<%BANNER%>

Effect of Mixture Component Characteristics on Property and Performance of Superpave Mixtures

Permanent Link: http://ufdc.ufl.edu/UFE0044513/00001

Material Information

Title: Effect of Mixture Component Characteristics on Property and Performance of Superpave Mixtures
Physical Description: 1 online resource (124 p.)
Language: english
Creator: Chun, Sang Hyun
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: asphalt -- characteristics -- mixture -- performance -- property -- superpave
Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Civil Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This study was conducted to evaluate the effect of mixture component characteristics (i.e. DASR and IC) on properties and performance of Superpave mixtures, specializing in the development of a set of implementable gradation and volumetric criteria, and Hot-Mix-Asphalt (HMA) mixture property predictive relationships based on mixture component characterization.  Four DASR-IC model parameters including DASR porosity, DF, EFT, and CFA/FFA have formed the DASR-IC criteria to effectively address the two primary components (i.e. DASR and IC) of asphalt mixtures that play a major role on properties and performance. Field performance evaluation of different Superpave mixtures was conducted to identify the relationships between the four DASR-IC parameters and field performance. Based on results analyzed, it was found that the introduction of DASR-IC criteria as performance-related design parameters to current mix design guidelines and specifications will lead to better and more consistent field rutting and cracking performance of Superpave mixtures.  In addition, the DASR-IC criteria will also provide a more rational method to consider the effect of DASR and IC on mixture behavior which strongly affects HMA fracture properties. Therefore, it is expected that this criteria will have a potential to identify the effect of mixture gradation and volumetric characteristics on mixture fracture properties which is more reliably related to performance of asphalt mixtures. Relationships able to predict initial fracture energy and creep rate, which are the properties known to govern the change in material property over time and are also required for performance model predictions, were developed in this study.  Furthermore, conceptual relationships were identified to describe changes in these properties over time (aging) by including the effect of the non-healable permanent damage related to load and moisture. This can serve as the foundation for further development of improved models to predict mixture properties as a function of age in the field based on additional field data and laboratory studies using more advanced laboratory conditioning procedures. The verified relationships will also serve to provide reliable inputs for prediction of service life using pavement performance prediction models.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Sang Hyun Chun.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Roque, Reynaldo.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044513:00001

Permanent Link: http://ufdc.ufl.edu/UFE0044513/00001

Material Information

Title: Effect of Mixture Component Characteristics on Property and Performance of Superpave Mixtures
Physical Description: 1 online resource (124 p.)
Language: english
Creator: Chun, Sang Hyun
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: asphalt -- characteristics -- mixture -- performance -- property -- superpave
Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Civil Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: This study was conducted to evaluate the effect of mixture component characteristics (i.e. DASR and IC) on properties and performance of Superpave mixtures, specializing in the development of a set of implementable gradation and volumetric criteria, and Hot-Mix-Asphalt (HMA) mixture property predictive relationships based on mixture component characterization.  Four DASR-IC model parameters including DASR porosity, DF, EFT, and CFA/FFA have formed the DASR-IC criteria to effectively address the two primary components (i.e. DASR and IC) of asphalt mixtures that play a major role on properties and performance. Field performance evaluation of different Superpave mixtures was conducted to identify the relationships between the four DASR-IC parameters and field performance. Based on results analyzed, it was found that the introduction of DASR-IC criteria as performance-related design parameters to current mix design guidelines and specifications will lead to better and more consistent field rutting and cracking performance of Superpave mixtures.  In addition, the DASR-IC criteria will also provide a more rational method to consider the effect of DASR and IC on mixture behavior which strongly affects HMA fracture properties. Therefore, it is expected that this criteria will have a potential to identify the effect of mixture gradation and volumetric characteristics on mixture fracture properties which is more reliably related to performance of asphalt mixtures. Relationships able to predict initial fracture energy and creep rate, which are the properties known to govern the change in material property over time and are also required for performance model predictions, were developed in this study.  Furthermore, conceptual relationships were identified to describe changes in these properties over time (aging) by including the effect of the non-healable permanent damage related to load and moisture. This can serve as the foundation for further development of improved models to predict mixture properties as a function of age in the field based on additional field data and laboratory studies using more advanced laboratory conditioning procedures. The verified relationships will also serve to provide reliable inputs for prediction of service life using pavement performance prediction models.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Sang Hyun Chun.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Roque, Reynaldo.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044513:00001


This item has the following downloads:


Full Text

PAGE 1

1 EFFECT OF MIXTURE COMPONENT CHARACTERISTICS ON PROPERTY AND PERFORMANCE OF SUPERPAVE MIXTURES By SANGHYUN CHUN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQ UIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Sanghyun Chun

PAGE 3

3 To my beloved wife, Eunsong Lee and lovely son, Woobin Chun

PAGE 4

4 ACKNOWLEDGMENTS It is a great pleasure for me to thank and acknowledge the indiv iduals who advised and supported during the course of my doctoral program. First of all, I would like to express my heartfelt appreciation to my committee chairman, Dr. Reynaldo Roque for his invaluable guidance and support throughout my studies at the Uni versity of Florida. I would not have been able to reach this great moment of my life without his encouragement, advice, and mentoring. I would like to extend my gratitude to other committee members, Dr. Mang Tia, Dr. Dennis R Hiltunen, Dr. Peter G. Ifju, and Dr. Claude Villiers, for their support to accomplish my doctoral study. I also would like to express my sincere appr eciation to Mr. George Lopp Dr. Jian Zou Dr. Weitao Li, Michael Bekoe, and Cristian Cocconcelli for their constructive advice and assis tance on laboratory testing data analysis and field works Many thanks would also be given to all former and current students in the material s group of the Department of Civil and Coastal Engineeri ng at the University of Florida for their help and friendship. Special thanks would go to Dr. Sungho Kim, Dr. Jaeseung Kim, Dr. Chulseung Koh, and Hyungsuk Lee for their encouragement and friendship. Lastly, I would like to thank my parents, Kijoon Chun and Wollim Myung, parents in law, Hwayoung Lee and C hunghee Shin, my sister, Seieun Chun, my brother, Sangkeun Chun, my wife, Eunsong Lee and my son Woobin Chun for their enduring trust encouragement and support

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 12 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTERS 1 INTRODUCTION ................................ ................................ ................................ .... 16 1.1 Background ................................ ................................ ................................ ....... 16 1.2 Hypothesis ................................ ................................ ................................ ........ 18 1.3 Objectives ................................ ................................ ................................ ......... 19 1.4 Scope ................................ ................................ ................................ ................ 19 1.5 Research Approach ................................ ................................ .......................... 20 2 CHARACTERIZATION OF MIXTURE GRADATION AND RESULTING VOLUMETRIC PROPERTIES (DOMINANT AGGREGATE SIZE RANGE INTERSTITIAL COMPONENT (DASR IC) MO DEL) ................................ ............... 22 2.1 Background ................................ ................................ ................................ ....... 22 2.2 Dominant Aggregate Size Range (DASR) ................................ ........................ 22 2.3 DASR Porosity ................................ ................................ ................................ .. 23 2.4 Interstitial Component (IC) of Mixture Gradation ................................ ............... 24 2.5 Disruption Factor (DF) ................................ ................................ ...................... 25 2.6 Effective Film Thickness (EFT) ................................ ................................ ......... 26 2.7 Ratio between Coarse Portion and Fine Portion of Fine Aggregates (CFA/FFA) ................................ ................................ ................................ ........... 29 2.8 Summary ................................ ................................ ................................ .......... 30 3 IMPLEMENTATION OF BINDER AND MIXTURE TESTS ON FIELD CORES FOR SUPERPAVE MIXTURES IN FLORIDA ................................ ......................... 31 3.1 Background ................................ ................................ ................................ ....... 31 3.2 Binder Recovery and Binder Tests ................................ ................................ ... 31 3.2.1 Binder Recovery ................................ ................................ ...................... 32 3.2.2 Penetration Test ................................ ................................ ...................... 32 3.2.3 Viscosity Test ................................ ................................ .......................... 33 3. 2.4 Dynamic Shear Rheometer Test (DSR) ................................ .................. 35 3. 2.5 Bending Beam Rheometer Test (BBR) ................................ .................... 37

PAGE 6

6 3. 2.6 Multiple Stress Creep Recovery Test (MSCR) ................................ ........ 39 3.3 Mixture Tests ................................ ................................ ................................ .... 42 3.3.1 Test Specimen Preparation ................................ ................................ ..... 42 3.3.1.1 Measuring, Cataloguing, and Inspecting ................................ ........ 42 3.3.1.2 Cutting ................................ ................................ ............................ 43 3.3.1.3 Gage Points Attachment ................................ ................................ 44 3.3.2 Test Procedure ................................ ................................ ........................ 45 3.3.2.1 Resilient Modulus Test ................................ ................................ ... 45 3.3.2.2 Creep Test ................................ ................................ ..................... 46 3.3.2.3 Tensile Strength Test ................................ ................................ ..... 48 3.3.3 Superpave IDT Test Results ................................ ................................ ... 50 3.3.4 Moisture Damaged Projects ................................ ................................ .... 58 3.4 Summary ................................ ................................ ................................ .......... 63 4 EVALUATION OF FIELD MIXTURE PERFORMANCE USING DASR IC MODEL PARAMETERS ................................ ................................ ......................... 64 4.1 Background ................................ ................................ ................................ ....... 64 4.2 Implementation of Gradation Analysis for Superpave Mixtures ........................ 64 4.3 Evaluation o f Gradation Analysis Results for Superpave Projects .................... 67 4.4 Evaluation of Field Mixture Performance ................................ .......................... 69 4.4.1 Field Performance: R utting ................................ ................................ ...... 70 4.4.2 Field Performance: Cracking ................................ ................................ ... 74 4.5 Summary ................................ ................................ ................................ .......... 78 5 IDENTIFICATION OF PREDICTIVE MIXTURE PROPERTY RELATIONSHIPS AND MODEL DEVELOPMENT ................................ ................................ .............. 80 5.1 Background ................................ ................................ ................................ ....... 80 5.2 Existing Material Pro perty Models in the HMA FM E Model ............................. 81 5.2.1 AC Stiffness Aging Sub Model ................................ ................................ 81 5.2.2 Fracture Energy Limit Aging Sub Model ................................ .................. 83 5.3 Key Elements for Material Property Relationships ................................ ............ 85 5.3.1 DASR IC Model Parameters ................................ ................................ ... 86 5.3.2 Initial Material Properties ................................ ................................ ......... 87 5.3.3 Factors Related to Non Healable Permanent Damage ........................... 88 5.4 Developm ent of Predictive Material Property Relationships ............................. 88 5.4.1 Relationships for Initial Material Properties ................................ ............. 89 5.4.1.1 Initial Frac ture Energy Relationship ................................ ............... 90 5.4.1.2 Initial Creep Rate Relationship ................................ ....................... 92 5.4.2 Models for Changes in Material Properties ................................ .............. 98 5.4.2.1 AC Stiffness Model ................................ ................................ ......... 99 5.4.2.2 Fracture Energy Limit Model ................................ ........................ 100 5.5 Summary ................................ ................................ ................................ ........ 103

PAGE 7

7 6 EVALUATION OF DASR IC CRITERIA USING HMA FM E MODEL ................... 104 6.1 Background ................................ ................................ ................................ ..... 104 6.2 Enhanced HMA Fracture Mechanics Based Model (HMA FM E Model) ......... 104 6.3 Input Module ................................ ................................ ................................ ... 105 6.4 Mod el Prediction Results ................................ ................................ ................ 108 6.5 Relationships between DASR IC Criteria and Model Prediction Results ........ 110 6.6 Summary ................................ ................................ ................................ ........ 113 7 CLOSURE ................................ ................................ ................................ ............ 114 7.1 Summary and Findings ................................ ................................ ................... 114 7.2 Conclusions ................................ ................................ ................................ .... 115 7.3 Recommendations and Future Works ................................ ............................. 116 APPENDIX: DETERMINISTIC PROCEDURE FOR ESTIMATION OF CRACK INITIATION TIME BASED ON CRACK RATING DA TA ................................ ........ 118 LIST OF REFERENCES ................................ ................................ ............................. 121 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 124

PAGE 8

8 LIST OF TABLES Table page 1 1 Mixture information of Superpave projects evaluated ................................ .......... 20 3 1 Asphalt binder used for Superpave projects evaluated ................................ ........ 31 3 2 Mixture information for 11 Superpave projects ................................ ..................... 50 3 3 Project information for moisture damaged sections ................................ ............. 59 4 1 Mixture information of Superpave projects analyzed ................................ ............ 65 4 2 DASR IC parameters calculated for Superpave projects ................................ ..... 65 4 3 Crack initiation time and cracking status determined for Superpave projects ...... 75 6 1 Summary of input data characteristics for the HMA FM E mod el ...................... 106 6 2 Data used for model prediction ................................ ................................ .......... 107 6 3 Predicted top down cracking performance using HMA FM E model .................. 109

PAGE 9

9 LIST OF FIGURES Figure page 1 1 DASR and IC for three different types of mixture (After Kim et al., 2006) ............ 17 1 2 Overall research approach flowchart ................................ ................................ .... 21 2 1 Mixture components for calculation of DASR porosity (Kim et al., 2006) ............. 24 2 2 Configuration of different DF values (Guarin, 2009) ................................ ............. 26 2 3 Effective film thickness vs. theoretical film thickness ................................ ........... 27 2 4 Conceptual drawing of film thickness effect ................................ ......................... 28 2 5 Determination of CFA/FFA ................................ ................................ ................... 29 3 1 Penetration test results for Superpave projects ................................ ................... 33 3 2 Viscosity t est r esults for Superpave p rojects ................................ ........................ 34 3 3 Change in viscosity with aging for Superpave projects ................................ ........ 35 3 4 G ................................ 36 3 5 S(t), 60 seconds loading time at 12 C (10.4 F) for Superpave projects ............ 38 3 6 m value, 60 seconds loading time at 12 C (10.4 F) for Superpave projects ..... 38 3 7 MSCR average recovery at 64 C (147.2 F) for Superpave projects .................. 40 3 8 MSCR nonrecoverable compliance at 64 C (147.2 F) for Superpave projects .. 41 3 9 Measuring, cataloguing, and inspecting work for field cores ................................ 43 3 10 Cut s pecimens for Superpave IDT t est s ................................ ............................... 43 3 11 Cutting machine used in t his study ................................ ................................ ....... 44 3 12 Gage points attachment ................................ ................................ ....................... 44 3 13 Superpave IDT tests ................................ ................................ ............................ 45 3 14 Power model of creep compliance ................................ ................................ ....... 47 3 15 Determination of f racture e nergy and d issipated c reep s train e nergy to f ailure ... 49 3 16 Change in resilient modulus over time ................................ ................................ 52

PAGE 10

10 3 17 Change in creep compliance over time ................................ ................................ 53 3 18 Change in creep rate over time ................................ ................................ ............ 54 3 19 Change in tensile strength over time ................................ ................................ .... 55 3 2 0 Change in failure strain over time ................................ ................................ ......... 56 3 21 Change in fracture energy over time ................................ ................................ .... 57 3 22 Change in air voids ov er time ................................ ................................ ............... 60 3 23 Initial rate of reduction in normalized fracture energy over time ........................... 61 3 24 Initial rate of reduct ion in normalized energy ratio over time ................................ 62 4 1 Initial fracture energy and creep rate for Superpave projects evaluated .............. 66 4 2 Field rutting performance for Superpave projects evaluated ................................ 71 4 3 Rutting performance evaluation using DASR IC parameters ............................... 73 4 4 Determination of observed crack initiation time for Project 1 and 2 ...................... 74 4 5 Field cracking performance for Superpave projects evaluated ............................ 75 4 6 Cracking performance evaluation using DASR IC parameters ............................ 77 5 1 Schematic plot for AC stiffness at surface vs. age ................................ ............... 82 5 2 Schematic plot for creep rate vs. age ................................ ................................ ... 83 5 3 Schematic plot for normalized change in AC stiffness vs. age ............................. 84 5 4 Schematic plot for FE limit vs. age ................................ ................................ ....... 85 5 5 Two material property relationships ................................ ................................ ..... 86 5 6 FE limit aging curve at different initial FE (k 1 =3 (Roque et al. 2010)) ................... 88 5 7 Flowchart for development of predictive material property relationships .............. 89 5 8 Relationship between initial fracture energy and DASR porosity ......................... 90 5 9 Relationship between initial fracture energy and disruption factor ....................... 91 5 10 Relationship between initial fracture energy and effective film thickness ............. 91 5 11 Predicted vs. measured initial fracture energy ................................ ..................... 92

PAGE 11

11 5 1 2 Relationship between initial creep rate and DASR porosity ................................ 93 5 1 3 Relationship between initial creep rate and DF ................................ .................... 94 5 1 4 Relationship between initial creep rate and EFT ................................ .................. 94 5 1 5 Relationship between initial creep rate and CFA/FFA ................................ .......... 95 5 16 Effect of polymer modification on relationship between initial creep rate and CFA/FFA ................................ ................................ ................................ ............ 96 5 17 Relationship between initial creep rate and viscosity ................................ ........... 97 5 18 Relationship between initial creep rate and effe ctive asphalt content .................. 97 5 19 ................................ ............. 98 5 2 0 Proposed AC stiffness model ................................ ................................ ............. 100 5 2 1 Proposed FE limit model ................................ ................................ .................... 102 6 1 General f ramework of the HMA FM E m odel ................................ ..................... 105 6 2 Predicted c rack a mount i ncrease o ver t ime u sing HMA FM E m odel ................ 110 6 3 Relationships between DASR IC criteria and field cracking performance .......... 112 A 1 Determination of observed crack initiation time for Project 1 and 2 ................... 118 A 2 Determination of observed crack initiation time for Project 3 and 4 ................... 118 A 3 Determination of observed crack initiation time for Project 5 and 6 ................... 119 A 4 Determination of observed crack initiation time for Project 7 and 8 ................... 119 A 5 Determination of observed crack initiation time for Project 9 and 10 ................. 120 A 6 Determination of observed crack initiation time for Project 11 and 12 ............... 120

PAGE 12

12 LIST OF ABBREVIATION S AASHTO American Association of State Highway and Transpo rtation Officials AC Asphalt Concrete APT Accelerated Pavement Testing ASTM American Society for Testing and Materials BBR Bending Beam Rheometer CFA/FFA Ratio of Coarse Fine Aggregate to Fine Fine Aggregate DASR Dominant Aggregate Size Range DF Disruption Factor DSR Dynamic Shear Rheometer EAC Effective Asphalt Content EFT Effective Film Thickness ESAL Equivalent Single Axle Load FDOT Florida Department of Transportation FWD Falling Weight Deflectometer JMF Job Mix Formula HMA Hot Mix Asphalt HMA FM E Enha nced Hot Mix Asphalt Fracture Mechanics Based Model IA Independent Assurance IC Interstitial Component IDT Indirect Tension Test ITLT Indirect Tensi on Test at Low Temperatures IV Interstitial Volume MEPDG Mechanistic Empirical Pavement Design Guide MSCR M ultiple Stress Creep Recovery

PAGE 13

13 NCHRP National Cooperative Highway Research Program PAV Pressur e Aging Vessel PG Performance Grade QA Quality Assurance QC Quality Control TCE Trichloroethylene TF Theoretical Film Thickness UF University of Florida VMA Voids in Mineral Aggregate

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFECT OF MIXTURE COMPONENT CHARACTERISTICS ON PROPERTY AND PERFORMANCE OF SUPERPAVE MIXTURES By Sanghyun Chun August 2012 Chair: Reynaldo Roque Major: Civil Engineering This study was conducted to evaluate the effect of mixture component characteristics (i.e. DASR and IC) on properties and perform ance of Superpave mixtures, specializing in the development of a set of implementable gradation and volumetric criteria, and Hot Mix Asphalt (HMA) mixture property predictive relationships based on mixture component characterization. Four DASR IC model pa rameters including DASR porosity, DF, EFT, and CFA/FFA have formed the DASR IC criteria to effectively address the two primary components (i.e. DASR and IC) of asphalt mixtures that play a major role on properties and performance. Field performance evaluat ion of different Superpave mixtures was conducted to identify the relationships between the four DASR IC parameters and field performance Based on results analyzed, it was found that the introduction of DASR IC criteria as performance related design param eters to current mix design guidelines and specifications will lead to better and more consistent field rutting and cracking performance of Superpave mixtures. In addition, the DASR IC criteria will also provide a more rational method to consider the effe ct of DASR and IC on mixture behavior which strongly affects HMA

PAGE 15

15 fracture properties. Therefore, it is expected that this criteria will have a potential to identify the effect of mixture gradation and volumetric characteristics on mixture fracture properti es which is more reliably related to performance of asphalt mixtures. Relationships able to predict initial fracture energy and creep rate, which are the properties known to govern the change in material property over time and are also required for perform ance model predictions, were developed in this study. Furthermore, conceptual relationships were identified to describe changes in these properties over time (aging) by including the effect of the non healable permanent damage related to load and moisture This can serve as the foundation for further development of improved models to predict mixture properties as a function of age in the field based on additional field data and laboratory studies using more advanced laboratory conditioning procedures. The verified relationships will also serve to provide reliable inputs for prediction of service life using pavement performance prediction models.

PAGE 16

16 CHAPTER 1 INTRODUCTION 1.1 Background It is now generally agreed that aggregate gradation is one of the most imp ortant factors that affects the properties and performance of asphalt mixtures. Having suitable gradation characteristics including appropriate aggregate particle size distribution and resulting volumetric properties is obviously important to ensure good f ield mixture performance. Therefore, aggregate related parameters have been studied to identify their effects on observed field performance of asphalt mixtures. Although different parameters, including effective film thickness and other volumetric paramete rs were found to affect mixture performance, consensus has not been reached regarding rational design guidelines and criteria, especially as related to the selection of the best aggregate blend to achieve optimal performance. The Superpave field monitoring project recently completed at the University of Florida has determined that existing mix design criteria included in Superpave system such as Voids in Mineral Aggregate (VMA), gradation control points, and effective asphalt content do not capture all crit ical aspects of gradation and resulting volumetric properties found to be most strongly related to field mixture performance (Roque et A l., 2011). Thereby, Superpave mixture performance varied significantly among mixtures that met all existing design and c onstruction specification criteria. Therefore, there was a need to identify and verify additional criteria that can assure better and more consistent Superpave mixture performance It was also found that differences in performance could not explained by di fferences in binder properties between mixtures.

PAGE 17

17 It appeared that differences in performance were primarily controlled by differences in gradation and resulting volumetric properties between mixtures. According to previous work conducted by University of F lorida researchers gradation characteristics of mixture can be expressed by separating the gradation into two major components: Dominant Aggregate Size Range (DASR) and Interstitial Component (IC) (After Kim et al., 2006). It has been shown that parameter s describing the characteristics of these components, which were determined based on packing theory and particle size distributions, seemed to be well correlated to mixture performance. K im et al. (2006) indicated in their research that porosity can be use d as a criterion to ensure contact between DASR particles within the asphalt mixture to provide adequate interlocking and resistance to deformation and fracture. The work has clearly shown that DASR porosity can be used as an indicator which reflects the c haracteristics of coarse aggregate structure. The schematic of the DASR and IC concept for three different types of mixtures is illustrated in Figure 1 1. (a) SMA Mixture (b) Coarse Dense Mixture (c) Fine Dense Mixture Figure 1 1. DASR and IC for three different types of mixture (After Kim et al., 2006) The work has also concluded that properties and characteristics of IC will strongly influence rutting and cracking resistance of asphalt mixtures. For the purpose of IC

PAGE 18

18 characterization, a new parameter called Disruption Factor (DF) was conceived and developed by Guarin (2009) to evaluate the potential of IC particles to disrupt the DASR structure. It was found that DF appeared to be one good indicator to describe IC characteristics with respect to volumetric distribution of IC. However, DF only considers volumetric distribution of IC to determine the potential of the finer portion of the mixture s gradation (i.e., IC) to disrupt the DASR structure. Therefore, there was a need to identify and dev elop additional parameters to more effectively characterize the IC of asphalt mixture with regard to stiffening effect of IC on mixture and structure of IC. 1.2 Hypothesis Since there was a need to address more aspects of IC characteristics to be tter capture the effects of IC on key mixture properties and expected performance, two new parameters, effective film thickness (EFT) and ratio of coarse fine aggregate to fine fine aggregate (CFA/FFA), were added in addition to DF in this study for furthe r IC characterization. It was expected that these two parameters will provide a more rational way to identify mixture behavior as mastic which is considered to strongly affect Hot Mix Asphalt (HMA) fracture properties. Therefore, the addition of EFT and CF A/FFA will give a more potential to identify the effect of IC characteristics on properties which is more reliably related to performance of asphalt mixtures. The following two hypotheses were made in this study. Interstitial component (IC) characte ristics related to the aggregate structure and binder distribution within the IC have an important effect on mixture fracture properties as well as on pavement cracking performance. Gradation and volumetric parameters that effectively characterize the DASR and IC can be used to predict mixture fracture properties.

PAGE 19

19 1.3 Objectives The primary objective of this study is to evaluate the effect of mixture component characteristics (i.e., DASR and IC) on properties and performance of Superpave mixtures. Detailed objectives are summarized as follows. Further develop the DASR IC criteria to more effectively characterize both coarse (DASR) and fine (IC) portions of mixture gradation and resulting volumetric properties that play a major role on mixture performance. E valuate and verify the DASR IC criteria as an effective and implementable set of gradation and volumetric criteria for mixture design and construction specification that can help to assure better and more consistent field mixture performance. Identify key mixture component characteristics associated with mixture fracture properties and changes in these properties (i.e., fracture energy and creep rate) over time. Develop predictive relationships for mixture fracture properties, specifically fracture energy a nd creep rate, which have shown to strongly affect (or influence) pavement cracking performance in the field. Identify and further develop improved forms of the HMA fracture property aging model in an effort to more accurately predict pavement cracking per formance. 1.4 Scope Eleven Superpave monitoring project field sections were evaluated including different types of gradation, aggregate, and asphalt binder. It is noted that fairly wide ranges of Superpave mixture were evaluated in this study. All mixture data were obtained or determined from field cores including gradation, binder properties, volumetric properties, and mixture properties. The standard University of Florida (UF) Superpave Indirect Tension Tests (IDT) were performed at 10 C and 20 C, to ob tain the HMA fracture properties of the different Superpave mixtures used for evaluation. Enhanced Hot Mix Asphalt Fracture Mechanics Based Model (HMA FM E) was

PAGE 20

20 employed to evaluate criteria developed as a performance prediction model. Table 1 1 summarizes the mixture information of Superpave projects evaluated in this study. Table 1 1. Mixture information of Superpave projects evaluated Project (UF) ID Aggregate Type Binder Type Mixture Type Top Bottom Top Bottom 1 Granite PG 67 22 PG 67 22 9.5C 19.0C 2 Granite PG 64 22 PG 64 22 12.5C 19.0C 3 Limestone PG 67 22 PG 67 22 12.5C 19.0C 4 Limestone PG 67 22 PG 67 22 9.5C 19.0C 6 Limestone PG 64 22 N/A 12.5F N/A 7 Limestone PG 64 22 PG 64 22 12.5F 12.5F 8 Limestone PG 76 22 PG 76 22 12.5C 12.5C 9 Granite ARB 5 PG 64 22 FC 6 12.5F 10 Granite ARB 5 PG 64 22 FC 6 12.5F 11 Granite PG 76 22 PG 64 22 12.5C 12.5C 12 Granite ARB 5 PG 64 22 FC 6 12.5F Note: Mixture Type: C = Coarse M ixtures, F = Fine M ixtures, N/A = Not A pplicable 1.5 Research Approa ch This research is mainly focused on evaluating the effect of mixture component characteristics on properties and performance of Superpave mixtures in order to develop more implementable performance related criteria and predictive mixture property relatio nships. The overall research approach to accomplish the objectives of this study is shown in Figure 1 2. Details for each phase of this research are described in the following sections. Development of DASR IC criteria: (1) Identify key mixture component ch aracteristics; (2) Identify and develop parameters to effectively characterize mixture component characteristics identified; (3) Evaluate relationships between parameters identified and field mixture performance; (4) Develop implementable performance relat ed criteria using parameters evaluated (i.e., DASR IC criteria). Development of predictive mixture property relationships: (1) Identify key mixture properties to be evaluated; (2) Evaluate relationships between properties identified and mixture performance ; (3) Evaluate relationships between parameters employed in DASR IC criteria and mixture fracture properties; (4) Identify key elements associated with initial mixture properties and changes in these properties

PAGE 21

21 over time (aging); (5) Develop predictive rel ationships for initial mixture properties; (6) Develop improved forms of mixture property aging model. Evaluation of criteria developed: (1) Evaluate criteria developed using performance prediction model (HMA FM E model); (2) Validate and refine criteria d eveloped using additional field and laboratory data. Figure 1 2. Overall research approach flowchart

PAGE 22

22 CHAPTER 2 CHARACTERIZATION OF MIXTURE GRADATION AN D RESULTING VOLUMETR IC PROPERTIES ( DOMINANT AGGREGATE S IZE RANGE INTERSTITIAL COMPONENT ( DASR IC ) MO DEL) 2.1 Background R esearch recently conducted at the University of Florida has concluded that the gradation of mixtures can be characterized by separating the gradation into two major components: The Dominant Aggregate Size Range (DASR) and the Interstit ial Components (IC) (Kim et al., 2006, Guarin, 2009, Roque et al., 2011). It was also shown that parameters describing the characteristics of two components, which were determined based on packing theory and particle size distributions, seemed to be well c orrelated to mixture performance. These parameters are DASR porosity, Disruption Factor (DF), Effective Film Thickness (EFT), and ratio of Coarse Fine Aggregate to Fine Fine Aggregate (CFA/FFA), which are used to address the following aspects of gradation characteristics: DASR porosity: coarse aggregate interlocking Disruption Factor (DF) : v olumetric distribution of the IC Effective Film Thickness (EFT) : stiffening effect of IC on mixture CFA/FFA: structure of the IC Detailed descriptions with regard to the definition and calculation procedure of each parameter are included in the following sections. 2.2 Dominant Aggregate Size Range (DASR) The concept and theoretical development of DASR, which was defined as the interactive range of particle sizes that forms the dominant structural network of aggregate, was introduce by Kim et al. (2006). According to the DASR approach, there is an interactive range of particle sizes that primarily contributes to aggregate interlocking in asphalt mixtures. Particle sizes interacting with each other will form the

PAGE 23

23 primary structure to resist deformation and fracture. Particle sizes smaller than the DASR will serve to fill the voids between DASR particles, called the interstitial volume. The IC particles combined with binder form a secondary structure to help resist deformation and fracture, and it is the primary source of adhesion and resistance to tension. Particle sizes larger than the DASR will simply float in the DASR matrix and will not play a major role in the aggregat e structure. The DASR, which is determined by conducting particle interaction analysis based on packing theory, can be composed of one size or multiple sizes. It was concluded that the DASR should be composed of coarse enough particles, and that all contig uous particle sizes determined to be interactive can be considered as part of the DASR. 2.3 DASR Porosity Porosity has been widely used in the field of soil mechanics as a dimensionless parameter that indicates the relative ratio of voids to total volume. It has been determined that the porosity of granular materials should be no greater than 50 % for particles to have contact with each other (i.e. to be interactive) (Lambe and Whitman, 1969). Research conducted by Kim et al. (2006) indicated that porosity can be used as a criterion to ensure contact between DASR particles within the asphalt mixture to provide adequate interlocking and resistance to deformation and fracture. The basic principles related to the calculation of DASR porosity are as follows. The Voids in Mineral Aggregate (VMA) of asphalt mixtures, which indicates the volume of available space between aggregates in a compacted mixture is comparable to volume of voids in soil. Porosity can be calculated for any DASR by assuming that a mixture has certain effective asphalt content and air voids (i.e. VMA) for a given gradation (Figure 2 1). Finally DASR porosity can be calculated using Equation 2 1.

PAGE 24

24 Figure 2 1. Mixture components for calculation of DASR porosity (Kim et al., 2006) (2 1) W here, DASR = DASR porosity, V V(DASR) = volume of voids within DASR, V T(DASR) = total volume available for DASR particles, V ICAGG = volume of I C aggregates, VMA = voids in mineral aggregate, V TM = total volume of mixture, V AGG>DASR = volume of particles bigger than DASR. 2.4 Interstitial Component (IC) of Mixture Gradation As illustrated in Figure 1 1, the interstitial component is the material i ncluding asphalt fine aggregates, and air voids that exists within the interstices of the DASR, and volume of this material is considered as the interstitial volume (IV). Research conducted by Guarin (2009) concluded that properties and characteristics of the IC will strongly influence the rutting and cracking resistance of asphalt mixtures. The IC should fill the voids within the aggregates larger than the IC without disrupting the DASR structure. As the DASR IC model assumes that the particles bigger tha n the DASR are floating in the DASR structure, it would be reasonable to accept that

PAGE 25

25 the effect of the DASR voids structure could be utilized to evaluate the total voids structure for the IC including the particles bigger than the DASR. Information on the IC characteristics is fundamental to understand and predict how the IC will fit into the IV and consequently to determine whether the DASR structure would be disrupted by the IC. The characteristics of the IC are expected to have a strong influence on key mixture properties including fracture energy and creep rate, as well as property changes due to aging. Therefore, it was expected that DASR IC parameters would correlate well with the mixture performance, including rutting and cracking. 2.5 Disruption Fact or (DF) A new parameter called the Disruption Factor (DF) was conceived and developed by Guarin (2009) to determine the potential of the finer portion of the mixture s gradation to disrupt the DASR structure. It was shown in laboratory studies that the DF can effectively evaluate the potential of IC particles to disrupt the DASR structure. DF can be calculated using the following equation. (2 2) Guarin also proffered an optimal DF range to attain better rutting and cracking performance of asphalt mixtu re. According to the DF approach, the IC aggregates would not be involved in transmitting load between the DASR aggregates if the DF is low. I n this case, the DASR structure would get no additional support or benefit from the IC particles. In the case of h igh DF, mixture performance would be negatively affected because the DASR structure would be disrupted by the IC aggregates. Lastly, if the DF is in the optimal range, better mixture performance would be expected because the IC aggregates will be involved in resisting shear stresses with the DASR structure.

PAGE 26

26 Therefore, it is expected that the DF will appear to be one good indicator to describe the IC characteristics with respect to the volumetric distribution of IC particles, and a link between the DF and ma terial properties which are related to the performance of asphalt mixtures. Figure 2 2 is a pictorial representation of the different configurations of DF values: low, optimal, and high. (a) DF < Optimal DF range (b) DF = Optimal DF range (c) DF > Opt imal DF range Figure 2 2. Configuration of different DF values (Guarin, 2009) 2.6 Effective Film Thickness (EFT) The film thickness of asphalt mixtures has been used to help explain aging phenomena, and many researchers have attempted to evaluate the relat ionship between film thickness and mixture performance. Kandhal and Chakraborty (1996) have shown that this parameter can be utilized as an indicator to characterize the durability and fatigue resistance of asphalt mixtures. However, it is still controvers ial with regard to its application in the mix design of HMA. More importantly, Superpave system does not have any requirements or guidelines regarding film thickness. Typically, apparent film thickness (or theoretical film thickness TF ), which is calculat ed by dividing the effective binder volume by the surface area of the aggregates, has been used for film thickness analysis. However, many researchers have questioned the relevance of this concept because it may not represent the distribution of binder in

PAGE 27

27 the mixture. Nukunya et al. (2001) introduced a new concept of effective film thickness (EFT), which can be calculated by using the effective volumetric properties of asphalt mixture. They concluded that the effective volumetric properties including the EF T seem to effectively evaluate aging effects and correlated well with mixture properties. In this study, the EFT was selected to act as a surrogate to stiffening effect of interstitial component on mixture Figure 2 3 is a pictorial representation of the d ifference between EFT and TF. The EFT can be calculated by using the following equation. (2 3) Where, V be = effective volume of asphalt binder, SA = surface area of fine aggregate, W T = total weight of mixture, PF AGG = percent fine aggregate by mass of total mixture, P b = asphalt content percent by mass of total mixture, Abs = absorption, P AGG = percent aggregate by mass of total mixture, G b = Specific gravity of asphalt binder. (a) Effective Film Thickness (b) Theoretical Film Thickness Figure 2 3. Effective film thickness vs. theoretical film t hickness

PAGE 28

28 Adequate interstitial volume is important for mixtures to have sufficient strain tolerance, which can be controlled by having an acceptable range of effective film thickness (EFT). EFT of asphalt m ixtures is related to the stiffening effect of IC on mixture, and t he fineness of the IC aggregates is the prim ary factor to control the EFT. In this study, the parameter EFT was evaluated and used to establish a set of performance related design criteria and predictive relationship s for fracture properties of asphalt mixtures. It was expected that the EFT is associated with the time dependent response and brittleness of asphalt mixtures. For example, higher EFT results in higher creep rate and higher fract ure energy. Figure 2 4 shows a schematic that conceptually illustrates how EFT affects mixture properties for two cases which have same component materials. Note: Then A > B therefore, E A < E B Figure 2 4. Conceptual drawing of film thickness effect The white color portion of Figure 2 4 show s the asphalt binder part, while the gray color portion represents the aggregates. In the cas e of thicker EFT represented by c ase 1, material will tolerate higher strain (i.e. less brittle) than the thinner EFT (case 2) and it

PAGE 29

29 will generally be broken by micro damage development with high strain at low stress level. However in the case of thinner EFT described by case 2, material wi ll exhibit less strain tolerance and failed in a brittle manner with low strain at high stress (local stress) level Therefore, mixtures should have an acceptable range of EFT for adequate strain tolerance and EFT can be controlled by limiting fineness of fine aggregate portion (i.e. IC) of mixture s gradation 2.7 Ratio between Coarse Portion and Fine Portion of Fine Aggregates (CFA/FFA) Preliminary analyses indicated that the fineness of the fine aggregate portion of the interstitial components was stro ngly related to effective film thickness. However, EFT does not reflect the effect of particle interaction within the IC, which could be one important factor for IC characterization. Therefore, a new parameter CFA/FFA, which is the ratio between the coarse portion and fine portion of the IC particles, was introduced to characterize the structure of the IC of mixture s gradation. Figure 2 5. Determination of CFA/FFA

PAGE 30

30 In this study, CFA/FFA was used as an indicator to represent the fineness and aggregate st ructure of the IC. It was hypothesized that CFA/FFA was related to the creep response or time dependent response of asphalt mixture. Figure 2 5 describes the basic principle of determining the CFA/FFA. 2.8 Summary For DASR IC characterization, two existing parameters including DASR porosity and DF, and two more parameters including EFT and CFA/FFA were newly added, especially for further IC characterization. Finally, f our parameters identified including DASR por osity, DF, EFT, and CFA/FFA have formed the DA SR IC criteria to effectively address the two primary components of asphalt mixtures (i.e. both coarse (DASR) and fine ( IC) portions of mixture gradation and resulting volumetric properties ) that play a major role on properties and performance These param eters (i.e. four DASR IC parameters) were used for evaluation conducted in this study.

PAGE 31

31 CHAPTER 3 IMPLEMENTATION OF BI NDER AND MIXTURE TES TS ON FIELD CORES FO R SUPERPAVE MIXTURES I N FLORIDA 3.1 Background Binder and mixture tests on field cores were cond ucted to determine binder and mixture properties for different Superpave projects in Florida. The information obtained was used to establish reasonable and effective mixture design guidelines and criteria, performance related laboratory properties, and par ameters, and predictive mixture property relationships. All binder tests were performed according to FDOT test methods, and HMA fracture mechanics model was used to analyze mixture test results. 3.2 Binder Recovery and Binder Tests Asphalt binder recoverie s and binder tests were conducted for cores obtained from different Superpave projects. Binder tests, including penetration test at 25 C viscosity test at 60 C dynamic shear rheometer (DSR) test, bending beam rheometer (BBR) test, and multiple stress c reep recovery (MSCR) test, were performed in this study. The binder testing plan is summarized below, and Table 3 1 represents the asphalt binder used on the Superpave projects evaluated. Penetration test at 25 C (77 F) Viscosity test at 60 C (140 F) D ynamic shear rheometer (DSR) test at 25 C (77 F) Bending beam rheometer (BBR) test at 12 C (10.4 F) Multiple stress creep recovery (MSCR) test at 64 C (147.2 F) Table 3 1. Asphalt binder used for Superpave projects evaluated Project 1 2 3 4 6 7 8 9 10 11 12 Layer A PG 67 22 PG 64 22 PG 67 22 PG 67 22 PG 64 22 PG 64 22 PG 76 22 ARB 5 ARB 5 PG 76 22 ARB 5 Layer B PG 67 22 PG 64 22 PG 67 22 PG 67 22 N/A PG 64 22 PG 76 22 PG 64 22 PG 64 22 PG 64 22 PG 64 22 Note: N/A = N ot Applicable

PAGE 32

32 3.2.1 Binder Recovery Asphalt recovery was performed by using the solvent extraction method for cut cores obtained from the different Superpave projects, including Superpave top and bottom layers which were denoted as layer A and B, respectiv ely. Trichloroethylene (TCE) was used as a solvent for binder recovery and the test procedure was carefully followed to minimize any additional aging of the binder during the binder recovery operation according to FDOT test methods. 3.2.2 Penetration Test The penetration test is one of the oldest and simplest empirical tests used to measure the consistency of asphalt binder. In general, penetration test is performed at 25 C which is considered approximately representative value of average service temperature for asphalt pavement. The depth of penetration is measured in units of 0.1 mm and reported in penetration units For example, if the penetrat ion depth of the needle is 8 mm, the penetration number of asphalt binder is 80 The description and practice of standard penetration test method is designated and reported in AASHTO T 49 and ASTM D 5. Penetration tests were conducted at 25 C. Figure 3 1 represents penetration tes t results from binder recovered for the Superpave projects evaluated Results show that penetration measured for binder extracted from top layer denoted as layer A generally has lower value than for binder obtained from bottom layer denoted as layer B. Thi s was expected because the effect of oxidative aging for top layer is generally more severe than bottom layer. Binders obtained from top layer of Project 9 and 10, which were rubber modified binder (ARB 5) exhibited especially lower penetration.

PAGE 33

33 Note: (C ) = Cracked, (U) = Uncracked Figure 3 1. Penetration test results for Superpave projects 3.2.3 Viscosity Test Viscosity represents the resistance to flow of a fluid and it can be simply defined as the ratio of shear stress to shear rate. As opposed to othe r empirical tests including penetration test, viscosity is a fundamental property. However, viscosity is generally measured at only one temperature, so it does not cover the full range of construction and service conditions. Viscosity test is usually perfo rmed at 60 C which is approximately considered to be representative of the maximum in service surface temperature of asphalt pavement. The description and practice of standard absolute viscosity test method is described in AASHTO T 202 and ASTM D 2171. Fi gure 3 2 exhibits current viscosity measured from extracted binder and Figure 3 3 shows the change in viscosity over time for the Superpave projects evaluated

PAGE 34

34 Note: (C) = Cracked, (U) = Uncracked Figure 3 2 Viscosity t est r esults for Superpave p rojects (a) Layer A

PAGE 35

35 (b) Layer B Note: (C) = Cracked, (U) = Uncracked Figure 3 3. Change in viscosity with aging for Superpave projects Due to more severe effect of oxidative aging caused by higher surface temperature, the top layer showed higher viscosity as wel l as higher rate of increase in viscosity than the bottom layer. Specifically, top layer (Layer A) of Project 8 through 12 which included polymer modified (PG 76 22) and rubber modified binder (ARB 5) sections indicated higher viscosity with around six to nine years of aging in the field. Also, as indicated in Figure 3 3 (a), these sections showed higher rate of increase in viscosity with aging. 3. 2.4 Dynamic Shear Rheometer Test (DSR) The dynamic shear r heometer (DSR) test is used in the Superpave system to characterize the viscous and elastic behavior of asphalt binder at intermediate and high service temperatures. The DSR measures the complex shear modulus G and p hase angle of asphalt binder to determine the characteristics of elastic and viscous components at pavement service temperatures. Specifically, G and measured are

PAGE 36

36 utilized as the indicat ors to predict two HMA distresse s : rutting and fatigue cracking. The descrip tion and practice of standard DSR test method is designated and reported in AASHTO TP 5. In the Superpave asphalt binder specification, two parameters have been chosen (G /sin and G sin ) for evaluation of rutting and fatigue cracking, respectively. S in ce the Superpav e project s investigated have six to eleven years of service period from the construction, all recovered binders obtained were considered as PAV aged binders. As the DSR test for PAV aged binder, samples were tested by using 8mm spindle at in termediate temperature determined based on the PG grade of original binder used. Figure 3 4 represents the parameter G sin for all Superpave projects evaluated Note: (C) = Cracked, (U) = Uncracked Figure 3 4. G sin 10 rad/sec at 25 C (77 F) for Superpave projects Figure 3 4 shows that all binders met the Superpave specification requirement for a maximum G sin o f 5000 kPa except for the top layer of Project 9 (ARB 5) and the

PAGE 37

37 top and bottom layer of Project 10 (Top: ARB 5. Bottom: PG 64 22). G sin is typically considered as an indicator of resistance to fatigue cracking because it indicates an amount of energy d issipated meaning that higher G sin is related to higher energy loss. However, based on the results shown in Figure 3 4, and considering the cracking performance, it appeared questionable whether the parameter G sin was consistently correlated with cra cking performance of mixtures. 3. 2.5 Bending Beam Rheometer Test (BBR) The bending beam rheometer (BBR) test is used in the Superpave system to determine the propensity of asphalt binders to thermal cracking at low temperatures. The BBR calculates the cree p stiffness of asphalt binder ( S(t) ) and the rate of change of the stiffness (m value). The creep stiffness ( S(t) ) is related to the thermal stresses developed in the HMA pavement as a result of thermal contraction, while the slope of the stiffness curve, m value, is associated with the ability of HMA pavement to relieve thermal stresses. In other words, m value is an indicator of the binder s ability to relax stresses by asphalt binder flow. The Superpave binder specification requires a maximum limit of cr eep stiffness and the minimum limit of m value. The description and practice of standard BB R test method is designated and reported in AASHTO TP 1 T he BBR tests for PAV aged binder samples were tested at PG grade temperature according to their original sp ecification. Figures 3 5 and 3 6 represent the parameters S(t) and m value as a result of the BBR testing for all Superpave project sections, respectively. Figure 3 5 shows that all binders met the Superpave specification requirement for a maximum S(t) of 300 MPa. Figure 3 6 indicates that all binders also met the Superpave specification requirement for a minimum m value of 0.3 except for the top layers of Project 9 (ARB 5) and Project 10 (ARB 5).

PAGE 38

38 Note: (C) = Cracked, (U) = Uncracked Figure 3 5. S(t), 60 seconds loading time at 12 C (10.4 F) for Superpave projects Note: (C) = Cracked, (U) = Uncracked Figure 3 6 m value, 60 seconds loading time at 12 C (10.4 F) for Superpave projects The BBR test results including S(t) and m value are typically eva luated to determine the propensity of binder for thermal cracking. However, based on the results

PAGE 39

39 shown by Figure 3 5 and 3 6, it appeared also questionable whether the parameters S(t) and m value were consistently correlated with cracking performance of mi xtures. 3. 2.6 Multiple Stress Creep Recovery Test (MSCR) The multiple stress creep recovery (MSCR) test is used to identify the presence of elastic response in the asphalt binder and the change of elastic response under shear creep and recovery using two d ifferent stress levels at a specified temperature. In general, the percent recovery of asphalt binders in the MSCR test is affected by the type and amount of polymer used in the polymer modified asphalt binder. Thus, it can be used as an indicator for dete rmining whether polymer was utilized. In addition, non recoverable creep compliance has been used as an indicator of the asphalt binder s resistance to permanent deformation under repeated load. D Angelo et al. (2009 2010 ) found that rutting is typically reduced by half as the non recoverable creep compliance is reduced by half. The description and practice of standard MSCR test method is desig nated and reported in AASHTO TP 70 07 and ASTM D 7405 The MSCR test w as conduct ed by u sing an 8 mm spindle at th e environmental grade temperature (64 C) for the State of Florida. Figure 3 7 and 3 8 represent the MSCR test results including average recovery and non recoverable compliance for all Superpave project sections, respectively. Figure 3 7 clearly shows that MSCR average percent recovery can distinguish the presence of polymers in asphalt binders. In general, percent recovery of polymer modified binders wa s greater than bas e binders including PG 64 22 and PG 6 7 22 for both stress levels. Rubber modified binde rs also show ed relatively high percent recovery than base binders. Since the percent recovery indicates the elastic response of asphalt binder,

PAGE 40

40 polymer modified binders (PG 76 22) appear to exhibit higher elastic response and less sensitivity to change of stress level. (a) Layer A (b) Layer B Figure 3 7 MSCR average recovery at 64 C (147.2 F) for Superpave projects

PAGE 41

41 (a) Layer A (b) Layer B Figure 3 8 MSCR nonrecoverable compliance at 64 C (147.2 F) for Superpave projects Based on Figure 3 8, polymer and rubber modified binders normally show ed lower non recoverable compliance than base binders for both stress levels. According to

PAGE 42

42 D Angelo et al. (2009), nonrecoverable compliance can be used for evaluating the rutting resistance of asphalt binder. Howev er, on the basis of t he results analyzed, it seem ed questionable whether it is consistently correlated with rutting performance of mixtures in the field. 3.3 Mixture Tests Superpave IDT tests were performed on field cores obtained from the Superpave proje cts evaluated to determine mixture properties including modulus, creep compliance, strength, failure strain, and fracture energy and to identify the change in key mixture properties as a function of age in the field. Tests were performed at 10 C and 20 C 3.3.1 Test Specimen Preparation Specimens were prepared for laboratory testing using field cores obtained from Superpave projects evaluated. Specific gravity (G mb ) test was conducted on each cut cores and air voids were calculated using the G mb and origi nal (first time of coring ) maximum specific gravity (G mm ). It should be noted that G mm could change with time, especially for moisture damaged projects. For moisture damaged projects, air voids determined using original G mm are probably conservatively low (i.e. true air voids of moisture damaged projects are likely higher than air voids calculated using original G mm ). C ores of similar air voids were grouped for s tandard Superpave IDT tests. 3.3.1.1 Measuring, Cataloguing, and Inspecting Each core obtained w as cleaned and the layer of each different asphalt mixture was properly identified, measured, and catalogued with appropriate markings to prevent any confusion. For quality control purposes, cores were inspected and compared to construction information to verify the presence of different mixtures and thicknesses. Figure 3 9 shows the measuring, cataloguing, and inspecting work for field cores.

PAGE 43

43 Figure 3 9 Measuring, cataloguing, and inspecting work for field cores 3.3.1.2 Cutting Once the data was proper ly logged and verified, the core was sliced to obtain test specimens for Superpave top and bottom layers for testing purposes. A cutting device, which has a diamond cutting saw and a special attachment to hold the cores, was used to slice the cores into sp ecimens of desired thickness. Because the saw uses water to keep the blade wet, the cut specimens were placed in the dehumidifier for at least two days (i.e. 48 hours) to negate the moisture effects in testing. Figure 3 1 0 represents the cut specimens prep ared for Superpave IDT tests and Figure 3 11 shows the cutting machine used i n this study Figure 3 10. Cut s pecimens for Superpave IDT t es t s

PAGE 44

44 Figure 3 11 Cutting machine used in this study 3.3.1.3 Gage Points Attachment Gage points were attached to th e specimens using a steel template, a vacuum pump setup, and a strong adhesive. Four gage points (5/16 inch diameter by 1/8 inch thick) were placed with epoxy on each side of the specimens at distance of 19 mm (0.75 in.) from the center, along the vertical and horizontal axes. Figure 3 12 shows the gage point attachment procedure. Figure 3 12 Gage points attachment

PAGE 45

45 During this process, the loading axis previously marked on the specimens was checked and clarified. This procedure helped for the placement o f specimen in the testing chamber and assured proper loading of the specimen. 3.3.2 Test Procedure One set of Superpave IDT tests including resilient modulus, creep compliance, and strength test were performed on each specimen for the Superpave projects ev aluated to determine modulus, creep compliance, strength, failure strain, and fracture energy at 10 C and 20 C These test results provide the properties to identify changes in key mixture properties over time with aging environment in the field. In addi tion, as it mentioned previously, this information was also critical to identify material properties and prediction model evaluation, and to calibrate and validate the pavement performance pr ediction model The material testing system (MTS) used for this s tudy, and test configuration of Superpave IDT test set up are shown in Figure 3 13. Figure 3 13 Superpave IDT tests 3.3.2.1 Resilient Modulus Test The resilient modulus is defined as the ratio of the applied stress to the recoverable strain when repeat ed loads are applied. The test was conducted according

PAGE 46

46 to the system developed by Roque et al (1997) to determine the resilient modulus and the The resilient modulus test was performed in a load control led mode by applying a repeated have rsine waveform load to the specimen for a 0.1 second followed by a rest period of 0.9 seconds. The load was selected to keep the horizontal resilient deformations within the linear viscoelastic range, wh ere h orizontal deformations are typically between 1 0 0 to 180 micro inches during the test The resilient modulus and Pois can be calculated by the following equations, which were developed based on three dimensional finite element analysis by Roque and Buttlar (1992). The equation is in corporated in the Superpave Indirect Tensi on Test at Low Temperatures (ITLT) computer program, which was developed by Roque et al (1997) (3 1) (3 2) W here, M R = R esilient modulus P = M aximum load GL = G age length H = H orizontal deformation t = T hickness D = D iameter C cmpl = 0.6354 (X/Y) 1 0.332 = Poisson s ratio and (X/Y) = R atio of horizontal to vertical deformation 3.3.2.2 Creep Test Creep compliance is a function of time dependent strain over stress The creep compliance curve was originally developed to predict thermally induced stress in asphalt pavement. However, it can also be used to evaluate the rate of damage accumulation of asphalt mixture. As shown in Figure 3 14 D 0 D 1 and m value are mix ture parameters obtained from creep compliance tests. Although D 1 and m value are related to each

PAGE 47

47 other, D 1 is more related to the initial portion of the creep compliance curve, while m value is more related to the longer term portion of the creep complian ce curve. Figure 3 14 Power model of creep compliance The creep test was performed in the load controlled mode by applying a static load in the form of a step function to the specimen and then holding it for 1000 seconds. The magnitude of load applied w as selected to maintain the accumulated horizontal deformations in the linear viscoelastic range, which is below the total horizontal

PAGE 48

48 deformation of 750 micro inches. Although the range of horizontal deformation at 100 seconds can vary depending upon test temperature, specimen type, and the level of aging a horizontal deformation of 100 to 130 micro inches at 100 seconds is generally considered to be acceptable. The Superpave Indirect Tensi on Test at Low Temperatures (ITLT) computer program was used to det ermine cr eep properties of the mixtures by analyzing the load and deformation data. computed by the following equations. (3 3) (3 4) W here, D(t) = C reep compliance at time t ( 1/psi ), H, t, D, C cmpl GL, P, and (X/Y) are same as described above 3.3.2.3 Tensile Strength Test Failure limits including tensile strength, failure strain, and fracture energy were determined from strength test. These properties can be used for estimating t he cracking resistance of the asphalt mixtures. The strength test was conducted in a displacement control led mode by applying a constant rate of displacement of 50 mm/min until the specimen failed. The maximum tensile strength is calculated as the followin g equation. (3 5) W here, S t = M aximum indirec t tensile strength P = F ailure load at first crack C sx = 0.948 0.01114 (b/D) 0.2693 +1.436(b/D) b = T hickness D = D iameter and = Poisson s ratio

PAGE 49

49 F racture energy and dissipated creep strain energy can be determined f rom the strength te st and the resilient modulus test. Fracture energy is the total energy necessary to induce fracture. Dissipated creep strain energy (DCSE) is the absorbed energy that damages the specimen, and dissipated creep strain energy to failure is the absorbed energ y to fracture (DCSE f ). As shown in the Figure 3 15 fracture energy and DCSE f can be determined as described below. The ITLT program also calculates fracture energy automatically. (3 6) (3 7) (3 8) (3 9) Where, S t = Tensile strength, and f = Failure strain. Figure 3 15 Determination of f racture e nergy and d issipated c reep s train e nergy to f ailure

PAGE 50

50 In addition, a parameter, Energy Ratio (ER), which represents the asphalt potential for top down cracking was recently developed by Roque et al. (200 4 ). This parameter allows the evaluation of cracking performance on different pavement structures by incorporating the effects of mixture properties and pavement structural characteristics. The ener gy ratio is expressed in the e quation shown below The ITLT program calculates energy ratio automatically. (3 10) W here, DCSE f = D issipated c reep s train e nergy (in KJ/m 3 ) DCSE min = M inimum d issipated c reep s train e nergy for adequate cracking performance (in KJ/m 3 ) D 1 and m = C reep parameters 3.1 (6.36 S t ) + 2.46 10 8 in whic h = T ensile stress of asphalt layer (in psi) and S t = T ensile strength (in MPa) 3.3.3 Superpave IDT Test Results Table 3 2 summarizes the mixture information for the Superpave projects evaluated. Table 3 2 Mixture information for 11 Superpave project s Project (UF) ID Year Aged Binder Type Mixture Type Traffic Level Top Bottom Top Bottom 1 11 PG 67 22 PG 67 22 9.5C 19.0C D/5 2 11 PG 64 22 PG 64 22 12.5C 19.0C D/5 3 11 PG 67 22 PG 67 22 12.5C 19.0C D/5 4 11 PG 67 22 PG 67 22 9.5C 19.0C E/6 6 12 PG 64 22 N/A 12.5F N/A C/4 7 12 PG 64 22 PG 64 22 12.5F 12.5F C/4 8 9 PG 76 22 PG 76 22 12.5C 12.5C D/5 9 7 ARB 5 PG 64 22 FC 6 12.5F C/4 10 7 ARB 5 PG 64 22 FC 6 12.5F B/4 11 6 PG 76 22 PG 64 22 12.5C 12.5C E/6 12 6 ARB 5 PG 64 22 FC 6 12.5F C/4

PAGE 51

51 The test results obtained from the Superpave IDT were analyzed using the ITLT computer program developed at the University of Florida. A comprehensive analysis of test results was conducted to identify the trend of changes in key mixture properties includi ng fracture energy, creep rate, resilient modulus, creep compliance, tensile strength, and failure strain as a function of age and environment for different Superpave mixtures. Results of resilient modulus (M R ) which is a measure of elastic stiffness ar e presented in Figure 3 16 These include initial and current values of resilient modulus obtained from field cores indicating the trend in resilient modulus over time for the Superpave projects. For most cases, resilient modulus decreased over time, which clearly indicates the presence of permanent damage and the existence of incomplete healing beyond after some level of aging. The top layer (Layer A) generally exhibited higher rates of reduction in resilient modulus than the bottom layer (Layer B). This r eflects that the effect of permanent damage induced by traffic load is more severe for top layer than bottom layer. (a) Layer A

PAGE 52

52 (b) Layer B Figure 3 16 Change in resilient modulus over time Creep complian ce results are shown in Figure 3 17 Creep compliance is related to the ability of a mixture to relax stresses. In general, higher creep compliance indicates that mixtures can relax stresses faster than mixtures with lower creep compliance, which is critical for evaluating thermal stresses. However, higher cr eep compliance may also be an indication of permanent damage, and the reduction in creep compliance is expected if there is no permanent damage effect. (a) Layer A

PAGE 53

53 (b) Layer B Figure 3 17 Change in creep compliance over time Creep rate, or the rate of c hange of creep compliance, is related to rate of damage. Figure 3 1 8 shows the creep rate results. For mixtures not affected by moisture damage (Project 1 through 8), creep rate of the top and bottom layers generally decreased over time, which indicates th at oxidative aging had a predominant effect on change in creep rate. (a) Layer A

PAGE 54

54 (b) Layer B Figure 3 18. Change in creep rate over time However, for mixtures affected by moisture damage (Project 9 through 12), three cases (Layer B of Project 9, 10, and 1 2) show ed clear increase as well as two cases (Layer A of Project 9 and 11) exhibit ed slight increase of creep rate over time as opposed to the effect of oxidative aging, which indicates the effect of non healable permanent damage induced by moisture. The other three cases exhibit ed clear decrease of creep rate over time. Tensile strength indicates the maximum tensile stress that the mixture can sustain before fracture. Figure 3 19 shows the tensile strength results, which exhibit a similar trend as the res ults of resilient modulus. It was also determined that tensile strength decreased over time, indicating the presence of permanent damage and the existence of incomplete healing after a certain level of aging. The top and bottom layers of Project 8 exhibit ed a lower rate of reduction in tensile strength which appears to be related to the effect of polymer modification. However, the top layer of Project 11, which also used a polymer modified binder, exhibited an

PAGE 55

55 unusually high rate of reduction in tensile st rength over time, which seemed to be associated with the effect of moisture damage. (a) Layer A (b) Layer B Figure 3 19. Change in tensile strength over time Failure strain characterizes the brittleness of a mixture. This value is related to the severity of aging condition and the mixture susceptibility to aging, especially oxidative

PAGE 56

56 aging. Figure 3 20 shows initial (less than six months after construction) and current failure strain for the Superpave projects evaluated. As expected, the rate of reduction in failure strain for top layer was generally greater than for the bottom layer. The top and bottom layers of Project 2 and the top layer of Project 6 exhibited the highest rate of reduction in fai lure strain of sections 1 to 8. (a) Layer A (b) Layer B Fi gure 3 20. Change in failure strain over time

PAGE 57

57 Projects 9 through 12, which showed evidence of moisture damage, also exhibited a high reduction in failure strain. High initial and current air voids as well as the increase in air voids over time caused by th e moisture damage for these sections may have accelerated the effect of oxidative aging, so the mixture embrittled within a relatively short period of time. (a) Layer A (b) Layer B Figure 3 21 Change in fracture energy over time

PAGE 58

58 Fracture energy reflects the mixture s resistance to damage without fracturing. It has been identified as a good indicator of cracking performance of asphalt pavements having similar pavement structure, traffic and environmental condition. Fracture energy results for the eleven Su perpave projects are presented in Figure 3 21. As expected, fracture energy has decreased over time. This observation was the basis for the fracture energy aging model introduced in the NCHRP Project 01 42A. Based on the results shown in Figure 3 21, it se emed clear that higher initial FE results in higher rate of reduction in FE with aging. Also, the top layers exhibited relatively higher rates of reduction in FE than the bottom layers for most projects. However, the rate of reduction in FE for projects sh owing evidence of moisture damage (Project 9, 10, 11 and 12) exhibited unusually high rates of reduction in FE regardless of the initial FE magnitude and layer depth. 3.3.4 Moisture Damaged Projects Moisture damage of asphalt mixture is a major distress mo de that can result in significant costs for repair and rehabilitation. The effect of moisture on asphalt mixture involves various factors acting simultaneously including the effect of moisture susceptibility of asphalt mixture, stresses induced by traffic load, environmental condition, and moisture. Many researchers have tried to identify relationships between asphalt mixture properties and moisture (Schmidt et al. 1972, Fwa et al. 1995 and Lottman 198 6). However, the mechanism and effect of moisture damag e have not yet been fully identified or verified. During the inspection process of the cores obtained, evidence of moisture damage was visually identified in the form of stripping for Project 9, 10, 11, and 12. Stripping was particularly prominent at th e interface between top and bottom Superpave layers. These

PAGE 59

59 projects are relatively new pavements (six to seven years of age). Specifically, as shown in Table 3 3 use of granite aggregates was common to all moisture damaged projects. No moisture damage was observed for projects produced with limestone aggregate. Three of the four projects (Project s 9, 10, and 12) had fine graded mixtures with rubber modified binder (ARB 5) in the top layer and PG 64 22 binder in the bottom layer. The fourth project had a co arse graded mixture with SBS modified binder in the top layer and PG 64 22 binder in the bottom layer. Table 3 3 summarizes project information and the moisture damaged sections are highlighted. Table 3 3. Project information for moisture damaged sections Project (UF) ID Year Aged Aggregate Type Binder Type Mixture Type Traffic Level Top Bottom Top Bottom 1 11 Granite PG 67 22 PG 67 22 9.5C 19.0C D/5 2 11 Granite PG 64 22 PG 64 22 12.5C 19.0C D/5 3 11 Limestone PG 67 22 PG 67 22 12.5C 19.0C D/5 4 1 1 Limestone PG 67 22 PG 67 22 9.5C 19.0C E/6 6 12 Limestone PG 64 22 N/A 12.5F N/A C/4 7 12 Limestone PG 64 22 PG 64 22 12.5F 12.5F C/4 8 9 Limestone PG 76 22 PG 76 22 12.5C 12.5C D/5 9 7 Granite ARB 5 PG 64 22 FC 6 12.5F C/4 10 7 Granite ARB 5 PG 64 22 FC 6 12.5F B/4 11 6 Granite PG 76 22 PG 64 22 12.5C 12.5C E/6 Note: N/A = Not Applicable Several unique trends were identified for moisture damaged sections with regard to the change in fracture energy and air voids over time. Relatively high initial and/or current air voids were measured on field cores obtained from layer A for Project 9 and 10. In some cases (Layer A of Project 9 and 11 and layer B of Project 9, 11, and 12), the air voids increased over time which appear ed to be related to the disp lacement of material caused by moisture damage Figure 3 22 shows the change in air voids for Superpave projects evaluated including layers A and B, respectively.

PAGE 60

60 (a) Layer A (b) Layer B Note: WP denotes the Wheel Path Figure 3 22. Change in air voids o ver time In addition, as shown in Figure 3 23 much greater rate of reduction in normalized fracture energy was obtained for moisture damaged sections. Rate of reduction of normalized fracture energy over time was calculated to account for the difference i n age between moisture damaged sections and other sections. Figure 3 23 shows the initial

PAGE 61

61 rate of reduction in normalized fracture energy for Superpave projects evaluated including layer A and B, respectively. (a) Layer A (b) Layer B Note: dFE n (t)/dt at t =0 denotes the initial rate of reduction in normalized fracture energy Figure 3 23. Initial rate of reduction in normalized fracture energy over time

PAGE 62

62 Birgisson et al. (2004) indicated that the Energy Ratio (ER) can be used to evaluate the effect of moistur e damage on changes in fracture resistance of asphalt mixtures. (a) Layer A (b) Layer B Note: dER n (t)/dt at t=0 denotes the initial rate of reduction in normalized energy ratio Figure 3 24 Initial rate of reduction in normalized energy ratio over time

PAGE 63

63 Fi gure 3 24 clearly shows that a much greater reduction in normalized ER was observed in moisture damaged sections. In other words, the effect of moisture dramatically reduced the fracture resistance of asphalt mixtures, and the ER was capable of detecting t he effect of moisture damage. Figure 3 24 clearly indicates that the ER is very sensitive to, and therefore able to capture the effects of moisture damage. As expected, high rate of reduction in normalized ER with aging was identified for moisture damaged sections. 3.4 Summary Binder and mixture tests were performed on field cores to determine key binder and mixture properties to identify the change in these properties as a function of age and environment in the field for different Superpave projects in Flo rida. Test results were carefully analyzed and further used to develop implementable mixture design criteria (i.e. DASR IC criteria), predictive mixture property relationships, and material property prediction model evaluation.

PAGE 64

64 CHAPTER 4 EVALUATION OF FIE LD MIXTURE PERFORMAN CE USING DASR IC MODEL PARAMETERS 4 .1 Background Four DASR IC model parameters including DASR porosity, DF, EFT, and CFA/FFA have formed the DASR IC criteria to effectively address the two primary components (i.e. DASR and IC) of asphal t mixtures that play a major role on properties and performance. Field p erformance evaluation of different Superpave mixtures was conducted to identify the relationships between the four DASR IC parameters and field performance 4 .2 Implementation of Grada tion Analysis for Superpave Mixtures Gradation analysis was conducted for different Superpave mixtures evaluated as part of Superpave monitoring pro jects sponsored by FDOT. Table 4 1 summarizes mixture information of Superpave projects analyzed. The actual reliability level for all binder s true grades is 98 % for the seven day average high temperature and the one day low temperature. In this study, extensive sampling was made by taking field cores from different Superpave mixtures constructed throughout th e state of Florida. It is noted that in place gradations were determined from field cores. These gradations are not simply job mix formula (JMF) gradations that may or may not be representative of the final results in the field. In addition, asphalt binder recovery was performed by using the solvent extraction method, and all mixture volumetric properties required were obtained from cut cores.

PAGE 65

65 Table 4 1. Mixture information of Superpave projects analyzed Project (UF) ID US Route County Aggregate Typ e Binder Type Mixture Type 1 I 10 WB Madison Granite PG 67 22 9.5C 2 I 75 SB Hamilton Granite PG 64 22 12.5C 3 I 75 SB Hamilton Limestone PG 67 22 12.5C 4 I 10 EB Duval Limestone PG 67 22 9.5C 6 US 301 SB Marion Limestone PG 64 22 12.5F 7 FL Turnp ike NB Palm Beach Limestone PG 64 22 12.5F 8 I 10 WB Leon Limestone PG 76 22 12.5C 9 SR 121 SB Alachua Granite ARB 5 FC 6 10 SR 16 EB Bradford Granite ARB 5 FC 6 11 I 295 SB Duval Granite PG 76 22 12.5C 12 SR 73 SB Calhoun Granite ARB 5 FC 6 Note: 1. Mixture Type: C = Coarse M ixtures, F = Fine M ixtures, N/A = Not A pplicable 2. WB = West Bound, SB = South Bound, EB = East Bound, and NB = North Bound All DASR IC parameters including DASR porosity, DF, EFT, and CFA/FFA were determined using th e in place gradations and mixture volumetric properties obtained from field cores for each project. Table 4 2 summarizes all parameters calculated. Figure 4 1 shows the initial values of fracture energy and creep rate, which are key mixture properties to c ontrol performance of asphalt pavement (Zhang et al. 2001), measured from field cores for Superpave projects evaluated. Table 4 2. DASR IC parameters calculated for Superpave projects Project (UF) ID DASR (mm) DASR Porosity (%) DF EFT (Micr ons) CFA/FFA 1 4.75 1.18 48.0 0. 6 4 19.5 0.31 2 4.75 60.2 1.02 37.1 0.60 3 4.75 2.36 43.8 0.52 23.5 0.35 4 4.75 2.36 47.5 0.56 32.5 0.46 6 4.75 1.18 56.2 0.92 13.7 0.29 7 9.5 1.18 50.2 0.86 12.7 0.30 8 4.75 2.36 48.8 0.60 28.3 0.42 9 4.7 5 1.18 51.0 0.69 15.2 0.29 10 9.5 1.18 50.3 0.71 14.4 0.31 11 4.75 1.18 40.6 0.56 24.8 0.39 12 4.75 1.18 61.3 0.76 30.4 0.24

PAGE 66

66 (a) Initial fracture energy (b) Initial creep rate Figure 4 1. Initial fracture energy and creep rate for Superpa ve projects evaluated Detailed descriptions of gradation analysis results for each project are included in the following section, including a brief introduction of material composition used, gradation characteristics, evaluation of DASR IC parameters calcu lated and mixture property characteristics.

PAGE 67

67 4 .3 Evaluation of Gradation Analysis Results for Superpave Projects Project 1 was constructed with 9.5 mm coarse graded Superpave mixture using granite aggregate. PG 67 22 binder was used for this project. Gradat ion analysis results indicated that Project 1 exhibited a continuous gradation pattern with good interaction within the DASR structure. All gradation parameters were within the acceptable range. Mixture test results indicated that initial fracture energy a nd creep rate were in a range associated with good performing mixtures. Project 2 was constructed using 12.5 mm coarse graded Superpave mixture. Granite aggregate was used with PG 64 22 binder. Project 2 exhibited high values of DASR porosity and DF with p oor interaction within the DASR structure. In addition, uncommonly high EFT and CFA/FFA were identified. It appeared that the gradation characteristics mentioned above resulted in the unusually high initial fracture energy and creep rate. Project 3 was con structed with 12.5 mm coarse graded Superpave mixture. Limestone aggregate was used along with PG 67 22 binder. As with Project 1, Project 3 exhibited a continuous gradation pattern with good interaction within the DASR structure. However, DF was relativel y low which led to high creep rate and is probably related to the poor cracking performance observed for Project 3. Project 4 was constructed with 9.5 mm coarse graded Superpave mixture. Limestone aggregate was used with PG 67 22 binder. Project 4 exhibite d an acceptable DASR porosity, but also had relatively high EFT and CFA/FFA with low DF, indicating that the IC aggregates would not be involved in transmitting load between the DASR aggregates. Mixture test results indicated relatively high initial creep rate, which

PAGE 68

68 appeared to be associated with the high EFT and CFA/FFA along with the low DF of Project 4. Project 6 was constructed with 12.5 mm fine graded Superpave mixture. Limestone aggregate was used along with PG 64 22 asphalt binder. Based on the grad ation analysis results, Project 6 had high DASR porosity and relatively high DF, indicating potentially poor mixture performance. In fact, relatively poor field performance for both rutting and cracking was identified based on the PCS data and field invest igation. Project 7 was constructed with 12.5 mm fine graded Superpave mixture. Limestone aggregate was used with PG 64 22 binder. Project 7 exhibited a continuous gradation pattern with good interaction within the DASR structure. However, the DASR porosity was in the marginal range with DF within the acceptable range. Acceptable values of initial fracture energy and creep rate were determined for mixture tests. Project 8 was constructed with 12.5 mm coarse graded Superpave mixture. Limestone aggregate was u sed along with an SBS modified binder. Project 8 exhibited a continuous gradation pattern with good interaction within the DASR structure. However, the DASR porosity was in the marginal range with DF within the acceptable range Project 8 mixture exhibited relatively high EFT and CFA/FFA which may be a ssociated with high damage rate However, mixture test results indicated that initial creep rate was within the range considered to be acceptable. This was probably due to the beneficial effect of polymer modi fication. More details regarding material property relationships will be introduced in Chapter 5

PAGE 69

69 Projects 9 through 12 were categorized as unusual projects from the standpoint of gradation effects and material property relationships because evidence of mo isture damage was observed. Results of gradation analysis for these sect ions are included in the Table 4 2. Also results of initial fracture energy and creep rate are included along with normal projects in Figure 4 1 However, moisture damaged sections ha ve to be dealt with differently from other sections for the performance evaluation. Therefore, moisture damaged sections, Project 9, 10, 11, and 12, were excluded from performance evaluation in this study. 4 .4 Evaluation of Field Mixture Performance Perfor mance evaluation of different Superpave mixtures was conducted to identify effects of gradation characteristics on field performance using the DASR IC model parameters. Field rutting and cracking performance data were collected as part of Superpave monitor ing projects conducted at the University of Florida. Structural deficiencies were not found for Superpave projects evaluated in this study. Based on moduli determined by backcalculation of falling weight deflectometer test (FWD) results obtained at multipl e times during the service life changes in base and subgrade moduli were not significant over time. Actual backcalculated base moduli var ied from 0.30 to 0.52 Gpa (44100 to 74900 psi) which indicate d competent base moduli for asphalt pavement. Gradation analysis results were used as gradation parameters for the evaluation. Based on the results introduced in prior research (Kim et al. 2006, Guarin 2009, Roque et al. 2011 ) and analyses performed in this study, acceptable range s of each parameter w ere identi fied : DASR porosity (%): 38 52 (48 52: Marginal DASR porosity) DF: 0.60 0.90

PAGE 70

70 EFT (microns): 12.5 25.0 CFA/FFA: 0.28 0.36 It is noted that binder properties obtained from field cores still met the Superpave specification requirement regarding six to eleven year old pavement for Superpave projects evaluated (Roque et al. 2011 ). In other words, parameters from binder tests did not appear to be consistently correlated with field performance of Superpave mixtures. Furthermore, b ased on the Quality Con trol ( QC ), Quality Assurance ( QA ), and Independen t Assurance (IA) data, all projects met relevant construction specification criteria including job mix formula (JMF) for gradation, air voids, binder content and in place density. 4 .4.1 Field Performance: Ru tting Comprehensive monitoring of field rutting performance was conducted for Superpave mixtures as part of Superpave monitoring projects sponsored by FDOT. Field rut depth measured from construction throughout the pavement s life using transverse profilog raph was used for the evaluation. The results are presented in terms of rut depth per ESALs (inch/ESALs 10 6 ) to normalize the effect of traffic volume between the different projects. In addition, the rut depth/ESALs at two years values were used for analys is to account for the fact that the rate of rutting generally decreases with time. Careful analysis of rut profiles with time has clearly shown that this phenomenon is associated with the fact that the rut basin continues to widen with continued loading, w hile rut depth does not increase very much. Figure 4 2 shows the rut depth/ESALs at two years for the Superpave projects evaluated. Projects 1, 2 and 6 exhibited relatively high rut depth/ESALs, which indicates poor field rutting performance, compared to o ther projects.

PAGE 71

71 Figure 4 2. Field rutting performance for Superpave projects evaluated Figure 4 3 shows the results of rutting performance evaluation using DASR IC parameters for Superpave projects. Figure 4 3 (a) clearly indicates that mixtures with good DASR porosity exhibited much lower rut depth/ESALs, which reflects good field rutting performance, than mixtures with high DASR porosity (i.e. outside the acceptable range). For mixtures with marginal DASR porosity, it was noted that mixtures could exhibi t either good or bad rutting performance. Based on Figure 4 3 (b), the range of DF appears to be correlated with DASR porosity. Mixtures with DF within the acceptable range exhibited lower rut depth/ESALs, except for Project 1. Project 1 had marginal DASR porosity which appeared to result in relatively bad performance Mixtures with high DF clearly exhibited high rut depth/ESALs. However, Projects 3 and 4 exhibited good rutting performance, even though the DF was relatively low. Gradation analysis results indicated that these two sections had good DASR porosity with good interaction within the DASR structure. Similar trends were also identified for relations between rut depth/ESALs and EFT and

PAGE 72

72 CFA/FFA (see Figures 4 3 (c) and (d)). It is noted that Project 8 exhibited good rutting performance with marginal DASR porosity which was probably related to the beneficial effect of polymer modification. (a) DASR porosity (b) Disruption factor

PAGE 73

73 (c) Effective film thickness (d) CFA/FFA Figure 4 3. Rutting performan ce evaluation using DASR IC parameters In summary, the characteristics of the coarse aggregate structure as reflected by the DASR porosity, is the most important parameter to control field rutting performance. It was observed that good IC characteristics i ncluding DF, EFT, and CFA/FFA could not

PAGE 74

74 overcome the problems associated with a mixture with DASR porosity outside the acceptable range. Therefore, the DASR porosity criteria introduced appeared to provide an effective tool that can accurately distinguish the field rutting performance of Superpave mixtures. 4 .4.2 Field Performance: Cracking Field cracking performance was also monitored for the Superpave projects as part of Superpave monitoring projects sponsored by FDOT. As the indicator of cracking perform ance evaluation, crack initiation time for each project was estimated using the approach based on the information obtained from comprehensive field investigation conducted by the University of Florida research team and the crack rating history data from th e pavement condition survey (PCS) performed by the FDOT (Roque et al. 2011 ) Figure 4 4 indicates the deterministic procedure used to estimate crack initiation time based on PCS da ta for Projects 1 and 2. Appendix A includes the deterministic procedure to estimate crack initi ati on time using crack rating data for al l Superpave projects evaluated. Figure 4 4. Determination of observed crack initiation time for Project 1 and 2

PAGE 75

75 Table 4 3 shows the crack initiation time and cracking status determined for Supe rpave projects evaluated and Figure 4 5 shows the crack initiation time estimated for the Superpave projects evaluated. Projects 1 and 7 exhibited relatively longer crack initiation times, which reflect better field cracking performance than other projects Table 4 3. Crack initiation time and cracking status determined for Superpave projects Project (UF ID) Age (year) PCS based Observed Decision Status t i (year) Status Status t i (year) 1 11 U 16 (P) U U 16 (P) 2 11 C 10 C C 10 3 11 C* 9 U C 9 4 11 U 28 (P) C* C < 11 6 11 C 11 C C 11 7 11 U 18 (P) U U 18 (P) 8 9 C 9 C C 9 Note: 1. U denotes Uncracked and C denotes Cracked 2. P denotes the value determined based on extrapolation 3. denotes the final decision when an incon sistency occurred between our observation at coring time and the PCS data Figure 4 5. Field cracking performance for Superpave projects evaluated Figure 4 6 represents the results of cracking performance evaluation using DASR IC par ameters for Superpav e projects. For cracking performance, IC characteristics

PAGE 76

76 including DF, EFT and CFA/FFA criteria are important. According to Figure 4 6 (a), Projects 1 and 7 exhibit ed relat ively good cracking performance with marginal DASR porosity, while Project 8 show ed early crack initiation time with marginal DASR porosity. (a) DASR porosity (b) Disruption factor

PAGE 77

77 (c) Effective film thickness (d) CFA/FFA Figure 4 6. Cracking performance evaluation using DASR IC parameters However, all three Projects (Project 1, 7 and 8) indicate d good ranges of DF as shown in Figure 4 6 (b). In this case, the EFT and CFA/FFA, which characterize binder

PAGE 78

78 distribution, fineness and aggregate structure of the IC, seem ed to play an important role o n cracking performance. Figure s 4 6 (c) and (d) show th at Projects 1 and 7 exhibit ed good ranges of both EFT and CFA/FFA. Project 8 exhibit ed higher ranges of EFT and CFA/FFA which potentially result ed in poor field cracking performance even though i t ha d a polymer modified binder. It is noted that mixtures sh ould have acceptable ranges of EFT and CFA/FFA for adequate strain tolerance. The EFT and CFA/FFA can be controlled by limiting fineness of the fine aggregate portion. In addition, Figures 4 6 (c) and (d) indicate that mixtures with high EFT and CFA/FFA cl early exhibited shorter crack initiation time, which reflects bad field cracking performance, than mixtures with good EFT and CFA/FFA. However, Project 3 and 6 exhibited bad cracking performance, even though the EFT and CFA/FFA were within the acceptable r anges. Gradation analysis results indicated that these two sections had either low (Project 3) or high (Project 6) DF (see Figure 4 6 (b)). T he effect of bad DF is probably related to poor cracking performance for Projects 3 and 6. On the basis of field cr acking performance evaluation, i t appear ed that the DF criteria should be considered in conjunction with EFT and CFA/FFA criteria to effectively distinguish the field cracking performance. 4 .5 Summary A comprehensive field performance evaluation for both r utting and cracking was conducted using the DASR IC model parameters to identify and verify performance related criteria. Based on results analyzed, it was expected that the introduction of DASR IC criteria as performance related design parameters to curre nt mix design guidelines and specifications will lead to better and more consistent field rutting and cracking performance of Superpave mixtures. In addition, the DASR IC criteria will also

PAGE 79

79 prov ide a more rational method to consider the effect of IC on mix ture behavior which strongly affects HMA fracture properties. Therefore, it is expected that this criteria will have a potential to identify the effect of mixture gradation and volumetric characteristics on properties which is more reliably related to perf ormance of asphalt mixtures.

PAGE 80

80 CHAPTER 5 IDENTIFICATION OF PR EDICTIVE MIXTURE PRO PERTY RELATIONSHIPS AND MODEL DEVELOPMENT 5 .1 Background The current lack of material property models that can accurately describe the changes in material properties o ver time under field conditions is probably the greatest deficiency in our ability to accurately predict pavement performance. Therefore, there is a need to evaluate existing material property models and develop improved models for use in the prediction of pavement performance Previous research has shown that fracture energy, which is associated with mixture s tolerance to damage, and creep rate, which is related to the rate of damage accumulation in the mixture, were key material properties that affect cr acking performance of asphalt pavements (Sedwick 1998, and Zhang et al. 2001). Thusly, two models, part of the Enhanced Hot Mix Asphalt Fracture Mechanics based Model (HMA FM E) developed during the completion of NCHRP Project 01 42A (Roque et al. 2010) an d directly associated with these two material properties were selected for evaluation in this study. These are the AC stif fness (creep compliance) aging sub model and the fracture energy limit aging sub model. Superpave monitoring projects, recently conduc ted at the University of Florida (UF), provided an unique opportunity to have material property data for several pavements throughout their early pavement life cycles, and to evaluate the selected material property models using this historical data. In add ition, four Dominant Aggregate Size Range Interstitial Component (DASR IC) model parameters were identified from the Superpave monitoring projects to be strongly related to cracking performance of asphalt mixtures: DASR porosity, disruption factor (DF), ef fective film thickness (EFT),

PAGE 81

81 and the ratio between the coarse and the fine portions of the fine aggregate (CFA/FFA) (Kim et al. 2006, Guarin 2009, and Roque et al. 2011). In this study, these DASR IC parameters, describing the characteristics of gradation and resulting volumetric properties, were used along with the historical material property data to develop improved predictive relationships for HMA fracture properties. These predictive relationships will help reduce the need for sophisticated laboratory mixture tests, thereby increasing the potential for implementation of more advanced pavement design systems such as the interim mechanistic empirical pavement design guide (MEPDG) recently adopted by American Association of State Highway and Transportatio n Officials (AASHTO). 5 .2 Existing Material Property Models in the HMA FM E Model The two previously mentioned material property models (i.e., the AC stiffness aging model and the fracture energy limit aging model) were re examined using this newly acquire d data. Each model was briefly described, and then used to predict the respective changes in the HMA material properties versus time in service. These predictions were then compared to the historical data obtained from the field sections as part of the Sup erpave monitoring projects. These comparisons were used to improve the existing models, by proposing modifications which increase the accuracy and reduce the error between predicted and observed values. 5.2.1 AC Stiffness Aging Sub Model The AC stiffness aging sub model was developed based on the global binder aging model (Mirza and Witczak, 1995) and the dynamic modulus model (Witczak and Fonseca, 1996) using a loading time of 0.1 seconds. In this model, the following empirical equation was identified to consider the effect of aging on mixture stiffness (S),

PAGE 82

82 (5 1) In this equation |E | 0 represents the unaged mixture stiffness, and t 0 correspond to the aged and unaged binder viscosities at 10 C, respectively. The general trend for the predicted change in mixture stiffness at surface of the pavement as a function of time or age is shown in Figure 5 1. T he stiffness S(t) increa ses with age and its rate of change decreases with time, where S 0 is the initial value and S max is the maximum value after being aged for a sufficiently long time. Figure 5 1 Schematic plot for AC stiffness at surface vs. age Using the AC stiffness agi ng model, creep compliance values were calculated at 1, 2, 5, 10, 20, 50, 100, 200, 500, and 1000 seconds for three temperatures (i.e. 0, 10, and 20 C) by taking the inverse of the AC stiffness values at the corresponding time for each of the temperatures This results in three 1000 second creep compliance curves. These isothermal creep compliance curves were then used to generate a master curve and the creep rate can be obtained from this master curve, as done by Buttlar et al., in 1998. The general trend for the predicted creep rate aging curve in Figure 5 2 show s that the creep rate CR(t) decreases as age increases at a decreasing rate.

PAGE 83

83 Figure 5 2 Schematic plot for creep rate vs. age The predicted trend in Figure 5 1 is generally consistent with obs ervations from the resilient modulus data during the early stages of pavement life. However, it is different from the data obtained at later stages of pavement life, during which it was found that the modulus actually decreases with age (Roque et al., 2011 ). It appears that oxidative aging alone is not sufficient to account for the change in AC stiffness over time. The effects of other factors on AC stiffness, such as load induced damage, moisture related damage, and healing potential, need to be considered for more accurate prediction. When the existing AC stiffness aging model is modified to include these key factors, the creep rate aging curve will be affected due to the dependence of creep rate on AC stiffness. 5.2.2 Fracture Energy Limit Aging Sub Model The fracture energy limit aging sub model, also developed as part of the HMA FM E model, is expressed in the following form. (5 2) Where, FE i is the initial fracture energy of the HMA and FE min is the minimum value of the FE after a sufficiently long aging period t inf FE min was estimated based on

PAGE 84

84 experimental data obtained from field cores to be approximately 0.2 kJ/m 3 and t inf is fixed at 50 years. The exponent k 1 is an aging parameter determined through field calibration (Roque et al., 2010), and S n (t) is the normalized change in stiffness (with respect to its initial value) at the surface of the AC layer, and is expressed as, (5 3) W here, S(t), S 0 and S max were defined when describing Figure 5 1. Figure 5 3 shows tha t S n (t) has the same form as Figure 5 1, and varies between zero and one. Figure 5 3 Schematic plot for normalized change in AC stiffness vs. age The predicted trend for the change in FE limit at the pavement surface as a function of pavement age is pre sented in Figure 5 4. As can be seen, the FE limit decreases with pavement age at a decreasing rate. This prediction generally agrees with field observations from measured FE limit data. However, the effects of load induced damage, moisture related damage, and healing potential on FE limit, which are not considered in the existing model, could affect the change in FE at later stages of pavement life (Roque et al., 2011). Therefore, it may be necessary to take these factors into account to improve the accura cy of the model. It is also noted that initial fracture

PAGE 85

85 energy (FE i ) for the existing model was predicted based on FE measured from field cores. Therefore, whereas the existing model requires at least one FE measurement, development of a model to predict i nitial FE would be of great value. Figure 5 4 Schematic plot for FE limit vs. age 5.3 Key Elements for M aterial Property Relationships The identi fication of appropriate m aterial property relationships of key mixture properties is important to accurately predict pavement performance. As mentioned previously, the current or existing material property models for AC stiffness, creep rate, and fracture energy, as included as part of the HMA FM E model, are capable of predicting changes in these mixture proper ties due only to oxidative aging. These current models have the following deficiencies: Load induced damage, moisture related damage, and healing that could also affect these material properties, especially in the later stages of pavement life were not con sidered. Determination of creep rate, as a function of pavement age, was based on the current AC stiffness aging model using a simple inverse relationship. Consequently, the creep rate predictions were relatively inaccurate. However, the overall trend of t he creep rate aging curve was generally correct. Therefore, measured creep rates obtained from field cores were needed to improve the accuracy of the model s prediction. At least one FE measurement is required to predict initial FE for use in the existing fracture energy aging model.

PAGE 86

86 Therefore, these existing mixture property models can be improved by incorporating these physical and environmental factors. Also, there is a need to develop models to predict initial fracture energy and creep rate, so that acc urate model predictions can be achieved without relying on physical measurements from field cores. As illustrated in Figure 5 5, two of material property relationships were identified and targeted for improvement and/or for additional development to meet these needs, and these are described below. Material property relationship I: This relationship associates or ties mixture gradation characteristics and volumetric properties to initial mixture properties, more specifically, the initial fracture energy and creep rate. Material property relationship II: This relationship improves the models by changing the mixture properties over time, which takes into account the effects of load induced damage, moisture related damage, and mixture healing, in addition to st andard oxidative aging. Figure 5 5 Two material property relationships 5.3.1 DASR IC Model Parameters As part of Superpave monitoring project recently completed at the University of Florida, field performance evaluation of different Superpave mixtures w as conducted to

PAGE 87

87 identify effects of gradation characteristics on field performance using DASR IC model. Field rutting and cracking performance data and the DASR IC model parameters determined based on the gradation analysis of these projects were used for the evaluation. Based on this evaluation, an acceptable range of each DASR IC model parameter was recommended for optimal mixture performance (Roque et al., 2011): DASR porosity (%): 38 52 (48 52: Marginal DASR porosity) DF : 0. 5 0 0.9 5 EFT ( microns ): 12.5 25.0 CFA/FFA: 0.28 0.36 It should be noted that acceptable range of DF was extended based on analyses conducted in this study for the purpose of development of predictive material property relationships. F urther analyses were undertaken to identi fy whether any distinctive relationship or pattern existed between key mixture properties (i.e., fracture energy limit and creep rate) and the DASR IC parameters within the recommended range. 5.3. 2 Initial Material Properties It was clear from the analysis that the initial value of material properties is one of the key elements that control their changes with aging (Roque et al., 2010). Figure 5 6 shows FE limit curves at three different values of initial fracture energy FE i for a constant k 1 value, and di fferentiate the effect of the initial FE magnitude on the FE limit curve as aging continues. As shown, the overall FE limit curves move upward as FE i increases. However, the initial reduction rate of the curves also increases. Therefore, models for initial material properties are key elements needed to further develop models for changes in these material properties (see Figure 5 5 ).

PAGE 88

88 Figure 5 6 FE limit aging curve at different initial FE (k 1 =3 (Roque et al. 2010)) 5.3. 3 Factors Related to Non Healable P ermanent Damage Existing material property models for changes in AC stiffness, fracture energy, and creep rate, included in HMA FM E model, considered oxidative aging as the only factor. However, the trends based on these existing models did not correlate well with the results of previous laboratory and field research (Roque et al., 2007, 2011). In this study, new concepts and modified models were developed to appropriately describe the known trends with respect to the changes in fracture energy and creep r ate with time, by including the effect of non healable permanent damage induced by loading and moisture, in addition to the effect of oxidative aging. 5.4 Development of Predictive Material Property Relationships Identification of mixture parameters that c ontrol mixture performance, including mixture characteristics, component properties/characteristics, and volumetric properties, led to the development of preliminary relationships that predict fundamental mixture properties (i.e. fracture energy and creep rate). DASR IC model parameters, which were introduced to describe the characteristics of gradation and resulting volumetric

PAGE 89

89 properties found to be the most strongly related to rutting and cracking performance of asphalt mixtures, were used to develop init ial property prediction models. In addition, conceptual relationships were identified to describe changes in these properties over time (aging). This can serve as the foundation for further development of improved models to predict mixture properties as a function of age in the field. T he verified relationships will also serve to provide reliable inputs for prediction of service life using pavement performa nce prediction models. Figure 5 7 shows the flowchart for development of predictive material property relationships. Figure 5 7 Flowchart for development of predictive material property relationships 5.4.1 Relationships for Initial Material Properties As mentioned before, initial values of material properties are key elements governing the changes in ma terial properties over time. Therefore, models for initial material properties are important for accurate prediction of their changes over time and for accurate prediction of overall pavement performance. Two relationships for initial

PAGE 90

90 material properties w hich are presented in the sub sections that follow were developed in this study. 5.4.1.1 Initial Fracture Energy Relationship The initial value of fracture energy is one key parameter that governs the trend of FE limit aging curve. Based on the analyses of mixture characteristics and component properties and the results of mixture testing, the relationship between initial fracture energy and DASR IC parameters were identified. Figures 5 8 through 5 10 present the relationships identified between initial fra cture energy and DASR porosity, DF, and EFT, respectively. As shown, the initial fracture energy generally decreases with increasing DASR porosity and disruption factor. Also, it can be seen that the initial fracture energy increases with increasing effect ive film thickness. As a result, it was identified that there are unique relationships between initial fracture energy and each individual DASR IC parameters, which control characteristics of the mixtures, particularly characteristics of the interstitial c omponent (IC) of mixture. It should also be noted that this trend appears to hold only when DASR porosity DF, and EFT are within the acceptable ranges. Figure 5 8 Relationship between initial fracture energy and DASR porosity

PAGE 91

91 Figure 5 9 Relationshi p between initial fracture energy and disruption factor Figure 5 10 Relationship between initial fracture energy and effective film thickness The relationships between initial fracture energy and the three individual parameters are either proportional o r inversely proportional Since there is a proportional relationship between the initial fracture energy and EFT, and inversely proportional relationships between initial fracture energy and DASR porosity and DF, respectively, an empirical relationship was further developed to relate initial fracture energy to all three parameters, which resulted in the following equation.

PAGE 92

92 (5 4) Where, FE i = Initial fracture energy, DASR = DASR porosity, DF = Disruption factor, EFT = Effective film thickness, a = 0.251, b = 0.034, c = 0.039, and d = 0.706. In this equation, a, b, c, and d are fitting parameters determined through linear regression. Figure 5 1 1 shows the initial fracture energy as calculated using the p redictive equation compared to the measure d values from the field cores. As shown, all data points in the figure are close to the line of equality, which indicates that the predictions compare well with the measured values. Figure 5 11 Predicted vs. mea sured initial fracture energy 5.4.1.2 Initial Creep Rate Relationship Creep rate, also known as the rate of creep compliance, is related to the rate of damage, which is considered to be a good indicator for evaluating the cracking performance of asphalt pa vement. Attempts were made to identify the relationship between the initial creep rate and each of the DASR IC parameters: DASR porosity, DF,

PAGE 93

93 and EFT. However, no distinct relationship or pattern was identified from the analysis. In other words, there was no relationship between these three parameters and the initial creep rate (see Figures 5 12 through 5 1 4 ). In these figures, Solid Diamond ( ) denotes unmodified mixtures for which all DASR IC parameters are within the ranges considered to be acceptable, Cross (x) denotes unmodified mixtures for which at least one of the DASR IC parameters is not within the ranges considered to be acceptable, and Square ( ) denotes polymer modified mixtures for which at least one of the DASR IC parameters is not withi n the ranges considered to be acceptable. However, prior research conducted at the University of Florida has shown that creep rate can be strongly influenced by parameters obtained using the DASR IC model, including properties and characteristics of the IC (Kim et al., 2006, Guarin, 2009, and Roque et al., 2011). Therefore, it was expected that parameters other than those discussed above could be identified to uniquely define the initial creep rate relationship. Figure 5 1 2 Relationship between initial c reep rate and DASR porosity

PAGE 94

94 Figure 5 1 3 Relationship between initial creep rate and DF Figure 5 1 4 Relationship between initial creep rate and EFT Preliminary analyses showed that the ratio between coarse and fine portions of fine aggregate (CFA/FFA) which is associated with effective film thickness, was strongly related to initial creep rate of mixture. Therefore, this new parameter representing the composition of interstitial component was used to identify the relationship between initial creep rat e and CFA/FFA presented in Figure 5 1 5 It

PAGE 95

95 appears that CFA/FFA is better correlated to initial creep rate than effective film thickness. This may be explained by the fact that CFA/FFA reflects the effect of particle interaction within the IC, whereas EFT does not. Figure 5 1 5 Relationship between initial creep rate and CFA/FFA It also can be seen fr om Figure 5 15 that the initial creep rate generally increases with increasing CFA/FFA. Similar to the initial fracture energy relationships, it is noted tha t this relationship appears to hold only when CFA/FFA are within the acceptable range. In addition, it was found that the initial creep rate relationship can be applied to mixtures with polymer modified bind er. As shown in Figure 5 16 the continuous line represents the initial creep rate relationship for mixtures with unmodified binder, while the dashed line represents the relationship for mixtures with polymer modified binder. It appears that polymer modification resulted in lower initial creep rate, even when not all DASR IC criteria were met. This observation is consistent with previous finding s from

PAGE 96

96 research that polymer modification generally helps to reduce the damage rate of asphalt mixtures (Kim et al., 2003). Figure 5 16 Effect of polymer modifi cation on relationship between initial creep rate and CFA/FFA Additional analyses were conducted to identify whether any clear pattern emerges between the initial creep rate and any binder property parameters, specifically viscosity, effective asphalt cont ent, and G* sin Figures 5 17 through 5 19 present the relationships between initial creep rate and each of the parameters evaluated, respectively. According to analysis results, it appeared that these binder parameters were not clearly correlated with i nitial creep rate. However, it was found that unmodified mixtures that met all DASR IC criteria exhibited relatively low initial creep rate as highlighted in Figures 5 17 through 5 19 Unmodified mixtures that did not meet all DASR IC criteria exhibited a broad range of creep rate, indicating that inadequate gradation can result in high damage rate even when binder properties are satisfactory.

PAGE 97

97 Also, modified mixtures exhibited low initial creep rate even though not all DASR IC criteria were met. This clearl y indicates that if all DASR IC parameters, including DASR porosity, DF, EFT, and CFA/FFA, are within the acceptable ranges, the unmodified mixtures will have relatively low initial creep rates which could result in better cracking performance. Figure 5 17 Relationship between initial creep rate and viscosity Figure 5 18 Relationship between initial creep rate and effective asphalt content

PAGE 98

98 Figure 5 19 Relationship between initial creep rate and G* sin However, it does not necessarily imply that binder properties are not important. Binder properties do play a major role on creep rate and fracture resistance when gradation deficiencies are not present. Also, the results indicated that polymer modifi cation is likely related to the initial creep rate magnitude. Therefore, both binder properties and gradation are important factors to control the initial creep rate. 5.4.2 Models for Changes in Material Properties The identification of appropriate trends with respect to the changes in key material properties (e.g. fracture energy limit and creep rate) over time is important for accurate prediction of cracking performance of asphalt mixtures. As discussed earlier, the current lack of material property model s of this type is probably the greatest shortcoming in the ability to accurately predict pavement performance. Possible goals to develop improved material property models are presented as follows. Adjust the changes in material property based on the change s in IC characteristics. Adjust the changes in material property by including the effects of non healable permanent damage related to traffic load and moisture (environment).

PAGE 99

99 However, due to the limited data available for this research, this part of the st udy was limited to the recommendation of new concepts for modification of the two candidate models under evaluation i.e. AC stiffness (creep compliance) aging model and fracture energy limit aging model. 5.4.2.1 AC Stiffness Model As mentioned before, the existing model in Figure 5 1 showed that the AC stiffness is continuously increasing with time. However, this trend does not coincide with the results of prior laboratory, field, and accelerated pavement testing (APT) research, which indicated that the st iffness generally reduces with time after a certain age (Roque et al 2007, 2011). Therefore, a new concept was proposed and modification of the existing model was designed to appropriately describe the known trend of this property. Fi gure 5 2 0 describes t he proposed modification including the effect of non healable permanent damage related to load and moisture on the change in AC stiffness with time. As shown in Figure 5 2 0 there is a critical time denoted as td which separates pavement life into two stag es. In the early stage, Stage I, there is no effect of non healable permanent damage (related to load and moisture) on changes in AC stiffness. In Stage I, the trend of change in AC stiffness is mainly controlled by oxidative aging This indicates that the healing process is a dominant factor as compared to the damage process in governing the trend of change in material property in this stage. A lso shown in Figure 5 2 0 the non healable permanent damage process takes control in Stage II. The non healable pe rmanent damage includes both load induced and moisture related damage, which tends to reduce the AC stiffness after the critical time. Clearly, determination of t d is the first task in finalizing the proposed AC stiffness

PAGE 100

100 model. The next challenge is how t o quantify the effect of load induced and moisture related damage on AC stiffness after the critical time. Figure 5 2 0 Proposed AC stiffness model 5.4.2.2 Fracture Energy Limit Model T he fracture energy limit aging model developed as part of the NCHRP P roject 01 42A showed that the FE limit generally decreases with aging, and eventually reaches some minimum value after a sufficiently long time (see Figure 5 4). However, the results of prior laboratory, field, and APT research did not coincide well with t he trend introduced in the existing model (Roque et al., 2007, 2011). Therefore, modification of the existing model was designed to reasonably describe the observed trend of this property. Figure 5 2 1 presents the proposed modification s including the effec t of the non healable permanent damage related to load and moisture on the change in fracture energy limit with time. Since the fracture energy limit in the existing model was associated with the AC stiffness in a normalized form (see Equation 5 2), the pr oposed modifications for the existing fracture e nergy limit model (see Figure 5 2 1 (b)) are related to the modified A C stiffness model (see Figure 5 2 1 (a)).

PAGE 101

101 As shown in Figure 5 2 1 pavement life was separated into two stages by the critical time t d : In S tage I when mixture healing potential is high, it was assumed that there is no permanent damage induced by load and moisture. Therefore, the existing relationship for surface AC stiffness S(t), which accounts for the change in AC stiffness due to only oxid ative aging, was used for determination of the normalized change in stiffness S n (t). As a result, the existing relationship for fracture energy limit can be used for this part of the model. However, during Stage II when the time is greater than t d a modif ied relationship for surface AC stiffness, termed S d (t), was required to consider the permanent damage effect on the change in AC st iffness with time (see Figure 5 2 1 (a)). As a result, the normalized change in the modified AC stiffness can be expressed us ing the following equation. (5 5) Where, S dn (t) is the normalized change in stiffness S d (t) defined for Stage II, and S d0 and S dmin are stiffness values when S d (t) is at t = t d and t = 50 years, respectively. Then, the modified FE limit aging function wa s introduced for Stage II by relating the normalized change in FE limit to the normalized change in stiffness S d (t) by a power of k 2 which is expressed in the following equation. (5 6) Where, FE d (t) is the fracture energy limit function defined for Sta ge II, FE d0 is the value when FE d (t) is at t = t d k 2 is aging parameter for Stage II, FE min denotes minimum FE, and S dn (t) was defined in Equation 5 5

PAGE 102

102 It can a lso be seen from Figure 5 2 1 due to the incorporation of the permanent damage effect, the modi fied fracture energy limit function (Stage II) has a higher rate of reduction than the existing function, which is closer to the trend of the change in FE over time actually measured from field cores. (a) Change in AC stiffness over time (b) Change in FE li mit over time Figure 5 2 1 Proposed FE limit model In the case of the proposed fracture energy function, it is important to determine the critical time t d which indicates the point when permanent damage starts affecting the change in fracture energy with time. Also, the effect of non healable permanent

PAGE 103

103 damage due to load and moisture must be quantified in addition to the effect of oxidative aging after the critical time. 5.5 Summary Material property relationships were identified using material properties measured from field cores over time and assigned DASR IC model parameters calculated. Unique relationships were identified between initial mixture properties (i.e. fracture energy and creep) and DASR IC parameters. It was found that the relationships for i nitial fracture energy and creep rate appeared to work best when all DASR IC parameters were within the ranges considered to be acceptable. In addition, based on evaluation of existing material property aging models, a new concept and modifications were pr oposed to improve these models for more accurate prediction of changes in material properties over time. Due to the limited data available for this study, a full model could not be developed. Nevertheless, procedures to continue and complete the developmen t of improved models were recommended for future works.

PAGE 104

104 CHAPT ER 6 EVALUATION OF DASR IC CRITERIA USING HM A FM E MODEL 6 .1 Background The primary purpose of this chapter is the evaluation of DASR IC criteria with regard to their relationships on fiel d mixture performance using performance prediction model recently developed at the University of Florida, Enhanced HMA Fracture Mechanics Based Model (HMA FM E Model). Performance prediction effort using HMA FM E Model is presented in Section 6.2 through 6 .4 including a brief introduction to this tool, its input module, and prediction results for Superpave projects evaluated. It is noted that Project 8 has a Portland cement concrete base. Therefore, this section was not included in this part of study. 6.2 Enhanced HMA Fracture Mechanics Based Model (HMA FM E Model) The HMA FM E model was developed by the UF research team as part of the NCHRP Project 0 1 42A to predict top down cracking performance in HMA layers. As indicated by the name, the model is an enha nced version of the existing HMA FM model, which is the result of the continuous efforts of the UF team over the past years (Zhang e t al. 2001, Roque et al. 2002, 2004, Sangpetngam et al. 2003, 2004, Kim et al. 2008). During the course of the NCHRP project the enhanced performance model was formed by developing and incorporating into the existing model appropriate sub models that account for effects of aging, healing, and transverse thermal stresses on top down cracking performance. Furthermore, the enhanc ed model was calibrated and validated using data from Flori da field sections (Roque et al. 2010). The enhanced top down cracking performance model has four major components, as shown in the general model framework presented in Figure 6 1, including :

PAGE 105

105 The lo ad response and load associated damage sub models that are used to predict step wise load induced damage. The thermal response and thermal associated damage sub models that are used to predict thermally induced damage. The damage recovery and accumulation process that is used to accumulate damage after taking into account healing effect. Once the accumulated damage reaches the threshold, a crack will initiate or propagate. The mixture properties sub models that were devised to account for changes in mixture damage, fracture, and healing properties due to aging. Details for each component are described elsewhere (Roque et al., 2010 Zou and Roque, 2011) Figure 6 1 General f ramework of the HMA FM E m odel The enhanced model was used to predict top down crac king perfor mance in HMA layers for Superpave projects evaluated The remaining parts of this section present the input module for the HMA FM E model, followed by model prediction results. 6.3 Input Module The inputs for the HMA FM E model are divided into five categories, including traffic, climate, structure, mixture damage and fracture properties, and mixture healing

PAGE 106

106 properties. Table 6 1 summarizes the input data characteristics for the HMA FM E model. Table 6 1. Summary of input data characteristics for the HMA FM E model Inputs Description Traffic Multi year data: Based on current ESALs/Year as measured (No growth is counted) Climate Multi year data: Based on typical one year data in Melrose, FL Structure Three layer AC modulus (multi year d ata) predicted from initial data: Gradation, binder type, volumetric information (for AC stiffness model) GB, SG moduli (current data) obtained from FWD data (No moisture effect) AC damage and fracture propert y Multi year data: Predicted using mix ture property aging sub models Adjusted based on current data from IDT tests AC healing propert y Multi year data: Predicted using the maximum healing potential aging sub model Analysis Type Deterministic analysis Further descriptions of d ata char acteristics for each input category are described as follows: Traffic: The traffic volume (in terms of million ESALs per year) for the year of field evaluation for each project was taken as the base value and applied to the corresponding pavement section for the entire simulation period, without considering annual traffic growth (see Table 6 2 (a) ). Climate: Hourly temperature variation at different depths in the asphalt concrete (AC) layer for one typical year in Melrose, FL was used for all projects for the entire simulation period. The year to year change in temperature was not considered. Structure: A three layer pavement structure was selected for the simulation (see also Table 6 2 (a)). T hickness for AC and Base was obtained from design values. Modulu s for Base and Subgrade were determined based on backcalculation of falling weight deflectometer (FWD) testing data obtained at the time of field evaluation. The change in Base and Subgrade moduli due to moisture variation were not considered because the f ormer can be ignored in Florida and of the latter has small effect on top down cracking. AC layer modulus was predicted using the AC stiffness aging sub model which requires gradation, binder type, and

PAGE 107

107 volumetric properties measured at the unaged condition (see Table 6 2 (b)). In other words, the change in AC modulus due to aging was taken into account for the entire simulation period. AC damage and fracture property: AC damage and fracture properties including creep rate, fracture energy limit (FE f ), were predicted using mixture property aging sub models, taking into account the measured properties from field cores obtained at the time of field evaluation (see Table 6 2 (c)). AC healing property: AC healing potential was predicted using the maximum healing potential aging sub model. Table 6 2. Data used for model prediction Project (UF) ID Layer Thickness (in) Layer Modulus (ksi) Yearly Traffic (10 6 ESALs) AC Base/Subbase Base/Subbase Subgrade 1 7.43 21.65 44.1 45.0 0.480 2 7.40 22.32 74.9 35.6 1.037 3 9.64 22.32 66.2 39.6 1.193 4 7.44 22.24 50.7 30.3 1.012 6 6.40 22.80 60.3 36.7 0.512 7 6.74 22.00 67.7 31.0 2.321 9 5.50 20.50 51.0 23.1 0.058 10 7.75 18.00 29.3 32.3 0.036 11 7.75 22.00 65.9 31.9 1.412 12 6.49 18.40 47.8 21.6 0.031 (a) Pavement str uctural and material property and traffic volume Project (UF) ID Percent Passing by Weight Binder Type V eff (%) V a (%) MAAT (F) 3/4 in 3/8 in #4 #200 1 100.0 99.0 64.0 5.1 PG 67 22 10.1 5.4 69 2 100.0 89.0 45.0 4.9 PG 64 22 10.0 4.2 68 3 100.0 9 0.0 67.0 4.4 PG 67 22 8.0 6.3 68 4 100.0 95.0 67.0 5.0 PG 67 22 9.2 6.8 68 6 100.0 88.0 75.0 5.0 PG 64 22 10.1 5.8 71 7 100.0 88.0 70.0 4.2 PG 64 22 7.6 6.8 75 9 100.0 90.0 73.0 5.5 ARB 5 7.3 7.3 69 10 100.0 88.0 67.0 5.4 ARB 5 7.8 9.2 69 11 100.0 87 .0 59.0 4.7 PG 76 22 5.3 5.8 68 12 100.0 90.0 59.0 3.5 ARB 5 8.5 5.7 66 (b) Mixture gradation, binder type, and volumetric property

PAGE 108

108 Project (UF) ID Age of Cores (Years) T=10 C T=20 C Master Curve M R (Gpa) S t (Mpa) FE f (Kpa) D (10 3 s) (1/Gpa) D (10 3 s) (1/Gpa) m D 1 (1/Gpa) 1 11 12.93 2.38 1.90 1.618 3.906 0.564 0.042 2 11 8.16 1.76 2.97 4.515 19.059 0.720 0.045 3 11 8.92 1.85 1.60 2.475 20.704 0.736 0.026 4 11 7.29 1.39 1.35 2.476 14.355 0.662 0.041 6 1 1 9.08 1.61 1.13 1.417 5.320 0.495 0.058 7 1 1 9.46 1.97 1.70 1.343 4.616 0.476 0.071 9 7 9.61 1.39 0.55 0.929 2.631 0.473 0.042 10 7 14.63 2.73 1.33 0.558 2.584 0.563 0.014 11 6 7.59 1.49 1.45 1.857 6.649 0.493 0.083 12 6 12.45 1.92 1.23 1.462 5.266 0.550 0.039 (c) Mixture fracture and damage proper ty 6.4 Model Prediction Results The predicted relative crack depth (CD r ), crack amount (CA), and crack status for top down cracking for each Superpave projects evaluated are presented in Table 6 3. It shows that at the time of field evaluation, two out of the ten sections (Project 3 and 4) had reached maximum crack amount to failure (CA max ), two (Project 6 and 7) just started to crack, three (Project 2, 9, and 11) did not but were about to crack, and the rest (Project 1, 10, and 12) were far from crack init iation. It can also be seen from Table 6 3 that project sections 3 and 4 had worse cracking performance in terms of shorter crack initiation time (t i ) and higher average crack growth rate (G t ) than the other sections, while project section 1 had the best p erformance. Further discussion of the prediction results, including the increment of top down crack amount over time and the relationships between the DASR IC criteria and top down cracking performance for individual projects, will be presented in the sect ion 6.5.

PAGE 109

109 Table 6 3. Predicted top down cracking performance using HMA FM E model Project (UF) ID Time of Evaluation (Year) HMA FM E Top Down Cracking Predictions CD r (%) CA (ft/100ft) Predicted Status CA i (ft/100ft) t i (Year) G t (ft/year) 1 11 2.2 14.5 U 22.2 16.9 63.7 2 11 2.9 19.2 U 22.3 12.8 100.0 3 11 50.0 330.0 C 17.1 6.9 153.0 4 11 50.0 330.0 C 22.2 5.9 198.8 6 11 10.7 70.5 C 25.8 9.5 101.3 7 11 8.2 54.3 C 24.5 10.0 108.5 9 7 4.1 27.4 U 30.0 7.7 97.7 10 7 1.5 10.1 U 21.3 14.7 66.1 11 6 2.8 18.8 U 21.3 6.8 133.8 12 6 1.5 9.7 U 25.4 15.8 63.6 Note: 1. In HMA FM E, CD r was defined as crack depth over AC layer thickness (in %), and CA max was determined to be 330ft/100ft when CD r is equal to 50 % 2. CA i is crack amount at crack initiation. Figure 6 2 represent s the predictions development of top down crack amount over time. As shown, Projects 3, 4, and 11 exhibited relatively bad cracking performance with shorter crack initiation time (t i ) and higher average crack growth rate (G t ) than rest of the projects, while Projects 1, 10, and 12 showed good field cracking performance. (a) Bad p erformance s ections

PAGE 110

110 (b) Intermediate p erformance s ections (c) Good p erformance s ections Figure 6 2 Predicted c rack a mount i ncrease o ver t ime u sing HMA FM E m odel 6.5 Relationships between DASR IC Criteria and Model Prediction Results A ttempt s were made to assess preliminary DASR IC criteria established based on the comprehensive field performance evaluation of Superpave projects, including DASR

PAGE 111

111 poro sity, DF, EFT, and CFA/FFA, for consistently enhanced cracking performance. The predicted top down cracking performance in terms of crack initiation time (t i ) and average crack growth rate (G t ) were used along with DASR IC parameters determined for each Su perpave projects For evaluation, all project s were group ed based on ranges of crack initiation time (t i ) and average crack growth rate (G t ) predicted by HMA FM E model. All project sections were also categorized considering ranges of DASR IC parameters pr eliminarily determined to identify and evaluate the relationship between DASR IC criteria and performance model prediction results For crack initiation time (t i year) : Bad performance section s : t i < 7 Inte rmediate performance section s : 7 t i < 13 Good performance section s : t i 13 For average crack growth rate (G t ft/year ): Bad performance section s : G t 130 Intermediate performance section s : 70 G t < 130 Good performance section s : G t < 70 For DASR IC criteria: Mixtures that met all D ASR IC criteria Mixtures that not all DASR IC criteria are met Figures 6 3 (a) and (b) show the relationships between performance model prediction results, including crack initiation time (t i ) and average crack growth rate (G t ), and the DASR IC criteria respectively According to res ult shown in Figure 6 3 (a), it was found that projects that met all DASR IC criteria exhibited relatively long crack initiation time (t i ) (i.e. good or intermediate ranges of crack initiation time) while projects

PAGE 112

112 that not al l DASR IC criteria are met showed a broad range of crack initiation time (t i ) indicating that inadequate DASR IC criteria can result in bad field cracking performance. (a) Crack initiation time (b) Average crack growth rate Figure 6 3. Relationships between D ASR IC criteria and field cracking performance

PAGE 113

113 In addition, result in Figure 6 3 (b) shows that projects that met all DASR IC criteria exhibited relatively low average crack growth rate (G t ) (i.e. good or intermediate ranges of average crack growth rate), while projects that not all DASR IC criteria are met showed a wide range of average crack growth rate (G t ) including range considered as bad performance. These results clearly indicate that if all DASR IC parameters, including DASR porosity, DF, EFT, and C FA/FFA, are within the acceptable ranges, mixtures will have relatively good cracking performance in the field. 6.6 Summary Preliminary DASR IC criteria established in this study was evaluated to identify and verify their relationships on field cracking pe rformance using model prediction results. Results indicated that acceptable ranges of DASR IC parameters will result in consistently enhanced field cracking performance Therefore, it appeared that the introduction of DASR IC criteria as performance relate d design parameters to current mix design guidelines and specifications will lead to better and more consistent field performance of Superpave mixtures. However, current criteria preliminarily established need to be further validated and refined using addi tional field and laboratory data.

PAGE 114

114 CHAPTER 7 CLOSURE 7 .1 Summary and Findings This study was conducted to evaluate the effect of mixture component characteristics (i.e., DASR and IC) on properties and performance of Superpave mixtures s pecializing in the develop ment of a set of implementable gradation and volumetric criteria, and Hot Mix Asphalt (HMA) mixture property predictive relationships based on mixture component characterization. Field performance evaluation for both rutting and c racking was conducted to identify and verify performance related criteria using parameters from DASR IC mixture gradation model. In addition, an evaluation of existing material property models was conducted to identify and develop improved material propert y relationships for more accurate prediction of cracking performance of asphalt mixtures. A summary of findings associated with these tasks are presented as follows: Mixtures having DASR porosity within the acceptable range clearly exhibited better field r utting performance than mixtures with high DASR porosity. Mixtures with marginal DASR porosity exhibited either good or bad rutting performance in the field. Mixtures with DF considered to be acceptable generally exhibited better cracking performance in th e field than mixtures with the either low or high DF. DASR porosity, which reflects the characteristics of coarse aggregate structure, is the most dominant parameter to control rutting performance. IC characteristics including DF, EFT, and CFA/FFA could no t overcome the problems associated with a mixture with DASR porosity outside the acceptable range. IC characteristics are more important than DASR porosity criteria to clearly differentiate field cracking performance. It appears that the DF criteria should be considered in conjunction with EFT and CFA/FFA criteria to effectively distinguish the field cracking performance. Unique relationships were identified between initial fracture energy and three DASR IC parameters: DASR porosity, DF, and EFT. Initial fr acture energy

PAGE 115

115 generally decreases with increasing DASR porosity and DF and increases with increasing EFT. A new parameter, namely the ratio of the coarse portion and fine portion of fine aggregates (CFA/FFA), representing the composition or aggregate struc ture of the interstitial component of mixture s gradation, was introduced to develop the relationship to predict initial creep rate. It was identified that the initial creep rate generally increased with increasing CFA/FFA. The effect of polymer modificati on helps to reduce the initial creep rate, which is the damage rate of an asphalt mixture. It was found that the relationships for initial fracture energy and creep rate appeared to work best when all DASR IC parameters were within the ranges considered to be acceptable. Existing material property models for AC stiffness, fracture energy, and creep rate considered oxidative aging as the only factor responsible for changes in these properties with time. However, it was found that the trends based on existing models did not correlate well with the results of prior laboratory and field research. 7 .2 Conclusions The following key conclusions were drawn based on the findings and results of this study: Mixture gradation characteristics and resulting volumetric pro perties that effectively characterize the DASR and IC have an important effect on mixture fracture properties as well as on pavement performance in the field. DASR IC parameters identified were able to appropriately describe the critical aspects of mixture gradation and volumetric properties that play a major role on mixture performance. Mixtures with acceptable range of DASR IC parameters clearly exhibited better mixture performance in the field than those with DASR IC parameters outside ranges identified as acceptable. Thus, the introduction of DASR IC criteria as performance related design parameters to current mix design guidelines and specifications will lead to better and more consistent field rutting and cracking performance of Superpave mixtures. Ade quate gradation characteristics will result in more consistent initial mixture properties and provide predictive relationships for mixture fracture properties that will enhance our ability to accurate ly predict pavement performance. Inadequate gradation ch aracteristics can result in improper mixture fracture properties even when binder properties are satisfactory. Binder properties do play

PAGE 116

116 a major role on creep rate and fracture resistance when gradation deficiencies are not present. In other words, binder properties alone cannot overcome deficiencies in gradation. Predictive mixture property relationships will help reduce the need for sophisticated laboratory mixture tests, thereby increasing the potential for implementation of these more advanced pavement design systems such as the interim MEPDG recently adopted by AASHTO. The verified relationships will also serve to provide reliable inputs for more accurate prediction of service life using pavement performance prediction models. Appropriate trends with re spect to the changes in fracture energy and creep rate over time were proposed by including the effects of the non healable permanent damage. The challenges are to quantify these effects in addition to the effect of oxidative aging after the critical time, and to determine the critical time, which indicates the point when permanent damage begins. 7 .3 Recommendations and Future Works Based on extensive evaluations performed in t his study, recommendations for further investigations regarding the effect of mix ture component characteristics on property and performance of Superpave mixtures are summarized below: Since the p reliminary criteria established in this study for consistently enhanced field mixture performance, including DASR porosity, disruption factor (DF), effective film thickness (EFT), and ratio between coarse and fine portion of the fine aggregate in the mixture (CFA/FFA), were only based on limited data from Superpave monitoring project there is a need to validate and refi ne the criteria developed with a thorough laboratory study for incorporation into asphalt mix design. An effective and clear implementable set of gradation and volumetric criteria need to be further established for an enhancement of mix design procedure, which should result in lon ger lasting asphalt pavements, with the benefit being cost savings and less disruption to the public due to less frequent construction cycles. Conceptual relationships identified to describe changes in mixture properties over time will serve as the foundat ion for further development of improved models based on additional field data and laboratory studies using more advanced laboratory conditioning procedures. Since healing has been determined to be one of the most critical factors that affect cracking perfo rmance of asphalt mixtures, further investigations are needed to assess the relationships between DASR IC criteria developed and healing

PAGE 117

117 characteristics of asphalt mixtures using the healing test recently developed in other FDOT research

PAGE 118

118 APPENDIX DE TERMINISTIC PROCEDUR E FOR ESTIMATION OF CRACK INITIATION TIM E BASED ON CRACK RATIN G DATA Figure A 1. Determination of observed crack initiation time for Project 1 and 2 Figure A 2. Determination of observed crack initiation time for Project 3 and 4

PAGE 119

119 Figure A 3. Determination of observed crack initiation time for Project 5 and 6 Figure A 4. Determination of observed crack initiation time for Project 7 and 8

PAGE 120

120 Figure A 5. Determination of observed crack initiation time for Project 9 and 10 Figure A 6. Determination of observed crack initiation time for Project 11 and 12

PAGE 121

121 LIST OF REFERENCES AASHTO Standard Method of Test for Bulk Specific Gravity of Compacted Bituminous Mixtures Using Saturated Surface Dry Specimens. AASHTO T 166, Washington, D. C 2001. AASHTO Standard Method of Test for Percent Air Voids in Compacted Dense and Open Bituminous Paving Mixtures. AASHTO T 269, Washington, D. C 2001. Asphalt Institute Superpave Mix Design Superpave Series No.2 (SP 2), Third Edition, A sph alt Institute, Lexington, KY, 2001. Birgisson, B., Roq ue, R., and Gale, C. Page Performance B ased F racture C riterion for E valuation of M oisture S usceptibility in H ot M ix A sphalt Transportation Research Record : Journal of the Transportation Research Board No. 1 891 2004, pp. 55 61. D Angelo, J., Dongre, R. N. Practical Use of Multiple Stress Creep Recovery Test: Characterization of Stylene Butadiene Stylene Dispersion and Other Additives in PMA Binders Transportation Research Record: Journal of the Trans portation Research Board No. 2126 2009, pp. 73 82. D Angelo, J. New High Temperature Binder Specification using Multistress Creep and Recovery. Transportation Research E Circular: Journal of the Transportation Research Board, No E C147 20 10 pp. 1 13 Fwa T. F., and Oh, C. B. Effect of Moisture Content on Measured Properties of Asphalt Mixtures Transportation Research Record : Journal of the Transportation Research Board, No. 1492, 1995, pp. 61 70 Guarin A Interstitial Component Characterization to Evaluate Asphalt Mixture Performance Ph.D. Dissertation, University of Florida, Gainesville, FL 2009 Kandhal P. S., and Chakraborty S. Effect of Asphalt Film Thickness on Short and Long Term A ging of Asphalt Paving Mixtures. Transportation Research Re cord: Journal of the Transportation Research Board, No 1535, 1996, pp. 83 90. Kim B., Roque R., and Birgisson B. Effect of Styrene Butadiene Styrene Modifier on Cracking Resistance of Asphalt Mixture Transportation Research Record: Journal of the Transpor tation Research Board No. 1829 2003, pp. 8 15 Kim, J., Roque, R., and Birgisson, B. Integration of Thermal Fracture in the HMA Fracture Model Journal of the Association of Asphalt Paving Technologists Vol. 77, 2008, pp. 631 662 Kim S ., Roque, R. and Bir gisson, B. Identification and Assessment of the Dominant Aggregate Size Range (DASR) of Asphalt Mixture Journal of the Association of Asphalt Paving Technologists Vol. 7 5, 2006, pp. 789 814.

PAGE 122

122 Lambe, T. W., and Whitman, R. V. Soil Mechanics John Wiley & Sons, New York, 1969. Lottman, R. P. Predicting Moisture Induced Damage to Asphaltic Concrete: Ten Year Field Evaluatio n. National Cooperative Highway Research Program Synthesis of Highway Practice 175 Transportation Research Board, Washington, D. C., unpublished manuscript 1986 Mirza M. W., and Witczak M. W. Development of Global Aging System for Short and Long Term Aging of Asphalt Cements. Journal of the Association of Asphalt Paving Technologists Vol. 64 1995, pp. 393 430 Nukunya, B., Roque, R ., Tia, M., and Birgisson, B. Evaluation of VMA and Other Volumetric Properties as Criteria for the Design and A cceptance of Superpave Mixtures. Journal of the Association of Asphalt Paving Technologists Vol. 7 0 2001, pp. 38 69 Roque, R., Birgisson, B., Drakos, C. A., and Dietrich, B. Development and Field Evaluation of Energy Based Criteria for Top Down Cracking P erformance of Hot Mix Asphalt Journal of the Association of Asphalt Paving Technologists Vol. 73, 2004, pp. 229 260 Roque, R., Birgisson, B ., Sangpetngam, B., and Zhang Z. Hot Mix Asphalt Fracture Mechanics: A Fundamental Crack Growth Law for As phalt mixtures Journal of the Association of Asphalt Paving Technologists Vol. 71, 2002, pp. 816 827 Roque, R., Buttlar, W. G. Ruth, B. E. Tia, M Dickson, S. W. and Reid, B Evaluation of SHRP Indirect Tension Tester to Mitigate Cracking in Asphalt Pavements and Overlay s. Final Report of Florida Department of Transportation, University of Fl orida, Gainesville, FL 1997 Roque, R., Chun, S., Zou, J., Lopp, G., and Villiers, C. Continuation of Superpave Projects Monitoring Final Report of Florida Department of Transportation, Universit y of Florida, Gainesville, FL 2011 Roque R., Guarin A., Wan g G., Zou J., and Mork H. Develop Methodologies/Proto cols to Assess Cracking Potential of Asphalt Mixtures Using Accelerated Pavement Testing Final Report of Florida Department of Transportation, University of Florida, Gainesville, FL 2007 Roque, R. Zou, J. Kim, Y. R. Baek, C. Thirunavukkarasu S., Un derwood B. S. and Guddati M. N. Top down Cracking of HMA Layers: Models for Initiation and Propagation NCHRP Web Only Document 162, National Cooperative Highway Research Program, Transportation Research Board of the National Academies 2010

PAGE 123

123 Sangpetng am B., Birgisson, B. and Roque, R. Development of Efficient Crack Growth Simulator Based on Hot mix Asphalt Fracture Mechanics Transportation Research Record: Journal of the Transportation Research Board No. 1832, 2003, pp. 105 112 Sangpetngam, B., Bir gisson, B., and Roque, R. Multilayer Boundary element Method for Evaluating Top down Cracking in Hot mix Asphalt Pavements Transportation Research Record: Journal of the Transportation Research Board No. 1896, 2004, pp. 129 137 Schmidt, J. and Graf, P. E The Effect of Water on the Resilient Modulus of Asphalt Treated Mixes Journal of the Association of Asphalt Paving Technologists Vol. 41, 1972, pp.118 162 Sedwick, S. C. Effect of Asphalt Mixture Properties and Characteristics on Surface Initiated Lo ngitudinal Wheel Path Cracking, Master s Thesis University of Florida, Gainesville F L 1998 Witczak M. W., and Fonseca O. A. Revised Predictive Model for Dynamic (Compl ex) Modulus of Asphalt Mixtures. Transportation Research Record: Journal of the Tran sportation Research Board No.1 540 1996, pp. 15 23 Zhang, Z., Roque, R., Birgisson, B., and Sangpetngam, B. Identification and Verification of a Suitable Crack Growth Law Journal of the Association of Asphalt Paving Technologists Vol. 70, 2001, pp. 206 241. Zou J. and Roque, R. Top Down Cracking: Enhanced Performance Model and Improved Understanding of Mechanisms Journal of the Association of Asphalt Paving Technologists Vol. 80 2011.

PAGE 124

124 BIOGRAPHICAL SKETCH Sanghyun Chun was born in Seoul, South K orea in 1977. He attended Kyunghee University and received a Bachelor of Engineering degree in civil engineering in 2003. In the middle of his undergraduate studies, he served as a sergeant in the Republic of Korean Army from 1998 to 2000. In February 2003 Sanghyun started a Master of Engineering program in civil engineering at the Kyunghee University. After finishing his master s degree, in February 2005, he joined the Chungsuk Engineering Co. Ltd. i n Seoul and he worked as a chief engineer in Road and Ai rport Division for two years. His academic pursuit led him to attend the Ph.D. program of pavement and materials group at the University of Florida in 2008 and worked as a graduate research assistant with doctoral advisor, Dr. Reynaldo Roque. After complet ing his Ph.D., he plans to work as academia, government agencies, or private companies in civil engineering.