You are on page 1of 204

Preliminary Design of Slender Reinforced

Concrete Highway Bridge Pier Systems

by

Aleksandar Kuzmanovic

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 Aleksandar Kuzmanovic (201 4)

Preliminary Design of Slender Reinforced


Concrete Highway Bridge Pier Systems
Aleksandar Kuzmanovic
Master of Applied Science
Graduate Department of Civil Engineering
University of Toronto
2014

Abstract
Feasible span-to-depth ratios for many modern bridge systems have been identified and
documented in literature. No such parameters have been adequately identified in terms of
proportioning bridge piers.
This thesis includes a study of 22 existing reinforced concrete highway bridges and their
respective pier systems to determine the state-of-the-art in design. The effect of different
geometric and material parameters such as concrete strength, reinforcement ratio and
slenderness ratio on the structural behavior of individual piers and multiple pier systems was
examined. Approximate methods, which may be used for the purposes of preliminary design are
discussed and reviewed. Serviceability and ultimate limit states design aids that can be used to
identify appropriate preliminary cross-sectional pier dimensions and reinforcement ratios for
individual piers given various slenderness ratios were developed. The structural behavior as well
as an approach to the preliminary design of multiple pier bridge systems is presented.

ii

Acknowledgements
First of all I would like to thank my mother for the encouragement and support she has given me
throughout the completion of this thesis.
Second, I would like to thank my better half, Monica, whose patience and undying support
throughout has made this process that much more enjoyable.
I would like to thank Robert Maksymec and the kind folks at Planmac for their help throughout
the completion of this thesis.
I would also like to thank all of my friends and colleagues in the research group. I feel there are
not enough words to truly describe the degree of talent, ambition, and generosity that I have
been so privileged to have been exposed to. That being said, there are specific individuals that
deserve special recognition.
I would like to thank David Hubbell, with whom I have discussed many technical challenges
throughout the completion of this thesis. His contribution has most certainly been appreciated.
Robert Botticchio and David Wang also deserve special recognition. I have had the pleasure of
having many insightful discussions with both of these gentlemen throughout this endeavour.
I would also like to give special recognition to Ainur Otarbayeva whose help towards the
completion of the bridge database has been greatly appreciated.
Lastly, I would like to thank Professor Paul Gauvreau. Over the past two years he has been both a
mentor and a role-model engineer. I am forever grateful in the contributions that he has made,
not only towards the completion of this thesis, but to my personal development as a future
professional engineer. Thank you Professor.

iii

ABSTRACT .............................................................................................................................................. II

ACKNOWLEDGEMENTS ..................................................................................................................... III

CHAPTER 1. INTRODUCTION ......................................................................................................... 1


1.1

Motivation............................................................................................................................................ 1

1.2

Methods of column design and analysis ................................................................................................ 4

1.2.1

Numerical methods .................................................................................................................................... 4

1.2.2

Graphical methods...................................................................................................................................... 5

1.2.3

Finite element methods.............................................................................................................................. 7

1.2.4

Other methods ........................................................................................................................................... 8

1.3

Objectives and content of thesis ......................................................................................................... 11

CHAPTER 2. REVIEW OF EXISTING BRIDGE PIERS .............................................................. 14


2.1

Study of bridge pier systems ............................................................................................................... 14

2.2

Trends in Design Parameters ............................................................................................................... 25

2.2.1

Slenderness ratio against geometric reinforcement ratio ........................................................................ 26

2.2.2

Slenderness ratio against tributary surface area of deck ......................................................................... 27

2.2.3

Slenderness ratio against pier height ....................................................................................................... 28

2.2.4

Slenderness ratio against compressive strength of concrete ................................................................... 29

2.2.5

Slenderness ratio against visual slenderness ratio ................................................................................... 30

2.3

Comparison of select bridges from study............................................................................................. 31

2.3.1

Big Qualicium compared to Kings Highway No.II ..................................................................................... 31

2.3.2

Reuss-Brcke Wassen ............................................................................................................................... 33

2.3.3

Shin Chon Bridge....................................................................................................................................... 35

iv

2.4

Range of design parameters ................................................................................................................ 36

2.5

Concluding remarks ............................................................................................................................ 36

CHAPTER 3. APPROXIMATE METHODS .................................................................................. 38


3.1

Bilinear stress-strain formulation ........................................................................................................ 38

3.1.1

3.2

Stress-strain model calculation comparison ............................................................................................. 40

Vianello's method ............................................................................................................................... 43

3.2.1

Case 1: eccentricity is proportional to buckled shape of column ............................................................. 43

3.2.2

Case 2: eccentricity is not proportional to buckled shape of column ...................................................... 47

3.2.3

Eccentricity proportional to buckled shape vs. constant eccentricity ...................................................... 49

3.3

Menn's Method .................................................................................................................................. 51

3.3.1

Sectional capacity and response............................................................................................................... 52

3.3.2

Menn's method: analysis and discussion.................................................................................................. 53

3.3.2.1

Rigorous analytical method ............................................................................................................. 54

3.3.2.2

Validating the rigorous analytical method ....................................................................................... 56

3.3.2.3

Influence of loading conditions, axial load, and slenderness ratio .................................................. 59

3.3.2.4

Influence of reinforcement ratio ..................................................................................................... 74

3.3.2.5

Influence of concrete strength......................................................................................................... 77

3.4

Recommendations using Menns method ......................................................................................... 79

3.4.1

Axial load .................................................................................................................................................. 80

3.4.2

Slenderness ratio ...................................................................................................................................... 80

3.4.3

Reinforcement ratio.................................................................................................................................. 81

3.4.4

Concrete strength ..................................................................................................................................... 81

3.4.5

Applied load versus imposed deformation ............................................................................................... 82

3.4.6

Summary ................................................................................................................................................... 82

3.5

Concluding remarks ............................................................................................................................ 83

CHAPTER 4. INDIVIDUAL BRIDGE PIERS ................................................................................ 85


4.1

Free-standing pier loading conditions.................................................................................................. 86

4.1.1

Wind load.................................................................................................................................................. 86

4.1.2

Dead load .................................................................................................................................................. 89

4.1.3

Pier with concentric load .......................................................................................................................... 89

4.1.4

Pier with eccentric load ............................................................................................................................ 90

4.1.5

Inferred eccentricity limit ......................................................................................................................... 96

4.2

Serviceability limit states design ....................................................................................................... 102

4.2.1

Identifying a critical cross section ........................................................................................................... 103

4.2.2

Identifying average crack spacing ........................................................................................................... 104

4.2.3

Identifying allowable stress in tensile reinforcement ............................................................................ 110

4.2.4

Identifying allowable sectional moment and axial force ........................................................................ 113

4.3

Ultimate limit states design .............................................................................................................. 121

4.3.1

Sectional limits and deformation limits .................................................................................................. 122

4.3.2

Developing design aids ........................................................................................................................... 123

4.4

Concluding remarks .......................................................................................................................... 129

CHAPTER 5. MULTIPLE PIER SYSTEMS ................................................................................. 131


5.1

Defining a multiple pier system ......................................................................................................... 131

5.2

System buckling load ........................................................................................................................ 133

5.2.1

Calculating the global stability factor of a system .................................................................................. 134

5.2.2

Limitations of Menn's method ............................................................................................................... 137

5.2.2.1

Case 1: both piers have the same flexural stiffness ....................................................................... 138

vi

5.2.2.2

Case 2: pier B is considerably stiffer than pier A............................................................................ 139

5.2.2.3

Calculating true system buckling load ........................................................................................... 140

5.2.2.4

Comparison between SAP2000 and spring model ......................................................................... 143

5.3

Parametric study of large pier systems .............................................................................................. 146

5.3.1

5.4

SAP2000 model ....................................................................................................................................... 146

5.3.1.1

Parametric study case 1: ................................................................................................................ 148

5.3.1.2

Parametric study case 2: ................................................................................................................ 151

5.3.1.3

Parametric study case 3: ................................................................................................................ 153

5.3.1.4

Parametric study case 4: ................................................................................................................ 155

5.3.1.5

Parametric study case 5: ................................................................................................................ 157

Concluding remarks .......................................................................................................................... 161

CHAPTER 6. CONCLUSIONS ....................................................................................................... 162


6.1

Review of existing bridge piers .......................................................................................................... 162

6.2

Approximate methods ...................................................................................................................... 162

6.3

Individual bridge piers ...................................................................................................................... 164

6.4

Multiple pier systems........................................................................................................................ 165

6.5

Design recommendations ................................................................................................................. 166

REFERENCES ...................................................................................................................................... 168

APPENDIX A SUPPLEMENTARY INFORMATION ................................................................... 171


Derivation of Vianellos method of successive approximations ...................................................................... 172

vii

Example Menns method and proposed rigorous analytical method............................................................ 175

Example Cantilever pier example for serviceability limit states ................................................................... 178

Serviceability limit states M-N interaction envelopes .................................................................................... 181

Table of bridge study data ............................................................................................................................. 182

Table of bridge references ............................................................................................................................. 184

APPENDIX B SERVICEABILITY LIMIT STATES DESIGN AIDS ............................................ 185

APPENDIX C ULTIMATE LIMIT STATES DESIGN AIDS ........................................................ 187

viii

LIST OF FIGURES
Figure 1-1. Graphical analysis of columns (L/h = 20) Adapted from Beal (1986).
Figure 2-1 Structural illustrations of the 22 studied bridges at a 1:6000 scale.

6
22

Figure 2-2 Slenderness ratio compared to geometric reinforcement ratio of individual piers for the bridges
studied.

26

Figure 2-3 Slenderness ratio compared to tributary surface area of deck of individual piers for the bridges
studied.

27

Figure 2-4 Slenderness ratio compared to pier height of individual piers for the bridges studied.

28

Figure 2-5 Slenderness ratio compared to compressive strength of concrete used in individual piers of
bridges studied.

29

Figure 2-6 Slenderness ratio compared to visual slenderness ratio of individual piers of bridges studied.

30

Figure 2-7 Comparison between circular and rectangular pier cross sections for which the rectangular cross
section has a thickness h that is equivalent to the diameter of the circular cross section.

32

Figure 2-8 Visual slenderness comparison between the Big Qualicium, King's Highway No.II and ReussBrcke Wassen piers.

34

Figure 3-1 Concrete stress-strain models: (a) bilinear, (b) parabolic. Adapted from fib Model Code (2010). 39
Figure 3-2 Cross section and calculation parameters for equivalent rectangular stress block method.

40

Figure 3-3 Cross section and calculation parameters for bilinear stress-strain method.

41

Figure 3-4 M-N interaction envelopes based on bilinear stress-strain model and equivalent stress block
method.

42

Figure 3-5 Structural model for axially loaded columns with initial eccentricity proportional to buckled
shape of column

43

Figure 3-6 Effective length factor for various column configurations.

46

Figure 3-7 Structural model for axially loaded columns with constant initial eccentricity

47

Figure 3-8 Load-deformation response comparison for cantilever columns with variable slenderness ratios
using different analytical approaches

49

Figure 3-9 Strain profiles: (a) crushing of extreme compressive fibre; (b) reduced state of strain - steel
yielding; (c) concrete cracking in extreme tensile fibre.

ix

51

Figure 3-10 Sectional capacity and moment curvature diagrams of pier columns under variable axial loads.52
Figure 3-11 Experimental setup configuration. Adapted from Baant and Kwon (1994).

57

Figure 3-12 Measured peak loads and mid-height deformations as reported by Baant and Kwon (1994). 59
Figure 3-13 Structural model for an axially loaded cantilever column with a horizontal point load.

60

Figure 3-14 Assumed curvature for virtual work calculation in rigorous analytical method.

61

Figure 3-15 M-N interaction envelope and column cross section

62

Figure 3-16 Lateral load limits and imposed lateral load deformation limits as per Menn's method and the
proposed rigorous analytical method.

64

Figure 3-17 Relationship between flexural stiffness and axial force.

65

Figure 3-18 Lateral load limits and imposed lateral deformation limits as per Menns method and the
proposed rigorous analytical method for a column with a slenderness ratio of 60.

66

Figure 3-19 (a) Lateral load limits and imposed lateral deformation limits as per Menns method and the
proposed rigorous analytical method for a column with slenderness ratio equal to 100, (b) flexural
stiffness as a function of axial force in column.

68

Figure 3-20 Lateral load limits and imposed lateral deformation limits as per Menns method and the
proposed rigorous analytical method for a column with a slenderness ratio equal to 140.
Figure 3-21 Failure modes as per Menn's method and the proposed rigorous analysis method.

69
71

Figure 3-22 Analysis method comparison: (a) applied lateral load limit , (b) imposed lateral deformation
limit.

72

Figure 3-23 Error in Menn's method in predicting maximum lateral load and maximum lateral imposed
deformation for a column with a slenderness ratio of 140.

76

Figure 3-24 Modulus of elasticity of concrete as a function of compressive strength of concrete.

78

Figure 3-25 M-N interaction envelopes for varying values of concrete strength.

79

Figure 3-26 Limits of recommended use for Menn's method.

82

Figure 4-1 Plan view of possible wind load conditions applied to a pier.

87

Figure 4-2 Statical models for free-standing pier loading conditions.

91

Figure 4-3 Statical model for the calculation of second-order effects in a free-standing column.

92

Figure 4-4 Pier eccentricity limits as per CSA A23.1 (2004) and Menn (1990).

94

Figure 4-5 Pier eccentricity limits as per CSA A23.1 (2004) and SiA 162 (1989).

95

Figure 4-6 M-N interaction diagram cut off at 0.75Acfc'.

96

Figure 4-7 Geometric reinforcement ratio range considered in scope of thesis.

98

Figure 4-8 Normalized M-N interaction envelopes and normalized inferred eccentricity.

99

Figure 4-9 Inferred initial eccentricity w0,min as a function of cross-sectional thickness h.

100

Figure 4-10 Inferred initial eccentricity compared to eccentricity recommended by Menn (1990).

101

Figure 4-11 Serviceability limit states analysis critical cross section.

103

Figure 4-12 Sectional configurations: (a) 5% centroidal clear cover, (b) 10% centroidal clear cover.

104

Figure 4-13 Maximum average crack spacing as a function of area ratio as per CHBDC 2006.

106

Figure 4-14 Maximum tensile stress in extreme tensile fibre at prescribed cracked condition compared to
tensile strength of concrete.

109

Figure 4-15 Serviceability limit states state of strain: (a) prior to crushing of extreme compressive fibre, (b)
after crushing of extreme compressive fibre.

112

Figure 4-16 Various M-N interaction envelopes based on limiting states of strain.

114

Figure 4-17 Serviceability limit states design aid example: (a) only first-order considerations, (b) first and
second-order considerations.

116

Figure 4-18 Serviceability limit states first-order model.

116

Figure 4-19 Serviceability limit states second-order statical model.

118

Figure 4-20 Serviceability limit states design aids.

120

Figure 4-21 factor for pier end-conditions: (a) fixed-fixed, (b) pin-fixed, (c) pin-pin. Adapted from Menn
(1990).

123

Figure 4-22 Ultimate limit states design aid example.

124

Figure 4-23 Design aids for ultimate limit states 1.

127

Figure 4-24 Design aids for ultimate limit states 2.

128

Figure 5-1 Pier type and system.

132

Figure 5-2 Two-pier system model with loading conditions and assumed deformations.

136

Figure 5-3 Buckled shapes of pier types considered.

137

Figure 5-4 Two pier system statical model.

138

xi

Figure 5-5 Spring model for a two pier system.

141

Figure 5-6 Two pier system schematic and calculated system buckling loads based on various methods. 144
Figure 5-7 Illustrative representation of SAP2000 model.

147

Figure 5-8 Error in Menn's method compared to standard deviation in system buckling load efficiency of a
pier system.

160

xii

LIST OF TABLES
Table 2-1 General information and structural illustrations of bridge pier systems.

15

Table 2-2 Cross sections of bents used in the 22 bridges studied.

23

Table 2-3 Comparison between Big Qualicium and King's Highway No.II piers.

31

Table 2-4 Comparison between Big Qualicium, King's Highway No.II and Reuss-Brcke Wassen piers.

33

Table 2-5 Range of relevant design parameters.

36

Table 3-1 Bilinear stress-strain design values. Adapted from fib Model Code 2010.

39

Table 3-2 Sectional capacity based on bilinear stress-strain relationship and equivalent rectangular stress
block method.

41

Table 3-3 Measured peak loads and mid-height deformations as reported by Baant and Kwon (1994).

58

Table 4-1 Wind load property values.

88

Table 4-2 Pier construction tolerances as prescribed by CSA A23.1.

94

Table 4-3 Slenderness ratio limits for serviceability limit states based on mechanical reinforcement ratio. 121
Table 5-1 Considered mechanical reinforcement ratios and their associated flexural stiffnesses EI.

148

Table 5-2 System buckling load analysis for parametric study case 1.

150

Table 5-3 System buckling load analysis for parametric study case 2.

152

Table 5-4 System buckling load analysis for parametric study case 3.

154

Table 5-5 System buckling load analysis for parametric study case 4.

156

Table 5-6 System buckling load analysis for parametric study case 5.

158

xiii

LIST OF SYMBOLS

1
1
c
c'
c,peak
s
sm
sy

cr
y
c
s
E

sect
s,allow

As
Asb
Ast
b
Ce
Cg
Ch
c
db
dt
E
EI
EIc
EIg
EIy
e
Fh
fc'
fcr
ft
ft,max
HW
Ht
h
h'
I

scaling factor for axial load in multiple pier system parametric study
equivalent stress block stress factor
equivalent stress block lever arm factor
factor used to calculate wcr, accounting for loading conditions
crushing strain strain at crushing of extreme compressive fibre
peak strain minimum strain at which fc occurs (bilinear model)
strain in steel reinforcement
average strain in tensile reinforcement
yield strain strain required to yield reinforcement
curvature
curvature at first cracking of concrete section
curvature at first yielding of a reinforcement layer
material resistance factor for concrete (0.75)
material resistance factor for steel (0.90)
factor of safety against buckling of system
slenderness ratio
factor accounting for end conditions of a pier
geometric reinforcement ratio
sectional density
allowable stress in tensile reinforcement
mechanical reinforcement ratio
total cross-sectional area of steel reinforcement
area of steel reinforcement in bottom cross-sectional layer (tensile)
area of steel reinforcement in top cross-sectional layer (compressive)
width of cross section
wind exposure coefficient
gust exposure coefficient
horizontal drag coefficient
compressive depth of cross section
nominal reinforcement bar diameter
tensile depth of cross section
youngs modulus of concrete
flexural stiffness
secant stiffness at first crushing of extreme compressive fibre
gross cross sectional flexural stiffness
reduced flexural stiffness
constant initial eccentricity of axial load
wind load per unit exposed frontal area of structure
compressive strength of concrete
cracking strength of concrete maximum stress before cracking occurs
strain in extreme tensile fibre of section
maximum stress in extreme tensile fibre of section
horizontal wind load linear resultant force
sum of horizontal forces transferred from superstructure to piers
thickness of cross section
distance between centroid of reinforcement layer and external face of section
moment of intertia
xiv

k
kb
kc
keff
kindi
L
M
MB
Mcr
Me
MR
MSLS
My
M*
m
mR
m*
Ncr,max
NR
N*
n
nE
nR
n*
pc
Q
QE
QE
q
qw
r
srm
t%
uG
u1G
Wself
w
wmax
wT
w0
w0,max
wtot
wtot,max
wcr

effective length factor


factor used to calculate wcr, accounting for type of coating on reinforcement
factor used to calculate srm (chbdc 2006)
effective spring stiffness
effective length factor of individual pier in system (parametric study factor)
height of pier
moment
moment at base of pier
cracking moment moment required to crack concrete cross section
moment due to axial load with constant eccentricity
moment capacity of section
serviceability limit states moment capacity
yield moment moment required to induce yielding of a reinforcement layer
moment demand
normalized moment
normalized moment capacity
normalized moment demand
maximum axial force under equilibrated conditions allowing for cracking
axial force capacity of section
axial load demand
normalized axial load
normalized euler buckling load
normalized axial force capacity
normalized axial load demand
ratio of area of steel provided for tension to area of concrete in tension
applied axial load
euler buckling load
reduced euler buckling load
reference wind pressure
uniformly distributed linear resultant of wind load
radius of gyration
average spacing of cracks
percent of area of cross section in tension
second-order longitudinal displacement of superstructure
first-order longitudinal displacement of superstructure
self weight
second-order lateral deformation of pier
maximum second-order later deformation of pier
lateral deformation at top of every pier in system
first-order lateral deformation of pier
maximum first-order lateral deformation of pier
total lateral deformation of pier
maximum total lateral deformation of pier
average crack width

xv

Chapter 1. Introduction
The primary goal of this thesis is to develop and validate a method that will allow designers to
efficiently design reinforced concrete bridge piers that are economic as well as aesthetically
pleasing. In order to optimize the use of materials the recommended pier designs will almost
certainly result in slender pier systems. When designing slender compression members, such as
piers, second-order effects of geometric and material non-linearity must be accounted for. In
order to create design parameters that are easier to manipulate, justifiable and conservative
assumptions have been made. These assumptions have been identified and documented in this
thesis. To best achieve the presented goals, this thesis consists of five parts: (1) an empirical study
of 22 reinforced concrete highway bridge piers designed in the industry over the last 50 years, (2)
a summary and validation of assumptions and simplifications made in design and analysis
procedures, including stress-strain formulations, (3) identification of relationships between
design parameters associated with slender reinforced concrete bridge piers, (4) assessment and
recommendations for individual reinforced concrete bridge pier designs at serviceability limit
states and ultimate limit states, as per requirements of the CAN/CSA-S6-06 Canadian Highway
Bridge Design Code (herein CHBDC 2006), and (5) design recommendations for preliminary
design and detailing of bridge pier systems.

1.1

Motivation

Reinforced concrete highway bridge piers and abutments comprise approximately 7% of the total
cost of a bridge structure; as such an appropriate pier design can greatly increase the
transparency of an overall bridge structure, intrinsically improving the aesthetic value of a
bridge, while having minimal impact on the overall cost (Menn 1990). Although slender pier
systems typically have lower material costs associated with them, in comparison to stockier pier

2
systems, they are seldom seen in North American highway bridge designs. As pier systems
become more slender, the complexity of designing these systems increases accordingly; this is
due to the influence of second-order effects. Piers are generally regarded as structures that
predominantly carry loads in axial compression. Stocky, or squat, reinforced concrete members
that are carrying compressive loads typically fail due to concrete crushing in the extreme
compressive fibre. Slender piers tend to deviate in this regard; they are susceptible to flexural
failure that may either be due to crushing of the extreme compressive fibre or instability of the
pier. Instability is a result of increasing second-order deformations and buckling effects resulting
from geometric and material nonlinearities (Baant and Kwon 1994). As such the design of
slender piers must strictly account for flexural failures due to the influence of second-order
effects.
Typical span-to-depth ratios for highway bridge superstructures have been identified extensively
in literature. These ratios have proven invaluable to designers in the critical preliminary design
stages which can often influence the type of superstructure that is ultimately used for a given
highway bridge design. Research completed to date has not yet adequately identified comparable
ratios in terms of the preliminary design stages of reinforced concrete highway bridge piers.
Furthermore the governing document used for highway bridge design in Canada, The CHBDC
2006, does not adequately address the design of reinforced concrete highway bridge piers with
slenderness ratios exceeding one-hundred.
It is arguable that slender pier systems should be designed whenever possible due to the intrinsic
aesthetic value and economic benefits associated with reductions in material costs. Well
designed slender piers can be used to provide flexibility to a system thus allowing temperature,
shrinkage and creep effects to be transmitted to the bridge abutments without the need for
bearings or intermediate expansion joints in the deck, all of which have intrinsic maintenance

3
costs associated with them (Menn 1990). Although there are evident benefits to designing slender
reinforced concrete bridge pier systems they are seldom seen in North American bridge designs
(Poston 1986).
Two factors that can arguably be attributed to controlling the design of reinforced concrete
highway bridge piers in North America are industry standardized designs and equipment, and
topography. A large portion of the pier designs in Canada, especially circular columns, are
typically designed based on an industry standard. Pre-fabricated industry standard steel forms
are often used for the construction of these columns. Both of these standardizations take an
element of control away from the designer and greatly constrict the extent of possible designs.
Slenderness ratio quantifies a column's mechanical slenderness (Barrera 2011); it is a
measurement relating the effective length of the column to the radius of gyration of the column
(Baant et al. 1991). The effective length is dependent on the free-standing length of a column
and its end fixity (Cranston 1972). The radius of gyration is a function of cross-sectional
dimensions and is dependent on the moment of inertia of the cross section and area of the cross
section (Rangan 1990). For columns of a rectangular cross section, slenderness ratio can be
directly equated as a relationship between effective length and thickness (Wong 1983). In any
case, the value of the slenderness ratio is linearly proportional to the effective length of the
column (Kwak and Kim 2007). With typical bridge designs, the pier heights are controlled by the
topography of the location and the required superstructure elevation. Thus it is reasonable to
assume that topography, an element that is outside of the designer's control, may in fact dictate
the length of the pier that is designed. If the pier length is indeed predominantly controlled by
topography, then the only control designers have of slenderness ratios comes through crosssectional dimensions and end connections. In terms of aesthetic value, the end connections have

4
little to no impact, thus the fundamental question to be answered is: "What are appropriate
cross-sectional dimensions for efficient pier designs"?

1.2

Methods of column design and analysis

In this section, existing methods of column design and analysis that have been documented in
literature are reviewed. Aspects of this body of knowledge that are currently under-researched
are identified and discussed. This thesis addresses these under-researched aspects.

1.2.1 Numerical methods


A commonly adopted approach towards modelling the behavior of slender reinforced concrete
columns is to use numerical methods to predict column load-deformation behavior based on
compatibility criterion.
Chuang and Kong (1998) proposed a numerical method for analyzing pin-ended columns
subjected to uniaxial bending. Chuang and Kong (1998) adopted a transformation concept, where
steel areas in compression and tension are replaced by equivalent concrete areas, in conjunction
with the load-deformation curve of a column to predict the failure load. The advantage of the
proposed method, over other numerical methods, extends from the use of the transformation
concept. Najami and Tayem (1993) showed that the application of the transformation concept
extended beyond simple elastic analysis, but was also applicable to inelastic section analysis; this
was done by transforming the concrete and steel, based on their respective secant moduli, thus
resulting in an equivalent homogeneous linearly elastic material. Chuang and Kong (1998)
compared their presented method against 84 empirically tested columns presented in 7 different
research papers. All of the columns were pin-ended and subjected to uniaxial bending by
applying an increasing axial load at a fixed eccentricity. The method proposed by Chuang and
Kong in its nature is iterative, as each successive increase in load must be calculated for until a

5
peak load is reached. The method accurately predicts the failure load of a slender column, with a
mean error within two percent of experimental results. The failure loads predicted by the method
are very close to the prediction of the P-Delta method that is commonly accepted, and were
found to be generally closer to experimental data results than predictions by the ACI Committee
318 (1995) and the British Standards Institution (2008).

1.2.2 Graphical methods


A less common approach towards modelling the behavior of slender reinforced concrete columns
is to use graphical methods. Graphical methods tend to increase freedom in design parameter
selection but are more computationally intensive, at least at the preliminary stages, than
numerical methods.
Beal (1995) presented a graphical method for calculating the peak load of slender eccentrically
loaded pin-ended rectangular columns. In his method he would calculate and plot load
eccentricity against section curvature for varying applied axial loads. The eccentricity and
curvature of a given column were normalized with respect to the thickness of section and the
applied axial load was normalized with respect to the maximum compressive load of the column.
A second set of lines was drawn to plot buckling deformation against section curvature for
varying effective slenderness ratios. The buckling deformation and curvature were normalized
with respect to the thickness of the section and the effective slenderness ratio was calculated as
the effective length of the column divided by the thickness of the column. For a rectangular cross
section, this can be directly translated to the conventional slenderness ratio parameter defined as
effective length divided by radius of gyration through the application of a scaling factor. Beal
(1995) proposed that in order for equilibrium to be achieved, the load eccentricity generated by
the internal stresses in a column must equal the eccentricity of the applied load, which in an
axially loaded column would be equal to the buckling deformation. Beal (1986) further proposed

6
that if the lines corresponding to buckling deformation for different effective slenderness ratios
are overlaid on the set of lines defining a section's behavior based on applied axial load, then the
points of tangential intersection correspond to failure load for a column of a given rectangular
cross section and effective slenderness ratio. Furthermore, if a column has an initial eccentricity,
possibly resultant of construction imperfection or applied moments, then the lines
corresponding to the buckling deformation of the column may be translated along the vertical
axis accordingly and a new set of failure loads may be identified. The plot presented by Beal
(1986), demonstrating his proposed graphical method is shown in Figure 1-1.

Figure 1-1. Graphical analysis of columns (L/h = 20) Adapted from Beal (1986).

The graphical method proposed by Beal (1986) offers several advantages over the numerical
method proposed by Chuang and Kong (1998). Once the initial curves are calculated, slenderness
parameters as well as initial eccentricities can easily be modified and a failure load can just as

7
easily be determined. The graphical method does however also have disadvantages. The
determination of failure load can only be determined as accurately as the resolution of the plots
permits. Also, the lines defining the section's behavior must be produced for finite increments of
axial load, thus the accuracy of failure load identification is limited by the size of load
increments. Presumably load increments can be reduced in order to offset the inaccuracies;
however, this would require the extensive computation of additional lines, which may offset the
merit of the initial increased efficiency.

1.2.3 Finite element methods


Finite element modelling is one of the fastest growing research topics in recent literature. Finite
element models, in terms of structural engineering, essentially involve the dissection of large
structural elements into many smaller elements and approximating the behavior of each
individual element by accounting for known boundary conditions, compatibility relations and
constitutive relations; typically this is accomplished through the use of shape functions or
interpolation functions (Bhavikati 2005)
Mancini et al. (1998) proposed a finite element model in order to simulate the increasing lateral
forces developed by the pier connection to the superstructure, and the transfer of forces to the
abutments. The model takes account of material and geometric nonlinearities through the use of
a companion numerical method. Mancini et al. (1998) concluded that taking into account the
effective restraint condition on top of a bridge pier significantly modifies the structural behavior
of the pier. It was also noted that the global safety factor of a bridge could be modified based on
the restraint conditions.
Kwak and Kim (2004) presented a finite element model that takes account of material
nonlinearity as well as geometric nonlinearity by using an initial stress matrix. Kwak and Kim
(2004) verified their model through comparison with previous analytical and experimental

8
studies. After their investigation, they reported two findings: (1) The use of high-strength
concrete in slender reinforced concrete columns is not as effective as in squat columns; and (2)
an increase in reinforcement ratio provides minimal increase in structural capacity of slender
reinforced concrete columns.

1.2.4 Other methods


With the exponentially increasing computational capacity of modern processors, the adaptation
of more computationally intensive methods towards modelling structures under complex loading
conditions has become ever more popular. Many of these methods have been adopted from other
fields of research where their necessity or applicability to specific projects is perhaps warranted.
Some of these modern methods may be more accurate than conventional methods; however, for
the purposes of practical design where construction errors and imperfections are not uncommon,
the complexity associated with the use of such methods may be enough reason to adopt simpler
methods. Some of the more frequently observed methods are discussed.
Poston (1986) proposed the use of a fibre model in order to accurately model second-order
effects associated with the structural behavior of slender reinforced concrete piers. Poston (1986)
more generally developed his method for any slender concrete member subjected to compressive
loads. The model takes into account P-delta effects, geometric nonlinearity, material
nonlinearity, and sustained load effects by using a modified linear formulation. Poston (1986)
used a stiffness method to calculate incremental displacement from applied loads, which were
then used to calculate incremental forces. The proposed fibre method is essentially a modified
finite element method. The difference between the method presented by Poston (1986) and
conventional finite element methods is the manner in which the individual elements are treated.
A conventional finite element method discretizes a larger complex body into many smaller
simplified bodies and attempts to model the overall behavior based on the effective sum of the

9
smaller bodies. The fibre model goes further to discretize the smaller elements into individual
fibres and model their individual behavior. The benefit associated with using the fibre method
comes from the fact that the fibres are two-dimensional elements that have a 2x2 stiffness matrix
associated with them, whereas a typical three-dimensional element would have a 3x3 stiffness
matrix. This presumably would decrease computational time. The disadvantage in the method
comes from the loss of transparency in the calculation process. By adding another level of
elements which eventually require assembly in order to describe the overall structural behavior,
it becomes even less obvious to a designer as to how the analysis is performed; this essentially
limits the level of control available during iterative design processes, potentially diminishing the
likelihood of arriving at a logical and efficient design.
Chuang et al. (1998) used a neural multilayer feedforward network to model the nonlinear
relationship between the numerous input parameters and actual ultimate capacity of a slender
reinforced concrete column. The premise of the neural network model is that input parameters
such as loads and dimensions are provided and fed into what is regarded as a hidden layer. Once
the input parameters are put in through the hidden layer they are used to calculate the output
parameters, which in this case may be such values as deformations or resulting second order
moments. The neural network method presented by Chuang et al. (1998) although interesting,
and perhaps innovative, suffers from some conspicuous drawbacks. The neural network method
doesn't actually provide any noticeable improvements in accuracy or efficiency over conventional
finite element method models. The method also has the inherent drawbacks associated with the
hidden layer. The hidden layer, as described by Chuang et al. (1998) is essentially a selfcalibrating set of equations. The hidden layer is the component of the neural network that is
trained through extensive input and output tests by providing already existing data to the neural
network. The problem with this type of model is that it is not certain what the governing

10
equations are inside the hidden layer and if they are truly valid for all sets of input parameters.
These equations are not derived on any basis of first principles but are simply fabricated
equations with the specific purpose of generating a specific set of output values given a specific
set of input values. The designer is thus effectively left at the mercy of this calibrated hidden
layer when assessing the validity of proposed designs.
Manzelli and Harik (1993) presented an approximate second-order hand calculation technique
for the analysis of cantilever compression members. The method is valid for both prismatic and
nonprismatic, members as well as members with varying longitudinal reinforcement. Manzelli
and Harik (1993) accounted for effects of creep, foundation rotation, and out-of-plumbness in
their method. The method involves the development of an approximated linear moment
curvature relationship for a given cross section and calculating the flexural stiffness of the
column given an applied axial load. Manzelli and Harik (1993) assumed a uniform curvature
distribution along the length of the column which they used to calculate the effective moment
capacity of the column. If a selected cross section and reinforcement ratio does not meet the
moment demand, iterations are necessary to select a new cross section or reinforcement ratio.
Although Manzelli and Harik (1993) proposed a method that allows engineers to effectively
design reinforced concrete bridge piers without the use of computer technology, the method is
still iterative and can easily become lengthy with more complicated pier geometry. Also, the
method does not provide any guidance to designers as to what the initial cross-sectional
dimensions should be or what an appropriate reinforcement ratio might be. Thus, apart from
introducing a simplified linearization of the nonlinear moment curvature relationship, this
method does little to improve upon already accepted reinforced concrete column design
methods, such as the widely accepted P-delta method.

11
Although the studies discussed provide several new and innovative ideas on how to analyze and
in some cases design slender reinforced concrete piers, they are somewhat limited in scope and
are usually only applicable to the modelling and analysis of such members. The data and
formulations presented in the studies are not helpful to designers who are interested in the
preliminary design of slender reinforced concrete bridge piers. As such, more simplified and
general methods or figures are needed in order to aid designers towards a more methodical
design approach. A more comprehensive study would include a thorough overview of slender
reinforced concrete compressive member behavior and overall system behavior. The need for a
generalized system in identifying appropriate member sizes and reinforcement ratios for
preliminary design is undeniable; presently available documents give no such guidance and as
such designers are typically left making decisions based on personal experiences, often resulting
in extensive iterations. Sectional and system capacities of slender reinforced concrete members
need to be properly addressed and quantified. The relationship between slenderness ratio,
reinforcement ratio, and column capacity needs to be identified; doing so would allow for the
documentation of adequate preliminary design recommendations. This thesis undertakes these
tasks with the intent of providing a document that designers may refer to for appropriate
preliminary design parameters associated with slender reinforced concrete bridge piers and
bridge pier systems.

1.3

Objectives and content of thesis

The primary objectives of this thesis are:


1.

to identify the influence of slenderness ratio, reinforcement ratio and strength of concrete
on the structural behavior and response of slender reinforced concrete bridge piers and
bridge pier systems,

12
2. to provide designers with guidance and preliminary design recommendations in terms of
slender reinforced concrete bridge piers and bridge pier systems,
The secondary objectives of this thesis, which support the primary objectives are:
3. to characterize the state-of-the-art of reinforced concrete bridge piers and bridge pier
systems,
4. to generalize the structural characteristics of a single bridge pier relative to a multiple
bridge pier system, and
5. to provide general recommendations related to aesthetics and constructability of slender
reinforced concrete bridge piers and bridge pier systems.
The content of this thesis encompasses the primary and secondary objectives presented above.
The content is discussed in logical sequence as pertained to design procedures, beginning with
material properties and stress-strain formulations, followed by approximate methods of
predicting structural response of slender reinforced concrete bridge piers, and ending with
design recommendations for slender reinforced concrete bridge piers and bridge pier systems.
Chapter 2 consists of a parametric design study of existing highway bridges in which designed
reinforced concrete piers are described. Pier height, slenderness ratio, and reinforcement ratio
are the primary parameters that are considered. Insight related to the efficiency, slenderness, and
aesthetic value are discussed and documented. In Chapter 3, the approximate bilinear stressstrain formulation for normal to medium strength concrete is presented and compared against
the equivalent stress block method presented in the CHBDC 2006. Further, Vianello's method of
successive approximations is discussed and related to the design of slender reinforced concrete
bridge piers as per Menn's method of reduced interaction diagrams. Menn's method is validated
by comparison with modified virtual work formulations and cross-validated by comparison with
reported laboratory results. In Chapter 4, the influence of axial load on the flexural stiffness for

13
varying reinforced concrete bridge piers of various reinforcement and slenderness ratios is
assessed and documented. Further, Menn's recommendations for slenderness ratio as a function
of reinforcement are reviewed, and appropriate serviceability limit states and ultimate limit
states preliminary design aids are developed. In Chapter 5, recommendations for preliminary
design of slender reinforced concrete multiple pier systems are made. The recommendations are
made based on considerations for the structural behavior of the system as a whole, rather than
behavior of individual piers. In the final chapter, the most important components and findings
from the preceding chapters are summarized. Potential areas of future work are addressed and
identified.
The most important contributions of this thesis are:

the development and conceptual validation of general preliminary design tools and
recommendations for the potential development of slender and efficient reinforced
concrete bridge pier systems.

a comprehensive comparative study of 22 existing highway bridges and their associated


reinforced concrete piers; the database serves as a characterization of the current stateof-the-art and as a valuable design tool that designers may reference.

The research presented in this thesis provides a means for designers to identify appropriate
preliminary design parameters for reinforced concrete bridge piers and pier systems. If optimal
design parameters are selected, the resulting pier designs will in almost all cases result in slender
piers, in which case second-order effects are a factor. The methods and design tools presented in
this thesis all take account of material and geometric nonlinearities which will almost certainly
be a controlling factor in the structural behavior of slender piers and slender pier systems.

14

Chapter 2. Review of Existing Bridge Piers


This chapter presents a study of 22 reinforced concrete bridges designed to date. The study
specifically looks at the design parameters associated with the piers. Section 2.1 summarizes the
general physical dimensions of the bridges. Section 2.2 discusses the specific design parameters
associated with the most critical pier and most slender pier in each system in order to identify
the current state-of-the-art of pier design. Section 2.3 compares relevant design parameters of
select bridges used in the study and discusses certain design choices and implications.

2.1

Study of bridge pier systems

Bridge pier systems consist of a series of piers that are interconnected through the bridge
superstructure. In such systems the piers collectively share their flexural stiffnesses. Generally the
distribution of flexural stiffness is such that the stiffer piers in a system provide additional
flexural stiffness to the more slender piers, at the expense of attracting more of the load.
Although the design of slender reinforced concrete bridge piers is not uncommon in many
continents around the world, such as Europe and Asia, the reinforced concrete highway bridge
design practice in North America has been shown to typically produce bulkier piers, seldom
exceeding slenderness ratios of more than 70 (Poston et al. 1986). Table 2-1 on the following page
summarizes the general physical characteristics and dimensions of the 22 bridge piers looked at
it in this study. Selected bridges have been broken down into component spans. These
separations have been done based on the location of expansion joints in the bridge
superstructure. Expansion joints in the superstructure effectively create a discontinuity, and thus
define the beginnings and ends of pier systems. Section 2.2 will discuss individual characteristics
of the piers in these systems, most specifically focused on identifying slenderness ratios and any
parameters that influence considerations for slenderness.

15

Table 2-1 General information and structural illustrations of bridge pier systems.
No.

Name and
Location

Year of
Design

Highway
No.401 & 2A
Interchange,
Canada

1970

Reuss-Brcke
Wassen,
Switzerland

1972

King's
Highway
No.2, Canada

1974

Structural Schematic

16

Islington
Avenue
Overpass,
Canada

1977

Highway 404
CNR
Overhead,
Canada

1979

CNR
Overhead
Highway
No.69,
Canada

1981

Turning
Roadway N.
to 409 E.,
Canada

1981

Highway 403
E.B. Express
over Highway
410 N.B.
Ramp S.W.,
Canada

1982

17

Highway 403
E.B. Express
over Highway
401 E.B.
Collector,
Canada

1982

10

King's
Highway 7
Underpass at
Dufferin
Street,
Canada

1983

11

Highway 401
Morningside
Ave.
Underpass,
Canada

1988

12

Highway 403
Upper
Middle Road
Underpass,
Canada

1990

13

Englehart
River Bridge
Highway 560
Crossing,
Canada

1990

14

Ramp 403/W
- QEW/E,
Canada

1990

18

15

State of
Hawaii,
Interstate
Route H-3
Windward
Viaduct,
United States

1990

16

Big
Qualicium
River Bridge
No. 3051 Steel
Alternative,
Canada

1995

19

17

I-93
Southbound
Viaduct
Concrete
Alternative,
United States

1996

18

Applewood
Crescent
Bridge,
Canada

2004

19

Caroni
Bridge,
Trinidad and
Tobago

2007

20

20

Pennsylvania
Turnpike
Commission
Expressway,
United States

1997

21

Ramp 401W
Collector 404N Over
Ramp 401W
Express,
Canada

2007

22

Shin Chon
Bridge, South
Korea

2007

21
Table 2-1 summarized the general information related to the bridges that were studied. The table
shows a structural schematic of the bridges in order to illustrate the overall characteristics of the
pier system. Pier heights and spans have been identified as well as the connection detailing at the
base of the piers and the connection detailing between the pier and superstructure. All of the
drawings which were illustrated in the preceding table are shown at a scale of 1:6000 on the
following page in Figure 2-1 in order to better illustrate the differences in sizes of these systems.
The general shapes of the bent cross sections are shown in Table 2-2 in order to illustrate the
types of cross sections that were used for the design of the studied piers. All cross sections have
been presented at a 1:250 scale.

22

Figure 2-1 Structural illustrations of the 22 studied bridges at a 1:6000 scale.

23
Table 2-2 Cross sections of bents used in the 22 bridges studied.
No.
1

Name
Highway No.401
& 2A
Interchange

Reuss-Brcke
Wassen

King's Highway
No.2

Islington Avenue
Overpass

Highway 404
CNR Overhead

CNR Overhead
Highway No.69

Turning
Roadway N. to
409 E.

10

11

Highway 403
E.B. Express over
Highway 410
N.B. Ramp S.W.
Highway 403
E.B. Express over
Highway 401
E.B. Collector
King's Highway
7 Underpass at
Dufferin Street
Highway 401
Morningside
Ave. Underpass

12

Highway 403
Upper Middle
Road Underpass

13

Englehart River
Bridge Highway
560 Crossing

14

Ramp 403/W QEW/E

Bent Cross Section

24

15

State of Hawaii,
Interstate Route
H-3 Windward
Viaduct

16

Big Qualicium
River Bridge No.
3051 Steel
Alternative

17

I-93 Southbound
Viaduct
Concrete
Alternative

18

Applewood
Crescent Bridge

19

Caroni Bridge

20

21

22

Pennsylvania
Turnpike
Commission
Expressway
Ramp 401W
Collector - 404N
Over Ramp
401W Express

Shin Chon
Bridge

Note: All bent cross sections are shown at a 1:250 scale.

25

2.2

Trends in Design Parameters

This section presents a comprehensive study of relevant design parameters of the 22 bridges
presented in the preceding section. These parameters are studied with the intent of identifying
particular trends that may be present in the design of highway bridge piers in the industry today.
Studying these design parameters will also provide for a better understanding of the present
state-of-the-art in reinforced concrete highway bridge pier design. All of the design parameters
that have been documented have been compared to the slenderness ratio of the bridge piers. In
order to keep the study unbiased towards particular bridges that may have more piers in their
system, only the most slender pier from each of the 22 bridges was documented. In all 22 bridges,
the pier with the greatest slenderness ratio was also the most critical as the tributary loaded
length on each pier of each bridge was very similar. The definition of slenderness ratio is
presented in Equation 2-1, where k is the effective length factor, L is the unsupported length of
the pier, and r is the radius of gyration of the pier.

Equation 2-1

The parameters that have been compared to the slenderness ratio of each pier were: (1) geometric
reinforcement ratio, (2) tributary loaded surface area of deck, (3) pier height, (4) concrete
strength, (5) and visual slenderness ratio. The parametric comparisons are made in subsections
2.2.1 through 2.2.5. The data points presented have been categorized based on the shapes of the
pier cross sections. Commentary on general findings is made following the presented figure in
each subsection. Tabulated data used to produce the plots can be found at the end of the thesis
in Appendix A.

26

2.2.1 Slenderness ratio against geometric reinforcement ratio

Figure 2-2 Slenderness ratio compared to geometric reinforcement ratio of individual piers for the bridges
studied.

Figure 2-2 shows the geometric reinforcement ratios of the most slender piers from each of the
bridges studies compared against their respective slenderness ratios . The piers range in
reinforcement ratio from a minimum of 0.4 to a maximum of approximately 4.7. The piers of the
Reuss-Brcke Wassen (Uri 1971) stand out as being exceptionally slender while having the
smallest reinforcement ratio amongst the 22 piers studied. Although it is recognized that 22 piers
may not be sufficient to establish industry trends, the collected data shows no discernible
correlation between reinforcement ratio and slenderness ratio in terms of the present state-ofthe art. Of the 22 piers studied, only 3 had geometric reinforcement ratios which significantly
exceeded 2%. Out of the 22 bridges studied, the four piers which exceeded a slenderness ratio
of 100 had reinforcement ratios that were 2.1% or lower, further suggesting that the merits of
added reinforcement are not capitalized upon in terms of designing slender piers.

27

2.2.2 Slenderness ratio against tributary surface area of deck

Figure 2-3 Slenderness ratio compared to tributary surface area of deck of individual piers for the bridges
studied.

Figure 2-3 shows the tributary surface area of the deck, in terms of loading, of each of the most
slender piers from the bridges studied, compared to the respective slenderness ratios of the
individual piers. The data suggests that there are no discernible trends between the two values
compared. This is contrary to what was expected, as a larger tributary surface area would imply a
larger dead load applied to the pier, which results in a more critical state for buckling failure. A
larger sample size of bridge piers may be required in order to validate the lack of correlation
between these two parameters in the present state-of-the-art.

28

2.2.3 Slenderness ratio against pier height

Figure 2-4 Slenderness ratio compared to pier height of individual piers for the bridges studied.

Figure 2-4 shows the heights of individual piers from the bridges studied compared to their
respective slenderness ratios . The primary intent of this particular comparison was to identify if
the slenderness ratios used in the industry are primarily driven by topographical needs. The
data does not show a confident correlation between the two parameters. A linear regression
would merit an R2 value of 0.307. This effectively implies that topographical conditions are not
necessarily a driving factor behind slenderness ratio in the present industry. If there truly is no
strong industry trend, it suggests that there may be room for improvement in terms of
conventional practises used for the design of reinforced concrete bridge piers. Further studies
should be done in order to better fill the gap in data between the tallest pier shown and the
shortest pier shown.

29

2.2.4 Slenderness ratio against compressive strength of concrete

Figure 2-5 Slenderness ratio compared to compressive strength of concrete used in individual piers of
bridges studied.

Figure 2-5 shows the compressive strength of concrete fc' used for the construction of individual
piers of the bridges studied compared against the individual slenderness ratios of the respective
piers. The primary purposes of this study were to determine whether or not concrete strength fc
is correlated with slenderness ratio in the present industry and to identify a reasonable range of
concrete strengths fc used to construct piers in the present industry. The minimum concrete
strength fc used amongst the 22 piers studied was 25 MPa, whilst the maximum was 50 MPa. The
25 MPa concrete was used in the construction of the Reuss-Brcke Wassen piers (Uri 1971), one
of the most slender amongst the 22 recorded. Although no discernible trend can be identified
between concrete strength fc and slenderness ratio , the data does suggest that concrete
strength fc does not provide noticeable benefit in terms of constructability of slender piers.

30

2.2.5 Slenderness ratio against visual slenderness ratio

Figure 2-6 Slenderness ratio compared to visual slenderness ratio of individual piers of bridges studied.

Figure 2-6 shows a comparison between the visual slenderness ratio L/h and slenderness ratio
of individual piers from the bridges studied. The visual slenderness ratio L/h is defined as the
ratio of pier height to its thickness in the longitudinal direction of the superstructure. The visual
slenderness provides a better means of quantifying the aesthetic slenderness of a pier, whereas
the slenderness ratio provides a means of measuring a piers susceptibility to a buckling failure.
A weak positive correlation exists between the two parameters. Since the two parameters do not
have a perfect linear relationship, it is implicit that the visual slenderness ratio L/h is to some
extent independent of the slenderness ratio . The limited independence of these two parameters
implies that aesthetic slenderness can be separate from mechanical slenderness. Section 2.3
discusses some specific characteristics of the four bridges identified in Figure 2-6 in order to
address design decisions that can be made, allowing for a greater visual slenderness ratio L/h,
while maintaining the same slenderness ratio .

31

2.3

Comparison of select bridges from study

This section compares and discusses the four bridges identified in Figure 2-6 from the preceding
section. The Big Qualicium (BC 1995), Kings Highway No.II (MTO 1974), Reuss-Brcke Wassen
(Uri 1971), and Shin Chong Bridge (Starossek 2009) all had specific characteristics associated with
the design of the piers which lead to a significantly different range in visual slenderness for very
similar values of slenderness ratio. It is the intent of this comparison to identify these specific
design choices such that designers may use and consider them when making preliminary design
decisions.

2.3.1 Big Qualicium compared to Kings Highway No.II


The Big Qualicium (BC 1995) and Kings Highway No.II (MTO 1974) piers have very similar
slenderness ratios ; respectively they are 120.7 and 121.6. The visual slenderness ratios or height
over thickness of the two piers are notably different; respectively they are 12.09 and 17.54. This
information is presented in Table 2-3 below.
Table 2-3 Comparison between Big Qualicium and King's Highway No.II piers.

Name
Big Qualicium
Kings Highway No.II

Slenderness Ratio
120.7
121.6

Visual Slenderness Ratio L/h


12.09
17.54

Both bridge piers have the same effective length factor k of 2.0; as such this is not the factor that
would differentiate the two. The difference is due to the shapes of cross sections. The Big
Qualicium (BC 1995) has a pier cross section that consists of two circular columns, whereas
Kings Highway No.II (MTO 1974) consists of a single rectangular cross section. The pier heights
are of comparable lengths, with the Big Qualicium (BC 1995) pier having a height of 30.2m and
the Kings Highway No.II (MTO 1974) pier having a height of 24.1 m. Given the same crosssectional thickness, a circular cross section will generally have a lower value for radius of gyration

32
r than a rectangular pier will. This is because a rectangle has more of its area distributed further
away from its geometric centroid and inherently will have a higher value of moment of inertia I
than a circle with the same thickness. This effectively implies that the radius of gyration of the
rectangular section will also be larger; as such inherently it will have a smaller slenderness ratio
given the same height of pier and the same pier to superstructure connection. Of course the area
of the circular section will be smaller than the rectangular cross section; however, increasing the
area of the circle would effectively require that the diameter and thus thickness of the pier be
increased, leading to a visually less slender pier. Figure 2-7 below shows the relationship between
visual slenderness ratio L/h and slenderness ratio for a pier with a rectangular cross section and
pier bents consisting of one to three circular cross sections arranged with their individual
centroids lying on the same axis. The plot was produced assuming an effective length factor k of
2.0 and a pier length of 32 m. The plot clearly demonstrates that for the same value of
slenderness ratio piers with circular cross sections will visually appear to be less slender. Also
this effect becomes more prominent as more circular piers are incorporated in the single pier
bent. This correlation is independent of the aspect ratio of the rectangular section.

Figure 2-7 Comparison between circular and rectangular pier cross sections for which the rectangular cross
section has a thickness h that is equivalent to the diameter of the circular cross section.

33
The comparison between the two pier cross sections has demonstrated that generally a pier with
a rectangular cross section can be designed to visually appear more slender than a pier with a
circular cross section given the same slenderness ratio .

2.3.2 Reuss-Brcke Wassen


Much like the Big Qualicium (BC 1995) and the Kings Highway No.II (MTO 1974), the ReussBrcke Wassen (Uri 1971) also has a very slender pier in its system. The tallest and most slender
pier has a slenderness ratio of 100.8. The pier is of similar height to the two other bridges piers,
with a height of 32.0 m. Visually, the pier appears to be much more slender than either of the
piers mentioned for the Big Qualicium (BC 1995) and Kings Highway No.II (MTO 1974). The
Reuss-Brcke Wassen (Uri 1971) pier has a visual slenderness ratio L/h of 29.09. A summary of
the slenderness ratios and visual slenderness ratios L/h for the three piers can be found under
Table 2-4.
Table 2-4 Comparison between Big Qualicium, King's Highway No.II and Reuss-Brcke Wassen piers.

Name
Big Qualicium
Kings Highway No.II
Reuss-Brcke Wassen

Slenderness Ratio
120.7
121.6
100.8

Visual Slenderness Ratio L/h


12.09
17.54
29.09

The large discrepancy in visual slenderness in this case is attributed to the connection between
the superstructure and the top of the pier. Both the Big Qualicium (BC 1995) and Kings Highway
No.II (MTO 1974) had hinged connections between the pier and superstructure, thus having an
effective length factor k of 2.0. The Reuss-Brcke Wassen (Uri 1971) is monolithically connected
to the superstructure, thus having an effective length factor k of 1.0. The reduced effective length
factor k allows for much greater visual slenderness ratios L/h to be achieved while retaining the
same slenderness ratio . Figure 2-8 on the following page shows a 1:300 scale drawing of the

34
profiles of the three piers discussed. The figure serves to demonstrate the differences in the
aesthetic and thus visual slenderness ratios L/h of the three piers relative to their very different
slenderness ratios .

Figure 2-8 Visual slenderness comparison between the Big Qualicium, King's Highway No.II and ReussBrcke Wassen piers.

Figure 2-8 demonstrates the aesthetic merits that can be attained by designing for a monolithic
connection between the pier and superstructure. Apart from aesthetic merits, economic merits
can also be attained, since monolithic connections eliminate the need for bearings which require
regular maintenance.
The comparison made with the Reuss-Brcke Wassen (Uri 1971) has demonstrated the general
aesthetic merits that can be attained by creating a monolithic connection. Reducing the effective

35
length factor k from 2.0 to 1.0 allows for the design of piers that appear significantly more slender
while retaining lower slenderness ratios .

2.3.3 Shin Chon Bridge


The last of the four bridge piers to be discussed is the tallest pier in the Shin Chon Bridge
(Starossek 2009). Of the four piers, it has the lowest slenderness ratio , with a value of 86.4;
however, it has the largest visual slenderness ratio with a value of 31.6. The designers were able to
attain such a large visual slenderness effectively by capitalizing on the two design principles
discussed in subsections 2.3.1 and 2.3.2. The pier cross section is of a hollow rectangular shape. A
hollow rectangle when compared to a solid rectangle works in the same way as a rectangle
compared to a circle. With a hollow rectangle, a larger portion of its cross sectional area is
further away from the centroid, as such, given the same cross sectional area, it has a larger radius
of gyration than a solid rectangle. Effectively this allows for lower values of slenderness ratio to
be attained for the same visual slenderness ratio L/h. In principle this works in the same way as a
comparison made between piers of rectangular and circular cross sections.
The pier was also monolithically connected to the superstructure, giving it an effective length
factor k of 1.0. The merit of attaining a lower value of effective length factor k through the use of
monolithic connections was discussed in subsection 2.3.2.
The general findings made from the preceding bridge pier comparisons can essentially be
summarized into two findings. These findings may prove to be good rules of thumb for designers
who seek to design piers that are slender and aesthetically pleasing. First, piers can be designed
to appear visually more slender by using cross sections for which more of the area is further away
from the cross sections geometric centroid, thus ensuring lower values for radius of gyration.
Second, by using monolithic connections the effective length factor k of a pier can be reduced,
thereby allowing for the design of a pier that visually appears more slender.

36
The following section summarizes the ranges of design parameters that were identified from the
22 piers studied. The range in parameters will be used to establish the scope of further chapters.

2.4

Range of design parameters

Table 2-5 summarizes the range in relevant design parameters, as identified in Chapter 1 of this
thesis, collected from the study of 22 existing bridge piers in order to identify the scope of
reasonable parameters to be considered in the remainder of this thesis.
Table 2-5 Range of relevant design parameters.

Design Parameter
concrete strength, fc
reinforcement ratio,
slenderness ratio,

2.5

Minimum Value Recorded


25 MPa
0.40 %
16.4

Maximum Value Recorded


50 MPa
4.71 %
121.6

Concluding remarks

This chapter presented 22 bridges and identified notable design parameters associated with the
most critical and, in the case of the 22 bridges studied, most slender piers. Section 2.1
summarized the general characteristics of the 22 bridges, such as span lengths, pier heights, and
end connections. Section 2.2 identified design trends in the present industry. From the
parameters compared, one of particular interest was the relationship between slenderness ratio
and visual slenderness ratio L/h. It was found that the two parameters could vary significantly,
thus implying that the aesthetic slenderness of a pier can be separated from the physical or
mechanical slenderness. Four bridges of the 22 were of particular interest: (1) The Big Qualicium,
(2) Kings Highway No.II, (3) Reuss-Brcke Wassen, and (4) Shin Chon Bridge. The four bridges
had piers with very different visual slenderness ratios L/h, but very similar slenderness ratios .
By studying these four bridges and making comparisons it was demonstrated that piers could be
designed to avoid slenderness effects such as buckling, but still retain an aesthetic or visual
slenderness. The two methods of accomplishing this were found to be the use of cross sections

37
with higher radii of gyration and through the use of monolithic connections which reduced the
effective length factor k. Lastly, the range in concrete strength fc, reinforcement ratio , and
slenderness ratio , amongst the 22 bridge piers studied was used as a representation of what may
be reasonable in the current state-of-the-art, helping establish the scope of further chapters.

38

Chapter 3. Approximate Methods


The purpose of this chapter is to present the approximate methods and their underlying
assumptions, on which further computations in this thesis are fundamentally based on. The
methods are validated through comparison with conventionally accepted design methods. A
bilinear stress-strain approximation for concrete is compared against the conventionally
accepted 1-1 equivalent stress block approximation. Vianello's method of successive
approximations is used to calculate second-order deformations of slender reinforced concrete
bridge piers; the derivation and limitations of the method are discussed. Menn's method of using
reduced states of strain in order to conservatively analyze slender reinforced concrete bridge
piers is presented and discussed. The failure loads of four hypothetical slender reinforced
concrete bridge piers are calculated using Menn's method and compared against the failure loads
calculated using method of virtual work.

3.1

Bilinear stress-strain formulation

The simplified bilinear stress-strain formulation used for all of the calculations presented in this
thesis was presented in the fib Model Code 2010. Since the purpose of the calculations presented
in this thesis is to provide designers with guidance for the preliminary design of reinforced
concrete bridge piers, the conservative simplification of assuming no tensile strength of concrete
has been made. The bilinear model simplifies the known parabolic stress-strain behavior of
conventional concrete by presenting appropriate values of peak stress value fc', the strain at
which peak stress is reached c,peak, and the maximum strain at which crushing of the concrete in
the extreme compressive fibre occurs c'.
A comparison between the bilinear stress-strain model, presented in the fib Model Code 2010, and
a conventional parabolic stress-strain relationship is presented in Figure 3-1. The figure was
adapted from the fib Model Code 2010.

39

Figure 3-1 Concrete stress-strain models: (a) bilinear, (b) parabolic. Adapted from fib Model Code (2010).

The intent of the bilinear model presented in Figure 3-1 (a) is to simplify the calculations
required to determine sectional forces based on strain profiles, while still providing a reasonable
degree of accuracy relative to the parabolic stress-strain model presented in Figure 3-1 (b).
The design values corresponding to the peak strain c,peak, the crushing strain c', and cylinder test
concrete strength fc' are presented in Table 3-1. The values presented are as shown in the fib
Model Code 2010.
Table 3-1 Bilinear stress-strain design values. Adapted from fib Model Code 2010.

Design
Property
fc' (MPa)
c,peak (%)

20
1.75

c' (%)

3.5

25

30

35

40

Design
Value
45
50

55
1.8

60
1.9

70
2.0

80
2.2

90
2.3

3.1

2.9

2.7

2.6

2.6

The next subsection serves to make a comparison between results that would be obtained using
the 1-1 equivalent stress block method and the proposed bilinear model. The intent is to
validate the bilinear approximation for the purposes of design.

40

3.1.1 Stress-strain model calculation comparison


The following subsection serves to make comparison between the bilinear stress-strain
relationship presented in the fib Model Code 2010 and the equivalent rectangular stress block
method presented in the CHBDC 2006. The equivalent rectangular stress method simplifies the
calculation of resultant forces by converting the integral of the parabolic stresses to a rectangular
stress block centered at the centroid of the parabolic stress distribution. A calculation
comparison is made between the two methods for a given cross section, given an assumed
ultimate limit states strain distribution. The cross-section, parameters, assumed strain
distribution and equivalent stress block for the equivalent rectangular stress block method are
presented in Figure 3-2. The same strain distribution was used for the purposes of the bilinear
stress-strain relationship; however, a different resultant stress distribution was attained. The
same representative diagram, for the bilinear stress-strain relationship is shown in Figure 3-3.
The equations for calculating 1 and 1 are presented in Equation 3-1 below.

Equation 3-1

Figure 3-2 Cross section and calculation parameters for equivalent rectangular stress block method.

41

Figure 3-3 Cross section and calculation parameters for bilinear stress-strain method.

The sectional capacity in terms of moment MR and axial load NR based on the two methods is
shown in Table 3-2. The proposed bilinear stress-strain relationship is shown to be in good
agreement with the results attained using equivalent rectangular stress block method. The
moment capacity MR is over-predicted by roughly 13% with the bilinear relationship while the
axial capacity NR is over-predicted by only 0.02%. This result suggests that the stress resultant
lever arms from the bilinear stress-strain relationship differ from what is suggested by the
equivalent rectangular stress block method, which is a function of the compression depth c and
the factor 1. No code material resistance factors i were used for the purposes of these
demonstrational calculations.
Table 3-2 Sectional capacity based on bilinear stress-strain relationship and equivalent rectangular stress
block method.

Design
Property
MR (kNm)
NR (kN)

Calculation Method
Bilinear
13 500
33 100

1-1
12 000
32 600

Error
(%)
12.50
0.02

In order to fully identify the differences in results obtained using the equivalent rectangular
stress block method and the bilinear model, a full M-N interaction diagram for the column cross

42
section presented in Figures 3-2 and 3-3 was calculated. The M-N interaction diagram is
presented in Figure 3-4 below.

Figure 3-4 M-N interaction envelopes based on bilinear stress-strain model and equivalent stress block
method.

Figure 3-4 shows that generally the equivalent rectangular stress block method calculates an M-N
interaction diagram with lower values than the bilinear model. The M-N interaction diagram
corresponding to the equivalent rectangular stress block calculation has been cut off at the
0.75Acfc as prescribed in CHBDC 2006. The conservative results predicted are to be expected.
Ibrahim and MacGregor (1997) showed that at the time of writing their paper the 1 and 1 factors
used by the ACI and CSA codes were generally unconservative for higher strength concretes, and
were only suitable for concrete with strength fc below 50 MPa. The present day 1 and 1 code
equations are similar to those proposed by Ibrahim and MacGregor (1997) and are intended to be
conservative for higher strength concretes. As direct consequence of the modern 1 and 1 code
equations, the M-N interaction diagram generated for lower strength concretes, such as the 25
MPa shown in Figure 3-4. With this comparison, the use of the bilinear stress-strain relationship
for the purposes of preliminary design has been validated.

43

3.2

Vianello's method

Italian engineer Vianello (1898) proposed a method for calculating second-order deformations,
Vianellos method of successive approximations. The method is widely accepted as a means to
approximating second-order deformations for axially loaded columns. The method has been
discussed by Menn (1990) in his book Prestressed Concrete Bridges. The method has also been
adapted in many modern bridge design codes, including the CHBDC 2006. This section discusses
the development, applications, and limitations of Vianello's method towards the design of
reinforced concrete highway bridge columns. The section also discusses the fundamental
assumptions and potential shortcomings of Vianellos method.

3.2.1 Case 1: eccentricity is proportional to buckled shape of column

Figure 3-5 Structural model for axially loaded columns with initial eccentricity proportional to buckled
shape of column

A structural model for axially loaded columns with an initial eccentricity is shown in Figure 3-5.
In this case, an initial eccentricity, w0, which follows the buckled shape of the column, is
assumed. The initial eccentricity can be resultant of many factors, such as imposed lateral loads,

44
imposed lateral deformations, or imperfections during construction. As axial load Q is applied, a
moment equal to Qw0 is generated; this moment is referred to as a first-order moment. The first
order moment generated causes the column to undergo further lateral deformations, until a total
maximum deformation equal to wtot,max is achieved. wtot is equal to the sum of the first-order
deformation w0 and the second-order deformation w. The magnitude of w along the length of the
column is affected by both material and geometric nonlinearities. The moment resulting from
the applied axial load Q and the second-order deformation w is known as a second-order
moment. Total moment M is the sum of the first-order moment and second-order moment and is
presented in Equation 3-2.

Equation 3-2

As similarly described by Salonga (2010), the term Qwo is referred to as the first-order moment
M0 since it corresponds to what is known as first-order analysis. First-order analysis is based on a
state of equilibrium that is calculated based on the initial geometry of the structure (Menn 1990).
The term Qw is representative of additional second-order moments M1 caused by member
deformation and change in flexural stiffness EI.
The magnitude of second-order deformations w, and thus second-order moments M1, is
dependent on the flexural stiffness EI of the column as well as the effective length of the column.
When members are stiff and short, second order deformations are small relative to first order
deformations and often a second-order analysis may not be necessary. When members are
slender, second-order deformations w are larger thus inducing larger second-order moments M1.
In the case of slender columns it is important to calculate second-order moments M1 when

45
undergoing design decisions, since the moment demand could otherwise be grossly
underestimated.
The calculation of second-order moments M1 involved the use of analysis methods that consider
both geometric and material nonlinearities. Geometric nonlinearity is the nonlinear loaddeformation response of a column, while material nonlinearity refers to the change in flexural
stiffness EI of the column due to excessive deformations. Some modern methods that consider
both geometric and material nonlinearities have been discussed in Chapter 1 of this thesis. Many
of the methods discussed are very computationally intensive and are often impractical from a
design standpoint. According to Vianello (1898), total deformations wtot can be calculated using
Equation 3-3 (Menn 1990), where w0 is the first-order deformation, Q is the applied axial load,
and QE is the Euler buckling load:

Equation 3-3

The Euler buckling load QE is presented in Equation 3-4, where EI if the flexural stiffness of the
column, k is the effective length factor, and L is the length of the column:

Equation 3-4

A summary of effective length factors k is given in Figure 3-6.


In Figure 3-6 the terms EXP., FIX., and MONO are representative of typically accepted
construction drawing naming conventions. The term EXP. often refers to the use of elastomeric
or pot bearings to form the connection between the pier and superstructure interface.

46
Elastomeric or pot bearing behavior can be approximately modelled as a roller; it allows for a
large degree of lateral and rotational deformation to take place at the pier-superstructure
interface, without the generation of internal forces. The term FIX. generally refers to a bearing
system that allows for a large degree of rotational deformation to take place at the piersuperstructure interface without the generation of internal forces. Lateral deformation will
however generate internal forces. This type of connection can be modelled as a pin connection.
The term MONO. refers to a monolithic connection. As the name implies, the connection does
not require the use of bearings; it does not allow for any lateral or rotational deformation to take
place without the generation of internal forces. This connection can be modelled as a fixed end.

Figure 3-6 Effective length factor for various column configurations.

47

3.2.2 Case 2: eccentricity is not proportional to buckled shape of column


The formulation presented in Equation 3-3 is limited in its applicability. In order for the equation
to be valid the value of w0 must be proportional to the buckled shape of the column along the
length of the column. It is often the case, both in experimental apparatus and in functioning
constructed bridges that axial load is applied to a column or pier at a fixed eccentricity. In the
case of constant eccentricity, magnifying the eccentricity by Vianello's factor would be incorrect;
this is because a constant eccentricity cannot be proportional to the buckled shape of the
column. The buckled shape proportionality requirement is a direct consequence of the derivation
of Vianello's method of successive approximations. The derivation is based on the method of
virtual work, which in this case is explicitly associated with a first-order deformation or
eccentricity that must be proportional to the buckled shape of the column. Details of the
derivation can be found in Appendix A.

Figure 3-7 Structural model for axially loaded columns with constant initial eccentricity

A structural model for eccentrically loaded cantilever columns is shown in Figure 3-7. Similar to
the case presented in Figure 3-5, moments M are induced by axial Q times the distance between

48
the line of action and the centroidal axis of the column. The first-order moment induced by Q
times the initial eccentricity e causes the column to experience further lateral deformations. The
first-order deformation w0 can be computed based on the initial flexural stiffness EI of the
column, which is uniform throughout the length of the column. The column will continue to
experience further deformations until a maximum deformation of wtot,max is achieved. The total
moment in this case is the sum of a constant moment Qe, a first-order moment Qw0, and a
second-order moment that is proportional to the buckled shape of the column Qw. The total
moment for this case is given by Equation 3-5.

Equation 3-5

In the constant initial eccentricity case, adopting Vianello's method to calculate second-order
deformations requires that the initial eccentricity e be used to calculate first-order deformations
w0 then be separated from Vianello's scaling factor as it is not proportional to the buckled shape
of the column. The total deformation in this case is given by Equation 3-6.

Equation 3-6

In the case where the initial eccentricity is proportional to the buckled shape of the column, it is
not necessary to separate any terms, and the total deformation can simply be calculated as was
presented in Equation 3-3.

49

3.2.3 Eccentricity proportional to buckled shape vs. constant eccentricity

Figure 3-8 Load-deformation response comparison for cantilever columns with variable slenderness ratios
using different analytical approaches

Figure 3-8 compares the load-deformation response of a column with an initial eccentricity w0
that is proportional to the buckled shape of the column (column A) versus a column with a
constant initial eccentricity e (column B). The cross-sectional dimensions and design parameters
are also presented in Figure 3-8. Columns with four different slenderness ratios covering a
range of reasonable design values are depicted. The value is shown in Equation 3-7.

Equation 3-7

Where k is the effective length factor, L is the length of the column, and r is the radius of
gyration of the cross section. All four of the sample columns had a fixed k value of 2. The length L
was varied in order to achieve the various slenderness ratio values. In both cases, the columns
had an initial eccentricity equal to one-three-hundredth of their respective effective buckling
lengths, as recommended by Menn (1990). The bold dashed line represents the sectional failure

50
envelope of the column cross section. The four columns plots demonstrate that a range exists,
where there is a significant variability in results depending on whether or not the initial
eccentricity is proportional to the buckled shape of the column or not.
Squat column case: For squat columns, such as may be modelled by a slenderness ratio
of 20, it was shown that whether or not the initial eccentricity is proportional to the buckled
shape does not affect the load-deformation response. The results are in agreement with the
general understanding that squat columns are insensitive to second-order effects.
Slender column case: As the slenderness ratio of the columns increases, considering
whether or not the initial eccentricity is proportional to the buckled shape becomes necessary.
With typical slender column slenderness ratios of around 100, the difference between loaddeformation responses of column A and column B becomes visible. In the slender column case,
column A has a maximum deformation w that is 18.5% smaller than column B and a maximum
axial load Q that is 17.2% greater.
Very slender column case: For very slender columns, in the slenderness ratio range of
approximately 180, the difference in load-deformation responses of column A and column B
diminishes. In the very slender column range, the magnitude of initial eccentricity becomes small
relative to second-order deformations w. With such a large slenderness ratio the column
becomes very sensitive to buckling, and may be regarded as unstable. A column with a concentric
load and no first-order deformation would have failed at a similar axial load Q due to buckling.
The preceding comments exemplify the importance of taking into consideration whether or not
initial eccentricities are proportional to the buckled shape when applying Vianello's equation
towards the calculation of total deformations. Vianellos method is only valid if the initial
eccentricity is proportional to the buckled shape of the column.

51

3.3

Menn's Method

Menn (1990) presented a conservative method for simplifying the calculation of second-order
effects for the design of slender piers. Although typical M-N interaction diagrams in modern
codes are based on states of strain that are limited by the crushing of concrete in the extreme
compressive fibre, as is depicted in Figure 3-9 (a), Menn (1990) recommends that when
calculating sectional forces, designers should limit the states of strain such that the strain in the
steel reinforcement s does not exceed the steel yield strain sy; this reduced state of strain is
depicted in Figure 3-9 (b). In his method, Menn (1990) makes two assumptions: (1) the reduced
state of strain can be used to calculate deformations as well as define sectional capacity and (2)
the flexural stiffness throughout the length of the column can be taken as constant and equal to
the stiffness corresponding to the reduced state of strain. Basing the deformations and sectional
capacity on one state of strain, ensures that a unified formulation of capacity and demand is
defined. Furthermore, by assuming that the flexural stiffness EI throughout the length of the
column is constant and equal to the minimum value as defined by the reduced state of strain,
Menn is able to easily employ Vianellos method to calculate second-order deformations, thus
allowing for a simple and elegant procedure for designing slender reinforced concrete columns.

Figure 3-9 Strain profiles: (a) crushing of extreme compressive fibre; (b) reduced state of strain - steel
yielding; (c) concrete cracking in extreme tensile fibre.

52

3.3.1 Sectional capacity and response


The sectional capacity and moment curvature plots that may be observed with a typical
reinforced concrete pier are presented in Figure 3-10; N is the axial force within the column,
resulting from applied axial loads Q, and M is the resulting moment. The curves presented in the
interaction diagram presented in Figure 3-10 represent combinations of axial force and moment
corresponding to the state of strain limitations presented in Figure 3-9. The axial load Q1
represents a possible magnitude of axial load that results in a state of equilibrium where the pier
can crack prior to the reinforcing steel yielding. The axial load Q2 represents a possible
magnitude of axial load that results in a state of equilibrium where the pier does not crack prior
to the reinforcing steel yielding; this is a case where the whole pier cross section is in a state of
compression. The values labelled Ai, Bi, and Ci refer to combinations of moment and axial force
where the concrete cracks in the extreme tensile fibre, steel reinforcement yields, and the
concrete crushes in the extreme compressive fibre, respectively.

Figure 3-10 Sectional capacity and moment curvature diagrams of pier columns under variable axial loads.

53
Menn (1990) suggests that for the purposes of design, the sectional capacity of the pier should be
limited to combinations of moment and axial force that lie on the reduced state of strain curve,
as presented in Figure 3-10. Menn (1990) also recommends the secant stiffnesses EI
corresponding to these combinations of moment and axial force should be used in calculating
deformations. In the case where the axial load on the column is large enough such that concrete
cracking cannot occur, using the reduced stiffness recommended by Menn (1990) produces the
same results that would be attained through using the flexural stiffness EI of the gross uncracked
section. The flexural stiffness EI of the gross uncracked rectangular section is given by Equation
3-8, where E is the Young's modulus of the concrete, and b and h are the width and height of the
cross section, respectively, with reference to the plane of bending.

Equation 3-8

In the case where the axial load on the column is of a magnitude that allows for cracking to occur
prior to steel reinforcement yielding, the flexural stiffness EI of the gross uncracked section and
the reduced stiffness can differ greatly. Most structural codes in Canada, including the CHBDC
2006, recommend that a flexural stiffness EI equal to 25% of the uncracked stiffness be used postcracking, for the purposes of analysis.

3.3.2 Menn's method: analysis and discussion


Menns simplified method is based on a lower bound minimum flexural stiffness EI and lower
bound reduced M-N interaction diagram, as such it is expected that using the method will
produce conservative results under most conditions. Menn employed his method typically to
design piers of similar proportions as the Reuss Brcke-Wassen piers. These piers can generally
be described as having low reinforcement and slenderness ratios in the range of 100. Under

54
conditions that greatly differ from those that Menn designed under, Menns method may become
unconservative. This subsection is intended to assess Menns method and better determine range
in design parameters for which it performs well and the range for which it performs poorly. Four
design parameters have been considered for identifying the limits of applicability of Menns
method: (1) slenderness ratio (2) reinforcement ratio , (3) magnitude applied axial load N, and
(4) concrete strength fc. Results obtained, using Menns method, are validated through
comparison with results obtained using a more rigorous analytical method. An explanation of the
more rigorous analytical method is presented in subsection 3.3.2.1. The proposed rigorous
analytical method is validated, through comparison with experimental results documented in
literature, in subsection 3.3.2.2. In subsection 3.3.2.3, a comparison between results obtained
using Menns method and the proposed rigorous analytical method is made. The subsection is
intended to answer 3 questions: (1) Does changing the loading configuration from applied lateral
load to imposed lateral deformation influence the validity of Menns method? (2) Is there a range
of slenderness ratios for which Menns method performs poorly? (3) Is there a range of
magnitudes of axial load for which Menns method performs poorly? Subection 3.3.2.4 serves to
identify the influence of reinforcement ratio on the performance of Menns method. Subsection
3.3.2.5 discusses the influence of concrete strength on the performance of Menns method. To
conclude, subsection 3.3.2.6 summarizes the findings made in the preceding subsections and
discusses general implications of the findings made. A summary of ranges in values of the
proposed design parameters, for which Menns method performs well, is provided.
3.3.2.1

Rigorous analytical method

This section describes the fundamental formulations and calculation procedures behind the
proposed rigorous analytical method. Following the description, a comparison outlining the
fundamental differences between the proposed analytical method and Menns method is
presented.

55
The proposed rigorous analytical method is fundamentally based on first-order theory and the
method of virtual work. The method accounts for both material and geometric nonlinearity in
the analysis procedure. These second-order effects are accounted for by partitioning the column
into a number of segments along its length and determining the first-order moment in each
segment. Based on the moment M, the flexural stiffness EI at each segment can be determined;
consequently the curvature at each segment can be determined. Using the method of virtual
work, the deformations along the length of the column can be calculated.
The deformation along the length of the column, coupled with any existing applied axial loads
will in turn generate additional second-order moments. The additional second-order moments
are added to the first-order order moments, and additional deformations along the length of the
column are calculated by determining the flexural stiffness EI and curvature at each segment in
the column, then employing the method of virtual work.
The process outlined is iteratively repeated until either one of two criteria is established: (1) a
deformation is converged upon and any successive iterations result in no additional deformations
or (2) deformations become excessive and thus the column becomes unstable.
The proposed rigorous analytical method offers some advantages over Menns method. The
method accounts for both material and geometric nonlinearity by partitioning the column into
segments and identifying flexural stiffness EI at each of the segments, thus accounting for a nonuniform distribution of flexural stiffness EI throughout the entirety of the column. Inherent in
this difference, is the possibility to design based on the full M-N interaction envelope rather than
the one defined by the reduced state of strain. The defining characteristics of Menns method and
the rigorous analytical method are outlined below.

56
Menns method
1) one

constant

flexural

Rigorous analytical method


stiffness

EI,

1) the column is partitioned into segments

corresponding to the reduced state of

and a variable flexural stiffness EI is

strain, is assumed throughout the

calculated throughout the length of the

length of the column

column

2) deformations are calculated based on

2) all deformations are calculated based

the flexural stiffness EI corresponding

on flexural stiffnesses EI that vary

to the reduced state of strain

throughout the length of the column

3) sectional capacity is limited based on

3) sectional capacity is not limited to

the reduced state of strain

those defined by the reduced state of


strain

A sample calculation comparison demonstrating the differences between the two methods is
shown in Appendix A.
The following section compares the difference in results attained using the two methods for a
cantilever column with an applied lateral load at its tip and discusses the resulting implications
for a cantilever column with an imposed lateral deformation at its tip. The differences in results
are discussed and conclusions pertaining to design procedures and potential shortcomings of
either of the two methods are documented.
3.3.2.2

Validating the rigorous analytical method

In order to validate the proposed rigorous analytical method, a comparison between column
laboratory test results obtained by Baant and Kwon (1994) and predictions made using the
proposed rigorous analytical method has been made. Baant and Kwon (1994) tested 26 separate
columns loaded with an eccentric axial load. Only the columns with the largest cross-sectional
dimensions were used for the purposes of this comparison. The columns were all double-pinended and had varying lengths of 292mm, 546mm and 800mm. Each column length had 3 test
specimens associated with it in order to observe any variability between successive tests. The

57
columns had a square cross section measuring 50.6mm on each side. The columns had a
geometric reinforcement ratio of 4.91%. The average compressive strength fc' of the concrete
was 28.96 MPa, measured at 28 days. Four bars were used for longitudinal reinforcement. The
steel yield strength fy was 552 MPa with a modulus of elasticity Es equal to 200,000 MPa. The
clear cover, measured to the centroid of the reinforcement layer, was 10 mm. The column was
loaded at an eccentricity of approximately 12.6 mm (Baant and Kwon 1994). A redrawn diagram
of the testing configuration as presented by Baant and Kwon (1994) is shown in Figure 3-11.

Figure 3-11 Experimental setup configuration. Adapted from Baant and Kwon (1994).

The rigorous analytical method was performed for an eccentrically loaded column model with
the load applied at an eccentricity of 12.6 mm. For the calculation procedure, the column model

58
was cut into ten segments along the length of the column. The peak load and total deformation
of the column were calculated and compared against the results reported by Baant and Kwon
(1994). A result comparison is shown in Table 3-3. The values presented in Table 3-3 have also
been plotted and shown in Figure 3-12 for illustrative purposes. The results predicted by the
rigorous analytical method are generally in good agreement with the results reported by Baant
and Kwon (1994). The rigorous analytical method generally predicted peak load with greater
accuracy than peak deformation; however, the column with a slenderness ratio of 35.6
demonstrates that the peak deformation is susceptible to a greater degree of variability than the
peak load, as such, some inaccuracy in predicting deformations was expected.
Table 3-3 Measured peak loads and mid-height deformations as reported by Baant and Kwon (1994).

Slenderness
ratio

Test peak load


(kN)

Test mid-height
deformation
(mm)

19.2

48.3
47.5
46.8
41.9
41.0
35.8
32.8
31.2
34.0

2.20
1.95
1.59
6.20
5.77
2.49
9.14
7.57
7.21

35.6

52.5

Virtual work
predicted
peak load
(kN)
46.5

Virtual work
predicted mid-height
deformation
(mm)
2.12

38.4

5.32

29.7

7.14

Figure 3-12 on the following page illustrates the data presented in Table 3-3, where the points
denoted by the grey "X" symbols refer to the values predicted using the rigorous analytical
method for the respective slenderness ratios . As mentioned earlier, the peak load predictions
and mid-height deformation predictions are in good agreement with the test results reported by
Baant and Kwon (1994). This comparison successfully demonstrated that the rigorous analytical
method is capable of predicting peak loads and deformations within a degree of accuracy that is
acceptable for design purposes, and can thus be used to validate Menns method.

59

Figure 3-12 Measured peak loads and mid-height deformations as reported by Baant and Kwon (1994).

3.3.2.3

Influence of loading conditions, axial load, and slenderness ratio

The following section discusses the analysis procedure and results attained regarding the
structural behavior of an axially loaded cantilever column with an applied lateral load H at its tip.
A comparison between the analysis method proposed by Menn (1990) and the proposed more
rigorous analytical method is made. As outlined in the preceding section, the fundamental
assumptions made in the two methods are widely different. The method proposed by Menn
(1990) is much simpler to employ and is vastly superior to the proposed rigorous analytical
method in terms of computational time required. Menns method, however, does not account for
both material and geometric nonlinearity; this is inherent in assuming a constant flexural
stiffness EI throughout the length of the column. It is the intent of this section to make a
comparison between the two methods, and to evaluate Menn's method and assess any inherent
potential analytical shortcomings resulting of its underlying simplified assumptions.
The inherent efficiency of Menn's method provides many advantages from a design perspective.
The method allows for the efficient identification of appropriate preliminary design parameters,
such as reinforcement ratio and cross-sectional dimensions. Since all of the simplifications and

60
assumption made in Menn's method are lower bound assumptions, it is expected that the
method will also produce conservative analysis results under most conditions. From a design
perspective, conservative results are perfectly acceptable as long as the degree of conservatism is
warranted. A structural model for the column, indicating the loading conditions is shown in
Figure 3-13.

Figure 3-13 Structural model for an axially loaded cantilever column with a horizontal point load.

As previously mentioned, Menn's method assumes a constant flexural stiffness throughout the
length of the column. In contrast, calculating the peak loads and deformations of the column
accurately with the proposed rigorous analytical method requires that the variable flexural
stiffness EI and curvature throughout the length of the column be accounted for. Diagrams for
moment and curvature throughout the length of the column are shown in Figure 3-13. The
column will have a constant flexural stiffness EI equal to the gross cross sectional stiffness up
until cracking occurs. First cracking will occur at a location along the column length where the
moment induced by the lateral point load H and the lever arm, in this case L, is equal to the
cracking moment Mcr. After cracking, the flexural stiffness EI of the column decreases
significantly, and as such the curvature will increase significantly. The rate of change of

61
curvature d/dx below the point of first cracking will also increase due to the continuous loss of
flexural stiffness in the cracked section of the column. The variable flexural stiffness EI was
modelled by partitioning the column into ten equal segments along its length L and assuming a
linear rate of change in curvature between the boundaries. A sample diagram of the column
model is shown in Figure 3-14.

Figure 3-14 Assumed curvature for virtual work calculation in rigorous analytical method.

S1 through S10 denote the segments along the length of the column. EI1 though EI10 denote the
respective flexural stiffnesses EI of each of the segments. Segments S1 through S4 have equal
flexural stiffness EI, which are equal to the gross uncracked section flexural stiffness EIg.
Segments S5 through S10 all have variable flexural stiffness EI which increases as the moment M in
the section increases. This results in an abrupt increase in curvature below the first cracking
segment. The moment-curvature diagram depicts the loss of flexural stiffness EI in the cracked
region. The tip deformation can be calculated with Equation 3-9, where wtop is the deformation at
the top of the column, i denotes the segment number, Mi is the induced moment in segment i, EIi

62
is the flexural stiffness of segment i, and Mxi is the moment obtained in segment i when a unit
lateral load is applied at the tip of the column.

Equation 3-9

Successive iterations of the method described above were performed for various values of axial
load Q in order to determine the magnitude of lateral load H that would result in failure and the
magnitude of first-order deformation w1,max resulting from the applied lateral load H. By
successively increasing Mi, second-order effects are taken into account.
Figure 3-14 is intended to illustrate typically expected curvature diagrams where no segment in
the column is in a state of strain for which reinforcement has yielded; sections are only cracked
or uncracked. In the event that a section of the column is in a state of strain where the
reinforcement has yielded, another abrupt change in the rate of change of curvature d/dx would
occur, similar to the transition between uncracked and cracked state.
Figure 3-15 depicts a column cross section and the corresponding M-N interaction envelope.

Figure 3-15 M-N interaction envelope and column cross section

63
The maximum values of lateral load H and the resulting first-order deformation w0,max for various
slenderness ratios , as calculated per Menn's method and proposed rigorous analytical method
for a column with the cross section depicted in Figure 3-15, are shown in Figure 3-16. The chosen
values of range from 60 to 140 in order to encompass a reasonable range, as determined in
Chapter 2. A slenderness ratio of 20 was also considered; it is referred to later as the results do
not differ significantly from those obtained for a column with a slenderness ratio of 60. The
cross-sectional parameters and material properties are the same as for the column presented in
Figure 3-8. The slenderness ratio was varied by changing the length of the column L rather than
the thickness h. The left-hand side graphs presented in Figure 3-16 show the maximum lateral
load H that will result in failure of the column as a function of applied axial load Q. The righthand side graphs in Figure 3-16 show the resulting first-order tip deformations at failure, and
thus also the maximum values of imposed deformation w0,max, for various applied axial loads Q,
which result in failure of the column.
The graphs demonstrate the differences and similarities in results between Menn's method and
the proposed rigorous analytical method. The results obtained from the two methods are in good
agreement under most conditions, and can vary under certain conditions. The intent of the
figures is to identify the potential conditions under which predictions made by Menn's method
differ from the predictions made by the proposed rigorous analytical method. The relationship
between axial force N and the flexural stiffness EI for the analyzed column model is shown in
Figure 3-17.

64

Figure 3-16 Lateral load limits and imposed lateral load deformation limits as per Menn's method and the
proposed rigorous analytical method.

65

Figure 3-17 Relationship between flexural stiffness and axial force.

In Figure 3-17 the curve denoted as gross stiffness (EIg) represents the gross flexural stiffness of
the column prior to cracking as a function of axial force N; this is represented by the point A1,
which can be found in the top-right moment-curvature diagram in the figure. The curve labelled
as yield stiffness (EIy) represents the flexural stiffness EI of the column, assuming that at least one
reinforcement layer has yielded, as a function of axial force N; this is represented by the point B1,
which can be found in the top-right moment curvature diagram in the figure. The last curve,
labelled as crushing stiffness (EIc'), represents the flexural stiffness of the column when the
extreme compressive fibre is at the crushing strain c'; this is represented by the point C1, which
can be found in the top-right moment curvature diagram in the figure. The value of Ncr,max
denotes the maximum amount of axial force, under equilibrated conditions, that allows for
cracking to occur in the column prior to the yielding of the reinforcement. Any values of axial
force exceeding Ncr,max will result in equilibrated states of strain for which both top and bottom
reinforcement layers are in a state of compression when the top reinforcement layer yields. The
values of axial force N denoted Qi refer to the values of axial load Q shown in Figure 3-16.

66
Case 1: column with a slenderness ratio equal to 60
The value denoted Q1 can found be found in Figure 3-18. For a column with a slenderness ratio
equal to 60, Q1 represents a value of axial force Q under which there is an abrupt decrease in
first-order deformation capacity of the column according to both of the presented analysis
methods.

Figure 3-18 Lateral load limits and imposed lateral deformation limits as per Menns method and the
proposed rigorous analytical method for a column with a slenderness ratio of 60.

The explanation for this behavior cannot be found in the relationship between axial force and
flexural stiffness EI alone. Figure 3-17 suggests that for an axial force N equal to Q1 additional
increases in axial force N will result in the greatest increases in yield stiffness EIy, which is
contrary to the behavior observed. The explanation for this behavior can however be found by
referring to the M-N interaction envelope of the column section, which is shown in Figure 3-15.
With an axial force N equal to Q1 the column is at its balance-point condition, and thus also at its
peak moment capacity. The moment in this system is predominantly generated by eccentricity of
the axial load Q, thus as the axial load Q increases, the maximum imposed lateral deformation
wo,max must decrease accordingly to satisfy the moment capacity of the column. For the full range
of axial loads considered, Menns method overpredicts the imposed lateral deformation capacity
of the column.

67
Case 2: column with a slenderness ratio equal to 100
In the case of a more slender column with a slenderness ratio equal to 100, there is an abrupt
increase in first-order deformation capacity of the column according to the proposed rigorous
analytical method, but no increase according to Menn's method. Unlike the case of a more squat
column with a slenderness ratio of 60, a column with a slenderness ratio of 100 tends to have
a behavior that is more controlled by second-order effects. Due to the second-order effects, the
lateral load H and associated maximum imposed lateral deformation w0,max curves will tend to
deviate from the M-N interaction envelope curve. The explanation for the abrupt increase in
imposed lateral deformation capacity according to the more rigorous analytical method can be
traced to the abrupt increase in yield stiffness EIy occurring at an axial force equal to Q1, as can be
seen in Figure 3-19 (b) on the following page. Menns method over-predicts the imposed lateral
deformation capacity of the column for the full range of axial loads considered. Since all of the
segments above the base segment will undoubtedly be in a state of strain where the extreme
compressive fibre has not reached the crushing strain, the smaller offset between the crushing
stiffness EIc' and the yield stiffness EIy will attribute to stiffer behavior of the column. Menn's
method does not portray the same behavior since it always assumes a constant stiffness equal to
the yield stiffness EIy throughout the length of the column. According to the more rigorous
analytical method, the column buckles under the same value of Euler buckling load QE as is
calculated using the reduced state of strain condition presented in Menn's method.
The value denoted Q2 in Figure 3-19 (a) represents a value of axial load Q at which the imposed
lateral deformation limits predicted by Menn's method and the method of virtual work are
approximately equal to one another.

68

Figure 3-19 (a) Lateral load limits and imposed lateral deformation limits as per Menns method and the
proposed rigorous analytical method for a column with slenderness ratio equal to 100, (b) flexural stiffness
as a function of axial force in column.

Under axial load Q2, there is a point loss in flexural stiffness EI and thus smaller imposed lateral
deformation limits w0,max are observed according to the proposed rigorous analytical method.
This was determined to be a result of the peak crushing stiffness EIc' being reached.

69
Case 3: column with a slenderness ratio equal to 140
In Figure 3-20 the value denoted Q4 shows the axial load Q, as calculated by the more rigorous
analytical method, at which there is an abrupt increase in lateral load H capacity of a column
with a slenderness ratio of 140.

Figure 3-20 Lateral load limits and imposed lateral deformation limits as per Menns method and the
proposed rigorous analytical method for a column with a slenderness ratio equal to 140.

At an axial force N equal to Q4 the predictions made using Menns method begin to greatly differ
from those predicted by the proposed rigorous analytical method. This phenomenon occurs due
to the failure mechanism, as proposed by the rigorous analytical method, occurring in a state
where the concrete section is cracked but the reinforcing steel not yet yielded. This suggests that
the column buckles with a flexural stiffness EI that is greater than the reduced flexural stiffness
EIy that is assumed by Menn. The true flexural stiffness EI of the column would have a value that
is between the gross uncracked flexural stiffness EIg of the column and the reduced flexural
stiffness EIy. Since additional axial load increases the flexural stiffness EI of the column, as was
shown in Figure 3-17, making the assumption that the flexural stiffness EI throughout the length
of the column is equal to the reduced flexural stiffness EIy produces results which are potentially
overly conservative. This is clearly exemplified in Figure 3-20. For extremely low values of axial
load, ranging up to the value Q4, indicated in Figure 3-20, Menns method tends to be moderately

70
unconservative in predicting the maximum lateral load H that can be applied before the column
fails. The unconservative predictions are associated with the fact that for extremely slender
columns with slenderness ratios in the range of 140, considering both material and geometric
nonlinearity becomes important, as failure modes are often resultant of column instability.
Menns method does not capture this phenomenon. For all values of axial load ranging up to the
value Q3, indicated in Figure 3-20, Menns method is very unconservative in predicting the
imposed lateral deformation capacity of the column.
Summary of case findings
For slenderness ratios ranging up to 100, Menns method was found to be conservative and
sufficiently accurate in terms of predicting maximum lateral load H that can be applied prior to
failure of the column. For exceptionally slender columns, with slenderness ratios in the range of
140, Menns method becomes unconservative; this is due to the effects of material and geometric
nonlinearity, which Menns method does not take into account. In all cases, Menns method
overpredicts the maximum imposed lateral deformation that can be applied before failure of the
column. This over-prediction in imposed lateral deformation is direct consequence of assuming
constant flexural stiffness EI that is equal to the reduced flexural stiffness EIy throughout the
length of the column. Assuming a reduced flexural stiffness EIy significantly increases the
flexibility of the column and magnitude of imposed deformations prior to failure.
States of strain at failure
The graphs presented in Figure 3-21 identify strain conditions upon failure and thus modes of
failure for different axial loads given various slenderness ratios . The curves are identical to the
ones presented in Figure 3-16 except for the addition of a curve representative of a column with a
slenderness ratio of 20.

71

Figure 3-21 Failure modes as per Menn's method and the proposed rigorous analysis method.

72
Select slenderness ratio data from the graphs presented in Figure 3-21 have been superposed on
one another and shown in Figure 3-22. The graphs for slenderness ratio of 20 have been omitted
as they are redundant with the graphs for slenderness ratio of 60. The superposition serves to
better make comparison between the graphs presented by placing them on a common set of axis,
whereas the separate graphs better serve to identify specific values on the individual sets of axes.

Figure 3-22 Analysis method comparison: (a) applied lateral load limit , (b) imposed lateral deformation
limit.

73
Figure 3-22 (a) depicts a large decrease in the lateral load H capacity of the columns as the
slenderness ratio is increased. The superposition of the graphs serves to better demonstrate the
magnitude of the difference in lateral load H capacity for different slenderness ratios . Although
there is a large decrease in the column's capacity to withstand a lateral load H as the column gets
taller, a clear increase in flexibility of the column can be observed in Figure 3-22 (b). A column
with a slenderness ratio of 100 is approximately twice as flexible as a column with a slenderness
ratio of 60, and a column with a slenderness ratio of 140 is approximately twice as flexible as a
column with slenderness ratio of 100. The explicit increase in flexibility of the columns as
slenderness ratio increases demonstrates the benefit of designing slender columns under
conditions for which imposed deformations are controlling factors. These conditions are however
seldom realized as controlling for ultimate limit states design, which Menns method caters
towards.
Summary of findings
This subsection assessed the range of slenderness ratio , range of applied axial loads Q, and
types of loading conditions in which Menns simplified method performs well. Menns method
was shown to be conservative within an acceptable degree for axial loads under which the
reduced M-N interaction closely matches the conventional non-reduced M-N interaction
diagram; generally, these are values of axial load that are below the balance point condition of
the reduced M-N interaction diagram demonstrated in Figure 3-16 with the value denoted Q1.
Menns method was found to perform conservatively and sufficiently accurately for slenderness
ratio value of up to 100; this is typical of the types of piers that Menn has designed, such as the
Reuss Brcke-Wassen (Uri 1971) presented in Chapter 2, with a slenderness ratio of 100.8.
Generally Menns simplified method should not be used to design piers where imposed lateral
deformations govern. In cases where imposed lateral deformations are significant a more
generalized method such as the proposed rigorous analytical method should be used.

74
The following subsection serves to assess the influence of various reinforcement ratios on the
performance of Menns method.
3.3.2.4

Influence of reinforcement ratio

The preceding subsection identified the range of axial loads Q, range of slenderness ratios , and
types of loading conditions for which Menns simplified method performs well. The analysis was
done for a mechanical reinforcement ratio of 0.625 which equates to a geometric reinforcement
ratio, assuming 400 MPa reinforcement steel strength, of 0.1. The relationship equating
mechanical reinforcement ratio to geometric reinforcement ratio is given in Equation 3-10, where
is the mechanical reinforcement ratio, is the geometric reinforcement ratio, fy is the strength
of steel, and fc is the concrete strength.

Equation 3-10

In this subsection, the 140 slenderness ratio column from the preceding section was analyzed
with mechanical reinforcement ratios ranging from 0.1 to 1.2; these equate to geometric
reinforcement ratios of 0.625 and 7.5, respectively. This range of reinforcement ratios
encompasses most of the scope as determined from the bridge database study presented in
Chapter 2. The bridge database had a minimum geometric reinforcement ratio of 0.4 and a
maximum of 4.71. The CHBDC 2006 allows for a maximum reinforcement of 8.0, thus a 7.5 limit
in this study is representative of what may be seen as a maximum reinforcement in the industry.
The difference in maximum applied lateral load H and maximum imposed lateral deformation
w0,max predicted using Menns method and the more rigorous analytical method was documented.
The difference is shown in Figure 3-23 on the following page and is regarded as a % error.

75
Figure 3-23 shows the % error in Menns method as a function of applied axial load Q for columns
with various mechanical reinforcement ratios . A squared-off grey region denoted
recommended design region is shown in the figure. The region indicates an area on the graph
in which Menns method has 15% or less error and conforms to the earlier recommendation of
only employing the method under circumstances where the applied axial load Q is lower than the
value of axial force N at the balanced condition in the reduced M-N interaction diagram.
The graphs demonstrate that, as long as the applied axial load Q is less than the axial force N at
the balanced point condition in the reduced M-N interaction diagram, Menns method is
typically within a margin of 15% accuracy when predicting the maximum lateral load H that can
be applied to a column before failure. The graphs indicate that as reinforcement ratio increases,
the maximum error in predicting the maximum applied lateral load H increases; however, the
maximum error occurs at an applied axial load Q that is well above the recommended reduced
balanced point force. The graphs further indicate that for lower values of reinforcement ratio,
Menns method is completely unacceptable when designing for imposed lateral deformations.

76

Figure 3-23 Error in Menn's method in predicting maximum lateral load and maximum lateral imposed
deformation for a column with a slenderness ratio of 140.

77
The same procedure was completed for a column with a slenderness ratio equal to 100; in this
case the % error in lateral load limit was always below 0%, implying that Menns method is
generally conservative in this case. The imposed lateral deformations showed a similar pattern as
for the case of the column with a slenderness ratio of 140, but the error was generally lower.
The following subsection discusses the influence of concrete strength fc on the relative
performance of Menns simplified method for the purposes of accurate preliminary design.
3.3.2.5

Influence of concrete strength

The preceding sections demonstrated that, in terms of slender column design, flexural behavior
rather than axial is the predominant factor. Generally, variations in the flexural stiffness EI of the
column due to variations in reinforcement ratio had the greatest influence in terms of dictating
whether Menns method was sufficiently accurate for preliminary design purposes. The CHBDC
2006 states that the modulus of elasticity of concrete Ec can be calculated as shown in Equation
3-11, where fc is the concrete compressive strength and c is the density of concrete.

Equation 3-11

If conforming to the recommendation of remaining below the axial force N corresponding to the
balanced point on the reduced M-N interaction diagram, then at ultimate limit states the section
will undoubtedly be in a state where a large portion of the section is cracked. Given that such a
large portion of the concrete section is cracked, the stiffness of the concrete will contribute
minimally to the flexural stiffness EI of the column. Figure 3-24 on the following page shows the
value of the modulus of elasticity of concrete Ec as a function of compressive strength of concrete
fc and a density c of 2450 kg/m3. A range in concrete strength fc from 25 MPa to 50 MPa was
selected to encompass the scope determined in Chapter 2.

78

Figure 3-24 Modulus of elasticity of concrete as a function of compressive strength of concrete.

Figure 3-24 indicates that there is only a 28.3% increase in the modulus of elasticity of concrete Ec
as the concrete strength fc is increased from 25 MPa to 50 MPa. Assuming that the section
remained completely uncracked, which is not the case, this would only result in a 28.3% increase
in flexural stiffness EI of the section, and this is with the assumption that the entire section
remains uncracked. Compared to increasing the concrete strength fc, increasing the mechanical
reinforcement ratio of the section from 0.1 to 1.2 results in a 280% increase in flexural stiffness
EI of the section. Since the calculation proceeding of Menns method are insensitive as to what
contributes to the reduced flexural stiffness EIy of the section, it can be concluded that the
concrete strength fc will have no significant influence on the performance of Menns method.
Furthermore, given that such minimal increases in the modulus of elasticity of the concrete are
shown in Figure 3-24, it can be anticipated that increasing the compressive strength of concrete
has minimal influence on the flexural capacity of a column. Figure 3-25 on the following page
shows the reduced M-N interaction diagrams for column cross sections with concrete strength fc
ranging from 25 MPa to 50 MPa.

79

Figure 3-25 M-N interaction envelopes for varying values of concrete strength.

Figure 3-25 clearly demonstrates that if the recommendation to stay below the axial force
corresponding to the balanced point is adhered to, then any increase in concrete strength fc
provides no significant increase in moment capacity. Furthermore, when designing for slender
columns, it is likely that the design in itself will be driven to axial forces lying in the region on the
interaction diagram that is below the balanced point. Slender column designs experiencing axial
forces which are located above the balanced point on the interaction diagram will be prone to
buckling and failure modes associated with instability.

3.4

Recommendations using Menns method

The preceding section thoroughly explored typical parameters associated with the design of
slender piers. This section serves to summarize the conclusions made in the preceding section
and identify the capabilities and limitations of Menns method, further defining when and how it
should be used from a preliminary design perspective.

80

3.4.1 Axial load


Based on the findings in preceding sections, it was determined that Menns method could
potentially be used to design for any value of applied axial load and still remain reasonably
conservative. This finding, however, is only general to relatively squat columns, such as those
with slenderness ratios of 60 or less. For more slender columns, it was generally found that
Menns method should only be adopted if the applied axial load is kept below the value of axial
force corresponding to the balanced point on the reduced M-N interaction diagram.

3.4.2 Slenderness ratio


In Chapter 2 a reasonable range of slenderness ratios for practical design was determined by
taking the maximum and minimum slenderness ratios of the piers comprising the 22 bridge pier
study. The slenderness ratios ranged from a minimum of 16.4 to a maximum of 121.6. As long as
the applied axial load remained below the axial force corresponding to the balanced point on the
reduced M-N interaction diagram, Menns method was found to conservatively predict the
maximum loads that could be applied at ultimate limit states for columns with slenderness ratios
ranging up to 100. In these cases the method was conservative by only a few percent.
For exceptionally slender columns, such as those with slenderness ratios in the range of 140,
Menns method was found to be generally unconservative in terms of predicting the maximum
loads that could be applied to the column. The method, however, was typically unconservative by
a maximum of 15% if the magnitude of applied axial load was below the axial force corresponding
to the balanced point condition on the reduced M-N interaction diagram. For the purposes of
preliminary design, results that are unconservative by 15% can be tolerated, thus making Menns
method a viable approach.

81

3.4.3 Reinforcement ratio


Column models with mechanical reinforcement ratios ranging from 0.1 to 1.2 were analyzed in
the preceding section. In terms of predicting the maximum load that could be applied to a
column at ultimate limit states, Menns method was generally found to be more accurate for
lower values of reinforcement ratio. If the applied axial load is kept below the axial force
corresponding to the balanced point on the reduced M-N interaction diagram, the accuracy in
Menns method increases as reinforcement ratio increases. This phenomenon occurs because the
greatest error occurs at magnitudes of axial load that are well above the axial force corresponding
to the balanced point on the reduced M-N interaction diagram. If the balanced point axial force
recommendation is adhered to, the predictions made by Menns method for any of the
reinforcement ratios considered were unconservative by at most 15%. It was demonstrated that,
for the purposes of preliminary design, Menns method could be used for columns with
mechanical reinforcement ratios ranging from 0.1 to 1.2.

3.4.4 Concrete strength


The calculations for Menns method are fundamentally based on using the reduced flexural
stiffness EIy of a column in order to calculate first-order and second-order deformations. The
method is effectively blind towards how that flexural stiffness EI is determined. It was shown that
increasing the concrete strength from fc from 25 MPa to 50 MPa could increase the flexural
stiffness EI of the column by at most 28.3% if the entire section remains uncracked. If the
recommendation to keep the magnitude of applied axial load below the axial force corresponding
to the balanced point on the reduced M-N interaction diagram is adhered to, a large portion of
the section will undoubtedly be cracked, as such the full 28.3% increase in flexural stiffness EI
cannot be realized; based on this, it is shown that concrete strength fc has no significant impact
on the results obtained using Menns method.

82

3.4.5 Applied load versus imposed deformation


In section 3.3 it was demonstrated that Menns method is generally unconservative when
calculating the maximum imposed deformation that can be tolerated prior to failure of a column.
The unconservative calculation of maximum imposed deformations is inherent of Menns
recommendation to make the flexural stiffness EI throughout the length column equal to the
reduced flexural stiffness EIy. Although assuming a lower flexural stiffness EI will typically lead to
conservative results when calculating maximum loads that can be applied, it has the opposite
effect on calculating the maximum imposed deformations. Assuming a lower flexural stiffness EI
effectively implies that lower section forces will be generated due to imposed deformations. This
finding suggests that Menns method is especially susceptible to error when calculating
maximum imposed deformations for more squat columns.

3.4.6 Summary
This subsection summarizes the limits of Menns method, as determined from the preceding
subsections. The limits defining when Menns method can appropriately be used are depicted in
Figure 3-26.

Figure 3-26 Limits of recommended use for Menn's method.

83
The recommendations presented in Figure 3-25 are based on calculations for which a maximum
tolerable error, either conservative or unconservative, of 15% was used. The figure was based on
the assumption that there are no imposed deformations applied to the column. As stated earlier,
for columns where imposed deformations govern, Menns method generally overestimates the
amount of deformation which can be imposed on the column prior to failure.

3.5

Concluding remarks

This chapter presented the bilinear concrete stress-strain relations for concrete documented in
the fib Model Code 2010. The bilinear stress-strain approximation was shown to be a sufficiently
accurate approximation of the generally accepted parabolic stress-strain relationship and is thus
appropriate for use in preliminary design. The methods employed by Menn (1990) in order to
calculate the deformation and capacity of a column subjected to axial load and moment were
discussed. A rigorous analytical method based on first-order theory and the method of virtual
work was introduced and validated through comparison with experimental results documented
in literature. Menn's methods were critiqued by comparing results obtained using the method to
results obtained using the proposed rigorous analytical method. Menns method was found to be
generally acceptable for calculating maximum loads that could be applied to a column at
ultimate limit states for columns with slenderness ratios ranging from effectively 0 to 140 as long
as the magnitude of axial load applied to the column was less than the axial force corresponding
to the balanced point on the reduced M-N interaction diagram. The performance of Menns
method was generally found to be insensitive to choice of reinforcement ratio and concrete
strength, as long as the magnitude of applied axial load was kept below the axial force
corresponding to the balanced point on the reduced M-N interaction diagram. Menns method
was generally found to perform poorly and produce unconservative results when calculating for
maximum imposed deformations that could be tolerated prior to the failure of a column.

84
In the next chapter, the methods which have been discussed and validated will be used
appropriately to develop design aids, which allow for efficient proportioning of cross-sectional
dimensions and reinforcement ratios at serviceability limit states and ultimate limits states. The
aids are intended to be used by designers for the purposes of preliminary design.

85

Chapter 4. Individual Bridge Piers


The following chapter discusses the analysis and design of individual bridge piers, either freestanding or in a system. The loading conditions as well as structural behavior of free-standing
bridge piers will generally differ from the loading conditions and structural behavior of bridge
piers in a system. The term free-standing refers to the bridge pier conditions prior to the
construction of the superstructure. The construction of the superstructure, or bridge deck,
effectively provides a rigid link between all of the piers in the system which are connected
through the deck. Although the loading conditions are typically much less severe during the freestanding component of a bridge pier's life, the pier is also generally much more flexible and as
such is more susceptible to lateral deformations due to applied loads. In the case of very slender
piers, the second order deformations due to applied loads while the pier is free-standing can
potentially govern the design of the pier in terms of reinforcement ratio and cross section. The
dominant loads applied to a bridge pier prior to the construction of the deck will be attributed to
wind and self-weight of the pier. The wind load acts laterally on the vertical face of the pier and
generates a moment at the base, as well as a first-order deformation at the tip. The dead load of
the pier acts as a gravity load and thus will induce second-order deformations if first-order
deformations or initial eccentricities are existent. As previously mentioned, first-order
deformations can result from wind loads, which when coupled with the self-weight of the pier
will result in second-order moments and second-order deformations. Initial eccentricity can also
exist due to imperfections in construction. Even minor eccentricities and first-order
deformations must be accounted for when analyzing very slender piers. All of these factors must
be combined and accounted for when designing for both serviceability and ultimate limit states.
This chapter discusses the specifics of loading conditions, serviceability limit states design and
ultimate limit states design of individual bridge piers, both from a free-standing and system
perspective.

86

4.1

Free-standing pier loading conditions

The following section discusses the loading conditions generally applied to the free-standing
pier. These loads primarily consist of lateral uniformly distributed loads due to wind and vertical
dead loads from self-weight of the pier. Independently, both of these loads can only induce firstorder effects, which are simple to account for. When combined with the wind load, the selfweight of the pier will generate second-order lateral deformations and in turn generate secondorder moments throughout the length of the pier. These second-order variables are generally
more complicated to account for. In the absence of wind load, which is seldom the case, secondorder effects may be generated if there is an initial eccentricity to the gravity load created by the
self-weight of the pier. These eccentricities can often result from imperfections in construction,
which are not uncommon, especially when constructing exceptionally slender bridge piers. The
following subsections will address the analysis and influence of wind load, self-weight, and initial
eccentricity.

4.1.1 Wind load


As mentioned earlier in this chapter, wind load generally acts as a uniformly distributed load on
a vertical face of the free-standing pier. The magnitude of wind load generally depends on the
geographic location of the structure and is prescribed in design codes that govern in the
jurisdiction under which the structures are designed. In the CHBDC 2006 wind loads are given as
pressures to be applied to any vertical face of an exposed substructure that face in the
longitudinal centreline of the superstructure and can be skewed by up to 60 degrees. When the
wind load direction is skewed, the load is to be resolved into components taken to act
perpendicularly to the end and side elevations of the substructure. These loads are assumed to
act horizontally at the centroids of the exposed areas of the end and side elevations of the
substructure.

87
For piers with a rectangular cross section, applying the wind load in a direction that is normal to
the vertical that is parallel to the weak axis of the pier will generally produce the most critical
conditions. Figure 4-1 shows a plan view diagram of wind load acting on the vertical longitudinal
face of a pier and the location of resultant components.

Figure 4-1 Plan view of possible wind load conditions applied to a pier.

The CHBDC 2006 provides a method of calculating the horizontal wind load to be applied to a
structure. The method is presented below.
The wind load resultant presented in Figure 4-1 above has been transformed to a linear unit load.
The transformation is presented in Equation 4-1.

Equation 4-1

Fh is the wind load per unit exposed frontal area of the structure, as presented in Equation 4-2.

Equation 4-2

88
In calculating Fh, q is the reference wind pressure determined from Table A3.1.1 in the CHBDC
2006, Ce is the wind exposure coefficient, Cg is the gust effect coefficient, and Ch is the horizontal
wind drag coefficient. The values of wind exposure coefficient Ce, gust effect coefficient Cg, and
horizontal wind drag coefficient Ch are presented in Section 3.10 of the CHBDC 2006.
The wind exposure coefficient Ce to be used is not to be less than 1.0 and shall be taken from
Table 3.8 of the CHBDC 2006. It may also be calculated as prescribed in Equation 4-3, where L is
the height of the pier.

Equation 4-3

The gust effect coefficient Cg to be used is 2.0 for bridges that are not sensitive to wind action.
For structures which are sensitive to wind action, the gust effect coefficient should not be used; a
more detailed dynamic analysis of wind action using an approved method is required.
The horizontal wind drag coefficient Ch is to be taken as 2.0.
Table 4-1 summarizes the values of factors required to calculate wind load per unit exposed area
Fh, as prescribed by CHBDC 2006.
Table 4-1 Wind load property values.

Property

Wind exposure coefficient (Ce)

Gust effect coefficient (Cg)


Horizontal wind drag coefficient (Ch)

Value Used
1.0 (Pier heights 0 to 10 metres)
1.1 (Pier heights over 10 to 16 metres)
1.2 (Pier heights over 17 to 25 metres)
1.3 (Pier heights over 25 to 37 metres)
1.4 (Pier heights over 37 to 54 metres)
1.5 (Pier heights over 54 to 76 metres)
1.6 (Pier heights over 76 to 100 metres)
2.0
2.0

89

4.1.2 Dead load


The dead load of the pier refers to the self-weight of the pier. The self-weight of the pier results
in a downward gravitational force. When the pier is perfectly straight and undeformed, this
results in a pure axial force which is simple to account and design for as it will only be governed
by ultimate limit states. A pier should be designed for two conditions in this situation: (1) to
resist the compressive stresses generated in the concrete, and thus prevent crushing, and (2) to
not buckle under the given axial load. When the pier is not perfectly straight, either due to
deformations or construction imperfections, second-order moments will be induced due to the
eccentricity of the load. In this case, the pier must be designed for both serviceability limit states
and ultimate limit states. Serviceability limit states design is generally governed by the control of
the spacing between cracks as well as the crack width. Both of these values must be kept within
certain limits under expected serviceability limit states loads.

4.1.3 Pier with concentric load


Under a purely axial load, the pier is also inherently under pure compression. Furthermore it is
under a state of uniform compression, that is to say the compressive stress is the same at any
point in the cross section. Typically a proper design will ensure that this compressive stress does
not exceed the reduced compressive strength of the concrete cfc'. Where c is the material
resistance factor for concrete and fc' is the nominal compressive strength of the concrete.
In order to prevent buckling, the axial load in the column must be kept below the Euler buckling
load of the column QE. The Euler buckling load QE for a free-standing column is given in
Equation 4-4, where L is the length of the column and EI is the flexural stiffness of the column.

Equation 4-4

90
In the case of the free standing pier with purely axial load, the flexural stiffness EI will be equal to
the gross cross-sectional flexural stiffness EIg throughout the entirety of the pier. For a pier with a
rectangular cross section, with a width b and a thickness h, the gross flexural stiffness is given by
Equation 4-5 where Ec is the Youngs modulus of the concrete.

Equation 4-5

4.1.4 Pier with eccentric load


When a pier is not perfectly normal to the ground, an eccentricity between the point of
connection at the base and the centre of mass of the pier will exist. As previously mentioned, this
eccentricity can either be due to induced lateral deformations or imperfections during
construction. This eccentricity will generate first-order moments throughout the pier, which will
further proceed to generate second-order moments and second-order deformations. In the case
of a free-standing cantilever pier, the greatest moment will be generated at the base.
The concentric load case previously mentioned is generally not a realistic representation of the
loading conditions that will be applied to a free-standing pier. Wind loads will undoubtedly be
applied to the vertical face of the pier, and will as such induce first-order deformations and firstorder moments at the base. These, when combined with the vertical gravity load resulting from
the self-weight of the pier, will generate second-order moments and second-order deformations.
Unlike an applied axial load which would be concentrated at the tip of the pier, the self-weight of
the pier induces vertical forces along the deformed shape acting downwards. The difference
between these two loading conditions is shown in Figure 4-2.

91

Figure 4-2 Statical models for free-standing pier loading conditions.

Calculating the moment for the free-standing pier with a point load at its tip is simple if the selfweight of the pier is omitted; this is however an unrealistic case, as the pier will always have a
self-weight associated with it. The moment at the base of the pier when including self-weight is
more difficult to calculate. Generally lateral deformations should be known along the entire
length of the pier. The integral of the force created by the self-weight of the pier and the lever
arm resulting from lateral deformations needs to be taken in order to determine the moment at
the base.
Menn (1990) stated that the moment generated by the self-weight of a free-standing pier may be
accounted for by applying one third of the self-weight at the tip of the column as a concentrated
point load. Figure 4-3 on the following page illustrates a statical model of Menn's
recommendation.

92

Figure 4-3 Statical model for the calculation of second-order effects in a free-standing column.

Menn's assumption of using one third of the self-weight of the pier is originally presented by
Timoshenko and Gere (1961). The total moment at the base of the column is a function of the
sectional density of the pier sect, the magnitude of applied lateral wind load qw, the length of the
pier L, and the flexural stiffness of the cross section EI. Menn further mentions that additional
loads can exist in an effectively free-standing column case when the pier is connected to the
superstructure through the use of expansion bearings. Generally expansion bearings are
modelled as rollers and as such cannot transfer any horizontal force or impose any deformations
from the pier to the superstructure and vice versa. However, Menn (1990) states that frictional
forces will exist between the expansion bearings and piers, which will transfer forces between
piers and superstructure. these forces are not general and are dependent on many factors such as
bearing type and manufacturer. Menn (1990) states that the frictional sliding forces, when
bearings are properly detailed and maintained, can be approximated as being equal to roughly
5% of the vertical dead load reaction from the superstructure. Given that this is an approximation
and is effectively dependent on the proper detailing and maintenance, it is recommended that a
more rigorous analysis be conducted for the finalized design of the piers.

93
Figure 4-3 illustrated a case where an applied lateral wind load created first-order deformations;
the same principles would hold true in the case of an initial eccentricity resulting in construction
imperfections. Menn (1990) recommends that a good approximation of the initial eccentricity of
a column, when calculating the capacity of the pier, is one-three-hundredth of the effective
buckling length, normalized with respect to the pier thickness. Menn's recommendation is
presented by Equation 4-6, where w0,max is the initial eccentricity, k is the effective length factor,
h is the thickness of the pier and L is the length of the pier.

Equation 4-6

Menn states that this eccentricity is assumed to be proportional to the buckled shape of the
column. Assuming that the initial eccentricity is proportional to the buckled shape allows for the
use of Vianello's method for calculating second-order deformations.
The magnitude of the initial eccentricity, as assumed by Menn is generally significantly greater
than those which are allowed for by construction tolerance limits prescribed by CSA A23.1 (2004).
CSA A23.1 governs tolerances for concrete construction in Canada. The prescribed construction
tolerance limit for alignment of concrete members is based on the member length. Taller
members or columns are generally expected to suffer greater misalignment or out-of-plumbness
during construction. Only in the case of exceptionally short columns does Equation 4-6, as
prescribed by Menn, suggest the use of initial eccentricity limits which are comparable to those
prescribed by CSA A23.1. Table 4-2 summarizes the construction tolerances for eccentricity as
adapted from those prescribed by CSA A23.1 (2004).

94
Table 4-2 Pier construction tolerances as prescribed by CSA A23.1.

Length of Pier (mm)

Eccentricity Tolerance (+/- mm)

0 - 2400
2400 - 4800
4800 - 9600
9600 - 14,400
14,400 - 19,200
19,200 - 57,600
above 57,600

5
8
12
20
30
50
as specified by designer

A comparison between assumed initial eccentricities as prescribed by Menn and the construction
tolerance limits prescribed by CSA A23.1 (2004) has been made in order to accurately identify the
difference between Menn's initial eccentricity assumption and the construction tolerance limits.
The comparison is presented in graphical format in Figure 4-4.

Figure 4-4 Pier eccentricity limits as per CSA A23.1 (2004) and Menn (1990).

Figure 4-4 demonstrates that Menn's eccentricity recommendation is significantly greater than
the construction limits prescribed by CSA A23.1 (2004). Although no further recommendations
are given by Menn (1990) for initial eccentricity, the Swiss Code prescribes limitations on this
exact recommended eccentricity.

95
The SiA 162 (1989) prescribes the use of an initial eccentricity that is equal to one-threehundredth of the critical length of a column and is proportional to the buckled shape of the
column, however it goes further to state that this initial eccentricity shall be no less than onefifteenth of the thickness of the column, and no greater than 0.1 m.
A comparison between initial eccentricity limits prescribed by the SiA 162 (1989) and the
construction tolerance limits prescribed by CSA A23.1 (2004) has been made and presented in
Figure 4-5. The minimum value for initial eccentricity, defined as one-fifteenth of the column
thickness, has been computed based on an assumed column thickness of 1100 mm.

Figure 4-5 Pier eccentricity limits as per CSA A23.1 (2004) and SiA 162 (1989).

Figure 4-5 shows that with the inclusion of the eccentricity limits prescribed by SiA 162 (1989),
the initial eccentricity prescribed by Menn (1990) is much closer to the construction tolerances
prescribed by CSA A23.1 (2004).

96

4.1.5 Inferred eccentricity limit


The preceding subsection discussed the behavior of eccentrically loaded piers and the
eccentricity limits imposed by the SiA 162 (1989) and the CSA A23,1 (2004). The CHBDC 2006 also
imposes a cut-off on the maximum axial force that can be used from produced M-N interaction
diagrams. The code states that no more than 75% of the compressive strength fc can be used at
any point in calculations for flexural members under compression. This effectively cuts off the MN interaction diagram at a value of 0.75Acfc. An axial force equal to 0.75Acfc will have a
maximum allowable moment associated with it on the M-N interaction envelope. For
demonstrational purposes, a sample M-N interaction envelope is shown in Figure 4-6.

Figure 4-6 M-N interaction diagram cut off at 0.75Acfc'.

The axial force NR and moment MR values corresponding to the axial force NR of 0.75Acfc have
been identified as N75 and M75, respectively. This axial force NR and moment MR are shown in
Figure 4-6. By cutting off the M-N interaction envelope at 0.75Acfc, a minimum moment M75 that
must be designed for is imposed. This moment M75 can effectively be used to infer minimum
eccentricity w0,min that must be designed for. The inferred minimum eccentricity is presented in

97
Equation 4-7 below, where N75 is the axial force equal to 0.75Acfc and M75 is the corresponding
moment.

Equation 4-7

For thicker sections and sections with generous reinforcement, the calculated eccentricity w0,min
could be quite large, whereas for thinner sections and sections with mild reinforcement, the
calculated eccentricity w0,min could be quite small. Effectively, the value of w0,min could range
anywhere from zero to infinity. In order to assess the extents of possible values for inferred
eccentricity w0,min the axial force NR, moment MR, and eccentricity w0,min have all been normalized
as presented in Equations 4-8, 4-9, and 4-10, where b is the width of the cross section, h is the
thickness of cross section, and fc is the compressive strength of concrete.

Equation 4-8

Equation 4-9

Equation 4-10

Using the normalized design values, a normalized set of M-N interaction envelopes for various
values of mechanical reinforcement ratio can be generated. The mechanical reinforcement
ratios which were considered in this thesis range in value from 0.1 to 1.2. This encompasses
most of the scope determined from the bridge study in Chapter 2. As mentioned previously, the
CHBDC 2006 prescribes a maximum geometric reinforcement ratio of 8% and a minimum as
presented in Equation 4-11, where As is the cross-sectional area of steel, Ag is the gross cross-

98
sectional area of the section, fy is the yield strength of steel, and fc is the compressive strength of
concrete.

Equation 4-11

Figure 4-7 below gives a visual representation of the range of geometric reinforcement ratio
that has been considered in the scope of this thesis. The figure assumes a yield strength fy of 400
MPa.

Figure 4-7 Geometric reinforcement ratio range considered in scope of thesis.

Based on the geometric reinforcement ratio , a set of M-N interaction envelopes were
developed. The envelopes are associated with mechanical reinforcement ratios which are
within the geometric reinforcement ratio scope, assuming a steel yield strength fy of 400 MPa.
The relationship between mechanical reinforcement ratio and geometric reinforcement ratio
is shown in Equation 4-12, where As is the cross-sectional area of steel, Ag is the gross crosssectional area, fy is the yield strength of steel, and fc is the compressive strength of concrete.

99
Equation 4-12

The developed M-N interaction envelopes, presenting the range of normalized inferred
eccentricity are shown in Figure 4-8 below.

Figure 4-8 Normalized M-N interaction envelopes and normalized inferred eccentricity.

Figure 4-8 demonstrates that by forming a line through all of the (M75,N75) coordinates on the
graph, it can be deduced that the rate of change in inferred eccentricity is constant with the rate
of change of mechanical reinforcement ratio . This is an important and interesting result, as it
implies that for a range of mechanical reinforcement ratios between 0.1 and 1.2 it is possible to
formulate a relationship between the mechanical reinforcement ratio and normalized inferred
initial eccentricity. The calculated relationship between mechanical reinforcement ratio and
normalized inferred initial eccentricity is presented in Equation 4-13 on the following page.

Equation 4-13

100

With the magnitude normalized inferred initial eccentricity being entirely dependent on the
thickness of the cross section h, it would be interesting to observe how the normalized inferred
initial eccentricity varies for different values of cross-sectional thickness h, and how it compares
to the maximum initial eccentricity limit prescribed by SiA 162 (1989) of 100 mm. Figure 4-9
below shows the relationship between inferred initial eccentricity w0,min and cross-sectional
thickness h for different values of mechanical reinforcement ratio .

Figure 4-9 Inferred initial eccentricity w0,min as a function of cross-sectional thickness h.

Figure 4-9 demonstrates the range of cross-sectional thicknesses h for which the SiA 162 (1989)
maximum initial eccentricity limit of 100 mm is in agreement with the inferred initial
eccentricity, based on the axial force N limitation of 0.75Acfc, prescribed in CHBDC 2006.
The maximum cross-sectional thickness h for which the SiA 162 (1989) limit remains lower than
the calculated inferred eccentricity is 1110 mm. For cross-sectional thicknesses exceeding 1110 mm,
the SiA (1989) prescribes design eccentricities which are lower than those inferred from the
CHBDC 2006.

101
As mentioned earlier, the recommendations made by Menn (1990) do not mention a 100 mm cutoff for the recommendation of proportioning the initial eccentricity as one-three-hundredth of
the critical buckling length of the pier. A comparison between Menns eccentricity
recommendation for different lengths of piers and calculated inferred eccentricity has been made
and presented in Figure 4-10.

Figure 4-10 Inferred initial eccentricity compared to eccentricity recommended by Menn (1990).

Figure 4-10 demonstrates the relationship between inferred initial eccentricity and the initial
eccentricity proposed by Menn (1990) through the association of pier length L with inferred
initial eccentricity w0,min. The primary vertical axis shows the inferred initial eccentricity w0,min,
whereas the secondary vertical axis shows the pier length L based on Menns initial eccentricity
recommendation of one-three-hundredth of the critical buckling length of the pier. An effective
length factor k of 2.0 has been used, as is the case for a free-standing pier. The dots plotted in
Figure 4-10 represent the values of cross-sectional thickness h and pier length L associated with
the 22 bridge piers studied in Chapter 2. The darker grey area in the graph indicates combination
of pier lengths L and cross-sectional thicknesses h for which Menns initial eccentricity
recommendation over-predicts, or is in other words conservative relative to the inferred initial

102
eccentricity. The lighter grey region, in contrast, indicates a combination of pier lengths L and
cross-sectional thicknesses h for which Menns initial eccentricity recommendation underpredicts, or is in other words unconservative relative to the inferred initial eccentricity. It is
evident that most of the 22 bridge piers studied have piers which fall in the range where Menns
method would under-predict the value for initial eccentricity, relative to the inferred initial
eccentricity. However, Figure 4-10 also indicates that most of the 22 bridges studied have piers
which fall in the range where it is effectively irrelevant whether or not the 100 mm cut-off, as
prescribed by SiA 162 (1989), is used in conjunction with Menns initial eccentricity
recommendation.
Since Menns initial eccentricity w0,min recommendation is in agreement with the SiA 162 (1989)
under most reasonable conditions, as identified through the 22 bridges studied in Chapter 2, it is
concluded that using the proposed recommendation of one-three-hundredth of the critical
buckling length is appropriate for the purposes of preliminary design procedures.

4.2

Serviceability limit states design

The design of piers at serviceability limit states design is generally based on a limit of crack width
and crack spacing in the extreme tensile fibre of the section. All presently developed methods for
calculating the crack width and crack spacing, including what is prescribed by the CHBDC 2006,
are based on the assumption that the stresses or strains in the reinforcement are known
beforehand (Menn 1990). The reality is that it is impossible to identify the exact stress or strain in
the reinforcement, as it is a function of many complex factors such as redistribution of sectional
forces, self-equilibrating states of stress, and restrained deformations (Menn 1990). In his book
Prestressed Concrete Bridges, renowned bridge designer Christian Menn states that any accuracy
that may be promised by an exact calculation of steel stresses under service conditions is in fact
illusory. As such, a more rational approach to designing for serviceability limit states is to make

103
simplifications based on rational models of structural behavior to calculate steel stresses. These
calculations are to be a rough conservative estimate rather than an "exact" solution (Menn 1990).

4.2.1 Identifying a critical cross section


In order to identify stresses in reinforcement, a conservative analysis of the most critical cross
section, as limited by the clear cover requirements prescribed by the CHBDC 2006, was used. The
section has been presented earlier in this thesis and makes the unrealistic but conservative
assumption that the reinforcing bars can have a cross-sectional area with no thickness. This
assumption effectively implies that the clear cover can be measured to the centroid of the
respective reinforcing layers. The cross section is presented in Figure 4-11.

Figure 4-11 Serviceability limit states analysis critical cross section.

The critical cross section has a thickness of 700 mm and a width of 2800 mm with a 70 mm
distance between the centroid of reinforcement and external face of the section. Effectively a
more critical section could have been analyzed were the clear cover increased or the thickness of
the section decreased; however, the scope of this thesis has been limited to two cross-sectional
configurations. The two configurations consist of a cross section in which the ratio between the
thickness of the cross section and the clear cover to the centroid of reinforcement is equal to 0.05
and a section for which the ratio is equal to 0.10. These configurations were selected to mimic

104
those presented by Menn (1990) in his book Prestressed Concrete Bridges. A diagram showing the
two configurations is shown in Figure 3-8.

Figure 4-12 Sectional configurations: (a) 5% centroidal clear cover, (b) 10% centroidal clear cover.

Adopting centroidal clear cover to sectional thickness ratios of 0.05 and 0.10 encompasses a
broad and applicable range of typical reinforcement covers that would be used in practice. For
the purposes of design, any value of the ratio below 0.05 would be more conservative and would
thus be accounted for by a design which is based on a 0.05 ratio. As mentioned earlier, given
CHBDC 2006 requirements, the thinnest cross section which can satisfy clear cover criteria, given
centroidal clear cover to sectional thickness ratio of 0.05, would be 700 mm in thickness. A 700
mm cross-thickness is well below what is conventionally used in practice for reinforced concrete
highway bridge piers; thus the proposed centroidal clear cover to sectional thickness ratio of 0.05
is in itself a conservative design assumption.

4.2.2 Identifying average crack spacing


At serviceability limit states the CHBDC 2006 controls the design of the pier through a
prescribed limit on crack width. In order to readily design a pier cross section, a state of strain is
needed such that stresses and sectional forces can be calculated. Translating the crack width to a
state of strain requires that the average spacing of cracks is known. In section 8.12.3.2 of the
CHBDC 2006 a formula for calculating average cracking spacing srm is provided. The formula is

105
presented in Equation 4-14, where srm is the average spacing of cracks, kc is a factor to be taken as
0.5 for sections in bending and 1.0 for sections in pure tension, pc is the ratio of area of steel
reinforcement provided for tension to area of concrete in tension, and db is the nominal bar
diameter.

Equation 4-14

Menn (1990) proposes that since it is impossible to truly know the value of all the parameters
required to calculate the average crack spacing it is more reasonable, for the purposes of design,
to make a conservative assumption. Menn proposes a value of 200 mm.
A brief parametric study was conducted in order to assess how the area ratio pc affects the
average cracking spacing srm that would be calculated using Equation 3-12 for a reasonably
dimensioned pier cross section. It was also the intent of the study to observe how Menn's 200
mm average crack spacing srm recommendation compares to what is prescribed by the CHBDC
2006. The results of this parametric study are presented in Figure 4-13.

106

Figure 4-13 Maximum average crack spacing as a function of area ratio as per CHBDC 2006.

The parametric study shows that the bar diameter db as well as area ratio pc have a significant
influence on average crack spacing srm. For small values of the area ratio pc, the bar diameter db
has no significant influence on the magnitude of average crack spacing srm. The choice of bar
diameter db does however dictate the maximum average crack spacing srm that can be attained
using Equation 4-14. The maximum average crack spacing for all of the shown bar diameters db,

107
ranging from 10M to 35M bars, is well above Menn's recommended 200 mm and is thus not of
any particular interest for the purposes of design.
Generally, lower the average crack spacing srm equates to larger stress in the reinforcement layers.
Smaller average crack spacing srm equates to a tighter concentrations of cracks and thereby larger
strains given a fixed crack width wcr.
According to the CHBDC 2006 proposed average cracking spacing srm, Menn's 200 mm
recommendation shows the greatest range of over-prediction for cross sections in which the bar
diameter size is small, as would be the case for the 10M bar reinforcement shown in the chart in
the top left of Figure 4-13. The figure shows that Menn's recommendation over-predicts the value
of average crack spacing srm, as calculated per CHDBDC 2006, when the area ratio pc becomes
greater than 0.04.
Even though 10M bars are never used for the purposes of longitudinal pier reinforcement in
highway bridges piers in Canada, a calculation was performed to determine the depth of the area
of the cross section that would be seen for the limiting pier cross section shown previously in
Figure 4-11. Determining the depth of the area in tension gives insight in determining how
realistic a case where the area ratio is above 0.04 would be. Inherently this provides further
means of determining how suitable Menn's 200 mm recommendation is for the purposes of
design, as prescribed by the CHBDC 2006. The calculation steps as well as figure depicting the
findings are shown on the following pages.

108
In determining the depth of the area in tension for the cross section shown in Figure 4-11, an area
ratio pc equal to 0.04 was assumed.
The following steps were taken in order to determine the depth of area in tension:
1) The area of concrete in tension Act was calculated.

2) Given the area of concrete in tension Act the depth of the area in tension dt can be
calculated:

3) Determine the % of section in tension t%:

109
4) Using the compression depth, a set of maximum stresses at the extreme tensile fibre ft,max,
assuming different compressive strengths of concrete fc' can be calculated:

Figure 4-14 summarizes the findings for concrete compressive strengths fc' ranging from
20MPa to 50MPa. The concrete tensile strength fcr was calculated based on Equation 4-15.
The Equation was attained the CHBDC 2006.

Equation 4-15

Figure 4-14 Maximum tensile stress in extreme tensile fibre at prescribed cracked condition compared to
tensile strength of concrete.

Figure 4-14 shows that the maximum tensile stresses that can be generated for the minimum
critical area ratio pc differ minimally from the tensile or cracking strength fcr for concrete

110
specimens ranging in compressive strengths fc' between 20 MPa and 50 MPa. For concrete
strengths up to 30 MPa, which is the most common concrete strength used for the piers of the 22
bridges studied, the stress in the extreme tensile fibre exceeds the cracking stress fcr by only
15.8%. It is emphasized that the value for maximum stress in the extreme tensile fibre ft,max is
derived based on the assumption that the bridge pier has maximum curvature for the given
value of area ratio pc, which would occur when the concrete stress in the extreme compressive
fibre has reached the compressive strength of the concrete fc'. In terms of practical design, this is
an impossible situation as it implies that the bridge pier serviceability limit states criteria as well
as the ultimate limit states criteria occur for the same state of strain in the cross section.
The preceding study demonstrates the extremity of measures that would have to be taken in
order to ensure that Menn's proposed average crack spacing srm of 200 mm is unconservative.
This in turn validates Menn's recommendation for the purpose of practical preliminary design.

4.2.3 Identifying allowable stress in tensile reinforcement


As described in the previous subsection, translating the average crack width wcr to a sectional
moment and axial force requires that the average crack spacing is known. Furthermore, in order
to translate the average crack spacing to an average crack width requires that the strain at the
tensile reinforcement layer sm is also known. As stated earlier, Menn (1990) explains that it is
impossible to know the actual average strain at the reinforcement, as it is a function of many
complex factors. Menn recommends that a good approximation of the average strain at the
tensile reinforcement layer sm is eighty percent of the general strain in the reinforcement.
Equation 4-16 demonstrates Menn's recommendation, where s is the stress in the reinforcement
and Es is the Youngs modulus of the steel.

Equation 4-16

111
With an average strain in the tensile reinforcement sm known, it is possible to calculate the
allowable stress in the reinforcement at serviceability limit states based on crack width limits.
The CHBDC 2006 provides a formula for calculating crack width. The formula is shown in
Equation 4-17 where kb is a factor to account for type of coating on the reinforcement, and is
equal to 1.2 for epoxy-coated reinforcement and 1.0 for all others. c is a factor used to account for
whether the cracking is caused by imposed deformations or applied load; the factor also accounts
for the minimum dimension of the cross section. srm is the average crack spacing, and sm is the
average strain in the tensile reinforcement. For the purposes of this thesis kb was taken to be 1.2
and c was taken to be 1.7 in order to account for the most critical conditions.

Equation 4-17

The crack width limitations for serviceability limit states are provided by the CHBDC 2006
section 8.12.3.1, Table 8.6. The crack width limit, as prescribed by the CHBDC 2006, for nonprestressed members with typical exposure conditions is 0.25 mm.
Given the 0.25 mm crack width wcr and the 200 mm average crack spacing srm the allowable stress
in the tensile reinforcement s,allow was calculated as shown below:

112
From the limit on allowable stress in the tensile reinforcement s,allow a limiting state of strain can
be identified if the critical cross section shown previously in Figure 4-11 is assumed. The limiting
state of strain condition is shown in Figure 4-15.

Figure 4-15 Serviceability limit states state of strain: (a) prior to crushing of extreme compressive fibre, (b)
after crushing of extreme compressive fibre.

Figure 4-15 portrays the state of strain limitations as dictated by the allowable stress limit in the
tensile reinforcement s,allow. This model formulation was used to develop the software algorithms
used for computation in this body of work. The figure shows two state of strain diagrams at
serviceability limit states. Figure 4-15 (a) shows a state of strain for which the extreme
compressive fibre has not reached the crushing strain of the concrete c'; for this condition the
range of possible states of strain can be understood by visualizing the plane of strain as rotating
on the level of the tensile reinforcement, denoted as the point of constant strain in Figure 4-15
(a). The plane of strain can rotate both clockwise and counter-clockwise. Once the extreme
compressive fibre reaches the crushing strain, which occurs after sufficient clockwise rotation of
the plane of strain that is depicted in Figure 4-15 (a), the remaining possible states of strain can
only be identified by rotating the plane of strain around a new point of constant strain located
at the extreme compressive fibre. The new point of constant strain point and rotational

113
criterion are depicted in Figure 4-15 (b). In the case presented in Figure 4-15 (b) the plane of
strain can only rotate counter-clockwise, as any clockwise rotation would result in larger than
allowable stresses in the tensile reinforcement.
All reasonable designs of reinforced concrete bridge piers will never result in a situation where
serviceability limit states conditions occur for the same state of strain as an ultimate limit states
condition, such as the crushing of the extreme compressive fibre. As such, the state of strain
depicted in Figure 4-15 (b) is not a practical consideration for the purposes of design, but is
merely presented to demonstrate all possible states of strain.

4.2.4 Identifying allowable sectional moment and axial force


Traversing all of the possible states of strain depicted in Figure 4-15 in the previous subsection
allows for the identification of all possible combinations of moment MSLS and axial force NSLS at
serviceability limit states. By traversing and plotting all combinations an effective serviceability
limit states M-N interaction envelope can be generated. The creation of such a lone envelope at
serviceability limit states, however, is not as beneficial as it is for ultimate limit states. As stated
previously, the sectional forces at serviceability limit states are typically generated due to
imposed deformations; this dictates that all of the internal forces will in turn be dependent on
the flexural stiffness of the pier. The flexural stiffness of the pier will be constant as long as it
remains uncracked; however, serviceability limit states allows for a degree of cracking to occur,
as was previously shown, and as such the flexural stiffness of the pier no longer remains constant
in any cracked regions. The flexural stiffness in all cracked regions along the length of the pier
will also vary depending on the curvature at the particular location. For demonstrational
purposes, a normalized serviceability limit states M-N interaction envelope for a mechanical
reinforcement ratio equal to 0.1 is shown in Figure 4-16. A set of serviceability limit states M-N
interaction envelopes for a range of mechanical reinforcement ratios is shown in Appendix A.

114

Figure 4-16 Various M-N interaction envelopes based on limiting states of strain.

For the purpose of demonstrational completeness, Figure 4-16 also shows other M-N interaction
envelopes for various limiting states of strain. The envelope denoted "crushing" refers to the
conventional M-N interaction envelope which is limited by the states of strain defined by the
extreme compressive fibre reaching the crushing strain c' of the concrete. The envelope denoted
"yielding" refers to the reduced M-N interaction envelope, as proposed by Menn (1990), where
the states of strain are limited by either layer of reinforcement reaching the yield strain y. The
envelope denoted "cracking" refers to the states of strain limited by the extreme tensile fibre of
the section reaching the cracking strain of the concrete; for the purposes of simplicity and
conservatism in calculations, this strain has been assumed as zero. Lastly, the envelope denoted
"SLS" refers to the serviceability limit states limited states of strain which were explained in the
preceding subsection. The shown envelopes have all been normalized with respect to the cross
section and compressive strength of the concrete as shown in Figure 4-16. It is also noted that the
sectional forces have accounted for material resistance factors c and s as prescribed by the
CHBDC 2006. The concrete was accounted for in the normalization equation. The material
resistance factors used were 0.75 for concrete and 0.9 for reinforcing steel.

115
In order to calculate the imposed deformation limits for a reinforced concrete bridge pier, the
method of virtual work can be employed. Given the serviceability limit states M-N interaction
envelope denoted "SLS" in Figure 4-16, a set of limiting sectional forces for the critical section of a
reinforced concrete bridge pier can be identified. Based on these sectional forces the moments M
and flexural stiffnesses EI throughout the length of the pier can be calculated. The curvature
throughout the length of the pier can be calculated by dividing the moment M by the stiffness EI.
The moment Mv throughout the length of the pier for a virtual system where a unit load is
applied at the point of interest for lateral deformation is calculated. By integrating the curvature
of the real system with the moment Mv for the virtual system, the lateral displacement can be
calculated. Using the calculated lateral displacement, the additional moment generated by the
applied axial load N and the displacement is subtracted from the maximum moment M attained
from the initial moment applied at the base. The process is then repeated until a solution is
converged upon. A calculation example assuming a cantilever pier model is presented in
Appendix A.
By performing the steps outlined in Appendix A for a series of piers with various applied axial
loads, slenderness ratios, and reinforcement ratios, plots intended to be used as design aids for
preliminary dimensioning of reinforced concrete pier cross sections. An example design aid plot
for a mechanical reinforcement ratio of 0.1 for a normalized applied axial load n equal to 0.1 is
shown in Figure 4-17 (a). The steps previously described only take into account first-order effects.
A similar method but with a second level of iterations was performed in order to take account of
second-order effects; the resulting second-order design aid plot is shown in Figure 4-17 (b) on the
following page.

116

Figure 4-17 Serviceability limit states design aid example: (a) only first-order considerations, (b) first and
second-order considerations.

The curve denoted "first-order" in Figures 4-17 (a) and (b) has been computed using a first-order
analysis of a cantilever pier. Effectively this implies that the analysis was done for a pier that is
fixed at its base and has a pin connection at the top. The pier also effectively has a lateral
deformation applied at the pinned connection but is restrained against any further lateral
deformation. The tip is, however, free to deform in the vertical direction. In a pier system with a
superstructure, this would imply that the superstructure is rigidly fixed and has infinite axial
stiffness, but no flexural stiffness. A model for this system can be seen in Figure 4-18.

Figure 4-18 Serviceability limit states first-order model.

117
Figure 4-18 represents the upper-bound solution for a pier system. When the superstructure
shrinks and imposes deformations on piers, the point where imposed deformation is effectively
zero is regarded as the centre of stiffness of the system. Typically the centre of stiffness of the
system is located closest to the piers which have the greatest flexural stiffness (Menn 1990). This
implies that the most flexible piers will typically experience the greatest imposed deformation.
This of course is not a general rule as arrangement of piers, location of expansion joints in
superstructure, as well as connectivity between the piers and superstructure all influence the
location of the system's centre of stiffness. If a pier is flexible enough relative to the remainder of
the piers in the system then the scenario can effectively be modelled as is shown in Figure 4-18.
It is important to recognize that although second-order deformations effectively cannot exist in
the traditional sense for the pier system shown in Figure 4-18, the effective Euler buckling load
QE of the system is lower than it would be for a pier with no imposed deformation. The true
buckling mode of the system will be a combination of the buckling mode for a perfectly straight
pin-ended cantilever with no imposed deformation and the buckling mode for a free-standing
cantilever with an imposed deformation. The value of the Euler buckling load QE will be
somewhere in between the Euler buckling loads QE for the two idealized cases.
The curve denoted "second-order" in Figure 4-17 (b) has been computed using a second-order
analysis of the cantilever pier. This analysis was performed for a free-standing cantilever pier.
This simulates a pier with an imposed lateral deformation that is not restrained against any
additional deformations resulting from second-order effects. For a pier system this is a
conservative assumption as it assumes that the superstructure provides no lateral or vertical
restraint, implying that the superstructure effectively has no axial or flexural stiffness. A statical
model for this system can be seen in Figure 4-19 where w0 is the imposed first-order deformation
and w is the corresponding second-order deformation.

118

Figure 4-19 Serviceability limit states second-order statical model.

Figure 4-17 demonstrated that for slenderness ratios ranging up to 80, the difference in
normalized deformation limit for the first-order and second-order analyses is negligible. The
difference between the two analyses methods continues to remain marginal up until a
slenderness ratio of approximately 110. Between slenderness ratios of approximately 80 and 110
second-order effects become more influential and as such the imposed deformation limit when
considering second-order effects is effectively reduced. With slenderness ratios exceeding 110,
the second-order pier enters a slenderness range where second-order effects become prominent.
The second-order analysis also limits the pier to a maximum slenderness ratio of 150 based on
the Euler buckling load QE of the pier.
It is poor design practice to have piers go into a state where second-order effects are prominent
at serviceability limit states. With the intent of being used as a serviceability limit states design
aid, the graph denoted first-order in Figure 4-17 should be cut off at the location where elastic
buckling can occur. The buckling load QE was calculated based on the flexural stiffness EI at the
base of the column which was assumed to be constant throughout the entire length of the pier.
This is a reasonable and conservative assumption.

119
Design aids corresponding to these prescribed conditions have been developed for normalized
axial forces n ranging from 0 to 0.5. If n exceeds 0.5, serviceability limit states design is likely not
necessary since the magnitude of compressive force prevents the section from cracking prior to
ultimate limit states. The design aids have been developed for axial force n increments of 0.1 and
are presented on the following page in Figure 4-20. The value identifies the difference in
mechanical reinforcement ratios between successive curves. Only the bottom-most two curves
in each plot have a value of 0.1. For convenience, the graphs are also presented in Appendix B.
The proposed design curves will always be conservative; however, since it is desirable to avoid
any second-order effects at serviceability limit states, the use of the proposed conservative curves
for the purpose of preliminary design is warranted.

120

Figure 4-20 Serviceability limit states design aids.

121
Table 4-3 below summarizes the maximum slenderness ratio that can be designed for at
serviceability limit states for different magnitudes of normalized axial force n, based on
mechanical reinforcement ratio .
Table 4-3 Slenderness ratio limits for serviceability limit states based on mechanical reinforcement ratio.

Normalized
Axial Force n
Normalized
Axial Force n
Normalized
Axial Force n
Normalized
Axial Force n
Normalized
Axial Force n
Normalized
Axial Force n

4.3

0.0
0.1
0.2
0.3
0.4
0.5

Max

Max

Max

Max

Max

Max

0.1
0.1
0.1
158
0.1
138
0.1
130
0.1
116

0.2
0.2
0.2
173
0.2
150
0.2
138
0.2
127

0.4
0.4
0.4
0.4
171
0.4
155
0.4
142

0.6
0.6
0.6
0.6
0.6
170
0.6
152

0.8
0.8
0.8
0.8
0.8
0.8
165

1.0
1.0
1.0
1.0
1.0
1.0
-

1.2
1.2
1.2
1.2
1.2
1.2
-

Ultimate limit states design

Unlike serviceability limit states design which requires a more complex analysis of the respective
flexural stiffness EI throughout the length of a pier in order to determine sectional forces based
on imposed deformations, design at ultimate limit states can be simplified using Menn's method
of reduced M-N interaction diagrams that was previously presented in Chapter 3 of this thesis. By
employing Menn's method it is possible to design a pier at ultimate limit states by calculating the
sectional forces based on the assumption that the entire pier has a constant uniform reduced
flexural stiffness EIy. Design aids similar to those presented for the serviceability limit states case
in the preceding section have been developed for the ultimate limit states case. The development
of these design aids is discussed and presented in this section.

122

4.3.1 Sectional limits and deformation limits


As discussed previously, the sectional forces in the pier can be limited to those corresponding to
the reduced M-N interaction envelope, where the reduced M-N interaction envelope is generated
by limiting the strain in either top or bottom reinforcement layer to the yield strain of the steel
y. Any deformations can also be calculated by assuming that the flexural stiffness EI throughout
the length of the pier is equal to a constant reduced flexural stiffness EIy corresponding to a
particular combination of axial force and moment from the assumed M-N interaction envelope.
Based on these reduced flexural stiffnesses EIy, the second-order deformations can be calculated
using Vianello's equation, which was presented in Chapter 3, assuming that a first-order
deformation is known.
Menn (1990) proposed using one three-hundredth of the effective length of a pier as the firstorder deformation limit when designing for ultimate limit states. In Section 3.1.6, Menn's
recommended deformation was assessed and verified to be appropriate and sufficiently
conservative.
With a known first-order deformation and a known limiting state of strain, all first-order and
second-order sectional forces can be calculated assuming an initial reinforcement ratio and axial
force or moment. Since piers are typically regarded as axial members, it is appropriate to first
assume an axial force and calculate the moment based on this assumed axial force. Equation 4-18
is used to calculate the normalized moment demand m* based on a normalized applied axial load
n*, where h is the thickness of the cross section, nE is the normalized Euler buckling load, w0,max is
the assumed first-order deformation, and is a factor accounting for end-conditions of the pier.

Equation 4-18

123
Figure 4-21 shows the appropriate values of to be used for piers with various end conditions, as
adapted from Menn (1990). A free-standing pier has an value of 1.0.

Figure 4-21 factor for pier end-conditions: (a) fixed-fixed, (b) pin-fixed, (c) pin-pin. Adapted from Menn
(1990).

4.3.2 Developing design aids


The following subsection explains the process used to develop appropriate ultimate limit states
design aids. The design aids are similar in nature to those previously developed and presented for
serviceability limit states in that they allow designers to identify suitable preliminary values for
cross-sectional dimensions and reinforcement ratio. Figure 4-22 on the following page depicts an
example of a developed design aid diagram for ultimate limit states.

124

Figure 4-22 Ultimate limit states design aid example.

The design aid example shown in Figure 4-22 was developed for a double pin-ended pier system
with a value of 1.0. The ratio h/h is equal to 0.05. Each of the curves in the figure represents a
different value of mechanical reinforcement ratio . The figure may be used to aid in preliminary
design decisions by identifying the maximum normalized axial force nR that may be applied to a
pier of a given slenderness ratio for a specific mechanical reinforcement ratio . The axis
labelled "modified slenderness ratio" is adopted from Menn (1990) and is a more suitable value
for design purposes than the traditional slenderness ratio as it allows for efficient identification
of cross-sectional thickness. Just as the serviceability limit states design aid should only be used
for the purposes of preliminary design, so should the presented ultimate limit states design aid.
The presented design aids do not guarantee the most efficient design parameters but rather a
reasonable conservative estimate. In all cases, a more rigorous process should be adopted for
finalized design.

125
The following steps outline the process that was taken in order to construct the ultimate limit
states design aid presented in Figure 4-22. The same process, but with different effective
reinforcement lever arm ratios h'/h and different values, was used to develop all of the ultimate
limit states design aids presented in Appendix C.
1) Cross-sectional dimensions, pier length, and reinforcement ratio are defined.
2) The end-conditions of the pier are defined and a corresponding factor is determined
from Figure 4-21.
3) A reduced M-N interaction envelope is generated for the cross section by abiding by the
state of strain limitations.
4) A normalized applied axial load n* is assumed.
5) Given the assumed normalized applied axial load n*, the corresponding normalized total
moment m* is calculated using Equation 4-18, assuming a maximum initial deformation
w0,max equal to one-three-hundredth of the effective length kL of the pier.
6) The normalized total moment m* is compared to the normalized moment capacity mR of
the pier given the applied normalized axial load n*. This is compared using the reduced
M-N interaction envelope.
7) If the normalized total moment m* is equal to the normalized moment capacity mR of the
pier, the corresponding applied axial load n* and slenderness ratio are plotted for the
given mechanical reinforcement ratio , as defined in Step 1. If the normalized total
moment m* is not equal to the normalized moment capacity mR of the pier then Steps 4
through 7 are repeated until the demand equals the capacity.
8) The pier length is varied in order to vary the slenderness ratio and Steps 4 through 7 are
repeated until all modified slenderness ratios ranging from 0 to 50 have been plotted for
the given mechanical reinforcement ratio .

126
9) The amount of reinforcement in the cross section is varied and a new mechanical
reinforcement ratio is calculated, then Steps 4 through 8 are repeated for the new
mechanical reinforcement ratio .
10) Step 9 is repeated until curves for all desired mechanical reinforcement ratios have
been plotted.
A full set of ultimate limit states design aids has been developed for pier models with different
end conditions and configurations. The design aids are presented in Figure 4-23 and Figure 4-24
on the following pages. The full set of ultimate limit states design aids are also presented in
Appendix C.

127

Figure 4-23 Design aids for ultimate limit states 1.

128

Figure 4-24 Design aids for ultimate limit states 2.

129

4.4

Concluding remarks

This chapter further discussed the design recommendations and proposed initial eccentricity
limits made by Menn (1990), and compared them to the corresponding construction tolerances
as prescribed by CSA A23.1 (2004). Menns eccentricity recommendations were generally greater
in magnitude that the CSA A23.1A (2004) prescribed construction tolerances. For exceptionally
tall members, Menns recommendations were found to be several orders of magnitude greater
than the CSA A23.1 (2004) construction tolerances. The SiA 162 (1989), the governing document
which Menns recommendations are based on, proposes a 100 mm cut-off on the maximum
eccentricity. With this cut-off imposed, Menns recommendations are in significantly better
agreement with the CSA A23.1 (2004). The possibility of an inferred eccentricity based on the
0.75fc M-N interaction envelope cut-off prescribed in the CHBDC 2006 was addressed. The
values of inferred eccentricity for different pier cross-sectional thickness and reinforcement
ratios were compared to calculated eccentricities based on Menns method. The piers of the 22
bridges studied in Chapter 2 were compared to Menns eccentricity recommendation and the
calculated inferred eccentricities. The eccentricity recommended by Menn, in comparison to the
calculated inferred eccentricity, under-predicted the initial eccentricity for most of the 22 bridges
studied. Only in the cases of the most slender of the 22 bridges studied did Menns initial
eccentricity recommendation produce greater values than the calculated inferred eccentricity. It
was determined that the initial eccentricity recommended by Menn was most suitable when
considering more slender piers.
Based on the studies performed for initial eccentricities, serviceability limit states and ultimate
limit states preliminary design aids were developed. The design aids allow designers to select
appropriate cross-sectional dimensions and reinforcement ratios based on the magnitude of
applied axial load and pier slenderness ratio. The maximum possible slenderness ratio based on

130
normalized axial force and mechanical reinforcement ratio in a pier was identified. From the
development of the ultimate limit states design aids it was evident that the use of additional
reinforcement in pier cross-sections became less beneficial for exceptionally slender piers, such
as those with slenderness ratios in the range of 140 or greater. For these exceptionally slender
piers, the failure mode was governed by elastic buckling, thus little benefit could be gained from
the additional post-cracking stiffness provided with the additional reinforcement.

131

Chapter 5. Multiple Pier Systems


The previous chapter discussed the design of individual piers, both free-standing and fully
restrained, and briefly touched upon the behavior of a multiple pier systems. The term "multiple
pier system" refers to a system in which two or more piers are connected through an axially rigid
superstructure. In a multiple pier system, the relative flexural stiffnesses EI of the individual piers
will affect the behavior of the system as a whole. Although it is necessary to undertake the
exercise of designing the piers of the system on an individual basis, as presented in Chapter 3,
further considerations must be taken for the behavior of the pier system as a whole. This chapter
builds upon the individual design of piers in order to account for the considerations that must be
made when designing for the behavior of a multiple pier system as a whole. The design
recommendations are based on conservative approximations and are thus only appropriate for
the purposes of preliminary design.

5.1

Defining a multiple pier system

There are effectively four different pier types that can be designed for, of which only three are
considered when designing for the combined system behavior. The four types are denoted: fixedfixed, fixed-pin, fixed-roller, and pin-pin. The fixed-roller type pier does not contribute to the
combined system behavior as it is effectively a free-standing pier, taking axial load but not
moment from the superstructure. For the scope of this thesis, the pin-pin type pier has not been
considered as it is a very uncommon and the additional considerations that must be made during
the construction process typically make this an inefficient design (Menn 1990). A diagram
portraying these pier types and defining which types contribute to the combined system behavior
is shown in Figure 5-1.

132

Figure 5-1 Pier type and system.

In Figure, 5-1 piers A though D are part of a floating system. A floating system is defined as a
system where the lateral movement of the bridge is resisted solely by the piers without any aid
provided by the abutments. This type of system is achieved through the use of expansion
bearings at the abutments, denoted EXP. Piers A, B, and C represent typical piers that are
commonly used in construction. Although the monolithic pier connection, represented by pier A,
is somewhat less common in smaller highway projects due to the construction process required,
it is the most economically efficient in the long run if designed correctly (Menn 1990). The fixedpin and fixed-roller pier types represented by piers B and C, respectively, require the use of
bearings which must be maintained and contribute to additional costs throughout the life of the
bridge. Lastly, the pin-pin pier type represented by pier D is not frequently used. This pier type is
unable to carry any moment, thus all forces are carried through a state of pure axial compression
or tension; as mentioned previously this pier type is not considered in the scope of this thesis.
The piers in the figure have been categorized into the four possible pier types. The piers are also
identified as "SYSTEM" or "INDIVIDUAL" piers in order to identify whether that type of pier
influences the behavior of the system, and in an equal regard whether the system influences the
behavior of the pier. The multiple pier system can thus be defined as consisting of piers A, B, and
D. It is emphasized that the end connections are denoted as MONO., FIX. and EXP. in order to
accurately correspond to the conventional semantics used in construction drawings.

133

5.2

System buckling load

Flexible structural systems, such as floating systems, must have sufficient longitudinal stiffness to
ensure that deformations and second-order moments are not excessive (Menn 1990). Menn
(1990) recommends the global stability of the stiffness an effective measure of longitudinal
stiffness. This can effectively be defined by Equation 5-1, where uG is the second-order
longitudinal displacement of the superstructure,

is the first-order longitudinal displacement

of the superstructure, and E is the inherent factor of safety against buckling of the system.

Equation 5-1

The second-order moments M resulting from the second-order longitudinal displacement of the
superstructure uG can be calculated in a similar fashion, and can be expressed as shown in
Equation 5-2.

Equation 5-2

According to Menn, the global stability factor E should not be less than 3.0 upon final design.
Menn further states that in the cases where the global stability factor E of the system is
calculated prior to the design of reinforcement, E should not be less than 2.0. It is assumed that
the pier reinforcement will be increased after final design, thus increasing the global stability
factor E to a minimum of 3.0. The global stability factor E should be checked after final design of
the piers has been completed in order to make sure that this requirement is met.

134

5.2.1 Calculating the global stability factor of a system


The buckling load of the system is reached when for an arbitrary longitudinal displacement of
the superstructure uG, the sum of the horizontal forces Ht transferred from the superstructure to
the piers is equal to zero (Menn 1990). Menn recommends that a procedure based on this
condition can be used in order to calculate the global stability factor E. The system buckling load
can be expressed as the product of the pier loads Q and the unknown global stability or safety
factor E (Menn 1990). The horizontal forces transferred from the superstructure to the piers Hi
can likewise be expressed as functions of the system buckling load EQ and the longitudinal
displacement of the superstructure uG. The horizontal forces carried by each of the piers in the
system are then added and their sum equated to zero (Menn 1990). This condition is expressed in
Equation 5-3.

Equation 5-3

Since the horizontal force Hi carried by each pier can be expressed as a function of the system
buckling EQ, then the global stability factor E can be calculated, assuming that the flexural
stiffness EI of each pier is known. Since Menn recommends that the specified minimum
reinforcement should initially be used and the global stability factor E for this case should be
above 2.0, it is reasonable to assume the flexural stiffness EI of each pier to be constant
throughout its length and equal to the reduced flexural stiffness EIy as defined in preceding
chapters. Assuming that the cross-sectional dimensions of the piers have been selected and that
the lengths of the piers are defined by topographical conditions, the reduced flexural stiffness EIy
of each pier, based on specified minimum reinforcement, can be calculated using the M-N
interaction envelope of each pier. This method of calculating the reduced flexural stiffness EIy
was presented in Chapter 3 of this thesis.

135
The two types of piers considered in this thesis are the fixed-pin connection and the fixed-fixed
connection. Menn (1990) derived the equation relating the later deformation wT at the top of
each pier to the amount of horizontal force H transferred from the superstructure to each pier. In
his derivation Menn showed that the amount of horizontal force H transferred from the
superstructure to each of these pier types is equal if the horizontal deformation at the top of each
pier is equal and the cross-sectional dimensions, heights, reinforcement ratio, and material
properties of the piers are identical.
The horizontal force H transferred from the superstructure to the pier, for either of the two pier
types considered, is shown in Equation 5-4. In Equation 5-4, wT is the total lateral deformation at
the upper end of the pier, Q is the vertical load applied by the superstructure at the top of the
pier, L is the length of the pier, and E is the global stability or safety factor of the system.

Equation 5-4

It is reasonable to assume the superstructure does not experience any short-term longitudinal
deformations, since generally the axial stiffness of the superstructure is much greater than the
flexural stiffness EI of the piers. As such, we can deduce that the lateral deformation at the top of
each pier in the system is the same. That is to say that the value of wT for each pier in the system
is identical. Given this condition we can use Equations 5-3 and 5-4 to derive Equation 5-5.

Equation 5-5

136
A diagram for a two-pier system showing the loading conditions and assumed deformations, as
prescribed by Menn (1990), is shown in Figure 5-2 below.

Figure 5-2 Two-pier system model with loading conditions and assumed deformations.

Menn's design recommendation can be simplified into two simple inequation criterion for
preliminary design and final design checks. These criterion are presented under Equation 5-6,
where Ep is the system stability factor for preliminary design and Ef is the system stability factor
for final design.

Equation 5-6

Although Menn's method of providing a means of determining the system stability factor E is
both simple and elegant, it makes one potentially unconservative assumption. The method
inherently calculates the system stability factor E by assuming that the buckled shape of each
pier is based on a specific single mode. This mode may have a buckling load QE associated with it
which is potentially higher than the true buckling load QE of the pier in the system. This
potentially unconservative assumption is explained in fuller detail in the next subsection.

137

5.2.2 Limitations of Menn's method


Menn's method of calculating system buckling load assumes that every pier type in the system
can only buckle in one mode. In reality each pier type can buckle in one of two different modes
or a combination of these two modes. For the two pier types considered in this thesis, a diagram
of the possible modes of buckling as well as a representation of the true buckled shape is shown
in Figure 5-3.

Figure 5-3 Buckled shapes of pier types considered.

The effective length factors k for the pier types and their respective upper bound and lower
bound buckling modes are shown in Figure 5-3. The figure shows that the upper bound buckling
modes have higher k values than the lower bound buckling modes for both pier types considered.
This statement is somewhat misleading as it effectively implies that the buckling mode with the
higher slenderness ratio is less conservative than the buckling mode with the lower slenderness

138
ratio . In a situation where the pier acts individually outside of the system, this statement would
always be incorrect; however, in a multiple pier system where the collective behavior of the piers
is considered this is not the case.
In order to explain how simply assuming that the buckling mode with the higher slenderness
ratio is never a conservative assumption when calculating the system buckling load, we can
consider a two pier system in which both piers have the same height L, the same applied load
from superstructure Q, the same material properties, the same reinforcement, and are both of
the fixed-pin pier type. A diagram of this two pier system is presented in Figure 5-4.

Figure 5-4 Two pier system statical model.

There are two cases that should be considered for the two pier system arrangement shown in
Figure 5-4. Case one considers that both piers have identical cross-sectional dimensions and thus
the same flexural stiffness EI. Case two considers that one of the two piers has a significantly
larger cross section and thus a considerably greater flexural stiffness EI.
5.2.2.1

Case 1: both piers have the same flexural stiffness

If we consider that both piers A and B in Figure 5-4 have the same flexural stiffness EI, then given
that the applied load Q on each pier is the same, the piers will deform identically. This implies
that neither pier will resist any imposed deformations that may be transferred from the other

139
pier. Any forces generated by longitudinal displacement of the superstructure will be transferred
equally to both piers. In this case, Menn's method of calculating system buckling load is perfectly
accurate, as the piers act the same way they would if they were independent of the system.
5.2.2.2

Case 2: pier B is considerably stiffer than pier A

If we consider a situation where pier B has significantly greater flexural stiffness EI than pier A,
then given that the applied vertical load Q on each pier is the same and the superstructure is
assumed axially rigid, the lateral deformation at the tips of both piers will be the same. Since the
piers have different flexural stiffnesses EI, the horizontal load transferred from longitudinal
displacements in the superstructure will not be the same. Pier B, the pier with greater flexural
stiffness EI, will take a greater component of the load. Effectively, pier B provides additional
flexural stiffness EI to pier A. Menn's method of calculating the system buckling load inherently
assumes that pier B could provide an infinite amount of flexural stiffness EI to pier A and thus
provide stiffness to the system; this is, however, incorrect. If we assume that the applied load Q
exceeds the individual Euler buckling load QE of pier A, then regardless of how stiff we make pier
B and thus the system, pier A will buckle. The confusion in the previous statement made that
"the buckling mode with the higher slenderness ratio is less conservative than the buckling mode
with the lower slenderness ratio " arises from the fact that this statement inherently assumes that
the effective flexural stiffness EI of the pier in both buckling modes is the same. Effectively the
lower bound buckling mode, although having a higher effective length factor k, only relies on the
individual flexural stiffness EI of the pier being considered. The upper bound buckling mode,
although having a lower effective length factor k, relies on the cumulative stiffness of the system.
By relying on the cumulative stiffness of the system, the upper bound buckling mode inherently
has a higher buckling load than the lower bound buckling mode, even though the effective
length factor k is greater.

140
Calculating the true buckling load of the system is complicated as it is dependent on many
factors, some of which themselves require complex calculations. Perhaps one of the most
complicated factors to consider is the distribution of flexural stiffness EI throughout the system
and the influence on buckled shape geometry of individual piers. A method for calculating the
true system buckling load of a two pier system was developed during the completion of this
thesis. The method is not intended to be used for design purposes. The method serves to
demonstrate the computational complexity involved in calculating the true system buckling load,
even for a small two pier system. Analysis of larger systems was completed in this thesis, and was
done using the structural analysis software program SAP2000. The method developed served to
validate the analysis results produced by a SAP2000 model, thus giving confidence in results
produced by the program for larger systems. The general steps required for determining the true
system buckling load of a two pier system are described below.
5.2.2.3

Calculating true system buckling load

For simplicity, the same system presented previously in Figure 5-4 can be considered. Assuming
that pier B is considerably stiffer than pier A, it was explained that the buckling load of the
system will be lower than what is calculated using Menn's method. The degree to which Menn's
calculated system buckling load is lower than the true system buckling load will depend entirely
on the magnitude of the difference between the flexural stiffnesses EI of the two piers. Since
calculating the true buckling load of the system effectively implies that we are calculating the
true buckling load of pier A, as it will have the lower buckling load, we must find a way to model
the influence of the system on pier A.
It is possible to equate the flexural stiffness provided by pier B to a lateral spring connected to
the tip of pier A. The axial stiffness keff of the model spring would have to be calculated based on

141
the contribution of the flexural stiffness EI of pier B and the lateral force generated at the tip of
pier A as it deforms laterally. A diagram of the spring model is shown in Figure 5-5 below.

Figure 5-5 Spring model for a two pier system.

If we make the assumption that the piers have constant flexural stiffness EI throughout their
length then the equal spring stiffness keff can be calculated from the familiar equation for a
cantilever column with an applied point load, as shown in Equation 5-7.

Equation 5-7

Assuming the flexural stiffness EI of pier B is constant is only valid if the pier has not cracked and
if calculating the ultimate limit states capacity of the pier, in which case it was shown that the
flexural stiffness EI can be assumed to be constant and equal to the reduced flexural stiffness EIy.
In cases where pier A is exceptionally less stiff in flexure than pier B, pier B may indeed remain
uncracked and thus, it is appropriate to set the effective stiffness of the spring keff to what is
shown in Equation 5-7; however, typically this is not the case. In cases where the flexural stiffness
of pier B varies along its length, an effective spring stiffness keff may be calculated using the more

142
rigorous analytical method, as was explained in Chapter 3. The spring stiffness keff will itself be a
factor of the magnitude of the lateral deformation of pier A, and as such will not be constant and
require successive iterations of the rigorous analytical method, proposed in order to calculate.
Given a means of calculating the effective spring stiffness keff, the deformed shape of the pier can
be calculated. This process is in itself iterative; to the best of the authors knowledge there is no
closed form solution to calculating the buckled shape of the column. Using Nathan's method
(1985) as described by Salonga (2010), it is possible to divide pier A into any number of segments
along its length, and approximately calculate the relative lateral deformation of each segment
based on the curvature of the preceding segment and the use of Taylor series expansion. The
greater the number of segments used, the greater the accuracy of the calculation. The calculation
that was performed for the purposes of this thesis was done on a 20 segment model.
By assuming a lateral deformation wtip at the tip of pier A, the effective spring stiffness keff can be
calculated, and thus the equal lateral point load H applied to the tip of the column can be
calculated. Based on the given lateral deformation wtip, the equal applied lateral point load H, the
applied axial load Q, and the height of the pier L, the moment at the base segment MB of the pier
can be calculated as shown in Equation 5-8.

Equation 5-8

The same procedure as described by Salonga (2010) is then adopted. Details of Salonga's method
can be found in Chapter 3, section 3.2.3 of his thesis titled "Innovative Systems for Arch Bridges
using Ultra High-Performance Fibre-Reinforced Concrete".
The single difference between the method employed in this thesis and that which was used by
Salonga (2010) is that Salonga calculates the deflections assuming only one buckled shape of the

143
column, which in his case was perfectly correct. For the purposes of this thesis, the second
buckled shape is the one of interest. By iteratively calculating individual free body diagrams for
each of the segments along the length of pier A and ensuring equilibrium, the second buckled
shape can be calculated.
The process explained is computationally intensive and is thus not a reasonable method to be
employed for the purposes of preliminary design. The description presented merely serves to
better demonstrate the many factors that must be considered in order to calculate the true
system buckling load.
5.2.2.4

Comparison between SAP2000 and spring model

A sample study of a two pier system model, for which system buckling results obtained using
SAP2000 were compared to results obtained with the spring model outlined in subsection 5.2.2.3,
is discussed in this section. The model was based on two piers with identical cross sections. It
was assumed that both piers had constant flexural stiffness EI equal to their respective reduced
flexural stiffness EIy throughout their lengths. This assumption was predominantly made in order
to ensure that the fundamental assumptions made by Menn (1990) were preserved. Only one of
the piers in the model was loaded axially with a force of 40 000 kN. The reduced flexural stiffness
EIy of the pier was calculated using the moment curvature diagram for the given applied axial
load of 40 000 kN. The applied axial load of 40 000 kN was then scaled by a factor that started
at zero and was increased until the system buckling load was reached. The flexural stiffness EI of
the pier that was not loaded was then increased. This increase was calculated based on
incremental increases in reinforcement. The calculated system buckling loads based on: the
spring model, SAP2000, and Menn's method, as well as the upper bound and lower bound limits
for the pier model are presented in Figure 4-6.

144

Figure 5-6 Two pier system schematic and calculated system buckling loads based on various methods.

Figure 5-6 demonstrates that the results obtained using the SAP2000 buckling analysis model
and the proposed spring model are in agreement. The proposed spring model and SAP2000
results represent the true system buckling load. Menn's method is also in agreement up until the
flexural stiffness of pier 2 EI2 becomes significantly greater than the flexural stiffness of pier 1 EI1.
The first SAP2000 point presented on the plot in Figure 5-6 represents the situation when both
pier 1 and 2 have the same reinforcement ratio and thus the same flexural stiffness EI. Only when
the flexural stiffness of pier 2 EI2 exceeds the flexural stiffness of pier 1 EI1 by a factor of 2.8 does
the system buckling load calculated based on Menn's method differ by more than 5% from the
true system buckling load. A difference in flexural stiffness EI of 2.8 times would require that pier

145
2 have 12 times the area of steel that pier 1 has. From a design perspective it is impractical for two
piers with equal heights and end conditions to have such a large difference in reinforcement. As
such Menn's method m,ay in fact be a very reasonable way of approximating the system buckling
load for the purposes of preliminary design.
The described study was performed with two goals in mind: (1) to provide confidence in and
validate SAP2000's ability to accurately calculate the system buckling load based on a buckling
analysis and (2) to provide a better understanding of the limitations of Menn's method. Both of
these goals were accomplished as the study demonstrated that the SAP2000 model was generally
in agreement with the proposed spring model and that Menn's method is sufficiently accurate as
long as the relative flexural stiffnesses EI between piers are not drastically different.
In the next section, larger pier system models are analyzed using SAP2000 and the results are
compared to those predicted by Menn's method. The intent is to perform a parametric study in
order to better understand the range of flexural stiffnesses EI that Menn's method is suitable for.
By performing this parametric study, recommendations can be given to designers in terms of
deciding when it is appropriate to use Menn's method for analysis and when a more rigorous
analysis may be required.
All of the recommendations are to be used for preliminary design purposes. The intent is to give
designers a good "starting point" in terms of cross-sectional dimensions and reinforcement. The
full behavior of multiple pier systems will greatly depend on the loading conditions, which are
impossible to generalize. As such, final analysis and design of the pier system should be
performed post selection of preliminary dimensions and reinforcement.

146

5.3

Parametric study of large pier systems

The preceding study was performed for a two pier system. Although the study provided useful
insight towards understanding how the true system buckling load changes as the flexural
stiffness EI of an axially loaded pier becomes negligible relative to flexural stiffness of the system,
further parametric study needs to be done in order to understand how the system buckling load
may be affected in larger pier systems. This section will discuss extensive parametric studies that
were done for a pier system model with six piers that are part of a floating system. All of the
models were analyzed in SAP2000. The primary goal of the study was to identify how the
combination of individual pier flexural stiffness EI and applied axial load Q on individual piers
affects the system buckling load. It is hypothesized that in order for Menn's method to provide a
good approximation of the true system buckling load, the flexural stiffness EI of each pier in a
system will have to be proportioned based on the respective axial load Q that is applied to that
pier relative to the axial loads Q applied to the remaining piers in the system. If this hypothesis is
validated it will provide a range of design values for which Menn's method may be employed for
the preliminary design of flexible bridge pier systems.

5.3.1 SAP2000 model


This subsection discusses, in detail, the SAP2000 model that was built for the purposes of the six
pier floating system parametric study. The model was built assuming idealized conditions in
terms of the behavior of the superstructure relative to the piers. The superstructure is modelled
such that it carries no moment and effectively has infinite axial stiffness. The purpose of the
superstructure is to ensure that all of the pier tips displace identically. Although, expansion and
contraction of the superstructure will occur, these deformations will be minimal relative to the
deformations that can be expected to be seen prior to buckling of the system. The assumption

147
made about the superstructure implies that the piers take all of the flexure in the system. Figure
5-7 shows an illustrative representation of the SAP2000 model.

Figure 5-7 Illustrative representation of SAP2000 model.

All of the piers in the SAP2000 model had a fixed height of 16 m and were fixed at their base. The
connection between the tops of piers and superstructure varied as either hinged or fixed;
respectively, this equates to independent effective length factors of 0.7 and 0.5. All of the spans
between piers were set to 25 m. The piers had variable flexural stiffnesses EIi, where i denotes the
pier number to which the flexural stiffness relates to. Flexural stiffnesses EIi could be one of
select values associated with specific values of mechanical reinforcement ratios at an applied
load Q of 40000 kN. The 40000 kN load was selected since it corresponds to roughly fifty percent
of the normalized axial load capacity nR of the cross section with the minimum mechanical
reinforcement ratio considered. Assuming that the flexural stiffness EI did not deviate from
this value greatly simplified the analysis procedure while keeping results attained acceptable for
the purposes of preliminary design. All of the piers were modelled with the same cross sectional
dimensions and an assumed concrete strength of 25 MPa. The values of mechanical
reinforcement ratio used and their corresponding flexural stiffnesses EI are tabulated in Table
5-1. The piers all had different applied axial load Qi, where i denotes the pier number to which the
axial load is applied to. It is noted here that the entire structure was modelled in SAP2000

148
assuming steel material with the appropriate flexural stiffnesses EI adjusted to equal those of the
corresponding reinforced concrete cross section.
Table 5-1 Considered mechanical reinforcement ratios and their associated flexural stiffnesses EI.

Property

Value

(%)
EI (106 kNm2)

0.1
4.99

0.2
5.71

0.4
7.19

0.6
8.78

0.8
10.45

1.0
12.19

1.2
13.98

3.6
36.97

The parametric studies performed can be grouped into five categories: (1) all applied loads Qi are
the same and flexural stiffnesses EIi are variable for all piers but one, (2) only one pier has an axial
load Q, and variable flexural stiffness EI and the remainder of the piers have a fixed flexural
stiffness EI, (3) all piers have the same flexural stiffness EI and all piers but one have identical
variable axial load Q applied to them, (4) all piers have axial loads Q applied to them such that
the ratio between applied load Q and the flexural stiffness EI is the same for all piers, and (5) a
study for a pier system with a variable number of piers and a wider range of mechanical
reinforcement ratios . All piers were modelled, unless otherwise stated, as having a fixed base
and a hinged upper end, or a fixed base and a monolithic upper end, thus equating to individual
effective length factors kindi of 0.7 and 0.5, respectively. All of the parametric studies performed
are summarized in the following subsections.
5.3.1.1

Parametric study case 1:

The following parametric study was performed with the condition that all six piers had the same
applied axial load Q while the flexural stiffnesses EI of all piers except the pier denoted "Pier 1"
were varied. Prior to the completion of this parametric study, it was hypothesized that the
discrepancy between the system buckling load calculated using Menn's method for pier systems
and the value obtained using the SAP2000 model would decrease as the difference between
flexural stiffnesses EI of the piers decreased.

149
All of the applied axial loads Qi, structural parameters, and analysis results are presented in Table
5-2 on the following page. The standard deviation in buckling efficiency of the six piers was
calculated. The standard deviation was used to determine if the hypothesis that Menn's method
becomes more accurate as the difference in flexural stiffnesses EI decreases is correct.

150

Table 5-2 System buckling load analysis for parametric study case 1.

Pier 1

Pier 2

Pier 3

Pier 4

Pier 5

Pier 6
System
Buckling
Load
(kN)

Trial Number
3

Q1 (kN)
EI1 (106 kNm2)
kindi,1
Q2 (kN)
EI2 (106 kNm2)
kindi,2
Q3 (kN)
EI3 (106 kNm2)
kindi3
Q4 (kN)
EI4 (106 kNm2)
kindi,4
Q5 (kN)
EI5 (106 kNm2)
kindi,5
Q6 (kN)
EI6 (106 kNm2)
kindi,6
SAP2000
Menn

79 210
4.99
0.7
79 210
5.71
0.7
79 210
7.19
0.7
79 210
8.78
0.7
79 210
10.45
0.7
79 210
12.19
0.7
79 210
79 210

92 499
4.99
0.7
92 499
7.19
0.7
92 499
8.78
0.7
92 499
10.45
0.7
92 499
12.19
0.7
92 499
13.98
0.7
92 499
92 499

103 410
4.99
0.7
103 410
8.78
0.7
103 410
10.45
0.7
103 410
12.19
0.7
103 410
13.98
0.7
103 410
13.98
0.7
103 410
103 410

111 755
4.99
0.7
111 755
10.45
0.7
111 755
12.19
0.7
111 755
13.98
0.7
117 405
13.98
0.7
111 755
13.98
0.7
111 755
111 755

117 417
4.99
0.7
117 417
12.19
0.7
117 417
13.98
0.7
117 417
13.98
0.7
117 417
13.98
0.7
117 417
13.98
0.7
117 417
117 417

Error (%)

0.00

0.00

0.00

0.00

0.00

0.0466

0.0557

0.0645

0.0720

0.0775

Buckling
Efficiency
Qi/QE,i
Standard
Deviation

The bar colored black represents the control pier, if applicable; this is typically P1.

151
Table 5-2 shows a measure of the buckling efficiency of each individual pier. The buckling
efficiency effectively measures what percent of the Euler buckling load QE, calculated based on
the individual effective length factor kindi, is applied to the pier. The buckling efficiency is defined
in Equation 5-9, where Qi is the applied axial load on pier i and QE,i is the Euler buckling load of
pier i.

Equation 5-9

The data presented in Table 5-2 indicates that for a wide range of flexural stiffness EI variations,
Menn's method calculated the system buckling load perfectly; thus, the hypothesis, Menns
method performing better as difference in flexural stiffnesses decreases, could not be validated.
The next parametric case study considers the situation where all piers have a constant applied
axial load Q. The flexural stiffness EI of a single pier is varied while all others are kept constant.
The intent of the study was to observe how, for a reasonable range of flexural stiffnesses EI, the
parameters associated with a single pier influence the accuracy of Menn's method.
5.3.1.2

Parametric study case 2:

The following parametric study was performed with the condition that only the pier denoted
"Pier 1" had an axial load Q applied on it. The flexural stiffnesses EI of all other piers was kept
constant corresponding to a mechanical reinforcement ratio of 1.2. The flexural stiffness EI of
"Pier 1" varied corresponding to values of mechanical reinforcement ratio ranging from 0.1 to
1.2. Prior to the completion of this study, it was hypothesized that the discrepancy in results
obtained using Menn's method and the SAP2000 analysis would be smallest when the flexural
stiffness EI of "Pier 1" was equal to the flexural stiffness EI of the remaining piers in the system.
Data and results are presented in Table 5-3 on the following page.

152

Table 5-3 System buckling load analysis for parametric study case 2.

Pier 1

Pier 2

Pier 3

Pier 4

Pier 5

Pier 6
System
Buckling
Load
(kN)

Trial Number
3

Q1 (kN)
EI1 (106 kNm2)
kindi,1
Q2 (kN)
EI2 (106 kNm2)
kindi,2
Q3 (kN)
EI3 (106 kNm2)
kindi,3
Q4 (kN)
EI4 (106 kNm2)
kindi,4
Q5 (kN)
EI5 (106 kNm2)
kindi,5
Q6 (kN)
EI6 (106 kNm2)
kindi,6
SAP2000
Menn

709 281
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
709 281
808 401

670 021
12.19
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
670 021
791 146

563 237
8.78
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
563 237
758 334

490 838
7.19
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
490 838
742 935

364 552
4.99
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
364 552
721 756

Error (%)

13.97

18.08

24.43

51.36

97.98

0.2633

0.2852

0.3327

0.3544

0.3791

Buckling
Efficiency
Qi/QE,i
Standard
Deviation

The bar colored black represents the control pier, if applicable; this is typically P1.

153
Table 5-3 shows that for the single loaded pier case, Menn's method was significantly less
accurate than for the case when all piers had an axial load applied to them. The reasoning for this
once again stems from the fact that Menn's method does not consider the combination of the
upper bound and lower bound buckling modes. Furthermore, Menn's method assumes that the
system is insensitive to the location of applied axial load, and that the load is effectively
distributed throughout all of the piers in the system. Generally this assumption is incorrect;
however, from a practical design standpoint it may be perfectly valid. It is realistically impossible
for a situation to occur where any pier in the system has no axial load applied to it. The dead load
from the superstructure alone contributes to a large portion of the load that is applied to a pier.
Given that Menn's method performs poorly in situations where it is assumed that no axial load is
applied to specific piers, but has been shown to perform exceptionally well when the same axial
load is assumed to be applied to all piers in the system, it would be worthwhile to get a better
understanding in terms of limiting the difference between axial loads that are applied to piers in
the system.
5.3.1.3

Parametric study case 3:

The following study was performed in order to determine how the accuracy of Menn's method is
affected by differences in applied axial loads in piers. The preceding study showed that in cases
where only one pier is loaded, Menn's method is generally inaccurate for approximating the
system buckling load of larger pier systems. The following study was performed by assuming a
scalar factor for all piers. The value associated with the pier denoted "Pier 1" was always fixed
as 1.0, the value was identical amongst the remaining piers and varied from 0.1 to 0.5. The value
of effectively determined what % of the axial load applied to "Pier 1" was applied to the rest of
the piers. The data and results of the analysis are presented in Table 5-4 on the following page.

154

Table 5-4 System buckling load analysis for parametric study case 3.

Pier 1

Pier 2

Pier 3

Pier 4

Pier 5

Pier 6
System
Buckling
Load
(kN)

Trial Number
3

Q1 (kN)
EI1 (106 kNm2)
1
Q2 (kN)
EI2 (106 kNm2)
2
Q3 (kN)
EI3 (106 kNm2)
3
Q4 (kN)
EI4 (106 kNm2)
4
Q5 (kN)
EI5 (106 kNm2)
5
Q6 (kN)
EI6 (106 kNm2)
6
SAP2000
Menn

211 207
13.98
1.0
115 486
13.98
0.5
115 486
13.98
0.5
115 486
13.98
0.5
115 486
13.98
0.5
230 972
13.98
0.5
230 972
230 972

269 160
13.98
1.0
107 664
13.98
0.4
107 664
13.98
0.4
107 664
13.98
0.4
107 664
13.98
0.4
107 664
13.98
0.4
269 160
269 467

322 341
13.98
1.0
96 702
13.98
0.3
96 702
13.98
0.3
96 702
13.98
0.3
96 702
13.98
0.3
96 702
13.98
0.3
322 341
323 360

399 380
13.98
1.0
79 876
13.98
0.2
79 876
13.98
0.2
79 876
13.98
0.2
79 876
13.98
0.2
79 876
13.98
0.2
399 380
404 201

519 547
13.98
1.0
51 955
13.98
0.1
51 955
13.98
0.1
51 955
13.98
0.1
51 955
13.98
0.1
51 955
13.98
0.1
519 547
538 934

Error (%)

0.00

0.11

0.32

1.21

3.73

0.0429

0.0599

0.0838

0.1186

0.1736

Buckling
Efficiency
Qi/QE,i
Standard
Deviation

The bar colored black represents the control pier, if applicable; this is typically P1.

155
Table 5-4 indicates that if a single pier is heavily loaded and all piers in the system have the same
flexural stiffness EI, then as long as there is a minimal load applied on the remainder of the piers,
Menn's method is sufficiently accurate in approximating the buckling load of the system. In the
case where five piers only had 10% of the applied axial load that was placed upon "Pier 1", Menn's
method only had an error of 3.73%. The data presented in Table 5-4 further continues to indicate
that the error in applying Menn's method decreases as the standard deviation in the buckling
efficiency of the system decreases. The next parametric study explores a scenario where there is a
large difference in the flexural stiffness EI of two pier groups in a system but the standard
deviation in buckling efficiency of the system is constantly zero. This will give further insight into
how reliable of a criterion standard deviation in buckling efficiency is for setting a range of
validity for employing Menn's method.
5.3.1.4

Parametric study case 4:

The following parametric study was performed in order to better gauge how accurate of a
criterion the standard deviation in buckling efficiency of a system is for setting the range of
validity for employing Menn's method of approximating the buckling load of a system. The study
was done such that three piers in the system had a single variable flexural stiffness EI and the
other three piers had another common variable flexural stiffness EI. The first set of three piers
had one variable applied axial load Q and the other set had another common variable applied
axial load Q. The loads that were applied were selected such that the buckling efficiency of all
piers was identical, thus the standard deviation in buckling efficiency was consistently zero. If
the standard deviation in buckling efficiency of the system truly is an adequate way to gauge
when the use of Menn's method is appropriate for determining the buckling load of the system,
then there would be no difference in the results obtained using Menn's method and the results
obtained using SAP2000 analysis in any of the following parametric study trials. All data and
results for this parametric study are presented in Table 5-5.

156

Table 5-5 System buckling load analysis for parametric study case 4.

Pier 1

Pier 2

Pier 3

Pier 4

Pier 5

Pier 6
System
Buckling
Load
(kN)

Trial Number
3

Q1 (kN)
EI1 (106 kNm2)
kindi,1
Q2 (kN)
EI2 (106 kNm2)
kindi,2
Q3 (kN)
EI3 (106 kNm2)
kindi,3
Q4 (kN)
EI4 (106 kNm2)
kindi,4
Q5 (kN)
EI5 (106 kNm2)
kindi,5
Q6 (kN)
EI6 (106 kNm2)
kindi,6
SAP2000
Menn

134 734
13.98
0.7
134 734
13.98
0.7
134 734
13.98
0.7
100 781
10.45
0.7
100 781
10.45
0.7
100 781
10.45
0.7
134 734
134 734

134 734
13.98
0.7
134 734
13.98
0.7
134 734
13.98
0.7
84 613
8.78
0.7
84 613
8.78
0.7
84 613
8.78
0.7
134 734
134 734

134 734
13.98
0.7
134 734
13.98
0.7
134 734
13.98
0.7
69 253
7.19
0.7
69 253
7.19
0.7
69 253
7.19
0.7
134 734
134 734

134 734
13.98
0.7
134 734
13.98
0.7
134 734
13.98
0.7
54 971
5.71
0.7
54 971
5.71
0.7
54 971
5.71
0.7
134 734
134 734

134 734
13.98
0.7
134 734
13.98
0.7
134 734
13.98
0.7
48 100
4.99
0.7
48 100
4.99
0.7
48 100
4.99
0.7
134 734
134 734

Error (%)

0.00

0.00

0.00

0.00

0.00

0.0000

0.0000

0.0000

0.0000

0.0000

Buckling
Efficiency
Qi/QE,i
Standard
Deviation

The bar colored black represents the control pier, if applicable; this is typically P1.

157
Table 5-5 confirmed the hypothesis that as long as the standard deviation in buckling efficiency
of the system was zero, there would be no difference between the results obtained using Menn's
method and the results obtained using the SAP2000 analysis. This implies that Menn's method is
perfectly accurate for calculating the system buckling load for these cases. Since the four
preceding studies were done for very different loading conditions and flexural stiffness
arrangements, an additional study was performed in order to try and map an error envelope
relative to the standard deviation in buckling efficiency of the system. This additional study
considers mechanical reinforcement ratios as large as 3.6, which is unrealistic for practical
design. The study also considers a pier system with a variable number of piers instead of the
previously considered six pier system in order to determine if the same standard deviation in
buckling efficiency of the system relates to the same percent error, regardless of the number of
piers in the system.
5.3.1.5

Parametric study case 5:

The following study was done assuming pier systems with variable numbers of piers. The
mechanical reinforcement ratios used among piers ranged from 0.1 to 3.6. The applied load Q
was common amongst all piers. The flexural stiffness EI of all piers was variable. The intent of the
study was to assess a wide range of values of standard deviation in buckling efficiency of the
system in order to determine if a generalized relationship between the standard deviation and
percent error of Menn's method can be mapped. The study also considered the variable number
of piers in order to determine whether the same generalized relationship between standard
deviation in buckling efficiency of the system and error in Menn's method, if found, is applicable
to all pier systems regardless of the number of piers. The analysis values, configuration of piers
and results obtained are presented in Table 5-6.

158

Table 5-6 System buckling load analysis for parametric study case 5.

Pier 1

Pier 2

Pier 3

Pier 4

Pier 5

Pier 6
System
Buckling
Load
(kN)

Trial Number
3

Q1 (kN)
EI1 (106 kNm2)
kindi,1
Q2 (kN)
EI2 (106 kNm2)
kindi,2
Q3 (kN)
EI3 (106 kNm2)
kindi,3
Q4 (kN)
EI4 (106 kNm2)
kindi,4
Q5 (kN)
EI5 (106 kNm2)
kindi,5
Q6 (kN)
EI6 (106 kNm2)
kindi,6
SAP2000
Menn

176 769
36.97
0.7
176 769
13.98
0.7

235 759
36.97
0.7
0
13.98
0.7
235 759
4.99
0.7

286 324
36.97
0.7
0
13.98
0.7
0
13.98
0.7
286 324
4.99
0.7

323 880
36.97
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
323 880
4.99
0.7

176 769
180 862

235 759
248 299

286 324
315 595

323 880
382 962

347 426
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
0
13.98
0.7
347 426
4.99
0.7
347 426
450 329

Error (%)

2.32

5.29

10.22

18.24

29.62

0.2696

0.3234

0.3500

0.3590

0.3544

Buckling
Efficiency
Qi/QE,i
Standard
Deviation

The bar colored black represents the control pier, if applicable; this is typically P1.

159
Table 5-6 shows that there is no specific relationship between the standard deviation in buckling
efficiency of a system and the error in results obtained using Menn's method to calculate system
buckling load. The data suggests that typically higher values of standard deviation in buckling
efficiency of a system can be tolerated for pier systems with fewer piers than can be tolerated for
larger pier systems. By comparing the error in trial number 4 to the error in trial number 5,
indication is given that it is possible for larger pier systems to have a larger percent error than
smaller pier systems even if the standard deviation in buckling efficiency of the system is higher.
The data also answers the question of whether there is a general relationship between the
standard deviation in buckling efficiency of a system and the error in results obtained using
Menn's method when the number of piers in a system is specified. The six pier system shown in
Table 5-6 under trial number 5 has an error of 29.62% with a standard deviation of 0.3544. The
same standard deviation was found in Table 5-3 trial number 4 for a six pier system with a
different loading configuration; however, in this case the error was 51.36%. This shows that, for
the same standard deviation in buckling efficiency of a system, Menn's method can have different
degrees of error associated with it, even if the number of piers in each system is the same.
All of the preceding parametric studies were also completed for cases where the piers were
monolithically connected to the superstructure. The results obtained were identical as long as
the loading conditions and buckling efficiency of each individual pier was the same. A separate
set of tables has not been presented in the interest of brevity.
Figure 5-8 on the following page shows a summary of the parametric study values relating the
standard deviation in buckling efficiency of the system to the error in using Menn's method to
calculate the system buckling load.

160

Figure 5-8 Error in Menn's method compared to standard deviation in system buckling load efficiency of a
pier system.

Figure 5-8 shows that typically as the standard deviation in the buckling efficiency of a pier
system increases the error in using Menn's method to approximate the system buckling load also
increases. The relationship between the standard deviation and percent error appears to be
exponential. From the data collected, it appears that as long as the standard deviation in
buckling efficiency of the pier system is below approximately 0.1, there is effectively no error in
using Menn's method to calculate the system buckling load. Further investigation should be
done in order to determine if there are potentially more reliable factors to base the limits of
Menn's method on.
It was demonstrated in the parametric studies that it was typically very difficult to increase the
standard deviation in buckling efficiency of the system unless it was assumed that some of the
piers had no axial load applied to them. As discussed earlier, this is an unrealistic case as there
will always be dead load. Table 5-4 trial number 5 shows a case where five piers took 10% of the
load that was assigned to the remaining pier in a six pier system with the same flexural stiffness
EI attributed to each pier, and the trial showed only 3.73% error in using Menn's method. This
trial was a very unrealistic case yet, Menn's method performed exceptionally well for the
purposes of preliminary design. Fundamentally, the conclusion drawn from this study is that

161
every effort should be made during the design process to ensure that the system is designed as
efficiently as possible in terms of buckling. That is to say every pier should be loaded based on a
percent of its respective Euler buckling load QE. As long as each pier in a system is loaded
accordingly, the standard deviation in buckling load efficiency of the system will be equal to zero.
The preceding study has demonstrated that Menn's method of calculating the system buckling
load, and in turn providing a means of designing pier cross-sectional dimensions and
reinforcement ratio has been validated. The method proves to be an efficient way of calculating
the system buckling load and is generally accurate as long as the standard deviation in buckling
efficiency of the system is zero.

5.4

Concluding remarks

This chapter discussed the structural behavior of multiple pier systems. Menn's method of
calculating the global stability factor E of the system and using it as a condition for preliminary
design of pier systems was discussed. The potentially unconservative aspects, associated with
assuming lower bound buckling modes, of Menns method were discussed. The standard
deviation in buckling efficiency of the pier system was shown to be a reasonable measure of
determining when the use of Menn's method is appropriate and when the influence of potentially
unconservative buckling modes can be ignored. It was demonstrated that generally if the
standard deviation in buckling efficiency of a pier system remains below 0.1, Menns proposed
method is sufficiently accurate for the purposes of preliminary design. This chapter identified the
limitations and extents of validity of Menns simplified method for designing multiple pier
systems. The next chapter consists of a parametric study of the current state-of-the-art bridge
pier design in the industry.

162

Chapter 6. Conclusions
This chapter summarizes the conclusions of the preceding chapters and makes suggestions for
consideration in future studies.

6.1

Review of existing bridge piers

In Chapter 2, a study of 22 existing bridge piers was summarized. Reasonable ranges in concrete
strength, reinforcement ratio, and slenderness ratio were identified based on the bridges studied.
The identified ranges were used to determine scope of parameters considered in the remainder of
the thesis.
The relationship between visual slenderness and slenderness ratio was found to be of particular
interest. It was determined that through conscientious decisions, a pier could be designed to look
more slender and be more resistant to failures governed by slenderness. The fundamental goal
should be to design a pier with a height to thickness ratio which is as large as possible while
having a slenderness ratio that is a small as possible. Two ways of doing that were determined
from the study of the 22 bridges. First, cross sections should be designed to have as much area as
possible as far away from the geometric centroid, thus increasing the radius of gyration of the
section. Second, monolithic connections should be made between pier and superstructure
whenever possible, thus reducing the effective length factor.
Future studies should consider expanding upon the presented 22 bridge database in order to
identify relationships between design trends in the industry with a greater degree of confidence.

6.2

Approximate methods

In Chapter 3, the bilinear stress-strain approximation for concrete was briefly reviewed and
found to be a sufficiently accurate practical simplification of the generally accepted equivalent
rectangular stress block approximation of the parabolic stress-strain relationship for concrete.

163
The influence of reinforcement ratio and concrete strength on individual pier capacity was
analyzed. For slender piers, where flexural behavior governs, reinforcement ratio greatly
influenced the capacity of the piers up until piers with extremely large slenderness ratios
(exceeding 140) were considered; at this point additional reinforcement became much less
influential. The concrete strength was found to be much less influential and any increase in
flexural capacity of the pier due to increase in concrete strength was negligible.
Menns approximate method of calculating first-order and second-order deformations was
presented and discussed. A more rigorous analytical method, which takes account of material
nonlinearity, was developed and validated. Results obtained using Menns method were then
compared to results obtained using the more rigorous analytical method in order to identify if
potential shortcomings in the assumptions made in Menns method in order to waive
consideration for the effects of material nonlinearity. Menns simplification was found to be
exceptionally accurate for practical pier designs and was found to be conservative in all cases
where loads were applied. For piers with imposed deformations, Menns method was found to be
generally unconservative. It was thus recommended that Menns simplified method not be used
for the design of piers where imposed deformations are governing. In terms of designing piers
where applied load govern, Menns method was found to be overly conservative if extremely
slender piers are considered. The slenderness ratios of piers for which Menns method was found
to be overly conservative were in the range of 140. A limiting range for when Menns method was
appropriate for the use of preliminary design procedures was identified and documented. The
chapter concluded that Menns method was appropriate for ultimate limit states design of piers
for a broad range of slenderness ratios, if imposed deformations are not governing.
Future studies should consider the development of simplified design methods that better take
into consideration imposed deformations.

164

6.3

Individual bridge piers

In Chapter 4, the serviceability and ultimate limit states design of individual bridge piers was
considered. Initial eccentricity recommendations for the purposes of preliminary design made by
Menn were reviewed and compared to those prescribed by the CSA A23.1 (2004) and CHDBC
2006. Menns initial eccentricity recommendations were found to be overly conservative relative
to the construction tolerance limits prescribed by CSA A23.1 (2004). A 100 mm maximum
eccentricity cut-off prescribed by SiA 162 (1989), the governing document upon which Menns
recommendation is based on, was applied to Menns initial eccentricity recommendation. With
the cut-off included, Menns initial eccentricity assumption was in significantly better agreement
with the construction tolerances prescribed by CSA A23.1 (2004).
A potential inferred eccentricity associated with the 0.75fc compressive stress limit prescribed in
the CHBDC 2006 was discussed and explained. The limits for inferred initial eccentricity were
calculated and compared to Menns initial eccentricity recommendation. The piers of the 22
bridges studied in Chapter 2 were used to assess Menns recommended eccentricity and the
calculated inferred eccentricity. For most of the 22 bridges studied, Menns recommendation lead
to an under-predicted initial eccentricity relative to the calculated inferred initial eccentricities.
Three bridges were shown to be in the range where Menns recommended initial eccentricity
over-predicted, or in other words was more conservative than the calculated inferred
eccentricity. The three bridges were: the Pennsylvania Turnpike Commission Expressway, the
Reuss-Brcke Wassen (Uri 1971), and the Shin Chon Bridge (Starossek 2009); each of which
exemplify state-of-the-art piers in terms of slenderness. From this analysis, it was deduced that
Menns initial eccentricity recommendation of one-three-hundredth of the critical buckling
length of the pier was appropriate for the purposes of preliminary design.

165
Serviceability limit states preliminary design aids for individual piers were developed, allowing
designers to efficiently select appropriate cross-sectional dimensions and reinforcement ratios
based on normalized expected lateral maximum deformations of the piers. The maximum
slenderness ratio that could be designed for at serviceability limit states were identified as a
function of axial force and reinforcement ratio.
Ultimate limit states preliminary design aids for individual piers were also developed. These were
developed based on principles of Menns simplified method, which was presented in Chapter 3.
The design aids were developed to consider factors prescribed by the CHBDC 2006. The design
aids consider piers with different types of connections between the pier and superstructure.
Future studies should consider expanding the design aids provided in this thesis to piers with
different cross-sectional shapes, as those presented here are only valid for piers with rectangular
cross sections.

6.4

Multiple pier systems

In Chapter 5, the structural behavior of multiple pier systems was discussed. Menns method of
calculating the global stability factor and buckling load of a pier system was discussed. Menns
preliminary design recommendations were presented.
The potential unconservative aspects of Menns method, in terms of considering multiple modes
of buckling, were discussed. It was shown that Menns method generally assumes the buckling
mode with a higher buckling load associated with it, which can potentially be a grossly
unconservative assumption.
The concept of buckling efficiency was presented as a means of gauging whether or not Menns
method would underestimate the true buckling load of the system. With a parametric study
performed using SAP2000 models, it was determined that by limiting the standard deviation in

166
buckling efficiency of all of the piers in a system to 0.1, designers can ensure that Menns
approximate method for designing pier systems is sufficiently accurate for the purposes of
preliminary design.
Future studies should consider further analyzing the relationship between buckling efficiency of
the system and accuracy of Menns method. Future studies may also want to consider identifying
a different set of parameters for gauging when the second buckling mode becomes prominent.

6.5

Design recommendations

This section summarizes the general design recommendations for designing slender bridge pier
systems. The recommendations are based on the studies performed and conclusions made
throughout this thesis. The recommendations are as follows:

Whenever possible, monolithic connections should be made between the pier top and
superstructure. This will reduce the effective length factor of the pier to the minimum
possible value of 1.0 for a floating system, and 0.5 for a fixed system. This ensures that a
pier cross sectional design which is aesthetically slender as possible and is as
economically efficient as possible can be realized.

When designing slender piers, with slenderness ratios in the range of 90 to 140, the cross
section should be designed such that the maximum axial load demand is below the axial
force capacity corresponding to the balanced point condition of the pier. Doing so will
allow for easier control and calculation of second-order effects.

For the design of slender piers, with slenderness ratios of approximately 100, the use of
additional reinforcement provides a significant increase in flexural stiffness, and as such
may be economically warranted. For the design of exceptionally slender piers, with
slenderness ratios of approximately 140, the use of additional reinforcement does not

167
provide significant flexural stiffness, since the pier will often be in a range where minimal
cracking occurs prior to buckling; as such, the use of additional reinforcement may not be
economically efficient.

When designing for serviceability limit states, the pier should be designed such that no
significant second-order effects are occurring. This ensures that deformations and
vibrations can properly be accounted for and designed for.

For preliminary design purposes, assuming an initial eccentricity that is proportional to


the buckled shape of the pier and equal to one-three-hundredth of the effective buckling
length of the pier is a good approximation.

When designing multiple pier systems, the axial capacity of each pier in the system
should be proportional to the anticipated applied axial load to each pier. Keeping the
standard deviation in buckling efficiency of the piers to 0.1 or below ensures that all piers
in the system have the same buckled shape at failure, and thus ensures that the system is
economically efficient.

168

References
ACI Committee 318. (1995). Building code requirements for structural concrete: (ACI 318-95).
Farmington Hills, MI: American Concrete Institute.
Barrera, A. C., Bonet, J. L., Romero, M. L., and Miguel, P. F. (2011). Experimental tests of
slender reinforced concrete columns under combined axial load and lateral force
Engineering Structures, 33, 3676-3689.
Baant, Z. P. and Kwon, Y. W. (1994). Failure of slender and stocky reinforced concrete
columns: tests of size effect Materials and Structures, 27, 79-90.
Baant, Z. P., Cedonlin, L., and Mazen, R.T. (1991). New method of analysis for slender
columns. ACI Structural Journal, 88(4), 391-402.
Beal, A. N. (1986). The design of slender columns Proceedings Institution of Civil
Engineers, Structures and Buildings, 81, 397-414.
Bhavikati, S.S. (2005). Finite Element Analysis. New Age International Ltd. New Delhi, India.
British Standards Institution. (2008). Eurocode 2: design of concrete structures: British
standard. London: BSi.
Canadian Standards Association (1994). Concrete Materials and Methods of Concrete
Construction. CSA A23.1-94, Rexdale, Ontario.
Canadian Standards Association (2006a). CAN/CSA-S6-06: Canadian Highway Bridge Design
Code. CSA, Mississauga, Ontario.
Chuang, P. H., Anthony, T. C., and Wu, X. (1998). Modeling the Capacity of Pin-Ended
Slender Reinforced Concrete Columns Using Neural Networks Journal of Structural
Engineering, 124, 830-838.
Chuang, P. H. and Kong, S. K. (1997). Failure loads of slender reinforced concrete columns
Proceedings Institution of Civil Engineers, Structures and Buildings, 122, 364-366.
Chuang, P. H. and Kong, S. K. (1998). Strength of Slender Reinforced Concrete Columns
Journal of Structural Engineering, 124, 992-998.

169
Cranston, W.B. (1972). Analysis and design of reinforced concrete columns. Res. Rep. 20.
Cement and Concrete Association., London, United Kingdom.
Fdration Internationale du Bton (fib)/International Federation for Structural Concrete. (2010).
fib Bulletin 66: Model Code, Final Draft Volume 2. Switzerland. Fdration Internationale
du Bton
Kwak, H. G. and Kim, J. K. (2005). Nonlinear behavior of slender RC columns (1). Numerical
formulation Construction and Building Materials, 20, 527-537.
Kwak, H. G. and Kim, J. K. (2007). P- effect of slender RC columns under seismic load
Engineering Structures, 29, 3121-3133.
Mancini, G., Martinez y Cabrera, F., Pisani, M. A., and Recupero, A. (1998). Behavior of
Nonlinearly Restrained Slender Bridge Piers Journal of Bridge Engineering, 3, 126 -131.
Manzelli, A. A., and Harik, I. E. (1993). Prismatic and Nonprismatic Slender Columns and
Bridge Piers Journal of Structural Engineering, 119, 1133-1149.
Menn, C. (1990). Prestressed Concrete Bridges. Translated and Edited by P. Gauvreau.Basel:
Birkhuser.
Nathan, N. D. (1985). Rational analysis and design of prestressed concrete beam columns
and wall panels Prestressed Concrete Institute Journal. 30(3), 82-133.
Poston, R. W. (1986). Nonlinear Analysis of Concrete Bridge Piers Journal of Structural
Engineering, 112, 2041-2056.
Poston, R. W., Diaz, M., and Breen, J.E. (1986). Design Trends for Concrete Bridge Pier s
ACI Structural Journal, 14-20.
Rangan, B. V. (1990). Strength of reinforced concrete slender columns ACI Special
Publication, No. SP-128, American Concrete Institute, Detroit, Michigan, 851-862.
Salonga, J. A. (2010). Innovative Systems for Arch Bridges using Ultra High-Performance
Fibre-Reinforced Concrete. University of Toronto, Toronto, Canada.
SIA 162 Betonbauten . (1989). Einf hrung in die Norm SIA 162. ETH-Z rich. Z rich: SIA,
Schweizerischer Ingenieur- und Architekten-Verein.

170
Starossek, U. (2009). Shin Chon Bridge, Korea Structural Engineering International, 19(1),
79-84.
Timo, T. K. and Mirza, S. A. (2005). Nonlinear EI Equation for Slender Reinforced Concrete
Columns ACI Structural Journal, 839-848.
Vianello, 1898. Graphische Untersuchung der Knickfestigkeit gerader Stbe. Zeitschrift des
Vereins Deutscher Ingenieure. 1436-1443.
Wong, H.H.A. Discussion of ref. 1. Proceedings Institution of Civil Engineers Part 2, 83,
490.
Zienkiewicz, O. C., and Taylor, R. L. (1989). The finite element method. 4th Ed., McGraw-Hill Inc.,
New York, N.Y.

171

Appendix A
Supplementary Information

172

Derivation of Vianellos method of successive approximations


The following is a derivation of Vianellos equation as presented by Menn, 1990:

The figure above shows an axially loaded column with a hinge at its base and a roller at its tip.
The column has an initial eccentricity of w0(x), with the maximum eccentricity denoted wmax
(occurring at mid-height). Application of load Q results in first or moments M0. The first order
moments M0 are a function of the columns initial geometry:

The first order moments induce additional displacements w1. These displacements can be
calculated using the method of virtual work.

Where Mx(s) are the moments obtained when a unit lateral load is applied at point x. If we
assume that the flexural stiffness EI is constant throughout the length of the column, the above
equation can be rewritten as

173

The additional deflections due to M1 calculated using the method of virtual work are

Additional terms wi(x) are obtained in a similar manner. The total deflection w(x) can be
expressed as an infinite series

If the additional displacement w1 is proportional to the initial displacement w0, then

Since K is a function of only the initial displacement, it will be constant; Menn has expressed this
as K=l2/c for convenience, where c is a constant. It follows that

It follows that

Thus the equation for w(x) can be transformed into

If 0<<1, the infinite series can be summed. The total deflection will thus be

174
If =1, deflection w(x) will be infinite. It thus follows that the condition =1 corresponds to the
critical load QE.
Since w0(x) is proportional to w(x) for all values of , it follows that w0(x) must be proportional to
the buckled shape of the column. The critical load is thus given by

Since a value of =1 corresponds to an axial load Q equal to QE, the equation for w(x) can be
rewritten as

This is Vianellos equation of successive approximations.


The preceding derivation was adapted from Christian Menns book Prestressed Concrete Bridges
(1986)

175

Example Menns method and proposed rigorous analytical method

Menn's method
1) The reduced state of strain flexural

Rigorous analytical method


1) A lateral load H is assumed:

stiffness EIy is determined from the


data depicted in Figure 2-12:

2) The moment demand M* induced by the


lateral load H is calculated along the

2) A lateral load H is assumed:

length of column and identified at the


top and bottom of every tenth of the

3) The first-order deformation w0 due to


the lateral load H is calculated using
the reduced state of strain stiffness
EIy:

length of the column:


Segment
S1

M*top(kNm)

0
184.5
S3
369.0
S4
553.5
S5
738.0
S6
922.6
S7
1107.1
S8
1291.6
S9
1476.1
S10
1660.6
3) The corresponding curvature
S2

M*bot(kNm)

184.5
369.0
553.5
738.0
922.6
1107.1
1291.6
1476.1
1660.6
1845.1
at the

4) The reduced Euler buckling load QE is

top and bottom of each segment Si is

calculated using the reduced state of

determined from previously calculated

strain flexural stiffness EIy:

M- diagrams:

176
Segment

top

bot

(rad/mm)

(rad/mm)

S1

0
3.43E-08
3.43E-08 6.86E-08
S3
6.86E-08 1.03E-07
S4
1.03E-07
1.37E-07
S5
1.37E-07
1.71E-07
S6
1.71E-07
2.06E-07
S7
2.06E-07 2.40E-07
S8
2.40E-07 2.74E-07
S9
2.74E-07 3.09E-07
S10
3.09E-07 3.43E-07
Being representative of the moment
S2

5) The

total

deformation

wtot

is

calculated using Vianello's method:

demand M* up until a segment Si cracks,


the curvature across the length of the
column is linear.
4) The virtual moment demand Mv* from a
unit lateral load applied at the tip of the
6) The total moment demand M*tot is
calculated and compared against the
reduced

state

of

strain

moment

capacity of the column MR-menn:

The moment demand M*tot exceeds


the moment capacity MR-menn by a
marginal 235 kNm. For the purposes
of this calculation example it is safe to
conclude that the maximum lateral
load H that can be applied to the
column is approximately 38 kN.

column is calculated:
Segment

Mv*top(kNm)

Mv*bot(kNm)

S1

0
2.223
2.223
4.446
S3
4.446
6.669
S4
6.669
8.892
S5
8.892
11.115
S6
11.115
13.338
S7
13.338
15.561
S8
15.561
17.784
S9
17.784
20.007
S10
20.007
22.230
5) According to underlying theory of the
S2

method of virtual work, the sum of the


integral of the virtual moment M*v and
the actual system curvature of all
segments

will

deformation
column:

equal

wtip

of

the

total

tip

the

cantilever

177
The lateral deformation w at each
segment is presented below:
Segment

wtop(mm)

wbot(mm)

S1

0.680
0.496
0.496
0.348
S3
0.348
0.233
S4
0.233
0.147
S5
0.147
0.085
S6
0.085
0.0436
S7
0.0436
0.0184
S8
0.0184
0.005
S9
0.005
0.001
S10
0.001
0.000
6) The additional moment due to the tip
S2

deformation wtip and the applied axial


load Q is calculated at each segment. The
moment is then added to moment
demand M* due to the applied lateral
load H.
7) Steps 2 through 6 of the procedure for
the method of virtual work are repeated
until successive repetitions of the steps
cease to produce additional deformations
or the base segment S10 reaches the gross
cross-sectional moment capacity MR-gross.
If the deformations converge and the
total moment calculated in Step 6 does
not

equal

the

gross

cross-sectional

moment capacity MR-gross then the lateral


load H applied in step 1 is increased and
the same procedure is repeated.

178

Example Cantilever pier example for serviceability limit states


The following steps outline the calculation process in determining the maximum imposed lateral
deformation that can be accommodated for by a cantilever pier under serviceability limit states.
1) An applied axial load N* to the system is identified. This is generally based on the
serviceability limit states loading conditions prescribed in governing codes, in this case
the CHBDC 2006. For this example a 18000 kN load was assumed.
2) Pier cross section and length are identified. In this case the cross section presented below.

3) Given the applied axial load N* of 9000 kN, a normalized axial load n* is calculated.

4) Given the normalized axial load n*, the maximum normalized moment m is determined
from the corresponding serviceability limit states M-N interaction envelope.

5) The normalized corresponding maximum moment m is applied at the base of a cantilever


pier with the corresponding applied axial load n* applied at the top of the pier. a diagram
of the loading conditions is shown on the following page.

179

6) The flexural stiffness EI along the length of the pier is identified and corresponding
curvatures are calculated.
7) A virtual system with an applied lateral unit load at the tip of the cantilever is created and
the corresponding virtual bending moment Mv diagram is calculated. A schematic is
shown below.

8) The virtual bending moment Mv is normalized with respect to the cross-sectional


dimensions and concrete strength.
9) The real curvature and virtual bending moment Mv are integrated and the lateral
deformation is found.
10) The additional moment due to the eccentricity of the applied axial load n* is subtracted
from the applied normalized moment m*, this new moment is denoted as m**.

180
11) Steps 1 through 9 are repeated assuming the new applied moment m**; this time the
additional moment due to eccentricity of applied axial load n* is added onto the original
applied moment m** since m** is lower than the allowed moment m.
12) Steps 1 through 10 are iteratively performed until the final moment, after the addition of
moment due to the eccentricity of applied axial load n*, equals the allowed moment m.

181

Serviceability limit states M-N interaction envelopes

182

Table of bridge study data


Year of
Design

Pier Cross
Section
Shape

1970

Circular

1972

Rectangular

1974

Rectangular

1977

Hollow
Rectangle

1979

Circular

1981

Circular

1981

Oval

1982

Circular

1982

Circular

1983

Rectangular

1988

Circular

Name of Bridge
Kings Highway
No.401 & 2A
Interchange
Reuss-Brcke
Wassen
Kings Highway
No.II
Islington Avenue
Overpass
Highway 404
CNR Overhead
CNR Overhead
Highway No.69
Turning
Roadway N. To
409 E.
Highway 403 E.B.
Express over
Highway 401 E.B.
Collector
Highway 403 E.B.
Express over
Highway 410
Highway 7
Underpass at
Dufferin Street
Highway 401
Morningside
Ave. Underpass

Slenderness
Ratio
(kL/r)

Reinforcement
Ratio
(%)

Unsupported Top
of Superstructure
Surface Area
(m2)

Pier
Height
(m)

Concrete
Strength
(MPa)

Visual
Slenderness
Ratio
(L/h)

Reference

66.11

4.71

447.6

7.561

35

8.26

Construction
Drawings

100.80

0.40

684.9

32.000

25

29.09

121.63

1.34

737.2

24.095

27.6

17.55

30.94

N/A

1189.2

13.725

41.3

5.71

91.87

2.84

412.5

10.507

27.6

11.48

39.53

1.02

497.9

4.447

30

4.94

36.79

1.12

863.4

10.352

30

5.14

Construction
Drawings

34.55

1.03

730.8

8.638

35

4.32

Construction
Drawings

40.18

1.08

573.0

7.534

35

5.03

Construction
Drawings

16.37

0.58

265.4

7.085

35

9.45

Construction
Drawings

37.25

2.04

1587.0

6.984

34.5

4.656

Construction
Drawings

Construction
Drawings
Construction
Drawings
Construction
Drawings
Construction
Drawings
Construction
Drawings

183

1990

Rectangular

1990

Rectangular

1990

Circular

1990

Hollow
Rectangle

1995

Circular

1996

Irregular
Circular

2004

Other

2007

Other

1997

Circular

2007

Circular

2007

Hollow
Rectangular

Highway 403
Upper Middle
Road Underpass
Englehart River
Bridge Highway
560 Crossing
Ramp 403/W QEW/E
State of Hawaii,
Interstate Route
H-3 Windward
Viaduct
Big Qualicium
River Bridge No.
3051 Steel
Alternative
I-93 Southbound
Viaduct Concrete
Alternative
Applewood
Crescent Bridge
Caroni Bridge
Pennsylvania
Turnpike
Commission
Expressway
Ramp 401W
Collector - 404N
Over Ramp
401W Express
Shin Chon
Bridge

17.44

1.30

906.7

7.475

30

6.23

Construction
Drawings

71.91

0.95

199.2

6.100

30

10.17

Construction
Drawings

41.30

1.06

680.3

12.383

35

5.16

Construction
Drawings

57.06

1.05

1300.0

43.683

30

10.24

Construction
Drawings

120.73

0.80

1571.7

30.183

27.6

12.09

Construction
Drawings

54.64

2.65

959.8

13.106

27.6

7.16

Construction
Drawings

57.93

1.98

667.7

8.690

50

7.24

49.10

1.17

224.4

8.232

30

8.86

110.17

2.05

1249.1

56.243

27.6

20.504

Construction
Drawings

28.26

1.81

317.7

5.298

50

6.14

Construction
Drawings

86.40

3.20

2142.0

88.400

N/A

31.57

Structural
Engineering
International

Construction
Drawings
Construction
Drawings

184

Table of bridge references


Bridge Name
Kings Highway No.401 & 2A
Interchange

Reference Document
DHO 1970. Kings Highway No. 401 & 2A Interchange. (Bridge Plans).
Scarborough: Department of Highways Ontario.

Reuss-Brcke Wassen

Uri 1971. Reuss-Brcke Wassen. (Bridge Plans). Altdorf: Tiefbauamt des


Kantons Uri.

Kings Highway No.II

MTO 1974. Englehart River Bridge at Englehart. (Bridge Plans). Englehart:


Ministry of Transportation Ontario.

Islington Avenue Overpass

Metro 1977. Islington Avenue Overpass (Bridge Plans). Toronto: Municipality


of Metropolitan Toronto Department of Roads and Traffic.

Highway 404 CNR Overhead

MTO 1979. Highway 404 CNR Overhead. (Bridge Plans). Toronto: Ministry of
Transportation Ontario.

CNR Overhead Highway No.69

MTO 1981. CNR Overhead HWY No. 69 Approx 3.5 km N. Parry Sound. (Bridge
Plans). Parry Sound: Ministry of Transportation Ontario.

Turning Roadway N. To 409 E.

MTO 1981. Turning Roadway N. To 409 E. (Bridge Plans). Toronto: Ministry of


Transportation Ontario.

Highway 403 E.B. Express over


Highway 401 E.B. Collector

MTO 1982. Hwy. 403 E.B. Express over Hwy. 401 E.B. Collectors. (Bridge Plans).
Toronto: Ministry of Transportation Ontario.

Highway 403 E.B. Express over


Highway 410

MTO 1982. Hwy. 403 E.B. Express over Hwy. 410 N.B. & Ramp S-W (Bridge No.
35). (Bridge Plans). Toronto: Ministry of Transportation Ontario.

Highway 7 Underpass at Dufferin


Street

MTO 1983. Highway 7N Underpass at Dufferin Street. (Bridge Plans). Toronto:


Ministry of Transportation Ontario.

Highway 401 Morningside Ave.


Underpass

MTO 1988. Hwy. 401/Morningside Ave. Underpass. (Bridge Plans). Toronto:


Ministry of Transportation Ontario.

Highway 403 Upper Middle Road


Underpass

MTO 1990. Hwy. 403/Upper Middle Road Underpass. (Bridge Plans). Oakville:
Ministry of Transportation Ontario.

Englehart River Bridge Highway 560


Crossing

MTO 1990. Englehart River Bridge Hwy. 560 Crossing. (Bridge Plans).
Englehart: Ministry of Transportation Ontario.

Ramp 403/W - QEW/E

MTO 1990. Ramp 403/W QEW/E (BR.41) Over QEW and Ramp QEW/S
403/E. (Bridge Plans). Toronto: Ministry of Transportation Ontario.

State of Hawaii, Interstate Route H3 Windward Viaduct

Hawaii 1990. Interstate Route H-3 Windward Viaduct. (Bridge Plans). Hawaii:
State of Hawaii.

Big Qualicium River Bridge No. 3051


Steel Alternative

BC 1995. Big Qualicium River Bridge No. 3051 (Steel Alternative). (Bridge
Plans). Vancouver: Province of British Columbia Ministry of Transportation
and Highways.

I-93 Southbound Viaduct Concrete


Alternative

NY 1996. I-93 Southbound Viaduct Concrete Alternative. (Bridge Plans).


Weschester County: State of New York Department of Transportation

Applewood Crescent Bridge

Vaughan 2004. Applewood Crescent Bridge (Bridge Plans). Vaughan: City of


Vaughan Engineering Department.

Caroni Bridge

NIDC 2007. Caroni Bridge Widening (Bridge Plans). San Juan: National
Infrastructure Development Company.

Pennsylvania Turnpike Commission


Expressway

PTC 2007. Pennsylvania Turnpike Commission MON/FAYETTE Expressway I70 to PA-51 SR 0043 Section 52B2 (Bridge Plans). Pennsylvania: Pennsylvania
Turnpike Commission.

Ramp 401W Collector - 404N Over


Ramp 401W Express

MTO 2007. Ramp 401W Collector 404N Over Ramp 401W Express DVPS.
(Bridge Plans). Toronto: Ministry of Transportation Ontario.

Shin Chon Bridge

Starossek, U. (2009). Shin Chon Bridge, Korea Structural Engineering


International, 19(1), 79-84.

185

Appendix B
Serviceability Limit States Design Aids

186

serviceability limit states design aids.

187

Appendix C
Ultimate Limit States Design Aids

188

ultimate limit states design aids 1.

189

ultimate limit states design aids 2.

You might also like