Citation
Advanced Vehicle Dynamics Modeling Approach in Traffic Microsimulation with Emphasis on Commercial Truck Performance and On-Board-Diagnostics Data

Material Information

Title:
Advanced Vehicle Dynamics Modeling Approach in Traffic Microsimulation with Emphasis on Commercial Truck Performance and On-Board-Diagnostics Data
Creator:
Ozkul, Seckin
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (166 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Civil Engineering
Civil and Coastal Engineering
Committee Chair:
WASHBURN,SCOTT STUART
Committee Co-Chair:
ELEFTERIADOU,AGELIKI
Committee Members:
SRINIVASAN,SIVARAMAKRISHNAN
YIN,YAFENG
GEUNES,JOSEPH PATRICK
Graduation Date:
8/9/2014

Subjects

Subjects / Keywords:
Acceleration ( jstor )
Engine transmissions ( jstor )
Engines ( jstor )
Motor vehicle traffic ( jstor )
Railroad passenger cars ( jstor )
Simulations ( jstor )
Speed ( jstor )
Trailers ( jstor )
Trucks ( jstor )
Vehicle dynamics ( jstor )
Civil and Coastal Engineering -- Dissertations, Academic -- UF
commercial-truck-performance -- obd -- obd-data-modeling-in-traffic-microsimulation -- on-board-diagnostics-data -- vehicle-dynamics-modeling
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Civil Engineering thesis, Ph.D.

Notes

Abstract:
In the U.S., the primary guide for conducting highway capacity and level of service analyses from planning through design is the Highway Capacity Manual (HCM). Passenger car equivalency (PCE) values are used to convert the mix traffic flow rates into passenger car only flow rates so that the level-of-service (LOS) on a roadway can be calculated correctly. Transportation practitioners utilize these PCE values from the HCM 2010 to account for the effects of commercial trucks on traffic flow operations for freeway and highway facilities. The current PCE values in the HCM 2010 used for freeways and multilane highways are based on a study performed in the mid-1990s (Webster & Elefteriadou, 1999), which utilized CORSIM (version 5.0). This version of CORSIM and the current version (6.3) both utilize a maximum acceleration versus speed table along with other leading traffic simulation programs. Furthermore, all of these traffic simulation programs do not account for the transmission gear changing capabilities of commercial trucks, which results in the overestimation of vehicle deceleration on grades. Therefore, this study describes the development of an advanced vehicle dynamics modeling approach and its implementation into a custom traffic microsimulation tool in order to obtain more accurate maximum acceleration values. These values can have a significant impact in calculating commercial truck PCE values, assessing the traffic stream conditions and capacity on grade, and designing commercial truck acceleration lanes. In addition, PCE prediction equations as well as updated commercial truck speed versus distance-grade graphs were developed as a part of this study. The results increase the accuracy of the HCM 2010 PCE values as well as the commercial truck speed versus distance-grade graphs. Additionally, this study also describes the development and implementation of a method for predicting and outputting second-by-second (1 Hz) values of selected OBD parameters to simulate real-time OBD data that can be obtained from an actual vehicle. The results can be used to increase the fidelity of microsimulation in modeling air quality impacts and serve as a test bed for in-vehicle software applications that utilize OBD data. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2014.
Local:
Adviser: WASHBURN,SCOTT STUART.
Local:
Co-adviser: ELEFTERIADOU,AGELIKI.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2015-02-28
Statement of Responsibility:
by Seckin Ozkul.

Record Information

Source Institution:
UFRGP
Rights Management:
Copyright Ozkul, Seckin. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
2/28/2015
Resource Identifier:
969977004 ( OCLC )
Classification:
LD1780 2014 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

ADVANCED VEHICLE DYNAMICS MODELING APPROACH IN TRAFFIC MICRO SIMULATION WITH EMPHASIS ON COMMERCIAL TRUCK PERFORMANCE AND ON BOARD DIAGNOSTICS DATA By SECKIN OZKUL A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIV ERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2014

PAGE 2

© 2014 Seckin Ozkul

PAGE 3

To my late father Ali Ruhi Ozkul, my mother Cemile Ozkul , my wife Pavla Ozkul, and my children Ella Sofie Ozkul and Sebastian Bora Ozkul

PAGE 4

4 ACKNOWLEDG E MENTS I am forever in debt of my mother and my late father and for making sacrifices from their lives and supplying me with the tools necessary to succeed in l ife. Additionally , I would like to thank my wife, Pavla Ozkul, for always being there for me and bringing to this world the two biggest joys of my life, Ella Sofie Ozkul and Sebastian Bora Ozkul. I also would like to thank my cousin Hande Capaner for sup porting me when I needed it most. In addition, I would like to extend my gratitude to my friends Yexi Guo, Evangelos Mintsis, Ben Reibach, Alexandra Kondyli , Roosbeh Nowrouzian and Don Watson for hosting me during my stays in Gainesville so that I can kee p on taking classes and doing my research without having to drive back and forth between Tampa and Gainesville. Finally, I would like to thank my advisor Dr. Scott Washburn for his support and guidance throughout my Ph.D. studies as well as the opportunity he has given me by including me in his research group .

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGEMENTS ................................ ................................ ............................... 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 14 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 17 Problem Statement ................................ ................................ ................................ . 20 Research Objective and Tasks ................................ ................................ ............... 22 2 LITERATURE REVIEW ................................ ................................ .......................... 24 Vehicle Dynamics Modeling ................................ ................................ .................... 24 PCE Calculation Methodologies ................................ ................................ ............. 31 Commercial Truck Speed Versus Distance Grade ................................ ................. 37 On Board Diagnostics Data Integration into Traffic Microsimulation ....................... 41 3 COMMERCIAL TRUCK PERFORMANCE ................................ ............................. 43 Methodology ................................ ................................ ................................ ........... 43 Commercial Truck Classification and AADT Data ................................ ............ 43 Commercial Truck Characteri stics Data ................................ ........................... 47 Custom Simulation Platform ................................ ................................ ............. 49 Improved Commercial Truck Acceleration Versus Speed Curves .................... 52 TruckSim ® ................................ ................................ ................................ ........ 54 Advanced/Full Vehicle Dynamics Modeling Approach ................................ ..... 59 Updated PCE Calculations U sing Full Vehicle Dynamics Modeling Approach ................................ ................................ ................................ ....... 66 Updated Commercial Truck Speed Versus Distance Grade Curves ................ 68 Results ................................ ................................ ................................ .................... 68 Advanced Vehicle Dynamics Modeling Approach ................................ ............ 68 PCE Estimation Equations ................................ ................................ ............... 69 Updated Commercial Truck Speed Versus Distance Grade Curves ................ 76 4 ON BOARD DIAGNOSTICS DATA ................................ ................................ ........ 84 Methodology ................................ ................................ ................................ ........... 85

PAGE 6

6 Results ................................ ................................ ................................ .................... 95 5 SUMMARY AND RECOMMENDATIONS ................................ ............................. 103 APPENDIX A WIM STATION SITE DATA ................................ ................................ .................. 108 B CUSTOM WIM STATION DATA PROCESSOR CONTROL WINDOWS .............. 132 C SAMPLE COMMERCIAL TRUCK CHARACTERISTICS RESULTS .................... 134 D TRUCK MAXIMUM ACCELERATION VERSUS SPEED CURVES ...................... 138 E COMMERCIAL TRUCK PERFORMANCE COMPARISON CURVES SWASHSIM WITH RAKHA MODEL VERSUS TRUCKSIM ® ................................ 146 F ADVANCED VEHICLE DYNAMICS APPROACH TO COMMERCIAL TRUCK MAXIMUM ACCELERATION MODELING IN THE CUSTOM SIMULATION TOOL EXAMPLE CALCULATION ................................ ................................ ........ 151 G TRANSMISSION GEAR CHANGING CAPABLE COMMERCIAL TRUCK PERFORMANCE COMPARISON CURVES SWASHSIM WITH FULL VEHICLE DYNAMICS MODEL VERSUS TRUCKSIM ® ................................ ........................ 153 LIST OF REFERENCES ................................ ................................ ............................. 159 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 166

PAGE 7

7 LIST OF TABLES Table page 3 1 Overall Truck Classification History as % of Truck AADT ................................ ... 45 3 2 Urban/Freeway Truck Classification History as % of Truck AADT ...................... 45 3 3 Urban/Multilane Highway Truck Classification Hist ory as % of Truck AADT ...... 45 3 4 Rural/Freeway Truck Classification History as % of Truck AADT ....................... 46 3 5 Rural/Multilane Highway Truck C lassification History as % of Truck AADT ....... 46 3 6 Heavy Vehicle Fleet Common Engine Characteristics (Power and Torque) ....... 47 3 7 Vehicl e Characteristics Data ................................ ................................ ............... 52 3 8 Variables used for Experimental Design in Simulation Runs .............................. 67 3 9 PCE Comparison by Terrain Type ................................ ................................ ...... 75 3 10 Terrain Type Specific Input Values ................................ ................................ ..... 75 4 1 2003 Honda Civic LX Vehicle Characteristics Data ................................ ............ 93 A 1 WIM Station 57 0291 (Okaloosa) Site Data ................................ ...................... 108 A 2 WIM Station 57 0291 (Okaloosa) Site Data per Truck Class ............................ 108 A 3 WIM Station 54 9901 (Jefferson) Site Data ................................ ...................... 109 A 4 WIM Station 54 9901 (Jefferson) Site Data per Truck Class ............................ 109 A 5 WIM Station 26 9904 (Alachua) Site Data ................................ ........................ 110 A 6 WIM Station 26 9904 (Alachua) Site Data per Truck Class .............................. 110 A 7 WIM Station 7 2 9905 (Duval) Site Data ................................ ............................ 111 A 8 WIM Station 72 9905 (Duval) Site Data per Truck Class ................................ .. 111 A 9 WIM Station 79 9906 (Volusia) Site Data ................................ ......................... 112 A 10 WIM Station 79 9906 (Volusia) Site Data per Truck Class ............................... 112 A 11 WIM Station 46 9907 (Bay) Site Data ................................ ............................... 113 A 12 WIM Station 46 9907 (Bay)Site Data per Truck Class ................................ ...... 113

PAGE 8

8 A 13 WIM Station 34 9909 (Levy) Site Data ................................ ............................. 114 A 14 WIM Station 34 9909 (Levy) Site Data per Truck Class ................................ ... 114 A 15 WIM Station 97 9913 (Turnpike) Site Data ................................ ....................... 115 A 16 WIM Station 97 9913 (Turnpike) Site Data per Truck Class ............................. 115 A 17 WIM Station 72 9914 (Duval) Site Data ................................ ............................ 116 A 18 WI M Station 72 9914 (Duval) Site Data per Truck Class ................................ .. 116 A 19 WIM Station 48 9916 (Escambia) Site Data ................................ ..................... 117 A 20 WIM Station 48 9916 (Escambia)Site Data per Truck Class ............................ 117 A 21 WIM Station 70 9919 (Brevard) Site Data ................................ ........................ 118 A 22 WIM Station 70 9919 (Brevard) Site Data per Truck Class .............................. 118 A 23 WIM Station 72 9923 (Duval) Site Data ................................ ............................ 119 A 24 WIM Station 72 9923 (Duval) Site Data per Truck Class ................................ .. 119 A 25 WIM Station 10 9926 (Hillsborough) Site Data ................................ ................. 120 A 26 WIM Station 10 9926 (Hillsborough) Site Data per Truck Class ....................... 120 A 27 WIM Station 16 9927 (Polk) Site Data ................................ .............................. 121 A 28 WIM Station 16 9927 (Polk) Site Data per Truck Class ................................ .... 121 A 29 WIM Station 79 9929 (Volusia) Site Data ................................ ......................... 122 A 30 WIM Station 79 9929 (Volusia) Site Data per Truck Class ............................... 122 A 31 WIM Station 97 9931 (Turnpike) Site Data ................................ ....................... 123 A 32 WIM Station 97 9931 (Turnpike) Site Data per Truck Class ............................. 123 A 33 WIM Station 97 9933 (Turnpike) Site Data ................................ ....................... 124 A 34 WIM Station 97 9933 (Turnpike) Site Data per Truck Class ............................. 124 A 35 WIM Station 97 9934 (Turnpike) Site Data ................................ ....................... 125 A 36 WIM Station 97 9934 (Turnpike) Site Data per Truck Class ............................. 125 A 37 WIM Station 2 9 9936 (Columbia) Site Data ................................ ...................... 126

PAGE 9

9 A 38 WIM Station 29 9936 (Columbia) Site Data per Truck Class ............................ 126 A 39 WIM Station 58 9937 (Santa Rosa) Site Data ................................ .................. 127 A 40 WIM Station 58 9937 (Santa Rosa) Site Data per Truck Class ........................ 127 A 41 WIM Station 50 9940 (Gadsden) Site Da ta ................................ ...................... 128 A 42 WIM Station 50 9940 (Gadsden) Site Data per Truck Class ............................ 128 A 43 WIM Station 87 9947 (Miami Dade) Site Data ................................ ................. 129 A 44 WIM Station 87 9947 (Miami Dade) Site Data per Truck Class ....................... 129 A 45 WIM Station 16 9948 (Polk) Site Data ................................ .............................. 130 A 46 WIM Station 16 9948 (Polk) Site Data per Truck Class ................................ .... 130 A 47 WIM Station 48 9949 (Escambia) Site Data ................................ ..................... 131 A 48 WIM Station 48 9949 (Escambia) Site Data per Truck Class ........................... 131 C 1 Commercial Truck Characteristics Results Year 2008 Urban Areas .......... 134

PAGE 10

10 LIST OF FIGURES Figure page 2 1 Illustration of Flow Density Relationship for PCE Calculation Methodology per Sumner et al. (1984) Figure reproduced from Webster and Elefteriad ou (1999). ................................ ................................ ................................ ................ 32 2 2 Webster and Elefteriadou Experimental Roadway Configuration ....................... 33 2 3 HCM 2010 Exhibit 11 A1 ................................ ................................ ................. 37 2 4 Effect of Length and Grade on Speed of Average Commercial Trucks on Multilane Highways ................................ ................................ ............................. 38 2 5 Effect of Length and Grade on the Speed of a 190 kg/KW Commercial Truck ... 40 3 1 FHWA Vehicle Classification Chart (TXDOT, 2013) ................................ ........... 44 3 2 TruckSim ® Run Control Window (TruckSim ® by Me chanical Simulation Corporation, http://carsim.com/products/trucksim/index.php, License obtained by UF) ................................ ................................ ................................ .. 55 3 3 TruckSim ® Vehicle Attributes Window (TruckSim ® by Mechanical Simulation Corporatio n, http://carsim.com/products/trucksim/index.php, License obtained by UF) ................................ ................................ ................................ .. 56 3 4 TruckSim ® Roadway Geometry Window (TruckSim ® by Mechanical Simulation Corporation, http://carsim.com/products/tr ucksim/index.php, License obtained by UF) ................................ ................................ ..................... 57 3 5 TruckSim ® Plot Outputs Window (TruckSim ® by Mechanical Simulation Corporation, http://carsim.com/products/trucksim/index.php, License obtained b y UF) ................................ ................................ ................................ .. 58 3 6 Torque Engine Speed Curve for a Paccar PX 7 Engine ................................ ..... 64 3 7 Torque Engine Speed Curve for a Paccar MX 13 Engine ................................ .. 64 3 8 Roadway Configuration used for Experimental Design Simulation Runs ........... 67 3 9 Commercial Truck Speed Versus Distance Grade Curves Up grade ................ 76 3 10 Commercial Truck Speed Versus Distance Grade Curves Downgrade ........... 77 3 11 Example Problem 1 Solution Using Composite Grade Procedure ................... 79 3 12 Example Problem 2 Solution Using Composite Grade Procedure ................... 80

PAGE 11

11 3 13 Example Problem 3 Upgrade Solution Using Composite Grade Procedure .... 82 3 14 Example Problem 3 Downgrade Solution Using Composite Grade Procedure ................................ ................................ ................................ ........... 83 4 1 Photo of an OBD II cable attached to the OBD II port of a vehicle Photo by Seckin Ozkul ................................ ................................ ................................ ....... 84 4 2 Aerial Photo of Freeway Segment for OBD Data Collection ............................... 87 4 3 Photo of the field data measurement vehicle 2003 Honda Civic LX Photo by Seckin Ozkul ................................ ................................ ................................ .. 88 4 4 Photo of the OBDLink SX © Scan Tool Photo by Seckin Ozkul ........................ 89 4 5 Screen capture of the OBDWiz © Diagnostics Software (OBDWiz © by OCTech, LLC, http://www.obdsoftware.net/OBDwiz.aspx, license obtained through purchase of OBDLink SX © Scan Tool cable) ................................ ......... 90 4 6 Aerial Photo of Arterial Segment for Second Set of OBD Data Collection .......... 91 4 7 Torque/Power Engine Speed Curves for 2003 Honda Civic LX (E Trailer , 2014) ................................ ................................ ................................ .................. 92 4 8 MAP*RPM versus VSP Field Data ................................ ................................ ..... 95 4 9 MAP*RPM versus Calculated Engine Load Field Data ................................ ....... 96 4 10 Newberry Road OBD Field Data MAP*RPM versus VSP ................................ ... 98 4 11 Newberry Road OBD Field Data MAP*RPM versus Calculated Engine Load .... 98 4 12 Comparison of Field and SwashSim MAP*RPM versus VSP Data .................... 99 4 13 Comparison of Field and SwashSim MAP*RPM versus Calculated Engine Load Data ................................ ................................ ................................ ......... 100 4 14 Comparison of Validation Data and SwashSim Engine Speed versus Vehicle Speed ................................ ................................ ................................ ............... 101 4 15 Comparison of Validation Data and SwashSim E ngine Speed, Selected Transmission Gear and Vehicle Speed ................................ ............................ 102 B 1 Control Window for Data File Manipulation ................................ ...................... 132 B 2 Control Window for Data Processor ................................ ................................ . 133 D 1 Maximum Acceleration vs. Speed Curve Single Unit Truck on a Level Grade ................................ ................................ ................................ ............... 138

PAGE 12

12 D 2 Maximum Acceleration vs. Speed Curve Class 8 Truck on a Level Grade .... 139 D 3 Maximum Acceleration vs. Speed Curve Class 9 Truck on a Level Grade .... 140 D 4 Maximum Acceleration vs. Speed Curve Class 11&12 Truck on a Level Grade ................................ ................................ ................................ ............... 141 D 5 Maximum Acceleration vs. Speed Curve Singe Unit Truck on a 5% Grade .. 142 D 6 Maximum Acceleration vs. Speed Curve Class 8 Truck on a 5% Grade ....... 143 D 7 Maximum Acceleration vs. Speed Curve Class 9 Truck on a 5% Grade ....... 144 D 8 Maximum Acceleration vs. Speed Curve Class 11&12 Truck on a 5% Grade ................................ ................................ ................................ ............... 145 E 1 Maximum Acceleration of a Single Unit Truck on a 1320 foot Link with 9% Grade ................................ ................................ ................................ ............... 146 E 2 Velocity of a Single Unit Truck on a 1320 foot Link with 9% Grade .................. 146 E 3 Maximum Acceleration of an In termediate Semi trailer on a 1760 foot Link with 6% Grade ................................ ................................ ................................ .. 147 E 4 Velocity of an Intermediate Semi trailer on a 1760 foot Link with 6% Grade .... 147 E 5 Maximum Acceleration of an Intermediate Semi trailer on a 1760 foot Link with 9% Grade ................................ ................................ ................................ .. 148 E 6 Velocity of an Intermediate Semi trailer on a 1760 foot Link with 9% Grade .... 148 E 7 Maximum Acceleration of an Interstate Semi trailer on a 2640 foot Link with 6% Grade ................................ ................................ ................................ ......... 149 E 8 Velocity of an Interstate Semi tr ailer on a 2640 foot Link with 6% Grade ......... 149 E 9 Maximum Acceleration of a Semi tractor+double trailer on a 1760 foot Link with 6% Grade ................................ ................................ ................................ .. 150 E 10 Velocity of a Semi tractor+double trailer on a 1760 foot Link with 6% Grade ... 150 F 1 Advanced Vehicle Dynamics Approach Example Calculations 1 ................... 151 F 2 Advanced Vehicle Dynamics Approach Example Calculations 2 ................... 152 G 1 Transmission Gear Change Capable Maximum Acceleration of a Single Unit Truck on a 1320 f oot Link with 6% Grade ................................ ........................ 153

PAGE 13

13 G 2 Transmission Gear Change Capable Velocity of a Single Unit Truck on a 1320 foot Link with 6% Grade ................................ ................................ .......... 153 G 3 Transmission Gear Change Capable Maximum Acceleration of a Single Unit Truck on a 1320 foot Link with 9% Grade ................................ ........................ 154 G 4 Transmission Gear Change Capable Velocity of a Single Unit Truck o n a 1320 foot Link with 9% Grade ................................ ................................ .......... 154 G 5 Transmission Gear Change Capable Maximum Acceleration of an Intermediate Semi trailer on a 1760 foot Link with 6% Grade .......................... 155 G 6 Transmission Gear Change Capable Velocity of an Intermediate Semi trailer on a 1760 foot Link with 6% Grade ................................ ................................ .. 155 G 7 Transmission Gear Change Capable Maxim um Acceleration of an Intermediate Semi trailer on a 1760 foot Link with 9% Grade .......................... 156 G 8 Transmission Gear Change Capable Velocity of an Intermediate Semi trailer on a 1760 foot Link with 9% G rade ................................ ................................ .. 156 G 9 Transmission Gear Change Capable Maximum Acceleration of an Interstate Semi trailer on a 2640 foot Link with 6% Grade ................................ ............... 157 G 10 Transmission Gear Change Capable Velocity of an Interstate Semi trailer on a 2640 foot Link with 6% Grade ................................ ................................ ....... 157 G 11 Transmission Gear Change Capable Maximum Acceleration of a Semi trac tor+double trailer on a 1760 foot Link with 6% Grade ................................ 158 G 12 Transmission Gear Change Capable Velocity of a Semi tractor+double trailer on a 1760 foot Link with 6% Grade ................................ ................................ .. 158

PAGE 14

14 LIST OF ABBREVIATIONS HCM Highway Capacity Manual IOV Internally Observable Variable MAP Manifold Absolute Pressure OBD On Board Diagnostics PCE Passenger Car Equivalen t RPM Revolutions per Minute VSP Vehicle Specific Power WIM Weigh in Motion

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ADVANCED VEHICLE DYNAMICS MODELING APPROACH IN TRAF FIC MICRO SIMULATION WITH EMPHASIS ON COMMERCIAL TRUCK PERFORMANCE AND ON BOARD DIAGNOSTICS DATA By Seckin Ozkul August 2014 Chair: Scott S. Washburn Major: Civil Engineering In the U.S. , the primary guide for conducting highway capacity and level of se rvice analyses from planning through design is the Highway Capacity Manual (HCM) . Passenger car equivalency ( PCE ) values are used to convert the mix traffic flow rates into passenger car only flow rates so that the level of service (LOS) on a roadway can be calculated correctly. Transportation practitioners utilize the se PCE values from the HCM 2010 to account for the effects of commercial trucks on traffic flow operations for freeway and highway facilities. The current PCE values in the HCM 2010 used for freeways and multilane highways are based on a study performed in the mid 1990s (Webster & Elefteriadou , 1999), which utilized CORSIM (version 5.0). This version of CORSIM and the current version (6.3) both utilize a maximum acceleration versus speed tab le along with other leading traffic simulation programs. Furthermore, all of these traffic simulation programs do not account for the transmission gear changing capabilities of commercial trucks, which results in the overestimation of vehicle deceleration on grades.

PAGE 16

16 Therefore, t his study describes the development of an advanced vehicle dynamics model ing approach and its implementation into a custom traffic microsimulation tool in order to obtain more accurate maximum acceleration value s. These values can have a significant impact in calculating commercial truck PCE values, assessing the traffic stream conditions and capacity on grade, and designing commercial truck acceleration lanes. In addition, PCE prediction equations as well as updated c ommercial tru ck speed versus distance grade graphs were developed as a part of this study . The results increase the accuracy of the HCM 2010 PCE values as well as the commercial truck speed versus distance grade graphs . Additionally, this study also describes the devel opment and implementation of a method for predicting and outputting second by second (1 Hz) values of selected OBD parameters to simulate real time OBD data that can be obtained from an actual vehicle. The results can be used to increase the fidelity of m icrosimulation in modeling air quality impacts and serve as a test bed for in vehicle software applications that utilize OBD data.

PAGE 17

17 CHAPTER 1 INTRODUCTION The American economy relies heavily on the movement of goods across the nation every day. In the yea r 2009, the total weight of shipments carried was 16,122 (millions of tons) and 10,868 (millions of tons) of this amount was carried by trucks ( Schmitt et al., 2010 ) . As impressive as these numbers already are, it is estimated that in the year 2040, the t otal weight of shipments carried would be 27,104 (millions of tons) and 18,445 (millions of tons) of these shipments would be made by truck ( Schmitt et al., 2010 ) . Currently, long haul truck traffic in the United States is concentrated on major routes con necting population centers, ports, border crossings and other major hubs of activity, and is mainly on interstate highways. The Freight Analysis Framework (FAF) estimates that by 2040 long haul freight truck traffic will increase dramatically on interstat e highways and other arterials throughout the United States ( Schmitt et al., 2010 ) . In addition , commercial trucks have a big influence on traffic operations due to their much larger size and heavier weight when compared to passenger cars. Therefore, the re is a need to be able to accurately quantify the impacts of these commercial trucks on traffic operations. The Highway Capacity Manual ( HCM ) is considered to be the primary guide for conducting highway capacity and level of service analyses from planning through design i n the United States of America . The current version of this manual is the HCM 2010 . The HCM 2010 analysis methodologies are based on traffic stream units of passenger cars , and passenger car equivalency ( PCE ) values are used to convert th e actual traffic stream units of vehicles to passenger cars. The PCE of a specific

PAGE 18

18 commercial truck is an estimate of how this commercial truck i s affecting the traffic stream in terms of passenger cars . As an example, if the PCE value of a commercial tr uck for a given traffic condition was 2, the effect it has on the traffic stream equals to the effect of 2 passenger cars for the same traffic conditions . Therefore, the commercial truck PCE values from the HCM 2010 are utilized to account for the effects of commercial trucks on traffic flow operations on freeway and multilane highway facilities (Ozkul et al., 2013). The current PCE values in the HCM 2010 used for freeways and multilane highways are based on a study performed in the mid 1990s (Webster & E lefteriadou , 1999) , and were based on simulation results from a now outdated traffic microsimulation tool. Related to the commercial truck PCEs, another tool that the HCM 2010 supplies for practitioners is truck speed versus distance grade curves. These c urves are travels on a known distance of roadway with a certain grade. Most commonly , these curves are used in order to determine a n equivalent grade that will result in the same final speed of commercial trucks as would the series of grades making up the composite grade. The speed versus distance grade curves are collectively presented in the HCM as an analysis chart that is composed of multiple speed versus length cur ves for differing grade values. This chart was generated for a typical commercial truck with a weight to horsepower ratio of 200 lb/hp and a maximum heavy vehicle speed o f 55 mi/h for entering a grade or 60 mi/h for accelerating on a grade.

PAGE 19

19 Currently ther e is an increased importance of vehicle emission and fuel rate estimations in the transportation field. Historically, the air quality modeling community has been using vehicle emissions and fuel rate estimations models that are based on vehicle specific p ower (VSP) , which is a function of vehicle acceleration, vehicle speed and grade of the roadway. However, as of late , there is a movement towards using fine grained vehicle operating data for such modeling purposes (i.e., Motor Vehicle Emissions Simulator (MOVES)). However, with microsimulation already providing these data, it is beneficial to be able to do both a traffic operations and air quality analysis in the same tool. Frey et al. ( 2010 ) indicate that the usage of direct engine data, such as revolut ions per minute (RPM) and manifold absolute pressure (MAP) , which can be n B oard D iagnostics (OBD) data port, leads to more accurate estimation of vehicle emissions and fuel rate than the models based only on VSP . This high er level of accuracy is due to the fact that the vehicle emissions and fuel rate estimation models that are developed by using the OBD data are based on more vehicle/engine param e ters compared to the VSP model s that are used to estimate vehicle emissions a nd fuel rate. The incorporation of OBD systems in vehicle s started in the early 1980s. Currently, the second generation of OBD systems (OBD II) are in effect for all vehicles that were produced beginning with 1996 model year vehicles . OBD systems allow t he status of different components of the vehicle to be monitored . This system status check can be performed by vehicle users or mechanics and only requires the usage of an OBD scan tool to a device with the corresponding OBD software ( SAE (Society of

PAGE 20

20 Auto motive Engineers ) J1978 , 1997; ISO ( International Organization for Standardization) , 1994 ; Draft ISO , 1995; SAEJ2178/3 , 1997; SAEJ2178/4 , 1997; SAE1979 , 1997; SAEJ1978 , 1997; SAEJ2190 , 1997). Also , the OBD port broadcasts real time data for numerous param eters , such as vehicle speed, engine speed ( revs /min also referred to as RPM ), manifold absolute pressure (MAP), engine torque, engine power, selected transmission gear , calculated engine load and many others. Problem Statement T he PCE values developed in th e Webster and Elefteriadou (1999) study were based strictly on simulation and a now outdated version of the CORSIM traffic microsimulation program ( version 5.0) . This version of CORSIM and the current version (6.3) of CORSIM both utilize a maximum accel eration versus speed table. In CORSIM 6.3 this table is stored under Record Type (RT) 173. The usage of this table is basically a very simplistic method of determining maximum commercial truck acceleration values (a lookup table with maximum acceleration values based simply on velocity ). Furthermore, the relationship between these table values and the grade adjustment factor is not accurate due to th is approach not being as mathematical as a vehicle dynamics approach . This makes it difficult to ensure t hat effect of grade on maximum acceleration is being properly accounted for. It should be noted that, the simplified maximum acceleration calculation methodology (maximum acceleration look up table) used in most commercially available traffic simulation to ols may have been a necessary compromise in the past due to software architecture and/or computational speed limitations , and possibly just to minimize the input data requirements. However, these issues are of much less concern climate.

PAGE 21

21 Besides CORSIM , other leading traffic simulation programs such as VISSIM, AIMSUN and Paramics also use a similar look up table approach to determining a . Another simulation tool, INTEGRATION, although much less used than the one s previously mentioned, uses an approach more advanced than a look up table for maximum acceleration calculations , but it is still more simplified . Furthermore , all of these traffic simulation programs , as well as CORSIM 6.3 , do not account for the transmission gear changing capabilities of commercial trucks, which generally results in the overestimation of vehicle slowing on grades . In addition, drivetrain performance has improved over the last 15 20 y ears; thus, power to weight ratios have also improved . Combined , these assumptions and shortcomings of the traffic simulation tools , along with a lower power to weight ratio having been used for commercial trucks, result in PCE values in the HCM 2010 to b e larger than what they really should be . This directly translates in to the traffic stream speeds and/or capacity values to be lower than what they really should be . Therefore, there is a n apparent need for the development of an advanced/full vehicle dyn amics modeling approach that is integrated into traffic microsimulation, in which maximum acceleration is calculated for each time step while the gear changing capability of vehicles are accounted for. The aforementioned changes in truck performance over t he last 20 years also necessitate an update to the HCM 2010 speed versus distance grade graphs. Additionally, t he estimation models for vehicle emissions and fuel rate are currently based on models of VSP, which are a function of vehicle acceleration, vehi cle velocity and roadway grade. Recent studies (Frey et al., 2010) showed that t he

PAGE 22

22 incorporation of more detailed vehicle operating/engine data will result in more accurate estimation models of vehicle emissions and fuel rate to be developed. However, the re currently is a limitation that is caused by traffic simulation tools that do not provide detailed engine parameters. These traffic simulation tools cannot support more detailed vehicle emissions/fuel estimations. Therefore, there is a need for the dev elopment and implementation of a method for predicting second by second (1 Hz) values of selected OBD parameters to simulate real time OBD data that can be obtained from an actual vehicle so that this new simulation tool can support the future integration of more accurate vehicle emissions and fuel rate estimations models. Research Objective and Tasks The objectives of this study were 1) to develop an advanced/full vehicle dynamics model ing approach that is integrated into traffic microsimulation, 2) the de termination of updated PCE estimation equations based on a simulation tool that uses the full vehicle dynamics model ing approach introduced in this dissertation , 3) the development of updated commercial truck speed versus distance grade graphs , and finally 4) the development and implementation of a method for predicting OBD parameters in traffic microsimulation . The tasks that were undertaken to support the completion of these objectives are as follows: Comprehensive literature review on the research object ives listed in this section. Develop ing components for the custom traffic micro simulation tool to support the implementation of the full vehicle dynamics model ing approach for determining maximum acceleration . Determining the most appropriate c ommercial t r uck c lassification s to accurately represent the current commercial truck fleet in the custom traffic microsimulation tool

PAGE 23

23 as well as WIM data analysis to be incorporated into SwashSim . Developing the fu ll vehicle dynamics model ing approach so that maximum acceleration for commercial trucks can be calculated for each time step without assuming peak power availability (i.e., power level is determined explicitly from engine and transmission parameter values specific to the given conditions). Validation of the full vehicle dynamics model ing approach using the TruckSim ® software program. Calculation of updated PCE values and developing PCE estimation equations based on the updated custom traffic microsimulatio n tool . Developing updated commercial truck speed versus distance grade curves for upgrades of 1 through 8% and downgrades of 1 through 6% using the full vehicle dynamics modeling approach. OBD field data collection from a freeway corridor as well as a mul tilane highway corridor in order to generate the OBD data models . OBD data model validations through comparisons of OBD field data with the SwashSim results . Developing an OBD data class to be incorporated into the custom traffic microsimulation tool so th at OBD data can be output in the simulation results.

PAGE 24

24 CHAPTER 2 LITERATURE REVIEW This chapter provides a review of all literature significantly relevant to the areas of focus for this dissertation. It is organized into the following sections: Vehicle Dyn amics Model ing , PCE Calculation Methodologies, Commercial Truck Speed Versus Distance Grade Curves, and On Board Diagnostics Data Integration into Traffic Microsimulation. Vehicle Dynamics Model ing The first HCM was published in 1950 and it was mainly focu sed on passenger car traffic ( Transportation Research Board, 1950) . After this first HCM publication, the transportation community realized that t he capacity effect s of commercial trucks on freeway a nd multilane highways need to be addressed . In 1965, th e second version of HCM was published by the Transportation Research Board (TRB). Unlike the 1950 version , this version ( Transportation Research Board , 1965) addressed the effects of commercial trucks on roadway traffic through the usage of correction fac tors that are multiplied by selected service volumes under ideal conditions. These correction factors were based on PCE values under a specific roadway and traffic scenario such as a known flow, link length and grade. In order to calculate the PCE values, traffic simulation studies that included vehicle maximum acceleration models were performed . As a part of a National Cooperative Highway Research Program (NCHRP ), St. John a nd Kobett (1978) came up with a commercial truck maximum acceleration model ing ap proach that was mainly based on the S ociety of Automotive E ngineers (SAE , 1965 ) commercial truck ability

PAGE 25

25 prediction procedure . The St. John and Kobett model ing approach consist of the following acceleration equations . (2 1 ) (2 2 ) (2 3 ) (2 4 ) (2 5 ) where, = Effective acceleration in ft/s 2 , = Parameter that is dependent on engine speeds (0.4 is recommended) , = Vehicle speed in ft/s , use if , = Power limited acceleration in ft/s 2 ; the bar indicates use of average available net horsepower , = 1 times the sign of (either + or ) , = Actual time required to shift transmission gears in seconds , = Maximum speed in lowest transmission gear ratio, in ft/s , = Acceleration in coasting at vehicle speed , in ft/s 2 , = Altitude correction factor converting sea level net horsepower to local elevation ( ); given formula is for gasoline engines ,

PAGE 26

26 = Altitude correction factor converting sea level aerodynamic drag to local elevation ( ) , = Gross weight in lb , = Rated net horsepower at sea level conditions , = Projected frontal area in ft 2 , = Acceleration due to gravity in ft/s 2 , and = Percent grade divided by 100 This approach was applied into traffic simula tion by St. John and Kobett. However, in the equations only the maximum speed in lowest transmission gear ratio is accounted for, which explain why there is only one/initial transmission gear changing observed in th e maximum acceleration versus speed curves developed using this methodology. Therefore, using this maximum acceleration model, one cannot observe the differing transmission gear changes of the commercial truck, but on ly the very initial shift point. Anoth er detailed simulation model ing approach was introduced by Linzer et al . (1979) , in which the St. John and Kobett modeling approach was improved in a series of adjustments so that the model replicates influences of grade, vehicle population and flow rate f or the data collected on selected highway sites in California. However, similar to the St. John and Kobett maximum acceleration model ing approach , the Linzer approach also fails to address the full transmission gear changing ability of commercial trucks a nd how this effects the overall maximum acceleration versus speed curves. Another study ( Akcelik & Biggs , 1987 ) developed three acceleration models, a two term sinusoidal, a three term sinusoidal and a polynomial model. However, in this

PAGE 27

27 study the authors have only considered passenger cars and chose not to include commercial trucks. Therefore, the results do not take into consideration the maximum acceleration profiles on grades as well as the available tractive effort calculations. Akcelik later on deve loped a more complete maximum acceleration model ing approach (Akcelik & Besley , 2001) and incorporated this into his own traffic simulation program aaSIDRA. Akcelik and Besley (2001) use d power to weight ratio o f the commercial trucks in their equations, which is a determining factor in commercial truck maximum acceleration models due to its emphasis on commercial truck performance. However, the authors introduced adjustment factors for commercial truck maximum acceleration rates instead of doing an in de pth vehicle dynamics modeling approach that includes transmission gear changes, overall gear ratios, transmission gear ratios, etc. The fact that the aaSIDRA traffic simulation maximum acceleration model ing approach is not the same as the full vehicle dyn amics modeling approach can also be observed from the resu lts Akcelik and Besley reported, specifically the maximum acceleration versus time graph. This graph does not include sudden acceleration shifts due to the transmission gear changing phenomenon as observed in actual field conditions . Another commercial truck maximum acceleration model ing approach was introduced by Archilla and De Cieza (1999) usi ng a study that was performed on Argentinean highways. The authors were able to analyze their data as be ing very similar to a California study performed by Ching and Rooney (1979). Th e Archilla and De Cieza commercial truck maximum acceleration model is based on simple force balance equations being fitted to the field data collected. However, even though t his maximum acceleration model ing approach is more advanced than earlier literature , it

PAGE 28

28 also fails to look into the calculation of the transmission gear changing capabilities using transmission gear ratios and is therefore also a simplified modeling approa ch . In another study, Searle (1999) used equations of motion to calculate the vehicle speed and distance traveled. The Sea rle vehicle dynamics model ing approach does not take into consideration the specific resistance forces on the vehicle and therefore , is unable to accurately calculate the trajectory of a specific vehicle in stop and go conditions or at high speeds. Bester (2000) also introduced a speed profile calculation methodology to the literature. However, t his modeling approach is only sensitive to vertical alignment and therefore does not take into consideration the effects of other important variables such as free flow speed, traffic volume, etc. As Bester states, the speed profiles are very crucial since they are used in traffic simulation pro grams (Hoban et al . , 1985), highway evaluation packages (Watanda et al . , 1985) and for assessment of roadway design consistency ( Leisch & Leisch , 1977). Bester addressed the transmission gear changing phenomenon of commercial trucks by introducing both st atic and rotating masses into the force formulation as introduced into literature by Smith (1970). Even though, Bester acknowledges the fact that the force formulations differ according to the speed that each different commercial truck is being driven at and the transmission gear that it is in, he makes an assumption to have one specific mass value for all commercial truck types in his calculations. Therefore, t his approach is too simplistic and will result in great variation between the commercial truck types and different speed and transmission gear combinations.

PAGE 29

29 Rakha et al . (2001) came up with a more advanced commercial truck maximum acceleration model ing approach to be incorporated into the traff ic microsimulation INTEGRATION. The INTEGRATION traffic simulation tool was developed as an integrated simulation and traffic assignment model ( Van Aerde , 1985 ; Van Aerde & Yagar , 1988 ; Van Aerde & Yagar , 1990) . Even though the INTEGRATION software was originally built as a mesoscopic simulation tool, with im provements such as the addition of car following and lane changing logics and dynamic traffic assignment routines , it was eventually developed into a microscopic traffic simulation tool ( Van Aerde et al . , 1996) . The Rakha et al . (2001) commercial truck ma ximum acceleration model ing approach is similar to the ones before it ( Fitch , 1994; Archilla & De Cieza , 1999 ) . However, the Rakha et al . (2001) model ing approach calculates the available tractive effort of the engine and also checks for maximum tractive effort , which accounts for the friction between the commercial truck s tires and the pavement. T o ensure that the tr a ctive effort does not approach infinity on lower speeds , the tractive effort is chosen as the minimum of the available tractive effort and the maximum tractive effort. The Rakha et al . (2001) method consists of the following three main tractive effort calculations: (2 6 ) (2 7 ) (2 8 ) where, = Tractive effort in N = Engine power in KW

PAGE 30

30 = Transmission efficiency = Maximum tractive effort in N = Vehicle mass on tractive axle in kg = coefficient of friction b etween tires and pavement = Tractive effort effectively acting on commercial truck in N The authors suggested the usage of typical transmission efficiency values between 0.89 and 0.94 per the SAE J2188 ( SAE, 1996) guidelines. The maximum acceleration i s then calculated by summing all of the resistance forces on the commercial truck and subtracting this from the available tractive effort and dividing the remainder by the total vehicle mass. However, the Rakha et al . (2001) model ing approach assumes cons tant peak power and does not directly account for transmission gear shifting , therefore not as realistic and accurate as it could be . Rakha and Lucic (2002) realizing the shortcomings of th e initial study, introduced a variable power factor ( ), which is a function of vehicle speed. This factor is estimated through calibration and gives the following formulation s : (2 9 ) (2 10 ) (2 11 ) w here , = Optimum speed (km/h) = weight to power ratio This model ing approach has been used in the literature by Rakha and Ahn (2004) , Rakha et al . ( 200 4 ) and Al Kaisy et al . (2005) . Howeve r, it should be noted that

PAGE 31

31 this maximum acceleration model ing approach is a lso a simplified approach that does not directly account for transmission gear changing . Another major contribution to the speed profile models was the development of a commercial t ruck speed profile model (TSPM) as a part of NCHRP Report 505 (Harwood et al. , 2003) . TSPM is implemented as an Excel spreadsheet and is mainly consistent with the St. John and Kobett (1978) methodology as discussed previously. PCE Calculation Methodologi es PCE value calculations play a very important role in the accuracy of the flow rate estimations of freeways and multilane highways and are greatly affected by the vehicle dynamics modeling of the simulation programs used . This portion of the dissertatio n presents different PCE calculation methodologies that exist in the literature. Elefteriadou et al . (1997) used speed as the performance measure to calculate PCE values in th is initial study. The methodology that was used i s generally based on the Sumner et al. (1984) study, for which the authors replaced vehicle miles with speed. In a later and more advanced study, Webster and Elefteriadou (1999 ) estimated commercial truck PCEs using traffic simulation and based their calculations on flow rate and densit freedom to maneuver and proximity to other vehicles and most importantly is consistent with the measure of effectiveness (MOE) for freeways used in the current HCM. The auth ors also acknowledged that other PCE calculation methodologies based on average travel speed, density and other parameters exist in the literature ( Linzer et al . , 1979 ; Huber , 1982 ; Sumner et al . , 1984 ; Krammes & Crowley , 1987 ; Okura & Sthapit , 1995).

PAGE 32

32 Webs ter and Elefteriadou (1999) used the PCE estimation technique developed by Sumner et al. (1984). The initial step is to generate a flow v ersus density curve by simulating a passenger car only traffic stream at nine different flow rates spread evenly betwe en zero vehicles per hour per lane (veh/h/ln) and capacity. The second step is generating a similar flow vs. density curve , but this time using the typical vehicle mix, including passenger cars and commercial trucks. The third step is to replace a certain number of passenger cars , of 5%, with an equal number of the subject commercial truck. Then the fourth step is to simulate the operations of this traffic mix at a selected traffic flow rate , , and obtain the resultant traffic density, which is denoted as Point C in Figure 2 1 . Figure 2 1 . Illustration of Flow Density Relationship for PCE Calculation Methodology per Sumner et al. (1984) Figure reproduced from Webster and Elefteriadou (1999) .

PAGE 33

33 The fifth step is to d raw a horizontal line from Point C to intersect with the mix traffic curve at Point B so that the value of is obtained. The final step is to use Equation 2 12 in order to calculate the PCEs for each subject vehicle. (2 12 ) The simulated freeway section that was used by the au thors is depicted in Figure 2 2 . Figure 2 2 . Webster and Elefteriadou Experimental Roadway Configuration Webster and Elefteriadou generated several PCE tables for different types of heavy vehicles and concluded by stating that the PCEs tend to increase with traffic flow, free flow speed, and grade/length of grade. In addition, the authors determined that PCEs tend to decrease with an increase in commercial truck percentage and number of lanes. The me th ods used in the Webster and Elefteriadou study agree the most with the current HCM methodology since it is more logical to base PCEs on density f or freeway operational analysis. D ensity is the performance measure used to define LOS (i.e., se rvice measur e) for freeways in the HCM 2010 . Werner and Morrall (1976) looked at calculating PCE values based on capacity and headway. The authors estimated commercial truck, bus and recreational vehicle PCEs, however their study mainly discusses the effects that rec reational vehicles (RVs) have on highway capacity. In addition, they discussed the methods referred to in the

PAGE 34

34 1965 HCM for determining PCEs and how the then new PCEs should be used for typical highway capacity computations. The authors state that the res ults obtained in their study strongly indicated that the 1965 HCM PCE speed curves and adjustment factors in the 1965 HCM require refinement, particularly at slower speeds. Therefore, the authors offered a revised set of values by applying basic traffic e ngineering relationships. Al Kaisy et al. (2002) investigated the hypothesis that the effect of heavy vehicles on traffic during congestion is greater than their effect under saturated conditions. The authors developed a new approach to derive PCEs using queue discharge flow (QDF) capacity as the equivalency criterion. The sites that were used for this study were an entrance ramp merge area and a long term freeway construction zone. Van Aerde and Yagar (1983) derived PCE values using a speed reduction met hod that was based on relative rates of speed reduction for each type of vehicle traveling in the main direction and for all vehicles in the opposing direction. The analysis of their data suggested that a general speed volume curve shape consisted of two distinct parts: a linear section depicting normal operating conditions and a nonlinear section depicting breakdown in flow as the capacity is approached. Therefore, the authors focused on the linear section of the speed volume curve since it mainly repres ents the entire range of practical operating volumes. A linear approximation was found by the authors to fit the data for each of the 10 th , 50 th and 90 th speed percentiles and a multiple linear regression model was structured in Equation 2 13 . (2 13 )

PAGE 35

35 Using this multiple linear regression mod el, the free flow speed and the speed reduction coefficients were estimated. The authors suggested that indicated the relative sizes of speed reductions for each vehicle type. The final PCE values for each vehicle type were then calcula ted by using Equation 2 1 4 , where is the coefficient for the vehicle type that the PCE is being calculated for. PCE n (2 14 ) The Van Aerde and Yagar paper, similar to the Linzer et al . (1979) study , estimates the PCEs by using speed relationships of the traffic stream . Cunagin and Messer (1983 ) used a methodology to calculate PCE values using ratio of delay experienced by a p assenger c ar due to a non p assenger c ar to the delay experienced by a p assenger c ar due to anot her p assenger c ar . In addition, the authors used speed distributions, traffic volume and vehicle type to come up with the PCE values . Similarly , Craus et al . (1980) also used delay f or calculating PCEs on two lane highways. The methodology was also base d on the Walker method, which assumes that faster vehicles in the traffic stream are not negatively affected as they overtake slower vehicles. This way the queues do not form. This method conflicts with the delay method, which basically assumes that the slower vehicles impact the faster vehicles , therefore generating queues. Th is is why a combination of the Walker and the delay method was used in the Craus et al . (1980) and Cunagin and Messer (1983) studies to calculate the PCE values, where the walker m ethod is used for lower volumes and the delay method is used for higher volumes. Another methodology was introduced by Fan (1990), for which the author used data that was obtained from a Singapore expressway. Fan focused on v/c ratios that

PAGE 36

36 are higher than 0.67, therefore congested traffic data was used rather than uncongested. The author then estimated the PCEs by using a multiple linear regression model that multiplies the ob served flow by the v/c ratios. Sumner et al. (1984) generated a methodology to c alculate PCE values between consecutive signalized intersections on urban arterial roads using Network Simulation ( NETSIM ) simulation program in order to obtain vehicle hours. The authors looked into the additional vehicle hours generated when commercial trucks were introduced into the traffic stream and based the PCE calculation s on this. Van Aerde and Yagar (1983) also looked at deriving PCE values using a platoon leadership and follower creation. The authors used the radar platoon technique for their da ta collection efforts. However, a peer reviewer of their original publication criticized this method since it was found to be very sensitive to the definition of what headway separation constitutes a different platoon. The authors replied by stating that they gathered a group of people from different backgrounds and there was unanimous agreement where the platoons star ted and ended for their study. Keller and Saklas (1984) looked at travel time to estimate PCE values. More specifically, the authors used signal timing for urban arterial networks in orde r to calculate the PCE values, which were calculated as the ratio of the total commercial truck travel times to passenger car travel times . Rakha et al . (2007) performed another PCE calculation study and use d density as the determining performance measure. He use d the INTEGRATION microsimulation program to run the traffic simulations and used the same methodology developed by Sumner et al . (1984) to calculate the PCE values.

PAGE 37

37 As one of the most recent studies , Cunha and Setti (2011) introduced a PCE calculation study performed in Brazilian highways that used density as its main measure. The authors used CORSIM 5.1 as their simulation tool and performed a two stage calibration using the commercial truck calibr ation factors introduced in Cunha et al . (2009 ) and also the car following parameters introduced in Ara ú jo (2007) . However, Cunha also acknowledges that CORSIM simulates the behavior of commercial trucks on grades using an average maximum acceleration fro m the predetermined look up table that is included in the software program. Commercial Truck Speed V ersus Distance G rade For a basic freeway segment analysis, if any one of the grades of the analysis segment is greater than 4% or the total length of the se gment is greater than 4000 feet, the composite grades procedure, as depicted in HCM 2010, Chapter 11 Appendix A, is recommended for use . Figure 2 3 presents HCM 2010 Exhibit 11 A1, labeled in the HCM as Performance Curves for 2 00 lb/hp Truck. Figure 2 3 . HCM 2010 Exhibit 11 A1

PAGE 38

38 The se curves are mainly based on the methodology introduced in literature by Schwender et al. (1957 ) . They are designed to estimate a single grade that substitute s the grad e combination of the segment and also outputs a final speed of a commercial truck as it reaches the end of a composite grade segment. The methodology Schwender et al . (1957) introduced was a capacity determination technique in mountainous terrain to accou nt for the effects of commercial trucks . The authors came up with a set of speed versus distance grade graphs that were later on incorporated into the HCM 1965 Highway Research Board (HRB) Special Report 87 ( HRB, 1965). Figure 2 4 depicts the speed versus distance grade graph as included in the HRB Special Report 87. This graph was based mainly on the Schwender et al . (1957) study as well as the study regarding commercial truck speeds on grades by Webb (1961). Figure 2 4 . Effect of Length and Grade on Speed of Average Commercial Trucks on Multilane Highways Leisch (1974) also acknowledged the Schwender et al . (1957) study in his own research regarding capacity analysis techniques of freeway faciliti es. Leisch suggested

PAGE 39

39 that an accompanying speed profile of a passenger car along with the commercial truck speed profile can help the designer to avoid any areas in freeway design, where there will be a big speed differential between the commercial trucks and the passenger cars. Walton and Lee (1975) also developed curves for commercial truck speed on grade as a part of their study. The authors collected speed and gross weight field data and by using the speed history records they obtained , speed versus di stance grade curves were developed using a stepwise regression analysis technique . The authors also looked at the driver experience factor from the results they obtained using the data from the survey results completed by commercial truck drivers. Archill a and De Cieza (1999) also introduced a then new commercial truck speed versus distance grade graph based on their commercial truck maximum acceleration model ing approach as discussed previously. However, even though the commercial truck maximum accelerat ion model ing approach they used in order to generate these speed versus distance grade graphs is more advanced than the ones used earlier in literature, this model ing approach also fails to look into the calculation of the transmission gear changing capabi lities using transmission gear ratios . The speed versus distance grade graph for a 190 kg/KW (~312 lb/ hp ) commercial truck that Archilla and De Cieza (1999) came up with is presented in Figure 2 5 .

PAGE 40

40 Figure 2 5 . Effect of Length and Grade on the Speed of a 190 kg/KW C ommercial T ruck The authors state that these commercial truck speed versus distance grade curve s that their model predict ed deviates from the HCM curve s . They explain this deviance by stat ing that the HCM values are obtained for a sea level elevation, whereas the Argentinean study is performed for an elevation of 2000 meters. Harwood et al. (2003) commercial truck maximum acceleration model ing approach also outputs a speed versus distance g rade graph through the TSPM Excel spreadsheet user interface and the speed profile calculations are performed using the commercial tru ck performance equations of TWOPAS ( St. John & Kobett , 1978 ; St. John & Harwood , 1986). TSPM accounts for desired speed, initial speed, weight to power ratio, frontal area of the commercial truck as well as the elevation above sea level and percent grade of the analysis segment. However, by analyzing the results, it was determined that the TSPM model ing approach only accoun ts for the initial

PAGE 41

41 transmission gear change to be observed in the maximum acceleration versus speed results. Rakha and Yu (2004) also introduced a study that models the commercial truck speed versus distance grade curves using the Rakha and Lucic (2002) co mmercial truck maximum acceleration model ing approach . However , this model ing approach is simplifi ed in its assumption of the tractive effort calculations and when the results are analyzed transmission gear changing is not observed in the maximum accelera tion versus speed graphs. On Board Diagnostics Data Integration into Traffic Micros imulation As discussed under the background section, Frey et al. (2010) states that models of fuel use and emission rates based on engine data are more predictive than those based only on VSP. This same study also found that a VSP based model for fuel use for a 2005 Chevrolet Cavalier had an R 2 of 0.87. However, a model based on the product of RPM and MAP had an R 2 of 0.99 . Thus, the direct use of internally observable var iables ( IOVs ) based on OBD data may lead to improved estimates of vehicle fuel use and emission rates. On the other hand, t he area of integrating OBD data into traffic simulation is fairly new in the literature. However , the interfacing of computer softwa re to the OBD system has been studied and various computer program s were developed in order to receive and store the OBD data received Studies such as Godavarty et al . (2000) and Dzhelarski and Alexiev (2005) worked on de vel oping computer code s that can read data from an OBD II data por t and display it through an OBD interface on a computer.

PAGE 42

42 Other studies like Ozbay et al . (2007) collected OBD data through the OBD port of the vehicle and made th ese data available to the simul ation runs that they performed using the microsimulation program P aramics . The authors then augmented P aramics to add communication ability between the vehicles in the traffic stream . This way, the vehicles can communicate th e OBD information to each oth er through a W i F i connection and warn each other of any issues that one of the vehicle s might be observing. Currently, there is a growing interest in developing applications that will use OBD data. In this effort, the U.S. Department of Energy recently s ponsored an event called ( Kalin , 2013 ) that has a main theme to use OBD data for driver feedback regarding fuel economy. In addition, General Motors is encouraging third party developers to develop da sh information systems (Programming Cars: Using the GM In Dash SDK, 2013 ). Furthermore, even though the current trend is that the user need s to be within the vehicle to observe the respective OBD data, in the near future OBD data will also be readily avai lable to users that are outside of the said vehicle. This new concept is made available through a device that connects to the OBD port and transmits data from the to drive). Users who choose to subscribe to this service can obtain information regarding how their driving impacts fuel usage and safety. The i2d device reports OBD data such vehicle speed, RPM, MAP and mass fuel flow (MFF) and the data becomes availabl e within minutes of its real time transmission to the server . (I2D, 2013)

PAGE 43

43 CHAPTER 3 COMMERCIAL TRUCK PERFORMANCE This chapter provides an overview for the methodolog ies used in order to develop a full vehicle dynamics model ing approach , incorporate it in to traffic microsimulation, generate new PCE estimation equations, and update the commercial truck speed versus distance grade curves. In addition, this chapter also provides the results obtained by performing these tasks . Methodology In the methodology s ubsection, the tasks that were undertaken in order to finalize the commercial truck performance chapter are discussed in detail. This subsection is organized in to the following subheadings: Commercial Truck Classification and AADT Data, Commercial Truck C haracteristics Data, Custom Simulation Platform, Improved Commercial Truck Acceleration Versus Speed Curves, TruckSim ® , Advanced/Full Vehicle Dynamics Modeling Approach, Updated PCE Calculations Using Full Vehicle Dynamics Modeling Approach , and Updated Co mmercial Truck Speed Versus Distance Grade Curves. Commercial Truck Classification and AADT Data In order to obtain commercial truck classification data from Florida freeways and multilane highways, provided the research team with a list of 24 Active Permanent Weigh in Motion (WIM) Stations that are located on Florida freeways and multilane h ighways. A DVD that contains FDOT Traffic Information ( AADT, total commercial truck volume , commercial truck classes, and corresponding volume s for each of the 24 WIM stations for the years 2008, 2009, and 2010 ) was also supplied by the FDOT Statistics Office . Th ese data w ere organized in tables (two data tables per

PAGE 44

44 page for each WIM station) and are shown in Table A 1 through Table A 48 in App endix A . Figure 3 1 presents the FHWA Vehicle Classification scheme and the corresponding vehicle class numbers used in this dissertation . Figure 3 1 . FHWA Vehicle Classification Chart ( TXDOT, 2013) Using these data, Table 3 1 provides the overall commercial truck classification history. In addition, Table 3 2 through Table 3 5 represent the commercial truck classification history based on area type and roadw ay type.

PAGE 45

45 Table 3 1 . Overall Truck Classification History as % of Truck AADT Truck Class Total Volume Per Truck Class % of AADT 5 26326 25.36% 6 6354 6.12% 7 1092 1.05% 8 10456 10.07% 9 55446 53.41% 10 605 0.5 9 % 11 2159 2.08% 12 1100 1.06% 13 275 0.26% TOTAL % 100.00% Table 3 2 . Urban/Freeway Truck Classification History as % of Truck AADT Truck Class Total Volume Per Truck Class % of AADT 5 15089 28.62% 6 3500 6.64% 7 671 1.27% 8 5883 11.16% 9 25482 48.33% 10 335 0.64% 11 1118 2.12% 12 470 0.89% 13 172 0.33% TOTAL % 100.00% Table 3 3 . Urban/Multilane Highway Truck Classification History as % of Truck AADT Truck Class Total Volume Per Truck Class % of AADT 5 2976 33.57% 6 1481 16.71% 7 314 3.54% 8 910 10.27% 9 3094 34.91% 10 41 0.46% 11 24 0.27% 12 13 0.15% 13 11 0.12% TOTAL % 100.00%

PAGE 46

46 Table 3 4 . Rural/Freeway Truck Classification History as % of Truck AADT Truck C lass Total Volume Per Truck Class % of AADT 5 5122 17.03% 6 791 2.63% 7 45 0.15% 8 2407 8.00% 9 20100 66.83% 10 173 0.58% 11 862 2.87% 12 532 1.77% 13 46 0.15% TOTAL % 100.00% Table 3 5 . Rural/Multilane Highway Truck Classification History as % of Truck AADT Truck Class Total Volume Per Truck Class % of AADT 5 3139 25.83% 6 582 4.79% 7 62 0.51% 8 1256 10.34% 9 6770 55.72% 10 56 0.46% 11 155 1.28% 12 85 0.70% 13 46 0.38% TOTAL % 100.00% By analyzing the dat a in Table 3 2 through Table 3 5, it was determined that commercial truck classifications 5, 6, 8, 9, 11 and 12 are the most prevalent commercial truck types on Florida multilane highways and freeways. Considering the characteristics and the similarities between some of these commercial truck types, four classes were generated, namely, Classes 5&6 combined under Single Unit Truck, Class 8 as Intermediate Semi trailer, Class 9 as Interstate Semi trailer, and Classes 11&12 combine d under Semi tractor+ double trailer .

PAGE 47

47 Commercial Truck Characteristics Data In order to obtain information on the current commercial truck fleet, physical and power characteristics for Mack, Peterbilt, Volvo , and Kenworth brand commercial trucks was obtained (Mack Trucks , 2013 & Peter bilt Trucks , 2013 & Volvo Trucks , 2013 & Kenworth Trucks , 2013). For the Mack brand, physical and power characteristics for models Pinnacle and Titan were obtained. For the Peterbilt brand, physical and power characteristics for models 386, 388, and 587 were obtained. For the Volvo brand, the physical and power characteristics for models VN 630, 780, and VNL 430 were obtained. Lastly, for the Kenworth brand , physical and p ower characteristics for models T700 , T800, and W900 were obtained. Table 3 6 summarizes the common engine characteristics of these commercial truck types. Table 3 6 . Heavy Vehicle Fleet Common Engine Characteristics (P ower and Torque) Heavy Vehicle Make/Model Typical Engine Type P o wer (hp) Torque (ft lb) Mack Pinnacle MP 7 MP 8 325 405 415 505 1260 1560 1460 1760 Mack Titan MP10 MCruise 515 605 1860 2060 Peterbilt 386 Paccar MX 385 485 1450 1 6 50 Peterbilt 388 Cummins ISX15 400 600 1450 1850 Peterbilt 587 Cummins ISX15 400 600 1 450 1850 Volvo 630 Volvo D13 375 500 1450 1750 Volvo 780 Volvo D16 500 550 1450 1850 Volvo 430L Volvo D11 325 405 1250 1450 Kenworth T 700 Paccar MX 385 485 1450 1 6 50 Kenworth T 800 Cummins ISX15 400 600 1450 1850 Kenworth W 900 Cummins ISX15 400 600 1450 1850

PAGE 48

48 In addition to the commercial truck classification and AADT data obtained using the FDOT Traffic Information DVD, the commercial truck characteristics data such as weight loadings, length of commercial truck, speed of commercial truck, etc. we re obtained by processing and analyzing the data obtained from the 24 WIM stations throughout the state. These data were obtained in raw WIM data format from FDOT for the years of 2008, 2009, 2010 and a portion of 2011 (up to September 2011) . The first s tep in the processing of the files was to use the PAT software program to convert the original binary files into ASCII format. This step is necessary since the PAT software is proprietary and the organization of the data in the original binary files is no t publis hed (PAT Reporter US A , 2007 ) . The next step was to process the ASCII files to obtain various statistics for each station such as speed for each class, average speed for all classes, weight for each class, average weight for all classes, frequency for each class, etc. Since there is a very large number of data files that need to be processed (approximately 44,000 for 4 years of data across 24 WIM stations), a custom WIM data processing program was developed in the C# programming environment to auto mate this process. This program has capabilities such as choosing a specific data folder, moving selected folders to a new folder, renaming the files of interest for analysis purposes and reading/processing the WIM files to produce results. In addition, once the files are analyzed, the program has the capability of allowing the user to select the WIM stations of interest, either one by one or in groups. Also, the program can distinguish between WIM stations by area and facility type and can aggregate res ults for full day or time of day analysis. Figure B 1 and Figure B 2 in Appendix B present the control windows for all of these capabilities of the C# data processing prog ram.

PAGE 49

49 Making full use of the capabilities of this program, two categories of results were obtained, namely, Full Day and Time of Day [Morning Peak (6 9 am), Mid Day (11 am 1 pm) and Evening Peak (4 7 pm)] so that the effects of full day versus time of day c ould be observed. In addition, the results were further divided by area type into Urban and Rural as well as by facility type into Multilane Highway and Freeway. A sample table of the results obtained is presented in Table C 1 under Appendix C . The data obtained in this table format r epresent the field conditions on Florida multilane highways and freeways, and the se values were used for generating the four commercial truck types and their characteristics during mi cro simulation modeling as described under the research approach section. Custom Simulation Platform At the UF TRC, CORSIM is typically the simulation tool of choice because this program is maintained at UF and the research team ha s full access and control of the source code . However, the current version of CORSIM (6.3) uses a simplistic method of determining maximum commercial truck acceleration values (a lookup table with maximum acceleration based simply on velocity) , as do the other major commercial sim ulation tools (Paramics, VISSIM, AIMSUN) . Furthermore, the relationship between these table values and the grade adjustment factor is convoluted is not accurate due to this approach not being as mathematical as a vehicle dynamics approach. This makes it difficult to ensure that effect of grade on maximum acceleration is being properly accounted for. Unfortunately, the software architecture of the current version of CORSIM is also not amenable to implementing a mo re comprehensive and accurate vehicle maxi mum acceleration model ing approach .

PAGE 50

50 Therefore, a custom traffic microsimulation tool , referred to as SwashSim , which has been under development for the past couple o f years , but not yet publicly available, was utilized for this project. This microsimulati on tool employs state of the art software architecture , which is object oriented and built on the C# / . NET framework programming model. This framework allows for a high level of extensibility and modularity. The new architecture also supports a high leve l of fidelity with respect to temporal and spatial modeling resolution. Much of the vehicle movement logic in SwashSim is the same as that employed in CORSIM 6 .3 , with the following exceptions . The custom traffic microsimulation tool uses the Modified Pitt car following model (Cohen, 2002) as following model. The discretionary lane changing logic has been enhanced by adding logic to bias (but not restrict) slower moving vehicles to the right side lanes. For example, for a 3 l ane roadway, the slowest vehicles in the traffic stream will generally be in the far right lane, the fastest vehicles will be in the far vehicles will be in the middle lane. However, unlike a lane restriction scenario, any o f these vehicles can still use other lanes, which might happen temporarily for conducting passing maneuvers. This logic particularly comes into play for the commercial truck vehicle types on grades. The commercial trucks generally have somewhat lower des ired speeds than the passenger cars, so they are more likely to be in the right or middle lane (of a 3 lane roadway) than in the left lane, but regardless of which lane they are initially in, as they begin to lose speed on a grade, they will look to move to the right side lanes. For a smaller commercial truck type (e.g., the single unit truck) on a moderate grade, it still may be able to maintain its desired speed and therefore will not be biased toward the right side lanes. Some other notable differences between the custom traffic microsimulation tool and CORSIM 6 include the following. A 0.1 second simulation time resolution instead of 1 second for CORSIM 6. Explicit modeling of vehicle paths from system entry to system exit. CORSIM 6 does not explicitl y model vehicle movements through an intersection area -the animation component of CORSIM 6, TrafVu, interpolates vehicle positions through the intersection areas based on estimated intersection vehicle entry and exit times from CORSIM 6.

PAGE 51

51 In order to devel op the full vehicle dynamics modeling approach and incorporate it into traffic microsimulation, the addition of new components to SwashSim along with revisions to some existing components was required. In general, because of the object oriented architectu re of the custom traffic microsimulation tool , there are many more possibilities of what c an be model ed versus CORSIM 6. In addition, the same scenarios that CORSIM 6 is currently capable of modeling, but with greater detail and accuracy , can also be mode led using the custom traffic microsimulation tool . As one example, in SwashSim vehicles and drivers are separate objects, whereas in CORSIM 6, there is a driver type property that is integral to the vehicle definition. By having separate vehicle and driv er objects, there is much more flexibility in the properties that can be assigned to both and how the two objects can be coupled together. One of the key features of the custom traffic microsimulation tool is the ability to model individual vehicle charact eristics and dynamics in great detail, which was specifically used for creating the advanced dynamics modeling approach and to model commercial truck characteristics and dynamics to replicate their behavior in the traffic stream. Additionally, passenger ca r and commercial truck characteristics data such as engine torque, transmission gear ratios, etc. were also incorporated into the SwashSim simulation program to ensure acc urate representation of the actual vehicles in the simulation process. These data are summarized in Table 3 7 .

PAGE 52

52 Table 3 7 . Vehicle Characteristics Data Passenger Car Single Unit Truck Intermediate Se mi Trailer Interstate Semi Trailer Semi tractor+double trailer Vehicle Height (ft) 4.46 10.00 10.00 10.00 10.00 Vehicle Width (ft) 5.74 7.00 8.00 8.00 8.00 Vehicle Length (ft) 16.00 29.00 55.00 68.50 74.60 Vehicle Weight (lb) 3 , 060 25,000 37,000 53,000 55,000 Maximum Torque (lb ft) 139 660 1650 1650 1650 Maximum Power (hp) 197 300 485 485 485 Wheel Radius (ft) 1.03 1.66 1.66 1.66 1.66 Differential Gear Ratio 4.77 4.40 3.50 3.50 3.50 Transmission Gear Ratios Gear 1 3.27 7.59 11.06 11.06 11.06 Gear 2 2.13 5.06 8.20 8.20 8.20 Gear 3 1.52 3.38 6.06 6.06 6.06 Gear 4 1.15 2.25 4.49 4.49 4.49 Gear 5 0.92 1.50 3.32 3.32 3.32 Gear 6 0.66 1.0 0 2.46 2.46 2.46 Gear 7 N/A 0.75 1.82 1.82 1.82 Gear 8 N/A N/A 1.35 1.35 1.35 Gear 9 N/A N/A 1.00 1.00 1.00 G ear 10 N/A N/A 0.74 0.74 0.74 Improved Commercial Truck Acceleration V ersus Speed Curves In order to compare the results of several advanced vehicle dynamics model ing approach es , three commercial truck maximum acceleration model ing approache s were consid ered for the simulation of Florida traffic conditions . As discussed under the literature review section, many different methods for determining vehicle maximum

PAGE 53

53 acceleration capabilities have been proposed in the literature. For evaluation purposes, t he r esearch team evaluated the most advanced/recent three methods. V ehicle dynamics model ing approach (Mannering & Washburn , 2012 & Wong , 2008 ). Refer HP Al Kaisy et al . (2005), which use a revised version of the model ing approach introduced into literature by Rakha and Lucic (2002) . Referred graphs since it was originally introduced by Rakha. Harwood et al. (2003), which is referred to as the TSPM method in graphs since this study developed the TSPM spreadsheet. The vehicle dynamics model ing approach relies on detailed calculations for engine generated tractive effort and maximum tractive effort (maximum force that can be handled at the tire pavement interface). This approach also requires detailed information on engine performance (power and torque) and transmission configuration (number of transmission gears and specific transmission gear ratios). For commercial trucks , the engine generated tractive effort (as opposed to maximum tractive effort) is almost alway s the controlling factor for maximum acceleration calculations . In addition to these forces, various resistance forces are also considered, since they counteract the tractive effort generated by the engine. These resistance forces are aerodynamic, rollin g, and grade resistance. Per the engine characteristics research performed as summarized in Table 3 6 , Paccar MX engines with a Horsepower Torque rating of 48 5 hp and 1650 lb ft were used for Intermediate/Interstate Semi trailers and Semi tractor+ double t railers. In addition, Paccar PX 6 300 engines with a Horsepower Torque rating of 300 hp and 660 lb ft were used for Single Unit trucks. uses formula s as outlined in Al Kaisy et al . (2005) , which is the latest version of this type of vehicle maximum acceleration model ing approach . These formulas are generally consistent with the conceptual approach used in the first

PAGE 54

54 method, but incorporate some simplifying assumptions to reduce the computational effort . The TSP M method uses the same methodology as highlighted in St. John and Kobett (1978) . Again, this methodology is generally consistent with the conceptual approach in the first method, but also uses simplifying assumptions , which ends up with only the initial t ransmission gear change being observed in the maximum acceleration versus speed graphs. After these three different methods were evaluated, the research team generated two sets of maximum acceleration vs. speed curves so that a comparison between the three methodologies could be obtained. The first set was developed for level grades and the second set for 5 % grades. These maximum acceleration versus speed curves are depicted in Figure D 1 through Figure D 8 in Appendix D . After the analysis of these curves, in order to improve the advanced vehicle dynamics model ing approach so that it can r eplicate the transmission gear changing capabilities of actual commercial truck oper ations and to prove that there is maximum acceleration versus speed difference compared to real life commercial truck operations from the other models , methodology (Al Kaisy et al. , 2005) was chosen as a test model ing approach . This choice was made due to this approach being a compromise between accuracy and computational effort when compared to the advanced vehicle dynamics model ing approach . TruckSim ® In order to replicate the actual commercial truck maximum acceleration performance on grades , a copy of the TruckSim ® software program was used . TruckSim ® provides detailed simulation of individual commercial trucks (not a traffic

PAGE 55

55 microsimulation tool) , based on mathematical models of the commercial powertrain (engine and transmission) a nd physical characteristics. TruckSim ® was used to determine the maximum acceleration capabilities of a commercial truck on varying grades and this information was used to validate the commercial truck maximum acceleration calculation methodology results as discussed under the previous section. This software allowed for the opportunity to further evaluate the commercial truck maximum acceleration curves previously developed by using the test commercial truck maximum acceleration model ing approach that was labeled Rakha . Figure 3 2 through Figure 3 5 depict some of the main control windows in TruckSim ® as well as the results that could be obtained. Figure 3 2 . TruckSim ® Run Control Window (TruckSim ® by Mechanical Simulation Corporation, http://carsim.com/products/trucksim/index.php , License obtained by UF)

PAGE 56

56 Through the run control window, the user can select th e commercial truck configuration that they want to analyze. In addition, the procedure/test that needs to be run is also set from this window by changing the roadway attributes, such as segment length, segment grade, etc. Figure 3 3 . TruckSim ® Vehicle Attributes Window (TruckSim ® by Mechanical Simulation Corporation, http://carsim.com/products/trucksim/index.php , License obtained by UF) By using the TruckSim ® vehicl e attributes window, the user can specify the commercial truck that they want to analyze, as well as the trailers and payloads that are associated with this specific commercial truck .

PAGE 57

57 Figure 3 4 . TruckSim ® Roadway Geometry Wi ndow (TruckSim ® by Mechanical Simulation Corporation, http://carsim.com/products/trucksim/index.php , License obtained by UF) By utilizing the TruckSim ® roadway geometry window, the user has the ability to set the type of roadway characteristics for their analyses, including grade, length, surface, etc.

PAGE 58

58 Figure 3 5 . TruckSim ® Plot Outputs Window (TruckSim ® by Mechanical Simulation Corporation, http://carsim.com/products/trucksim/index.php , License obtained by UF) The user can specify the plots that they would like to use in their analyses and get the TruckSim ® outputs accordingly. In addition, all of the pl ot outputs are available to be saved as a text file to the users discretion.

PAGE 59

59 Advanced /Full Vehicle Dynamics Model ing Approach Test runs were performed using Single Unit, Intermediate Semi traile r and Interstate Semi trailer / Semi tractor+double trailer truc ks by using both the revised custom traffic microsimulation tool maximum acceleration model ing approach (Al Kaisy et al ., 2005) and TruckSim ® . The se test runs were to compare the maximum acceleration and velocity outputs of these models and to observe if the Rakha method fail s to replicate the actual vehicle dynamics characteristics as observed by TruckSim ® . This way, if th is model ing approach failed to replicate the actual vehicle dynamics results from TruckSim ® , the need for a more advanced (full) veh icle dynamics model ing approach in traffic micro simulation is warranted . Figure E 1 through Figure E 10 in Appendix E present the results that were obtained by these test r uns. By examining the TruckSim ® and the SwashSim Rakha vehicle maximum acceleration model ing approaches initially chosen for testing, it was determined that the differences between the actual field commercial truck performance ( TruckSim ® ) and the m aximum acceleration model ing approach were significantly different . These differences were largely due to the fact that the Rakha maximum acceleration model ing approach does not take into account the transmission gear shifting characteristics of commercia l trucks , whereas the TruckSim ® model does. This causes inconsistencies between the maximum acceleration and velocities of the two model ing approache s and therefore affect s the traffic streams generated by either of these modeling approaches differently . Therefore, accounting for transmission gear shifting ability of commercial trucks i s a significant enough factor that it should be accounted for in the commercial truck maximum acceleration calculation model ing . This capability is not accounted for

PAGE 60

60 in an y traffic microsimulation program ( CORSIM, VISSIM, AIMSUN, PARAMICS, INTEGRATION, SIDRA, etc.) and the maximum acceleration model ing approache s used in these simulation programs are relatively simplistic. Therefore, one of the main contri butions of this di ssertation is introducing an advanced vehicle dynamics model ing approach as a part of a custom t raffic microsimulation program , which has the capability to account for the transmission gear shiftin g abilities of commercial trucks and show the transmission gear selected at each time step of the simulation. In this effort, transmission gear changing speeds for the passenger car and the four commercial truck types of interest was calculated using vehicle dynamics equations from Mannering and Washburn (2012). In addition, typical transmission gear ratios for these vehicles using transmission information from Wong (2008) and the Internet ( Allison & Eaton Transmissions , 2013 ) were obtained . An overview of the vehicle dynamics model ing approach that was developed and incorporated into SwashSim is given here. The approach at its most basic level determines maximum acceleration through the fundamental equation relating tractive force to resistance forces is depicted in E quation 3 1 . ( 3 1 ) The tractive force, F , referred to here as available tractive effort, is taken as the lesser of maximum tractive effort and engine generated tractive effort. Maximum tractive effort is a functi wheelbase, center of gravity, and weight) and the roadway coefficient of road adhesion. Maximum tractive effort represents the amount of longitudinal force that can be accommodated by the ti re pavement interface. Engine generated tractive effort is a

PAGE 61

61 function of engine torque, transmission and differential gearing, and drive wheel radius. For vehicles with low power to weight ratios, such as commercial trucks, maximum tractive effort is ver y rarely the governing condition. Thus, the maximum acceleration calculations for commercial trucks in SwashSim are coded to be based on engine generated tractive effort. The major resistance forces are aerodynamic, rolling, and grade. The equation for d etermining aerodynamic resistance is presented in Equation 3 2 . ( 3 2 ) w here , = aerodynamic resistance in lb , = air density in slugs/ft 3 , = coefficient of drag (unitless), = frontal area of the vehicle (projected area of the vehicle in the direction of travel) in ft 2 , and = speed of the vehicle in ft/s. The coefficient of rolling resistance for road vehicles operating on paved surfaces is approximated as presented in E quation 3 3 . ( 3 3 ) w he re , = coefficient of rolling resistance (unitless), and = vehicle speed in ft/s. The rolling resistance, in lb . , is simply the coefficient of rolling resistance multiplied by W cos , the vehicle weight acting normal to the roadway surfa ce. For

PAGE 62

62 most highway applications is very small, so it can be assumed that cos = 1, giving the equation for rolling resistance ( ) as presented in Equation 3 4 . ( 3 4 ) Grade resistance is s imply the gravitational force (the component parallel to the roadway) acting on the vehicle. The expression for grade resistance ( ) is as depicted in Equation 3 5 . ( 3 5 ) As in the development of t he rolling resistance formula, highway grades are usually very small, so sin tan . Thus, grade resistance is calculated as presented in Equation 3 6 . ( 3 6 ) w here , = grade, defined as the vertical rise per some specified horizontal distance in ft/ft. Grades are generally specified as percentages for ease of understanding. Thus a roadway that rises 5 ft vertically per 100 ft horizontally ( G = 0.05 and = 2.86°) is said to have a 5% grade. The relationship between vehicle speed and engine speed is summarized in Equation 3 7 . ( 3 7 ) w here , = vehicle speed in ft/ s, = engine speed in crankshaft revolutions per second ,

PAGE 63

63 = slippage of the drive axle, and = overall gear reduction ratio The overall gear reduction ratio is a function of the differential gear ratio and the transmission gear ratio, which is a function of the selected transmission gear for the running speed. This equation can be rearranged to solve for engine speed given the current vehicle speed (if vehicle speed is zero, engine speed is a function of throttle input). With the calculated engine speed, the torque being produced by the engine can be det ermined from the torque engine speed relationship. For example, assuming an engine speed of 2000 revs /min with the torque engine speed relationship (Paccar PX 7 Engine) shown in Figure 3 6 ( Peterbilt Engine PX7 , 2013 ), the result ing torque is 660 ft lb. In addition, Figure 3 7 shows the torque engine speed relationship for a Paccar MX 13 engine ( Peterbilt Engine MX13 , 2013 ).

PAGE 64

64 Figure 3 6 . Torque Engine Speed Curve for a Pacca r PX 7 Engine Figure 3 7 . Torque Engine Speed Curve for a Paccar MX 13 Engine

PAGE 65

65 Power is the rate of engine work, expressed in horsepower (hp), and is related to Equation 3 8 . ( 3 8 ) w here , = engine generated horsepower (1 horsepower equals 550 ft lb/s), = engine speed in crankshaft revolutions per second , and = engine torque in ft lb. The engine genera ted tractive effort reaching the drive wheels is given in Equation 3 9 . ( 3 9 ) w here , = engine generated tractive effort reaching the drive wheels in lb, = engine torque in ft lb. = overall gear reduction ratio, = mechanical efficiency of the drivetrain, and = radius of the drive wheels in ft. It should be noted that since torque and horsepower are directly related, if only a power engine speed relationship is available, this can be converted to a torque engine speed relationship by using Equation 3 9 . For determining vehicle maximum acceleration, Equation 3 1 is rearranged and an additional term, t be overcome during acceleration, is included as depicted in Equation 3 10 .

PAGE 66

66 ( 3 10 ) , referred to as the mass factor, is approximated as presented in Equation 3 11 . ( 3 11 ) For further clarification of how the advanced vehicle dynamics model ing approach calculates the maximum acceleration of a vehicle, a n example application of this approach is presented in Figure F 1 and Figure F 2 of Appendix F . Updated PCE Calculations Using Full Vehicle Dynamics Modeling Approach Due to the fact that the incorporation of the full vehicle dynamics model ing approach into traffic microsimulation affect s PCE values directly , this dissertation present s updated PCE values that are calculated by utilizing the advanced vehicle dynamics model ing approach developed and incorporated into SwashSim . As discussed under the literature review chapter , multiple methodologies to calculate PCE values have been introduced. After further analysis, i t was determined that the methods used in Webster and Elefteriadou (1999) are the most consistent with the current HCM methodology. A primary reason for this is because density is the performance measure used to define LOS for freeway operations in the HCM. Therefore, this methodology was selected for calculating the updated PCE values. The roadway configuration aspect of the experimental design that was used in this study is depicted in Figure 3 8 .

PAGE 67

67 Figure 3 8 . Roadway Configuration used for Experimental Design Simulation Runs Once this experimental roadway configuration was set up in SwashSim , the following variables in Table 3 8 were used to calculate the final PCE values for this study. Table 3 8 . Variables used for Experimental Design in Simulation Runs Variable Setting Level Number of lanes in analysis direction 2 lanes 3 lanes Roadway Grade Level 3% 6% Free Flow Speed (mi/h) 55 65 75 Segment Length (ft) 1320 2640 3960 5280 HV Percentage 5% 10% 15% 20% Flow Rate (veh/h/ln) 800 1200 1600

PAGE 68

68 It was calculated that six replications of each simulation scenario would be sufficien t to provide a 95% confidence interval (CI) for the obtained density values. Updated Commercial Truck Speed V ersus Distance Grade Curves Similar to the updated PCE values, the introduction of the advanced vehicle dynamics model ing approach raises a need to regenerate a n updated set of commercial truck speed versus distance grade curves. The calculation of the commercial truck speed versus distance grade curves was performed by using TruckSim ® , which uses a full vehicle dynamics modeling approach similar to the one that was developed a nd presented in this dissertation . In agreement with the HCM, the most prevalent (typical) truck type ( Interstate Semi Trailer ) was simulated for segment lengths of 10,000 feet on varying up grade values of 1 to 8 percent as wel l as downgrades values of 1 through 6 . The curves assume a desired speed of 65 mi/h while entering a grade and also assume desired speed values of 65, 60, 55, 50, 45 and 45 mi/h while traveling on downgrades of 1, 2, 3, 4, 5 and 6 percent, respectively. R esults The results subsection summarizes the results obtained by using the advanced vehicle dynamics model ing approach , the PCE estimation equations obtained and the updated commercial truck speed versus distance grade curves. Advanced Vehicle Dynamics Mod eling Approach Once the advanced vehicle dynamics model ing approach with transmission gear changing capability was finalized and integrated into SwashSim , simulations as discussed in the methodology section were run to ensure that the desired transmission gear changes were observed to replicate the field conditions . Figure G 1 through

PAGE 69

69 Figure G 12 in Appendix G present the transmission gear change capable SwashSim versus TruckSim ® results that were obtaine d in these runs. These results show that by introducing the transmission gear changing capabilities of the se vehicles into SwashSim , the maximum acceleration performance of commercial trucks that were modeled match very closely to those in TruckSim ® and th erefore , the commercial truck operation characteristics in the field . It is determined that t he level of accuracy of commercial truck maximum acceleration modeling using the advanced vehicle dynamics mode ling approa ch is superior to those of the earlier mo del ing approache s used in the literature. Through the usage of this approach future users will be able to observe the selected transmission gear of each vehicle in the traffic stream and the corresponding available power, torque and RPM for that specific transmission gear. Therefore , the introduction of the transmission gear changes is very vital to commercial truck performance modeling, due to the fact that the commercial trucks might possibly be able to pick up minor acceleration values on grades by bei ng able to drop multiple transmission gears (e.g. gradually decreasing selected transmission gear 10 to selected transmission gear 6) . This phenomenon was not taken into consideration in any of the previous traffic simulation programs and therefore affect s the traffic stream results on grades negatively. With the addition of the advanced vehicle dynamics modeling approach into SwashSim , a higher level of accuracy to commercial truck performance modeling was introduced into the literature. PCE Estimation E quations Once the experimental design was executed, which resulted in a total of 311,040 simulation runs, the selected PCE calculation methodology was applied to determine

PAGE 70

70 the PCE values for each specific commercial truck class. From these PCE values, a r egression analysis was used to develop equations to estimate the PCE values for each commercial truck type as a function of several explanatory variables. From the analysis, it was determined that there was not much difference in the PCE values between Cl ass 8 and Class 9 trucks. Although they have different average load conditions, their drivetrain and physical characteristics are very similar. And given that Class 9 trucks are much more prevalent in the traffic streams of freeways and multilane highway s than Class 8 trucks, it was decided to use just three separate commercial truck categories for the purposes of PCEs, for which Class 9 would represent the s 3 12 through 3 14 are the resulting PCE estimation equations fr om this study. The se equations correspond to Single Unit (Small) Trucks, Semi tractor+trailer (Medium) T rucks , and Semi tractor+double trailer (Large) T rucks, respectively . It should be noted that all calculated PCE values using the below equations shoul d be rounded to the nearest hundredth. ( 3 12 ) ( 3 13 ) ( 3 14 )

PAGE 71

71 w here , = Passenger car equivalent value for truck type i ( i = ST, MT, or LT for small truck, medium truck and large truck, respectively ) = Segment/Link length in ft. = Proportion of grade (i.e., % grade/100) = Free flow speed in ft/s = Number of lanes in analysis direction = Proportion of single unit trucks in traffic stream (i.e., % ST/100) = Proportion of medium trucks in traffic stream (i.e., % MT/100) = Proportion of large trucks in traffic stream (i.e., % LT/100) = Measured volume in veh/h/ln All of the variables in these three PCE models are statistically significant at a 95 % CI, with respective adjusted R 2 values of 0.7194, 0.7170, and 0.7195. Furthermore, the signs of all the model variables are logical. The PCE values increase as the magnitude of the grade increases and/or the length of the grade. These two variables are included in the model as an interaction term because it is the combined effect of these two variables that are important (e.g., the length of the segment is not important if the grade is level). Furthermore, the impact of this interaction on commercial t ruck performance is not strictly linear; thus, the polynomial form (squared and linear terms). Although the sign of the squared term is positive and the sign of the linear term is negative, the overall effect of this interaction

PAGE 72

72 will be positive. Eventua lly, if the grade is steep enough, or the grade is long enough, a commercial truck will reach its crawl speed, which will create a limit point of the impact of the commercial noted that th e study of th is dissertation did not account for extreme commercial truck performance/roadway combinations that could lead to a commercial truck not being able to at least maintain a crawl speed up the grade. Through testing of the three commercial truck types and various percent grade/length of grade combinations, an approximate value of the product of grade length and grade proportion where commercial trucks reached their crawl speed was identified. This value is 300 and is accounted for in the PCE equa tions through the minimum function. The PCE values increase with an increase in free flow speed. This is logical since as the free flow speed increases, finding acceptable gaps for lane changing maneuvers become more difficult. Note that the minimum free flow speed that should be used in these equations is 45 mi/h (66 ft/s). The PCE values decrease with the number of lanes. This result was as expected since with fewer lanes available for the passenger cars to make a passing maneuver, the impact of commer cial trucks on the traffic stream increases. The PCE values increase with the proportion of commercial trucks. This result is counter to the values in the HCM 2010, where the PCE values decrease with increasing commercial truck percentage. The HCM explai ns this relationship as being due to the tendency of commercial truck drivers to form platoons with one another in the traffic stream and that this platooning effect reduces the relative impact of each commercial truck. This in fact may be true; however, SwashSim does not employ logic to form

PAGE 73

73 platoons between multiple commercial trucks in the traffic stream, although some of this did occur through the lane biasing based on speed logic as discussed earlier. It should be noted that the previous version of C ORSIM that was used to develop the PCE values that are in the HCM 2010 also did not have the platooning logic . The effect of proportion of commercial trucks on PCE is also captured through an interaction term with the traffic flow rate. This effect is als o positive; that is, as the flow rate increases and/or the proportion of commercial trucks increases, the PCE will increase. This term essentially reflects the impact that the number of commercial trucks on the roadway will have on traffic stream operatio ns. In other words, a high commercial truck percent by itself may not have much impact on traffic stream operations if the overall traffic level is low. Note that the minimum flow rate that should be used in these equations is 100 veh/h/ln. T he updated P CE values that would be calculated using Equations 3 12 through 3 14 differ from the HCM PCE values as follows : They r epresent three different classes of trucks rather than one "typical" truck . They are different for each set of inputs and can account for six different variables. As expected , the PCE values for a small truck are less than the PCE value s for a medium truck, which is less than the PCE value s for a large truck given a specific set of input values that include the same percentage of these truck types (e.g. , 5% of ST, 5% of MT, and 5% of LT). It is possible to obtain a higher PCE value for smalle r truck type s compared to larger truck type s, if the overall percentage of the smaller truck s in the traffic stream is much higher than the percentage of larger trucks in the traffic stream. This is due to the effect of the proportion of each truck type on the respective PCE value. Considering the given variables for a roadway segment, the PCE value would be calculated by using Equations 3 12 through 3 14.

PAGE 74

74 Grade: 4% FFS : 95.33 ft/s Trucks: 3 % small, 6 % medium and 5 % large Segment length : 2640 ft L anes in analysis direction : 2 Flow rate: 1200 veh/h/ln Therefore, the development of these three types of PCE values for three categories of commercial trucks can be used to calculate a heavy vehicle adjustment factor ( ), according to Equation 3 15. ( 3 15 ) w here , = Proportion of truck type i in the traffic stream ( i = ST, MT, or LT for small truck, medium truck and large truck, respectively) = 0. 91

PAGE 75

75 The PCE estimation equations provide the ability to estimate PCE values as a function of six explanatory vari ables, and at a much finer resolution than those provided in the HCM in a tabular format. However, to give a general idea of what kind of PCE values are obtained for a certain set of variables, Tab le 3 9 presents the PCE values o btained for each commercial truck type for level, rolling, and mountaino us terrain. The other variable values, in addition to the variable values shown in the table that these PCE values are based on, are given in Table 3 10 . Tab le 3 9 . PCE Comparison by Terrain Type Vehicle Link Length (ft) PCE by Type of Terrain Level Rolling Mountainous 2 lanes 3 lanes 2 lanes 3 lanes 2 lanes 3 l anes Single Unit Truck 2640 1.41 1.32 1.81 1.73 2.48 2.40 5280 1.41 1.32 2.32 2.24 3.17 3.09 Intermediate/Interstate Semi Trailer 2640 1.44 1.36 1.86 1.79 2.60 2.52 5280 1.44 1.36 2.40 2.33 3.34 3.26 Semi tractor+d ouble t railer 2640 1.63 1.53 2.16 2.05 3.01 2.90 5280 1.63 1.53 2.73 2.62 3.78 3.68 Table 3 10 . Terrain Type Specific Input Values Input Values Level Rolling Mountainous Prop. Specific Truck Type 0.05 0.10 0.15 FFS (ft/s) 95.33 95.33 73.33 Prop. Grade 0.00 0.04 0.08 Flow Rate (veh/h/ln) 1200 Although it is difficult to directly compare the PCE values from this dissertation to those of the HCM 2010 because the HCM values are much more generalized, for similar input conditions the PCE values from this study are generally lower . This is due to the effect of the full veh icle dynamics modeling approach that was developed and

PAGE 76

76 incorporated into SwashSim as well as the higher power to weight ratios of the commercial trucks used in this study. Updated Commercial Truck Speed V ersus Distance Grade Curves Once the simulation runs were executed as described under the methodology section , the commercial truck speed versus distance grade curves were obtained fo r upgrade values of 1 through 8 as well as downgrade values of 1 through 6 . Figure 3 9 and Figure 3 10 p resent these curves , respectively . Figure 3 9 . Commercial Truck Speed Versus Distance Grade Curves Upgrade

PAGE 77

77 Figure 3 10 . Commercial Truck Speed Versus Distance G rade Curves Down grade To ensure that these graphs generate accurate final truck speed s after a segment length plus grade combination is traveled by the commercial truck , three example probl ems were generated . Example Problem 1 depicts a scenario where th e commercial truck enters an upgrade of 2% for 5000 feet, followed by an upgrade of 6% for another 5 , 000 feet. In this situation, an average grade approach cannot be utilized per the HCM 2010 since one portion of the grade is steeper than 4% and the total length of the grade is larger than 4 , 000 feet. Therefore, as depicted in Figure 3 11 , the following steps have been followed for the solution process.

PAGE 78

78 A vertical line is drawn at 5,000 feet to intersect the 2% upgrade curve (Poi nt 1). From Point 1 , a horizontal line is drawn to the y axis and the intersection point (Point 2) represents the speed of the commercial truck (63 mi/h) at the end of the 5,000 feet of 2% upgrade. This speed also represents the speed that the commercial t ruck will enter the 5,000 feet of 6% upgrade. Therefore, a horizontal line is drawn to the intersection of the curve for 6% (Point 3). From Point 3 , a vertical line is drawn to the x axis. This point (Point 4) represent s that the commercial truck enter t he 6 % upgrade as if it was on the upgrade for approximately 200 feet. Since the commercial truck will travel another 5,000 feet on the 6% upgrade, Point 5 is found by drawing a horizontal line of 5,000 feet from Point 4. A vertical line is then drawn from Point 5 to the 6% upgrade curve (Point 6). Followed by a horizontal line to the y axis (Point 7) . This point represents the speed of the commercial truck at the end of the two segment composite upgrade. This speed was found to be approximately 36 mi/h gr aphically by using Figure 3 11 and was also validated by coding this two segment scenario into TruckSim ® and running the simulation. TruckSim ® returned a final commercial truck speed of 36.3 mi/h, which is practically the same fi nal speed value as was found by the graphical method.

PAGE 79

79 Figure 3 11 . Example Problem 1 Solution Using Composite Grade Procedure Example Problem 2 represents a scenario where the commercial truck enters a three segment composi te up grade, with the first segment being an upgrade of 2% for 3,334 feet, followed by an upgrade of 5% for another 3,333 feet , and followed by an upgrade of 8% for another 3,333 feet. The speed of the commercial truck at the end of the three segment compo site grade was found to be approximately 29 mi/h by using the same graphical composite grade methodology as described above for Example Problem 1. This solution was presented in Figure 3 12 . Again, in order to validate the

PAGE 80

80 graph ical solution this three segment composite grade scenario was coded into TruckSim ® and the simulation was run. TruckSim ® returned a final commercial truck speed of 28.6 mi/h at the end of the three segment composite grade , which is practically the same fi nal speed value as presented by the graphical method. Figure 3 12 . Example Problem 2 Solution Using Composite Grade Procedure In addition, to ensure the accuracy of the commercial truck downgrade curves a third example prob lem, Example Problem 3, was generated. This problem presents a scenario where the commercial truck enters a three segment composite grade with the

PAGE 81

81 first grade being an upgrade of 3% for 3,334 f ee t, followed by a downgrade of 3% for 3,333 feet and followed by an upgrade of 4% for 3,333 feet. The speed of the commercial truck at the end of the three segment composite grade was found to be approximately 50 mi/h by using the same graphical composite grade methodology as described above for Example Problem 1 , with the difference that there are two graphs being used for this approach since one set of curves are for upgrades and the other for downgrades . Therefore, the solution starts with Figure 3 13 , since the problem starts with an u pgrade. However, once the downgrade starts, the downgrade solution is utilized as presented in Figure 3 14 , only to go back to Figure 3 13 when the last segment starts , which is an upgrade. Again, in ord er to validate this graphical solution th e three segment composite grade scenario was coded into TruckSim ® and the simulation was run. TruckSim ® returned a final commercial truck speed of 49.7 mi/h at the end of the three segment composite grade, which is practically the same final speed value as presented by the graphical method.

PAGE 82

82 Figure 3 13 . Example Problem 3 Upgrade Solution Using Composite Grade Procedure

PAGE 83

83 Figure 3 14 . Example Problem 3 Dow ngrade Solution Using Composite Grade Procedure

PAGE 84

84 CHAPTER 4 ON BOARD DIAGNOSTICS DATA Starting in 1996 , t he second specification for On Board Diagnostics ( OBD II ) became a mandatory component for all cars sold in the U.S . The OBD functionality of a vehi cle provides the capability for operational data on various vehicle sub systems to be collected through a standardized digital communications port . Figure 4 1 shows the photo of an OBD II cable that is attached to the OBD II port of a vehicle. R eal time data such as vehicle speed, engine revolutions per minute ( RPM ), engine fuel rate, calculated engine load, etc. can be obtained . Figure 4 1 . Photo of an OBD II cable attached to the OBD II port of a v ehicle Photo by Seckin Ozkul

PAGE 85

85 This chapter provides an overview of the methodology used to integrate the OBD class into traffic microsimulation . The motivation behind this study was to make the estimation of more accurate vehicle emissions and fuel consu mption rate s possible for future studies through the us e of direct OBD engine parameters since the current estimation models are based on a more limited number of factors . Therefore, the results of this study will provide the foundation which later resear ch can build upon to develop more accurate vehicle emissions and fuel consumption rate models for microsimulation . In addition, this chapter also provides the results obtained by performing this task . Methodology The amount of available OBD data that is r II port is vast. Among these available data are the following. engine coolant temperature short term fuel trim long term fuel trim fuel pressure intake manifold absolute pressure engine RPM vehicl e speed intake air temperature engine air flow rate throttle position oxygen sensor information run time since engine start fuel rail pressure barometric pressure distance traveled ambient air temperature engine oil temperature engine torque data

PAGE 86

86 However, for the purposes of this study, not all available OBD data are required to be coded into the traffic microsimulation program. Therefore, SwashSim is designed to calculate and output the following parameters, which factor into the emissions and fuel rate estimation models. engine speed ( revs/min , also referred to as RPM ) engine torque (lb ft) engine power (hp) calculated engine load (%) vehicle speed (mi/h) selected transmission gear overall gear ratio manifold absolute pressure ( kPa ) The integration of t he OBD data into SwashSim was performed through the development of a C# class for the OBD data generation . This class collect s the necessary variables for OBD data reporting such as RPM, torque, power, selected transmission gear, vehicle velocity, MAP, ca lculated engine load , etc. from other classes within the traffic simulation program and outputs the OBD data directly . Therefore, the user can either observe these variables through the animation screen by simply clicking on a vehicle of their choice or r eview the output file generated at the end of each run . As mentioned in the literature review chapter, Frey et al. (2010) conducted a study regarding the explanatory power and goodness of fit of alternative modeling approaches for predicting vehicle emissi ons and fuel rate. The Frey et al. (2010) study compared models based on use of EOVs with models that are based on IOVs . The study also mentioned that fuel use and emissions depend on what is happening inside the engine and therefore an IOV approach are more predictive than those based only on binned values of VSP as the determining factor . Frey et al. (2010) also reported in

PAGE 87

87 the study that the vehicle emissions and the fuel rate estimation models that are a function of the product of RPM and MAP have an R 2 value of 0.99 compared to the binned VSP approach model R 2 value of 0.87. Thus, an accurate OBD data ( especially RPM and MAP ) estimation capability in a traffic microsimulator would allow for future research to develop more accurate vehicle emi ssions and fuel rate estimation s . As t he first step in generating and implementing the OBD data models into SwashSim , an OBD data collection was performed on a six mile freeway corridor with a posted speed limit of 70 mi/h . This corridor is located on Int erstate 75 (I 75) between the Intersections of I 75 and SW Archer Road and I 75 and NW 39 th Avenue . Figure 4 2 depicts the location of the freeway data collection site in red . Figure 4 2 . Aerial Pho to of Freeway Segment for OBD Data Collection

PAGE 88

88 Th ese field OBD data measurements were conducted using a light duty passenger vehicle, a 2003 Honda Civic LX with a 4 cylinder 1.7 L engine and a 4 speed automatic transmission. Figure 4 3 depicts a photo of the field data measurement vehicle. Figure 4 3 . Photo of the field data measurement vehicle 2003 Honda Civic LX Photo by Seckin Ozkul An OBD scantool (i.e., OBD Link SX Scan © Tool ( Figure 4 4 ) with OBDWiz © diagnostic s software ( Figure 4 5 ) ) was used to record OBD parameters at

PAGE 89

89 approximately 1 Hz frequency including MAP, RPM, engine load, intake manifold absolute pressure , vehicle speed and engine tim e stamp. Figure 4 4 . Photo of the OBD Link SX © Scan Tool Photo by Seckin Ozkul

PAGE 90

90 Figure 4 5 . Screen capture of the OBDWiz © Diagnostic s Software ( OBDWiz © by O CTech, LLC , http://www.obdsoftware.net/OBDwiz.aspx , license obtained through purchase of OBD Link SX © Scan Tool cable) Another field data collection was also performed at an arterial corridor with a posted speed limit of 45 mi/h, located on Newberry Road (SR 26) between the Intersections of SR 26 and NW 98th Street and SR 26 and NW 122 nd Street. Figure 4 6 depicts the location of the arterial data collection site in red.

PAGE 91

91 Figure 4 6 . Aerial Photo of Arterial Segment for Second Set of OBD Data Collection During the data collection effort , a total of 3,355 seconds of valid OBD data were collected in the field for three different types of driving behavior ranging from non aggressive, mo derate, and aggressive. This stratification was deemed necessary for this study in order to observe and account for these driver type differences that affect the internal engine variables of interest (MAP and RPM). After the data collection effort, the im plementation of the OBD field data measurement vehicle into SwashSim was performed. This step required the definition of the field data measurement vehicle in SwashSim using the following vehicle characteristics data : Height Width Length Weight Torque/pow er engine speed curve

PAGE 92

92 Wheel radius Transmission gear ratios Figure 4 7 depicts the torque/power versus engine speed relationship for the data collection vehicle. Note that the lower (blue) curves in each pair of curves correspond to the data collection vehicle used in this study. The upper curves correspond to the same type of vehicle that included a high performance exhaust system. Figure 4 7 . Torque/Power Engine Speed Curves fo r 2003 Honda Civic LX (E Trailer, 2014) Table 4 1 summarizes the remainder of the attributes that were coded into SwashSim in order to replicate the OBD data collection vehicle.

PAGE 93

93 Table 4 1 . 2003 Honda Civic LX Vehicle Characteristics Data 2003 Honda Civic LX Vehicle Height (ft) 4.59 Vehicle Width (ft) 5.56 Vehicle Length (ft) 14.56 Vehicle Weight (lb) 2,474 Maximum Torque (lb ft) 1 05 Maximum Power (hp) 115 Wheel Radius (ft) 1.03 Differential Gea r Ratio 4.07 Transmission Gear Ratios Gear 1 2.722 Gear 2 1.516 Gear 3 0.975 Gear 4 0.674 After the implementation of the OBD data collection vehicle into SwashSim , the OBD field data were reorganized in a format where it can be used to estimate MAP, RPM, and calculated engine load values through regression analysis. The initial step was to estimate the MAP, and to do so a regression analysis of the product of MAP and RPM vs. VSP was performed. Since the engine speed calculations are derived from th e advanced vehicle dynamics approach under Chapter 3 of this dissertation , the only other unknown is VSP. Therefore, the calculation of VSP was performed by using Equation 4 1 . ( 4 1 ) w here , Vehicle Specific Power (kw/ton) Vehicle Speed (km/h) Vehicle Acceleration (km/h/s) Roadway Grade (%)

PAGE 94

94 The freeway and the multilane highway corridors that were used in this study w ere found to have a n average grade of 0.5% to be used in Equation 4 1 . In the earlier Frey et al. work ( 2008 ) Equation 4 1 was the sole basis for determining emission and fuel consumption rates . In this earlier work, the calculated VSP value (which serves as a surrogate for engine load) is compared to a table that subdivides the full range of VSP values into 14 categories (bins). The corresponding emission and fuel consumption rates are read f rom this table for the given VSP bin. This approach has it s limits given the use of just the VSP value and the aggregation of a continuous variable into discrete 2010 ) , which is the approach used in the work for this dissertation, uses VSP directly as a continuous variable in combination with other engine outputs, namely engine speed and MAP. This more disaggregate and multivariate approach leads to more accurate emission and fuel c onsumption estimates. Once the MAP is estimated as described previously , calculated engine load was also estimated through a regression analysis of the product of RPM and MAP versus calculated engine load. Therefore , since the product of MAP and RPM is kn own , the calculated engine load is estimated accordingly. Upon the completion of the model estimations, a validation was performed in order to ensure that SwashSim was able to replicate the field data. In this effort, the custom simulation tool was set up to replicate a varying desired speed scenario (30, 40, and 50 mi/h) similar to the Newberry Road ( Figure 4 6 ) field conditions . Lastly, c omparisons of the Newberry OBD field data were performed against the OBD values

PAGE 95

95 generated b y SwashSim . These comparisons along with the generated OBD models are discussed in depth under the results section. Results After performing the freeway and arterial OBD data collection effort as described under the methodology section , the project team d eveloped Figure 4 8 and Figure 4 9 in order to explain the relationship between the product of MAP and RPM versus VSP as well as the relationship between the product of MAP and RPM versus calculated engin e load . Figure 4 8 . MAP*RPM versus VSP Field Data

PAGE 96

96 Figure 4 9 . MAP*RPM versus Calculated Engine Load Field Data It was observed from these field data, as well as the Frey et al. (2010) study that the model to represent the product of MAP and RPM versus VSP is a power function, whereas the model to represent the product of MAP and RPM versus calculated engine load is a linear one. Running a regression analysis, the below relationship was derived fr om Figure 4 8 with an R 2 value of 0.72. ( 4 2 ) Since the RPM is calculated using the advanced vehicle dynamics equations as highlighted in Chapter 3 of this dissertation, the only unknown in Equation 4 2 is the MAP. Therefore, the estim ation of MAP in the custom traffic simulation tool was performed by coding Equation 4 3 into SwashSim . ( 4 3 )

PAGE 97

97 Once MAP was coded into the custom traffic simulation tool, the calculated engine l oad was determined by running a regression analysis using the data as depicted in Figure 4 9 . This linear model was found to have an R 2 value of 0. 76 and is depicted in Equation 4 4 . ( 4 4 ) Therefore, in order to obtain the calculated engine load value , Equation 4 4 is reorganized into Equation 4 5 . ( 4 5 ) Once the inclusion of these models into SwashSim is finalized and the simulation test runs confirmed the validity of the results returned by SwashSim , the validation process was initiated. In order to validate the OBD data models that were gene rated , simulation runs for a multilane highway of 1.5 miles w e re performed for varying desired speeds of 30 to 70 mi/h so that both the arterial and freeway conditions are covered . The se simulation runs were performed i n order to observe how well the fiel d OBD data set fit the OBD data that was generated using SwashSim. Figure 4 10 and Figure 4 11 depict the OBD data that was collected on Newberry Road for varying speeds of 30, 40 and 50 mi/h as a compari son to the overall OBD field data as depicted in Figure 4 8 and Figure 4 9 .

PAGE 98

98 Figure 4 10 . Newberry Road OBD Field Data MAP*RPM versus VSP Figure 4 11 . Newberry Road OBD Field Data MAP*RPM versus Calculated Engine Load

PAGE 99

99 The relationship between the product of MAP and RPM versus VSP that is depicted in Figure 4 10 was found to have an R 2 value of 0.72 and the relationship between the product of MAP and RPM versus calculated engine load that is depicted in Figure 4 11 was found to have an R 2 value of 0.76. In addition, as expected , the trends of the OBD data that were generated by SwashSim followed the same trends of the OBD data that were collected from the field . This comparison is depicted in Figure 4 12 and Figure 4 13 . Figure 4 12 . Comparison of Field and SwashS im MAP*RPM versus VSP Data

PAGE 100

100 Figure 4 13 . Comparison of Field and SwashSim MAP*RPM versus Calculated Engine Load Data It can be observed from Figure 4 12 that as they should, the SwashSim data points f ollow Equation 4 2 that was coded into it, which matches the field data equation for which the R 2 is 0. 72 . Also , by the analysis of Figure 4 13 it can be observed that the SwashSim data points follow Equation 4 4 that was coded in to it. Similar to the results above, the SwashSim results match ed the field data model that has an R 2 of 0.76 .

PAGE 101

101 Another verification between OBD field data and simulation results was performed to observe the relationship between the engine parameter RPM an d vehicle speed. Also, both of these variables are directly related to selected transmission gear. Figure 4 14 depicts the comparison performed in order to observe the engine speed (revs/min) versus vehicle speed for the field O BD data (depicted with blue markers) and the simulation (depicted with red markers) results. Figure 4 14 . Comparison of Validation Data and SwashSim Engine Speed versus Vehicle Speed It could be observed from Figure 4 14 that SwashSim does a very good job of replicating the 4 speed transmission of the OBD data collection vehicle. In addition, Figure 4 15 depicts the relationship between engine speed , selected transmission ge ar and vehicle speed for the field OBD data ( depicted with blue circular markers in the

PAGE 102

102 original document ) and the simulation ( depicted with red square markers in the original document ) results . Figure 4 15 . Comparison of Val idation Data and SwashSim Engine Speed, Selected Transmission Gear and Vehicle Speed Figure 4 15 clearly depicts that the results generated by SwashSim highly correlate with the field data and th us can be used as a tool for future research on OBD data based vehicle emissions and fuel consumption rate models to improve on the accuracy of the current VSP based models.

PAGE 103

103 CHAPTER 5 SUMMARY AND RECOMMENDATIONS In this research , three significant contributions to the field have been made . The development of an advanced/full vehicle dynamics modeling approach and its integration into traffic microsimulation. The development of updated PCE estimation formulas along with updated truck speed versus distance grade curves using the advanced/full vehicle dynamics modeling approach . The development and integration of OBD data estimation capability in to traffic microsimulation. Due to the fact that CORSIM 6.3 and other leading commercially available traffic simulation programs utilize a simplified ma ximum acceleration calculation methodology that does not take into account the transmission gear changing capabilities of commercial trucks, this study introduced the inclusion of an advanced/full vehicle dynamics modeling approach into SwashSim. As detail ed under Equation s 3 1 through 3 11 , the full vehicle dynamics model ing approach does not assume peak power and calculates engine speed, torque and power for every time step for a selected transmission gear. The selected transmission gear of the vehicle i s updated for each time step according to the gear changing criteria set and therefore is accounted for unlike other commercially available traffic simulation programs. This cap ability ensures that the deceleration on grades, especially for commercial tru cks, is not overestimated (as is generally being done in all commercially available simulation programs due to a full vehicle dynamics modeling approach not being utilized ), and the traffic flow operating conditions are not characterized as being better th an they are . With the addition of the advanced vehicle dynamics modeling

PAGE 104

104 approach into SwashSim, a higher level of accuracy to commercial truck performance modeling in simulation was developed . Due to the fact that the incorporation of the full vehicle dy namics modeling approach into traffic microsimulation improv es the commercial truck performance, the PCE values that are calculated by using microsimulation are also affect ed . Therefore, updated PCE values that are calculated by utilizing the advanced veh icle dynamics modeling approach developed and incorporated into SwashSim are also included as a part of this study . Once the experimental design was executed, which resulted in a total of 311,040 simulation runs, the selected PCE calculation methodology w as applied to determine the PCE values for each specific commercial truck class. From these PCE values, a regression analysis was used to develop equations to estimate the PCE values for each commercial truck type as a function of several explanatory vari ables ( segment length, proportion of grade, FFS, number of lanes, proportion of truck type, and flow rate ) . Equations 3 12 through 3 14 are the resulting PCE estimation equations as a result of this study. They correspond to Single Unit Trucks (Small), Se mi tractor+trailer trucks (Medium), and Semi tractor+double trailer trucks (Large), respectively. As expected, the PCE values obtained using the proposed advanced vehicle dynamics modeling approach incorporat ed into SwashSim were smaller than the current HCM values . Similar to the updated PCE values, the introduction of the advanced vehicle dynamics modeling approach raise d a need to revise the HCM commercial truck speed versus distance grade curves.

PAGE 105

105 Similar to the HCM, the most prevalent (typical) truck t ype (Interstate Semi Trailer) was simulated using TruckSim ® for segment lengths of 10,000 feet on varying upgrade values of 1 to 8 percent as well as downgrades values of 1 to 6. Figure 3 9 and Figure 3 10 present these updated commercial truck speed versus distance grade curves, respectively. The curves assume a desired speed of 65 mi/h while entering a grade and also assume desired speed values of 65, 60, 55, 50, 45 and 45 mi/h while traveling on dow ngrades of 1, 2, 3, 4, 5 and 6 percent, respectively. The reasoning behind the downgrade desired speed values was to prevent the runaway truck phenomenon. Improved commercial truck performance on grades are observed as highlighted in this dissertation du e to the fact that the trucks are more capable on grades since their gear changing capabilities are being accounted for in TruckSim ® . Lastly, with the goal of setting a baseline for more accurate vehicle emissions and fuel consumption estimations using tra ffic microsimulation in mind, more accurate OBD data, especially RPM and MAP, estimation capability in SwashSim was developed. T he integration of the OBD data into SwashSim was performed through the development of a separate C# class for the OBD data gene ration. This class collects the necessary variables for OBD data reporting such as RPM, torque, power, selected transmission gear, vehicle velocity, MAP, calculated engine load, etc. from other classes within the traffic simulation program and outputs the OBD data directly. Therefore, the user can either observe these variables through the animation screen by simply clicking on a vehicle of their choice or review the output file generated at the end of each run.

PAGE 106

106 Two separate OBD data collection efforts ha ve been performed for this study. The first data collection was performed at a freeway facility as depicted in Figure 4 2 . The second data collection effort was performed at a multilane highway facility as depicted in Figure 4 6 . U sing th ese data , OBD data models were estimated and incorporated into SwashSim. The OBD field data results were compared with the SwashSim results and it was determined that the se results replicate the field OBD data very clo sely. As mentioned before, t his OBD data generation capability in traffic microsimulation (SwashSim) also creates a foundation for future studies to build more accurate vehicle emissions and fuel use models that are based on these internal engine OBD data such as RPM, MAP, engine load, etc . As a result of the research efforts described in this dissertation, it is recommended that the H ighway C apacity and Q uality of S ervice (HCQS) committee of the Transportation Research Board (TRB) consider the updated PCE results along with the updated commercial truck speed versus distance grade results to be incorporated into the next update of the HCM. It is also recommended that future studies of OBD data estimations be performed using additional vehicle types ( e.g. , pi ck up trucks, sport utility vehicles, mini vans, high performance passenger cars , etc.) so that the overall simulation of traffic will be more representative of current real world traffic stream s. T he following future studies could be performed using Swash Sim (because commercial truck performance is modeled more realistically then in other simulation programs) in order to build on the concepts introduced into literature by this study.

PAGE 107

107 Commercial truck restricted lane studies . Any roadway that is being cons idered for a possibility of a restricted commercial truck lane addition can be more accurately analyzed by using the results as highlighted in this dissertation . Runaway truck studies. Similar to the above mentioned reasons, these types of studies will al so benefit from the results as highlighted in this dissertation. Commercial truck platoon ing studies and their effect on PCE values . There is a notion in transportation engineering that when the amount of trucks in the traffic stream increases up to a cer tain percentage, there is a platooning effect observed. With the usage of SwashSim along with a data collection effort , this platooning effect can be verified and the possible effect on PCE values can be analyzed. Development of vehicle emissions and fuel rate models based on OBD data . The current work discussed in this dissertation lays the foundation for this future study, since SwashSim is currently ready to receive vehicle emissions and fuel consumption estimation equations (from this future study) ba sed on OBD data in order to generate these results. SwashSim can also serve as a platform for testing in vehicle apps based on OBD data. This can be achieved through the addition of outputting the OBD data that is generated through the simulation through the USB port in a manner that is compatible with OBD port computer connection cables.

PAGE 108

108 APPENDIX A WIM STATION SITE DATA Table A 1 . WIM Station 57 0291 (Okaloosa) Site Data Year Total Truck Volume Passenger Cars Single Unit Tru cks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volum e % of AADT Volum e % of AADT Volume % of AADT Volum e % of AADT 2008 584 4.21 13290 95.79 365 2.63 215 1.55 4 0.03 13874 2009 586 4.1 0 13709 95.9 0 373 2.61 209 1.46 4 0.03 1 4295 2010 579 3.98 13978 96.02 376 2.58 201 1.38 3 0.02 14557 Table A 2 . WIM Station 57 0291 (Okaloosa) Site Data per Truck Class Truck Class Volume % of AADT 5 267 1.84 6 74 0.51 7 24 0.16 8 94 0.65 9 104 0.71 10 3 0.0 2 11 3 0.02 12 0 0 .00 13 0 0 .00 Total 569 3.91

PAGE 109

109 Table A 3 . WIM Station 54 9901 (Jefferson) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 5868 25.38 17252 74.62 883 3.82 4754 20.56 231 1 .00 23120 2009 5115 21.01 19230 78.99 896 3.68 4037 16.58 183 0.75 24345 2010 5318 21.07 19923 78.93 866 3.43 4263 16.89 1 89 0.75 25241 Table A 4 . WIM Station 54 9901 (Jefferson) Site Data per Truck Class Truck Class Volume % of AADT 5 658 2.61 6 120 0.48 7 5 0.02 8 314 1.24 9 3926 15.54 10 27 0.11 11 120 0.48 12 59 0.23 13 9 0.04 Total 5238 20.75

PAGE 110

110 Table A 5 . WIM Station 26 9904 (Alachua) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volum e % of AADT Volum e % of AAD T Volume % of AADT Volum e % of AADT 2008 13274 22.04 46951 77.96 2686 4.46 9925 16.48 662 1.1 0 60225 2009 11120 18.16 50111 81.84 2082 3.4 0 8511 13.9 0 527 0.86 61231 2010 10918 17.79 50449 82.21 2259 3.68 8138 13.26 522 0.85 61367 Table A 6 . WIM Station 26 9904 (Alachua) Site Data per Truck Class Truck Class Volume % of AADT 5 1854 3.02 6 208 0.34 7 24 0.04 8 749 1.22 9 7347 11.97 10 45 0.07 11 301 0.49 12 206 0.34 13 15 0.02 Total 10749 17.51

PAGE 111

111 Table A 7 . WIM Station 72 9905 (Duval) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AAD T 2008 11653 13.9 0 72189 86.1 0 2767 3.3 8442 10.07 444 0.53 83842 2009 10628 12.51 74316 87.49 2481 2.92 7765 9.14 382 0.45 84944 2010 10716 12.21 77055 87.79 2431 2.77 7890 8.99 395 0.45 87771 Table A 8 . WIM Station 72 99 05 (Duval) Site Data per Truck Class Truck Class Volume % of AADT 5 1826 2.08 6 441 0.5 0 7 21 0.02 8 932 1.06 9 6896 7.86 10 61 0.07 11 270 0.31 12 105 0.12 13 16 0.02 Total 10568 12.04

PAGE 112

112 Table A 9 . WIM Station 79 990 6 (Volusia) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 7916 8.55 84662 91.45 3250 3. 51 4407 4.76 259 0.28 92578 2009 7790 8.23 86874 91.77 3228 3.41 4335 4.58 227 0.24 94664 2010 7525 7.89 87844 92.11 2918 3.06 4368 4.58 238 0.25 95369 Table A 10 . WIM Station 79 9906 (Volusia) Site Data per Truck Class Truc k Class Volume % of AADT 5 2355 2.47 6 350 0.37 7 49 0.05 8 620 0.65 9 3716 3.9 0 10 31 0.03 11 151 0.16 12 75 0.08 13 9 0.01 Total 7356 7.72

PAGE 113

113 Table A 11 . WIM Station 46 9907 (Bay) Site Data Year Total Truck Volume Pa ssenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 1370 9.43 13158 90.57 545 3.75 818 5.63 7 0.05 14528 2009 1500 10.11 13335 8 9.89 616 4.15 877 5.91 7 0.05 14835 2010 1468 10.31 12770 89.69 592 4.16 867 6.09 9 0.06 14238 Table A 12 . WIM Station 46 9907 (Bay) Site Data per Truck Class Truck Class Volume % of AADT 5 458 3.22 6 87 0.61 7 8 0.06 8 24 2 1.7 0 9 618 4.34 10 7 0.05 11 6 0.04 12 2 0.01 13 2 0.01 Total 1430 10.04

PAGE 114

114 Table A 13 . WIM Station 34 9909 (Levy) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Tr ucks Total AADT Volume % of AADT Volum e % of AADT Volum e % of AADT Volume % of AADT Volum e % of AADT 2008 988 7.8 0 11673 92.2 0 470 3.71 510 4.03 8 0.06 12661 2009 900 7.31 11413 92.69 430 3.49 438 3.56 32 0.26 12313 2010 878 7.1 0 11486 92.9 0 404 3.27 438 3.54 36 0.29 12364 Table A 14 . WIM Station 34 9909 (Levy) Site Data per Truck Class Truck Class Volume % of AADT 5 338 2.75 6 48 0.39 7 3 0.02 8 146 1.18 9 289 2.34 10 3 0.02 11 2 0.02 12 5 0.04 13 29 0.23 Total 863 6.99

PAGE 115

115 Table A 15 . WIM Station 97 9913 (Turnpike) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AAD T Volume % of AADT Volume % of AADT 2008 No data available for 2008 39784 2009 3869 4.04 34976 90.04 1569 4.04 1958 5.04 342 0.88 38845 2010 3967 4.01 34850 89.78 1557 4.01 2077 5.35 334 0.86 38817 Table A 16 . WIM Station 97 9913 (Turnpike) Site Data per Truck Class Truck Class Volume % of AADT 5 1244 3.21 6 145 0.37 7 15 0.04 8 377 0.97 9 1686 4.34 10 17 0.04 11 160 0.41 12 92 0.24 13 83 0.21 Total 3819 9.83

PAGE 116

116 Table A 1 7 . WIM Station 72 9914 (Duval) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 9843 14.48 58118 85.52 35 01 5.15 5927 8.72 415 0.61 67961 2009 8566 13.01 57266 86.99 3088 4.69 5122 7.78 356 0.54 65832 2010 8290 12.73 56835 87.27 2866 4.4 0 5073 7.79 352 0.54 65125 Table A 18 . WIM Station 72 9914 (Duval) Site Data per Truck Class Truck Class Volume % of AADT 5 1825 2.8 0 6 913 1.4 0 7 23 0.04 8 605 0.93 9 4410 6.77 10 59 0.09 11 271 0.42 12 65 0.1 0 13 11 0.02 Total 8182 12.57

PAGE 117

117 Table A 19 . WIM Station 48 9916 (Escambia) Site Data Year Total Tru ck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 1667 5.42 29091 94.58 840 2.73 797 2.59 31 0.1 30758 2009 1873 5 .97 29486 94.03 1016 3.24 828 2.64 28 0.09 31359 2010 1990 6.31 29545 93.69 1041 3.3 918 2.91 32 0.1 31535 Table A 20 . WIM Station 48 9916 (Escambia) Site Data per Truck Class Truck Class Volume % of AADT 5 732 2.32 6 264 0. 84 7 29 0.09 8 192 0.61 9 713 2.26 10 14 0.04 11 19 0.06 12 9 0.03 13 2 0.01 Total 1974 6.26

PAGE 118

118 Table A 21 . WIM Station 70 9919 (Brevard) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 No data available for 2008 39500 2009 No data available for 2009 N/A 2010 No data available for 2010 N/A Table A 22 . WIM Station 70 9919 (Brevard) Site Data per Truck Class Truck Class Volume % of AADT 5 No data available 6 No data available 7 No data available 8 No data available 9 No data available 10 No data available 11 No data availa ble 12 No data available 13 No data available Total No data available

PAGE 119

119 Table A 23 . WIM Station 72 9923 (Duval) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks To tal AADT Volume % of AADT Volum e % of AADT Volum e % of AADT Volume % of AADT Volum e % of AADT 2008 No data available for 2008 N/A 2009 8635 14.27 51880 85.73 1936 3.2 0 6433 10.63 266 0.44 60515 2010 9308 15.22 51852 84.78 2012 3.29 6978 11.41 318 0.5 2 61160 Table A 24 . WIM Station 72 9923 (Duval) Site Data per Truck Class Truck Class Volume % of AADT 5 1520 2.49 6 336 0.55 7 10 0.02 8 554 0.91 9 6342 10.37 10 82 0.13 11 184 0.3 0 12 117 0.19 13 18 0.03 Total 9163 14.99

PAGE 120

120 Table A 25 . WIM Station 10 9926 (Hillsborough) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volum e % of AADT Volum e % of AA DT Volume % of AADT Volum e % of AADT 2008 No data available for 2008 13263 0 2009 9548 7.22 12267 8 92.78 3875 2.93 5554 4.2 0 119 0.09 13222 6 2010 9634 7.32 12197 9 92.68 3751 2.85 5751 4.37 132 0.1 0 13161 3 Table A 26 . WIM St ation 10 9926 (Hillsborough) Site Data per Truck Class Truck Class Volume % of AADT 5 3001 2.28 6 508 0.39 7 98 0.07 8 1779 1.35 9 3868 2.94 10 101 0.08 11 82 0.06 12 33 0.03 13 19 0.01 Total 9489 7.21

PAGE 121

121 Table A 27 . W IM Station 16 9927 (Polk) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 2182 14.5 0 1286 7 85.5 778 5.17 1395 9.27 9 0.06 15049 2009 1961 13.04 13080 86.96 693 4.61 1260 8.38 8 0.05 15041 2010 1966 13.25 12873 86.75 686 4.62 1276 8.6 0 4 0.03 14839 Table A 28 . WIM Station 16 9927 (Polk) Site Data per Truck Class Truck Class Volume % of AADT 5 448 3.02 6 201 1.35 7 9 0.06 8 192 1.29 9 1082 7.29 10 3 0.02 11 0 0 .00 12 2 0.01 13 3 0.02 Total 1940 13.06

PAGE 122

122 Table A 29 . WIM Station 79 9929 (Volusia) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 522 4.41 11321 95.59 397 3.35 126 1.06 0 0 .00 11843 2009 462 3.96 11204 9 6.04 356 3.05 105 0.9 0 1 0.01 11666 2010 461 4.03 10979 95.97 345 3.02 113 0.99 2 0.02 11440 Table A 30 . WIM Station 79 9929 (Volusia) Site Data per Truck Class Truck Class Volume % of AADT 5 292 2.55 6 34 0.3 0 7 3 0.03 8 91 0.8 0 9 19 0.17 10 2 0.02 11 0 0 .00 12 0 0 .00 13 2 0.02 Total 443 3.89

PAGE 123

123 Table A 31 . WIM Station 97 9931 (Turnpike) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volum e % of AADT Volum e % of AADT Volume % of AADT Volum e % of AADT 2008 No data available for 2008 35858 2009 5406 15.04 30535 84.96 1463 4.07 3512 9.77 431 1.2 0 35941 2010 5622 15.1 0 31613 84.9 1545 4.15 3671 9.8 6 406 1.09 37235 Table A 32 . WIM Station 97 9931 (Turnpike) Site Data per Truck Class Truck Class Volume % of AADT 5 1242 3.34 6 159 0.43 7 8 0.02 8 758 2.04 9 2890 7.76 10 21 0.03 11 239 0.64 12 162 0.44 13 5 0.01 T otal 5484 14.71

PAGE 124

124 Table A 33 . WIM Station 97 9933 (Turnpike) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 No data available for 2008 73500 2009 2888 3.7 0 75159 96.3 1866 2.39 960 1.23 62 0.08 78047 2010 2883 3.63 76539 96.37 1898 2.39 921 1.16 64 0.08 79422 Table A 34 . WIM Station 97 9933 (Turnpike) Site Data per Truck Class Truck Class Volume % of AADT 5 1547 1.95 6 213 0.27 7 61 0.08 8 296 0.37 9 618 0.78 10 10 0.01 11 30 0.04 12 20 0.03 13 4 0.01 Total 2799 3.54

PAGE 125

125 Table A 35 . WIM Station 97 9934 (Turnpike) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 5937 7.36 74720 92.64 34 52 4.28 2380 2.95 105 0.13 80657 2009 5193 6.45 75309 93.55 3019 3.75 2069 2.57 105 0.13 80502 2010 5164 6.19 78247 93.81 2895 3.47 2152 2.58 117 0.14 83411 Table A 36 . WIM Station 97 9934 (Turnpike) Site Data per Truck Clas s Truck Class Volume % of AADT 5 1786 2.14 6 632 0.76 7 387 0.46 8 587 0.7 9 1544 1.85 10 26 0.03 11 58 0.07 12 36 0.04 13 21 0.03 Total 5077 6.08

PAGE 126

126 Table A 37 . WIM Station 29 9936 (Columbia) Site Data Year Total Truc k Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volum e % of AADT Volum e % of AADT Volume % of AADT Volum e % of AADT 2008 No data available for 2008 20000 2009 4493 22.31 15647 77.69 691 3.43 3619 17.97 183 0.91 20140 2010 4759 23.24 15717 76.76 674 3.29 3893 19.01 192 0.94 20476 Table A 38 . WIM Station 29 9936 (Columbia) Site Data per Truck Class Truck Class Volume % of AADT 5 506 2.47 6 88 0.43 7 3 0.01 8 346 1.69 9 3521 17.2 0 10 25 0.12 11 138 0.67 12 47 0.23 13 8 0.04 Total 4682 22.86

PAGE 127

127 Table A 39 . WIM Station 58 9937 (Santa Rosa) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 No data available for 2008 12600 2009 570 4.43 12292 95.57 570 4.43 386 3 8 0.06 12862 2010 549 4.29 12251 95.71 549 4.29 380 2 .97 8 0.06 12800 Table A 40 . WIM Station 58 9937 (Santa Rosa) Site Data per Truck Class Truck Class Volume % of AADT 5 320 2.5 0 6 46 0.36 7 8 0.06 8 81 0.63 9 78 0.61 10 3 0.02 11 2 0.02 12 2 0.02 13 2 0.02 Total 542 4.24

PAGE 128

128 Table A 41 . WIM Station 50 9940 (Gadsden) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Vol ume % of AADT Volume % of AADT 2008 488 5.91 7778 94.09 223 2.7 265 3.21 0 0 8266 2009 No data available for 2009 7600 2010 544 6.77 7497 93.23 274 3.41 269 3.35 1 0.01 8041 Table A 42 . WIM Station 50 9940 (Gadsden) Site D ata per Truck Class Truck Class Volume % of AADT 5 225 2.8 0 6 38 0.47 7 3 0.04 8 74 0.92 9 194 2.41 10 2 0.02 11 0 0 .00 12 0 0 .00 13 1 0.01 Total 537 6.67

PAGE 129

129 Table A 43 . WIM Station 87 9947 (Miami Dade) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volum e % of AADT Volum e % of AADT Volume % of AADT Volum e % of AADT 2008 No data available for 2008 N/A 2009 No data available f or 2009 32783 2010 4605 14.12 28006 85.88 2854 8.75 1738 5.33 13 0.04 32611 Table A 44 . WIM Station 87 9947 (Miami Dade) Site Data per Truck Class Truck Class Volume % of AADT 5 1504 4.61 6 982 3.01 7 273 0.84 8 435 1.33 9 1280 3.93 10 22 0.07 11 5 0.02 12 2 0.01 13 4 0.01 Total 4507 13.83

PAGE 130

130 Table A 45 . WIM Station 16 9948 (Polk) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 No data available for 2008 N/A 2009 3107 13.94 19183 86.06 1099 4.93 1964 8.81 45 0.2 0 22290 2010 3040 13.71 19138 86.29 1120 5.05 1878 8.47 42 0.19 22178 Table A 46 . WIM Station 16 9948 (Polk) Site Data per Truck Class Truck Class Volume % of AADT 5 873 3.94 6 169 0.76 7 11 0.05 8 305 1.38 9 1561 7.04 10 11 0.05 11 22 0.1 0 12 17 0.08 13 3 0.01 Total 2972 13.41

PAGE 131

131 Table A 47 . WIM Station 48 9949 (Escambia) Site Data Year Total Truck Volume Passenger Cars Single Unit Trucks Combo Trailer Trucks Multi Trailer Trucks Total AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT Volume % of AADT 2008 No data available for 2008 N/A 2009 No data available for 2009 N/A 2010 5551 12.01 40684 87.99 1937 4.19 3462 7.49 153 0.33 46235 Table A 48 . WIM Station 48 9949 (Escambia) Site Data per Truck Class Truck Class Volume % of AADT 5 1505 3.26 6 298 0.65 7 17 0.04 8 687 1.49 9 2744 5.94 10 30 0.06 11 96 0.21 12 44 0.1 0 13 9 0.02 Total 5430 11.77

PAGE 132

132 APPENDIX B CUSTOM WIM STATION DATA PROCESSOR CONTROL WINDOWS Figure B 1 . Control Window for Data File Manipulation

PAGE 133

133 Figure B 2 . Control Window for Data Processor

PAGE 134

134 APPENDIX C SAMPLE COMMERCIAL TRUCK CHARACTERISTICS RESULTS Table C 1 . Commercial Truck Characte ristics Results Year 2008 Urban Areas Full Day Morning Peak Mid Day Evening Peak Number Of Vehicle in Class1 0 0 0 0 Number Of Vehicle in Class2 0 0 0 0 Number Of Vehicle in Class3 0 0 0 0 Number Of Vehicle in Class4 158380 30567 9412 23107 Numbe r Of Vehicle in Class5 2918200 571190 194292 466929 Number Of Vehicle in Class6 1387521 289196 115775 130237 Number Of Vehicle in Class7 291650 77969 27206 9644 Number Of Vehicle in Class8 887198 173944 57124 108365 Number Of Vehicle in Class9 7504628 1049607 451856 959193 Number Of Vehicle in Class10 64488 11130 5247 9075 Number Of Vehicle in Class11 339054 39130 3737 24827 Number Of Vehicle in Class12 106493 13097 2253 7248 Number Of Vehicle in Class13 32508 2850 1569 6521 Number Of Vehicle in Cl ass14 0 0 0 0 Number Of Vehicle in Class15 0 0 0 0 Total Number of Vehicles 13690120 2258680 868471 1745146 Avg Weight of Class1 Avg Weight of Class2 Avg Weight of Class3 Avg Weight of Class4 27.021 26.783 27.033 27.563

PAGE 135

135 Table C 1. Continued Full Day Morning Peak Mid Day Evening Peak Avg Weight of Class5 14.498 15.014 14.626 13.487 Avg Weight of Class6 31.385 33.218 32.707 27.643 Avg Weight of Class7 65.046 65.486 64.952 63.681 Avg Weight of Class8 37.758 40.132 37.259 3 3.997 Avg Weight of Class9 51.420 51.814 49.251 51.373 Avg Weight of Class10 61.722 62.261 62.819 60.557 Avg Weight of Class11 53.893 51.534 51.610 54.057 Avg Weight of Class12 54.036 53.800 56.323 52.981 Avg Weight of Class13 73.024 79.688 82.893 66. 492 Avg Weight of Class14 Avg Weight of Class15 Average Weight Overall 40.823 39.454 38.933 38.288 Percent Vehicles in Class1 0 0 0 0 Percent Vehicles in Class2 0 0 0 0 Percent Vehicles in Class3 0 0 0 0 Percent Vehicles in Class4 0. 0116 0.0135 0.0108 0.0132 Percent Vehicles in Class5 0.2132 0.2529 0.2237 0.2676 Percent Vehicles in Class6 0.1014 0.1280 0.1333 0.0746 Percent Vehicles in Class7 0.0213 0.0345 0.0313 0.0055 Percent Vehicles in Class8 0.0648 0.0770 0.0658 0.0621 Perce nt Vehicles in Class9 0.5482 0.4647 0.5203 0.5496 Percent Vehicles in Class10 0.0047 0.0049 0.0060 0.0052

PAGE 136

136 Table C 1. Continued Full Day Morning Peak Mid Day Evening Peak Percent Vehicles in Class11 0.0248 0.0173 0.0043 0.0142 Percent Vehicles in Cla ss12 0.0078 0.0058 0.0026 0.0042 Percent Vehicles in Class13 0.0024 0.0013 0.0018 0.0037 Percent Vehicles in Class14 0 0 0 0 Percent Vehicles in Class15 0 0 0 0 Avg Speed in Class1 Avg Speed in Class2 Avg Speed in Class3 Avg Speed in Class4 65.397 64.661 66.271 64.312 Avg Speed in Class5 65.314 65.028 65.232 65.091 Avg Speed in Class6 62.501 62.123 62.657 62.226 Avg Speed in Class7 58.960 58.597 59.343 58.525 Avg Speed in Class8 63.548 63.122 63.426 62.986 Avg Speed in Cl ass9 64.930 64.162 64.997 64.374 Avg Speed in Class10 65.628 64.994 65.832 64.948 Avg Speed in Class11 63.772 63.667 63.753 63.317 Avg Speed in Class12 66.095 65.509 67.225 66.126 Avg Speed in Class13 65.811 64.279 61.913 65.625 Avg Speed in Class14 Avg Speed in Class15 Average Speed Overall 64.540 63.858 64.471 64.286 Avg Length by Class1

PAGE 137

137 Table C 1. Continued Full Day Morning Peak Mid Day Evening Peak Avg Length by Class2 Avg Length by Class3 Avg Lengt h by Class4 40.229 40.136 39.726 40.830 Avg Length by Class5 28.371 28.313 28.580 27.860 Avg Length by Class6 29.407 29.567 29.581 29.224 Avg Length by Class7 27.427 27.742 27.207 27.613 Avg Length by Class8 56.597 54.708 55.504 58.675 Avg Length by C lass9 68.371 67.426 68.110 69.164 Avg Length by Class10 72.791 72.407 72.552 73.201 Avg Length by Class11 75.031 74.829 75.137 75.172 Avg Length by Class12 78.433 78.219 80.339 80.156 Avg Length by Class13 93.371 92.994 95.241 92.408 Avg Length by Cla ss14 Avg Length by Class15

PAGE 138

138 APPENDIX D TRUCK MAXIMUM ACCELERATION V ERSUS SPEED CURVES Figure D 1 . Maximum Acceleration vs. Speed Curve Single Unit Truck on a Level Grade -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0 20 40 60 80 100 120 Acceleration (ft/s 2 ) Speed (ft/s) TSPM Effective Acceleration HP-Torque Acceleration Rakha Acceleration

PAGE 139

139 Figure D 2 . Maximum Acceleration vs. Speed Curve Class 8 Truck on a Level Grade -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0 20 40 60 80 100 120 Acceleration (ft/s 2 ) Speed (ft/s) TSPM Effective Acceleration HP-Torque Acceleration Rakha Acceleration

PAGE 140

140 Figure D 3 . Maximum Acceleration vs. Speed Curve Class 9 Truck on a Level Grade -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0 20 40 60 80 100 120 Acceleration (ft/s 2 ) Speed (ft/s) TSPM Effective Acceleration HP-Torque Acceleration Rakha Acceleration

PAGE 141

141 Figure D 4 . Maximum Acceleration vs. Speed Curve Class 11&12 Truck on a Level Grade -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0 20 40 60 80 100 120 Acceleration (ft/s 2 ) Speed (ft/s) TSPM Effective Acceleration HP-Torque Acceleration Rakha Acceleration

PAGE 142

142 Figure D 5 . Maximum Acceleration vs. Speed Curve Singe Unit Truck on a 5% Grade

PAGE 143

143 Figure D 6 . Maximum Acceleration vs. Speed Curve Class 8 Truck on a 5% Grade -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0 20 40 60 80 100 120 Acceleration (ft/s 2 ) Speed (ft/s) TSPM Effective Acceleration HP-Torque Acceleration Rakha Acceleration

PAGE 144

144 Figure D 7 . Maximum Acceleration vs. Speed Curve Class 9 Truck on a 5% Grade -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 0 20 40 60 80 100 120 Acceleration (ft/s 2 ) Speed (ft/s) TSPM Effective Acceleration HP-Torque Acceleration Rakha Acceleration

PAGE 145

145 Figure D 8 . Maximum Acceleration vs. Speed Curve Class 11&12 Truck on a 5% Grade -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 0 20 40 60 80 100 120 Acceleration (ft/s 2 ) Speed (ft/s) TSPM Effective Acceleration HP-Torque Acceleration Rakha Acceleration

PAGE 146

146 APPENDIX E COMMERCIAL TRUCK PERFORMANCE COMPARISON CURVES SWASHSIM WITH RAKHA MODEL VERSUS TRUCKSIM ® Figure E 1 . Maximum Acceleration of a Single Unit Truck on a 1320 footLink with 9% Grade Figure E 2 . Ve locity of a Single Unit Truck on a 1320 foot Link with 9% Grade

PAGE 147

147 Figure E 3 . Maximum Acceleration of an Intermediate Semi trailer on a 1760 foot Link with 6% Grade Figure E 4 . Velocity of an Inte rmediate Semi trailer on a 1760 foot Link with 6% Grade

PAGE 148

148 Figure E 5 . Maximum Acceleration of an Intermediate Semi trailer on a 1760 foot Link with 9% Grade Figure E 6 . Velocity of an Intermediate Semi trailer on a 1760 foot Link with 9% Grade

PAGE 149

149 Figure E 7 . Maximum Acceleration of an Interstate Semi trailer on a 2640 foot Link with 6% Grade Figure E 8 . Velocity of an Interstate Semi traile r on a 2640 foot Link with 6% Grade

PAGE 150

150 Figure E 9 . Maximum Acceleration of a Semi tractor+double trailer on a 1760 foot Link with 6 % Grade Figure E 10 . Velocity of a Semi tractor+double trailer on a 1760 foot Link with 6 % Grade

PAGE 151

151 APPENDIX F ADVANCED VEHICLE DYNAMICS APPROACH TO COMMERCIAL TRUCK MAXIMUM ACCELERATION MODELING IN THE CUSTOM SIMULATION TOOL EXAMPLE CALCULATION Figure F 1 . Advanced Vehicle Dynamics Approach Example Calculations 1

PAGE 152

152 Figure F 2 . Advanced Vehicle Dynamics Approach Example Calculations 2

PAGE 153

153 APPENDIX G TRANSMISSION GEAR CHANG ING CAPABLE COMMERCIAL TRUCK PERFORMANCE COMPARISON CURVES SWASHSIM WITH FULL VEHICLE DYNAMI CS MODEL VERSUS TRUCKSIM ® Figure G 1 . Transmission Gear Change Capable Maximum Acceleration of a Single Unit Truck on a 1320 foot Link with 6% Grade Figure G 2 . Transmission Gear Change Capable Velocity of a Single Unit Truck on a 1320 foot Link with 6% Grade

PAGE 154

154 Figure G 3 . Transmission Gear Change Capable Maximum Acceleration of a Single Unit Truck on a 1320 foot Link with 9% Grade Figure G 4 . Transmission Gear Change Capable Velocity of a Single Unit Truck on a 1320 foot Link with 9% Grade

PAGE 155

155 Figure G 5 . Transmission Gear Change Capable Maximum Acceleration of an Intermediate Semi trailer on a 1760 foot Link w ith 6% Grade Figure G 6 . Transmission Gear Change Capable Velocity of an Intermediate Semi trailer on a 1760 foot Link with 6% Grade

PAGE 156

156 Figure G 7 . Transmission Gear Change Capable Maximum Accelera tion of an Intermediate Semi trailer on a 1760 foot Link with 9% Grade Figure G 8 . Transmission Gear Change Capable Velocity of an Intermediate Semi trailer on a 1760 foot Link with 9% Grade

PAGE 157

157 Figure G 9 . Transmission Gear Change Capable Maximum Acceleration of an Interstate Semi trailer on a 2640 foot Link with 6% Grade Figure G 10 . Transmission Gear Change Capable Velocity of an Interstate Semi trailer on a 2640 foo t Link with 6% Grade

PAGE 158

158 Figure G 11 . Transmission Gear Change Capable Maximum Acceleration of a Semi tractor+double trailer on a 1760 foot Link with 6 % Grade Figure G 12 . Transmission Gear Change C apable Velocity of a Semi tractor+double trailer on a 176 0 foot Link with 6 % Grade

PAGE 159

159 LIST OF REFERENCES Akçelik, R., & Besley, M. (2001). Acceleration and deceleration models. In 23rd Conference of Australian Institutes of Transport Research (CAITR 2001), M onash University, Melbourne, Australia , 10 12. Akçelik, R., & Biggs, D. C. (1987). Acceleration profile models for vehicles in road traffic. Transportation Science , 21 (1), 36 54. Al Kaisy, A. F., Hall, F. L., & Reisman, E. S. (2002). Developing passenger car equivalents for heavy vehicles on freeways during queue discharge flow. Transportation Research Part A: Policy and Practice , 36 (8), 725 742. Al Kaisy, A., Jung, Y., & Rakha, H. (2005). Developing passenger car equivalency factors for heavy vehicles dur ing congestion. Journal of transportation engineering , 131 (7), 514 523. Allison Transmission (n.d.d). Tab Models . Retrieved August 14, 2013 from http://allisontransmis sion.com/commercial/transmissions/#tab models Allison Transmission (n.d.d). Transmission SA5341 . Retrieved August 14, 2013 from http://www. allisontransmission.com/servlet/DownloadFile?Dir=publications/pubs& FileToGet=SA5341EN.pdf Araújo, J. J. (2007). Study of Heavy Vehicles Impact on Highway Infrastructure through Microscopic Traffic Simulation (in Portuguese) . Retrieved December 18, 2013 fr om http://www.teses.usp.br/index.php?option=com_jumi&fileid=17&Itemid=160&id=0 7A3A96AAA70&lang=en . Archilla , A. R., & De Cieza, A. O. F. (1996). Truck performance on Argentinean highways. Transportation Research Record: Journal of the Transportation Research Board , 1555 (1), 114 123. Bester , C. J. (2000). Truck speed profiles. Transportation Research Record: Journal of the Transportation Research Bo ard, 1701(1), 111 115. Ching, P. Y., & Rooney, F. D. (1979). Truck Speeds on Grades in California (No. FHWA CA TO 79 1 Final Rpt.). Cohen, S. L. (2002). Application of car following systems to queue discharge problem at signalized intersections. Transporta tion Research Record: Journal of the Transportation Research Board , 1802 (1), 205 213. Cohen, S. L. (2002). Application of car following systems in microscopic time scan simulation models. Transportation Research Record: Journal of the Transportation Resear ch Board , 1802 (1), 239 247.

PAGE 160

160 Craus, J., Polus, A., & Grinberg, I. (1980). A revised method for the determination of passenger car equivalencies. Transportation Research Part A: General , 14 (4), 241 246. Cunagin, W. D., & Messer, C. J. (1983). Passenger car e quivalents for rural highways (discussion) (No. HS 036 187). Cunha, A. L., Bessa Jr, J. E., & Setti, J. R. (2009). Genetic algorithm for the calibration of vehicle performance models of microscopic traffic simulators. In Progress in Artificial Intelligence (pp. 3 14). Springer Berlin Heidelberg. Cunha, A. L., & Setti, J. R. (2011). Truck Equivalence Factors for Divided, Multilane Highways in Brazil. Procedia Social and Behavioral Sciences , 16 , 248 258. Dzhelekarski, P., & Alexiev, D. (2005). Reading and int erpreting diagnostic data from vehicle obdii system. ELECTRONICS , 5 , 12 22. Eaton Transmission s (n.d.d). Truck Transmissions . Retrieved August 14, 2013 from http://www .eaton.com/Eaton/ProductsServices/Truck/Transmissions/index.htm Elefteriadou, L., Torbic, D., & Webster, N. (1997). Development of passenger car equivalents for freeways, two lane highways, and arterials. Transportation Research Record: Journal of the Tra nsportation Research Board , 1572 (1), 51 58. E Trailer (n.d.d). Honda Civic MF 15712 . Retrieved June 5, 2014 from http://www.etrailer.com/Exhaust/Honda/Civic/2 003/MF15712.html?vehicleid=200 3204853 . Fan, H. S. (1990). Passenger car equivalents for vehicles on Singapore expressways. Transportation Research Part A: General , 24 (5), 391 396. Fitch, J. W. (1993). Motor truck engineering handbook. Training , 2013 , 08 0 5. Frey, H. C., Zhang, K., & Rouphail, N. M. (2008). Fuel use and emissions comparisons for alternative routes, time of day, road grade, and vehicles based on in use measurements. Environmental science & technology, 42 (7), 2483 2489. Frey, H. C., Zhang, K. , & Rouphail, N. M. (2010). Vehicle specific emissions modeling based upon on road measurements. Environmental science & technology , 44 (9), 3594 3600. Godavarty, S., Broyles, S., & Parten, M. (2000). Interfacing to the on board diagnostic system. In Vehicular Technology Conference, 2000. IEEE VTS Fall VTC 2000. 52nd (Vol. 4). IEEE.

PAGE 161

161 Harwood, D. W., Torbic, D. J., Richard, K. R., Glauz, W. D., & Elefeteriadou, L. (2003) . NCHRP 505 Review of Truck Characteristics as Factors in Roadway Design . Washingto n, DC: Transportation Research Board. Highway Capacity Manual (1965). HRB Special report 87. HRB, National Research Council, Washington, DC , 397 . Hoban, C. J., Fawcett, G. J., & Robinson, G. K. (1985). A model for simulating traffic on two lane rural roads : user guide and manual for TRARR Version 3.0 (No. ATM 10A). Huber , M. J. (1982). Estimation of passenger car equivalents of trucks in traffic stream (discussion and closure). Transportation Research Record 869 , 60 70 . I2D (n.d.d). I2D demo . Retrieved Marc h 14, 2013 from http://demo.i2d.co/xportal/xmain?xpid=i2d&xlang=en . International Organization for Standardization (1994). Road vehicles Diagnostic systems: CARB requirements for interchange of digital information . ISO 9141 1 . International Organization for Standardization (1995). Road vehicles Diagnostic systems Keyword protocol 2000 Part 2 . Draft International Standard, ISO/DIS 14230 3 . 1 97. Kalin, I. (2013, Febr uary 5). Apps for Vehicles Challenge Finalists Announced . Retrieved February 12, 2013 from http://energy.gov/articles/apps vehicles challenge finalists announced . Keller , E. L., & Saklas, J. G. (1984). Passenger car equivalents from network simulation. Journal of Transportation Engineering , 110 (4), 397 411. Kenworth Trucks (n.d.d). Trucks . Retrieved December 18, 2013 from http://www.kenworth.com/trucks.aspx Krammes, R. A., & Crowley, K. W. (1986). Passenger car equivalents for trucks on level freeway segments. Transportation Research Record , 1091 , 10 17. Leisch, J. E. (1974). Capacity Analysis Techniques for Design and Op eration of Freeway Facilities (No. FHWA RD 74 24 Final Rpt.). Leisch, J. E., & Leisch, J. P. (1977). New concepts in design speed application. Transportation Research Recor d, 631 , 4 14 . Linzer, E. M., Roess, R. P., & McShane, W. R. (1979). Effect of trucks , buses, and recreational vehicles on freeway capacity and service volume. Transportation Research Record , 699 , 17 26.

PAGE 162

162 Mack Trucks (n.d.d). Trucks . Retrieved December 18, 2013 from http://ww w.macktrucks.com/default.aspx?pageid=9111 Mannering, F.L., and Washburn, S.S. (2012). Principles of Highway Engineering and Traffic Analysis (5th ed). Hoboken, NJ: John Wiley and Sons, Inc. Okura, I., & Sthapit, N. (1995). Passenger car equivalents of hea vy vehicles for uncongested motorway traffic from macroscopic approach. In P roceedings japan society of civil engineers (pp. 73 82 ). Dotoku gakkai. Ozbay , K., Nassif, H., & Goel, S. (2007, June). Propagation characteristics of dynamic information collecte d by in vehicle sensors in a vehicular network. In Intelligent Vehicles Symposium, 2007 IEEE (pp. 1089 1094). IEEE. Ozkul, S., Washburn, S. S., & McLeod, D. S. (2013). Revised Version of the Automobile Level of Service Methodology for Urban Streets in the Highway Capacity Manual 2010. Transportation Research Record: Journal of the Transportation Research Board , 2395 (1), 66 72. International Road Dynamics, Inc., Revision B . Spring Grove, IL. Peterbilt (n.d.d). Pete rbilt Engine MX13 . Retrieved August 19, 2013 from http://www.peterbilt.com/resources/Engine%20Spec%20Sheets/2013%20MX%2 0Spec%20Sheet%20121212.pdf Peterbilt (n.d.d). Peterbilt Engine P X 7 . Retrieved August 19, 2013 from http://www.peterbilt.com/resources/Engine%20Spec%20Sheets/2013%20PX7%2 0Spec%20Sheet%20121212.pdf Peterbilt Trucks (n.d.d). Trucks . Retrieved December 18, 2013 from http://www.peterbilt.com/products/trucks/ Rakha, H., Lucic, I., Demarchi, S. H., Setti, J. R., & Aerde, M. V. (2001). Vehicle dynamics model for predicting maximum truck acceleration levels . Journal of transportation engineering, 127 (5), 418 425. Rakha, H., & Lucic, I. (2002). Variable power vehicle dynamics model for estimating truck accelerations. Journal of transportation engineering , 128 (5), 412 419. Rakha, H., & Ahn, K. (2004). Integration modeling framework for estimating mobile source emissions. Journal of transportation engineering , 130 (2), 183 193. Rakha, H., & Yu, B. (2004). Truck performance curve s reflective of truck and pavement characteristics. Journal of transportation engineering , 130 (6), 753 767.

PAGE 163

163 Rakha, H., Ingle, A., Hancock, K., & Al Kaisy, A. (2007). Estimating truck equivalencies for freeway sections. Transportation Research Record: Journ al of the Transportation Research Board , 2027 (1), 73 84. Rakha, H., Snare, M., & Dion, F. (2004). Vehicle dynamics model for estimating maximum light duty vehicle acceleration levels. Transportation Research Record: Journal of the Transportation Research B oard , 1883 (1), 40 49. Schmitt, R., Sprung, M., Rick, C., & Sedor, J. (2010). Freight Facts and Figures 2010 (No. FHWA HOP 10 058). Schwender, H. C., Normann, O. K., & Granum, J. O. (1957). New Methods of Capacity Determination for Rural Roads in Mountainou s Terrain. Highway Research Board Bulletin , 167 , 10 37. Searle, J. (1999). Equations for Speed, Time and Distance for Vehicles under Maximum Acceleration (No. 1999 01 0078). Smith, G. L. (1970). Commercial Vehicle Performance and Fuel Economy. SAE Paper SP 355 . Society of Automotive Engineers. Society of Automotive Engineers (1975). Truck Ability Prediction Procedure: SAE J688 . Society of Automotive Engineers. Society of Automotive Engineers (1996). Commercial truck and bus SAE recommended procedure for veh icle performance prediction and charting : SAE J2188 . Society of Automotive Engineers. Society of Automotive Engineers (1997). OBD II Scan Tool On Board Diagnostics and multiplex Technical Reports : SAE J1978 . Society of Automotive Engineers. Society of Au tomotive Engineers (1997). Diagnostics for Light and Medium Duty Vehicles Standard Manual : SAE J2178/3 . Society of Automotive Engineers. Society of Automotive Engineers (1997). Diagnostics for Light and Medium Duty Vehicles Standard Manual : SAE J2178/4 . So ciety of Automotive Engineers. Society of Automotive Engineers (1997). E/E diagnostic test modes Diagnostics for Light and Medium Duty Vehicles Standard Manual : SAE J1979 . Society of Automotive Engineers. Society of Automotive Engineers (1997). OBD II sc an tool Diagnostics for Light and Medium Duty Vehicles Standard Manual : SAE J1978 . Society of Automotive Engineers. Society of Automotive Engineers (1997). Enhanced E/E Diagnostic Test Modes Diagnostics for Light and Medium Duty Vehicles Standard Manua l : SAE J2190 . Society of Automotive Engineers.

PAGE 164

164 St John, A. D., & Kobett, D. R. (1978). NCHRP Report 185: Grade Effects on Traffic Flow Stability and Capacity. Transportation Research Board , Washington D.C. St John, A. D., & Harwood, D. W. (1986). TWOPAS us Federal Highway Administration , 7 . Sumner , R., Hill, D., & Shapiro, S. (1984). Segment passenger car equivalent values for cost allocation on urban arterial roads. Transportation Research Part A: General , 18 (5), 399 406. Talbot, D. (2013, Febru ary 1). Programming Cars: Using the GM In Dash SDK , December 2013. Retrieved December 18, 2013 from http://www.codeguru.com/csharp/programming cars using the gm in dash sdk.htm . Texas DOT. (n.d.d). FHWA Vehicle Classification Chart . Retrieved April 26, 2013 from http://onlinemanuals.txdot.gov/txdotmanuals/tri/im ages/FHWA_Classification_Ch art_FINAL.png Transportation Research Board. (1950). Highway Capacity Manual . TRB, National Research Council. Washington, D.C. Transportation Research Board. (1965). Highway Capacity Manual . TRB, National Research Council. Washi ngton, D.C. Transportation Research Board. (1985). Highway Capacity Manual . TRB, National Research Council. Washington, D.C. Transportation Research Board. (2000). Highway Capacity Manual . TRB, National Research Council. Washington, D.C. Transportation Re search Board. (2010). Highway Capacity Manual . TRB, National Research Council. Washington, D.C. Van Aerde, M. (1985). Modelling of traffic flows, assignment and queueing in integrated freeway/traffic signal networks , Ph.D. Thesis , Department of Civil Engin eering, University of Waterloo, Waterloo, Canada. Van Aerde, M., Hellinga, B., Baker, M., & Rakha, H. (1996). INTEGRATION: An overview of traffic simulation features. Transportation Research Records . Department of Civil Engineering, Queen's University, Kin gston, Canada. Van Aerde, M., and Yagar, S. (1983). Capacity, Speed and Platooning Vehicle Equivalents for Two Lane Rural Highways. Transportation Research Record 971 . Transportation Research Board. Washington, DC.

PAGE 165

165 Van Aerde, M., & Yagar, S. (1988). Dynam ic integrated freeway/traffic signal networks: A routing based modelling approach. Transportation Research Part A: General , 22 (6), 445 453. Van Aerde, M., & Yagar, S. (1988). Dynamic integrated freeway/traffic signal networks: Problems and proposed solutio ns. Transportation Research Part A: General , 22 (6), 435 443. Van Aerde, M. W., & Yagar, S. (1990, May). Combining traffic management and driver information in integrated traffic networks. In Road Traffic Control, 1990., Third International Conference on (p p. 11 15). IET. Volvo Trucks (n.d.d). Trucks . Retrieved December 18, 2013 from http://www.volvotrucks.com/trucks/na/en us/Pages/Volvo%20Trucks%20USA.aspx Walton, M. C., & Lee, C. E. (1975). Speed of Vehicles on Grades (No. CFHR 3 8 73 20 1F Final Rpt.). University of Texas at Austin, Center for Highway Research. Watanda, T. (1985). Models for Predicting Vehicle Speeds and Operating Costs Based on Mechanistic Principle s; Theory Quantification. International Bank for Reconstruction and Development . Washington, D.C. Webb, G.M. (1961). Truck Speeds on Grades. Traffic Bulletin 2 . California Division of Highways. Webster, N., & Elefteriadou, L. (1999). A simulation study of truck passenger car equivalents (PCE) on basic freeway sections. Transportation Research Part B: Methodological , 33 (5), 323 336. Werner, A., & Morrall, J. F. (1976). Passenger car equivalencies of trucks, buses, and recreational vehicles for two lane rural highways. Transportation Research Record , (615). Wong. J.Y. (2008). Theory of Ground Vehicles (4th Ed) . Hoboken, NJ: John Wiley and Sons, Inc.

PAGE 166

166 BIOGRAPHICAL SKETCH Seckin Ozkul was born and raised in Istanbul, Turkey, where he attended the prestigious Ka dikoy Anatolian High School . He moved to the United States after finishing h igh school and received his B achelor of Science in Civil Engineering Magna Cum Laude from Auburn University in Auburn, Alabama. During his undergraduate studies, he also held a s ummer internship at Vratsinas Construction Co mpany (VCC) in he accepted a position to join Sprinkle Consulting, Inc., a civil engineering company that is l ocated in Lutz, Florida. During his 3+ y ears at Sprinkle, Seckin worked on a multitude of civil engineering transportation and site drainage projects. Also during his time at Sprinkle, he received his Master of Civil Engineering at the University of South Florida in Tampa, Florida . After his c onsulting engineering work experience , Seckin decided to register to the University of Florida Transportation Engineering Ph.D. program in August 2010 as a research assistant with full research funding. Together with his advisor Dr. Scott Washburn, Seckin worked on multiple projects that focused on level of service analysis updated commercial heavy vehicle passenger car equivalency value calculations using a custom traffic microsimulation tool , and fi nally the implementation of on board diagnostics data parameters into traffic simulation. Seckin has received multiple honors during his Ph.D. studies at the University of Florida including 2 nd and 3 rd place poster competition results at the local and regi onal level as well as an International Road Federation (IRF) Fellow award in 2014. In addition Seckin has served as the vice president of the ITE student chapter at the University of Florida.