-strand ducts

40

$ 2240 (98.8%) 14 27-strand ducts

$ 2520 (95.2%) 14 27-strand ducts

$ 2830 (94.2%) 18 27-strand ducts

$ 3250 (94.3%) 22 27-strand ducts

45 $ 2860 (95.4%) 20 27-strand ducts $ 3280 (95.1%) 24 27-strand ducts

128

C.3 Precast Segmental Span-by-Span Box-Girder

Table C-3. Summary of results of precast segmental span-by-span box-girder analysis
Span-todepth ratio 15 30 $ 2040 (100%) 6 27-strand ducts Span length (m) 40 $ 2280 (100%) 6 37-strand ducts 50 $ 2380 (100%) 6 37-strand ducts

20 $ 2040 (100%) 6 37-strand ducts

$ 2250 (98.6%) 8 37-strand ducts

$ 2350 (98.4%) 8 37-strand ducts

25

$ 2280 (100%) 8 37-strand ducts

$ 2390 (100%) 9 37-strand ducts

129

130

C.4 Sensitivity with Respect to Changes in Construction Cost Breakdown

$8,000 Cost variation 0.1% (3.4%) $6,000 Superstructure as % of total construction cost 20%

Total construction cost per deck area ($/m2) $4,000

0.2% (5.2%)

30%

0.3% (6.9%) 0.3% (8.6%) 0.4% (9.0%) 0.4% (10%) 0.5% (12%) 0.5% (14%)

40% 50% 54.5% (Menn) 60% 70% 80%

$2,000

$0 0 5 10 15 L/h 20 25 30 35

Figure C-1. Cast-in-place on falsework box-girder with L=50m

$8,000 Cost variation 1.7% (1.7%) Superstructure as % of total construction cost 20%

$6,000 Total construction cost per deck area ($/m2) $4,000 3.4% (3.5%) 4.3% (4.5%) 4.8% (5.0%) 5.1% (5.4%) 6.0% (6.4%) 6.8% (7.3%) 40% 50% 54.5% (Menn) 60% 70% 80%

2.6% (2.6%)

30%

$2,000

$0 25 30 35 L/h 40 45

Figure C-2. Cast-in-place on falsework solid slab with L=25m

131
$8,000 Superstructure as % of total construction cost 20% $6,000 Total construction cost per deck area ($/m2) $4,000 0.5% (1.3%) 30% 0.3% (0.8%)

Cost variation

0.6% (1.7%) 0.8% (2.1%) 0.8% (2.3%) 0.9% (2.5%) 1.1% (3.0%) 1.2% (3.4%)

40% 50% 54.5% (Menn) 60% 70% 80%

$2,000

$10 15 20 L/h 25 30

Figure C-3. Precast segmental span-by-span box-girder with L=40m