Development of Passenger Car Equivalents for Basic

Freeway Segments
Anthony Ingle

Thesis submitted to the faculty of the Virginia Polytechnic Institute and State University
in partial fulfillment of the requirements for the degree of
Master of Sciences
In
Civil Engineering

Hesham Rahka, Chair

Antonio Trani
Kyoungho Ahn

July 8, 2004
Blacksburg, Virginia
Keywords: passenger car equivalent, highway capacity manual, weight to power ratio

Development of Passenger Car Equivalents for Basic

Freeway Segments
Anthony Ingle

Abstract
Passenger car equivalents (PCEs) are used in highway capacity analysis to convert a
mixed vehicle flow into an equivalent passenger car flow. This calculation is relevant to
capacity and level of service determination, lane requirements, and determining the effect
of traffic on highway operations. The most recent Highway Capacity Manual 2000
reports PCEs for basic freeway segments according to percent and length of grade and
proportion of heavy vehicles. Heavy vehicles are considered to be either of two
categories: trucks and buses or RVs. For trucks and buses, PCEs are reported for a typical
truck with a weight to power ratio between 76.1 and 90.4 kg/kW (125 and 150 lb/hp).
The weight to power ratio is an indicator of vehicle performance. Recent development of
vehicle dynamics models make it possible to define PCEs for trucks with a wider variety
of weight to power ratios. PCEs were calculated from the relative impact of trucks on
traffic density using the simulation model INTEGRATION. The scope of this research
was to evaluate PCEs for basic freeway segments for trucks with a broader range of
weight to power ratios. Such results should make freeway capacity analysis more
accurate for mixed vehicle flow with a non-typical truck population. In addition, the
effect of high proportion of trucks, pavement type and condition, truck aerodynamic
treatment, number of freeway lanes, truck speed limit, and level of congestion was
considered. The calculation of PCEs for multiple truck weight to power ratio populations
was not found to be different from single truck weight to power ratio populations. The
PCE values were tabulated in a compatible format to that used in the Highway Capacity
Manual 2000.

Ingle

Acknowledgments
I would like to thank the U.S. Department of Transportation, Federal Highway
Administration, and National Highway Institute for providing assistance to my pursuit of
a Masters degree with the award of the Dwight David Eisenhower Transportation
Fellowship. The fellowship has provided me the excellent opportunity to pursue my
educational dreams at the school of my choice. I also appreciate the support of the Via
Department of Civil and Environmental Engineering at Virginia Polytechnic Institute and
State University. I appreciate the guidance given me by my advisor Dr. Hesham Rahka
and the other members of my advisory committee. My final acknowledgment is owed to
God who is my creator and responsible for all my accomplishments.

iii

Ingle

Table of Contents
Abstract... ii
Acknowledgements iii
Table of Contents... iv
List of Figures vi
List of Tables.. vii
Chapter 1. Introduction 1
1.1 Background 1
1.2 Research Objectives and Research Significance... 4
1.3 Paper Layout.. 6
Chapter 2. Literature Review... 7
2.1 PCEs in the 1965 HCM. 7
2.2 PCEs Based on Delay 8
2.3 PCEs in the TRB Circular 212... 9
2.4 PCEs Based on Speed 11
2.5 PCEs in the 1985 HCM. 15
2.6 PCEs Based on v/c Ratio... 15
2.7 PCEs Based on Headways. 16
2.8 PCEs Based on Queue Discharge Flow. 18
2.9 PCEs Based on Density. 19
Chapter 3. Methodology 21
3.1 Overview 21
3.2 Variables Considered. 24
3.3 Plan of Research 24
3.3.1 Physical Network 24
3.3.2 Integration Inputs 27
3.3.3 Simulation Outputs. 27
3.3.4 Examination of the Effect of Weight to Power Ratio. 28
3.3.5 Examination of the Effect of Engine Power... 31
3.3.6 Examination of the Effect of Proportion of Trucks 31
3.3.7 Examination of the Effect of Pavement Type and Condition. 32
3.3.8 Examination of the Effect of Aerodynamic Treatment.. 32
3.3.9 Examination of the Effect of Three Lane Segments... 33
3.3.10 Examination of the Effect of Truck Speed Limit..33
3.3.11 Examination of the Effect of Level of Congestion... 33
3.4 Data Analysis. 34
Chapter 4. Results for Single Truck Populations35
4.1 Background 35
4.2 Speed-flow-density relationships.. 38
4.3 Variability of PCE by Weight to Power Ratio.. 41
4.4 Variability of PCE by Length of Grade. 46
4.5 Variability of PCE by Percent Grade.... 48
4.6 Variability of PCE by Engine Power. 50
4.7 Variability of PCE by Proportion of Trucks.. 52
4.8 Variability of PCE by Pavement Type and Condition.. 54

iv

Ingle

4.9 Variability of PCE by Aerodynamic Treatment 57

4.10 Variability of PCE for Three Lane Segments.. 59
4.11 Variability of PCE for Truck Speed Limit.. 61
4.12 Variability of PCE by Level of Congestion. 61
4.13 Combined Effects of Variables on PCEs. 63
4.14 Conclusions.. 69
Chapter 5. Results for Multiple Truck Populations... 72
5.1 Background 72
5.2 Speed-flow-density relationships.. 76
5.3 Variability of PCE by Weight to Power Ratio.. 78
5.4 Variability of PCE by Length of Grade. 82
5.5 Variability of PCE by Percent Grade.... 84
5.6 Variability of PCE by Proportion of Trucks.. 84
5.7 Variability of PCE by Pavement Type and Condition.. 86
5.8 Variability of PCE by Aerodynamic Treatment 89
5.9 Variability of PCE for Three Lane Segments 91
5.10 Variability of PCE for Truck Speed Limit.. 93
5.11 Variability of PCE by Level of Congestion. 93
5.12 Conclusions.. 96
Chapter 6. Conclusions and Recommendations..99
6.1 Recommended PCE Table. 99
6.2 Variables Not Found to Have a Significant Effect on PCEs. 101
6.3 Variables Found to Have a Significant Effect on PCEs 102
6.3.1 Recommendation to Account for the Effect of
Weight to Power Ratio 103
6.3.2 Recommendation to Account for the Effect of
Length of Grade.. 105
6.3.3 Recommendation to Account for the Effect of
Percent Grade.. 105
6.3.4 Recommendation to Account for the Effect of
Proportion of Trucks... 105
6.3.5 Recommendation to Account for the Effect of
Pavement Type and Condition 106
6.3.6 Recommendation to Account for the Effect of
Truck Speed Limit.. 108
6.3.7 Recommendation to Account for the Effect of
Level of Congestion 110
6.3.8 Recommendation to Account for the Combined
Effect of Variables.. 112
6.4 Summary of Conclusions and Recommendations. 113
References.. 115
Appendix. 118
Vita.. 122

Ingle

List of Figures
Figure 1. Idealized link structure for INTEGRATION simulations (not to scale).. 26
Figure 2. Truck survey results at I-81 Troutville weigh station in Virginia 29
Figure 3. Speed-flow-density relationships, base vehicle flow. 39
Figure 4. Speed-flow-density relationships, mixed vehicle flow single truck
population of 83.7 kg/kW (137.5 lb/hp) 40
Figure 5. PCE variability by weight to power ratio for selected grades and LOS C
congestion.. 42
Figure 6. PCE variability by length of grade for average single truck population,
83.7 kg/kW (137.5 lb/hp) and LOS C congestion. 47
Figure 7. PCE variability by percentage grade for average single truck population,
83.7 kg/kW (137.5 lb/hp) and LOS C congestion. 49
Figure 8. PCE variability by proportion of trucks for average single truck population,
83.7 kg/kW (137.5 lb/hp) and LOS C congestion. 53
Figure 9. PCE variability by level of congestion for average single truck population, 83.7
kg/kW (137.5 lb/hp)..
64
Figure 10. Truck distribution for J and K multiple truck populations. 73
Figure 11. Truck distribution for L, M, N, and O multiple truck populations. 75
Figure 12. Speed-flow-density relationships, mixed vehicle flow multiple truck
population of 83.7 kg/kW (137.5 lb/hp) and standard deviation of 25.12
kg/kW (41.25 lb/hp).. 77
Figure 13. PCE variability by multiple truck populations at LOS C congestion 79
Figure 14. PCE variability by length of grade for N multiple truck population,
83.7 kg/kW (137.5 lb/hp) and standard deviation of 25.12 kg/kW
(41.25 lb/hp) at LOS C congestion 83
Figure 15. PCE variability by percentage grade for N multiple truck population,
83.7 kg/kW (137.5 lb/hp) and standard deviation of 25.12 kg/kW
(41.25 lb/hp) at LOS C congestion 85
Figure 16. PCE variability by proportion of trucks for N multiple truck population,
83.7 kg/kW (137.5 lb/hp) and standard deviation of 25.12 kg/kW
(41.25 lb/hp) at LOS C congestion 87
Figure 17. PCE variability by level of congestion for N multiple truck population,
83.7 kg/kW (137.5 lb/hp) and standard deviation of 25.12 kg/kW
(41.25 lb/hp) at LOS C congestion 95

vi

Ingle

List of Tables
Table 1. LOS categories for basic freeway segments as used in the HCM 2000 (1).. 3
Table 2. Complete list of variables considered 25
Table 3. Refined list of variables considered.. 26
Table 4. PCEs for trucks and buses on upgrades, average single truck population
83.7 kg/kW (137.5 lb/hp) and LOS C congestion. 44
Table 5. Ratio of PCE for light and heavy single truck populations to PCE for
average single truck population. 45
Table 6. Percentage difference in calculated PCE for different engine powers
compared to 75th percentile engine power 354 kW (475 hp). 51
Table 7. PCEs for trucks and buses on upgrades with different pavement types,
average single truck population 83.7 kg/kW (137.5 lb/hp) and
LOS C congestion.. 55
Table 8. Percentage difference in calculated PCE for different pavement types
compared to asphalt fair pavement. 55
Table 9. Percentage difference in calculated PCE for different aerodynamic
treatments compared to full aerodynamic treatment.. 58
Table 10. Percentage difference in calculated PCE for three lane freeway segments
compared to two lane freeway segments. 60
Table 11. Percentage difference in calculated PCE for a truck speed limit at
88.5 km/hr (55 mi/h) compared to no separate truck speed limit
112.6 km/hr (70 mi/h). 62
Table 12. Percentage difference in calculated PCE for different levels of congestion,
average single truck population 83.7 kg/kW (137.5 lb/hp). 65
Table 13. Case study 1 and 2 for combined effect of variables... 68
Table 14. PCEs for trucks and buses on upgrades, N multiple truck population, 83.7
kg/kW (137.5 lb/hp) and standard deviation of 25.12 kg/kW (41.25 lb/hp)
at LOS C congestion80
Table 15. Percentage difference in calculated PCE for N multiple truck population
compared to the average single truck population 81
Table 16. PCEs for trucks and buses on upgrades with different pavement types,
N multiple truck population, 83.7 kg/kW (137.5 lb/hp) and standard
deviation of 25.12 kg/kW (41.25 lb/hp) at LOS C congestion88
Table 17. Percentage difference in calculated PCE for different pavement types
compared to asphalt fair pavement. 88
Table 18. Percentage difference in calculated PCE for different aerodynamic
treatments compared to full aerodynamic treatment... 90
Table 19. Percentage difference in calculated PCE for three lane freeway segments
compared to two lane freeway segments. 92
Table 20. Percentage difference in calculated PCE for a truck speed limit at
88.5 km/hr (55 mi/h) compared to no separate truck speed limit
112.6 km/hr (70 mi/h). 94
Table 21. Percentage difference in calculated PCE for different levels of congestion,
N multiple truck population, 83.7 kg/kW (137.5 lb/hp) and standard
deviation of 25.12 kg/kW (41.25 lb/hp).. 97

vii

Ingle

Table 22. PCEs for basic freeway segments for trucks and buses on upgrades.. 100
Table 23. PCE scale factors to account for the effect of weight to power ratio.. 104
Table 24. PCE scale factors to account for the effect of pavement type and condition.. 107
Table 25. PCE scale factors to account for the effect of a truck speed limit... 109
Table 26. PCE scale factors to account for the effect of level of congestion.. 111

viii

Ingle

Chapter 1. Introduction
1.1 Background
Highway capacity is expressed in passenger cars per hour per lane. This is a
measure of the maximum throughput of vehicles that can be expected to pass a point on a
section of highway under prevailing roadway, traffic, and control conditions. The
presence of large and/or low performance vehicles in the traffic stream results in a
reduction of the allowable throughput. The Highway Capacity Manual (HCM) reasons
that the reduction in allowable throughput is due to the fact that heavy vehicles take up
more space and more importantly that heavy vehicles have lower performance, especially
on grades. Traffic volumes containing a mix of vehicle types must be converted into an
equivalent flow of passenger cars using passenger car equivalents (PCEs). The procedure
in the HCM allows that freeway traffic volumes containing a mix of vehicle types be
adjusted by the use of a heavy vehicle factor, fHV, into an equivalent flow rate of
passenger cars. The heavy vehicle adjustment factor is based on the passenger car
equivalence of trucks, buses, and recreation vehicles (RVs). According to the HCM 2000
(1), the heavy vehicle adjustment factor is found as
f HV =

1
1 + PT (ET 1) + PR (E R 1)

(1.1)

where PT and PR are the proportion of trucks/buses and RVs in the traffic stream
respectively, and ET and ER are the PCEs for trucks/buses and RVs respectively. The
HCM considers trucks and buses to have the same PCE because trucks are generally the
only heavy vehicle type present in the traffic stream.
PCEs are given in the HCM 2000 for extended freeway segments as well as for
specific grades. An extended freeway segment may contain a number of upgrades,
1

Ingle

downgrades, and level segments without a single uniform grade. As long as no single
grade is of 3% or greater and longer than 0.4 km (0.25 mi) or no single grade is of less
than 3% and longer than 0.8 km (0.5 mi), extended segment analysis may be applied. For
grades that exceed these criteria, a specific grade analysis of heavy vehicles should be
applied due to their significant effect on traffic flow. For extended freeway segments,
PCEs are given according to the terrain type: level, rolling, or mountainous. No
consideration is given as to the truck type, proportion of trucks, or freeway facility type.
PCEs for specific grades however, are given with regard to percent of grade, length of
grade, and proportion of heavy vehicles. In addition, the HCM 2000 provides a method
for calculating PCEs on composite grades.
Level of service (LOS) is a familiar term in highway capacity analysis. This term,
originally introduced in the 1965 HCM, correlates the drivers perception of operating
conditions with traffic flow parameters such as density, speed, volume to capacity (v/c)
ratio, and flow rate. LOS operating conditions are divided into five categories, A through
E. Each of these categories represents a range of traffic flow parameters that reflect a
drivers perceptions of quality of service. Quality of service includes speed, travel time,
freedom to maneuver, traffic interruptions, comfort, and convenience. The HCM 2000
primarily defines LOS in terms of density; the LOS categories for basic freeway
segments are provided in Table 1. These LOS categories are based on a free-flow speed
of 120 km/h (75 mi/h).

Ingle

Table 1. LOS categories for basic freeway segments as suggested in the HCM 2000 (1).

LOS

Maximum Density

A
B
C
D
E

(pc/km/ln) (pc/ml/ln)
7
11
11
18
16
26
22
35
28
45

Minimum Speed
(km/h)
120
120
114.6
99.6
85.7

(mi/h)
75
74.8
70.6
62.2
53.3

Maximum v/c

0.34
0.56
0.76
0.90
1.00

Maximum
Flow Rate
(pc/h/ln)
820
1350
1830
2170
2400

Ingle

1.2 Research Objectives and Research Significance

The objective of this research was to evaluate PCEs for basic freeway segments
considering variables beyond the scope of the current HCM 2000. Specifically, the
variables that were examined are variance in truck weight to power ratio, truck power,
high proportion of trucks, pavement type and condition, truck aerodynamic treatment,
number of freeway lanes, truck speed limit, and level of congestion.
An objective of this research was to verify the hypothesis that a single truck
population, containing only a typical weight to power ratio, will perform the same as a
multiple truck population, containing a mix of weight to power ratios. The effects of all
variables on single truck population PCEs were compared with the effects of all variables
on multiple truck population PCEs. Different truck populations were simulated by
changing the standard deviation of weight to power ratio around the average. Variability
in truck power was tested for the average single weight to power ratio truck population. It
was hypothesized that PCEs will be higher for trucks with lower engine power. Another
objective of this research was to verify a trend observed in the HCM 2000; that the PCE
for a realistic truck population decreases with increasing proportion of trucks. In this
research, an even higher proportion of trucks was examined than the limits of the HCM
2000, which considers up to 25% proportion of trucks.
Pavement type and condition have an impact on truck performance. An objective
of this research was to verify the hypothesis that PCEs developed for trucks will vary by
pavement type and condition. Poor pavement condition was expected to result in
increased PCE values compared to the same truck population operating on good
pavement condition. The development of PCEs by pavement condition could result in a

Ingle

capacity justification for pavement resurfacing or reconstruction. Another objective was

to verify the hypothesis that truck aerodynamic treatment will have a significant effect on
calculated PCEs. Both pavement type and condition as well as aerodynamic treatment
affect the vehicle dynamics of the truck populations.
An objective of this research was to verify the hypothesis that trucks on freeways
with more than two directional lanes will have lower PCEs than trucks on freeways with
only two directional lanes. This hypothesis is based on the availability of passing being
increased for freeways with more than two directional lanes. In addition, the effect of a
truck lane restriction to the rightmost two lanes was considered. Another objective was to
verify the hypothesis that the imposition of a truck speed limit below the speed limit of
all other vehicles results in higher PCEs. Many states throughout the nation have imposed
truck lane restrictions and truck speed limits. The results of this research could supply a
capacity justification for or against these practices.
Lastly, it was hypothesized that increasing level of congestion will result in
increasing PCEs. The prevailing traffic density represents the level of congestion on the
freeway segment. As congestion increases, the amount of interaction of trucks with
passenger cars and other trucks should increase. The negative aspects of this interaction
were expected to result in increasing PCEs for increasing prevailing traffic density.
The significance of this research is that PCEs developed herein can be utilized to
enhance the HCM 2000. Improvements to the accuracy of highway capacity analysis
should increase justification for capacity improvements. PCEs developed as part of this
research were tabulated in a form compatible with the current HCM 2000, with the hope
that they can be incorporated into a future edition of the HCM.

Ingle

1.3 Paper Layout

Chapter 2 is a literature review of previous research concerning the development
of PCEs. It is organized both by subject methodology and chronologically, progressing
from PCEs applied in the 1965 HCM to those applied in the current HCM 2000. The
literature review is a substantial resource for anyone interested in the historical
development of PCEs. Many alternative methods for calculating PCEs are presented with
full references.
Chapter 3 describes the methodology for this research. It describes the approach
to calculating PCEs and the parameters needed for the traffic simulation program
INTEGRATION. Specific attention is given to examining the effect of the variables
outlined in the research objectives.
Chapter 4 is a presentation of the results specific to single truck populations.
Chapter 5 is a presentation of the results specific to multiple truck populations.
Chapter 6 presents the conclusions and recommendations drawn from the results
of this research.

Ingle

Chapter 2. Literature Review

Various methods have been used to calculate PCEs throughout the evolution of
highway capacity analysis. These methods have been applied both for two lane highways
and multilane highways or freeways.
2.1 PCEs in the 1965 HCM
The 1965 HCM, which was the second edition of the HCM, formally introduced
the concept of LOS and the definition of PCE (2). In the 1965 HCM, PCE was defined as
The number of passenger cars displaced in the traffic flow by a truck or a bus, under the
prevailing roadway and traffic conditions (3). The 1965 HCM used relative speed
reduction to define PCEs for two lane highways and quantified this by the relative
number of passings known as the Walker method. For multilane highways, PCEs were
based on the relative delay due to trucks (4).
The relative delay due to trucks was calculated using the Walker method for twolane highways in conjunction with gradability curves and field observations. Gradability
curves relate speed distribution verses grades of specific length and percent. Steeper
grades and longer grades result in a more drastic speed reduction. Cunagin and Messer
(4) suggested that the gradability curve used to calculate PCEs in the 1965 HCM. was
based on a truck with a weight to power ratio of 197.9 kg/kW (325 lb/hp), which was
considered typical for trucks of the time. However, Roess and Messer (13) emphasized
that the normal truck assumed in the 1965 HCM was of 121.8 kg/kW (200 lb/hp).
Regardless of which truck was assumed, the gradability curve was eventually considered
obsolete for vehicle performance calculations and was updated in subsequent years.

Ingle

PCEs for multilane highways based on relative delay may be found as

ET =

(D

ij

DB )

[2.1]

DB

where Dij is the delay to passenger cars due to vehicle type i under condition j and DB is
the base delay to standard passenger cars due to slower passenger cars.
PCEs in the 1965 HCM were reported for grades of specific length and percent,
proportion of trucks, and LOS grouped as A through C or D and E. As expected, the
highest PCE was reported for the longest and steepest grade with the highest proportion
of trucks and the lowest LOS. However, in many cases the PCE for a given grade and
LOS decreased with increasing proportion of trucks. This result has been obtained by
many other researchers, as mentioned later.
2.2 PCEs Based on Delay
In 1983, Cunagin and Messer (4) used an extension of the 1965 HCM method to
calculate PCEs for multilane highways based on relative delay. In their approach, they
used a combination of the Walker method of relative number of passings and the relative
delay method. They recognized that on multilane highways, passing or overtaking
vehicles are inhibited only by concurrent flow traffic. PCEs were calculated as
ET =

(OTi

(OTLPC

VOLi )[(1 SPM ) (1 SPB )]

VOLLPC )[(1 SPPC ) (1 SPB )]

[2.2]

where OTi is the number of overtakings of vehicle type i by passenger cars, VOLi is the
volume of vehicle type i, OTLPC is the number of overtakings of lower performance
passenger cars by passenger cars, VOLLPC is the volume of lower performance passenger
cars, SPM is the mean speed of the mixed traffic stream, SPB is the mean speed of the base

Ingle

traffic stream with only high performance passenger cars, and SPPC is the mean speed of
the traffic stream with only passenger cars.
Since at low traffic volumes faster vehicles will not likely be impeded in
overtaking other vehicles, equation [2.2] was used with the omission of the bracketed
expression. However at higher traffic volumes, such as near capacity, slower overtaking
vehicles will impede faster vehicles. This results in queue formation in the passing lane.
In their research, Cunagin and Messer applied a linear combination of equation [2.2] with
and without the bracketed expression for intermediate volumes.
Cunagin and Messer examined three different grade conditions, flat, moderate,
and steep. In addition, they examined proportion of trucks and volume levels
corresponding to each of the five LOS categories. The PCEs developed by Cunagin and
Messer increased for proportion of trucks and volume levels in flat and moderate grade
conditions. However, in steep grade conditions, the PCEs decreased for increasing
proportion of trucks.
2.3 PCEs in the TRB Circular 212
The TRB Circular 212 titled Interim Materials on Highway Capacity was
published in 1980, as an effort to summarize the current knowledge in highway capacity
and to identify needs for immediate research before the completion of the planned third
edition of the HCM (5). PCEs reported in TRB Circular 212 were developed based on the
constant v/c method. An article published by Linzer et al (6) in 1979 describes the
constant v/c method, whereby PCEs are calibrated such that the mixed traffic flow will
produce the same v/c ratio as a passenger car only flow.

Ingle

The research by Linzer et al made use of design charts resulting from

microsimulation done by the Midwest Research Institute (MRI) under the direction of St
John and Glauz (7) in 1976. The design chart relates the percent grade, mixed vehicle
flow, and percent reference trucks to percent capacity (equivalent to v/c ratio). The PCE
is formulated as
ET =

q B q M (1 PT )
q M * PT

[2.3]

where qB is the equivalent passenger car only flow rate for a given v/c ratio, qM is the
mixed flow rate, and PT is the proportion of trucks in the mixed traffic flow.
St John and Glauz introduced the concept of percent reference trucks to account
for the variability of truck performance characteristics by truck type. This was
accomplished by aggregating all truck types into a single reference truck. The Ohio
Department of Transportation provides an excellent copy of the most common vehicle
classification scheme on their website (20). The Federal Highway Administration
(FHWA) also follows this vehicle classification scheme whereby trucks are considered to
be vehicle types 5 through 13; the FHWA vehicle classification scheme is also available
online (21). For any given truck population, St John and Glauz derived weight factors to
compute the percent reference trucks. The derived weight factors were based on the
performance of each truck type relative to the slowest speed truck. The higher the weight
factor, the worse performing the subject truck is compared to the slowest speed truck.
The equation for percent reference trucks is

Percent reference trucks = PT (3.16 p10 + 1.41 p9 + 0.14 p8 + 0.06 p7 )

10

[2.4]

Ingle

where PT is the total proportion of trucks and pi is the proportion of index truck type i out
of the total proportion of trucks.
The typical truck used in calculation of PCEs for the TRB Circular 212 by Linzer
et al was of 182.7 kg/kW (300 lb/hp), slightly less than the 197.9 kg/kW (325 lb/hp) truck
used in the 1965 HCM, and reflecting the increased performance of trucks since the
1960s. In addition, a light truck of 91.4 kg/kW (150 lb/hp) and a heavy truck of 213.2
kg/kW (350 lb/hp) were used to calculate PCEs. Truck performance curves were used
from research conducted by Pennsylvania State University, with initial truck speed of
88.5 km/h (55 mi/h). Since the research by Linzer et al calculated PCEs for truck
populations with a single weight to power ratio, the percent reference trucks method
proposed by the MRI was used by assuming that only trucks of the given weight to power
ratio existed.
Results of the constant v/c method for calculating PCEs indicated that PCEs did
not change significantly for changes in the v/c ratio or the freeway design speed. For this
reason, PCEs reported in the TRB Circular 212 were given according to percent grade,
length of grade, and percent trucks just as they had been in the 1965 HCM. In addition
however, PCEs were calculated for freeways with six or more lanes as well as typical
freeways with four lanes. The need to calculate PCEs for different freeway sizes (number
of lanes) arose from cases of high proportion trucks and/or steep grades. PCEs developed
by Linzer et al exhibit a decrease for increasing percent trucks.
2.4 PCEs Based on Speed
As an extension to his research on truck performance on upgrades in 1976, St
John (8) proposed a non-linear truck factor. This non-linearity addressed the successively

11

Ingle

smaller impact of trucks on the traffic stream as the proportion of trucks increased. He
reasoned that as the proportion of trucks increases platoons may form and the interaction
with cars may be reduced. In addition, St John asserted that the effect of multiple truck
types highlights the need for a non-linear truck factor. The truck factor was based on a
speed flow relationship. He introduced the concept of equivalence kernel, which accounts
for the incremental effect of trucks in a traffic stream and is used to calculate PCEs.
Hu and Johnson (9) described how to use the 1965 HCM to find PCEs based on
speed in a report published in 1981. According to this report, PCEs are used to convert a
mixed vehicle flow into a passenger car only flow with the same operating speed. They
used equation [2.3] developed by Linzer et al (6) to calculate the PCE. Operating speeds
were based on the design charts obtained by research performed by the MRI, as described
in the section on the TRB Circular 212. Hu and Johnson did not use specific grade
adjustments, but rather developed their PCEs based on extended freeway segments.
In 1982, Huber (10) derived equation [2.3] in a different functional form to relate
PCE to the flow of a passenger car only traffic stream and a mixed vehicle traffic stream.
The effect of trucks is quantified by relating the traffic flows for an equal LOS. Any
equivalent LOS or impedance could be chosen for the equality. If for example, density
was used to define the equal LOS criteria, the flow-density relationship could be used to
relate the traffic flows at an equal density value. Hubers basic equation is formulated as

ET =

1
PT

qB

1 + 1
qM

[2.5]

where PT is the proportion of trucks in the mixed traffic flow, qB is the base flow rate
(passenger cars only), and qM is the mixed flow rate. Huber used the assumption of equal
average travel time as the measure of LOS. Equal average travel time on a one-mile
12

Ingle

segment is equivalent to the inverse of the average speed. The consequence of his
assumption of equal speed is that PCEs decrease as volumes increases. A slow moving
truck will have a smaller impact on the average speed when the total volume is higher.
Huber found this result objectionable and suggested that equal total travel time be used as
a measure of LOS. He formulated equal total travel time as the volume in vehicles per
hour multiplied by the average travel time in hours per mile. By this representation, equal
total travel time is equivalent to equal density because it describes equal vehicle
occupancy on the roadway in vehicles per mile. The calculation of PCE by equal density
is discussed later.
In 1984, Sumner et al (11) expanded the relationship described by Huber to
calculate the PCE of a single truck in a mixed traffic stream, which includes multiple
truck types. This calculation requires an observed base flow, mixed flow, and flow with
the subject vehicles. The equal LOS or impedance measure would cut across all three
flow curves. The relationship described by Sumner et al is formulated as

ET =

1 qB qB

+1

P q s qM

[2.6]

where P is the proportion of subject vehicles that is added to the mixed flow and
subtracted from the passenger car proportion, qB is the base flow rate (passenger cars
only), qM is the mixed flow rate, and qS is the flow rate including the added subject
vehicles. Sumner et al used total travel time in terms of vehicle hours as the equal
measure of LOS. In this case total travel time was applied to urban arterial roads and
measured in terms of vehicle hours, which is not equivalent to density.
Using the formulation in equation [2.6], Elefteriadou et al (3) calculated PCEs for
freeways, two-lane highways, and arterials in 1997 based on equal speed. The researchers
13

Ingle

also examined the impact of prevailing traffic flow, proportion of trucks, truck type (by
length and weight to power ratio), length and percent grade, and number of freeway lanes
in their evaluation. Their analysis was based on specific truck types, and not truck
populations. The results of the analysis by Elefteriadou et al indicated that PCEs remain
mostly unchanged for increasing traffic flow on freeway segments while PCEs remain
unchanged or slightly increase with increasing proportion of trucks. The report did not
indicate the impact of number of freeway lanes on the PCE.
In 1984, Van Aerde and Yagar (12) developed a methodology to calculate PCE
based on relative rate of speed reduction. This PCE was intended for use in average speed
analysis of capacity, which is unique to two lane highways. Field observations and
known speed-flow relationships were used to calibrate a multiple linear regression model
that estimates the percentile speed based on the free speed and speed reduction
coefficients for each vehicle type. A linear speed-flow model was chosen because the
speed-flow relationship within the bounds of practical operating volumes was found to be
nearly linear. The multiple linear regression model is
Percentile speed = free speed + C1(number of passenger cars) + C2(number of
trucks) + C3(number of RVs) + C4(number of other vehicles) + C5(number of
opposing vehicles)

[2.7]

where coefficients C1 to C5 are the relative sizes of speed reductions for each vehicle
type. Although this model was formulated for two lane highways with opposing traffic
flow, it could be applied to multilane highways by setting the coefficient C5 to zero.
Using the speed reduction coefficients, the PCE for a vehicle type n is calculated as

En =

Cn
C1

[2.8]

14

Ingle

where Cn is the speed reduction coefficient for vehilce type n and C1 is the speed
reduction coefficient for passenger cars.
2.5 PCEs in the 1985 HCM
Based on the recommendations of Roess and Messer (13), PCEs in the 1985 HCM
were calculated for trucks of 60.9, 121.8, and 182.7 kg/kW (100, 200, and 300 lb/hp)
with 121.8 kg/kW (200 lb/hp) being considered the normal truck population. The
consideration of freeway size, introduced in the TRB Circular 212, was retained in the
1985 HCM. The shift of the typical truck from 182.7 to 121.8 kg/kW (300 to 200 lb/hp)
was inspired by indications that the average truck population on freeways was between
76.1 and 103.5 kg/kW (125 and 170 lb/hp). Besides this change, the approach to
calculating PCE based on v/c ratio in the TRB Circular 212 remained unchanged in the
1985 HCM. Just as in the TRB Circular 212, PCEs were greatest for long steep grades,
but decreased for increasing proportion of trucks.
2.6 PCEs Based on v/c Ratio
After the publication of the 1985 HCM, the constant v/c method for calculating
PCE subsided. The constant v/c method was most appropriate when LOS was defined
primarily in terms of v/c ratio; however, since LOS is now defined primarily by density,
the constant v/c method is no longer favorable. Traffic streams with an equal v/c ratio
will not necessarily have equal density and speed and therefore LOS. However, Fan (14)
applied this method in 1989 to calculate PCEs for expressways in Singapore. He reasoned
that although density was used to define LOS for freeways, capacity analysis performed
with PCEs would still be desirable to be based on the v/c ratio. The functional form of his
relationship was a multiple linear regression equation whereby the v/c ratio was related to

15

Ingle

the PCE multiplied by the observed flow of each vehicle type. The target v/c ratio to
compute PCE was at 0.67 to 1.0, corresponding to LOS D or E. Fan pointed out that for
capacity analysis it would be unimportant to calculate PCEs at v/c ratios well below
capacity. The results of the research by Fan were PCEs for multiple vehicle types.
2.7 PCEs Based on Headways
Realizing one of the primary effects of heavy vehicles in the traffic stream is that
they take up more space, headways have been used for some of the most popular methods
to calculate PCEs. In 1976, Werner and Morrall (15) suggested that the headway method
is best suited to determine PCEs on level terrain at low levels of service. The PCE is
calculated as

ET = M PC PT
HB

[2.9]

where HM is the average headway for a sample including all vehicle types, HB is the
average headway for a sample of passenger cars only, PC is the proportion of cars, and PT
is the proportion of trucks. In their study, Werner and Morrall used the headway method
for low speed trucks and the conventional speed method of the 1965 HCM for higher
speed trucks. One question arises as to use of the headway method for low speed trucks
when low speeds generally occur on upgrades rather than on level terrain. The results of
the study by Werner and Morrall replicated PCEs in the 1965 HCM for higher speed
trucks. PCEs were categorized by percent grade, length of grade, and LOS grouped A and
B, C, or D and E.
In 1982, in an article by Cunagin and Chang (16) it was revealed that the presence
of trucks in the traffic stream of a freeway result in increased average headways. The
largest headways involved trucks following trucks, and the headways increased for larger
16

Ingle

truck types. Seguin et al (17) formulated the spatial headway method for calculating
PCEs in 1982. This method defines the PCE as the ratio of the mean lagging headway of
a subject vehicle divided by the mean lagging headway of the basic passenger car and is
formulated as
ET =

H ij

[2.10]

HB

where Hij is the mean lagging headway of vehicle type i under conditions j and HB is the
mean lagging headway of passenger cars. The lagging headway is determined from the
rear bumper of the lead vehicle to the rear bumper of the following vehicle and therefore
includes the following vehicles length.
The constant volume to capacity method, equal density method, and spatial
headway method were compared in 1986 in an article by Krammes and Crowley (2). The
authors concluded that the spatial headway method was most appropriate for level
freeway segments. Krammes points out that the spatial headway method not only
accounts for the accepted effect of trucks due to size and lower performance, but also the
psychological impact of trucks on drivers of other vehicles. This impact is in the form of
aerodynamic disturbances, splash and spray, sign blockage, offtracking, and underride
hazard.
Spatial headway is considered to be a surrogate measure for density. Both of
which reflect the freedom of maneuverability in a traffic stream. A modification to
equation [2.5] put forth by Huber to calculate PCE based on flow rate allows the
calculation of PCE based on headway. The equation uses the lagging headway because it
is the following vehicles perception of maneuverability that affects the PCE.
Contradictory to the findings of Cunagin and Chang, the lagging headway for trucks

17

Ingle

following trucks was found to be significantly less than the lagging headway for cars
following trucks. Therefore, in contrast to the recommended equation [2.10] by Seguin,
Krammes and Crowley suggest that PCE should be calculated as

ET = [(1 PT )H TP + pH TT ] / H P

[2.11]

where PT is the proportion of trucks, HTP is the lagging headway of trucks following
passenger cars in the mixed vehicle stream, HTT is the lagging headway of trucks
following trucks in the mixed vehicle stream, and HP is the lagging headway of cars
following either vehicle type in the mixed vehicle stream. An additional improvement
over equation [2.10] recommended by Seguin is that the proportion of trucks is
considered in equation [2.11]. Krammes and Crowley believe that an increase in the
proportion of trucks will result in higher PCEs because the opportunity for interaction
between cars and trucks will increase.
A drawback of the headway method is that it must be assumed that drivers are
exhibiting steady state, in lane behavior. It would be hard therefore to separate the
headways observed from drivers who are either not in steady state, or are not maintaining
the lane (continuously following the same vehicle). Specific to multilane highways, it is
less likely that cars will continue to follow trucks given the first opportunity to pass.
2.8 PCEs Based on Queue Discharge Flow
In 2002, Al-Kaisy et al (22) published a report describing the calculation of PCE
using measurements of queue discharge flow. Their hypothesis was that the effect of
trucks on traffic is greater during congestion than during under saturated conditions. The
congested condition is represented by queue discharge flow, where the v/c ratio is equal
to one. A primary assumption of their work was that queue discharge flow capacity is

18

Ingle

constant except for the effect of trucks in the traffic stream. Al-Kaisy et al used field
observations and linear programming to determine the PCE. For the case studies in their
analysis, Al-Kaisy et al did not find a relationship between PCE and the proportion of
trucks. However, they theorized that the PCE should decrease with increasing proportion
of trucks because the interactive effect of trucks on trucks may be less than the effect of
trucks on passenger cars.
2.9 PCEs Based on Density
As mentioned before, Huber (10) introduced the concept of using equal density to
relate mixed flow rate and base flow rate for calculation of PCE in equation [2.5]. The
drawback of Hubers computation is that it assumes the mixed vehicle flow contains
passenger cars and only one type of truck. However, the formulation by Sumner et al (11)
in equation [2.6] allows the calculation of the PCE of a single truck in a mixed vehicle
stream including multiple truck types. As applied to freeways, density is the most
common equal measure of LOS, and Webster and Elefteriadou (18) used this method to
calculate PCEs for trucks in 1999. Their approach was to use simulation modeling to
calculate the flow verses density relationships. Again, the researchers examined the
impact of prevailing traffic flow, proportion of trucks, truck type (by length and weight to
power ratio), length and percent grade, and number of freeway lanes in their evaluation.
The results of the analysis by Webster and Elefteriadou indicated that PCEs increase with
increasing traffic flow on freeway segments and decrease with increasing proportion of
trucks and number of lanes. The most important conclusion is that truck type, as defined
by length and weight to power ratio, is critical for determination of PCEs.

19

Ingle

In 2003, Demarchi and Setti (19) published an article describing the limitations of
deriving PCEs for traffic streams with multiple truck types. In an algebraic derivation,
they proved that PCEs developed for a single truck type in a mixed traffic flow
containing multiple truck types using equation [2.6] do not fully account for the
interaction between trucks. They reasoned that considered separately, the PCE value for
the subject vehicle is normally underestimated, because the marginal impact decreases as
the proportion of subject vehicles in the stream increases. Conversely, the impact of
trucks already in the mixed vehicle stream is overestimated because their actual
proportion should be smaller than it is prior to addition of the subject vehicles.
Demarchi and Setti suggested that a possible workaround to avoid the errors
associated with calculating the PCE for each truck separately is to calculate an aggregate
PCE formulated as

ET =

qB

1 + 1

Pi qM

[2.12]

where Pi is the proportion of trucks of type i out of all trucks n in the mixed traffic flow,
qB is the base flow rate (passenger cars only), and qM is the mixed flow rate. This
equation is basically equation [2.5] put forth by Huber and modified for multiple truck
types in the mixed traffic stream. This approach, using an aggregate PCE, seems to have
been adopted in the 1994, 1997, and 2000 editions of the HCM. PCEs in the HCM 2000
are reported by percent grade, length of grade, and percent trucks. The PCEs exhibit a
decrease for increasing proportion of trucks.

20

Ingle

Chapter 3. Methodology
3.1 Overview
Among the methods that have been employed to calculate PCEs, the equal density
method was selected for this research. A primary advantage of the equal density method
is that density is used in the HCM 2000 to define LOS. Density is an indicator of freedom
to maneuver in the traffic stream. In addition, density measurements are commonly made
on freeways using presence type detectors. The common use of density measurements in
the field makes it most practical for calculation of PCEs.
The HCM 2000 (1) defines PCE as The number of passenger cars displaced by a
single heavy vehicle of a particular type under specified roadway, traffic, and control
conditions. The PCE is also referred to as the number of passenger cars that would use
the same amount of freeway capacity as a single truck or bus. It is common in this sense
of the definition to equate a mixed vehicle flow to a passenger car only vehicle flow
using the PCE. Equations [3.1] through [3.3] show the derivation of an aggregate PCE.
A traffic stream may contain any number of trucks of type i, in total amounting to
n type trucks. Each of these truck types may be either a separate vehicle class or a distinct
weight to power ratio, thus representing different vehicle sizes and performance
characteristics. The flow verses density relationship may be obtained for this traffic
stream by measuring the density and flow rate of a given number of simulations of the
traffic stream. In a similar way, the flow verses density relationship for a traffic stream
containing only passenger cars may be obtained. Figure 6 shows the flow verses density
relationships for this situation.

21

Ingle

Flow-Density (0.8 km)

\
3000.0

Base Flow

Flow (veh/h/ln)

2500.0
2000.0

Mixed Flow

kE

1500.0

qB

1000.0

qM

500.0
0.0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

Density (veh/km/ln)
qm 5% trucks 6% grade

qb 6% grade

Figure 6. Flow-density relationship for a 0.8 km (0.5 mi) grade.

22

35.0

Ingle

The flow of a passenger car only traffic stream may be related to the flow of a mixed
traffic stream using an aggregate PCE and formulated as

q B = 1 i P i q M + i Pi ET q M
n

[3.1]

where qB is the base flow rate (passenger cars only), Pi is the proportion of trucks of type
i, qM is the mixed vehicle flow rate, and ET is the aggregate PCE.
Dividing through by qM and subtracting 1 from each side yields

qB
n
1 = i Pi (ET 1)
qM
Finally, dividing through by

[3.2]

P and adding 1 to each side yields

n

qB
1 + 1 = ET

i Pi q M
1

[3.3]

The general method for calculating the PCE according to equation [3.3] is as
follows. First generate a flow verses density relationship for the base vehicle stream by
simulating passenger cars only. Simulations of the traffic stream were obtained using the
microscopic traffic simulator INTEGRATION. The simulation was conducted at 5
different flow rates, as shown in Table 1, and corresponding to the maximum service
flow rate for each LOS category from the HCM 2000. Second, generate a flow verses
density relationship for the mixed vehicle stream, replacing passenger cars with an equal
number of trucks from the subject truck population. The proportion of trucks in this
research was varied from 2 to 100 percent. Third, interpolate between observed values to
obtain the base flow rate and mixed vehicle flow rate at an equal density value. Initially
an equal density value of 12.4 pc/km/ln (20 pc/mi/ln), corresponding to a density at LOS
C, was used. Fourth, calculate the PCE according to equation [3.3].

23

Ingle

3.2 Variables Considered

As mentioned in the research objectives, a multitude of variables were considered,
which may have an impact on the development of PCEs. The physical characteristics of
the simulated network contain variables such as percent grade, length of grade, number of
freeway lanes, and pavement type and condition. In addition, the simulated traffic
characteristics contain variables such as flow rate, proportion of trucks, and variable truck
populations. Table 2 summarizes the total number of variables which could be
considered. The list of variables from Table 2 was refined for the investigative phase of
the research. A smaller list of variables was created, as shown in Table 3, from which the
simulations were performed. The refined list of variables made the research work load
less, while still accomplishing the research objective of determining the relationships
between variables. Final simulations were conducted to fill in the blanks within the data.
3.3 Plan of Research
The plan of research is summarized in the following sections.
3.3.1 Physical Network
The physical network was initially constructed using the 2, 4, and 6 percent grade
and the six different length of grade variables, resulting in 18 combinations. The structure
of the freeway section modeled contained two links in each chain. The first link was the
subject grade, having a length that varies between 0.4 and 2.4 km (0.25 and 1.5 mi) and a
grade that varies between 1 and 6 percent. The second link was always 0.11 km (0.07 mi
or 370 ft) long and was used to obtain measurements of traffic flow and density. To
economize simulation time, a parallel link structure was created in the simulation model.
This structure is idealized in Figure 1.

24

Ingle

Table 2. Complete list of variables considered.

Pavement Type and

Condition

Flow Rate (pc/h/ln)

% Trucks

Single and Multiple

Truck Populations
(lkg/kW) [lb/hp]

1
2
3
4
5
6

0.4 [0.25]
0.8 [0.50]
1.2 [0.75]
1.6 [1.00]
2.0 [1.25]
2.4 [1.50]

2
3

Asphalt-good
Asphalt-fair
Asphalt-poor
Conc-excellent
Conc-good
Conc-poor
Snow Covered

820 (LOS A)
1350 (B)
1830 (C)
2170 (D)
2400 (E)

0
2
4
5
6
8
10
15
20
25
30
40
50
60
70
80
90
100

38.1 [62.5]
53.3 [87.5]
68.5 [112.5]
83.7 [137.5]
99.0 [162.5]
114.2 [187.5]
129.4 [212.5]
144.6 [237.5]
159.9 [262.5]
Multiple w/
avg 83.7 &
stdev 10%
Multiple w/
avg 83.7 &
stdev 20%
Multiple w/
avg 83.7 &
stdev 30%

18

12

# of Lanes

Length of Grade
(km) [mi]

# of
Variables

Traffic Characteristics

% Grade

Physical Network

25

Ingle

Table 3. Refined list of variables considered.

% Trucks

Single and
Multiple Truck
Populations
(lkg/kW) [lb/hp]

Flow Rate
(pc/h/ln)

1
2
3
4
5
6

0.4 [0.25]
0.8 [0.50]
1.2 [0.75]
1.6 [1.00]
2.0 [1.25]
2.4 [1.50]

Asphalt-fair

820 (LOS A)
1350 (B)
1830 (C)
2170 (D)
2400 (E)

0
5
10
25
100

83.7 [137.5]
Multiple w/
avg 83.7 &
stdev 30%

# of Lanes

Pavement Type
and Condition

Length of Grade
(km) [mi]

# of
Variables

Traffic Characteristics

% Grade

Physical Network

Measurement Link
Subject Grade
4%
4%

2%
0.8 km
(0.5 mi)
2%
1.6 km (1.0 mi)

0.11 km
(0.07 mi)

Figure 1. Idealized link structure for INTEGRATION simulations (not to scale).

26

Ingle

Locating the flow and density measurement after the subject grade results in the
worst case measurement of the effect of trucks for the up grade. Considering the first and
second links as a hill, this measurement occurs at the crest of the hill. Among the 36
combinations of grade percent and length to test, the variable number of lanes as well as
pavement type and condition further expands the possible combinations of the physical
network. The initial physical network was comprised of only two lane freeway segments,
since the examination of three lane freeway segments was conducted later. The initial
physical network was also comprised of only asphalt-fair condition pavement. The
examination of other pavement types was also conducted later.
3.3.2 INTEGRATION Inputs
The traffic simulator, INTEGRATION, requires a number of input variables. The
free-flow speed used for this simulation was set to 120.7 km/h (75 mi/h) and the
saturation flow rate per lane was set to 2400 veh/h, which is the capacity of a freeway
lane according to the HCM 2000. The vehicle speed coefficient of variation was set to
0.08, a dimensionless quantity recommended based on previous research experience. The
speed at capacity was set to 85.8 km/h (53.3 mi/h), which is the minimum speed
corresponding to LOS E in the HCM 2000. The jam density was set to 139.8 veh/km (225
veh/mi), which is five times the density at capacity. Vehicle headways were simulated to
be 100% random. The characterization of subject truck populations as well as pavement
type and condition are inputs described in their respective sections below
3.3.3 Simulation Outputs
Traffic flow and density measurements were obtained as an output from the
simulation on 15 minute intervals since service flow rates for determining capacity are

27

Ingle

generally based on 15 minute observations. The capability of the traffic simulator allows
ten 15 minute periods to be simulated for the prescribed network size. The first interval
was discarded because it may take this amount of time for trucks to reach steady flow
across the final link. The remaining nine 15 minute intervals were averaged to obtain the
hourly average flow rate and density on the final link. To provide a suitable sample size
to overcome randomness in the results, the simulations were replicated with a different
random number seed. Up to three sets of nine samples, for a total of 27 samples, were
obtained from the traffic simulator.
3.3.4 Examination of the Effect of Weight to Power Ratio
A primary hypothesis of this research was that truck populations containing a
single weight to power ratio perform differently than multiple truck populations
containing a mix of weight to power ratios. This hypothesis was tested by comparing the
PCEs calculated for a single truck population with the PCEs calculated for a multiple
truck population with the same average weight to power ratio. In addition, it was
hypothesized that the PCE for single truck populations varies significantly due to the
weight to power ratio. Rakha and Lucic (23) found, in 2002, in a random sample of
trucks at the Troutville weigh station along I-81 in Virginia that the average weight to
power ratio was 79.2 kg/kW (130 lb/hp) with a standard deviation of 27.4 kg/kW (45
lb/hp). Figure 2 shows the results of the truck survey along I-81; it includes histograms of
the truck weight and truck power separately as well a histogram of the truck weight to
power ratio. The HCM 2000 similarly states that several studies have indicated that the
average weight to power ratio is between 76.1 and 90.4 kg/kW (125 and 150 lb/hp).

28

Ingle
35%
Frequency

30%
25%
20%
15%
10%
5%
0%

2268

6804

11340

15876

20412

24947

29483

34019

Friday Avg.

0%

5%

5%

16%

14%

20%

14%

24%

38555

2%

Survey

0%

1%

6%

21%

11%

18%

10%

31%

1%

Truck Weight (kg)

Frequency

40%
30%
20%
10%
0%

10000

20000

30000

40000

50000

60000

70000

80000

90000

Friday Avg.

0%

5%

5%

16%

14%

20%

14%

24%

2%

Survey

0%

1%

6%

21%

11%

18%

10%

31%

1%

35%

35%

30%

30%

25%

25%

Frequency

Frequency

Truck Weight (lb)

20%
15%
10%

20%
15%
10%
5%

5%
0%

0%

149 187 224 261 298 336 373 410 448 485

Survey 0%

1%

4% 11% 13% 29% 32% 8%

1%

0%

200 250 300 350 400 450 500 550 600 650

Survey 0%

1%

0%

25%

25%

20%

20%

Frequency

Frequency

1%

Truck Power (hp)

Truck Power (kW)

15%

15%
10%

10%
5%

5%
0%

4% 11% 13% 29% 32% 8%

15

30

46

61

76

91 106 122 137 152 167

Survey 0% 1% 12% 20% 12% 20% 20% 10% 4% 1% 0%

W eight-to-Pow e r Ra tio (kg/kW )

0%

25

50

75 100 125 150 175 200 225 250 275

Survey 0% 1% 12% 20% 12% 20% 20% 10% 4% 1% 0%

W eight-to-Pow er Ratio (lb/hp)

Figure 2. Truck survey results at I-81 Troutville weigh station in Virginia.

29

Ingle

To examine the effect of weight to power ratio, the simulation was initially run
using nine single truck populations. These were comprised of only a single truck type,
with a weight to power ratio that ranged from 38.1 to 159.9 kg/kW (62.5 to 262.5 lb/hp)
in increments of 5.2 kg/kW (25 lb/hp). This range represents the midpoint of the bins in
which the I-81 truck sample results were classified. This set of simulations was carried
out for 5 flow rates corresponding to the LOS flow rates. The truck proportion was tested
at 5 and 10% proportion of trucks. It was not necessary in the initial investigation to
create the density flow relationship over all truck proportions, nor was it necessary to
change the pavement type and condition from the initial asphalt-fair pavement.
Truck performance characteristics are determined internally to the traffic
simulator using vehicle dynamics models. The necessary inputs to define the truck
populations were obtained from an article by Rahka et al (24) in 2001. The important
vehicle dynamics values include: vehicle length of 16 m (52.5 ft), proportion of mass on
tractive axle of 0.36, transmission efficiency of 0.88, drag coefficient of 0.58, frontal area
of 10.7 m2 (115.2 ft2), first rolling resistance constant of 0.0328, and second rolling
resistance constant of 4.575. It is important to note that the passenger car population in all
simulations was the default passenger car used by the Environmental Protection Agency
for vehicle performance and emissions modeling.
Further simulations were conducted considering more realistic truck populations
that contain multiple truck weight to power ratios. This set of simulations was used to test
the hypothesis that multiple truck populations perform differently than single truck
populations. Multiple truck populations were created by varying the standard deviation of
the truck weight to power ratio around the average weight to power ratio. Therefore a

30

Ingle

normal distribution of trucks was created. The standard deviation was varied from 0 to
40% of the average weight to power ratio. Other non-normal multiple truck populations
were tested, but it was determined that the normal multiple truck population was just as
accurate without being overly complicated.
3.3.5 Examination of the Effect of Engine Power
The variability of engine power within a single weight to power ratio was also
considered. It was hypothesized that trucks with a lower engine power would have higher
PCEs even though they may have the same weight to power ratio. This hypothesis was
tested by varying the engine power from 242 to 392 kW (325 to 525 hp) in increments of
38 kW (50 hp). This range represents the midpoint of the bins in which a majority of the
trucks from the I-81 truck survey were classified. The truck weight was set as necessary
to obtain a weight to power ratio of 83.7 kg/kW (137.5 lb/hp). The truck weights ranged
from 20,270 to 32,743 kg (44,687 to 72,187 lb), which can be observed from figure 2 to
fall within the range of the most frequently observed truck weights from the survey
results.
3.3.6 Examination of the Effect of Proportion of Trucks
The HCM 2000 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 proportion of trucks, the proportion of trucks was considered in
increments of 10% all the way to 100% trucks. At the lower proportion of trucks (below
25%), the percentage categories used in the HCM 2000 were matched for clarity. An
objective of this research was to verify that the PCE decreases as the proportion of trucks

31

Ingle

increases. Although this trend is observed in the HCM 2000, it has not been tested at high
proportion of trucks.
3.3.7 Examination of the Effect of Pavement Type and Condition
It was hypothesized that poor pavement type and condition will result in increased
PCEs. The pavement type and condition for the entire network can be changed with the
modification the rolling coefficient and the coefficient of friction in the vehicle dynamics
input. The values of these parameters were obtained from an article on vehicle dynamics
by Rahka et al (24). The simulation model was run under seven different pavement types
and conditions: concrete pavement excellent condition, concrete pavement good
condition, concrete pavement poor condition, asphalt pavement good condition, asphalt
pavement fair condition, asphalt pavement poor condition, and snow covered. These
simulations were performed on a single truck population as well as a multiple truck
population. The default pavement type was asphalt pavement fair condition.
3.3.8 Examination of the Effect of Truck Aerodynamic Treatment
It was hypothesized that truck aerodynamic treatment will have a significant
effect on PCEs. The aerodynamic treatment of trucks affects the vehicle drag coefficient.
The default drag coefficient was 0.58. Rakha and Lucic (23) found, in 2002, in a random
sample of trucks along I-81 in Virginia that 55% of trucks had full aerodynamic
treatment, 15% had partial aerodynamic treatment, and 29% had no aerodynamic
treatement. For these simulations, the pavement type and condition was maintained as
fair asphalt, but the vehicle drag coefficient was changed to reflect the aerodynamic
treatment.

32

Ingle

3.3.9 Examination of the Effect of Three Lane Segments

It was hypothesized that PCEs for freeways with more than two directional lanes
will be lower than PCEs for freeways with only two directional lanes. To examine this
hypothesis, the number of lanes was increased from two to three. This change was
applied to a single truck as well as a multiple truck population. It was also of interest to
examine the effect of lane restrictions. Simulations were conducted with a lane restriction
in effect, which limited trucks to the two rightmost lanes. This lane restriction was
instituted in the form of a lane bias, which made the leftmost lane appear to trucks to
operate slower than the other lanes.
3.3.10 Examination of the Effect of Truck Speed Limit
It was hypothesized that the imposition of a speed limit on trucks will increase the
PCEs. The default PCE was calculated with no separate speed limit for trucks. However,
in many states the truck speed limit is regulated to as much as 24.1 km/h (15 mi/h) below
the speed limit for other vehicles. This lowers the speed of trucks approaching a grade
and thus increases the effect of the grade on trucks. For the default simulations, the speed
limit for all vehicles was 112.6 km/h (70 mi/h). The actual vehicle speeds for this case
were limited to 8 km/h (5 mi/h) above the speed limit. For the truck speed limit
simulations, a speed limit was imposed on trucks at 88.5 km/h (55 mi/h).
3.3.11 Examination of the Effect of Level of Congestion
It was hypothesized that increasing level of congestion will result in increasing
PCEs. It has already been established that simulations will be conducted at 5 different
traffic flow rates, corresponding to the LOS categories shown in Table 1. However,
simulation at these 5 flow rates does not in itself quantify the effect of level of

33

Ingle

congestion. The effect of level of congestion is considered in the equal density value that
is used to calculate the PCE. As mentioned in section 3.1, an equal density value of 12.4
pc/km/ln (20 pc/mi/ln), corresponding to density at LOS C, was used initially. To
examine the effect of level of congestion, the equal density values used to calculate PCE
were varied in increments of 3.1 pc/km/ln (5 pc/mi/ln) from 9 to 25 pc/km/ln (15 to 40
pc/mi/ln). An important note is that the effect of prevailing traffic volume can be
examined without further simulations.
3.4 Data Analysis
The simulation results were used to calculate the PCE following the methodology
described in section 3.1. In simulations to test the effect of weight to power ratio,
proportion of trucks, or level of congestion, the calculated PCE was compared with the
HCM 2000. In simulations to test the effect of pavement type and condition, aerodynamic
treatment, three lane segments, or truck speed limit, the calculated PCE was compared
with the default simulation PCEs. The default simulation was a single truck population
with a weight to power ratio of 83.7 kg/kW (137.5 lb/hp), operated on fair asphalt
pavement with full aerodynamic treatment and no separate truck speed limit. Percentage
differences that exceeded 10% were considered significant for investigation. If no
significant difference existed, the variable was considered to have no impact on the
development of PCEs.

34

Ingle

Chapter 4. Results for Single Truck Populations

4.1 Background
A simplified method for determining the PCE of trucks in a traffic stream is to
assume that all trucks have the same characteristics. A realistic truck population may
contain any number of trucks with different operating characteristics. Even if all trucks
on the highways had the same engine power, which they do not, there would still be
differences in the trucks. Loaded trucks have a higher weight to power ratio than empty
trucks. Trucks also have variability in the type of aerodynamic features used. Many
interstate, long-haul, trucks employ full aerodynamic features to permit the maximum
acceleration. Still others may employ partial or no aerodynamic treatments.
Although variability does exist in realistic truck populations, using only a single
truck population is a popular way to simplify calculations. This method has been applied
in many previous research studies. This was especially popular before the mainstream use
of computer simulators because the interaction between trucks was not as clearly known.
Simplifying assumptions had to be made to make use of field data, for which the exact
truck population was unknown. Even today, the HCM 2000 provides PCEs that were
calibrated from a mix of trucks that ranged within the average truck weight to power
ratio. This population may be a simplification of a realistic truck population, but it has
been determined accurate enough for meaningful results. Simplifications and
generalizations are common and useful practices in traffic engineering to make work
more efficient.
In this research, single truck populations were defined to have different weight to
power ratios. This not only represents the variability in engine power, but also the

35

Ingle

difference between loaded and unloaded trucks. Nine different weight to power ratios
were used, ranging from 38.1 to 159.9 kg/kW (62.5 to 262.5 lb/hp) in increments of 5.2
kg/kW (25 lb/hp). These truck populations represented the midpoint of the survey bins
from the random truck survey along I-81 in Virginia (23). One objective of the single
truck population tests was to determine how significant the impact of truck weight to
power ratio is. The single truck population that most closely matched the PCEs provided
in the HCM 2000 was chosen for further investigative research. The variability of the
single truck population by length and percent grade was used to validate the chosen
population. Further variables for investigation included engine power, proportion of
trucks, pavement type and condition, aerodynamic treatment, three lane segments, truck
speed limit, and level of congestion.
The variables selected for the single truck population investigations were used to
verify the hypotheses of this research. It was first hypothesized that truck weight to
power ratio has a significant effect on the PCE. In accordance with this hypothesis, it was
expected that the PCE for single truck populations with a low weight to power ratio will
be much less than the PCE for single truck populations with a high weight to power ratio.
Section 4.3 presents the results of the variability of the PCE by truck weight to power
ratio. It was also hypothesized that variability in engine power within the same weight to
power ratio would significantly affect the PCE. Trucks with low engine power were
expected to have higher PCEs than trucks with high engine power but the same weight to
power ratio. Section 4.6 presents the results of the variability of the PCE by engine
power. It was hypothesized that the PCE for a single truck population decreases with
increasing proportion of trucks. This relationship exists in the current HCM 2000, but it

36

Ingle

was examined for a higher proportion of trucks in this research. Section 4.7 presents the
results of the variability of PCE by proportion of trucks. Another hypothesis of this
research was that pavement type and condition significantly influence the PCE. It was
expected that poor pavement would result in higher PCEs in comparison to the same
single truck population on better pavement. The results of the variability of PCE by
pavement type and condition are presented in section 4.8.
It was hypothesized that truck aerodynamic treatment would significantly affect
the PCE. Single truck populations with only partial or no aerodynamic treatment were
expected to have higher PCEs than single truck populations with full aerodynamic
treatment. Section 4.9 presents the results of the variability of PCE by truck aerodynamic
treatment. Another hypothesis of this research was that PCEs for three lane freeway
segments would be much lower than PCEs for two lane freeway segments. The results of
the variability of PCE for three lane segments are presented in section 4.10. It was
hypothesized that the institution of a truck speed limit below the speed limit of other
vehicles would result in significantly higher PCEs. The results of the variability of PCE
for a truck speed limit are presented in section 4.11. A final hypothesis of this research
was that the level of congestion will significantly affect the calculated PCE. Under this
hypothesis, it was expected that increasing levels of congestion will result in increasing
calculated PCEs for the same single truck population. Section 4.12 presents the
variability of PCE by level of congestion. Section 4.13 presents an examination of the
combined effects of the variables on the calculated PCE. Whereas the preceding sections
of this chapter examine the effect of the variables independently, the combined effects of
the variables is a more complicated issue.

37

Ingle

4.2 Speed-Flow-Density Relationships

The first task in the traffic simulations was to obtain speed, flow, and density
data. These data were used to construct the speed-flow-density relationships. The data
were only collected for the uncongested flow regime. Figure 3 shows the speed-flowdensity relationships for the base vehicle flow on a 0.8 km (0.5 mi) grade. Separate
curves were plotted for the different percent grades, 2, 4, and 6%. It can be observed that
the speed-flow-density relationships for passenger cars are the same regardless of the
different percent grades. The observed relationships between speed, flow, and density are
as would be expected for freeway segments in the uncongested flow regime.
The speed-flow-density relationships were also constructed for the mixed traffic,
passenger cars and a single truck population. Figure 4 shows the speed-flow-density
relationships for a single truck population of 83.7 kg/kW (137.5 lb/hp) on a 0.8 km (0.5
mi) grade. Separate curves were plotted for different proportion of trucks, 5 and 10% as
well as the different percent grades. The stated proportion of trucks was chosen because
5% is low enough to have high PCE values for comparisons and 10% is close, although
below, the actual proportion of trucks observed on many freeways. A separate speed,
flow, or density relationship could be plotted for each combination of the roadway and
traffic characteristics. It was decided to focus on these roadway and traffic characteristics
for the 0.8 and 1.6 km (0.5 and 1.0 mi) grades for most of this research.
The speed-flow-density relationships are a primary tool to proof check the
simulation results. These relationships are also useful in the calculation of the PCE,
although not graphically. The mixed vehicle flow-density relationship in figure 4
illustrates that for an equal density value, flow decreases with increasing proportion of

38

Ingle
Speed-Flow (0.8 km)

140

140

120

120

100

100

Speed (km/h)

Speed (km/h)

Speed-Density (0.8 km)

80
60

80
60

40

40

20

20
0

0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

0.0

35.0

500.0

1000.0

1500.0

2000.0

2500.0

Flow (ve h/h/ln)

De nsity (pc/km/ln)

Flow-Density (0.8 km)

2500.0

Flow (veh/h/ln)

2000.0

1500.0

1000.0

500.0

0.0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

De nsity (ve h/km/ln)

Figure 3. Speed-flow-density relationships, base vehicle flow.

39

S peed (kph)

Legend
200
100
0
0.0

Qb 2% grade
5.0

10.0

Qb 4% grade
15.0

20.0

Density (vpkm pln)

25.0

Qb 6% grade
30.0

35.0

Ingle
Speed-Flow (0.8 km)

140

140

120

120

100

100
Speed (km/h)

Speed (km/h)

Speed-Density (0.8 km)

80
60

80
60

40

40

20

20
0

0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

0.0

35.0

500.0

1000.0

1500.0

2000.0

2500.0

Flow (ve h/h/ln)

De nsity (veh/km/ln)

Flow-Density (0.8 km)

2500.0

Legend
1500.0

Qm 5% trucks 4% grade

Qm 5% trucks 2% grade

150
Speed (kph)

Flow (veh/h/ln)

2000.0

1000.0

500.0

100
50

Qm 10% trucks 2% grade

Qm
5% trucks 6% grade
0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

De ns ity (vpk m pln)

0.0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

Qm 10% trucks 4% grade

Qm 10% trucks 6% grade

De nsity (veh/km/ln)

Figure 4. Speed-flow-density relationships, mixed vehicle flow single truck population of 83.7 kg/kW (137.5 lb/hp).

40

Ingle

trucks or percentage grade. This trend is also observed in the speed-density and speedflow relationships at equal values of density and flow respectively. The general trends
observed in the mixed vehicle speed-flow-density relationships are as would be expected
for freeway segments in the uncongested flow regime.
4.3 Variability of PCE by Weight to Power Ratio
The weight to power ratio of the single truck populations was found to have a
significant effect on the calculated PCE; percentage differences between different
populations exceeded 10%. The PCE of trucks was computed for each truck population at
an equal density value of 12.4 pc/km/ln (20 pc/ml/ln), corresponding to LOS C. Figure 5
shows the predicted PCE for specific grade length and percent combinations plotted
against the corresponding weight to power ratio. There is much less variability in the
predicted PCE for short and mildly sloped grades as compared to long and steeply sloped
grades. The PCE for a 0.8 km (0.5 mi), 2% grade ranges from 1.5 to 3.0; however, the
PCE for a 1.6 km (1.0 mi), 6% grade ranges from 2.0 to 8.0. For the same slope grade,
the PCE and variability of PCE is higher for longer grades. This means that the effect of
the weight to power ratio in truck populations is more pronounced for longer grades.
The 83.7 kg/kW (137.5 lb/hp) truck population most closely matches the PCE
provided in the HCM 2000 across all of the grade length and percent combinations. This
weight to power truck population usually intersects or is very near to the horizontal line
in figure 5 which represents PCEs in the HCM 2000. The 83.7 kg/kW (137.5 lb/hp) truck
population confirms the assertion made in the HCM 2000 that the PCEs provided therein
were calculated for a truck population with an average weight to power ratio between
76.1 and 90.4 kg/kW (125 and 150 lb/hp). This observation also validates the

41

Ingle
1.6 km 2% grade

8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

P CE

P CE

0.8 km 2% grade

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0

0.0
35.0

55.0

75.0

95.0

115.0

135.0

0.0
35.0

155.0

55.0

75.0

5% trucks

5% trucks

HCM 2000

8.0

7.0

7.0

6.0

6.0

5.0

5.0
PCE

P CE

8.0

4.0

3.0

2.0

2.0

1.0

1.0
75.0

95.0

115.0

135.0

0.0
35.0

155.0

55.0

75.0

7.0

6.0

6.0

5.0

5.0
P CE

P CE

8.0

7.0

4.0

3.0

2.0

2.0

1.0

1.0
95.0

115.0

135.0

135.0

155.0

HCM 2000

155.0

0.0
35.0

55.0

75.0

95.0

115.0

W eight to Pow er (kg/kW )

W eight to Pow er (kg/kW )

5% trucks

155.0

4.0

3.0

75.0

115.0

1.6 km 6% grade

8.0

55.0

135.0

HCM 2000

95.0

5% trucks

HCM 2000

0.8 km 6% grade

0.0
35.0

155.0

W eight to Pow er (kg/kW )

W eight to Pow er (kg/kW )

5% trucks

135.0

4.0

3.0

55.0

115.0

1.6 km 4% grade

0.8 km 4% grade

0.0
35.0

95.0

W eight to Pow er (kg/kW )

W eight to Pow er (kg/kW )

HCM 2000

5% trucks

HCM 2000

Figure 5. PCE variability by weight to power ratio for selected grades and LOS C congestion.

42

Ingle

PCE calculation procedure used in this research because it accurately replicates the PCEs
provided in the HCM 2000. Further investigations in this research focused on truck
populations with an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp). Table 4
shows the calculated PCEs for the average single truck population.
The PCE calculated for different weight to power ratio single truck populations
provides sufficient evidence to suggest that weight to power ratio significantly affects the
PCE. The PCEs calculated for two single truck populations were compared with the
PCEs calculated for the average weight to power ratio single truck population. The first
single truck population had a weight to power ratio of 68.5 kg/kW (112.5 lb/hp) and
represents a lighter than average weight to power ratio truck population. This weight to
power ratio is slightly below the range of 76.1 to 90.4 kg/kW (125 to 150 lb/hp) that the
HCM 2000 is said to represent. The second single truck population had a weight to power
ratio of 106.6 kg/kW (175 lb/hp) and represents a heavier than average weight to power
ratio truck population. These two truck populations were selected because they represent
potential realistic truck populations that are lighter and heavier than the average truck
population.
The ratio of the examined truck population PCE to the average truck population
PCE was calculated for comparison. Table 5 shows the calculated ratio of the PCEs for
these two truck populations. It can be observed that for the lighter than average single
truck population, the PCE is on average a fraction of 0.8 of the average single truck
population PCE for 5% trucks. However, as the percentage of trucks increases, the ratio
becomes closer to 1.0, which no longer represents a significant difference. For the heavier
than average single truck population, the PCE is on average a fraction

43

Ingle

Table 4. PCEs for trucks and buses on upgrades, average single truck population 83.7 kg/kW (137.5 lb/hp) and LOS C congestion.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22

0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00

2
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.0
2.5
3.0
3.0
3.0
3.5
4.0
2.5
3.0
3.5
4.0
4.5
4.5
5.0
3.5
3.5
5.5
5.5
6.5
6.5
7.5
3.5
4.5
7.5
7.0
8.5
8.5
9.0
5.0
8.0
8.0
9.0
9.5
10.0
10.5

4
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.0
2.5
2.5
2.5
2.5
3.0
3.0
2.0
2.5
2.5
3.5
3.5
4.0
4.0
2.5
3.5
4.0
4.5
5.0
5.0
5.0
3.0
4.5
5.5
6.0
6.5
7.0
6.5
5.0
7.5
7.5
8.0
8.0
8.0
9.5

44

5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
3.0
3.5
2.0
2.5
3.5
3.5
4.0
4.0
4.5
2.5
3.0
3.5
4.0
4.0
5.0
5.5
3.0
4.5
5.5
6.0
6.0
6.5
6.5
4.5
7.0
7.5
8.0
8.0
8.0
8.5

6
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
3.0
2.0
2.5
3.0
3.5
3.5
4.0
4.0
2.5
3.0
3.5
4.0
4.5
4.5
5.0
3.0
4.0
5.0
6.0
6.0
5.5
6.0
4.0
6.5
7.0
7.0
7.0
7.5
7.5

8
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
3.0
2.0
2.5
3.0
3.0
3.5
3.5
3.5
2.0
3.0
3.0
3.0
3.5
3.5
4.5
2.5
4.0
4.5
5.0
5.0
5.5
5.5
4.0
5.5
6.0
6.0
6.0
6.5
6.5

ET
Percentage of Trucks and Buses
10
15
20
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.0
2.5
2.5
2.0
2.0
2.0
2.0
2.5
2.5
2.0
3.0
2.5
2.5
3.0
3.0
2.5
3.0
3.0
2.5
3.0
3.0
2.5
3.5
3.0
2.5
2.5
2.0
2.0
2.5
2.5
2.0
3.0
3.0
2.5
3.0
3.0
3.0
3.5
3.0
3.0
3.5
3.5
3.0
4.0
3.5
3.0
2.5
2.5
2.5
3.5
3.0
3.0
4.0
3.5
3.5
4.5
3.5
3.5
4.5
4.0
3.5
5.0
4.0
3.5
5.0
4.5
3.5
3.5
3.5
3.0
5.0
4.5
4.0
5.5
4.5
4.0
5.5
5.0
4.5
6.0
5.0
4.5
6.0
5.0
4.5
6.0
5.0
4.5

25
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
2.0
2.0
2.5
3.0
3.0
3.0
3.0
2.0
2.5
3.0
3.0
3.5
3.5
3.5
3.0
3.5
4.0
4.0
4.0
4.0
4.5

30
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.0
2.5
2.5
2.5
2.5
3.0
3.0
2.0
2.5
3.0
3.0
3.0
3.0
3.5
2.5
3.5
3.5
4.0
4.0
4.0
4.0

40
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
2.5
2.0
2.5
2.5
3.0
3.0
3.0
3.0
2.5
3.0
3.5
3.5
3.5
3.5
3.5

50
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.0
2.5
2.5
2.5
2.5
2.5
2.5
2.5
3.0
3.0
3.0
3.0
3.0
3.0

60
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
2.5
2.0
2.5
3.0
3.0
3.0
3.0
3.0

Ingle

Table 5. Ratio of PCE for light and heavy single truck populations to PCE for average single truck population.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

Ratio to Average Single Truck PCE, 83.7 kg/kW (137.5 lb/hp)

Light Truck Pop. 68.5 kg/kW (112.5 lb/hp)
Heavy Truck Pop. 106.6 kg/kW (175 lb/hp)
Percentage of Trucks and Buses
5
10
25
100
5
10
25
100
1.0
1.0
1.0
0.7
1.3
1.0
1.0
1.0
0.8
1.0
1.0
0.7
1.3
1.3
1.3
1.0
0.8
1.0
1.0
0.7
1.3
1.3
1.3
1.0
0.8
1.0
1.0
0.7
1.3
1.7
1.3
1.0
0.8
1.0
1.0
0.7
1.3
1.7
1.3
1.0
0.5
1.0
1.0
0.7
1.3
1.7
1.3
1.3
1.0
1.0
0.8
1.0
1.3
1.3
1.0
1.0
0.8
0.8
1.0
1.0
1.4
1.2
1.3
1.0
0.8
0.7
1.0
1.0
1.5
1.2
1.5
1.3
0.7
0.8
0.8
1.0
1.6
1.3
1.4
1.3
0.8
0.8
0.8
1.0
1.3
1.3
1.4
1.3
0.8
0.8
0.8
1.0
1.4
1.3
1.4
1.7
0.8
1.0
1.0
1.0
1.2
1.2
1.5
1.3
0.8
0.9
1.0
0.8
1.3
1.3
1.6
1.0
0.8
0.9
0.8
0.8
1.3
1.4
1.5
1.0
0.8
0.8
0.8
0.8
1.3
1.2
1.5
1.0
0.8
0.8
0.7
0.8
1.3
1.2
1.3
1.0
0.8
0.7
0.9
0.8
1.2
1.2
1.4
1.0
0.8
0.9
0.9
0.8
1.3
1.3
1.4
1.1

45

Ingle

of 1.3 of the average single truck population PCE for 5 or 10% trucks. The largest ratio of
the PCE occurs for 25% trucks, at an average of 1.4. These ratios could be used as a scale
factor to multiply the average single truck population PCEs by. This could account for
the effect of lower or higher weight to power ratio single truck populations.
4.4 Variability of PCE by Length of Grade
The PCE was observed to increase as the length of grade increased. This trend
exists in the PCEs provided in the HCM 2000, and it was observed for the truck
population with an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp). The
variability of PCE by length of grade was used to validate the PCE calculation procedure
used in this research because length of grade is a major determining factor in the PCEs
provided in the HCM 2000. The PCEs provided in the HCM 2000 are given for a range
of grades between 0.40 and 2.41 km (0.25 and 1.5 mi). The range of grades considered
for this research was extended to 0.40 to 9.65 km (0.25 to 6 mi) for illustrative purposes.
Figure 6 shows the variability of the PCE over this range of grade lengths for 5 and 10%
trucks. It is encouraging that the trends observed from the results of this research are
similar to the trends observed in the HCM 2000.
A primary observation from figure 6 is that the PCE changes more drastically by
length of grade for steeper grades. This trend is sensible, and closely ties into the
variability of the PCE by percent grade. In addition, it can be observed that the PCE and
variability in PCE is higher for the 5% proportion of trucks as compared to 10%
proportion of trucks. This diminishing effect of the proportion of trucks was a major
reason for choosing to examine such a low proportion of trucks in this research. It is clear
that at low percent grades, the PCE changes very little from one grade length to the next.

46

Ingle
2% grade 10% trucks

8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0
P CE

P CE

2% grade 5% trucks

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0

0.0
0.40

1.40

2.40

3.40

4.40

5.40

6.40

7.40

8.40

0.0
0.40

9.40

1.40

2.40

3.40

Length of Grade (km)

HCM 2000

5% Trucks

7.0

7.0

6.0

6.0

5.0

5.0

PCE

PCE

8.0

4.0

3.0

2.0

2.0

1.0

1.0

3.40

4.40

5.40

6.40

7.40

8.40

0.0
0.40

9.40

1.40

2.40

3.40

HCM 2000

7.0

6.0

6.0

5.0

5.0

PCE

PCE

8.0

7.0

4.0

3.0

2.0

2.0

1.0

1.0

3.40

4.40

5.40

6.40

7.40

8.40

Length of Grade (km)

5% Trucks

7.40

8.40

9.40

7.40

8.40

9.40

HCM 2000

5.40

6.40

HCM 2000

4.0

3.0

2.40

9.40

6% grade 10% trucks

8.0

1.40

4.40

10% Trucks

6% grade 5% trucks

0.0
0.40

8.40

Le ngth of Gra de (km)

Le ngth of Gra de (km)

5% Trucks

7.40

4.0

3.0

2.40

6.40

4% grade 10% trucks

4% grade 5% trucks

1.40

5.40

10% Trucks

8.0

0.0
0.40

4.40

Length of Grade (km)

9.40

0.0
0.40

1.40

2.40

3.40

4.40

5.40

6.40

Length of Gra de (km)

HCM 2000

10% Trucks

HCM 2000

Figure 6. PCE variability by length of grade for average single truck population, 83.7 kg/kW
(137.5 lb/hp) and LOS C congestion.

47

Ingle

This is illustrated by the almost constant value of PCE over all lengths of grade for 2%
grades. However, for steeper grades the change in PCE between successive lengths of
grades up to 3.22 km (2 mi) increased by an average of 0.6. For grades longer than 3.22
km (2 mi) up to 9.65 km (6 mi), the PCE did not increase by very much if at all.
4.5 Variability of PCE by Percent Grade
The PCE was observed to increase as the percent grade increased. This trend was
also used to validate the selected truck population because it is a determining factor used
in the PCEs provided in the HCM 2000. The PCEs provided in the HCM 2000 only range
between 1 and 6%; however, grades up to 8% were tested in this research for
investigative purposes. Figure 7 shows the variability of the PCE over percent grades
ranging from 1 to 8%. It is clear that the single truck population with an average weight
to power ratio of 83.7 kg/kW (137.5 lb/hp) closely matches the trend in PCEs from the
HCM 2000. The curves in Figure 7 only apply to the 5% proportion of trucks because
this proportion trucks shows more variability in PCE than 10% proportion of trucks. It
can be observed, regardless of the length of grade, that the PCE for 1% grades is always
around 1.5. For longer grades, the PCE varies more dramatically by percent grade,
creating a larger range of PCE values for the longest grades. The average difference in
PCE from one percent grade to the next is 0.6. A uniform incremental increase does not
truly exist however. In fact, the change in PCE from 1% to 2% grade is always less
incrementally than the change in PCE from 5% to 6% grade. This illustrates the
detrimental effect of steep grades on heavy vehicles. Since the PCEs calculated for 8%
grades are larger than the PCEs for 6% grades, it must not be assumed that PCEs for

48

Ingle
0.8 km grade

8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

P CE

P CE

0.4 km grade

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

Percentage Grade
5% trucks

HCM 2000

5% trucks

8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

P CE

P CE

HCM 2000

1.6 km grade

1.2 km grade

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
1

5% trucks

Percentage Grade

Percentage Grade

5% trucks

HCM 2000

HCM 2000

2.4 km grade

2.0 km grade
8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

P CE

PCE

Percentage Grade

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
1

Percentage Grade

Percentage Grade

5% trucks

5% trucks

HCM 2000

HCM 2000

Figure 7. PCE variability by percentage grade for average single truck population, 83.7 kg/kW
(137.5 lb/hp) and LOS C congestion.

49

Ingle

grades beyond the range of the HCM 2000 are constant, but rather that they continue to
increase with percent grade.
4.6 Variability of PCE by Engine Power
Engine power was only found to significantly affect the PCE for a low engine
power single truck population; the percentage difference in the PCE was as much as 13%.
This partially confirms a hypothesis of this research. However, it was not only expected
that PCEs will be higher for low engine powers, but also that PCEs will be lower for high
engine powers. Five different engine powers were tested in this research, representing a
range of engine powers observed in the truck survey along I-81 in Virginia. The truck
weight corresponding to each engine power was set so that the weight to power ratio of
the trucks remained 83.7 kg/kW (137.5 lb/hp). Table 6 shows the variability of the PCE
for the different truck powers as a percentage difference compared to the PCE for the
default truck power. The default truck power for this research was 354 kW (475 hp)
representing the 75th percentile truck power in the survey.
The lowest truck power simulated was 242 kW (325 hp) representing the 10th
percentile truck power in the survey. From table 6 it can be observed that the PCEs for
10% proportion of trucks for this truck power were significantly higher (13%) than the
PCEs for the average single truck population with the default engine power. However,
the PCEs at 5, 25, and 100% proportion of trucks were not significantly higher (6%) than
the PCEs for the average single truck population. The other truck powers simulated
represented the 22nd, 45th, 75th, and 95th percentile truck powers from the truck survey.
For the 22nd percentile truck power, the PCEs were again significantly different only for
10% proportion of trucks. For the 45th and 95th percentile truck powers, none of the PCEs

50

Ingle

Table 6. Percentage difference in calculated PCE for different engine powers compared to 75th percentile engine power 354 kW (475 hp).
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

10th Percentile Power, 242 kW (325 hp)

5
0%
0%
0%
0%
0%
0%
25%
20%
33%
14%
0%
13%
0%
0%
0%
0%
0%
0%
6%

10
0%
0%
33%
33%
33%
33%
0%
20%
0%
0%
17%
0%
20%
14%
13%
0%
11%
0%
13%

25
0%
0%
0%
0%
0%
0%
0%
0%
25%
0%
0%
0%
25%
20%
0%
0%
0%
0%
4%

51

100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

% Error From 75th Percentile Power, 354 kW (475 hp)

22nd Percentile Power, 280 kW (375 hp)
45th Percentile Power, 317 kW (425 hp)
Percentage of Trucks and Buses
5
10
25
100
5
10
25
100
0%
0%
0%
0%
0%
0%
0%
0%
0%
33%
0%
0%
0%
0%
0%
0%
0%
33%
0%
0%
0%
0%
0%
0%
0%
33%
0%
0%
0%
0%
0%
0%
0%
33%
0%
0%
0%
0%
0%
0%
0%
33%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
17%
0%
0%
0%
17%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
17%
0%
0%
0%
17%
0%
0%
0%
20%
0%
0%
0%
20%
0%
0%
11%
0%
20%
0%
0%
14%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
8%
0%
0%
0%
0%
0%
0%
0%
-8%
0%
0%
0%
-8%
-10%
0%
0%
2%
11%
1%
0%
0%
2%
0%
0%

95th Percentile Power, 392 kW

5
0%
0%
0%
0%
0%
0%
0%
0%
17%
0%
0%
0%
0%
0%
0%
-8%
0%
-8%
0%

10
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

25
0%
0%
0%
0%
0%
0%
0%
0%
0%
-20%
0%
0%
0%
20%
0%
0%
-14%
0%
-1%

Ingle

were significantly different from the PCEs calculated for the average single truck
population.
4.7 Variability of PCE by Proportion of Trucks
The proportion of trucks in the traffic stream was found to have a significant effect on the
calculated PCE only at low proportion of trucks. This trend is especially truefor longer
and steeper grades. Figure 8 shows the variability of the PCE over proportion of trucks
ranging from 2% to 100%. The PCE table in the HCM 2000 only provides PCEs for up to
25% proportion of trucks. It can be observed that the PCE is highest for low proportion
trucks, and that the PCE decreases and levels off as the proportion trucks increases. The
PCE past 60% trucks shows very little variability.
The variability in PCE for proportion of trucks between 25 and 60% provides
significant evidence that the PCE table in the HCM 2000 should be expanded to include
up to 60% trucks. If one were to assume that on a 1.6 km (1 mi), 6% grade, the PCE for
60% trucks was equal to that for 25% trucks, the PCE would be 3.0 as calculated in this
research. However, the calculated PCE for 60% trucks on such a grade is actually 2.5.
The relative difference of 0.5 in the PCE seems small, but is actually 20% larger than it
should be. If the freeway traffic volume is assumed to be 3000 veh/h over two lanes, the
proportion of the traffic that is trucks would be 1800 veh/h. Using the PCE of 3.0 would
result in an estimate of the equivalent throughput of passenger cars as 6600 pc/h.
However, using the PCE of 2.5 would result in an estimate of the equivalent throughput
of passenger cars that is 16% lower, as 5700 pc/h.
The observed decreasing trend in PCE with increasing proportion of trucks
confirms a hypothesis of this research. An explanation for this observed trend is that as
52

Ingle
1.6 km 2% grade

8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

P CE

P CE

0.8 km 2% grade

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
0

10

20

30

40

50

60

70

80

90

100

10

20

30

2% grade

HCM 2000

70

80

90

100

70

80

90

100

70

80

90

100

70

80

90

100

HCM 2000

1.6 km 4% grade

8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

P CE

P CE

60

2% grade

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
0

10

20

30

40

50

60

70

80

90

100

10

20

30

40

50

60

Percentage Trucks

Percentage Trucks

4% grade

4% grade

HCM 2000

HCM 2000

1.6 km 6% grade

0.8 km 6% grade
8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

P CE

P CE

50

Percentage Trucks

0.8 km 4% grade

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

10

20

30

40

50

60

70

80

90

100

10

20

30

Percentage Trucks

40

50

60

Percentage Trucks

HCM 2000

6% grade

6% grade

0.8 km 8% grade

HCM 2000

1.6 km 8% grade

8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

P CE

P CE

40

Percentage Trucks

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

10

20

30

40

50

60

70

80

90

100

10

20

30

40

50

60

Percentage Trucks

Percentage Trucks

8% grade

8% grade

HCM 2000

HCM 2000

Figure 8. PCE variability by proportion of trucks for average single truck population, 83.7 kg/kW
(137.5 lb/hp) and LOS C congestion.
53

Ingle

the proportion of trucks increases, the interaction of trucks with cars decreases because
trucks will form platoons climbing the grade. The interaction of trucks with trucks is
negligible since the trucks perform the same, having the same weight to power ratio.
Providing PCEs that apply for proportion of trucks beyond 25% is desirable since the
proportion of trucks on many freeway segments currently exceeds 25% at times.
4.8 Variability of PCE by Pavement Type and Condition
Pavement type and condition was found to have a significant effect on the
calculated PCE, with percentage differences exceeding 10%. This confirms a hypothesis
of this research. The pavement type and condition affects the rolling coefficient and
pavement friction coefficient. These vehicle dynamics coefficients affect the maximum
acceleration of all vehicles and especially trucks. The default pavement type for
simulations in this research was asphalt in fair condition. This is a reasonable to slightly
conservative estimate for the pavement condition for many interstate freeway segments.
The PCE of trucks was calculated for six additional pavement types and
conditions. The calculated PCE for these different pavement types and conditions for the
truck population with an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp) is
shown in table 7. Table 8 shows the variability of the PCE for these different pavement
types and conditions calculated as the percent difference in PCE from the default
pavement type. Poor asphalt and poor concrete pavement always had higher or equal
PCEs as compared with fair asphalt. The PCE for both poor pavement types was on
average 9% higher than for the fair asphalt. However, the effect of the poor pavement
types was even more severe for a lower proportion of trucks such as 5 or 10% trucks. The
PCE for poor asphalt pavement at 10% trucks averaged 17% higher than the PCE for fair

54

Ingle

Table 7. PCEs for trucks and buses on upgrades with different pavement types, average single truck population 83.7 kg/kW (137.5 lb/hp) and
LOS C congestion.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50

Asphalt Poor
5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
3.0
4.0
4.0
4.5
4.5
3.0
5.0
6.0
6.5
6.5
7.0

10
1.5
2.0
2.0
2.0
2.0
2.0
2.0
3.0
3.0
3.5
3.5
3.5
3.0
4.0
4.5
5.0
5.0
5.0

25
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.5
2.5
2.5
3.0
2.5
3.0
3.5
3.5
3.5
3.5

Concrete Poor
100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0

5
1.5
2.0
2.0
2.0
2.0
2.0
2.5
3.0
3.5
4.0
4.5
4.5
3.0
4.5
5.5
6.0
6.0
6.5

10
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
3.0
3.0
3.5
3.5
3.0
4.0
4.5
4.5
4.5
5.0

25
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.5
2.5
2.5
2.5
2.5
3.0
3.0
3.0
3.5
3.5

100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0

5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.5
3.0
3.5
3.5
3.5
3.0
4.0
5.0
5.5
6.0
6.0

ET
Asphalt Good
Concrete Good
Percentage of Trucks and Buses
10
25
100
5
10
25
1.5
1.5
1.0
1.5
1.5
1.5
1.5
1.5
1.0
1.5
1.5
1.5
1.5
1.5
1.0
1.5
1.5
1.5
1.5
1.5
1.0
1.5
1.5
1.5
1.5
1.5
1.0
1.5
1.5
1.5
1.5
1.5
1.0
1.5
1.5
1.5
2.0
2.0
1.5
2.0
2.0
2.0
2.5
2.0
1.5
2.5
2.5
2.0
2.5
2.0
1.5
3.5
2.5
2.0
2.5
2.0
1.5
3.5
3.0
2.5
3.0
2.5
1.5
4.0
3.0
2.5
3.0
2.5
1.5
4.0
3.0
2.5
2.5
2.0
1.5
3.0
2.5
2.0
3.5
2.5
1.5
4.0
3.5
3.0
4.0
3.0
2.0
5.0
4.0
3.0
4.0
3.0
2.0
5.5
4.0
3.0
4.5
3.0
2.0
6.0
4.5
3.0
4.5
3.0
2.0
6.0
4.5
3.5

Concrete Excellent
100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0

5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.5
3.0
3.0
3.0
3.5
2.5
4.0
4.5
5.0
5.5
5.5

10
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.5
2.5
2.5
3.0
2.5
3.5
4.0
4.0
4.0
4.5

25
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.5
2.0
2.5
3.0
3.0
3.0
3.0

Snow Covered
100
1.0
1.0
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0

5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
5.5
6.0

10
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5
3.5

25
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0

Table 8. Percentage difference in calculated PCE for different pavement types compared to asphalt fair pavement.
Upgrade
(%)

Length
(km)
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

Length
(mi)
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

Asphalt Poor
5
33%
0%
0%
0%
0%
0%
25%
20%
33%
14%
13%
13%
0%
11%
9%
8%
8%
8%
11%

10
0%
33%
33%
33%
33%
33%
0%
20%
0%
17%
17%
17%
20%
14%
13%
11%
11%
0%
17%

25
0%
0%
0%
0%
0%
0%
0%
0%
25%
0%
0%
20%
25%
20%
17%
17%
0%
0%
7%

Concrete Poor
100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

5
0%
0%
0%
0%
0%
0%
25%
20%
17%
14%
13%
13%
0%
0%
0%
0%
0%
0%
6%

55

10
0%
33%
33%
33%
33%
33%
0%
0%
0%
0%
17%
17%
20%
14%
13%
0%
0%
0%
14%

25
0%
0%
0%
0%
0%
0%
0%
0%
25%
0%
0%
0%
25%
20%
0%
0%
0%
0%
4%

100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

5
0%
-25%
-25%
-25%
-25%
-25%
0%
0%
0%
0%
-13%
-13%
0%
-11%
-9%
-8%
0%
-8%
-10%

% Error From Asphalt Fair

Asphalt Good
Concrete Good
Percentage of Trucks and Buses
10
25
100
5
10
25
0%
0%
-33%
0%
0%
0%
0%
0%
-33%
-25%
0%
0%
0%
0%
-33%
-25%
0%
0%
0%
0%
-33%
-25%
0%
0%
0%
0%
-33%
-25%
0%
0%
0%
0%
-33%
-25%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-17%
0%
0%
17%
-17%
0%
-17%
-20%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-25%
-11%
0%
20%
0%
0%
0%
-9%
0%
0%
-11%
0%
0%
-8%
-11%
0%
0%
-14%
0%
0%
0%
-14%
-10%
-14%
0%
-8%
-10%
0%
-3%
-3%
-13%
-8%
-2%
0%

Concrete Excellent
100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

5
0%
-25%
-25%
-25%
-25%
-25%
0%
0%
0%
-14%
-25%
-13%
-17%
-11%
-18%
-17%
-8%
-15%
-15%

10
0%
0%
0%
0%
0%
0%
0%
-20%
-17%
-17%
-17%
0%
0%
0%
0%
-11%
-11%
-10%
-6%

25
0%
0%
0%
0%
0%
0%
-25%
0%
0%
-20%
-20%
0%
0%
0%
0%
0%
-14%
-14%
-5%

Snow Covered
100
-33%
-33%
-33%
-33%
-33%
-33%
0%
0%
0%
0%
0%
0%
0%
-25%
0%
0%
0%
0%
-13%

5
267%
175%
175%
175%
175%
175%
175%
120%
83%
57%
38%
38%
83%
22%
0%
-8%
-8%
-8%
96%

10
133%
133%
133%
133%
133%
133%
75%
40%
17%
17%
17%
17%
40%
0%
-13%
-22%
-22%
-30%
52%

25
33%
33%
33%
33%
33%
33%
0%
0%
0%
-20%
-20%
-20%
0%
-20%
-33%
-33%
-43%
-43%
-2%

Ingle

asphalt pavement across the selected grades. As the proportion of trucks increased
however, the effect of poor pavement was less.
The effect of good pavement was only significant at 5% trucks. Overall, the
calculated PCE for good concrete pavement was 2% lower than the PCE for the default
pavement type. Similarly, the calculated PCE for good asphalt pavement was on average
5% lower than the PCE for fair asphalt. The PCE on good pavements exhibited much less
variability because the value of the rolling coefficient and pavement friction coefficient
for good pavement is very close to the value of the coefficients for the default, fair
asphalt, pavement. Excellent concrete pavement had a much improved rolling coefficient
and pavement friction coefficient compared to the default pavement. For this reason, the
PCE for excellent concrete pavement was always lower than the PCE for fair asphalt. On
average, the PCE for excellent concrete was 8% lower than the PCE for fair asphalt;
however, the effect of excellent concrete pavement decreased as the proportion of trucks
increased.
Snow covered pavement resulted in PCEs that were as much as 267% higher than
the PCEs calculated for the default pavement. Snow covered pavement does not depend
on the pavement type beneath, but represents 5.08 cm (2 in) of snow cover. The effect of
snow covered pavement decreased as the proportion of trucks increased. The effect of
snow cover was not primarily limited only to trucks; it also affected the speed of
passenger cars. For this reason, the PCEs exhibited a different trend than for the other
pavement types. The PCEs for snow covered pavement were constant regardless of the
percent or length of grade. Snow cover dramatically increases the pavement rolling
coefficient and decreases the friction coefficient. Essentially, the worst effect of the snow

56

Ingle

cover applied to all grade combinations. In the case of long 6% grades, the PCEs
calculated for snow covered pavement were actually less than the PCEs calculated for the
default pavement. This represents the situation that passenger cars were also slowed
down due to the snow cover. Since the entire traffic stream slowed down, the PCE of
trucks was not as high as it was for the other pavement types and conditions where
passenger cars in the traffic stream remained at higher speeds. Roadway surface
coefficients were not obtained for rainy (wet pavement) conditions, but the effect of wet
pavement on the PCEs is expected to be similar with a different magnitude.
4.9 Variability of PCE by Truck Aerodynamic Treatment
Truck aerodynamic treatment was found only to have a significant effect on the
calculated PCE for 10% proportion of trucks, with a percentage difference of 14%. This
partially confirms a hypothesis of this research. However, it was expected that the effect
of aerodynamic treatment would also be significant at 5% trucks, as it was for the
pavement type and condition presented in section 4.8. The aerodynamic treatment of
trucks affects the vehicle drag coefficient, which in turn affects the maximum
acceleration. The default aerodynamic treatment was full aerodynamic treatment, with a
drag coefficient of 0.58. Two other aerodynamic treatment options were examined, one
with partial aerodynamic aids on the roof and the other with no aerodynamic aids.
Table 9 shows the percentage difference of the PCE for the different aerodynamic
treatments as compared to full aerodynamic treatment. It is not expected that the
calculated PCE for either of the two aerodynamic treatment options should be lower than
the default PCE because the effect of these treatments decreases the maximum vehicle
acceleration obtainable. The calculated PCE for partial aerodynamic aids on the roof was

57

Ingle

Table 9. Percentage difference in calculated PCE for different aerodynamic treatments compared to full aerodynamic treatment.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

% Error From Full Aerodynamic Treatment

Partial Aerodynamic Aids
No Aerodynamic Aids
Percentage of Trucks and Buses
5
10
25
100
5
10
25
0%
0%
0%
0%
0%
0%
0%
-25%
0%
0%
0%
0%
33%
0%
-25%
0%
0%
0%
0%
33%
0%
0%
33%
0%
0%
0%
33%
0%
0%
33%
0%
0%
0%
33%
0%
0%
33%
0%
0%
0%
33%
0%
0%
0%
0%
0%
25%
0%
0%
20%
0%
0%
0%
20%
0%
0%
17%
0%
0%
0%
17%
0%
25%
0%
0%
0%
0%
14%
0%
0%
0%
0%
0%
0%
0%
17%
0%
0%
0%
0%
0%
0%
17%
0%
0%
20%
0%
0%
0%
20%
0%
0%
0%
20%
0%
0%
14%
20%
0%
0%
0%
0%
0%
13%
0%
0%
0%
0%
0%
0%
0%
0%
8%
0%
-14%
0%
0%
0%
0%
-8%
0%
0%
0%
-8%
0%
0%
-1%
7%
0%
0%
4%
14%
3%

58

100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

Ingle

on average 4% higher than the default PCE. The greatest effect of the aerodynamic
treatment was at 10% trucks, and the effect decreased as the proportion of trucks
increased. Similarly, the calculated PCE for no aerodynamic aids was on average 6%
higher than the default PCE. Again, the effect of aerodynamic treatment was only
significant for 10% proportion of trucks. A realistic truck population may contain a
variety of aerodynamic treatments; these simulation results cover a worst case scenario, a
single truck population with no aerodynamic treatments.
4.10 Variability of PCE for Three Lane Segments
The calculated PCE for three lane freeway segments was found only to be
significantly different from two lane freeway segments at 5% proportion of trucks. This
partially confirms a hypothesis of this research. Two different situations were examined,
one in which trucks were allowed to use any of the three lanes and one in which trucks
were restricted from using the leftmost lane. Truck restrictions are common for many
three lane freeway segments. Table 10 shows the percentage difference of the PCE for
three lane segments as compared to two lane segments. The PCEs for three lane freeway
segments are generally lower because the third lane allows the opportunity for cars to
pass while trucks in the middle lane are passing other trucks. The calculated PCE for
three lane segments averages -9% for 5% proportion of trucks. This average is not
significant, but a preponderance of the observations for 5% proportion of trucks are
significantly different (having absolute differences greater than 10%). At higher
proportion of trucks, the PCE for three lane segments is hardly different than for two lane
segments. The truck lane restriction had no effect on the calculated PCE as compared
with three lane segments without truck lane restriction. This suggests that a truck lane

59

Ingle

Table 10. Percentage difference in calculated PCE for three lane freeway segments compared to two lane freeway segments.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

% Error From Two Lane Fwy Segments

Three Lanes w/o Restrictions
Three Lanes w/ Restrictions
Percentage of Trucks and Buses
5
10
25
100
5
10
25
0%
0%
0%
0%
0%
0%
0%
-25%
0%
0%
0%
-25%
0%
0%
-25%
0%
0%
0%
-25%
0%
0%
-25%
0%
0%
0%
-25%
0%
0%
-25%
0%
0%
0%
-25%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
20%
0%
0%
0%
0%
0%
0%
0%
-17%
0%
0%
0%
-17%
0%
0%
0%
-20%
0%
0%
0%
-20%
-13%
0%
0%
0%
-13%
0%
0%
-13%
0%
0%
0%
-13%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-11%
0%
0%
0%
0%
0%
0%
-9%
0%
0%
0%
-9%
0%
0%
-8%
-11%
0%
0%
-8%
-11%
0%
-17%
-11%
-14%
0%
-8%
-11%
-14%
-15%
-10%
-14%
0%
-15%
-10%
-14%
-9%
-3%
-3%
0%
-9%
-3%
-3%

60

100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

Ingle

restriction provides no significant capacity increase as compared with no truck lane

restriction. However, this must be considered in light of the fact that trucks in the
simulation and realistically tend only to use the two rightmost lanes on grades anyway.
4.11 Variability of PCE for Truck Speed Limit
The institution of a truck speed limit was found to significantly affect the
calculated PCE; percentage differences exceeded 10%. This confirms a hypothesis of this
research. The default situation was an equal speed limit for all vehicles at 112.6 km/h (70
mi/h). The simulated truck speed limit restricted trucks to 88.5 km/h (55 mi/h). Table 11
shows the percentage difference of the PCE for the truck speed limit as compared with
the PCE without a separate truck speed limit. The calculated PCE for the truck speed
limit was on average 32% higher than the calculated PCE without a truck speed limit.
The greatest effect of the truck speed limit occurs at 5% proportion of trucks. As the
proportion of trucks increases, the effect of the truck speed limit decreases. However, the
truck speed limit still produces a significant increase in the calculated PCE for 25%
proportion of trucks. At times, the calculated PCE for the truck speed limit was double
the PCE without a separate truck speed limit. For 100% proportion of trucks, the truck
speed limit actually had no substantial effect.
4.12 Variability by Level of Congestion
The level of congestion was found to significantly affect the calculated PCE;
percentage differences exceeded 10%. Average density was used in this research as a
determinant of the level of congestion. It was found that an average density value of 12.4
pc/km/ln (20 pc/ml/ln) resulted in PCEs that most closely matched those provided in the
HCM 2000. This average density value corresponds to LOS C. However, it is also
61

Ingle

Table 11. Percentage difference in calculated PCE for a truck speed limit at 88.5 km/hr (55 mi/h)
compared to no separate truck speed limit 112.6 km/hr (70 mi/h).

Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

% Error From No Truck Speed Limit

Truck Speed Limit
Percentage of Trucks and Buses
5
10
25
100
67%
33%
33%
0%
50%
67%
33%
0%
75%
100%
33%
0%
75%
100%
33%
0%
100%
100%
33%
0%
100%
100%
33%
0%
50%
25%
0%
0%
80%
40%
25%
0%
83%
33%
25%
0%
71%
33%
0%
0%
63%
50%
0%
0%
63%
33%
0%
0%
33%
40%
25%
0%
22%
29%
20%
0%
27%
25%
17%
0%
17%
11%
17%
0%
25%
11%
0%
0%
15%
10%
0%
0%
56%
47%
18%
0%

62

Ingle

desirable to quantify the effect of the level of congestion on the PCE because many
freeway segments do not operate at LOS C. Figure 9 shows the variability of the
calculated PCE for different average density values, ranging from 9.3 to 24.9 veh/km/ln
(15 to 40 veh/mi/ln). This range of average density values represents a LOS that ranges
from B to E.
It can be observed from figure 9 that for grades of 2 or 4%, the calculated PCE
does not vary by much. However, for steeper grades such as 6 or 8%, the calculated PCE
varies significantly by the average density value. Table 12 shows the percentage
difference of the PCE calculated for the different average density values compared to the
PCE calculated for an average density value of 12.4 pc/km/ln (20 pc/ml/ln). The greatest
difference occurs for average density values within the LOS D regime. This means that
PCE is most critical at LOS D congestion. The PCEs calculated for LOS B congestion are
mostly lower, and the PCEs calculated for LOS D or E congestion are mostly higher than
the PCEs calculated at LOS C congestion. The effect of level of congestion may be
accounted for by multiplying the PCEs calculated at LOS C congestion by a scale factor.
4.13 Combined Effect of Variables on PCEs
The individual effect of each of the variables presented in the previous sections
was examined with respect to the default scenario. This default scenario consisted of a
single truck population with an average weight to power ratio of 83.7 kg/kW (137.5
lb/hp), operating on fair asphalt pavement, with full aerodynamic features, with no
separate truck speed limit, and a level of congestion corresponding to LOS C. The
variability of the PCE was considered independent of any interaction that may occur
when these variables are combined. Demarchi and Setti (19) discussed the limitations of

63

Ingle
1.6 km 2% grade

0.8 km 2% grade
LOS B

LOS C

LOS D

LOS E

10.0

9.0

9.0

8.0

8.0

7.0

7.0

6.0

6.0
PCE

PCE

10.0

LOS B

5.0

LOS C

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0
6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

26.0

6.0

28.0

8.0

10.0

12.0

5% Trucks

18.0

20.0

22.0

24.0

26.0

28.0

10% Trucks

1.6km 4% grade

LOS D

LOS E

LOS B

10.0

9.0

9.0

8.0

8.0

7.0

7.0

6.0

6.0

PCE

PCE

10.0

LOS C

16.0

5% Trucks

10% Trucks

0.8 km 4% grade
LOS B

14.0

Density (veh/km/ln)

Density (veh/km/ln)

5.0

LOS C

LOS D

LOS E

5.0

4.0

4.0

3.0

3.0
2.0

2.0

1.0

1.0

0.0

0.0
6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

26.0

6.0

28.0

8.0

10.0

12.0

5% Trucks

LOS B

LOS C

16.0

18.0

5% Trucks

10% Trucks

20.0

22.0

24.0

26.0

28.0

10% Trucks

1.6 km 6% grade

0.8 km 6% grade
10.0

14.0

Density (veh/km/ln)

De nsity (ve h/km/ln)

LOS D

LOS E

LOS B

10.0

9.0

9.0

8.0

8.0

7.0

7.0

6.0

6.0

PCE

PCE

LOS E

5.0

4.0

0.0

5.0

LOS C

LOS D

LOS E

5.0

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

26.0

6.0

28.0

8.0

10.0

12.0

5% Trucks

10.0

LOS C

16.0

5% Trucks

10% Trucks

0.8 km 8% grade
LOS B

14.0

18.0

20.0

22.0

24.0

26.0

28.0

Density (veh/km/ln)

Density (veh/km/ln)

10% Trucks

1.6 km 8% grade

LOS D

LOS E

LOS B

10.0

9.0

9.0

8.0

8.0

7.0

7.0

6.0

6.0

PCE

PCE

LOS D

5.0

LOS D

LOS E

5.0

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0

0.0

LOS C

0.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

Density (veh/km/ln)
5% Trucks

26.0

28.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

26.0

28.0

Density (veh/km/ln)

10% Trucks

5% Trucks

10% Trucks

Figure 9. PCE variability by level of congestion for average single truck population 83.7 kg/kW
(137.5 lb/hp).

64

Ingle

Table 12. Percentage difference in calculated PCE for different levels of congestion, average single truck population 83.7 kg/kW (137.5
lb/hp).
Upgrade
(%)

Density
(veh/mi/ln)
(veh/km/ln)
% Trucks
Length
Length
(km)
(mi)
0.40
0.25
0.80
0.50
1.21
0.75
1.61
1.00
2.01
1.25
2.41
1.50
0.40
0.25
0.80
0.50
1.21
0.75
1.61
1.00
2.01
1.25
2.41
1.50
0.40
0.25
0.80
0.50
1.21
0.75
1.61
1.00
2.01
1.25
2.41
1.50
0.40
0.25
0.80
0.50
1.21
0.75
1.61
1.00
2.01
1.25
2.41
1.50
average=

15
9.3

20
12.4

% Error From LOS C Congestion, 12.4 veh/km/lm (20 veh/mi/ln)

25
30
35
15.5
18.6
21.8
10
5
10
5
10
5

10

-25%

0%

0%

0%

0%

33%

0%

33%

-25%

0%

0%

0%

0%

33%

0%

-17%

-20%

0%

0%

17%

20%

-14%

-17%

0%

0%

14%

-22%

-14%

0%

0%

-25%

-22%

0%

-21%

-20%

-25%

-18%

-22%

-14%

65

40
24.9
10

10

25%

33%

0%

0%

33%

25%

33%

0%

0%

17%

20%

33%

20%

17%

20%

0%

29%

17%

29%

17%

14%

0%

11%

14%

22%

29%

22%

29%

22%

14%

0%

8%

11%

17%

11%

17%

11%

8%

0%

0%

0%

7%

10%

14%

10%

14%

20%

21%

20%

0%

0%

6%

9%

13%

18%

19%

18%

19%

18%

0%

0%

8%

16%

14%

21%

23%

23%

13%

9%

Ingle

calculating the PCE of a specific truck type out of a population that contained multiple
truck types. They found that the combined impact of multiple truck types was not
accounted for by simply taking the algebraic sum of the impact of each truck type
separately. Their recommendation to workaround this problem was to consider only an
aggregate PCE which would represent all truck types present. This recommendation was
adopted in this research with respect to multiple truck populations, as presented in
chapter 5.
A similar consideration may be made for the interaction among the different
roadway and traffic characteristics considered in this research. However, it is
unreasonable to calculate an aggregate PCE for each combination of the roadway and
traffic characteristics examined in this research, or further, for all the roadway and traffic
characteristics that exist. Therefore an aggregate PCE that accounts for the roadway and
traffic characteristics was not calculated. Instead it was decided to determine what error
might be expected if the individual effects of the variables are combined algebraically. A
set of simulations was conducted for two case studies that involved variability of the
weight to power ratio, pavement type and condition, and level of congestion
simultaneously. Both case studies were examined at 10% proportion of trucks. The first
case study considered a single truck population with a weight to power ratio of 106.6
kg/kW (175 lb/hp), operating on poor concrete pavement, and a level of congestion
corresponding to LOS D. Each of these variables tends to increase the PCE when
considered individually. The second case study considered a single truck population with
a weight to power ratio of 68.5 kg/kW (112.5 lb/hp), operating on excellent concrete

66

Ingle

pavement, and a level of congestion corresponding to LOS B. Each of these variables

tends to decrease the PCE when considered individually.
For each case study, flow and density data were obtained as output from traffic
simulations. The equivalent passenger car throughput (base flow) was calculated as
q B = (1 PT ) q M + PT ET SF q M

[4.1]

where qB is the base flow rate (passenger cars only), PT is the proportion of all trucks, qM
is the mixed vehicle flow rate, ET is the aggregate PCE of trucks, and SF are any number
of scale factors which are recommended to account for the effect of roadway and traffic
characteristics. A set of scale factors are recommended in chapter 6, which were
determined based upon the results presented in this chapter. The scale factors are based
on a ratio of the PCE; they are multiplicative such that they will either increase or
decrease the PCE calculated for the default situation depending on whether the variable
in question has an increasing or decreasing effect on the PCE. It was assumed for the case
study that the effects of different variables on the PCE could be combined without
considering any interactive effects. A consequence of this assumption is that the scale
factor for a variable that tends to increase the PCE may counteract the scale factor for a
variable that tends to decrease the PCE. However, this consequence was not examined in
the two case studies presented here because in either case study, all of the variables tend
to have the same effect on the PCE.
The base flow was calculated using the default flow data at a level of congestion
corresponding to LOS B and LOS D. The only scale factor applied to the default flow
data was a scale factor that accounts for the effect of the level of congestion. Table 13
shows the flow data and calculated base flow for the default scenario as well as the two

67

Ingle

Table 13. Case study 1 and 2 for combined effect of variables.

Default Flows
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.25
Flow (veh/h/ln)
Density (veh/km/ln)
0.80
0.50
Flow (veh/h/ln)
Density (veh/km/ln)
1.21
0.75
Flow (veh/h/ln)
Density (veh/km/ln)
1.61
1.00
Flow (veh/h/ln)
Density (veh/km/ln)
2.01
1.25
Flow (veh/h/ln)
Density (veh/km/ln)
2.41
1.50
Flow (veh/h/ln)
Density (veh/km/ln)
0.40
0.25
Flow (veh/h/ln)
Density (veh/km/ln)
0.80
0.50
Flow (veh/h/ln)
Density (veh/km/ln)
1.21
0.75
Flow (veh/h/ln)
Density (veh/km/ln)
1.61
1.00
Flow (veh/h/ln)
Density (veh/km/ln)
2.01
1.25
Flow (veh/h/ln)
Density (veh/km/ln)
2.41
1.50
Flow (veh/h/ln)
Density (veh/km/ln)
0.40
0.25
Flow (veh/h/ln)
Density (veh/km/ln)
0.80
0.50
Flow (veh/h/ln)
Density (veh/km/ln)
1.21
0.75
Flow (veh/h/ln)
Density (veh/km/ln)
1.61
1.00
Flow (veh/h/ln)
Density (veh/km/ln)
2.01
1.25
Flow (veh/h/ln)
Density (veh/km/ln)
2.41
1.50
Flow (veh/h/ln)
Density (veh/km/ln)

Base Flow

10% single trucks 83.3 kg/kW

(137.5 lb/hp), fair asphalt
pavement
LOS B
LOS C
LOS D

Base Flow

at density =
at density = 19.6
9.3 veh/km/ln veh/km/ln
(pc/h/ln)
(pc/h/ln)

Case Study 1
Base Flow
10% single truck
at density =
106.6 kg/kW (175
lb/hp), poor concrete 19.6
veh/km/ln
pavement
LOS C
LOS D
(pc/h/ln)

Error

Case Study 2
10% single truck
68.5 kg/kW (112.5
from default lb/hp), excellent
base flow concrete pavement
LOS B
LOS C

Base Flow

Error

at density =
9.3 veh/km/ln
(pc/h/ln)

from default
base flow

814
7.0

1352
12.6

1846
19.0

1065

2044

1350
15.0

1835
24.0

1976

-3%

817
8.0

1349.0
13.8

963

-10%

819
7.2

1344
12.6

1840
19.0

1049

2042

1353
16.5

1842
26.2

1868

-9%

823
8.0

1351.0
14.0

960

-8%

817
7.2

1346
12.8

1833
18.8

1040

2051

1347
16.1

1822
25.2

1894

-8%

820
8.1

1351.8
14.0

952

-8%

820
7.3

1349
12.8

1829
18.7

1035

2052

1358
16.8

1856
23.9

1924

-6%

817
8.1

1356.1
14.0

954

-8%

820
7.3

1356
13.0

1826
18.7

1035

2056

1361
17.2

1847
23.2

1929

-6%

816
8.1

1355.0
14.0

950

-8%

827
7.4

1352
13.0

1830
18.7

1032

2057

1349
16.5

1818
22.3

1982

-4%

828
8.2

1347.3
14.3

948

-8%

818
7.2

1348
13.2

1828
20.3

1069

2033

1348
14.6

1825
23.6

2049

1%

821
8.0

1348.0
13.8

964

-10%

821
7.6

1345
14.1

1828
21.3

1063

2057

1350
16.1

1843
25.8

2079

1%

822
8.0

1350.4
13.9

990

-7%

824
7.9

1348
14.6

1833
22.1

1078

2107

1356
16.5

1829
24.0

2251

7%

823
8.1

1351.6
14.1

1008

-7%

819
8.0

1352
14.8

1831
22.5

1063

2079

1350
16.7

1829
23.1

2278

10%

820
8.1

1341.8
14.1

1002

-6%

822
8.0

1346
14.9

1820
22.4

1063

2068

1348
17.1

1832
22.6

2277

10%

819
8.2

1342.0
13.9

998

-6%

823
8.2

1342
15.1

1834
23.0

1041

2042

1350
15.9

1831
22.5

2354

15%

817
8.2

1347.9
14.0

998

-4%

825
7.6

1349
14.2

1824
22.2

1063

2003

1347
14.5

1826
24.0

2180

9%

824
8.1

1346.2
13.9

992

-7%

817
8.3

1356
16.1

1835
25.2

1054

2031

1353
15.8

1823
24.9

2394

18%

821
8.1

1352.9
14.0

1054

0%

821
8.8

1350
17.0

1823
25.9

1055

2051

1345
16.5

1833
25.4

2479

21%

826
8.2

1352.7
14.1

1079

2%

822
9.0

1342
17.4

1815
26.2

1072

2102

1349
16.8

1832
23.2

2699

28%

820
8.1

1350.7
13.9

1109

3%

812
9.2

1341
17.7

1821
26.5

1049

2078

1345
16.5

1854
23.7

2702

30%

816
8.1

1354.0
14.2

1105

5%

822
9.3

1356
18.1

1833
26.8

1082

2153

1351
16.3

1802
21.7

2964

38%

815
8.1

1352.2
14.1

1132

5%

68

Ingle

case studies. The base flow for the first and second case studies was calculated using
scale factors to account for the effect of weight to power ratio, pavement type and
condition, and level of congestion. The percent error was calculated between the case
study base flow and the default base flow. For the first case study, the error in the
calculated base flow was as much as 38% on a 2.41 km (6 mi), 6% grade. However, if an
up to 15% error in the calculated base flow is acceptable, the combined effect of the
variables (use of multiple scale factors) may be applied for up to a 2.41 km (1.5 mi), 4%
grade or a 0.4 km (0.25 mi) 6% grade. The shaded area in table 13 represents the range of
grades where the combined effect of the variables results in too great of errors. For the
second case study, the error in the calculated base flow was as much as 10% on a 0.40
(0.25 mi), 2% grade. The combined effect of the variables may be applied for all of the
grade length and percent combinations examined for the second case study.
The two case studies do not constitute a comprehensive review of the combined
effect of roadway and traffic characteristics. However, the case studies do illustrate that
there is an applicable range where the use of multiple scale factors may be used if a
certain error in the calculated base flow is acceptable. It may be prudent that further
research should focus on the applicable range of the combined effect of the roadway and
traffic characteristics on the PCE.
4.14 Conclusions
Based on the results presented for single truck populations, the following
conclusions have been made. First, truck weight to power ratio has a significant effect on
the calculated PCE. A scale factor may be applied to the PCEs calculated for the average
weight to power ratio to account for the effect of weight to power ratio. Second, the PCE

69

Ingle

of trucks remains relatively constant for grades longer than 3.22 km (2 mi), thus the range
of grade lengths provided in the HCM 2000 should be extended to include a 3.22 km (2
mi) grade. Third, since the PCE of trucks continues to increase as the percentage grade
increases to 8%, PCEs should be provided for steeper grades and assumed to continue to
increase as the grade increases. Fourth, engine power only significantly affects the PCE
at a low engine power and a low proportion of trucks. It is sufficient to determine PCEs
for different weight to power ratios, but to ignore the variability of engine power within a
weight to power ratio. Fifth, since the PCE of trucks remains relatively constant for a
proportion of trucks greater than 60%, PCEs should be provided for up to 60% proportion
of trucks. Sixth, pavement type and condition significantly affects the calculated PCE.
Adjustments for pavement type and condition should be made, especially for a low
proportion of trucks. Seventh, aerodynamic treatment significantly affects the calculated
PCE for a low proportion of trucks only; it is sufficient not to make an adjustment for
aerodynamic treatment. Eighth, the PCE for three lane segments is significantly lower
than the PCE for two lane segments only for a low proportion of trucks, thus an
adjustment for the number of lanes is not necessary. In addition, lane restrictions have no
impact on the calculated PCE. Ninth, a separate truck speed limit below the speed limit of
other vehicles causes a significant increase in the calculated PCE. A scale factor may be
applied to the PCEs calculated for trucks without a separate speed limit in order to
account for the effect of a truck speed limit. Tenth, the PCEs provided in the HCM 2000
were calculated for LOS C congestion. A scale factor may be applied to the PCEs to
account for the effect of level of congestion. Eleventh, the combined affect of roadway
and traffic characteristics results in errors in the calculated base flow that are not entirely

70

Ingle

accounted for by the combination of individual scale factors. However, with the
exception of very steep grades, the error due to the combined effects is acceptable.
Further study should focus on the applicability of multiple scale factors to account for the
combined effects of roadway and traffic characteristics.

71

Ingle

Chapter 5. Results for Multiple Truck Populations

5.1 Background
In the previous chapter, the calculation of the PCE of trucks was simplified by
assuming that all trucks had the same characteristics. However, a realistic truck
population may contain any number of trucks with different operating characteristics.
Multiple truck populations were investigated as part of this research in order to examine
the validity of the single truck population simplification. Multiple truck populations were
defined to contain trucks with different weight to power ratios. This not only represents
the variability in engine power, but also the difference between loaded and unloaded
trucks. Pursuant the finding from the single truck population tests, that a truck population
with an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp) most closely matched
the PCEs from the HCM 2000, multiple truck populations were created to have the same
average weight to power ratio. Six different multiple truck populations were created.
Two of the multiple truck populations, referred to as J and K, were created to
model the distribution observed in the I-81 truck survey in Virginia (23). The J
population was created to match the exact bin frequencies from the I-81 truck survey.
However, due to aggregating the data into bins of 5.2 kg/kW (25 lb/hp), the average
weight to power ratio of this population was 93.5 kg/kW (153.5 lb/hp). The K population
was modified to closely match the distribution observed in the I-81 truck survey, while
maintaining an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp). Figure 10
shows the distribution of trucks in the J and K populations. It can be observed from figure
10 that the population distributions are similar to a normal distribution with skew to the
right.

72

Ingle

Multiple Truck Population Distribution

100
90
80
Percentage

70
60
50
40
30
20
10
0
38.1

53.3

68.5

83.7

99.0

114.2

129.4

Weight to Power Ratio (kg/kW)

Single Truck

J Trucks

K Trucks

Figure 10. Truck distribution for J and K multiple truck populations.

73

144.6

159.9

Ingle

Four of the multiple truck populations, referred to as L, M, N, and O, were created

to model a normal distribution with an average weight to power ratio of 83.7 kg/kW
(137.5 lb/hp). These normal distributions were created by varying the standard deviation
of the weight to power ratio in increments of 10% of the average weight to power ratio.
Thus the standard deviation of the L population was 10% of the average weight to power
ratio or 8.37 kg/kW (13.75 lb/hp) and so on up to 40% of the average weight to power
ratio for the O population. Figure 11 shows the obtained multiple truck population
distributions. These distributions also contain a skew to the right, which was created to
account for the criteria that the lowest weight to power truck simulated was 38.1 kg/kW
(62.5 lb/hp). Of the normal multiple truck population distributions, the N population most
closely matched the results of the I-81 truck survey, having a standard deviation of 25.12
kg/kW (41.25 lb/hp).
In a similar way as the single truck population investigations, the multiple truck
populations were examined to see that they closely matched the PCEs provided in the
HCM 2000. The variability of the multiple truck population by length and percent grade
was used to validate the chosen population. High proportion of trucks, pavement type and
condition, aerodynamic treatment, three lane segments, truck speed limit, and level of
congestion were then examined for the multiple truck population.
The variables mentioned above were used to verify the hypotheses of this
research. The primary hypothesis in testing multiple truck populations was that the
calculated PCEs would be the same as those calculated for a single truck population with
the same weight to power ratio. Section 5.3 presents the results of the comparison
between single and multiple truck populations. The other hypotheses that were

74

Ingle

Multiple Truck Population Distribution

100
90
80
Percentage

70
60
50
40
30
20
10
0
38.1

53.3

68.5

83.7

99.0

114.2

129.4

144.6

Weight to Power Ratio (kg/kW)

Single Truck

L Trucks

M Trucks

N Trucks

O Trucks

Figure 11. Truck distribution for L, M, N, and O multiple truck populations.

75

159.9

Ingle

tested for single truck populations were also tested for multiple truck populations. It was
hypothesized that the PCE for a multiple truck population decreases with increasing
proportion of trucks. Section 5.6 presents the results of the variability of PCE by
proportion of trucks. Another hypothesis of this research was that pavement type and
condition significantly influence the PCE, with poor pavement types having higher PCEs
in comparison to the default pavement type. The results of the variability of PCE by
pavement type and condition are presented in section 5.7. It was hypothesized that truck
aerodynamic treatment would significantly affect the PCE. Section 5.8 presents the
results of the variability of PCE by truck aerodynamic treatment. Another hypothesis of
this research was that PCEs for three lane freeway segments would be much lower than
PCEs for two lane freeway segments. The results of the variability of PCE for three lane
segments are presented in section 5.9. It was hypothesized that the institution of a truck
speed limit below the speed limit of other vehicles would result in significantly higher
PCEs. The results of the variability of PCE for a truck speed limit are presented in section
5.10. A final hypothesis of this research was that the level of congestion will significantly
effect the calculated PCE. Section 5.11 presents the variability of PCE by level of
congestion.
5.2 Speed-Flow-Density Relationships
The speed-flow-density relationships for the uncongested flow regime were
created using flow and density output data obtained from the traffic simulations. The base
vehicle speed-flow-density relationships were already presented in figure 3 and remained
unchanged for the multiple truck population research. Figure 12 shows the speed-flowdensity relationships for the multiple truck population N, with an average

76

Ingle
Speed-Flow (0.8 km)

140

140

120

120

100

100
Speed (km/h)

Speed (km/h)

Speed-Density (0.8 km)

80
60

80
60

40

40

20

20
0

0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

0.0

35.0

500.0

1000.0

1500.0

2000.0

2500.0

Flow (ve h/h/ln)

De nsity (ve h/km/ln)

Flow-Density (0.8 km)

2500.0

2000.0

1500.0

Qm 5% trucks 4% grade

Qm 5% trucks 2% grade

150
Speed (kph)

Flow (veh/h/ln)

Legend

1000.0

500.0

100
50

Qm 10% trucks 2% grade

Qm
5% trucks 6% grade
0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

De ns ity (vpk m pln)

0.0
0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

Qm 10% trucks 4% grade

Qm 10% trucks 6% grade

De nsity (ve h/km/ln)

Figure 12. Speed-flow-density relationships, mixed vehicle flow multiple truck population of 83.7 kg/kW (137.5 lb/hp) and standard
deviation of 25.12 kg/kW (41.25 lb/hp).
77

Ingle

weight to power ratio of 83.7 kg/kW (137.5 lb/hp) and standard deviation of 25.12 kg/kW
(41.25 lb/hp). It can be observed from the flow-density relationship that for an equal
density value, flow decreases with increasing proportion of trucks and even more so for
increasing percentage grade. This trend is also observed in the speed-density and speedflow relationships at equal values of density and flow respectively. The general trends
observed in the mixed vehicle speed-flow-density relationships are as would be expected
for freeway segments in the uncongested flow regime.
5.3 PCE for Single vs Multiple Truck Populations
The PCEs calculated for a multiple truck population were not significantly
different from the PCEs calculated for a single truck population; percentage differences
rarely exceeded 10%. This disproves a hypothesis of this research, that a multiple truck
population would perform differently than a single truck population. Figure 13 shows the
calculated PCE for multiple truck populations plotted relative to the PCE from the HCM
2000 for specific grades. The M multiple truck population most closely matches the
HCM 2000 PCEs for the range of values presented in figure 13. However, as mentioned
in section 5.1, the N multiple truck population most closely matched the I-81 truck survey
results. For this reason, the N multiple truck population was chosen for further
investigative study. Overall, the calculated PCEs for the multiple truck populations were
rarely more than 0.5 in absolute difference from the HCM 2000 PCEs.
The calculated PCEs for the N multiple truck population are provided in table 14.
The difference between the single and multiple truck population is characterized in table
15, which shows the percentage difference between the PCEs calculated for the average
single truck population and the PCEs calculated for the N multiple truck population. The

78

Ingle
1.6 km 2% grade

6.0

6.0

5.0

5.0

4.0

4.0
PCE

P CE

0.8 km 2% grade

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

Multiple Truck Population

5% trucks

5% trucks

HCM 2000

0.8 km 4% grade

HCM 2000

6.0

5.0

5.0

4.0

4.0

3.0

P CE

PCE

1.6 km 4% grade

6.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

Multiple Truck Population

5% trucks

Multiple Truck Population

HCM 2000

5% trucks

HCM 2000

1.6 km 6% grade

0.8 km 6% grade
6.0

6.0

5.0

5.0

4.0

4.0
P CE

P CE

Multiple Truck Population

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
J

5% trucks

Multiple Truck Population

Multiple Truck Population

5% trucks

HCM 2000

HCM 2000

Figure 13. PCE variability by multiple truck populations at LOS C congestion.

79

Ingle

Table 14. PCEs for trucks and buses on upgrades, N multiple truck population, 83.7 kg/kW (137.5 lb/hp) and standard deviation of 25.12
kg/kW (41.25 lb/hp) at LOS C congestion.

Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50

10

15

ET
Percentage of Trucks and Buses
20
25
30

40

50

60

70

80

90

100

1.5
2.0
2.0
2.0
2.0
2.0

1.5
1.5
2.0
2.0
2.0
2.0

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

1.5
1.5
1.5
1.5
1.5
1.5

2.0
3.0
3.5
4.0
4.0
4.0

2.0
2.5
3.0
3.0
3.0
3.0

2.0
2.0
2.5
2.5
2.5
2.5

2.0
2.0
2.5
2.5
2.5
2.5

2.0
2.0
2.0
2.5
2.5
2.5

1.5
2.0
2.0
2.0
2.5
2.0

1.5
2.0
2.0
2.0
2.0
2.0

1.5
2.0
2.0
2.0
2.0
2.0

1.5
1.5
2.0
2.0
2.0
2.0

1.5
1.5
2.0
2.0
2.0
2.0

1.5
1.5
2.0
2.0
2.0
2.0

1.5
1.5
2.0
2.0
2.0
2.0

3.0
4.5
5.5
6.0
6.0
6.0

2.5
3.5
4.5
4.5
4.0
4.0

2.5
3.0
3.0
3.5
3.0
3.0

2.5
3.0
3.0
3.0
3.0
3.0

2.0
2.5
3.0
3.0
3.0
3.0

2.0
2.5
2.5
2.5
3.0
2.5

2.0
2.5
2.5
2.5
2.5
2.5

2.0
2.5
2.5
2.5
2.5
2.5

2.0
2.0
2.5
2.5
2.5
2.5

2.0
2.0
2.0
2.5
2.5
2.5

1.5
2.0
2.0
2.5
2.5
2.5

1.5
2.0
2.0
2.0
2.5
2.5

80

Ingle

Table 15. Percentage difference in calculated PCE for N multiple truck population compared to the average single truck population.

Upgrade
(%)

Length
(km)
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

Length
(mi)
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average=
average=
average=

10

15

% Error From Single Truck Population

Percentage of Trucks and Buses
20
25
30

40

50

60

70

80

90

100

0%
-33%
-33%
0%
0%
0%

0%
0%
-33%
-33%
-33%
-33%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
-50%
-50%

0%
-20%
0%
-14%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
0%
0%
0%
0%

0%
0%
-25%
0%
0%
0%

0%
0%
0%
-25%
-25%
0%

0%
0%
0%
0%
-25%
0%

0%
-33%
0%
0%
0%
0%

0%
-33%
0%
0%
0%
0%

0%
0%
-33%
-33%
-33%
-33%

0%
0%
-33%
-33%
-33%
-33%

0%
0%
-33%
-33%
-33%
-33%

0%
0%
-33%
-33%
-33%
-33%

0%
0%
14%
0%
14%
14%
2%

-25%
-20%
0%
0%
14%
14%
-2%

0%
-25%
0%
0%
0%
0%
-3%

-33%
0%
-25%
-25%
-25%
-25%
-15%

-33%
0%
0%
-25%
-25%
-25%
-13%

0%
0%
0%
-25%
-25%
-25%
-12%

0%
0%
0%
0%
-25%
-25%
-16%

0%
0%
0%
0%
0%
8%
-5%
-6%
-2%

81

0%
0%
-13%
0%
11%
20%
-6%

0%
0%
0%
0%
0%
0%
-3%

0%
0%
0%
17%
0%
17%
0%

0%
0%
0%
0%
0%
0%
-2%

Ingle

average difference is -6% considering all observations, and as little as -2% if observations
for the very high proportion of trucks (70 to 100% trucks) are ignored. A proportion of
trucks beyond 70% is not very likely, but was used in this research for investigative
purposes. The closest match between the single and multiple truck populations occurs for
PCEs calculated for between 30 to 60% trucks. This is a very meaningful result because
freeways operating with between 30 to 60% trucks may typify current freeways with
heavy truck use. It was expected that the effect of multiple truck populations would be
most dramatic for freeways with heavy truck use; however this observation supports the
contrary. These data present no significant evidence that the multiple truck population
performs differently than the single truck population.
5.4 Variability of PCE by Length of Grade
The PCE was observed to increase as the length of grade increased. This trend
exists in the PCEs provided in the HCM 2000, and it was observed for the average single
truck population results presented in section 4.4. Figure 14 shows the variability of the
PCE over lengths of grade ranging from 0.40 to 2.41 km (0.25 to 1.5 mi). In all but one of
the plots, the PCE is constant for grades 1.6 km (1 mi) and longer. This suggests that
grades longer than 1.6 km (1 mi) have no additional effect on truck performance. The
HCM 2000 similarly shows this trend by providing PCEs for length of grade criteria 1.6
km (1 mi) and longer for grades 4, 5, or 6%. However, for grades 2 and 3%, the length of
grade criteria includes grades 2.4 km (1.5 mi) and longer. There is slightly more
variability in the PCE by length of grade for the N multiple truck population as compared
to the average single truck population. This variability is evidence of the interaction of
different truck types among other trucks and passenger cars in the traffic stream.

82

Ingle
2% grade

7.0

7.0

6.0

6.0

5.0

5.0

4.0

4.0

P CE

P CE

2% grade

3.0

3.0

2.0

2.0

1.0

1.0

0.0
0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

2.00

2.20

0.0
0.40

2.40

0.60

0.80

1.00

10% Trucks

HCM 2000

6.0

5.0

5.0

4.0

4.0

P CE

P CE

6.0

3.0

2.0

1.0

1.0

1.20

1.40

1.60

1.80

2.00

2.20

0.0
0.40

2.40

0.60

0.80

1.00

6.0

5.0

5.0

4.0

4.0

3.0

2.0

1.0

1.0

1.40

1.60

1.80

2.00

2.00

2.20

2.40

2.00

2.20

2.40

HCM 2000

2.20

2.40

0.0
0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

Length of Grade (km)

Length of Grade (km)

5% Trucks

1.80

3.0

2.0

1.20

2.40

6% grade

P CE

P CE

1.60

10% Trucks

HCM 2000

7.0

1.00

1.40

5% Trucks

6.0

0.80

1.20

Length of Grade (km)

7.0

0.60

2.20

HCM 2000

Length of Grade (km)

6% grade

0.0
0.40

2.00

3.0

2.0

1.00

1.80

4% grade
7.0

0.80

1.60

5% Trucks

7.0

0.60

1.40

Length of Grade (km)

4% grade

0.0
0.40

1.20

Length of Grade (km)

10% Trucks

HCM 2000

HCM 2000

Figure 14. PCE variability by length of grade for N multiple truck population, 83.7 kg/kW (137.5
lb/hp) and standard deviation of 25.12 kg/kW (41.25 lb/hp) at LOS C congestion.

83

Ingle

However, the trends observed in figure 14 are not significantly different than the trends
observed in the single truck population results to suggest that the multiple truck
population performs differently.
5.5 Variability of PCE by Percent Grade
The PCE was observed to increase as the percent grade increased. This trend was
also used to validate the selected multiple truck population because it is a determining
factor used in the PCEs provided in the HCM 2000. Figure 15 shows the variability of the
PCE over percent grades ranging from 1 to 6%. It is clear that the N multiple truck
population with an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp) and
standard deviation of 25.12 kg/kW (41.25 lb/hp) very closely matches the trend in PCEs
from the HCM 2000. Regardless of the length of grade, the PCE for 1% grades is always
1.5. In addition, the relationship between PCE and percent grade is almost the same for
all grades 1.6 km (1 mi) and longer. This supports the observation made in section 5.5
that the PCE is the same for grades 1.6 km (1 mi) and longer. The average difference in
PCE from one percent grade to the next is 0.4. A uniform incremental increase of 0.5 is
most clearly observed for higher proportion of trucks, like 20 or 25% trucks. For the 5%
proportion of trucks, the incremental increase is more likely to be either 0.5 or 1.0 from
one percent grade to the next.
5.6 Variability of PCE by Proportion of Trucks
The proportion of trucks in the traffic stream was found to have a significant
effect on the calculated PCE only at low proportion of trucks. As the proportion of trucks
increases, the PCE decreases. This confirms a hypothesis of this research. The decreasing
trend in PCEs with proportion of trucks is especially evident for longer and steeper
84

Ingle
0.8 km grade

7.0

7.0

6.0

6.0

5.0

5.0

4.0

4.0

P CE

P CE

0.4 km grade

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
1

5% trucks

HCM 2000

5% trucks

HCM 2000

1.2 km grade

7.0

7.0

6.0

6.0

5.0

5.0

4.0

4.0

P CE

P CE

1.6 km grade

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

Percentage Grade
5% trucks

Percentage Grade

HCM 2000

5% trucks

2.0 km grade

HCM 2000

2.4 km grade

7.0

7.0

6.0

6.0

5.0

5.0

4.0

4.0

P CE

P CE

Percentage Grade

Percentage Grade

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

Percentage Grade
5% trucks

Percentage Grade

HCM 2000

5% trucks

HCM 2000

Figure 15. PCE variability by percentage grade for N multiple truck population, 83.7 kg/kW
(137.5 lb/hp) and standard deviation of 25.12 kg/kW (41.25 lb/hp) at LOS C
congestion.

85

Ingle

grades. Figure 16 shows the variability of the PCE over proportion of trucks ranging from
5% to 100%. It can be observed that the PCE is highest for low proportion trucks, and
that the PCE decreases and levels off as the proportion trucks increases. The PCE past
60% trucks shows very little variability. In fact, the calculated PCE has no variability
between 40 and 60% trucks. Since this range may represent an upper limit on the
reasonable proportion of trucks, PCEs provided up to 40% trucks would sufficiently
describe the variability of PCE by proportion of trucks. In comparison to the single truck
population, variability of PCE by proportion of trucks is less for the multiple truck
population. This is contrary to what was expected in this research. It seems that the
multiple truck population has a smoothing effect, which results in less variability as
compared to the single truck population.
5.7 Variability of PCE by Pavement Type and Condition
Pavement type and condition was found to have a significant effect on the
calculated PCE; percentage differences exceeded 10%. This confirms a hypothesis of this
research. The default pavement type for simulations in this research was fair asphalt
pavement. Table 16 shows the calculated PCEs for the different pavement types and
conditions for the N multiple truck population. Table 17 shows the variability of the PCE
for different pavement types and conditions calculated as the percent difference from the
default pavement type. Poor asphalt and poor concrete pavement always had higher or
equal PCEs as compared with fair asphalt. The PCEs for the poor asphalt pavement was
on average 11% higher than the PCEs for the fair asphalt pavement, and the PCEs for the
poor concrete pavement was on average 6% higher. For either of these pavement types,
there was much less variability for 100% trucks as compared to a lower proportion of

86

Ingle
1.6 km 2% grade

7.0

7.0

6.0

6.0

5.0

5.0

4.0

4.0

P CE

P CE

0.8 km 2% grade

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

10

20

30

40

50

60

70

80

90

100

10

20

30

Percentage Trucks

70

80

90

100

70

80

90

100

70

80

90

100

1.6 km 4% grade

7.0

7.0

6.0

6.0

5.0

5.0

4.0

4.0

P CE

P CE

60

HCM 2000

2% grade

0.8 km 4% grade

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
0

10

20

30

40

50

60

70

80

90

100

10

20

30

40

50

60

Percentage Trucks

Percentage Trucks

4% grade

4% grade

HCM 2000

0.8 km 6% grade

HCM 2000

1.6 km 6% grade

7.0

7.0

6.0

6.0

5.0

5.0

4.0

4.0

P CE

P CE

50

Percentage Trucks

HCM 2000

2% grade

40

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0

10

20

30

40

50

60

70

80

90

100

10

20

30

40

50

60

Percentage Trucks

Percentage Trucks

6% grade

6% grade

HCM 2000

HCM 2000

Figure 16. PCE variability by proportion of trucks for N multiple truck population, 83.7 kg/kW
(137.5 lb/hp) and standard deviation of 25.12 kg/kW (41.25 lb/hp) at LOS C
congestion.

87

Ingle

Table 16. PCEs for trucks and buses on upgrades with different pavement types, N multiple truck population, 83.7 kg/kW (137.5 lb/hp) and
standard deviation of 25.12 kg/kW (41.25 lb/hp) at LOS C congestion.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50

Asphalt Poor
5
2.0
2.0
2.0
2.5
2.5
2.5
2.5
3.5
4.0
4.5
4.5
4.5
3.0
5.5
6.5
6.5
6.5
6.5

10
1.5
2.0
2.0
2.0
2.0
2.0
2.0
3.0
3.5
3.5
3.5
3.5
3.0
4.0
4.5
5.0
4.5
4.5

25
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
2.5
2.5
3.0
3.5
3.5
3.0
3.5

Concrete Poor
100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5

5
2.0
2.0
2.0
2.0
2.5
2.5
2.0
3.0
3.5
4.0
4.5
4.5
3.0
5.0
6.0
6.0
6.0
6.0

10
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
3.0
3.5
3.5
3.5
3.0
4.0
4.5
4.5
4.0
4.0

25
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
2.5
3.0
3.0
3.0
3.0
3.0

100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
1.5
2.0
2.5
2.0
2.5
2.5

ET
Asphalt Good
Percentage of Trucks and Buses
5
10
25
100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
1.5
1.5
2.0
2.0
2.0
1.5
2.5
2.5
2.0
1.5
3.0
2.5
2.0
1.5
3.5
3.0
2.5
2.0
4.0
3.0
2.5
2.0
4.0
3.0
2.5
2.0
3.0
2.5
2.0
1.5
4.5
3.5
2.5
2.0
5.5
4.0
3.0
2.0
5.5
4.0
3.0
2.0
5.5
4.0
3.0
2.0
5.5
4.0
3.0
2.0

Concrete Good
5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.5
3.5
4.0
4.0
4.0
3.0
4.5
5.5
5.5
5.5
5.5

10
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.5
3.0
3.0
3.0
3.0
2.5
3.5
4.0
4.5
4.0
4.0

25
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.5
2.5
2.5
2.5
2.0
3.0
3.0
3.0
3.0
3.0

Concrete Excellent
100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
1.5
2.0
2.0
2.0
2.0
2.0

5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.5
3.0
3.0
3.5
3.5
2.5
4.0
5.0
5.5
5.5
5.5

10
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.5
2.5
3.0
3.0
3.0
2.5
3.5
4.0
4.0
4.0
4.0

25
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.5
2.5
2.0
2.5
3.0
3.0
3.0
3.0

100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
1.5
2.0
2.0
2.0
2.0
2.0

Table 17. Percentage difference in calculated PCE for different pavement types compared to asphalt fair pavement.
Upgrade
(%)

Length
(km)
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

Length
(mi)
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

Asphalt Poor
5
33%
0%
0%
25%
25%
25%
25%
17%
14%
13%
13%
13%
0%
22%
18%
8%
8%
8%
15%

10
0%
33%
33%
0%
0%
0%
0%
20%
17%
17%
17%
17%
20%
14%
0%
11%
13%
13%
12%

25
0%
0%
33%
33%
33%
33%
0%
25%
0%
0%
0%
0%
0%
0%
17%
17%
0%
17%
12%

Concrete Poor
100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
33%
0%
25%
25%
0%
0%
5%

88

5
33%
0%
0%
0%
25%
25%
0%
0%
0%
0%
13%
13%
0%
11%
9%
0%
0%
0%
7%

10
0%
33%
33%
0%
0%
0%
0%
0%
0%
17%
17%
17%
20%
14%
0%
0%
0%
0%
8%

25
0%
0%
0%
33%
33%
33%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
6%

100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
25%
0%
0%
0%
1%

% Error From Asphalt Fair

Asphalt Good
Percentage of Trucks and Buses
5
10
25
100
0%
0%
0%
0%
-25%
0%
0%
0%
0%
0%
0%
0%
-25%
-25%
0%
0%
-25%
-25%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-17%
0%
0%
0%
-14%
-17%
-20%
-25%
-13%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-20%
0%
0%
0%
-17%
0%
0%
-11%
0%
0%
-8%
-11%
0%
0%
-8%
0%
0%
-20%
-8%
0%
0%
-20%
-8%
-5%
-3%
-4%

Concrete Good
5
0%
-25%
0%
0%
0%
0%
0%
-17%
0%
0%
0%
0%
0%
0%
0%
-8%
-8%
-8%
-4%

10
0%
0%
0%
-25%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-11%
0%
0%
0%
-2%

25
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-20%
0%
0%
0%
0%
0%
-1%

Concrete Excellent
100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-20%
-20%
-2%

5
0%
-25%
-25%
-25%
-25%
-25%
0%
-17%
-14%
-25%
-13%
-13%
-17%
-11%
-9%
-8%
-8%
-8%
-15%

10
0%
0%
0%
-25%
-25%
-25%
0%
0%
-17%
0%
0%
0%
0%
0%
-11%
-11%
0%
0%
-6%

25
0%
0%
0%
0%
0%
0%
0%
0%
-20%
-20%
0%
0%
-20%
-17%
0%
0%
0%
0%
-4%

100
0%
0%
0%
0%
0%
0%
0%
0%
-25%
0%
0%
0%
0%
0%
0%
0%
-20%
-20%
-4%

Ingle

trucks. The variability in PCE was significant for the poor asphalt pavement at 5, 10, and
25% proportion of trucks. The calculated PCE for good concrete pavement was not
significantly different than the PCE calculated for the default pavement at any proportion
of trucks. Similarly, the calculated PCE for good asphalt pavement was on average only
5% lower than the PCEs for fair asphalt pavement, and the difference was only
significant at 5% proportion of trucks. The PCE for excellent concrete pavement was on
average 7% lower than the PCE for fair asphalt. This represents a significant decrease in
the PCE due to the increased pavement condition. Again, the greatest variability in the
PCE calculated for excellent concrete pavement was for 5 or 10% proportion of trucks.
5.8 Variability of PCE by Truck Aerodynamic Treatment
Truck aerodynamic treatment was found only to have a significant effect on the
calculated PCE for 10% proportion of trucks. This partially confirms a hypothesis of this
research. However, it was expected that the effect of aerodynamic treatment would also
be significant at 5% trucks, as it was for the pavement type and condition presented in
section 5.7. This same result was observed for the average single truck population
presented in section 4.8. It is not expected that the calculated PCE for either of the two
aerodynamic treatment options should be lower than the default PCE because the effect
of these treatments decreases the maximum vehicle acceleration obtainable. Table 18
shows the percentage difference of the PCE for the different aerodynamic treatments as
compared to the full aerodynamic treatment. The calculated PCE for partial aerodynamic
aids on the roof was on average 1% higher than the default PCE. A significant difference
(10%) between the PCEs calculated for partial aerodynamic treatment and the default,
full, aerodynamic treatment was only observed at 10% proportion of trucks. The

89

Ingle

Table 18. Percentage difference in calculated PCE for different aerodynamic treatments compared to full aerodynamic treatment.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

% Error From Full Aerodynamic Treatment

Partial Aerodynamic Aids
No Aerodynamic Aids
Percentage of Trucks and Buses
5
10
25
100
5
10
25
0%
0%
0%
0%
33%
0%
0%
0%
33%
0%
0%
0%
33%
0%
0%
33%
0%
0%
0%
33%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
25%
0%
33%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
17%
0%
0%
13%
17%
0%
0%
0%
0%
0%
13%
17%
0%
0%
20%
-20%
0%
0%
20%
0%
11%
14%
0%
0%
11%
14%
0%
9%
0%
0%
0%
9%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-20%
0%
13%
0%
-8%
0%
0%
-20%
0%
0%
0%
1%
7%
-1%
-2%
6%
8%
2%

90

100
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

Ingle

calculated PCE for no aerodynamic aids was on average 4% higher than the default PCE,
but the difference in the PCEs was significant for 10% proportion of trucks. Although
realistic truck populations contain a variety of aerodynamic treatments, the simulation
results suggest that even the worst case, a multiple truck population with no aerodynamic
treatments, does not have significantly higher PCEs across the full range of proportion of
trucks.
5.9 Variability of PCE for Three Lane Segments
The calculated PCE was found only to be significantly different (10%) for three
lane freeway segments as compared to two lane freeway segments for 5% proportion of
trucks. This partially confirms a hypothesis of this research. Two different situations were
examined, one in which trucks were allowed to use any of the three lanes and one in
which trucks were restricted from using the leftmost lane. Table 19 shows the percentage
difference of the PCE for three lane segments as compared to two lane segments. The
effect of three lane segments is only significant at 5% proportion of trucks. For 5%
proportion of trucks, the calculated PCE for three lane segments without lane restrictions
averages 8% less than the PCE for two lane segments. However, at a higher proportion of
trucks, the PCE for three lane segments is hardly different than for two lane segments.
The truck lane restriction had no effect on the calculated PCE as compared with three
lane segments without truck lane restriction. This suggests that a truck lane restriction
provides no significant capacity increase as compared with no truck lane restriction.
However, this must be considered in light of the fact that trucks in the simulation and
realistically tend only to use the two rightmost lanes on grades anyway.

91

Ingle

Table 19. Percentage difference in calculated PCE for three lane freeway segments compared to two lane freeway segments.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

% Error From Two Lane Fwy Segments

Three Lanes w/o Restrictions
Three Lanes w/ Restrictions
Percentage of Trucks and Buses
5
10
25
100
5
10
25
100
0%
0%
0%
0%
0%
0%
0%
0%
-25%
0%
0%
0%
-25%
0%
0%
0%
0%
33%
0%
0%
-25%
33%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-25%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-17%
0%
0%
0%
0%
0%
0%
0%
-14%
-17%
-20%
0%
-14%
0%
-20%
0%
-13%
0%
0%
0%
-13%
0%
0%
0%
-13%
0%
0%
0%
-13%
0%
0%
0%
-13%
0%
0%
0%
-13%
0%
0%
0%
0%
0%
0%
33%
0%
0%
0%
33%
0%
0%
0%
0%
0%
0%
0%
0%
-9%
-11%
0%
0%
0%
-11%
0%
0%
-17%
-11%
0%
0%
-8%
-11%
0%
0%
-17%
-13%
0%
-20%
-17%
-13%
0%
-20%
-17%
-13%
0%
-20%
-17%
-13%
0%
-20%
-8%
-3%
-1%
0%
-8%
-1%
-1%
0%

92

Ingle

5.10 Variability of PCE for Truck Speed Limit

The institution of a truck speed limit was found to significantly affect the
calculated PCE; percentage differences exceeded 10%. This confirms a hypothesis of this
research. The default situation was an equal speed limit for all vehicles at 112.6 km/h (70
mi/h). The simulated truck speed limit restricted trucks to 88.5 km/h (55 mi/h). Table 20
shows the percentage difference of the PCE for the truck speed limit as compared to
without a truck speed limit. The calculated PCE for the truck speed limit was on average
32% higher than the calculated PCE without a truck speed limit. It is interesting to note
that for 100% trucks, the truck speed limit actually had no affect. The maximum
difference was a calculated PCE that was 150% higher than the PCE without a truck
speed limit. The institution of a truck speed limit below the passenger car speed limit
results in a significant capacity reduction. However, the effect of the truck speed limit
decreases as the proportion of trucks increases.
5.11 Variability by Level of Congestion
The level of congestion was found to significantly affect the calculated PCE;
percentage differences exceeded 10%. Average density was used in this research as a
determinant of the level of congestion. It was found that an average density value of 12.4
pc/km/ln (20 pc/ml/ln) resulted in PCEs that most closely matched those provided in the
HCM 2000. This average density value corresponds to LOS C. However, it is also
desirable to quantify the effect of the level of congestion on the PCE because many
freeway segments do not operate at LOS C. Figure 17 shows the variability of the
calculated PCE for different average density values ranging from 9.3 to 25.9 veh/km/ln
(15 to 40 veh/mi/ln). This range of average density values represents a LOS that ranges

93

Ingle

Table 20. Percentage difference in calculated PCE for a truck speed limit at 88.5 km/hr (55 mi/h)
compared to no separate truck speed limit 112.6 km/hr (70 mi/h).

Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average =

% Error From No Truck Speed Limit

Truck Speed Limit
Percentage of Trucks and Buses
5
10
25
100
67%
33%
33%
0%
75%
67%
33%
0%
100%
100%
33%
0%
125%
75%
33%
0%
125%
75%
33%
0%
150%
75%
67%
0%
50%
25%
0%
0%
50%
40%
25%
0%
57%
33%
0%
0%
50%
33%
20%
0%
50%
50%
20%
0%
63%
50%
20%
0%
33%
40%
0%
0%
33%
29%
0%
0%
27%
11%
17%
0%
17%
11%
17%
0%
25%
25%
17%
-20%
25%
25%
17%
0%
62%
44%
21%
-1%

94

Ingle
1.6 km 2% grade

0.8 km 2% grade
LOS B

LOS C

LOS D

LOS E

LOS B

10.0

9.0

9.0

8.0

8.0

7.0

7.0

6.0

6.0

P CE

P CE

10.0

5.0

LOS C

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0
6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

26.0

6.0

28.0

8.0

10.0

12.0

5% Trucks

LOS B

LOS C

16.0

5% Trucks

10% Trucks

18.0

20.0

22.0

24.0

26.0

28.0

10% Trucks

1.6km 4% grade

0.8 km 4% grade
10.0

14.0

Density (veh/km/ln)

Density (veh/km/ln)

LOS D

LOS E

LOS B

10.0

LOS C

LOS D

LOS E

9.0

9.0

8.0

8.0

7.0

7.0

6.0
PCE

6.0
PCE

LOS E

5.0

4.0

0.0

5.0

5.0

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

26.0

6.0

28.0

8.0

10.0

12.0

5% Trucks

LOS B

LOS C

16.0

5% Trucks

10% Trucks

18.0

20.0

22.0

24.0

26.0

28.0

10% Trucks

1.6 km 6% grade

0.8 km 6% grade
10.0

14.0

Density (veh/km/ln)

Density (veh/km/ln)

LOS D

LOS E

LOS B

10.0

9.0

9.0

8.0

8.0

7.0

7.0

6.0

6.0
P CE

P CE

LOS D

5.0

LOS C

LOS D

LOS E

5.0

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0
0.0

0.0
6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

28.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

24.0

26.0

28.0

Density (veh/km/ln)

Density (veh/km/ln)

5% Trucks

26.0

5% Trucks

10% Trucks

10% Trucks

Figure 17. PCE variability by level of congestion for N multiple truck population, 83.7 kg/kW
(137.5 lb/hp) and standard deviation of 25.12 kg/kW (41.25 lb/hp) at LOS C
congestion.

95

Ingle

from B to E. It can be observed from figure 17 that for 2% grades, the calculated PCE
does not vary by much. However, for steeper grades such as 4 or 6%, the calculated PCE
varies significantly by the average density value. Table 21 shows the percentage
difference of the PCE calculated for the different average density values compared to the
PCE calculated for an average density value of 12.4 pc/km/ln (20 pc/ml/ln). The greatest
difference occurs for an average density values within the LOS D regime. This means
that PCE is most critical at LOS D congestion.
5.12 Conclusions
Based on the results presented for multiple truck populations, the following
conclusions have been made. First, multiple truck populations do not perform
significantly different from single truck populations to produce different PCEs. The
assumption of a single truck population is therefore a valid simplification. Second, the
PCE of trucks remains relatively constant for grades longer than 1.6 km (1 mi), thus the
range of grade lengths provided in the HCM 2000 are adequate. Third, since the PCE of
trucks remains constant for a proportion of trucks greater than 60%, PCEs should be
provided for up to 60% proportion of trucks. Fourth, pavement type and condition
significantly effects the calculated PCE for a low proportion of trucks only; adjustments
for pavement type and condition should only be made for a low proportion of trucks.
Fifth, aerodynamic treatment significantly affects the calculated PCE for a low proportion
of trucks only; adjustments for aerodynamic treatment should only be made for a low
proportion of trucks. Sixth, the PCE for three lane segments is significantly lower than
the PCE for two lane segments only for a low proportion of trucks, thus an adjustment for
the number of lanes should be made only for a low proportion of trucks. In addition, lane

96

Ingle

Table 21. Percentage difference in calculated PCE for different levels of congestion, N multiple truck population, 83.7 kg/kW (137.5 lb/hp) and
standard deviation of 25.12 kg/kW (41.25 lb/hp).

Upgrade
(%)

Density
(veh/mi/ln)
(veh/km/ln)
% Trucks
Length
(km)

0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41
0.40
0.80
1.21
1.61
2.01
2.41

15
9.3

20
12.4

% Error From LOS C Congestion, 12.4 veh/km/lm (20 veh/mi/ln)

25
30
35
15.5
18.6
21.8
10
5
10
5
10
5

10

0%

0%

0%

0%

25%

33%

25%

33%

0%

-25%

0%

0%

25%

0%

25%

-17%

-20%

0%

0%

17%

20%

-25%

-17%

0%

0%

13%

-11%

-14%

0%

0%

-25%

-22%

0%

0%

40
24.9
10

10

25%

33%

0%

33%

0%

25%

0%

25%

0%

33%

20%

33%

20%

33%

20%

17%

25%

17%

25%

17%

13%

17%

22%

14%

33%

29%

33%

29%

33%

29%

8%

0%

17%

11%

17%

11%

8%

0%

Length
(mi)
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
0.25
0.50
0.75
1.00
1.25
1.50
average=

-13%

-16%

97

0%

0%

18%

14%

26%

18%

26%

18%

19%

16%

Ingle

restrictions have no impact on the calculated PCE. Seventh, a separate truck speed limit
below the speed limit of other vehicles causes a significant increase in the calculated
PCE. A scale factor may be applied to the PCEs calculated for trucks without a separate
speed limit in order to account for the effect of a truck speed limit. Eighth, the PCEs
provided in the HCM 2000 were calculated for LOS C congestion. A scale factor may be
applied to the PCEs to account for the effect of level of congestion.

98

Ingle

Chapter 6. Conclusions and Recommendations

6.1 Recommended PCE Table
Based on the results and conclusions presented in this research, a recommended
PCE table has been created as shown in table 22. This PCE table has incorporated the
recommendations presented in sections 6.2 and 6.3 with regard to the variables
investigated in this research. The PCEs in table 22 have been calculated for a single truck
population with an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp), with full
aerodynamic aids, fair asphalt pavement, and LOS C congestion. The PCEs listed in table
22 are very similar to the PCEs provided in the HCM 2000 and are in a compatible
format with the HCM 2000. It can be observed that grades up to 3.22 km (2 mi) in length
and grades up to 8% have been added to the PCE table. In addition, the proportion of
trucks has been expanded to include up to 60% proportion of trucks. The incorporation of
these criteria fulfills the recommendations presented in this research. The recommended
scale factor tables presented in section 6.3 may be applied to this table to account for the
effect of the variables found to significantly affect the PCE.
To illustrate the significance of this research in calibrating PCEs for an expanded
set of roadway and traffic characteristics, a set of example problems has been worked out.
The example problems are included in the appendix. Example problem 1 is a realistic
case study of I-81 in Virginia. It examines a 3.22 km (2 mi) grade of 4% on northbound I81 between milepost 126 and 128. The per lane traffic volume is assumed to be 1300
veh/h/ln, with a known 32% proportion of trucks. The traffic density at this traffic
volume corresponds to a LOS C congestion. This example makes use of the PCEs
provided in table 22. The PCE recommended in this research is more accurate than that

99

Ingle

Table 22. PCEs for basic freeway segments for trucks and buses on upgrades.
Upgrade
(%)

Length
(km)

Length
(mi)

0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22
0.40
0.80
1.21
1.61
2.01
2.41
3.22

0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00
0.25
0.50
0.75
1.00
1.25
1.50
2.00

2
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.0
2.5
3.0
3.0
3.0
3.5
4.0
2.5
3.0
3.5
4.0
4.5
4.5
5.0
3.5
3.5
5.5
5.5
6.5
6.5
7.5
3.5
4.5
7.5
7.0
8.5
8.5
9.0
5.0
8.0
8.0
9.0
9.5
10.0
10.5

4
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.0
2.5
2.5
2.5
2.5
3.0
3.0
2.0
2.5
2.5
3.5
3.5
4.0
4.0
2.5
3.5
4.0
4.5
5.0
5.0
5.0
3.0
4.5
5.5
6.0
6.5
7.0
6.5
5.0
7.5
7.5
8.0
8.0
8.0
9.5

5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
3.0
3.5
2.0
2.5
3.5
3.5
4.0
4.0
4.5
2.5
3.0
3.5
4.0
4.0
5.0
5.5
3.0
4.5
5.5
6.0
6.0
6.5
6.5
4.5
7.0
7.5
8.0
8.0
8.0
8.5

100

6
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
3.0
2.0
2.5
3.0
3.5
3.5
4.0
4.0
2.5
3.0
3.5
4.0
4.5
4.5
5.0
3.0
4.0
5.0
6.0
6.0
5.5
6.0
4.0
6.5
7.0
7.0
7.0
7.5
7.5

8
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
3.0
2.0
2.5
3.0
3.0
3.5
3.5
3.5
2.0
3.0
3.0
3.0
3.5
3.5
4.5
2.5
4.0
4.5
5.0
5.0
5.5
5.5
4.0
5.5
6.0
6.0
6.0
6.5
6.5

ET
Percentage of Trucks and Buses
10
15
20
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.0
2.5
2.5
2.0
2.0
2.0
2.0
2.5
2.5
2.0
3.0
2.5
2.5
3.0
3.0
2.5
3.0
3.0
2.5
3.0
3.0
2.5
3.5
3.0
2.5
2.5
2.0
2.0
2.5
2.5
2.0
3.0
3.0
2.5
3.0
3.0
3.0
3.5
3.0
3.0
3.5
3.5
3.0
4.0
3.5
3.0
2.5
2.5
2.5
3.5
3.0
3.0
4.0
3.5
3.5
4.5
3.5
3.5
4.5
4.0
3.5
5.0
4.0
3.5
5.0
4.5
3.5
3.5
3.5
3.0
5.0
4.5
4.0
5.5
4.5
4.0
5.5
5.0
4.5
6.0
5.0
4.5
6.0
5.0
4.5
6.0
5.0
4.5

25
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
2.0
2.0
2.5
3.0
3.0
3.0
3.0
2.0
2.5
3.0
3.0
3.5
3.5
3.5
3.0
3.5
4.0
4.0
4.0
4.0
4.5

30
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.0
2.5
2.5
2.5
2.5
3.0
3.0
2.0
2.5
3.0
3.0
3.0
3.0
3.5
2.5
3.5
3.5
4.0
4.0
4.0
4.0

40
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
2.5
2.0
2.5
2.5
3.0
3.0
3.0
3.0
2.5
3.0
3.5
3.5
3.5
3.5
3.5

50
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.0
2.5
2.5
2.5
2.5
2.5
2.5
2.5
3.0
3.0
3.0
3.0
3.0
3.0

60
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
2.0
2.0
2.0
2.0
2.0
1.5
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.5
2.5
2.5
2.5
2.5
2.0
2.5
3.0
3.0
3.0
3.0
3.0

Ingle

provided in the HCM 2000 because it is closer to the known proportion of trucks and it
has been calibrated for the correct length of grade. The best PCE provided in the HCM
2000 was for any grade greater than 1.6 km (1 mi) grade with 25% proportion of trucks.
The calculated base flow (passenger cars only) using the PCE provided in the HCM 2000
is 2132 pc/h/ln. Using the PCE recommended in this research results in a calculated base
flow that is 10% lower, at 1924 pc/h/ln.
6.2 Variables Not Found to Significantly Affect PCEs
A primary hypothesis of this research was that multiple truck populations would
not perform differently than a single truck population with the same average weight to
power ratio. This hypothesis was confirmed. The variables listed in the research
methodology were tested for both single and multiple truck populations. The
relationships for these variables were found to be the same for single truck populations as
compared to multiple truck populations. Based on these results, it is recommended that
the simplification of simulating a single truck population with an average weight to
power ratio can confidently be used.
The PCE was not found to be significantly affected by variability of engine power
within a single weight to power ratio truck population; percentage differences were less
than 10%. The PCEs for 10% proportion of trucks for the two lowest engine powers
simulated, the 10th and 22nd percentile engine powers, were significantly higher (13%)
than the PCEs for the average truck with the 75th percentile engine power. However, this
trend did not apply over the whole range of applicable proportion of trucks. In addition,
the PCE was not found to be significantly different for the 45th or 95th percentile engine
power, with percentage differences less than 10%. Based on this result, it has been

101

Ingle

concluded that variability of engine power within a single weight to power ratio is not an
important consideration for determining the PCE of trucks. No recommendations were
made to account for the effect of engine power.
Truck aerodynamic treatment on the whole did not significantly affect PCEs. The
PCEs for 10% proportion of trucks with partial and no aerodynamic aids were
significantly higher (14%) than the PCEs calculated for full aerodynamic aids. However,
this trend did not apply over the whole range of applicable proportion of trucks. There
were not significant differences in the calculated PCEs for either 5 or 25% proportion of
trucks. Since the trend is not wholly applicable, a recommendation to account for the
effect of truck aerodynamic treatment has not been made.
Similarly, the effect of three lane segments on the whole was not significant. The
PCEs for 5% proportion of trucks were considerably lower (9%) for three lane segments
in comparison to two lane segments. However, for a slightly higher proportion of trucks,
the differences were not considered significant. Since this trend is not wholly applicable
to a realistic range of proportion of trucks, a recommendation to account for the effect of
three lane segments has not been made. It was also found in the investigations that a truck
lane restriction, limiting trucks to using the two rightmost out of three lanes, did not
change the PCEs for three lane segments in any way.
6.3 Variables Found to Significantly Affect PCEs
Several of the variables hypothesized to significantly affect the PCE were found
to have a significant effect on the PCE; percentage differences exceeded 10%. Of the
variables tested in this research, weight to power ratio, length of grade, percent grade,
proportion of trucks, pavement type and condition, truck speed limit, and level of

102

Ingle

congestion were found to have a significant effect on the calculated PCE. The effect of
the length of grade and percent grade was well known at the onset of the research;
however, these trends were used to validate the calculated PCEs. Based on the results of
the research into the variability of PCE for each of these variables, a set of
recommendations has been established for each variable.
6.3.1 Recommendation to Account for the Effect of Weight to Power Ratio
Truck weight to power ratio was found to have a significant effect on the
calculated PCE, with percentage differences that exceeded 10%. A single or multiple
truck population with an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp) was
found to most closely match the PCEs provided in the HCM 2000. Table 23 presents
scale factors that may be applied to the PCE table 22 for the average weight to power
ratio truck population. These scale factors account for the effect of weight to power ratio
for a lighter or heavier than average weight to power ratio truck population. The light
weight to power ratio is 68.5 kg/kW (112.5 lb/hp) and results in PCEs that are as much as
22% lower than the PCEs for the average weight to power ratio truck population. The
heavy weight to power is 106.6 kg/kW (175 lb/hp) and results in PCEs that are as much
as 30% higher than the PCEs for the average weight to power ratio truck population. The
light weight to power ratio represents the 40th percentile and the heavy weight to power
ratio represents the 85th percentile from a survey of trucks along I-81 in Virginia.
Example problem 2 in the appendix illustrates the significance of the weight to
power ratio scale factor applied to a case study of I-81 in Virginia. The example problem
is the same as example 1 except the truck population is assumed to be heavier than
average, with a weight to power ratio approximately 106.6 kg/kW (175 lb/hp). The

103

Ingle

Table 23. PCE scale factors to account for the effect of weight to power ratio.

Upgrade (%)
2
> 2-4
> 4-6

Weight to Power Ratio Scale Factors

Light Truck 68.5 kg/kW (112.5 lb/hp)
Heavy Truck 106.6 kg/kW (175 lb/hp)
Percentage of Trucks and Buses
5
10
25
100
5
10
25
100
0.8
1.0
1.0
0.7
1.4
1.4
1.3
1.1
0.8
0.8
0.9
0.8
1.4
1.4
1.3
1.1
0.8
0.8
0.9
0.8
1.4
1.4
1.5
1.1

104

Ingle

calculated base flow using the PCE provided in the HCM 2000 is 2132 pc/h/ln. Using the
PCE recommended in this research and the scale factor from table 23 to account for the
heavier than average truck population results in a calculated base flow of 2444 pc/h/ln,
15% higher than the estimate from the HCM 2000.
6.3.2 Recommendation to Account for the Effect of Length of Grade
The relationship between the length of grade and the PCE was used to validate the
PCEs calculated in this research. In addition, it was observed that the PCE for a single
truck population continues to increases for lengths of grade up to 3.22 km (2 mi). Based
on this observation, it is recommended that PCEs be provided for grades up to 3.22 km (2
mi) in length. The PCE table 22 has incorporated this recommendation..
6.3.3 Recommendation to Account for the Effect of Percent Grade
The relationship between the percent grade and the PCE was also used to validate
the PCEs calculated in this research. It was observed that the PCE for a single truck
population continues to increases for grades of 8%. Based on this observation, it is
recommended that PCEs be provided for grades up to 8%. The PCE table 22 has
incorporated this recommendation.
6.3.4 Recommendation to Account for the Effect of Proportion of Trucks
The proportion of trucks was found to dramatically affect the calculated PCE. The
most variability of PCE by proportion of trucks occurs for a very low proportion of
trucks. The PCE decreases as the proportion of trucks increases. Beyond 60% proportion
of trucks, the PCE was found to be mostly constant across all lengths and percentage
grades. Based on these observations, it is recommended that PCEs be provided for up to
60% proportion of trucks. This is an extension of the proportion of trucks currently

105

Ingle

provided in the HCM 2000; however, expanding the proportion of trucks will make the
PCE table more applicable for freeways with a large amount of truck traffic. Many
freeways in the nation have a proportion of trucks that exceeds 25%. This
recommendation has been incorporated into the PCE table 22.
6.3.5 Recommendation to Account for the Effect of Pavement Type and Condition
The pavement type and condition was found to significantly affect the calculated
PCE. The PCEs for poor pavement types are as much as 14% higher than the PCEs for
the default (fair asphalt) pavement. Similarly, the PCEs for good or excellent pavement
types are as much as 15% lower than the PCEs for the default pavement. This constitutes
a capacity justification for maintenance, reconstruction, or rehabilitation projects that will
improve or maintain the condition of pavements. Poor pavements had PCEs that were on
average 9% higher than the default PCEs, and good pavements had PCEs that were on
average 2 to 5% lower than the default PCEs. Snow covered pavement had PCEs that
were constant regardless of the length or percent grade. An explanation for this
observation is that snow covered pavement resulted in reduced speeds for all vehicles
types, not primarily trucks as the other pavement types did. The PCEs for snow covered
pavement were as much as 267% higher than the default PCEs. However, in the case of
steep grades, the PCEs for snow covered pavement were lower than the default PCEs.
This was a result of the speed reduction that applied to all vehicles types. Table 24
presents scale factors that may be applied to the PCE table 22 for the average truck
population to account for the effect of pavement type and condition.
Example problem 3 in the appendix illustrates the significance of the pavement
type and condition scale factor applied to a case study of I-81 in Virginia. The example

106

Ingle

Table 24. PCE scale factors to account for the effect of pavement type and condition.
Upgrade
(%)
2
> 2-4
> 4-6

Asphalt Poor
5
1.0
1.2
1.1

10
1.3
1.1
1.1

Concrete Poor
25
1.0
1.0
1.1

5
1.0
1.2
1.0

10
1.3
1.1
1.1

25
1.0
1.0
1.0

107

Pavement Type and Condition Scale Factors

Asphalt Good
Concrete Good
Percentage of Trucks and Buses
5
10
25
5
10
25
0.8
1.0
1.0
0.8
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
0.9
1.0
1.0
0.9
1.0
1.0

Concrete Excellent
5
0.8
0.9
0.9

10
1.0
0.9
0.9

25
1.0
0.9
1.0

Snow Covered
5
2.8
1.9
1.1

10
2.3
1.3
0.9

25
1.3
0.9
0.7

Ingle

problem is the same as example 1 except the pavement condition is assumed to be poor
asphalt rather than fair asphalt which is the default type. The calculated base flow using
the PCE provided in the HCM 2000 is 2132 pc/h/ln. Using the PCE recommended in this
research and the scale factor from table 24 results in a calculated base flow that is 5%
lower, at 2028 pc/h/ln. Although the PCE scale factor to account for the detrimental
effect of poor pavement is greater than 1, the resulting base flow remains lower than that
estimated using the HCM 2000 method. This occurs because the basic PCE
recommended for this case is lower than the PCE provided in the HCM 2000. The impact
of pavement type and condition scale factors is more influential at a lower proportion of
trucks.
6.3.6 Recommendation to Account for the Effect of Truck Speed Limit
The institution of a truck speed limit was found to significantly affect the
calculated PCE. The PCEs for a truck speed limit of 88.5 km/h (55 mi/h) are as much as
100% higher than the PCEs for a speed limit of 112.6 km/h (70 mi/h) that applies to all
vehicles. Although the regulation of truck speeds has not been examined from a safety
perspective, this is evidence of a major capacity reduction due to truck speed regulation.
Table 25 presents scale factors that may be applied to the PCE table 22 for the average
truck population to account for the effect of a truck speed limit.
Example problem 4 in the appendix illustrates the significance of the speed limit
scale factors. The example problem is the same as example problem 1 except the truck
speed limit is assumed to be limited to 88.5km/h (55 mi/h). The calculated base flow
using the PCE provided in the HCM 2000 is 2132 pc/h/ln. Using the PCE recommended
in this research and the scale factor from table 25 results in a calculated base flow that is

108

Ingle

Table 25. PCE scale factors to account for the effect of a truck speed limit.
Upgrade
(%)
2
> 2-4
> 4-6

Truck Speed Limit Scale Factors

88.5 km/h (55 mi/h) Truck Speed Limit
Percentage of Trucks and Buses
5
10
25
1.8
1.8
1.3
1.7
1.4
1.1
1.2
1.2
1.1

109

Ingle

5% lower, at 2028 pc/h/ln. The same explanation from section 6.3.5 applies here; the
PCE recommended for the basic case is lower than the PCE provided in the HCM 2000.
The impact of speed limit scale factors is much more influential at a lower proportion of
trucks.
6.3.7 Recommendation to Account for the Effect of Level of Congestion
The level of congestion was found to significantly affect the calculated PCE;
percentage differences exceeded 10%. The PCEs calculated for an average density value
of 12.4 pc/km/ln (20 pc/ml/ln), corresponding to LOS C, most closely matched the PCEs
provided in the HCM 2000. However, the most critical level of congestion for calculating
the PCE is in the LOS D regime. Table 26 presents scale factors that may be applied to
the PCE table 22 for the average truck population to account for the effect of level of
congestion.
Example problem 5 in the appendix illustrates the significance of the level of
congestion scale factors. The example is the same as example problem 1 except the
traffic volume is assumed to be 1500 veh/h/ln, such that the level of congestion
corresponds to LOS D. The calculated base flow using the PCE provided in the HCM
2000 is 2460 pc/h/ln. Using the PCE recommended in this research and the scale factor
from table 26 results in a calculated base flow that is equal to that found using the HCM
2000 method. The level of congestion scale factor does increase the PCE; however, since
the basic recommended PCE is lower than the HCM 2000 PCE, the scale factor only
increases the recommended PCE to a value equal to the PCE provided in the HCM 2000.
Therefore, the resulting flows are equal.

110

Ingle

Table 26. PCE scale factors to account for the effect of level of congestion.

Level of Congestion
Density
LOS
(veh/km/ln) (veh/mi/ln)
B
9.3
15
C
12.4
20
C
15.5
25
D
18.6
30
D
21.8
35
E
24.9
40

PCE Scale
Factor
0.8
1.0
1.1
1.2
1.2
1.1

111

Ingle

6.3.8 Recommendation to Account for the Combined Effect of Variables

Scale factors presented in the previous sections can confidently be used to
individually account for the effect of different roadway and traffic characteristics. In
addition, when considering the combined effect of different roadway and traffic
characteristics, multiple scale factors may be used for low to moderately steep grades and
a low proportion of trucks. When applying multiple scale factors, an error in the
calculated base flow must be expected because the scale factors themselves do not
account for combined effects. For heavier than average truck populations with 10%
proportion of trucks operating on poor pavement and a level of congestion corresponding
to LOS D, the use of multiple scale factors should be limited to grades less than or equal
to 3.22 km (2 mi) and 4% or 0.40 km (0.25 mi) and 6%. For lighter than average truck
populations with 10% proportion of trucks operating on good or excellent pavement and
a level of congestion corresponding to LOS B, the use of multiple scale factors can be
applied to all grades. It is recommended that further research focus on defining the
applicable range where multiple scale factors can be used or other methods to quantify
the combined effect of roadway and traffic characteristics such as those investigated in
this research.
Example problem 6 in the appendix illustrates the significance of using multiple
scale factors to account for the combined effect of variables. The example problem is the
basic case from example problem 1, but it includes each of the modifications used in
example problems 2 through 5. The calculated base flow using the PCE provided in the
HCM 2000 is 2460 pc/h/ln. Using the PCE and multiple scale factors recommended in
results in a calculated base flow of 3634 pc/h/ln, which is 48% than the HCM 2000

112

Ingle

estimate. In this example problem, the combined effect of variables is overestimated by

using multiple scale factors. It is recommended that no more than a couple scale factors
be used in combination because the multiplicative effect of each may exaggerate the
result. The use of multiple scale factors to account for the combined effects of variables
should be approached with caution.
6.4 Summary of Conclusions and Recommendations
This research resulted in the calculation of PCEs for trucks for basic freeway
segments. The equal density method was successfully used to calculate PCEs that
replicated those found in the HCM 2000. The PCEs for a single or multiple truck
population with an average weight to power ratio of 83.7 kg/kW (137.5 lb/hp) most
closely matched the PCEs in the HCM 2000. It was found that PCEs calculated for a
multiple truck population were not much different from PCEs calculated for a single
truck population with the same average weight to power ratio. The PCEs provided in the
HCM 2000 were calculated for LOS C congestion, however a set of scale factors has
been recommend to account for different levels of congestion. Weight to power ratio was
found to significantly affect the calculated PCE; a set of scale factors has been
recommend to account for different weight to power ratios. Variability of engine power
within the weight to power ratio was not found to significantly affect the calculated
PCEs. It was recommended that the PCE table be expanded to include lengths of grade up
to 3.22 km (2 mi) and percent grades up to 8%. In addition, it was recommended to
expand the proportion of trucks to 60% proportion of trucks to make the PCE table more
applicable to freeways with large amounts of truck traffic. Pavement type and condition
were found to significantly affect the calculated PCE, and a set of scale factors has been

113

Ingle

recommended to account for different pavement types and conditions. Truck

aerodynamic treatment was not found to significantly affect the calculated PCE over the
whole range of applicable values. In a similar fashion, three lane segments were not
found to significantly affect the calculated PCEs in comparison to two lane segments. It
was found that truck lane restrictions for three lane segments had no effect on the
calculated PCEs. The institution of a truck speed limit at 88.5 km/h (55 mi/h) was found
to greatly increase the calculated PCEs; a set of scale factors has been recommended to
account for this effect. The application of scale factors to account for the combined
effects of the variables examined in this research should be applied carefully. An error
should be expected due to the combined effect of variables, especially for steep grades
with a heavier than average truck population.

114

Ingle

References
1. Highway Capacity Manual 2000. TRB, National Research Council, Washington,
DC., 2000.
2. Krammes, R., and Crowley, K. Passenger Car Equivalents for Trucks on Level
Freeway Segments. In Transportation Research Record 1091. TRB, National
Research Council, Washington, DC., 1986, pp. 10-17.
3. Elefteriadou, L., Torbic, D., and Webster, N. Development of Passenger Car
Equivalents for Freeways, Two-Lane Highways, and Arterials. In Transportation
Research Record 1572. TRB, National Research Council, Washington, DC.,
1997, pp. 51-58.
4. Cunagin, W., and Messer, C. Passenger Car Equivalents for Rural Highways. In
Transportation Research Record 905. TRB, National Research Council,
Washington, DC., 1983, pp. 61-68.
5. May, A. Traffic Flow Fundamentals. Prentice-Hall, Englewood Cliffs, New
Jersey, 1990, pp. 247-253.
6. Linzer, E., Roess, R., and McShane, W. Effect of Trucks, Buses, and
Recreational Vehicles on Freeway Capacity and Service Volume. In
Transportation Research Record 699. TRB, National Research Council,
Washington, DC., 1979, pp. 17-24.
7. St John, A., and Glauz, W. Speeds and Service on Multilane Upgrades. In
Transportation Research Record 615. TRB, National Research Council,
Washington, DC., 1976, pp. 4-9.
8. St John, A. Nonlinear Truck Factor for Two-Lane Highways. In Transportation
Research Record 615. TRB, National Research Council, Washington, DC., 1976,
pp. 49-53.
9. Hu, Y., and Johnson, R. Passenger Car Equivalents of Trucks in Composite
Traffic. Report DTFH-61-80-C-00127, FHWA, U.S. Department of
Transportation, 1981.
10. Huber, M. Estimation of Passenger Car Equivalents of Trucks in Traffic
Stream. In Transportation Research Record 869. TRB, National Research
Council, Washington, DC., 1982, pp. 60-70.
11. Sumner, R., Hill, D., and Shapiro, S. Segment Passenger Car Equivalent Values
for Cost Allocation on Urban Arterial Roads. In Transportation Research, Vol.
18A, No. 5/6, 1984, pp. 399-406.

115

Ingle

12. Van Aerde, M., and Yagar, S. Capacity, Speed, and Platooning Vehicle
Equivalents for Two-Lane Rural Highways. In Transportation Research Record
971. TRB, National Research Council, Washington, DC., 1984, pp. 58-67.
13. Roess, R., and Messer, C. Passenger Car Equivalents for Uninterrupted Flow:
Revision of Circular 212 Values. In Transportation Research Record 971. TRB,
National Research Council, Washington, DC., 1984, pp. 7-13.
14. Fan, H. Passenger Car Equivalents for Vehicles on Singapore Expressways. In
Transportation Research, Vol. 24A, No. 5, 1990, pp. 391-396.
15. Werner, A., and Morrall, J. Passenger Car Equivalencies of Trucks, Buses, and
Recreational Vehicles for Two-Lane Rural Highways. In Transportation
Research Record 615. TRB, National Research Council, Washington, DC., 1976,
pp. 10-17.
16. Cunagin, W., and Chang, C. Effects of Trucks on Freeway Vehicle Headways
Under Off-Peak Flow Conditions. In Transportation Research Record 869. TRB,
National Research Council, Washington, DC., 1982, pp. 54-59.
17. Seguin, E., Crowley, K., and Zweig, W. Passenger Car Equivalents on Urban
Freeways. Report DTFH61-80-C-00106, FHWA, U.S. Department of
Transportation, 1982.
18. Webster, N., and Elefteriadou, L. A Simulation Study of Truck Passenger Car
Equivalents (PCE) on Basic Freeway Sections. In Transportation Research, Vol.
33B, 1999, pp. 323-336.
19. Demarchi, S., and Setti, J. Limitations of PCE Derivation for Traffic Streams with
More Than One Truck Type. TRB 2003 Annual Meeting CD-ROM.
20. Ohio Department of Transportation. (11/17/2003). FHWA Vehicle Classification
Scheme F Report. [online], Available:
http://www.dot.state.oh.us/techservsite/availpro/Traffic_Survey/SchemeF/FHWA
_Scheme_F_Report.PDF
21. Federal Highway Administration. Office of Highway Policy Information.
(11/17/2003). FHWA Vehicle Types. [online], Available:
http://www.fhwa.dot.gov/policy/ohpi/vehclass.htm
22. Al-Kaisy, A., Hall, F., and Reisman, E. Developing Passenger Car Equivalents
for Heavy Vehicles on Freeways During Queue Discharge Flow. In
Transportation Research, Vol. 36A, 2002, pp. 725-742.
23. Rahka, H., and Lucic, I. Variable Power Vehicle Dynamics Model for
Estimating Truck Accelerations. 2002.

116

Ingle

24. Rakha, H, Lucic, I, Demarchi, S, Setti, J, and Van Aerde, M. Vehicle Dynamics
Model for Predicting Maximum Truck Acceleration Levels. In Journal of
Transportation Engineering, 2001, pp. 418-425.

117

Ingle

Appendix
Example Problem 1
Northbound I-81 in Virginia between milepost 126 and 128 is a 3.22 km (2.0 mi) grade of
4%. The per lane traffic volume is assumed to be 1300 veh/h/ln, with a known 32%
proportion of trucks. The traffic density at this traffic volume corresponds to a LOS C
congestion. The truck population is assumed to be equal to the average weight to power
ratio. The pavement type and condition is assumed to be fair asphalt. The speed limit for
both cars and trucks is 104.6 km/.h (65 mi/h).
The base flow is calculated as
q B = q M * (1 PT ) + q M * PT * ET * SF

[A1.1]

where qB is the base flow rate (passenger cars only), PT = 32% is the proportion of trucks,
qM = 1300 veh/h/ln is the mixed vehicle flow rate, ET is the PCE, and SF are any number
of scale factors recommended in chapter 6 (applied to the proposed method only).
a) HCM 2000 method
ET = 3.0 (for 25% trucks on a 1.6 km (1.0 mi), 4% grade)
q B = 1300 * (1 0.32) + 1300 * 0.32 * 3.0 = 2132 pc/h/ln

[A1.2]

b) Proposed method
ET = 2.5 (for 30% trucks on a 3.22 km (2.0 mi), 4% grade)
q B = 1300 * (1 0.32) + 1300 * 0.32 * 2.5 = 1924 pc/h/ln

[A1.3]

Using the PCE and scale factors proposed from this research results in a base flow that is
208 pc/h/ln or 10% lower than the estimate from the HCM 2000.

118

Ingle

Example Problem 2
Same situation as example problem 1 except the truck population is assumed to be
heavier than average, a weight to power ratio approximately 106.6 kg/kW (175 lb/hp).
a) HCM 2000 method
q B = 2132 pc/h/ln

b) Proposed method
ET = 2.5 (for 30% trucks on a 3.22 km (2.0 mi), 4% grade)
SF2 = 1.5 (adjustment for heavy truck)
q B = 1300 * (1 0.32) + 1300 * 0.32 * 2.5 * 1.5 = 2444 pc/h/ln

[A2.1]

Using the PCE and scale factor proposed from this research results in a base flow that is
312 pc/h/ln or 15% higher than the estimate from the HCM 2000.
Example Problem 3
Same situation as example problem 1 except the pavement type and condition is assumed
to be poor asphalt.
a) HCM 2000 method
q B = 2132 pc/h/ln

b) Proposed method
ET = 2.5 (for 30% trucks on a 3.22 km (2.0 mi), 4% grade)
SF3 = 1.1 (for poor asphalt pavement)
q B = 1300 * (1 0.32) + 1300 * 0.32 * 2.5 * 1.1 = 2028 pc/h/ln

[A3.1]

Using the PCE and scale factor proposed from this research results in a base flow that is
104 pc/h/ln or 5% lower than the estimate from the HCM 2000.

119

Ingle

Example Problem 4
Same situation as example problem 1 except the truck speed limit is assumed to be
88.5km/h (55 mi/h).
a) HCM 2000 method
q B = 2132 pc/h/ln

b) Proposed method
ET = 2.5 (for 30% trucks on a 3.22 km (2.0 mi), 4% grade)
SF4 = 1.1 (for truck speed limit at 88.5 km/h (55 mi/h))
q B = 1300 * (1 0.32) + 1300 * 0.32 * 2.5 * 1.1 = 2028 pc/h/ln

[A4.1]

Using the PCE and scale factor proposed from this research results in a base flow that is
104 pc/h/ln or 5% lower than the estimate from the HCM 2000.
Example Problem 5
Same situation as example problem 1 except the traffic volume is assumed to be 1500
veh/h/ln. The corresponding density is approximately 20 veh/km/ln (32 veh/mi/ln),
corresponding to LOS D.
a) HCM 2000 method
q B = 1500 * (1 0.32) + 1500 * 0.32 * 3.0 = 2460 pc/h/ln

[A5.1]

b) Proposed method
ET = 2.5 (for 30% trucks on a 3.22 km (2.0 mi), 4% grade)
SF5 = 1.2 (for level of congestion corresponding to 20 veh/km/ln (32 veh/mi/ln))
q B = 1500 * (1 0.32) + 1500 * 0.32 * 2.5 * 1.2 = 2460 pc/h/ln

[A5.2]

Using the PCE and scale factor proposed from this research results in a base flow that is
equal to the estimate from the HCM 2000.

120

Ingle

Example Problem 6
A combination of example problems 1 through 5. The basic grade criterion is used from
example problem 1, a 3.22 km (2.0 mi) grade of 4%. The truck population is assumed to
be heavier than average, although the proportion of trucks remains 32%. The pavement
type and condition is assumed to be poor asphalt. A truck speed limit is assumed to be in
effect at 88.5km/h (55 mi/h). The traffic volume is assumed to be 1500 veh/h/ln, with a
traffic density corresponding to LOS D.
a) HCM 2000 method
q B = 1500 * (1 0.32) + 1500 * 0.32 * 3.0 = 2460 pc/h/ln

[A6.1]

b) Proposed method
ET = 2.5 (for 30% trucks on a 3.22 km (2.0 mi), 4% grade)
SF2 = 1.5 (adjustment for heavy truck)
SF3 = 1.1 (for poor asphalt pavement)
SF4 = 1.1 (for truck speed limit at 88.5 km/h (55 mi/h))
SF5 = 1.2 (for level of congestion corresponding to 20 veh/km/ln (32 veh/mi/ln))
q B = 1500 * (1 0.32) + 1500 * 0.32 * 2.5 * 1.5 * 1.1 * 1.1 * 1.2 = 3634 pc/h/ln

[A6.2]

Using the PCE and multiple scale factors proposed from this research results in a base
flow that is 1174 pc/h/ln or 48% higher than the estimate from the HCM 2000.

121

Ingle

Vita
Anthony Ingle
EDUCATION

M.S. Civil Engineering, Anticipated July 2004

Virginia Polytechnic Institute and State University (Virginia Tech),
Blacksburg, VA
GPA: 3.95/4.0
Federal Highway Administration Dwight D. Eisenhower
Transportation Fellow
B.S. Civil Engineering, May 2003
Michigan State University, East Lansing, MI
GPA: 3.98/4.0
Study Abroad: International Practice of Civil Engineering in
Volgograd, Russia

PROFESSIONAL
EXPERIENCE

Undergraduate Research Assistant, Michigan State University, East

Lansing, MI
September 2002 to April 2003
Intern Transportation Engineer, Alfred Benesch and Company,
Lansing, MI
Summers 1999, 2000, 2001, and 2002

HONORS &
ACTIVITIES

Virginia Tech & Michigan State University Institute of Transportation

Engineers, American Society of Civil Engineers
Michigan State University College of Engineering Student Ambassador
Speaker for the Michigan State University Engineering Alumni Board
of Directors
Knights of Columbus service organization
Virginia Tech Via Department of Civil Engineering Masters
Fellowship
Michigan Associated General Contractors Scholarship
Michigan State University College of Engineering Scholarships
Michigan Society of Professional Engineers Scholarship
Michigan American Society of Civil Engineers Scholarship

122