A thesis presented to. the faculty of. In partial fulfillment. of the requirements for the degree. Master of Science. Matthew J. Scheer.

Transcription

1 Impact of Pavement Thickness on Load Response of Perpetual Pavement A thesis presented to the faculty of the Russ College of Engineering and Technology of Ohio University In partial fulfillment of the requirements for the degree Master of Science Matthew J. Scheer August Matthew J. Scheer. All Rights Reserved

2 2 This thesis titled Impact of Pavement Thickness on Load Response of Perpetual Pavement by MATTHEW J. SCHEER has been approved for the Department of Civil Engineering and the Russ College of Engineering and Technology by Shad M. Sargand Russ Professor of Civil Engineering Dennis Irwin Dean, Russ College of Engineering and Technology

3 3 ABSTRACT SCHEER, MATTHEW J., M.S., August 2013, Civil Engineering Impact of Pavement Thickness on Load Response of Perpetual Pavement Director of Thesis: Shad M. Sargand This thesis studies the performance of perpetual pavement structures constructed by the Ohio Department of Transportation. The pavement responses, collected from sensors installed in three separate perpetual pavement test sections on U.S. Route 23 in Delaware, Ohio during controlled vehicle load testing, were the main analysis for the study. To complement the pavement responses, a software analysis of the test sections was performed using PerRoad. The main pavement responses measured by test section instrumentation were strain at the bottom of the asphalt layer as well as in the base layer of the asphalt structure, subgrade pressure, pavement deflection, and subgrade deflection. Pavement responses were compared with fatigue endurance thresholds in order to evaluate the longevity of the pavement test sections. Additionally, the influences of several variables, including axle configuration, speed, and tire pressure, were analyzed to further understand their effects on pavement responses. Although controlled vehicle load testing was conducted during periods of colder temperatures, it was discovered that all of the pavement responses analyzed for the three sections were less than their respective fatigue endurance thresholds. Additionally, speed and axle configuration had a significant influence on the pavement responses. As testing speeds were increased the pavement responses decreased in magnitude. Furthermore, testing utilizing a tandem axle truck with dual tires produced reduced pavement responses in comparison with

4 4 testing utilizing a single axle truck with wide-based tires even though the tandem axle truck carried a greater load. Tire pressure did not have a significant effect on pavement responses recorded in lower portions of the pavement structure. The PerRoad analysis performed, using inputs corresponding to the U.S. Route 23 test sections, revealed that, in theory, all three of the test sections would have in-service lives in excess of their design life of 50 years. An optimized pavement thickness for perpetual pavement systems in Ohio was not recommended due to a need for controlled vehicle load testing to be conducted during warmer periods since pavement responses are typically greater at higher pavement temperatures.

5 5 ACKNOWLEDGEMENTS I would like to first acknowledge and thank the civil engineering staff along with professors from outside of the civil engineering department from which I have received instruction. All of you have provided education and guidance that has led to my success at Ohio University and have assuredly provided the necessary training for me to succeed in my future career. I would like to specifically thank my research supervisor, Sam Khoury. Sam motivated me to take advantage of the civil engineering graduate program here at Ohio University along with being a part of the momentous perpetual pavement research I have been working on. The master s degree and experience I have gained will prove to be beneficial through my civil engineering career. I cannot thank Sam enough for the support he has provided me during my graduate years at Ohio University. Furthermore, I would like to thank my advisor, Dr. Sargand, for providing me with this opportunity and sharing his knowledge with me. I would not be in the thriving position I am today without Dr. Sargand. Finally, I would like to acknowledge Dr. Masada, Dr. Steinberg, and Dr. Melkonian for participating as members of my thesis defense committee and for the advice I received from them pertaining to my work. An acknowledgement needs to be made for fellow graduate students who have spent countless hours with me during the construction and testing of my research project. There are many of you but I would like to specifically recognize Jayson Gray, Ben Jordan, and John Ubbing. Without your help, the work I have been doing would not have been possible.

6 6 Lastly, but most importantly, I want to thank my parents. I only wish I could display to them my appreciation for the opportunities they have provided me and the wisdom and guidance they have given to me throughout my life.

7 7 TABLE OF CONTENTS Page Abstract... 3 Acknowledgements... 5 List of Tables List of Figures Chapter 1: Introduction General Statement Objectives Outline Chapter 2: Literature Review Perpetual Pavement Pavement Reactions Fatigue Cracking Rutting Perpetual Pavement Design Subgrade Aggregate Base Fatigue Resistant Layer... 39

8 Intermediate Layer Surface Layer Perpetual Pavement Structural Trends Fatigue Endurance Limits Fatigue Endurance Limit Research Fatigue Endurance Limit Research in the Lab Carpenter, Ghuzlan, and Shen Ning, Molenaar, Van de Ven, and Shaopeng Fatigue Endurance Limit Research in the Field NCAT Test Track Texas Collective Analysis Perpetual Pavement Research China Marquette Oregon Advanced Transportation Research and Engineering Laboratory Wisconsin Kansas... 62

9 2.3.7 National Center for Asphalt Technology New York Accelerated Pavement Loading Facility Interstate 77, Canton, Ohio U.S. Route 30, Wayne County, Ohio Summary Chapter 3: U.S. Route 23 Project Background Project Site Description Design of Test Sections Instrumentation of Test Sections Strain Gage Installation Pressure Cell Installation Thermocouple Installation Linear Variable Differential Transducer Installation Strain Gage Rosette Hole Installation Controlled Vehicle Load Testing Chapter 4: CVL Testing Pavement Response Strain Response of the Pavement Longitudinal Strain in the FRL

10 Single and Tandem Axle Comparison Influence of Speed on Longitudinal Strain in the FRL Influence of Tire Pressure on Longitudinal Strain in the FRL Longitudinal Strain in the Base Layer Influence of Speed on Longitudinal Strain in the Base Layer Influence of Tire Pressure on Longitudinal Strain in the Base Layer Transverse Strain in the Base Layer Influence of Speed on Transverse Strain in the Base Layer Influence of Tire Pressure on Transverse Strain in the Base Layer Comparisons Between Strain Responses Subgrade Pressure Influence of Speed on Subgrade Pressure Influence of Tire Pressure of Subgrade Pressure Deflection Pavement Deflection Influence of Speed on Pavement Deflection Influence of Tire Pressure on Pavement Deflection Subgrade Deflection Influence of Speed on Subgrade Deflection

11 Influence of Tire Pressure on Subgrade Deflection Pavement and Subgrade Deflection Comparison Comparison of Pavement Structures Chapter 5: PerRoad Analysis Loading Conditions Accuracy of WIM Scales Vehicle Classification Distribution Performance Criteria Results Chapter 6: Conclusions and Recommedations Conclusions Recommendations References Appendix A: LVDT Calibration Appendix B: LVDT Case and Reference Rod Diagrams Appendix C: Strain Gage Rosette Hole Diagrams Appendix D: Instrumentation Diagrams Appendix E: Truck Loadings and Dimensions Appendix F: Testing Temperature Data

12 Appendix G: Strain Gage Data Appendix H: LVDT and Pressure Cell Data Appendix I: Lateral Tire Offset Data Appendix J: WIM Accuracy Data Appendix K: PerRoad Inputs

13 13 LIST OF TABLES Table 3.1 Layer Specifications Table 3.2 Layer Thicknesses Table 4.1 Percentage of Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for the 11 Inch Section Table 4.2 Percentage of Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for the 13 Inch Section Table 4.3 Percentage of Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for the 15 Inch Section Table 4.4 Maximum Longitudinal Strain in the FRL for the 11 Inch Section (µɛ) Table 4.5 Maximum Longitudinal Strain in the FRL for the 13 Inch Section (µɛ) Table 4.6 Maximum Longitudinal Strain in the FRL for 15 Inch Section (µɛ) Table 4.7 Average Maximum Longitudinal Strain in the FRL Comparison Between Single Axle and Tandem Axle for the 11 Inch Section (µɛ) Table 4.8 Average Maximum Longitudinal Strain in the FRL Comparison Between Single Axle and Tandem Axle for the 13 Inch Section (µɛ) Table 4.9 Average Maximum Longitudinal Strain in the FRL Comparison Between Single Axle and Tandem Axle for the 15 Inch Section (µɛ) Table 4.10 Longitudinal Strain in the Base Layer for the 11 Inch Section (µɛ) Table 4.11 Maximum Longitudinal Strain in the Base Layer for the 13 Inch Section (µɛ)

14 14 Table 4.12 Maximum Longitudinal Strain in the Base Layer for the 15 Inch Section (µɛ) Table 4.13 Maximum Transverse Strain in the Base Layer for the 11 Inch Section (µɛ) Table 4.14 Maximum Transverse Strain in the Base Layer for the 13 Inch Section (µɛ) Table 4.15 Maximum Transverse Strain in the Base Layer for the 15 Inch Section (µɛ) Table 4.16 Longitudinal Compared to Transverse Strains in the Base Layer for the 11 Inch Section (µɛ) Table 4.17 Percentage Transverse Strain in Base Layer of Longitudinal Strain in Base Layer for the 11 Inch Section Table 4.18 Longitudinal Compared to Transverse Strains in the Base Layer for the 13 Inch Section (µɛ) Table 4.19 Percentage Transverse Strain in Base Layer of Longitudinal Strain in Base Layer for the 13 Inch Section Table 4.20 Longitudinal Compared to Transverse Strains in the Base Layer for the 15 Inch Section (µɛ) Table 4.21 Percentage Transverse Strain in Base Layer of Longitudinal Strain in Base Layer for the 15 Inch Section Table 4.22 Maximum Longitudinal Strain Compared Between the FRL and Base Layer for Single Axle Truck Testing for the 11 Inch Section (µɛ)

15 15 Table 4.23 Percentage Longitudinal Strain in Base Layer of Longitudinal Strain in FRL for Single Axle Truck Testing for the 11 Inch Section Table 4.24 Maximum Longitudinal Strain Compared Between the FRL and Base Layer for Single Axle Truck Testing for the 13 Inch Section (µɛ) Table 4.25 Percentage Longitudinal Strain in Base Layer of Longitudinal Strain in FRL for Single Axle Truck Testing for the 13 Inch Section Table 4.26 Maximum Longitudinal Strain Compared Between the FRL and Base Layer for Single Axle Truck Testing for the 15 Inch Section (µɛ) Table 4.27 Percentage Longitudinal Strain in Base Layer of Longitudinal Strain in FRL for Single Axle Truck Testing for the 15 Inch Section Table 4.28 Maximum Subgrade Pressure for the 11 Inch Section (psi) Table 4.29 Maximum Subgrade Pressure for the 13 Inch Section (psi) Table 4.30 Maximum Subgrade Pressure for the 15 Inch Section (psi) Table 4.31 Maximum Pavement Deflection for the 11 Inch Section (mils) Table 4.32 Maximum Pavement Deflection for the 13 Inch Section (mils) Table 4.33 Maximum Pavement Deflection for the 15 Inch Section (mils) Table 4.34 Maximum Subgrade Deflection for the 11 Inch Section (mils) Table 4.35 Maximum Subgrade Deflection for the 13 Inch Section (mils) Table 4.36 Maximum Subgrade Deflection for the 15 Inch Section (mils) Table 4.37 Comparison Between Pavement and Subgrade Deflection for the 11 Inch Section (mils)

16 16 Table 4.38 Percentage Subgrade Deflection of Pavement Deflection for the 11 Inch Section Table 4.39 Comparison Between Pavement and Subgrade Deflection for the 13 Inch Section (mils) Table 4.40 Percentage Subgrade Deflection of Pavement Deflection for the 13 Inch Section Table 4.41 Comparison Between Pavement and Subgrade Deflection for the 15 Inch Section (mils) Table 4.42 Percentage Subgrade Deflection of Pavement Deflection for the 15 Inch Section Table 4.43 Maximum Pavement Response for the 11 Inch Section Table 4.44 Maximum Pavement Response for the 13 Inch Section Table 4.45 Maximum Pavement Response for the 15 Inch Section Table 5.1 Manual and WIM Counts by Classification Table 5.2 Vehicle Type Distribution Table 5.3 Seasonal Information (The Weather Channel, 2013) Table 5.4 Elastic Modulus and Poison s Ratio Table 5.5 Performance Criteria Table 5.6 Fatigue Transfer Function Empirical Constants (Timm & Priest 2006) Table Inch Section PerRoad Results Table Inch Section PerRoad Results Table Inch Section PerRoad Results

17 17 Table A.1 Calibration Data (LVDT 1-1 through 4-2) Table A.2 Calibration Data (LVDT 5-1 through 8-2) Table A.3 LVDT Calibration Factors Table F.1 Testing Temperatures for the 11 Inch Section (12/18/12) Table F.2 Testing Temperatures for the 13 Inch Section (12/19/12) Table F.2 Testing Temperatures for the 15 Inch Section (11/29/12) Table G.1 Maximum Strains for Testing on the 11 Inch Section Involving a Tire Pressure of 80 psi (µɛ) Table G.2 Maximum Strains for Testing on the 11 Inch Section Involving a Tire Pressure of 110 psi (µɛ) Table G.3 Maximum Strains for Testing on the 11 Inch Section Involving a Tire Pressure of 125 psi (µɛ) Table G.4 Maximum Strains for Testing on the 13 Inch Section Involving a Tire Pressure of 80 psi (µɛ) Table G.5 Maximum Strains for Testing on the 13 Inch Section Involving a Tire Pressure of 110 psi (µɛ) Table G.6 Maximum Strains for Testing on the 13 Inch Section Involving a Tire Pressure of 125 psi (µɛ) Table G.7 Maximum Strains for Testing on the 15 Inch Section Involving a Tire Pressure of 80 psi (µɛ) Table G.8 Maximum Strains for Testing on the 15 Inch Section Involving a Tire Pressure of 110 psi (µɛ)

18 18 Table G.9 Maximum Strains for Testing on the 15 Inch Section Involving a Tire Pressure of 125 psi (µɛ) Table G.10 Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for Testing on the 11 Inch Section (µɛ) Table G.11 Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for Testing on the 13 Inch Section (µɛ) Table G.12 Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for Testing on the 15 Inch Section (µɛ) Table G.13 Additional Maximum Strains for Testing on the 13 Inch Section (µɛ) Table G.14 Additional Maximum Strains for Testing on the 15 Inch Section (µɛ) Table G.15 Maximum Strains Produced by the Steer Axle of the Tandem Axle Truck for Testing on the 13 Inch Section (µɛ) Table G.16 Maximum Strains Produced by the Steer Axle of the Tandem Axle Truck for Testing on the 15 Inch Section (µɛ) Table G.17 Additional Maximum Strains Produced by the Steer Axle of the Tandem Axle Truck for Testing on the 13 Inch Section (µɛ) Table G.18 Additional Maximum Strains Produced by the Steer Axle of the Tandem Axle Truck for Testing on the 15 Inch Section (µɛ) Table H.1 Maximum Displacements and Pressures for Testing on the 11 Inch Section Involving a Tire Pressure of 80 psi Table H.2 Maximum Displacements and Pressures for Testing on the 11 Inch Section Involving a Tire Pressure of 110 psi

19 19 Table H.3 Maximum Displacements and Pressures for Testing on the 11 Inch Section Involving a Tire Pressure of 125 psi Table H.4 Maximum Displacements and Pressures for Testing on the 13 Inch Section Involving a Tire Pressure of 80 psi Table H.5 Maximum Displacements and Pressures for Testing on the 13 Inch Section Involving a Tire Pressure of 110 psi Table H.6 Maximum Displacements and Pressures for Testing on the 13 Inch Section Involving a Tire Pressure of 125 psi Table H.7 Maximum Displacements and Pressures for Testing on the 15 Inch Section Involving a Tire Pressure of 80 psi Table H.8 Maximum Displacements and Pressures for Testing on the 15 Inch Section Involving a Tire Pressure of 110 psi Table H.9 Maximum Displacements and Pressures for Testing on the 15 Inch Section Involving a Tire Pressure of 125 psi Table I.1 Average Lateral Tire Offset (in.) Table J.1 Accuracy Data for Classes One Through Four in 15 Minute Intervals Table J.2 Accuracy Data for Classes Five Through Eight in 15 Minute Intervals Table J.3 Accuracy Data for Classes 9 Through 12 in 15 Minute Intervals Table K.1 U.S. Route 23 Yearly Volume by Classification Table K.2 Average Monthly Temperature in Delaware, Ohio

20 20 LIST OF FIGURES Figure 3.1 Location of SHRP Test Road (Geology, 2013) Figure 3.2 Location of Test Sections (picture taken by ODOT) Figure 3.3 KM Strain Gage (photo taken by Jaime Hernandez) Figure 3.4 PM Strain Gage (photo taken by Jaime Hernandez) Figure 3.5 Instrumentation Layout in the Field Figure 3.6 Wire Installation in DGAB Figure 3.7 Wire Installation on Asphalt Surface (photo 4 taken by Jaime Hernandez).. 81 Figure 3.8 Strain Gage Installation (photos 2, 3, and 5 taken by Jaime Hernandez) Figure 3.9 Surface Strain Gage Installation Figure 3.10 Geokon Earth Pressure Cell Figure 3.11 Subgrade Pressure Cell Installation Figure 3.12 RST Total Earth Pressure Cell (photo taken by Jaime Hernandez) Figure 3.13 FRL Pressure Cell Installation (photos taken by Jaime Hernandez) Figure 3.14 Thermocouple and Strain Gage Alignment Figure 3.15 LVDT Figure 3.16 LVDT Case and Reference Rod Diagram Figure 3.17 LVDT Reference Rod Installation Figure 3.18 LVDT Case Installation Figure 3.19 Strain Gage Rosette Hole Diagram Figure 3.20 Strain Gage Rosette Hole Installation Figure 3.21 Instrumentation Layout... 95

21 21 Figure 3.22 Measurement of Lateral Offset Figure 3.23 ODOT Truck used for CVL Testing Figure 3.24 Axle configurations Figure Inch Section Temperature Profile Figure Inch Section Temperature Profile Figure Inch Section Temperature Profile Figure 4.4 Single Axle, Longitudinal Strain Response Figure 4.5 Single Axle, Transverse Strain Response Figure 4.6 Tandem Axle, Longitudinal Strain in the FRL Response at Five MPH Figure 4.7 Tandem Axle, Longitudinal Strain in the FRL Response at 55 MPH Figure 4.8 Tandem Axle, Strain Response with Steer Axle Producing Maximum Strain Figure 4.9 Strain Response Above the Neutral Axis of the Pavement Figure 4.10 Maximum Longitudinal Strain in FRL versus Speed Figure 4.11 Maximum Longitudinal Strain in FRL versus Tire Pressure Figure 4.12 Maximum Longitudinal Strain in Base Layer versus Speed Figure 4.13 Maximum Longitudinal Strain in the Base Layer versus Tire Pressure Figure 4.14 Maximum Transverse Strain in the Base Layer versus Speed Figure 4.15 Maximum Transverse Strain in in the Base Layer versus Tire Pressure Figure 4.16 Ratio of Transverse to Longitudinal Strain for the 13 Inch Section Figure 4.17 Ratio of Transverse to Longitudinal Strain for the 15 Inch Section Figure 4.18 Single Axle, Subgrade Pressure Response

22 22 Figure 4.19 Tandem Axle, Subgrade Pressure Response Figure 4.20 Maximum Subgrade Pressure versus Speed Figure 4.21 Maximum Subgrade Pressure versus Tire Pressure Figure 4.22 Single Axle, Pavement Deflection Response Figure 4.23 Tandem Axle, Pavement Deflection Response Figure 4.24 Tandem Axle, Subgrade Deflection Response Figure 4.25 Maximum Pavement Deflection versus Speed Figure 4.26 Maximum Pavement Deflection versus Tire Pressure Figure 4.27 Maximum Subgrade Deflection versus Speed Figure 4.28 Maximum Subgrade Deflection versus Tire Pressure Figure 4.29 Maximum Longitudinal Strain Obtained in the FRL for Each Test Section at a Testing Speed of Five MPH and a Tire Pressure of 110 PSI Figure 5.1 WIM Scales Figure 5.2 WIM Scales Figure 5.3 Vehicle Type Distribution Figure 5.4 Loading Conditions Figure 5.5 Structural and Seasonal Information Figure B.1 11 Inch Section LVDT Case and Reference Rod Figure B.2 13 Inch Section LVDT Case and Reference Rod Figure B.3 15 Inch Section LVDT Case and Reference Rod Figure C.1 13 Inch Section Round Strain Gage Rosette Hole Figure C.2 15 Inch Section Round Strain Gage Rosette Hole

23 23 Figure C.3 13 Inch Section Square Strain Gage Rosette Hole Figure C.4 15 Inch Section Square Strain Gage Rosette Hole Figure D.1 11 Inch Section Instrumentation Figure D.2 13 and 15 Inch Section Instrumentation Figure E.1 Single Axle Wide-Based Tire (Empty) Figure E.2 Single Axle Wide-Based Tire (Max Load) Figure E.3 Tandem Axle Dual Tire (Empty) Figure E.4 Tandem Axle Dual Tire (Max Load)

24 24 CHAPTER 1: INTRODUCTION 1.1 General Statement Pavement systems with extended service lives have become imperative to the highway industry in the United States. The current rate at which pavement roadways diminish to a state where replacement or rehabilitation is necessary has become problematic for the economy and progressive traffic demands. The construction involved due to the poor conditions of these pavement systems has not only created an increase in cost but also user delays and, therefore, fuel consumption and likelihood of crashes as well. The United States transportation system is essential to travelers and businesses across the country. Currently, the system includes over four million miles of roads carrying approximately 4.7 trillion passenger miles and 3.7 trillion ton miles of freight per year (National Atlas of the United States, 2013). The expenses for maintaining this vast amount of roadway, a lot of which is in poor condition, is enormous. Portillo (2008) estimates the yearly expenditures to rebuild and rehabilitate pavement structures across the country are in excess of $110 billion. A significant amount of this expenditure could be avoided through the development of a longer lasting pavement system. The perpetual pavement system, having proper pavement designs and resulting in long lasting structures, is a development which could save state and federal departments of transportation (DOTs) tremendous amounts of budget and time. Perpetual pavements are pavement structures designed to carry the heaviest vehicle loadings without the need for major structural rehabilitation or reconstruction for at least 50 years (Asphalt

25 25 Pavement Alliance, 2002). These pavement structures obtain their longevity from resisting the failure modes of bottom-up fatigue cracking and structural rutting. They achieve this through an increased thickness and a multiple layer system where thickness and stiffness are varied depending on the type of distress the layer is intended to resist. Properly designed perpetual pavement structures can provide superior engineering performance creating minimal user-delays and economic efficiency. Fatigue endurance thresholds are used to evaluate these perpetual pavement systems without requiring the need to wait through the system s entire service life before an analysis can be made. It was discovered that if certain pavement reactions remain below their specific thresholds, the pavement structure, in theory, can withstand an infinite amount of loading cycles without the development of damages. During loading, reactions at critical localities can be monitored and compared with these threshold limits and if the reactions remain below the threshold limit, damages that will transpire during the structure s lifespan will be diminished (Timm & Newcomb, 2006). Both bottom-up fatigue cracking and structural rutting can be analyzed through fatigue endurance threshold limits and will be the main source for analysis throughout this report. A supplementary analysis for perpetual pavement systems can be made using the computer software program, PerRoad. PerRoad performs a layered elastic analysis with a Monte Carlo Simulation on pavement structures. It was developed by the Asphalt Pavement Alliance to estimate stresses, strains, and deflections within the pavement structure using loading conditions, seasonal temperatures, pavement layer thicknesses and properties, and limiting pavement responses. Pavement responses are compared with

26 26 threshold limits in order to predict if they will be detrimental to the pavement s structure. Fatigue transfer functions can also be applied to estimate accumulating damages and, therefore, lifespan. Finally, a cost analysis can also be performed. The PerRoad program is useful for analyzing pavement structures under the likely conditions they will experience during their service lives. In order to properly implement perpetual pavements in Ohio, the Ohio Department of Transportation (ODOT) constructed four experimental perpetual pavement sections in Delaware, Ohio on U.S. Route 23 in hopes of discovering an optimized pavement thickness. The test sections exploited three different pavement thicknesses of 11, 13, and 15 inches. Additionally, Ohio University instrumented portions of the test sections with sensors to further ODOT s analysis of the structures. The main source for analysis was controlled vehicle load (CVL) testing where trucks of known loadings were driven over sections utilizing instrumentation and measurements recorded by the sensors. CVL testing involved varying tire pressures, axle configurations, and speeds in order to analyze the effects of these factors. In this thesis, pavement responses collected through sensors installed in the experimental test sections are evaluated in order to analyze the performance of the pavement and the perpetual nature of the various sections. Additionally, an evaluation was made on the effects of variables including axle configuration, speed, and tire pressure. Furthermore, the pavement test sections were analyzed using the computer program PerRoad which modeled and evaluated the sections using realistic conditions.

27 27 The objective of these analyses was to investigate various perpetual pavement structures implementing a variety of pavement thicknesses. 1.2 Objectives The principal objectives of this study are to: Investigate various perpetual pavement structures through varying the thickness of pavement layers in instrumented field test sections; Propose an optimum thickness for perpetual pavements in Ohio; Study possible relationships between tire pressure and pavement load response. Analyze the effect of speed and pavement load response; Observe and report the influence of tire pressure on speed and pavement load response trends; Study the effect of speed on tire pressure and pavement load response trends; Analyze pavement load responses with respect to axle configuration and loading; Compare longitudinal and transverse strains throughout the depth of various pavement thicknesses; and Evaluate the perpetual nature of the test sections using the computer program PerRoad. 1.3 Outline Chapter 1 introduces and defines the objectives of this study. Chapter 2 is a literature review including the concept, failure modes, structure, fatigue endurance limits, and existing research on perpetual pavement systems.

28 28 Chapter 3 describes the pavement test sections and instrumentation installation along with testing procedures. Chapter 4 includes the results and analysis obtained from CVL field testing conducted on the pavement test sections. Chapter 5 illustrates the PerRoad computer software including results obtained from analyses involving the U.S. Route 23 pavement test sections. Chapter 6 draws conclusions and recommendations based on the results presented.

29 29 CHAPTER 2: LITERATURE REVIEW With today s increasing traffic demands and tight economy, the construction of roadways is far too costly to have a design life of only 20 years. Tarefder and Bateman (2009) state that, currently, 25% of roads in this country must be replaced or at least rehabilitated between only four or five years of service life. A 20-year design life has been an American Association of State Highway and Transportation Officials (AASHTO) guideline ever since the interstate program began (Asphalt Pavement Association of Oregon, 2005). This is due to the fact that, until recently, sophisticated future traffic demand prediction models have not been available. Over the last ten years additional knowledge on the performance of high-volume roads has become available. There have also been misconceptions that perpetual pavements, due to higher initial expenditures, are more costly than standard flexible pavements. Contrary to misbelief, savings attained in the long run by perpetual pavement systems can far exceed additional costs presented during initial construction (Asphalt Pavement Association of Oregon, 2005). In fact, these pavement systems, with thorough design and consistent maintenance, can have life expectancies of 50 years or more. 2.1 Perpetual Pavement Perpetual pavements originated from the realization that forever increasing the thickness of the pavement to continue to provide structurally stable roads with everincreasing traffic volumes was too costly (Newcomb, Willis, & Timm, 2010). It was discovered that the heaviest loads could be indefinitely supported without further structure. There comes a point when additional pavement becomes overly-adequate to

30 30 provide for the heaviest loads, creating unnecessary uses of resources and preventable added costs. Designing perpetual pavements is better approached by analyzing pavement reactions in the context of pressure, strain, and displacement (Newcomb et al., 2010). Perpetual pavements are defined as an asphalt pavement designed and built to last longer than 50 years without requiring major structural rehabilitation or reconstruction, and needing only periodic surface renewal in response to distresses confined to the top of the pavement (Asphalt Pavement Alliance, 2002, p. 5). These innovative pavement structures are designed to provide an infinite structural life under the heaviest vehicle loadings. Only surface layers would intermittently require milling and replacement to reduce surface distresses, a quick repair with minimal delays and restoration costs (Sargand, Figueroa, & Romanello, 2008). Perpetual pavement designs have become practical because of their engineering performance and economical efficiency. Perpetual pavement structures are more endurable and, can be maintained at minimum costs. The little maintenance that is required incorporates minimal user-delays. They are high performance, superior pavement structures. When designed appropriately, perpetual asphalt pavements can be more economical than conventionally designed asphalt concrete pavements (Newcomb et al., 2010). Sargand et al. (2008) approximated that perpetual pavements can save state and federal DOTs billions of dollars. They also mention that adequately designed and constructed systems can outlast rigid pavement structures. Although not recognized until recently, long-life asphalt pavements have been around since the 1960s (Newcomb & Hansen, 2006). Many sections of pavement have

31 31 received perpetual pavement awards for having extended service lives. These sections had thicker designs, whether full-depth or deep-strength, and were well-constructed resulting in successful performances (Estes, 2005). Asphalt pavements built directly on the subgrade soil are called full-depth pavements while asphalt pavements that are built on a thin granular base course are deemed deep-strength pavements (Buchner, Newcomb, & Huddleston, 2000). Using the same concepts, pavement systems are now purposefully designed to be perpetual. Perpetual pavements should be monitored periodically in order to confine damages within the surface of the pavement, avoiding total reconstruction. Deterioration of perpetual pavements typically results in the surface layer due to cracking or rutting (Nunn & Ferne, 2001). Traffic and weathering will cause damage to appear in the surface layer every 15 to 20 years (Hatch, 2008). This damage includes top-down fatigue cracking, thermal cracking, rutting, and surface wear (Asphalt Pavement Alliance, 2002). In order to remain economical, the layer should be replaced once distresses in the surface of the pavement reach critical levels. If distresses in the surface extend down into the structural layers of the pavement, reconstruction of the entire structure will be necessary (Buchner et al., 2000). Resurfacing may also be essential when improvements are needed for noise, driving comfort, or safety (Newcomb et al., 2010). Replacing the surface layer can be done overnight with minimal traffic disturbance. Perpetual pavement designs can be used for any pavement with the desire to minimize reconstruction and rehabilitation costs while reducing user delays. Perpetual pavements are especially applicable for higher traffic volume roadways due to their

32 32 ability to minimize user delays. They are also valid design considerations for major airports where delays due to construction are extremely costly (Asphalt Pavement Alliance, 2002). Perpetual pavements are typically only used for high volumes of traffic, but designs for roads with lower traffic volumes can be vindicated as well (Buchner et al., 2000). Utilizing perpetual pavement designs have advantages for most scenarios Pavement Reactions Perpetual pavement systems are successful due to the fact that they prevent two major distresses, rutting and fatigue cracking. Most approaches for designing perpetual pavement structures are focused on limiting the pavement s response to bottom-up fatigue cracking and structural rutting (Newcomb et al., 2010). The damage accumulated by these two modes of failure must be limited to practically zero for a design to be effective Fatigue Cracking One of the major components instigating structural deterioration of pavements is fatigue cracking. Bottom-up fatigue cracking is the process where heavy traffic loading repeatedly creates high tensile strains in the bottom of the asphalt layer due to bending. (Newcomb et al., 2010; Retrepo-Velez, 2011; Willis, 2009). These repeated strains eventually exceed the fatigue resistance of the asphalt concrete and result in cracking. This cracking eventually propagates through the asphalt structure to the surface. Fatigue life is ultimately determined by the frequency and severity of horizontal strains experienced at the base of the asphalt layer (Newcomb et al., 2010).

33 33 Fatigue cracking is detrimental to a pavement structure. Once initiated, fatigue cracking will accelerate pumping, rutting, and overall deterioration of the asphalt layer (Willis, 2009). If the pavement s environment includes freezing and thawing cycles water that has seeped into crack voids will expand under freezing conditions causing the crack to expand and eventually produce a pot hole. Although fatigue cracking occurs in the asphalt layer of the pavement structure it ultimately affects all the other components of the structure as well. Fatigue cracking provides openings for water and moisture to seep into the unbound layers of the pavement structure, thus altering their material properties (Willis, 2009; Willis & Timm, 2009). Once a pavement structure has a cracked asphalt layer and a weakened subgrade or bases from moisture damage, reconstruction will be necessary. Fatigue cracking is prevented in perpetual pavement by either increasing the thickness of the pavement structure or developing a more flexible base layer which can better resist strain repetitions (Newcomb et al., 2010). By reducing the strain at the base of the asphalt, the critical location becomes the surface where deep structural maintenance is avoided. This does not mean surface cracks can be overlooked due to their effect on ride quality and noise, but it does mean maintenance costs will be significantly reduced (Newcomb et al., 2010) Rutting Rutting is a permanent deformation that develops when the strength of the pavement structure is inadequate (Newcomb et al., 2010). Rutting deformation can

34 34 present itself in one or multiple layers. Structural rutting occurs in the aggregate base or subgrade while surface rutting occurs in the surface layer of the asphalt layer. Rutting potential is dependent on temperature. At lower temperatures, the asphalt binder is stiff and less vulnerable to rutting while at higher temperatures the binder is soft and behaves viscoelastically, and therefore, more susceptible to rutting. Brown (2002) found, while studying data collected from all of the test sections instrumented at the NCAT Test Track, that the rate of rutting was minimized to nearly zero when air temperatures dropped below 82 F. Structural rutting is of main concern when designing perpetual pavements since surface rutting is restricted to the top of the asphalt layer and can easily be remediated through resurfacing. Although rarely a problem in modern long-life pavements, extensive rehabilitation or reconstruction will be required for pavements experiencing structural rutting (Newcomb et al., 2010). Structural rutting arises in the subgrade. It evolves from the asphalt insufficiently spreading the load, thus inadequately decreasing the pressure being applied on the subgrade causing it to deform (Nunn & Ferne, 2001). In essence, the asphalt layer is not adequately protecting the subgrade. Rutting potential can be minimized using different techniques. Studies have shown a threshold effect where rutting potential significantly decreases after a certain thickness is reached (Nunn & Ferne, 2001). Structural rutting can be prevented not only by designing a thicker pavement structure but also by increasing the stiffness in portions of the structure as well (Newcomb et al., 2010). A stiff elastic binder is sought when considering rutting deformation. While analyzing data collected from the NCAT Test

35 35 Track, Brown (2002) and his associates discovered that rutting was decreased by 60 percent when higher high-temperature binder grades of even only one increment were used. Providing a mix design that allows for better compaction is an additional solution for limiting rutting potential. Using the same data as Brown, Timm (2006) notes that lower air voids resulted in less rutting especially when using higher high temperature grade binders. Perpetual pavement systems limit rutting to the surface layers which can be easily alleviated through resurfacing Perpetual Pavement Design Perpetual pavement requires proper structural design as well as the use of appropriate materials. The main difference between perpetual pavements and conventional asphalt pavements is the sound design methodology used with perpetual pavements. Deep structural rutting and fatigue cracking are considered to be the most destructive pavement distresses (Newcomb et al., 2010). Perpetual pavements are specifically designed to resist these distresses. The optimal pavement structure is obtained by using a multiple layer system where thickness and stiffness are varied depending on the type of distress the layer is intended to diminish (Tarefder & Bateman, 2012). The structural and mix designs are essential to creating a long-life pavement. The material selection involved with perpetual pavements is tremendously important. Due to the fact that each layer is designed to resist various distresses, the materials considered, mix design, and laboratory testing should be unique for each layer (Buchner et al., 2000). All layers of the pavement system should possess good constructability characteristics as well as, to the greatest extent, be impermeable to avoid

36 36 stripping damage along with other moisture problems (Newcomb & Hansen, 2006). The pavement structure must also be thick enough to provide the structural integrity needed to prevent structural rutting, fatigue cracking, and permanent deformation while also having the durability to resist damage from the environment (Willis & Timm, 2009). While durability is essential for every layer, the stiffness of each layer is dependent on whether the layer needs to be resistant to rutting or fatigue cracking. The pavement system must provide sufficient stiffness in the upper regions of the pavement to avoid rutting and flexibility and adequate total pavement thickness in lower regions to resist bottom-up fatigue cracking (Newcomb & Hansen, 2006). When developing the surface layer, it is important to consider shear strains (Newcomb & Hansen, 2006). While each layer plays a specific role, the most important aspect of the system is the total pavement thickness. The overall performance of the pavement is not just a function of design. It must be understood that design life is also affected by construction and maintenance practices. Construction should be done with a high commitment to quality and a large amount of attention to detail. The density of the pavement is crucial for extended fatigue life. Lack of density can be caused by poor compaction during construction (Newcomb et al., 2010). Furthermore, segregation needs to be avoided because it can cause permeable pockets to develop which may lead to moisture damage. Volumetrics must be tightly controlled and proper mixing should be achieved in order to ensure a quality product. Debonding is another important issue when analyzing the pavement s longevity. Debonding can be a result of erratic asphalt construction temperatures due to delays or

37 37 slow construction (Newcomb et al., 2010). If the pavement system contains debonded layers, fatigue cracking may develop at the interface (Tarefder & Bateman, 2012). Good construction practices will result in good pavement performance. Currently, most engineers are using Mechanistic-Emprical (M-E) Design when developing perpetual pavement systems. The M-E design process utilizes the concept of restricting fatigue cracking to the surface layer by restraining tensile strains at the base of the asphalt structure. Timm and Davis (2009) mention that there are two design procedures currently utilize the M-E Design process. The first procedure is the Mechanistic Empirical Pavement Design Guide (MEPDG) and the second is a computer software program, PerRoad. Timm and Davis discovered that, although the PerRoad software was relatively less complex than the MEPDG, both programs developed similar pavement structures when using identical input data Subgrade The subgrade is the foundation for the entire pavement system and is crucial to the overall performance. The subgrade should provide adequate stiffness because it provides resistance to deflection allowing rollers to produce a firm compaction of all layers (Asphalt Pavement Alliance, 2002). A weak foundation will fail to provide an adequate working platform for construction equipment as well as enough resistance to allow the various asphalt layers to be compacted to their desired densities (Newcomb et al., 2010). A properly designed and constructed subgrade will not become overstressed and will minimize rutting development caused by construction traffic.

38 38 A pavement may fail prematurely due to extended periods of saturation in the foundation. Drainage features may be added in certain climates to avoid moisture problems (Asphalt Pavement Alliance, 2002). If the subgrade is weakened due to moisture, it may be wise to slightly add to the thickness of the surface course the next time the pavement is resurfaced (Asphalt Pavement Alliance, 2002). Although increasing the total thickness is not desired, this would help to ensure that the pavement system will remain perpetual in nature. The subgrade not only provides loading support but it also hinders volume variability due to moisture and freeze-thaw changes (Newcomb et al., 2010). Typically, subgrade moisture content values are particularly high after a thaw cycle, suggesting that the subgrade is the weakest in the late spring (Asphalt Pavement Alliance, 2002). Subgrades should be analyzed and pavement structures should be designed using the maximum moisture content in the subgrade soil. The subgrade is typically designed as a compacted subgrade but may also include chemical stabilization. Lime, cement, or fly ash procedures are frequently used to stabilize the subgrade if additional support is necessary (Newcomb et al., 2010; Tarefder & Bateman, 2009). According to Newcomb et al. (2010) the subgrade should have a California Bearing Ratio (CBR) of at least six in order to provide proper construction support and to prevent overstressing of the subgrade during the pavements life. Soils with lower values of CBR should incorporate the use of a Dense Graded Aggregate Base (DGAB).

39 Aggregate Base A DGAB is necessary when the subgrade cannot provide adequate support. In such a case, the DGAB provides this additional support along with other properties that the subgrade should have. Aggregate bases attenuate stresses on the foundation and provide additional support to allow for proper compaction of asphalt layers (Newcomb et al., 2010). It accomplishes this through its load spreading ability provided by the aggregates resilient modulus, or stiffness, Moreover, aggregate bases offer supplementary drainage and protection of the subgrade against environmental damages. Aggregate bases should only be designed to the extent of providing necessary foundation support. While analyzing data collected from instrumented, perpetual pavement test sections in Oregon, the Asphalt Pavement Association of Oregon (2005) found that increasing the thickness of the DGAB was not significantly influential in preventing fatigue cracking. The thickness and material properties of each asphalt layer is the best approach for optimizing the pavements fatigue life Fatigue Resistant Layer The structural adequacy of the pavement relies on the fatigue behavior of the fatigue resistant layer (FRL). The main function of the FRL is to provide durability and resistance to fatigue cracking. The FRL is the bottom layer of the asphalt structure, so if damages occur, total reconstruction is unavoidable. Cracking tendencies need to be mitigated in the bottom layer. Increasing the asphalt content of the bottom layer provides further protection against fatigue cracking (Buchner et al., 2000; Newcomb et al., 2010). Carpenter and Shen (2006) estimate that

40 40 the FRL should have a half percent of asphalt binder content increase. Although the asphalt content can become too high, additional asphalt, up to a certain point, impedes the development and progression of fatigue cracks (Asphalt Pavement Alliance, 2002). The increased amount of binder avoids the initiation of cracks by making the layer more flexible and, thus, more resistant to repeated loading. Since the layer is designed to reduce fatigue cracking through flexibility, it is essential that rut-testing of the material is performed to warrant that the equipment driving directly on the layer during construction does not cause permanent damage (Asphalt Pavement Alliance, 2002). The FRL should be between three and four inches to avoid extending base material properties into higher portions of the asphalt system (Newcomb et al., 2010). Although increasing the binder content in the bottom layer of the asphalt structure provides additional fatigue protection, Newcomb and Hansen (2006) suggest that increasing the total pavement thickness is considerably more beneficial than solely relying on increasing the binder content. Furthermore, Tarefder and Bateman (2012) report that the necessity for a rich bottom fatigue layer diminishes once the total pavement thickness starts to exceed 12 inches. They also report that, when using an FRL, the layer does not have the stiffness required to make contributions to the pavements rutting strength and that using an FRL layer is not an economical option. Higher air voids in the FRL will result in reduced fatigue life. The FRL should have a max in-place density of between 96% and 98% or 2% to 4% air voids (Buchner et al., 2000). Typical standard designs are only compacted to between 6 and 7% air voids (Carpenter & Shen, 2006). Newcomb and Hansen (2006) suggest using a fine graded

41 41 mix to permit for lower air voids. This was also a consideration made by Willis (2009) while he analyzed data collected from Phase III of the NCAT Test Track. He discovered that fine graded mixtures, when compared to coarse graded mixtures, seemed to be more resistant to cracking. The reduction in air voids allows the layer to be further compacted, thus increasing its resistance to fatigue and improving durability (Newcomb et al., 2010). Increasing the volume of voids that are filled with binder (VMA) is a critical component to the pavement s durability and flexibility. Moisture susceptibility problems can be of concern with the FRL because the layer is likely to have extended periods of contact with water (Newcomb et al., 2010). Moisture can lead to stripping, a decrease in bond strength between the aggregate and binder, which will gradually decrease the strength of the pavement. Proper mix designs and compaction during construction are essential to combatting moisture susceptibility. It is also important to realize that the stiffness of the base layer will increase with time which will intensify its bottom-up fatigue cracking potential (Nunn & Ferne, 2001). Engineers should take both of these factors into account when designing perpetual pavement structures Intermediate Layer The intermediate layer of the perpetual pavement system provides both rutting resistance and durability. The intermediate layer is designed to carry most of the traffic load. Both binder and aggregate are of importance when designing the intermediate layer where higher stresses may induce shear failure, creating ruts. The intermediate layer is typically the thickest layer in the system, sometimes consisting of two layers. This means

42 42 sections of the layer will lie on both sides of the neutral axis, exposing it to both tension and compression. Rutting is prevented by using an appropriate high-temperature grade binder. A report by Willis (2009) on the findings from Phase III of the NCAT Test Track said that PG 76 polymer modified binders had a rutting rate half of what the unmodified PG 67 binder had. No effects on strain were seen. A solid aggregate skeleton needs to be provided in the intermediate layer. Aggregates in the intermediate layer should provide strong internal friction (Buchner et al., 2000). Internal friction and, thus, stability performance is achieved by stone-on-stone contact within the coarse aggregate and can be obtained by using a crushed aggregate and a large nominal maximum-size (Newcomb et al., 2010). Aggregates with small nominal maximum-size can be used as long as stone-on-stone contact is upheld (Buchner et al., 2000). Additionally, densely graded mixtures are recommended for the intermediate layer (Newcomb et al., 2010). Reclaimed asphalt pavements (RAP) are sometimes used to increase rutting resistance by stiffening the pavement (Newcomb et al., 2010). Willis (2009) discovered that, on the NCAT Test Track, sections containing RAP performed better than sections without. The same effect is seen through the use of recycled asphalt shingles (RAS). Using recycled materials not only provides additional structural support but is environmentally friendly too.

43 Surface Layer The surface layer provides protection from direct environmental and traffic exposure, and, therefore, requirements for the surface layer depend on the environment and traffic conditions it will be exposed to. It is subjected to the highest temperature variation and, therefore, the highest thermal stresses (Newcomb & Hansen, 2006). Thermal cracking, a result of high thermal stresses exceeding the ability of the material to relieve deformation, is prevented through the proper selection of the low temperature grade when selecting a binder. The surface layer should be impermeable and resistant to rutting and wear (Buchner et al., 2000). It needs to provide not only resistance to rutting and wear but good friction, minimization of splash and spray, and mitigation of tirepavement noise as well (Newcomb et al., 2010). While the structural design life of perpetual pavements should be 50 years, the wearing course should have a design life of approximately 20 years (Willis, 2009). The surface layer is not meant to prevent topdown cracking but these damages can easily be eliminated through asphalt overlays or inlays. The surface layer is the only part of the system that has additional aspects to consider past its structural features. This layer must be safe for the user to drive on, provide comfort for the driver, and reduce traffic noise for the public around the roadway. Safety can be provided through friction or skid-resistance by using a low-polish aggregate with proper microtexture (Newcomb & Hansen, 2006). Proper water drainage will reduce splash and spray increasing visibility and reducing hydroplaning. A proper balance of friction through aggregate microtexture will keep noise to a minimum while

44 44 still providing a safe surface (Newcomb & Hansen, 2006). This will also provide comfort for the driver. The surface layer should be designed so that it can be easily renewed when damage arises that may increase driver discomfort (Tarefder & Bateman, 2009). By designing a thicker pavement, strains in the bottom of the asphalt structure are reduced and the critical strains are then relocated and remain in the wearing surface. The surface layer is subjected to the highest vertical compressive stresses and is susceptible to rutting. An increase of one increment for the high-temperature grade of the binder should be used for additional rut protection (Asphalt Pavement Alliance, 2002). Also, the use of crushed aggregate is recommended for additional internal friction (Newcomb & Hansen, 2006). This is especially important since the thickness of the surface course is thin which restricts the nominal maximum size of the aggregate. There are various mixtures that can be implemented for the surface layer. A stone matrix asphalt (SMA) is recommended for roads experiencing large amounts of truck traffic. SMA mixes are highly rut resistant, durable, and wear resistant (Newcomb et al., 2010). They are also impermeable providing the system with moisture control. They use a polymer-modified binder along with mineral fillers or fibers, which provide additional stiffness for cracking protection along with an impermeable effect. (Newcomb & Hansen, 2006; Newcomb et al., 2010). Newcomb and Hansen (2006) suggest the addition of fibers to create a thicker film coat for the aggregate and prevent drain-down. The aggregate gradation should be carefully controlled during the production of an SMA mix. A gap gradation of crushed aggregate will provide interlock and resistance to rutting

45 45 (Newcomb & Hansen, 2006). An open-graded friction course (OGFC) can be beneficial in wetter climates. This design allows for drainage, therefore improving friction during wet conditions. This is accomplished by using one size of aggregate resulting in air voids ranging between 18 and 22% (Newcomb et al., 2010). Tarefder and Bateman (2012) point out that an OGFC is not considered a structural layer so it will not provide any resistance to structural distresses. A standard superpave dense graded mixture can be used when traffic volumes are lower. Dense graded superpave mixes are the most economical and easiest to produce (Newcomb & Hansen, 2006). However, while testing on the NCAT Test Track, Willis (2009) noted that section using an SMA surface course were more resistant to raveling along with other degradations that come along with aging when compared to sections using superpave Perpetual Pavement Structural Trends In 2012, Tarefder and Bateman compiled a prominent amount of perpetual pavement research reports from around the country to analyze and possibly determine the optimal perpetual pavement design. When analyzing structures of perpetual pavement test sections, they discovered that surface layers had thicknesses, on average, of two inches with a 2.5 inch transitional layer. Intermediate layers were, on average, eight inches thick and FRLs, if they were used, were five inches. Aggregate bases were also not always used but when they were they had an eight inch average thickness. Tarefder and Bateman also discovered that researchers used an FRL 60% of the time and aggregate bases were used 80% of the time. When considering the surface layer, an

46 46 SMA surface course was used approximately 30% of the time while an OGFC surface was used 15% of the time. 2.2 Fatigue Endurance Limits A perpetual pavement system is often analyzed using some threshold level. During loading, if reactions at critical localities within the pavement remain below this threshold, damages occurring during the pavement s life will be minimized (Timm & Newcomb, 2006). Tensile strain produced at the bottom of the asphalt is typically used for this threshold. The fatigue endurance limit (FEL) is a strain limit for the bottom of the asphalt layer that researchers have been using to make assumptions at early stages of the pavement s life on whether or not the pavement will be perpetual. The FEL has become an important consideration for any multi-layered perpetual pavement design. Many researchers suggest that involving an FEL while designing a perpetual pavement system is the utmost important aspect (Thompson & Carpenter, 2006a). The Asphalt Pavement Alliance (2002) alleged that the FEL is the classification of an ideal pavement system that will not involve reconstruction. Pavement is considered to have failed by many engineers when more than a half inch of rutting is produced along the wheel path or fatigue cracking reaches 20%. (Newcomb et al., 2010). Portillo (2008) states that if the tensile strain applied to the pavement structure is contained within the endurance limit then the asphalt should be able to resist over a billion loading cycles before failure through fatigue cracking is

47 47 reached. The FEL correlates to stresses that approach the endurance limit for the pavement structure. The concept behind the FEL is that little or no fatigue damage will develop if strains are kept below the FEL (Carpenter, Ghuzlan, & Shen, 2003). In essence, fatigue life is significantly extended and may possibly be infinite if strains at the bottom of the asphalt layer are retained below this limit. When conceptualizing this idea graphically, the FEL is when a plot of stress compared to load cycles causing failure essentially becomes zero (Prowell & Brown, 2006). At lower strain levels, the damage produced by each loading cycle is low enough to allow the pavement s healing process to nearly replace all of the energy lost while being deformed (Carpenter et al., 2003). This means that, if strains are kept small enough, sufficient time will be available during rest periods for damages to recover. There may exist a strain level where the healing process will offset damages. As with most properties of flexible pavement, the FEL is not a constant. Modifiers within the binder, the aggregate used as well as it gradation, the grade of the binder, and volumetric properties of the mixture all have an effect on the FEL value (National Cooperative Highway Research Program, 2010). Once an FEL is defined, the pavement should be designed so that the majority of strains experienced by the pavement are lower than the limit in order to avert structural damage. Pavements built thicker than what is required to retain strains below the FEL will not further the pavement s service life (National Cooperative Highway Research Program, 2010). Identifying critical limits where stresses, strains, or displacements in the

48 48 pavement initiate cracking or deep rutting, allows for pavement to be designed so that it can withstand the heaviest anticipated traffic while not over-designing the structure (Newcomb et al., 2010). Fatigue limits can be established for more than just strain located at the bottom of the asphalt layer. Engineers typically design perpetual pavement structures to limit compressive strain located at the top of the subgrade or aggregate base layer to 200 micro strains (µɛ) (Willis & Timm, 2009). Furthermore, Romanello suggested (2007) that, in order to evaluate the pavement s potential for structural rutting, the subgrade should not exceed 20 mils of deflection and that, in order to monitor the strain distribution throughout the depth of the pavement, the base layer should not exceed 50 µɛ of horizontal strain Fatigue Endurance Limit Research Both lab and field work have pointed towards the idea that there exists thresholds where if reactions are kept below this level, damage will not accumulate. Numerous asphalt pavement structures have been in use for 40 years or longer without any signs of bottom-up fatigue cracking suggesting that an FEL is a concept that can be validated (National Cooperative Highway Research Program, 2010) Fatigue Endurance Limit Research in the Lab Carpenter, Ghuzlan, and Shen Carpenter, Ghuzlan, and Shen (2003) performed laboratory research in order to discover whether or not an FEL exists, possibly determine its precise value, and analyze different variables that might alter the FEL. Their testing included ten mixtures with

49 49 varying gradations, asphalt contents, and air voids. The samples were obtained directly from the truck, were reheated and compacted, and then tested using an IPC fatigue machine. A dissipated energy ratio analysis was performed which quantifies the amount of energy a mixture is capable of handling. Carpenter, Ghuzlan, and Shen made several significant discoveries. First and foremost, they noticed that the rate at which damage was accumulated began to diminish around 100 µɛ. They predicted the FEL to be in the range of 70 µɛ to 90 µɛ and noted that reducing the FEL below 70 µɛ would not provide any additional fatigue life and therefore is unnecessary. They discovered that the FEL was different for different mixes but as long as strain in the bottom of the asphalt layer remained below 70 µɛ, material variability would have no effect on fatigue life. Distinct differences were found between higher and lower strain levels regardless of the mix being tested and accumulated damage was disproportionately less than traditional theoretical strain level fatigue testing extrapolations Ning, Molenaar, Van de Ven, and Shaopeng Ning and associates (2010) also performed laboratory research in order to discover more about the FEL. They used two types of tests, a uniaxial tension test and a four-point bending test. Their samples included three kinds of binders tested at three different temperatures, 41 F, 68 F, and 104 F. Eight different strain levels between 50 µɛ and 200 µɛ were also implemented. They found that the FEL would fluctuate depending on temperature and frequency. Also, by comparing results when testing to failure with those from testing

50 50 until the sample reached half its initial stiffness, they concluded that FEL values can be determined without fatigue testing to failure. It was noted that FEL values could be calculated using initial stiffness, frequency, and temperature Fatigue Endurance Limit Research in the Field Little understanding of HMA performance is known at lower levels of strain because fatigue testing is typically performed at higher strain levels in the lab. Furthermore, a report by the Advanced Asphalt Technologies (2008) suggests that further field research is needed on the FEL because current fatigue damage models do not incorporate healing or damage recovery that is taking place in between loading cycles. Fatigue curves developed by current AASHTO models have misrepresented field conditions when strains produced are limited (Carpenter et al., 2003). In other words, fatigue behavior, when subjected to low strain levels, does not follow typical model relationships. The current model does not even support the idea of an FEL. Lab testing also leaves out factors such as the environment, loading variation, and rest periods which can alter estimated FEL values NCAT Test Track Willis and Timm (2009) wrote a report on field-based strain thresholds using instrumented perpetual pavement test section on the NCAT Test Track in Alabama. They wanted to discover if there was a relationship between field measured and laboratory strain thresholds. They also sought to recommend an FEL based on both experimental methods. Their testing included data recorded from 2006 NCAT Test Track sections, lab tests performed on samples obtained from the same test sections, and a PerRoad analyses.

51 51 Willis and Timm (2009) discovered that sections of the test track failed early due to fatigue when experiencing strains greater than 125 µɛ at the bottom of the asphalt layer. While some sections eventually failed due to fatigue even while experiencing strains lower than 70 µɛ, this information was deemed to be irrelevant due to the fact that the sections were constructed upon poor subgrades. Sections experiencing strain below 10 µɛ were considered to be overdesigned and thus 10 µɛ was set as a lower bound value for the FEL. Lab testing resulted in strain thresholds that were higher than thresholds revealed from the test track, with lab tests resulting in strain thresholds as high as 220 µɛ. Willis and Timm found no clear correlation between lab and field testing. The Transportation Research Board (2010) also did a study using test sections from the NCAT Test Track. They had the objectives for determining if an FEL exists, measuring the FEL, and determining effects on the FEL. Their testing, like Willis and Timm, included laboratory work on samples obtained from test sections that were analyzed during field testing. One aggregate gradation was used with variations of two different binder grades and two different binder content levels. Beam fatigue and uniaxial tension tests were performed in the lab. The Transportation Research Board (2010) discovered FEL values ranging between 70 µɛ and 200 µɛ depending on mix. No conclusions were made on the effect of binder content and binder grade because trends varied depending on testing method. All FEL values were found to be above 70 µɛ. A reasonable shift factor between field and lab measurements was determined to be ten although shift factors ranged between 4 and 76. The Transportation Research Board noted that the FEL can reasonably be

52 52 estimated as strain at which the pavement structure can withstand 500 million traffic load repetitions during a 40 year period. Prowell and Brown (2006) performed a small study in order to determine a fieldbased FEL at the NCAT Test Track. After analyzing field measurements, they supported the concept of an FEL and found that fatigue cracking was prevented when strains at the bottom of the asphalt layer were kept below 100 µɛ. They also suggested that the extrapolation model recommended by AASHTO was not appropriate. Willis (2009) composed his thesis centered on an analyses of field based strain thresholds in perpetual pavement test sections at the NCAT Test Track. He found that some sections had received strain readings at the base of the asphalt layer greater than 70 µɛ without fatigue damage even after being exposed to 19 million ESALs. His conclusions were that perpetual pavements could possibly withstand strains greater than 70 µɛ or even 100 µɛ Texas A field study was performed by Quintus (2006) by utilizing field performance data of existing pavement structures. Quintus analyzed existing pavement structures by comparing their survival rates to the amount of loading, thus strain, they were receiving. He wanted to determine if field performance data supported the concept of an FEL and determine a reasonable FEL. Quintus revealed that survival rates began to dramatically increase as 100 µɛ and below thresholds levels were seen. He predicted the FEL to be 65 µɛ with a 95% confidence level. He also deemed the FEL to be an acceptable design premise.

53 Collective Analysis Advanced Asphalt Technologies (2008) wrote a report analyzing existing information and research completed by experts around the country. They examined lab, case, field studies, and alternative methods. Their objectives were to improve the MEPDG by validating that the FEL exists, addressing potential variances between lab and field measured FEL values, and developing a methodology for determining FEL and incorporating it within the MEPDG. Their results concluded that HMA clearly demonstrates endurance limit behavior but the value will alter depending on environmental and loading settings. They discovered that multiple properties have influences on allowable strain including pavement temperature, length of time between traffic loads, and mix composition. A report by Thompson and Carpenter (2006b) makes many reasonable considerations when determining an FEL. They suggest that for typical mixes and binders a reasonable FEL value is between 70 and 100 µɛ and that if FEL data is not available a conservative value of 70 µɛ should be used. When considering an FEL, Thompson and Carpenter recommend using an 18 kip single axle load along with the asphalt moduli for the hottest temperature of the year. Finally, they note that strain levels above the FEL may by sustained by the pavement for short duration during the year. 2.3 Perpetual Pavement Research The research being done by experts around the world on perpetual pavements is essential to them becoming the new standard for roadway design. The design and analysis of asphalt perpetual pavement structures is very complex due to many factors

54 54 including varying temperature, combinations of material layers resulting in intricate mechanistic reactions, and dynamic loadings (Jincheng, Lin, & Shijie, 2012). With such diverse conditions, the true strain distribution at the base of the asphalt structure cannot be determined solely analytically China In China, during the month of June (2005), three perpetual pavement test sections were constructed and instrumented in order to provide better understanding concerning perpetual pavement design (Yang et al., 2006). The testing conducted in China, as presented by Yang and colleagues, included three full-depth pavement systems and two control sections located on an expressway in Shandong Providence. The sections were built on top of a lime stabilized soil with a CBR of eight and the water table was only five feet below the ground. One of two full-depth sections had a total pavement thickness of 15 inches while the other had a total pavement thickness of 20 inches. The five additional inches in the 20 inch section were incorporated in the intermediate layer. Each section utilized an FRL by increasing the asphalt content in the base layer by 0.6%. The project also consisted of two control sections that were used for comparison, one having a 13 inch thickness and the other having a six inch thickness. The instrumentation of the sections, as described by Yang (2006), included pressure cells located on top of the subgrade, longitudinal and transverse strain gages at the bottom of the asphalt layer, and temperature probes. They spaced the strain gages into three rows in order to capture the natural variation of vehicle wheel path. Located along the wheel path, directly before the instrumented sections, were Weigh-in-motion

55 55 (WIM) stations to record live traffic volumes, weights, and axle configurations and axle sensors to pinpoint load locations. The researchers conducted Controlled Vehicle Loading (CVL) testing in order to better analyze the perpetual pavement structures. Two axle configurations were used including a single axle dual tire and a tridem axle dual tire. The single axle was tested using loadings of approximately 20, 27, 34, and 40 kips while the tridem axle was tested using loadings of approximately 94, 108, 121, and 135 kips (Jincheng et al., 2012). Jincheng reported that the tire pressures were around 116 psi and testing was conducted at 18 mph. Yang noted that five passes of each testing condition were performed. As expected, strain increased as section thickness was reduced or loading increased (Timm et al., 2010). Timm discussed how strains recorded in the perpetual sections were significantly less than those found in the standard sections. Additionally, strain measured in all the perpetual sections were ordinarily low enough for bottom-up fatigue cracking to be limited when using an FEL of 125 µɛ. The results of all three perpetual sections indicate that they were likely to be perpetual and the 20 inch section was believed to be overdesigned Marquette Marquette University instrumented a single perpetual pavement test section constructed on an interchange of Interstate 43 in downtown Milwaukee. The area was prone to heavy amounts of traffic so the section was perfect to analyze pavement responses under real traffic conditions (Hornyak, 2010). Hornyak (2010) reported that the section was constructed with a total pavement thickness of 13 inches consisting of a

56 56 four inch base layer, seven inch intermediate layer, and a two inch SMA surface course. Furthermore, the asphalt layer was built on top of a three layer aggregate base which included four inches of open graded aggregate, followed by six inches of dense graded aggregate, and finally, 18 inches of crushed aggregate. The section was instrumented with various dynamic sensors, as described by Hornyak (2010). Strain gages were arranged in three separated groups of three rows perpendicular to traffic along the wheel path. Row one contained three longitudinal strain gages spaced 12 inches from each other. Row two consisted of two transverse strain gages while row three comprised of three transverse strain gages both with 24 inch spacing s. In between each of the groups of strain gages were two earth pressure cells. One pressure cell was located on the natural subgrade while the other was on the DGAB. Temperature probes were installed in the pavement as well as the natural subgrade. The probes measured temperatures at one inch intervals throughout the depth of the asphalt layer. Surface temperature was attained using an infrared thermometer. A wheel wander system was installed consisting of piezo strips positioned in a sideways N pattern as well as a weigh-in-motion (WIM) sensor placed directly after the pressure cells. Results from the project were expansive due to the fact that data was collected year round. In order to evaluate the performance of the pavement, Hornyak (2010) analyzed data for 100 randomly selected vehicles ranging from class four to ten during one week in September of Moderate temperatures were seen during this month with an average of 59 F. Hornyak reported that a majority of results obtained from this sample were below 12 µɛ. When examining all of the data that had been collected, he

57 57 found that peak strain readings were recorded in the month of August when pavement temperatures were the highest. He also discovered that the strain gages never measured a strain greater than 15 µɛ and only 10% of strains exceeded 5 µɛ Oregon The Oregon Department of Transportation (DOT) along with Oregon State University performed a field study in 2005 of a perpetual pavement test section constructed on Interstate 5. A journal article on the project written by Estes (2005) revealed that the section was not originally intended to be perpetual but after analysis was performed by Oregon State it was discovered that the section possessed perpetual qualities. The objectives of the field study were to gain valuable information to further guidelines for future designs of perpetual pavement systems (Estes, 2005). Scholz (2006) described the section as having a total pavement thickness of 12 inches with two inches in the surface layer and a ten inch base layer. He further defined the section as being built on top of an eight inch rubberized joint reinforced concrete pavement (JRCP) as well as a nine inch aggregate base. The section was compared to a standard 12 inch section constructed directly on 16 inches of aggregate (Scholz et al., 2006). A large quantity of instrumentation was implemented into the pavement section in Oregon. Scholz (2006) illustrated the instrumentation as 12 strain gages placed directly on the aggregate base, six longitudinal and six transverse, and 12 placed on the rubberized JRCP, also including six in the longitudinal direction and six in the transverse. He described the gages as being located along the wheel path and having an orientation

58 58 of four rows of three with two foot spacing s between them. Four temperature probes were used to record pavement temperature and a WIM sensor along with a classifier loop were installed before the instrumentation in order to record loading information and location (Scholz et al., 2006). Estes (2005) disclosed that data was collected from the section for a year and Scholz (2006) discovered the following results. Longitudinal strain gages experienced compressive strain directly before and after being loaded during which they experienced tensile strain. Transverse strain gages did not experience any compressive strain. Longitudinal strain s magnitude was discovered to be greater in the standard section. Pavement seemed to be behaving elastically in the longitudinal direction but had a more viscous response transversely. Longitudinal and transverse strains recorded have been less than the 70 µɛ threshold Advanced Transportation Research and Engineering Laboratory The Advanced Transportation Research and Engineering Laboratory (ATREL) at the University of Illinois Urbana-Champaign branch campus has an Accelerated Transportation Loading Assembly (ATLAS) for evaluating pavement systems under realistic loading conditions. The loading facility is capable of inducing 10,000 loading cycles on an experimental pavement test pad per day (Al-Qadi, Wang, Yoo, & Dessouky, 2008). The assembly is capable of producing various loads along with multiple tire configurations. Al-Qadi et al. (2008) wrote a report containing an evaluation of testing performed on two perpetual pavement test sections built at the ATREL. They described the test

59 59 sections as having total pavement thicknesses of approximately of 16.5 inches. One of the test sections utilized an SMA surface course and an FRL with an increased binder content of 5.1%. The second test section incorporated a standard superpave surface layer and a base layer with a binder content of 4.5%. Moreover, two control sections were tested as well; one had a total pavement thickness of about ten inches and the other, six inches. Al-Qadi detailed that all four sections were constructed on top of a lime modified subgrade. Instrumentation of the test sections, as illustrated by Al-Qadi, included H-type strain gages located on the top of the subgrade, both in the longitudinal and transverse directions, and thermocouples throughout the thickness of the asphalt layer and subgrade. Testing of the sections at the ATREL was performed using a load of approximately ten kips with two tire configurations including a dual-tire and a wide-base tire (Al-Qadi et al., 2008). Al-Qadi stated that the tires were filled to a tire pressure of 100 psi and two speeds of five mph and ten mph were implemented in the study. He noted that although test sections were outside and exposed to various temperatures, strain measurements were normalized to 77 F. Al-Qadi et al. (2008) reported numerous findings from the testing results they obtained. He noticed that longitudinal strain included both a compressive and tensile zone unlike the transverse gages which only experienced tensile strain. Longitudinal strain was also measured to be greater in magnitude than transverse strain. This difference seemed to decrease as the thickness of the pavement increased. Strains recorded in the perpetual test sections were reduced when compared to those found in the standard sections although strain in the perpetual sections were only slightly lower than

60 60 strains in the ten inch standard section. Al-Qadi et al. discovered that strain was higher for the wide-base tire configuration compared to the dual-tire and that when using the wide-base tire, strain was effected more by loading. He also noted that strain decreased as speed increased. When comparing differences between the three various pavement thicknesses, Al-Qadi et al. found that tensile strain was strongly influenced by the thickness of the asphalt layer Wisconsin In 2003, the Wisconsin DOT performed an evaluation of two perpetual pavement test sections constructed on an entrance ramp to a weigh station along Interstate 94. Their objectives included using evaluations made from test sections to provide valuable information on perpetual pavement systems and to develop an optimal design methodology (Battaglia, Bischoff, Ryan, & Reichelt, 2010). They also wanted to analyze the effect that various materials in different pavement layers had on the structure. Battaglia et al. (2010) describes the two test sections, which were built in 2003, as consisting of an 11 inch asphalt layer comprising of three layers constructed on top of a four inch open graded base course, all of which was constructed on 17 inches of dense graded crushed aggregate. The subgrade soil was reported to have medium strength qualities. A total of 16 strain gages were installed in one of the test sections including eight between the open graded base and bottom layer of asphalt and eight between the base and intermediate asphalt layers (Battaglia et al., 2010). In each of the layers four of the gages were oriented in the longitudinal direction and four were oriented in the longitudinal direction. Unfortunately, Battaglia et al. reported that only three of the strain

61 61 gages survived, two between the bottom and intermediate asphalt layers and one between the open graded base course and bottom asphalt layer. Battaglia et al. (2010) conducted Falling Weight Deflectometer (FWD) testing in the months of September and October using 5, 9, and 12 kips as impact loads. It was discovered that strains were higher at the bottom of the asphalt layer than they were in between the bottom and intermediate asphalt layers. The largest strain at the base of the asphalt layer, while testing with 12 kips, was recorded as 85 µɛ while the largest strain recorded between the bottom and intermediate asphalt layers was 37 µɛ. Truck testing was also conducted by Battaglia et al. (2010) in the months of April and July where pavement temperatures ranged between 80 F and 91 F and 90 F and 103 F respectively. It was discovered that higher temperatures resulted in higher strain values. Using the same load and a speed of 55 mph, strain measurements at the bottom of the asphalt layer revealed a 35 µɛ difference between testing conducted in April compared to July with April recording a strain of 15 µɛ and July with 50 µɛ. It was also discovered that lower speeds resulted in higher strains. In July the same truck produced a strain of 69 µɛ when traveling at 26 mph and 48 µɛ when traveling at 54 mph. After analyzing all of the results Battaglia et al. reported that there seemed to be a linear correlation between strain and load. Also, the perpetual pavement system seemed to function effectively with adequate foundation support and good structural condition. After being opened to traffic, which was solely heavy truck traffic traveling though the weigh station, visual surveys were periodically made through 2009 by Battaglia et al. (2010). By 2009, the pavement began to experience alligator cracking in

62 62 the wheel paths with little transverse and longitudinal cracking. Some aggregate popouts were noted and rutting was less than 1/10 inches after six years of service. The pavement system was deemed to have a good ride quality and cores taken from the sections exposed no bottom-up cracking Kansas In order to investigate the practical use of perpetual pavement structures, the Kansas DOT constructed and instrumented four perpetual pavement structures on U.S. Route 75. The four sections were constructed in 2005 on top of a subgrade consisting of a high plasticity clay with a resilient modulus of 2,500 psi (Portillo, 2008). Portillo provides that the total pavement thickness of the sections ranged from 11 inches to 16 inches. Thickness differences were implemented in the base layer and a variety of binder grades were used in the intermediate and base layers. He adds that each section was constructed on top of six inches of lime stabilized subgrade. Each test section included the following instrumentation as described by Portillo (2008). Strain gages were positioned on top of the lime treated subgrade. Four longitudinal and transverse strain gage pairs were installed directly on the wheel path while four pairs were installed 6 inches to the right of the wheel path. One pressure cell was positioned directly on the wheel path. Temperature probes were located at middepth of each asphalt layer. Testing was conducted seven times between July 2005 and October 2007 using a class five single axle dump truck (Portillo, 2008). Portillo states that for each testing session, three sets of five passes were accomplished for each pavement section consisting

63 63 of three different speeds; mph, mph, and mph. Wheel loads ranged within approximately a kip of ten kips for each session. Portillo (2008) discovered that speed and temperature had large effects on strain readings. The highest strains were recorded in August while the lowest strains were recorded in October. Additionally, there seemed to be a linear relationship between strain and temperature. Romanoschi (2006) noted that strains did not seem to have a linear relationship with speed. As speed decreased the rate at which strains increased intensified. Abnormally, as described by Portillo, transverse strains were greater in magnitude than longitudinal strains. Although strains were typically lower than 70 µɛ, sections with pavement thickness s less than 16 inches experienced some longitudinal strains greater than 70 µɛ when the truck was traveling at 20 mph along with transverse strain exceeding 70 µɛ at all three speed increments (Portillo, 2008). Portillo reported that pressure seemed to decrease with increasing speeds and temperatures with a maximum pressure recording of six psi. Strain measurements produced by the steering axle were typically between 50% and 70% of the strains produced by the rear axles suggesting that assessing damages caused by the steering axles ought to be considered (Romanoschi, Gisi, Portillo, & Dumitru, 2008) National Center for Asphalt Technology The National Center for Asphalt Technology (NCAT) Test Track is a pavement testing facility at Auburn University. The NCAT Test Track is a very realistic way to subject pavements to lading conditions at an accelerated rate (Brown et al., 2002). The track allows several different pavement sections to be tested at the same time. Brown

64 64 states that the facility includes 46 test sections. Various studies have been conducted at the NCAT Test Track since its inception in As part of Phase III testing at the NCAT Test Track, eleven perpetual pavement test sections were constructed in order to evaluate the structural performance effects of thickness and polymer additions. Six sections were constructed with three varying thicknesses, 17, 15, and 13 inches. Each thickness incorporating a section using unmodified binder and a section using polymer modified binder (Timm, 2009; Willis, 2009). Timm mentions that two additional sections were built, one with an FRL and one without. These additional sections also incorporated an SMA surface course. The final three sections included two with total pavement thicknesses of 15 inches and one with a total pavement thickness of 13 inches with an SMA surface course (Willis, 2009). A report composed by Robbins and Timm (2008) illustrated the instrumentation used for Phase III. Pressure plates were installed on the surface of the subgrade and base layer. Twelve strain gages, both in the longitudinal and transverse orientations, were placed at the base of the asphalt layer both along the wheel path and offset to both sides. Three temperature probes were installed to monitor pavement temperature on the surface, at mid-depth, and base of the asphalt layer along with the granular base. Robbins and Timm (2008) described testing performed on four different dates. They reported that loading was applied to the sections by a 12 kip steer axle, a 40 kip tandem axle, and a 20 kip single axle. Testing was performed at speeds of 15, 25, 35, 45, and 55 mph. Included in the report was surface temperature data which ranged between 70 F and 125 F during testing.

65 65 According to Robbins and Timm (2008), longitudinal and transverse strain decreased in a logarithmic fashion with increasing speeds. They discovered that the rate at which the strain changed due to speed amplified as the temperature increased. Hence, at higher temperatures, speed was more influential on the strain induced at the bottom of the asphalt layer. This trend existed for both axle types. Robbins and Timm also revealed that strain exponentially increased with pavement temperature increases for both the transverse and longitudinally oriented gages. Equations to represent the relationship between temperature and strain were developed and resulted in high correlations between equations and experimental data. This was also reported in Willis s (2009) thesis along with documentation stating that the two 15 inch sections had still experienced no fatigue cracking after receiving over 19 million ESALs. Additionally, although all sections had recorded strains above 70 µɛ, the 14 inch section utilizing the SMA surface course and the FRL was experiencing the lowest strain values New York In 2008, Ohio University and the New York DOT instrumented a perpetual pavement test section in hopes of validating New York DOT designs and to further investigate the concept of perpetual pavements. Sargand, Khoury, and Morrison (2012) describe the test section as a 7.5 inch asphalt layer constructed on top of nine inches of rubberized Portland cement concrete (PCC). A standard section was instrumented for comparison purposes (Bendana, Sargand, & Hernandez, 2009). Sargand et al. (2012) explains the instrumentation as three transverse and five longitudinal strain gages in the base layer and two transverse and six longitudinal strain

66 66 gages in the intermediate layer. Moreover, two pressure cells were installed on the subgrade surface and two deep and shallow linear variable displacement transducers (LVDT) were installed to reference subgrade displacement and total pavement system displacement. Finally, thermocouples were placed in the base and intermediate layers to monitor pavement temperature. CVL testing was conducted in October of 2008 and May of 2009 (Bendana et al., 2009). Overall, Sargand et al. (2012) found that strain results were typically below 70 µɛ. Bendana et al. (2009) reported that strains in the perpetual pavement test section were lower than strains in the standard test section. He found it interesting, however, that just after construction the standard section resulted in lower strain values than the perpetual section. Also, pavement deflections were discovered to be higher in the standard test section when compared to results from the perpetual test section and in both sections deflections were increased in May when pavement temperatures were higher. Perpetual section results contained more variation in pavement response between May and October than the standard section. Sargand et al. commented that during visual inspection between 2008 and 2011, no damage requiring reconstruction was noted even after the pavement had been subjected to heavy truck loading and weathering. Sargand et al. also mentions that during FWD testing strain resulting in the base layer was around 100 µɛ when using a 16 kip impact load. The pavement temperature was recorded at 84 F during FWD testing.

67 Accelerated Pavement Loading Facility The Accelerated Pavement Loading Facility (APLF) is an indoor pavement testing facility located at the Ohio University Lancaster branch campus. The facility is capable of representing tire loading and environmental effects. Tire loading can be simulated at speeds of up to five mph and at loads of up to 30 kips (Hernandez, 2010). Hernandez (2010) denotes that the facility is unique in that it can simulate various environmental effects including air temperatures ranging between 10 F and 130 F in addition to any percentage of humidity. He adds that the facility also incorporates the use of underground pipes to allow the subgrade moisture to be manipulated. Ohio University evaluated four perpetual pavement sections at the APLF. The perpetual pavement sections built at the APLF were used to evaluate the impact of various perpetual pavement thicknesses. Sargand, Figueroa, Edwards, and Al- Rawashdeh (2009) describe the sections as including an A6-A7 subgrade soil, a six inch DGAB, and the use of FRLs. They continue by adding that the total pavement thicknesses of the sections were 13, 14, 15, and 16 inches, which were accomplished by altering the thickness of the intermediate layer. It was noted that different types of Warm-Mix Asphalt (WMA) were used in the surface layer but had negligible effects on the perpetual pavement behavior. Instrumentation was installed in the sections in order to experimentally investigate the sections. Hernandez (2010) explains the instrumentation as including two longitudinal and two transverse strain gages spaced 18 inches apart located at the bottom

68 68 of the asphalt layer. Additionally, one pressure cell was placed on top of the subgrade and deep and shallow LVDTs were used. Testing was performed at various temperatures using multiple loads. The pavement sections were subjected to loadings of 6, 9, and 12 kips at temperatures of 40 F, 70 F, and 104 F (Hernandez, 2010). Sargand et al. (2009) state that the test sections were tested before and after being subjected to 10,000 loading cycles of a nine kip wheel load. They also add that both the loading cycles and testing were performed at five mph. Many response trends were discovered by Sargand et al. (2009) as a result of their experiment. Transverse and longitudinal strains in the FRL were only slightly higher as the thickness of the structure decreased independent of temperature and loading. These facts lead them to believe that similar effects would be seen by increasing the base structure as increasing the asphalt structure. Longitudinal strains remained below 70 µɛ except for a few outliers during test runs with 12 kips of loading at 104 F. Increases in longitudinal strain were seen when increasing either load or temperature. Pressure, like strain, increased with increasing temperature and/or load ranging between approximately two and ten psi. Pressure difference between section thicknesses was small. Trends along with strain and pressure magnitudes were similar before and after 10,000 passes of nine kip wheel load indicating the sections endured the 10,000 passes well. Deflections of the subgrade increased with increasing temperature and/or load especially for thicker sections. Deflections were reduced after 10,000 passes of 9 kips. Hernandez s (2010) report made additional observations by adding that the longitudinal strain went from

69 69 compressive to tensile and back to compressive again while being loaded whereas the transverse strain was only subjected to tensile strain. Also, longitudinal strain seemed to increase more rapidly as temperature increased while the transverse strain had a linear relationship with temperature creating an increased difference between the two at higher temperatures. Finally, he added that temperature had less effect on strain as pavement thickness was increased Interstate 77, Canton, Ohio The first instrumented perpetual pavement section constructed in Ohio was located in Canton, Ohio on Interstate 77. Sargand and Figueroa (2010) recounted that the section had a total pavement thickness of inches, which included an FRL and an SMA surface course. The section was reported to have been instrumented with three longitudinal and three transverse strain gages at the bottom of the FRL, two pressure cells located on the subgrade surface, and thermocouples installed near the strain gages. CVL testing of the section took place on December 15 of 2003 (Sargand & Figueroa, 2010). Sargand and Figueroa (2010) denoted the testing as a single axle dual tire producing a 13.5 kip load at speeds of 5, 30, 40, and 50 mph. The pavement temperature that evening was measured at 36 F. No further testing has been completed because of high traffic volumes on the section. Due to minimal testing only a few conclusions could be made by Sargand and Figueroa (2010). The first was an increase in strain with decreasing speeds. The maximum longitudinal and transverse strains, 35.6 µɛ and 24.4 µɛ respectively, were recorded at a creep speed of five mph. Strains received at 50 mph were slightly lower at

70 µɛ and 20.0 µɛ for longitudinal and transverse respectively. Pressures measured by the pressure cells were minimal for all increments of speed U.S. Route 30, Wayne County, Ohio In the fall of 2005, the Ohio DOT and the Ohio Research Institute for Transportation and the Environment (ORTIE) constructed a perpetual pavement test section on U.S. Route 30 in Wayne County, Ohio. Sargand, Khoury, Romanello, and Figueroa (2006) label the route as being a rural freeway. Three identical test sections were instrumented in order to analyze performance and design assumptions. Sargand, Romanello, and Figueroa (2008) made the following descriptions on the three instrumented test sections. The sections had total pavement thicknesses of inches including a four inch FRL and an SMA surface course. The asphalt layer was constructed on a six inch crushed DGAB all of which was supported by a subgrade with a CBR of between four and six. Strain gages were installed in the FRL in order to monitor fatigue resistance as well as in the intermediate layer to evaluate the pavement s potential for cracking. Strain gages were oriented in the longitudinal and transverse directions. Shallow and deep LVDTs were used to measure displacement of the total pavement system or subgrade. Subgrade pressure was measured through the use of pressure cells installed on the subgrade surface. An automated weather station was constructed to monitor air temperature, precipitation, wind speed and direction, relative humidity, and incoming solar radiation. CVL testing was conducted in December of 2005, July of 2006, and May of Testing performed in December of 2005 was done prior to the roadway being opened to

71 71 traffic. A standard Ohio DOT single and tandem axle truck, typically used for salt spreading and snow removal, were the testing vehicles (Sargand et al., 2008). Sargand et al. reported pavement temperatures of 30 F to 35 F were experienced in December and 95 F to 126 F in July. Pavement temperatures ranging between 56 F and 74 F were reported by Restrepo-Velez (2011) for the May round of testing. Varying between approximately 20 to 30 kips and 35 to 50 kips were axle loadings used for the tandem and single axle, respectively (Restrepo-Velez, 2011; Sargand et al., 2008). Testing was conducted in a similar fashion during all three sessions. A large amount of quality results were found for the U.S. Route 30 project since testing was conducted at various speeds, temperatures, loads, and axle configurations. Although the tandem axle truck carried a greater load than the single axle truck, strains were found to be higher for the single axle truck than for the tandem axle truck while both axle types produced higher strains at lower speeds as depicted by Restrepo-Velez (2011). However, this was not the case for transverse strains recorded in the intermediate layer. She discovered that strains were higher in the FRL than the intermediate layer. Also, Restrepo-Velez found that longitudinal strains experienced tension and compression while transverse strain experienced only tension. Sargand et al. (2006) reported that pressures in the subgrade were higher due to the tandem axle than the single axle. Furthermore, pressures typically decreased with increasing truck speeds. Although no deflections were recorded exceeding 20 mils, greater deflections were produced as temperature was increased (Romanello, 2007; Sargand et al., 2008). Romanello (2007) added that this effect was less profound in the subgrade. Sargand et al. (2008) concluded

72 72 that although some strains were recorded above the 70 µɛ threshold, they were produced at abnormally slow speeds and heavy loads and therefore under typical conditions the pavement would rarely be subjected to strain values in the FRL exceeding the FEL Summary Many similar discoveries and trends can be observed among research previously described in this chapter. Instrumentation almost always included strain gages located at the bottom of the asphalt layer along with pressure cells installed on the subgrade surface. Furthermore, temperature probes were used to monitor pavement temperature. Some researchers also implemented strain gages in the intermediate layer and/or displacement sensors. All of the research showed an increase in pavement response, strain, pressure, and displacement, with increasing pavement temperatures. Additionally, a large amount of the findings reported an exponential relationship between strain and temperature. Researchers tended to find an increase in strain with a decrease in speed as well with some researchers denoting a linear relationship. Strain results typically decreased with decreasing loads as well as an increase in total pavement thickness. If multiple axle configurations were used, it seemed as though tandem axles, even when carrying higher loads, generated lower strains. This was not necessarily the case when analyzing pressure and displacement data.

73 73 CHAPTER 3: U.S. ROUTE 23 PROJECT BACKGROUND 3.1 Project Site Description As part of the ODOT s Partnered Research Exploration Program, ODOT and the Ohio Research Institute for Transportation and the Environment (ORITE) continued research with intentions of developing optimized perpetual pavement designs in Ohio through the instrumentation of four pavement test sections with varying thicknesses on U.S. Route 23 in Delaware, Ohio. The project was focused on analyzing the impact of varying total pavement thicknesses by monitoring data obtained during controlled vehicle load (CVL) testing. Research findings should help provide further understanding of the optimal design thickness for perpetual pavements in order to minimize their life cycle costs. The Strategic Highway Research Program (SHRP) Test Road on U.S. Route 23 in Delaware County s location is displayed in Figure 3.1. The location is indicated with a green point.

74 74 Figure 3.1 Location of SHRP Test Road (Geology, 2013) The U.S. Route 23 Project consisted of four asphalt concrete perpetual test sections. Two of the test sections were constructed on North Waldo Road as indicated by the red arrow in Figure 3.2. The two sections were constructed adjacently to each other, section 39BN803 in the northbound lane and section 39BS803 in the southbound lane. The two remaining test sections were constructed on the northbound lanes of the U.S. Route 23 mainline, indicated by the green arrow in Figure 3.2. These sections were also adjacent to each other with section 39D168 residing in the driving lane and section 39P168 in the passing lane.

75 75 Figure 3.2 Location of Test Sections (picture taken by ODOT) 3.2 Design of Test Sections The perpetual pavement sections were designed to have service lives of 50 years or longer without major rehabilitation by ODOT. Designs were developed using the Mechanistic Empirical Pavement Design Guide (MEPDG) levels one, two, and three and Marshall Mix Designs based on data previously obtained from existing sections on Route 30 in Wayne County and the Accelerated Pavement Load Facility (APLF). A practical perpetual pavement design was ultimately used. Although layer thicknesses may vary between sections, as described later, Table 3.1 lists specifications for each layer used.

76 76 Table 3.1 Layer Specifications Layer Description PG Binder Grade % Binder Content Surface Layer Fine Graded Polymer Asphalt Concrete PG 76-22M 7.6 Intermediate Layer Asphalt Concrete, 19 mm PG Base Layer Asphalt Concrete PG Fatigue Resistant Layer Fatigue Resistant Base Layer PG Table 3.2 provides layer thicknesses for each section. Each section was constructed on a six inch DGAB. The subgrade for all four sections had a soil classification of A-6. In addition, the subgrade for sections 39BS803 and 39BN803 had been chemically stabilized through lime treatment. Testing conducted by Ohio University revealed that the subgrade beneath sections 39D168 and 39P168 had an average resilient modulus of 20 ksi while the subgrade for section 39BS803 and 39BN803 had an average resilient modulus of 40 ksi. It is important to note that after construction of the test sections, core samples obtained revealed that section 39P168 had been constructed with a total pavement thickness of 16 inches, a three inch difference from what had been designed. Due to this discrepancy, data obtained from this section was not presented in this report. Table 3.2 Layer Thicknesses Layer 39D168 39P186 39BS803 39BN803 Surface Layer (in.) Intermediate Layer (in.) Base Layer (in.) Fatigue Resistant Layer (in.) Total Pavement Thickness (in.)

77 Instrumentation of Test Sections In order to evaluate pavement responses while being exposed to dynamic loading and various environmental conditions, numerous sensors were installed throughout the pavement structure. ORITE developed a comprehensive instrumentation plan including the use of thermocouples to measure pavement temperature throughout the depth of the pavement and air temperature. Linear Variable Differential Transformers (LVDT) were installed in order to measure deflections of the pavement structure as well as the subgrade exclusively and strain gages to monitor horizontal pavement strains. Pressure cells recorded subgrade and FRL pressures. Horizontal strain gages were installed in every asphalt layer in the longitudinal, transverse, and 45 positions. Furthermore, for each section excluding 39BN803, a square and round hole were cored through the entire depth of the pavement along the instrumentation line and strain gage rosettes were bonded to their walls to acquire additional strain measurements including the vertical direction Strain Gage Installation Strain gages were used for the U.S. Route 23 project to measure pavement strains. Strain gages measure strain by linking the change in electrical resistance of the strain gage s semiconductor or metallic foil pattern to the amount of strain it is experiencing (Portillo, 2008). As the semiconductor or metallic foil pattern is stretched or compressed its electrical resistance is increased or decreased respectively which then can be correlated to a value of strain. Proper installation of strain gages is essential for the sensor to work properly.

78 78 Two different strain gages were used for the U.S. Route 23 project. A majority of the gages were embedment strain gages of type KM-100HAS, provided by Tokyo Sokki Kenkyujo Co. These gages had a 350 ohm full bridge resistance and are displayed in Figure 3.3. The second type of strain gages used were asphalt strain gages of type PMFLS-60, also provided by Tokyo Sokki Kenkyujo Co. In contrast, these gages had a 120 ohm quarter bridge resistance. The PM gages are shown in Figure 3.4. Figure 3.3 KM Strain Gage (photo taken by Jaime Hernandez)

79 79 Figure 3.4 PM Strain Gage (photo taken by Jaime Hernandez) For all of the sections excluding section 39BN803, four longitudinal, one transverse, and one 45 oriented strain gage were installed at the bottom of the FRL. Also, the base layer was instrumented with three longitudinal and transverse strain gages accompanied by two 45 strain gages. Finally, the intermediate layer contained two sets of a longitudinal, transverse, and 45 strain gage. Section 39BN803 utilized six longitudinal strain gages in the FRL and three longitudinal and three transverse strain gages in the base layer. It did not contain any gages in the intermediate layer. A detailed layout of the entire instrumentation plan is presented at the conclusion of section 3.3. Preceding the paving of the asphalt layer receiving instrumentation, strain gage location layouts were painted using LVDT stakes as reference points. This is depicted in Figure 3.5. After their locations were marked, strain gage lead wires were stretched to their proper pull boxes on the side of the road and buried in the DGAB (Figure 3.6), when instrumenting the FRL, or taped to the asphalt surface (Figure 3.7), when instrumenting the base and intermediate layers. In order to prevent the paver from damaging or shifting the strain gages they were embedded and buried, by hand, in a thin

80 80 layer of asphalt (Figure 3.8) before equipment laid down the entire lift. Asphalt used to embed and bury the strain gages was attained directly from the truck and sifted in order to remove large aggregates that might have punctured the strain gages or lead wires (Figure 3.8). For further protection of the instrumentation, during compaction, two passes of the compaction rollers were made before vibratory roller compaction was used. All of the strain gages installed were found to be working properly after the pavement had cooled. Figure 3.5 Instrumentation Layout in the Field

81 81 Figure 3.6 Wire Installation in DGAB Figure 3.7 Wire Installation on Asphalt Surface (photo 4 taken by Jaime Hernandez)

82 82 Figure 3.8 Strain Gage Installation (photos 2, 3, and 5 taken by Jaime Hernandez) Due to the thin nature of the surface layer, a slightly different approach was used when installing these gages. In all of the sections except for section 39BN803, one longitudinal, transverse, and 45 oriented strain gage was installed. This was accomplished by leaving a thin piece of metal approximately the size of the strain gage in the proper location and orientation while the surface layer was paved. Figure 3.9 illustrates the metal plates after the surface layer had been paved. After the paving was complete and the asphalt had cooled, these metal plates were removed and a strain gage was bonded in its place. The gages used for the surface layer were 120 ohm, quarter bridge, type WFLM-60, provided by Tokyo Sokki Kenkyujo Co.

83 83 Figure 3.9 Surface Strain Gage Installation Pressure Cell Installation For the U.S. Route 23 project, pressure cells were utilized in order to monitor vertical pressure. Subgrade pressure is of high importance when evaluating perpetual pavement systems. For this purpose, two Geokon 3500 pressure cells were installed on the subgrade surface for each section. Figure 3.10 displays the Geokon 3500 pressure cell. These pressure cells are created to measure changes in stress applied to their surface (Hornyak, Crovetti, Newman, & Schabelski, 2007). These pressure cells measure pressure though the use of a pressure transducer connected to liquid, a high stiffness oil, confined between two steel plates (Portillo, 2008). When a load is applied to the flat circular surface, the plates deflect causing the fluid pressure to increase.

84 84 Figure 3.10 Geokon Earth Pressure Cell In order to avoid construction equipment driving directly over the earth pressure cells, they were installed after the DGAB had been graded. At this point, holes were excavated down to the subgrade along with trenches leading off the road large enough to encompass the earth pressure cells and their lead wires. The pressure cells were installed using a layer of sand above and below them to provide a uniform surface before being buried by the aggregate base. This process is illustrated in Figure Figure 3.11 Subgrade Pressure Cell Installation

85 85 In addition to the subgrade pressure cells, two pressure cells were installed on the surface of the FRL in every section except for 39BN803. These pressure cells were different from those used on the subgrade. They were Total Earth Pressure Cells purchased from RST Instruments (Figure 3.12). A unique feature of these pressure cells was a kink in the tube allowing the cell to lie flat on the hard, DGAB surface while their handles were buried in the DGAB. Similarly, the pressure cells were first surrounded by sand to provide them with a uniform surface. Then, directly before paving took place, the pressure cells were hand compacted in a thin layer of sifted asphalt in a similar fashion as the strain gages. Their lead wires were installed similar to that of the strain gages as well. Figure 3.13 displays the instrumentation of the FRL pressure cells. Figure 3.12 RST Total Earth Pressure Cell (photo taken by Jaime Hernandez)

86 86 Figure 3.13 FRL Pressure Cell Installation (photos taken by Jaime Hernandez) Thermocouple Installation Two thermocouples were installed in the instrumentation line of every asphalt layer receiving instrumentation. Single point thermocouples purchased from Measurement Instruments were used to measure pavement temperature of the various asphalt layers. Thermocouples rely on the thermoelectric principle where two different materials, when exposed to a temperature gradient, generate two different voltages which are correlated to the temperature (Portillo, 2008). An additional thermocouple was used to measure air temperature. Thermocouples were installed in the same way as strain gages. In Figure 3.14, item 1 is a photograph of the thermocouples implemented in the Route 23 project along with items 2 through 4 which are pictures displaying the typical arrangement of strain gages and thermocouples.

87 87 Figure 3.14 Thermocouple and Strain Gage Alignment Linear Variable Differential Transducer Installation The LVDTs utilized were used to measure vertical deflections and therefore the pavement s potential for rutting and an indicator of subgrade performance. LVDTs measure displacement through three coils linked together by a magnetic core. The primary coil is located on one side of the magnetic core while the two secondary coils are positioned next to each other on the other side of the magnetic core. While the magnetic core is centered between the secondary coils the voltage transferred to the primary coil from the secondary coils is identical but as displacement is induced and the magnetic core moves further from one of the secondary coils, and towards the other, the voltage in the closer secondary coil increases while the voltage in the other secondary coil decreases (Portillo, 2008). This change of voltages transferred by the two secondary coils to the

88 88 primary coil correlates to a displacement measurement. The LVDTs used for the U.S. Route 23 project were purchased from Trans-Tek. An example of one is displayed in Figure Before the LVDTs were installed, they were calibrated. Calibration data can be found in Appendix A. Figure 3.15 LVDT A diagram of the LVDT case and reference rod for the 13 inch section is displayed in Figure Diagrams of the LVDT cases and reference rods for all sections can be found in Appendix B. As shown in the Figure, the LVDT was connected to a plate inside the LVDT case. The LVDT case was installed in the asphalt concrete so that when the surface of the pavement displaced in the vertical direction, the LVDT case and thus the LVDT shifted as well. During these vertical displacements the LVDTs measured off of deep and shallow reference rods. Deep LVDT reference rods extended 11 feet below the subgrade surface where displacements, at this point, were insignificant. Displacements measured off of deep reference rods represented the displacement of the entire pavement system. Shallow reference rods, as shown in Figure 3.16, extended to the subgrade surface. By subtracting the displacements measured from the shallow

89 89 reference rods from the displacements measured from the deep reference rods the displacement of the subgrade was obtained. Figure 3.16 LVDT Case and Reference Rod Diagram LVDT installation began after the DGAB had been placed. At this point, the LVDT locations were surveyed (items 3 and 4 of Figure 3.17) and holes were drilled approximately ten feet deep in the subgrade in order to install the deep LVDT reference rods (item 1 of Figure 3.17). PVC piping was used to prevent the hole from collapsing (item 2 of Figure 3.17) and reference rods were pounded another foot or two into the bottom of the hole (item 5 of Figure 3.17). Next, the deep reference rods were grouted

90 90 into place. Holes were then excavated down to the subgrade surface and shallow LVDT plates and reference rods were placed, as shown in item 4 of Figure Surveys were then made so that the location of the LVDT reference rods could be located after the completion of paving. Figure 3.17 LVDT Reference Rod Installation After the completion of paving, cores were made through the pavement in order to gain access to the LVDT reference rods. Then, rod extensions and caps were connected to the existing rods to elevate them to the proper height. LVDTs were housed in cases that were connected using bolts and proproxy adhesive in the core holes. The caps of the LVDT cases were flush with the pavement surface. Figure 3.18 illustrates the installation process of the LVDT cases.

91 91 Figure 3.18 LVDT Case Installation Strain Gage Rosette Hole Installation A unique feature implemented in the U.S. Route 23 project was the use of vertical rosettes installed along the walls of a round and square hole in section 39BS803, 39P168, and 39D168. The round holes had diameters of six inches and the square holes had widths of six inches. Strain gage rosettes were installed in the holes on the walls parallel and perpendicular with traffic. Two strain gages rosettes per wall were placed in the base layer while one strain gage rosette was placed in the intermediate and surface layers. Surface layer strain gage rosettes were not implemented in the round holes. A diagram for the square rosette hole in the 13 inch section is displayed in Figure The square rosette holes were built with identical dimensions. Dimensions of the base layer strain gage rosettes varied depending on the thickness of the base layer. Diagrams for square and round rosette holes for all sections containing them can be found in Appendix C. The gages used for the rosette holes were 120 ohm, quarter bridge, type WFLM-30, provided by Tokyo Sokki Kenkyujo Co. Due to limitations involved with the thickness of the surface layer, type WFLM-10 gages were used for the surface layer.

92 92 Figure 3.19 Strain Gage Rosette Hole Diagram The rosette hole installation process began after the DGAB was placed. At this point, PVC pipes were buried in the DGAB from the location of the rosette holes to the edge of the road. This was done so that strain gage wires could be snaked underneath the pavement after the rosette holes were constructed. Next, after the completion of paving, six inch cores were made in the locations of the rosette holes (item 1 of Figure 3.20). A grinder was used to square off the square rosette holes as shown in item 2 of Figure Then, strain gage rosettes were bonded to the walls of the holes in their proper locations

93 93 through the use of a rubber mat (items 3 and 4 of Figure 3.20). For the square holes the strain gage rosettes were held in place with a clamp while the adhesive dried while a balloon was used to keep pressure on the strain gage rosettes in the round holes (item 5 of Figure 3.20). Finally, lead wires were snaked through the PVC piping. Figure 3.20 Strain Gage Rosette Hole Installation A diagram illustrating the entire instrumentation layout for the 13 and 15 inch sections is presented in Figure Instrumentation diagrams for the 11 inch section is shown in Appendix D. Yellow represents strain gages which can be found in the fatigue resistant, base, and intermediate layers. Beneath the strain gage locations in the profile view an L, T, or 45 is given indicating the orientation of the strain gage with L, T, and 45 representing longitudinal, transverse, and 45 respectively. Red represents thermocouples. Two thermocouples were located within each layer s strain gage lineup as shown in Figure Deep and shallow referenced LVDTs can be found in either

94 94 direction of the strain gages and thermocouples. LVDTs are represented by the color blue. Finally, located at the end of the instrumentation line are the strain gage rosette holes with the round hole preceding the square hole. The strain gage rosette holes are indicated with turquois. Surface layer strain gages are not presented in Figure 3.21 but were aligned directly before the first set of LVDTs. As shown in Figure 3.21, all of the sensors reside in the wheel path of the lane.

95 Figure 3.21 Instrumentation Layout 95

96 Controlled Vehicle Load Testing Controlled vehicle load (CVL) testing was performed as part of the U.S. Route 23 project to evaluate the various pavement structures under real traffic conditions. Many variables were controlled while performing CVL tests. For example, data was collected from the sensors while being subjected to a known load, axle type, tire pressure, and speed. Although lateral offset of the tire load could not be completely controlled, it was monitored through the use of sand prints as shown in Figure CVL testing was performed using two different drivers driving ODOT trucks typically used for snow removal and salt spreading. Figure 3.23 is a photo of the ODOT truck utilized. For every variation of load, axle type, tire pressure, and speed, at least five tests were performed and the three tests with the least amount of lateral offset measured were used for analysis. Figure 3.22 Measurement of Lateral Offset

97 97 Figure 3.23 ODOT Truck used for CVL Testing The first round of CVL testing was conducted during the end of November and throughout December of For this round of testing, two types of axle configurations were used with the maximum load the ODOT trucks were capable of carrying. The first truck utilized a dual tire, tandem axle while the second truck consisted of a single axle with a single wide-based tire. The two axle configurations are depicted in Figure The maximum loading resulted in the tandem and single axle types carrying approximately a 37 kip and 29 kip axle load, respectively. The exact loadings and dimensions used can be found in Appendix E. This round of testing consisted of tire pressures of 80, 110, and 125 psi and truck speeds of 5, 35, and 55 mph. Figure 3.24 Axle configurations

98 98 CHAPTER 4: CVL TESTING PAVEMENT RESPONSE Truck testing was conducted on the perpetual pavement test sections during the months of November and December Three separate days were used for testing of the 11, 13, and 15 inch sections but testing for each section did not extend to multiple days. Henceforth, sections 39BN803, 39BS803, and 39D168 will be identified as the 11, 13, and 15 inch sections respectively. Each section was tested throughout the course of a single day, typically between 10:00 AM and 3:00 PM. Although pavement temperatures remained fairly consistent during testing some variations were seen throughout the course of testing for each section and between testing days. The following provides information on air and pavement temperatures measured throughout the course of testing and should be taken into consideration when considering data analyzed in this chapter. The 11 inch section was tested on December 18, Figure 4.1 displays the air and pavement temperatures recorded during testing. Temperatures were measured every 15 minutes. Testing was conducted throughout the day at a tire pressure of 125 psi, followed by 110 psi, and finally 80 psi. For each tire pressure, testing began at 5 mph, then 30 mph, and finally 55 mph. Tandem and single axle trucks made test runs consecutively. Testing temperatures can be found in tabular form in Appendix F.

99 99 Figure Inch Section Temperature Profile The 13 inch section was tested on December 19, Figure 4.2 illustrates the air and pavement temperatures measured during testing. Temperatures were recorded every 15 minutes. Testing was conducted throughout the day at a tire pressure of 125 psi, followed by 110 psi, and finally 80 psi. For each tire pressure, testing began at 5 mph, then 30 mph, and finally 55 mph. Tandem and single axle trucks made test runs consecutively.

100 100 Figure Inch Section Temperature Profile The 15 inch section was tested on November 29, Figure 4.3 shows the air and pavement temperatures measured throughout the progression of testing. Temperatures were, once again, measured every 15 minutes. Testing was conducted throughout the day at a tire pressure of 125 psi, followed by 110 psi, and finally 80 psi. Unlike the 11 and 13 inch sections, for each tire pressure, testing began at 55 mph, then 30 mph, and finally 5 mph. Tandem and single axle trucks made test runs sequentially.

101 101 Figure Inch Section Temperature Profile 4.1 Strain Response of the Pavement The pavement s strain response was dependent on many factors. One of these factors included the orientation of the strain gage. A strain gage aligned parallel with traffic, or in the longitudinal direction, tended to receive compression strains as the truck tire approached and moved away from the location of the strain gage. However, while the tire load was passing over the longitudinal strain gage it produced a tensile strain. The compression and tensile strains experienced by a longitudinally oriented strain gage are displayed in Figure 4.4 which represents the longitudinal strains measured in the FRL of the 11 inch section during a single axle test run.

102 102 Figure 4.4 Single Axle, Longitudinal Strain Response As seen in Figure 4.4, Compression strains were recorded as negative values whereas tensile strains were recorded as positive values. The first tensile strain peak received by the strain gage was due to the steer axle of the truck while the second tensile strain peak was due to the loading produced by the single axle at the rear of the truck. Figure 4.4 shows that both the steer and single axle tensile strain peaks are preceded and proceeded by smaller but notable compression strain peaks. Although it is important to note that the pavement responded with both tensile and compression strains in the longitudinal direction, the critical value, which will be analyzed throughout these results, was the maximum strain recorded. For this scenario, the maximum strain recorded was tensile and was approximately 42 με. As noted in Figure 4.4, the maximum strain

103 103 produced by the steer axle was approximately 22 µɛ; about 52% of the maximum tensile strain produced by the rear single axle, 42 µɛ. The tire load produced by the steer axle, which utilized a regular tire, was measured to be 6.10 kips; about 43% of the tire load produced by the rear single axle wide-base tire configuration which was measured at kips. Conversely, strain gages oriented perpendicular to the flow of traffic, or in the transverse direction, tended to receive only compressive or tensile strains depending on the depth of the gage. The transverse strains measured in the base layer of the 11 inch section during a single axle test run are shown in Figure 4.5. As seen in Figure 4.5, as both the steer and single rear axle passed over the strain gage, only a tensile strain peak was recorded by the strain gage.

104 104 Figure 4.5 Single Axle, Transverse Strain Response The tandem axle truck was another factor influencing the pavement s strain response. Figure 4.6 displays the longitudinal strains recorded in the FRL of the 11 inch section during a tandem axle test run. It is important to note that this test run was performed at five mph. As displayed in Figure 4.6, test runs involving the tandem axle truck involve three tensile strain peaks, one for the steer axle and one for each of the axles that comprised the tandem axle. Figure 4.6 also shows that each tensile strain peak, like the tensile peaks produced by the single axle, were led and followed by a compression strain peaks although the compression strain did not return to zero between the tandem axles of the tandem axle configuration or between the steer axle and first axle of the tandem axle configuration. It is interesting to note that the highest value of

105 105 compression strain measured was between the tandem axle configuration. Also, the maximum strains produced by the each of the axle in the tandem axle configuration were nearly identical. This is because the tire loads for each of the axles producing the tandem axle were almost identical. As noted in Figure 4.6, the maximum tensile strain produced by the steer axle was approximately 19 µɛ; about 83% of the maximum tensile strain produced by the tandem axle, 23 µɛ. The tire load produced by the steer axle was measured to be 6.60 kips; about 75% of the dual-tire loads produced by the tandem axle which was measured at 8.8 kips. It is important to remember that the tandem axle truck had dual-tires spreading the load in the rear of the truck between eight tires while the single axle truck used wide based tires spreading the load in the rear of the truck between only two tires. Figure 4.6 Tandem Axle, Longitudinal Strain in the FRL Response at Five MPH

106 106 The speed at which the tandem axle truck was traveling during testing had some effects on the pavement s strain response. Figure 4.7 shows longitudinal strains measured in the FRL of the 11 inch section during a tandem axle test run conducted at 55 mph. As displayed in Figure 4.7, the maximum tensile strain produced by the first axle of the tandem axle configuration was notably higher than the strain produced by the second axle. This difference was also noted for tandem axle testing conducted at 30 mph but was not as substantial. Surprisingly, even at high truck speeds, a compression strain peak was observed between the axles of the tandem axle configuration. Both of these observations were prevalent for all tandem axle testing at higher speeds. Figure 4.7 Tandem Axle, Longitudinal Strain in the FRL Response at 55 MPH

107 107 Maximum strains measured in the base, intermediate, and surface layers were in some instances produced by the steer axle of the tandem axle truck. On the rare occasion, this was also true for maximum longitudinal strains produced in the FRL of the 11 inch section. Longitudinal strains recorded in the base layer of the 13 inch section for a tandem axle test run are displayed in Figure 4.8. Figure 4.8 shows that, for this test run, the maximum strain produced occurred while the steer axle was crossing the location of the gage. This phenomenon only transpired during tandem axle truck testing. Maximum strains during single axle truck testing were always produced by the single axle. Figure 4.8 Tandem Axle, Strain Response with Steer Axle Producing Maximum Strain

108 108 To further analyze when the steer axle and when the tandem axle of the tandem axle truck were producing the maximum strains, Tables 4.1, 4.2, and 4.3 were created. The tables represent the percentage of maximum strains that were produced by the tandem axle for orientations and layers applicable to each section. Additionally, the tables were created using one tandem axle truck test run for each speed and tire pressure combination which had a lateral tire offset of two inches or less. This was done in order to minimize the effects of lateral tire offset. As the strain gage location became closer to the surface the maximum longitudinal strain was produced by the tandem axle less frequently. Additionally, the maximum longitudinal strain seemed to be produced by the tandem axle less frequently as tire pressure increased. For strain gages oriented in the transverse direction, the maximum strain was typically produced by the tandem axle except for gages located in the base layer. Table 4.1 Percentage of Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for the 11 Inch Section Tire Pressure (psi) FRL Longitudinal Strain Base Layer Longitudinal Strain Base Layer Transverse Strain 80 94% 22% 11% % 11% 0% % 0% 0%

109 Table 4.2 Percentage of Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for the 13 Inch Section 109 Tire Pressure (psi) FRL Longitudinal Strain FRL Transverse Strain Base Layer Longitudinal Strain Base Layer Transverse Strain Intermediate Layer Longitudinal Strain Intermediate Layer Transverse Strain % 100% 78% 22% 33% 100% % 100% 67% 33% 33% 67% % 100% 22% 0% 0% 100% Table 4.3 Percentage of Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for the 15 Inch Section Tire Pressure (psi) FRL Longitudinal Strain FRL Transverse Strain Base Layer Longitudinal Strain Base Layer Transverse Strain Intermediate Layer Longitudinal Strain Intermediate Layer Transverse Strain Surface Layer Longitudinal Strain Surface Layer Transverse Strain % 100% 100% 56% 83% 100% 67% 100% % 100% 100% 67% 50% 100% 0% 67% % 100% 100% 89% 67% 100% 33% 100% Finally, as the location of the strain gage moves from below (FRL and base layer) to above (intermediate and surface layers) the neutral axis of the pavement the maximum strain switched from being tensile to compressive. Longitudinally oriented strain gages above the neutral axis receive slight tensile strains directly before and after being loaded while receiving a larger compressive strain peak while being loaded. Transversely oriented strain gages above the neutral axis received only compressive strain peaks. Figure 4.9 displays a chart of strains recorded by a longitudinal strain gage in the surface layer of the 15 inch section during a single axle truck test run.

110 110 Figure 4.9 Strain Response Above the Neutral Axis of the Pavement This section will analyze trends found when comparing maximum strains produced at different tire pressures and speeds in various pavement structures. These general strain response trends should be acknowledged while reviewing the results. First, for strain gages located below the neutral axis of the asphalt concrete, the longitudinal strain in the pavement was initially compressed before becoming tensile and finally compressed again during dynamic loading. Longitudinal strain gages located above the neutral axis received strains similarly but in the opposite directions. Transverse strains did not experience strains in opposing directions. As speed increased from five mph, the first axle of the tandem axle configuration began to produce the maximum tensile strain.

111 111 In the FRL of the 11 inch section and the base, intermediate, and surface layers of all of the sections, the maximum strain was, in certain scenarios, produced by the steer axle of the tandem axle truck Longitudinal Strain in the FRL The strain in the FRL was one of the most important pavement responses in respect to testing of the perpetual pavement test sections. Using the concept of the FEL, as discussed in Chapter 2, the strain measured in the FRL was directly correlated to a fatigue life deemed acceptable for perpetual pavements. A conservative FEL of 70 µɛ was used for the following analyses of the perpetual nature of the test sections. The 11 inch section received the highest longitudinal strains in the FRL in comparison with the 13, and 15 inch sections. Table 4.4 shows the average maximum longitudinal strain and maximum longitudinal strain measured in the FRL for testing conducted on the 11 inch section. It compares strains found at various speeds and tire pressures for both the single axle wide-based tire and tandem axle dual tire trucks. As seen in Table 4.4 the maximum longitudinal strain recorded in the FRL for the 11 inch section was µɛ. This value was captured during single axle truck testing at a truck speed of five mph and a tire pressure of 110 psi. Although this value was lower than the FEL, 70 µɛ, it was of concern due to the low pavement temperatures observed during testing. These strain measurements are expected to increase for warmer pavement temperatures. Table 4.4 presents average longitudinal strains in the FRL ranging between 32 and 48 µɛ for tests involving the single axle truck and 19 and 25 µɛ for tandem axle truck testing. Table 4.4 shows that the highest longitudinal strain in the FRL for both

112 112 axle configurations at every speed was measured during testing at a tire pressure of 110 psi except for testing at 30 mph with the tandem axle truck where the average and maximum longitudinal strain was observed to be maximized at 80 psi. The highest average and maximum longitudinal strain in the FRL for the 11 inch section were consistently seen during five mph testing for each tire pressure. Table 4.4 Maximum Longitudinal Strain in the FRL for the 11 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) The 13 inch section received significantly lower longitudinal strains in the FRL than the 11 inch section. Table 4.5 presents a summary of longitudinal strains measured in the FRL during testing on the 13 inch section. The maximum FRL, longitudinal strain obtained in the 13 inch section was µɛ which was considerably less than the 11 inch section and the FEL, 70 µɛ. The maximum longitudinal strain in the FRL for the 13 inch section was discovered when testing at five mph with a tire pressure of 125 psi and the tandem axle truck. Table 4.5 shows that average longitudinal strain in the FRL for the 13 inch section ranged between 22 and 32 µɛ for the single axle truck and 11 and 17

113 113 µɛ for the tandem axle truck. Similar to the 11 inch section, Table 4.5 also shows that strains produced by the single axle truck were consistently higher than strains produced by the tandem axle truck. Longitudinal strain in the FRL were also, once again, highest when testing at five mph for each tire pressure, although, they were maximized between all three tire pressures for each speed. Table 4.5 Maximum Longitudinal Strain in the FRL for the 13 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Surprisingly, the 15 inch sections received slightly higher longitudinal strain in the FRL compared to the 13 inch section, although, they were still substantially lower than measurements in the 11 inch section. One possible explanation for higher resulting strains in the 15 inch section compared to the 13 inch section could be the difference in subgrade resilient modulus. The 13 inch section was constructed on chemically stabilized subgrade which had a resilient modulus of 40 ksi while the 15 inch section was constructed on a subgrade which had a resilient modulus of 20 ksi. Table 4.6 displays the maximum and average maximum longitudinal strains recorded in the FRL for the 15 inch

114 114 section. The table shows that the maximum strain discovered was µɛ which was considerably lower than the FEL, 70 µɛ. The maximum strain was measured during five mph testing using the single axle truck at a tire pressure of 80 psi. Average longitudinal strain in the FRL ranged between 24 and 43 µɛ for testing utilizing the single axle truck and 18 and 23 µɛ with the tandem axle truck. Once more, the highest strains were obtained during five mph for each tire pressure, although, a variety of tire pressures acquired the highest strains for each speed as shown in Table 4.6. Table 4.6 Maximum Longitudinal Strain in the FRL for 15 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Single and Tandem Axle Comparison Although the tandem axle dual-tire truck was carrying an axle load of 37 kips compared to the single axle wide-based tire truck which was carrying an axle load of 29 kips, the tandem axle load was split between more axles and tires resulting in it producing much lower longitudinal strains in the FRL. This can be perceived through Tables 4.7, 4.8, and 4.9 which are comparisons of longitudinal strain in the FRL produced by the

115 115 tandem and single axle trucks for each section. One test run with less than two inches of lateral tire offset for each speed and tire pressure was selected in order to alleviate the effects of wheel wander. Additionally, for scenarios in the 11 inch section when maximum strains were produced by the steer axle of the tandem axle truck, the strain produced by the tandem axle was used for this comparison. Tables 4.7, 4.8, and 4.9 show that the tandem axle truck produced longitudinal strains in the FRL which were between 46 to 58 percent of the strains produced by the single axle truck for the 11 inch section, 49 to 54 percent for the 13 inch section, and 55 and 74 percent for the 15 inch section. Consistent differences were seen at five mph for all three sections. There seemed to be no correlation between the tandem and single axle truck longitudinal strain in the FRL difference with respect to tire pressure, speed, or the section thickness. Table 4.7 Average Maximum Longitudinal Strain in the FRL Comparison Between Single Axle and Tandem Axle for the 11 Inch Section (µɛ) Tire Pressure Single Axle Wide-Base Tire Tandem Axle Dual Tire Percent Tandem Axle of Single Axle (psi) 5 mph % % % 30 mph % % % 55 mph % % %

116 116 Table 4.8 Average Maximum Longitudinal Strain in the FRL Comparison Between Single Axle and Tandem Axle for the 13 Inch Section (µɛ) Tire Pressure Single Axle Wide-Base Tire Tandem Axle Dual Tire Percent Tandem Axle of Single Axle (psi) 5 mph % % % 30 mph % % % 55 mph % % % Table 4.9 Average Maximum Longitudinal Strain in the FRL Comparison Between Single Axle and Tandem Axle for the 15 Inch Section (µɛ) Tire Pressure Single Axle Wide-Base Tire Tandem Axle Dual Tire Percent Tandem Axle of Single Axle (psi) 5 mph % % % 30 mph % % % 55 mph % % %

117 Influence of Speed on Longitudinal Strain in the FRL The longitudinal strain in the FRL typically decreased as speed increased. This can be seen in Tables 4.4, 4.5, and 4.6 along with Figure Figure 4.10 is a chart of strain versus speed for both axle configurations and all three tire pressures in the 11 inch, 13 inch, and 15 inch sections. Once again, one run with little lateral tire offset was used to create the results as well as strains produced only by the tandem axle of the tandem axle truck. The figure clearly shows a decrease in longitudinal strain in the FRL as speed increased for all tire pressures, in each section, and for both axle configurations. The influence of speed on strain was most prevalent in the 15 and 11 inch sections while its influence in the 13 inch section was slightly less. Longitudinal strain in the FRL decreased more with speed for testing conducted with the single axle truck. Figure 4.10, when analyzing the 11 and 13 inch sections for both axle configurations and the 15 inch section with the tandem axle configuration, shows that at lower tire pressures the longitudinal strain in the FRL seemed to stabilize at higher speeds, thus testing conducted between 30 and 55 mph. For lower tire pressures, specifically the 80 psi tire pressure, the strain continued to decrease between 30 and 55 mph. This observation was not made for single axle truck testing on the 15 inch where all tire pressures displayed a decrease in strain between 30 and 55 mph.

118 Figure 4.10 Maximum Longitudinal Strain in FRL versus Speed 118

119 Influence of Tire Pressure on Longitudinal Strain in the FRL Tire pressure seemed to have minimal effects on longitudinal strain located in the FRL. The effect of tire pressure is displayed in Figure In order to reduce the effects of lateral tire offset, one run for each tire pressure and speed with less than two inches of lateral tire offset was used to generate all sets of results. Additionally, only tandem axle generated strains were used in order to create an accurate comparison. Figure 4.11 plots strain against tire pressure for all three sections and all testing speeds. The figure showed that tire pressure had little influence on the longitudinal strain in the FRL, especially at lower testing speeds, with virtually no influence occurring at a speed of five mph. Some inconsistencies were discovered for the single axle truck traveling at speeds of 30 and 55 mph but the differences were still fairly minimal. Tire pressure appeared to have the strongest influence on the 11 inch section and its influence was reduced as pavement thickness increased.

120 Figure 4.11 Maximum Longitudinal Strain in FRL versus Tire Pressure 120

121 Longitudinal Strain in the Base Layer Strains in the base layer were monitored during CVL testing for two reasons. One reason was to analyze the propagation of strain through the pavement. As discussed in Chapter 2, Romanello (2007) provided a horizontal strain threshold of 50 με for the base layer in order to provide an indication of the pavement structure s strain distribution. The second reason, although critical strains and therefore the critical location of fatigue cracking was in the FRL, was to monitor the base layer s potential for cracking. Longitudinal strains found in the base layer, like longitudinal strain in the FRL, were greatest in the 11 inch section. Table 4.10 shows the maximum and average maximum longitudinal strain measured in the base layer for the 11 inch section. The maximum longitudinal strain discovered in the base layer was less than the 50 με threshold limit. Table 4.10 shows that the maximum longitudinal strain recorded in the base layer for the 11 inch section was µɛ during single axle truck testing at five mph with a tire pressure of 125 psi. Table 4.10 displays that the average base layer strain for the 11 inch section ranged between 9 and 19 µɛ for the single axle truck and 6 and 11 µɛ for the tandem axle truck. Similar to the FRL, longitudinal strains were greater in the base layer when testing with the single axle truck, although, correlations between the tandem axle truck and single axle truck were avoided since maximum strains produced in the base layer by the tandem axle truck were occasionally the result of the steer axle. All maximum and average maximum longitudinal strains in the base layer of the 11 inch section were discovered while testing at five mph for each tire pressure. Additionally, strains provided in Table 4.10 typically increased with tire pressure.

122 122 Table 4.10 Longitudinal Strain in the Base Layer for the 11 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Similar to longitudinal strain measurements in the FRL, the 13 inch section received the lowest longitudinal strains in the base layer. Table 4.11 provides a summary of longitudinal strain in the base layer for the 13 inch section. The maximum strain discovered was µɛ as shown in Table This maximum strain was discovered at five mph, 110 psi tire pressure, and testing with the single axle truck and was less than the 50 με threshold. Furthermore, average longitudinal strains in the base layer ranged between 7 and 14 µɛ for the single axle truck and 4 and 7 µɛ for the tandem axle truck. Average maximum longitudinal strains in the base layer of the 13 inch sections were greatest when testing at five mph for each tire pressure. Additionally, average maximum longitudinal strains were maximized at 125 psi for each speed except for single axle truck testing at 55 mph where it was maximized at 110 psi.

123 Table 4.11 Maximum Longitudinal Strain in the Base Layer for the 13 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) The 15 inch section received longitudinal strain results in the base layer that were slightly higher than the 13 inch section. Table 4.12 offers maximum and average maximum longitudinal strain in the base layer for the 15 inch section. The largest maximum longitudinal strain in the base layer for the 15 inch section was less than 50 με. It was discovered to be µɛ which was obtained during single axle truck testing at five mph with a tire pressure of 125 psi. Table 4.12 shows that average strain measurements ranged between 9 and 15 µɛ for the single axle truck and 5 and 8 µɛ for the tandem axle truck. Slightly different from the 11 and 13 inch sections, Table 4.12 illustrates an increase in average maximum strain between tire pressures of 80 and 110 psi but then a slight decrease between 110 and 125 psi. This trend occurred for each testing speed. When comparing each tire pressure, it was discovered that the maximum average longitudinal strain was found at a five mph testing speed.

124 Table 4.12 Maximum Longitudinal Strain in the Base Layer for the 15 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Influence of Speed on Longitudinal Strain in the Base Layer Like the longitudinal strain in the FRL, the longitudinal strain in the base layer decreased as speed increased. This is presented in both Tables 4.10, 4.11, and 4.12 and Figure Figure 4.12, for the 11, 13, and 15 inch sections, relates longitudinal strain in the base layer to speed without the effects of lateral tire offset and is categorized by tire pressure and axle configuration. Once again, only tandem axle produced strains were used for the tandem axle truck data. Another observation similar to that of the longitudinal strain in the FRL was that the longitudinal strain in the base layer tended to stabilize at higher speeds when testing at higher tire pressures but only occurred in the 11 and 13 inch sections as shown in Figure Although, as the section thickness was increased the correlation became more linear for all three tire pressures with a purely linear trend resulting in the 15 inch section. This was especially true for testing with the single axle truck while testing with the tandem axle truck remained more consistent.

125 Figure 4.12 Maximum Longitudinal Strain in Base Layer versus Speed 125

126 Influence of Tire Pressure on Longitudinal Strain in the Base Layer Although tire pressure had minimal impacts on tandem axle truck testing, a clear increase in strain was witnessed as tire pressures were increased for the single axle truck. Figure 4.13 was created to relate longitudinal strain in the base layer and tire pressure for each test section and testing speed. Similar to earlier comparisons, the lateral tire offset factor was removed by analyzing runs with little wheel wander and tandem axle produced strains were used for tandem axle truck test runs. The chart showed a fairly steady increase in strain with tire pressure for the single axle truck until the tire pressure reached its maximum of 125 psi. At this point the data appeared to become erratic. This may have been due to inconsistent dynamic loadings during testing involving tires inflated to their maximum. Additionally, truck tire were inflated to 125 psi at cold conditions in the garage and tire pressures as high as 140 psi were seen after the tires had been heated during test runs. When the truck tires were inflated to tire pressure of 125 psi or more, they were so highly inflated that a bouncing effect may have been occurring as the trucks drove over the instrumentation. During this bouncing effect some strain gages may have been receiving additional dynamic loading as the truck tire landed directly on top of their location while other gages may have been subjected to the entire tire load due to the tire bouncing over the strain gage location. Figure 4.13 shows that the influence of tire pressure was reduced as pavement thickness was increased.

127 Figure 4.13 Maximum Longitudinal Strain in the Base Layer versus Tire Pressure 127

128 Transverse Strain in the Base Layer Transverse strain gages were implemented in the base layer of sections involved in the U.S. Route 23 project for similar purposes to those of the longitudinal strain gages installed in the base layer. They monitor cracking potential of the base layer and illustrate the propagation of stress and strain through the pavement. A transverse orientation was chosen in addition to the longitudinal orientation because the critical strain was not always in the longitudinal direction and, in some cases, resulted in the transverse direction. The horizontal strain threshold established by Romanello (2007) can be implemented for transverse strains as well. Transverse strain results in the base layer were greatest in the 11 inch section. A summary of the maximum and average maximum transverse strains found in the base layer is displayed in Table The highest transverse strain measured in the base layer of the 11 inch section was µɛ which was recorded during single axle truck testing with a tire pressure of 125 psi at five mph and was less than 50 µɛ. Transverse strains in the base layer for the 11 inch section varied between 6 and 13 µɛ for testing conducted with the single axle truck and 3 and 10 µɛ for testing conducted with the tandem axle truck. For each tire pressure, the maximum and average maximum transverse strain was greatest at five mph. With respect to tire pressure, the maximum and average maximum transverse strain always occurred at 110 or 125 psi tire pressure for each testing speed.

129 Table 4.13 Maximum Transverse Strain in the Base Layer for the 11 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Although longitudinal strains in the 13 inch section were found to be lower than the 11 and 15 inch sections, the transverse strains in the base layer were typically higher than the other two sections especially when testing with the single axle truck. A summary of transverse strain discovered in the base layer of the 13 inch section is presented in Table The maximum transverse strain occurred during single axle truck testing at five mph with a tire pressure of 125 psi and was µɛ. It was less than 50 µɛ. Single axle, average transverse strain ranged between 8 and 15 µɛ while transverse axle, average transverse strains ranged between 3 and 12 µɛ. For the 13 inch section, as displayed in Table 4.14, average transverse strains were typically maximized at a tire pressure of 125 psi with a one discrepancy where the average transverse strain while testing with the single axle truck at 55 mph was maximized at 110 psi tire pressure. Additionally, average maximum transverse strains in the base layer, for each tire

130 130 pressure, occurred at five mph except for the tandem axle truck implementing a tire pressure of 125 psi. Table 4.14 Maximum Transverse Strain in the Base Layer for the 13 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Transverse strains in the base layer of the 15 inch section were slightly reduced compared to those of the 13 inch section. Table 4.15 provides the maximum and average maximum transverse strains in the base layer of the 15 inch section. As seen in Table 4.15, the maximum strain obtained was µɛ during single axle truck testing at five mph with its tires inflated to 110 psi and was less than 50 µɛ. Average transverse strains in the base layer of the 15 inch section fluctuated between 8 and 14 µɛ for single axle truck testing and 4 and 8 µɛ for tandem axle truck testing. Table 4.15 shows that strains were largest at a tire pressure of 125 psi except for testing conducted with the single axle truck at 5 and 30 mph. Also, transverse strains in the base layer of the 15 inch section were greatest at five mph for each tire pressure.

131 Table 4.15 Maximum Transverse Strain in the Base Layer for the 15 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Influence of Speed on Transverse Strain in the Base Layer Similar to other strain responses previously discussed, transverse strain in the base layer decreased with increasing speeds. Figure 4.14 relates transverse strain found in the base layer to speed. Test runs used for the figure incorporated little or no lateral tire offset and only tandem axle produced strains were used. For the 11 inch section the transverse strain in the base layer resulting from the single axle truck seemed to decrease with speed but became stable once a speed of 30 mph was achieved. This was especially true for the higher tire pressures, 110 and 125 psi. For the 13 and 15 inch sections, as shown in Figures 4.14, transverse strains in the base layer typically linearly decreased with increasing testing speeds for the single axle truck and also for the tandem axle truck.

132 Figure 4.14 Maximum Transverse Strain in the Base Layer versus Speed 132

133 Influence of Tire Pressure on Transverse Strain in the Base Layer Similar to the longitudinal strains in the base, transverse strains in the base layer tended to increase with increasing tire pressures. Figure 4.15 relates transverse strain in the base layer to tire pressure without the influence of wheel wander or the steer axle of the tandem axle truck. Figure 4.15, while observing the 11 and 15 inch sections, showed that as tire pressure increased, transverse strains in the base layer increased but began to stabilize as the tire pressure reached upper limits. This was especially true for testing using the single axle truck. There tended to be minimal changes in strain measurements between tire pressure of 110 and 125 psi. Unlike other sections, the 13 inch section transverse strains in the base layer were fairly constant between 80 and 110 psi tire pressure but increased significantly between 110 and 125 psi tire pressure when conducting single axle truck testing. Overall, tire pressure had a minimal influence on transverse strain in the base layer especially during testing involving the tandem axle truck.

134 Figure 4.15 Maximum Transverse Strain in in the Base Layer versus Tire Pressure 134

135 Comparisons Between Strain Responses Transverse strains in the base layer were typically discovered to be less than longitudinal strains in the base layer for the 11 and 15 inch sections, although, for the 13 inch section transverse strains were typically found to be greater than longitudinal strains. Tables 4.16, 4.18, and 4.20 compare longitudinal and transverse strains in the base layer of the 11, 13, and 15 inch sections respectively. Once more, the tables do not incorporate influences of lateral wheel offset. Furthermore, due to the mixture of maximum strains produced by either the tandem axle or steer axle of the tandem axle truck, strains produced by the tandem axle were used for this comparison. For the 11 inch section, as portrayed in Table 4.16, when testing both the single axle truck and tandem axle truck, longitudinal strains were greater in magnitude than transverse strains independent of tire pressure, speed, or axle configuration. A switch of critical strains from longitudinal to transverse was prevalent in the 13 inch section when testing with the single axle truck as shown in table This was an important discovery because the critical strains for these tests in the base layer were the transverse strains. In the 15 inch section, as displayed in Table 4.20, all of the longitudinal strains were greater than the transverse strains. Tables 4.17, 4.19, and 4.21 further compare longitudinal and transverse strains in the base layer by providing the percentage transverse of longitudinal strain for each section. In the 11 inch section, for the single axle truck, percentages ranged between 60 and 76 percent, although, most percentages fell below 70. A much larger range was seen for the truck implementing the tandem axle in the 11 inch section, although, percentages

136 136 were smaller. Many of the percentages in Table 4.19, or the 13 inch section, were greater than 100 due to the fact that, for these cases, the transverse strain was larger than the longitudinal strain. This observation was only true for tests involving the single axle truck. Once again, in the 13 inch section, it was discovered that the tandem axle had percentages lower than the single axle truck. Percentages discovered for the 15 inch section ranged between 88 and 99 percent for the single axle truck which was a significant increase from the 11 inch section. It was interesting to note that percentages increased for the tandem axle truck as the section thickness increased. No clear correlations were noticed when relating the difference between transverse and longitudinal strain to speed or tire pressure for any of the sections. Table 4.16 Longitudinal Compared to Transverse Strains in the Base Layer for the 11 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) Transverse Longitudinal Transverse Longitudinal Transverse Longitudinal Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load)

137 Table 4.17 Percentage Transverse Strain in Base Layer of Longitudinal Strain in Base Layer for the 11 Inch Section Tire Pressure (psi) 5 Speed (mph) Single Axle Truck % 66.03% 69.60% % 60.90% 72.03% % 64.69% 69.05% Tandem Axle Truck % 24.96% 37.05% % 30.08% 30.98% % 38.52% 60.63% 137 Table 4.18 Longitudinal Compared to Transverse Strains in the Base Layer for the 13 Inch Section (µɛ) Tire Pressure (psi) Speed (mph) Transverse Longitudinal Transverse Longitudinal Transverse Longitudinal Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load)

138 Table 4.19 Percentage Transverse Strain in Base Layer of Longitudinal Strain in Base Layer for the 13 Inch Section Tire Pressure (psi) 5 Speed (mph) Single Axle Truck % % % % % 94.96% % % % Tandem Axle Truck % 79.85% 77.02% % 74.10% 76.03% % 84.75% 86.93% 138 Table 4.20 Longitudinal Compared to Transverse Strains in the Base Layer for the 15 Inch Section (µɛ) Tire Pressure (psi) Speed (mph) Transverse Longitudinal Transverse Longitudinal Transverse Longitudinal Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load)

139 Table 4.21 Percentage Transverse Strain in Base Layer of Longitudinal Strain in Base Layer for the 15 Inch Section Tire Pressure (psi) 5 Speed (mph) Single Axle Truck % 90.69% 88.74% % 90.83% 88.75% % 95.61% 98.93% Tandem Axle Truck % 72.14% 81.14% % 82.25% 79.55% % 85.27% 74.37% 139 In order to obtain a better understanding of which orientation, longitudinal or transverse, typically receives the maximum strain throughout the depth of the pavement, Figures 4.16 and 4.17 were generated. The figures represent the average ratio of transverse strain to longitudinal throughout the depth of the pavement for the 13 and 15 inch sections. For this comparison the lateral tire offset factor was reduced as previously described and the tandem axle truck was divided into steer and tandem axle results. The purple line in the figures represents where the transverse and longitudinal strains were, on average, the same. Points above this line represent areas in the pavement where the transverse strain was critical while points below this line represent areas in the pavement where the longitudinal strain was critical.

140 140 Figure 4.16 Ratio of Transverse to Longitudinal Strain for the 13 Inch Section Figure 4.17 Ratio of Transverse to Longitudinal Strain for the 15 Inch Section

141 141 As shown in the figures, the ratio hovers around equal for the single axle truck and steer axle of the tandem axle truck, although, the ratio seemed to increase as the depth in the pavement increased. In Figure 4.17, representing the 15 inch section, the ratio representing the single and steer axle at a depth of one inch breaks the trend and increases. At a depth of only one inch the strain gages are experiencing increased dynamic loads. Additionally, the load has not spread significantly at a depth of only one inch and, therefore, the strain gage may not be experiencing the entire load. The ratio representing a depth of one inch for the 13 inch section could not be generated due to a strain gage not functioning properly. For the tandem axle of the tandem axle truck, the ratio was exceptionally high at a depth of three inches. This could have been due to the large width of the dual tires on the tandem axle truck producing high transverse strains at depths closer to the surface. Although, a decrease in the tandem axle ratio was observed at a depth of one inch where dynamic loading may have an increased influence on the longitudinal gages and, due to the lack of load spreading at this depth, the gages may not be experiencing the entire load distribution of the dual tire used on the tandem axle. It is important to note that small offsets from the wheel path when installing instrumentation will influence the ratios presented in Figures 4.16 and Another factor to remember was that while testing with the tandem axle truck the tandem axle was being aligned with the wheel path causing the steer axle to be slightly offset. This was a possible factor for explaining the increased steer axle ratio compared to the single axle ratio. The offset steer axle tire was possibly not passing directly over

142 142 the thin contact area of the longitudinal gages while still passing over a portion of the wider contact area of the transverse gages, therefore, increasing the transverse to longitudinal ratio. The longitudinal strains measured in the base layer were significantly lower than the longitudinal strains measured in the FRL in each of the three sections. This was due to strain gages located in the base layer being closer to the neutral axis and therefore experiencing less bending. Tables 4.22, 4.24, and 4.26 compare longitudinal strain values obtained in the base and FRL for single axle truck testing in the 11, 13, and 15 inch sections respectively. Wheel wander effects were removed for this comparison. Tandem axle truck testing was also removed from this analysis since, in the base layer, it was no longer consistently producing the strains with its tandem axle load. Tables 4.23, 4.25, and 4.27 show the longitudinal strain measured in the base layer as a percentage of the longitudinal strain measured in the FRL. The difference in longitudinal strain between the two layers ranged between 28 and 39 percent, 35 and 43 percent, and 33 and 39 percent for the 11, 13, and 15 inch sections respectively. Table 4.23 shows that the difference increased as speed decreased and also as tire pressure increased for the 11 inch section. As speed increased in the 11 inch section the rate at which the difference between longitudinal strain in the base and FRL was increasing with tire pressure decreased, as shown in Table 4.8. This was not the case for the 13 and 15 inch section as displayed in Tables 4.25, and In the 11 inch section, the largest difference in longitudinal strain between the layers was µɛ which occurred during five mph testing with a tire pressure of 125 psi. This difference decreased in the 13 and

143 inch sections as strains produced in these sections were lower. The smallest difference was discovered to be µɛ in the 11 inch section and was recorded during testing at 55 mph with a tire pressure of 80 psi. Table 4.22 Maximum Longitudinal Strain Compared Between the FRL and Base Layer for Single Axle Truck Testing for the 11 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) FRL Base Layer FRL Base Layer FRL Base Layer Single Axle Wide-Base Tire (29 Kip Axle Load) Table 4.23 Percentage Longitudinal Strain in Base Layer of Longitudinal Strain in FRL for Single Axle Truck Testing for the 11 Inch Section Tire Pressure 5 Speed (mph) Single Axle Truck % 31.84% 28.32% % 35.41% 31.08% % 37.21% 35.34%

144 Table 4.24 Maximum Longitudinal Strain Compared Between the FRL and Base Layer for Single Axle Truck Testing for the 13 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) FRL Base Layer FRL Base Layer FRL Base Layer Single Axle Wide-Base Tire (29 Kip Axle Load) Table 4.25 Percentage Longitudinal Strain in Base Layer of Longitudinal Strain in FRL for Single Axle Truck Testing for the 13 Inch Section Tire Pressure 5 Speed (mph) Single Axle Truck % 35.32% 35.78% % 42.51% 36.54% % 41.46% 39.28% Table 4.26 Maximum Longitudinal Strain Compared Between the FRL and Base Layer for Single Axle Truck Testing for the 15 Inch Section (µɛ) Tire Speed (mph) Pressure (psi) FRL Base Layer FRL Base Layer FRL Base Layer Single Axle Wide-Base Tire (29 Kip Axle Load)

145 Table 4.27 Percentage Longitudinal Strain in Base Layer of Longitudinal Strain in FRL for Single Axle Truck Testing for the 15 Inch Section Tire Pressure 5 Speed (mph) Single Axle Truck % 33.75% 35.50% % 36.38% 36.77% % 36.29% 37.31% Subgrade Pressure Subgrade pressure, like strain, was plotted against time for each test run in order to discover the maximum subgrade pressure produced. These maximum values were then used to evaluate the performance of the pavement and to compare with each other in order to find trends relating to speed, tire pressure, and pavement thickness. Figure 4.18 displays a plot of subgrade pressure versus time for a single axle truck test run. As seen in Figure 4.18, a smaller subgrade pressure peak was produced as the steer axle crossed over the pressure cell and a maximum subgrade pressure peak was produced as a result of the single axle. A plot of subgrade pressure versus time for a tandem axle test run is presented in Figure Figure 4.19 shows three peaks, one for the steer axle and one for each of the axles involved in the tandem axle configuration. The plot shows that the second axle of the tandem axle configuration produced the maximum subgrade pressure. This was typically true for all tire pressures, speeds, and pavement structures. Unlike strains, maximum subgrade pressures were always produced by the tandem axle of the tandem axle truck.

146 146 Figure 4.18 Single Axle, Subgrade Pressure Response Figure 4.19 Tandem Axle, Subgrade Pressure Response

147 147 Subgrade pressure was monitored during testing of the U.S. Route 23 project in order to evaluate the pavement s resistance to structural rutting in the subgrade. A compressive strain endurance limit was provided by Willis and Timm (2009) as 200 µɛ. By correlating this strain limit with the section s subgrade resilient modulus, a pressure endurance limit was obtained. The 11 and 13 inch sections had a subgrade resilient modulus of 40 ksi and therefore a pressure endurance limit of eight psi while the 15 inch section had a subgrade resilient modulus of 20 ksi resulting in a pressure endurance limit of four psi. If pressures recorded in these sections remain below their pressure endurance limits then, in theory, structural rutting in the subgrade should be avoided for an infinite number of loading cycles. Subgrade pressure in the 11 inch section had many of the same trends as strain measurements. Table 4.28 summarizes subgrade pressures obtained for CVL testing conducted on the 11 inch section. The maximum subgrade pressure obtained was psi which occurred during single axle truck testing at 80 psi with a truck speed of five mph. The endurance limit calculated for the 11 and 13 inch sections was eight psi, therefore, during testing the subgrade pressure of the 11 inch section did not exceed the endurance limit. Table 4.28 shows that at five mph, subgrade pressure decreases with increasing tire pressure but as speed increases, the maximum subgrade pressure was found to be at higher tire pressures. For each tire pressure, the maximum average subgrade pressure was found during testing at five mph. The table also shows that subgrade pressures were higher for single axle truck testing than for tandem axle truck testing. Average maximum subgrade pressures for the 11 inch section ranged between

148 and psi for single axle truck testing and between and psi for tandem axle truck testing. Table 4.28 Maximum Subgrade Pressure for the 11 Inch Section (psi) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) The subgrade pressure in the 13 inch section was less than half of that of the 11 inch section. Table 4.29 provides maximum and average maximum subgrade pressure measurements for the 13 inch section. The maximum subgrade pressure recorded during testing was psi which was discovered during single axle truck testing at five mph using a tire pressure of 80 psi. The maximum subgrade pressure of psi was significantly less than the pressure endurance limit of eight psi. Table 4.29 shows that, like the 11 inch section, subgrade pressures were much lower for the tandem axle truck than for the single axle truck. Additionally, average maximum subgrade pressures ranged between and psi for single axle truck testing and between and psi for tandem axle truck testing.

149 149 Table 4.29 Maximum Subgrade Pressure for the 13 Inch Section (psi) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Subgrade pressures in the 15 inch section were greater than subgrade pressure in the 13 inch section but still significantly less than those found in the 11 inch section, although, its pressure endurance limit was significantly less as well. Table 4.30 presents a summary of subgrade pressures discovered in the 15 inch section. The maximum subgrade pressure found in the 15 inch sections was psi which was lower than the 15 inch section pressure endurance limit of four psi. This maximum subgrade pressure was discovered during single axle truck testing at five mph with a tire pressure of 80 psi. Unlike the 11 and 13 inch sections, the tandem axle configuration produced similar strains to those produced by the single axle configuration in the 15 inch section. Table 4.30 displays the range of the average maximum subgrade pressures to have been to psi and to psi for the single and tandem axle configurations, respectively.

150 150 Table 4.30 Maximum Subgrade Pressure for the 15 Inch Section (psi) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Influence of Speed on Subgrade Pressure Subgrade pressure typically decreased as truck speed increased. This is shown in Figure 4.20 which is a chart of subgrade pressure versus speed for the 11, 13, and 15 inch sections. Effects of wheel wander have been removed from the figure. The influence of speed was highly influential in the 11 inch section compared to the 13 and 15 inch sections. As shown in Figure 4.20, speed had less of an effect on subgrade pressure for the tandem axle truck. No correlations between tire pressure and the speed-subgrade pressure trend were observed.

151 Figure 4.20 Maximum Subgrade Pressure versus Speed 151

152 Influence of Tire Pressure of Subgrade Pressure Tire pressure tended to have negligible effects on subgrade pressure. Effects were more noticeable in the 11 inch section compared to the 13 and 15 inch sections, although, they were still minimal. This was especially obvious for subgrade pressure data obtained during tandem axle truck testing as shown in Figure Figure 4.21 relates subgrade pressure to tire pressure for each section without the influence of significant lateral tire offset. Similar to many of the other pavement responses already discussed in this chapter, subgrade pressure tended to become irregular when the truck tires were inflated to 125 psi.

153 Figure 4.21 Maximum Subgrade Pressure versus Tire Pressure 153

154 Deflection Pavement and subgrade deflections were once again analyzed in the same manner as pavement strains and subgrade pressures. Plots of deflection versus time were used to discover maximum deflections which were then used to evaluate pavement performance and form trends related to tire pressure and speed. Figure 4.22 is a plot of pavement deflection versus time for a single axle truck test run. Similar to subgrade pressure, the pavement deflected for both the steer and single axle but a significant increase was seen for deflections produced by the single axle. This was true for the maximum pavement and subgrade deflections at all tire pressures, speeds, and pavement thicknesses. Figure 4.22 Single Axle, Pavement Deflection Response

155 155 For the tandem axle truck, plots of pavement deflection versus time were also similar to those of subgrade pressure versus time. Three pavement deflection peaks were seen with the second axle of the tandem axle configuration producing the maximum pavement deflection. Figure 4.23 displays a plot of pavement deflection versus time for a tandem axle truck test run. When analyzing the subgrade deflection it was discovered that the second axle of the tandem axle configuration did not always produce the maximum subgrade deflection as shown in Figure Figure 4.24 is a plot of subgrade deflection versus time for a tandem axle test run. Maximum subgrade deflections were typically similar between each axle of the tandem axle configuration. Similar to subgrade pressure responses, maximum deflections recorded were always produced by the tandem axle of the tandem axle truck. Figure 4.23 Tandem Axle, Pavement Deflection Response

156 156 Figure 4.24 Tandem Axle, Subgrade Deflection Response Deflections were monitored in order to evaluate the pavement s load transfer and overall performance. Deflection measurements were also an indicator for subgrade performance and structural rutting. As discussed in Chapter 2, Romanello (2007) established a vertical deflection limit for subgrade deflection in order to monitor the pavement system s potential for structural rutting. The value established was 20 mils Pavement Deflection It was discovered that the pavement deflections observed during CVL testing had many of the same correlations as strain and subgrade pressure. Table 4.31 displays the maximum and average maximum pavement deflections obtained during testing on the 11 inch section. The maximum pavement deflection attained was 4.84 mils during single

157 157 axle truck testing at five mph with a tire pressure of 80 psi. Table 4.31 shows that, although values were fairly similar, the tandem axle truck produced pavement deflections that were smaller in magnitude than the pavement deflections produced by the single axle truck. Average maximum pavement deflections for single axle truck testing ranged between 3.42 and 4.39 mils and tandem axle truck testing ranged between 2.56 and 3.18 mils. The highest average maximum pavement deflection, for each tire pressure, always transpired at a testing speed of five mph. Table 4.31 Maximum Pavement Deflection for the 11 Inch Section (mils) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Pavement deflections in the 13 inch section were comparable to those found in the 11 inch section. Table 4.32 provides a summary of maximum pavement deflections recorded for the 13 inch section. The maximum pavement deflection in the 13 inch section was found during single axle truck testing at five mph with an 80 psi tire pressure and was 4.65 mils. Like the 11 inch section, table 4.32 shows that the tandem axle truck produced similar but slightly smaller pavement deflections than the single axle truck. A

158 158 range of 3.10 to 4.50 mils and 2.37 to 3.59 mils was seen for single and tandem axle average maximum pavement deflections respectively. For each tire pressure, the highest average maximum pavement deflection occurred at a testing speed of five mph. Additionally, for many of the testing speeds, the highest average maximum pavement deflection was produced during testing involving a tire pressure of 80 psi. Table 4.32 Maximum Pavement Deflection for the 13 Inch Section (mils) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Pavement deflection in the 15 inch section did not resemble that of the 11 and 13 inch sections. The pavement deflections found in the 15 inch section were noticeably larger. Table 4.33 presents maximum and average maximum pavement deflections found while testing the 15 inch section. The maximum pavement deflection in the 15 inch section was 7.54 which, like the other two sections, was discovered while testing with the single axle truck with 80 psi tire pressure and traveling at five mph. Additionally, the tandem axle truck generated pavement deflections comparable to those produced by the single axle truck but were marginally less. Single axle truck testing produced a range of

159 159 average maximum pavement deflections between 3.94 and 4.91 mils while the tandem axle truck testing produced a range between 3.24 and 4.90 mils. A testing speed of five mph consistently resulted in the greatest average maximum pavement deflection for each tire pressure. Table 4.33 Maximum Pavement Deflection for the 15 Inch Section (mils) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Influence of Speed on Pavement Deflection Pavement deflection decreased with increasing speeds. This is shown in Figure 4.25 which is a chart of pavement deflection versus speed for each section without the effects of lateral tire offset beyond two inches. Unlike the other pavement responses, the influence of speed on pavement deflection was consistent between the single and tandem axle trucks as seen in the figures.

160 Figure 4.25 Maximum Pavement Deflection versus Speed 160

161 Influence of Tire Pressure on Pavement Deflection Tire pressure tended to have insignificant effects on pavement deflection in all three pavement structures. Figure 4.26 compare pavement deflection to tire pressure using one run for each tire pressure, axle configuration, and speed that had a lateral tire offset of less than two inches. The influence of tire pressure on pavement deflection seemed to be inconsistent, although, many of the transitions to higher tire pressures showed slight decreases in pavement deflection.

162 Figure 4.26 Maximum Pavement Deflection versus Tire Pressure 162

163 Subgrade Deflection Table 4.34 summarizes the maximum and average maximum subgrade deflections recorded for both axle types, various tire pressures, and various speeds. The maximum subgrade deflection was discovered to be 3.65 mils during single axle truck testing at five mph utilizing a tire pressure of 80 psi. This value was significantly less than the 20 mil vertical subgrade deflection established by Romanello (2007). As seen in Table 4.34, subgrade deflection was typically the highest when the truck tires were inflated to 80 psi. Subgrade deflections were smaller when testing with the tandem axle truck than when testing with the single axle truck. Average subgrade deflections ranged between 2.20 and 2.92 mils for the single axle truck and between 1.82 and 2.28 mils for the tandem axle truck thus showing consistency between test runs. The greatest average maximum subgrade deflection was always observed at five mph testing. Table 4.34 Maximum Subgrade Deflection for the 11 Inch Section (mils) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load)

164 164 Contrary to other pavement responses, subgrade deflections in the 13 inch section were greater than subgrade deflections in the 11 inch section. Table 4.35 summarizes subgrade deflections discovered while testing the 13 inch section. The highest subgrade deflection recorded during testing was less than 20 mils. The largest magnitude of subgrade deflection in the 13 inch section was discovered while testing with the single axle truck with 80 psi tire pressures traveling at five mph and was 3.65 mils. Subgrade deflections obtained during tandem axle truck testing were lower than when single axle truck testing, although, the differences were found to be minimal. The single axle truck produced average subgrade deflections between 2.27 and 3.54 mils while the tandem axle truck produced average subgrade deflections between 1.97 and 3.12 mils. Average maximum subgrade deflections were greatest while testing at five mph. Table 4.35 Maximum Subgrade Deflection for the 13 Inch Section (mils) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Subgrade deflections observed in the 15 inch section, much like pavement deflections, were greater than those observed in the 11 and 13 inch sections. Table 4.36

165 165 presents the maximum and average maximum subgrade deflections obtained during testing on the 15 inch section. The maximum subgrade deflection observed was 6.13 mils which was less than 20 mils and occurred during single axle truck testing at five mph with a tire pressure of 80 psi. Tandem axle produced subgrade deflections were found to be slightly lower than single axle produced subgrade deflections. A range of 1.43 mils, between 3.08 and 4.51 mils, was seen for the single axle subgrade deflections compared to a range of 1.39 mils, between 2.77 and 4.16 mils, for the tandem axle subgrade deflections. Once more, the greatest average maximum subgrade deflections were discovered during five mph testing. Table 4.36 Maximum Subgrade Deflection for the 15 Inch Section (mils) Tire Speed (mph) Pressure (psi) Average Max Average Max Average Max Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load) Influence of Speed on Subgrade Deflection A decrease in subgrade deflection was discovered for increasing truck testing speeds. Figure 4.27 relates subgrade deflection to speed for the 11, 13, and 15 inch sections. Lateral tire offset factors greater than two inches were removed from the

166 166 figures. Although there were a few exceptions, subgrade deflection typically decreased with increasing speeds as seen in the figures. Furthermore, there seemed to be an increase of this effect as pavement thickness was increased.

167 Figure 4.27 Maximum Subgrade Deflection versus Speed 167

168 Influence of Tire Pressure on Subgrade Deflection Tire pressure had unpredictable effects on subgrade deflection. Figure 4.28 is a plot of subgrade deflection versus tire pressure for all three sections. No correlations between tire pressure and subgrade deflection were seen and influences seemed to be minimal. Testing at warmer temperatures will provide additional information on the extent of the influence of tire pressure on subgrade deflection.

169 Figure 4.28 Maximum Subgrade Deflection versus Tire Pressure 169

170 Pavement and Subgrade Deflection Comparison Comparisons between pavement and subgrade deflections were found to be fairly consistent between all three sections, tire pressures, speeds, and axle configurations. Tables 4.37, 4.39, and 4.41 compare average pavement and subgrade deflections excluding the influence of lateral tire offset greater than two inches for each section. Tables 4.38, 4.40, and 4.42 display what percentage the subgrade deflection was of the total pavement deflection. Differences seemed to be higher but more consistent when testing with the tandem axle truck. Differences in the 13 and 15 inch sections were typically greater than differences in the 11 inch section. Table 4.37 Comparison Between Pavement and Subgrade Deflection for the 11 Inch Section (mils) Tire Speed (mph) Pressure (psi) Pavement Subgrade Pavement Subgrade Pavement Subgrade Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load)

171 171 Table 4.38 Percentage Subgrade Deflection of Pavement Deflection for the 11 Inch Section Tire Pressure 5 Speed (mph) Single Axle Truck % 87.03% 71.70% % 60.56% 66.36% % 63.52% 70.34% Tandem Axle Truck % 66.67% 80.67% % 75.28% 81.51% % 64.57% 76.68% Table 4.39 Comparison Between Pavement and Subgrade Deflection for the 13 Inch Section (mils) Tire Speed (mph) Pressure (psi) Pavement Subgrade Pavement Subgrade Pavement Subgrade Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load)

172 172 Table 4.40 Percentage Subgrade Deflection of Pavement Deflection for the 13 Inch Section Tire Pressure 5 Speed (mph) Single Axle Truck % 80.00% 78.90% % 77.94% 78.61% % 73.02% 56.27% Tandem Axle Truck % 88.78% 88.48% % 84.25% 85.89% % 84.62% 85.02% Table 4.41 Comparison Between Pavement and Subgrade Deflection for the 15 Inch Section (mils) Tire Speed (mph) Pressure (psi) Pavement Subgrade Pavement Subgrade Pavement Subgrade Single Axle Wide-Base Tire (29 Kip Axle Load) Tandem Axle Dual Tire (37 Kip Axle Load)

173 173 Table 4.42 Percentage Subgrade Deflection of Pavement Deflection for the 15 Inch Section Tire Pressure 5 Speed (mph) Single Axle Truck % 77.62% 79.29% % 76.18% 77.70% % 77.23% 76.77% Tandem Axle Truck % 84.11% 86.88% % 86.20% 86.01% % 85.17% 85.67% 4.4 Comparison of Pavement Structures The CVL testing indicated that the 13 inch section performed better than the 11 and 15 inch sections with the 11 inch section performing significantly worse. Although in theory the 15 inch section should have performed better than the 13 inch section, the 13 inch section was constructed on top of a subgrade with a 40 ksi resilient modulus while the 15 inch section s subgrade resilient modulus was 20 ksi resulting in the 13 inch section performing better than the 15 inch section. Tables 4.43, 4.44, and 4.45 specify the maximum FRL longitudinal strain, maximum base strain, subgrade pressure, and subgrade deflection for each section. The tables also identify the axle configuration, tire pressure and speed that produced each maximum pavement response. All three sections did not acquire longitudinal strain in the FRL greater than the FEL of 70 µɛ nor did any of the sections have maximum base strains, subgrade pressures, or subgrade deflections exceeding their respected endurance limits. Tables 4.43, 4.44, and 4.45 show that all maximum values obtained were done so during testing involving the single axle truck at a

174 174 speed of five mph. Many of the maximum strain values were obtained during testing involving higher tire pressure except for the maximum longitudinal strain in the FRL of the 15 inch section which was acquired during testing with 80 psi tire pressure. All of the maximum subgrade pressures and pavement deflections were discovered during 80 psi tire pressure testing. In regards to subgrade deflection, the 15 inch section recorded the highest at 6.13 mils. Subgrade deflection was the only maximum response of which the 15 inch section was greater than the 11 inch section. Table 4.43 Maximum Pavement Response for the 11 Inch Section Pavement Max Value Endurance Limit Response FRL longitudinal Strain (µɛ) Base Strain (µɛ) Subgrade Pressure Axle Single Axle Wide-Based Tire Single Axle Wide-Based Tire Pressure 125 psi Speed Configuration Single Axle Wide-Based Tire 110 psi 5 mph 5 mph (psi) Tire 80 psi 5 mph Subgrade Deflection (mils) 3.65 Single Axle Wide-Based 20 Tire 80 psi 5 mph

175 175 Table 4.44 Maximum Pavement Response for the 13 Inch Section Pavement Response Max Value Endurance Limit FRL longitudinal Strain (µɛ) Axle Tire Configuration Pressure Single Axle Wide-Based Tire 125 psi 5 mph Speed Single Axle Base Strain Wide-Based (µɛ) Tire 110 psi 5 mph Subgrade Pressure (psi) Single Axle Wide-Based 8 Tire 80 psi 5 mph Subgrade Deflection (mils) Single Axle Wide-Based Tire 80 psi 5 mph Table 4.45 Maximum Pavement Response for the 15 Inch Section Pavement Response Max Value Endurance Limit FRL longitudinal Strain (µɛ) Axle Configuratio Tire Pressure Single Axle Wide-Based Tire 80 psi 5 mph Speed Single Axle Base Strain Wide-Based (µɛ) Tire 125 psi 5 mph Subgrade Pressure (psi) Single Axle Wide-Based 4 Tire 80 psi 5 mph Subgrade Deflection (mils) Single Axle Wide-Based Tire 80 psi 5 mph Finally, in order to further compare the three test sections, Figure 4.29 displays the maximum longitudinal strain obtained in the FRL for each test section when testing at a speed of five mph and a tire pressure of 110 psi. As shown in Figure 4.29, the 11 inch section recorded the greatest maximum strain in the FRL at µɛ while the 13 and 15

176 176 inch sections received significantly lower maximum strains in the FRL at µɛ and µɛ respectively. Typically as total pavement thickness increases the strain measurements in the FRL decrease but due to the variation in subgrade resilient moduli between the 13 and 15 inch sections, the 13 inch section recorded a slightly smaller FRL longitudinal strain than the 15 inch section. Figure 4.29 Maximum Longitudinal Strain Obtained in the FRL for Each Test Section at a Testing Speed of Five MPH and a Tire Pressure of 110 PSI

177 177 CHAPTER 5: PERROAD ANALYSIS In order to further evaluate the perpetual nature of the test sections constructed on U.S. Route 23 in Delaware, Ohio, an analysis of the test sections was performed using the mechanistic based software program, PerRoad. PerRoad was designed by NCAT for the Asphalt Pavement Alliance (APA). The software relies on layered elastic theory and Monte Carlo simulation. It integrates loading and seasonal data with flexible pavement s layer material properties and thicknesses (Romanello 2007). Additionally, the software incorporates endurance limits and strain-based transfer functions to predict damage accumulation and, therefore, the in-service life of the pavement structure being analyzed. 5.1 Loading Conditions In order to accurately analyze the U.S. Route 23 test sections, loading properties were entered into the PerRoad program through a vehicle classification distribution. The program requires the percent annual average daily truck traffic (%AADTT) for vehicle classifications 4 through 13. These percentages were discovered for the test sections by Weigh-in-motion (WIM) scales installed on the northbound lanes of U.S. Route 23. Although the traffic seen for the test sections on the northbound lanes of U.S. Route 23 will be significantly different than for the test sections on North Waldo Road, the traffic on the northbound lanes of U.S. Route 23 is considerably higher than that of North Waldo Road creating a conservative analysis for the test sections on North Waldo Road. Weigh-in-motion (WIM) is a technology used for detecting various traffic parameters of moving vehicles. They are designed to capture and record axle weights and are capable of measuring at high traffic speeds without requiring the vehicle to stop.

178 178 This makes the weighing process more efficient. Loads are not the only parameter WIM systems can collect. They can also obtain speed, classification, and volume information as well. As scale availability and accuracy has increased over the last few years, WIM sensors have become more widely used. A prominent WIM sensor application in pavement design, monitoring and research is the amount of axle loading, especially from trucks, the proposed roadway is going to have to endure. With a static scale, the axle is placed on the scale and the force it produces is constant while the vehicle remains motionless. WIM technology has the ability to measure the actual loads being applied to a roadway by a moving vehicle which more accurately represents what the pavement is subjected to (Bushman, 1998). The actual load applied by a vehicle involves more than solely the weight of the vehicle. Bushman (1998) states, as a vehicle travels, the dynamic load applied to the road varies significantly due to the vehicle bouncing, acceleration or deceleration, and shifting of the load either physically or just in its distribution through the suspension system. WIM systems measure the combination of all these loading influences. In Delaware, Ohio, on US Route 23, a Mettler-Toledo WIM system was installed on a section of test road at the time of construction in The scales are used to continuously monitor traffic in all four pavement lanes. Each vehicle crossing the WIM load plates in the pavement generates data including gross weight, classification, date and hour of crossing, and the weight and spacing of individual vehicle axles. WIM sensors on US Route 23 in Delaware, Ohio were calibrated on Thursday, May 10, Figures 5.1 and 5.2 are photographs of the actual sensors.

179 179 Figure 5.1 WIM Scales Figure 5.2 WIM Scales

180 Accuracy of WIM Scales In September of 2012 the test road section on US Route 23 in Delaware, Ohio was reconstructed with experimental, perpetual asphalt pavement. The design of the perpetual pavement was based on data collected from the WIM sensors. To verify the accuracy of the WIM system, manual counts by classification were recorded for a three hour period on the morning of Thursday, May 17, Due to the high volumes of traffic, counts were only made for the driving lane. Manual counts were compared to counts provided by the WIM sensors and analyzed. Table 5.1 shows total counts by classification for the entire three hour period of both manual and WIM along with the corresponding percent error. Tables attached in Appendix J at the end of the report separates Table 5.1 into five minute intervals. Table 5.1 Manual and WIM Counts by Classification Class Class 1 - Motorcycles Class 2 - Passenger Cars Class 3 - Trucks and Vans Class 4 - Buses Class 5-2 Axles, 6 Tire Single Units Class 6-3 Axles, Single Unit Class 7-4 or More Axles, Single Unit Class 8-3 to 4 Axles, Single Trailer Class 9-5 Axles, Single Trailer Class 10-6 or More Axles, Single Trailer Class 11-5 or Less Axles, Multi-Trailer Class 12-6 Axles, Multi-Trailers Total Manual Count WIM Count Percent Error % % % % % % % % % % % % %

181 181 Table 5.1 displays no abnormalities for the accuracy of the WIM system. A high percent error of % was discovered for motorcycles due to motorcycles only having one wheel per axle making it easy for them to miss the two WIM scales by either driving in between or around them. This is why the manual count was excessively higher than the WIM count. This error is irrelevant for the traffic study needed for the PerRoad analysis which only includes vehicles of class four or higher. A high percent error can also be seen in Table 5.1 for classes four and six because differentiating between the two was difficult since they both have three axles. It seemed like, for some measurements, the WIM scales were classifying a vehicle as class four when it was actually a class six. Class five had a significant percent error of 24.00% because two-axle single unit trucks were sometimes misrepresented as passenger class vehicles. Errors can not only be attributed to the accuracy of the WIM system since there may be error associated with the manual counts obtained Vehicle Classification Distribution In order to obtain a vehicle classification distribution for the test sections on U.S. Route 23, AADTTs for vehicle classifications 4 through 13 were generated by ODOT from data collected by the WIM scales for 2010, 2011, and These three years were then averaged, converted into percentages and entered into the PerRoad program. Table 5.2 provides the %AADTTs used and Figure 5.3 is a screen capture of the PerRoad Vehicle Type Distribution screen. Data used to calculate %AADTTs is provided in Appendix K.

182 182 Table 5.2 Vehicle Type Distribution Class %AADTT % % % % % % % % % % Figure 5.3 Vehicle Type Distribution

183 183 After the Vehicle Type Distribution data was entered into the PerRoad program, the program used standard values for the average number of axles per vehicle classification to generate a current axle load distribution, as shown in Figure 5.4. Figure 5.4 displays PerRoad s Loading Conditions screen. In addition to the current axle load distribution, ODOT WIM scale measurements of general traffic data on U.S. Route 23 were needed. Once again, these values were obtained from data collected in 2010, 2011, and Figure 5.4 Loading Conditions

184 Structural and Seasonal Information In order for the PerRoad software to correlate modulus values for a variation of temperatures seen during different seasons, the duration and mean air temperature for each season were obtained using average monthly temperatures for Delaware, Ohio from The Weather Channel (2013). Table 5.3 displays the seasonal information used. Appendix K provides monthly average temperatures for Delaware, Ohio. Table 5.3 Seasonal Information (The Weather Channel, 2013) Season Winter Fall Summer Spring Duration (weeks) Average Temperature ( F) Due to the fact that the PerRoad program only allows a pavement structure of five layers or less, for the U.S. Route 23 analysis, the surface and intermediate layers were combined. The layer thicknesses and specifications were provided in Chapter 3. Typical design values for the Modulus of Elasticity and Poison s Ratio at 70 F were provided by Sargand et al. (2008) and are shown in Table 5.4 along with the values chosen for the PerRoad analysis. Figure 5.5 displays the Structural and Seasonal Information screen.

185 185 Table 5.4 Elastic Modulus and Poison s Ratio Values Provided by Sargand et al. (2008) Input Values for PerRoad Layer Modulus of Modulus of Poison's Ratio Elasticity, E (ksi) Elasticity, E (ksi) Poison's Ratio Surface/Intermediate Base Fatigue Resistant DGAB Subgrade /40* 0.45 *Since the modulus of the subgrade was previously obtained, it was used. Figure 5.5 Structural and Seasonal Information 5.3 Performance Criteria In order to evaluate the performance of the perpetual pavement test sections, performance criteria were assigned at four different locations in the PerRoad program. By assigning performance criteria, such as horizontal strain, vertical stress, or vertical

186 186 deflection, and a corresponding threshold limit, the PerRoad program computed the percent of pavement responses that fell below the given threshold limit for a variety of loading and temperature conditions it developed. The program also returned a value for each condition and, therefore, a maximum value was obtained for each pavement response. A performance criterion was assigned to evaluate pavement deflection, horizontal strain in the base layer and FRL, and pressure on the subgrade surface. Table 5.5 provides the location, performance criteria, and threshold limit used. Table 5.5 Performance Criteria Location Performance Threshold Layer Position Criteria Limit Base Layer Bottom Horizontal Strain 50 µɛ FRL Bottom Horizontal Strain 70 µɛ DGAB Bottom Vertical Deflection 20 mils DGAB Bottom Vertical Pressure 8/4 psi The final PerRoad function utilized was fatigue transfer functions for strain performance criteria. The fatigue transfer functions within the PerRoad program calculated the fatigue life of the pavement, or the number of years until damage occurs (D=0.1), for test sections which were analyzed to have strain responses greater than their respective threshold limits. The PerRoad software required the input of the empirical constants, k 1 and k 2, in order to complete fatigue transfer functions. Priest and Timm (2006) conducted a study using the NCAT Test Track in order to develop these empirical

187 187 constants. The empirical constants they generated were used for this analysis and are shown in Table 5.6. Table 5.6 Fatigue Transfer Function Empirical Constants (Timm & Priest 2006) Empirical Constant Base Layer FRL K K Results The following provides results obtained from the PerRoad analysis using the structures utilized in the U.S. Route 23 test sections. It is important to note that the PerRoad program includes variability in the layer thicknesses and moduli that would realistically occur between different construction projects. The analysis should be used in consideration for future construction projects. Table 5.7 displays the results obtained for the 11 inch section. The table shows that 97.72% of the strains generated in the FRL were less than 70 µɛ resulting in a fatigue life of 325 years. The 11 inch section had the shortest fatigue life compared to the 13 and 15 inch sections. The maximum horizontal strain calculated in the FRL during the analysis was µɛ. In the base layer, the maximum horizontal strain found was µɛ and 99.94% of the strains calculated were less than 50 µɛ. The table also shows that 99.40% of the vertical deflections calculated at the bottom of the DGAB were less than 20 mils and the maximum deflection found was mils. Additionally, the

188 188 maximum subgrade pressure discovered during the analyses was 6.82 psi and all subgrade pressures calculated were less than eight psi. The 11 inch section had the highest values of subgrade deflection and pressure compared to the 13 and 15 inch sections. Table Inch Section PerRoad Results Pavement Response Maximum Pavement Percent Below Years to D=0.1 Response Threshold Horizontal Strain in the Base Layer µɛ Horizontal Strain in the FRL µɛ Vertical Deflection of the Subgrade mils NA Subgrade Pressure 6.82 psi NA Table 5.8 shows the results for the 13 inch section obtained from the PerRoad analysis and indicates that the maximum horizontal strain in the FRL found during the analysis was µɛ and 99.70% of the strains were less than 70 µɛ. The fatigue life corresponding to fatigue cracking occurring in the FRL was calculated to be 386 years. All of the horizontal strains calculated for the base layer were less than 50 µɛ and the maximum strain obtained was µɛ, as shown in Table 5.8. The 13 inch section received a maximum subgrade vertical deflection of mils during the analysis and 99.60% of the deflections were less than 20 mils. Additionally, the 13 inch section obtained a maximum subgrade pressure of 5.28 psi which was less than the threshold limit of eight psi.

189 189 Table Inch Section PerRoad Results Pavement Response Maximum Pavement Percent Below Years to D=0.1 Response Threshold Horizontal Strain in the Base Layer µɛ Infinite Horizontal Strain in the FRL µɛ Vertical Deflection of the Subgrade mils NA Subgrade Pressure 5.28 psi NA The 15 inch section was analyzed to have the longest fatigue life. Table 5.9 shows the 15 inch section PerRoad results. The maximum horizontal strain calculated in the FRL was µɛ and 99.60% of the strains obtained from the analysis were less than 70 µɛ. The fatigue life corresponding to horizontal strain in the FRL was 402 years. Table 5.9 shows that the maximum horizontal strain discovered in the base layer during the analysis was µɛ and 99.96% of the strains generated were less than 50 µɛ. The PerRoad program found that 99.66% of the vertical deflections produced at the bottom of the DGAB were less than 20 mils and the maximum deflection obtained was mils. Furthermore, the maximum subgrade pressure calculated by the software was 3.87 psi which was less than the 15 inch sections threshold limit of four psi. The 15 inch section received the lowest values of subgrade deflection and pressure compared to the 11 and 13 inch sections.

190 190 Table Inch Section PerRoad Results Pavement Response Maximum Pavement Percent Below Years to D=0.1 Response Threshold Horizontal Strain in the Base Layer µɛ Horizontal Strain in the FRL µɛ Vertical Deflection of the Subgrade mils NA Subgrade Pressure 3.87 psi NA

191 191 CHAPTER 6: CONCLUSIONS AND RECOMMEDATIONS The following draws conclusions and makes recommendations on objectives previously presented using CVL testing results and the PerRoad analysis and should provide a better understanding on perpetual pavement designs in Ohio and suggestions for future research in the perpetual pavement field. Additionally, these conclusions will advance the knowledge considering the effects of axle configuration, speed, and tire pressure on the load response of pavement structures. Pavement thickness, when designing a perpetual pavement structure, is one of the most important aspects of a proper design. The thickness of the asphalt layer, as part of a perpetual pavement system, limits two of the major distresses typically causing failure in a pavement section, bottom-up fatigue cracking and structural rutting. Although a thicker asphalt layer provides better protection against these failure modes, by applying the concept of the fatigue endurance limit, there comes a point where the pavement thickness becomes overdesigned and, therefore, an unnecessary use of resources. This study intended to determine an adequate pavement thickness, for perpetual pavements in Ohio, by analyzing load responses from various test sections utilizing different pavement thicknesses. Field measurements captured by sensors installed in the U.S. Route 23 test sections can effectively evaluate the performance of the pavement when they are compared to fatigue endurance thresholds. In order to minimize bottom-up fatigue cracking, a conservative threshold of 70 µɛ at the bottom of the asphalt layer was used. Structural rutting was evaluated using a vertical strain threshold at the top of the subgrade

192 192 of 200 µɛ. To further the investigation, subgrade displacements were compared to 20 mils and horizontal strains in the base layer were considered high if they exceeded 50 µɛ (Romanello 2007). The influence of variables including axle configuration, speed, and tire pressure on various load responses were analyzed. Additionally, comparisons were made between horizontal strains found in the FRL and base layer as well as between transverse and longitudinal orientations where applicable. In order to minimize the effects of wheel wander, these analyses were performed using CVL test runs with a lateral tire offset of two inches or less. Furthermore, due to the fact that different axles on the tandem axle truck produced various pavement responses, for these analyses, only pavement responses produced by the tandem axle of the tandem axle truck were used. A PerRoad analysis of the U.S. Route 23 test sections was performed to evaluate their perpetual nature. Using realistic properties determined for the test sections such as temperature, pavement layer thicknesses and properties, and loading conditions, PerRoad performed a linear elastic analysis on the test sections resulting in theoretical stresses, strains, and deflections the pavement structures would obtain during their service lives. These results where then compared with limiting pavement responses and fatigue transfer functions were used to calculate accumulating damages. 6.1 Conclusions The following conclusions were made based on the obtained results: Many factors influenced the shape of the pavement s strain response. Strains produced in the longitudinal direction below the neutral axis of the pavement

193 193 were tensile during direct loading from the tire but as the tire approached and moved away from the given location, the strains were compressive. Strains produced in the longitudinal direction above the neutral axis of the pavement were compressive during direct loading from the tire but as the tire approached and moved away from the given location, the strains were tensile. Strains produced in the transverse direction did not experience this opposing effect during the loading approach and departure. The tandem axle of the tandem axle truck did not always produce the maximum horizontal strain. As the location of the strain became closer to the surface, the maximum longitudinal strain was more likely to be produced by the steer axle. Transversely oriented strains were typically produced by the tandem axle although many exceptions were discovered in the base layer. When analyzing longitudinal strains in the FRL, the 11 inch section obtained the highest magnitudes while the 13 inch section obtained the lowest magnitudes which were slightly lower than the 15 inch section. The 11 and 13 inch sections had a subgrade that was previously chemically stabilized resulting in a resilient modulus of 40 ksi while the 15 inch section s subgrade had a resilient modulus of 20 ksi. Both the 13 and 15 inch sections received strains that were considerably lower than 70 µɛ while the 11 inch section received strains verging on the 70 µɛ threshold which was of some concern due to the low pavement temperature observed during testing. Although, all of the maximum longitudinal strains in the

194 FRL were obtained during five mph testing which is not the more typical speed of 55 mph on this section of U.S. Route The 11 inch section obtained the highest longitudinal and transverse strains in the base layer followed by the 15 inch section and the 13 inch section received the lowest. All three sections obtained results that were lower than the threshold limit provided by Romanello (2007) of 50 µɛ. The maximum longitudinal and transverse strains in the base layer for all three sections was consistently seen at a testing speed of five mph. For the 13 and 15 inch sections, transverse strains were compared to longitudinal strains throughout the depth of the pavement. This was an important comparison since it provides a better understanding of which orientation is critical at different depths within a pavement structure. For data pertaining to the single axle truck and the steer axle of the tandem axle truck, the ratio relating transverse strain to longitudinal strain remained approximately equal throughout the depth of the pavement, although, the ratio seemed to increase as the depth in the pavement increased. The data portraying the tandem axle of the tandem axle truck behaved similarly except the ratio spiked to an exceptionally high value at a depth of three inches. For both sections at the bottom of the asphalt layer, the transverse strain tended to be slightly greater than the longitudinal strain except for tandem axle measurements made in the 15 inch section where the longitudinal strain were slightly higher than the transverse strains.

195 195 Subgrade pressure results revealed that the 13 inch section was experiencing the lowest magnitudes of subgrade pressure and the 11 inch section was experiencing significantly higher subgrade pressures compared to the 13 and 15 inch sections. All subgrade pressure measurements in all three sections were correlated to be less than the vertical strain threshold limit of 200 µɛ. Additionally, the maximum subgrade pressures obtained for all three sections were done so at a testing speed of five mph. In terms of pavement and subgrade deflection, both measurements were similar for the 11 and 13 inch sections and were greater for the 15 inch section. In all three sections, all of the maximum subgrade deflections obtained were less than 20 mils. Furthermore, the maximum pavement and subgrade deflections measured for all three sections were recorded during five mph testing. It was discovered that subgrade deflections were usually between 68% and 89% of the corresponding total pavement deflections with single axle truck data resulting in percentages at the lower end of the range and tandem axle truck data resulting in percentages at the higher end of the range. For all pavement responses measured, including longitudinal strain in the FRL, horizontal strain in the base layer, subgrade pressure, pavement deflection, and subgrade deflection, results tended to decrease as the truck testing speed was increased. This trend was least prevalent in the 13 inch section for strain measurements obtained in both the FRL and base layer. Additionally, at higher tire pressures, the strain measured in the FRL and base layers of all three sections

196 196 tended to stabilize between 30 and 55 mph, although, as the pavement thickness increased, strain measurements continued to decrease between 30 and 55 mph even at higher tire pressures. When analyzing the influence of tire pressure on pavement load response it was discovered that the influence was minimal and any influence seen was diminished as pavement thickness increased. A minimal increase in horizontal strain in the base layer of all three sections was seen as tire pressure increased, however, the results tended to become erratic at the highest tire pressure, 125 psi. Although the tandem axle dual-tire truck was carrying an axle load greater than that of the single axle wide-based tire truck, 37 kips and 29 kips respectively, the tandem axle load was split between more axles and tires resulting in it producing lower pavement responses. However, pavement and subgrade deflections were similar when a contrast was made between the single axle and tandem axle truck. The influences observed for speed and tire pressure were more prevalent for the single axle truck except when analyzing pavement and subgrade deflections where the influences were similar between the two axle configurations. For each pavement response, including longitudinal strain in the FRL, longitudinal and transverse strain in the base layer, subgrade pressure, and pavement and subgrade deflection, the lowest pavement responses were measured in the 13 inch section in comparison to the 15 inch sections. It was expected that pavement responses would be lower for the thicker test section but due to the

197 197 variation in subgrade moduli, the 13 inch section received lower pavement responses than the 15 inch section. A PerRoad analysis revealed that all three sections had lifespan expectancies greatly exceeding 50 years. In all three sections the horizontal strain produced in the FRL controlled the life expectancy of the pavement. The 11, 13, and 15 inch sections received predicted lifespans of 324, 385, and 402 years respectively. Pavement responses, including horizontal strain in the FRL and base layer, vertical deflection of the subgrade, and subgrade pressure, were calculated to be greater than their corresponding threshold limit less than one percent of the time. One exception was the horizontal strain in the FRL of the 11 inch section which was calculated to be greater than 70 µɛ 2.28% of the time. 6.2 Recommendations A firm conclusion on an optimized pavement thickness for perpetual pavements in Ohio should not be made until further testing is conducted during higher pavement temperatures where pavement responses are augmented. Although, after analyzing the CVL testing conducted at colder pavement temperatures and the PerRoad analysis performed, the recommendation may be that a pavement thickness of 13 inches can be used for systems utilizing a stiffer subgrade with a high resilient modulus but an increase in thickness may be required for subgrades of lower resilient moduli. For proposed pavement structures at sites with subgrades of lower resilient moduli, a cost analysis should be performed to

198 determine whether the subgrade should be stabilized or the asphalt layer should be designed with an increased pavement thickness. 198 Due to the similarity between transverse and longitudinal strain responses, both orientations of strain gages should be used for future test sections especially when testing involving axles employing single tires is proposed. The literature review conducted for this report revealed that testing perpetual pavement systems without the use of an FRL while maintaining the pavement thickness may be beneficial. Additionally, the use of a DGAB could prove to be unnecessary in situations where the subgrade has a higher resilient modulus than the DGAB. Further investigation into the accuracy of the PerRoad computer program should be conducted before PerRoad analysis results are deemed acceptable.

199 199 REFERENCES Al-Qadi, I. L., Wang, H., Yoo, P. J., & Dessouky, S. H. (2008). Dynamic analysis and in situ validation of perpetual pavement response to vehicular loading. Transportation Research Record: Journal of the Transportation Research Board, 2087(1), Asphalt Pavement Alliance. (2002). Perpetual pavements: A synthesis. (APA 101). Lanham, MD: National Asphalt Pavement Association, Asphalt Institute, State Asphalt Pavement Associations. Asphalt Pavement Association of Oregon. (2005). Perpetual pavement concept renders 20-year standard obsolete. Centerline, 4(2), 1 2. Battaglia, I. K., Bischoff, D., Ryan, J., & Reichelt, S. (2010). Evaluation of a hot mix asphalt perpetual pavement. (FEP-01-10). Madison, WI: Wisconsin Department of Transportation Bendana, J., Sargand, S. M., & Hernandez, J. A. (2009). Comparison between perpetual and standard asphalt concrete pavement sections on NY I-86. Paper presented at the Proceedings for the International Conference on Perpetual Pavements. Columbus, OH. Brown, E. R., Cooley, L. A., Hanson, D., Lynn, C., Powell, B., Prowell, B., & Watson, D. (2002). NCAT test track design, construction, and performance. (NCAT Report 02-12). Auburn, AL: National Center for Asphalt Technology.

200 Buchner, M., Newcomb, D., & Huddleston, J. (2000). Perpetual pavements. Asphalt: The Magazine of the National Asphalt Institute, 15(3). 200 Bushman, R., Pratt, A.J. (1998). Weigh in motion technology - Economics and performance. Paper presented at the North American Travel Monitoring Exposition and Conference, Charlotte, NC. Carpenter, S. H., & Shen, S. (2006). Fatigue characteristics of rich bottom bases (RBB) for structural design of perpetual pavements. Paper presented at the Proceedings for the International Conference on Perpetual Pavements, Columbus, OH. Carpenter, S. H., Ghuzlan, K. A., & Shen, S. (2003). A fatigue endurance limit for highway and airport pavements. Paper presented at the Transportation Research Board 2003 Annual Meeting (Paper No ), Washinton, D.C. Estes, T. (2005). Oregon answers perpetual pavement analysis with a field test. Maximum Asphalt, 75(11), Geology (2013). Ohio map collection. Retrieved March 28, 2013, from Hatch, N. (2008). Perpetual pavement: monitoring performance in real time. (No ). Madison, WI: Wisconsin Department of Transportation.

201 201 Hernandez, J. A. (2010). Evaluation of the response of perpetual pavement at accelerated pavement loading facility: Finite element analysis and experimental investigation. (Unpublished MS thesis), Ohio University, Athens, OH. Hornyak, N. J. (2010). Perpetual pavement analysis for the Marquette interchange instrumentation project. (Unpublished MS thesis), Marquette University, Milwaukee, WI. Hornyak, N. J., & Crovetti, J. A. (2008). Marquette interchange perpetual pavement instrumentation project phase II final report. (WHRP 08-04). Milwaukee, WI: Transportation Reasearch Center, Marquette University. Jincheng, W., Lin, W., & Shijie, M. (2012). Modeling mechanical response of a perpetual pavement test road. Journal of Performance of Contrusted Facilities, 2(26), doi: /(asce)cf National Atlas of the United States. (2013). Transportation of the United States. Retrieved May 13, 2013, from National Cooperative Highway Research Program. (2008). An experimental plan validation of an endurance limit for HMA pavements. (Document 134). Sterling, VA: Advanced Asphalt Technologies.

202 202 National Cooperative Highway Research Program. (2010). Validating the fatigue endurance limit for hot mix asphalt. (Report 646). Washington, DC: Transportation Research Board. Newcomb, D. E., & Hansen, K. R. (2006). Mix type selection for perpetual pavements. Lanham, MD: National Asphalt Pavement Association. Newcomb, D. E., Willis, R., & Timm, D. H. (2010). Perpetual asphalt pavements a synthesis. (IM 40) Lanham, MD: Asphalt Pavement Alliance. Ning, L. I., Molenaar, A. A. A., Van de Ven, M. F. C., & Shaopeng, W. U. (2010). Estimation of the fatigue endurance limit of HMAC for perpetual pavements. Journal of Wuhan University of Technology-Mater, 25(4), doi: /s Nunn, M., & Ferne, B. W. (2001). Design and assessment of long-life flexible pavements. Transportation Research Circular, 503, ISSN: Portillo, M. M. (2008). Measured and theoretical reponse of perpetual pavement structures. (Unpublished MS thesis), University of Texas at Arlington, Arlington, TX. Priest, A. L., & Timm, D. H. (2006). Methodology and calibration of fatigue transfer functions for mechanistic-empirical flexible pavement design. (NCAT Report 06-03). Auburn, AL: National Center for Asphalt Technology.

203 203 Prowell, B. D., & Brown, E. R. (2006). Methods for determining the endurance limit using beam fatigue tests. Paper presented at the Proceedings for the International Conference on Perpetual Pavements, Columbus, OH. Quintus, H. L. Von. (2006). Application of the endurance limit premise in mechanisticempirical based pavement design procedures. Paper presented at the Proceedings for the International Conference on Perpetual Pavements, Columbus, OH. Retrepo-Velez, A. M. (2011). Long-term performance of asphalt concrete perpetual pavement WAY-30 project. (Unpublished MS thesis), Ohio University, Athens, OH. Robbins, M. M., & Timm, D. H. (2008). Temperature and velocity effects on a flexible perpetual pavement. Paper presented at the 3rd Internation Conference on Accelerated Pavement Testing, Madrid, Spain. Romanello, M. T. (2007). Load response analysis of the WAY-30 test pavements: US Route 30, Wayne County, Ohio. (Unpublished MS thesis), Ohio University, Athens, OH. Romanoschi, S. A., Gisi, A. J., & Dumitru, C. (2006). The dynamic response of Kansas perpetual pavements under vehicle loading. Paper presented at the Proceedings for the International Conference on Perpetual Pavements, Columbus, OH.

204 204 Romanoschi, S. A., Gisi, A. J., Portillo, M. M., & Dumitru, C. (2008). First findings from the Kansas perpetual pavements experiment. Transportation Research Record, 2068, Sargand, S., Figueroa, J. L., Edwards, W., & Al-Rawashdeh, A. S. (2009). Performance assessment of warm mix asphalt (WMA) pavements. (Report No. FHWA/OH- 2009/08). Athens, OH: Ohio Research Institute for Transportation and the Environment, Ohio Department of Transportation. Sargand, S., Figueroa, J. L., & Romanello, M. (2008). Instrumentation of the WAY-30 test pavements. (Report No. FHWA/OH-2008/7). Athens, OH: Ohio Research Institute for Transportation and the Environment, Ohio Department of Transportation. Sargand, S. M., & Figueroa, J. L. (2010). Monitoring and modeling of pavement response and performance task A: Ohio. (pp ). (Report No. FHWA/OH- 2010/3A). Athens, OH: Ohio Research Institute for Transportation and the Environment, Ohio Department of Transportation. Sargand, S. M., Khoury, I. S., & Morrison, J. (2012). Monitoring and modeling of pavement response and performance task b: New York (Vol. 1). (Report No. FHWA/OH-2012/08A). Athens, OH: Ohio Research Institute for Transportation and the Environment, Ohio Department of Transportation.

205 205 Sargand, S. M., Khoury, I. S., Romanello, M. T., & Figueroa, J. L. (2006). Seasonal and load response instrumentation of the WAY-30 perpetual pavements. Paper presented at the Proceedings for the International Conference on Perpetual Pavements, Columbus, OH. Scholz, T. V., Huddleston, J., Hunt, E. A., Lundy, J. R., & Shippen, N. C. (2006). Instrumentation and analysis of a perpetual pavement on an interstate freeway in oregon. Paper presented at the Proceedings for the International Conference on Perpetual Pavements, Columbus, OH. Tarefder, R. A., & Bateman, D. (2009). Determining the optimal perpetual pavement structure. Ohio Research Institute for Transportation and the Environment. Paper presented at the Procedings for the International Conference on Perpetual Pavement. Columbus, OH. Tarefder, R. A., & Bateman, D. (2012). Design of optimal perpetual pavement structure. Journal of Transportation Engineering, (February), doi: /(asce)te Thompson, M. R., & Carpenter, S. H. (2006a). Perpetual pavement design: an overview. Paper presented at the Proceedings for the International Conference on Perpetual Pavements, Columbus, OH.

206 206 Thompson, M. R., & Carpenter, S. H. (2006b). Considering hot-mix-asphalt fatigue endurance limit in full-depth mechanistic-empirical pavement design. Paper presented at the Proceedings for the International Conference on Perpetual Pavements, Columbus, OH. Timm, D. H. (2009). Design, construction and instrumentation of the 2006 test track structural study. (NCAT Report 09-01). Auburn, AL: National Center for Asphalt Technology. Timm, D. H., & Davis, K. P. (2009). Perpetual pavement design using the MEPDG and PerRoad. Paper presented at the Proceedings for the Internation Conference on Perpetual Pavement, Columbus, OH. Timm, D. H., & Newcomb, D. E. (2006). Perpetual pavement design for flexible pavements in the US. International Journal of Pavement Engineering, 7(2), doi: / Timm, D. H., Robbins, M. M., Huber, G., & Yang, Y. (2010). Analysis of perpetual pavement experiment sections in china. Paper presented at the 90th Annual Meeting of the Transportation Research Board, Washington, D.C. Timm, D., Selvaraj, I., Brown, R., West, R. C., Priest, A., Powell, B., & Zhang, J. (2006). Phase II NCAT test track results. (NCAT Report 06-05). Auburn, AL: National Center for Asphalt Technology.

207 207 The Weather Channel (2013). Monthly weather for Delaware. Retrieved from Willis, J. R., & Timm, D. H. (2009). Field-based strain thresholds for flexible perpetual pavement design. (NCAT Report 09-09). Auburn, AL: National Center for Asphalt Technology. Willis, J. R. (2009). Field-based strain thresholds for flexible perpetual pavement design. (Unpublished Ph.D. dissertation), Auburn University, Auburn, AL. Willis, R., Timm, D., West, R., Powell, B., Robbins, M., Taylor, A., Smit, A., et al. (2009). Phase III NCAT test track findings. (NCAT Report 09-08). Auburn, AL: National Center for Asphalt Technology. Yang, Y., Gao, X., Lin, W., Timm, D. H., Priest, A. L., Huber, G. A., & Andrewski, D. A. (2006). Perpetual pavement design in China. Paper presented at the Proceedings for the International Conference on Perpetual Pavements. Columbus, OH.

208 208 APPENDIX A: LVDT CALIBRATION Table A.1 Calibration Data (LVDT 1-1 through 4-2) Displacement (in) 1-1 (V) 1-2 (V) 2-1 (V) 2-2 (V) 3-1 (V) 3-2 (V) 4-1 (V) 4-2 (V)

209 209 Table A.2 Calibration Data (LVDT 5-1 through 8-2) Displacement (in) 5-1 (V) 5-2 (V) 6-1 (V) 6-2 (V) 7-1 (V) 7-2 (V) 8-1 (V) 8-2 (V)

210 210 Table A.3 LVDT Calibration Factors Sensor Calibration Factor (V/in) Calibration Factor (mil/mv) LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT LVDT

211 211 APPENDIX B: LVDT CASE AND REFERENCE ROD DIAGRAMS Figure B.1 11 Inch Section LVDT Case and Reference Rod

212 Figure B.2 13 Inch Section LVDT Case and Reference Rod 212

213 Figure B.3 15 Inch Section LVDT Case and Reference Rod 213

214 214 APPENDIX C: STRAIN GAGE ROSETTE HOLE DIAGRAMS Figure C.1 13 Inch Section Round Strain Gage Rosette Hole

215 Figure C.2 15 Inch Section Round Strain Gage Rosette Hole 215

216 Figure C.3 13 Inch Section Square Strain Gage Rosette Hole 216

217 Figure C.4 15 Inch Section Square Strain Gage Rosette Hole 217

218 218 APPENDIX D: INSTRUMENTATION DIAGRAMS Figure D.1 11 Inch Section Instrumentation

219 Figure D.2 13 and 15 Inch Section Instrumentation 219

220 220 APPENDIX E: TRUCK LOADINGS AND DIMENSIONS Figure E.1 Single Axle Wide-Based Tire (Empty) Figure E.2 Single Axle Wide-Based Tire (Max Load)

221 221 Figure E.3 Tandem Axle Dual Tire (Empty) Figure E.4 Tandem Axle Dual Tire (Max Load)

222 222 APPENDIX F: TESTING TEMPERATURE DATA Table F.1 Testing Temperatures for the 11 Inch Section (12/18/12) Time of Day Air Temp ( C) TC1 ( C) TC2 ( C) TC3 ( C) TC4 ( C) 10: : : : : : : : : : : : : : : : : : : : : : : : : Note: T1 and T2 were located in the FRL and T3 and T4 were located in the base layer

223 223 Table F.2 Testing Temperatures for the 13 Inch Section (12/19/12) Time of Day Air Temp ( C) TC1 ( C) TC2 ( C) TC3 ( C) TC4 ( C) TC5 ( C) TC6 ( C) 10: : : : : : : : : : : : : : : : : : : : : : Note: T1 and T2 were located in the FRL, T3 and T4 were located in the base layer, and T5 and T6 were located in the intermediate layer

224 224 Table F.2 Testing Temperatures for the 15 Inch Section (11/29/12) Time of Day Air Temp ( C) TC1 ( C) TC2 ( C) TC3 ( C) TC4 ( C) TC5 ( C) TC6 ( C) 10: : : : : : : : : : : : : : : : : : : : : : : : : Note: T1 and T2 were located in the FRL, T3 and T4 were located in the base layer, and T5 and T6 were located in the intermediate layer

225 225 APPENDIX G: STRAIN GAGE DATA Table G.1 Maximum Strains for Testing on the 11 Inch Section Involving a Tire Pressure of 80 psi (µɛ) Run Longitudinal Strain Gages Located in the FRL KM-001 KM-002 KM-003 PM-001 PM-002 PM-003 KM-004 KM-006 PM-005 KM-005 PM-004 PM-006 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Longitudinal Strain Gages Located in the Base Layer Transverse Strain Gages Located in the Base Layer Tandem Axle 55 MPH Test Runs

226 Table G.2 Maximum Strains for Testing on the 11 Inch Section Involving a Tire Pressure of 110 psi (µɛ) 226 Run Longitudinal Strain Gages Located in the FRL KM-001 KM-002 KM-003 PM-001 PM-002 PM-003 KM-004 KM-006 PM-005 KM-005 PM-004 PM-006 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Longitudinal Strain Gages Located in the Base Layer Transverse Strain Gages Located in the Base Layer Tandem Axle 55 MPH Test Runs

227 Table G.3 Maximum Strains for Testing on the 11 Inch Section Involving a Tire Pressure of 125 psi (µɛ) 227 Run Longitudinal Strain Gages Located in the FRL KM-001 KM-002 KM-003 PM-001 PM-002 PM-003 KM-004 KM-006 PM-005 KM-005 PM-004 PM-006 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Longitudinal Strain Gages Located in the Base Layer Transverse Strain Gages Located in the Base Layer Tandem Axle 55 MPH Test Runs

228 Table G.4 Maximum Strains for Testing on the 13 Inch Section Involving a Tire Pressure of 80 psi (µɛ) 228 Run Longitudinal Strain Gages Located in the FRL Longitudinal Strain Gages Located in the Base Layer KM-001 KM-002 KM-003 PM-003 KM-005 KM-008 PM-005 KM-007 PM-004 PM-006 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Transverse Strain Gages Located in the Base Layer Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs

229 Table G.5 Maximum Strains for Testing on the 13 Inch Section Involving a Tire Pressure of 110 psi (µɛ) 229 Run Longitudinal Strain Gages Located in the FRL Longitudinal Strain Gages Located in the Base Layer KM-001 KM-002 KM-003 PM-003 KM-005 KM-008 PM-005 KM-007 PM-004 PM-006 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Transverse Strain Gages Located in the Base Layer Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs

230 Table G.6 Maximum Strains for Testing on the 13 Inch Section Involving a Tire Pressure of 125 psi (µɛ) 230 Run Longitudinal Strain Gages Located in the FRL Longitudinal Strain Gages Located in the Base Layer KM-001 KM-002 KM-003 PM-003 KM-005 KM-008 PM-005 KM-007 PM-004 PM-006 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Transverse Strain Gages Located in the Base Layer Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs

231 Table G.7 Maximum Strains for Testing on the 15 Inch Section Involving a Tire Pressure of 80 psi (µɛ) 231 Run Longitudinal Strain Gages Located in the FRL Longitudinal Strain Gages Located in the Base Layer PM-001 PM-002 PM-003 KM-001 PM-005 KM-005 KM-008 PM-004 PM-006 KM-007 Single Axle Five MPH Test Runs NA NA NA Tandem Axle Five MPH Test Runs NA NA NA Single Axle 33 MPH Test Runs NA NA NA Tandem Axle 33 MPH Test Runs Transverse Strain Gages Located in the Base Layer NA NA NA Single Axle 55 MPH Test Runs NA NA NA Tandem Axle 55 MPH Test Runs NA NA NA

232 Table G.8 Maximum Strains for Testing on the 15 Inch Section Involving a Tire Pressure of 110 psi (µɛ) 232 Run Longitudinal Strain Gages Located in the FRL Longitudinal Strain Gages Located in the Base Layer PM-001 PM-002 PM-003 KM-001 PM-005 KM-005 KM-008 PM-004 PM-006 KM-007 Single Axle Five MPH Test Runs NA NA NA Tandem Axle Five MPH Test Runs NA NA NA Single Axle 33 MPH Test Runs NA NA NA Tandem Axle 33 MPH Test Runs Transverse Strain Gages Located in the Base Layer NA NA NA Single Axle 55 MPH Test Runs NA NA NA Tandem Axle 55 MPH Test Runs NA NA NA

233 Table G.9 Maximum Strains for Testing on the 15 Inch Section Involving a Tire Pressure of 125 psi (µɛ) 233 Run Longitudinal Strain Gages Located in the FRL Longitudinal Strain Gages Located in the Base Layer PM-001 PM-002 PM-003 KM-001 PM-005 KM-005 KM-008 PM-004 PM-006 KM-007 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Transverse Strain Gages Located in the Base Layer Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs

234 Table G.10 Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for Testing on the 11 Inch Section (µɛ) 234 Run Speed (mph) Longitudinal Strain Gages Located in the FRL Longitudinal Strain Gages Located in the Base Layer Transverse Strain Gages Located in the Base Layer KM-001 KM-002 KM-003 PM-001 PM-002 PM-003 KM-004 KM-006 PM-005 KM-005 PM-004 PM psi Test Runs psi Test Runs psi Test Runs

235 Table G.11 Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for Testing on the 13 Inch Section (µɛ) Run Longitudinal Strain Gages Transverse Strain Gages Speed Located in the Base Layer Located in the Base Layer (mph) KM-005 KM-008 PM-005 KM-007 PM-004 PM psi Test Runs psi Test Runs psi Test Runs Table G.12 Maximum Strains Produced by the Tandem Axle of the Tandem Axle Truck for Testing on the 15 Inch Section (µɛ) Run Longitudinal Strain Gages Transverse Strain Gages Speed Located in the Base Layer Located in the Base Layer (mph) PM-005 KM-005 KM-008 PM-004 PM-006 KM psi Test Runs psi Test Runs psi Test Runs

236 236 Table G.13 Additional Maximum Strains for Testing on the 13 Inch Section (µɛ) FRL Intermediate Intermediate Speed Run Truck Transverse Longitudinal Transverse (mph) PM-002 KM-011 KM-013 KM-010 KM psi Test Runs 3 5 Single Tandem Single Tandem Single Tandem psi Test Runs 1 5 Single Tandem Single Tandem Single Tandem psi Test Runs 1 5 Single Tandem Single Tandem Single Tandem Note: Tandem axle truck data refers to maximum strain produced by the tandem axle of the tandem axle truck.

237 237 Table G.14 Additional Maximum Strains for Testing on the 15 Inch Section (µɛ) FRL Intermediate Intermediate Surface Surface Speed Run Truck Transverse Longitudinal Transverse Longitudinal Transverse (mph) KM-002 KM-013 KM-011 KM-010 KM-012 WFLM-043 WFLM psi Test Runs 1 5 Single Tandem Single Tandem Single Tandem psi Test Runs 1 5 Single Tandem Single Tandem Single Tandem psi Test Runs 3 5 Single Tandem Single Tandem Single Tandem Note: Tandem axle truck data refers to maximum strain produced by the tandem axle of the tandem axle truck.

238 Table G.15 Maximum Strains Produced by the Steer Axle of the Tandem Axle Truck for Testing on the 13 Inch Section (µɛ) Run Longitudinal Strain Gages Transverse Strain Gages Speed Located in the Base Layer Located in the Base Layer (mph) KM-005 KM-008 PM-005 KM-007 PM-004 PM psi Test Runs NA psi Test Runs psi Test Runs

239 Table G.16 Maximum Strains Produced by the Steer Axle of the Tandem Axle Truck for Testing on the 15 Inch Section (µɛ) Run Longitudinal Strain Gages Transverse Strain Gages Speed Located in the Base Layer Located in the Base Layer (mph) PM-005 KM-005 KM-008 PM-004 PM-006 KM psi Test Runs NA NA NA psi Test Runs NA NA NA psi Test Runs NA NA NA Table G.17 Additional Maximum Strains Produced by the Steer Axle of the Tandem Axle Truck for Testing on the 13 Inch Section (µɛ) FRL Intermediate Intermediate Speed Run Transve Longitudinal Transverse (mph) PM-002 KM-011 KM-013 KM-010 KM psi Test Runs NA NA NA psi Test Runs NA NA NA psi Test Runs NA NA NA

240 Table G.18 Additional Maximum Strains Produced by the Steer Axle of the Tandem Axle Truck for Testing on the 15 Inch Section (µɛ) 240 FRL Intermediate Intermediate Surface Surface Speed Run Transve Longitudinal Transverse Longitudinal Transverse (mph) KM-002 KM-013 KM-011 KM-010 KM-012 WFLM-043 WFLM psi Test Runs psi Test Runs psi Test Runs

241 241 APPENDIX H: LVDT AND PRESSURE CELL DATA Table H.1 Maximum Displacements and Pressures for Testing on the 11 Inch Section Involving a Tire Pressure of 80 psi Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs NA NA Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs NA Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

242 Table H.2 Maximum Displacements and Pressures for Testing on the 11 Inch Section Involving a Tire Pressure of 110 psi 242 Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

243 Table H.3 Maximum Displacements and Pressures for Testing on the 11 Inch Section Involving a Tire Pressure of 125 psi 243 Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

244 Table H.4 Maximum Displacements and Pressures for Testing on the 13 Inch Section Involving a Tire Pressure of 80 psi 244 Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs 1 NA NA NA Tandem Axle Five MPH Test Runs 1 NA NA NA Single Axle 33 MPH Test Runs 1 NA NA NA Tandem Axle 33 MPH Test Runs 1 NA NA Single Axle 55 MPH Test Runs NA Tandem Axle 55 MPH Test Runs 1 NA NA Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

245 Table H.5 Maximum Displacements and Pressures for Testing on the 13 Inch Section Involving a Tire Pressure of 110 psi 245 Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs 1 NA NA NA Tandem Axle Five MPH Test Runs 1 NA NA NA Single Axle 33 MPH Test Runs 1 NA NA NA Tandem Axle 33 MPH Test Runs 1 NA NA Single Axle 55 MPH Test Runs 1 NA NA NA Tandem Axle 55 MPH Test Runs Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

246 Table H.6 Maximum Displacements and Pressures for Testing on the 13 Inch Section Involving a Tire Pressure of 125 psi 246 Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs NA Tandem Axle 55 MPH Test Runs Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

247 Table H.7 Maximum Displacements and Pressures for Testing on the 15 Inch Section Involving a Tire Pressure of 80 psi 247 Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

248 Table H.8 Maximum Displacements and Pressures for Testing on the 15 Inch Section Involving a Tire Pressure of 110 psi 248 Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

249 Table H.9 Maximum Displacements and Pressures for Testing on the 15 Inch Section Involving a Tire Pressure of 125 psi 249 Run Displacements (mils) Pressure (psi) LV-001 LV-002 LV-003 LV-004 PC-001 PC-002 Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs Note: LV-001 and LV-004 were deep LVDTs and LV-002 and LV-003 were shallow LVDTs

250 250 APPENDIX I: LATERAL TIRE OFFSET DATA Table I.1 Average Lateral Tire Offset (in.) 11 Inch Section 13 Inch Section 15 Inch Section Run Pressure (psi) Single Axle Five MPH Test Runs Tandem Axle Five MPH Test Runs Single Axle 33 MPH Test Runs Tandem Axle 33 MPH Test Runs Single Axle 55 MPH Test Runs Tandem Axle 55 MPH Test Runs

251 251 APPENDIX J: WIM ACCURACY DATA Table J.1 Accuracy Data for Classes One Through Four in 15 Minute Intervals Class 1 - Motorcycles Class 2 - Passenger Cars Class 3 - Trucks and Vans Class 4 - Buses Time Manual WIM Percent Manual WIM Percent Manual WIM Percent Manual WIM Percent (AM) Count Count Error Count Count Error Count Count Error Count Count Error 7:00-7: % % % % 7:05-7: % % % % 7:10-7: % % % % 7:15-7: % % % % 7:20-7: % % % % 7:25-7: % % % % 7:30-7: % % % % 7:35-7: % % % % 7:40-7: % % % % 7:45-7: % % % % 7:50-7: % % % % 7:55-8: % % % % 8:00-8: % % % % 8:05-8: % % % % 8:10-8: % % % % 8:15-8: % % % % 8:20-8: % % % % 8:25-8: % % % % 8:30-8: % % % % 8:35-8: % % % % 8:40-8: % % % % 8:45-8: % % % % 8:50-8: % % % % 8:55-9: % % % % 9:00-9: % % % % 9:05-9: % % % % 9:10-9: % % % % 9:15-9: % % % % 9:20-9: % % % % 9:25-9: % % % % 9:30-9: % % % % 9:35-9: % % % % 9:40-9: % % % % 9:45-9: % % % % 9:50-9: % % % % 9:55-10: % % % % Totals % % % %

252 252 Table J.2 Accuracy Data for Classes Five Through Eight in 15 Minute Intervals Time (AM) Class 5-2 Axles, 6 Tire Single Units Class 6-3 Axles, Single Unit Class 7-4 or More Axles, Single Unit Class 8-3 to 4 Axles, Single Trailer Manual Count WIM Count Percent Error Manual Count WIM Count Percent Error Manual Count 7:00-7: % % % % 7:05-7: % % % % 7:10-7: % % % % 7:15-7: % % % % 7:20-7: % % % % 7:25-7: % % % % 7:30-7: % % % % 7:35-7: % % % % 7:40-7: % % % % 7:45-7: % % % % 7:50-7: % % % % 7:55-8: % % % % 8:00-8: % % % % 8:05-8: % % % % 8:10-8: % % % % 8:15-8: % % % % 8:20-8: % % % % 8:25-8: % % % % 8:30-8: % % % % 8:35-8: % % % % 8:40-8: % % % % 8:45-8: % % % % 8:50-8: % % % % 8:55-9: % % % % 9:00-9: % % % % 9:05-9: % % % % 9:10-9: % % % % 9:15-9: % % % % 9:20-9: % % % % 9:25-9: % % % % 9:30-9: % % % % 9:35-9: % % % % 9:40-9: % % % % 9:45-9: % % % % 9:50-9: % % % % 9:55-10: % % % % Totals % % % % Manual Count WIM Count Percent Error WIM Count Percent Error

253 253 Table J.3 Accuracy Data for Classes 9 Through 12 in 15 Minute Intervals Time (AM) Class 9-5 Axles, Single Trailer Class 10-6 or More Axles, Single Trailer Class 11-5 or Less Axles, Multi-Trailer Manual Count WIM Count Percent Error Manual Count WIM Count Percent Error 7:00-7: % % % % 7:05-7: % % % % 7:10-7: % % % % 7:15-7: % % % % 7:20-7: % % % % 7:25-7: % % % % 7:30-7: % % % % 7:35-7: % % % % 7:40-7: % % % % 7:45-7: % % % % 7:50-7: % % % % 7:55-8: % % % % 8:00-8: % % % % 8:05-8: % % % % 8:10-8: % % % % 8:15-8: % % % % 8:20-8: % % % % 8:25-8: % % % % 8:30-8: % % % % 8:35-8: % % % % 8:40-8: % % % % 8:45-8: % % % % 8:50-8: % % % % 8:55-9: % % % % 9:00-9: % % % % 9:05-9: % % % % 9:10-9: % % % % 9:15-9: % % % % 9:20-9: % % % % 9:25-9: % % % % 9:30-9: % % % % 9:35-9: % % % % 9:40-9: % % % % 9:45-9: % % % % 9:50-9: % % % % 9:55-10: % % % % Totals % % % % Manual Count WIM Count Percent Error Class 12-6 Axles, Multi-Trailers Manual Count WIM Count Percent Error

254 254 APPENDIX K: PERROAD INPUTS Table K.1 U.S. Route 23 Yearly Volume by Classification Class %AADTT % % % % % % % % % % Table K.2 Average Monthly Temperature in Delaware, Ohio Month Average Temperature ( F) January 29 February 33 March 42 April 54 May 63 June 72 July 75 August 73 September 67 October 55 Novemeber 44 December 33

255 !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Thesis and Dissertation Services!