Live Load Models For Long Span Bridges

University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln

Civil Engineering Theses, Dissertations, and
Civil Engineering
Student Research
12-2009
Live Load Models for Long Span Bridges

Marta Lutomirska
[email protected]
Follow this and additional works at: http://digitalcommons.unl.edu/civilengdiss

Part of the Civil Engineering Commons
Lutomirska, Marta, "Live Load Models for Long Span Bridges" (2009). Civil Engineering Theses, Dissertations, and Student Research. 1.
http://digitalcommons.unl.edu/civilengdiss/1
This Article is brought to you for free and open access by the Civil Engineering at DigitalCommons@University of Nebraska - Lincoln. It has been
accepted for inclusion in Civil Engineering Theses, Dissertations, and Student Research by an authorized administrator of
DigitalCommons@University of Nebraska - Lincoln.
LIVE LOAD MODELS
FOR LONG SPAN BRIDGES
by
Marta Lutomirska
A DISSERTATION
Presented to the Faculty of

The Graduate College at the University of Nebraska
In Partial Fulfillment of the Requirements
For the Degree of Doctor of Philosophy
Major: Engineering
Under the Supervision of Professor Andrzej S. Nowak
Lincoln, Nebraska
December, 2009
LIVE LOAD MODELS
FOR LONG SPAN BRIDGES
Marta Lutomirska, PhD
University of Nebraska, 2009
Advisor: Andrzej S. Nowak
In the doctoral dissertation a live load model for long span structures was derived.
The live load model is valid for spans between 600 ft and 5000 ft and it is intended to
reflect current traffic patterns, quantities of trucks and their weights. The live load models
available were developed for short and medium span bridges. Those models were not
appropriate for long span bridges due to different types of structure and critical traffic
patterns. Live load on long spans depends on traffic mix. One heavily overloaded truck
does not have significant influence. Moreover, the continuous increase in the number of
the trucks, their weights, and high percentage of overweight trucks led to a search for the
newest traffic data. The database includes variety of sites within many different states. A
numerical procedure was developed to process the database and simulate traffic jam
situations. From the simulation the values of uniformly distributed load were derived.
Trucks were kept in actual order, as recorded in the WIM surveys. Results of the
simulations were plotted as a cumulative distribution function of uniformly distributed
load for considered span lengths. For longer spans, uniformly distributed load decreases
and is closer to the mean value. The bias factors were calculated for the heaviest 75-year
combination of vehicles. The 75-year uniformly distributed loads were derived from
extrapolated distributions. It was stated that for most of the bridges current live load HL-
93 is appropriate. It was also noticed that some bridges, characterized by high ADTT and
increased percentage of overloaded loaded vehicles, require special attention and
application of increased design live load. The developed live load model is recommended
to be taken into consideration in the bridge design code.

© Marta Lutomirska
All rights reserved
2009
iii
DEDICATION
To My Family
iv
ACKNOWLEDGMENTS
I wish to express my deepest gratitude to Professor Andrzej S. Nowak, my

academic advisor and chairperson of my dissertation committee, for his support and
guidance throughout my graduate study at the University of Nebraska. My appreciations
also go to a member of my doctoral committee, Dr. George Morcous, who advised and
encouraged me in my study. Special thanks go to Dr. Maria M. Szerszen and Dr.
Elizabeth G. Jones, members of my reading committee for their effort and time spent on
reviewing my dissertation and offering helpful suggestions.
The Department of Civil Engineering at the University of Nebraska-Lincoln is

acknowledged for the study and research opportunities they provided.
I am deeply grateful to my husband, Tomasz, my parents, and whole family for

their great support and encouragement.
v
TABLE OF CONTENTS
ABSTRACT .................................................................. Error! Bookmark not defined.

DEDICATION .............................................................................................................. iv
ACKNOWLEDGMENTS ...............................................................................................v
LIST OF FIGURES ..................................................................................................... viii
LIST OF TABLES ....................................................................................................... xii
CHAPTER 1 INTRODUCTION......................................................................................1
1.1. PROBLEM STATEMENT ......................................................................................1
1.2. OBJECTIVE AND BENEFITS OF THE STUDY ..........................................................2
1.3. ORGANIZATION OF THE DISSERTATION ..............................................................4
1.4. PRIOR INVESTIGATIONS .....................................................................................5
CHAPTER 2 LIVE LOAD IN CURRENT DESIGN CODES........................................ 11
2.1. INTRODUCTION ............................................................................................... 11
2.2. INTERNATIONAL PROVISIONS FOR LIVE LOADING ............................................ 12
2.3. PROVISIONS FOR DYNAMIC LOAD FACTOR....................................................... 20
2.4. PROVISIONS FOR MULTILANE REDUCTION FACTORS ........................................ 21
2.5. COMPARISON OF EQUIVALENT UNIFORMLY DISTRIBUTED LOADS..................... 23
CHAPTER 3 STRUCTURAL RELIABILITY PROCEDURES .................................... 29
3.1. INTRODUCTION ............................................................................................... 29
3.2. STANDARD VARIABLES AND PROBABILITY DISTRIBUTIONS ............................... 30
3.3. LIMIT STATE FUNCTION ................................................................................... 33
3.4. RELIABILITY INDEX......................................................................................... 35
3.5. MONTE CARLO METHOD SIMULATION TECHNIQUE .......................................... 37
3.6. NORMAL PROBABILITY PAPER ......................................................................... 39
CHAPTER 4 TRAFFIC DATA ..................................................................................... 42
4.1. REGULATIONS OF TRUCK TYPES, TRUCK SIZES AND WEIGHT LIMITS ................ 42
4.2. DATA COLLECTION METHODOLOGY (WIM) .................................................... 49
4.3. WEIGH IN MOTION DATABASE ........................................................................ 51
CHAPTER 5 DEVELOPMENT OF LIVE LOAD MODEL .......................................... 61
5.1. INTRODUCTION ............................................................................................... 61
vi
5.2. MODEL BASED ON AVERAGE 5-AXLE TRUCK ..................................................61
5.3. MODEL BASED ON LEGAL LOAD TRUCKS ........................................................ 65
5.4. MODEL BASED ON TRAFFIC JAM SIMULATION USING WIM DATA .................... 66
CHAPTER 6 MULTIPLE PRESENCE ......................................................................... 77
6.1. INTRODUCTION ............................................................................................... 77
6.2. STUDIES ON PRESENCE OF MULTIPLE TRUCKS .................................................78
6.3. MULTIPLE PRESENCE OF TRUCKS BASED ON THE VIDEO FILES OF TRAFFIC ....... 81
6.4. APPROACHES TO MULTILANE REDUCTION FACTORS ......................................... 86
6.5. CONCLUSIONS................................................................................................. 88
CHAPTER 7 DYNAMIC FACTOR ..............................................................................89
7.1. INTRODUCTION ............................................................................................... 89
7.2. STUDIES ON PARAMETERS AFFECTING DYNAMIC BRIDGE RESPONSE ................ 90
7.1. BRIDGE-VEHICLE INTERACTION MODEL AND DERIVATION OF DYNAMIC FACTOR
....................................................................................................................... 98
7.2. CONCLUSIONS ............................................................................................... 105
CHAPTER 8 RELIABILTY ANALYSIS OF SUSPENSION BRIDGE....................... 106
8.1. RELIABILITY ANALYSIS PROCEDURE ............................................................. 106
8.2. SELECTION OF REPRESENTATIVE STRUCTURE, ELEMENT AND LIMIT STATE
FUNCTION ................................................................................................................ 107
8.3. NOMINAL RESISTANCE .................................................................................. 108
8.4. LOAD MODEL ............................................................................................... 113
8.5. RELIABILITY RESISTANCE MODELS ............................................................... 120
8.6. LOAD MODEL ............................................................................................... 126
8.7. RELIABILITY ANALYSIS ................................................................................. 126
CHAPTER 9 SUMMARY & CONCLUSIONS ........................................................... 129
CHAPTER 10 RECCOMENDATIONS ...................................................................... 132
REFERENCES ............................................................................................................ 134
APPENDIX A CDF OF UDL FOR ALL TRUCK COMBINATIONS ........................ 141
APPENDIX B CDF OF MAXIMUM DAILY UDL ................................................... 153
APPENDIX C CDF OF maximum weekly UDL......................................................... 161
vii
LIST OF FIGURES
Figure 2.1. HL-93 Live Loading in AASHTO LRFD Code [2007]. Truck and Lane Load.
...................................................................................................................................... 12
Figure 2.2. HL-93 Live Loading in AASHTO LRFD Code [2007]. Tandem and Lane
Load. ............................................................................................................................. 12
Figure 2.3. HL-93 Live Loading in AASHTO LRFD Code [2007]. Alternative Load for
Negative Moment between points of contraflexure and reaction at interior piers. ........... 13
Figure 2.4. OHBD Live Loading [1991]. OHBD Truck. ................................................ 13
Figure 2.5. OHBD Live Loading [1991]. OHBD Truck and Lane Load. ........................ 14
Figure 2.6. CAN/CSA-S6-00 Live Loading [2000]. CL-W Truck .................................. 15
Figure 2.7. CAN/CSA-S6-00 Live Loading [2000]. CL-W Lane Load ........................... 15
Figure 2.8. BS 5400 Live Loading curve HA UDL [2006] ............................................. 16
Figure 2.9. Dimensions of HB vehicle. .......................................................................... 17
Figure 2.10. Eurocode 1 [2002]. Load Model 1 .............................................................. 18
Figure 2.11. ASCE Loading on Log Scale ..................................................................... 19
Figure 2.12. Equivalent Unfactored Loads, w/o IM, w/o multilane factors. .................... 24
Figure 2.13. Equivalent Factored Loads, w/o IM, w/o multilane factors. ........................ 25
Figure 2.14. Equivalent Unfactored Loads, with IM, w/o multilane factors. ................... 25
Figure 2.15. Equivalent Factored Loads, with IM, w/o multilane factors. ....................... 26
Figure 2.16. Equivalent Factored Loads, with IM, with multilane factors for 4 traffic
lanes. ............................................................................................................................. 26
Figure 3.1 PDF and CDF of a normal random variable .................................................. 32
Figure 3.2 PDF’s of load, resistance, and safety margin ................................................. 34
Figure 3.3 Reliability index defined as the shortest distance in the space of reduced
variables ........................................................................................................................ 36
Figure 3.4 Normal Distribution Function on the Normal Probability Paper. ................... 41
Figure 4.1. FHWA 13-category scheme ......................................................................... 43
Figure 4.2. Longer Combination Vehicles (LCV’s) ........................................................ 45
Figure 4.3. States allowing various Longer Combination Vehicles ................................. 45
Figure 4.4. New Bridge Formula - regulation of vehicles' length and weight .................. 46
viii
Figure 4.5. Number of trucks by weight (in thousands of trucks). Transportation statistics
annual report, December 2006. ...................................................................................... 47
Figure 4.6. Freight Tonnage Moved by Truck (FHWA) ................................................. 48
Figure 4.7. Time variation of total truck weight statistic. (Gindy, M., Nassif, H.H., 2006)
...................................................................................................................................... 49
Figure 4.8. WIM data collection .................................................................................... 51
Figure 4.9. CDF's of GVW by axles Oregon I-5 Woodburn ........................................... 58
Figure 4.10. CDF's of GVW by axles Oregon I-84 Emigrant Hill................................... 59
Figure 4.11. CDF's of GVW by axles NY I-495 EB ....................................................... 60
Figure 5.1. Critical loading. Traffic jam scenario. .......................................................... 62
Figure 5.2. CDF’s of GVW for 5-axle trucks. New York WIM Data. ............................. 62
Figure 5.3. CDF of GVW for 5-axle and 11-axle trucks (Nowak, A.S., Laman, J. and
Nassif, H., 1994)............................................................................................................ 63
Figure 5.4. Percentage of vehicles by number of axles. FHWA WIM Data. ................... 63
Figure 5.5. Percentage of vehicles by number of axles. (Kim, S-J., Sokolik, A.F., and
Nowak, A.S., 1997) ....................................................................................................... 64
Figure 5.6. CDF’s of GVW for all types of vehicles in Oregon. ..................................... 64
Figure 5.7. AASHTO LRFD legal load trucks, Type 3-3 Units. ..................................... 65
Figure 5.8. Simulation of trucks moving throughout span length.................................... 66
Figure 5.9. Clearance - Gap and Spacing - Headway Concepts ...................................... 67
Figure 5.10. Interstate semitrailer WB-20 (AASHTO Geometric Design of Highways and
Streets) .......................................................................................................................... 67
Figure 5.11. HL-93 proposed for long span bridges ....................................................... 69
Figure 5.12. Mean value of uniformly distributed load ................................................... 72
Figure 5.13. Daily maximum mean value of uniformly distributed load ......................... 72
Figure 5.14. Weekly maximum mean value of uniformly distributed load ...................... 73
Figure 5.15. Bias (mean max 75 year to nominal value of UDL) .................................... 73
Figure 5.16. Bias (mean max 75 year to nominal value of UDL) .................................... 74
Figure 5.17. Bias (mean max 75 year to nominal value of UDL) assumed designed UDL
of 0.85 k/ft ..................................................................................................................... 74
Figure 5.18. Bias for heavily loaded localizations, assumed designed UDL of 1.25 k/ft . 75
ix
Figure 5.19. Coefficient of variation of daily maximum uniformly distributed load ....... 75
Figure 5.20. Coefficient of variation of weekly maximum uniformly distributed load .... 76
Figure 5.21. Proposed coefficient of variation of uniformly distributed load .................. 76
Figure 6.1. Traffic loading pattern used for multiple truck presence statistics. ................ 79
Figure 6.2. Variation of multiple truck presence statistics with respect to truck volume.
Gindy and Nassif (2006). ............................................................................................... 80
Figure 6.3. Variation of multiple truck presence statistics with respect to bridge span
length. Gindy and Nassif (2006). ................................................................................... 81
Figure 6.4. Video 10, time: 00:00:58..............................................................................83
Figure 6.5. Video 1, time: 00:05:28 ............................................................................... 84
Figure 6.9. Multilane load in design codes AASHTO LRFD Code (2007), OHBDC
(1991), CAN/CSA-S6-00 [2000], and ASCE (1981). ..................................................... 87
Figure 6.10. Multilane load in Eurocode 1. ....................................................................87
Figure 6.11. Multilane load in actual observation. .......................................................... 87
Figure 7.1. Distribution of fundamental bridge frequencies (Cantieni 1984)....................... 90
Figure 7.2. Fundamental frequency versus span length (Cantieni 1984) ............................. 91
Figure 7.3. Fundamental frequency versus span length (Paultre 1992) ............................... 91
Figure 7.4. Impact factor versus vehicle speed and road surface condition (Wang, Shahawy,
and Huang 1993) ............................................................................................................ 93
Figure 7.5. Effects of vehicle suspension on the measured bridge response (Biggs and Suer
1955) ............................................................................................................................. 95
Figure 7.6. Impact versus span length (Fleming and Romualdi 1961) ................................ 96
Figure 7.7. Impact versus span length (Cantieni 1984)...................................................... 96
Figure 7.8. Meshed model of the bridge........................................................................... 99
Figure 7.9. Truck model in FEM ..................................................................................... 99
Figure 7.10. Vehicle-bridge interacting force. .............................................................. 101
Figure 7.11. Force due to moving truck versus time. Plot from ABAQUS. ...................... 101
Figure 7.12. Bending modes ......................................................................................... 102
x
Figure 7.13. Torsion modes. ......................................................................................... 102
Figure 7.14. Maximal deflection due to moving truck .................................................... 103
Figure 7.15. Deflection due to a truck moving 40mil/hr versus time. ............................... 104
Figure 7.16. Deflection due to a truck moving at crawling speed versus time. .................. 104
Figure 8.1 Distribution of strains for the pure axial loading .......................................... 109
Figure 8.2 Distribution of strains distribution for the balance failure point B, the end of
the compression control zone ....................................................................................... 109
Figure 8.3 End of compression block of concrete in the bottom flange ......................... 110
Figure 8.4 End of compression block of concrete in the web. ....................................... 111
Figure 8.5 End of compression block of concrete in the top flange. .............................. 111
Figure 8.6 Stress - Strain Relationship for Reinforcing Steel ........................................ 112
Figure 8.7. Geometry of Cooper River Bridge ............................................................. 114
Figure 8.8. Load combinations ...................................................................................... 115
Figure 8.9. Envelope of bending moments for bridge tower for w=0.64k/ft .................. 116
Figure 8.10. Envelope of bending moments for bridge tower for w=0.80 k/ft ............... 117
Figure 8.13 Bias factor for compressive strength of concrete ....................................... 121
Figure 8.14 Coefficient of variation for compressive strength of concrete .................... 122
Figure 8.15 CDF’s of yield strength for Reinforcing Steel Bars, Grade 60 ksi .............. 123
Figure 8.16 Recommended material parameters for reinforcing steel bars, Grade 60 ksi
.................................................................................................................................... 124
Figure 8.17 Statistical Parameters of Resistance .......................................................... 125
Figure 8.18 Force and Moment results on the bridge tower for different live loads....... 127
Figure 8.19 Reliability indexes due to different live load ............................................. 128
xi
LIST OF TABLES
Table 2.1. Number of design lanes vs. road width in OHBDC [1991] ............................ 14
Table 2.2. Number of design lanes vs. road width in CAN/CSA-S6-00 .......................... 15
Table 2.3. Characteristic values of load for successive road lanes .................................. 18
Table 2.4. Dynamic allowance in AASHTO LRFD [2007] ............................................ 20
Table 2.5. Dynamic allowance in CAN/CSA-S6-00 ....................................................... 21
Table 2.6. Comparison of Multilane Reduction Factors.................................................. 22
Table 2.7. Multilane Reduction Factors for BS 5400 ...................................................... 22
Table 2.8. Values of Equivalent Unfactored Loads, w/o IM, w/o multilane factors......... 27
Table 2.9. Values of Equivalent Factored Loads, w/o IM, w/o multilane factors. ........... 27
Table 2.10. Values of Equivalent Unfactored Loads, with IM, w/o multilane factors...... 27
Table 2.11. Values of Equivalent Factored Loads, with IM, w/o multilane factors. ........ 28
Table 2.12. Values of Equivalent Factored Loads, with multilane factors for 4 traffic
lanes. ............................................................................................................................. 28
Table 3.1. Relationship between vertical scale on Normal Probability Paper and
Probability ..................................................................................................................... 40
Table 4.1. Conversion chart for vehicles’ class and number of axles .............................. 44
Table 4.2. Summary of WIM Data ................................................................................. 53
Table 4.3. Vehicles by axle in Oregon ........................................................................... 54
Table 4.4. Vehicles by traffic lane in Oregon ................................................................. 54
Table 4.5. Vehicles by axle in Florida ............................................................................ 55
Table 4.6. Vehicles by traffic lane in Florida ................................................................. 55
Table 4.7. Vehicles by axle in Indiana ........................................................................... 56
Table 4.8. Vehicles by traffic lane in Indiana ................................................................. 56
Table 4.9. Vehicles by axle in New York ....................................................................... 57
Table 4.10. Vehicles by traffic lane in New York........................................................... 57
Table 5.1. Statistical parameter for proposed uniformly distributes live load .................. 69
Table 5.2. Summary of simulated data ........................................................................... 70
Table 5.3. Heaviest truck combinations for 600 ft on I-495 WB ..................................... 71
Table 6.1. Presence of multiple trucks and their location on the road lanes .................... 78
xii
Table 8.1 Recommended Statistical Parameters for Compressive Strength, f c ' (Nowak
A.S. et al., 2008) .......................................................................................................... 122
Table 8.2 Statistical Parameters of Fabrication Factor. ................................................. 124
xiii
CHAPTER 1
INTRODUCTION
1.1. PROBLEM STATEMENT
The live load models available were developed for short and medium span
bridges. This doctoral dissertation deals with the development of a live load model for
long span structures. The developed live load model is valid for spans between 600 ft and
5000 ft. In contrast to short and medium spans, a long span live load must include the
possibility of multiple trucks being present.
The continuous increase in the number of the trucks and their weights led to a
review of traffic data for live load. Observing traffic statistics helps to realize the rate of
those changes, their importance, and to draw some conclusions regarding design. In the
last 30 years, the number of the vehicle miles logged annually on American highways has
increased 225%, with heavy truck traffic increasing 550%. Some percentage of trucks
runs overweight, particularly if it is to their economic advantage. Therefore, a new live
load model for long span bridges had to be developed and it had to be based on the
newest traffic data obtained from highway and bridge administrators.
During the AASHTO LRD calibration, the live load model for short and medium
span bridges was developed based on a set of truck weight and load effect statistics that
were presumed to be valid for any typical bridge site in the U.S. The live load model may
not represent the actual loading conditions at a particular bridge site or bridges in a state.
Nowadays, several states are using Weigh-In-Motion (WIM) systems to collect vast
1
amounts of truck weight and traffic data that can be used to obtain site-specific and state-
specific live load models for bridge design and load capacity evaluation. This could allow
individual states to adjust the AASHTO live load factors to take into consideration the
particular truck traffic conditions throughout a state, a region, or for a particular route.
Site-specific or state-specific live load models may be developed based on actual truck
weight and traffic data collected at the site or within the state. Traffic varies for different
sites within each state. As a result, site specific models depending on average daily truck
traffic and participation of heavily loaded vehicles seem to be more practical.
Since early publications by the American Association of State Highway Officials
(AASHO), live load was modeled as an HS20 truck. As an addition to truck load, the
uniformly distributed load of 0.64 kip/ft was introduced in 1944. Since then the original
definition of HS-20 has been changed. The concentrated load was substituted with three
axial forces representing a truck. In contrast, the uniformly distributed load has never
been updated. It is still used in the current AASHTO LRFD Code as it was in 1944. The
derivation of uniformly distributed load is not clear. To amend this, a new approach to
model uniformly distributed load had to be developed and new value of uniformly
distributed load had to be proposed. Current multilane reduction factors and dynamic
allowance also may not be appropriate for long span bridges. Review of those topics was
necessary.
1.2. OBJECTIVE AND BENEFITS OF THE STUDY
The objective in this study was to develop a live load model for long span bridges.
The model is valid for spans between 600 ft and 5000 ft. It is intended to reflect current
traffic patterns, quantities of trucks and their weights. The newest available traffic
database from a variety of sites within many different states is used. Based on the
analysis of traffic records (weigh-in-motion and videos) the design live load is developed
and recommended to be taken into consideration in the bridge design code. Reliability
analysis is used to verify the developed live load model.
In accordance with the stated objective, the first stage was to study previous
research and current international codes’ provisions on the topic. The second stage of the
2
research was the collection of state of the art traffic data from highway and bridge
administrators. The data obtained had to be analyzed and filtered out from erroneous
readings of measurement instruments. Then a new uniformly distributed load was
derived. The value of the new live load is based on three models: an average 5-axle truck,
legal load trucks and simulation of a traffic jam using WIM data. Such an extensive
actual weigh in motion database has never been used in the derivation of live load for
long span bridges. Most of the previous studies were based on measurements from
limited numbers of sites within one state. The magnitude of the database obtained for the
scope of this research has to be underlined. A derivation of uniformly distributed load
from WIM data required developing a numerical procedure of calculation to process the
extensive database. Cumulative distribution functions were plotted for all data, as well as
for maximum daily and maximum weekly uniformly distributed load. New uniformly
distributed load was proposed. Statistical parameters for live load (bias and coefficient of
variation) are derived. Relationship between site characteristics (ADTT, percentage of
overloaded loaded vehicles) and calculated values of uniformly distributed loads were
studied. The problems of multilane reduction factors and the dynamic factor were also
discussed.
The final step of this dissertation was reliability analysis. Reliability analysis was
performed in order to assess how the increase in live load influences reliability indexes.
An exemplary suspension bridge, the bridge component and a limit state function that are
the most influenced by live load were selected. The calculations were performed for the
current AASHTO LRFD design live load and increased load values obtained from real
traffic data. For the scope of this study, new statistical parameters for uniformly
distributed load were used and statistical parameters of resistance were derived based on
the newest material, fabrication and professional factors.
The outcome of this research is the recommendation of a live load model for long
span bridges. There is a recommended value of uniformly distributed load for bridges
carrying low and average ADTT. In addition, there is a recommendation for an increase
of the live load model for the long span bridges in heavily loaded urban and industrial
areas.
3
1.3. ORGANIZATION OF THE DISSERTATION
Chapter 1 of this dissertation is an introduction to the research conducted. It

presents a problem statement, objectives and the scope of the study, as well as the
benefits and limitations of this research. The chapter also includes a review of prior
research on the topic.
Chapter 2 reviews current international bridge design codes regarding live load,
dynamic load and multiple presence factors provisions. Uniformly distributed load for
wide range of spans is calculated and compared.
Chapter 3 presents the principals of the reliability theory that were applied in this
study. Definitions of standard variables, probability distributions, limit state functions
and reliability index are introduced. Methods of use of the normal probability paper and
simulation techniques are described.
Chapter 4 describes the study of traffic data, regulations of truck types, truck sizes
and weight limits, as well as the weight-in-motion technology and database to be used in
this dissertation.
Chapter 5 describes development of the live load model. A model based on an
average 5-axle truck, a model based on legal load trucks, and a model based on traffic
jam simulation using WIM data are presented. New values of uniformly distributed load
are derived and proposed to be applied in the code. New statistical parameters for live
load are calculated.
In the Chapter 6, the problem of the presence of multiple trucks on a bridge is
discussed. A short review of current studies, analysis of video recordings of traffic jam
situations, and discussion on different approaches to the problem are presented.
Chapter 7 presents the problem of dynamic factor. Parameters affecting bridge
dynamic response are discussed. Exemplary bridge-vehicle interaction is modeled and the
dynamic factor is derived.
In the Chapter 8, the reliability procedures for a long span bridge are developed.
The analysis is performed on a bridge tower of the selected suspension bridge. Reliability
indexes for bridge loaded increasing values of live load are calculated.
4
Chapter 9 presents the summary and conclusions of research performed for the
scope of this dissertation. As well, recommendations are specified.
1.4. PRIOR INVESTIGATIONS
1.4.1. Prior Investigations on Live Loading on Short and Medium Span Bridges
Live load models for short and medium span bridges were of interest to many
researchers. Most of the studies performed on live load models were based on truck data
obtained within programs carried out by the Ontario Ministry of Transportation since the
early 1970s. This was the vastest database available until now.
Nowak and Hong (1991) formulated a procedure to calculate maximum moments
and shears for various time periods. The maximum load effects for various time periods
from one day to 75 years were derived from extrapolated distributions. Single and two
lane bridges are considered. For one lane traffic a single truck governs for shorter spans,
and two following trucks govern for longer spans. For two lanes of traffic, the maximum
effect is obtained for two trucks with fully correlated weights, travelling side-by-side. It
has also been concluded that the bias factor (ratio of the mean to nominal value) is larger
for smaller spans.
Kim, Sokolik, and Nowak (1997) studied actual truck loads on selected bridges in
the Detroit area. The measurements were taken by using a weight in motion system. It
was observed that truck loads are strongly site specific. The observed truck weights were
often heavier than legal limits. The maximum observed truck weights were up to 250
kips, causing maximum moments two times larger than AASHTO load and resistance
factor design values. Gindy and Nassif (2006) formulated a similar conclusion based on
data from New Jersey. It was found out that maximum gross vehicle weight reaches a
value of 225 kips and it shows a steady increase at an annual growth rate is 1.2%.
Nowak, Laman, and Nassif (1994) published a research report on the effect of
truck loading on bridges. The WIM measurements were taken on seven bridges in
Michigan. The researchers developed procedures for evaluation of live load spectra on
steel girder bridges with regard to fatigue. The deteriorating capacity of bridge was
5
evaluated as a function of the rate of corrosion. It has been proved that WIM
measurements show the unbiased truck weights, which are 30-50 percent larger than
extreme values obtained at weight stations. The WIM data is unbiased because the drivers
are not aware of the measurements and they do not make an effort to avoid the scales.
The WIM measurements from Michigan have also been used to study dynamic load,
Nassif and Nowak (1995). It was found out that the dynamic load factor decreases with
increased static loads, and that larger values of DLF are observed in exterior girders due
to relatively smaller static load effect. Derivation of the dynamic load model is described
by Hwang and Nowak (1991).
1.4.2. Prior Investigations on Live Loading on Long Span Bridges
The most widely known researcher in the field of live loading on long span brides
is Peter G. Buckland (1978, 1980, and 1991). He concluded that traffic loading on long
span bridges can be accurately represented in the traditional manner, by one set of
uniform and concentrated loads. One of his findings was that uniform load per foot
reduces as the load length is increased. However, unlike many other studies he found out
that concentrated load increases as the loaded length increases. Four uniform loading
curves were developed for different loading cases. The load cases were distinguished
depending on the percentage of “heavy vehicles”: 2.4, 7.4, 30.0, or 100 percent, where
“heavy vehicles” are defined as trucks and buses over 12000 lb. These loading curves
were recommended by the ASCE Committee as vehicle loading of long-span bridges, in
1981. They are known unofficially as the ASCE Loading. However, they have never been
applied into the design codes.
Peter G. Buckland had also made a valuable observation regarding several loaded
lanes. He stated that if a single lane has a certain load on it, than the additional lanes
would increase the load in the lane closer to the curb, as trucks gravitate towards it.
However, load in the additional lanes can be reduced.
In the paper by Buckland (1991) the comparison of North American and British
live loads on long-span bridges is presented. The loads are compared as equivalent
uniformly distributed loads, calculated as an equivalent shear and bending moment for
6
simply supported beams. This approach can be successfully used for short and medium
span bridges. However, its application to long spans can be questioned, since long span
bridges cannot be constructed as simply supported beams. This method of deriving the
equivalent load can be used exclusively for comparison of codes.
1.4.3. Prior Investigations on Structural Reliability
The theory of structural reliability have been investigated and described by many
researchers. Several books and publications provide available knowledge regarding
reliability theory, for instance, Thorf-Christtensen and Baker (1982), Ang and Tang
(1984), Madsen, Krenk and Lind (1986); Thorf-Christtensen and Murotsu (1986), Ayyub
and McCuen (1997), Murzewski (1989), Nowak and Collins (2000) and Wolinski and
Wrobel (2001). The application of reliability theory has resulted in the improvement of
structural design in terms of safety, serviceability and durability. However, it was not
until the late 70’s when safety factors based on load and resistance uncertainties were
proposed for introduction into the codes. In the United States, it was the building design
code (Galambos and Ravindra 1978, and Ellingwood 1980, 1982), and in Canada, it was
the Ontario highway bridge design code (Nowak and Lind 1979). Since then, reliability
techniques have been increasingly used in modern design codes. Nowadays many
researchers keep working on further development of new methods of structural reliability
analysis, among them A.H.-S. Ang, O. Ditlevesen, R.E. Melchers, H. Nielsen- Faber,
A.S. Nowak, R. Rackwitz, and P. Thoft-Christiansen.
For many years the random nature of various parameters influencing structural
safety has been of interest to engineers. Until they gathered more knowledge about the
laws of nature, they used to assure structural safety through ‘trial and error’ and intuition.
Mathematical theories available nowadays describe material and structural behavior
sufficiently enough to give a rational basis for structural safety evaluations (Nowak and
Collins, 2000). Early publications that quantified and presented a mathematical
formulation of structural safety problems were published by Mayer (1926) and
Wierzbicki (1936). They recognized that load and resistance parameters have random
characteristics, and that each structure has a finite and limited probability of failure. Their
7
concept of a structural reliability problem has been subsequently adopted in the
precursory publication for that field by Freudenthal (1956). In the 1960s, a new trend in
using probabilistic concepts in the analysis of limit capacity and structural resistance was
developed. It was first presented by researchers Augusti and Baratta (1973, 1973).
Subsequently, important work was done by Corotis and Nafday (1989) analyzing the
limit capacity of beam-column frame structures using the principles of conditional
probability.
The extensive development of practical tools and efficient methods for evaluating
the probability of structural failure has been made in the last 30 decades. Initially, the
probability of failure was defined by multidimensional integral functions of distributions
and it was cumbersome to evaluate. Pioneering studies on the first practical application of
reliability analysis were performed by Cornell and Lind in the late 1960s and early 1970s.
Their approach estimated the limit state function at mean values of random parameters
and used a linearized limit state function. A milestone was the estimation of the
probability of failure proposed by Hasofer and Lind (1974). The simplified procedure
involved a nonlinear mathematical programming problem with boundary conditions (an
estimated limit state for all variables and a defined probability in the so called “design
point”). The extension of the Hasofer and Lind approach and the transformation of
uncorrelated random variables of various distributions into standardized normal
distributions were proposed by Rackwitz and Fiessler (1978). The developed numerical
procedure of the design point estimation used to be called the Rackwitz-Fiessler
procedure. Hohenbichler and Rackwitz (1988) used the Rosenblatt (1952) transformation
procedure for the transformation of dependent (correlated) random variables from and
into standardized form, which is currently one of the major tools used in modern
reliability analysis. Another commonly used transformation is Nataf’s transformation,
which was presented in work by Kiureghian and Liu (1986). Commonly used methods of
reliability analysis are based on the approximation of the limit state function at the design
point using first or second order functions (FORM and SORM). Advanced SORM have
been elaborated by researchers such as Fiessler, Neumann and Rackwitz (1979), Breitung
(1984), Nowak and Collins (2000). Adhikari (2004) systemized and published all of the
earlier proposed SORM procedures. Simulation techniques are another approach to
8
estimating probability of failure. The most popular is the Monte Carlo Method simulation
technique (Thoft-Christensen and Baker, 1982; Hart, 1982). Determination of the
probability of structural failure with the use of simulation techniques has limited
accuracy, and a huge number of numerical simulations are required to achieve a high
accuracy of results. This method becomes very useful and an especially practical tool in
cases where physical testing is expensive.
The structural system reliability is a field of interest for many researchers. Bridges
usually consist of a combination of series and parallel systems. Identification of collapse
mode and degree of correlation between members is very difficult or often even
impossible to evaluate. Moses (1982) proposed incremental load approach and suggested
a procedure for identifying collapse mode for both ductile and brittle components. The
identification of collapse mode was also discussed by Rashedi and Moses (1988).
Reliability models applied to bridge evaluation were addressed by Nowak and
Tharmabala (1988). Moses and Verma (1987) used a load and resistance approach to
evaluate the strength of bridges with reliability principles. Tantawi (1986) developed a
grid nonlinear analysis program to calculate the moment-carrying capacity of a bridge.
Than Zhou (1987) developed an integration sampling technique to calculate the system
reliability of a bridge. Both Tantawi and Zhou, found that bridge system reliability is
higher than girder reliability.
Practical procedures for system reliability analysis were suggested by Nowak and
Zhou (1990), Zhou and Nowak (1990), and Tabsh and Nowak (1991). The procedures
assumed that the ultimate load carrying capacity is equal to the weight of a truck which
causes a collapse. A collapse was defined as an excessive, non-acceptable deflection.
Estes and Frangopol (1999) assumed that failure occurs when failure occurs in three of
five adjacent girders. Several major contributions were also made by researchers such as
Ditlevsen (1982, 1996), Grigoriu (1982, 1983), and Rackwitz (1985). Ditlevsen used
conditional probability to calculate bounds of the probability of failure, Grigoriu
discussed a parallel system with brittle elements, and Rackwitz recognized the effect of
correlation on system performance.
The present requirements for civil engineering structures primarily focus on
structural safety. New structural design codes are calibrated using the limit state analysis
9
approach to assure safety standards and to provide the required reliability of new design
structures. However, there is a major gap between the development of reliability
techniques and their application to structural engineering design and evaluation. The
system reliability analysis requires an efficient structural analysis procedure, as the
calculations have to be repeated many times. Therefore, a more comprehensive method to
assess the system reliability of a bridge needs to be developed.
10
CHAPTER 2
LIVE LOAD IN CURRENT DESIGN CODES
2.1. INTRODUCTION
While approaching the problem of live load for bridges it is necessary to review
the current codes and perform a comparative analysis. The codes were selected with the
objective to present various approaches to design live load, use of multilane factors, and
dynamic impact allowance.
The live load for bridges can be represented in many ways, including a uniform
load and a combination of truck(s), as for example in AASHTO LRFD Bridge Design
Specifications (2007), CAN/CSA-S6-00 Canadian Highway Bridge Design Code (2000),
and Eurocode 1 (2002). Non-uniform loading curves are used in the British Standard
5400 (2006) and ASCE Recommended Design Loads for Bridges (1981).
The design live loads specified in AASHTO LRFD Code (2007), OHBDC (1991),
CAN/CSA-S6-00 [2000], Eurocode [2002], and ASCE (1981) were briefly summarized
in the following paragraphs. The comparison of equivalent uniformly distributed loads
for a variety of spans is presented in Paragraph 2.5.
11
2.2. INTERNATIONAL PROVISIONS FOR LIVE LOADING
2.2.1. AASHTO LRFD Code [2007]
The American Association of State Highway and Transportation Officials

Standard Specifications for Highway Bridges define HL-93 live load as the extreme force
effect taken as the larger of following:
1) The AASHTO LRFD 3-axle design truck, and uniformly distributed design lane
load of 0.64 klf. Figure 2.1.
2) The AASHTO LRFD Design Tandem, and uniformly distributed design lane load
of 0.64 klf. Figure 2.2.
3) For negative moment between points of contraflexure and reaction at interior
piers, combination of two design trucks spaced at minimum of 50 ft, and
uniformly distributed design lane load of 0.64 klf, should be considered. All of the
forces reduced to 90 %. Figure 2.3.
32.0 kip 32.0 kip
8.0 kip
0.640 klf
14’ 14’ - 30’
Figure 2.1. HL-93 Live Loading in AASHTO LRFD Code [2007].

Truck and Lane Load.
25.0 kip 25.0 kip
0.640 klf
4’

Tandem and Lane Load.
12
28.8 kip 28.8 kip 28.8 kip 28.8 kip
7.2 kip 7.2 kip

0.576 klf
14’ 14’ - 30’ ≥ 50’ 14’ 14’ - 30’

Alternative Load for Negative Moment between points of contraflexure
and reaction at interior piers.
The loads shall occupy 10 ft transversally within a design lane. Truck wheels are
assumed to be spaced 6.0 ft transversally.
2.2.2. OHBDC [1991]
The Ontario Highway Bridge Design Code (Third edition) determines live load as
a truck or a combination of truck and lane load, whichever produces the maximum load
effect:
1) The OHBD Truck, which is a 5-axle truck. Figure 2.4.
2) The OHBD Lane Load consists of an OHBD Truck with each axle reduced to
70%, and superimposed centrally within the width of a 3.0 m (10 ft) wide
uniformly distributed load of 10.0 kN/m (0.685 kip/ft). Figure 2.5.
36.0 kip 36.0 kip 45.0 kip 36.0 kip
13.5 kip
12’ 4’ 20’ 24’
Figure 2.4. OHBD Live Loading [1991]. OHBD Truck.
13
25.2 kip 25.2 kip 31.5 kip 25.2 kip
9.4 kip
0.685 kip/ft
12’ 4’ 20’ 24’
Figure 2.5. OHBD Live Loading [1991]. OHBD Truck and Lane Load.
Table 2.1. Number of design lanes vs. road width in OHBDC [1991]
Width Number of lanes
Wc ≤ 6.0 m 1
6.0 m < Wc ≤ 10.0 m 2
10.0 m < Wc ≤ 13.5 m 3
13.5 m < Wc ≤ 17.0 m 4
17.0 m < Wc ≤ 20.5 m 5
20.5 m < Wc ≤ 24.0 m 6
24.0 m < Wc ≤ 27.5 m 7
27.5 m < Wc 8
2.2.3. CAN/CSA-S6-00 [2000]
The Canadian Highway Bridge Design Code applies CL-W loading, which
consists of the truck or the lane load:
1) The CL-W Truck is 5-axle truck. The number "W" indicates the gross load of the
truck in kN. For the design of a national highway network, loading not less than
CL-625 shall be used. Figure 2.6.
2) The CL-W Lane Load consists of CL-W Truck with each axle reduced to 80%,
and a superimposed uniformly distributed lane load of 9.0 kN/m (0.617 kip/ft),
that is 3.0 m (10 ft) wide. Figure 2.7.
14
39.3 kip 33.7 kip
28.1 kip 28.1 kip (0.28 W)
(0.2 W) (0.2 W) (0.24 W)
11.2 kip
(0.08 W)
12’ 4’ 22’ 22’
Figure 2.6. CAN/CSA-S6-00 Live Loading [2000]. CL-W Truck
31.5 kip 27.0 kip

22.5 kip 22.5 kip (0.224 W)
(0.16 W) (0.16 W) (0.192 W)
9.0 kip
(0.064 W)
0.617 kip/ft
12’ 4’ 22’ 22’
Figure 2.7. CAN/CSA-S6-00 Live Loading [2000]. CL-W Lane Load
Table 2.2. Number of design lanes vs. road width in CAN/CSA-S6-00

Width Number of lanes
Wc ≤ 6.0 m 1
6.0 m < Wc ≤ 10.0 m 2
10.0 m < Wc ≤ 13.5 m 2 or 3 (check both)
13.5 m < Wc ≤ 17.0 m 4
17.0 m < Wc ≤ 20.5 m 5
20.5 m < Wc ≤ 24.0 m 6
24.0 m < Wc ≤ 27.5 m 7
27.5 m < Wc 8
2.2.4. BS 5400 [2006]
According to British Standard the structures and its elements shall be designed to
resist the more severe effects of either design HA loading or design HA loading
15
combined with design HB loading. HA loading represents normal traffic in Great Britain.
HB loading is an abnormal vehicle unit loading. Both loadings include impact.
HA loading consists of uniformly distributed load (UDL), and a knife edge load
(KEL), or a single-wheel load. Live Loading curve HA UDL is shown in the Figure 2.8.
Value of nominal uniformly distributed load (UDL) equal to:
• for loaded lengths up to and including 50 m:
W = 336 (1/L) 0.67 [kN]
• for loaded lengths in excess of 50 m but less than 160 0m:
W = 36 (1/L) 0.1 [kN]
• for loaded lengths above 1600 m, the UDL shall be in agree with the relevant
authority.
The nominal knife edge load (KEL) per lane shall be taken as 120 kN (27 kip).
The single nominal wheel load alternative to UDL and KEL is one 100 kN (22.5 kip)
wheel placed on a carriageway and uniform distribution over a circular contact area
assuming an effective pressure of 1.1 N/mm2.
Figure 2.8. BS 5400 Live Loading curve HA UDL [2006]

16
HB loading defines the minimum number of HB loading units that shall be
considered, which is 30. This number may be increased up to 45 if so directed by the
relevant authority. Figure 2.9 below shows the plan and axle arrangement for one unit of
HB loading. One axle is 10 kN (2.25 kip), i.e. 2.5 kN (0.56 kip) per wheel. The overall
length of the HB vehicle shall be taken as 10, 15, 20, 25, or 30 m for inner axle spacing
of 6, 11, 16, 21, or 26 m respectively. The effects of the most severe case shall be
adopted.
Figure 2.9. Dimensions of HB vehicle.
2.2.5. Eurocode 1 [2002]
The Eurocode 1 Part 2 is applicable to bridges with spans from 5 to 200 m (17 to
667 ft), and carriageway width up to 42 m (140 ft). It presents four models for
determining the main vertical loads from traffic:
1) Load Model 1 (LM1) consists of concentrated and uniformly distributed loads

(Figure 2.10) which cover most of the effects of the traffic of trucks and cars. The
code specifies live load to be used on each traffic lane, therefore there is no need
to introduce multilane factors. It is used for general and local verifications.
17
Table 2.3. Characteristic values of load for successive road lanes
Location Axle load Qik UDL qik
Lane number 1 300 kN (67.5 kip) 9 kN/m2 (0.188 kip/ft2)
Lane number 2 200 kN (45.0 kip) 2.5 kN/m2 (0.052 kip/ft2)
Lane number 3 100 kN (22.5 kip) 2.5 kN/m2 (0.052 kip/ft2)
Other lanes 0 2.5 kN/m2 (0.052 kip/ft2)
Remaining area 0 2.5 kN/m2 (0.052 kip/ft2)
Figure 2.10. Eurocode 1 [2002]. Load Model 1
2) Load Model 2 (LM2) consists of a single axle load of 400 kN, which covers the
dynamic effects of the normal traffic on short structural members. The distance
between wheels is 2 m. The contact surface of each wheel should be taken as a
rectangle of sides 0.35 m and 0.60 m. When relevant, only one wheel of 200 kN
may be taken into account.
3) Load Model 3 (LM3) consists of sets of axle loads representing special (carrying
heavy loads) vehicles, which can travel on routes permitted for abnormal loads. It
is intended for general and local verifications.
18
4) Load Model 4 (LM4) represents crowd loading of 5.0 kN/m2. It is intended only
for general verifications and it is particularly relevant for bridges in or near town
areas.
2.2.6. ASCE Loading [1981]
The live load known unofficially as the ASCE Loading is a result of the studies
performed by Peter G. Buckland, which was recommended for long span bridge by the
American Society of Civil Engineers Committee on Loads and Forces on Bridges. ASCE
(1981) specifies three levels of live load for highway bridges depending on the average
percentage of heavy vehicles in traffic flow: 7.5%, 30%, and 100% heavy vehicles of the
total vehicle population. "Heavy vehicles" were defined as buses and trucks over 12 000
lbs. It has been proved that the loading can be represented by a uniform load and a
concentrated load (Figure 2.11) to give moments and shears with a sufficient degree of
accuracy.
2.74
P U
U(100% HV )
158.6
U (k/ft)
P
1
102.4
P (kips)
56.2 U (30% HV )
U (7.5% HV )
0 0
50 100 200 400 800 1600 3200 6400
Loaded Length (ft)
Figure 2.11. ASCE Loading on Log Scale

19
2.3. PROVISIONS FOR DYNAMIC LOAD FACTOR
There is considerable variation in the treatment of dynamic load effects by bridge

design codes in different countries. The most common approach is to apply dynamic
response as a fraction or multiple of the response that would be obtained if the same
forces or loads were applied statically. The objective of this simple approach is to not
increase complexity to the designer.
The American Association of State Highway and Transportation Officials
Standard Specifications for Highway Bridges AASHTO LRFD [2007] define that
dynamic load allowance shall be applied to static load effects of the truck or tandem, as a
percentage specified in the table below. It shall not be applied to pedestrian loads or to
the design lane load.
Table 2.4. Dynamic allowance in AASHTO LRFD [2007]

Component Dynamic allowance, IM
Deck joints,
75%
all limit states
All other components,
15%
fatigue and fracture limit states
All other components,
33%
other limit states
Until recently AASHTO Standard Specifications for Highway Bridges used to

specify dynamic load effects in terms of an impact factor that is a function of bridge span.
However relation between dynamic load factor and bridge span is a controversial issue
between researchers (see Paragraph 7.2.).
The Canadian Highway Bridge Design Code CAN/CSA-S6-00 [2000] and
OHBDC [1991] apply dynamic allowance as a percentage to static effects of the CL-W
Truck for the number of axles considered in the design lane, as shown in the table below.
20
Table 2.5. Dynamic allowance in CAN/CSA-S6-00
for deck joints 0.50
1 axle of CL-W Truck 0.40
2 axles of CL-W Truck 0.30
3 or more axle of CL-W Truck 0.25
According to British Standard BS 5400 [2006] and the Eurocode 1 [2002] the
effects of vibration due to live load are not required to be considered. Their effect has
already been taken into consideration in definition of design loading.
The ASCE model (1981) does not have any allowance for dynamic load on the
ground that the worst loading occurs with stationary bumper-to-bumper traffic.
2.4. PROVISIONS FOR MULTILANE REDUCTION FACTORS
For multilane bridges, the multiple lane reduction factors are specified in most of
the codes. The approaches to multilane factors vary significantly. They are shown in
Table 2.6 and Table 2.7. The British Standard BS 5400 developed the most compound
procedure of selection of multilane factor, which depend not only on the number of lanes,
but also on loaded length, number and width of notional lanes (Table 2.7). Eurocode does
not define multilane reduction factor, but it gives the load values to be applied on
successive road lanes directly (Table 2.3). The multilane reduction factors are further
discussed in CHAPTER 6 of this dissertation.
21
Table 2.6. Comparison of Multilane Reduction Factors
Number of Lanes
Code
1 2 3 4 5 6 or more
AASHTO LRFD (2007) 1.20 1.00 0.85 0.65 0.65 0.65
OHBDC (1983, 1991) 1.00 0.90 0.80 0.70 0.60 0.55
CAN/CSA-S6-00 (2000) 1.00 0.90 0.80 0.70 0.60 0.55
ASCE (1981) 1.00 0.70 0.40 0.40 0.40 0.40
Table 2.7. Multilane Reduction Factors for BS 5400
Number of Lanes
loaded length [m] 1 2 3 4 or more
0<L≤20 α1 α1 0.6 0.6 α1
20<L≤40 α2 α2 0.6 0.6 α2
40<L≤50 1.0 1.0 0.6 0.6
50<L≤112, N<6 1.0 7.1/√L 0.6 0.6
50<L≤112, N≥6 1.0 1.0 0.6 0.6
L>112, N<6 1.0 0.67 0.6 0.6
L>112, N≥6 1.0 1.0 0.6 0.6
α1 = 0.274 bL
α2 = 0.0137{bL(40-L)+3.65(L-20)}
where bL is the notional lane width
N is total number of notional lanes on the bridge. For a bridge carrying one-way traffic only,
the value N shall be multiplied by 2.
22
2.5. COMPARISON OF EQUIVALENT UNIFORMLY DISTRIBUTED LOADS
In this paragraph, the resulting equivalent uniformly distributed loads for a variety
of design codes and span lengths are plotted and compared. Figure 2.12 and Figure 2.14
compare the equivalent unfactored uniform loads. More valid comparison is to compare
factored loads; these are shown in Figure 2.13 and Figure 2.15. For AASHTO LRFD
(2007) live load factor is 1.75, for OHBDC (1991) it is 1.40, for CAN/CSA-S6-00 it is
1.70, for BS 5400 it is 1.50, and for Eurocode it is 1.35. ASCE studies made no reference
to load factors to be used with its recommended loading, but since a factor of 1.80 has
been used by (Buckland 1991), the same value has been adopted for this comparison.
Equivalent loads including dynamic loads are shown in Figure 2.14, Figure 2.15,
and Figure 2.16. The design loads are increased by the dynamic load factor, DLF, which
has a value as described for each code in the paragraphs above. Since live loadings in
British Standards and Eurocode include dynamic load, they have been used only for
comparison of loads including dynamic load.
Results of the comparison show that variation between the unfactored values of
live load in different codes is significant. European values double those of North
America. The application of load factors slightly reduces the differences. The comparison
of four loaded lanes shows that the differences are reduced even more, Figure 2.16. The
importance of load factors and multilane factors cannot be underestimated.
To obtain plots of UDL, the maximum bending moment (Mmax) was calculated for
simple spans from 400 through 5000 ft. Then, the equivalent uniformly distributed load
UDL was determined from the following formula:
UDL = 8 ⋅ ( M max ) / L2 (2.1)

where: L is span length.
It should be noted that most of the codes are limited to certain lengths of span,
and they are extrapolated beyond these points partially for interest and partially because
these codes are also occasionally used for longer spans due to a lack of adequate guides.
23
2.00 AASHTO HL93
OHBDC 1991
CAN/CSA-S6-00
ASCE 7.5%HV
1.50
equivalent UDL [kip/ft]
ASCE 30% HV
ASCE 100% HV
1.00
0.50
0.00
0 1000 2000 3000 4000 5000
loaded length [ft]
2.00 AASHTO HL93

OHBDC 1991
CAN/CSA-S6-00
ASCE 7.5%HV
1.50
ASCE 30% HV
ASCE 100% HV
1.00
0.50
0.00
0 1000 2000 3000 4000 5000
loaded length [ft]

Figure 2.12. Equivalent Unfactored Loads, w/o IM, w/o multilane factors.
24
3.00 AASHTO HL93 x 1.75
OHBDC 1991 x 1.40
CAN/CSA-S6-00 x 1.70
2.50
ASCE 7.5%HV x 1.80
ASCE 30% HV x 1.80
2.00 ASCE 100% HV x 1.80
1.50
1.00
0.50
0.00
0 1000 2000 3000 4000 5000
loaded length [ft]

Figure 2.13. Equivalent Factored Loads, w/o IM, w/o multilane factors.
AASHTO HL93
2.50
OHBDC 1991
CAN/CSA-S6-00
ASCE 7.5%HV
2.00
ASCE 30% HV
ASCE 100% HV
BS 5400 HA loading
1.50 Eurocode LM1
1.00
0.50
0.00
0 1000 2000 3000 4000 5000
loaded length [ft]
Figure 2.14. Equivalent Unfactored Loads, with IM, w/o multilane factors.
25
3.00 AASHTO HL93 x 1.75
OHBDC 1991 x 1.40
2.50
ASCE 7.5%HV x 1.80
ASCE 30% HV x 1.80
2.00 ASCE 100% HV x 1.80

BS 5400 HA loading x 1.50
1.50 Eurocode LM1 x 1.35
1.00
0.50
0.00
0 1000 2000 3000 4000 5000
loaded length [ft]
Figure 2.15. Equivalent Factored Loads, with IM, w/o multilane factors.
6.00
5.00
4.00
3.00
AASHTO HL93 x 1.75

2.00
OHBDC 1991 x 1.40
BS 5400 HA loading x 1.50
1.00
Eurocode LM1 x 1.35
0.00
0 1000 2000 3000 4000 5000
loaded length [ft]
Figure 2.16. Equivalent Factored Loads, with IM, with multilane factors
for 4 traffic lanes.
26
Table 2.8. Values of Equivalent Unfactored Loads, w/o IM, w/o multilane factors.
length [ft] OHBDC 1991 CAN/CSA-S6-00 HL-93
500 1.151 1.067 0.928
1000 0.918 0.842 0.784
1500 0.841 0.767 0.736
2000 0.802 0.729 0.712
2500 0.779 0.707 0.698
3000 0.763 0.692 0.688
3500 0.752 0.681 0.681
4000 0.744 0.673 0.676
4500 0.737 0.667 0.672
5000 0.732 0.662 0.669
Table 2.9. Values of Equivalent Factored Loads, w/o IM, w/o multilane factors.
length [ft] OHBDC 1991 CAN/CSA-S6-00 HL-93
500 1.612 1.813 1.624
1000 1.286 1.431 1.372
1500 1.177 1.303 1.288
2000 1.123 1.240 1.246
2500 1.090 1.202 1.221
3000 1.068 1.176 1.204
3500 1.053 1.158 1.192
4000 1.041 1.144 1.183
4500 1.032 1.134 1.176
5000 1.025 1.125 1.170
Table 2.10. Values of Equivalent Unfactored Loads, with IM, w/o multilane factors.
length [ft] OHBDC 1991 CAN/CSA-S6-00 HL-93 BS 5400 Eurocode
500 1.336 1.179 1.023 1.603 2.120
1000 1.045 0.898 0.832 1.449 1.985
1500 0.948 0.804 0.768 1.375 1.941
2000 0.900 0.757 0.736 1.328 1.918
2500 0.870 0.729 0.717 1.294 1.905
3000 0.851 0.711 0.704 1.268 1.896
3500 0.837 0.697 0.695 1.246 1.889
4000 0.827 0.687 0.688 1.228 1.884
4500 0.819 0.679 0.683 1.212 1.881
5000 0.812 0.673 0.678 1.198 1.878
27
Table 2.11. Values of Equivalent Factored Loads, with IM, w/o multilane factors.
OHBDC 1991 CAN/CSA-S6-00 HL-93 BS 5400 Eurocode
500 1.871 2.004 1.790 2.404 2.862
1000 1.463 1.526 1.455 2.173 2.680
1500 1.327 1.367 1.343 2.063 2.620
2000 1.259 1.288 1.288 1.993 2.589
2500 1.219 1.240 1.254 1.942 2.571
3000 1.191 1.208 1.232 1.902 2.559
3500 1.172 1.185 1.216 1.869 2.550
4000 1.157 1.168 1.204 1.842 2.544
4500 1.146 1.155 1.194 1.818 2.539
5000 1.137 1.144 1.187 1.797 2.535
Table 2.12. Values of Equivalent Factored Loads, with multilane factors for 4 traffic
lanes.
length [ft] OHBDC 1991 CAN/CSA-S6-00 HL-93 BS 5400 Eurocode
500 5.238 5.612 4.655 5.771 5.309
1000 4.097 4.274 3.783 5.216 4.945
1500 3.716 3.828 3.493 4.952 4.823
2000 3.526 3.605 3.348 4.782 4.762
2500 3.412 3.471 3.261 4.660 4.726
3000 3.336 3.382 3.202 4.564 4.702
3500 3.282 3.318 3.161 4.486 4.684
4000 3.241 3.271 3.130 4.420 4.671
4500 3.209 3.234 3.106 4.364 4.661
5000 3.184 3.204 3.086 4.314 4.653
28
CHAPTER 3
STRUCTURAL RELIABILITY PROCEDURES
3.1. INTRODUCTION
The structures and their components should be designed to have a desirable level
of reliability, which would assure their good performance to account for actions applied
during construction and service. For this purpose civil engineering uses a probabilistic
evaluation of reliability. The design of new structures as well as the evaluation of existing
structures requires verification of limit states, which when exceeded lead to structural
failure (ultimate limit states) or make use of the structure impossible (serviceability limit
states).
The actions (loads, Q ) and structural resistance (capacity, R ) are the variables that
decisively influence the state of a structure. They include uncertainties coming from
mechanical material properties, geometry of a structure, loads, etc. Those uncertainties
can be measured only with the use of probability. Therefore, the design of structures is a
process in which decisions are made under uncertainty and limits. Their rational
treatment, and agreement between real-input data and a mathematical model of
phenomenon, is a concern of structural reliability.
Unreliability of a structure is a state in which a structure does not fulfill design
requirements related to its function and desirable performance. It could be a collapse of a
structure, failure or other deficiency in a structural resistance, unfulfilled service demands
29
of a structure, i.e. excessive deformations, excessive vibrations, etc. Structures usually
have a number of possible failure scenarios. For most of the structures it is impossible to
examine all their failure modes. Therefore, representative failure scenarios have to be
chosen. The analysis usually includes an estimation of structural reliability with respect
to specified failure modes. All modes must be treated separately. Thus, reliability of a
structure is the probability that the system will not reach a specified failure mode related
to a specified limit state during a specified period of time.
3.2. STANDARD VARIABLES AND PROBABILITY DISTRIBUTIONS
The key probabilistic characteristics of a random variable are described in terms

of mean, variance and standard deviation. A distribution function would complete the
description of the probabilistic characteristics of random variables, but sometimes it
remains unknown. There are two types of random variables: discrete and continuous. A
discrete random variable may take on only discrete values. Its probability is given by the
probability mass function, PX ( xi ) . A continuous random variable can take on a
continuous range of values, and its probability is defined by the probability density
function (PDF), f X (x) .
Mean
The mean (expected value) is an average of all observations on a random variable. It is
also defined as the first moment about the origin. For the continuous random variables,
the mean ( µ ) can be computed as:
+∞
µ = ∫x⋅
−∞
f X ( x ) dx (3.1)
For the discrete random variables, the mean is given by:

n
µ = ∑ xi ⋅ PX ( xi ) (3.2)
i =1
If all n observations are given equal weights ( PX ( xi ) = 1 / n ), then the mean for a discrete
random variable is given by:
30
1 n
X= ⋅ ∑ xi (3.3)
n i =1
Variance
The variance ( σ 2 ) is the second moment about the mean, and it is computed as follows:
+∞
σ 2 = ∫ ( x − µ ) 2 ⋅ f X ( x)dx (3.4)
−∞
For the discrete variable, the variance is computed as:

n
σ 2 = ∑ ( xi − µ ) 2 ⋅ PX ( xi ) (3.5)
i =1
If all n observations are given equal weights ( PX ( xi ) = 1 / n ), the variance is as follows:

n
1
σ2 = ⋅ ∑ ( xi − X ) 2 (3.6)
n − 1 i =1
Standard Deviation
The standard deviation ( σ 2 ) of a probability distribution is defined as the square root of
the variance.
Coefficient of Variation
The coefficient of variation ( V ) is a dimensionless quantity defined as:
σ
V =
µ (3.7)
Probability distributions
There are many types of discrete and continuous distributions. The most commonly used
are the continuous distributions: uniform, normal, lognormal, exponential, and gamma. In
this section the normal distribution is presented, because this is the only distribution used
in this dissertation. Further details about distributions can be found, for instance, in
Nowak and Collins (2000).
Normal distribution (Gaussian distribution) is the most widely used probability
distribution. It has a probability density function given by:
31
1 1 x−µ 2
f X ( x) = ⋅ exp[ − ⋅ ( ) ]
σ ⋅ 2⋅π 2 σ (3.8)
This function is graphically represented as shown in Figure 3.1.
FX(x)
random variable X
Figure 3.1 PDF and CDF of a normal random variable
The standard normal distribution is a special case of Gaussian distribution, with

parameters µ X = 0 and σ X = 1.0 . Its PDF is denoted as φ (z) , its CDF is denoted as
Φ(z) , and they are defined as follows:
1 1
φ ( z) = ⋅ exp(− ⋅ z 2 ) = f Z ( z ) (3.9)
2 ⋅π 2
z
1 1
Φ( z ) = ∫
−∞ 2⋅π
⋅ exp(− ⋅ z 2 )dz
2 (3.10)
Central limit theorem states that the sum of a large number of independent observations,
without a dominating distribution type, approaches an approximate normal distribution.
The higher the number of observations the better is the approximation. This theorem is
one of the most important in probability theory. The sum of variables is often used to
model total load acting on a structure, which can be approximated as a normal variable.
32
Mathematically it can be expressed that the sum of n random variables, X 1 , X 2 ,..., X n ,
is equal to function Y having normal distribution:
Y = X 1 + X 2 + ... + X n (3.11)
µY = µ X + µ X + ... + µ X
1 2 n (3.12)
σY 2 =σ X 2 +σ X
1 2
2
+ ... + σ X n
2
(3.13)
3.3. LIMIT STATE FUNCTION
In most design codes, the structural design is based on the concept of limit states.
The philosophy of limit state design assumes equilibrium between applied loads and
structural response of the structure (capacity, resistance). Therefore, a specified set of
load and resistance factors is required for each limit state formulated for different
possible scenarios of structural behavior during construction as well as service life.
Three types of limit states are typically used with reference to structural reliability
analysis:
1. Ultimate limit states (ULSs), which represents the loss of structural capacity.
2. Serviceability limit states (SLSs), which represents failure due to deterioration
of functionality.
3. Fatigue limit states (FLSs), which represents the loss of strength for a structural
component under the action of repeated loading.
The limit state defines the boundary between the desired and undesired
performance of a structure, between situations when the structure is safe (a safety margin
exists) and the structure is not safe (failure occurs). The probability of the desired
performance of a structure is equal to the safety margin ( PS ). The probability of an
undesired performance of a structure is equal to the probability of failure ( P f ). Failure
and non-failure states fulfill the entire probabilistic sample space ( Ω ). Therefore, the
probability of occurrence is P ( Ω ) = 1 , so:
33
P ( Ω ) = P f + Ps = 1 ⇒ Ps = 1 − P f (3.14)
PDF
Reliability analysis usually begins with the formulation of a limit state function
(performance function). All loads are being incorporated into one variable ( Q ) and the
R-Q, safety Q, load effect R, resistance
resistance of Probability
PDFthe structure is being incorporated into one variable ( R ). In the general case, margin
the performance function
PDF of a system can be related to any possible failure scenario or
ofof capacity
any limiting state and defined as a function Failure and demand:
R-Q, safety R-Q,loadsafety
Q,margin effect Q, load effect
R, resistance R, resistance
PDF
margin g (R, Q) = R − Q PDF PDF
(3.15)
R-Q, safety PDF PDF Q, load effect R
R-Q, safety
where Probability
R is capacity representing resistance of a structural
R-Q, safetysystem
margin
Q, load effector a structural
R, resistance element, margin
Probability R-Q,R-Q,safetysafety Q, loadeffect
Q, load effect R, resistanceR, resistance
and Q is demand of Failure
margin
of Failurerepresenting load effect in a structure or a structural component. margin
margin
Both R and Q are continuous random variables having a probability density

Probability
Probability
Probability(PDF). The quantity R − Q is alsoofProbability
function a
Failure
random variable with of
its Failure
own
of Failure Probability PDF as shown
Probability
of Failure 0 ofFailure
of Failure
in the Figure 3.2.
0
PDF
PDF
PDF
0
0
Figure 5-8 PDFs of load, resistance, and safety margin.
0
0 0
R- Q Q 0 R
R-Q,safety
R-Q, safety Q, load effect R,resistance
resistance
Resistance
Safety
margin
margin Q, load effect Load R,
5-8 PDFs of load, resistance, and safety
PDF margin Figure Figure 5-8 maPD
Figure 5-8 PDFs of load, resistance, and safety margin. Figure 5-8 PDFsofload, resistance, andsafetymargin.
0 R-Q, safety Q, load effect R, resistance 5-8 PDFs of load, resistance, and safety margin.
margin Figure 5-8 PDFs of load, resistance, and safetyFiguremargin.
Probability
Probability
Probability
ofofFailure
of Failure Probability
Failure of Failure

0
Fig
00 random value
Figure 3.2 PDF’s of load, resistance, and safety margin
Figure 5-8
Figure PDFsofofload,
5-8 PDFs load,resistance,
resistance,and
andsafety
safetymargin.
margin.
The performance function usually is a function of capacity and demand variables (

X 1 , X 2 ,..., X n ), and it adopts a format:
g ( X ) = g ( X 1 , X 2 ,..., X n )
(3.16)
where X i is the collection of input parameters.
34
From the definition of the performance function, it can be derived that when
g ( X ) < 0 , it indicates failure and, when g ( X ) ≥ 0 , it indicates acceptable performance.
The performance defined as g ( X ) = 0 is called the failure surface. The corresponding
probability of failure can be defined as the integral of the joint density function of the
variables over the negative domain of g ( X ) (Thoft-Christensen and Baker 1982):
Pf = ∫ ....∫ f
g ( X ) <0
X 1 , X 2 ,...,X n (x1 , x 2 ,..., xn )dx1 ⋅ ... ⋅ dxn
(3.17)
where f X is the joint probability density function of X 1 , X 2 ,..., X n

There is almost never sufficient data to define the joint probability density
function for all basic variables, thus the equation (3.17) is very difficult to evaluate. Even
knowing the joint density function, the necessary multi-dimensional integration may be
extremely time-consuming. In practice, a direct calculation of the probability of failure
becomes inefficient. Therefore, indirect procedures such as the reliability index are used.
3.4. RELIABILITY INDEX
The evaluation reliability index, also called safety index, is an effective measure
of the probability of failure. There are several methods to calculate reliability of structural
components: the first-order reliability methods (FORM), the advanced first-order second-
moment methods (FOSM), simulation techniques, etc. In the late 1960s, the first-order
second-moment formulation was developed and advanced by Cornell (1967) and Ang
and Cornell (1974). Further advances in these methods were made by Hasofer and Lind
(1974) and Rackwitz and Fiessler (1978). The FOSM methods can be used to solve many
practical problems. The concept of second-moment is often used in practical
quantification of safety and reliability. It has been extensively used in calibrations of
structural design codes. The FOSM approach can be put into several categories with
regard to accuracy of results, required input data, computing cost, or simplicity of
formulation.
35
In 1974, Hasofer and Lind introduced the definition of the reliability index as the
shortest distance from the origin to the limit state function in a system of reduced
variables coordinates (Figure 3.3). Using geometry the reliability index can be calculated
as:
µ R − µQ
β= (3.18)
σ R2 + σ Q2
g(ZR,ZQ) = 0
ZR limit state function
SAFE
µ R-µ Q β ZQ
σQ µ R-µ Q
σQ
FAILURE
Figure 3.3 Reliability index defined as the shortest distance in the space of reduced
variables
All the variables should be expressed in non-dimensional forms. Thus, the

reduced variables ( Z R , Z Q ) have to be introduced:
R − µR
ZR =
σR
R − µQ (3.19)
ZQ =
σQ
The resistance ( R ) and the load ( Q ) can also be expressed in the form of reduced
variables:
36
R = µR + ZR ⋅σ R
Q = µQ + Z Q ⋅ σ Q (3.20)
Therefore the limit state function g ( R, Q) = R − Q in terms of reduced variables can be

rewritten as:
g (Z R , Z Q ) = µ R + Z R ⋅ σ R − µ Q + Z Q ⋅σ Q
(3.21)
The reliability index recognizes the importance of uncertainty in load effects and
strength. It incorporates the four key parameters, resistance and load with their mean
values and standard deviations, µ R , µ Q and σ R , σ Q , respectively.
The limit state function used in this dissertation is linear. In case it was nonlinear,
iteration would be required to find the design point in reduced variable space such that β
corresponds to the shortest distance. Moreover, the Hasofer-Lind approach evaluates the
reliability index for uncorrelated random variables. Thus, if the initial variables are
correlated they must be transformed into uncorrelated random variables.
The probability of failure can be calculated using the formula (3.22). The
calculation can give exact results, if the random variables are normally distributed and
uncorrelated. Otherwise it provides only an approximation.
Pf = Φ ( − β )
(3.22)
where Φ is the standard normal distribution function.
3.5. MONTE CARLO METHOD SIMULATION TECHNIQUE
Simulation is the process of replacing reality with theoretical and experimental

models. Theoretical simulation is also called numerical or computer experimentation. It is
a practical tool that allows obtaining data, either instead of or in addition to real-world
data. Simulating a phenomenon numerically assumes events occur a finite number of
times. The frequency of occurrence of an event in the entire set of simulations
approximates its probability of occurrence. This relatively straightforward concept often
requires complex procedures. The most commonly used simulation technique is the
Monte Carlo Method (Thoft-Christensen and Baker, 1982; Hart, 1982). Many other
37
simulation techniques work similarly to the Monte Carlo Method. However instead of
generating a large number of random values for variables, the values are arbitrary and
selected according to specified rules.
The Monte Carlo Method is commonly used to predict the behavior of structural
elements and systems from the probability point of view, and without physical testing. It
is used to evaluate the probability of structural failure and, indirectly, the reliability
index. A large amount of random numbers corresponding to random variables is
numerically generated. This large number of repetitions is particularly valuable in solving
problems involving rare events, where physical testing could be very expensive. This
method may be used not only to study performance of a structural system for a prescribed
set of design variables, but also to measure sensitivity in system performance due to
variations of some parameters. Therefore, for engineering purpose, it can be used to
determine optimal design.
The Monte Carlo simulation method consists of the following steps. In the first
step a simulation of the uniformly distributed random numbers u1 , u 2 ,..., u n between 0
and 1 is performed. They can be generated by computer programs using a built-in option.
Then, the standard normal random values can be calculated using generated numbers and
information about the types of distributions and statistical parameters of distributions
(mean value and standard deviation) of each design variable. The standard normal
random number zi is calculated using the equation:
z i = Φ − 1 (u i ) (3.23)
where:
Φ −1 is the inverse of the standard normal cumulative distribution function
Using standard random values (mean value µ X and standard deviation σ X ), the
values xi of sample random numbers can be generated for the random normal variable X
, as:
xi = µ X + z i ⋅ σ X (3.24)
38
It is important to simulate an efficient number of sets, such that the variation of
the design parameters in a single simulation will not influence the solution of the entire
process of simulations.
Performed Monte Carlo simulations allow for estimation of the probability of
failure. The probability of failure is defined as the ratio between the numbers of times the
criterion for the failure is achieved ( n ), to the total number of simulations ( N ). Each
simulated value has the same weight.
n total number of simulations when g( X ) < 0
Pf = = (3.25)
N total number of simulations of g( X )
where:
g ( X ) defines the performance function with the limit state g ( X ) = 0
g ( X ) < 0 defines the state of failure
3.6. NORMAL PROBABILITY PAPER
Normal probability paper is used to present cumulative distribution functions

(CDF) in a convenient way. Cumulative distribution functions for the normal distribution
are “S-shape” function. Normal probability paper redefines the vertical scale so that the
normal CDF can be plotted as a straight line and allows for an easy evaluation of the
most important statistical parameters as well as type of distribution function. More
detailed information about normal probability paper can be found in textbooks (Nowak
and Collins 2000, Benjamin and Cornell 1970). The basic variable is presented on the
horizontal. The vertical axis, being the standard normal variable, represents the distance
from the mean value in terms of standard deviations. It can also be implied as the
corresponding probability of being exceeded. The relationship between the standard
normal variable and probability is given in Table 3.1.
39
Table 3.1. Relationship between vertical scale on Normal Probability Paper
and Probability
Distance from the mean value in terms of
Corresponding probability
standard deviations
4 0.9999683
3 0.99865
2 0.9772
1 0.841
0 0.5
-1 0.159
-2 0.0228
-3 0.00135
-4 0.0000317
The shape of the resulting curve representing CDF allows for analysis of the test
data plotted on the normal probability paper.
The basic properties of the normal probability paper:

A straight line represents a normal distribution function.
The mean value and standard deviation read directly from the graph.
The mean value is at the intersection of the normal CDF and horizontal axis.
The standard deviation can also be read as shown in Figure 3.4.
40
Standard Normal CDF
Variable F(x)
3
mean
0
-1
standard deviations
-2
Figure 3.4 Normal Distribution Function on the Normal Probability Paper.
41
CHAPTER 4
TRAFFIC DATA
4.1. REGULATIONS OF TRUCK TYPES, TRUCK SIZES AND WEIGHT LIMITS
Federal and state regulations limit the weight and dimensions of vehicles on U.S.
highways. These restrictions have important impacts on highway construction costs,
maintenance costs, and highway safety issues. Current Federal law includes the following
limits:
- 20 000 pounds - maximum gross weight upon any one axle
- 34 000 pounds - maximum gross weight on tandem axles
- 80 000 pounds - maximum gross vehicle weight
- 102 inches - maximum vehicle width
- 48-feet - minimum vehicle length for a semi-trailer in a truck-tractor/semi-
trailer combination
- 28 feet - minimum vehicle length for a semi-trailer or trailer operating in a
truck-tractor/semi-trailer/trailer combination.
The types of the vehicles in use on American roads are classified by FHWA into 13-
categories, as show in Figure 4.1. Classes 1-3 are passenger vehicles, classes 4-7 are
single unit trucks and buses, classes 8-10 are combination trucks, classes 11-13 are multi-
trailer trucks.
42
Figure 4.1. FHWA 13-category scheme
43
Table 4.1. Conversion chart for vehicles’ class and number of axles
Average Number of Axles
Vehicle Class
per Vehicle
1 2
2 2
3 2
4 2.2
5 2
6 3
7 4
8 4
9 5
10 6
11 4
12 6
13 7
Several states issue overweight permits and allow higher truck loads. For example
the state of Michigan, from where some publications on field test results are used in this
study, allows trucks up to 164,000 pounds. The states which allow various longer
combination vehicles are presented in Figure 4.3. Types of longer combination vehicles
(Figure 4.2) are:
- Rocky Mountain Double (common maximum weight – 105,500 - 137,800 lbs)
- Turnpike Double (common maximum weight – 105,500 - 129,000 lbs)
- B-train Double Trailer Combination (common maximum weight – 105,500 -
147,000 lbs)
- Triple Trailer Combination (common maximum weight – 105,500 - 131,000 lbs)
44
Rocky Mountain Double Turnpike Double
8-Axle B-Train Double Trailer Combination
Triple Trailer Combination
Figure 4.2. Longer Combination Vehicles (LCV’s)
States Allowing LCVs’ States Allowing Triples
States Allowing Turnpike Doubles States Allowing Rocky Mountain Doubles
Figure 4.3. States allowing various Longer Combination Vehicles
45
Federal size and weight studies were established in 1982, and since then no
significant changes have been made. However, several proposals to make changes in
these regulations were presented. The most recent studies are the TRB Special Report
267 "Regulation of Weights, Lengths, and Widths of Commercial Motor Vehicles" and
the U.S. Department of Transportation "Comprehensive Truck Size and Weight Study:
Volume I Summary Report". Both documents discuss existing regulations and give
recommendations on their improvement. Elimination of the federal 80 000 pounds weight
limit on Interstate highways is recommended. It is proposed that the gross weight should
be governed by appropriate axle weight limits and the bridge formula (Figure 4.4). The
maximum weight (in pounds) carried on a group of two or more consecutive axles would
not exceed that given by the following formulas:
- W = 1000*(2L+26) for L≤24 ft
- W = 1000*(L/2+62) for L>24 ft
6.0
5.0 for L≤24 ft

for L>24 ft
vehicle weight [kip/ft]
4.0
3.0
2.0
1.0
0.0
0 20 40 60 80 100 120
vehicle length [ft]
Figure 4.4. New Bridge Formula - regulation of vehicles' length and weight
46
In the Figure 4.5, the data from the Transportation Statistics Annual Report
December 2006 of U.S. Department of Transportation is shown. As can be observed, the
number of trucks has a trend of rapid growth. The number of heavy trucks is growing
faster than number of light trucks. The number of light trucks (under 10 000 pounds)
increased 73 percent between 1992 and 2005, and the number of heavy trucks (greater
than 10 000 pounds) increased 112 percent. In 2005 heavy trucks constituted 8% of the
volume of trucks, while in the 1990’s they were only 4% of the volume. In 2005, 95.3
million light trucks traveled 1.060 trillion vehicle-miles, and 8.5 million heavy trucks
traveled 222.29 billion vehicle-miles.
120000
95300
100000
68100 79760
80000
55193 light trucks
60000
heavy trucks
40000
20000 5400 8500

4007 4700
0
1992 1997 2002 2005
Figure 4.5. Number of trucks by weight (in thousands of trucks).

Transportation statistics annual report, December 2006.
According to Texas Transportation Institute “Over the next 20 years, truck tonnage is
expected to increase at a rate more than five times that of population growth.”
47
25000
Tons (millions) 20000
15000
10000
5000
0
1995 2000 2005 2010 2015 2020 2025 2030 2035
Year
Figure 4.6. Freight Tonnage Moved by Truck (FHWA)
Figure 4.7 presents time variation of total truck weight statistic between the years
1993 and 2003. The data is expressed as: mean value (µ), 95th percentile (W95, 95
percent of the trucks weigh less), and maximum observed total truck weight (Max). The
study was made for the state of New Jersey, which has lower limits than the state of
Michigan. However, the observed maximum gross vehicle weight is high, and it reaches a
value of 225 kips (1000 kN). The maximum truck weight shows steady increase at an
annual growth rate of 1.2%.
48
Figure 4.7. Time variation of total truck weight statistic.
(Gindy, M., Nassif, H.H., 2006)
4.2. DATA COLLECTION METHODOLOGY (WIM)
Weigh in Motion (WIM) Technology had its beginnings in the early 1950s when
the U.S. Bureau of Public Roads, the Virginia State Department of Highways, and the
Williams Construction Company installed a load cell WIM system on the Henry G.
Shirley Memorial Highway. From these early beginnings, WIM technology and
application continued to advance and spread across the nation. In 1990, the American
Society for Testing and Materials (ASTM) published the first Standard Specification for
Highway Weigh-in-Motion (WIM) Systems with User Requirements and Test Methods
(Designation: E 1318-90). This document was revised in 1994, and again in 2002 to the
version (Designation: E 1318-02) that is used today.
ASTM Designation: E 1318-02 defines WIM as “the process of measuring the
dynamic tire forces of a moving vehicle and estimating the corresponding tire loads of the
49
static vehicle”. In addition to the collection of dynamic tire forces, a variety of ancillary
traffic data can also be obtained through the use of WIM systems: traffic volume, speed,
directional distribution, lane distribution, date and time of passage, axle spacing, axle
weight, and vehicle classification. Of all data collection methodologies, WIM data
collection requires the most sophisticated technology for data collection sensors, the most
controlled operating environment (smooth, level pavement), as well as the highest
equipment set-up and calibration costs. The primary reason for sophistication in
technology and its high costs comes from a need to determine static weight from a
dynamic measurement. In standard weigh scale application, vehicles are stopped on a
static scale and are measured without any interaction between the vehicle and the
roadway. In WIM applications, a variety of forces are acting on the vehicle, including the
force of gravity as well as dynamic effects of influences such as: roadway roughness,
vehicle speed, vehicle acceleration and deceleration, out of balance tires and wheels, tire
inflation pressure, suspension, aerodynamics and wind; and other dynamic factors.
Several different technologies are available for WIM data collection systems. The
most commonly used are: piezoelectric cables, bending plate, load cell, quartz cables, and
Bridge WIM systems. ASTM Standard Designation: E 1318-02 distinguishes four types
of WIM systems (Type I, II, III, and IV) based on application and performance
requirements for data collection. Each type of WIM system has been specified to perform
its indicated functions within specific tolerances. The piezoelectric sensors, which are the
most common, offer acceptable accuracy ±15%. Strain based and load cell WIM systems
are much more expensive, but they provide more accuracy. Recently, piezo-quartz
sensors were introduced in the United States. They are less sensitive to temperature
changes and generally more accurate. The majority of WIM data collection is done with
permanently installed weight sensors, although the data is not always collected
continuously.
In order to assure unbiased data, WIM sites should be localized away from weight
stations and be unknown to truck drivers’. WIM equipment should be subject to a regular
maintenance and calibration. To limit erroneous data, it is recommended to avoid sites
with numerous traffic stoppages (speed >10 mph), close to exits, and with rough surfaces.
50
Highway agencies have recognized the advantages of having automated data
collection systems that can provide information on truck weights and truck traffic
patterns for economic analysis, traffic management and various other purposes.
Therefore, the quality and quantity of WIM data has greatly improved in recent years,
and new WIM technologies continue to be developed. Due to the weigh-in-motion
technologies, unbiased truckloads are being collected at normal highway speeds, in large
quantity, and without truck driver’s knowledge (Figure 4.8).
Figure 4.8. WIM data collection
4.3. WEIGH IN MOTION DATABASE
For the scope of this research, the weigh in motion database was obtained from
the project NCHRP 12-76 and measurements on the Throggs Neck Bridge in New York.
The database includes newest (2001-2006) WIM database for a variety of sites:
California (6), Florida (5), Indiana (6), Mississippi (5), New York (7+2), Oklahoma (16),
51
and Oregon (4). The variety of sites is important, because the truck traffic changes
depending on the site location: interstate or non-interstate highways, rural or urban areas,
and state. Traffic data varies also depending on time of day, day of the week, season of
the year, and direction. Therefore, it is important that the WIM database is collected
continuously (mostly one-year data), on many traffic lanes and in both directions
(usually). The database contains date and time, lane, number of axles, spacing between
axles, axle weight, speed, and vehicle category. A summary of WIM data, including site
localizations, number of lanes, and types of sensors used is presented in Table 4.2.
Distribution of vehicles by axles and traffic lanes are presented in Table 4.3 - Table 4.10.
The statistics is presented for four states selected for simulations: Oregon, Indiana,
Florida, and New York. Cumulative distribution functions of gross vehicle weights by
axles (GVW) were plotted in Figure 4.9-Figure 4.11. It can be noticed, that for New York
I-495, the heaviest vehicles are 6-axles. Those are construction debris, gravel and garbage
haulers. They often drive overloaded above 150 kips and occasionally above 200 kips,
while NYSDOT routine permit trucks that are legal up to 120 kips.
The WIM technology is known to have certain traffic data quality problems. The
errors are due to physical and software-related failures of equipment and transmission,
the difference between the dynamic weight measured and the actual static scale weight,
as well as the effect of tire pressure, size and configuration of the WIM results.
Therefore, the data quality checks have to be implemented to detect and fix/eliminate
erroneous data before processing. A standardized procedure to filter out errors is applied
to WIM data from various sites. According to Traffic Monitoring Guide (2001),
reasonableness checks were performed on the axle weights and spacing. The limits were
200 to 20,000 kilograms (0.44 to 4.41 kips) for axle weights and 0.5 to 15 meters (1.6 to
49.2 feet) for axle spacing. Moreover, all obvious errors such as zero readings for number
of axles or speed were eliminated. The percentage of filtered out data varies for different
sites, which depends on condition of equipment, its regular maintenance and
recalibration.
52
Table 4.2. Summary of WIM Data
# of Traffic # of WIM
State Site ID Route Both Dir WIM Type
Lanes Lanes
CA 0001 Lodi
CA 0003 Antelope
CA 0004 Antelope
CA 0059 LA710
CA 0060 LA710
CA 0072 Bowman
FL 9916 US-29 4 4 Y P
FL 9919 I-95 4 4 Y P
FL 9926 I-75 6 4 Y BP
FL 9927 SR-546 4 4 Y BP
FL 9936 I-10 4 4 Y P
IN 9511 I-65 4 4 Y P
IN 9512 I-74 4 4 Y SLC
IN 9532 US-31 4 4 Y P
IN 9534 I-65 6 6 Y P
IN 9544 I-80/I-94 6 6 Y P
IN 9552 US-50 2 2 Y P
MS 2606 I-55 4 4 Y P
MS 3015 I-10 4 4 Y P
MS 4506 I-55 4 4 Y P
MS 6104 US-49 2 2 Y P
MS 7900 US-61 4 4 Y P
NY 8280 I-84 4 4 Y P
NY 8382 I-84 4 4 Y P
OR Woodburn I-5 3 2 N SLC
OR Emigrant Hill I-84 2 1 N SLC
OR Lowell OR 58 2 2 N SLC
OR Bend US 97 2 1 N SLC
NY 9121 I-81 2 2 Y P
NY 2680 8 4 4 Y P
NY I-495 Y
P – Piezo, BP – Bending Plate, SLC – Single Load Cell
53
Table 4.3. Vehicles by axle in Oregon
axles total I-5 Woodburn (NB) I-84 Emigrant Hill (WB) OR 58 Lowell (WB) US 97 Bend (NB)
2 44507 4.6% 36959 6.04% 3333 1.56% 2273 2.48% 1942 3.28%
3 71365 7.3% 42009 6.87% 9242 4.34% 9807 10.70% 10307 17.40%
4 62025 6.4% 31066 5.08% 14728 6.91% 8032 8.76% 8199 13.84%
5 575846 59.0% 350107 57.22% 140520 65.97% 57123 62.30% 28096 47.44%
6 70639 7.2% 46792 7.65% 14441 6.78% 3789 4.13% 5617 9.48%
7 85658 8.8% 58407 9.55% 19003 8.92% 5725 6.24% 2523 4.26%
8 60907 6.2% 43947 7.18% 10041 4.71% 4764 5.20% 2155 3.64%
9 3500 0.4% 2253 0.37% 894 0.42% 131 0.14% 222 0.37%
10 697 0.1% 210 0.03% 404 0.19% 26 0.03% 57 0.10%
11 345 0.0% 46 0.01% 239 0.11% 20 0.02% 40 0.07%
12 277 0.0% 34 0.01% 172 0.08% 6 0.01% 65 0.11%
sum 975766 611830 213017 91696 59223
Table 4.4. Vehicles by traffic lane in Oregon

lane I-5 Woodburn (NB) I-84 Emigrant Hill (WB) OR 58 Lowell (WB) US 97 Bend (NB)
1 552388 90.28% 213017 100.00% 51404 56.06% 59223 100.00%
2 59442 9.72% 40292 43.94%
3
4
sum 611830 213017 91696 59223
54
Table 4.5. Vehicles by axle in Florida
axles total Florida 9916 Florida 9919 Florida 9926 Florida 9927 Florida 9936
2 2400362 25.9% 482051 65.98% 229680 10.11% 1408095 36.45% 127986 19.70% 152550 8.37%
3 779456 8.4% 67058 9.18% 124347 5.47% 402640 10.42% 115466 17.77% 69945 4.12%
4 747991 8.1% 24388 3.34% 156406 6.88% 430324 11.14% 74984 11.54% 61889 3.50%
5 5156752 55.6% 148424 20.31% 1720367 75.69% 1554039 40.23% 297593 45.81% 1436329 82.08%
6 143470 1.5% 7011 0.96% 37909 1.67% 56038 1.45% 11271 1.74% 31241 1.76%
7 18988 0.2% 888 0.12% 2847 0.13% 6990 0.18% 6296 0.97% 1967 0.11%
8 8495 0.1% 410 0.06% 867 0.04% 2475 0.06% 4051 0.62% 692 0.03%
9 15755 0.2% 422 0.06% 405 0.02% 2533 0.07% 11977 1.84% 418 0.02%
10
11
12
sum 9271269 730652 2272828 3863134 649624 1755031
Table 4.6. Vehicles by traffic lane in Florida

lane Florida 9916 Florida 9919 Florida 9926 Florida 9927 Florida 9936
1 249213 34.11% 915451 40.28% 897806 23.24% 229181 35.28% 818460 46.64%
2 74085 10.14% 237260 10.44% 1119347 28.98% 88830 13.67% 114341 6.52%
3 407354 55.75% 181535 7.99% 0 0.00% 62778 9.66% 112767 6.43%
4 938583 41.30% 0 0.00% 268830 41.38% 709463 40.42%
5 1044850 27.05%
6 801070 20.74%
sum 730652 2272829 3863073 649619 1755031
55
Table 4.7. Vehicles by axle in Indiana
axles total Indiana 9511 Indiana 9512 Indiana 9532 Indiana 9534 Indiana 9544 Indiana 9552
2 2527382 27.0% 44867 10.11% 16938 8.38% 738274 57.08% 1509944 24.18% 82330 9.91% 135029 37.48%
3 513522 5.5% 15135 3.41% 7400 3.66% 91398 7.07% 353721 5.67% 26523 3.19% 19345 5.37%
4 571231 6.1% 12517 2.82% 3170 1.57% 120486 9.31% 385840 6.18% 26400 3.18% 22818 6.33%
5 5654115 60.3% 364519 82.13% 171368 84.79% 334457 25.86% 3928062 62.91% 675794 81.38% 179915 49.93%
6 95770 1.0% 6534 1.47% 2993 1.48% 7359 0.57% 60292 0.97% 15688 1.89% 2904 0.81%
7 7547 0.1% 193 0.04% 159 0.08% 836 0.06% 3740 0.06% 2381 0.29% 238 0.07%
8 2967 0.0% 59 0.01% 54 0.03% 486 0.04% 1501 0.02% 829 0.10% 38 0.01%
9 945 0.0% 10 0.00% 14 0.01% 100 0.01% 396 0.01% 414 0.05% 11 0.00%
10 355 0.0% 8 0.00% 18 0.01% 46 0.00% 225 0.00% 52 0.01% 6 0.00%
11 131 0.0% 4 0.00% 6 0.00% 20 0.00% 89 0.00% 12 0.00% 0 0.00%
12 88 0.0% 0 0.00% 0 0.00% 21 0.00% 67 0.00% 0.00% 0 0.00%
sum 9374053 443846 202120 1293483 6243877 830423 360304
Table 4.8. Vehicles by traffic lane in Indiana

lane Indiana 9511 Indiana 9512 Indiana 9532 Indiana 9534 Indiana 9544 Indiana 9552
1 375357 84.57% 185645 91.85% 823175 63.64% 3232127 51.76% 412695 49.70% 360304 100.00%
2 68489 15.43% 16475 8.15% 470308 36.36% 2657717 42.57% 393046 47.33% 0 0.00%
3 0 0.00% 0 0.00% 354033 5.67% 24682 2.97% 0
4 0 0.00% 0 0.00% 0 0.00% 0
sum 443846 202120 1293483 6243877 830423 360304
56
Table 4.9. Vehicles by axle in New York
axles total I-495 EB I-495 WB NY 9121 NY 2680
2 114115 41.6% 56459 38.58% 57656 45.02% 108563 7.88% 34752 25.34%
3 40359 14.7% 21102 14.42% 19257 15.04% 135919 9.87% 23644 17.24%
4 19297 7.0% 10605 7.25% 8692 6.79% 74822 5.43% 14154 10.32%
5 82959 30.2% 47294 32.32% 35665 27.85% 1010780 73.39% 52845 38.54%
6 17426 6.4% 10716 7.32% 6710 5.24% 44357 3.22% 9720 7.09%
7 217 0.1% 131 0.09% 86 0.07% 1758 0.13% 1113 0.81%
8 27 0.0% 20 0.01% 7 0.01% 542 0.04% 560 0.41%
9 20 0.0% 19 0.01% 1 0.00% 335 0.02% 212 0.15%
10 0 0.0% 0 0.00% 0.00% 182 0.01% 98 0.07%
11 0 0.0% 0.00% 0.00% 18 0.00% 25 0.02%
12 0 0.0% 0.00% 0.00% 8 0.00% 4 0.00%
13 0.0% 0.00% 0.00% 2 0.00% 1 0.00%
sum 274420 146346 128074 1377284 137127
Table 4.10. Vehicles by traffic lane in New York

lane I-495 EB I-495 WB NY 9121 NY 2680
1 52703 36.01% 43278 33.79% 618289 44.89% 46008 33.55%
2 90625 61.93% 78891 61.60% 88037 6.39% 3148 2.30%
3 3018 2.06% 5905 4.61% 74831 5.43% 8298 6.05%
4 0 0.00% 0.00% 596129 43.28% 79677 58.10%
sum 146346 128074 1377286 137131
57
I-5 Woodburn (NB)
3.00
2.00
Inverse of Standard Normal Distribution
1.00 2-axle
3-axle
4-axle
5-axle
6-axle
0.00
7-axle
0 20 40 60 80 100 120 140 160 180 200 220 240
8-axle
9-axle
10-axle
-1.00 11-axle
12-axle
-2.00
-3.00
GVW [kips]
Figure 4.9. CDF's of GVW by axles Oregon I-5 Woodburn
58
I-84 Emigrant Hill (WB)
3.00
2.00
2-axle
1.00
3-axle
4-axle
5-axle
6-axle
0.00
7-axle
0 20 40 60 80 100 120 140 160 180 200 220 240
8-axle
9-axle
10-axle
-1.00
11-axle
12-axle
-2.00
-3.00
GVW [kip]
Figure 4.10. CDF's of GVW by axles Oregon I-84 Emigrant Hill
59
CDF's of GVW by axles
NY I-495 EB
5.00
4.00
3.00
2.00
2-axle
1.00 3-axle
4-axle
5-axle
0.00
6-axle
0 20 40 60 80 100 120 140 160 180 200 220 240
7-axle
-1.00 8-axle
9-axle
-2.00
-3.00
-4.00
-5.00
Figure 4.11. CDF's of GVW by axles NY I-495 EB
60
CHAPTER 5
DEVELOPMENT OF LIVE LOAD MODEL
5.1. INTRODUCTION
For long span bridges, the extreme live load is governed by the traffic jam
scenario. The live load is modeled as the uniformly distributed lane load and additional
axle load or single truck for deck components. Development of live load model is based
on three approaches. Two of them can be classified as initial studies. The first of them is
based on a 5-axle average truck and the second one is based on AASHTO LRFD legal
load trucks. The third approach is detailed study based on truck WIM Data.
5.2. MODEL BASED ON AVERAGE 5-AXLE TRUCK
For computation of the live load on the most loaded lane, the following traffic model
has been assumed:
- Traffic jam situation, Figure 5.1.
- Left lane loaded only with average trucks.
- Average trucks are 5-axle trucks, which are the most popular among truck
types, Figure 5.4 and Figure 5.5. In the FHWA WIM Data, vehicle categories 1-
3 FHWA are omitted, therefore the percentage of 2-axle vehicles is relatively
low.
61
- An average 5-axle truck:
- is 45 ft long
- weights 55 kips, Figure 5.2 and Figure 5.3.
- Clearance distance is 10 to 15 ft, therefore spacing between the last axle of one
truck and first axle of the following truck is 20-25 ft.
Live load due to such a combination of vehicles is equal to:

55 kip / 70 ft = 0.79 kip/ft for clearance distance of 15 ft
55 kip / 65 ft = 0.85 kip/ft for clearance distance of 10 ft
Figure 5.1. Critical loading. Traffic jam scenario.

5.00
4.00
3.00
inverse of standard normal distribution
2.00
1.00
0.00
0 20 40 60 80 100 120 140 160 180 200
-1.00
-2.00
-3.00
-4.00
-5.00
GVW [kips]
Figure 5.2. CDF’s of GVW for 5-axle trucks. New York WIM Data.
62
Figure 5.3. CDF of GVW for 5-axle and 11-axle trucks
(Nowak, A.S., Laman, J. and Nassif, H., 1994)
70%
60% Oregon
50% Florida
40% Indiana
30% New York
20%
10%
0%
2 3 4 5 6 7 8 9 10 11 12
number of axles
Figure 5.4. Percentage of vehicles by number of axles. FHWA WIM Data.
63
50%
43.2%
40%
30%
22.4%
20%
10.9%
10% 6.5% 5.1% 6.3%
2.6% 1.5% 0.8% 0.5%
0%
2 3 4 5 6 7 8 9 10 11 12
number of axles
Figure 5.5. Percentage of vehicles by number of axles.

(Kim, S-J., Sokolik, A.F., and Nowak, A.S., 1997)
If we would like to consider all types of trucks, not only 5-axles, the result would
be similar. The mean value of GVW is above 50 kips, Figure 5.6.
5.00
4.00
3.00
2.00 US 97 Bend
I-84 Emigrant Hill
1.00 OR 58 Lowell
I-5 Woodburn
0.00
0 20 40 60 80 100 120 140 160 180 200 220
-1.00
-2.00
-3.00
-4.00
-5.00
GVW [kip]
Figure 5.6. CDF’s of GVW for all types of vehicles in Oregon.

64
5.3. MODEL BASED ON LEGAL LOAD TRUCKS
The second approach to model live load is based on vehicles called "legal load
types", which are developed in AASHTO. Some of the states use them for rating instead
of the traditional HS-20 load. These vehicles were selected to match the federal bridge
formula (known as Formula B) for vehicles up to 80 000 lb. While HS-20 was intended
to be appropriate for all span ranges, three legal AASHTO vehicles (Type 3 Unit, Type
3S2 Unit, Type 3-3 Unit) are supposed to adequately model short vehicles and a
combination of vehicles for short, medium and long spans respectively.
To model traffic jam situations on long span bridge the Type 3-3 Units have been
placed in a lane with the clearance distance of 10 to 15 ft. Therefore spacing between the
last axle of one truck and first axle of the following truck is 20-25 ft. Figure 5.7. Gross
Vehicle Weight of a Type 3-3 Unit is 80 kips and total length of a Type 3-3 Unit is 54 ft,
therefore:
80 kips / (54+25) ft = 1.01 kip/ft for clearance distance of 15 ft
80 kips / (54+20) ft = 1.08 kip/ft for clearance distance of 10 ft
Since the value obtained in this way is based on heavy trucks and it is very
conservative, its value can be multiplied by factor 0.75. This approach derives from basic
philosophy used to develop lane load of 0.64 kip/ft.
0.75 x 1.01 kip/ft = 0.76 kip/ft
0.75 x 1.08 kip/ft = 0.81 kip/ft
12 kip 12 kip 12 kip 16 kip 14 kip 14 kip 12 kip 12 kip 12 kip 16 kip 14 kip 14 kip
15.0’ 4.0’ 15.0’ 16.0’ 4.0’ 15.0’ 4.0’ 15.0’ 16.0’ 4.0’
54.0’ 20.0’ - 25.0’ 54.0’

Type 3-3 Unit Type 3-3 Unit
Figure 5.7. AASHTO LRFD legal load trucks, Type 3-3 Units.
65
5.4. MODEL BASED ON TRAFFIC JAM SIMULATION USING WIM DATA
The considered WIM data was obtained from NCHRP 12-76, described in
CHAPTER 4. The available data served as a basis for simulation of a traffic jam
situation. Starting with the first truck, all consecutive trucks were added with a fixed
clearance distance between them until the total length reached the span length (Figure
5.8). Then, the total load of all trucks was calculated and divided by the span length to
obtain the first value of the average uniformly distributed load. Next, the first truck was
deleted, and one or more trucks were added so that the total length of trucks covers the
full span length and the new value of the average uniformly distributed load was
calculated. The calculations were performed for span lengths 600, 1000, 2000, 3000,
4000, and 5000 ft. Trucks were kept in actual order, as recorded in the WIM surveys.
Clearance distance is assumed to be about 15 ft, while spacing between the last axle of
one truck and first axle of the following truck is 25 ft. Clearance concept is as defined as
in Figure 5.9, and according to literature it varies between 6 and 21 ft. Spacing between
the last axle of one truck and first axle of the following truck is clearance plus distance
from first and last axles to corresponding bumpers, based on the most common 5-axle
truck WB-20 defined in “AASHTO Geometric Design of Highways and Streets. Only the
most loaded lane was considered. It was assumed that in a traffic jam situation, light
vehicles are using faster lanes, therefore, vehicles of the 1-3 FHWA category were
omitted.
Figure 5.8. Simulation of trucks moving throughout span length
66
Figure 5.9. Clearance - Gap and Spacing - Headway Concepts
Figure 5.10. Interstate semitrailer WB-20

(AASHTO Geometric Design of Highways and Streets)
To perform an extensive number of simulations on the huge databases for each

site and different span lengths, a numerical procedure based on Visual Basic coding were
developed. WIM data, such as weight and length of the trucks in actual order, had to be
inputted to the program. The desired clearance distance was added automatically before
computation. Running the computation was an extremely time consuming process.
Results of the simulations were plotted as a cumulative distribution function
(CDF) of uniformly distributed load for considered span lengths. The CDF’s for all
simulated data are presented in Appendix A, CDF’s of daily maximum combinations in
Appendix B, and CDF’s of weekly maximum combination in Appendix C.
Calculations were performed for one month, three, four, six, nine months or one
year data depending on localization. Even though it is a long period of time, it is small
compared to the actual life time of the structure. The uniformly distributed loads
67
corresponding to longer period of time, 75 years, were calculated by extrapolation of
simulated results. The extrapolated distributions are shown for maximum daily and
weekly combinations (Appendix B and Appendix C). Let N be the number of truck
combinations in time period T and assume that the traffic will remain the same. For T =
75 years, N will be larger 900 times for one month data, 300 times for 3 month, 75 times
for one year data etc. For example, for a site with 400,000 truck combinations monthly,
this will result in N = 360 million truck combinations. The probability corresponding to
N is 1/N. For 360 million, it is 1/360,000,000=2.8x10-9, which is 5.83 on the vertical
scale of CDF plot. Probability corresponding to extrapolated maximum daily truck is
3.65x10-5, and to maximum weekly truck is 2.56x10-4. The number of truck combinations
N, probabilities 1/N, and inverse normal distribution corresponding to 75 years periods
are shown in Table 5.2.
From the results of simulations, the statistical parameters of live load were
obtained. It was noticed that mean value of uniformly distributed load oscillates between
value 0.50 and 0.75 k/ft (Figure 5.12). The value 0.75 k/ft is close to those obtained in
two previous models. The mean daily and weekly maximum can be found in Figure 5.12
and Figure 5.13. For longer spans uniformly distributed load decreases and is closer to
mean value. This observation confirms that for a long loaded span, one heavily
overloaded truck does not have significant influence. This is because the load depends on
a mix of traffic. The bias factors (ratio of mean to nominal) were calculated for the
heaviest 75-year combination of vehicles. The 75-year uniformly distributed loads were
derived from extrapolated distributions. In Figure 5.16, it can be noticed that the bias
factor values for some sites do not exceed 1.25, which is similar as for short and medium
spans, as shown in the NCHRP Report 368 (1999). It is recommended to use HL-93 also
for those long spans (Figure 5.11). To keep bias value below 1.0, it would be necessary to
increase design value of uniformly distributed load to 0.85 k/ft (Figure 5.17). Brides in
localizations with high ADDT and high percentage of overloaded trucks, such as those in
New York, will require development of site specific models. For some sites with very
heavy traffic, the bias factor reaches value 2.0 (Figure 5.16). Therefore, for those bridges
the uniformly distributed load should be higher. It was found that to not exceed bias 1.25,
the uniformly distributed load should be 1.2 kip/ft (Figure 5.18). The heaviest truck
68
combinations were observed on I-475, Throggs Neck Bridge in New York. They have
been presented in Table 5.3.
The coefficient of variation is calculated from the slope of transformed CDF.
Figure 5.19 and Figure 5.20 present coefficient of variation of daily and weekly
maximum uniformly distributed load. Daily maximum uniformly distributed load has
more variation due to weekends. Lighter traffic during weekends can be observed in
lower tail of CDF’s. Estimated coefficients of variation were derived from weekly
maximum values, excluding sites from Yew York. Calculated statistical parameters for
uniformly distributed load are summarized in Table 5.1.
1.50
AASHTO HL93
1.00
0.68 kip/ft
0.50
0.00
0 1000 2000 3000 4000 5000
loaded length [ft]
Figure 5.11. HL-93 proposed for long span bridges
Table 5.1. Statistical parameter for proposed uniformly distributes live load
span length Bias CoV
600 – 1000 ft 1.25 0.10
> 1000 ft 1.20 0.08
69
Table 5.2. Summary of simulated data
State Site ID Route Number of truck Time 75-years
combinations period probability
FL 9916 US-29 247,449 1 year 5.39x10-8
FL 9919 I-95 222,368 3 months 1.50x10-8
FL 9927 SR-546 225,868 1 year 5.90x10-8
FL 9936 I-10 188,990 1 year 7.05x10-8
IN 9512 I-74 167,630 1 month 6.63x10-9
IN 9534 I-65 266,333 1 month 4.17x10-9
IN 9544 I-80/I-94 406,418 1 month 2.73x10-9
OR Woodburn I-5 552,390 4 months 8.04x10-9
OR Emigrant Hill I-84 213,019 4 months 2.09x10-8
OR Lowell OR 58 51,406 4 months 8.65x10-8
OR Bend US 97 59,225 3 months 7.50x10-8
NY 9121 I-81 300,500 6 months 2.22x10-8
NY 2680 8 45,030 9 months 2.23x10-7
NY I-495WB 43,200 1 month 2.57x10-8
NY I-495EB 52,618 1 month 2.11x10-8
sum 3,042,444
70
Table 5.3. Heaviest truck combinations for 600 ft on I-495 WB
number Total
W L W L W L W L W L W GVW
of axles length
[kip] [ft] [kip] [ft] [kip] [ft] [kip] [ft] [kip] [ft] [kip] [ft] [kip]
5 20.04 17.39 41.41 4.27 35.68 22.31 38.77 4.27 37.44 0.00 0.00 73.3 173.3
2 27.97 19.69 32.82 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 44.7 60.8
5 33.26 17.72 36.12 4.27 33.04 23.62 25.11 3.94 28.41 0.00 0.00 74.6 155.9
3 30.40 16.08 26.87 4.27 27.75 0.00 0.00 0.00 0.00 0.00 0.00 45.4 85.0
2 28.19 22.97 42.73 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 48.0 70.9
5 35.68 15.42 37.44 4.27 38.55 34.45 33.04 3.94 32.38 0.00 0.00 83.2 177.1
5 33.48 12.14 43.83 3.94 38.99 28.87 43.61 3.94 39.87 0.00 0.00 74.0 199.8
5 38.11 13.12 25.33 4.27 20.48 31.50 19.60 3.94 23.57 0.00 0.00 77.9 127.1
2 33.26 15.75 39.87 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 40.8 73.1
561.9 1123.1 2.00 k/ft
2 11.23 18.70 21.15 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 43.7 32.4
5 20.04 17.39 41.41 4.27 35.68 22.31 38.77 4.27 37.44 0.00 0.00 73.3 173.3
2 27.97 19.69 32.82 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 44.7 60.8
5 33.26 17.72 36.12 4.27 33.04 23.62 25.11 3.94 28.41 0.00 0.00 74.6 155.9
3 30.40 16.08 26.87 4.27 27.75 0.00 0.00 0.00 0.00 0.00 0.00 45.4 85.0
2 28.19 22.97 42.73 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 48.0 70.9
5 35.68 15.42 37.44 4.27 38.55 34.45 33.04 3.94 32.38 0.00 0.00 83.2 177.1
5 33.48 12.14 43.83 3.94 38.99 28.87 43.61 3.94 39.87 0.00 0.00 74.0 199.8
5 38.11 13.12 25.33 4.27 20.48 31.50 19.60 3.94 23.57 0.00 0.00 77.9 127.1
564.9 1082.4 1.92 k/ft
5 38.99 14.11 35.68 4.27 31.94 17.39 28.85 3.94 31.28 0.00 0.00 64.8 166.7
5 37.00 12.14 40.09 4.27 33.04 32.48 30.84 3.94 37.22 0.00 0.00 77.9 178.2
5 37.67 10.50 22.47 4.27 24.01 26.57 15.42 10.17 18.28 0.00 0.00 76.6 117.8
2 29.07 16.08 38.11 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 41.1 67.2
3 43.61 18.70 21.37 4.59 27.53 0.00 0.00 0.00 0.00 0.00 0.00 48.3 92.5
2 34.14 21.33 26.21 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 46.4 60.4
5 40.75 13.45 30.62 4.27 25.33 30.51 22.03 3.94 22.47 0.00 0.00 77.3 141.2
5 35.46 10.50 22.03 4.27 20.70 22.64 20.04 4.27 20.48 0.00 0.00 66.7 118.7
3 36.56 19.69 37.00 4.27 31.06 0.00 0.00 0.00 0.00 0.00 0.00 49.0 104.6
548.1 1047.4 1.91 k/ft
71
0.80 OR 58 Lowell
I-84 Emigrant Hill
0.70 US 97 Bend
I-5 Woodburn
0.60
Florida 9936
Florida 9916
0.50
UDL [k/ft]
Florida 9927
0.40 Florida 9919
Indiana 9511
0.30 Indiana 9512
Indiana 9534
0.20
Idiana 9544
NY I-495 EB
0.10
NY I-495 WB
0.00 NY 2680
0 1000 2000 3000 4000 5000 6000 NY 9121
span length [ft]
Figure 5.12. Mean value of uniformly distributed load
1.60 OR 58 Lowell
I-84 Emigrant Hill

1.40
US 97 Bend
I-5 Woodburn
1.20
Florida 9936
1.00 Florida 9916

UDL [k/ft]
Florida 9927
0.80
Florida 9919
Indiana 9512
0.60
Indiana 9534
0.40 Idiana 9544
NY I-495 EB
0.20
NY I-495 WB
0.00 NY 2680
0 1000 2000 3000 4000 5000 6000 NY 9121
span length [ft]
Figure 5.13. Daily maximum mean value of uniformly distributed load
72
1.60 OR 58 Lowell
I-84 Emigrant Hill

1.40
US 97 Bend
1.20 I-5 Woodburn
Florida 9916
1.00
UDL [k/ft]
Florida 9936
0.80 Florida 9927
Florida 9919
0.60
NY 2680
NY 9121
0.40
0.20
0.00
0 1000 2000 3000 4000 5000 6000
span length [ft]
Figure 5.14. Weekly maximum mean value of uniformly distributed load
2.60
2.40
2.20
2.00
1.80 OR US-97
OR I-5
1.60
OR 58
1.40
OR I-84
Bias
1.20 FL 9919
1.00 FL 9927
0.80 FL 9936
IN 9534
0.60
IN 9544
0.40
IN 9512
0.20
0.00
0 1000 2000 3000 4000 5000
span length
Figure 5.15. Bias (mean max 75 year to nominal value of UDL)
73
2.60
2.40
2.20
2.00
1.80
1.60
1.40 FL 9916
Bias
NY 9121
1.20
NY 2680
1.00
NY I-495 EB
0.80 NY I-495 WB
0.60
0.40
0.20
0.00
0 1000 2000 3000 4000 5000
span length
2.60
2.40
2.20
2.00
1.80 OR US-97
OR I-5
1.60
OR 58
1.40
OR I-84
Bias
1.20 FL 9919
1.00 FL 9927
0.80 FL 9936
IN 9534
0.60
IN 9544
0.40
IN 9512
0.20
0.00
0 1000 2000 3000 4000 5000
span length

assumed designed UDL of 0.85 k/ft
74
2.60
2.40
2.20
2.00
1.80
1.60
1.40 FL 9916
Bias
NY 9121
1.20
NY 2680
1.00
NY I-495 EB
0.80 NY I-495 WB
0.60
0.40
0.20
0.00
0 1000 2000 3000 4000 5000
span length
Figure 5.18. Bias for heavily loaded localizations, assumed designed UDL of 1.25 k/ft
0.30 OR 58 Lowell
OR I-84 Emigrant Hill
0.25 OR US 97 Bend
OR I-5 Woodburn
0.20 Florida 9916
Florida 9936
UDL [k/ft]
Florida 9927
0.15
Florida 9919
Indiana 9512
0.10
Indiana 9534
Indiana 9544
0.05 NY I-495 EB
NY I-495 WB
0.00 NY 2680
0 1000 2000 3000 4000 5000 6000 NY 9121
span length [ft]
Figure 5.19. Coefficient of variation of daily maximum uniformly distributed load
75
0.30 OR 58 Lowell
OR I-84 Emigrant Hill

0.25
OR US 97 Bend
UDL [k/ft] 0.20 OR I-5 Woodburn
0.15 Florida 9916
Florida 9936
0.10
Florida 9927
0.05 Florida 9919
NY 2680
0.00
NY 9121
0 1000 2000 3000 4000 5000 6000
span length [ft]
Figure 5.20. Coefficient of variation of weekly maximum uniformly distributed load
0.25 OR 58 Lowell
OR US 97 Bend
0.20
OR I-5 Woodburn
0.15 Florida 9916

UDL [k/ft]
Florida 9936
0.10 Florida 9927
Florida 9919
0.05
0.00
0 1000 2000 3000 4000 5000 6000
span length [ft]
Figure 5.21. Proposed coefficient of variation of uniformly distributed load
76
CHAPTER 6
MULTIPLE PRESENCE
6.1. INTRODUCTION
Determination of live loading for multiple traffic lanes is very important for
appropriate bridge design. Each traffic lane can be loaded with different live load. The
more traffic lanes the more difference in distribution of loading. Trucks tend to use right
lanes loading them heavily, while passenger cars use faster left lanes. Structural
components are strongly influenced by the location of trucks on the bridge. Those
carrying the right lane of traffic are usually subjected to more load cycles and fatigue
than the components closer the left lane. However, it has to be remembered that traffic
can be deviated and truck can be directed to left lanes, for example during maintenance
works on bridge. Presence of truck on multiple traffic lanes at the same time is critical
from the bridge design point of view.
In this chapter, a short review of current studies on presence of multiple trucks,
study of video recording of traffic jam situations, and discussion on different approaches
to multilane reduction factors are presented. Multiple presence factors applied in
international codes are discussed in CHAPTER 2.
77
6.2. STUDIES ON PRESENCE OF MULTIPLE TRUCKS
In 1993 Andrzej S. Nowak, Hani Nassif, and Leo DeFrain performed a study on
the occupation of road lanes and presence of multiple truck. For the study, the database of
over 600 000 trucks on two lanes in each direction was collected using weigh-in-motion
equipment. It was found that 70-90% of trucks use the right lane (Table 6.1), and 65-70
percent of trucks are 5-axle trucks. The researchers had also found out that less than 2%
of trucks appear simultaneously with another truck in the lane, 4-8% side by side in
tandem or behind, with distance between front axles less than 50 ft (Table 6.1).
Table 6.1. Presence of multiple trucks and their location on the road lanes
Eastbound Westbound
Interstate
highway side by side side by side
right left in lane right left in lane
tandem behind tandem behind
I-94 1582 151 14 28 48 2073 349 14 86 93
91,3% 8,7% 0,8% 1,6% 2,8% 85,6% 14,4% 0,6% 3,6% 3,8%
U.S.-23 1685 371 24 40 76 1247 601 2 38 34
82,0% 18,0% 1,2% 1,9% 3,7% 67,5% 32,5% 0,1% 2,1% 1,8%
The results were later confirmed by the investigation by Gindy and Nassif (2006).
Data used in this study was collected over an 11-year period between 1993 and 2003 by
the New Jersey Department of Transportation. The study gives detailed analysis of the
relation between multiple presences of trucks and four parameters: truck volume, area
type, road type, and bridge span length. An increase in truck volume results in an increase
of all multiple presence cases and a decrease in the frequency of single loading events
(Figure 6.2). The area (urban or rural) and vicinity of industry affect the frequency of
multiple truck presence. Heavier volume sites tend to be located in urban areas, and as a
consequence more cases of multiple trucks can be observed. Increasing bridge spans also
gives more opportunities for trucks to occur simultaneously (Figure 6.3). The frequency
of staggered events increases faster for shorter spans and at a steadier pace for longer
spans. Span length has almost no influence on the frequency of side-by-side trucks.
78
Assumptions of traffic patterns made for multiple presence analysis following
(see Figure 6.1):
- single – only one truck is present on the bridge
- following – two trucks on the same lane with a varying clearance distance
- side-by-side – two trucks in adjacent lanes with an overlap at least one-half the body
length of the leading truck
- staggered – two trucks in adjacent lanes with an overlap at less than one-half the body
length of the leading truck
Figure 6.1. Traffic loading pattern used for multiple truck presence statistics.
Comparing results of the study made by Andrzej S. Nowak, Hani Nassif, and Leo
DeFrain in 1993 and the study made by Mayrai Gindy and Hani H. Nassif in 2006, some
divergences can be observed. Some of them can be caused by the differences in the
definition of “side-by-side” and “staggered” cases. Trucks which are considered
“staggered” in one case can be considered as “side by side” in the other. However,
joining those two cases and making a sum of those two values, we obtain relatively close
results, 7.3 % for the study from 1993 and 6% for study from 2006. Regarding the
occurrence of following trucks, values obtained in 1993 vary between 0.1 and 1.2%, and
values obtained in 2006 vary between 1 and 8%. Those values cannot be compared
79
because of differences in the assumptions and definition of “following” trucks.
Furthermore, the bridges taken into consideration in the study from 2006 were longer,
which increases the probability of this occurring. The probability that the trucks will
occur as following trucks increases with the span length.
In the report NCHRP 12-76 (2008), it is stated that multi-presence probabilities
for permit trucks are different from those for normal traffic. The likelihood of permit
trucks exceeding the authorized weight as well as the likelihood of the presence of
multiple permit trucks is reduced.
Figure 6.2. Variation of multiple truck presence statistics with respect to truck volume.
Gindy and Nassif (2006).
80
Figure 6.3. Variation of multiple truck presence statistics with respect to bridge span
length. Gindy and Nassif (2006).
6.3. MULTIPLE PRESENCE OF TRUCKS BASED ON THE VIDEO FILES OF

TRAFFIC
FHWA has provided a DVD including video files monitoring seven traffic
situations in different sites, at different times and days of the week. The localization of
the sites was not specified. Total time of all sixteen video recordings is 2 hours 6 minutes
and 30 seconds. The list of video recordings is following:
- (1) 05.30.2008, Friday, 12.17 pm (33min44sec)
- (2) 05.29.2008, Thursday, 6.08pm (4min43sec)
81
- (3) 05.22.2008, Thursday, 11.22am (0min22sec)
- (4) 05.22.2008, Thursday, 11.20am (0min22sec)
- (5) 05.14.2008, Wednesday, 12.52pm (4min51sec)
- (9) 05.06.2008, Tuesday, 7.59am (3min49sec)
- (10) 05.01.2008, Thursday, 12.20pm (1min04sec)
- (11) 04.04.2008, Friday, 4.42pm (9min28sec)
- (12) 04.04.2008, Friday, 4.43pm (8min54sec)
- (13) 04.04.2008, Friday, 4.43pm (0min36sec)
- (14) 04.04.2008, Friday, 4.43pm (19min26sec)
- (15) 04.04.2008, Friday, 4.43pm (22min53sec)
- (16) 04.04.2008, Friday, 4.43pm (8min06sec)
The recordings show dense traffic jam situations, some of them being the result of
traffic accidents. They allow for making some observations and conclusions regarding
traffic patterns and the presence of multiple trucks moving at a crawling speed. This is
the critical case from the point of view of live loading on bridges. Despite the fact the
recordings are taken on highways, the recorded situations can be related to bridges as
well. One of the most important observations is that even in very dense traffic it is very
common to observe cars or pick-up among heavy vehicles (Figure 6.4).
82
Figure 6.4. Video 10, time: 00:00:58
The video file number 1 contains the longest and the most interesting material. It
has a registered traffic accident on a highway having four lanes in one direction. The
accident takes place on the second lane (counting from the external side). The second and
the third lanes remain completely stopped by crushed cars and emergency vehicles for
approximately half an hour. Passing cars are using the first and the fourth lanes. For a
short period of time three lanes are blocked, and the passing cars can use only the fourth
lane, that intensifies jam-packed traffic. For the majority of time we can observe that the
moving lanes contain a mixture of trucks and cars. However, we can also observe some
situations with multiple-presence of trucks occupying three or four lanes at the same time
(Figure 6.5 and Figure 6.6). We can also observe a situation when one lane is almost
exclusive occupied by trucks (Figure 6.6). Those cases should also be taken into
consideration during the evaluation of design live load. There is no information about
trucks’ weight. However, it can be assumed that only a limited number of the trucks are
correlated, and while some of them are fully loaded some percentage of them can be
empty.
83

84
Some of the video files show traffic jam situations at roads junction that cannot be
representative. The number of cars among trucks is increased because some of them are
entering or exiting the highway (Figure 6.7 and Figure 6.8). This observation calls our
attention to the importance of the appropriate selection of sites.
85
Video recordings from 11 to 16 show traffic situation caused by the same

accident. Three lanes are blocked and the moving vehicles are using only one lane.
However, warning signs posted adequately ahead result in the vehicles forming one lane
and do not cause multilane traffic jam situation in the vicinity of the place of the accident.
An ordered one lane of traffic is a mixture of trucks and cars. This situation does not
allow for the observation of multiple truck presence situations.
6.4. APPROACHES TO MULTILANE REDUCTION FACTORS
There are many approaches to multilane reduction factors. International codes

vary significantly in this matter (see CHAPTER 2), and all of them are much more
simplified than actual situation.
Most of the design codes decrease the value of uniformly distributed load as the
number of traffic lanes increases. The load value is the same on all traffic lanes, Figure
6.9.
86
Figure 6.9. Multilane load in design codes AASHTO LRFD Code (2007),
OHBDC (1991), CAN/CSA-S6-00 [2000], and ASCE (1981).
According to the Eurocode one of the lanes is loaded more than the others, Figure
6.10.
Figure 6.10. Multilane load in Eurocode 1.
The observations indicate that the actual traffic is distributed differently for each
lane of traffic, Figure 6.11.
Figure 6.11. Multilane load in actual observation.
87
To reflect actual traffic situation would be very difficult and time consuming for
designers. Therefore, the approach applied in the Eurocode seems the most practical one.
Each of the lanes should be considered as the most loaded one, while the other lanes carry
equal loading.
6.5. CONCLUSIONS
Presence of multiple trucks depend on factors such as: truck volume (light,
average, heavy) area type (urban/rural), road type (major/minor), bridge span length, law
enforcement, and traffic flow control, which can cause heavy truck queues. Therefore,
multilane reduction factors could be very site specific.
Based on video recordings of traffic it was confirmed that for the majority of time
we can observe that the moving lanes contain a mixture of trucks and cars. However,
situations when one lane is almost exclusively occupied by trucks or trucks occupy three
or four lanes at the same time are also possible. Multiple reduction factors for design live
load should account for those the most critical loading cases.
In the available WIM database, the vehicles of 1-3 categories have not been
registered. Therefore, it does not allow for simulations and derivation of multilane factors
for all traffic lanes. Simulation of the traffic on the most loaded lane was possible with
assumption, that it is occupied exclusively by trucks. In the traffic jam situations, the
passenger vehicles are assumed to move to faster lanes.
It was concluded that the multilane reduction factors have to be an object of
additional extensive studies. They have to account for intensive traffic jam situations, as
those registered in the video recordings. As well, different distribution of loading on
multiple traffic lanes has to be considered.
Since no new multilane reduction factors were proposed, those from the current
AASHTO Code are used in this dissertation.
88
CHAPTER 7
DYNAMIC FACTOR
7.1. INTRODUCTION
The scope of this chapter is studying the origin and adequacy of the application of the
dynamic load factor in bridge design. There is considerable variation in the treatment of
dynamic load effects by bridge design codes in different countries (see CHAPTER 2). The
most common approach is to apply dynamic response as a fraction or multiple of the
response that would be obtained if the same forces or loads were applied statically. The
objective of this simple approach is to not increase complexity for the designer. This is the
approach specified in the current AASHTO Code. The live load model itself does not
account for dynamics, but the dynamic amplification is added additionally as a percentage to
static effects.
In this chapter there is a short review of the research studies. As well, a developed
exemplary vehicle-bridge interaction model and the derivation of dynamic factor is
presented. In the modeling a finite element software ABAQUS was used. Three-axle
AASHTO truck HS-20 travelling over a 120 ft steel girder bridge is modeled. The truck is
assumed to travel with the velocity of 40 miles per hour and with the crawling speed. The
final result is comparison of the static the dynamic deflections, and derivation of dynamic
factor for this specific case.
89
7.2. STUDIES ON PARAMETERS AFFECTING DYNAMIC BRIDGE RESPONSE
The dynamic response of a bridge to a crossing vehicle is a complex problem affected

by the dynamic characteristics of the bridge, the vehicle, and by the bridge surface
conditions. Many of the parameters interact with one another, further complicating the issue.
Consequently, many research studies have reported seemingly conflicting conclusions. Based
on the review of past research, the effects of various parameters on the dynamic response of
bridges to vehicular loading are discussed in this paragraph.
Bridge Fundamental Frequency
The fundamental frequency of vibration for a bridge due to vertical loading has a
significant effect on the dynamic response. If the frequencies of the bridge and vehicle
converge, the dynamic response induced may be large. A majority of the fundamental
frequencies for typical bridges are in the range of 2 to 5 Hz (Figure 7.1), which corresponds
to the body bounce response frequency range of a truck.
Figure 7.1. Distribution of fundamental bridge frequencies

(Cantieni 1984)
90
Field measurements and values obtained from analytical modeling show relation
between frequencies and bridge span (Figure 7.2 and Figure 7.3).
Figure 7.2. Fundamental frequency versus span length

(Cantieni 1984)
Figure 7.3. Fundamental frequency versus span length

(Paultre 1992)
91
Bridge Damping
From basic dynamic principles, higher levels of damping reduce the dynamic
response in bridges and low levels of damping in a bridge are expected to result in high
dynamic amplification. However, damping affects impact differently at different locations
within the bridge as a result of varying modal contributions (Huang, Wang, and Shahawy,
1992). Damping values for bridges obtained from field testing can also vary considerably
based on the method of testing, level of loading, and different methods used for evaluating
damping. Reported values of damping for different types of bridges are as follows:
- concrete bridges - 2 to 10 % (Tilly, 1978)
- steel bridges - 2 to 6 %( Tilly, 1978), 0.4 to 1.3 % (Billing, 1984)
- composite steel-concrete bridges - 5 to 10 % (Tilly, 1978)
- prestressed concrete bridges - 1 to 2.2 % (Billing, 1984)
- timber bridges - 3 to 4 % (Ritter, 1995)
Roadway Roughness and Approach Condition
The bridge approaches and roughness of the roadway surface have a significant
influence on the magnitude of the dynamic response. Not only do the impact forces increase
for increased roughness, but also vehicle speed affects the influence of roughness. The faster
vehicle speed has greater impact on rougher surfaces than on better maintained ones. The
results of the study by Wang, Shahawy, and Huang (1993) can be seen in Figure 7.4.
92
Figure 7.4. Impact factor versus vehicle speed and road surface condition
(Wang, Shahawy, and Huang 1993)
Experimental tests have also shown that the most severe wheel impact forces are
likely to occur adjacent to the bridge approaches, i.e., shortly after a vehicle enters the bridge
(Tilly 1978). In many experimental investigations, wooden planks were placed in the path of
the test vehicle. It was supposed to represent surface irregularities such as dropped objects or
packed snow on the roadway. Dynamic response was higher with the planks and the planks
were exciting wheel hop in the test vehicles, although excitation of the higher vibration mode
associated with wheel hop is also speed dependent.
Vehicle Speed
For most heavy trucks, natural frequencies of the vehicle typically occur in two
frequency ranges: between approximately 2 and 5 Hz for the "body bounce" response and
between approximately 10 and 15 Hz for the "wheel hop" response. However, depending on
vehicle speed, roadway surface irregularities may be effective in exciting both modes of
response (Cantieni 1983).
93
Weight of Vehicles
Many studies have shown that as the weight of the crossing vehicle increases, the
magnitude of the dynamic response expressed as a percentage of the static load decreases.
The explanation of this fact is that when the dynamic forces increase with increasing
vehicle weight, the static load increases more rapidly with increasing weight. Thus, the
impact ratio of dynamic force to live load decreases with increasing vehicle weight and
impact factors obtained from the measured dynamic response of lightly loaded vehicles
will be relatively large.
Number of Vehicles
The dynamic load factors associated with multiple vehicles are lower than those for
single vehicles. This is most likely because the total static load is larger (similar to having a
heavier vehicle) compared to the associated dynamic load, and the dynamic responses from
the two individual vehicles are likely to be at least somewhat out of phase with each other.
Vehicle suspension
The vehicle frequency ranges are a function of the suspension systems. The body
bounce frequencies in vehicles with air suspensions are lower than those for steel leaf-spring
suspensions, with measured frequencies in the 1.5 to 2 Hz range. Worn dampers in the
suspension systems also dramatically increased the dynamic wheel forces.
In Figure 7.5 the influence of suspension is presented. On one deflection trace the
vehicle had its normal suspension characteristics, in the other the springs were blocked so
that the truck rode directly on the axles. The increase in response is evident for the unsprung
condition.
94
Figure 7.5. Effects of vehicle suspension on the measured bridge response
(Biggs and Suer 1955)
Bridge Span
There are conflicting opinions regarding dynamic response and bridge span. While
some investigations have shown a general trend of decreasing impact in conjunction with
increasing span (Figure 7.6, Fleming and Romualdi 1961), other investigations have
concluded that considerable scatter exists in the results and there is poor correlation of impact
and span (Figure 7.7, Cantieni 1983). Some researchers have concluded that impact is not a
function of bridge span (Coussy et al. 1989). However, it should be concluded that as span
length influences bridge fundamental frequency, it also indirectly influences bridge
dynamics.
95
Figure 7.6. Impact versus span length (Fleming and Romualdi 1961)
Figure 7.7. Impact versus span length (Cantieni 1984)
96
Bridge Type and Geometry
Most of the studies on dynamic response were performed on simple-span multi-girder

bridges. The investigations concluded that the total number of girders has little influence on
the maximum impact factors for each girder and that the impact at the interior supports was
larger than at external locations. The dynamic response of simple-span bridges is higher than
that for similar continuous bridges.
Cable-stayed and suspension bridges are more complicated to assess than for
beam/girder bridges, mostly due to the influence of cables’ dynamics. Dynamic response
quantities are sensitive to damping, which is difficult to determinate in these types of bridges,
and may be different for different vibration modes. In analytical investigations it was found
that, with a good road surface, impact factors were generally less than 0.20. However, for
rough surfaces, impact forces increased dramatically. (Khalifa 1992)
Dynamic response of continuous and cantilever thin-walled box girder bridges under
multi-vehicle loading was analytically investigated by Wang, Huang, and Shahawy (1996)
and Huang, Wang, and Shahawy (1995a). It was found that the vibration characteristics of
the continuous and cantilever box girder bridges are quite different. For cantilever bridges,
the most important factor affecting impact is the vehicle speed, and they are much more
susceptible to vibration than continuous bridges. This is due to the abrupt change in loading
due to span discontinuities, when no support exists between cantilevers. For continuous
bridges, both vehicle speed and surface roughness are significant. End diaphragms were
found to provide lateral support and significantly reduce the response of the box girder
bridges. The beneficial effect of a midspan diaphragm is relatively small.
Dynamic behavior depends on curvature of the bridge. It was found that the dynamic
response in horizontally curved bridges is influenced by centrifugal accelerations, thus,
vehicle speed is particularly important. Impact forces are higher in the outer elements of the
curved bridges. Impact forces are insensitive to curvature for radii greater than 4 000 ft (1219
97
m) and markedly influenced by curvature for radii less than 800 ft (244 m). Research done by
Galdos et al. (1993) and Schilling et al. (1992) and Huang, Wang, and Shahawy (1995b).
7.1. BRIDGE-VEHICLE INTERACTION MODEL AND DERIVATION OF

DYNAMIC FACTOR
7.1.1. Bridge and Vehicle Model
To study vehicle-bridge interaction, with the use of ABAQUS 6.6.1 software a finite
element model of bridge and vehicle has been developed.
The bridge chosen for the analysis is a 120 ft steel girder bridge. It is modeled with 3D shell
elements:
Steel girders:
- five steel W40x264 (profile properties: area A = 77.60 in2, depth of the section
d=40 in, web thickness: tw = 0.96 in, flange width bf = 11.93 in, flange thickness
tf =: 1.73 in, moment of inertia Ix-x=19400 in4)
- spaced 64 in
- fy = 60 ksi
- steel diaphragms every 30 ft
- 2360 elements, 2715 nodes
Concrete slab:
- thickness: 7.5 in
- width:312 in
- fc’=:8 ksi
- 4464 elements, 4640 nodes
Supports:
- left support - pinned
- right support - roller
Connection between girders and slab: tied (compatibility of all degrees of freedom)
Bridge model meshed with S4R elements is shown in Figure 7.8.
98
Figure 7.8. Meshed model of the bridge
A vehicle model was developed based on three-axle AASHTO HS-20 truck, which is
the design vehicle in the AASHTO Specifications. Two cases are considered, the truck HS-
20 is assumed to travel with the velocity of 40 miles per hour and with the crawling speed.
Figure 7.9, shows FEM model of moving masses. To simplify the analysis each wheel is
modeled with one DOF in the vertical direction. Detailed model would include a seven
degree of freedom system. Tractor and semitrailer would have individually assigned two
DOFs corresponding to: vertical displacement (yi) and rotation about the transverse axis (θi).
Moreover each wheel-axle set would be provided with one DOF in the vertical direction (yi).
Five sprung masses would be: the tractor, semitrailer, steer wheel-axle set, tractor wheel-axle
set, and trailer wheel axle set. A more detailed model could include rotations about
longitudinal axis (Φi).In this case the total independent DOFs would be eleven.
Figure 7.9. Truck model in FEM
99
7.1.1. Bridge-Vehicle Interaction Model
There are two main sources of vibrations induced by the vehicle into the bridge. One
of them is the settling of the approach slab of the bridge. The bump at the bridge entrance
causes so called “wheel hop” that could be approximated by impulse loading. A vehicle
vibrates with a frequency between 10 and 15 Hz. Transmitted force is significant, however it
fades relatively fast converging into “body bounce” that occurs at lower frequencies. The
other source of vibrations is an undulating roadway surface. The condition of driving a
vehicle over an undulating roadway surface can be approximately idealized as an SDOF
system under harmonic loading provided the roadway varies as a sine wave. These kinds of
vibrations are called “body bounce” and usually occur at frequencies between 2 and 5 Hz.
The loading is characterized by the roughness amplitude, roughness wavelength, and vehicle
speed. If the shock-absorbing elements of the vehicle suspension system are worn, then the
vehicle's damping is fairly low and the response is quite large. Under such conditions, nearly
resonant response can develop.
Since the truck model is simplified and the scope of the project requires comparison
of the deflections in the middle of the span, the interaction model includes only “body
bounce”, which depends on the relation between suspension system and the truck. Tire
stiffness and “wheel hop” response are neglected.
The vertical interaction force acting on a bridge consists of the static interaction force
Fw and the variation of the interaction force ∆Fw.
Fb = Fw + ∆Fw.
Fw = M g
dy(t ) dy1
∆F = ci ⋅ ( − ) + k i ⋅ ( y (t ) − y1 (t ))
dt dt
The interaction force is approximated by a sinusoidal shape, as shown in the

Figure 7.10. The force Fw transmitted by the rear wheels is 16 kips and its variation ∆Fw is
equal to ±10%. The front wheels carrying 4 kips each are also assumed to produce force
variation ±10%. The effects of damping are small, and they will be neglected in the analysis.
100
20
15
10
0
0 1 2
t
Figure 7.10. Vehicle-bridge interacting force.
Figure 7.11 is a plot from ABAQUS that shows forces being applied to the bridge versus
time. It can be noticed when the following truck axes are entering and leaving the bridge.
Their amplitudes are interfering and adding to each other. Those forces could also be
canceling each other. Such a situation of adding amplitudes was simulated on purpose, to
obtain the maximum dynamic deflections in the middle span.
Figure 7.11. Force due to moving truck versus time. Plot from ABAQUS.
101
7.1.1. Results of analysis
Mode shapes obtained in the analysis are shown in Figure 7.12 (bending modes) and
Figure 7.13 (torsion modes). The natural frequencies are following:
- first bending mode 2.12 Hz
- second bending mode: 8.19 Hz
- third bending mode: 17.10 Hz
- first torsion mode: 3.17 Hz
- second torsion modes: 9.09 Hz
Figure 7.12. Bending modes
Figure 7.13. Torsion modes.
102
The deflection due to dead load, weight of the concrete slab and the steel profiles, is 2.74 in.
Maximal deflection due to moving truck is 0.69 in (Figure 7.14). It corresponds to the truck
being located almost, but not in the span center.
Figure 7.14. Maximal deflection due to moving truck
Figure 7.15 and Figure 7.16 show deflections in the middle of the span due to
moving and stationary trucks. Maximum deflection due to moving truck (0.69 in) versus
maximum deflection due to stationary truck (0.65 in) shows 6% difference.
103
Figure 7.15. Deflection due to a truck moving 40mil/hr versus time.
Figure 7.16. Deflection due to a truck moving at crawling speed versus time.
104
7.2. CONCLUSIONS
The study of the topic and the developed FEM model allowed for estimating the
magnitude of the dynamic load factor and to draw some conclusions. It was concluded that
the current dynamic load factor of 0.33 is too high for bridges with longer spans. It may be
applicable in short bridges, when vibration due to “wheel hop” on the approach slab is
significant. However, for longer bridges where the influence of the approach slab decreases
and vibrations of many vehicles interfere with each other, the dynamic load factor could be
smaller. Even for the exemplary case of medium span bridge presented in this thesis the
dynamic load factor is only 6%. The FEM model was built using ABAQUS 6.6.1 software.
Finite element problem modeling, including moving load and defined interaction between
surfaces, are non linear and very time demanding. To be adequate and draw more
conclusions further studies should be performed, including a wide range of bridge types,
spans and roadway conditions. Also, a more elaborate truck model should be performed,
accounting for both “body bounce” as well as “wheel hop”. Moreover, analytical studies
should be confirmed with field tests on the representative structures.
In this dissertation, a traffic jam situation was assumed to develop the live load
model. Therefore there is no allowance for dynamic. However, to not introduce confusion
among designers it is recommended do keep dynamic factor as it is for short and medium
span bridges, which results in very small value for long span bridges.
105
CHAPTER 8
RELIABILTY ANALYSIS OF SUSPENSION BRIDGE
8.1. RELIABILITY ANALYSIS PROCEDURE
Reliability analysis was performed to verify the live load model for long span
bridges. The reliability procedure includes the following steps:
1. Selection of a representative bridge and its component

A representative suspension bridge and the structural element that is the most
influenced by live load were selected.
2. Limit state function
The limit state function was defined as the exceeding of the ultimate bending
moment capacity by the cross-section and Strength I combination.
3. Nominal resistance model
In order to find nominal resistance, software for the tower cross section was
developed.
4. Reliability resistance models
The material, fabrication and professional factors were established. Evaluation
of material, fabrication and professional factors was based on the database
available in the literature of and described by Nowak et al. (2008) and
Ellingwood et al. (1980). Based on the developed software for nominal
106
resistance and the Monte Carlo simulation method, the statistical parameters
for resistance were obtained.
5. Load model
The three-dimensional FEM model of a bridge was created using Robot
Millennium software. The model was based on the actual Cooper Bridge
design made by PB World. Cross sectional axial force and bending moment
along the tower height due to dead load and live load were derived. The
statistical models for load components are defined.
6. Reliability Indices
Reliability indices were calculated in order to assess how they are influenced
by the increase in the values of live load. Reliability analysis was performed
for the several forces possible and moment in the bridge tower due to live load
cases.
8.2. SELECTION OF REPRESENTATIVE STRUCTURE, ELEMENT AND

LIMIT STATE FUNCTION
The Cooper River Bridge in South Carolina was chosen to be a representative

long span bridge for this analysis. It is a suspension bridge designed by PB Word in 2001.
For long span bridges dead load is the main loading. Dead load is also critical for
most of the structural components, such as the deck and cables. For the scope of this
dissertation, the bridge component that is the most influenced by live loading had to be
selected. The bridge tower was chosen to be such an element. Live load from all spans is
transferred through cables to the bridge tower. When not all spans are loaded evenly, the
bending moment due to live load is much higher than the bending moment due to dead
load. The bending moment due to dead load of opposite spans almost negate each other.
The limit state function was defined as the exceeding of the ultimate bending
moment capacity by the cross-section. The tower cross-section just above the deck level
was selected.
107
8.3. NOMINAL RESISTANCE
The nominal resistance can be represented by interaction diagram of force and

moment for the eccentric loaded bridge tower. To plot interaction diagram, representing
all possible cases of force and bending moment combinations, a MathCad software was
used. The mathematical description of the behavior for the exemplary bridge tower was
developed based on procedure for columns described by Lutomirski (2009). The
assumptions are based on linear strains in a distribution over the cross section and the
mechanical behavior of reinforcement. The procedure takes into account the geometry of
the tower, all layers of the reinforcement in the cross-section and the characteristics of the
materials (steel and concrete).
The first step in the procedure is the calculation of the initial compression block.
The initial compression block represents the compression over the full cross-section and
the lowest layer of the reinforcement yields due to compression. It means that strains in
concrete at the top of cross section reach value of ε m = 0.003 , while the strains in the
lowest reinforcement bar represent yielding strains due to compression in steel
ε s = ε y = 0.00207 (Figure 8.2). The position of the neutral axis can be calculated based
on the linear strain distribution and strain compatibility assumptions. As a result of the
first step the size of the compression block is much bigger than the size of the cross-
section. The size of the initial compression block can be calculated as follows:
εm ⋅ dI
a I = β1 (8.1)
εm −ε y
where:
dI is the distance of the lowest layer of the reinforcement to the top of the cross-section.
The position of the neutral axis in the initial step can be calculated from:
aI
cI = (8.2)
β1
108
β
Figure 8.1 Distribution of strains for the pure axial loading
The characteristic point of end of compression control zone and beginning of

tension control zone call balance failure point can be derived. It happens when the strains
in the bottom layer of reinforcement reaches the yielding strains for the steel εs=εy
(Figure 8.2). The size of the compression block in balance failure is represented:
εm ⋅dI
a B = β1 (8.3)
εm + εy
ε
Figure 8.2 Distribution of strains distribution for the balance failure point B, the end of
the compression control zone
109
The entire force and moment interaction diagram is calculated using the
decreasing size of the compression block of concrete, from the initial aI up to the point
when the compression block does not exist ( a = 0 ). For each reinforcement layer and for
each size of the compression block the strains in reinforcement are computed from the
equation:
 di 
ε s (i, a) = ε m ⋅ 1 −  (8.4)
 c(a ) 
where:
i number of ith reinforcement bar in cross-section
a size of compression block of concrete
εm extreme compressive strain in concrete equal to 0.003
di distance of ith reinforcement layer for the top of cross-section
c (a ) position of neutral axis due to changing size of compression block
In the procedure, four characteristic cases there can be distinguished. The first
case is when the position of neutral axis is outside of the cross section (Figure 8.1). The
second case is when the end of compression block of concrete is in the bottom flange of
the cross section (Figure 8.3). Two next cases correspond to the end of compression
block of concrete localized in the webs and in the top flange (Figure 8.4 and Figure 8.5).
ε
β
Figure 8.3 End of compression block of concrete in the bottom flange
110
ε
β
Figure 8.4 End of compression block of concrete in the web.
Figure 8.5 End of compression block of concrete in the top flange.
Figure 8.6 shows material behavior of reinforcing steel, and it is described by the
following equation:
 − fy for ε s < −ε y
E ⋅ ε for − ε y ≤ ε s ≤ ε y

fs =  s s (8.5)
 fy for ε y < ε s ≤ ε m
 0 for εs > εm
where:
fy
Es = (8.6)
εy
111
Figure 8.6 Stress - Strain Relationship for Reinforcing Steel
Having calculated strains in every reinforcement bar, it is possible to evaluate

forces for each reinforcement layer.
 As ⋅ f y for ε s (i, a ) ≥ ε y

P(i, a ) =  As ⋅ E s ⋅ ε s (i, a ) for − ε y ≤ ε s (i, a ) ≤ ε y (8.7)
 −A ⋅ f for ε s (i, a ) < −ε y
 s y
where:
i number of ith reinforcement bar in cross-section
a size of compression block of concrete
The resultant reinforcement force is calculated using:

Psteel (a ) = ∑ P (i, a ) (8.8)
i
The force in concrete is based on the size of compression block of concrete.

PC (a ) = 0.85 ⋅ A( a ) ⋅ f c ' (8.9)
For each size of the compression block, the resistance force of the cross section is
expressed by the sum of all forces acting in the cross-section:
PTotal ( a ) = ∑ P (i , a ) + Pc ( a ) (8.10)
i
112
For each size of the compression block, the bending moment resistance is equal to
sum of all the forces in the section multiplied by the corresponding force arm to the
centroid of the cross-section:
 h 
M Total (a) = ∑ P(i, a) ⋅  − d i  + 0.85 ⋅ A(a) ⋅ f ' c ⋅YC (a) (8.11)
i  2 
8.4. LOAD MODEL
For the scope of this dissertation Cooper River Bridge was modeled using Robot
Millennium software. The three dimensional model was based on the actual design made
by PB World. Shell elements were used to model concrete slab in the bridge. 840 3-D
beam elements were used - with 12 degrees of freedom u1x , u1y , u1z , φ1x , φ1y , φ1z , u2x
, u2y , u2z , φ2x , φ2y , φ2z - to represent all others members: towers leg, tower, girders and
the diaphragms. The cables elements were used to model suspension cables.
Total length of the structure is 3296 ft. Main span is 1546 ft long, two spans are
650 ft and two spans are 225 ft long. Geometry of the bridge is shown in Figure 8.7.
Bridge towers are 568 ft high, both of them have 368 ft above deck level.
113
Figure 8.7. Geometry of Cooper River Bridge
In the analysis all loading cases have been considered. For five spans, there are 31
loading combinations. They are shown in Figure 8.8.
All of the live load combinations were used to calculate load effect on the bridge
tower. The 31 combinations were used four times; for the different value of loading: 0.64
k/ft, 0.80 k/ft, 1.00 k/ft and 1.20 k/ft as a value of lane live load loading.
The resulting envelopes of bending moments due to various combinations of live
load for bridge tower are shown in Figure 8.9 - Figure 8.12.
114
Figure 8.8. Load combinations
115
500
Possition on Pylon (ft)

400 LL1
LL2
LL3
LL4
LL5
LL6
300 LL7
LL8
LL9
LL10
LL11
LL12
200
LL13
LL14
LL15
LL16
LL17
LL18
100
LL19
LL20
LL21
LL22
LL23
0 LL24
-400000 -300000 -200000 -100000 0 100000 200000 300000 LL25
Mz (kips-ft) LL26
LL27
LL28
LL29
-100 LL30
LL31
-200
Figure 8.9. Envelope of bending moments for bridge tower for w=0.64k/ft
116
500

400 LL1
LL2
LL3
LL4
LL5
LL6
300 LL7
LL8
LL9
LL10
LL11
LL12
200
LL13
LL14
LL15
LL16
LL17
LL18
100
LL19
LL20
LL21
LL22
LL23
0 LL24
-500000 -400000 -300000 -200000 -100000 0 100000 200000 300000 400000 LL25
Mz (kips-ft) LL26
LL27
LL28
LL29
-100 LL30
LL31
-200
Figure 8.10. Envelope of bending moments for bridge tower for w=0.80 k/ft
117
500

400 LL1
LL2
LL3
LL4
LL5
LL6
300 LL7
LL8
LL9
LL10
LL11
LL12
200
LL13
LL14
LL15
LL16
LL17
LL18
100
LL19
LL20
LL21
LL22
LL23
0 LL24
-600000 -500000 -400000 -300000 -200000 -100000 0 100000 200000 300000 400000 500000 LL25
Mz (kips-ft) LL26
LL27
LL28
LL29
-100 LL30
LL31
-200
118
500

400 LL1
LL2
LL3
LL4
LL5
LL6
300 LL7
LL8
LL9
LL10
LL11
LL12
200
LL13
LL14
LL15
LL16
LL17
LL18
100
LL19
LL20
LL21
LL22
LL23
0 LL24
-800000 -600000 -400000 -200000 0 200000 400000 600000 LL25
Mz (kips-ft) LL26
LL27
LL28
LL29
-100 LL30
LL31
-200
119
8.5. RELIABILITY RESISTANCE MODELS
Due to various categories of uncertainties, the resistance of a structural

component R , can be considered as a random variable being a product of nominal
resistance Rn and three factors: the materials factor, fabrication factor and professional
factor:
R = Rn M F P (8.12)
The materials factor represents material properties, in particular strength and
modulus of elasticity. The fabrication factor represents the dimensions and geometry of
the component, including cross-sectional area, moment of inertia, and section modulus.
The professional factor represents the approximations involved in the structural analysis
and idealized stress/strain distribution models. The professional factor is defined as the
ratio of the test capacity to analytically predicted capacity (the actual in-situ performance
to the model used in calculations).
The statistical parameters for material factors used in this dissertation were based
on the project "Reliability-Based Calibration for Structural Concrete" (Nowak A.S. et al.,
2008). Because the quality of materials such as reinforcing steel and concrete has
improved over the years, the materials factors have been updated based on a new test
database. There is no new information regarding two other factors, F and P . Therefore,
in most cases, statistical parameters for F and P are taken from the previous study
(Ellingwood et al. 1980).
8.5.1. Material Factor
The material factors for concrete were based on the study within the project
"Reliability-Based Calibration for Structural Concrete" (Nowak A.S. et al., 2008). Figure
8.13 and Figure 8.14 show the bias factor and coefficients of variation for all types of
concrete and all nominal compressive strengths of concrete. In both figures there is a
trend line of changing parameter with respect to concrete compressive strength f c ' .
120
Recommended values are summarized in Table 8.1. In this dissertation, the concrete
compressive strength of bridge tower is 7000 psi. Statistical parameters assumed are: bias
factor λ = 1.13 and coefficient of variation V = 0.12.
1.6
1.5
1.4
1.3
1.2
λ 1.1
1.0
0.9
light weight ordinary, ready mix ordinary, plant cast
0.8
high strength approximation
0.7
0.6
2000
3000
4000
5000
6000
7000
8000
9000
10000
11000
12000
13000
fc'
Figure 8.13 Bias factor for compressive strength of concrete

(Nowak A.S. et al., 2008)
121
0.18
0.17
0.16
0.15
0.14
0.13
0.12
0.11
0.10
V 0.09
0.08
0.07
0.06
0.05
0.04
0.03 light weight ordinary, ready mix ordinary, plant cast
0.02
high strength approximation
0.01
0.00
2000
3000
4000
5000
6000
7000
8000
9000
10000
11000
12000
13000
fc'
Figure 8.14 Coefficient of variation for compressive strength of concrete

Table 8.1 Recommended Statistical Parameters for Compressive Strength, f c '

Concrete Grade fc '
fc' (psi) λ V
4000 1.24 0.15
6000 1.15 0.125
8000 1.11 0.11
12,000 1.08 0.11
The material factors for reinforcement steel were based on the study within the
project "Reliability-Based Calibration for Structural Concrete" (Nowak A.S., Szerszen
122
M.M., et al., 2008). Data included the yield strength for the reinforcing steel bars with the
nominal yield strength of 60 ksi, and different bar sizes from No.3 to No.14. The
recommended values of statistical parameters are: bias factor λ = 1.13 and coefficient of
variation V = 0.03. Those recommended values have been used in this dissertation for the
bar sizes No.9 and No.11, which were used in the calculations. Plots of the cumulative
distribution functions (CDF) of yield strength of every reinforcement size and
recommended parameters are shown in Figure 8.15 and Figure 8.16.
λ = 1.13
3
V = 1.03
2
Normal Inverse Probability
-1
-2
-3
-4
50 60 70 80 90 100 110
Steel rebar size: Yield strength of rebars [ksi]

#3 #4 #5 #6 #7 #8
#9 #10 #11 #14 All diameters Approximation
Figure 8.15 CDF’s of yield strength for Reinforcing Steel Bars, Grade 60 ksi
123
4
3 λ = 1.13
V = 1.03
2
Normal Inverse Probability
-1
-2
-3
-4
50 60 70 80 90 100 110
Yield strength of rebars [ksi]

All diameters Normal Approximation Lognormal Approximation
Figure 8.16 Recommended material parameters for reinforcing steel bars, Grade 60 ksi
8.5.1. Fabrication Factor
The statistical parameters of fabrication factor are based on previous studies by

Ellingwood et al., 1980. Due to lack of data for bridge towers, the parameters for
columns were assumed. They are summarized in Table 8.2.
Table 8.2 Statistical Parameters of Fabrication Factor.

Material Item Bias factor V
Radius of column
Concrete 1.005 0.04
Reinforcement cover
Reinforcement area
Steel 1.00 0.015
Reinforcement diameter
124
8.5.1. Professional Factor
The statistical parameters of professional factor are based on the previous study
performed by Ellingwood et al. (1980). The bridge tower behaves as an eccentrically
loaded column. Therefore, the statistical parameters of professional factors for columns
were used in this dissertation. The professional factors were chosen for tied columns: bias
factor is λ 1.00 and coefficient of variation is V=0.08.
8.5.1. Statistical Parameters of Resistance
Statistical parameters of resistance were obtained by 10000 Monte Carlo

Simulation: coefficient of variation V=0.16 and bias factor λ=1.17.
250000
mean+std
nominal
200000
mean
mean-std
150000 Balance
Force [kips]
Failure
100000
50000
0
0 200000 400000 600000 800000 1000000
Bending Moment [k-ft]
Figure 8.17 Statistical Parameters of Resistance
125
8.6. LOAD MODEL
The case considered in this study is combination Strength I, which is combination

of dead load and live load. This combination has the highest load factor for live load.
For time varying loads two random variables for arbitrary-point-in-time and
maximum 50-year load components can be considered. In this study live load is the only
varying load and it is assumed to reach the maximum 50-year load value. Therefore there
is no need to use Turkstra’s rule.
Variation in the dead load, which is caused by variation of the weight of materials
(concrete and steel), variation of dimensions, and idealization of analytical models affects
statistical parameters of resistance. The assumed statistical parameters for dead load are
based on the data available in literature (Ellingwood et al. 1980; Nowak 1999). They
include for cast-in-place concrete elements a bias factor of 1.05 and coefficient of
variation of 0.10.
Variation of live load is derived in Paragraph 5.4 of this dissertation.
8.7. RELIABILITY ANALYSIS
Reliability analysis was performed for the considered bridge. The element
selected for the analysis was the bridge tower subjected to the bending moment and the
axial force. The bridge was design by the PB World for the design live load specified in
AASHTO LRFD CODE. The limit state function was selected as the Strength I load
combination according to AASHTO LRFD Code. There are only two major load
components: dead load and live load. However, on the real structure in a specific
localization other loads, such as wind load, influence the bridge behavior, the selected
limits state was chosen to demonstrate the sensitivity of reliability index on long span
bridges due to change of live load.
Reliability indexes were calculated for bridge loaded with AASHTO design live
load 0.64 k/f t and three other possible load cases of 0.80, 1.00 and 1.20 k/ft. For every
value of lane load the 31 load cases were analyzed according to paragraph 8.4. Each load
case, Figure 8.8, generates separate loading case to the bridge tower with different
126
eccentricity condition. The results vary depending on eccentricity. None of them exceeds
the balance failure zone, Figure 8.18 . It means that all the cases are in the compression
control and the strength reduction factor is φ=0.75, specified in AASHTO LRFD Section
5.
Nominal Resistance
180000 Balance Failure
Force [kips] Nominal Loads w=0.64 k/ft
Nominal Loads w=0.80 k/ft
160000
140000
120000
100000
80000
60000
40000
20000
0
-750000 -500000 -250000 0 250000 500000 750000
Moment [k-ft]
Figure 8.18 Force and Moment results on the bridge tower for different live loads
Figure 8.19 shows the results of reliability indexes due to different live loads. It
can be noticed that values of β are in high range for the cases with the eccentricity of
loads very small for all possible loading conditions. For the cases with the eccentricities
approaching the balance failure the reliability indexes are decreasing with the increase of
the actual loading.
127
7.00
6.00
5.00
4.00
0.64
β
0.80
3.00
1.00
1.20
2.00
1.00
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
Uniformly Distributed Load [k/ft]
Figure 8.19 Reliability indexes due to different live load
128
CHAPTER 9
SUMMARY & CONCLUSIONS
In the study, a live load model for long span structures was derived. The live load
model is valid for spans between 600 ft and 5000 ft and it is intended to reflect current
traffic patterns, quantities of trucks and their weights. The developed live load model is
recommended to be taken into consideration in the bridge design code.
Preliminary study was performed by reviewing previous research and current
provisions of international codes on the topic. Equivalent uniformly distributed load is
calculated and compared. The new live load was developed based on three models: an
average 5-axle truck, legal load trucks and simulation of a traffic jam using WIM data.
The newest available traffic database from a variety of sites within many different states
was obtained. The magnitude of the database has to be underlined, because such an
extensive actual weigh in motion database has never been used in the derivation of live
load for any kind bridges. A numerical procedure was developed to filtered out WIM data
from erroneous readings and to simulate traffic jam situations. From the simulation the
values of uniformly distributed load were derived for a variety of span lengths and site
localizations. In the developed procedure, starting with the first truck, all consecutive
trucks were added with a fixed clearance distance between them until the total length
reached the span length. Then, the total load of all trucks was calculated and divided by
the span length to obtain the first value of the average uniformly distributed load. Next,
the first truck was deleted, and one or more trucks were added so that the total length of
trucks covers the full span length and the new value of the average uniformly distributed
129
load was calculated. Trucks were kept in actual order, as recorded in the WIM surveys.
Results of the simulations were plotted as a cumulative distribution function of uniformly
distributed load for considered span lengths. The obtained mean value oscillates between
values of 0.50 and 0.75 k/ft. Cumulative distribution functions were also plotted for
maximum daily and maximum weekly combinations of trucks. For longer spans,
uniformly distributed load decreases and is closer to the mean value. This observation
confirms that for a long loaded span, one heavily overloaded truck does not have
significant influence. This is because the load depends on a mix of traffic. The bias
factors (ratio of mean to nominal) were calculated for the heaviest 75-year combination
of vehicles. The 75-year uniformly distributed loads were derived from extrapolated
distributions. It was noticed that the bias factor values for most of the sites do not exceed
1.25, which is similar as in short and medium spans, as shown in the NCHRP Report 368
(1999). It is recommended to use HL-93 also for those long spans. Two other models, an
average 5-axle truck and a legal load trucks model led to similar conclusion. It was
noticed that for some sites, with very heavy traffic, the bias factor reaches a value of 2.0.
Those sites are characterized by high ADTT (usually over 3000) or increased percentage
of overloaded loaded vehicles (over about 10%). Bridges located in such sites require
special attention and application of increased design live load. The value of the design
load should be agreed with the owner of the structure. For some bridges considered, for
example in the area of New York, it was found that the uniformly distributed load should
be 1.25 kip/ft in order to obtain bias lower than 1.25. Statistical parameters for live load
are: for spans 600-1000 ft bias 1.25 and coefficient of variation 0.10, for spans longer
than 1000 ft bias 1.20 and coefficient of variation 0.08.
In this dissertation, the problem of multilane reduction factors was discussed.
Multilane factors were found to be very site specific, as with the live load. Video
recordings of traffic confirmed that for the majority of the time we can observe that the
moving lanes contain a mixture of trucks and cars. However, situations when one lane is
almost exclusively occupied by trucks or trucks occupy three or four lanes at the same
time are also possible. Multiple reduction factors for design live load should account for
those most critical loading cases. In the WIM database available, the vehicles of 1-3
categories 1-3 have not been registered. Therefore, it did not allow for simulations and
130
derivation of multilane factors for all traffic lanes. Simulation of the traffic on the most
loaded lane was possible with an assumption, that in traffic jam situations passenger
vehicles merge to the left and the right lane remains occupied exclusively by trucks. It
was stated that equal reduction of load on all traffic lane does not reflect the actual
situation. At least one of the lanes should be loaded more than the others. It was
concluded that the multilane reduction factors have to be an objective of additional
extensive studies.
The study of dynamic factor was performed for the research. It was concluded that
the current dynamic load factor of 0.33 is too high for bridges with longer spans. It may be
applicable in short bridges, when vibration due to “wheel hop” on the approach slab is
significant. For longer bridges, where the influence of the approach slab decreases and
vibrations of many vehicles interfere with each other, the dynamic load factor could be
smaller. The assumption of a traffic jam situation to develop live load model induces no
dynamic allowance. However, to not introduce confusion among designers it is
recommended do keep dynamic factor as it is for short and medium span bridges, which
results in very small value for long span bridges.
131
CHAPTER 10
RECCOMENDATIONS
1. The developed live load model is valid for long spans between 600 ft and 5000 ft.
2. For long spans it is recommended to use HL-93 load, specified in AASHTO
LRFD Code (2007), uniformly distributed load of 0.64 k/ft plus design truck or
tandem. The bias factor calculated for the heaviest 75-year combination of
vehicles did not exceed 1.25, which is similar as in short and medium spans, as
shown in the NCHRP Report 368 (1999).
3. For some sites characterized by high ADTT or increased percentage of
overloaded vehicles (over 10%) bias factor reaches a value of 2.0. Those bridges
require application of increased site specific design live load, which should be
agreed with the owner of the structure.
4. It was proposed to use dynamic factor as specified in AASHTO LRFD Code
(2007). Developed live load model assumes traffic jam situation and does not
allow for dynamic. However, to not introduce confusion among designers it is
recommended do keep dynamic factor as it is for short and medium span bridges,
which results in very small value for long span bridges.
5. It was proposed to use multilane reduction factors as specified in AASHTO
LRFD Code (2007). It is recommended to perform further studies in this field.
132
6. Statistical parameters for live load are:
- for spans 600-1000 ft: bias 1.25 and coefficient of variation 0.10,
- for spans longer than 1000 ft: bias 1.20 and coefficient of variation 0.08.
The developed live load model is recommended to be taken into consideration in the
AASHTO LRFD Bridge Design Specifications.
133
REFERENCES
[1]. AASHTO LRFD Bridge Design Specifications. (2007). American Association of

State Highway and Transportation Officials, Washington, D.C.
[2]. AASHTO Geometric Design of Highways and Streets (2001). American

Association of State Highway and Transportation Officials, Washington, D.C.
[3]. CAN/CSA-S6-00 Canadian Highway Bridge Design Code, CSA International,

Toronto, Canada.
[4]. Manual for Condition Evaluation and load and Resistance Factor Rating (LRFD) of
Highway Bridges. 2003. American Association of State Highway and
Transportation Officials, Washington, D.C.
[5]. OHBDC (1991), Ontario Highway Bridge Design Code, Ontario Ministry of
Transportation, Downsview, Ontario.
[6]. ASCE Committee on Loads and Forces on Bridges (1981). “Recommended Design
Loads for Bridges”, ASCE Journal of Structural Engineering, Vol. 107, No. 7,
December 1981, pp. 1161-1213.
[7]. Ang, A. H-S., and Tang, W.H., “Probability Concepts in Engineering Planning and
Design,” Volume I: Basic Principles, John Wiley & Sons, new York, 1975
[8]. Ang, A. H-S., and Tang, W.H., “Probability Concepts in Engineering Planning and
Design,” Volume II: Decision, Risk, and Reliability, John Wiley & Sons, New
York, 1984
[9]. Augusti G., Barrata B. “Limit and Shakedown Analysis of Structures with
Stochastic Strengths,” Proceedings of the Second SMiRT Conference, Berlin, 1973
[10]. Augusti G., Barrata B. “Plastic Shakedown of Structures with Stochastic Local
Strengths,” Proceedings of the Symposium on Resistance Ultimate Deformability of
Structures, IABSE, Lisbon, 1973
[11]. Ayyub B.M., Halder A. “Practical Structural Reliability Techniques,” Journal of

Structural Engineering, ASCE, Vol. 110, No. 8, 1984, pp. 1707-1724
134
[12]. Ayyub B.M., McCuen R.H. “Probability, Statistics, & Reliability for Engineers,”
CRC Press Boca Raton, New York, 1997
[13]. Benjamin J.R., Cornell C.A. “Probability, Statistics and Decision for Civil
Engineers,” McGraw-Hill, New York, 1970
[14]. Biggs, J.M., Suer, H.S. 1955. “Vibration Measurements on Simple-Span Bridges”,
Highway Research Board, Bulletin 124.
[15]. Breitung K. “Asymptotic Approximations for Multinormal Integrals,” Journal of

Engineering Mechanics, ASCE, Vol. 110, 1984, pp. 357-366
[16]. Buckland, P.G., Navin, F.P.D, Zidek, J.V., and McBryde, J.P, 1978, “Traffic
loading of long-span bridges”, Transportation Research Record, Vol. 665, pp. 146-
154.
[17]. Buckland, P.G., Navin, F.P.D, Zidek, J.V., and McBryde, J.P., 1980, “Proposed
Vehicle Loading of Long-Span Bridges”, Journal of Structural Division, ASCE,
Vol. 106, pp. 915-932.
[18]. Buckland, P.G., 1991, “North American and British Long-Span Bridge Loads”,
Journal of Structural Division, ASCE, Vol. 117, pp. 2972-2987.
[19]. Cantieni, R. 1984. “Dynamic Load Testing of Highway Bridges”, Transportation

Research Record No 950, pp. 141-148.
[20]. Cebon, D. 1999. “Handbook of vehicle-road interaction”, Department of

Engineering, University of Cambridge. London.
[21]. Cebon, D. 1993. “Interaction between heavy vehicles and roads”, Department of
Engineering, University of Cambridge. London.
[22]. Cornell, C. A., “Bounds on the Reliability of Structural Systems,” Journal of the
Structural Division, ASCE, 93 (ST1), 171-200, 1967
[23]. Corotis R.B., Nafday A.M. “Structural Systems Reliability using Linear
Programming and Simulations,” Journal of Structural Engineering, ASCE, Vol. 115,
No. 10, 1989, pp. 2435-2447
135
[24]. Ellingwood, B.; Galambos, T.V.; MacGregor, J.G., and Cornell, C.A.,
“Development of a Probability Based Load Criterion for American National
Standard A58,” NBS Special Publication 577, Washington, DC, National Bureau of
Standards, 1980
[25]. Ellingwood, B., MacGregor, J.G., Galambos, T.V., and Cornell, C.A., “Probability
Based Load Criteria: Load factors and Load Combinations,” Journal of Structural
Engineering, ASCE, Vol. 108, No. ST5, 1982
[26]. Fiessler B., Naumann H-J., Rackwitz R. “Quadratic Limit States in Structural
Reliability,” Journal of Engineering Mechanics, ASCe, Vol. 105, 1979, pp. 661-676
[27]. Fleming, F.J., Romualdi, J.P. 1961. “Dynamic Response of Highway Bridges”,
Journal of the Structural Division, Vol. 87.
[28]. Fraudenthal A.M. “Safety and Probability of Structural Failure,” Transactions,

ASCE, Vol. 121, 1956, pp. 1337-1375
[29]. Galambos, T.V., and Ravindra, M.K., “Load and Resistance Factor Design,”
Journal of Structural Devision, ASCE, ST9, Proc. Paper 14008, Sep. 1978
[30]. Galdos, N.H., Schelling, D.R., Sahin, M.A. 1993. “Methodology for Impact Factor
of Horizontally Curved Box Bridges”, Journal of Structural Engineering, Vol. 119,
No. 6, pp. 1917-1934.
[31]. Gindy, M., Nassif, H.H., 2006, “Multiple Presence Statistics for Bridge Live Load
Based on Weigh-in-Motion Data”, TRB annual meeting
[32]. Hasofer A.M., Lind N.C. “An Exact and Invariant First Order Reliability Format,”
Journal of Structural Mechanics Division, ASCE, Vol. 100, No. 1, 1974
136
[33]. Hohenbichler M., Rackwitz R. “Improvement of Second-Order Reliability Estimates
by Importance Sampling,” Journal of the Engineering Mechanics, ASCE, Vol. 114,
1988, pp. 2195-2199
[34]. Hwang, E-S. and Nowak, A.S., “Simulation of Dynamic Load for Bridges”, ASCE
Journal of Structural Engineering, Vol. 117, No. 5, May 1991, pp. 1413-1434.
[35]. Khalifa, M.A. 1992. “Parametric Study of Cable-Stayed Bridge Response Due to
Traffic-induced Vibration”, Computers & Science, Vol. 47, No. 2, pp. 321-339.
[36]. Khisty, C.J. Prentice Hall, Englewood Cliffs, New Jersey, “Transportation
Engineering”, 1990.
[37]. Kim, S-J., Sokolik, A.F., and Nowak, A.S., 1997, “Measurement of Truck Load on
Bridges in the Detroit Area”, Transportation Research Record, No.1541, pp. 58-63.
[38]. Kiureghian A., Liu P-L. “Structural Reliability under Incomplete Probability
Information,” Journal of Engineering Mechanics, ASCE, Vol. 112, 1986, pp. 85-104
[39]. Kulicki J.M., Prucz Z., Clancy C.M., Mertz D.R., Nowak A.S., “Updating the
calibration report for AASHO LRFD”, NCHRP 20-7/186, January 2007
[40]. Lutomirski, T.A., “Reliability models for circular concrete columns”, Doctoral
Dissertation, 2009.
[41]. MacAdam, C.C., Fancher, P.S., Hu, G.T., Gillespie, T.D. 1980. “A computerized
model for simulating the breaking and steering dynamics of trucks, tractor-
semitrailers, doubles, and triples combinations”, MVMA Project 1197.
[42]. McLean, D.I., Marsh, M.L. 1998. “Dynamic Impact Factors for bridges”, NCHRP
Synthesis 266.
[43]. Nassif, H.H., and Nowak, A.S., 1995, “Dynamic Load Spectra for Girder Bridges”,
Transportation Research Record, No. 1476, pp. 69-83.
[44]. Navin, FPD, 1982, “Road vehicles on bridges”, Canadian Journal of Civil
Engineering, Vol. 9, pp. 468-476.
137
[45]. Nowak, A.S., “Calibration of LRFD Bridge Design Code,” NCHRP Report 368,
Transportation Research Board, Washington, D.C., USA, 1999
[46]. Nowak, A.S., and Lind , N.D., “Practical Bridge Code Calibration,” Journal of
Structural Division, ASCE, pp. 2497-2510, December 1979
[47]. Nowak, A.S., 1993, “Live Load Model for Highway Bridges”, Journal of Structural
Safety, Vol. 13, Nos. 1+2, December, pp. 53-66.
[48]. Nowak, A.S., “Calibration of LRFD Bridge Design Code”, Final Report, NCHRP
Project 12-33, December 1993.
[49]. Nowak, A.S., 1994, “Load Model for Bridge Design Code”, Canadian Journal of
Civil Engineering, Vol. 21, pp. 36-49.
[50]. Nowak, A.S., 1995, “Calibration of LRFD Bridge Code”, ASCE Journal of
Structural Engineering, Vol. 121, No. 8, pp. 1245-1251.
[51]. Nowak, A.S. and Hong, Y-K., “Bridge Live Load Models”, ASCE Journal of
Structural Engineering, Vol. 117, No. 9, Sept. 1991, pp. 2757-2767.
[52]. Nowak, A.S., Hong, Y-K and Hwang, E-S., “Modeling Live Load and Dynamic
Load for Bridges”, Transportation Research Record, No. 1289, 1991, pp. 110-118.
[53]. Nowak, A.S., Kim, S-J., Laman, J., Saraf, V. and Sokolik, A.F., “Truck Loads on
Selected Bridges in the Detroit Area”, Final Report submitted to MDOT, November
1994.
[54]. Nowak, A.S., Laman, J. and Nassif, H., “Effect of Truck Loading on Bridges”, Final
Report submitted to MDOT, December 1994.
[55]. Nowak, A.S. and Collins, K.R., “Reliability of Structures,” McGraw-Hill New
York, New York, USA, 2000
138
[56]. MacGregor J.G., Mirza S.A., Ellingwood B., “Statistical Analysis of Resistance of
Reinforced and Prestressed Concrete Members”, ACI Structural Journal, Title
no.80-16, May-June, 1983, pages 167 to 176.
[57]. Madsen H.O., Krenk S., Lind N.C. “Methods of Structural Safety,” Prentice-Hall,
1986
[58]. Murzewski J. “Reliability of Engineering Structures” (in Polish), Arkady, Warsaw,

Poland, 1989
[59]. Paultre, P., Chaallal, O., Proulx, J. 1992. “Bridge Dynamics and Dynamic
Amplification Factors: A Review of Analytical and Experimental Findings”,
Canadian Journal of Civil Engineering, Vol. 19, No. 2, pp. 260-278.
[60]. Rackwitz, R., Fiessler, B. “Structural Reliability under Combined Random Load
Sentences,” Computers & Structures, Vol. 9, 1978, pp. 489-494
[61]. Rosenblatt, M. “Remarks on Multivariate Transformations,” The Annuals of

Mathematical Statistics, Vol. 23, 1952, pp. 470-472
[62]. Schelling, D.R., Galdos, N.H., Sahin, M.A. 1992. “Evaluation of Impact Factors for
Horizontally Curved Steel Box Bridges”, Journal of Structural Engineering, Vol.
118, No. 11, pp. 3203-3221.
[63]. Sivakumar, B., Ghosn, M., Moses, F., “Protocols for Collecting and Using Traffic
Data in Bridge Design,” Final Report, NCHRP Project 12-76, July 2008.
[64]. Thoft-Christiansen, P., Baker, M.J. “Structural Reliability Theory and Its
Applications,” Springer-Verlag, Berlin, Heidelberg, New York, 1982
[65]. Thoft-Christiansen, P., Murotsu, Y. “Application of Structural Systems Reliability

Theory,” Springer-Verlag, Berlin, Heidelberg, New York, Tokio, 1986
139
[66]. Wang, T.L., Shahawy, M., Huang, D.Z. 1993. Dynamic Response of Highway
Truck Due to Road Surface Roughness. Computers and Structures, Vol. 49, No. 6,
pp. 1055-1067.
[67]. Wang, T.L, Hunag, D.Z. 1992. “Cable-stayed Bridge Vibration Due to Road Surface
Roughness”, Journal of Structural Engineering, Vol. 188, No. 5, pp. 1354-1374.
[68]. Wang, T.L., Huang, D.Z., Shahawy, M. 1996. “Dynamic Behavior of Continuous
and Cantilever Thin-Walled Box Girder Bridges”, Journal of Bridge Engineering,
Vol. 1, No. 2, pp. 67-75.
[69]. Winkler, C.B., Fancher, P.S., MacAdam, C.B. 1983. “Parametric Analysis of heavy
duty truck dynamic stability”, UMTRI-83-13.
[70]. Wolinski Sz., Wrobel K. “Reliability of Engineering Structures” (in Polish),

Publishers of Rzeszow Technical University (Oficyna Wydawnicza Politechniki
Rzeszowskiej), 2001
[71]. Zhu, X.Q., Law, S.S. 2002. “Dynamic Load on Continuous Multi-lane Bridge Deck
from moving vehicles”, Journal of Sound and Vibration, Vol. 251, No. 4, pp. 697-
716.
[72]. Transportation Statistics Annual Report, December 2006, U.S. Department of

Transportation.
[73]. “Regulation of Weights, Lengths, and Widths of Commercial Motor Vehicles”.

Special Report 267. TRB, National Research Council, Washington, DC, 2002.
[74]. “Comprehensive Truck Size and Weight Study. Volume I: Summary Report”.
Publication FHWA-PL-00-029. FHWA, U.S. Department of Transportation,
Washington, DC, 2000.
[75]. Truck Driver's Guidebook, Michigan Center for Tuck Safety, 2008.
[76]. “Model for Highway Bridge Impact”, Journal of Structural Engineering, Vol. 116,
No. 7, pp. 1772-1793.
[77]. Traffic Monitoring Guide, U.S. Department of Transportation, Federal Highway

Administration, 2001.
140
APPENDIX A
CDF OF UDL FOR ALL TRUCK COMBINATIONS
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.1. CDF of UDL for Oregon I-5 Woodburn, lane 1
141
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.2. CDF of UDL for Oregon I-84 Emigrant Hill, lane 1
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.3. CDF of UDL for Oregon OR 58 Lowell, lane 1
142
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.4. CDF of UDL for Oregon OR 58 Lowell, lane 2

5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.5. CDF of UDL for Oregon US 97 Bend, lane 1
143
1st lane
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.6. CDF of UDL for Florida 9916, lane 1

5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
144
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00

5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
145
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.10. CDF of UDL for Indiana 9534, lane 1

5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
146
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00

5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
147
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00

5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.15. CDF of UDL for New York 9121, lane 1
148
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00

5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
149
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00

5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.19. CDF of UDL for New York I-495 EB, lane 1
150
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
151
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.22. CDF of UDL for New York I-495 WB, lane 1
5.00
4.00
3.00
2.00
600 ft
1.00
1000 ft
2000 ft
0.00
3000 ft
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
4000 ft
-1.00
5000 ft
-2.00
-3.00
-4.00
-5.00
Figure A.23. CDF of UDL for New York I-495 WB, lane 2
152
APPENDIX B
CDF OF MAXIMUM DAILY UDL
4.0
3.0
600 ft
1000 ft
2.0
2000 ft
3000 ft
4000 ft
1.0
5000 ft
extrpolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
-3.0
Figure B.1. CDF of maximum daily UDL for Oregon I-5 Woodburn, lane 1
153
4.0
3.0
600 ft
1000 ft
2.0
2000 ft
3000 ft
1.0 4000 ft
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
-3.0
Figure B.2. CDF of maximum daily UDL for Oregon I-84 Emigrant Hill, lane 1
4.0
3.0
600 ft
1000 ft
2.0 2000 ft
3000 ft
4000 ft
1.0
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
-3.0
Figure B.3. CDF of maximum daily UDL for Oregon OR 58 Lowell, lane 1
154
4.0
3.0
600 ft
1000 ft
2.0
2000 ft
3000 ft
1.0 4000 ft
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
-3.0
Figure B.4. CDF of maximum daily UDL for Oregon US 97 Bend, lane 1
4.0
3.0
2.0
600 ft
1000 ft
1.0
2000 ft
3000 ft
4000 ft
0.0
0.00 0.50 1.00 1.50 2.00 5000 ft
extrapolated
-1.0
-2.0
-3.0
Figure B.5. CDF of maximum daily UDL for Florida 9916, lane 1
155
4.0
3.0
600 ft
1000 ft
2.0
2000 ft
3000 ft
4000 ft
1.0
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
-3.0
4.0
3.0
2.0
600 ft
1000 ft
1.0
2000 ft
3000 ft
4000 ft
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 5000 ft
extrapolated
-1.0
-2.0
-3.0
156
4.0
3.0
600 ft
2.0 1000 ft
2000 ft
3000 ft
1.0 4000 ft
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
-3.0
4.0
3.0
2.0
600 ft
1000 ft
1.0
2000 ft
3000 ft
4000 ft
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 5000 ft
extrapolated
-1.0
-2.0
-3.0
Figure B.9. CDF of maximum daily UDL for Indiana 9534, lane 1
157
4.0
3.0
2.0
600 ft
1000 ft
1.0
2000 ft
3000 ft
4000 ft
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 5000 ft
extrapolated
-1.0
-2.0
-3.0
4.0
3.0
2.0
600 ft
1000 ft
1.0
2000 ft
3000 ft
4000 ft
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 5000 ft
extrapolation
-1.0
-2.0
-3.0
158
4.0
3.0
600 ft
2.0 1000 ft
2000 ft
3000 ft
1.0 4000 ft
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40
-1.0
-2.0
-3.0
Figure B.12. CDF of maximum daily UDL for New York 9121, lane 1
4.0
3.0
600 ft
1000 ft
2.0 2000 ft
3000 ft
4000 ft
1.0
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40
-1.0
-2.0
-3.0
Figure B.13. CDF of maximum daily UDL for New York 2680, lane 1
159
4.00
3.00
600 ft
1000 ft
2.00
2000 ft
3000 ft
1.00 4000 ft
5000 ft
extrapolated
0.00
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20
-1.00
-2.00
-3.00
Figure B.14. CDF of maximum daily UDL for New York I-495 EB, lane 1
4.0
3.0
600 ft
1000 ft
2.0
2000 ft
3000 ft
4000 ft
1.0
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40
-1.0
-2.0
-3.0
Figure B.15. CDF of maximum daily UDL for New York I-495 WB, lane 1
160
APPENDIX C
CDF OF MAXIMUM WEEKLY UDL
4.0
3.0
2.0 600 ft
1000 ft
2000 ft
1.0 3000 ft
4000 ft
5000 ft
0.0 extrapolated
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
Figure C.1. CDF of maximum weekly UDL for Oregon I-5 Woodburn, lane 1
161
4.0
3.0
600 ft
1000 ft
2.0
2000 ft
3000 ft
4000 ft
1.0
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
Figure C.2. CDF of maximum weekly UDL for Oregon I-84 Emigrant Hill, lane 1
4.0
3.0
600 ft
1000 ft
2.0 2000 ft
3000 ft
4000 ft
5000 ft
1.0
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
Figure C.3. CDF of maximum weekly UDL for Oregon OR 58 Lowell, lane 1
162
4.0
3.0
600 ft
1000 ft
2000 ft
2.0
3000 ft
4000 ft
5000 ft
1.0
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
Figure C.4. CDF of maximum weekly UDL for Oregon US 97 Bend, lane 1
4.0
3.0
2.0
600 ft
1000 ft
1.0
2000 ft
3000 ft
4000 ft
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 5000 ft
extrapolated
-1.0
-2.0
-3.0
Figure C.5. CDF of maximum weekly UDL for Florida 9916, lane 1
163
4.0
3.0
2.0 600 ft
1000 ft
2000 ft
1.0 3000 ft
4000 ft
5000 ft
0.0 extrapolated
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
4.0
3.0
600 ft
1000 ft
2.0
2000 ft
3000 ft
1.0 4000 ft
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
-3.0
164
4.0
3.0
2.0 600 ft
1000 ft
2000 ft
1.0 3000 ft
4000 ft
5000 ft
0.0 extrapolated
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-1.0
-2.0
4.0
3.0
2.0
600 ft
1000 ft
1.0
2000 ft
3000 ft
4000 ft
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 5000 ft
extrapolatred
-1.0
-2.0
-3.0
Figure C.9.CDF of maximum weekly UDL for New York 9121, lane 1
165
4.0
3.0
600 ft
1000 ft
2.0
2000 ft
3000 ft
1.0 4000 ft
5000 ft
extrapolated
0.0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40
-1.0
-2.0
-3.0
Figure C.10. CDF of maximum weekly UDL for New York 2680, lane 1
166

Live Load Models For Long Span Bridges

Uploaded by

Copyright:

Available Formats

Live Load Models For Long Span Bridges

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Live Load Models For Long Span Bridges

Uploaded by

Copyright:

Available Formats

University of Nebraska - Lincoln

DigitalCommons@University of Nebraska - Lincoln

Live Load Models for Long Span Bridges

Follow this and additional works at: http://digitalcommons.unl.edu/civilengdiss

Presented to the Faculty of

Under the Supervision of Professor Andrzej S. Nowak

FOR LONG SPAN BRIDGES

Marta Lutomirska, PhD

University of Nebraska, 2009

Advisor: Andrzej S. Nowak

simulations were plotted as a cumulative distribution function of uniformly distributed

increased percentage of overloaded loaded vehicles, require special attention and

to be taken into consideration in the bridge design code.

I wish to express my deepest gratitude to Professor Andrzej S. Nowak, my

The Department of Civil Engineering at the University of Nebraska-Lincoln is

I am deeply grateful to my husband, Tomasz, my parents, and whole family for

ABSTRACT .................................................................. Error! Bookmark not defined.

1.1. PROBLEM STATEMENT

1.2. OBJECTIVE AND BENEFITS OF THE STUDY

Chapter 1 of this dissertation is an introduction to the research conducted. It

1.4. PRIOR INVESTIGATIONS

1.4.2. Prior Investigations on Live Loading on Long Span Bridges

1.4.3. Prior Investigations on Structural Reliability

LIVE LOAD IN CURRENT DESIGN CODES

2.2.1. AASHTO LRFD Code [2007]

The American Association of State Highway and Transportation Officials

32.0 kip 32.0 kip

14’ 14’ - 30’

Figure 2.1. HL-93 Live Loading in AASHTO LRFD Code [2007].

25.0 kip 25.0 kip

Figure 2.2. HL-93 Live Loading in AASHTO LRFD Code [2007].

7.2 kip 7.2 kip

14’ 14’ - 30’ ≥ 50’ 14’ 14’ - 30’

Figure 2.3. HL-93 Live Loading in AASHTO LRFD Code [2007].

2.2.2. OHBDC [1991]

36.0 kip 36.0 kip 45.0 kip 36.0 kip

12’ 4’ 20’ 24’

Figure 2.4. OHBD Live Loading [1991]. OHBD Truck.

12’ 4’ 20’ 24’

2.2.3. CAN/CSA-S6-00 [2000]

12’ 4’ 22’ 22’

Figure 2.6. CAN/CSA-S6-00 Live Loading [2000]. CL-W Truck

31.5 kip 27.0 kip

12’ 4’ 22’ 22’

Figure 2.7. CAN/CSA-S6-00 Live Loading [2000]. CL-W Lane Load

Table 2.2. Number of design lanes vs. road width in CAN/CSA-S6-00

2.2.4. BS 5400 [2006]

Figure 2.8. BS 5400 Live Loading curve HA UDL [2006]

Figure 2.9. Dimensions of HB vehicle.

2.2.5. Eurocode 1 [2002]

1) Load Model 1 (LM1) consists of concentrated and uniformly distributed loads

Figure 2.10. Eurocode 1 [2002]. Load Model 1

2.2.6. ASCE Loading [1981]

Figure 2.11. ASCE Loading on Log Scale

There is considerable variation in the treatment of dynamic load effects by bridge

Table 2.4. Dynamic allowance in AASHTO LRFD [2007]

Until recently AASHTO Standard Specifications for Highway Bridges used to

2.4. PROVISIONS FOR MULTILANE REDUCTION FACTORS

AASHTO LRFD (2007) 1.20 1.00 0.85 0.65 0.65 0.65