UNDERSTANDING THE DYNAMICS OF TRUCK TRAFFIC ON
FREEWAYS BY EVALUATING TRUCK PASSENGER CAR
EQUIVALENT (PCE) IN THE HIGHWAY CAPACITY MANUAL
(HCM) 2010
By
Susi Marlina
Associate Eng. (A.Md.) in Civil Engineering, University of Lampung, Indonesia 1997
B.Sc. in Civil Engineering, University of Bandar Lampung, Indonesia 2000
M.Sc. in Civil Engineering, University of Colorado at Denver, USA 2006
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Civil Engineering
2012
2012 by Susi Marlina
All rights reserved
This thesis for the Doctor of Philosophy
degree by
Susi Marlina
has been approved
by
Bruce N. Janson
Wesley Marshall
Yuk Lee
Angela Bielefeldt
Juan Robles
Marlina, Susi (Ph.D, Civil Engineering)
Understanding the Dynamics of Truck Traffic on Freeways by Evaluating Truck
Passenger Car Equivalent (PCE) in the Highway Capacity Manual (HCM) 2010
Thesis directed by Professor Bruce N. Janson
ABSTRACT
Truck traffic causes significant problems, including congestion, delay, crashes,
pollution, energy consumption, and road damage in many regions because trucks
are larger in size and heavier than passenger cars. In addition, trucks have limited
performance, especially on grades and curves for accelerating and decelerating. A
common treatment of truck traffic procedures for highway capacity and level of
service (LOS) determination is to multiply trucks by a passenger car equivalent
(PCE) based on Highway Capacity Manual (HCM) guidelines. A deficiency exists in
the current PCE estimation because a truck does not perform like multiple cars. The
main objective of this study is to understand the dynamics of truck traffic and develop
a new truck PCE using a simulationbased approach.
Speedflowdensity relationship was used for this study because speedflowdensity
is able to measure traffic flow and the randomness in the traffic stream and its
validity has been proven by empirical research with field observations. The basic
concept of PCE derivation uses a deterministic model of traffic flow from
Greenshields Model proposed by Huber and developed by Sumner et al. Their
concept is based on an equal speeddensity relationship. That method was
compared to Demarchi and Settis to estimate truck PCE based on equal flow
density relationship. Those PCE values were used to evaluate the current HCM
2010. Statistical analysis was conducted to investigate the differences between
calculating truck PCE using equal flowdensity and equal speeddensity
methodologies.
A number of different scenarios and variables were considered: truck proportion,
grade percentages, length of grades, number of lanes, and truck lane restrictions.
New truck PCE estimation tables are proposed for future revisions of the Highway
Capacity Manual.
This abstract accurately represent the content of the candidates thesis. I
recommend its publication.
Approved
Bruce N. Janson
DEDICATION
I dedicate this dissertation to my parents and family for their unfaltering
understanding, praying and support to reach my dreams. I also dedicate this
dissertation to my extended family, friends and colleagues, both in Indonesia and in
the USA, who gave me an appreciation of learning and taught me the values of
perseverance, humility, persistence, and passion.
ACKNOWLEDGEMENT
I am sincerely grateful to the Almighty Creator of the World.
The ideas, models and outcomes contained in this dissertation could not have
been achieved without the support of many people.
I wish to convey my deepest gratitude to Professor Bruce N. Janson, who
served as my advisor and mentor. With his guidance, the majority of the ideas
and work contained in this dissertation were finely sharpened and organized.
His energy has given an enormous boost to my work. Most of all, I wish to
sincerely thank him for his enthusiasm toward research and his
encouragement to persevere. Without his support, this dissertation could not
been possible.
I also appreciate the guidance given me by the members of my advisory
committee: Dr. Yuk Lee, Dr. Wesley Marshall, Dr. Angela Bieldfeldt, and Dr.
Juan Robles.
I give my deepest thanks to my parents and siblings for their endurance and
unfaltering prayers.
I also wish to express my gratitude to my host family and friends Katie
Paganucci, Jay Harker, Valerie Kiltzer, Zachary Held, Jeremy Harker, and
David. They have provided me an alternate home here.
I also wish to thank Kathie Haire and Cindy Collip for supportive assistance
and provocative discussions. Thanks to all of my peers at Parson Brinckerhoff
for their positive encouragement and patience with me, especially Jim Daves.
He provided the fortunate opportunity to work on the 170 simulation research
project as part of subcontract of University of Colorado Denver.
A million thanks goes to Jim Root, David Krutsinger, Karl Bucholz, Alvin
Stamp, Patricia Batuna, Ririen Indriani, and Ronny H. Purba for their support
of all my effort to get it done! I have also enjoyed advice from C. Jon Chen to
my intellectual development.
I also would like to acknowledge and thank Katherine Carol, Adriana Carlson,
and Mikelle Learned for the opportunity to live and work together at the
beginning of my doctorate program.
Sincere thanks goes to Krista Nordback for her efforts with proofreading my
thesis and reviewing my dissertation .Kudos to Pamela Fischhaber for her
brainstorming with me in preparation of my defense.
Finally, I would deliver my gratitude to the Colorado Department of
Transportation (CDOT) especially Region 1, University of Bandar Lampung 
Indonesia, and the Indonesian community and friends both in the USA and in
Indonesia.
Susi Marlina
TABLE OF CONTENTS
LIST OF TABLES..............................................................x
LIST OF FIGURES...........................................................xii
Chapter
1. Introduction............................................................1
1.1 Background.........................................................1
1.2 Study Objectives...................................................2
1.3 Significant Studies................................................3
1.4 Simulation.........................................................6
1.5 Organization of the Dissertation...................................7
2. Literature Review.......................................................9
2.1 Review of Truck Equivalencies......................................9
2.1.1 PCEs Based on Flow Rates and Density............................10
2.1.2 PCEs Based on Headways..........................................12
2.1.3 PCEs Based on Queue Discharge Flow..............................15
2.1.4 PCEs Based on Speed.............................................15
2.1.5 PCEs Based on Delays............................................16
2.1.6 PCEs Based on V/C Ratio.........................................19
2.1.7 PCEs Based on VehicleHours.....................................19
2.1.8 PCEs Based on Platoon Formation.................................20
2.1.9 PCEs Based on Travel Time.......................................21
2.2 Truck Equivalencies in the 2010 HCM...............................22
2.3 Traffic Flow Parameter Concepts...................................24
2.4 Level of Service (LOS)............................................26
2.5 Vehicle Classification............................................28
2.6 Simulation Models Study.......................................31
3. Methodology............................................................32
3.1 PCE Methods Theoretical Analysis Evaluation...................32
3.2 Simulation........................................................34
3.2.1 Decision Support Methodology (DSM)..............................34
3.2.2 VISSIM Micro Simulation.........................................38
3.2.3 Calibration, Validation and Verification........................43
3.3 Steps Performed to Develop Truck Equivalence Factors..............45
3.3.1 Equal FlowDensity..............................................45
3.3.2 Equal SpeedFlow................................................48
3.4 Scenarios.........................................................51
4. Analysis and Results...................................................55
4.1 Overview..........................................................55
4.2 SpeedFlowDensity Relationship...................................56
viii
4.2.1 Trucks Proportion...............................................56
4.2.2 Length of Grade.................................................62
4.2.3 Grade Percentages...............................................67
4.3 PCE Results......................................................72
4.3.1 Variance PCE by Trucks Proportion...............................72
4.3.2 Variance PCE by Length of Grade.................................81
4.3.3 Variance PCE by Grade Percentages...............................90
4.3.4 Variance PCE by Number of Lane.................................105
4.3.4.1 Truck Proportion.............................................105
4.3.4.2 Length of Grade...............................................114
4.3.4.3 Grade Percentages.............................................122
4.3.5 Variance PCE by Trucks Restrictions............................136
4.3.5.1 Truck Proportion.............................................137
4.3.5.2 Length of Grade...............................................145
4.3.5.3 Grade Percentages.............................................153
5. Statistical Analysis.................................................168
5.1 Overview.........................................................168
5.2 Truck Proportion.................................................169
5.2.1 Mean, Standard Deviation and Other Statistics..................169
5.2.2 Independent Sample ttest......................................171
5.2.3 Percent Differences............................................173
5.3 Length of Grade.................................................176
5.3.1 Mean, Standard Deviation and Other Statistics..................177
5.3.2 Independent Sample ttest......................................178
5.3.3 Percent Differences............................................180
5.4 Grade Percentages................................................184
5.4.1 Mean, Standard Deviation and Other Statistics..................185
5.4.2 Independent Sample ttest......................................186
5.4.3 Percent Differences............................................188
6. Additional Study.....................................................192
6.1 Overview.........................................................192
6.2 Fuel Consumption and Emissions...................................193
6.3 Summary..........................................................197
7. Conclusions and Recommendations......................................198
7.1 Conclusions......................................................198
7.2 Recommendations..................................................204
REFERENCES............................................................214
APPENDIX..............................................................219
IX
LIST OF TABLES
Table
2.1 PCEs for Heavy Vehicles in General Terrain Segments (HCM, 2010)...........22
2.2 PCEs for Trucks and Buses (Et) on Specific Downgrades (HCM, 2010)........23
2.3 PCEs for Trucks and Buses (Et) on Upgrades (HCM, 2010)...................23
2.4 SpeedDistance Relationship for Acceleration of Heavy Truck (AASHTO A
Policy on Geometric Design of Highways and Streets, 2000)....................26
2.5 Vehicle Weight and Power (Traffic Engineering Handbook, 2009)...........31
3.1 Decision Support Methodology (DSM) Result................................36
3.2 Car Following VISSIM Parameters.........................................42
3.3 Lane Changing VISSIM Parameters.........................................42
3.4 Interstate 70 Simulation Calibration and Validation.....................44
3.5 Truck Percentages Scenarios..............................................52
3.6 Length of Grade Scenarios................................................53
3.7 Grade Percentages Scenarios..............................................54
5.1 Mean and Standard Deviation of PCE.......................................170
5.2 Other Statistics........................................................170
5.3 Independent ttest between 2Lane and 3Lane............................171
5.4 Independent ttest between 3Lane without and with Trucks Restrictions..172
5.5 Truck Proportion Percent Difference in PCEs using Demarchi & Settis Method
173
5.6 Truck Proportion Percent Difference in PCEs using Sumner et al.s Method.... 175
5.7 Mean and Standard Deviation of PCE......................................177
5.8 Other Statistics........................................................178
5.9 Independent ttest between 2Lane and 3Lane............................179
5.10 Independent ttest between 3Lane without and with Trucks Restrictions.180
5.11 Length of Grade Percent Difference in PCEs using Demarchi & Settis Method
181
5.12 Length of Grade Percent Differences in PCEs using Sumner et al.........183
5.13 Mean and Standard Deviation of PCE.....................................185
5.14 Other Statistics.......................................................186
5.15 Independent ttest between 2Lane and 3Lane...........................187
5.16 Independent ttest between 3Lane without and with Trucks Restrictions.188
5.17 Grade Percentages Percent Difference in PCEs using Demarchi & Settis 189
5.18 Grade Percentages Percent Difference in PCEs using Sumner et al.s......190
6.1 Fuel Consumption.........................................................194
6.2 Emissions...............................................................196
7.1 PCE ranges of Truck Proportions..........................................200
x
7.2 PCE ranges of Length of Grades..................................201
7.3 PCE ranges of Grade Percentages.................................202
7.4 Summary Percent Differences of PCE for Three Lanes vs Two Lanes.202
7.5 PCE Recommendation based on Truck Proportion for 0.5 mile......204
7.6 PCE Recommendation based on Truck Proportion for 1.0 mile......206
7.7 PCE Recommendation based on Length of Grade for 5% Trucks......207
7.8 PCE Recommendation based on Length of Grade for 10% Trucks.....209
7.9 PCE Recommendation based on Grade Percentages for 5% Trucks.....210
7.10 PCE Recommendation based on Grade Percentages for 10% Trucks...212
XI
LIST OF FIGURES
Figure
2.1 GENERALIZED SPEEDFLOWDENSITY RELATIONSHIP ON
UNINTERRUPTEDFLOW FACILITIES (HCM, 2010)........................25
2.2 LOS FOR BASIC FREEWAY (HCM, 2010)............................27
2.3 THREE TYPES OF FREEWAY (HCM 2010)............................28
2.4 FHWA VEHICLE CLASSIFICATIONS (FHWA, 1985)....................30
3.1 RELATIONSHIP BETWEEN SPEED AND DENSITY (Greenshield, 1934)...33
3.2 RELATIONSHIP BETWEEN SPEED (v) AND FLOW (q) (Greenshield, 1934)...34
3.3 FUNDAMENTAL TRAFFIC FLOW DIAGRAMS CASE STUDY AT I95 SOUTH
FLORIDA (SIUHI AND MUSSA, 2007)..................................41
3.4 FLOW DENSITY RELATIONSHIP FOR BASE VEHICLES............45
3.5 FLOW DENSITY RELATIONSHIP FOR MIXED VEHICLES...........46
3.6 INTERPOLATION FLOW DENSITY RELATIONSHIP FOR qB and qM......47
3.7 FLOW DENSITY RELATIONSHIP FOR BASE VEHICLES............48
3.8 FLOW DENSITY RELATIONSHIP FOR MIXED VEHICLES...........49
3.9 FLOW DENSITY RELATIONSHIP FOR SUBJECT VEHICLES.........49
3.10 INTERPOLATION FLOW DENSITY RELATIONSHIP FOR qB, qM and qs.50
4.1 SPEEDFLOWDENSITYRELATIONSHIP FOR 0.5 Ml  5% TRUCKS...58
4.2 SPEEDFLOWDENSITYRELATIONSHIP FOR 0.5 Ml  10% TRUCKS..59
4.3 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.0 Ml  5% TRUCKS...60
4.4 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.0 Ml  10% TRUCKS..61
4.5 SPEEDFLOWDENSITYRELATIONSHIP FOR 3.0 Ml 5% TRUCKS...63
4.6 SPEEDFLOWDENSITYRELATIONSHIP FOR 4.0 Ml 5% TRUCKS...64
4.7 SPEEDFLOWDENSITYRELATIONSHIP FOR 3.0 Ml 10% TRUCKS..65
4.8 SPEEDFLOWDENSITYRELATIONSHIP FOR 4.0 Ml 10% TRUCKS..66
4.9 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.5 Ml 5% TRUCKS...68
xii
4.10 SPEEDFLOWDENSITYRELATIONSHIP FOR 2.0 Ml 5% TRUCKS.....69
4.11 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.5 Ml 10% TRUCKS....70
4.12 SPEEDFLOWDENSITYRELATIONSHIP FOR 2.0 Ml 10% TRUCKS....71
4.13 PCE VARIABILITY BY TRUCKS PROPORTION 0.5 Mi2% GRADE.....73
4.14 PCE VARIABILITY BY TRUCKS PROPORTION 0.5 Mi4% GRADE.....74
4.15 PCE VARIABILITY BY TRUCKS PROPORTION 0.5 Mi 6% GRADE...75
4.16 PCE VARIABILITY BY TRUCKS PROPORTION 0.5 Mi 8% GRADE...76
4.17 PCE VARIABILITY BY TRUCKS PROPORTION 1.0 Mi 2% GRADE...77
4.18 PCE VARIABILITY BY TRUCKS PROPORTION 1.0 Mi 4% GRADE...78
4.19 PCE VARIABILITY BY TRUCKS PROPORTION 1.0 Mi 6% GRADE...79
4.20 PCE VARIABILITY BY TRUCKS PROPORTION 1.0 Mi 8% GRADE...80
4.21 PCE VARIABILITY BY LENGTH OF GRADE 2% GRADE..............82
4.22 PCE VARIABILITY BY LENGTH OF GRADE 4% GRADE..............83
4.23 PCE VARIABILITY BY LENGTH OF GRADE 6% GRADE..............84
4.24 PCE VARIABILITY BY LENGTH OF GRADE 8% GRADE..............85
4.25 PCE VARIABILITY BY LENGTH OF GRADE 2% GRADE..............86
4.26 PCE VARIABILITY BY LENGTH OF GRADE 4% GRADE..............87
4.27 PCE VARIABILITY BY LENGTH OF GRADE 6% GRADE..............88
H OF GRADE 8% GRADE 89
PERCENTAGES  0.25 Mi 91
PERCENTAGES  0.50 Mi 92
PERCENTAGES  0.75 Mi 93
PERCENTAGES  1.0 Mi 94
PERCENTAGES  1.25 Mi 95
PERCENTAGES  1.50 Mi 96
PERCENTAGES  2.0 Mi 97
PERCENTAGES  0.25 Mi 98
PERCENTAGES  0.50 Mi 99
PERCENTAGES  0.75 Mi 100
PERCENTAGES  1.0 Mi 101
PERCENTAGES  1.25 Mi 102
PERCENTAGES  1.50 Mi 103
xiii
4.42 PCE VARIABILITY BY
4.43 PCE VARIABILITY BY
4.44 PCE VARIABILITY BY
4.45 PCE VARIABILITY BY
4.46 PCE VARIABILITY BY
4.47 PCE VARIABILITY BY
4.48 PCE VARIABILITY BY
4.49 PCE VARIABILITY BY
4.50 PCE VARIABILITY BY
4.51 PCE VARIABILITY BY
4.52 PCE VARIABILITY BY
4.53 PCE VARIABILITY BY
4.54 PCE VARIABILITY BY
4.55 PCE VARIABILITY BY
4.56 PCE VARIABILITY BY
4.57 PCE VARIABILITY BY
4.58 PCE VARIABILITY BY
4.59 PCE VARIABILITY BY
4.60 PCE VARIABILITY BY
4.61 PCE VARIABILITY BY
4.62 PCE VARIABILITY BY
4.63 PCE VARIABILITY BY
4.64 PCE VARIABILITY BY
4.65 PCE VARIABILITY BY
4.66 PCE VARIABILITY BY
4.67 PCE VARIABILITY BY
4.68 PCE VARIABILITY BY
4.69 PCE VARIABILITY BY
4.70 PCE VARIABILITY BY
4.71 PCE VARIABILITY BY
4.72 PCE VARIABILITY BY
4.73 PCE VARIABILITY BY
GRADE PERCENTAGES 
TRUCK PROPORTION0.
TRUCK PROPORTION0.
TRUCK PROPORTION0.
TRUCK PROPORTION0.
TRUCK PROPORTION 1.
TRUCK PROPORTION 1.
TRUCK PROPORTION 1.
TRUCK PROPORTION 1.
LENGTH OF GRADE 2%
LENGTH OF GRADE 4%
LENGTH OF GRADE 6%
LENGTH OF GRADE 8%
LENGTH OF GRADE 2%
LENGTH OF GRADE 4%
LENGTH OF GRADE 6%
LENGTH OF GRADE 8%
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
GRADE PERCENTAGES 
TRUCK PROPORTION0.
2.0 Mi............104
5 Mi 2% GRADE.....106
5 Mi 4% GRADE.....107
5 Mi 6% GRADE.....108
5 Mi 8% GRADE.....109
0 Mi 2% GRADE.....110
0 Mi 4% GRADE.....111
0 Mi 6% GRADE.....112
0 Mi 8% GRADE.....113
GRADE.............114
GRADE.............115
GRADE.............116
GRADE.............117
GRADE.............118
GRADE.............119
GRADE.............120
GRADE.............121
0.25 Mi...........122
0.50 Mi...........123
0.75 Mi...........124
1.0 Mi............125
1.25 Mi...........126
1.50 Mi...........127
2.0 Mi............128
0.25 Mi...........129
0.50 Mi...........130
0.75 Mi...........131
1.0 Mi............132
1.25 Mi...........133
1.50 Mi...........134
2.0 Mi............135
5 Mi 2% GRADE.....137
XIV
4.74 PCE VARIABILITY BY TRUCK PROPORTION 0.5 M
4.75 PCE VARIABILITY BY TRUCK PROPORTION 0.5 M
4.76 PCE VARIABILITY BY TRUCK PROPORTION 0.5 M
4.77 PCE VARIABILITY BY TRUCK PROPORTION 1.0 M
4.78 PCE VARIABILITY BY TRUCK PROPORTION 1.0 M
4.79 PCE VARIABILITY BY TRUCK PROPORTION 1.0 M
4.80 PCE VARIABILITY BY TRUCK PROPORTION 1.0 M
4.81 PCE VARIABILITY BY LENGTH OF GRADE
4.82 PCE VARIABILITY BY LENGTH OF GRADE
4.83 PCE VARIABILITY BY LENGTH OF GRADE
4.84 PCE VARIABILITY BY LENGTH OF GRADE
4.85 PCE VARIABILITY BY LENGTH OF GRADE
4.86 PCE VARIABILITY BY LENGTH OF GRADE
4.87 PCE VARIABILITY BY LENGTH OF GRADE
4.88 PCE VARIABILITY BY LENGTH OF GRADE
4% GRADE....138
6% GRADE....139
8% GRADE....140
2% GRADE....141
4% GRADE....142
6% GRADE....143
8% GRADE....144
2% GRADE..........145
4% GRADE..........146
6% GRADE..........147
8% GRADE..........148
2% GRADE..........149
4% GRADE..........150
6% GRADE..........151
8% GRADE...............152
4.89 PCE VARIABILITY BY GRADE PERCENTAGES 0.25 Mi..............153
4.90 PCE VARIABILITY BY GRADE PERCENTAGES 0.50 Mi..............154
4.91 PCE VARIABILITY BY GRADE PERCENTAGES 0.75 Mi..............155
4.92 PCE VARIABILITY BY GRADE PERCENTAGES 1.0 Mi...............156
4.93 PCE VARIABILITY BY GRADE PERCENTAGES 1.25 Mi..............157
4.94 PCE VARIABILITY BY GRADE PERCENTAGES 1.50 Mi..............158
4.95 PCE VARIABILITY BY GRADE PERCENTAGES 2.0 Mi...............159
4.96 PCE VARIABILITY BY GRADE PERCENTAGES 0.25 Mi..............160
4.97 PCE VARIABILITY BY GRADE PERCENTAGES 0.50 Mi..............161
4.98 PCE VARIABILITY BY GRADE PERCENTAGES 0.75 Mi..............162
4.99 PCE VARIABILITY BY GRADE PERCENTAGES 1.0 Mi...............163
4.100 PCE VARIABILITY BY GRADE PERCENTAGES 1.25 Mi.............164
4.101 PCE VARIABILITY BY GRADE PERCENTAGES 1.50 Mi.............165
Figure 4.102 PCE VARIABILITY BY GRADE PERCENTAGES 2.0 Mi.......166
xv
1.
Introduction
1.1 Background
Truck traffic causes significant problems including congestion, delay, crashes,
pollution, energy consumption, and road damage in many regions locally, nationally
and internationally. Those problems happen because trucks are larger in size and
heavier than passenger cars. In addition, trucks have limited performance, especially
on grades and curves for accelerating and decelerating.
A common treatment of truck effects on traffic flow in many traffic engineering and
transportation planning procedures (other than in simulation models) for highway
capacity and level of service (LOS) determination is to multiply the number of trucks
by a passenger car equivalent (PCE) that is based on Highway Capacity Manual
(HCM) procedures. The HCM provides the concepts and guidelines to measure the
capacity and quality of service for transportation facilities. HCM 2010 defines PCE as
the number of passenger cars that will result in the same operational conditions as a
single heavy vehicle of a particular type under specified roadway, traffic, and control
conditions. For instance, the equivalence factor of trucks for signalized intersection is
2.0. The common uses of truck PCEs in traffic engineering and transportation
analysis include 2 PCEs for each single unit truck (SUT) and 3 PCEs for each
longer combination vehicle (LCV). This PCE may be very approximate. In the HCM
1
procedures, the PCE values account for roadway grades and truck percentages. The
effect of truck traffic under uncongested conditions is expected to be small on level
terrain. As congestion increases, the truck traffic is expected to affect the traffic
stream significantly, particularly on the grades.
The deficiency of current PCE estimation is because a truck does not perform like
multiple cars. For example, an LCV may only travel at a crawl speed up a steep
incline, whereas five or ten cars travel up the same incline with faster speed.
Similarly, the blocking effect of an LCV on urban streets is likely to be much greater
than the blocking effect of multiple cars.
Therefore, there is a need to better quantify the representation of trucks in traffic
models such as HCM and other modeling procedures.
1.2 Study Objectives
The main goal of this study is to understand the dynamics of truck effects in traffic
through improved truck passenger car equivalents using a simulationbased
approach. Thus the objectives of this study are:
1. Quantify the impacts of various proportions of trucks, grade percentages, and
length of grade with different vehicle types on truck PCE using micro simulation,
considering variables beyond the scope of the current HCM 2010.
2
2. Compile and compare PCEs obtained from simulation outputs using both equal
density and equal speed formulas and HCM 2010.
3. Verify the hypotheses that:
a. When the proportion of trucks is higher in the traffic stream, the PCE
should decrease.
b. When the grade percentages increase with 5% and 10% trucks in the
traffic stream, the PCE should be increased.
c. When the length of the grade is greater than 2.0 miles with 5% and 10%
trucks in the traffic stream, the PCE should be greater.
d. When the number of lanes increases from two to three lanes, the PCE
should decrease.
4. Quantify the effects of truck lane restrictions in the left most lanes on PCE
values.
5. Propose new truck PCE estimation tables for future revisions of the Highway
Capacity Manual.
1.3 Significant Studies
The significance of this study is that the truck PCE values developed can be utilized
to improve the HCM 2010. Since heavy vehicles like trucks can have major impacts
in the traffic stream, such as slower movement, bigger size, longer headways and
3
more frequent gaps, and PCE is used as an equivalency factor. However, many
studies suggested ways to improve the current method of computing PCE:
Geistefeldt (2009) estimated PCE based on capacity variability. The study used a
simulation for a more detailed investigation of different factors affecting PCE values,
such as truck type or heavy vehicle percentage.
A study by Rakha et al. (2007) used a simulation and many variables to measure the
PCE. The simulation results demonstrated that the proportion of trucks in the traffic
stream has a significant impact on the heavyvehicle PCE at low proportions, and the
trend of decreasing PCE is especially true for longer and steeper grades.
Alkaisy et al. (2006) investigated the limitations and appropriate use of HCM PCE
factors for heavy vehicles on freeways and multilane highways. The authors found
that the effect of heavy vehicles on a traffic stream is greater during congestion than
uncongested conditions. He recommended the use of a more realistic PCE factor for
heavy vehicles considering queues and congestion analysis and strongly suggested
practical consideration that a set of PCE factors at specific grades be developed for
congested freeways and included in the capacity analysis procedures.
Demarchi and Setti (2003) studied the limitation of PCE derivation with more than
one truck type. They found that in analysis the current PCE derivations are only able
to account for the impact caused by one heavyvehicle class, and in practice there is
4
an error in the conversion of an observed mixed flow to an equivalent flow rate
expressed in PCE per hour. Thus, they suggested further studies on the derivation of
PCEs for trucks are needed.
Benekohal and Zhao (1999) studied delaybased PCE for trucks at signalized
intersections and found that the constant PCE recommended in the HCM
overestimated the impact of single unit trucks and the capacity reduction effects of
combination trucks. They suggested that further studies are needed to cover different
traffic and geometric conditions as well as other heavy vehicle types.
Elefteriadou and Webster (1997) used the FRESIM simulation model to develop the
scenarios with up to 6 percent upgrade. The values obtained in the research were
similar to HCM values on level and slight grades but significantly lower for long and
steep grades. Further research is recommended to validate the PCE values
reported.
Molina (1987) recommended further research into the development of PCE values
for large trucks at signalized intersections, particularly the effect of turning
maneuvers and grades.
Roess & Messer (1984) reviewed the various approaches for calibration and
interpretation of PCE values and recommended revision to PCE values for multilane
uninterrupted flow.
5
A study by Craus, et al. (1979) suggested further research on proposed truck
equivalency, especially for climbing lanes.
1.4 Simulation
Traffic simulation has become a vital tool for the design of robust and stochastic
complex technical systems, including for traffic engineering and transportation
planning. A simulation model is able to represent these effects more accurately
without traffic disruption with many different scenarios. Hence, the simulationbased
highway capacity concept was used to evaluate PCE value.
Simulation is defined as a numerical technique for conducting experiments on a
digital computer, which may include stochastic characteristics, be microscopic or
macroscopic in nature, and involve mathematical models that describe the behavior
of a transportation system over extended periods of real time (May, 1989).
Simulation is the representation of the actual conditions by means of mathematical or
physical approximation. This representation (in models like VISSIM, CORSIM, and
TransModeler) uses software based on mathematical models specifically for traffic
simulation. Simulation features consist of timevarying periods, human performance,
and system wide analyses including: spillback, spillover, and bottlenecks.
6
As VISSIM provides adequate analysis outputs or measures of effectiveness, for this
study VISSIM, a microscopic, behaviorbased traffic simulation program is used.
1.5 Organization of the Dissertation
The organization of the dissertation steps sequentially beginning with a historical
look of PCE and concluding with the recommendation of a new truck PCE.
Chapter 2 contains a review of PCE studies organized based on the methods used,
traffic flow theory, vehicle classification, and simulation.
Chapter 3 describes the theoretical analysis development of truck equivalencies from
Greenshields to Demarchi Setti methods, Federal Highway Administration
(FHWA)s Decision Support Methodology (DSM), and steps performed to estimate
truck equivalence factor. It also describes the VISSIM traffic microsimulation,
especially the Wiedeman 1999 method of psychophysical driving behavior as well
as lane changing parameters.
Chapter 4 presents the speedflowdensity relationship analysis from VISSIM
simulation output. Truck PCE values using Demarchi and Settis and Sumners
methods based on truck proportion, length of grade, grade percentages, number of
lane and truck restrictions.
7
Chapter 5 describes the statistical tests used to evaluate PCE computed based on
both equal flow density and equal speed density approached. The tests included are
test of normality, independence ttest, and percent error using the mean, variance,
and standard deviation from the groups of PCE results. SPSS software is used to do
the statistical analysis in this study.
Chapter 6 is an additional analysis and research of literature on air quality, including
fuel consumptions and emissions, to understand the effect of trucks traffic on air
quality. The assumptions and formula for fuel consumption and emission is based on
Synchro simtraffic software analysis.
Chapter 7 describes the conclusions and recommendations drawn from the results of
this study.
8
2.
Literature Review
2.1 Review of Truck Equivalencies
Truck equivalencies were first mentioned in the 1950 Highway Capacity Manual
(HCM), which stated that trucks on twolane highways on level terrain have the same
effect as two passenger cars (PC). This estimate was based on the number of
passenger cars passing trucks compared to the number of passenger cars passing
passenger cars.
The second edition of the HCM was published in 1965, and passenger car equivalent
(PCE) was introduced. PCE was defined as the number of passenger cars
displaced in the traffic flow by a truck or bus, under the prevailing roadway and traffic
conditions (HCM, 1965). The current definition of PCE in the HCM 2010 is similar,
the number of passenger cars that will result in the same operational conditions as
a single heavy vehicle of a particular type under specified roadway, traffic, and
control conditions (HCM, 2010).
The calculation approach used separated speed distribution for twolane highways
and relative delay, due to trucks on various grades for multilane highways, and was
based on Walkers Method in the 1965 HCM. Many studies have been carried out
since then. Those studies have examined PCE based on field data, simulations or
combination field data and simulations.
9
According to the HCM 2010, the heavy vehicle adjustment factor is found as:
1
fm = 1 + pt.(et 1) + pr.(er 1) (2'
Where
fHV = heavy vehicle adjustment factor
PT = truck proportion
PR = recreational vehicles (RVs) proportion
Er = truck PCE
Et = recreational vehicles (RVs) PCE
2.1.1 PCEs Based on Flow Rates and Density
Measuring the traffic flow is critical on the roadway whether the road is congested or
not. In transportation engineering, traffic flow rate is the term used to indicate the
equivalent hourly rate of vehicles passing a point per unit of time. Many studies of
truck equivalencies have developed using flow rates as the basis of estimation. PCE
is computed based on mixed vehicle flow, percentage of grade, and truck volume to
capacity ratio (John and Glauz, 1976):
et =
Rb Rm (1 Pt)
qM x PT
(2.2)
Where
10
qB= equivalent passenger car only flow rate for a given v/c ratio
qM= mixed flow rate
PT = truck proportion in the mixed traffic flow
Et = truck PCE
Hubers model estimated PCEvalues for vehicle under freeflowing, multilane
conditions by considering the relationship between some measure of impedance
along a length of roadway and the flow rate along the same roadway for two different
traffic streams. Sumner et al. (1982) further developed Hubers method by including
more than one truck type in the traffic stream.
Hubers basic equation:
Where
qB= equivalent passenger car only flow rate
qM = mixed flow rate
qs = additional subject flow rate
PT= truck proportion in the mixed traffic flow
(2.3)
Sumner formula:
(2.4)
11
VP= proportion of subject vehicles
Et = truck PCE
Demarchi and Setti (2003) suggested the PCEs formula to eliminate the possible
error for mixed heavy vehicles in the traffic stream, including interaction between
multiple trucks types:
Where
Pt= proportion of trucks of type i out of all trucks n in the mixed traffic flow
qB = base flow rate (passenger cars only)
qM = mixed flow rate
Et = passenger equivalent of trucks
2.1.2 PCEs Based on Headways
Headway is defined as the time in seconds between two successive vehicles as they
pass a point of the roadway. It can be measured either from front axle or front
bumper (HCM, 2010). Understanding the nature of headway from truck movement
on the roadway is another approach to calculate truck equivalencies. Many
researchers have used headway as the basis of estimation.
Werner and Morrall (1976) recommended determining PCE using headways when
the roadway is sufficiently congested on level terrain:
(2.5)
12
(2.6)
Where
Hm= average headway for all vehicles
Hm= average headway for passenger car
PT= truck proportion
Pc = passenger car proportion
Et = truck PCE
Using the spatial headway methodology, Seguin, et.al.(1982) formulated PCE as the
ratio of average headway for vehicle types: average truck headway divided by the
average headway for passenger cars:
HLJ
PCEij = (2.7)
pcj
Where
PCEij = the PCE of vehicle Type /'under Conditions j
Hij= average headway for vehicle Type /
Hpcj = the average headway for passenger car for Conditions j.
13
Cunagin and Chang (1982) determined the effect of the presence of heavy trucks on
freeway traffic streams using time headway based on headway type, lane width, and
traffic volume as shown on Equation 2.8. They conclude that the presence of trucks
in the traffic stream is accompanied by an increase in the mean headway. The
lagging headway is measured from the rear bumper of the lead vehicle to the rear
bumper of the following vehicle.
Hu
Et = 7T (28)
Hb
Where
Htj = the mean lagging headway of vehicle type /'under conditions j
Hb = the mean lagging headway of passenger cars.
Et = truck PCE
Krammes and Crowley (1986) suggested that:
Et = [(1 Pt).Htp + p. Htt ] /Hp (2.9)
Where
PT = the proportion of trucks
Htp = the lagging headway of trucks following passenger cars
H = the lagging headway of trucks following trucks
HP = the lagging headway of cars following either vehicle type
Et= truck PCE
14
Anwar et al. (2011) presented a statistical approach for determining the PCE values
for singleunit trucks and combination trucks using the concept of spatial lagging
headways, the distance from the rear bumper of leading vehicle to the rear bumper
of the following vehicle, measured from real traffic data.
2.1.3 PCEs Based on Queue Discharge Flow
A queue is an accumulation of vehicles upstream of a bottleneck when the vehicles
are moving slowly or standing still. It occurs when the demand of vehicles passes
exceeds available capacity. AlKaisy et al. (2002) quantified the calculation of PCE
using queue discharge flow (QDF) based on the assumption that QDF capacity
observation can be expected to show minimal variation if the traffic stream is uniform
and consists of passenger cars only. They found that the effect of heavy vehicles on
a freeway is greater when it is operating in oversaturated conditions. In addition, it
was found that PCE both during dry or rainy days and during the presence of
roadside maintenance work are not significantly different.
2.1.4 PCEs Based on Speed
Another measure used by many researchers in their studies of PCE is the rate of
motion of vehicles in a distance per unit of time or speed. Van Aerde and Yagar
(1983) developed a methodology to estimate PCE based on the relative rates of
speed for each type of vehicle traveling in the main direction and for all vehicles
combined traveling in the opposing direction. They found that PCE decreases for
15
higher speed percentiles. The speed analysis using the linear regression model
structure is:
Percentile speed = free speed + Cx (number of passenger cars)
+ C2 (number of passenger trucks) + C3 (number of RVs)
+ C4 (number of other vehicles)
+ C5 (number of opposing vehicles) (2.10)
Where
cx to c5 = the coefficients of speed reductions for each vehicle type.
Using the speed reduction coefficients, the PCE for a vehicle type n is calculated as
Where
Cn= speed reduction coefficient for vehicle type n
C1= speed reduction coefficient for passenger cars
En = truck PCE
2.1.5 PCEs Based on Delays
The HCM denotes delay as the additional travel time experienced by a driver,
passenger, or pedestrian (HCM, 2010). The PCE values were determined by using
Walker spatialheadway and equivalentdelay methods. A basic assumption in the
16
Walker method is that faster vehicles are not hindered in passing as they overtake
slower vehicles, so queues do not form. In contrast, in the equivalentdelay method,
it assumed that faster vehicles are always hindered by slower vehicles, such that
queues form. Using that premise, PCE values calculated using a linear combination
of the Walker and equivalentdelay in each intermediate volume level yields:
QOTi/VOLj) [1/SPM\ [1/SPB ])
{OTlpc /VOLlpc) [1 /SPPC] [1/SPB ])
Where
OTi = the number of overtakings of vehicle type /'by passenger cars
VOL = the volume of vehicle type /
/
07 = the number of overtakings of lower performance passenger cars by
passenger cars
VOL = the volume of lower performance passenger cars
SP = the mean speed of the mixed traffic stream
M
SPp= the mean speed of the base traffic stream with only high performance
passenger cars
SPpc = the mean speed of the traffic stream with only passenger cars
Et= truck PCE
17
Craus et al. (1979) developed a passenger car equivalent for trucks ratio of delay
time caused by one truck to the delay time caused by one passenger car. This
method takes the oppositelane traffic into consideration. The following equation
reflects the actual disturbance and delay caused by trucks to other traffic:
Where
E = truck PCE
dkt= average delay time caused by one truck
dkp= average delay time caused by one passenger car
Cunagin and Messer (1983) developed PCE estimation based on speed distribution,
traffic volumes, and vehicle types. Their method estimates PCEs using the ratio of
delay experienced by a passenger car due to nonpassenger vehicles to the delay
experienced by a passenger car due to other passenger cars:
(2.11)
(2.12)
Where
Et= PCE of vehicle Type /'under Conditions j
Dij= delay to passenger cars due to vehicle Type /'under Conditions j
18
Dbase= delay to standard passenger cars due to slower passenger cars
2.1.6 PCEs Based on V/C Ratio
Fan (1989) studied PCE for expressways in Singapore using volumetocapacity
(V/C) ratio instead of density or level of service because these freeways operate at
LOS E. The study focused on congested flow conditions or V/C ratio above 0.67 and
mentioned that it is unnecessary to calculate PCEs at uncongested flow conditions.
Using multiple linear regressions by multiplying the observed flow by the V/C ratios,
he found that commercial vehicles such as light and heavy trucks, buses, and trailers
generally have higher PCE values compared with the PCEs used in US and UK for
the level terrain.
2.1.7 PCEs Based on VehicleHours
Hourly traffic volumes are used for determining the length and magnitude of peak
periods, evaluating capacity, and assessing geometric design and traffic control.
Sumner et al. (1984) determined a method of calculating PCE values between
consecutive signalized intersections on urban arterial roads using microscopic
simulation, NETSIM. The values are derived from the vehiclehours of road utilization
that are added when large vehicles are introduced to the traffic stream. The study
concluded that PCE is lower for better levels of service, specifically PCE values at
LOS B are less than the PCE values at LOS D.
19
2.1.8
PCEs Based on Platoon Formation
Platooning occurs when the fast vehicles catch the slower vehicles such trucks,
buses and recreational vehicles, and the fast vehicles are not able to pass. This
often occurs on rural twolane highways or on upgrade multilane highways where
high traffic flow make lane changing and overtaking difficult. Van Aerder and Yagar
(1983) studied PCE in both platoon leadership and follower creation. Large vehicles
have a tendency to be platoon leaders. They analyzed large vehicles using the ratio
of percent leads, by vehicle type, and to percentage of total mainline traffic to obtain
PCE by normalizing those ratios to the original ratio of passenger cars.
In the study, there were relative effects of trucks in the creation of platoon followers.
The follower production rates were higher in the highvolume area for all types of
vehicles and both directions of travel, which results in smaller PCE. Here is the
model of number of followers using separate multiple linear terms:
Number of followers = B0+ B1 (cars) + B2 (trucks) + B3 (RVs) + B4 (other vehicles) +
B5 (opposing vehicles, mainline vehicles) (2.13)
Where
B0 = the constant number of followers in a platoon on twolane highways,
20
Bi to B5 = indicate the rate at which the number of followers increases for each traffic
volume component
B] to B4 = the number of additional followers produced per vehicle.
B5 = the number of followers produced per opposing vehicle at a mainline volume of
1.000 vph.
Thus,
PCEs= Bn/ Bi
(2.14)
Where
PCEs = Passenger Car Equivalent
Bn = the constant number of followers in a platoon on twolane highways,
Bi to B5 = indicate the rate at which the number of followers increases for each traffic
volume component
6? to B4 = the number of additional followers produced per vehicle.
B5 = the number of followers produced per opposing vehicle at a mainline volume of
1.000 vph.
2.1.9 PCEs Based on Travel Time
Travel time is the total time spent by vehicles from one point to another point under
prevailing conditions. It includes acceleration, deceleration, and stopped time. Keller
and Saklas (1984) utilized a macroscopic traffic simulation, TRANSYT/7N for
estimating PCE for large vehicles operating on urban arterial networks as a function
21
of traffic volume, vehicle classification, and signal setting. The premise was that the
capacity reducing effect of the larger vehicles is related directly to the additional
delay from such vehicles when compared to the all passenger cars. The outcomes
showed that PCE values increase as vehicle get larger and as signalization
approaches the maximum. The PCE values were estimated over a wide range of
flow rates to approximately simulate levels of service from A to F. In addition, the
PCE values were relatively constant for most of LOSs until the volume approach
LOS F at which the PCE values significantly increase.
2.2 Truck Equivalencies in the 2010 HCM
In the 2010 HCM, PCEs are given based on the percent and length of grade and
proportion of heavy vehicles, and terrain, as shown on Tables 2.1,2.2, and 2.3. The
extended freeway segment includes specific upgrades and downgrades.
Table 2.1 PCEs for Heavy Vehicles in General Terrain Segments (HCM, 2010)
Vehicle Level PCE bv TvDe of Terrain Rolling Mountainous
Trucks and buses, ET 1.5 2.5 4.5
RVs, Â£ 1.2 2.0 4.0
22
Table 2.2 PCEs for Trucks and Buses (Â£r) on Specific Downgrades (HCM, 2010)
Downgrade Length of Prooortion of Trucks and Buses
(%) Grade (mi) 5% 10% 15% >20%
<4 All 1.5 1.5 1.5 1.5
45 <4 1.5 1.5 1.5 1.5
>4 2.0 2.0 2.0 1.5
>56 <4 1.5 1.5 1.5 1.5
>4 5.5 4.0 4.0 3.0
>6 <4 1.5 1.5 1.5 1.5
>4 7.5 6.0 5.5 4.5
Table 2.3 PCEs for Trucks and Buses (ET) on Upgrades (HCM, 2010)
Upgrade
(/>
<2
>23
>34
>45
>56
>6
Length (mi) 2% 4% Prooortion of Trucks and Buses 5% 6% 8% 10% 15% 20% >25%
All 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
0.000.25 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
>0.250.50 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
>0.500.75 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
>0.751.00 2.0 2.0 2.0 2.0 1.5 1.5 1.5 1.5 1.5
>1.001.50 2.5 2.5 2.5 2.5 2.0 2.0 2.0 2.0 2.0
>1.50 3.0 3.0 2.5 2.5 2.0 2.0 2.0 2.0 2.0
0.000.25 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
>0.250.50 2.0 2.0 2.0 2.0 2.0 2.0 1.5 1.5 1.5
>0.500.75 2.5 2.5 2.0 2.0 2.0 2.0 2.0 2.0 2.0
>0.751.00 3.0 3.0 2.5 2.5 2.5 2.5 2.0 2.0 2.0
>1.001.50 3.5 3.5 3.0 3.0 3.0 3.0 2.5 2.5 2.5
>1.50 4.0 3.5 3.0 3.0 3.0 3.0 2.5 2.5 2.5
0.000.25 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
>0.250.50 3.0 2.5 2.5 2.5 2.0 2.0 2.0 2.0 2.0
>0.500.75 3.5 3.0 3.0 3.0 2.5 2.5 2.5 2.5 2.5
>0.751.00 4.0 3.5 3.5 3.5 3.0 3.0 3.0 3.0 3.0
>1.00 5.0 4.0 4.0 4.0 3.5 3.5 3.0 3.0 3.0
0.000.25 2.0 2.0 1.5 1.5 1.5 1.5 1.5 1.5 1.5
>0.250.30 4.0 3.0 2.5 2.5 2.0 2.0 2.0 2.0 2.0
>0.300.50 4.5 4.0 3.5 3.0 2.5 2.5 2.5 2.5 2.5
>0.500.75 5.0 4.5 4.0 3.5 3.0 3.0 3.0 3.0 3.0
>0.751.00 5.5 5.0 4.5 4.0 3.0 3.0 3.0 3.0 3.0
>1.00 6.0 5.0 5.0 4.5 3.5 3.5 3.5 3.5 3.5
0.000.25 4.0 3.0 2.5 2.5 2.5 2.5 2.0 2.0 1.0
>0.250.30 4.5 4.0 3.5 3.5 3.5 3.0 2.5 2.5 2.5
>0.300.50 5.0 4.5 4.0 4.0 3.5 3.0 2.5 2.5 2.5
>0.500.75 5.5 5.0 4.5 4.5 4.0 3.5 3.0 3.0 3.0
>0.751.00 6.0 5.5 5.0 5.0 4.5 4.0 3.5 3.5 3.5
>1.00 7.0 6.0 5.5 5.5 5.0 4.5 4.0 4.0 4.0
Note: Interpolation for percentage of trucks and buses is recommended to the nearest 0.1.
23
2.3 Traffic Flow Parameter Concepts
The Highway Capacity Manual 2010 defines basic concepts for uninterruptedflow
facilities as in volume, flow rate, speed, density, headway, and capacity. The
definitions are following:
Volume the total number of vehicles or other roadway users that pass over
a given point or section of a lane or roadway during a given time interval,
often 1h.
Flow rate the equivalent hourly rate at which vehicles or other roadway
users pass over a given point or section of a lane or roadway during a given
time interval of less than 1 hour, usually 15 minutes.
Speed a rate of motion expressed as distance per unit of time.
Density the number of vehicles occupying a given length of a lane or
roadway at a particular instant.
Headway the time between successive vehicles as they pass a point on the
roadway, measured from the same common feature of both vehicles (for
example, the front axle or the front bumper).
Figure 2.1 shows the relationship among density, speed and flow rate, and
headway and spacing. The flowdensity graph is placed directly below the speed
density because of the common horizontal scales. Similarly, the speedflow
graph is place directly next to the speeddensity graph because of the common
vertical scales. The figure illustrates that a zero flow rate occurs under two
different conditions: stable and unstable flows. Stable condition or
undersaturated flow is when there are no vehicles on the segment, the density
and flow rate is zero. Unstable condition or oversaturated flow is when density
becomes very high, all vehicles must stop, no movement can occur, and speed
24
declines due to the vehicle interaction. When capacity is reached, the density
and speed is in the maximum flow rate.
Density (veh/mi/In)
Flow (veh/h/ln)
>
5
o
tL
LEGEND
Undersaturatcd flow
Oversaturated flow
Density (veh/mi/ln)
Figure 2.1 GENERALIZED SPEEDFLOWDENSITY RELATIONSHIP ON
UNINTERRUPTEDFLOW FACILITIES (HCM, 2010)
Truck acceleration speed especially on the ramps and on the steep grades, affects
the passenger car equivalent. As speed decreases the truck PCE increases. Table
2.4 shows the relationship between speed and stopping distance for heavy trucks.
25
Table 2.4 SpeedDistance Relationship for Acceleration of Heavy Truck (AASHTO 
A Policy on Geometric Design of Highways and Streets, 2000)
3% Grade 5 % Grade
Distance (feet)' Speed (mph) 500 27 1.000 29 1,500 31 Distance Speed 500 22 1,000 24 1,500 25
). Distance from a stop condition
Bassok et al. (2009) concluded that (regardless of the method) truck speeds are
consistently slower than passenger vehicle speeds, though the size difference is
dependent on the chosen methodology, specific facility, time of day and congested
direction. Overall, they found that heavy vehicle speed is ten percent slower than
passenger car speed on freeways.
2.4 Level of Service (LOS)
LOS is commonly used to measure traffic operation of roadway network, including
freeways, ramps, surface streets and intersections. It is a grading system based on
speed, delay, and travel time. LOS is defined into 6 categories ranging from A,
ideal conditions to F, extreme delays, as shown on Figure 2.2.
26
Figure 2.2 LOS FOR BASIC FREEWAY (HCM, 2010)
The HCM 2010 describes in more detail the types of traffic flow on basic freeway
segments, which for undersaturated, queue discharge, and oversaturated flow
conditions as shown in Figure 2.3:
Undersaturated flow represents conditions under which the traffic stream is
unaffected by upstream or downstream bottlenecks.
Queue discharge flow represents traffic flow that has just passed through a
bottleneck and is accelerating back to drivers desired speeds for the
prevailing conditions. As long as another downstream bottleneck does not
exist, queue discharge flow is relatively stable until the queue is fully
discharged.
Oversaturated flow represents the conditions within a queue that has backed
up from a downstream bottleneck. These flow conditions do not reflect the
prevailing conditions of the site itself, but rather the consequences of a
downstream problem. All oversaturated flow is considered to be congested.
27
UNDEttSATURATID FLOW
80
70
~60
C
> 50
E
Â£30
*20
10
0
V t j __________ . r uT^r" ,
it's 
___ . /
~~
queue DISCHARGE FlOW
.1/
OVEnSATURATED FLOW
0 500 1,000 1,500 2,000
Flow Rate (veh/h/ln)
Note: 1405, Los Angeles, Calif,
Source: California Department of Transportation, 2008,
2,500
Figure 2.3 THREE TYPES OF FREEWAY (HCM 2010)
2.5 Vehicle Classification
Federal Highway Administration (FHWA) divides vehicles into 15 categories.
Therefore, classification data are necessary to define in the beginning of studies,
since they would be useful for predicting the commodity flow, the loads, and freight
movements as shown on Figure 2.4. FHWA vehicle classes with definitions are
following:
1. Motorcycles All two or threewheeled motorized vehicles. Typical vehicles
in this category have saddle type seats and are steered by handlebars rather
than steering wheels. This category includes motorcycles, motor scooters,
mopeds, motorpowered bicycles, and threewheel motorcycles.
2. Passenger Cars All sedans, coupes, and station wagons manufactured
primarily for the purpose of carrying passengers and including those
passenger cars pulling recreational or other light trailers.
3. Other TwoAxle, FourTire Single Unit Vehicles All twoaxle, fourtire,
vehicles, other than passenger cars. Included in this classification are
pickups, panels, vans, and other vehicles such as campers, motor homes,
ambulances, hearses, carryalls, and minibuses. Other twoaxle, fourtire
singleunit vehicles pulling recreational or other light trailers are included in
this classification.
28
4. Buses All vehicles manufactured as traditional passengercarrying buses
with two axles and six tires or three or more axles. This category includes
only traditional buses (including school buses) functioning as passenger
carrying vehicles.
5. TwoAxle, SixTire, SingleUnit Trucks All vehicles on a single frame
including trucks, camping and recreational vehicles, motor homes, etc., with
two axles and dual rear wheels.
6. ThreeAxle SingleUnit Trucks All vehicles on a single frame including
trucks, camping and recreational vehicles, motor homes, etc., with three
axles.
7. Four or More Axle SingleUnit Trucks All trucks on a single frame with
four or more axles.
8. Four or Fewer Axle SingleTrailer Trucks All vehicles with four or fewer
axles consisting of two units, one of which is a tractor or straight truck power
unit.
9. FiveAxle SingleTrailer Trucks All fiveaxle vehicles consisting of two
units, one of which is a tractor or straight truck power unit.
10. Six or More Axle SingleTrailer Trucks All vehicles with six or more axles
consisting of two units, one of which is a tractor or straight truck power unit.
11. Five or fewer Axle MultiTrailer Trucks All vehicles with five or fewer
axles consisting of three or more units, one of which is a tractor or straight
truck power unit.
12. SixAxle MultiTrailer Trucks All sixaxle vehicles consisting of three or
more units, one of which is a tractor or straight truck power unit.
13. Seven or More Axle MultiTrailer Trucks All vehicles with seven or more
axles consisting of three or more units, one of which is a tractor or straight
truck power unit.
In addition, FHWA provide the following criteria for reporting information on trucks:
1. Truck tractor units traveling without a trailer will be considered singleunit
trucks.
2. A truck tractor unit pulling other such units in a "saddle mount" configuration
will be considered one singleunit truck and will be defined only by the axles
on the pulling unit.
3. Vehicles are defined by the number of axles in contact with the road.
Therefore, "floating" axles are counted only when in the down position.
4. The term "trailer" includes both semi and full trailers.
29
a. ass
GROUP
FHWA VEHICLE CLASSIFICATION
0ESCRPTCXN
NO. Of AXUES
MOTORCYCLES
ALL CARS CARS
CARSW.' 1 AXLE TRALER
CARS w: 2AXLE TRAlER
PCX UPS 4 VANS
1 4 2 AXLE TRAUERS
2.3,44
YTTTT
BUSES
243
2.AXLE. SiNGLE JN (T
3 AXLE, SNGLE UNIT
4.axle,single unit
Â£ ^3

2 AXLE. tractor.
1.AXLE TRALER (241)
2 AXLE, TRACTOR.
2 AXLE TRALER <242
3 AXLE. TRACTOR.
1 AXLE TRAILER <341}
tn
o
Z)
DU
I
Â£
X
w
3. AXLE. TRACTOR.
2AXLE TRAILER .342}
3. AXLE. TRUCK
W,' 2AXLE TRALER
TRACTOR W: SINGLE TRALER
5
Â£
447
5.AXLE MJLTITRALER
6AXLE MJLTfTRALER
ANY 7 OR MORE AXLE
6
7 ormoro
14
NOT USED
15
UNKNOWN VEJllCLE TYPE
Figure 2.4 FHWA VEHICLE CLASSIFICATIONS (FHWA, 1985)
30
Another major consideration is the operating characteristics of different vehicle types
and dimensions, turning radii and offtracking, resistance to motion, power
requirements, acceleration performance, and deceleration performance. Motor
vehicles, including passenger cars, trucks, vans, buses, recreational vehicles, and
motorcycles, have unique weight, length, size, and operational characteristics as
shown on Table 2.5.
Table 2.5 Vehicle Weight and Power (Traffic Engineering Handbook, 2009)
Motor Vehicles Empty Weight with Driver (lb) Nominal Power (hp) WeighttoPower Ratio (Ib/hp)
Passenger car 3,400 105 32.4
Large pickup truck 4,200 175 24.0
Twoaxle, sixtire truck 10,000 175 57.1
Tractorsemitrailer 25,000 325 76.9
2.6 Simulation Models Study
Since various microscopic models exist for the simulation of traffic flow, several
micro simulation software packages were reviewed for feasibility in calculating PCEs.
Bloomberg and Dale (2000) observed that overall CORSIM and VISSIM are more
similar than they are different. Both models are designed to model any combination
of surface street and freeway facilities, including most signal control and other
operational strategies. Both models provide detailed and focused output, both in
tabular format and via animated graphics. The main differences between the two
models are in vehicle and driver behavior, primarily in the carfollowing and gap
acceptance logic.
31
3. Methodology
3.1 PCE Methods Theoretical Analysis Evaluation
Among all the truck equivalencies approaches described on the literature reviews,
the speedflowdensity method is used for this study because:
Speedflowdensity is able to measure traffic flow and the randomness in the
traffic stream, and its validity has been proven by empirical research with field
observation.
Speed is a performance measure experienced directly by drivers on the
roadway and provides vivid pictures of traffic flow.
Density is a critical parameter for uninterruptedflow facilities and
characterizes the quality of traffic operations. It describes the proximity of
vehicles to one another and reflects the freedom to maneuver within the
traffic stream (HCM, 2010).
Flow rate represents the number of vehicles that pass a certain section per
time unit, commonly in vehicle per hour.
The basic concept of this PCE derivation was proposed by Huber using a
deterministic model of traffic flow from Greenshields Model. Greenshield developed
an uninterrupted traffic flow model based on the relationship of speed, flow and
density. As shown on Figure 3.1, the relationship between speed and density are
32
relatively linear. When density is zero, the flow is zero, and speed approaches free
flow speed, since there is no traffic on the roadway as in Figure 3.2. When density
increases, the flow increases as well until it reaches some maximum flow conditions,
called the jam density, which occurs when all vehicles must stop and the speed and
flow rate becomes zero. Sumner et al. developed the relationship described by
Huber which includes multiple truck types, followed by Demarchi and Setti. The
limitation of deriving PCE in the traffic stream with multiple truck types has found in
Demarchi and Settis study. To avoid the limitation of not counting the interaction
between trucks, Demarchi and Setti developed a new derivation of PCE.
Figure 3.1 RELATIONSHIP BETWEEN SPEED AND DENSITY (Greenshield, 1934)
33
Figure 3.2 RELATIONSHIP BETWEEN SPEED (v) AND FLOW (q) (Greenshield,
1934)
3.2 Simulation
3.2.1 Decision Support Methodology (DSM)
Federal Highway Administration (FHWA) produced a manual and worksheet that can
be used to select and identify the appropriate type traffic analysis tool for operational
improvements. Decision Support Methodology (DSM) was applied into this study to
select the correct measurable approach for the freight movement traffic on freeway.
Next is the following DSM process for this particular study:
34
Step 1: Analysis Context
Step 2: Geographic Scope
Step 3: Facility Type
Step 4: Travel Mode
Step 5: Management Strategy/Application
Step 6: Traveler Response
Step 7: Performance Measure
Step 8: Tool/Cost Effectiveness
Step 9: Criteria Weights
Among the 7 tools in DSM, the two appropriate tools for this study are either Micro
simulation or a Travel Demand Model (TDM) as shown on Table 3.1. Travel Demand
could not apply for this analysis because the objective of this study for understanding
the dynamic truck traffic in traffic operational analysis, not in dynamic assignment
equilibrium of origin and destination trip planning pattern. The most appropriate tool
is Microsimulation since it is based on driver carfollowing and lanechanging
behavior. Vissim micro simulation is the tool used in this study.
35
Table 3.1 Decision Support Methodology (DSM) Result
Criftiia Weights "5" i "i
eqjatftt/Critcria (0 wt Titutnl, 5 = mmt rvliranll Cirtarii IUI*711LC4 >VlrJitdSuH5tilt Column $ x Cflwjm 7
Skrtcii PEml tdm Aulilical 4HCM} Traffic Opt Xlarro Sim Me so Sim XllCTD Sim Sketch Plan TDM Analytical {EÂ£CM> Traffic Opt Mirra Sim Meu Sinn Micro Sim
0 5 25 0 54 50 n 50 HI 125 0 254 254 250 258 250
1 5 ma to ?y 15 Â£5 23 21 K5 so IB 73 142 m M2
2 Fwi%Tflw 5 M 24 If 2? 23 a 72 \n St fl 124 134 tei
S 7f4V*IM#fe 55 25 S* 3* a 126 1 SC Cl 126 104 m
1 5 IS W a 4 17 17 18 01 53 38 38 41 81 31
5 t r Tr*vl*rI>^riM PwfwirierFW Metiuei TcdrCotf Dlf4i*rtr*t55 5 5 5 83 Â£0 33 ca 25 i r 4W K w M 110 S if E4 a 13 a a a JW HD 1C7 125 w 2475 101 4125 (52 142 744 Ci P5 EG WO 34 14C IK ido
VEIQHTED TOTALS Mui ImI riMjerki: SIS 838 F7Z4 111 332 7B3 1238
1. HtefO 5=i(Vi
2. TDM
The DSM results tool uses the following FHWA definition:
Travel Demand Models
Predicting travel demand, traffic operations, and impacts in response to
operational strategies requires specific analytical capabilities, such as the
prediction of travel demand and the consideration of destination choice, mode
choice, timeofday travel choice, and route choice, as well as the
representation of traffic flow in the highway network. These attributes are
found in the structure and orientation of travel demand models, mathematical
models that forecast future travel demand from current conditions, and future
projections of household and employment characteristics. Travel demand
models were originally developed to determine the benefits and impacts of
major highway improvements in metropolitan areas. Today, travel demand
models are used in more wideranging tasks, including development of
transportation master plans, evaluation of proposed landuse changes, initial
design of transportation facilities, evaluation of air quality impacts, and
assessment of future transportation needs. However, these tools were not
designed to evaluate travel management strategies, such as ITS and
operational strategies. Travel demand models have only limited capabilities
to accurately estimate changes in operational characteristics (such as speed,
delay, and queuing) resulting from implementation of ITS/operational
strategies. These inadequacies generally occur because of the poor
representation of the dynamic nature of traffic in travel demand models.
Microscopic simulation
Microscopic simulation models simulate the movement of individual vehicles,
based on theories of carfollowing and lanechanging. Typically, vehicles
enter a transportation network using a statistical distribution of arrivals (a
stochastic process), and are tracked through the network over small time
intervals (e.g., one second or fraction of a second). Typically, upon entry,
each vehicle is assigned a destination, a vehicle type, and a driver type. In
many microscopic simulation models, the traffic operational characteristics of
each vehicle are influenced by vertical grade, horizontal curvature, and
superelevation, based on relationships developed in prior research. The
primary means of calibrating and validating microscopic simulation models is
through the adjustment of driver sensitivity factors. Computer time and
storage requirements for microscopic models are large, usually limiting the
network size and the number of simulation runs that could be completed.
37
3.2.2
VISSIM Micro Simulation
VISSIM model uses inputs such as lane assignments and geometries, intersection
turning movement volumes, vehicle speeds, percentages of vehicles by type, and
pretimed and/or actuated signal timing. There are two approaches to modeling the
traffic in VISSIM: static route and dynamic assignment. The basic concept for the
static route approach is Car Following Logic or named Wiedemann 1974 for urban
arterials, which assumes that the driver of a faster moving vehicle starts to
decelerate as the driver reaches his individual perception threshold of a slower
moving vehicle. Since the driver cannot exactly determine the speed of that vehicle,
his speed will fall below that vehicles speed until he starts to slightly accelerate
again after reaching another perception threshold (PTV AG manual 2008, page 26).
Another approach is Wiedemann 1999 for freeways. The car following and lane
change logic is similar to urban arterial, and includes vehicle headways. The
dynamic assignment model in VISSIM chooses routes based on dynamic factors,
such as the impacts of variable message signs or the potential traffic diversion into
neighborhoods for networks up to the size of medium sized cities. When using
dynamic assignment travel demand is not specified by using vehicle input on
selected links with a given volume but in the form of an origindestination matrix.
Traffic starting at a parking lot is similar to traffic generated by vehicle inputs, but the
composition of the traffic explicitly is specified for the parking lot.
The approach used in VISSIM is based on the Wiedemann model of psychophysical
driving behavior:
38
Free driving is when no influence of preceding is vehicles observable.
 Approaching is the process of adapting the drivers own speed to the
lower speed of a preceding vehicle.
Following is when the driver follows the preceding car without any
conscious acceleration or deceleration.
Braking is the application of medium to high deceleration rates if the
distance falls below the desired safety distance.
For the purposes of this study, the Wiedemann 1999 method for interurban or
freeway traffic is used, such that:
CCO (Standstill distance) defines the desired distance between stopped cars
with no variation.
CC1 (Headway time) is the time in seconds that a driver wants to keep.
dxjafe CCO + CC1 *v.
Where:
v = given speed
dx safe = safety distance or clear space desired by the driver (ft/ses)
CC2 (Following variation) is how much more distance than the desired safety
distance a driver allows before he intentionally moves closer to the car in
front.
CC3 (Threshold for entering Following) defines how many seconds before
reaching the safety distance the driver starts to decelerate.
CC4 and CC5 (Following thresholds) control the speed differences during
the Following state.
CC6 (Speed dependency of oscillation) influences the distance on speed
oscillation while in following process.
39
CC7 (Oscillation acceleration) is actual acceleration during the oscillation
process.
CC8 (Standstill acceleration) is desired acceleration when starting from
standstill
CC9 (Acceleration) is desired acceleration.
Lookback Distance is the distance that a vehicle can see backwards in order
to react to other vehicles behind (within the same link).
Similar to Siuhi and Mussa (2007) findings the VISSIM default parameters used in
trial simulation runs for checking any coding error for their study, showed that the
default model parameters were incorrectly emulating the existing traffic flow
characteristics. The necessitated calibrating the model by tuning carfollowing and
lanechanging parameters while comparing simulation data with field data. The fine
tuning process involved iterative parameter changing and simulating until simulated
speeds closely matched speeds observed in the fields. Furthermore, the simulated
values were verified against the observed field values as indicated in the
fundamental flow diagrams shown in Figure 3.3.
40
Density (vpmpl)
Figure 3.3 FUNDAMENTAL TRAFFIC FLOW DIAGRAMS CASE STUDY AT 195
SOUTH FLORIDA (SIUHI AND MUSSA, 2007)
Due to unrealistic vehicle behavior in comparison to the field observation when using
VISSIM default parameters, several cars following and lane changing parameters
have been changed, based on intuitive engineering knowledge and best practices
(Marlina and Janson, 2011). These car following parameters include CC1 (headway
time) or the time in seconds that a driver wants to maintain from the car ahead, and
CC2 (following variation) or desired safety distance that a driver allows before
moving closer to the car ahead. A value of 1.12 seconds was used for CC1 instead
of the default value of 0.90 seconds. A value of 14.50 feet was used for CC2 since
drivers are more cautious when the roadway is congested, and the maximum
deceleration for cooperative braking is 20 ft/sec2. The VISSIM default value for
41
waiting time before diffusion of 60 seconds was increased to 300 seconds to allow
for more congested traffic. When distance between vehicles increases, it will have
more impact on capacity. Tables 3.2 and 3.3 present car following and lane changing
parameters used in the VISSIM model and the default values.
Table 3.2 Car Following VISSIM Parameters
Car Following Parameter Default Used
CCO (Standstill Distance) ft 4.92 4.92
CC1 (Headway Time) seconds 0.90 1.12
CC2 (Following Variation) ft 13.12 14.50
CC3 (Threshold for entering 'Following') 8.00 8.00
CC4 (Negative 'Following' Threshold) 0.35 0.35
CC5 (Positive 'Following' Threshold) 0.35 0.35
CC6 (Speed Dependency of Oscillation) 11.44 11.44
CC7 (Oscillation Acceleration) ft/s2 0.82 0.82
CC8 (Standstill Acceleration) ft/s2 11.48 9.65
CC9 (Acceleration at 50 mph) ft/s2 4.92 4.92
Table 3.3 Lane Changing VISSIM Parameters
Lane Change Default Own Trailing Veh Used Own Used Trailing Veh
Maximum Deceleration ft/sc2 13.12 9.84 13.12 9.84
1 ft/s2 per distance 200.00 200.00 100.00 100.00
Accepted Deceleration ft/s2 3.28 1.64 3.28 1.64
Waiting Time Before Diffusion Seconds 60 300
Min. Headway (front/rear) 1.64 1.64
Safety Distance Reduction Factor 0.6 0.1
Maximum Deceleration for Cooperative Braking ft/sc2 9.84 20.0
42
The VISSIM model in this study uses vehicle inputs at the start of the network and
routing decisions throughout the network to determine vehicle flows. In the simulation
process, the following main assumption was made: the freeway was considered
homogenous meaning that all vehicles entering the segment exit at the end of the
segment and no traffic enters and exits in the middle of the segment.
3.2.3 Calibration, Validation and Verification
In order to get a reasonable match between the observed and expected traffic, the
calibration was performed. Calibration is a process to determine whether the
conceptual simulation model is realistic or accurate representation of the actual
condition of the system study or not. In addition, verification was performed to ensure
the input data and model is appropriate for the study conditions. In Federal Highway
Administration guidelines for applying traffic micro simulation modeling software,
freeway traffic volume calibration is required to have an accurate model using either
percent differences or Geoffrey E. Havers (GEH) statistics formula. For these
purposes, less than 10 percent differences and 5 GEH statistics are required in this
study. A GEH is a nonlinear standard measure of the goodness of fit. When GEH
value is less than 5.0, it is considered a good match between the model and the
actual traffic counts. Applying a stochastic model, it is not likely to get exact volume
matches, but very close. In Marlina and Janson (2011) Interstate 70 Simulation study
show the calibration and validation of traffic volumes performed in table 3.4. The
network and VISSIM model for this study using I70 model calibrated for all
parameters.
43
Table 3.4 Interstate 70 Simulation Calibration and Validation
Segment ID Actual Input Volume (V) Model Output Volume (E) Differences1 GEH Statistic1
EB Direction
228 2476 2437 1.6% 0.39
232 2812 2756 2.0% 0.53
233 2776 2741 1.3% 0.33
234 2708 2691 0.6% 0.17
235 2510 2463 1.9% 0.47
238 2607 2623 0.6% 0.16
239 3083 3057 0.8% 0.23
240 2906 2910 0.1% 0.04
241 2980 2989 0.3% 0.08
243 3337 3252 2.5% 0.74
244 2578 2585 0.3% 0.07
WB Direction
228 1450 1424 1.8% 0.34
232 1516 1530 0.9% 0.18
233 1516 1462 3.6% 0.70
234 1095 1078 1.6% 0.25
235 1163 1154 0.8% 0.13
238 1167 1182 1.3% 0.22
239 1167 1155 1.0% 0.18
240 939 946 0.7% 0.11
241 1266 1272 0.5% 0.08
243 1167 1174 0.6% 0.10
244 1450 1470 1.4% 0.25
differences between models estimated volumes and actual volumes
2GEH statistics: E = model estimated volume and V = Field Counts
GEH =
(E Vf
{E + V)!2
44
3.3 Steps Performed to Develop Truck Equivalence Factors
The methods performed to estimate PCE are based on Equal FlowDensity
(Dermarchi & Setti, 2003) and Equal SpeedDensity (Sumner et al., 1984).
3.3.1 Equal FlowDensity
The following procedures were used to determine the impedance of the basevehicle
flow rate (qB) and mixedvehicle flow rate (qM) in the traffic stream:
1. Establish a speedflowdensity relationship for the base vehicles or
passenger car only. This is obtained from VISSIM microscopic simulation.
The flow rates correspond to the maximum service flow rate for each LOS
Figure 3.4 FLOW DENSITY RELATIONSHIP FOR BASE VEHICLES
45
2. Similar to previous step, generate a speedflowdensity relationship for the
mixed vehicle stream with an equal number of trucks.
0 10 10 30 40 50 60
PeniHy (veh/mi/ln)
Figure 3.5 FLOW DENSITY RELATIONSHIP FOR MIXED VEHICLES
3. Interpolate between observed values to obtain the base flow rate and mixed
vehicle flow rate at an equal density value. Initially, use the density at LOS C,
at an equal value of22 pc/mi/ln.
46
Flow Density (0.5 Miles) 5% Trucks
2500
0
20
30
Density (wh/mi/ln)
50
60
Figure 3.6 INTERPOLATION FLOW DENSITY RELATIONSHIP FOR qB and qM
4. Calculate PCE using below equation based on the same level of impedance
as shown on Figure 3.3;
Where
Pt= proportion of trucks of type i out of all trucks n in the mixed traffic flow
qB = base flow rate (passenger cars only)
qM = mixed flow rate (passenger cars + trucks)
Et = passenger equivalent of trucks
47
3.3.2
Equal SpeedFlow
The following procedures were used to determine the impedance of base vehicles
flow rate (qB), mixed vehicles flow rate (qM), and subject vehicles (qs) in the traffic
stream:
1. Generate the relationship between impedance and flow rate for the
passenger car only (qB). The results are obtained from VISSIM microscopic
simulation. The flow rates correspond to the maximum service flow rate for
each LOS category from HCM 2010.
Figure 3.7 FLOW DENSITY RELATIONSHIP FOR BASE VEHICLES
2. Similar to Step 1, generate a speedflowdensity relationship for the mixed
vehicle stream with an equal number of trucks, containing (1 p) passenger
cars and trucks.
48
Figure 3.8 FLOW DENSITY RELATIONSHIP FOR MIXED VEHICLES
3. Similar to Step 1, generate the relationship between impedance measure and
flow rate for the stream resulting from replacing Ap passenger cars with Ap
subject vehicles in the mixed vehicle stream.
Figure 3.9 FLOW DENSITY RELATIONSHIP FOR SUBJECT VEHICLES
49
4. Interpolate between observed values to obtain the base flow rate (qB), mixed
flow rate (qM), and subject flow rate (qs) at an equal density value. Initially, the
density at LOS C, at an equal value of 22 pc/mi/ln.
Speed Flow (0.5 Miles) 5% Trucks
Figure 3.10 INTERPOLATION FLOW DENSITY RELATIONSHIP FOR qB, qM and
qs
5. Calculate PCE using below equation based on the same level of impedance
as shown on Figure 3.4;
Et =
JL (3*
VP \qs
^)+l
50
Where
qB= equivalent passenger car only flow rate
qM = mixed flow rate
qs = additional subject flow rate
PT= truck proportion in the mixed traffic flow
VP= proportion of subject vehicles
Et = truck PCE
3.4 Scenarios
To analyze the impact of the proportion of trucks in the traffic stream, truck
proportions from 5% to 100% trucks were simulated for lengths of grade of 0.5 and
1.0 miles. The network segment was constructed for 2%, 4%, 6%, and 8% grades.
All of these combinations were applied to road segments with 2lanes, 3lanes
without trucks restrictions, and 3lanes with trucks restrictions, resulting in 264
combinations as shown on the table 3.5.
51
Table 3.5 Truck Percentages Scenarios
% Truck Length of grade (mi) Grade Description
5 0.5 2% 2Lane
10 1.0 4% 3Lane
20 6% 3Lane with truck restriction
30 8%
40
50
60
70
80
90
100
11 2 4 3 264
To analyze the impact of length of grade, grade distances of 0.25 to 5 miles were
simulated, with 5% and 10% trucks in the traffic stream. The network segment was
constructed for 2%, 4%, 6%, and 8% grades. All of these combinations were applied
to road segments with 2lanes, 3lanes without trucks restrictions, and 3lanes with
trucks restrictions, resulting in 240 combinations as shown on the table 3.6.
52
Table 3.6 Length of Grade Scenarios
Length of grade (mi) Truck Grade Description
0.25 5% 2% 2Lane
0.50 10% 4% 3Lane
0.75 6% 3Lane with truck restriction
1.00 8%
1.25
1.50
2.00
3.00
4.00
5.00
10 2 4 3 240
To analyze the impact of the grade, grade percentages from 1% to 10% were
simulated, with 5% and 10% trucks in the traffic stream and lengths of grade from
0.25 to 2 miles. All of these combinations were applied to road segments with 2
lanes, 3lanes without trucks restrictions, and 3lanes with trucks restrictions,
resulting in 420 combinations as shown on the table 3.7.
53
Table 3.7 Grade Percentages Scenarios
% Grade Length of grade (mi) Truck Description
1 0.25 5% 2Lane
2 0.50 10% 3Lane
3 0.75 3Lane with truck restriction
4 1.00
5 1.25
6 1.50
7 2.00
8
9
10
10 7 2 3 420
54
4.
Analysis and Results
4.1 Overview
Federal Highway Administration classified vehicles into 13 categories. Heavy
vehicles fall from category four to thirteen based on number of axles, length, width,
and power units. To simplify multiply truck types in the traffic stream, in this study
trucks are divided into 3 categories and specified into the VISSIM model:
1. Single Unit Trucks (SUT) (vehicle type from numbers 4 to 7, FHWA) with
length of 33.51 feet and width of 4.92 feet.
2. Medium Trucks (vehicle type from numbers 8 to 10, FHWA) with length of
41.59 feet and width of 4.92 feet.
3. Longer Combination Vehicles (LCV) (vehicle type from numbers 11 13,
FHWA) with length of 54.14 feet and width of 4.92 feet.
In addition passenger car is categorized as vehicle type number 2 according to
FHWA vehicles classification.
VISSIM requires a number of input variables. Desired speed is an important
parameter affecting traffic flow in VISSIM. The desired speed distributions specified
in this model were: 50 70 mph for passenger cars and 40 60 mph for trucks. The
level of congestion and roadway geometry dictates the vehicle speeds within these
55
ranges. A speed decision is placed on the entry link. The simulation was conducted
at five different saturation flow rates corresponding to the maximum service flow
rates in the HCM 2010: 820 pc/h/ln (LOS A), 1310 pc/h/ln (LOS B), 1750 pc/h/ln
(LOS C), 2110 pc/h/ln (LOS D), and 2400 pc/h/ln (LOS E). The VISSIM model in this
study was used the static routing decision to direct vehicles. The random seed for
generating identical result used is 42. It is the default value in VISSIM simulation
parameter. The base and mix flow rates, speed, and density were obtained as the
output from VISSIM simulation.
4.2 SpeedFlowDensity Relationship
Figures 4.1 to 4.8 show the speedflowdensity relationship obtained from VISSIM
traffic simulation. Separate curves were plotted for different truck percentages,
various lengths of grades, different percent grades (2%, 4%, 6%, and 8%), and
different mixtures of base and mixed vehicles in the traffic stream.
4.2.1 Trucks Proportion
The speeddensityrelationship was obtained from VISSIM for different truck
proportion, from 5 to 100 percent trucks, with 10% increments when mixed vehicles
occur in the traffic stream, as well as for base vehicle. The roadway characteristics
are used for the 0.5 and 1.0 mi.
56
Figures 4.1 and 4.2 illustrate speeddensityrelationship for 0.5 mi with 5% and 10%
trucks. Figures 4.3 and 4.4 illustrate speeddensityrelationship for 1.0 mi length of
grade with 5% and 10% trucks.
57
Ul
00
Speed Densitv (0.5 Miles) 5% Trucks
0
,e, 7T
a JU
S, J;u
0 10 2 qM2%grade cM4 qEJ 2% [fade c D 4* 0 3 (v % grade a grade 0 4 eh/mS/ln) qM 6% gra giac 0 5 de 11,1 qM e qEJ 0 4 894pade K pads
Speed Flow (0.5 Miles) 5% Trucks
0 500 1000 1500 2000 1500
Flow vih/h/ln)
I qM t% grade grade m qM 6& grade 1 qM S% grade
* qB2%grtfc qB4% grade qB6% grade ~ qBflK grade
Figure 4.1 SPEEDFLOWDENSITYRELATIONSHIP FOR 0.5 Ml
5% TRUCKS
Ul
CD
Speed Density (0,5 Miles) 10% Trucks
. so 1 I* 4Q 0
5 30 .
Â£ 20
i
Speed Flow {0,5 Miles) 10% Trucks
[ 20
iO
I11!1
0 500 1000 1500 2000 2500
Flow
* qM 2*4 jrsde qM4% gr#de qM5S p*de ^qM8H gr*de
 qBHijjpsde qS+S grade q6OH grade grade
Figure 4.2 SPEEDFLOWDENSITYRELATIONSHIP FOR 0.5 Ml 10% TRUCKS
CD
O
Speed Density (1,0 Mi les) 5% Trucks
50 1 7
E 30 V
1 (radc q2* paje 7 2 Oe qw 49$ gr gn 0 3 rsrty (ve h/mif sde ^qVlG ce 0 A % grace 1 q & grade t O 5 MfiS padc BBifc grade
Flow Density (1.0 Miles) 5% Trucks
qM 2*? srade A qMOft Erode 1 ( qMSH grade
* qft29fcg"a:le * qB4?& grace qfl S* grade qSSK grade
Figure 4.3 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.0 Ml
5% TRUCKS
Speed Density(1.0 Miles) 10% Trucks
1* qMJ'X srede m qMGSfr Â£r*&t   jrcd*
* qB 2% gride qB4Â§tgr*>c
0 ID 20*0 40 >0
Demit y (veh/imi/Jnli
9 qVI 1% grade qM4% jjraie . qM6% trade tMAK grade
* qiJMgrarfe q84%gra ^^qBfitf grade tBSftgade
Speed Flow (1,0 Miles) 10% Trucks
)M zk grids m q m 4% grade qw ** gradeorvi 8% grade
+ gredc ~kq04H grsdtqDOi grads q0 6Hg*ce
Fiaure 4.4 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.0 Ml
10% TRUCKS
4.2.2
Length of Grade
The lengths of grade were measured from 0.25 to 5 miles for 5% and 10% trucks in
mixed traffic. Figures 4.5 and 4.6 illustrate the example of speeddensityrelationship
for 3.0 and 4.0 miles length of grade with 5% trucks. Figures 4.7 and 4.8 illustrate the
example of speeddensityrelationship for 3.0 and 4.0 miles length of grade with 10%
trucks.
62
CD
CD
Speed Density{3.0 Miles] 5%Trucks
ffr0e A qM4K.fr*de qwe* rwJe^qMW *rad*
n82%yBde > qS4i% grjde 1 qB'frfr pade qB6Hr*de
Speed Flow (3.0 Miles) 5% Trucks
"qMS* padf
" gB Jit trade *qBW grade 1QSSK grade qBB?i grade
Fiaure 4.5 SPEEDFLOWDENSITYRELATIONSHIP FOR 3.0 Ml 5% TRUCKS
Speed Flow {4.0 Miles) 594 Trucks
o 
U SDU HUM 3SINJ
Fhv (rth/h/ki!
qM $i >d* tiM^V (r1 kh 6K (f tde qMr* fil*
qB3Kr*^ 4 q0 Â£% pid*jridt qB S'tpM*
Fiaure 4.6 SPEEDFLOWDENSITYRELATIONSHIP FOR 4.0 Ml 5% TRUCKS
CD
Ul
Speed Density{3.0 Miles) 10% Trucks
qM 2ft grade qM 4ft grade qM &ft grade qM S ft grade
qBZft gf*d* V qBU94ff*flhs qB6Hir*Je cB SH ptdt
> qM 2% grade U qM4lfc grade qM6Kgrade 11 qM SH grade
=*=flB ZH gr*de grade 1 qB6*ft grade qB8ft grade
Fiaure 4.7 SPEEDFLOWDENSITYRELATIONSHIP FOR 3.0 Ml 10% TRUCKS
CD
CD
Flow Density (4.0 Miles}10% Trucks
*qW2H rk qM4S. grack qMi&K jridt tMfl% gride
qfilSfc (fade Â§ qfi4%grflde qBÂ£Sigr*de ttftflit grade
Fiaure 4.8 SPEEDFLOWDENSITYRELATIONSHIP FOR 4.0 Ml 10% TRUCKS
4.2.3
Grade Percentages
The grade percentages were measured from 1 to 10 percent, for 5% and 10% trucks
in the traffic stream. Figure 4.9 illustrates the example of speeddensityrelationship
for 1.5 mile length of grade with 5% trucks.
Figure 4.10 illustrates the example of speeddensityrelationship for 2.0 miles length
of grade with 5% trucks.
Figure 4.11 illustrates the example of speeddensityrelationship for 1.5 miles length
of grade with 10% trucks.
Figure 4.12 illustrates the example of speeddensityrelationship for 2.0 miles length
of grade with 10% trucks.
67
Spttd Flow [X.S Miks 5% Truck*
Sp**d DtniRy il,5 M lltif 5ft Trutki tfu:! 
c j^tS
sjjjj'1 aw rt ra
l * "*r aw n piA  u 1* 
I ^lurtrDN ai rtr*i* *;*"
u> urtpM* 4ri* * 3*HiyS*
 aAfii wrtp***
fa a & i o a &*** a a o a a *t a 1 33 ** U* r5l
Flaw Dtfiafty (1.5 Milti] Sttlhidta
r gfct r"**
*'**"* fT*
gwniF*
jWAr^
 c* w I<31
*qtArt
o*l*i pki
4 SS6 1446 i.164 2464 ii
cwsvdhV^N
rtf'll!
*M**F'ai
npurtriH
rt ri*
. aw^i pii
^ #Wrtf*i#i
aw ia rii
i y ^sh r*=*
a^Apdl
ailAp^*
* *rt*nd*
*>*
anfiii
*!r
Fiaure 4.9 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.5 Ml 5% TRUCKS
 uiniy (2,u Mil*a) Trucks
adptai
ism a* r
>au^ri
r"
rpMpn
uiSrM
*l MP pii
Iai S rt*
Urtr**
 iftj'ij*
f*t
mfcpM*
n^r***
gin r*
awn*
ai tA r>s<
lew Density [2.0 Milts) S% Trucks
2CK 13U L in AptM UU J%p*3* nrn iiifSytw 4WFGjr*a m *(#* tunpiii 4 m >;*!* 4l}fcr*4fr **4% 1**44 ^"^lr*dt #> 4nrox la <**a>asH*ias
Uf
1 t b t U M 1 EMWlhl 3 G 3 i 4 0 4 a a
Fiaure 4.10 SPEEDFLOWDENSITYRELATIONSHIP FOR 2.0 Ml 5% TRUCKS
Fiaure 4.11 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.5 Ml 10% TRUCKS
01
Fiaure 4.12 SPEEDFLOWDENSITYRELATIONSHIP FOR 2.0 Ml 10% TRUCKS
YL
4.3 PCE Results
4.3.1 Variance PCE by Trucks Proportion
The HCM 2010 considers proportion of trucks only up to 25 percent. However, on
many freeways in the United States, the proportion of trucks exceeds 25 percent. To
test the effect of high truck percentages, the trucks proportion was considered in
increments of 10%, started from 5 percent all the way to 100 percent trucks. At the
lower proportion of trucks (from 0% to approximately 30% or > 25%), the percentage
categories used in the HCM 2010 were matched for comparison.
One aim of this study is to verify that the PCE decreases as the proportion of trucks
increases. Even though this trend is observed in the HCM 2010, it has not been
tested at high trucks percentages. An additional substantial aim is compare
Demarchi & Setti (flowdensity based) and Sumner et at. (speedflow based)
methods for calculating PCEs.
72
Figure 4.13 shows PCE variability with truck proportion for 0.5 mile and 2% grade
using Demarchi & Setti and Sumner et al. methodologies.
0.5 mile 2% grade
Percentage Trucks (%)
2Lane HCM 2010
(a) Demarchi & Settis Method
0.5 mile 2% grade
2Lane
HCM 2010
(b) Sumner et als Method
Figure 4.13 PCE VARIABILITY BY TRUCKS PROPORTION 0.5 Mi 2% GRADE
73
Figure 4.14 shows PCE variability with truck proportion for 0.5 mile and 4% grade
using Demarchi & Setti and Sumner methodologies.
0.5 mile 4% grade
2Lane HCM 2010
(a) Demarchi & Settis Method\
0.5 mile 4% grade
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.14 PCE VARIABILITY BY TRUCKS PROPORTION 0.5 Mi 4% GRADE
74
Figure 4.15 shows PCE variability with truck proportion for 0.5 mile and 6% grade
using Demarchi & Setti and Sumner et al. methodologies.
0.5 mile 6% grade
2Lane HCM 2010
(a) Demarchi & Settis Method
0.5 mile 6% grade
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.15 PCE VARIABILITY BY TRUCKS PROPORTION 0.5 Mi 6% GRADE
75
Figure 4.16 shows PCE variability with truck proportion for 0.5 mile and 8% grade
using Demarchi & Setti and Sumner et al. methodologies.
0.5 mile 8% grade
2Lane HCM 2010
(a) Demarchi & Settis Method
0.5 mile 8% grade
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.16 PCE VARIABILITY BY TRUCKS PROPORTION 0.5 Mi 8% GRADE
76
Figure 4.17 shows PCE variability with truck proportion for 1.0 mile and 2% grade
using Demarchi & Setti and Sumner et al. methodologies.
1.0 mile 2% grade
2Lane HCM 2010
(a) Demarchi & Settis Method
1.0 mile 2% grade
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.17 PCE VARIABILITY BY TRUCKS PROPORTION 1.0 Mi 2% GRADE
77
Figure 4.18 shows PCE variability with truck proportion for 1.0 mile and 4% grade
using Demarchi & Setti and Sumner et al. methodologies.
1.0 mile 4% grade
2Lane HCM 2010
(a) Demarchi & Settis Method
1.0 mile 4% grade
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.18 PCE VARIABILITY BY TRUCKS PROPORTION 1.0 Mi 4% GRADE
78
Figure 4.19 shows PCE variability with truck proportion for 1.0 mile and 6% grade
using Demarchi & Setti and Sumner et al. methodologies.
1.0 mile 6% grade
2Lane HCM 2010
(a) Demarchi & Settis Method
1.0 mile 6% grade
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.19 PCE VARIABILITY BY TRUCKS PROPORTION 1.0 Mi 6% GRADE
79
Figure 4.20 shows PCE variability with truck proportion for 1.0 mile and 8% grade
using Demarchi & Setti and Sumner et al. methodologies.
1.0 mile 8% grade
2Lane HCM 2010
(a) Demarchi & Settis Method
1.0 mile 8% grade
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.20 PCE VARIABILITY BY TRUCKS PROPORTION 1.0 Mi 8% GRADE
80
The proportion of trucks in the traffic stream was found to have a significant effect on
the calculated PCE. For higher truck percentages, the PCE decreases and levels off,
except for Figure 4.15b. That is because of the limitation of Sumner et al.s method
which does not include multiple trucks interaction. This trend is justified because as
the truck percentages increases, the interaction between trucks and passenger car
decreases. The trucks will form platoons climbing the grade. The interaction among
trucks would be negligible as they have same performances and operations.
The variation in PCE for proportion of trucks between 25 and 50 percents provides
substantial evidence. Hence, the HCM 2010 should be extended to include 50
percent trucks. Above 50 percent of trucks, the PCE shows very little variability.
4.3.2 Variance PCE by Length of Grade
The HCM 2010 considers length of grade up to 1 mile for specific upgrades.
Nevertheless, on many freeways in the United States, the length of grade exceeds 1
mile. To test the effect of length of grade, grades measure in increments of 1 mile,
started from 0.25 mile all the way to 5 miles length. At the lower percent grades (from
0% to approximately 6% or > 6%), the percent grades categories used in the HCM
2010 were matched for comparison. The aim is to test the hypothesis that as the
length of grade increases, the PCE increases and to compare the results using
Dermarchi & Settis and Sumner et al.s methodologies.
81
Figure 4.21 shows PCE variability with length of grade for 2% grade and 5% trucks
using Demarchi & Setti and Sumner et al. methodologies. At the low grade lengths,
the simulations agree well with the HCM model, but the PCE continues to increase
significantly.
(a) Demarchi & Settis Method
(b) Sumner et als Method
Figure 4.21 PCE VARIABILITY BY LENGTH OF GRADE 2% GRADE
82
Figure 4.22 shows PCE variability with length of grade for 4% grade and 5% trucks
using Demarchi & Setti and Sumner et al. methodologies.
4% grade 5% trucks
Length of Grade (Mi)
2Lane HCM 2010
(a) Demarchi & Settis Method
4% grade 5% trucks
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.22 PCE VARIABILITY BY LENGTH OF GRADE 4% GRADE
83
Figure 4.23 shows PCE variability with length of grade for 6% grade and 5% trucks
using Demarchi & Setti and Sumner et al. methodologies.
6% grade 5% trucks
2Lane HCM 2010
(a) Demarchi & Settis Method
6% grade 5% trucks
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.23 PCE VARIABILITY BY LENGTH OF GRADE 6% GRADE
84
Figure 4.24 shows PCE variability with length of grade for 8% grade and 5% trucks
using Demarchi & Setti and Sumner et al. methodologies.
8% grade 5% trucks
Length of Grade (Mi)
2Lane HCM 2010
(a) Demarchi & Settis Method
8% grade 5% trucks
2Lane HCM 2010
(b) Sumner et als Method
Figure 4.24 PCE VARIABILITY BY LENGTH OF GRADE 8% GRADE
85

PAGE 1
UNDERSTANDING THE DYNAMICS OF TRUCK TRAFFIC ON FREEWAYS BY EVALUATING TRUCK PASSENGER CAR EQUIVALENT (PCE) IN THE HIGHWAY CAPACITY MANUAL (HCM) 2010 By Susi Marlina Associate Eng. (A.Md.) in Civil Engineering, University o f Lampung, Indonesia 1997 B.Sc. in Civil Engineering, University of Bandar Lampun g, Indonesia 2000 M.Sc. in Civil Engineering, University of Colorado at D enver, USA 2006 A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Doctor of Philosophy Civil Engineering 2012
PAGE 2
2012 by Susi Marlina All rights reserved
PAGE 3
This thesis for the Doctor of Philosophy degree by Susi Marlina has been approved by Bruce N. Janson Wesley Marshall Yuk Lee Angela Bielefeldt Juan Robles
PAGE 4
Marlina, Susi (Ph.D, Civil Engineering) Understanding the Dynamics of Truck Traffic on Freeways by Evaluating Truck Passenger Car Equivalent (PCE) in the Highway Capacity M anual (HCM) 2010 Thesis directed by Professor Bruce N. Janson ABSTRACT Truck traffic causes significant problems, including congest ion, delay, crashes, pollution, energy consumption, and road damage in ma ny regions because trucks are larger in size and heavier than passenger cars. In a ddition, trucks have limited performance, especially on grades and curves for accelerat ing and decelerating. A common treatment of truck traffic procedures for highway capacity and level of service (LOS) determination is to multiply trucks by a passen ger car equivalent (PCE) based on Highway Capacity Manual (HCM) guidelin es. A deficiency exists in the current PCE estimation because a truck does not perfor m like multiple cars. The main objective of this study is to understand the dynamic s of truck traffic and develop a new truck PCE using a simulationbased approach. Speedflowdensity relationship was used for this study because speedflowdensity is able to measure traffic flow and the randomness in t he traffic stream and its validity has been proven by empirical research with fiel d observations. The basic concept of PCE derivation uses a deterministic model of traffic flow from GreenshieldÂ’s Model proposed by Huber and developed by Sumner et al. Their concept is based on an equal speeddensity relationship That method was compared to Demarchi and SettiÂ’s to estimate truck PCE based on equal flowdensity relationship. Those PCE values were used to evalu ate the current HCM 2010. Statistical analysis was conducted to investigate t he differences between calculating truck PCE using equal flowdensity and equal speeddensity methodologies. A number of different scenarios and variables were consi dered: truck proportion, grade percentages, length of grades, number of lanes, a nd truck lane restrictions. New truck PCE estimation tables are proposed for future revisions of the Highway Capacity Manual. This abstract accurately represent the content of the can didateÂ’s thesis. I recommend its publication. Approved Bruce N. Janson
PAGE 5
DEDICATION I dedicate this dissertation to my parents and family f or their unfaltering understanding, praying and support to reach my dreams. I also dedicate this dissertation to my extended family, friends and colleague s, both in Indonesia and in the USA, who gave me an appreciation of learning and taught me the values of perseverance, humility, persistence, and passion.
PAGE 6
ACKNOWLEDGEMENT I am sincerely grateful to the Almighty Creator of the World. The ideas, models and outcomes contained in this dissertat ion could not have been achieved without the support of many people. I wish to convey my deepest gratitude to Professor Bruce N. Janson, who served as my advisor and mentor. With his guidance, the ma jority of the ideas and work contained in this dissertation were finely sharp ened and organized. His energy has given an enormous boost to my work. Most of a ll, I wish to sincerely thank him for his enthusiasm toward research and his encouragement to persevere. Without his support, this diss ertation could not been possible. I also appreciate the guidance given me by the members of my advisory committee: Dr. Yuk Lee, Dr. Wesley Marshall, Dr. Angel a Bieldfeldt, and Dr. Juan Robles. I give my deepest thanks to my parents and siblings for th eir endurance and unfaltering prayers. I also wish to express my gratitude to my host family and friends Katie Paganucci, Jay Harker, Valerie Kiltzer, Zachary Held, Jere my Harker, and David. They have provided me an alternate home here. I also wish to thank Kathie Haire and Cindy Collip for supportive assistance and provocative discussions. Thanks to all of my peers at Par son Brinckerhoff for their positive encouragement and patience with me especially Jim Daves. He provided the fortunate opportunity to work on the I70 simulation research project as part of subcontract of University of Colorado Denver. A million thanks goes to Jim Root, David Krutsinger, Kar l Bucholz, Alvin Stamp, Patricia Batuna, Ririen Indriani, and Ronny H Purba for their support of all my effort to get it done! I have also enjoyed a dvice from C. Jon Chen to my intellectual development. I also would like to acknowledge and thank Katherine Ca rol, Adriana Carlson, and Mikelle Learned for the opportunity to live and wo rk together at the beginning of my doctorate program.
PAGE 7
Sincere thanks goes to Krista Nordback for her efforts wit h proofreading my thesis and reviewing my dissertation .Kudos to Pamela F ischhaber for her brainstorming with me in preparation of my defense. Finally, I would deliver my gratitude to the Colorado Department of Transportation (CDOT) especially Region 1, University of Bandar Lampung Indonesia, and the Indonesian community and friends bo th in the USA and in Indonesia. Susi Marlina
PAGE 8
viii TABLE OF CONTENTS LIST OF TABLES .................................... ................................................... ........... x LIST OF FIGURES ................................... ................................................... ........ xii Chapter 1. Introduction ................................... ................................................... ...............1 1.1 Background ..................................... ................................................... ......1 1.2 Study Objectives ................................ ................................................... ...2 1.3 Significant Studies............................ ................................................... .....3 1.4 Simulation .................................... ................................................... .........6 1.5 Organization of the Dissertation ................. ..............................................7 2. Literature Review ............................... ................................................... ..........9 2.1 Review of Truck Equivalencies..................... ............................................9 2.1.1 PCEs Based on Flow Rates and Density ........... .................................10 2.1.2 PCEs Based on Headways ........................ .........................................12 2.1.3 PCEs Based on Queue Discharge Flow ............ .................................15 2.1.4 PCEs Based on Speed .......................... .............................................15 2.1.5 PCEs Based on Delays .......................... ............................................16 2.1.6 PCEs Based on V/C Ratio ...................... ............................................19 2.1.7 PCEs Based on VehicleHours ................... ........................................19 2.1.8 PCEs Based on Platoon Formation .............. ......................................20 2.1.9 PCEs Based on Travel Time..................... ..........................................21 2.2 Truck Equivalencies in the 2010 HCM................ ....................................22 2.3 Traffic Flow Parameter Concepts ................. ..........................................24 2.4 Level of Service (LOS) .......................... .................................................26 2.5 Vehicle Classification ........................... ..................................................2 8 2.6 Simulation Models Study ........... ................................................... ..31 3. Methodology .................................... ................................................... ..........32 3.1 PCE Methods Theoretical Analysis Evaluation ....... ................................32 3.2 Simulation .................................... ................................................... .......34 3.2.1 Decision Support Methodology (DSM) ............. ...................................34 3.2.2 VISSIM Micro Simulation ..................... ...............................................38 3.2.3 Calibration, Validation and Verification ..... ..........................................43 3.3 Steps Performed to Develop Truck Equivalence Factors. .......................45 3.3.1 Equal FlowDensity ........................... ..................................................4 5 3.3.2 Equal SpeedFlow ............................ ..................................................4 8 3.4 Scenarios ...................................... ................................................... ......51 4. Analysis and Results .............................. ................................................... ....55 4.1 Overview ........................................ ................................................... .....55 4.2 SpeedFlowDensity Relationship ................. .........................................56
PAGE 9
ix 4.2.1 Trucks Proportion .............................. .................................................56 4.2.2 Length of Grade.............................. ................................................... .62 4.2.3 Grade Percentages ............................ ................................................67 4.3 PCE Results..................................... ................................................... ...72 4.3.1 Variance PCE by Trucks Proportion ................ ...................................72 4.3.2 Variance PCE by Length of Grade ............... .......................................81 4.3.3 Variance PCE by Grade Percentages ............. ....................................90 4.3.4 Variance PCE by Number of Lane ................ .................................... 105 4.3.4.1 Truck Proportion ............................ ............................................... 105 4.3.4.2 Length of Grade ........................... ................................................... ... 114 4.3.4.3 Grade Percentages .......................... .................................................. 122 4.3.5 Variance PCE by Trucks Restrictions ................ ............................... 136 4.3.5.1 Truck Proportion ............................ ............................................... 137 4.3.5.2 Length of Grade ........................... ................................................... ... 145 4.3.5.3 Grade Percentages .......................... .................................................. 153 5. Statistical Analysis .............................. ................................................... ..... 168 5.1 Overview ........................................ ................................................... ... 168 5.2 Truck Proportion ................................ .................................................. 169 5.2.1 Mean, Standard Deviation and Other Statistics.. ............................... 169 5.2.2 Independent Sample ttest .................... ............................................ 171 5.2.3 Percent Differences ........................... ............................................... 173 5.3 Length of Grade ............................... ................................................... 176 5.3.1 Mean, Standard Deviation and Other Statistics.. ............................... 177 5.3.2 Independent Sample ttest .................... ............................................ 178 5.3.3 Percent Differences ........................... ............................................... 180 5.4 Grade Percentages .............................. ................................................ 18 4 5.4.1 Mean, Standard Deviation and Other Statistics.. ............................... 185 5.4.2 Independent Sample ttest .................... ............................................ 186 5.4.3 Percent Differences ........................... ............................................... 188 6. Additional Study ............................... ................................................... ........ 192 6.1 Overview ........................................ ................................................... ... 192 6.2 Fuel Consumption and Emissions .................. ...................................... 193 6.3 Summary ....................................... ................................................... ... 197 7. Conclusions and Recommendations ................... ........................................ 198 7.1 Conclusions ..................................... ................................................... 198 7.2 Recommendations ................................ ............................................... 204 REFERENCES ........................................ ................................................... ... 214 APPENDIX .......................................... ................................................... ........ 219
PAGE 10
x LIST OF TABLES Table 2.1 PCEs for Heavy Vehicles in General Terrain Segments (HCM, 2010) ...............22 2.2 PCEs for Trucks and Buses ( E T ) on Specific Downgrades (HCM, 2010) ............23 2.3 PCEs for Trucks and Buses ( E T ) on Upgrades (HCM, 2010) .......................... ....23 2.4 SpeedDistance Relationship for Acceleration of Hea vy Truck (AASHTO A Policy on Geometric Design of Highways and Streets, 2000) ..................................26 2.5 Vehicle Weight and Power (Traffic Engineering Han dbook, 2009) .....................31 3.1 Decision Support Methodology (DSM) Result ........ ............................................36 3.2 Car Following VISSIM Parameters ............... ................................................... ...42 3.3 Lane Changing VISSIM Parameters ............... ................................................... 42 3.4 Interstate 70 Simulation Calibration and Validat ion ............................................44 3.5 Truck Percentages Scenarios ....................... ................................................... ..52 3.6 Length of Grade Scenarios ..................... ................................................... ........53 3.7 Grade Percentages Scenarios ..................... ................................................... ...54 5.1 Mean and Standard Deviation of PCE ............. ................................................ 17 0 5.2 Other Statistics ................................ ................................................... ............. 170 5.3 Independent ttest between 2Lane and 3Lane ... ............................................ 171 5.4 Independent ttest between 3Lane without and wi th Trucks Restrictions ........ 172 5.5 Truck Proportion Percent Difference in PCEs using D emarchi & SettiÂ’s Method .................................................. ................................................... ......................... 173 5.6 Truck Proportion Percent Difference in PCEs using S umner et al.Â’s Method .... 175 5.7 Mean and Standard Deviation of PCE ............. ................................................ 17 7 5.8 Other Statistics ................................ ................................................... ............. 178 5.9 Independent ttest between 2Lane and 3Lane ... ............................................ 179 5.10 Independent ttest between 3Lane without and with Trucks Restrictions ...... 180 5.11 Length of Grade Percent Difference in PCEs using Demarchi & SettiÂ’s Method .................................................. ................................................... ......................... 181 5.12 Length of Grade Percent Differences in PCEs using Sumner et al. ................ 183 5.13 Mean and Standard Deviation of PCE............. ............................................... 185 5.14 Other Statistics ............................... ................................................... ............ 186 5.15 Independent ttest between 2Lane and 3Lane ............................................ 187 5.16 Independent ttest between 3Lane without and with Trucks Restrictions ...... 188 5.17 Grade Percentages Percent Difference in PCEs using D emarchi & SettiÂ’s..... 189 5.18 Grade Percentages Percent Difference in PCEs using S umner et al.Â’s .......... 190 6.1 Fuel Consumption ............................... ................................................... .......... 194 6.2 Emissions ....................................... ................................................... .............. 196 7.1 PCE ranges of Truck Proportions................. ................................................... 200
PAGE 11
xi 7.2 PCE ranges of Length of Grades ................ ................................................... .. 201 7.3 PCE ranges of Grade Percentages ................ .................................................. 202 7.4 Summary Percent Differences of PCE for Three Lanes vs Two Lanes ............ 202 7.5 PCE Recommendation based on Truck Proportion for 0 .5 mile........................ 204 7.6 PCE Recommendation based on Truck Proportion for 1 .0 mile........................ 206 7.7 PCE Recommendation based on Length of Grade for 5% Trucks .................... 207 7.8 PCE Recommendation based on Length of Grade for 10% Trucks .................. 209 7.9 PCE Recommendation based on Grade Percentages fo r 5% Trucks ............... 210 7.10 PCE Recommendation based on Grade Percentages fo r 10% Trucks ........... 212
PAGE 12
xii LIST OF FIGURES Figure 2.1 GENERALIZED SPEEDFLOWDENSITY RELATIONSHIP ON UNINTERRUPTEDFLOW FACILITIES (HCM, 2010) ......... .....................................25 2.2 LOS FOR BASIC FREEWAY (HCM, 2010) ............. ...........................................27 2.3 THREE TYPES OF FREEWAY (HCM 2010) ............. ........................................28 2.4 FHWA VEHICLE CLASSIFICATIONS (FHWA, 1985) ..... ...................................30 3.1 RELATIONSHIP BETWEEN SPEED AND DENSITY (Greenshi eld, 1934) ........33 3.2 RELATIONSHIP BETWEEN SPEED (v) AND FLOW (q) (Gre enshield, 1934) ...34 3.3 FUNDAMENTAL TRAFFIC FLOW DIAGRAMS CASE STUDY AT I95 SOUTH FLORIDA (SIUHI AND MUSSA, 2007) ................... .................................................41 3.4 FLOW Â– DENSITY RELATIONSHIP FOR BASE VEHICLES ............................45 3.5 FLOW Â– DENSITY RELATIONSHIP FOR MIXED VEHICLES ...........................46 3.6 INTERPOLATION FLOW Â– DENSITY RELATIONSHIP FOR q B and q M ............47 3.7 FLOW Â– DENSITY RELATIONSHIP FOR BASE VEHICLES ............................48 3.8 FLOW Â– DENSITY RELATIONSHIP FOR MIXED VEHICLES ...........................49 3.9 FLOW Â– DENSITY RELATIONSHIP FOR SUBJECT VEHICLES ......................49 3.10 INTERPOLATION FLOW Â– DENSITY RELATIONSHIP FOR q B q M and q S .....50 4.1 SPEEDFLOWDENSITYRELATIONSHIP FOR 0.5 MI 5 % TRUCKS ............58 4.2 SPEEDFLOWDENSITYRELATIONSHIP FOR 0.5 MI 10 % TRUCKS ...........59 4.3 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.0 MI 5 % TRUCKS ............60 4.4 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.0 MI 1 0% TRUCKS ..........61 4.5 SPEEDFLOWDENSITYRELATIONSHIP FOR 3.0 MI 5% TRUCKS .............63 4.6 SPEEDFLOWDENSITYRELATIONSHIP FOR 4.0 MI 5% TRUCKS .............64 4.7 SPEEDFLOWDENSITYRELATIONSHIP FOR 3.0 MI 10 % TRUCKS ...........65 4.8 SPEEDFLOWDENSITYRELATIONSHIP FOR 4.0 MI 10 % TRUCKS ...........66 4.9 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.5 MI 5% TRUCKS .............68
PAGE 13
xiii 4.10 SPEEDFLOWDENSITYRELATIONSHIP FOR 2.0 MI 5 % TRUCKS ...........69 4.11 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.5 MI 1 0% TRUCKS .........70 4.12 SPEEDFLOWDENSITYRELATIONSHIP FOR 2.0 MI 1 0% TRUCKS .........71 4.13 PCE VARIABILITY BY TRUCKS PROPORTION Â– 0.5 Mi Â– 2% GRADE .........73 4.14 PCE VARIABILITY BY TRUCKS PROPORTION Â– 0.5 Mi Â– 4% GRADE .........74 4.15 PCE VARIABILITY BY TRUCKS PROPORTION Â– 0.5 Mi Â– 6% GRADE .........75 4.16 PCE VARIABILITY BY TRUCKS PROPORTION Â– 0.5 Mi Â– 8% GRADE .........76 4.17 PCE VARIABILITY BY TRUCKS PROPORTION Â– 1.0 Mi Â– 2% GRADE .........77 4.18 PCE VARIABILITY BY TRUCKS PROPORTION Â– 1.0 Mi Â– 4% GRADE .........78 4.19 PCE VARIABILITY BY TRUCKS PROPORTION Â– 1.0 Mi Â– 6% GRADE .........79 4.20 PCE VARIABILITY BY TRUCKS PROPORTION Â– 1.0 Mi Â– 8% GRADE .........80 4.21 PCE VARIABILITY BY LENGTH OF GRADE Â– 2% GRADE ............................82 4.22 PCE VARIABILITY BY LENGTH OF GRADE Â– 4% GRADE ............................83 4.23 PCE VARIABILITY BY LENGTH OF GRADE Â– 6% GRADE ............................84 4.24 PCE VARIABILITY BY LENGTH OF GRADE Â– 8% GRADE ............................85 4.25 PCE VARIABILITY BY LENGTH OF GRADE Â– 2% GRADE ............................86 4.26 PCE VARIABILITY BY LENGTH OF GRADE Â– 4% GRADE ............................87 4.27 PCE VARIABILITY BY LENGTH OF GRADE Â– 6% GRADE ............................88 4.28 PCE VARIABILITY BY LENGTH OF GRADE Â– 8% GRADE ............................89 4.29 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi ............................91 4.30 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi ............................92 4.31 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi ............................93 4.32 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi ..............................94 4.33 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi ............................95 4.34 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi ............................96 4.35 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi ..............................97 4.36 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi ............................98 4.37 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi ............................99 4.38 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi .......................... 100 4.39 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi ............................ 101 4.40 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi .......................... 102 4.41 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi .......................... 103
PAGE 14
xiv 4.42 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi ............................ 104 4.43 PCE VARIABILITY BY TRUCK PROPORTION Â– 0.5 Mi 2 % GRADE ............ 106 4.44 PCE VARIABILITY BY TRUCK PROPORTION Â– 0.5 Mi 4 % GRADE ............ 107 4.45 PCE VARIABILITY BY TRUCK PROPORTION Â– 0.5 Mi 6 % GRADE ............ 108 4.46 PCE VARIABILITY BY TRUCK PROPORTION Â– 0.5 Mi 8 % GRADE ............ 109 4.47 PCE VARIABILITY BY TRUCK PROPORTION Â– 1.0 Mi 2 % GRADE ............ 110 4.48 PCE VARIABILITY BY TRUCK PROPORTION Â– 1.0 Mi 4 % GRADE ............ 111 4.49 PCE VARIABILITY BY TRUCK PROPORTION Â– 1.0 Mi 6 % GRADE ............ 112 4.50 PCE VARIABILITY BY TRUCK PROPORTION Â– 1.0 Mi 8 % GRADE ............ 113 4.51 PCE VARIABILITY BY LENGTH OF GRADE Â– 2% GRADE .......................... 114 4.52 PCE VARIABILITY BY LENGTH OF GRADE Â– 4% GRADE .......................... 115 4.53 PCE VARIABILITY BY LENGTH OF GRADE Â– 6% GRADE .......................... 116 4.54 PCE VARIABILITY BY LENGTH OF GRADE Â– 8% GRADE .......................... 117 4.55 PCE VARIABILITY BY LENGTH OF GRADE Â– 2% GRADE .......................... 118 4.56 PCE VARIABILITY BY LENGTH OF GRADE Â– 4% GRADE .......................... 119 4.57 PCE VARIABILITY BY LENGTH OF GRADE Â– 6% GRADE .......................... 120 4.58 PCE VARIABILITY BY LENGTH OF GRADE Â– 8% GRADE .......................... 121 4.59 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi .......................... 122 4.60 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi .......................... 123 4.61 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi .......................... 124 4.62 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi ............................ 125 4.63 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi .......................... 126 4.64 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi .......................... 127 4.65 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi ............................ 128 4.66 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi .......................... 129 4.67 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi .......................... 130 4.68 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi .......................... 131 4.69 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi ............................ 132 4.70 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi .......................... 133 4.71 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi .......................... 134 4.72 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi ............................ 135 4.73 PCE VARIABILITY BY TRUCK PROPORTION Â– 0.5 Mi 2 % GRADE ............ 137
PAGE 15
xv 4.74 PCE VARIABILITY BY TRUCK PROPORTION Â– 0.5 Mi 4 % GRADE ............ 138 4.75 PCE VARIABILITY BY TRUCK PROPORTION Â– 0.5 Mi 6 % GRADE ............ 139 4.76 PCE VARIABILITY BY TRUCK PROPORTION Â– 0.5 Mi 8 % GRADE ............ 140 4.77 PCE VARIABILITY BY TRUCK PROPORTION Â– 1.0 Mi 2 % GRADE ............ 141 4.78 PCE VARIABILITY BY TRUCK PROPORTION Â– 1.0 Mi 4 % GRADE ............ 142 4.79 PCE VARIABILITY BY TRUCK PROPORTION Â– 1.0 Mi 6 % GRADE ............ 143 4.80 PCE VARIABILITY BY TRUCK PROPORTION Â– 1.0 Mi 8 % GRADE ............ 144 4.81 PCE VARIABILITY BY LENGTH OF GRADE 2% GRADE ........................... 145 4.82 PCE VARIABILITY BY LENGTH OF GRADE 4% GRADE ........................... 146 4.83 PCE VARIABILITY BY LENGTH OF GRADE 6% GRADE ........................... 147 4.84 PCE VARIABILITY BY LENGTH OF GRADE 8% GRADE ........................... 148 4.85 PCE VARIABILITY BY LENGTH OF GRADE 2% GRADE ........................... 149 4.86 PCE VARIABILITY BY LENGTH OF GRADE 4% GRADE ........................... 150 4.87 PCE VARIABILITY BY LENGTH OF GRADE 6% GRADE ........................... 151 4.88 PCE VARIABILITY BY LENGTH OF GRADE 8% GRADE ........................... 152 4.89 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi .......................... 153 4.90 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi .......................... 154 4.91 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi .......................... 155 4.92 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi ............................ 156 4.93 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi .......................... 157 4.94 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi .......................... 158 4.95 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi ............................ 159 4.96 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi .......................... 160 4.97 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi .......................... 161 4.98 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi .......................... 162 4.99 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi ............................ 163 4.100 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 M i ........................ 164 4.101 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 M i ........................ 165 Figure 4.102 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi ............... 166
PAGE 16
1 1. Introduction 1.1 Background Truck traffic causes significant problems including congestio n, delay, crashes, pollution, energy consumption, and road damage in ma ny regions locally, nationally and internationally. Those problems happen because trucks a re larger in size and heavier than passenger cars. In addition, trucks have limi ted performance, especially on grades and curves for accelerating and decelerating A common treatment of truck effects on traffic flow in m any traffic engineering and transportation planning procedures (other than in simu lation models) for highway capacity and level of service (LOS) determination is to m ultiply the number of trucks by a passenger car equivalent (PCE) that is based on High way Capacity Manual (HCM) procedures. The HCM provides the concepts and guide lines to measure the capacity and quality of service for transportation faciliti es. HCM 2010 defines PCE as the number of passenger cars that will result in the sam e operational conditions as a single heavy vehicle of a particular type under specified roadway, traffic, and control conditions. For instance, the equivalence factor of trucks f or signalized intersection is 2.0. The common uses of truck PCEs in traffic engineeri ng and transportation analysis include 2 PCEÂ’s for each single unit truck (SUT) and 3 PCEÂ’s for each longer combination vehicle (LCV). This PCE may be very a pproximate. In the HCM
PAGE 17
2 procedures, the PCE values account for roadway grades and truck percentages. The effect of truck traffic under uncongested conditions is exp ected to be small on level terrain. As congestion increases, the truck traffic is expecte d to affect the traffic stream significantly, particularly on the grades. The deficiency of current PCE estimation is because a tru ck does not perform like multiple cars. For example, an LCV may only travel at a crawl speed up a steep incline, whereas five or ten cars travel up the same incl ine with faster speed. Similarly, the blocking effect of an LCV on urban stree ts is likely to be much greater than the blocking effect of multiple cars. Therefore, there is a need to better quantify the rep resentation of trucks in traffic models such as HCM and other modeling procedures. 1.2 Study Objectives The main goal of this study is to understand the dynamic s of truck effects in traffic through improved truck passenger car equivalents using a simulationbased approach. Thus the objectives of this study are: 1. Quantify the impacts of various proportions of trucks, grade percentages, and length of grade with different vehicle types on truck P CE using micro simulation, considering variables beyond the scope of the current HCM 2010.
PAGE 18
3 2. Compile and compare PCEÂ’s obtained from simulation outputs using both equal density and equal speed formulas and HCM 2010. 3. Verify the hypotheses that: a. When the proportion of trucks is higher in the traff ic stream, the PCE should decrease b. When the grade percentages increase with 5% and 10% trucks in the traffic stream, the PCE should be increased. c. When the length of the grade is greater than 2.0 m iles with 5% and 10% trucks in the traffic stream, the PCE should be greater. d. When the number of lanes increases from two to thre e lanes, the PCE should decrease. 4. Quantify the effects of truck lane restrictions in the left most lanes on PCE values. 5. Propose new truck PCE estimation tables for future revisions of the Highway Capacity Manual. 1.3 Significant Studies The significance of this study is that the truck PCE values developed can be utilized to improve the HCM 2010. Since heavy vehicles like trucks can have major impacts in the traffic stream, such as slower movement, bigger si ze, longer headways and
PAGE 19
4 more frequent gaps, and PCE is used as an equivalency fa ctor. However, many studies suggested ways to improve the current method of computing PCE: Geistefeldt (2009) estimated PCE based on capacity varia bility. The study used a simulation for a more detailed investigation of differ ent factors affecting PCE values, such as truck type or heavy vehicle percentage. A study by Rakha et al. (2007) used a simulation and man y variables to measure the PCE. The simulation results demonstrated that the propo rtion of trucks in the traffic stream has a significant impact on the heavyvehicle PCE a t low proportions, and the trend of decreasing PCE is especially true for longer and steeper grades. Alkaisy et al. (2006) investigated the limitations and appropriate use of HCM PCE factors for heavy vehicles on freeways and multilane highwa ys. The authors found that the effect of heavy vehicles on a traffic stream is greater during congestion than uncongested conditions. He recommended the use of a more realistic PCE factor for heavy vehicles considering queues and congestion analysis an d strongly suggested practical consideration that a set of PCE factors at speci fic grades be developed for congested freeways and included in the capacity analysis pro cedures. Demarchi and Setti (2003) studied the limitation of PCE derivation with more than one truck type. They found that in analysis the current PCE derivations are only able to account for the impact caused by one heavyvehicle class, an d in practice there is
PAGE 20
5 an error in the conversion of an observed mixed flow t o an equivalent flow rate expressed in PCE per hour. Thus, they suggested further studies on the derivation of PCEs for trucks are needed. Benekohal and Zhao (1999) studied delaybased PCE for trucks at signalized intersections and found that the constant PCE recommende d in the HCM overestimated the impact of single unit trucks and the cap acity reduction effects of combination trucks. They suggested that further studies ar e needed to cover different traffic and geometric conditions as well as other heavy veh icle types. Elefteriadou and Webster (1997) used the FRESIM simul ation model to develop the scenarios with up to 6 percent upgrade. The values obtai ned in the research were similar to HCM values on level and slight grades but sig nificantly lower for long and steep grades. Further research is recommended to validate the PCE values reported. Molina (1987) recommended further research into the d evelopment of PCE values for large trucks at signalized intersections, particularly the effect of turning maneuvers and grades. Roess & Messer (1984) reviewed the various approaches for calibration and interpretation of PCE values and recommended revision t o PCE values for multilane uninterrupted flow.
PAGE 21
6 A study by Craus, et al. (1979) suggested further research on proposed truck equivalency, especially for climbing lanes. 1.4 Simulation Traffic simulation has become a vital tool for the desig n of robust and stochastic complex technical systems, including for traffic engineeri ng and transportation planning. A simulation model is able to represent th ese effects more accurately without traffic disruption with many different scenarios. Hence, the simulationbased highway capacity concept was used to evaluate PCE value. Simulation is defined as Â“a numerical technique for con ducting experiments on a digital computer, which may include stochastic characteristi cs, be microscopic or macroscopic in nature, and involve mathematical models th at describe the behavior of a transportation system over extended periods of real t imeÂ” (May, 1989). Simulation is the representation of the actual condit ions by means of mathematical or physical approximation. This representation (in models l ike VISSIM, CORSIM, and TransModeler) uses software based on mathematical models s pecifically for traffic simulation. Simulation features consist of timevarying periods, human performance, and system wide analyses including: spillback, spillover, and bottlenecks.
PAGE 22
7 As VISSIM provides adequate analysis outputs or measures o f effectiveness, for this study VISSIM, a microscopic, behaviorbased traffic simulat ion program is used. 1.5 Organization of the Dissertation The organization of the dissertation steps sequentiall y beginning with a historical look of PCE and concluding with the recommendation of a new truck PCE. Chapter 2 contains a review of PCEÂ’ studies organized b ased on the methods used, traffic flow theory, vehicle classification, and simulatio n. Chapter 3 describes the theoretical analysis development o f truck equivalencies from Greenshields to Demarchi Â– Setti methods, Federal Highw ay Administration (FHWA)Â’s Decision Support Methodology (DSM), and steps p erformed to estimate truck equivalence factor. It also describes the VISSIM traf fic microsimulation, especially the Wiedeman 1999 method of psychophysical dri ving behavior as well as lane changing parameters. Chapter 4 presents the speedflowdensity relationship analysis from VISSIM simulation output. Truck PCE values using Demarchi and S ettiÂ’s and SumnerÂ’s methods based on truck proportion, length of grade, gr ade percentages, number of lane and truck restrictions.
PAGE 23
8 Chapter 5 describes the statistical tests used to evaluate PCE computed based on both equal flow density and equal speed density approach ed. The tests included are test of normality, independence ttest, and percent err or using the mean, variance, and standard deviation from the groups of PCE results. SPSS software is used to do the statistical analysis in this study. Chapter 6 is an additional analysis and research of lit erature on air quality, including fuel consumptions and emissions, to understand the effect of trucks traffic on air quality. The assumptions and formula for fuel consumptio n and emission is based on Synchro simtraffic software analysis. Chapter 7 describes the conclusions and recommendations dr awn from the results of this study.
PAGE 24
9 2. Literature Review 2.1 Review of Truck Equivalencies Truck equivalencies were first mentioned in the 1950 Hig hway Capacity Manual (HCM), which stated that trucks on twolane highways on le vel terrain have the same effect as two passenger cars (PC). This estimate was based on the number of passenger cars passing trucks compared to the number of pa ssenger cars passing passenger cars. The second edition of the HCM was published in 1965, a nd passenger car equivalent (PCE) was introduced. PCE was defined as Â“the number of passenger cars displaced in the traffic flow by a truck or bus, under t he prevailing roadway and traffic conditionsÂ” (HCM, 1965). The current definition of PCE in the HCM 2010 is similar, Â“the number of passenger cars that will result in the same operational conditions as a single heavy vehicle of a particular type under specified roadway, traffic, and control conditionsÂ” (HCM, 2010). The calculation approach used separated speed distribut ion for twolane highways and relative delay, due to trucks on various grades for mul tilane highways, and was based on WalkerÂ’s Method in the 1965 HCM. Many studies h ave been carried out since then. Those studies have examined PCE based on fiel d data, simulations or combination field data and simulations.
PAGE 25
10 According to the HCM 2010, the heavy vehicle adjustment f actor is found as: = n r n r nr Where = heavy vehicle adjustment factor = truck proportion = recreational vehicles (RVs) proportion = truck PCE = recreational vehicles (RVs) PCE 2.1.1 PCEs Based on Flow Rates and Density Measuring the traffic flow is critical on the roadway whe ther the road is congested or not. In transportation engineering, traffic flow ra te is the term used to indicate the equivalent hourly rate of vehicles passing a point per un it of time. Many studies of truck equivalencies have developed using flow rates as the basis of estimation. PCE is computed based on mixed vehicle flow, percentage of g rade, and truck volume to capacity ratio (John and Glauz, 1976): = n r nr Where
PAGE 26
11 = equivalent passenger car only flow rate for a given v/c ratio = mixed flow rate = truck proportion in the mixed traffic flow = truck PCE HuberÂ’s model estimated PCEvalues for vehicle under f reeflowing, multilane conditions by considering the relationship between some m easure of impedance along a length of roadway and the flow rate along th e same roadway for two different traffic streams. Sumner et al. (1982) further develo ped HuberÂ’s method by including more than one truck type in the traffic stream. HuberÂ’s basic equation : = nr Sumner formula: = nr Where = equivalent passenger car only flow rate = mixed flow rate = additional subject flow rate = truck proportion in the mixed traffic flow
PAGE 27
12 = proportion of subject vehicles = truck PCE Demarchi and Setti (2003) suggested the PCEÂ’s formula to eliminate the possible error for mixed heavy vehicles in the traffic stream, inclu ding interaction between multiple trucks types: = nr Where = proportion of trucks of type i out of all trucks n in t he mixed traffic flow = base flow rate (passenger cars only) = mixed flow rate = passenger equivalent of trucks 2.1.2 PCEs Based on Headways Headway is defined as the time in seconds between two s uccessive vehicles as they pass a point of the roadway. It can be measured either f rom front axle or front bumper (HCM, 2010). Understanding the nature of head way from truck movement on the roadway is another approach to calculate truck eq uivalencies. Many researchers have used headway as the basis of estimation. Werner and Morrall (1976) recommended determining P CE using headways when the roadway is sufficiently congested on level terrain:
PAGE 28
13 = # n$r Where = average headway for all vehicles = average headway for passenger car = truck proportion = passenger car proportion = truck PCE Using the spatial headway methodology, Seguin, et.al.(1 982) formulated PCE as the ratio of average headway for vehicle types: average truck headway divided by the average headway for passenger cars: % & = & '(& n)r Where % & = the PCE of vehicle Type i under Conditions j & = average headway for vehicle Type i '(& = the average headway for passenger car for Conditions j
PAGE 29
14 Cunagin and Chang (1982) determined the effect of th e presence of heavy trucks on freeway traffic streams using time headway based on headwa y type, lane width, and traffic volume as shown on Equation 2.8. They conclude tha t the presence of trucks in the traffic stream is accompanied by an increase in the mean headway. The lagging headway is measured from the rear bumper of t he lead vehicle to the rear bumper of the following vehicle. = & n*r Where & = the mean lagging headway of vehicle type i under conditions j = the mean lagging headway of passenger cars. = truck PCE Krammes and Crowley (1986) suggested that : = + n r ., / 0, n1r Where = the proportion of trucks = the lagging headway of trucks following passenger cars H TT = the lagging headway of trucks following trucks = the lagging headway of cars following either vehicle t ype = truck PCE
PAGE 30
15 Anwar et al. (2011) presented a statistical approach fo r determining the PCE values for singleunit trucks and combination trucks using the co ncept of spatial lagging headways, the distance from the rear bumper of leading vehicle to the rear bumper of the following vehicle, measured from real traffic data. 2.1.3 PCEs Based on Queue Discharge Flow A queue is an accumulation of vehicles upstream of a bott leneck when the vehicles are moving slowly or standing still. It occurs when the de mand of vehicles passes exceeds available capacity. AlKaisy et al. (2002) quant ified the calculation of PCE using queue discharge flow (QDF) based on the assumptio n that QDF capacity observation can be expected to show minimal variation if the traffic stream is uniform and consists of passenger cars only. They found that the ef fect of heavy vehicles on a freeway is greater when it is operating in oversatur ated conditions. In addition, it was found that PCE both during dry or rainy days and du ring the presence of roadside maintenance work are not significantly differen t. 2.1.4 PCEs Based on Speed Another measure used by many researchers in their studi es of PCE is the rate of motion of vehicles in a distance per unit of time or spe ed. Van Aerde and Yagar (1983) developed a methodology to estimate PCE based o n the relative rates of speed for each type of vehicle traveling in the main d irection and for all vehicles combined traveling in the opposing direction. They fou nd that PCE decreases for
PAGE 31
16 higher speed percentiles. The speed analysis using the li near regression model structure is: 2345367893:.33;<=433:.33;> ? n@ABCDEFGHIID@JDEKHEIr > L n@ABCDEFGHIID@JDEMEAKNIr> O n@ABCDEFPQIr > R n@ABCDEFFMSDETDSUKVDIr > W n@ABCDEFFGGFIU@JTDSUKVDIr nXr Where > ? to > W = the coefficients of speed reductions for each vehicle type. Using the speed reduction coefficients, the PCE for a veh icle type n is calculated as: Y = % Y % ? nr Where % Y = speed reduction coefficient for vehicle type n % ? = speed reduction coefficient for passenger cars Y = truck PCE 2.1.5 PCEs Based on Delays The HCM denotes delay as the additional travel time ex perienced by a driver, passenger, or pedestrian (HCM, 2010). The PCE values we re determined by using Walker spatialheadway and equivalentdelay methods. A basic assumption in the
PAGE 32
17 Walker method is that faster vehicles are not hindered i n passing as they overtake slower vehicles, so queues do not form. In contrast, in the equivalentdelay method, it assumed that faster vehicles are always hindered by sl ower vehicles, such that queues form. Using that premise, PCE values calculated using a linear combination of the Walker and equivalentdelay in each intermedia te volume level yields: = n Z[ 0QZ\ r + 0] / + 0] / r n Z[ ^" 0QZ\ ^" r + 0] " / + 0] / r nr Where Z[ = the number of overtakings of vehicle type i by passenger cars VOL i = the volume of vehicle type i OT LPC = the number of overtakings of lower performance passeng er cars by passenger cars VOL LPC = the volume of lower performance passenger cars SP M = the mean speed of the mixed traffic stream SP B = the mean speed of the base traffic stream with onl y high performance passenger cars SP PC = the mean speed of the traffic stream with only passen ger cars E T = truck PCE
PAGE 33
18 Craus et al. (1979) developed a passenger car equivalent for trucks ratio of delay time caused by one truck to the delay time caused by one p assenger car. This method takes the oppositelane traffic into consideratio n. The following equation reflects the actual disturbance and delay caused by trucks t o other traffic: = `a `' nr Where = truck PCE `a = average delay time caused by one truck `' = average delay time caused by one passenger car Cunagin and Messer (1983) developed PCE estimation bas ed on speed distribution, traffic volumes, and vehicle types. Their method estimate s PCEs using the ratio of delay experienced by a passenger car due to nonpassen ger vehicles to the delay experienced by a passenger car due to other passenger ca rs: = b & b cdef b cdef nr Where = PCE of vehicle Type i under Conditions j b & = delay to passenger cars due to vehicle Type i under Conditions j
PAGE 34
19 b cdef = delay to standard passenger cars due to slower passenge r cars 2.1.6 PCEs Based on V/C Ratio Fan (1989) studied PCE for expressways in Singapore usin g volumetocapacity (V/C) ratio instead of density or level of service becaus e these freeways operate at LOS E. The study focused on congested flow conditions or V/C ratio above 0.67 and mentioned that it is unnecessary to calculate PCEs at un congested flow conditions. Using multiple linear regressions by multiplying the obser ved flow by the V/C ratios, he found that commercial vehicles such as light and heavy tr ucks, buses, and trailers generally have higher PCE values compared with the PC Es used in US and UK for the level terrain. 2.1.7 PCEs Based on VehicleHours Hourly traffic volumes are used for determining the l ength and magnitude of peak periods, evaluating capacity, and assessing geometric design and traffic control. Sumner et al. (1984) determined a method of calculat ing PCE values between consecutive signalized intersections on urban arterial ro ads using microscopic simulation, NETSIM. The values are derived from the ve hiclehours of road utilization that are added when large vehicles are introduced to t he traffic stream. The study concluded that PCE is lower for better levels of service, specifically PCE values at LOS B are less than the PCE values at LOS D.
PAGE 35
20 2.1.8 PCEs Based on Platoon Formation Platooning occurs when the fast vehicles catch the slower vehicles such trucks, buses and recreational vehicles, and the fast vehicles are not able to pass. This often occurs on rural twolane highways or on upgrade multilane highways where high traffic flow make lane changing and overtaking di fficult. Van Aerder and Yagar (1983) studied PCE in both platoon leadership and fol lower creation. Large vehicles have a tendency to be platoon leaders. They analyzed larg e vehicles using the ratio of percent leads, by vehicle type, and to percentage of total mainline traffic to obtain PCE by normalizing those ratios to the original ratio of passenger cars. In the study, there were relative effects of trucks in the creation of platoon followers. The follower production rates were higher in the high volume area for all types of vehicles and both directions of travel, which results in smaller PCE. Here is the model of number of followers using separate multiple l inear terms: Number of followers = B 0 + B 1 (cars) + B 2 (trucks) + B 3 (RVs) + B 4 (other vehicles) + B 5 (opposing vehicles, mainline vehicles) (2.13) Where B 0 = the constant number of followers in a platoon on t wolane highways,
PAGE 36
21 B 1 to B 5 = indicate the rate at which the number of followe rs increases for each traffic volume component B 1 to B 4 = the number of additional followers produced per v ehicle. B 5 = the number of followers produced per opposing vehicl e at a mainline volume of 1,000 vph. Thus, PCEs = Bn / B 1 (2.14) Where PCEs = Passenger Car Equivalent B n = the constant number of followers in a platoon on tw olane highways, B 1 to B 5 = indicate the rate at which the number of followe rs increases for each traffic volume component B 1 to B 4 = the number of additional followers produced per v ehicle. B 5 = the number of followers produced per opposing vehicl e at a mainline volume of 1,000 vph. 2.1.9 PCEs Based on Travel Time Travel time is the total time spent by vehicles from one point to another point under prevailing conditions. It includes acceleration, decelera tion, and stopped time. Keller and Saklas (1984) utilized a macroscopic traffic simulation TRANSYT/7N for estimating PCE for large vehicles operating on urban ar terial networks as a function
PAGE 37
22 of traffic volume, vehicle classification, and signal sett ing. The premise was that the capacity reducing effect of the larger vehicles is related directly to the additional delay from such vehicles when compared to the all pass enger cars. The outcomes showed that PCE values increase as vehicle get larger an d as signalization approaches the maximum. The PCE values were estimated ov er a wide range of flow rates to approximately simulate levels of service fro m A to F. In addition, the PCE values were relatively constant for most of LOSs un til the volume approach LOS F at which the PCE values significantly increase. 2.2 Truck Equivalencies in the 2010 HCM In the 2010 HCM, PCEs are given based on the percent a nd length of grade and proportion of heavy vehicles, and terrain, as shown on Tables 2.1, 2.2, and 2.3. The extended freeway segment includes specific upgrades and d owngrades. Table 2.1 PCEs for Heavy Vehicles in General Terrain Se gments (HCM, 2010)
PAGE 38
23 Table 2.2 PCEs for Trucks and Buses ( E T ) on Specific Downgrades (HCM, 2010) Table 2.3 PCEs for Trucks and Buses ( E T ) on Upgrades (HCM, 2010)
PAGE 39
24 2.3 Traffic Flow Parameter Concepts The Highway Capacity Manual 2010 defines basic concepts for uninterruptedflow facilities as in volume, flow rate, speed, density, headw ay, and capacity. The definitions are following: Volume Â– the total number of vehicles or other roadwa y users that pass over a given point or section of a lane or roadway during a given time interval, often 1h. Flow rate Â– the equivalent hourly rate at which vehicles or other roadway users pass over a given point or section of a lane or roa dway during a given time interval of less than 1 hour, usually 15 minutes. Speed Â– a rate of motion expressed as distance per unit of time. Density Â– the number of vehicles occupying a given length o f a lane or roadway at a particular instant. Headway Â– the time between successive vehicles as they pass a po int on the roadway, measured from the same common feature of bo th vehicles (for example, the front axle or the front bumper). Figure 2.1 shows the relationship among density, speed and flow rate, and headway and spacing. The flowdensity graph is placed dire ctly below the speeddensity because of the common horizontal scales. Similar ly, the speedflow graph is place directly next to the speeddensity graph be cause of the common vertical scales. The figure illustrates that a zero flow rate occurs under two different conditions: stable and unstable flows. Stable condition or undersaturated flow is when there are no vehicles on the segment, the density and flow rate is zero. Unstable condition or oversaturat ed flow is when density becomes very high, all vehicles must stop, no movement can occur, and speed
PAGE 40
25 declines due to the vehicle interaction. When capacity i s reached, the density and speed is in the maximum flow rate. Figure 2.1 GENERALIZED SPEEDFLOWDENSITY RELATIONS HIP ON UNINTERRUPTEDFLOW FACILITIES (HCM, 2010) Truck acceleration speed especially on the ramps and on the steep grades, affects the passenger car equivalent. As speed decreases the truck PC E increases. Table 2.4 shows the relationship between speed and stopping d istance for heavy trucks.
PAGE 41
26 Table 2.4 SpeedDistance Relationship for Acceleration of Heavy Truck (AASHTO A Policy on Geometric Design of Highways and Streets, 2000 ) Bassok et al. (2009) concluded that (regardless of the m ethod) truck speeds are consistently slower than passenger vehicle speeds, though th e size difference is dependent on the chosen methodology, specific facility, ti me of day and congested direction. Overall, they found that heavy vehicle speed i s ten percent slower than passenger car speed on freeways. 2.4 Level of Service (LOS) LOS is commonly used to measure traffic operation of road way network, including freeways, ramps, surface streets and intersections. It is a grading system based on speed, delay, and travel time. LOS is defined into 6 ca tegories ranging from Â“AÂ”, ideal conditions to Â“FÂ”, extreme delays, as shown on Figu re 2.2.
PAGE 42
27 Figure 2.2 LOS FOR BASIC FREEWAY (HCM, 2010) The HCM 2010 describes in more detail the types of traf fic flow on basic freeway segments, which for undersaturated, queue discharge, a nd oversaturated flow conditions as shown in Figure 2.3: Undersaturated flow represents conditions under which the traffic stream is unaffected by upstream or downstream bottlenecks. Queue discharge flow represents traffic flow that has ju st passed through a bottleneck and is accelerating back to driversÂ’ desired speeds for the prevailing conditions. As long as another downstream b ottleneck does not exist, queue discharge flow is relatively stable until t he queue is fully discharged. Oversaturated flow represents the conditions within a que ue that has backed up from a downstream bottleneck. These flow conditio ns do not reflect the prevailing conditions of the site itself, but rather the consequences of a downstream problem. All oversaturated flow is considered to be congested.
PAGE 43
28 Figure 2.3 THREE TYPES OF FREEWAY (HCM 2010) 2.5 Vehicle Classification Federal Highway Administration (FHWA) divides vehicles int o 15 categories. Therefore, classification data are necessary to define in the beginning of studies, since they would be useful for predicting the commodity f low, the loads, and freight movements as shown on Figure 2.4. FHWA vehicle classes with definitions are following: 1. Motorcycles All two or threewheeled motorized vehicles. Typica l vehicles in this category have saddle type seats and are steered by ha ndlebars rather than steering wheels. This category includes motorcycles, moto r scooters, mopeds, motorpowered bicycles, and threewheel motorcycles. 2. Passenger Cars All sedans, coupes, and station wagons manufactured primarily for the purpose of carrying passengers and incl uding those passenger cars pulling recreational or other light tra ilers. 3. Other TwoAxle, FourTire Single Unit Vehicles All twoaxle, fourtire, vehicles, other than passenger cars. Included in this classifi cation are pickups, panels, vans, and other vehicles such as campers, motor homes, ambulances, hearses, carryalls, and minibuses. Other twoaxle, fourtire singleunit vehicles pulling recreational or other ligh t trailers are included in this classification.
PAGE 44
29 4. Buses All vehicles manufactured as traditional passengercar rying buses with two axles and six tires or three or more axles. Th is category includes only traditional buses (including school buses) functioning as passengercarrying vehicles. 5. TwoAxle, SixTire, SingleUnit Trucks All vehicles on a single frame including trucks, camping and recreational vehicles, motor homes, etc., with two axles and dual rear wheels. 6. ThreeAxle SingleUnit Trucks All vehicles on a single frame including trucks, camping and recreational vehicles, motor homes, etc. with three axles. 7. Four or More Axle SingleUnit Trucks All trucks on a single frame with four or more axles. 8. Four or Fewer Axle SingleTrailer Trucks All vehicles with four or fewer axles consisting of two units, one of which is a tractor o r straight truck power unit. 9. FiveAxle SingleTrailer Trucks All fiveaxle vehicles consisting of two units, one of which is a tractor or straight truck power unit. 10. Six or More Axle SingleTrailer Trucks All vehicles with six or more axles consisting of two units, one of which is a tractor or strai ght truck power unit. 11. Five or fewer Axle MultiTrailer Trucks All vehicles with five or fewer axles consisting of three or more units, one of which is a tractor or straight truck power unit. 12. SixAxle MultiTrailer Trucks All sixaxle vehicles consisting of three or more units, one of which is a tractor or straight truck po wer unit. 13. Seven or More Axle MultiTrailer Trucks All vehicles with seven or more axles consisting of three or more units, one of which is a tractor or straight truck power unit. In addition, FHWA provide the following criteria for reporting information on trucks: 1. Truck tractor units traveling without a trailer wi ll be considered singleunit trucks. 2. A truck tractor unit pulling other such units in a "sa ddle mount" configuration will be considered one singleunit truck and will be def ined only by the axles on the pulling unit. 3. Vehicles are defined by the number of axles in conta ct with the road. Therefore, "floating" axles are counted only when in the down position. 4. The term "trailer" includes both semiand full tr ailers.
PAGE 45
30 Figure 2.4 FHWA VEHICLE CLASSIFICATIONS (FHWA, 1985 )
PAGE 46
31 Another major consideration is the operating characteri stics of different vehicle types and dimensions, turning radii and offtracking, resistan ce to motion, power requirements, acceleration performance, and deceleratio n performance. Motor vehicles, including passenger cars, trucks, vans, buses, recre ational vehicles, and motorcycles, have unique weight, length, size, and operati onal characteristics as shown on Table 2.5. Table 2.5 Vehicle Weight and Power (Traffic Engineer ing Handbook, 2009) 2.6 Simulation Models Study Since various microscopic models exist for the simulation of traffic flow, several micro simulation software packages were reviewed for fea sibility in calculating PCEs. Bloomberg and Dale (2000) observed that overall CORS IM and VISSIM are more similar than they are different. Both models are desig ned to model any combination of surface street and freeway facilities, including most signal control and other operational strategies. Both models provide detailed an d focused output, both in tabular format and via animated graphics. The main dif ferences between the two models are in vehicle and driver behavior, primarily in the carfollowing and gap acceptance logic.
PAGE 47
32 3. Methodology 3.1 PCE Methods Theoretical Analysis Evaluation Among all the truck equivalencies approaches described on the literature reviews, the speedflowdensity method is used for this study becaus e: Speedflowdensity is able to measure traffic flow a nd the randomness in the traffic stream, and its validity has been proven by empi rical research with field observation. Speed is a performance measure experienced directly by d rivers on the roadway and provides vivid pictures of traffic flow. Density is a critical parameter for uninterruptedflow facilities and characterizes the quality of traffic operations. It descr ibes the proximity of vehicles to one another and reflects the freedom to man euver within the traffic stream (HCM, 2010). Flow rate represents the number of vehicles that pass a certain section per time unit, commonly in vehicle per hour. The basic concept of this PCE derivation was proposed b y Huber using a deterministic model of traffic flow from GreenshieldÂ’s M odel. Greenshield developed an uninterrupted traffic flow model based on the re lationship of speed, flow and density. As shown on Figure 3.1, the relationship betwe en speed and density are
PAGE 48
33 relatively linear. When density is zero, the flow is zer o, and speed approaches free flow speed, since there is no traffic on the roadway as in Figure 3.2. When density increases, the flow increases as well until it reaches so me maximum flow conditions, called the jam density, which occurs when all vehicles must stop and the speed and flow rate becomes zero. Sumner et al. developed the r elationship described by Huber which includes multiple truck types, followed by Demarchi and Setti. The limitation of deriving PCE in the traffic stream with multiple truck types has found in Demarchi and SettiÂ’s study. To avoid the limitation of n ot counting the interaction between trucks, Demarchi and Setti developed a new deri vation of PCE. Figure 3.1 RELATIONSHIP BETWEEN SPEED AND DENSITY ( Greenshield, 1934)
PAGE 49
34 Figure 3.2 RELATIONSHIP BETWEEN SPEED (v) AND FLOW ( q) (Greenshield, 1934) 3.2 Simulation 3.2.1 Decision Support Methodology (DSM) Federal Highway Administration (FHWA) produced a manu al and worksheet that can be used to select and identify the appropriate type traffic analysis tool for operational improvements. Decision Support Methodology (DSM) was app lied into this study to select the correct measurable approach for the freight movement traffic on freeway. Next is the following DSM process for this particular stud y:
PAGE 50
35 Step 1 : Analysis Context Step 2 : Geographic Scope Step 3 : Facility Type Step 4 : Travel Mode Step 5 : Management Strategy/Application Step 6 : Traveler Response Step 7 : Performance Measure Step 8 : Tool/Cost Effectiveness Step 9 : Criteria Weights Among the 7 tools in DSM, the two appropriate tools for this study are either Microsimulation or a Travel Demand Model (TDM) as shown on Table 3.1. Travel Demand could not apply for this analysis because the objective of this study for understanding the dynamic truck traffic in traffic operational analysis, n ot in dynamic assignment equilibrium of origin and destination trip planning pattern. The most appropriate tool is Microsimulation since it is based on driver carfollowi ng and lanechanging behavior. Vissim micro simulation is the tool used in thi s study.
PAGE 51
36 Table 3.1 Decision Support Methodology (DSM) Result
PAGE 52
37 The DSM results tool uses the following FHWA definitio n: Travel Demand Models Predicting travel demand, traffic operations, and impact s in response to operational strategies requires specific analytical capab ilities, such as the prediction of travel demand and the consideration of d estination choice, mode choice, timeofday travel choice, and route choice, as well as the representation of traffic flow in the highway network These attributes are found in the structure and orientation of travel dema nd models, mathematical models that forecast future travel demand from curren t conditions, and future projections of household and employment characteristics. Travel demand models were originally developed to determine the be nefits and impacts of major highway improvements in metropolitan areas. Toda y, travel demand models are used in more wideranging tasks, including development of transportation master plans, evaluation of proposed landuse changes, initial design of transportation facilities, evaluation of air quality impacts, and assessment of future transportation needs. However, thes e tools were not designed to evaluate travel management strategies, suc h as ITS and operational strategies. Travel demand models have onl y limited capabilities to accurately estimate changes in operational characterist ics (such as speed, delay, and queuing) resulting from implementation o f ITS/operational strategies. These inadequacies generally occur because o f the poor representation of the dynamic nature of traffic in tra vel demand models. Microscopic simulation Microscopic simulation models simulate the movement of ind ividual vehicles, based on theories of carfollowing and lanechanging. Typically, vehicles enter a transportation network using a statistical distri bution of arrivals (a stochastic process), and are tracked through the network o ver small time intervals (e.g., one second or fraction of a second). Typically, upon entry, each vehicle is assigned a destination, a vehicle type, an d a driver type. In many microscopic simulation models, the traffic operationa l characteristics of each vehicle are influenced by vertical grade, horizonta l curvature, and superelevation, based on relationships developed in pri or research. The primary means of calibrating and validating microscopic simu lation models is through the adjustment of driver sensitivity factors. Co mputer time and storage requirements for microscopic models are large, usu ally limiting the network size and the number of simulation runs that cou ld be completed.
PAGE 53
38 3.2.2 VISSIM Micro Simulation VISSIM model uses inputs such as lane assignments and ge ometries, intersection turning movement volumes, vehicle speeds, percentages of veh icles by type, and pretimed and/or actuated signal timing. There are t wo approaches to modeling the traffic in VISSIM: static route and dynamic assignment. The basic concept for the static route approach is Car Following Logic or named Wiedemann 1974 for urban arterials, which assumes that Â“the driver of a faster movin g vehicle starts to decelerate as the driver reaches his individual percepti on threshold of a slower moving vehicle. Since the driver cannot exactly determine th e speed of that vehicle, his speed will fall below that vehicleÂ’s speed until he starts to slightly accelerate again after reaching another perception thresholdÂ” ( PTV AG manual 2008, page 26) Another approach is Wiedemann 1999 for freeways. The ca r following and lane change logic is similar to urban arterial, and includes ve hicle headways. The dynamic assignment model in VISSIM chooses routes based on dynamic factors, such as the impacts of variable message signs or the potenti al traffic diversion into neighborhoods for networks up to the size of medium size d cities. When using dynamic assignment travel demand is not specified by using veh icle input on selected links with a given volume but in the form of an origindestination matrix. Traffic starting at a parking lot is similar to traffic g enerated by vehicle inputs, but the composition of the traffic explicitly is specified for the parking lot. The approach used in VISSIM is based on the Wiedeman n model of psychophysical driving behavior:
PAGE 54
39 Free driving is when no influence of preceding is vehicles observab le. Approaching is the process of adapting the driverÂ’s own speed to the lower speed of a preceding vehicle. Following is when the driver follows the preceding car without an y conscious acceleration or deceleration. Braking is the application of medium to high deceleration ra tes if the distance falls below the desired safety distance. For the purposes of this study, the Wiedemann 1999 met hod for interurban or freeway traffic is used, such that: CC0 (Standstill distance) defines the desired distance between stopped cars with no variation. CC1 (Headway time) is the time in seconds that a driver wants to keep. Where: v = given speed dx_safe = safety distance or clear space desired by the d river (ft/ses) CC2 (Following variation) is how much more distance than the desired safety distance a driver allows before he intentionally moves clos er to the car in front. CC3 (Threshold for entering Â‘FollowingÂ’) defines how many seconds before reaching the safety distance the driver starts to decelerat e. CC4 and CC5 (Â‘FollowingÂ’ thresholds) control the speed differences during the Â‘FollowingÂ’ state. CC6 (Speed dependency of oscillation) influences the distance on speed oscillation while in following process.
PAGE 55
40 CC7 (Oscillation acceleration) is actual acceleration during the oscillation process. CC8 (Standstill acceleration) is desired acceleration when starting from standstill CC9 (Acceleration) is desired acceleration. Lookback Distance is the distance that a vehicle can see b ackwards in order to react to other vehicles behind (within the same lin k). Similar to Siuhi and Mussa (2007) findings Â“the VISSIM default parameters used in trial simulation runs for checking any coding error fo r their study, showed that the default model parameters were incorrectly emulating th e existing traffic flow characteristics. The necessitated calibrating the model by tuning carfollowing and lanechanging parameters while comparing simulation da ta with field data. The fine tuning process involved iterative parameter changing an d simulating until simulated speeds closely matched speeds observed in the fieldsÂ”. F urthermore, the simulated values were verified against the observed field values a s indicated in the fundamental flow diagrams shown in Figure 3.3.
PAGE 56
41 Figure 3.3 FUNDAMENTAL TRAFFIC FLOW DIAGRAMS CASE S TUDY AT I95 SOUTH FLORIDA (SIUHI AND MUSSA, 2007) Due to unrealistic vehicle behavior in comparison to th e field observation when using VISSIM default parameters, several cars following and lane changing parameters have been changed, based on intuitive engineering kno wledge and best practices (Marlina and Janson, 2011). These car following paramet ers include CC1 (headway time) or the time in seconds that a driver wants to maintain from the car ahead, and CC2 (following variation) or desired safety distance that a driver allows before moving closer to the car ahead. A value of 1.12 seconds w as used for CC1 instead of the default value of 0.90 seconds. A value of 14 .50 feet was used for CC2 since drivers are more cautious when the roadway is congested, a nd the maximum deceleration for cooperative braking is 20 ft/sec 2 The VISSIM default value for
PAGE 57
42 waiting time before diffusion of 60 seconds was increased to 300 seconds to allow for more congested traffic. When distance between vehicl es increases, it will have more impact on capacity. Tables 3.2 and 3.3 present car following and lane changing parameters used in the VISSIM model and the default values. Table 3.2 Car Following VISSIM Parameters Car Following Parameter Default Used CC0 (Standstill Distance) Â– ft 4.92 4.92 CC1 (Headway Time) Â– seconds 0.90 1.12 CC2 (Following Variation) Â– ft 13.12 14.50 CC3 (Threshold for entering 'Following') 8.00 8.00 CC4 (Negative 'Following' Threshold) 0.35 0.35 CC5 (Positive 'Following' Threshold) 0.35 0.35 CC6 (Speed Dependency of Oscillation) 11.44 11.44 CC7 (Oscillation Acceleration) Â– ft/s 2 0.82 0.82 CC8 (Standstill Acceleration) Â– ft/s 2 11.48 9.65 CC9 (Acceleration at 50 mph) Â– ft/s 2 4.92 4.92 Table 3.3 Lane Changing VISSIM Parameters Lane Change Default Own Trailing Veh Used Own Used Trailing Veh Maximum Deceleration ft/sc2 13.12 9.84 13.12 9.84 1 ft/s2 per distance 200.00 200.00 100.00 100.00 Accepted Deceleration ft/s2 3.28 1.64 3.28 1.64 Waiting Time Before Diffusion Seconds 60 300 Min. Headway (front/rear) 1.64 1.64 Safety Distance Reduction Factor 0.6 0.1 Maximum Deceleration for Cooperative Braking ft/sc2 9.84 20.0
PAGE 58
43 The VISSIM model in this study uses vehicle inputs at the st art of the network and routing decisions throughout the network to determine v ehicle flows. In the simulation process, the following main assumption was made: the fr eeway was considered homogenous meaning that all vehicles entering the segmen t exit at the end of the segment and no traffic enters and exits in the middle o f the segment. 3.2.3 Calibration, Validation and Verification In order to get a reasonable match between the obser ved and expected traffic, the calibration was performed. Calibration is a process to d etermine whether the conceptual simulation model is realistic or accurate repr esentation of the actual condition of the system study or not. In addition, verifi cation was performed to ensure the input data and model is appropriate for the study conditions. In Federal Highway Administration guidelines for applying traffic micro sim ulation modeling software, freeway traffic volume calibration is required to have a n accurate model using either percent differences or Geoffrey E. Havers (GEH) statistics f ormula. For these purposes, less than 10 percent differences and 5 GEH stat istics are required in this study. A GEH is a nonlinear standard measure of the goodness of fit. When GEH value is less than 5.0, it is considered a good match betw een the model and the actual traffic counts. Applying a stochastic model, it is not likely to get exact volume matches, but very close. In Marlina and Janson (2011) In terstate 70 Simulation study show the calibration and validation of traffic volumes p erformed in table 3.4. The network and VISSIM model for this study using I70 mode l calibrated for all parameters.
PAGE 59
44 Table 3.4 Interstate 70 Simulation Calibration and V alidation 1 Differences between models estimated volumes and actual volumes 2 GEH statistics: E = model estimated volume and V = Field Counts
PAGE 60
45 3.3 Steps Performed to Develop Truck Equivalence Fa ctors The methods performed to estimate PCE are based on Eq ual FlowDensity (Dermarchi & Setti, 2003) and Equal SpeedDensity (Sum ner et al., 1984). 3.3.1 Equal FlowDensity The following procedures were used to determine the i mpedance of the basevehicle flow rate (q B ) and mixedvehicle flow rate (q M ) in the traffic stream: 1. Establish a speedflowdensity relationship for the base vehicles or passenger car only. This is obtained from VISSIM microscop ic simulation. The flow rates correspond to the maximum service flow ra te for each LOS category from the HCM 2010. Figure 3.4 FLOW Â– DENSITY RELATIONSHIP FOR BASE VEH ICLES
PAGE 61
46 2. Similar to previous step, generate a speedflowde nsity relationship for the mixed vehicle stream with an equal number of trucks. Figure 3.5 FLOW Â– DENSITY RELATIONSHIP FOR MIXED VE HICLES 3. Interpolate between observed values to obtain the b ase flow rate and mixed vehicle flow rate at an equal density value. Initially, use the density at LOS C, at an equal value of22 pc/mi/ln.
PAGE 62
47 Figure 3.6 INTERPOLATION FLOW Â– DENSITY RELATIONSHI P FOR q B and q M 4. Calculate PCE using below equation based on the sam e level of impedance as shown on Figure 3.3 : Where = proportion of trucks of type i out of all trucks n in t he mixed traffic flow = base flow rate (passenger cars only) = mixed flow rate (passenger cars + trucks) = passenger equivalent of trucks
PAGE 63
48 3.3.2 Equal SpeedFlow The following procedures were used to determine the i mpedance of base vehicles flow rate (q B ), mixed vehicles flow rate (q M ), and subject vehicles (q S ) in the traffic stream: 1. Generate the relationship between impedance and fl ow rate for the passenger car only (q B ). The results are obtained from VISSIM microscopic simulation. The flow rates correspond to the maximum se rvice flow rate for each LOS category from HCM 2010. Figure 3.7 FLOW Â– DENSITY RELATIONSHIP FOR BASE VEH ICLES 2. Similar to Step 1, generate a speedflowdensity r elationship for the mixed vehicle stream with an equal number of trucks, containing (1 Â– p) passenger cars and trucks.
PAGE 64
49 Figure 3.8 FLOW Â– DENSITY RELATIONSHIP FOR MIXED VE HICLES 3. Similar to Step 1, generate the relationship bet ween impedance measure and flow rate for the stream resulting from replacing p passenger cars with p subject vehicles in the mixed vehicle stream. Figure 3.9 FLOW Â– DENSITY RELATIONSHIP FOR SUBJECT VEHICLES
PAGE 65
50 4. Interpolate between observed values to obtain the b ase flow rate (q B ), mixed flow rate (q M ), and subject flow rate (q S ) at an equal density value. Initially, the density at LOS C, at an equal value of 22 pc/mi/ln. Figure 3.10 INTERPOLATION FLOW Â– DENSITY RELATIONSH IP FOR q B q M and q S 5. Calculate PCE using below equation based on the sam e level of impedance as shown on Figure 3.4 :
PAGE 66
51 Where = equivalent passenger car only flow rate = mixed flow rate = additional subject flow rate = truck proportion in the mixed traffic flow = proportion of subject vehicles = truck PCE 3.4 Scenarios To analyze the impact of the proportion of trucks in t he traffic stream, truck proportions from 5% to 100% trucks were simulated for l engths of grade of 0.5 and 1.0 miles. The network segment was constructed for 2%, 4%, 6%, and 8% grades. All of these combinations were applied to road segments with 2lanes, 3lanes without trucks restrictions, and 3lanes with trucks restrict ions, resulting in 264 combinations as shown on the table 3.5.
PAGE 67
52 Table 3.5 Truck Percentages Scenarios % Truck Length of grade (mi) Grade Description 5 0.5 2% 2Lane 10 1.0 4% 3Lane 20 6% 3Lane with truck restriction 30 8% 40 50 60 70 80 90 100 11 2 4 3 264 To analyze the impact of length of grade, grade dista nces of 0.25 to 5 miles were simulated, with 5% and 10% trucks in the traffic stream. The network segment was constructed for 2%, 4%, 6%, and 8% grades. All of these combinations were applied to road segments with 2lanes, 3lanes without trucks restr ictions, and 3lanes with trucks restrictions, resulting in 240 combinations as shown on the table 3.6.
PAGE 68
53 Table 3.6 Length of Grade Scenarios Length of grade (mi) Truck Grade Description 0.25 5% 2% 2Lane 0.50 10% 4% 3Lane 0.75 6% 3Lane with truck restriction 1.00 8% 1.25 1.50 2.00 3.00 4.00 5.00 10 2 4 3 240 To analyze the impact of the grade, grade percentages from 1% to 10% were simulated, with 5% and 10% trucks in the traffic stream and lengths of grade from 0.25 to 2 miles. All of these combinations were appli ed to road segments with 2lanes, 3lanes without trucks restrictions, and 3lanes wi th trucks restrictions, resulting in 420 combinations as shown on the table 3. 7.
PAGE 69
54 Table 3.7 Grade Percentages Scenarios % Grade Length of grade (mi) Truck Description 1 0.25 5% 2Lane 2 0.50 10% 3Lane 3 0.75 3Lane with truck restriction 4 1.00 5 1.25 6 1.50 7 2.00 8 9 10 10 7 2 3 420
PAGE 70
55 4. Analysis and Results 4.1 Overview Federal Highway Administration classified vehicles into 13 categories. Heavy vehicles fall from category four to thirteen based on n umber of axles, length, width, and power units. To simplify multiply truck types in the traffic stream, in this study trucks are divided into 3 categories and specified into t he VISSIM model: 1. Single Unit Trucks (SUT) (vehicle type from numbers 4 to 7, FHWA) with length of 33.51 feet and width of 4.92 feet. 2. Medium Trucks (vehicle type from numbers 8 to 10, FHWA) with length of 41.59 feet and width of 4.92 feet. 3. Longer Combination Vehicles (LCV) (vehicle type fro m numbers 11 Â– 13, FHWA) with length of 54.14 feet and width of 4.92 f eet. In addition passenger car is categorized as vehicle type number 2 according to FHWA vehicles classification. VISSIM requires a number of input variables. Desired s peed is an important parameter affecting traffic flow in VISSIM. The desi red speed distributions specified in this model were: 50 Â– 70 mph for passenger cars and 40 Â– 60 mph for trucks. The level of congestion and roadway geometry dictates the v ehicle speeds within these
PAGE 71
56 ranges. A speed decision is placed on the entry link. The simulation was conducted at five different saturation flow rates corresponding t o the maximum service flow rates in the HCM 2010: 820 pc/h/ln (LOS A), 1310 pc/h /ln (LOS B), 1750 pc/h/ln (LOS C), 2110 pc/h/ln (LOS D), and 2400 pc/h/ln (LO S E). The VISSIM model in this study was used the static routing decision to direct vehi cles. The random seed for generating identical result used is 42. It is the defa ult value in VISSIM simulation parameter. The base and mix flow rates, speed, and d ensity were obtained as the output from VISSIM simulation. 4.2 SpeedFlowDensity Relationship Figures 4.1 to 4.8 show the speedflowdensity relation ship obtained from VISSIM traffic simulation. Separate curves were plotted for di fferent truck percentages, various lengths of grades, different percent grades (2 %, 4%, 6%, and 8%), and different mixtures of base and mixed vehicles in the traff ic stream. 4.2.1 Trucks Proportion The speeddensityrelationship was obtained from VISSIM for different truck proportion, from 5 to 100 percent trucks, with 10% in crements when mixed vehicles occur in the traffic stream, as well as for base vehicle. The roadway characteristics are used for the 0.5 and 1.0 mi.
PAGE 72
57 Figures 4.1 and 4.2 illustrate speeddensityrelationsh ip for 0.5 mi with 5% and 10% trucks. Figures 4.3 and 4.4 illustrate speeddensityrelat ionship for 1.0 mi length of grade with 5% and 10% trucks.
PAGE 73
58 Figure 4.1 SPEEDFLOWDENSITYRELATIONSHIP FOR 0.5 MI 5% TRUCKS
PAGE 74
59 Figure 4.2 SPEEDFLOWDENSITYRELATIONSHIP FOR 0.5 MI 10% TRUCKS
PAGE 75
60 Figure 4.3 SPEEDFLOWDENSITYRELATIONSHIP FOR 1.0 MI 5% TRUCKS
PAGE 76
61 Figure 4.4 SPEED FLOW DENSITY RELATIONSHIP FOR 1.0 MI 10% TRUCKS
PAGE 77
62 4.2.2 Length of Grade The lengths of grade were measured from 0.25 to 5 mil es for 5% and 10% trucks in mixed traffic. Figures 4.5 and 4.6 illustrate the exampl e of speeddensityrelationship for 3.0 and 4.0 miles length of grade with 5% trucks. Figures 4.7 and 4.8 illustrate the example of speeddensityrelationship for 3.0 and 4.0 miles length of grade with 10% trucks.
PAGE 78
63 Fig ure 4.5 SPEED FLOW DENSITY RELATIONSHIP FOR 3.0 MI 5% TRUCKS
PAGE 79
64 Figure 4.6 SPEED FLOW DENSITY RELATIONSHIP FOR 4.0 MI 5% TRUCKS
PAGE 80
65 Figure 4.7 SPEED FLOW DENSITY RELATIONSHIP FOR 3.0 MI 10% TRUCKS
PAGE 81
66 Figure 4.8 SPEED FLOW DENSITY RELATIONSHIP FOR 4.0 MI 10% TRUCKS
PAGE 82
67 4.2.3 Grade Percentages The grade percentages were measured from 1 to 10 per cent, for 5% and 10% trucks in the traffic stream. Figure 4.9 illustrates the exam ple of speeddensityrelationship for 1.5 mile length of grade with 5% trucks. Figure 4.10 illustrates the example of speeddensityrel ationship for 2.0 miles length of grade with 5% trucks. Figure 4.11 illustrates the example of speeddensityrel ationship for 1.5 miles length of grade with 10% trucks. Figure 4.12 illustrates the example of speeddensityrel ationship for 2.0 miles length of grade with 10% trucks.
PAGE 83
68 Figure 4.9 SPEED FLOW DENSITY RELATIONSHIP FOR 1.5 MI 5% TRUCKS
PAGE 84
69 Figure 4.10 SPEED FLOW DENSITY RELATIONSHIP FOR 2.0 MI 5% TRUCKS
PAGE 85
70 Figure 4.11 SPEED FLOW DENSITY RELATIONSHIP FOR 1.5 MI 10% TRUCKS
PAGE 86
71 Figure 4.1 2 SPEED FLOW DENSITY RELATIONSHIP FOR 2.0 MI 10% TRUCKS
PAGE 87
72 4.3 PCE Results 4.3.1 Variance PCE by Trucks Proportion The HCM 2010 considers proportion of trucks only up to 2 5 percent. However, on many freeways in the United States, the proportion of t rucks exceeds 25 percent. To test the effect of high truck percentages, the trucks prop ortion was considered in increments of 10%, started from 5 percent all the way to 100 percent trucks. At the lower proportion of trucks (from 0% to approximately 30 % or 25%), the percentage categories used in the HCM 2010 were matched for compar ison. One aim of this study is to verify that the PCE decreases as the proportion of trucks increases. Even though this trend is observed in the HCM 2010, it has not been tested at high trucks percentages. An additional substant ial aim is compare Demarchi & Setti (flowdensity based) and Sumner et al. (speedflow based) methods for calculating PCEs.
PAGE 88
73 Figure 4.13 shows PCE variability with truck proportion f or 0.5 mile and 2% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.13 PCE VARIABILITY BY TRUCKS PROPORTION Â– 0.5 Mi Â– 2% GRADE nr # $r% n r # $r%
PAGE 89
74 Figure 4.14 shows PCE variability with truck proportion f or 0.5 mile and 4% grade using Demarchi & Setti and Sumner methodologies. (a) Demarchi & SettiÂ’s Method\ (b) Sumner et alÂ’s Method Figure 4.14 PCE VARIABILITY BY TRUCKS PROPORTION Â– 0.5 Mi Â– 4% GRADE n r # $r% n r # $r%
PAGE 90
75 Figure 4.15 shows PCE variability with truck proportion f or 0.5 mile and 6% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.15 PCE VARIABILITY BY TRUCKS PROPORTION Â– 0.5 Mi Â– 6% GRADE nr "# $r% n r # $r%
PAGE 91
76 Figure 4.16 shows PCE variability with truck proportion f or 0.5 mile and 8% grade using Demarchi & Setti and Sumner et al. methodolog ies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.16 PCE VARIABILITY BY TRUCKS PROPORTION Â– 0.5 Mi Â– 8% GRADE n r n! "# $r% nr n! # $r%
PAGE 92
77 Figure 4.17 shows PCE variability with truck proportion f or 1.0 mile and 2% grade using Demarchi & Setti and Sumner et al. methodolo gies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.17 PCE VARIABILITY BY TRUCKS PROPORTION Â– 1.0 Mi Â– 2% GRADE nr # $r% n r # $r%
PAGE 93
78 Figure 4.18 shows PCE variability with truck proportion f or 1.0 mile and 4% grade using Demarchi & Setti and Sumner et al. methodolog ies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.18 PCE VARIABILITY BY TRUCKS PROPORTION Â– 1.0 Mi Â– 4% GRADE n r "# $r% nr # $r%
PAGE 94
79 Figure 4.19 shows PCE variability with truck proportion f or 1.0 mile and 6% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.19 PCE VARIABILITY BY TRUCKS PROPORTION Â– 1.0 Mi Â– 6% GRADE n r "# $r% nr # $r%
PAGE 95
80 Figure 4.20 shows PCE variability with truck proportion f or 1.0 mile and 8% grade using Demarchi & Setti and Sumner et al. methodolog ies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.20 PCE VARIABILITY BY TRUCKS PROPORTION Â– 1.0 Mi Â– 8% GRADE n r n! "# $r% nr n! "# $r%
PAGE 96
81 The proportion of trucks in the traffic stream was foun d to have a significant effect on the calculated PCE. For higher truck percentages, the P CE decreases and levels off, except for Figure 4.15b. That is because of the limita tion of Sumner et al.Â’s method which does not include multiple trucks interaction. This t rend is justified because as the truck percentages increases, the interaction between trucks and passenger car decreases. The trucks will form platoons climbing the grad e. The interaction among trucks would be negligible as they have same performance s and operations. The variation in PCE for proportion of trucks between 25 and 50 percents provides substantial evidence. Hence, the HCM 2010 should be ext ended to include 50 percent trucks. Above 50 percent of trucks, the PCE shows v ery little variability. 4.3.2 Variance PCE by Length of Grade The HCM 2010 considers length of grade up to 1 mile fo r specific upgrades. Nevertheless, on many freeways in the United States, the length of grade exceeds 1 mile. To test the effect of length of grade, grades me asure in increments of 1 mile, started from 0.25 mile all the way to 5 miles length. At the lower percent grades (from 0% to approximately 6% or > 6%), the percent grades cat egories used in the HCM 2010 were matched for comparison. The aim is to test th e hypothesis that as the length of grade increases, the PCE increases and to com pare the results using Dermarchi & SettiÂ’s and Sumner et al.Â’s methodologies
PAGE 97
82 Figure 4.21 shows PCE variability with length of grade for 2% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. At the low grade lengths, the simulations agree well with the HCM model, but t he PCE continues to increase significantly. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.21 PCE VARIABILITY BY LENGTH OF GRADE Â– 2% GRADE r#&'()!% "# $r% r#&'()!% "# $r%
PAGE 98
83 Figure 4.22 shows PCE variability with length of grade for 4% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.22 PCE VARIABILITY BY LENGTH OF GRADE Â– 4% GRADE r#&'()!% # $r% r#&'()!% # $r%
PAGE 99
84 Figure 4.23 shows PCE variability with length of grade for 6% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.23 PCE VARIABILITY BY LENGTH OF GRADE Â– 6% GRADE n r#&'()!% # $r% n r#&'()!% "# $r%
PAGE 100
85 Figure 4.24 shows PCE variability with length of grade for 8% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.24 PCE VARIABILITY BY LENGTH OF GRADE Â– 8% GRADE n r#&'()!% n! # $r% n r#&'()!% n! "# $r%
PAGE 101
86 Figure 4.25 shows PCE variability with length of grade for 2% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.25 PCE VARIABILITY BY LENGTH OF GRADE Â– 2% GRADE r#&'()!% "# $r% r#&'()!% "# $r%
PAGE 102
87 Figure 4.26 shows PCE variability with length of grade for 4% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.26 PCE VARIABILITY BY LENGTH OF GRADE Â– 4% GRADE r#&'()!% "# $r% r#&'()!% "# $r%
PAGE 103
88 Figure 4.27 shows PCE variability with length of grade for 6% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.27 PCE VARIABILITY BY LENGTH OF GRADE Â– 6% GRADE n r#&'()!% "# $r% r#&'()!% # $r%
PAGE 104
89 Figure 4.28 shows PCE variability with length of grade for 8% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.28 PCE VARIABILITY BY LENGTH OF GRADE Â– 8% GRADE n r#&'()!% n! "# $r% n r#&'()!% n! "# $r%
PAGE 105
90 The PCE was observed to increase as the length of grade increases and matches the PCE provided in the HCM 2010. 4.3.3 Variance PCE by Grade Percentages The HCM 2010 mentions the effect of heavy vehicles on tr affic flow depends on terrain and grade conditions as well as traffic compositi ons. The HCM 2010 considers trucks equivalent for specific upgrades up to 6 per cent. Some freeways such Interstate 70 in Colorado have higher grade. To t est the effect of percent grade, grades measure in increments of 1%, started from 1 per cent all the way to 10 percent grades. The aim is to verify that the PCE incre ases as the percent grade increases and to compare the results using Dermarchi & Se ttiÂ’s and Sumner et al.Â’s methodologies.
PAGE 106
91 Figure 4.29 shows PCE variability with grade percentage s for 0.25 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.29 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi n r)! !" r r nr)! !" r r
PAGE 107
92 Figure 4.30 shows PCE variability with grade percentage s for 0.50 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.30 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi n r)! !" r r nr)! !" r r
PAGE 108
93 Figure 4.31 shows PCE variability with grade percentage s for 0.75 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.31 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi nr)! !" r r nr)! !" r r
PAGE 109
94 Figure 4.32 shows PCE variability with grade percentage s for 1.0 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.32 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi n r)! !" r r nr)! !" r r
PAGE 110
95 Figure 4.33 shows PCE variability with grade percentage s for 1.25 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.33 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi n r)! !" r r nr)! !" r r
PAGE 111
96 Figure 4.34 shows PCE variability with grade percentage s for 1.50 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi Â– SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.34 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi nr)! !" r r nr)! !" r r
PAGE 112
97 Figure 4.35 shows PCE variability with grade percentage s for 2.0 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.35 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi n r)! !" r r n r)! !" r r
PAGE 113
98 Figure 4.36 shows PCE variability with grade percentage s for 0.25 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.36 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi nr)! !" r r nr)! !" r r
PAGE 114
99 Figure 4.37 shows PCE variability with grade percentage s for 0.50 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.37 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi nr)! !" r r n r)! !" r r
PAGE 115
100 Figure 4.38 shows PCE variability with grade percentage s for 0.75 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.38 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi n r)! !" r r nr)! !" r r
PAGE 116
101 Figure 4.39 shows PCE variability with grade percentage s for 1.0 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.39 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi nr)! !" r r n r)! !" r r
PAGE 117
102 Figure 4.40 shows PCE variability with grade percentage s for 1.25 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.40 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi nr)! !" r r nr)! !" r r
PAGE 118
103 Figure 4.41 shows PCE variability with grade percentage s for 1.50 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.41 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi n r)! !" r r n r)! !" r r
PAGE 119
104 Figure 4.42 shows PCE variability with grade percentage s for 2.0 miles and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi Â– SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.42 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi
PAGE 120
105 From figures 4.36 to 4.42 presents the PCE values in s awtooth lines due to the randomness of simulation such as complex vehicular interact ions and their maneuvers. At the lower percent grades (from 0% to a pproximately 6% or > 6%), the percent grades categories used in the HCM 2010 were ma tched for comparison. The PCE was observed to increase as the percent grade increase d. 4.3.4 Variance PCE by Number of Lane The HCM 2010 does not calculate truck PCE based on numb er of lanes. It applies for any number of lanes. To test the effect of differen t number of lanes, this study compared the PCE values on twolane and threelane ro ad segments. On the VISSIM network the number of lanes was set for twolan es and threelanes using three consideration of truck percentages, length of gra de, and grade percentages. The comparison applied by using Dermachi & Setti and S umner et al. methodologies. One of the aims of this study is to verify that freeways w ith more than two directional lanes will have lower PCE than freeways with only two directional lanes such Interstate 70. 4.3.4.1 Truck Proportion Increasing one lane on the network from twoto three lane is estimated by truck proportion with the same variables of scenarios consider ed in this study.
PAGE 121
106 Figure 4.43 shows PCE variability with truck proportion f or 0.5 mile and 2% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.43 PCE VARIABILITY BY TRUCK PROPORTION Â– 0 .5 Mi 2% GRADE nr r nr r
PAGE 122
107 Figure 4.44 shows PCE variability with truck proportion f or 0.5 mile and 4% grade using Demarchi Â– Setti and Sumner methodologies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.44 PCE VARIABILITY BY TRUCK PROPORTION Â– 0 .5 Mi 4% GRADE nr r nr r
PAGE 123
108 Figure 4.45 shows PCE variability with truck proportion f or 0.5 mile and 6% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.45 PCE VARIABILITY BY TRUCK PROPORTION Â– 0 .5 Mi 6% GRADE nr r nr r
PAGE 124
109 Figure 4.46 shows PCE variability with truck proportion for 0.5 mile and 8% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.46 PCE VARIABILITY BY TRUCK PROPORTION Â– 0 .5 Mi 8% GRADE nr n! r nr n! r
PAGE 125
110 Figure 4.47 shows PCE variability with truck proportion f or 1.0 mile and 2% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.47 PCE VARIABILITY BY TRUCK PROPORTION Â– 1 .0 Mi 2% GRADE n r r n r r
PAGE 126
111 Figure 4.48 shows PCE variability with truck proportion f or 1.0 mile and 4% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.48 PCE VARIABILITY BY TRUCK PROPORTION Â– 1 .0 Mi 4% GRADE n r r nr r
PAGE 127
112 Figure 4.49 shows PCE variability with truck proportion for 1.0 mile and 6% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.49 PCE VARIABILITY BY TRUCK PROPORTION Â– 1 .0 Mi 6% GRADE n r r n r r
PAGE 128
113 Figure 4.50 shows PCE variability with truck proportion for 1.0 mile and 8% grade using Demarchi & Setti and Sumner et al. methodolog ies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.50 PCE VARIABILITY BY TRUCK PROPORTION Â– 1 .0 Mi 8% GRADE nr n! r nr n! r
PAGE 129
114 4.3.4.2 Length of Grade Figure 4.51 shows PCE variability with length of grade for 2% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.51 PCE VARIABILITY BY LENGTH OF GRADE Â– 2% GRADE r#&'()!% r r#&'()!% r
PAGE 130
115 Figure 4.52 shows PCE variability with length of grade for 4% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.52 PCE VARIABILITY BY LENGTH OF GRADE Â– 4% GRADE r#&'()!% r r#&'()!% r
PAGE 131
116 Figure 4.53 shows PCE variability with length of grade for 6% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.53 PCE VARIABILITY BY LENGTH OF GRADE Â– 6% GRADE r#&'()!% r r#&'()!% r
PAGE 132
117 Figure 4.54 shows PCE variability with length of grade for 8% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.54 PCE VARIABILITY BY LENGTH OF GRADE Â– 8% GRADE r#&'()!% n! r r#&'()!% n! r
PAGE 133
118 Figure 4.55 shows PCE variability with length of grade for 2% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.55 PCE VARIABILITY BY LENGTH OF GRADE Â– 2% GRADE r#&'()!% r r#&'()!% r
PAGE 134
119 Figure 4.56 shows PCE variability with length of grade for 4% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.56 PCE VARIABILITY BY LENGTH OF GRADE Â– 4% GRADE r#&'()!% r r#&'()!% r
PAGE 135
120 Figure 4.57 shows PCE variability with length of grade for 6% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi Â– SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.57 PCE VARIABILITY BY LENGTH OF GRADE Â– 6% GRADE r#&'()!% r r#&'()!% r
PAGE 136
121 Figure 4.58 shows PCE variability with length of grade for 8% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.58 PCE VARIABILITY BY LENGTH OF GRADE Â– 8% GRADE r#&'()!% n! r r#&'()!% n! r
PAGE 137
122 4.3.4.3 Grade Percentages Figure 4.59 shows PCE variability with grade percentage s for 0.25 mi grade and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.59 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi r)! !" r r)! !" r
PAGE 138
123 Figure 4.60 shows PCE variability with grade percentage s for 0.50 mi grade and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.60 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi r)! !" r r)! !" r
PAGE 139
124 Figure 4.61 shows PCE variability with grade percentag es for 0.75 mi grade and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.61 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi r)! !" r r)! !" r
PAGE 140
125 Figure 4.62 shows PCE variability with grade percentage s for 1.0 mi grade and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.62 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi r)! !" r r)! !" r
PAGE 141
126 Figure 4.63 shows PCE variability with grade percentage s for 1.25 mi grade and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.63 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi r)! !" r r)! !" r
PAGE 142
127 Figure 4.64 shows PCE variability with grade percentage s for 1.50 mi grade and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.64 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi r)! !" r)! !"
PAGE 143
128 Figure 4.65 shows PCE variability with grade percentage s for 2.0 mi grade and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.65 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi r)! !" r)! !"
PAGE 144
129 Figure 4.66 shows PCE variability with grade percentage s for 0.25 mi grade and 10% trucks using Demarchi & Setti and Sumner et al. met hodologies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.66 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi r)! !" r r)! !" r
PAGE 145
130 Figure 4.67 shows PCE variability with grade percentage s for 0.50 mi grade and 10% trucks using Demarchi & Setti and Sumner et al. met hodologies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.67 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi r)! !" r r)! !" r
PAGE 146
131 Figure 4.68 shows PCE variability with grade percentage s for 0.75 mi grade and 10% trucks using Demarchi & Setti and Sumner et al. met hodologies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.68 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi r)! !" r r)! !" r
PAGE 147
132 Figure 4.69 shows PCE variability with grade percentage s for 1.0 mi grade and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.69 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi r)! !" r r)! !" r
PAGE 148
133 Figure 4.70 shows PCE variability with grade percentage s for 1.25 mi grade and 10% trucks using Demarchi & Setti and Sumner et al. met hodologies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.70 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi r)! !" r r)! !" r
PAGE 149
134 Figure 4.71 shows PCE variability with grade percentage s for 1.50 mi grade and 10% trucks using Demarchi & Setti and Sumner et al. met hodologies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.71 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi r)! !" r)! !"
PAGE 150
135 Figure 4.72 shows PCE variability with grade percentage s for 2.0 mi grade and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.72 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi n r)! !" n r)! !"
PAGE 151
136 For all cases of various truck proportions, different len gth of grade and variety of grades, PCE for freeways with more than two directional lanes are lower than PCE for freeways with only two directional lanes. Using both methods: equal flowdensity and equal speeddensity of calculating PCE shows that PCE for two lanes is higher than for three lanes except for only one Figure 4.59a with 0.25 mile and 5%. The justificatio n for the only exception is because of the randomness of simulation. 4.3.5 Variance PCE by Trucks Restrictions Truck lane restrictions become common and feasible policie s on freeways to improve mobility, efficiency, and safety. The HCM 2010 has not co nsidered lane restrictions for trucks to measure PCE values. In order to understan d the truck traffic dynamic, this study is applied truck lane restriction on the left most lanes. One of the aims of this study is to test the hypothesis that truck passenger car equivalents are impacted by truck lane restriction.
PAGE 152
137 4.3.5.1 Truck Proportion Figure 4.73 shows PCE variability with truck proportion f or 0.5 mile and 2% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.73 PCE VARIABILITY BY TRUCK PROPORTION Â– 0 .5 Mi 2% GRADE nr rrr r nr rrr r
PAGE 153
138 Figure 4.74 shows PCE variability with truck proportion f or 0.5 mile and 4% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.74 PCE VARIABILITY BY TRUCK PROPORTION Â– 0 .5 Mi 4% GRADE n r rrr r nr rrr r
PAGE 154
139 Figure 4.75 shows PCE variability with truck proportion f or 0.5 mile and 6% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.75 PCE VARIABILITY BY TRUCK PROPORTION Â– 0 .5 Mi 6% GRADE n r rrr r nr rrr r
PAGE 155
140 Figure 4.76 shows PCE variability with truck proportion f or 0.5 mile and 8% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.76 PCE VARIABILITY BY TRUCK PROPORTION Â– 0 .5 Mi 8% GRADE nr n! rrr r nr n! rrr r
PAGE 156
141 Figure 4.77 shows PCE variability with truck proportion f or 1.0 mile and 2% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.77 PCE VARIABILITY BY TRUCK PROPORTION Â– 1 .0 Mi 2% GRADE n r rrr r nr rrr r
PAGE 157
142 Figure 4.78 shows PCE variability with truck proportion f or 1.0 mile and 4% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.78 PCE VARIABILITY BY TRUCK PROPORTION Â– 1 .0 Mi 4% GRADE n r rrr r n r rrr r
PAGE 158
143 Figure 4.79 shows PCE variability with truck proportion f or 1.0 mile and 6% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.79 PCE VARIABILITY BY TRUCK PROPORTION Â– 1 .0 Mi 6% GRADE n r rrr r n r rrr r
PAGE 159
144 Figure 4.80 shows PCE variability with truck proportion f or 1.0 mile and 8% grade using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.80 PCE VARIABILITY BY TRUCK PROPORTION Â– 1 .0 Mi 8% GRADE n r n! rrr r n r n! rrr r
PAGE 160
145 4.3.5.2 Length of Grade Figure 4.81 shows PCE variability with length of grade for 2% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.81 PCE VARIABILITY BY LENGTH OF GRADE 2% GRADE r#&'()!% rrr r r#&'()!% rrr r
PAGE 161
146 Figure 4.82 shows PCE variability with length of grade for 4% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.82 PCE VARIABILITY BY LENGTH OF GRADE 4% GRADE r#&'()!% rrr r r#&'()!% rrr r
PAGE 162
147 Figure 4.83 shows PCE variability with length of grade for 6% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.83 PCE VARIABILITY BY LENGTH OF GRADE 6% GRADE r#&'()!% rrr r r#&'()!% rrr r
PAGE 163
148 Figure 4.84 shows PCE variability with length of grade for 8% grade and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.84 PCE VARIABILITY BY LENGTH OF GRADE 8% GRADE r#&'()!% n! rrr r r#&'()!% n! rrr r
PAGE 164
149 Figure 4.85 shows PCE variability with length of grade for 2% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.85 PCE VARIABILITY BY LENGTH OF GRADE 2% GRADE r#&'()!% rrr r r#&'()!% rrr r
PAGE 165
150 Figure 4.86 shows PCE variability with length of grade for 4% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.86 PCE VARIABILITY BY LENGTH OF GRADE 4% GRADE r#&'()!% rrr r r#&'()!% rrr r
PAGE 166
151 Figure 4.87 shows PCE variability with length of grade for 6% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.87 PCE VARIABILITY BY LENGTH OF GRADE 6% GRADE r#&'()!% rrr r r#&'()!% rrr r
PAGE 167
152 Figure 4.88 shows PCE variability with length of grade for 8% grade and 10% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.88 PCE VARIABILITY BY LENGTH OF GRADE 8% GRADE r#&'()!% n! rrr r r#&'()!% n! rrr r
PAGE 168
153 4.3.5.3 Grade Percentages Figure 4.89 shows PCE variability with grade percentage s for 0.25 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.89 PCE VARIABILILTY BY GRADE PERCENTAGES Â– 0.25 Mi r)! !" rrr r r)! !" rrr r
PAGE 169
154 Figure 4.90 shows PCE variability with grade percentage s for 0.50 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.90 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi r)! !" rrr r r)! !" rrr r
PAGE 170
155 Figure 4.91 shows PCE variability with grade percentage s for 0.75 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.91 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi r)! !" rrr r r)! !" rrr r
PAGE 171
156 Figure 4.92 shows PCE variability with grade percentage s for 1.0 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.92 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi r)! !" rrr r r)! !" rrr r
PAGE 172
157 Figure 4.93 shows PCE variability with grade percentage s for 1.25 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.93 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi r)! !" rrr r r)! !" rrr r
PAGE 173
158 Figure 4.94 shows PCE variability with grade percentage s for 1.50 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.94 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi r)! !" rrr r)! !" rrr
PAGE 174
159 Figure 4.95 shows PCE variability with grade percentage s for 2.0 mile and 5% trucks using Demarchi & Setti and Sumner et al. methodologi es. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.95 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi r)! !" rrr r)! !" rrr
PAGE 175
160 Figure 4.96 shows PCE variability with grade percentage s for 0.25 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.96 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.25 Mi r)! !" rrr r r)! !" rrr r
PAGE 176
161 Figure 4.97 shows PCE variability with grade percentage s for 0.50 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.97 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.50 Mi r)! !" rrr r r)! !" rrr r
PAGE 177
162 Figure 4.98 shows PCE variability with grade percentage s for 0.75 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.98 PCE VARIABILITY BY GRADE PERCENTAGES Â– 0.75 Mi r)! !" rrr r r)! !" rrr r
PAGE 178
163 Figure 4.99 shows PCE variability with grade percentage s for 1.0 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.99 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.0 Mi r)! !" rrr r r)! !" rrr r
PAGE 179
164 Figure 4.100 shows PCE variability with grade percenta ges for 1.25 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.100 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.25 Mi r)! !" rrr r r)! !" rrr r
PAGE 180
165 Figure 4.101 shows PCE variability with grade percenta ges for 1.50 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.101 PCE VARIABILITY BY GRADE PERCENTAGES Â– 1.50 Mi r)! !" rrr r)! !" rrr
PAGE 181
166 Figure 4.102 shows PCE variability with grade percenta ges for 2.0 mile and 10% trucks using Demarchi & Setti and Sumner et al. methodol ogies. (a) Demarchi & SettiÂ’s Method (b) Sumner et alÂ’s Method Figure 4.102 PCE VARIABILITY BY GRADE PERCENTAGES Â– 2.0 Mi n r)! !" rrr n r)! !" rrr
PAGE 182
167 By varying truck restrictions at the third left most lanes, it was found that at higher truck proportions the PCE was different both in two lan es and three lanes with and without truck restrictions. As the length of grade incre ases and the percent grade increases, there are no changes in PCE with the higher number of lanes and trucks restrictions.
PAGE 183
168 5. Statistical Analysis 5.1 Overview Statistical analysis was conducted to investigate the diff erences between PCE values using equal flowdensity and equal speeddensity met hodologies for different scenarios of truck proportions, length of grade, percentag es of grade, number of lane, and truck lane restriction from the VISSIM simulat ion output. The Statistical Package for the Social Sciences (SPSS) version 13.0 softwa re package was used for the analysis. The followings are statistical analysis performed: 1. Mean, standard deviation, and other statistics are the basis of statistical analysis performed. 2. Independent Sample ttest was performed for the sig nificance of the difference between the means of two independent sampl es. The degrees of freedom are assumed that two groups of variances are eq ual. In this study, the test groups are for: (a) PCE of 2lane and 3lane and (b) PCE of 3lane without trucks restrictions and with trucks restrictions. T he hypothesis testing is used twosided with 95% confidence level.
PAGE 184
169 3. Percent Error was performed to compare the results of two different PCEs: (1) PCE of 2Lane and 3Lane without trucks restrictio ns, and (2) PCE of 2lane with trucks restrictions. 5.2 Truck Proportion For statistical analysis, the truck proportion is divided in to 12 groups as follows: 1. Group 1: 2Lane 0.5 mile no truck restrictions Â– equ al flow density 2. Group 2: 2Lane 0.5 mile no truck restrictions Â– equ al speed density 3. Group 3: 3Lane 0.5 mile no truck restrictions Â– equ al flow density 4. Group 4: 3Lane 0.5 mile no truck restrictions Â– equ al speed density 5. Group 5: 3Lane 0.5 mile with truck restrictions Â– e qual flow density 6. Group 6: 3Lane 0.5 mile with truck restrictions Â– e qual speed density 7. Group 7: 2Lane 1.0 mile no truck restrictions Â– equ al flow density 8. Group 8: 2Lane 1.0 mile no truck restrictions Â– eq ual speed density 9. Group 9: 3Lane 1.0 mile no truck restrictions Â– equ al flow density 10. Group 10: 3Lane 1.0 mile no truck restrictions Â– e qual speed density 11. Group 11: 3Lane 1.0 mile with truck restrictions Â– equal flow density 12. Group 12: 3Lane 1.0 mile with truck restrictions Â– equal speed density 5.2.1 Mean, Standard Deviation and Other Statistics Tables 5.1 and 5.2 show mean, standard deviation, and other statistic values from each groups classified.
PAGE 185
170 Table 5.1 Mean and Standard Deviation of PCE Group Mean N Std. Deviation 1 1.667 44 0.7062 2 2.591 44 1.0891 3 1.456 44 0.4881 4 2.580 44 1.1590 5 1.486 44 0.5289 6 2.600 44 1.1517 7 1.988 44 1.1947 8 3.439 44 0.8884 9 1.700 44 0.8529 10 3.749 44 1.2956 11 1.756 44 0.8574 12 3.459 44 1.0845 Total 2.373 528 1.2464 Table 5.2 Other Statistics PCE Group N Valid 528 528 Mean 2.373 6.50 Median 1.825 6.50 Mode 2.1 1 Std. Deviation 1.2464 3.455 Variance 1.553 11.939 Skewness 0.843 0.000 Std. Error of Skewness 0.106 0.106 Kurtosis 0.522 1.217 Std. Error of Kurtosis 0.212 0.212 Range 4.8 11 Minimum 1.0 1 Maximum 5.8 12 Sum 1,252.7 3,432
PAGE 186
171 5.2.2 Independent Sample ttest The hypotheses for comparison of the means in a twogrou p ttest of truck proportion are as follows: Null Hypotheses ( H 0 ): 1 = 2 (the PCE means of 2lane and 3Lane are the same) Alternative Hypotheses ( H 1 ): 1 2 (the PCE means of 2lane and 3Lane are different) Table 5.3 shows that all combinations of PCE of 2lan e and 3Lane are significantly different using both methods, except groups 2 and 4 a re not significantly different with a pvalue greater than 0.05. Table 5.3 Independent ttest between 2Lane and 3La ne Descriptions F pvalue t df Sig. twotailed Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Groups 1 & 3 2Lane and 3Lane (Equal Flow Density Method) 0.5 Mile Equal variances assumed 4.20 0.04 1.63 86.00 0.11 0.21 0.13 0.05 0.47 Groups 2 & 4 2Lane and 3Lane (Equal Speed Density Method) 0. 5 Mile Equal variances assumed 1.37 0.24 0.04 86.00 0.96 0.01 0.24 0.47 0.49 Groups 7 & 9 2Lane and 3Lane (Equal Flow Density Method) 1.0 Mile Equal variances assumed 4.99 0.03 1.30 86.00 0.20 0.29 0.22 0.15 0.73 Groups 8 & 10 2Lane and 3Lane (Equal Speed Density Method) 1. 0 Mile Equal variances assumed 23.6 0.00 1.31 86.00 0.19 0.31 0.24 0.78 0.16
PAGE 187
172 The hypotheses for comparison of the means in a twogrou p ttest of truck proportion are as follows: Null Hypotheses ( H 0 ): 1 = 2 (the PCE means of 3lane without and with trucks restrictions are the same) Alternative Hypotheses ( H 1 ): 1 2 (the PCE means of 3lane without and with trucks restrictions are different) Table 5.4 shows that all groups combinations are no si gnificant different with pvalue greater than 0.05, but groups 8 and 10 are significan tly different. Table 5.4 Independent ttest between 3Lane without and with Trucks Restrictions Descriptions F pvalue t df Sig. twotailed Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Groups 1 & 3 3Lane without & with truck restrictions (Equal Flow D ensity) 0.5 Mi Equal variances assumed 0.15 0.70 0.28 86.00 0.78 0.03 0.11 0.25 0.19 Groups 2 & 4 3Lane without & with truck restrictions (Equal Speed Density) 0.5 Mi Equal variances assumed 0.02 0.89 0.08 86.00 0.93 0.02 0.25 0.51 0.47 Groups 7 & 9 3Lane without & with truck restrictions (Equal Flow D ensity) Â– 1.0 Mi Equal variances assumed 0.05 0.83 0.30 86.00 0.76 0.06 0.18 0.42 0.31 Groups 8 & 10 3Lane without & with truck restrictions (Equal Speed Density) Â– 1.0 Mi Equal variances assumed 5.23 0.02 1.14 86.00 0.26 0.29 0.25 0.22 0.80
PAGE 188
173 5.2.3 Percent Differences Applying different trucks percentages from 5% to 100%, th e PCE values for threelane freeway segments is lower with the averages percent differences of 10% for 0.5 mile long of the segment and 9% for 1.0 mile usi ng equal density methodology as shown on Table 5.5. In addition PCEs with truck lane restrictions are not too different from lanes without truck restrictions with the averages percent differences of 8% for 0.5 mile long of the segment and 7% for 1.0 mile. Table 5.5 Truck Proportion Percent Difference in PCEs usi ng Demarchi & SettiÂ’s Method Upgrade (%) Truck Proportion (%) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions 0.5 Mi 1.0 Mi 0.5 Mi 1.0 Mi 2 5 2% 3% 11% 6% 10 13% 8% 8% 11% 20 6% 2% 10% 18% 30 3% 4% 6% 13% 40 8% 7% 10% 14% 50 6% 2% 7% 2% 60 3% 3% 5% 1% 70 4% 1% 5% 1% 80 3% 7% 4% 5% 90 3% 5% 4% 4% 100 3% 5% 4% 5% 4 5 25% 27% 25% 35% 10 22% 41% 10% 23% 20 12% 34% 12% 20% 30 17% 19% 17% 17% 40 9% 6% 14% 16% 50 7% 5% 12% 13% 60 12% 6% 12% 11% 70 10% 9% 11% 14% 80 9% 8% 10% 12% 90 7% 12% 9% 11% 100 7% 11% 7% 10%
PAGE 189
174 Table 5.5 (Cont.) Upgrade (%) Truck Proportion (%) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions 0.5 Mi 1.0 Mi 0.5 Mi 1.0 Mi 6 5 13% 13% 8% 13% 10 18% 24% 15% 24% 20 26% 35% 24% 27% 30 35% 30% 33% 23% 40 31% 20% 29% 15% 50 9% 3% 8% 1% 60 15% 8% 14% 1% 70 13% 7% 13% 0% 80 12% 2% 11% 4% 90 9% 2% 9% 4% 100 10% 3% 10% 2% 8 5 19% 16% 16% 16% 10 31% 30% 29% 30% 20 24% 31% 22% 25% 30 27% 32% 25% 27% 40 15% 7% 14% 6% 50 3% 8% 2% 10% 60 1% 8% 0% 12% 70 1% 7% 0% 11% 80 1% 7% 0% 10% 90 1% 6% 0% 9% 100 2% 5% 1% 7% Average 10% 9% 8% 7% Applying different trucks percentages from 5% to 100% and using speed density methodology, the PCEs value for threelane freeway segm ents are hardly different from the twolane segments with averages percent diffe rences of 0.8 for 0.5 mile long segments. PCEs are higher for threelane freeway se gments than for twolane with the average percent error of 7% for 1.0 as shown on Table 5.6.
PAGE 190
175 Table 5.6 Truck Proportion Percent Difference in PCEs usi ng Sumner et al.Â’s Method Upgrade (%) Truck Proportion (%) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions 0.5 Mi 1.0 Mi 0.5 Mi 1.0 Mi 2 5 12% 4% 2% 9% 10 3% 12% 11% 8% 20 9% 13% 11% 27% 30 8% 14% 9% 27% 40 21% 13% 22% 26% 50 22% 10% 22% 20% 60 19% 10% 20% 20% 70 25% 11% 25% 20% 80 23% 5% 23% 13% 90 29% 7% 29% 13% 100 27% 5% 27% 11% 4 5 4% 1% 4% 0% 10 10% 4% 4% 5% 20 4% 3% 4% 6% 30 5% 2% 5% 6% 40 2% 6% 5% 6% 50 3% 6% 5% 6% 60 11% 5% 11% 6% 70 8% 7% 11% 8% 80 8% 7% 11% 8% 90 4% 4% 11% 8% 100 4% 4% 4% 8% 6 5 2% 5% 2% 5% 10 18% 20% 18% 17% 20 15% 24% 15% 18% 30 9% 24% 9% 18% 40 12% 26% 12% 20% 50 23% 22% 23% 17% 60 21% 19% 21% 14% 70 21% 19% 20% 14% 80 21% 22% 20% 16% 90 21% 22% 21% 16% 100 20% 28% 20% 20%
PAGE 191
176 Table 5.6 (Cont.) Upgrade (%) Truck Proportion (%) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions 0.5 Mi 1.0 Mi 0.5 Mi 1.0 Mi 8 5 2% 2% 1% 2% 10 28% 20% 29% 0% 20 34% 17% 34% 4% 30 31% 14% 31% 7% 40 29% 18% 29% 2% 50 31% 20% 31% 1% 60 29% 20% 30% 1% 70 30% 20% 31% 1% 80 29% 20% 30% 1% 90 29% 20% 30% 1% 100 30% 20% 30% 1% Average 0.8% 7% 0.3% 2% In addition PCEs with truck lane restriction are a litt le higher than those for lanes without truck restrictions with the average percent diffe rences of 0.3% for 0.5 mile long of the segment, but this average is negligible. I n contrast PCEs for lanes with truck lane restriction are lower than PCEs for lanes with out truck restrictions with the average percent differences of 2% for 1.0 mile long o f the segment, yet this average is not substantial. 5.3 Length of Grade For statistical analysis, the length of grade is divided i nto 12 groups as follows: 1. Group 1: 2Lane for 5% trucks without truck restrictio ns Â– equal flow 2. Group 2: 2Lane for 5% trucks without truck restrictio ns Â– equal speed 3. Group 3: 3Lane for 5% trucks without truck restrictio ns Â– equal flow
PAGE 192
177 4. Group 4: 3Lane for 5% trucks without truck restrictio ns Â– equal speed 5. Group 5: 3Lane for 5% trucks with truck restrictions Â– equal flow 6. Group 6: 3Lane for 5% trucks with truck restrictions Â– equal speed 7. Group 7: 2Lane for 10% trucks without truck restrict ions Â– equal flow 8. Group 8: 2Lane for 10% trucks without restrictions Â– equal speed 9. Group 9: 3Lane for 10% trucks without restrictions Â– equal flow 10. Group 10: 3Lane for 10% trucks without restrictions Â– equal speed 11. Group 11: 3Lane for 10% trucks with truck restriction s Â– equal flow 12. Group 12: 3Lane for 10% trucks with truck restriction s Â– equal speed 5.3.1 Mean, Standard Deviation and Other Statistics Tables 5.7 and 5.8 show mean, standard deviation, and other statistic values from each groups classified. Table 5.7 Mean and Standard Deviation of PCE Group Mean N Std. Deviation 1 4.604 40 1.8474 2 4.186 40 1.6975 3 3.810 40 1.2575 4 3.965 40 1.6992 5 3.734 40 1.3264 6 3.885 40 1.7593 7 5.058 40 2.6109 8 4.971 40 2.9291 9 2.463 40 0.6039 10 4.456 40 2.3496 11 2.528 40 0.6018 12 4.436 40 2.3690 Total 4.008 480 2.0293
PAGE 193
178 Table 5.8 Other Statistics PCE Group N Valid 480 480 Mean 4.008 6.50 Median .0926 .158 Mode 3.299 6.50 Std. Deviation 4.6 1(a) Variance 2.0293 3.456 Skewness 4.118 11.942 Std. Error of Skewness 1.139 .000 Kurtosis .111 .111 Std. Error of Kurtosis 1.349 1.217 Range .222 .222 Minimum 12.5 11 Maximum 1.0 1 Sum 13.5 12 Total 1923.8 3120 5.3.2 Independent Sample ttest The hypotheses for comparison of the means in a twogro up ttest of PCE of length of grade are as follows: Null Hypotheses ( H 0 ): 1 = 2 (the PCE means of 2lane and 3Lane are the same) Alternative Hypotheses ( H 1 ): 1 2 (the PCE means of 2lane and 3Lane are different) Table 5.9 shows that the alternative hypotheses using eq ual flowdensity formula shows means the PCE means of 2lane and 3Lane are diff erent. Groups 2 and 4
PAGE 194
179 and groups 8 and 10 are not significantly different, with pvalue greater than 0.05. The null hypotheses are accepted using equal speeddensit y method. Table 5.9 Independent ttest between 2Lane and 3La ne Descriptions F Sig. t df Sig. twotailed Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Groups 1 & 3 2Lane and 3Lane (Equal Flow Density) Â– 5% Trucks Equal variances assumed 8.46 0.00 2.25 78.00 0.03 0.79 0.35 0.09 1.50 Groups 2 & 4 2Lane and 3Lane (Equal Speed Density) 5% Trucks Equal variances assumed 0.04 0.85 0.58 78.00 0.56 0.22 0.38 0.54 0.98 Groups 7 & 9 2Lane and 3Lane (Equal Flow Density) 10% Trucks Equal variances assumed 48.87 0.00 6.12 78.00 0.00 2.59 0.42 1.75 3.44 Groups 8 & 10 2Lane and 3Lane (Equal Speed Density) 10% Trucks Equal variances assumed 0.94 0.34 0.87 78.00 0.39 0.52 0.59 0.67 1.70 The hypotheses for comparison of the means in a twogro up ttest of length of grade are as follows: Null Hypotheses ( H 0 ): 1 = 2 (the PCE means of 3lane without and with trucks restrictions are the same) Alternative Hypotheses ( H 1 ): 1 2 (the PCE means of 3lane without and with trucks restrictions are different)
PAGE 195
180 Table 5.10 shows that all groups combinations are not si gnificantly different with pvalue greater than 0.05. Hence, when on the freeway seg ment applies both no trucklane restrictions and trucklane restrictions, the PCE value s were not too difference. Table 5.10 Independent ttest between 3Lane without and with Trucks Restrictions Descriptions F Sig. t df Sig. twotailed Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Groups 1 & 3 3Lane without & with truck restrictions (Equal Flow D ensity) Â– 5% Trucks Equal variances assumed 1.39 0.24 0.26 78.00 0.79 0.08 0.29 0.50 0.65 Groups 2 & 4 3Lane without & with truck restrictions (Equal Speed Density) 5%Trucks Equal variances assumed 0.04 0.84 0.21 78.00 0.84 0.08 0.39 0.69 0.85 Groups 7 & 9 3Lane without & with truck restrictions (Equal Flow D ensity) Â– 10% Trucks Equal variances assumed 0.00 1.00 0.48 78.00 0.63 0.06 0.13 0.33 0.20 Groups 8 & 10 3Lane without & with truck restrictions (Equal Speed Density) Â– 10%Trucks Equal variances assumed 0.00 0.96 0.04 78.00 0.97 0.02 0.53 1.03 1.07 5.3.3 Percent Differences Applying various lengths of grade from 0.25 mile to 5 miles, the PCE values for threelane freeway segments are lower with the average s percent differences of 12% with 5% trucks and 44% with 10% trucks using equal de nsity methodology as shown on Table 5.11. Since the absolute differences ar e more than 10% in both cases, the length of grade substantially impacts truck PC E values. In addition PCEs
PAGE 196
181 with truck lane restriction are not too different from lanes without truck restrictions with the averages percent differences of 13% with 5% trucks and 42% with 10% trucks. Table 5.11 Length of Grade Percent Difference in PCEs using Demarchi & SettiÂ’s Method Upgrade (%) Length (mi) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Proportion of Trucks and Buses 5% 10% 5% 10% 2% 0.25 2% 23% 11% 7% 0.50 2% 13% 11% 7% 0.75 2% 20% 11% 20% 1.00 28% 33% 32% 33% 1.25 3% 28% 6% 28% 1.50 4% 28% 1% 28% 2.00 24% 33% 22% 33% 3.00 28% 37% 26% 37% 4.00 32% 41% 30% 41% 5.00 35% 47% 33% 47% 4% 4% 0.25 11% 11% 57% 11% 0.50 25% 22% 25% 22% 0.75 9% 27% 41% 27% 1.00 6% 30% 17% 27% 1.25 22% 44% 39% 44% 1.50 26% 46% 42% 46% 2.00 27% 48% 44% 48% 3.00 30% 51% 47% 51% 4.00 33% 53% 49% 53% 5.00 36% 55% 51% 55%
PAGE 197
182 Table 5.11 (Cont.) Upgrade (%) Length (mi) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Proportion of Trucks and Buses 5% 10% 5% 10% 6% 0.25 25% 20% 16% 20% 0.50 11% 39% 4% 25% 0.75 6% 45% 2% 38% 1.00 1% 50% 1% 48% 1.25 17% 56% 17% 56% 1.50 30% 59% 30% 59% 2.00 32% 64% 32% 64% 3.00 34% 67% 34% 67% 4.00 36% 67% 36% 67% 5.00 38% 68% 38% 68% 8% 0.25 42% 23% 83% 3% 0.50 17% 43% 11% 38% 0.75 0% 42% 4% 42% 1.00 16% 50% 16% 48% 1.25 1% 50% 1% 50% 1.50 5% 54% 5% 53% 2.00 8% 57% 8% 56% 3.00 15% 61% 15% 61% 4.00 21% 67% 21% 67% 5.00 24% 68% 24% 68% Average 12% 44% 13% 42% Applying various lengths of grade from 0.25 mile to 5 miles, the PCE values for threelane freeway segments are lower with the average s percent differences of 6% with 5% trucks and 8% with 10% trucks as shown on Table 5.12.
PAGE 198
183 Table 5.12 Length of Grade Percent Differences in PCE s using Sumner et al. Upgrade (%) Length (mi) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Proportion of Trucks and Buses 5% 10% 5% 10% 2% 0.25 12% 17% 15% 46% 0.50 12% 3% 15% 11% 0.75 12% 9% 15% 9% 1.00 12% 6% 19% 6% 1.25 3% 7% 7% 7% 1.50 3% 7% 6% 7% 2.00 0% 6% 3% 6% 3.00 0% 5% 3% 5% 4.00 0% 5% 3% 5% 5.00 0% 4% 3% 4% 4% 0.25 9% 22% 48% 22% 0.50 0% 7% 0% 7% 0.75 3% 6% 1% 6% 1.00 2% 6% 4% 7% 1.25 1% 4% 3% 4% 1.50 1% 4% 3% 4% 2.00 1% 4% 4% 4% 3.00 1% 4% 3% 4% 4.00 1% 4% 4% 4% 5.00 1% 4% 4% 4% 6% 6% 0.25 27% 12% 52% 12% 0.50 16% 20% 15% 25% 0.75 15% 18% 15% 19% 1.00 16% 18% 16% 18% 1.25 15% 18% 15% 18% 1.50 14% 18% 14% 18% 2.00 15% 18% 15% 18% 3.00 15% 18% 15% 18% 4.00 15% 19% 15% 19% 5.00 15% 19% 15% 19%
PAGE 199
184 Table 5.12 (Cont.) Upgrade (%) Length (mi) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Proportion of Trucks and Buses 5% 10% 5% 10% 8% 0.25 19% 10% 29% 15% 0.50 4% 7% 6% 8% 0.75 0% 2% 0% 2% 1.00 2% 1% 2% 2% 1.25 1% 2% 1% 2% 1.50 1% 1% 1% 1% 2.00 1% 1% 1% 1% 3.00 2% 1% 2% 1% 4.00 2% 0% 2% 0% 5.00 2% 32% 2% 32% Average 6% 8% 9% 9% Since the absolute differences are less than 10% in bo th cases, the length of grade does not substantially impact truck PCE values, using spee d density methodology. In addition PCEs with truck lane restriction are a little l ower than lanes without truck restrictions with the averages percent differences of 9% with 5% trucks and 9% with 10% trucks. 5.4 Grade Percentages For statistical analysis, the grade proportion is divided i nto 12 groups as follows: 1. Group 1: 2Lane for 5% trucks without truck restrictio ns Â– equal flow 2. Group 2: 2Lane for 5% trucks without truck restrictio ns Â– equal speed 3. Group 3: 3Lane for 5% trucks without truck restrictio ns Â– equal flow 4. Group 4: 3Lane for 5% trucks without truck restrictio ns Â– equal speed
PAGE 200
185 5. Group 5: 3Lane for 5% trucks with truck restrictions Â– equal flow 6. Group 6: 3Lane for 5% trucks with truck restrictions Â– equal speed 7. Group 7: 2Lane for 10% trucks without truck restrict ions Â– equal flow 8. Group 8: 2Lane for 10% trucks without restrictions Â– equal speed 9. Group 9: 3Lane for 10% trucks without restrictions Â– equal flow 10. Group 10: 3Lane for 10% trucks without restrictions Â– equal speed 11. Group 11: 3Lane for 10% trucks with truck restriction s Â– equal flow 12. Group 12: 3Lane for 10% trucks with truck restriction s Â– equal speed 5.4.1 Mean, Standard Deviation and Other Statistic s Tables 5.13 and 5.14 show mean, standard deviation, an d other statistic values from each groups classified. Table 5.13 Mean and Standard Deviation of PCE Group Mean N Std. Deviation 1 3.983 28 1.6458 2 3.621 28 1.4554 3 3.618 28 1.2374 4 3.395 28 1.4867 5 3.596 28 1.3029 6 3.310 28 1.5642 7 4.234 28 2.0483 8 3.924 28 1.8754 9 2.351 28 0.6007 10 3.599 28 1.6756 11 2.437 28 0.6042 12 3.571 28 1.7023 Total 3.470 336 1.5672
PAGE 201
186 Table 5.14 Other Statistics PCE Group N Valid 336 336 Mean 3.470 6.50 Median 0.0855 0.189 Mode 3.000 6.50 Std. Deviation 4.6 1 Variance 1.5672 3.457 Skewness 2.456 11.952 Std. Error of Skewness 0.810 0.000 Kurtosis 0.133 0.133 Std. Error of Kurtosis 0.014 1.217 Range 0.265 0.265 Minimum 7.8 11 Maximum 1.0 1 Sum 8.8 12 Total 1,165.9 2,184 5.4.2 Independent Sample ttest The hypotheses for comparison of the means in a twogrou p ttest of PCE of grade percentages are as follows: Null Hypotheses ( H 0 ): 1 = 2 (the PCE means of 2lane and 3Lane are the same) Alternative Hypotheses ( H 1 ): 1 2 (the PCE means of 2lane and 3Lane are different) Table 5.15 shows that the alternative hypotheses using eq ual flowdensity formula shows means the PCE of 2lane and 3Lane are different Groups 2 and 4 and groups 8 and 10 are not significantly different, with pvalue greater than 0.05. The null hypotheses are accepted using equal speeddensity me thod.
PAGE 202
187 Table 5.15 Independent ttest between 2Lane and 3L ane Descriptions F pvalue t df Sig. twotailed Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Groups 1 & 3 2Lane and 3Lane (Equal Flow Density) Â– 5% Trucks Equal variances assumed 5.27 0.03 0.94 54.00 0.35 0.37 0.39 0.42 1.15 Groups 2 & 4 2Lane and 3Lane (Equal Speed Density) 5% Trucks Equal variances assumed 0.00 1.00 0.57 54.00 0.57 0.23 0.39 0.56 1.01 Groups 7 & 9 2Lane and 3Lane (Equal Flow Density) 10% Trucks Equal variances assumed 32.4 0.00 4.67 54.00 0.00 1.88 0.40 1.07 2.69 Groups 8 & 10 2Lane and 3Lane (Equal Speed Density) 10% Trucks Equal variances assumed 0.27 0.60 0.68 54.00 0.50 0.33 0.48 0.63 1.28 The hypotheses for comparison of the means in a twogro up ttest of length of grade are as follows: Null Hypotheses ( H 0 ): 1 = 2 (the PCE means of 3lane without and with trucks restrictions are the same) Alternative Hypotheses ( H 1 ): 1 2 (the PCE means of 3lane without and with trucks restrictions are different) Table 5.16 shows that all groups combinations are not si gnificantly different with pvalue greater than 0.05. Hence, when on the freeway se gment applies both no trucklane restrictions and trucklane restrictions, there are no differences in PCE values.
PAGE 203
188 Table 5.16 Independent ttest between 3Lane without and with Trucks Restrictions Descriptions F pvalue t df Sig. twotailed Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Lower Upper Groups 1 & 3 3Lane without & with truck restrictions (Equal Flow D ensity) Â– 5% Trucks Equal variances assumed 1.19 0.28 0.06 54.00 0.95 0.02 0.34 0.66 0.70 Groups 2 & 4 3Lane without & with truck restrictions (Equal Speed Density) 5%Trucks Equal variances assumed 0.04 0.85 0.21 54.00 0.84 0.09 0.41 0.73 0.90 Groups 7 & 9 3Lane without & with truck restrictions (Equal Flow D ensity) Â– 10% Trucks Equal variances assumed 0.03 0.87 0.53 54.00 0.60 0.09 0.16 0.41 0.24 Groups 8 & 10 3Lane without & with truck restrictions (Equal Speed Density) Â– 10%Trucks Equal variances assumed 0.01 0.93 0.06 54.00 0.95 0.03 0.45 0.88 0.93 5.4.3 Percent Differences Applying various grade percentages of 2%, 4%, 6%, and 8% PCE values for threelane freeway segments are lower with the average per cent differences of 5% for 5% trucks and 38% for 10% trucks, using equal density methodol ogy as shown on Table 5.17. Hence, the calculated PCEs for threelane freeways are substantially different only at 10% trucks percentages. In addition PC Es with truck lane restrictions are a little lower for lanes without truck restrictions w ith the averages percent differences of 4% for 5% trucks and 35% for 10% trucks
PAGE 204
189 Table 5.17 Grade Percentages Percent Difference in PCE s using Demarchi & SettiÂ’s Upgrade (%) Length (mi) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Proportion of Trucks and Buses 5% 10% 5% 10% 2% 0.25 2% 23% 11% 7% 0.50 2% 13% 11% 7% 0.75 2% 20% 11% 20% 1.00 28% 33% 32% 33% 1.25 3% 28% 6% 28% 1.50 4% 28% 1% 28% 2.00 24% 33% 22% 33% 4% 0.25 11% 11% 57% 11% 0.50 25% 22% 25% 22% 0.75 9% 27% 41% 27% 1.00 6% 30% 17% 27% 1.25 22% 44% 39% 44% 1.50 26% 46% 42% 46% 2.00 27% 48% 44% 48% 6% 0.25 25% 20% 16% 20% 0.50 11% 39% 4% 25% 0.75 6% 45% 2% 38% 1.00 1% 50% 1% 48% 1.25 17% 56% 17% 56% 1.50 30% 59% 30% 59% 2.00 32% 64% 32% 64% 8% 0.25 42% 23% 83% 3% 0.50 17% 43% 11% 38% 0.75 0% 42% 4% 42% 1.00 16% 50% 16% 48% 1.25 1% 50% 1% 50% 1.50 5% 54% 5% 53% 2.00 8% 57% 8% 56% Average 5% 38% 4% 35%
PAGE 205
190 Applying various grade percentages of 2%, 4%, 6%, and 8% PCEs values for threelane freeway segments are lower with the average per cent differences of 7% for 5% trucks and 8% for 10% trucks using speed density methodolog y as shown on Table 5.18. The calculated PCEs for threelane freeways are not substantially different for both 5% and 10% of trucks percentages. Table 5.18 Grade Percentages Percent Difference in PCE s using Sumner et al.Â’s Upgrade (%) Length (mi) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Proportion of Trucks and Buses 5% 10% 5% 10% 2% 0.25 12% 17% 15% 46% 0.50 12% 3% 15% 11% 0.75 12% 9% 15% 9% 1.00 12% 6% 19% 6% 1.25 3% 7% 7% 7% 1.50 3% 7% 6% 7% 2.00 0% 6% 3% 6% 4% 0.25 9% 22% 48% 22% 0.50 0% 7% 0% 7% 0.75 3% 6% 1% 6% 1.00 2% 6% 4% 7% 1.25 1% 4% 3% 4% 1.50 1% 4% 3% 4% 2.00 1% 4% 4% 4% 6% 0.25 27% 12% 52% 12% 0.50 16% 20% 15% 25% 0.75 15% 18% 15% 19% 1.00 16% 18% 16% 18% 1.25 15% 18% 15% 18% 1.50 14% 18% 14% 18% 2.00 15% 18% 15% 18%
PAGE 206
191 Table 5.18 (Cont.) Upgrade (%) Length (mi) % Error from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Proportion of Trucks and Buses 5% 10% 5% 10% 8% 0.25 19% 10% 29% 15% 0.50 4% 7% 6% 8% 0.75 0% 2% 0% 2% 1.00 2% 1% 2% 2% 1.25 1% 2% 1% 2% 1.50 1% 1% 1% 1% 2.00 1% 1% 1% 1% Average 7% 8% 11% 9% In addition PCEs with truck lane restriction are a littl e lower for lanes without truck restrictions with the average percent differences of 11 % for 5% trucks and 9% for 10% trucks.
PAGE 207
192 6. Additional Study 6.1 Overview Additional analysis was performed to estimate the air po llution impacts of trucks performance in the traffic stream. The air quality ef fects by air pollutants produced by both passenger car and truck vehicular traffic types were studied using parameters identified within this specific study. Several factors, su ch as vehicle characteristics, driving conditions, and weather conditions are factors t hat influence air pollution from roadway traffic. The United States Environmental Protection Agency (USEP A), under the requirements of the 1970 Clean Air Act (CAA), as amend ed in 1977 and 1990 (Clean Air Act Amendments [CAAA]), has set national air quality standards for six common pollutants: Carbon monoxide (CO) Nitrogen dioxides (NO 2 ) Ozone (O 3 ) (with nitrogen oxides (NO x ) and volatile organic compounds [VOCs] as precursors) Particulate matter (PM) (PM 10 Â– less than 20 microns in particle diameter; PM 2.5 Â– less than 2.5 microns in particle diameter) Lead (Pb) Sulfur dioxide (SO 2 )
PAGE 208
193 Among these six pollutants, carbon monoxide, nitrogen d ioxide, hydrocarbons, and carbon dioxide are the major exhaust pollutants from ro ad traffic. In this additional analysis, carbon monoxide (CO), nitrogen oxides (NO x ), and volatile compounds (VOC) were estimated from fuel consumption. 6.2 Fuel Consumption and Emissions Fuel consumption refers to the amount of gasoline used by all types of vehicles. Table 6.1 shows a simple representation of fuel consu mption for the various scenarios. The objective is to understand the air polluti on impact of truck traffic based on truck proportion, length of grade, and grad e percentages. Fuel consumption is derived from vehicle miles travel (V MT). VISSIM network performance output provides total network VMT, total d elay, and total stops. Fuel consumption in this study is specified for gasoline and is calculated using the following formula from the Synchro software package. S ynchro is a macroscopic model widely used by transportation engineering consul tants and public agencies. The formula is: F = Total Travel x k1 + Total Delay x k2 + Stops x k3 Where, k1 = 0.075283 Â– (0.0015892 x speed) + 0.000015066 x spe ed 2 k2 = 0.7329 k3 = 0.0000061411 x speed 2 F = fuel consumed in gallons
PAGE 209
194 Table 6.1 Fuel Consumption Scenarios Average Speed (mph) Total Travel (VMT) Total Delay (hrs) Fuel Consumption (gallons) Fuel (miles/ gallon) Truck Proportions 5% Trucks 2% Grade 2Lane No Trucks Restrictions 51.325 1748.426 3.742 61.149 0.0350 3Lane No Trucks Restrictions 55.485 1749.206 1.193 59.452 0.0340 3Lane With Trucks Restrictions 55.527 1749.702 1.169 59.457 0.0340 10% Trucks 2% Grade 2Lane No Trucks Restrictions 50.01 1748.262 4.34 61.725 0.0353 3Lane No Trucks Restrictions 54.553 1749.145 1.429 59.511 0.0340 3Lane With Trucks Restrictions 54.617 1749.499 1.392 59.502 0.0340 Length of Grade 10% trucks 2.0 Mi 4% Grade 2Lane No Trucks Restrictions 49.113 6885.527 19.697 245.605 0.0357 3Lane No Trucks Restrictions 54.045 6911.837 6.902 235.917 0.0341 3Lane With Trucks Restrictions 54.141 6912.023 6.674 235.783 0.0341 10% trucks 2.0 Mi 6% Grade 2Lane No Trucks Restrictions 48.835 6883.592 20.491 246.339 0.0358 3Lane No Trucks Restrictions 53.88 6911.089 7.29 236.134 0.0342 3Lane With Trucks Restrictions 54.03 6911.559 6.919 235.916 0.0341 Grade Percentages 10% trucks 1.0 Mi 5% Grade 2Lane No Trucks Restrictions 51 3495.596 7.928 122.635 0.0351 3Lane No Trucks Restrictions 55.175 3497.381 2.74 119.045 0.0340 3Lane With Trucks Restrictions 55.198 3497.468 2.714 119.034 0.0340 10% trucks 1.0 Mi 10% Grade 2Lane No Trucks Restrictions 46.629 3492.399 14.365 129.052 0.0370 3Lane No Trucks Restrictions 53.204 3497.115 5.057 120.433 0.0344 3Lane With Trucks Restrictions 53.107 3497.073 5.178 120.516 0.0345
PAGE 210
195 Vehicles produce carbon monoxide (CO), nitrogen oxides (NO x ), and volatile compounds (VOC). Several factors can influence vehicle emi ssion rates (Meyer, 2001): 1. Vehicle parameters including vehicle classification, mod el and year, accrued vehicle mileage, fuel delivery system, emission control system onboard computer control system, control system tampering, and in spection and maintenance history. 2. Fuel parameters including fuel type, oxygen content o f fuel, fuel volatility, sulfur content, benzene content, olefin and aromatic co ntent, lead and metals content, and trace sulfurcatalyst effect. 3. Environment factors including altitude, humidity, am bient temperature, diurnal temperature range, road grade, cold and hot start engine mode, average vehicle speed, engine modal activities, load, trip length and number of trips per day, and driver behavior. Emissions were calculated based simply on fuel consumptio n multiplied by the following factors : CO = F x 69.9 g/gal NO x = F x 13.6 g/gal VOC = F x 16.2 g/gal The simplified rates from the Synchro formula are base d on an unpublished letter to the Federal Highway Administration from Oak Ridge Nat ional Labs (Husch and Albeck, 2006)
PAGE 211
196 Table 6.2 Emissions Scenarios CO (g) NOx (g) VOC (g) Truck Proportions 5% Trucks 2% Grade 2Lane No Trucks Restrictions 4274 832 991 3Lane No Trucks Restrictions 4156 809 963 3Lane With Trucks Restrictions 4156 809 963 10% Trucks 2% Grade 2Lane No Trucks Restrictions 4315 839 1000 3Lane No Trucks Restrictions 4160 809 964 3Lane With Trucks Restrictions 4159 809 964 Length of Grade 10% trucks 2.0 Mi 4% Grade 2Lane No Trucks Restrictions 17168 3340 3979 3Lane No Trucks Restrictions 16491 3208 3822 3Lane With Trucks Restrictions 16481 3207 3820 10% trucks 2.0 Mi 6% Grade 2Lane No Trucks Restrictions 17219 3350 3991 3Lane No Trucks Restrictions 16506 3211 3825 3Lane With Trucks Restrictions 16491 3208 3822 Grade Percentages 10% trucks 1.0 Mi 5% Grade 2Lane No Trucks Restrictions 8572 1668 1987 3Lane No Trucks Restrictions 8321 1619 1929 3Lane With Trucks Restrictions 8321 1619 1928 10% trucks 1.0 Mi 10% Grade 2Lane No Trucks Restrictions 9021 1755 2091 3Lane No Trucks Restrictions 8418 1638 1951 3Lane With Trucks Restrictions 8424 1639 1952 The limitation of the Synchro fuel and emissions calcula tions model is for signalized intersections. The Joslin (2010) study discussed Â“Synchro emissi ons model neglects many of the factors affecting vehicle emissions. It does no t directly account for various vehicle types (semitrucks, buses, etc.) or driver beh aviors (acceleration and deceleration rates, etc). The model is based solely on variable that are used directly in, or can be determined from, the calculation of sig nalized intersection level of
PAGE 212
197 service based on the HCM 2000 procedures. Therefore, t he effects of heavy vehicles are only indirectly accounted for in the model by the fact that the total delay at an intersection may increase (and speeds may decrease) as the percentage of heavy vehicles increaseÂ”. To overcome the Synchro omissions, VISSI M software was utilized to generate the corridor operation based on t ruck performance. The VISSIM operations output was utilized in the emission calculati ons. 6.3 Summary The results from the fuel consumption and emissions tab les are as follows: 1. The number of lanes affects fuel consumption. With th e same distance and number of vehicles in the traffic stream, a twolane f reeway consumed fuel more than a threelane freeway. Since fuel is calculate d based on the speed, a twolane freeway has a lower average speed than a threelane freeway. 2. The higher the percentage of trucks in the traffic s tream, the more fuel is used and the more emissions are produced. 3. The longer the distance, the more fuel is used and the more emissions produced. 4. The steeper the grade, the more pollutants are pr oduced. 5. There is no substantial difference in fuel consump tion and emissions between 3lane freeways with and without trucks. This trend is j ustified because although the speed is higher in the left most lane w ithout the truck in the lane and the speed in the other twolanes is lower, and th erefore the average speed for all lanes is the same in either case.
PAGE 213
198 7. Conclusions and Recommendations 7.1 Conclusions HCM 2010 presents PCE values for heavy vehicles in gener al terrain segments, for trucks and buses on specific downgrades, and for trucks and b uses on upgrades. The primary objective of this study is to understand the dynamics of truck traffic by quantifying truck passenger car equivalents using VISSIM mi cro simulation. Truck PCE is calculated using equal flow density and equal spe ed density methodologies based on a speedflowdensity relationship. A number of different scenarios and variables are considered as following: trucks proportion, grade percentages, length of grades, number of lanes, and truck lane restrictions. The findings show that: 1. The proportions of trucks have a significant effect on the calculated PCE with greater than 95% confidence. The PCE values decrease no nlinearly as the truck proportion increases over a range of 2%, 4%, 6%, a nd 8% grades. This trend is justified because as the truck percentages in creases, the interaction between trucks and passenger cars decrease.
PAGE 214
199 2. For truck proportions higher than 50%, the PCE demo nstrates very little variability. The substantial variation occurs for a truck pr oportion between 25 and 50 percents. Thus HCM 2010 should be extended to in clude up to 50 percent trucks. 3. As the number of lanes increases in the traffic strea m, the PCE value decreases. Using an Independent tstatistic test, number o f lanes are significantly affected PCE with greater than 95% confid ence. Freeways with 3 lanes have significantly lower PCE than freeways with 2 lanes. 4. There is no significant difference of the PCE calcula ted with and without trucks restrictions on the third most lanes with greater t han 95% confidence. This trend is justified because, regardless of legal restri ctions, the trucks in reality generally use two right most lanes and only use t he third lane for passing other vehicles. 5. The PCE increases as the length of grade increases. Th is trend matches the trend in the HCM 2010. 6. The truck PCE remains constant for distances longer th an 1.5 miles. Thus, the range of grade in the HCM 2010 is adequate. 7. The PCE increases as the percent grade increases. This tr end matches the trend in the HCM 2010. 8. As the length of grade increases and the percent gr ade increases, PCE does not change with higher number of lanes and restricted t ruck lanes. 9. The PCE values ranges:
PAGE 215
200 a. Truck Proportion Table 7.1 presents the summary PCE values ranges of tr uck proportion with 0.5 and 1.0 miles segment. Table 7.1 PCE ranges of Truck Proportions Truck Proportions 0.5 Mi Grade Trucks PCE Ranges HCM 2010 Methods Equal Density Equal Speed Two Lanes without Restrictions 2% 1.50 1.11 1.80 1.38 1.78 4% 1.50 2.00 1.23 2.67 1.49 1.94 6% 2.50 3.50 1.38 3.50 3.50 4.34 8% 2.50 3.50 1.20 4.48 3.03 4.13 Three Lanes without Restrictions 2% 1.50 1.14 1.83 1.00 1.57 4% 1.50 2.00 1.15 2.00 1.56 2.02 6% 2.50 3.50 1.23 3.03 3.21 3.43 8% 2.50 3.50 1.18 3.61 3.99 4.06 Three Lanes with Restrictions 2% 1.50 1.15 2.00 1.00 1.75 4% 1.50 2.00 1.14 2.00 1.56 2.02 6% 2.50 3.50 1.24 3.20 3.22 3.44 8% 2.50 3.50 1.19 3.78 4.00 4.07 Truck Proportions 1.0 Mi Grade Trucks PCE Ranges HCM 2010 Methods Equal Density Equal Speed Two Lanes without Restrictions 2% 2.00 2.50 1.19 2.65 2.05 2.64 4% 2.50 3.00 1.28 4.64 2.73 3.05 6% 3.50 5.00 1.24 5.35 3.48 4.71 8% 4.00 5.50 1.19 5.76 4.53 5.43 Three Lanes without Restrictions 2% 2.00 2.50 1.14 2.74 2.18 2.32 4% 2.50 3.00 1.15 3.40 2.82 2.87 6% 3.50 5.00 1.20 4.64 4.44 4.47 8% 4.00 5.50 1.25 4.82 5.45 5.52 Three Lanes with Restrictions 2% 2.00 2.50 1.14 2.82 2.07 2.23 4% 2.50 3.00 1.13 3.10 2.60 2.84 6% 3.50 5.00 1.27 4.64 4.35 4.47
PAGE 216
201 8% 4.00 5.50 1.28 4.82 4.49 5.52 b. Lengths of Grade Table 7.2 presents the summary PCE values ranges of le ngth of grades with 5% and 10% trucks. Table 7.2 PCE ranges of Length of Grades Descriptions Grade Trucks PCE Ranges Trucks PCE Ranges 5% Trucks 10% Trucks HCM 2010 Methods HCM 2010 Methods Equal Density Equal Speed Equal Density Equal Speed Two Lanes without Restrictions 2% 1.50 2.50 1.80 4.21 1.78 3.36 1.50 2.50 1.63 3.61 1.58 3.42 4% 1.50 4.00 1.80 5.76 1.92 5.91 1.50 4.00 1.80 5.76 1.92 5.91 6% 1.50 5.00 2.43 7.45 2.10 7.02 1.50 5.00 2.33 9.83 2.0 11.19 8% 2.50 5.50 2.54 7.80 2.13 7.10 2.50 5.50 2.5410.21 2.5713.50 Three Lanes without Restrictions 2% 1.50 2.50 1.83 2.74 1.57 3.36 1.50 2.50 1.25 1.91 1.54 3.24 4% 1.50 4.00 2.00 3.70 1.75 5.87 1.50 4.00 1.61 2.62 1.49 5.68 6% 1.50 5.00 3.03 4.64 1.52 5.97 1.50 5.00 1.87 3.14 2.18 9.09 8% 2.50 5.50 3.61 5.95 1.73 7.26 2.50 5.50 1.96 3.23 2.31 9.15 Three Lanes with Restrictions 2% 1.50 2.50 2.00 2.82 1.52 3.24 1.50 2.50 1.43 1.91 1.00 3.24 4% 1.50 4.00 1.83 3.00 1.00 5.69 1.50 4.00 1.61 2.62 1.49 5.68 6% 1.50 5.00 2.82 4.64 1.00 5.97 1.50 5.00 1.87 3.14 2.03 9.09 8% 2.50 5.50 4.64 5.95 1.51 7.26 2.50 5.50 2.45 3.29 2.18 9.04 c. Grade Percentages Table 7.3 presents the summary PCE values ranges of gr ade percentages with 5% and 10% trucks.
PAGE 217
202 Table 7.3 PCE ranges of Grade Percentages Descriptions Grade Trucks PCE Ranges Trucks PCE Ranges 5% Trucks 10% Trucks HCM 2010 Methods HCM 2010 Methods Equal Density Equal Speed Equal Density Equal Speed Two Lanes without Restrictions 2% 1.50 2.50 1.80 3.61 1.78 2.74 1.50 2.50 1.63 2.85 1.58 2.82 4% 1.50 4.00 1.80 5.07 1.92 5.07 1.50 4.00 1.80 5.07 1.92 5.07 6% 1.50 5.00 2.43 6.81 2.10 5.98 1.50 5.00 2.33 8.83 2.07 8.55 8% 2.50 5.50 2.54 6.49 2.13 6.32 2.50 5.50 2.54 7.53 2.57 7.11 Three Lanes without Restrictions 2% 1.50 2.50 1.83 2.74 1.57 2.74 1.50 2.50 1.25 1.91 1.54 2.65 4% 1.50 4.00 2.00 3.70 1.75 5.02 1.50 4.00 1.61 2.62 1.49 4.85 6% 1.50 5.00 3.03 4.64 1.52 5.10 1.50 5.00 1.87 3.14 2.18 7.00 8% 2.50 5.50 3.61 5.95 1.73 6.40 2.50 5.50 1.96 3.23 2.31 7.03 Three Lanes with Restrictions 2% 1.50 2.50 2.00 2.82 1.52 2.65 1.50 2.50 1.43 1.91 1.00 2.65 4% 1.50 4.00 1.83 3.00 1.00 4.88 1.50 4.00 1.61 2.62 1.49 4.85 6% 1.50 5.00 2.82 4.64 1.00 5.10 1.50 5.00 1.87 3.14 2.03 7.00 8% 2.50 5.50 4.64 5.95 1.51 6.40 2.50 5.50 2.45 3.29 2.18 7.03 10. Summary of percent differences is shown on Table 7. 4: Table 7.4 Summary Percent Differences of PCE for Thre e Lanes vs Two Lanes Descriptions % Difference from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Truck Proportions Length of Grade 0.5 Mi 1.0 Mi 0.5 Mi 1.0 Mi Demarchi Setti Formula based on equal density 10% 9% 8% 7% Huber Sumner Formula based on equal speed 0.8% 7% 0.3% 2% Length of Grade Proportion of Trucks and Buses 5% 10% 5% 10% Demarchi Setti Formula based on equal density 12% 44% 13% 42% Huber Sumner Formula based on equal speed 6.1% 8% 9.3% 9%
PAGE 218
203 Table 7.4 (Cont.) Descriptions % Difference from Two Lanes Freeway Segments Three Lanes without Restrictions Three Lanes with Restrictions Grade Percentages Proportion of Trucks and Buses 5% 10% 5% 10% Demarchi Setti Formula based on equal density 5% 38% 4% 35% Huber Sumner Formula based on equal speed 7.4% 8% 11.2% 9% In order to better understand the impact of truck traff ic on fuel consumption and air pollution, additional analysis was done using Synchro. Two main findings of this analysis are reported below: 1. The number of lanes affects fuel consumption. With th e same distance and number of vehicles in the traffic stream a twolane fr eeway consumed fuel more than a threelane freeway, because the average vehi cle speed in a twolane is lower than in a threelane freeway. 2. There is no substantial difference in fuel consump tion and emissions between threelane freeways with and without trucks. This trend is justified because the speed is higher in the right lane without the tru ck lane restriction and the speed in the other twolanes is lower. Thus, the avera ge speed is for all lanes are the same in either case.
PAGE 219
204 7.2 Recommendations The PCE calculated using the equal flow density method is consistent and better compared to equal speed density method. Therefore, the proposed recommended tables are based on PCE values from equal flow density method. Table 7.5 contains the proposed PCE values recommended to improve truck eq uivalency estimation in the HCM 2010 for truck proportion of 0.5 mile long. Table 7.5 PCE Recommendation based on Truck Proportio n for 0.5 mile Truck Proportion (%) Grade Percentages (%) 2Lane 0.5 Mi HCM 2010 PCE 3 Lane no trucks restrictions 0.5 Mi 3 Lane with truck restrictions 0.5 Mi PCE PCE PCE 5 2% grade 1.80 1.5 1.83 2.00 10 1.63 1.5 1.42 1.50 20 1.35 1.5 1.43 1.48 30 1.25 1.5 1.29 1.32 40 1.19 1.28 1.30 50 1.17 1.23 1.25 60 1.15 1.19 1.21 70 1.13 1.18 1.19 80 1.13 1.16 1.17 90 1.11 1.15 1.16 100 1.11 1.14 1.15 5 4% grade 2.67 2 2.00 2.00 10 1.92 2 1.50 1.72 20 1.67 1.5 1.48 1.48 30 1.58 1.5 1.32 1.32 40 1.43 1.30 1.24 50 1.35 1.25 1.19 60 1.38 1.21 1.21 70 1.32 1.19 1.18 80 1.28 1.17 1.16 90 1.25 1.16 1.14 100 1.23 1.15 1.15
PAGE 220
205 Table 7.5 (Cont.) Truck Proportion (%) Grade Percentages (%) 2Lane 0.5 Mi HCM 2010 PCE 3 Lane no trucks restrictions 0.5 Mi 3 Lane with truck restrictions 0.5 Mi PCE PCE PCE 5 6% grade 3.50 3.5 3.03 3.20 10 2.46 2.5 2.02 2.10 20 2.23 2.5 1.65 1.70 30 2.20 2.5 1.43 1.46 40 2.00 1.39 1.41 50 1.46 1.32 1.34 60 1.49 1.27 1.28 70 1.45 1.26 1.27 80 1.40 1.23 1.24 90 1.39 1.26 1.28 100 1.38 1.24 1.25 5 8% grade 4.48 3.5 3.61 3.78 10 3.36 3 2.30 2.39 20 2.24 2.5 1.70 1.75 30 2.00 2.5 1.47 1.50 40 1.63 1.38 1.40 50 1.36 1.32 1.34 60 1.30 1.29 1.30 70 1.26 1.24 1.26 80 1.24 1.23 1.24 90 1.21 1.20 1.21 100 1.20 1.18 1.19 Table 7.6 contains the proposed PCE values recommended to improve truck equivalency estimation in the HCM 2010 for truck proport ion of 1.0 mile long.
PAGE 221
206 Table 7.6 PCE Recommendation based on Truck Proportio n for 1.0 mile Truck Proportion (%) Grade Percentages (%) 2Lane 1.0 Mi HCM 2010 PCE 3 Lane no trucks restrictions 1.0 Mi 3 Lane with truck restrictions 1.0 Mi PCE PCE PCE 5 2% grade 2.65 2.5 2.74 2.82 10 2.14 2 1.96 1.91 20 1.46 2 1.48 1.71 30 1.31 2 1.35 1.48 40 1.19 1.27 1.36 50 1.26 1.23 1.29 60 1.23 1.19 1.24 70 1.19 1.18 1.20 80 1.24 1.16 1.18 90 1.21 1.15 1.16 100 1.20 1.14 1.14 5 4% grade 4.64 3 3.40 3.00 10 4.00 3 2.35 3.10 20 2.55 2.5 1.68 2.05 30 1.83 2.5 1.48 1.51 40 1.45 1.36 1.23 50 1.36 1.29 1.18 60 1.32 1.24 1.17 70 1.32 1.21 1.14 80 1.28 1.18 1.13 90 1.32 1.16 1.17 100 1.29 1.15 1.15 5 6% grade 5.35 5 4.64 4.64 10 3.73 3.5 2.82 2.82 20 3.00 3.5 1.96 2.19 30 2.33 3.5 1.64 1.79 40 1.88 1.51 1.60 50 1.45 1.41 1.48 60 1.45 1.34 1.45 70 1.39 1.29 1.38 80 1.28 1.25 1.34 90 1.25 1.23 1.30 100 1.24 1.20 1.27
PAGE 222
207 Table 7.6 (Cont.) Truck Proportion (%) Grade Percentages (%) 2Lane 1.0 Mi HCM 2010 PCE 3 Lane no trucks restrictions 1.0 Mi 3 Lane with truck restrictions 1.0 Mi PCE PCE PCE 5 8% grade 5.76 5.5 4.82 4.82 10 4.17 4.5 2.91 2.91 20 3.00 4 2.06 2.24 30 2.50 4 1.71 1.83 40 1.71 1.59 1.62 50 1.36 1.47 1.50 60 1.30 1.41 1.46 70 1.26 1.35 1.40 80 1.23 1.31 1.35 90 1.20 1.28 1.31 100 1.19 1.25 1.28 Tables 7.7 contains the proposed PCE values recommende d to improve truck equivalency estimation in the HCM 2010 for length of gr ade for 5% trucks. Table 7.7 PCE Recommendation based on Length of Grad e for 5% Trucks HCM 2010 Length of grade (mi) Simulation Length of grade (mi) Grade Percentages (%) 2Lane 5% Trucks HCM 2010 PCE 3Lane no trucks restrictions 5% Trucks 3Lane with trucks restrictions 5% Trucks PCE PCE PCE 00.25 0.25 2% 1.80 1.5 1.83 2.00 >0.250.5 0.50 1.80 1.5 1.83 2.00 >0.50.75 0.75 1.80 1.5 1.83 2.00 >0.751.0 1.00 2.14 2 2.74 2.82 >1.01.50 1.25 2.67 2.5 2.74 2.82 >1.5 1.50 2.85 2.5 2.74 2.82 2.00 3.61 2.74 2.82 3.00 3.81 2.74 2.82 4.00 4.01 2.74 2.82 5.00 4.21 2.74 2.82
PAGE 223
208 Table 7.7 (Cont.) HCM 2010 Length of grade (mi) Simulation Length of grade (mi) Grade Percentages (%) 2Lane 5% Trucks HCM 2010 PCE 3Lane no trucks restrictions 5% Trucks 3Lane with trucks restrictions 5% Trucks PCE PCE PCE 00.25 0.25 4% 1.80 1.5 2.00 2.82 >0.250.5 0.50 2.67 2.5 2.00 2.00 >0.50.75 0.75 3.13 3 3.40 1.83 >0.751.0 1.00 3.61 3.5 3.40 3.00 >1 1.25 4.64 4 3.60 2.82 1.50 4.85 3.60 2.82 2.00 5.07 3.70 2.82 3.00 5.30 3.70 2.82 4.00 5.53 3.70 2.82 5.00 5.76 3.70 2.82 00.25 0.25 6% 2.43 1.5 3.03 2.82 >0.250.5 0.50 3.07 3 3.41 3.20 >0.50.75 0.75 3.86 4 3.61 3.78 >0.751.0 1.00 4.70 4.5 4.64 4.64 >1 1.25 5.62 5 4.64 4.64 1.50 6.60 4.64 4.64 2.00 6.81 4.64 4.64 3.00 7.02 4.64 4.64 4.00 7.23 4.64 4.64 5.00 7.45 4.64 4.64 00.25 0.25 8% 2.54 2.5 3.61 4.64 >0.250.5 0.50 4.33 3.5 3.61 4.82 >0.50.75 0.75 4.64 4.5 4.64 4.82 >0.751.0 1.00 5.76 5 4.82 4.82 >1 1.25 6.00 5.5 5.95 5.95 1.50 6.24 5.95 5.95 2.00 6.49 5.95 5.95 3.00 7.00 5.95 5.95 4.00 7.53 5.95 5.95 5.00 7.80 5.95 5.95
PAGE 224
209 Table 7.8 contains the proposed PCE values recommended to improve truck equivalency estimation in the HCM 2010 for length of gr ade for 10% trucks. Table 7.8 PCE Recommendation based on Length of Grad e for 10% Trucks HCM 2010 Length of grade (mi) Simulation Length of grade (mi) Grade Percentages (%) 2Lane 10% Trucks HCM 2010 PCE 3Lane no trucks restrictions 10% Trucks 3Lane with trucks restrictions 10% Trucks PCE PCE PCE 00.25 0.25 2% 1.63 1.5 1.25 1.52 >0.250.5 0.50 1.63 1.5 1.42 1.52 >0.50.75 0.75 1.80 1.5 1.43 1.43 >0.751.0 1.00 2.14 2 1.43 1.43 >1.01.50 1.25 2.67 2.5 1.91 1.91 >1.5 1.50 2.67 2.5 1.91 1.91 2.00 2.85 1.91 1.91 3.00 3.03 1.91 1.91 4.00 3.22 1.91 1.91 5.00 3.61 1.91 1.91 00.25 0.25 4% 1.80 1.5 1.61 1.61 >0.250.5 0.50 2.67 2.5 2.09 2.09 >0.50.75 0.75 3.13 3 2.30 2.30 >0.751.0 1.00 3.61 3.5 2.51 2.62 >1 1.25 4.64 4 2.62 2.62 1.50 4.85 2.62 2.62 2.00 5.07 2.62 2.62 3.00 5.30 2.62 2.62 4.00 5.53 2.62 2.62 5.00 5.76 2.62 2.62 00.25 0.25 6% 2.33 1.5 1.87 1.87 >0.250.5 0.50 3.86 3.5 2.36 2.90 >0.50.75 0.75 4.70 4 2.57 2.90 >0.751.0 1.00 5.62 4.5 2.79 2.90 >1 1.25 6.60 5 2.90 2.90 1.50 7.67 3.14 3.14 2.00 8.83 3.14 3.14 3.00 9.44 3.14 3.14 4.00 9.57 3.14 3.14 5.00 9.83 3.14 3.14
PAGE 225
210 Table 7.8 (Cont.) HCM 2010 Length of grade (mi) Simulation Length of grade (mi) Grade Percentages (%) 2Lane 10% Trucks HCM 2010 PCE 3Lane no trucks restrictions 10% Trucks 3Lane with trucks restrictions 10% Trucks PCE PCE PCE 00.25 0.25 8% 2.54 2.5 1.96 2.45 >0.250.5 0.50 4.33 3.5 2.45 2.67 >0.50.75 0.75 4.64 4.5 2.67 2.67 >0.751.0 1.00 5.76 5 2.89 3.00 >1 1.25 6.00 5.5 3.00 3.00 1.50 7.00 3.23 3.29 2.00 7.53 3.23 3.29 3.00 8.37 3.23 3.29 4.00 9.89 3.23 3.29 5.00 10.21 3.23 3.29 Table 7.9 contains the proposed PCE values recommended to improve truck equivalency estimation in the HCM 2010 for grade percen tages for 5% trucks. Table 7.9 PCE Recommendation based on Grade Percenta ges for 5% Trucks HCM 2010 Length of grade (mi) Simulation Length of grade (mi) Grade Percentages (%) 2Lane 5% Trucks HCM 2010 PCE 3 Lane no trucks restrictions 5% Trucks 3 Lane with trucks restrictions 5% Trucks PCE PCE PCE 00.25 0.25 2% 1.80 1.5 1.83 2.00 >0.250.5 0.50 1.80 1.5 1.83 2.00 >0.50.75 0.75 1.80 1.5 1.83 2.00 >0.751.0 1.00 2.14 2 2.74 2.82 >1.01.50 1.25 2.67 2.5 2.74 2.82 >1.5 1.50 2.85 2.5 2.74 2.82 2.00 3.61 2.74 2.82
PAGE 226
211 Table 7.9 (Cont.) HCM 2010 Length of grade (mi) Simulation Length of grade (mi) Grade Percentages (%) 2Lane 5% Trucks HCM 2010 PCE 3 Lane no trucks restrictions 5% Trucks 3 Lane with trucks restrictions 5% Trucks PCE PCE PCE 00.25 0.25 4% 1.80 1.5 2.00 2.82 >0.250.5 0.50 2.67 2.5 2.00 2.00 >0.50.75 0.75 3.13 3 3.40 1.83 >0.751.0 1.00 3.61 3.5 3.40 3.00 >1 1.25 4.64 4 3.60 2.82 1.50 4.85 3.60 2.82 2.00 5.07 3.70 2.82 00.25 0.25 6% 2.43 1.5 3.03 2.82 >0.250.5 0.50 3.07 3 3.41 3.20 >0.50.75 0.75 3.86 4 3.61 3.78 >0.751.0 1.00 4.70 4.5 4.64 4.64 >1 1.25 5.62 5 4.64 4.64 1.50 6.60 4.64 4.64 2.00 6.81 4.64 4.64 00.25 0.25 8% 2.54 2.5 3.61 4.64 >0.250.5 0.50 4.33 3.5 3.61 4.82 >0.50.75 0.75 4.64 4.5 4.64 4.82 >0.751.0 1.00 5.76 5 4.82 4.82 >1 1.25 6.00 5.5 5.95 5.95 1.50 6.24 5.95 5.95 2.00 6.49 5.95 5.95
PAGE 227
212 Table 7.10 contains the proposed PCE values recommended to improve truck equivalency estimation in the HCM 2010 for grade percen tages for 10% trucks. Table 7.10 PCE Recommendation based on Grade Percenta ges for 10% Trucks HCM 2010 Length of grade (mi) Simulation Length of grade (mi) Grade Percentages (%) 2Lane 10% Trucks HCM 2010 PCE 3 Lane no trucks restrictions 10% Trucks 3 Lane with trucks restrictions 10% Trucks PCE PCE PCE 00.25 0.25 2% 1.63 1.5 1.25 1.52 >0.250.5 0.50 1.63 1.5 1.42 1.52 >0.50.75 0.75 1.80 1.5 1.43 1.43 >0.751.0 1.00 2.14 2 1.43 1.43 >1.01.50 1.25 2.67 2.5 1.91 1.91 >1.5 1.50 2.67 2.5 1.91 1.91 2.00 2.85 1.91 1.91 00.25 0.25 4% 1.80 1.5 1.61 1.61 >0.250.5 0.50 2.67 2.5 2.09 2.09 >0.50.75 0.75 3.13 3 2.30 2.30 >0.751.0 1.00 3.61 3.5 2.51 2.62 >1 1.25 4.64 4 2.62 2.62 1.50 4.85 2.62 2.62 2.00 5.07 2.62 2.62 00.25 0.25 6% 2.33 1.5 1.87 1.87 >0.250.5 0.50 3.86 3 2.36 2.90 >0.50.75 0.75 4.70 4 2.57 2.90 >0.751.0 1.00 5.62 4.5 2.79 2.90 >1 1.25 6.60 5 2.90 2.90 1.50 7.67 3.14 3.14 2.00 8.83 3.14 3.14 00.25 0.25 8% 2.54 2.5 1.96 2.45 >0.250.5 0.50 4.33 3.5 2.45 2.67 >0.50.75 0.75 4.64 4.5 2.67 2.67 >0.751.0 1.00 5.76 5 2.89 3.00 >1 1.25 6.00 5.5 3.00 3.00 1.50 7.00 3.23 3.29 2.00 7.53 3.23 3.29
PAGE 228
213 To test the proposed PCE tables, a set of problems in t he appendix was created. The data set is based on realistic conditions: Interstate 70 in Colorado, at Empire Junction between MP 231.5 and 232.5, grades of 8%. T he traffic is 1,750 veh/h/lane (2lane both directions) for weekend winter Sunday tr affic, with approximately 6% to 8% of trucks percentages. The speed limit for both passen ger cars and trucks is 55 mph. Using the PCE proposed from this research shows in the appendix that base flow is higher than the HCM method for twolane free ways by two percent as number of lane decreases, the PCE increases. The base fl ow from PCE proposed is less than HCM method for threelane freeways by five p ercent as number of lane increases, the PCE decreases. When trucklane restriction ap plies, the PCE shows very little variability.
PAGE 229
214 REFERENCES AASHTO. (2004). A Policy on Geometric Design of Highways and Streets. AlKaisy, Ahmed. (2006). Passenger Car Equivalents for Heavy Vehicles at Freeways and Multilane Highways : Some Critical Issues ITE Journal. AlKaisy, A., Hall, F., and Reisman, E. (2002). Developing Passenger Car Equivalents for Heavy Vehicles on Freeways During Queue Discharge Flow. Transportation Research Vol. 36A. Anwaar, Ahmed., Boxel, Dan Van., Volovski, Mathew., Ana stasopoulos, Panagiotis Ch., and Sinha, Kumares C. (2011). Using Lagging Headways to Estimate Passenger Car Equivalents on Basic Freeway Sections. Journal of Transportation of the Institute of Transportatin Eng ineers, Volume 1, Issue 2. Bassok, A., Outwater, M.l., Frkonja, J., and Johnson, C (2009) Speed Difference between Cars and Trucks on Freeways Metrans Transportation Center. Benekohal, R.F., and Zhao, W. (1999). Delaybased Passenger Car Equivalents for trucks at Signalized Intersections. Transportation Research A. Volume 34. pp 437457. Bloomberg, L., and Dale, J. (2000). Comparison of VISSIM and CORSIM Traffic Simulation Models on a Congested Network. Transportation Research Record No. 1727. Transportation Research Board. Washington, DC. Cambridge Systematics, I. (2005). An Initial Assessment of Freight Bottlenecks on Highways. Federal Highway Administration Office of Transportat ion Policy Studies. Cate, Matthew A., and Urbanik II, Thoma. (2004). Another View of Truck Lane Restrictions. Transportation Research Record No. 1867. Transportat ion Research Board. Washington, DC. Craus, J., Polus, A., and Grinberg, I. (1979). A Revised Method for the Determination of Passenger Car Equivalencies Transportation Research A. Volume 14A. pp 241246.
PAGE 230
215 Cunagin, W., and Chang, C. (1982). Effects of Trucks on Freeway Vehicle Headways Under OffPeak Flow Conditions. Transportation Research Record No. 869. Transportation Research Board. Washington, DC. Cunagin, W., and Messer, C. (1983). Passenger Car Equivalents for Rural Highways. Transportation Research Record No. 905. Transportat ion Research Board. Washington, DC. Development Research Partner, Denver Metro Chamber of C ommerce, & Metro Denver Economic Development Corporation. (2007, Apri l) The Impact of I70 Congestion on Colorado Denver to Grand Junction. Demarchi, S.H., & Setti, J.R. (2003). Limitation of PassengerCar Equivalent Derivation For Traffic Streams with More Than One Truck Type. Transportation Research Record No. 1852. Transportat ion Research Board. Washington, DC. Elefteriadou, L., and Webster, N. (1998). A Simulation Study of Truck Passenger Car Equivalents (PCE) on Basic Freeway Sections. Transportation Research Part B (1999). Elliot, Alan C., and Woodward, Wayne A. (2007). Statistical Analysis Quick Reference Guidebook with SPSS Examples. London. Osage Publication Ltd. Fan, H. (1990). Passenger Car Equivalents for Vehicles on Singapore Exp ressways Transportation Research Vol. 24A. Federal Highway Administration. (2004). Traffic Analysis Toolbox Volume I: Traffic Analysis Tools Primer. Publication No. FHWAHRT04038. Federal Highway Administration. (2004). Traffic Analysis Toolbox Volume II: Decision Support Methodology for Selecting Traffic Analysis Tools Publication No. FHWAHRT04039. Federal Highway Administration. (2004). Traffic Analysis Toolbox Volume III: Guidelines for Applying Traffic Microsimulation Modelin g Software. Publication No. FHWAHRT04040. Federal Highway Administration. (2007). Traffic Analysis Toolbox Volume IV:Guidelines for Applying CORSIM Microsimulation Mode ling Software. Publication No. FHWAHOP07079. Federal Highway Administration. (2004). Traffic Analysis Toolbox Volume V:Traffic Analysis Tools Case Studies Â– Benefits and Application. Publication No. FHWAHOP06005.
PAGE 231
216 Federal Highway Administration. (2007). Traffic Analysis Toolbox Volume VI: Definition, Interpretation, and Calculation of Traf fic Analysis Tools Measures of Effectiveness. Publication No. FHWAHOP08054. Federal Highway Administration. (2008). Traffic Analysis Toolbox Volume VII: Predicting Performance with Traffic Analysis Tools. Publication No. FHWAHOP08055. Federal Highway Administration. (1985). FHWA Vehicle Types Retrieved December 30, 2011 from the Federal Highway Administration websi te: http://www.fhwa.dot.gov/policy/ohpi/vehclass.htm Field, Andy. (2005). Discovering Statistics Using SPSS London. Osage Publication Ltd. Geistefeldt, Justin. (2009). Estimation of Passenger Car Equivalents Based on Capacity Variability Transportation Research Record No. 2130. Transportation Research Board. Washington, DC. Greenshields, B.D. (1933). The Photographic Method of Studying Traffic Behavior Proceedings of the 13 th Annual Meeting of the Highway Research Board. Greenshields, B.D. (1935). A Study of Highway Capacity Proceedings Highway Research Record. Volume 14. Washington, DC. Hayter, Anthony. J. (2002). Probability and Statistic For Engineers and Scientists. Pacific Group, CA. Duxbury. Highway Capacity Manual (HCM) 2010. Transportation Research Board. Washington, DC. Huber, M. (1982). Estimation of Passenger Car Equivalents of Trucks in Traff ic Stream Transportation Research Record No. 869. Transportat ion Research Board. Washington, DC. Husch, D & Albeck, J. (2006). Traffic Signal Software Â– User Guide Synchro Studio 7 Synchro plus SimTraffic and 3D Viewer. Sugarland, TX. Trafficware, Ltd. Ingle, Anthony. (2004). Development of Passenger Car Equivalents for Basic Freeway Segment Thesis. Virginia Polytechnic Institute and State Univ ersity. John, A., & Glauz, W. (1976). Speed and Service on Multilane Upgrades Transportation Research Record No. 615. Transportati on Research Board. Washington, DC.
PAGE 232
217 Joslin, Devin, C. (2010). Fuel and Emissions Modeling Calculations in Transportati on Analysis Software Master Report. University of Colorado Denver. Keller, E.L, and Saklas, J.G. (1984). Passenger Car Equivalents from Network Simulation. Journal of Transportation Engineering 110 No. 4. Krammes, R., and Crowley, K. Passenger Car Equivalents for Trucks on Level Freeway Segments. Transportation Research Record No. 1091. Transportation Research Board. Washington, DC. May, Adolf D. (1990). Traffic Flow Fundamentals PrenticeHall Inc. Marlina, S., & Janson, B. (2011). Understanding the Dynamic of Truck Traffic Using VISSIM MicroSimulation of Zipper Lane Option on Int erstate 70 ITE Western Distric Annual Meeting. Anchorage, Alaska. Mc Shane, William R., & Roess, Roger P. (1990). Traffic Engineering New Jersey. Prentice Hall, Inc. Meyer, Michael D. (2001). Urban Transportation Planning: A DecisionOriented Approach New York. McGrawHill. Middleton, D., Venglar, S., Quiroga, C., Lord, D., & Jasek, D. (2006). Strategies for Separating Trucks from Passenger Vehicles: Final Report. Research Report FHWA/TX07/046632. College Station: Texas Transportation Institute. Molina, C.J. (1987). Development of Passenger Car Equivalencies for Large Trucks at Signalized Intersections. ITE Journal. November. NHTSA. (2006). Traffic Safety Facts. Department of Transportation. Norusis, Marija J. (2005). SPSS 13.0 Statistical Procedures Companion. New Jersey. Prentice Hall, Inc. Planung Tranport Verker AG. (2008, July) Vissim 5.10 User Manual. Germany Rakha, H., Ingle, A., Hancock, K., & AlKaisy, A. (2007 ). Estimating Truck Equivalencies for Freeway Sections Transportation Research Record No. 2027. Transportation Research Board. Washington, DC. Roess, Roger P., and Messer, Carrol J. Passenger Car Equivalents for Uninterrupted Flow: Revision of Circular 212 Values Transportation Research Record No. 971. Transportation Research Board. Washington, DC.
PAGE 233
218 Seguin, E., Crowley, K., and Zweig, W. (1982). Passenger Car Equivalents on Urban Freeways. Report DTFH6180C00106. FHWA. US Department of Transportation. Siuhi, Saidi., and Mussa, Renatus. (2007). Simulation Analysis of TruckRestricted and HighOccupancy Vehicle Lanes Transportation Research Record No. 2012. Transportation Research Board. Washington, DC. Sumner, R., Hill, D., & Shapiro, S. (1984). Segment Passenger Car Equivalent Values for Cost Allocation on Urban Arterial Roads. Transportation Research Volume 18A No. 5/6. Traffic Engineering Handbook. 6 th Edition. Institute of Transportation Engineers. (2009). Transportation Research Board. (2003). Review of Truck Characteristics as Factors in Roadway Design. NCHRP Report No. 505. Van Aerde, M., and Yagar, S. (1983). Capacity, Speed, and Platooning Vehicle Equivalents for TwoLane Rural Highways. Transportation Research Record No. 971. Transportation Research Board. Washington, DC. Werner, A., and Morrall, J. Passenger Car Equivalencies of Trucks, Buses, and Recreational Vehicles for TwoLane Rural Highways. Transportation Research Record No. 615. Transportation Research Board. Washington, DC. Yun, S.., White, W.W., Lamb, D.R., & Wu, Y. (2005). Accounting for the Impact of Heavy Truck Traffic in VolumeDelay Functions in Transport ation Panning Model. Transportation Research Record No. 1931. Transportat ion Research Board. Washington, DC.
PAGE 234
219 APPENDIX Data: Interstate 70 in Colorado, at Empire Junction between MP 231.5 and 232.5 (the distances used in the calculations are 0.5 and 1.0 miles) grades of 8%. The assumed traffic is 1,750 veh/h/lane (2lane both di rections) weekend winter Sunday traffic, with approximately 6% to 8% trucks (the truck proportions used in the calculations are 5% and 10%). The traffic density corresponds to acceptable level of service (LOS C or worse ). The speed limit for both passenger cars and truck is 55 mph. Using Demarchi and SettiÂ’s formula, the base flow is cal culated: < g n r g g Where = proportion of trucks of type i out of all trucks n in t he mixed traffic flow = base flow rate (passenger cars only) = mixed flow rate (passenger cars + trucks) = passenger equivalent of trucks
PAGE 235
220 Example Problem 1 PCE for 0.5 mile Â– 5% trucks 1. PCE from HCM 2010 = 3.5 = 1750 x (1 Â– 0.05) + (1750 x 0.05 x 3.5) = 1969 pc/h/ ln 2. PCE Proposed a. TwoLanes = 4.48 = 1750 x (1 Â– 0.05) + (1750 x 0.05 x 4.48) = 2055 pc/h /ln b. ThreeLane No Trucks Restrictions = 3.61 = 1750 x (1 Â– 0.05) + (1750 x 0.05 x 3.61) = 1978 pc/h /ln c. ThreeLane With Trucks Restrictions = 3.78 = 1750 x (1 Â– 0.05) + (1750 x 0.05 x 3.78) = 1993 pc/h /ln
PAGE 236
221 Example Problem 2 PCE for 1.0 mile Â– 5% trucks 1. PCE from HCM 2010 = 5.5 = 1750 x (1 Â– 0.05) + (1750 x 0.05 x 5.5) = 2144 pc/h/ ln 2. PCE Proposed a. TwoLanes = 5.76 = 1750 x (1 Â– 0.05) + (1750 x 0.05 x 5.76) = 2167 pc/h /ln b. ThreeLane No Trucks Restrictions = 4.82 = 1750 x (1 Â– 0.05) + (1750 x 0.05 x 4.82) = 2084 pc/h /ln c. ThreeLane With Trucks Restrictions = 4.82 = 1750 x (1 Â– 0.05) + (1750 x 0.05 x 4.82) = 2084 pc/h /ln
PAGE 237
222 Example Problem 3 PCE for 0.5 mile Â– 10% trucks 1. PCE from HCM 2010 = 3.0 = 1750 x (1 Â– 0.10) + (1750 x 0.10 x 3.0) = 2100 pc/h/ ln 2. PCE Proposed a. TwoLanes = 3.36 = 1750 x (1 Â– 0.10) + (1750 x 0.10 x 3.36) = 2163 pc/h /ln b. ThreeLane No Trucks Restrictions = 2.30 = 1750 x (1 Â– 0.10) + (1750 x 0.10 x 2.30) = 1978 pc/h /ln c. ThreeLane With Trucks Restrictions = 2.39 = 1750 x (1 Â– 0.10) + (1750 x 0.10 x 2.39) = 1993 pc/h /ln
PAGE 238
223 Example Problem 4 PCE for 1.0 mile Â– 10% trucks 1. PCE from HCM 2010 = 4.5 = 1750 x (1 Â– 0.10) + (1750 x 0.10 x 4.5) = 2363 pc/h/ ln 2. PCE Proposed a. TwoLanes = 4.17 = 1750 x (1 Â– 0.10) + (1750 x 0.10 x 4.17) = 2305 pc/h /ln b. ThreeLane No Trucks Restrictions = 2.91 = 1750 x (1 Â– 0.10) + (1750 x 0.10 x 2.91) = 2084 pc/h /ln c. ThreeLane With Trucks Restrictions = 2.91 = 1750 x (1 Â– 0.10) + (1750 x 0.10 x 2.91) = 2084 pc/h /ln
