<%BANNER%>

Characterizing Healable and Non-Healable Micro-Damage in Asphalt Mixture

MISSING IMAGE

Material Information

Title:
Characterizing Healable and Non-Healable Micro-Damage in Asphalt Mixture
Physical Description:
1 online resource (122 p.)
Language:
english
Creator:
Simms, Reebie K
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Civil Engineering, Civil and Coastal Engineering
Committee Chair:
Roque, Reynaldo
Committee Members:
Hiltunen, Dennis R
Tia, Mang
Subhash, Ghatu

Subjects

Subjects / Keywords:
asphalt -- brittleness -- damage -- healing -- load -- micro -- mixture -- modulus -- permanent -- repeated -- resilient -- strain
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:
In this study, healable and non-healable micro-damage were characterized by investigating changes in resilient modulus during repeated load of asphalt mixture. A common assumption in the asphalt pavement community is that micro-cracks (micro-damage) are fully healable, while macro-cracks (macro-damage) are non-healable. It is now well recognized that micro-damage healing may strongly influence fatigue life of flexible pavements during the cracking process. Micro-damage and micro-damage healing processes, however, are not fully understood. To achieve a better understanding of these phenomena, a test was developed to allow for both evaluation and quantification of the effects of micro-damage and micro-damage healing in asphalt mixture. The healing potential test consists of repeated load damage tests (resilient modulus tests) during the damage phase followed by a healing phase during which resilient modulus tests are performed on a limited basis to monitor modulus recovery (healing). A parameter, modulus at fracture, was introduced to characterize relative brittleness of asphalt mixture. Relationships between modulus at fracture and load level were established to aid in the selection of appropriate load magnitudes for inducing micro-damage in asphalt mixture. Rate of healing was found to not be constant, but rather changed with time at a decreasing rate for any given mixture. As a result, a healing rate parameter was defined to allow for comparison between mixtures. Results showed that, in general, expected trends were observed. Mixtures tested at higher temperatures healed at a faster rate than those tested at lower temperatures and mixtures subjected to less oxidative aging healed at faster rates than those subjected to more oxidative aging. In addition, a mechanism involving differential thermal contraction was proposed to explain non-healable micro-damage development in asphalt mixture. For the mixtures and conditioning procedures used in this study, results were found to be inconclusive when assessing non-healable micro-damage development through temperature cycling of asphalt mixture. Results did, however, indicate the presence of load-associated non-healable micro-damage in asphalt mixture.
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 Reebie K Simms.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Roque, Reynaldo.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045724:00001

MISSING IMAGE

Material Information

Title:
Characterizing Healable and Non-Healable Micro-Damage in Asphalt Mixture
Physical Description:
1 online resource (122 p.)
Language:
english
Creator:
Simms, Reebie K
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Civil Engineering, Civil and Coastal Engineering
Committee Chair:
Roque, Reynaldo
Committee Members:
Hiltunen, Dennis R
Tia, Mang
Subhash, Ghatu

Subjects

Subjects / Keywords:
asphalt -- brittleness -- damage -- healing -- load -- micro -- mixture -- modulus -- permanent -- repeated -- resilient -- strain
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:
In this study, healable and non-healable micro-damage were characterized by investigating changes in resilient modulus during repeated load of asphalt mixture. A common assumption in the asphalt pavement community is that micro-cracks (micro-damage) are fully healable, while macro-cracks (macro-damage) are non-healable. It is now well recognized that micro-damage healing may strongly influence fatigue life of flexible pavements during the cracking process. Micro-damage and micro-damage healing processes, however, are not fully understood. To achieve a better understanding of these phenomena, a test was developed to allow for both evaluation and quantification of the effects of micro-damage and micro-damage healing in asphalt mixture. The healing potential test consists of repeated load damage tests (resilient modulus tests) during the damage phase followed by a healing phase during which resilient modulus tests are performed on a limited basis to monitor modulus recovery (healing). A parameter, modulus at fracture, was introduced to characterize relative brittleness of asphalt mixture. Relationships between modulus at fracture and load level were established to aid in the selection of appropriate load magnitudes for inducing micro-damage in asphalt mixture. Rate of healing was found to not be constant, but rather changed with time at a decreasing rate for any given mixture. As a result, a healing rate parameter was defined to allow for comparison between mixtures. Results showed that, in general, expected trends were observed. Mixtures tested at higher temperatures healed at a faster rate than those tested at lower temperatures and mixtures subjected to less oxidative aging healed at faster rates than those subjected to more oxidative aging. In addition, a mechanism involving differential thermal contraction was proposed to explain non-healable micro-damage development in asphalt mixture. For the mixtures and conditioning procedures used in this study, results were found to be inconclusive when assessing non-healable micro-damage development through temperature cycling of asphalt mixture. Results did, however, indicate the presence of load-associated non-healable micro-damage in asphalt mixture.
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 Reebie K Simms.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Roque, Reynaldo.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045724:00001


This item has the following downloads:


Full Text

PAGE 1

1 CHARACTERIZING HEALABLE AND NON HEALABLE MICRO DAM A GE IN ASPHALT MIXTURE By REEBIE SIMMS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGR EE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Reebie Simms

PAGE 3

3 To my mother

PAGE 4

4 ACKNOWLEDGEMENTS First and foremost, I would like to thank my advisor and committee chair, Dr. Reynaldo Roque, fo r his guidance and knowledge during the past four years. The hours upon hours of meetings and discussions we had have made me into the engineer I am today. My opinions were always heard, and while I am sure that not all of my ideas were good ones, he made me feel as if each and every one was valued. I am beyond grateful and will miss the dynamic that the two of us shared. I would also like to thank my committee members, Dr. Mang Tia, Dr. Dennis Hiltunen, and Dr. Ghatu Sub h ash for their support as well. Als o, special thanks to Dr. Charles Glagola for his mentorship long before I embarked on the journey of attaining my Ph.D. Thanks to my colleagues both former and current, for their support and friendship during my stay at the University of Florida. I have n ever meet a group of people as nice and as caring as those in the materials group, and I wish each and every one of them well. Special thanks goes to the Florida Department of Transportation for providing both the technical and financial support that made the majority of this research possible. Last but not least, I would like to thank my mother for whom I owe my very existence to. Her support and patience is not forgotten in the least bit.

PAGE 5

5 TABLE OF CONTENTS Page ACKNOWLEDGEMENTS ................................ ................................ ................................ ............. 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FI GURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 12 CHAPTER INTRODUCTION ................................ ................................ ................................ .................. 14 1 1.1 Background ................................ ................................ ................................ ...................... 14 1.2 Hypothesis ................................ ................................ ................................ ....................... 15 1.3 Objectives ................................ ................................ ................................ ........................ 16 1.4 Scope ................................ ................................ ................................ ................................ 16 1.5 Research Approach ................................ ................................ ................................ .......... 17 LITERATURE REVIEW ................................ ................................ ................................ ....... 21 2 2.1 Background ................................ ................................ ................................ ...................... 21 2.2 Damage and Healing Processes ................................ ................................ ....................... 21 2.2.1 Damage Process ................................ ................................ ................................ ... 22 2.2.1.1 Tradition al fatigue approach ................................ ................................ 22 2.2.1.2 Endurance limit approach ................................ ................................ ..... 23 2.2.1.3 Dissipated energy approach ................................ ................................ .. 24 2.2.1.4 HMA fracture mechanics approach ................................ ...................... 25 2.2.2 Healing Process ................................ ................................ ................................ ... 27 2.3 Experimental Evaluation of Damage and Healing ................................ .......................... 28 2.3.1 Energy Threshold Concept ................................ ................................ .................. 28 2.3.2 Fatigue Endurance Limit ................................ ................................ ..................... 28 2.3.3 Plateau Value ................................ ................................ ................................ ....... 29 2.4 Permanent Micro Damage ................................ ................................ ............................... 30 2.5 Low Temperature Cracking in Asphalt Mixtur e ................................ ............................. 31 2.6 Summary ................................ ................................ ................................ .......................... 32 TEST MATERIALS AND S PECIMEN PREPARATION ................................ .................... 33 3 3.1 Materials ................................ ................................ ................................ .......................... 33 3.2 Test Specimen Preparation ................................ ................................ .............................. 34 3.2.1 Batching and Mixing ................................ ................................ ........................... 34 3.2.2 Compaction ................................ ................................ ................................ .......... 34 3.2.3 Long Term Oven Aging ................................ ................................ ...................... 35 3.2.4 Determination of Air Void Content ................................ ................................ ..... 35

PAGE 6

6 3.2.5 Cutting of the Compacted Core ................................ ................................ ........... 37 3.2.6 Gage Point Attachment ................................ ................................ ........................ 37 DEVELOPMENT OF THE HEALING POTENTIAL TEST ................................ ................ 43 4 4.1 Background ................................ ................................ ................................ ...................... 43 4.2 Analysis of Loading Characteristics ................................ ................................ ................ 43 4.2.1 Load Mode ................................ ................................ ................................ ........... 44 4.2.2 Load Shape ................................ ................................ ................................ .......... 44 4.2.3 Load Level ................................ ................................ ................................ ........... 45 4.2.4 Rest Period ................................ ................................ ................................ ........... 45 4.3 Modulus at Fracture ................................ ................................ ................................ ......... 46 4.4 Determination of Load Level and Rest Period ................................ ................................ 48 4.5 The Healing Potential Test ................................ ................................ .............................. 50 4.6 Summary ................................ ................................ ................................ .......................... 51 TEST PROCEDURE A ND DATA INTERPRETATION METHODS ................................ 64 5 5.1 Overview ................................ ................................ ................................ .......................... 64 5.2 Data Acquisition ................................ ................................ ................................ .............. 64 5.3 Testing Procedure ................................ ................................ ................................ ............ 65 5.4 Damage Phase Data Interpretation ................................ ................................ .................. 65 5.5 Healing Phase Data Interpretation ................................ ................................ ................... 67 NON HEALABLE MICRO DAMAGE IN ASPHALT MIXTURE ................................ ..... 76 6 6.1 Background ................................ ................................ ................................ ...................... 76 6.2 Diff erential Thermal Contraction in Asphalt Mixture ................................ ..................... 76 6.3 Experimental Investigation into Temperature Associated Micro Damage ..................... 77 6. ................................ 79 6.4.1 ................................ ................................ ..................... 80 6.4.2 atio ................................ ................................ ...................... 82 6.5 Load Induced Permanent Micro Damage ................................ ................................ ....... 84 6.6 Summary ................................ ................................ ................................ .......................... 84 COMPONENTS OF MICRO DAMAGE AND MICRO DAMAGE HEALING ................. 97 7 CLOSURE ................................ ................................ ................................ .............................. 99 8 8.1 Summary and Findings ................................ ................................ ................................ .... 99 8.2 Conclusions ................................ ................................ ................................ .................... 100 8.3 Recommendations ................................ ................................ ................................ .......... 100 APPENDIX A ASPHALT MIXTURE INFORMATION ................................ ................................ ............ 102 B BULK SPECIFIC GRAVITY RESULTS ................................ ................................ ............ 105

PAGE 7

7 C SUPERPAVE IDT TEST RESULTS ................................ ................................ ................... 108 D HEALING POTENTIAL TEST RESULTS ................................ ................................ ......... 110 E HEALING RATE PARAMETER ................................ ................................ ........................ 115 LIST OF REFERENCES ................................ ................................ ................................ ............. 117 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 122

PAGE 8

8 LIST OF TABLES Table Page 3 1 Aggregate sources for dense graded mixture ................................ ................................ .... 38 3 2 Aggregate sources for open graded mixture ................................ ................................ ...... 38 4 1 Grouping of asphalt mixture using M f ................................ ................................ ............... 53 4 2 Load level and rest period for initial trial testing ................................ ............................... 53 4 3 Asphalt mixture testing conditions ................................ ................................ .................... 54 5 1 Data acquisition times for damage phase ................................ ................................ ........... 70 5 2 Data acquisition times for healing phase ................................ ................................ ........... 70 6 1 Average linear coefficient of thermal expansion fo r materials used in asphalt mixture ... 86 6 2 Temperature cycling conditions for asphalt mixture ................................ ......................... 86 6 3 Comparison of strain to fai lure for Superpave IDT and HPT ................................ ............ 8 6 A 1 Dense gradation job mix formula (JMF) ................................ ................................ ......... 103 A 2 Open gradation job mix formula (JMF) ................................ ................................ ........... 103 A 3 Dense gradation batch weights (cumulative) ................................ ................................ ... 104 A 4 Open gradation batch weights (cumulative) ................................ ................................ .... 104 B 1 Bulk specific gravity test results for open graded mixture with SBS modified binder ... 106 B 2 Bulk specific gravity test results for open graded mixture wit h asphalt rubber binder ... 107 C 1 Superpave IDT test results ................................ ................................ ............................... 109

PAGE 9

9 LIST OF FIGURES Figure Page 3 1 Aggregate gradation for dense graded mixture ................................ ................................ 38 3 2 Aggregate gradation for open graded mixture ................................ ................................ ... 39 3 3 Mechanical mix er ................................ ................................ ................................ ............... 39 3 4 Superpave gyratory compactor ................................ ................................ .......................... 40 3 5 Wire mesh configuration ................................ ................................ ................................ .... 40 3 6 Water displacement method ................................ ................................ .............................. 41 3 7 Masonry saw ................................ ................................ ................................ ...................... 41 3 8 Gage poi nts ................................ ................................ ................................ ....................... 42 4 1 Accumulation of delayed elasticity (static load) ................................ ................................ 55 4 2 Accumulation of delayed elasticity (repeated load) ................................ ........................... 55 4 3 Fatigue curve due to repeat load ................................ ................................ ........................ 56 4 4 Haversine waveform with varying rest period duration ................................ .................... 56 4 5 Illustr ation of britt leness index. ................................ ................................ ......................... 57 4 6 Comparison of relative brittleness ................................ ................................ ..................... 58 4 7 Modulus at fracture ................................ ................................ ................................ ............ 58 4 8 Ranking of asphalt mixtures by M f ................................ ................................ .................... 59 4 9 Effect of load level with 0.9 second rest period (DUS20) ................................ ................. 60 4 10 Effect of load level with 0.1 second rest period (DUS20) ................................ ................. 60 4 11 Effect of load level with 0.4 second rest period (DUS20) ................................ ................. 60 4 12 Effect of rest period with 40% load level (DUS20) ................................ ........................... 60 4 13 Effect of rest period with 25% load level (DUS20) ................................ ........................... 61 4 14 Effect of rest period with 35% load level (DUS20) ................................ ........................... 61 4 15 Effect of rest period with 30% load level (DUS20) ................................ ........................... 61

PAGE 10

10 4 16 Effect of rest per iod with 40% load level (DUS20) ................................ ........................... 61 4 17 Original M f load level relationship ................................ ................................ ................... 62 4 18 Relationship between M f and load level for dense graded mixture ................................ ... 62 4 19 Relationship between M f and load level for open graded mixture ................................ .... 63 4 20 Schematic of the healing pot ential test ................................ ................................ .............. 63 5 1 HPT setup. ................................ ................................ ................................ ......................... 71 5 2 Typical load and deformation response for the healing potential test ............................... 72 5 3 Typical horizontal deformation measurements ................................ ................................ .. 73 5 4 Typical resilient modulus data during the damage phase ................................ .................. 73 5 5 Fitting of data from healing phase ................................ ................................ ..................... 74 5 6 Determination of MR undamaged ................................ ................................ ............................. 74 5 7 Typical percent age vs. time with fit ................................ ................................ ................... 75 6 1 Testing plan for assessing permanent micro damage in asphalt mixture .......................... 87 6 2 Low t emperature bath. ................................ ................................ ................................ ...... 87 6 3 Specimen in polyethylene plastic bag ................................ ................................ ................ 88 6 4 Masking tape around specimen edges ................................ ................................ ................ 88 6 5 Specimens conditioning in water bath ................................ ................................ ............... 89 6 6 HPT results before and after 1 hour of thermal conditioning ................................ ............ 89 6 7 HPT results before and after 16 hours of thermal conditioning ................................ ......... 90 6 8 HPT results before and after four 16 hour cycles of thermal conditioning ....................... 90 6 9 Evolution of resilient modulus when performing successive HPTs (w/thermal conditioning) ................................ ................................ ................................ ...................... 91 6 10 Evolution of resilient modulus when performing successive HPTs (w/o thermal conditioning) ................................ ................................ ................................ ...................... 91 6 11 Damage rate for unconditioned and conditioned specimens ................................ ............. 92 6 12 Failure of asphalt mixtu re ................................ ................................ ................................ .. 93

PAGE 11

11 6 13 DCSE concept ................................ ................................ ................................ .................... 93 6 14 Comparing accumulated DCSE and DCSE f ................................ ................................ ...... 94 6 15 Permanent strain resulting from HPT (20C) ................................ ................................ .... 94 6 16 Evolution of resilient modulus when performing successive HPTs (10C) ...................... 95 6 17 Permanent strain resulting from HPT performed at 10C and 20C ................................ 95 6 18 Determination of resilient modulus using a calculated Poisson's ratio .............................. 96 6 19 Calculated Poisson's ratio during the HPT ................................ ................................ ........ 96 7 1 Effects responsible for changes in resilient modulus during the HPT ............................... 98 D 1 HPT results for dense graded unmodified mixture (STOA) ................................ ............ 111 D 2 HPT results for dense graded unmodified mixture (LTOA) ................................ ........... 111 D 3 HPT results for dense graded modified mixture (STOA) ................................ ................ 112 D 4 HPT results for dense graded modified mixture (LTOA) ................................ ............... 112 D 5 HPT results for open graded SBS modified mixture (STOA) ................................ ......... 113 D 6 HPT results for open graded SBS modified mixture (LTOA) ................................ ......... 113 D 7 HPT results for open graded ARB modified mixture (STOA) ................................ ........ 114 D 8 HPT results for open graded ARB modified mixture (LTOA) ................................ ....... 114 E 1 Healing rate parameter for dense graded mixture ................................ ........................... 116 E 2 Healing rate parameter for open graded mixture ................................ ............................. 116

PAGE 12

12 Abstract of Dissertation Present ed to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHARACTERIZING HEALABLE AND NON HEALABLE MICRO DAMAGE IN ASPHALT MIXTURE By Reebie Simms August 2013 Chair: Reynaldo Roque Major: Civil Engineering In this study, healable and non healable mic r o damage were characterized by investigating changes in resilient modulus during repeated load of asphalt mixture. A common assumption in the asphalt pavement com munity is that micro cracks (micro damage) are fully healable, while macro cracks (macro damage) are non healable. It is now well recognized that micro damage healing may strongly influence fatigue life of flexible pavements d uring the cracking process. M i cro damage and micro damage healing processes, however, are not fully understood To achieve a better understanding of these phenomena, a test was developed to allow for both evaluation and quantification of the effects of micro damage and micro damage he aling in asphalt mixture. The healing potential test consists of repeated load damage tests (resilient modulus tests) during the damage phase followed by a healing phase during which resilient modulus tests are performed on a limited basis to m onitor modul us recovery (healing). A parameter, modulus at fracture was introduced to characterize relative brittleness of asphalt mixture. Relationships between modulus at fracture and load le vel were established to aid in the selection of appropriate load magnitude s for inducing micro damage in asphalt mixture. Rate of healing was found to not be constant but rather changed with time at a decreasing rate for any

PAGE 13

13 given mixture. As a result a healing rate parameter was defined to allow for comparison between mixture s. Results showed that, in general, expected trends were observed. Mixtures tested at higher temperatures healed at a faster rate than those tested at lower temperatures and mixtures subjected to less oxidative aging healed at faster rates than those subje cted to more oxidative aging. In addition, a mechanism involving differential thermal contraction was proposed to explain non healable micro damage development in asphalt mixture. For the mixtures and conditioning pro cedures used in this study, results wer e found to be inconclusive when assessing non healable micro damage development through temperature cycling of asphalt mixture. Results did, however, indicate the presence of load associated non healable micro damage in asphalt mixture.

PAGE 14

14 CHAPTER 1 INTRODUCTION 1.1 Bac kground In flexible pavement, fatigue cracking is one of the most common and detrimental modes of pavement distress. For the most part, mechanisms explaining fatigue cracking macro crack initiation and propag ation are understood. What pre cedes macro cr ack initiation and propagation, however, is not. Conceptually speaking, it is easy to visualize fatigue crack development as series of micro cracks (micro damage) which propagate and then join together to form macro cracks. The true nature of the micro dam age process, however, is unknown. This is damage do not yet exist making the understanding of the micro damage mechanism(s) a very difficult and challenging task. It is now well recognized that micro dama ge healing may strongly influence fatigue life during micro crack initiation and propagation. As a result, many researchers within the asphalt pavement community have observed and verified healing effects in asphalt mixture and have developed different way s of evaluating such effects. Kim and Kim (1997) monitored effective layer moduli using the stress wave method by measuring wave transients after varying periods of accelerated loading. Results indicated that the analysis technique used in the study provid ed a sensitive means of evaluating the changes in asphalt surface layer properties during fatigue loading and rest periods. Effective modulus of the asphalt layer was found to increase as the rest period between loading cycles increased. Daniel and Kim (20 01) evaluated healing of laboratory generated asphalt concrete mixtures using the impact resonance test method. The impact resonance method was able to detect a decrease in dynamic modulus of elasticity as damage accumulated and an increase in dynamic modu lus of elasticity after specimens were a llowed to rest. Si et al. (2002 ) monitored changes in pseudo stiffness both during repeated load

PAGE 15

15 cycles and after rest periods. The authors indicated that pseudo stiffness was found to decrease consistently with an i ncreasing number of stress loading cycles and concluded that subsequent recovery of pseudo stiffness after rest periods indicated that healing took place. Degree of healing was found to be a function of binder properties. Liu et al. (2011) investigated hea ling of porous asphalt concrete by monitoring changes in resilient modulus. Samples were damaged using indirect tensile fatigue tests and healed using heat induction. Zou et al. (2012) studied the effects of age hardening and healing on top down cracking p erformance of full scale accelerated pavement test sections. The authors concluded that ext ensively aged test sections lost the ability to heal and consequently, became more susceptible to cracking. While it is clear a great deal of effort has been given to investigating micro damage and micro damage healing, missing from these efforts is a clearly defined approach to examine the observed phenomena. As such, this study seeks to provide a systematic method of both evaluating and quantifying the effects of m icro damage and micro damage healing in asphalt mixture to allow for better understanding of the micro damage and micro damage healing processes. 1.2 Hypothesis The following hypotheses were made in this study: Both healable and non healable micro damage may d evelop in asphalt mixture Healable micro damage is primarily load associated and cohesive in nature Differential thermal contraction in asphalt mixture will result in non healable damage which is adhesive in nature Once non healable micro damage is present additional micro damage induced by way of load may be non healable, as well

PAGE 16

16 1.3 Objectives The pr imary objective of this study was to characterize healable and non healable micro damage by investigating changes in effective modulus. Detailed objectives are s ummarized as follows: Identify appropriate procedures for inducing healable micro damage in asphalt mixture by way of load Identify appropriate procedures for inducing permanent micro damage in asphalt mixture by way of temperature cycling Develop an appro priate and systematic methodology to evaluate healing potential of asphalt mixture Use the developed system to differentiate healable and non healable micro damage in asphalt mixture 1.4 Scope This study was initiated because micro damage and micro damage mech anisms are not fully understood. To achieve a better understanding of these phenomena, a systematic method is needed to both evaluate and quantify the effects of micro damage and micro damage healing in asphalt mixture. In order to assess the relative cont ribution of micro d amage accumulation and micro damage healing to overall cracking performance, it is necessary to devise a way of isolating each effect. Previous works have shown that once a macro crack has initiated, it is no longer capable of being heal ed. As a result, this study was limited to micro damage analysis only. This study was divided into two phases. The first phase primarily focused on development of a test method to allow for evaluation of the effects of micro damage and micro damage healing in asphalt mixture, while the second phase focused on utilizing the developed test to differentiate healable and non healable micro damage in asphalt mixture.

PAGE 17

17 For Phase I, two mixture types, one dense graded and one open graded, were used. Binder types P G 67 22 (unmodified) and PG 76 22 (SBS polymer modified) were used for the dense graded mixture, while binder types PG 76 22 and ARB 12 were used for the open graded mixture. Each mixture was tested at three test temperatures: 0C, 10C, and 20C Aging co nditions included both short term oven aging (STOA) and long term oven aging (LTOA). For Phase II, evaluation was limited t o dense graded mixture ; the mixture and binder types being those of Phase I. Test temperature varied throughout this phase, while ag ing was limited to the LTOA condition only. 1.5 Research Approach This research is primarily focused on developing a test method to evaluate healing potential of asphalt mixture and using said test to differentiate healable and non healable micro damage in as phalt mixture. The overall research approach used to meet the objectives of this study was divided into five main tasks, which are described in detail below: Literature Review: Knowledge regarding the assessment of healing in asphalt binder and mixture was reviewed. Definitions of micro damage and micro damage healing along with current understandings of healing mechanisms were reviewed as well. Particular interest was given to existing approaches used to characterize and/or quantify micro damage and micro damage healing. Test Method Development: Includes development of a healing potential test method and associated data reduction and interpretation methods for evaluation of healing effects in asphalt mixture. Said test should be capable of inducing micro da mage in the asphalt mixture without inducing macro damage and in addition, should be able to evaluate healing potential of asphalt mixture. Development of such a test should include identification of a) a ppropriate mode of loading (static or repeated) b) a ppropriate loading procedure (magnitude and duration) c) r easonable testing and data interpretation time and d) a dequate frequency of data acquisition Laboratory Tests (Phase I): Use results from standard Superpave IDT and repeated load damage tests to id entify appropriate load levels and rest periods for use during the damage phase of the healing potential test. Once appropriate load levels and rest periods are identified, perform the healing potential test for a broad range of asphalt mixtures subjected to various conditions to evaluate healing potential. Lastly, compare healing

PAGE 18

18 potential test results to expected trends as a means of validating the test. This process is illustrated in Figure 1 1 Establish Conditioning Procedures: Identify appropriate met hods for subjecting mixtures to various forms of temperature cycling to highlight differential thermal contraction between asphalt binder and aggregate. Moisture effects (presence of water) will not be considered. Laboratory Tests (Phase II): Subject aspha lt mixtures to predetermined methods of temperature cycling in an effort to induce micro damage through loss of adhesion between the asphalt matrix and aggregate. Perform the healing potential test on damaged specimens to evaluate healing potential. Lastly use the healing potential test to differentiate healable and non healable micro damage This process is illustrated in Figure 1 2.

PAGE 19

19 Figure 1 1 Research approach flowchart for Phase I Perform trial repeated load damage tests with constant load level and varying rest period Does selected load level/rest period combin ation result in macro damage for allotted time? Increase rest p eriod duration Use optimum rest period duration in further analyses Perform Superpave IDT tests to obtain key damage/fracture properties Establish relationships between key damage/fracture properties and load level Us e load level given from relationship to perform healing potential test Compare results to expected trends NO YES

PAGE 20

20 Figure 1 2 Research approach flowchart for Phase II Establish resilient modulus reference by performing healing potential test Establish procedures for thermal conditioning Subject mixture to temperature cycle Perform healing potential test after thermal conditioning Compare results from initial healing potential test to healing test performed after thermal conditioning Permanent micro damage induced? Increase cond itioning tim e and/or ad d additional temperature cycles End YES NO

PAGE 21

21 CHAPTER 2 LITERATURE REVIEW 2.1 Background Loosely speaking, damage can be defined as a loss of structural integrity, either locally (at the material level) or globally (at the structural level ). This in turn, can be thought of as a reduction in stiffness. A recovery in stiffness is then thought to be healing. Damage and healing are best described by Bhasin et al. (2011): The growth of a micro crack is associated with the creation of new fractur e surfaces. A precursor to the growth of a micro crack is damage in the vicinity of the crack tip. This damage is associated with the deformation and rearrangement of molecules in the failure zone. The mechanism of self healing can therefore be regarded as a reversal of these processes. Many researchers characterize healing in asphalt mixture by simply an increase in stiffness because the healing mechanism is not clearly understood. Because the healing mechanism is not fully understood, it is difficult to t ruly define micro damage and micro damage healing. Therefore, in this paper, damage will be characterized by a reduction in resilient modulus, and healing will be characterized by a recovery in resilient modulus. Additionally, in crack initiation or growth. It shoul d be noted, however, that terminology used in the following sections of this literature review is that of each individual work, which may or may not adhere to the definitions of damage and healing given above. 2.2 D amage and Healing Processes Many empirical an d mechanistic approaches have been investigated in order to evaluate and develop appropriate models for predicting asphalt pavement performance with respect to

PAGE 22

22 fatigue cracking. In the following sections, a number of these approaches are presented in an ef fort to identify similarities and differences amongst the different approaches with the hope of achieving a greater understanding of the micro damage and micro damage healing processes. 2.2.1 D amage Process 2.2.1.1 Traditional fatigue approach Monismith et al. (1985) de veloped what is perhaps the most commonly used fatigue failure criterion. The failure criterion involves establishing a relationship between tensile strain and number of load cycles to failure and takes the form: (2 1) Where: N f is the number of cycles to failure; K, n, and b are regression coefficients from laboratory testing; t is the tensile strain; and E is the stiffness of the mixture. Using this relationship, failure is said to occur when 50% reduction in initial modulus is reached. Traditional fatigue models like the one expressed above, however, have been shown to be inadequate when predicting fatigue life due to the complex nature of the damage process. One problem is the assumption of a constant n valu e regardless of asphalt mixture type, with mixture variables for the most part being included in the K term. Various researchers have shown that K and n are interrelated ( Myre, 1992 ; Thompson and Carpenter, 2006 ). Also, simple fatigue laws such as the one developed by Monismith et al. do not consider the effects of geometry change (stress redistribution due to micro cracks and macro cracks) in asphalt mixture. In other words,

PAGE 23

23 the state of the mixture never changes and damage is assumed to not affect respons e. In addition, in traditional fatigue analysis, pavement distress is assumed to be continuous and cumulative, meaning all loads regardless of magnitude inflict distress. These observations along with others tend to result in an underprediction of field fa tigue life when using traditional fatigue analyses. Laboratory based fatigue life has also been shown to require shift factors ranging from 3 to 13 to match actual fatigue cracking developed in the field ( Finn et al., 1977 ; Bonnaure et al., 1982; Francken and Clauwaert, 1987; Groenendijk et al., 1997 ). This implies that important factors are not being properly accounted for in these approaches. Perhaps the most important factor not accounted for, is that of healing. 2.2.1.2 Endurance limit approach Monismith et al. (1970) has long proposed the idea of there being an endurance limit, a strain level below which none or very little fatigue damage develops. The fatigue endurance limit (FEL) is defined as the repeated flexural strain level below which damage is not cumul ative for hot asphalt mixture (HMA) and was proposed to be 70 microstrain by Monismith et al. If such a limit were to exist, it would imply that HMA layers experiencing strain levels lower than the established limit would never fail due to fatigue. Data fr om Monismith and associates indicated that a strain level around 70 microstain appeared to result in a noticeably long fatigue life Data was limited, but the authors concluded that the endurance limit was real and made an argument for 70 microstrain being a representative value. However, there is nothing unique or fundamental about strain in itself. So, the assumption of 70 microstrain as an endurance limit for all mixtures seems misplaced. Perhaps rather than seeking to define what the FEL is in terms of absolute measure, the more important realization is in how such a limit comes to exist or in other words, what conditions are needed such that the FEL can be observed.

PAGE 24

24 2.2.1.3 Dissipated energy approach For an undamaged nonlinear viscoelastic material, the plot o f measured stress versus strain is a hysteresis loop. The area within the hysteresis loop for an undamaged material is the dissipated energy. Since asphalt concrete is a viscoelastic material, it is highly time and history dependent, meaning that the dissi pated energy in one loading cycle depends on the dissipated energy in previous loading cycles. According to Ghuzlan and Carpenter (2000) not all dissipated energy is responsible for damage. Only the relative amount of dissipated energy created by each add itional load cycle will produce further damage, excluding dissipated energy due to plastic deformation and heat dissipation. To account for this, Ghuzlan and Carpenter developed the dissipated energy ratio concept. The ratio of dissipated energy change (RD EC) is the ratio of dissipated energy change between any two loading cycles divided by the number of cycles between them and is written as follows: (2 2) Where: RDEC n is the average ratio of dissipa ted energy change at cycle n; DE m and DE n are dissipated energies at cycle m and n, respectively; and m and n are loading cycles m and n, respectively. It is thought that the RDEC provides a true indication of the damage being done to the mixture from one determining how much of it causes damage. A typical plot of RDEC versus load consists of three zones. Of particular interest is zone II, where the RDEC is more or less constant. This zone is

PAGE 25

25 referred to as the plateau stage. During this period, a constant percentage of input energy is being converted into damage. This stage is followed by a rapid increase in RDEC indicating that fatigue failure has occurred (zone III). Zone III represents a critical condition in which the asphalt material is no longer capable of withstanding further external loading due to accumulation of internal micro damage which ultimately results in macro crack formation. The RDEC value corresponding to the number of l oading cycles at 50% reduction in stiffness is defined as the Plateau Value (PV). The RDEC concept is promising because the plateau value has been found to be a fundamental property of the asphalt mixture, meaning it is independent of mode of loading and t esting condition (Shen and Carpenter, 2005). However, the approach still manages to incorporate a 50% reduction in initial modulus, as is used in traditional fatigue analysis, as a failure criterion. According to the findings by Shen and Carpenter, there e xists a unique relationship between PV and N f (fatigue life at 50% stiffness reduction). In other work (Ghuzlan and Carpenter, 2000), 50% reduction in initial modulus has been shown to be inconsistent when predicting failure. Interestingly, the plateau val ue is defined as the RDEC corresponding to the number of load cycles at a 50% reduction in initial modulus (N f ) while failure is defined as the point where the RDEC increases rapidly (zone III). This seems to be a contradiction within itself. Furthermore, failure is determined by visual inspection which may prove difficult due to significant scatter of data. 2.2.1.4 HMA fracture mechanics approach Zhang et al. (2001) realized that to fully understand cracking in asphalt pavements, the use of fracture mechanics is necessary to fundamentally and accurately describe the cracking process. Therefore, a viscoelastic fracture mechanics based crack growth law for asphalt mixture was developed (Zhang et al, 2001; Roque et al. 2002). Central to the crack growth law is the

PAGE 26

26 co ncept of a damage threshold as a failure criterion for the initiation and propagation of macro cracks. Roque et al. (2002) proposed that damage below the threshold (i.e. micro damage) is fully healable, while damage occurring above the threshold (i.e. macr o cracks/macro damage) is not healable. Under the crack growth law, cracking is viewed as a discontinuous, stepwise process unlike traditional fatigue theory in which the cracking process is seen as continuous and cumulative. From the crack growth law, Roq ue and associates developed the HMA fracture mechanics model to predict macro crack initiation and growth. The HMA fracture mechanics model monitors the energy input, the dissipated creep strain energy (DCSE), induced in predetermined zones where crack gro wth may be of interest. The accumulated DCSE per load cycle is determined as follows: (2 3) Where: AVE is the average stress applied; and p max is the creep strain rate from the Superpave IDT creep test. A crack will propagate if the accumulat ed DCSE in the zone of interest exceeds the total DCSE limit of the mixture. The advantage of the model is that as cracks either initiate or propagate, the stresses induced redistribute to reflect the changing geometry unlike in traditional fatigue analysi s in which the state or integrity of the mixture is assumed to never change. In addition, the model accounts for healing, which is described in terms of recovered dissipated creep strain energy, to more accurately describe the cracking mechanism.

PAGE 27

27 2.2.2 Healing P rocess Some researchers describe micro damage healing using two steps: (1) crack wetting (closing of micro cracks) and (2) intrinsic healing (strength gain). Intrinsic healing is then divided into two additional components: instantaneous strength gain due to interfacial cohesion between the surfaces of the wetted crack interface and time dependent strength gain due to randomization of molecules across the wetted crack interface (Bhasin et al., 2011). According to Little and Bhasin (2007), healing should not be confused with viscoelastic recovery due to rearrangement of molecules within the bulk of the material which occurs even when damage has not been induced with the material. m acroscopic recovery, or rate of healing, within a material as the combination of two functions: h (t)]. The net macroscopic recovery, R(t), is defined below: (2 4) Where: R h (t) is the intrinsic healing function; is the time variable. R h (t) is controlled by inherent material properties such as surface free energy of the binder and molecular morphology and e

PAGE 28

28 Bhasin et al., 2011). 2.3 Experimental Evaluation of D amage and Healing 2.3.1 Energy Threshold Concept As stated previously, Zhang et al. (2001) proposed the existence of an energy threshold below which damage is fully healable and above which damage is non healable To verify the threshold concept, the researchers evaluated healing potential of both micro damage and macro damage by conducting a series of repeated load damage tests. In the first stage, repeated load (below the threshold) was applied for 600 load appl ications. Specimens were then healed by being placed in an oven at high temperature for 12 hours. After healing, specimens were then loaded again in the same manner as before. Results indicated that there was no change is resilient deformation measurements before and after healing. As a result, the authors concluded that micro damage in the asphalt mixture was fully healable. In the second stage, repeated load (above the threshold) was applied for 600 load applications. Healing of the mixture was attempted by placing specimens in an oven at high temperature for 12 hours. Afterwards, specimens were then loaded again until failure was reached. Results indicated that resilient deformation measurements before and after healing were significantly different. As a result, the authors concluded that micro damage was not healable since resilient deformation continued to increase (non 2.3.2 Fatigue Endurance Limit The existence of a fatigue endurance limit (FEL) was verified by conducting lengthy fatigue tests at low strain levels (Carpenter et al., 2003; Shen and Carpenter, 2005; Thompson and Carpenter, 2006). These researchers found that fatigue behavior at normal strain levels (~300 to

PAGE 29

29 mixtures deviated from the typical linear relationship between strain and number of load cycles to failure (traditional fatigue curve) for normal strain levels and reached a flat line behavior at low strain levels. The point at which the traditional fatigue curve began to flatten out was deemed the FEL. Data showed that strains below the FEL resulted in extremely long fatigue lives as compared to those predicted by traditional fatigue analysis. The au thors suggested that the FEL represented the threshold between damage and healing in asphalt mixture. Damage resulting from strain levels above the FEL is greater than the healing potential of the mixture, while damage resulting from strain levels below th e FEL is either equal or less than the healing potential of the mixture, thus allowing for full healing of the mixture to occur. Test results also indicated that the FEL is not unique, but rather mixture dependent. The authors concluded that a value of 70 microstrain is conservative, but that the FEL can vary from 70 to 100 microstrain. 2.3.3 Plateau Value Thompson and Carpenter (2006) realized the failing of traditional fatigue analysis (strain vs. load cycles to failure at normal strain levels) in accurately de scribing mixture performance, but saw potential in the FEL concept (strain vs. load cycles to failure at low strain levels). Thompson and Carpenter pointed out that the strain load cycles to failure relationship resulting from fatigue testing is not unique but that the dissipated energy (RDEC) approach mentioned in previous sections does provide a unique relationship between damage and loads to failure. As such, the FEL and dissipated energy concepts were incorporated in further research. Using the IPC fou r point bending beam, Carpenter and Shen (2006) ran a series of healing tests at ( section 2.2.1) decreased meaning longer fatigue life. The authors also developed an a dditional parameter, the PV L a proposed PV value for the FEL. The authors suggested that the healing data could be extrapolated to the PV L arguing that as strain approaches the PV L the time

PAGE 30

30 required to reverse damage would decrease. For strain levels at the PV L there would be an equal amount of damage accumulation and damage recovery. 2.4 Permanent Micro D amage Within the asphalt community, it is generally accepted that micro damage is healable and macro damage is non healable, or permanent. In the research efforts mentioned in previous sections, it is implied that load induced micro damage is permanent only if the mixture is no longer capable of being healed. Stated differently, load induced damage is completely healable as long as the mixture has not been s ubjected to significant oxidative aging and embrittlement. To date, it appears that only a distinction between healable micro damage and non healable macro damage has been made. Most recently, Roque et al. (2012 a ) were successful in inducing permanent micr o damage in asphalt mixture through a combination of long term oven aging (LTOA) and cyclic pore pressure conditioning (CPPC). In their work, the researchers evaluated the effects of laboratory heating, cyclic pore pressure, and cyclic loading on changes i n fracture properties of asphalt mixture. Using established procedures, specimens were aged, vacuum saturated, and then subjected to triaxial cyclic pore pressure conditioning. In both cases vacuum saturation and pore pressure conditioning water was th e permeant. When considering both heat oxidation conditioning (HOC) and CPPC as compared to HOC only, the researchers reported an observed overall reduction in effective modulus (stiffness). This overall reduction in effective modulus was equated to perman ent micro damage. However, Kim and Coree (2005) have suggested that vacuum saturation might lead to irreversible damage since the only way to get water into internal voids is by rupturing the binder or mastic medium between the voids. Meaning, the reductio n in effective modulus observed by Roque and associates might not be a result of HOC and CPPC,

PAGE 31

31 but rather the result of premature damage through vacuum saturation. It appears that to date, no studies have proposed other mechanisms for possible sources of p ermanent micro damage 2.5 Low Temperature Cracking in Asphalt Mixture When examining cracking performance of asphalt mixture at low temperature, Deme and Young (1987) observed quite an interesting occurrence. When plotting breaking stress (tensile strength) versus stiffness (modulus at low temperature), breaking stress was found to decrease after peaking. Stated differently, tensile strength did not continue to increase with decreasing temperature, as one would expect, rather tensile strength was found to dec rease after peaking with decreasing temperature. El Hussein and El Halim (1993) observed a similar effect when investigating the effect of temperature change on bond performance of asphalt mixtures. In their work, they found that specimens exposed to low t emperature ( 30C) for an extended period of time (14 days) resulted in a decrease in tensile strength, which was attributed to internal damage in the asphalt mixture caused by differential contraction of the asphalt aggregate system. In addition, the auth ors found that freeze thaw conditioning resulted in more stripping damage evident by a further reduction in tensile strength. It was suggested that during the freeze thaw process, cracks induced by way of differential thermal contraction acted as a path fo r water to flow into the asphalt aggregate interface thereby enabling stripping to occur. Th e authors did not explictly state whether the damage was healable or non healable, but one could infer that damage induced in such a manner would logically be perma nent, and thus non healable. It should be noted that the work done by El Hussein and El Halim utilizes freeze thaw in the presence of water to examine the effects of temperature change on asphalt mixture behavior, but this work seeks to do the same without the presence of water.

PAGE 32

32 It is possible that the phenomena observed by both Deme and Young and El Hussein and El Halim are the result of the development of internal micro damage due to differential thermal contraction of the asphalt mixture and will be inve stigated further in this study. 2.6 Summary From this literature review it can be seen that the re is no consensus on micro damage and micro damage healing processes. This is because micro damage and micro damage healing processe s are not fully understood. M icr o damage and micro damage healing processes are not fully understood partially because the tools necessary to verify the existence of the phenomena do not yet exist. As stated previously and as shown in this literature review, most research efforts to dat e view micro damage as being fully healable and macro damage as being non healable, or permanent. Four approaches were presented including the traditional fatigue approach, endurance limit approach, dissipated energy approach, and the HMA fracture mechanic s approach. All of the approaches, with the exception of the traditional fatigue approach which does not consider healing, share one thing in common. Central to these approaches is the concept of a threshold by which damage and healing are separated. And, it is with this shared commonality that the basis of this study is found.

PAGE 33

33 CHAPTER 3 TEST MATERIALS AND S PECIMEN PREPARATION 3.1 Materials This section provides information regarding the materials used to produce laboratory generated specimens for the purposes of thi s study. Materials were selected to be representative of those used in the state of Florida. Using Superpave (A sphalt Institute, 2001 ) and FDOT Specifications Sections 334 and 337 (FDOT 2010 a ), two gradations of asphalt mixture, one dense graded and one o pen graded (FC 5 mixture), were designed. Dense graded asphalt mixtures are typically used for structural purposes, while open graded asphalt mixtures usually serve as wearing courses which provide frictional characteristics such as skid resistance. Georgi a granite, one of the more commonly used aggregate types, was used to produce all dense graded asphalt mixtures, while Florida oolitic limestone was used to produce all open graded asphalt mixtures. Aggregate source information is shown in Table 3 1 for de nse graded mixture and Table 3 2 for open graded mixture. All dense graded mixtures were designed to have a 12.5mm nominal maximum aggregate size. Open graded mixtures were designed according to FDOT specification Section 337 for FC 5 mixture type (FDOT 2 010 a ). The gradation, or particle size distribution, for the dense graded mixture can be found in Figure 3 1 and Figure 3 2 for the open graded mixture. Detailed information regarding the two gradations can be found in Appendix A. In all, three asphalt bi nder types were used in this study: a control binder (PG 67 22), a polymer modified binder (PG 76 22), and a rubber modified binder (ARB 12). Binder was obtained from CITGO Asphalt Refining Company. The PG 76 22 binder was achieved by blending PG 67 22 and Styrene Butadiene Styrene (SBS) polymer. The ARB 12 is a blend of PG 67 22 with 12 percent ground tire rubber by weight. Each mixture type was designed with

PAGE 34

34 two binder types. Dense graded mixtures were designed according to the Superpave volumetric mix de sign method with a 4.80 percent and 4.82 percent binder content for the control binder (PG67 22) and the modified binder (PG76 22), respectively. Open graded mixtures were designed in accordance wi th FDOT specification (FDOT 2010 a ) with a 6.40 percent bind er content for both binder types (PG76 22 and ARB 12). 3.2 Test Specimen Preparation 3.2.1 Batching and Mixing Aggregates and asphalt binder were oven heated at the desired mixing temperature for approximately 3 hours before mixing. For the unmodified mixture and m odified mixture, this temperature was determined to be 315F and 325F, respectively. After being heated, the aggregates and binder were removed from the oven and mixed together using a mechanical mixer ( Figure 3 3 ) until the aggregates were coated well. T he resulting 4500g mixtures were then placed in pans and put into an oven set at the mixing temperature and short term oven aged for 2 hours. To ensure a uniformly aged mixture, each mixture was stirred an hour into the STOA conditioning process. 3.2.2 Compacti on After STOA conditioning, the 4500g mixtures were removed from the oven and compacted to a 150mm diameter core using the Servopac Superpave Gyratory Compactor (SGC) shown in Figure 3 4 It should be noted that even though dense graded mixtures were desig ned to have a 4.0 percent air void content, mixtures were actually compacted to achieve a 7.0 0.5 percent air void content for the sliced specimen ( section 3.2.5). It is believed that a 7.0 percent air void content more accurately represents air void con tent in field pavements immediately after construction.

PAGE 35

35 3.2.3 Long T erm O ven A ging In addition to being short term oven aged, mixtures were also subjected to LTOA conditioning for 5 days at 185 0.5F. LTOA conditioning simulates aging of pavements after 7 to 10 years of service. Since cracking is a long term phenomenon usually taking place 5 to 7 years after construction, it is more likely that it will be easier to induce micro damage, healable or otherwise, once mixtures have been aged appropriately through L TOA conditioning. Because of the high temperature specified for LTOA conditioning, there is always a possibility that the gyratory compacted cores could fall apart during the conditioning process. To ensure an intact specimen during LTOA conditioning, a pr ocedure developed by Varadhan (2004) was used. While the procedure was actually intended for open graded gyratory compacted cores, it was used for the dense graded specimens produced in this study as well. As described in the procedure, gyratory compacted specimens were wrapped with a wire mesh with openings of 1/8 and 3/8 inch for the dense graded and open graded compacted cores, respectively. The 1/8 inch openings allow for good air circulation while preventing smaller aggregates from falling through the mesh. Two metal bands were then clamped to the mesh and core as shown in Figure 3 5 Special care was taken not to apply too much pressure when attaching the bands. Lastly, the specimens were placed on porous metal plates and then placed into an oven. 3.2.4 Dete rmination of Air Void Content After the gyratory compacted specimens were allowed to reach ambient temperature, the bulk specific gravity (G mb ) of each compacted core was measured and the resulting air void content was calculated. When determining G mb of c ompacted asphalt mixture, it is standard practice to use water displacement methods ( Figure 3 6) such as those outlin ed in AAHSTO T166 (AASHTO, 2012a ) and ASTM D2726 (ASTM, 2011 a ). For the most part, when considering fine graded mixtures like the dense gra ded mixtures used in this study, these

PAGE 36

36 methods result in accurate G mb However, when testing coarser mixtures such as open graded mixture and stone matrix asphalt (SMA), it has been found that these methods result in erroneous measurement of G mb ( Cooley et al., 2002 ; Buchanan and White, 2 005 ). Therefore, for the open graded mixture used in this study, G mb was determined using the Corelok device as suggested by Cooley et al. (2002) and Buchanan and White (2005). The procedure for determination of G mb using t he Corelok device can be found in ASTM D6752 (ASTM, 20 11 b). Bulk specific gravity results for open graded mixture are presented in Appendix B. Using equation 3 1 air void content was calculated as follows: (3 1) Where: V a is percent air voids, G mb is the bulk specific gravity; and G mm is the theoretical maximum specific gravity of the asphalt mixture. For the dense graded asphalt mixture, the target air void content of the gyratory compacted core was 7.5 percent and 7.0 percent (0.5 percent) for the sliced specimen ( section 3.2.5). Air void content tends to be approximately 0.5 percent lower in the mid core region where the sliced specimen is ultimately taken from, so t he target air void content of the slice was 7.0 percent. It is believed that a 7.0 percent air void content more accurately represents air void content in field pavements immediately after construction. All sliced specimens were the arranged into groups of three to minimize variability in terms of air void content.

PAGE 37

37 3.2.5 Cutting of the Compacted Core The 150 millimeter diameter cylindrical cores were cut using a water cooled masonry saw as shown in Figure 3 7 From one compacted core, two sliced test specimens we re obtained. Test specimen dimensions were 150 millimeters in diameter and 1.5 inches and 2.0 inches in thickness for the dense graded and open graded mixtures, respectively. Since water is used to keep the blade of the saw cool, specimens were dried by be ing placed in a dehumidifier at a constant relative humidity of 50 percent for at least 48 hours before any further testing. 3.2.6 Gage Point Attachment Four gage points (5/16 inch diameter by 1/8 inch thick) were attached to each face of the test specimen with epoxy using the apparatus shown in Figure 3 9 Gage points were placed 1.5 inches apart from one another in both the vertical and horizontal axes.

PAGE 38

38 Table 3 1 Aggregate sources for dense graded mixture Type of Materi al FDOT Code Producer Pit No. # 78 Stone 43 Junction City Mining GA 553 # 89 Stone 51 Junction City Mining GA 553 W 10 Screenings 20 Junction City Mining GA 553 Local Sand V. E. Whitehurst & Sons Starvation Hill Table 3 2 Aggregate sources for open graded mixture Type of Material FDOT Code Producer Pit No. S1A Stone 41 White Rock Quarries 87 339 S1B Stone 53 White Rock Quarries 87 339 Screenings 22 White Rock Quarries 87 339 Filler Figure 3 1 Aggregate gradation for dense graded mixture

PAGE 39

39 Figure 3 2 Aggregate gradation for open graded mixture Figure 3 3 Mechanical mixer

PAGE 40

40 Figure 3 4 Superpa ve gyratory compactor Figure 3 5 Wire mesh configuration

PAGE 41

41 Figure 3 6 Water displacement method Figure 3 7 Masonry saw

PAGE 42

42 A B Figure 3 8 Gage point s. A) A pparatus B) A ttachment

PAGE 43

43 CHAPTER 4 DEVELOPMENT OF THE HEALING POTENTIAL TEST 4.1 Background Through comprehensive literature review, it was determined that the Superpave IDT testing system should be utilized throughout this research because of its practicality and ability to provide reasonable and accurate damage/fracture properties of asphalt mixture including open graded asphalt mixture (Koh and Roque, 2010a; Koh and Roque 2010b). Another advantage of the Superpave IDT over other testing syste ms is that field core samples can be used The Superpave IDT test is comprised of three standard tests: the resilient modulus, creep, and strength test. Ultimately, the goal of this research is to characterize micro damage and micro damage healing through changes in effective stiffness. To do so, characterizing damage is most logically done by evaluating modulus. The type of modulus chosen can significantly affect interpretation of the results. In this study, it was concluded that resilient modulus (M R ) was appropriate for both assessing damage and damage recovery since it is a convenient way to measure the effective stiffness of the asphalt mixture. Using resilient modulus, the elastic behavior, or the ability of the material to rebound to its original stat e, can be approximated. Asphalt mixture is not a purely elastic material; it is a viscoelastic material. However, any time dependent response of the material can be accounted for during rest periods between repeated load applications. 4.2 Analysis of Loading C haracteristics When considering how to appropriately induce load associated micro damage in asphalt mixture, several factors need to be considered; among these are load mode, load shape, load magnitude, and rest period duration. Each of these factors will have an effect on the amount of

PAGE 44

44 micro damage accumulation and thus, healing of the resulting micro damage. Each of the four factors is discussed in the following sections. 4.2.1 Load Mode One concern regarding damaging of asphalt mixture is selection of load mod e. Static and repeate d loading are two common load modes used to evaluate damage and fracture of asphalt mixture. Static load methods tend to result in significantly reduced testing time, but delayed elasticity may cause error in resilient modulus calculat ions, as it is nearly impossible to separate delayed elastic response from actual damage/damage recovery ( Figures 4 1 and 4 2). Roque et al. (1997a) showed that delayed elasticity is present for static loading conditions even when load durations are very s hort. Thus a repeated load approach was selected for damaging of asphalt mixture. By using repeated load, it then becomes possible to incorporate rest periods between load applications to minimize delayed elasticity. With the selection of the repeated loa d mode, three factors including load shape, load level (magnitude) and duration of rest period, need to be considered. 4.2.2 Load Shape A haversine waveform was selected for testing. According to Seed et al. (1955), a haversine load pulse with a 0.1 second durat ion was found to replicate the loading time within a pavement structure resulting from the pass of a vehicle tire traveling at approximately 88 km/h. Servo hydraulic load systems are capable of applying both haversine and square load shapes, however, in th e case of square load shapes, it is difficult for the MTS machine to instantaneously apply and remove load through hydraulic transmission. In contrast, the loading process of the haversine waveform is more gradual, allowing for more accurate control of loa d magnitude.

PAGE 45

45 4.2.3 Load Level Determination of an appropriate load level should result in development of significant damage within the steady state damage range as shown in Figure 4 3. By remaining in the steady state damage range, the tertiary state, where macr o crack initiation and growth occurs, is avoided. However, remaining in this range does not necessarily ensure that appropriate load levels are being used. From a practical point of view, loads too high may result in excessively high rates of damage which would make testing difficult to control and results hard to analyze. In addition, high loads might also lead to stress concentration at the loading strip which might result in failure of this region (Varadhan 2004). On the other hand, loads too low in magn itude might result in excessively long testing times to induce the same amount of damage as higher loads. One need also consider stiffness or brittleness of the asphalt mixture. M aterials tested at low er temperatures will exhibit lower strain to failure th an more ductile materials or those tested at higher temperatures, which means these materials may not be able to sustain loads of great magnitude while avoiding macro crack initiation. 4.2.4 Rest Period Traditionally, the standard Superpave IDT resilient modulus test uses a haversine waveform with a 0.1 second load pulse followed by a 0.9 second rest period. It is important, however, that rest period duration be studied because rest period can have a huge effect on healing. Shorter rest periods may result in high damage rates, while longer rest periods will yield relatively long testing times and excessive healing with little or no accumulation of damage. From a practical point of view, the most effective rest period duration should be one long enough to allow rec overy of most of the delayed elasticity, yet be short enough to minimize healing during damage accumulation. Rest periods lasting 0.1, 0.4, and 0.9 second s ( Figure 4 4) were examined to see their effect on healing.

PAGE 46

46 4.3 Modulus at Fracture Traditionally, in the field of rock mechanics a n energy based approach is used to characterize brittleness of rocks. The brittleness index, I B is defined as the ratio of specific elastic energy (S) accumulated in the materia l up to the point of fracture and total specific ene rgy ( W) consumed up to that point (Coates and Parsons, 1966; Hucka and Das, 1974). This concept is illustrated in Figure 4 5(a) and in the following equation: (4 1) Where: I B is the brittleness index; S is the specific elastic energy; and W is the total specific energy. Roque et al. (2012 b ) hypothesized that appropriate load levels used to induce micro dama ge in asphalt mixture might depend on brittleness of the mixture. Using the HMA fracture mechanics concept developed at the University of Florida (Zhang et al., 2001; Roque et al., 2002), the equation above can be rewritten using fracture energy (FE), diss ipated creep strain energy (DCSE), and elastic energy (EE) without sacrificing any meaning of equation 4 1 ( Figure 4 5(b)). This new equation in terms of asphalt mixture fracture properties is written below: (4 2) Where: I B,HMA is the brittleness index based on HMA fracture mechanics; EE is the elastic energy; FE is the fracture energy; and

PAGE 47

47 DCSE is the dissipated creep strain energy of the asphalt mixture. W hen asphalt mixtures were ranked using the brit tleness index based on the HMA fracture m echanics concept (I B, HMA ), no logical trend appeared to emerge from the data. The index failed to sort mixtures by brittleness, be it by temperature or degree of heat oxidation (STOA or LTOA). Results seem ed to indicate that this energy based approach was not appropriate for asphalt mixture. Consequently, there was a need to develop a new approach. Figure 4 6 illustrates a typical strength failure strain relationship f or asphalt mixture. From this illustration, it can be said that Mixture 1 is more brittle than Mixture 2. This is based on the observation that strength is higher and failure strain is lower as test temperature is reduced. With this in mind, a new paramete r modulus at fracture was conceived The modulus at fracture ( M f ) is illustrated in Figure 4 7 and is written as follows : (4 3) Where: M f is the modulus at fracture ; S t is the strength in kPa; and f Figure 4 8 shows mixtures ranked in terms of M f As seen in the figure, this approach successfull y separated relative brittleness of asphalt mixture with temperature and degree of oxidative aging Based on these results, M f was deemed to be a more appropriate parameter for assessing relative brittleness of asphalt mixture.

PAGE 48

48 4.4 Determination of L oad L evel and R est P eriod Load level as a percent of the ultimate load from the Superpave IDT strength test is a pra ctical starting point in the determination of appropriate load levels for damaging due to consistency and simplicity of interpretation of strength tes t results. Since dimensions can vary slightly from specimen to specimen, load magnitudes were corrected using the following equations established in previous research (Roque and Buttlar, 1992; Buttlar and Roque, 1994; Roque et al., 1997b ): (4 4) Where: P applied is the applied load; S t is the strength from the Superpave IDT strength test; d is the specimen diameter; t is the specimen thickness; A is the selected load level; and CSX is a stress correction f actor to account for thr ee dimensional stress state. For initial trial testing, repeated load damage tests were performed on asphalt mixtures encompassing a range of brittleness, load level and rest period duration to identify appropriate load levels and rest periods for inducing micro damage in asphalt mixture. For these initial tests, mixtures were categorized into three groups based on broad ranges of M f as shown in Table 4 1. Two mixtures from each of the three groups were selected and tested based on the combinations presented in Table 4 2. A total of 210 repeated load damage tests were performed: 14 combinations of mixture, brittleness, and rest period*5 load levels*3 specimen replicates. All damage tests were performed for up to 40 minutes. Results f or the dense graded (D), unmodified

PAGE 49

49 (U), short term oven aged (S) mixture tested at 20C are presented in Figures 4 9 through 4 16. Results are presented in terms of normalized resilient modulus to eliminate specimen to specimen variation. In general, resu lts were in good agreement with expectations. Figures 4 10 and 4 16 clearly indicate that higher load level and/or shorter rest period result in greater rate of reduction in resilient modulus. It was also observed that higher load level and/or shorter rest period caused greater effects of heating and reversal of steric hardening, which may be d ifficult to separate from induced micro damage. From a practical point of view, higher load levels and/or shorter rest periods may cause excessively high rates of dam age making testing diffic ult to control ( Figures 4 10, 4 14 and 4 16). As stated earlier, another concern associated with short rest periods is delayed elasticity. With shorter rest periods, delayed elasticity is not given ample time to recover. This resu lts in an underestimatio n of resilient modulus when micro damage is allowed to recover or heal. Ideally, rest period duration should be long enough to allow most of the delayed elasticity to recover, yet short enough to minimize healing while inducing dama ge. As can be seen from Figures 4 9 and 4 12, the use of lower load level and/or longer rest period yield relatively long testing times, which is not desired; that is, it may take an excessively long time to accumulate a sufficient amount of damage in the asphalt mixture In addition, Figure 4 9 shows similar reduction in resilien t modulus for load levels of 20 percent to 30 percent of the ultimate strength for the 0.9 second rest period. This again appears to be evidence that healing occurring during the l onger rest periods is having a dominant effect on response. Based on these results, it was concluded that a 0.4 second rest period was most appropriate for all mixtures tested. With a 0.4 second rest period, delayed elasticity is less of an issue and testi ng times are not terribly long. From the results of Figures 4 9 through 4 16 load

PAGE 50

50 levels of 25, 30, and 45 percent of P fail were initially thought to be appropriate for the ductile, semi ductile and brittle grou ps, respectively as the se results suggested that there may have been a linear relationship between M f and load level for the six mixtures tested, as shown in Figure 4 17. This relationship, however, prove d to be inconsistent when additional damage tests were performed on various mixtures (those oth er than the six previously tested) when predicting load level for inducing appropriate amounts of micro damage. So, the relationship was used only as a general guide for selection of load level. Upon further refinement and analysis of load level, it was di scovered that the relationship be tween M f and load level was in fact logarithmic, not linear as previously thought. The modified relationship is shown in Figure 4 18 for the dense graded mixture and Figure 4 19 for the open graded mixture, and is based on all mixtures tested. It should be noted that load levels determined from this relationship do not necessarily result in equal damage rate for all mixtures, but rather result in sufficient damage being induced; that being, 10 to 25 percent reduction in resi lient modulus. From the thorough analysis of load level, it was ob served that mixtures with high M f were much more sensitive to chang es in load than those with low M f This phenomenon was also observed by Roque et al. (2012 a ) when evaluating effects of oxi dative aging on changes in effective modulus. As a result, to ensure sufficient reduction in resilient modulus, it is recom mended that mixtures with high M f values be tested at 40 percent of P fail or below. If a specific reduction in resilient modulus is d esired, load levels should be adjusted as necessary. 4.5 The Healing Potential Test The developed healing test consisted of two phases: a damage phase and a healing phase. For purposes of practicality, each phase was l imited to 30 min, but no restrictions were placed on testin g time as long as enough time was given to induce proper damage during the damage phase and allow recovery of damage during the healing phase. During the damage phase, loading was

PAGE 51

51 applied continuously while during the healing phase, load was applied only at selected times to obtain sufficient resilient modulus measurements to obtain healing rate. Load levels applied during the damage phase were determined based on brittleness of the asphalt mixture. To minimize potential damage in the asph alt mixture during the healing pha se, load level was reduced by 5 percent from load levels applied during the damage phase. The overall healing potential test is illustrated in Figure 4 20. Goals of the damage phase are summarized below: Induce sufficient damage to evaluate healing effect B e in the linear damage range (i.e., no induced macro damage) Allow l ittle or no healing during rest periods between load applications Accumulate l ittle or no delayed elasticity Have s hortest testing time possible Goals o f the he aling phase are summarized as follows : Induce little or no additional damage Have s hortest testing time possible T esting conditions for all asphalt mixture s can be seen in Table 4 3 In all, 72 healing potential tests were performed: 4 mixture typ es*2 levels of oxidative aging* 3 test temperatures* 3 specimen replicates. All healing potential test results can be found in Appendix D 4.6 Summary A test was developed to evaluate healing behavior of asphalt mixture The healing potential test (HPT) consi sts of two phases: a damage phase and a healing phase. During the damage phase, load is applied continuously such that micro damage is accumulated while remaining below the damage threshold ( section 2.2.1.4 ) During the healing phase, accumulated micro dam age is allowed to heal by ceasing loading of the asphalt mixture. Careful consideration was given to selection of appropriate load magnitudes and rest period duration by conducting a series of repeated load damage tests in which load magnitudes, rest perio d duration, and overall

PAGE 52

52 testing t ime were evaluated A parameter, modulus at fracture was developed to characterize relative brittleness of asphalt mixture Relationships between modulus at fracture and load level were established to further aid in the se lection of appropriate load magnitudes for inducing micro damage in asphalt mixture.

PAGE 53

53 Table 4 1 Gro uping of asphalt mixture using M f Group Mixture Condition M f Load Level (% of S t ) 1 Ductile M f < 1.0 20 to 40 2 Semi Ductile 1.0 < M f < 2.5 25 to 45 3 Brittle M f > 2.5 30 to 50 Table 4 2 Load level and rest period for initial trial testing Rest Period Mixture Condition Mixture Type Load Level (% of St) 0.9 sec Ductile DUS20 ORS20 20 25 30 35 40 Semi Ductile DML10 ORL10 25 30 35 40 45 Brittle DUL00 OMS00 30 35 40 45 50 0.4 sec Ductile DUS20 ORS20 20 25 30 35 40 Semi Ductile DML10 ORL10 25 30 35 40 45 Brittle DUL00 OMS00 30 35 40 45 50 0.1 sec Ductile DUS20 ORS20 20 25 30 35 40 Semi Ductile DML10 ORL10 25 30 35 40 45 Brittle DUL00 OMS00 30 35 40 45 50

PAGE 54

54 Table 4 3 Asphalt mixture testing conditions Mixture Type Aging Condition Temperature ( C) Dense Graded (PG 67 22) STOA 0 10 20 LTOA 0 10 2 0 Dense Graded (PG 76 22) STOA 0 10 20 LTOA 0 10 20 Open Graded (PG 76 22) STOA 0 10 20 LTOA 0 10 20 Open Graded (ARB 12) STOA 0 10 20 LTOA 0 10 20

PAGE 55

55 Figure 4 1 Accumulation o f delayed elasticity (static load) Figure 4 2 Accumulation of delayed elasticity (repeated load)

PAGE 56

56 Figure 4 3 Fatigue curve due to repeat load Figure 4 4 Havers ine waveform with varying rest period duration

PAGE 57

57 A B Figure 4 5 Illustration of brittleness index. A) I B B) I B,HMA

PAGE 58

58 Figure 4 6 Comparison of relative brittleness Figure 4 7 Modulus at fracture

PAGE 59

59 Note: O open graded, D dense graded, M modified binder (PG 76 22), U unmodified binder (PG 67 22), R asphalt rubber binder (ARB 12), L long term oven aging, S short term oven aging, 00 test temperature 0C, 10 test temperature 10C, and 20 test temperature 20C Figure 4 8 Ranking of asphalt mixtures by M f 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 Mixture Types

PAGE 60

60 Figure 4 9 Effect of load level with 0.9 second rest period (DUS20) Figure 4 10 Effect of load level with 0.1 second rest period (DUS20) Figure 4 11 Effect of load level with 0.4 second rest period (DUS20) Figure 4 12 Effect of rest period with 40% load level (DUS20)

PAGE 61

61 Figure 4 13 Effect of rest period with 25% load level (DUS20) Figure 4 14 Effect of rest period with 35% load level (DUS20) Figure 4 15 Effect of res t period with 30% load level (DUS20) Figure 4 16 Effect of rest period with 40% load level (DUS20)

PAGE 62

62 Figure 4 17 Original M f load level relationship Figure 4 18 Relationship between M f and load level for dense graded mixture y = 5.05x + 22.98 R = 0.97 0 10 20 30 40 50 0 1 2 3 4 5 Load Level (% St) Modulus at Fracture (M f ) y = 4.3928ln(x) + 37.034 R = 0.8424 0 10 20 30 40 50 0 1 2 3 4 5 Load Level (% S t ) Modulus at Fracture (M f )

PAGE 63

63 Figure 4 19 Relat ionship between M f and load level for open graded mixture Figure 4 20 Schematic of the healing potential te st y = 8.1607ln(x) + 28.476 R = 0.9385 0 10 20 30 40 50 0 1 2 3 4 5 Load Level (% S t ) Modulus at Fracture (M f )

PAGE 64

64 CHAPTER 5 TEST PROCEDURE AND DATA INTERPRETATION METHODS 5.1 Overview This chapter includes information regarding data acquisition, test procedure, and data interpretation for the healing potential test. The healing potential test utilizes the Superpave IDT system and as such, equipment and specimen preparation procedures used by Roque and Buttlar (19 92) and Roque et al. (1997b) were employed Standard Superpave IDT tests (resilient modulus, creep, and strength tests) preceded all healing potential tests to obtain essential magnitudes for testing. Setup for the healing potential test can be seen in Figure 5 1 5.2 Data Acquisition To adequately observe changes in effective modul us, it is imperative that data be acquired at a sufficient ly high rate. When performing the standard Superpave resilient modulus test, data is acquired at a rate of 500 points per cycle This is based on a load cycle duration of 1.0 second (0.1 second load pulse 0.9 second rest period). For the developed healing potential test, a 0.1 second load p ulse followed by a 0.4 second rest period, for a total load cycle duration of 0.5 second, was found to be an appropriate loading sequence. Since this modification reduced the overall load cycle duration from 1.0 second to 0.5 second, data is acquired at a rate of 1000 points per second or 500 points per cycle. Data acquisition occurred for both the damage phase and healing phase, separately. During the damage phas e, repeated loading was applied continuously; it is however unnecessary to record every load cycle during this phase With this in mind, data was recorded for six consecutive loading cycles at the times specified in Table 5 1. During the healing phase, l oad was applied only at discrete times as a means of obtaining recovery of effective modulus

PAGE 65

65 Thus, data was only acquired and recorded at these times which are shown in Table 5 2. Because more healing will take place during the beginning of the healing ph ase, it is important that emphasize be given to earlier portions of the healing phase. To deemphasize the minimal healing effect at longer healing times, data was acquired at times such that they were more or less equally spaced a part on natural log scale. 5.3 Testing Procedure The testing procedures are summarized in the following steps: 1) Superpave IDT specimens were prepared following procedures provided in Roque and Buttlar (1992) and Roque et al. (1997b). 2) Gage points (5/16 inch diameter by 1/8 inch thick) we re affixed to each specimen face with epoxy and allowed to dry for at least 24 hours prior to testing. 3) Specimens were placed in an environmental chamber set a t the desired testing temperature and left to condition for at least 6 hours. 4) Strain gages were m ounted on the gage points on each specimen face to measure horizontal and vertical deformations. 5) The test specimen was placed into the Superpave IDT loading frame. A seating load of approximately 10 pounds was applied to the test specimen to ensure proper contact between the specimen and loading heads. 6) The specimen was then loaded by applying a repeated haversine waveform (0.1 second load and 0.4 second rest period) for 30 min to induce micro damage in the specimen. The load was determined from the modulus at fracture versus load level relationship presented in Chapter 4. Test data were recorded by the computer software (data acquisition program) 19 times as directed in Table 5 1. Data was acquired at a rate of 500 points per cycle for 6 loading cycles. 7) The specimen was then unloaded and the healing phase began. The specimen was loaded by a pplying a repeated haversine waveform (0.1 second loading and 0.4 second rest period) for 6 cycles. The load level was reduced by 5 percent to minimize potential for addit ional micro damage to develop during the healing phase. Data was recorded at a rate of 500 points per cycle at 6 discrete times during the 30 min healing phase as shown in Table 5 2. 5.4 Damage Phase Data Interpretation Damage is observed by an apparent reduc tion in resilient modulus. The resilient modulus of an asphalt mixture is a stress strain relationship and can be defined as follows:

PAGE 66

66 (5 1) Where: r is the repeated stress; and r is the recoverable axial strain. Using equations developed by Roque et al. (1997 b ), resilient modulus for the standard Superpave IDT test, can be determined as follows: (5 2) Where: M R is the resilient modulus; GL is the gage length; P is peak load; D is the specimen diameter; t is the specimen thickness; and C cpml is a correction factor. Here, resilient mod ulus is calculated by using total recoverable deformation, which includes bot h the instantaneous and time dependent recoverable deformation during the unloading and rest period portion of each loading cycle ( Figure 5 2 atio from the standard Superpa ve IDT resilient modulus test was

PAGE 67

67 healing potential test may not be reliable since damage is introduced into the specimen (section 6.4) Typical horizontal deformation versus time data is presented in Figure 5 3 Using measured total recoverable deformation, resilient modulus was calculated at 19 times throughout the damage phase. Typical resi lient modulus data during the dam age phase is shown in Figure 5 4 Results were normalized with the initial resilient modulus value determined at the beginning of the damage phase to eliminate specimen to specimen variation. As reported by Grant (2001), mi xtures typically stabilized (exhibit ed linear behavior) after about 2 min of loading. The nonlinear behavior observed at the beginning of the damage phase is thought to be primarily caused by local elevation of temperature in the specimen and reversal of s teric hardening ( Mouillet et al., 2012; Tabatabee et al, 2012; Santagata et al., 2013 ). To avoid the nonlinear behavior observed at the beginning of the test, which is thought not to be damage, linear regression analysis was performed on data obtained from 2 min to 30 min. Damage rate is taken as the slope of the regression line shown in Figure 5 4 5.5 Healing Phase Data Interpretation Unlike the damage phase in which damage is characterized by a reduction in resilient modulus, in the healing phase healing i s characterized by a recovery or increase in resilient modulus. Analyzing data from the healing phase, data was used to both evaluate and quantify healing potential of the asphalt mixture. For the healing phase, a generalized power law function was used fo r fitting of the data. The power model function used here is the same used to fit data in the Superpave IDT creep curve fitting process This power mod el function is presented in equation 5 3. Typical healing phase data and data fitting are presented in Fi gure 5 5 (5 3) Where:

PAGE 68

68 MR(t) is the normalized resilient modulus at time t; MR 0 is the normalized resilient modulus at t equals zero; and MR 1 and m are power model parameters. Once the healing data was fit to the power model fun ction, the slope of the healing phase curve at time anytime t was obtained by taking the derivative of the power model function with respect to time as shown in following equation: (5 4) Where: is the instantaneous slope at any time t; and MR 1 and m are power model parameters as defined in equation 5 3. It was observed that as resilient modulus recovered, the slope of the healing phase curve decreased. It is reasonable to hypothesize that the slope of the healing phase curve will approach and reach zero only when the mixture is fully healed. In other words, the healing phase curve will reach asymptotic behavior when the mixture is fully healed thereby indicating that the mixture has reached its ful l healing potential. Because full healing did not always occur within the allotted 30 minute healing phase, the instantaneous slopes of the healing phase curve at times 10, 20 and 30 minutes were used to predict the normalized undamaged resilient modulus ( MR undamaged ); that is, the value at which full healing occurs. An example of this process is shown in Figure 5 6 The normalized undamaged resilient modulus is circled in red in Figure 5 6 In theory, the normalized undamaged resilient modulus should be eq ual to 1.0 specimen but may not be equal to 1.0 because of the presence of steric hardening.

PAGE 69

69 After MR undamaged was determined, the percentage of healing with time was calculated using equation 5 6 The percentage of healing can be cons idered a measure of how close a particular mixture is to being fully healed. (5 5) Where: MR(t) is the normalized resilient modulus at any time t; MR 0 is the normalized resilient modulus at time equal to zero; and MR undamaged is the undamaged normalized resilient modulus. A typi cal percentage of healing versus time relationship is presented in Figure 5 7 This relationship shows that healing rate (defined as the slope of percentage of healing time relationship) is time dependent. In other words, healing rate decreases as time increases. Since healing rate was found to be time depen dent, no single healing rate could be obtained f or each mixture As a result, a parameter was established to allow for relative comparison of healing poten tial among mixtures. T defined as the coefficient of the logarithmic function used to fit the percentage of healing time relationship ( Figure 5 7 ). Healing rate parameters for all mixtures tested are presented in Appendix E It should be noted that healing rate results were in agreement with those found by Kim and Roque (2006). In general, healing potential test results were reasonable and in good agreement with expected trends thereby supporting validity of the healing pot ential test. Results indicated that mixtures tested at higher temperatures had greater healing potential than those tested at lower temperatures. Likewise, STOA mixtures had greater healing potential than LTOA mixtures. For dense graded mixture, mixture wi th polymer modified binder had greater h ealing potential than mixture with unmodified binder.

PAGE 70

70 Table 5 1 Data acquisition times for damage phase Acquisition # Time (sec) 1 0.67 2 6.67 3 13.3 4 26.67 5 40 6 60 7 80 8 120 9 160 10 240 11 320 12 400 13 600 14 800 15 1000 16 1200 17 1400 18 1600 19 18000 Table 5 2 Data acquisition times for healing phase Acquisition # 1 2 3 4 5 6 Time (sec) 0 120 300 600 1200 1800

PAGE 71

71 A B Figure 5 1 HPT setup A) F ront view B) S ide view

PAGE 72

72 Note: V T is the total recoverable vertical deformation and H T is the total recoverable horizontal deformation Figure 5 2 Typical load and deformation response for the healing potential test

PAGE 73

73 Figure 5 3 Typical horizontal deformation measurements Figure 5 4 Typical resilient modulus data during the damage phase y = 0.003x + 0.9339 R = 0.9349 0 0.2 0.4 0.6 0.8 1 1.2 0 5 10 15 20 25 30 35 Normalized MR Time(min)

PAGE 74

74 Figure 5 5 Fitting of data from healing phase Figure 5 6 Determination of MR undamaged 0.75 0.8 0.85 0.9 0.95 1 0 5 10 15 20 25 30 35 Normalized MR Time(min ) 0C 10C 20C y = 0.7901+0.0956x 0.1380 y = 0.8541+0.0185x 0.4211 y = 0.7895+0.0256x 0.2897 y = 30.258x + 0.9637 R = 0.9968 0.880 0.890 0.900 0.910 0.920 0.930 0.940 0.950 0.960 0.970 0.0000 0.0005 0.0010 0.0015 0.0020 0.0025 Normalized MR Instantaneous Slope 30 min. 20 min. 10 min.

PAGE 75

75 Figure 5 7 Typical percentage vs. time with fit 0 20 40 60 80 100 0 5 10 15 20 25 30 35 Percentage of Healing Time(min) y = log 10 (a*x 0.75 +1) a healing rate parameter

PAGE 76

76 CHAPTER 6 NON HEALABLE MICRO DAMAGE IN ASPHALT MIXTURE 6.1 Background Until this point, micro damage development in asphalt mixture has been de scribed as being a completely reversible process. Micro damage in asphalt mixture has also been assumed to be purely l oad associated in nature These views are not unique to this research, but are shared by others as well ( Chapters 1 and 2). I n this study, a new mechanism to explain non healable micro damage one involving differential thermal contraction, was proposed. 6.2 Differential Thermal Contraction in Asphalt Mixture It is known that asphalt mixture properties change over time (Roque et al., 2011; Roque et al., 2012a). These changes to some extent can be attributed to load, heat oxidation, and moisture. associated binder hardening. However, aging can be thought of more liberally as any detrimental effect on asphalt mixture properties during pavement life With this definition, the three effects listed above can then be seen as having part in damaging asphalt mixture over time To more effectively simulate asphalt mixture aging in the field, it is important that all factors contributing to the aging process be considered in laboratory settings. One potential source of aging in field pavements may be the effect of temperature related internal micro damage; more specifically, the effe ct of differential thermal contraction between the asphalt binder and aggregate components Coefficients of thermal contraction for asphalt binder are much higher than that of its aggregate counterpart ( Table 6 1) As such, d uring cooling, the asphalt bin der will try to contract more than the aggregate due to the considerable difference in coefficient of thermal contraction. The binder, however, is limited by the contraction of the aggregate and as a result, tensile

PAGE 77

77 thermal stresses will develop within the asphalt binder and shear stress will develop along the asphalt aggregate interface. At a cert ain point, these stresses are unable to relax leading to cracking of the asphalt binder. The resulting tensile thermal stresses should not be confused with global tensile thermal stresses that can result in thermal transverse pavement cracking. It is thought that the difference in coefficient of thermal contraction between the asphalt binder and aggregate may lead to the development of permanent micro damage throug h loss of adhesion. It should be noted that while pavements in the state of Florida do not experience particularly low temperatures, pavement s do, however, experience large temperature differentials and high rates of oxidative aging ( Lytton et al., 1993; Z ou and Roque, 2011 ). 6.3 Experimental Investigation into Temperature Associated Micro D amage A n attempt was made to establish appropriate procedures for inducing permanent micro damage in asphalt mixture through temperature cycling. Healing potential tests wer e conducted on specimens conditioned at two extreme temperatures: 30C and +20C. The low temperature ( 30C) was chosen in an attempt to replicate the behavior observed by Deme and Young of decreasing tensile strength after peaking at low temperature ( se ction 2.5) All HPTs were performed at 20C in an effort to promote healing of any induced micro damage (permanent or oth erwise) Recall, as shown in section 5.5 and Appendix E healing rate was found to be greatest for the mixtures tested at 20C. Testing conditions for thermal conditioning of the asphalt mixture can be seen in Table 6 2 and an overall testing plan is shown in Figure 6 1. A more detailed research approach is also shown in Figure 1 2. It should be noted that t here are several test methods w hich include freeze thaw con ditioning in some form or other, m ost of which involve vacuum saturating specimens before the freeze thaw process (AASHTO, 2007; FDOT, 2010b) However these tests were designed to assess moisture susceptibility of the asphalt m ixture not temperature related damage per se In reality, there are two possible ways of

PAGE 78

78 evaluating the effects of freeze thaw: (1) with the presence of water and (2) without the presence of water. The second which has not been studied, may be more realis tic for unsaturated pavement s ubjected to colder temperatures and/or large temperature differentials. Consequently, it was the scenario chosen for this study For this phase of the study, testing was limited to unmodified, dense graded, LTOA mixture. LTOA mixture is more susceptible to cracking, thus making it more likely for micro damage to develop. Because, temperatures below 0C were needed to condition specimens, a of the freeze thaw conditioning process as shown in Figure 6 2. Since the low temperature bath operates using oil, special care was given to isolate specimens from the oil using polyethylene plastic as shown in Figure 6 3. Masking tape was also wrapped ar ound the specimen edges to prevent puncture of the polyethylene plastic during conditioning as shown in Figur e 6 4. After specimens were immersed at 30C for the desired conditioning time ( Table 6 2 ), a water bath as shown in Figure 6 5. Again, polyethylene plastic was used to prevent specimens from becoming saturated with water. Once water is introduced into the asphalt aggregate system, the damage mechanism is changed as water expands upon freezing resulting in d amage of a different nature. It is important to note that specimens were not allowed to gradually reach the desired conditioning temperatures, but rather were immediately exposed to the two extreme temperatures. It was ield the greatest probability of inducing non healable micro damage in asphalt mixture. Results for the three conditioning times are shown in Figures 6 6 through 6 8. In Figures 6 7 and 6 8, a clear reduction in initial modulus was observed before and afte r conditioning. This observation seemed to indicate the presence of permanent micro damage in the asphalt mixture.

PAGE 79

79 In the case of the single 16 hour conditioning period, further loading was applied to the specimens to evaluate how micro damage would evolve over time. HPTs were performed once a day with at least 24 hours in between successive tests in an effort to promote healing of the asphalt mixture. Figure 6 9 shows the change in resilient modulus over the 10 day process. To verify that the continued red uction in initial modulus was indeed due to permanent micro damage, the same process was conducted on specimens not subjected to thermal conditioning. As seen in Figure 6 10 the same effect was observed for specimens without thermal conditioning meaning t hat any reduction in resilient modulus could not be explained by thermal conditioning alone. Damage rate ( section 5.4 ) was also examined to see if any differences could be observed between the unconditioned and conditioned specimens, but alas as seen in Fi gure 6 1 1 n o significant difference was found And while thermal conditioning was found to have some effect on resilient modulus (Figures 6 6 through 6 8) thermal conditioning alone did not explain the reduction in initial modulus observed when performin g successive HPTs ( Figure 6 9 ) 6.4 Given the unexpected results presented in Figure 6 10, it was immediately questioned whether or not the reduction in initial modulus observed from day to day w as indeed the result of permanent micro damage or if the effect was just an apparent one due to analysis of the results. It was hypothesized that the reduction in initial resilient modulus observed from day to day could be due to either : 1) micro damage ac cumulation from repeated load either within the asphalt mastic or at the interface between the asphalt mastic and aggregate or 2) reorientation of aggregate particles due to accumulation of large permanent deformation during repeated load. With diametrical ly loaded specimens such as those used in the HPT and Superpave IDT, ult

PAGE 80

80 because measurements are taken from head to head or end to end unlike with the Superpave IDT in which measurements are taken over a smaller area where stresses and strains are known to be fairly uniform. I n the procedures established by Roque and Buttl ar (1992) and Roque et al. that is calculated using measured horizontal and vertical deformations. When initially developing the HPT, a decision was made to assu modulus because it was felt that the HPT may not have been reliable once damage was introduced into the asphalt mixture (section 5.4). Assuming the data presented in Figure 6 10, is the result of micro damage accumulation, then the decision to change in the asphalt material or aggre gate structure However, if the data present ed in Figure 6 10 is the result of aggregate reorientation due to accumulation of large permanent deformation, then allowing using deformations measured during repeated load may be more appropriate given the significant cha nge in aggregate structure that would occur. In the following 6.4.1 Constant Poiss Using an assumed (determined from the standard Superpave IDT resilient modulus test ) the asphalt mixture was loaded in two ways: with rest periods (successive HPTs) and without rest periods (conventional fatigue) It should be no ted that testing was performed at 10 C in an effort to limit overall strain. As shown in Figure 6 12 independent of loading procedure failure of the mixture was found to occur at the same reduction in normalized resilient mod ulus (approximately 40 percen t). While k nowing the value of resilient modulus at which

PAGE 81

81 failure occurs is useful, even more useful, is knowing how close a particular mixture is to failure during repeated loading over time. Previous work (Zhang et al 2001; Roque et al., 2002 ) has show n DCSE f (the energy threshold) in addition to fracture energy (FE), to be a fundamental property of asphalt mixture; meaning, that regardless of mode of loading (strength, cyclic, or creep), DCSE f remains the same (Figure 6 1 3 ). In the HMA fracture mechani cs approach (section 2.2.1.4), damage is characterized by DCSE Using the HMA fracture mechanics model the accumulated DCSE per load cycle (equation 2 3) can be predicted for a given asphalt mixture (Figure 4 5(b)). Comparing accumulated DCSE (DCSE/cycle number of load cycles) during the HPT to DCSE f of the mixture, it was discovered that the two values were close (Figure 6 1 4 ) In theory, the se two values should be exactly equal, but in reality it is very unlikely that they would ever be equal for a few reasons. The first has to do with machine error. I t is very difficult for servo hydraulic systems to instantaneously apply and remove load through hydraulic transmission so in actuality the haversine waveform used in the HPT is never applied perfectly. S econd, t he DCSE result s shown in Figure 6 1 4 are based on IDT properties (equations 6 1 through 6 3) in which failure strain among specimen replicates can vary significantly yet an average value is used in all calculations Third, elastic energy used in t he determination of DCSE f is not actually measured, but rather is approximated using resilient modulus. While these potential errors may not be great on the ir own together they may explain the small difference in accumulated DCSE and DCSE f observ ed in Fi gure 6 14 In any case, the two values do not differ significantly and the DCSE threshold is known to be exceeded in either case simply because fracture is known to have occurred (Figure 6 12 )

PAGE 82

82 (6 1) (6 2) (6 3) Where, FE is the total energy required to fracture the mixture; f is the failure strain; S ( ) is the stress at a given strain level; EE is the e lastic energy ; S t is the tensile strength from the Superpave IDT strength test; and DCSE f is the accumulated dissipated creep strain energy to failure. 6.4.2 When developing the HPT, one of the more pressing issues was making sure that su fficient micro damage was induced into the asphalt mixture such that recovery of the mic ro damage could be measured. This issue was resolved by inducing as much micro damage as possible while still remaining below the damage threshold (section 2.2.1.4). Be cause of this, little attention was given to strain level. Recall, from Chapter 2 that micro damage is considered to be completely healable if below the damage threshold. Upon further investigation of the results presented in Figure 6 10, it was noticed th at deformation accumulated during any given HPT was relatively large. Figure 6 15 shows accumulated permanent deformation (deformation which was not recovered) for each day of testing. These results are particularly noteworthy because the permanent strain accumulated during just one HPT at times exceeds the strain required to fail the mixture when performing the Superpave IDT strength test (Appendix C). In other words, even

PAGE 83

83 though more strain was accumulated when performing the HPT, failure of the mixture d id not occur. In an effort to limit overall strain, testing was als o performed at 10C. Figure 6 16 shows HPT results for mixture tested at 10C. As seen in the figure, the trend of decreasing initial resilient modulus is still evident, but not as great as that at 20C. Comparing permanent deformations for the tests perform ed at 10C and 20C (Figure 6 17 ), it is clear that permanent deformation was significantly reduced by performing te sts at 10C. However, strain accumulated to failure for the HPT was st i ll greater than that for the IDT strength test (Table 6 3) and perhaps more importantly, the trend of decreasing initial resilient modulus (Figure 6 16 ) was still evident. As stated previously, it was hypothesized the apparent reduction in initial resilien t modulus observed when performing successive HPTs might be the result of decisions made regarding the determination of resili ent modulus As a result, the data presented in Figure 6 16 was reanalyzed usi As shown in Figure 6 18 ratio was allowed to change during repeated load, then the reduction in i nitial modulus observed from day to day in Figure 6 16 was no longer present These results could be interpreted to mean that there was no t an y micro damage accumulated duri ng the HPT or at least very little and would support the theory of aggregate reorientation during repeated load However, the se results would also imply that micro damage accumulation does not precede fracture which seems ve ry unlikely. Fig ure 6 19 during the HPT. As seen in the f It was concluded that th is in fact not a repre sentation of the actual material, but was rather a reflection of micro damage accumulated during repeated load since tensile response is more affected by damage than compressive response with the HPT In addition, it was noticed

PAGE 84

84 sed during the healing phase of the HPT, which would support the theory of micro damage healing and refute the theory of aggregate reorientation, as a reversal of aggregate reorientation seems very unlikely. 6.5 Load In duced Permanent Micro Damage The analyses performed in sections 6.4.1 and 6.4.2 highlight the importance of selecting a with diametrically loaded specimens. Given the findings of the above analyses it was concluded that the reducti on in initial modulus observed when performing successive HPTs was the result of load induced permanent micro damage calculations of resilient modulus was correct Even though strain accumula ted d uring the HPT far exceeded that accumulated during the IDT strength test DCSE accumulated to failure was approximately the same for the two test methods. In addition, failure of the mixture was found to occur at the same reduction in resilient modulu s independent of how the mixture was loaded. It was felt that these observations together supported the hypothesis that the reduction in initial modulus observed when performing successive HPTs was due to micro damage accumulation from repeated load either within the asphalt mastic or at the asphalt aggregate interface. Given previous efforts ( Zhang et al., 2001; Roque et al., 2002 ; Kim and Roque, 2006) have shown load induced micro damage to be completely healable, these findings may suggest the existence of a micro damage threshold. 6.6 Summary A new mechanism to explain micro damage development in asphalt mixture was proposed. It was proposed that differential thermal contraction in asphalt mixture could result in permanent micro damage in asphalt mixture. Wh ile results indicated that subjecting asphalt mixture to temperature cycling did have an effect on resilient modulus measurements

PAGE 85

85 temperature cycling a lone did not explain the development of permanent micro damage in asphalt mixture when performing the HP T. R esults did however, indicate the presence of load induced permanent micro damage which may suggest the existence of a micro damage threshold.

PAGE 86

86 Table 6 1 Average linear coefficient of thermal expansion for materials used in asphalt mixture Material Linear Coefficient of Thermal Expansion ( 10 6 /C) Granites and gneisses* 6.5 to 8.5 Dense, crystalline porous limestones* 3.5 to 6.0 Asphalt binder ~ 600 Asphalt mixture 20 to 30 *Lamond and Pi elert, 2006; 2003 Table 6 2 Temperature cycling conditions for asphalt mixture Thermal Condition Temperature Range (C) Number of Thermal Cycles Conditioning Time (hr) Freeze Thaw 30 to 20 1 1 Freeze Thaw 30 to 20 1 16 Freeze T haw 30 to 20 4 16 Table 6 3 Comparison of strain to failure for Superpave IDT and HPT Temperature (C) f, IDT f, HPT 10 1336.78 10055.14 20 2838.78 34578.57

PAGE 87

87 Figure 6 1 Te sting plan for assessing permanent micro damage in asphalt mixture A B Figure 6 2 Low temperature bath. A) Equ ipment. B) S pecimens conditioning in bath.

PAGE 88

88 Figure 6 3 Specimen in polyethylene plastic bag Figure 6 4 Masking tape around specimen edges

PAGE 89

89 Figure 6 5 Specimens c onditioning in water bath Figure 6 6 HPT results before and after 1 hour of thermal conditioning

PAGE 90

90 Figure 6 7 HPT results before and after 16 hours of thermal conditioning Figure 6 8 HPT results before and after four 16 hour cycles of thermal conditioning

PAGE 91

91 Figure 6 9 Evolution of resilient modulus when performing successive HPTs (w/thermal conditioning) Figure 6 10 Evolution of resilient modulus when performing successive HPTs (w/o thermal conditioning)

PAGE 92

92 Figure 6 11 Damage rate for unconditioned and conditioned specimens A 0.000 0.002 0.004 0.006 0.008 0.010 1 2 3 4 5 6 7 8 9 10 11 Damage Rate Day unconditioned conditioned 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 60 70 Normalized MR Time(min) Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 point of failure

PAGE 93

93 B Figure 6 12 F ailure of asphalt mixture. A) Using HPT B) Using conventional fatigue Figure 6 13 DCSE concept 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 0 50 100 150 200 250 Normalized MR Time(min) point of failure

PAGE 94

94 Note: D dense graded, U unmodified binder (PG 67 22), L long term oven aging, 10 test temperature 10C, and 20 test temperature 20C Figure 6 14 Comparing accumulated DCSE and DCSE f Figure 6 15 Permanent strain resulting from HPT (20C) 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 DUL20 DUL10 DCSE (kJ/m 3 ) Mixture Type DCSE_accumulated DCSE_final 0 1000 2000 3000 4000 5000 6000 7000 1 2 3 4 5 6 7 8 9 Permanent Deformation ( in) Day

PAGE 95

95 Figure 6 16 Evolution of resilient modul us when performing successive HPTs (10C) Figure 6 17 Permanent strain resulting from HPT performed at 10C and 20C 0 1000 2000 3000 4000 5000 6000 7000 1 2 3 4 5 6 7 8 9 Permanent Deformation ( in) Day 20C 10C

PAGE 96

96 Figure 6 18 Determination of resilient modulus using a calculated Poisson 's ratio Figure 6 19 Calculated Poisson's ratio during the HPT 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 0 10 20 30 40 50 60 70 Normalized MR Time(min) Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 0.30 0.32 0.34 0.36 0.38 0.40 0.42 0.44 0.46 0.48 0.50 0 10 20 30 40 50 60 70 Poisson's Ratio Time(min) Day 1 Day 2 Day 3 Day 4 Day 5 Day 6

PAGE 97

97 CHAPTER 7 COMPONENTS OF MICRO DAMAGE AND MICRO DAMAGE HEALING In this study, healable and non healable micro damage in asphalt mixture were characterized by changes in resilient modulus. Piecing together the findings of th is work, change in resilient modulus observed during the healing potentia l test can be attributed to five phenomena: 1) Reversal of steric hardening/local elevation of temperature : The beginning of the HPT is marked by a rapid nonlinear decrease in resilient modulus. Common wisdom states that this behavior is most likely due to rev ersal of steric ha rdening ( Mouillet et al., 2012; Tabatabee et al, 2012; Santagata et al., 2013 ) To a lesser extent, there a re those who believe that a rise in temperature takes place as well with perhaps Di Benedetto et al. (2011) having done the most extensive work on the matter. The effects of both steric hardening and temperature change are thought to be completely reversib le. 2) Micro damage accumulation: After about 2 minutes of repeated load, rate of reduction in resilient modulus decreases and a linear rate of reduction is observed Micro damage accumulated during this period is thou ght to be reversible. 3) Healing of micro d amage: After a period of continuous, repeated load during the damage phase, healing of the asphalt mixture is marked by a nonlinear increase in resilient modulus during the healing phase. 4) Permanent micro damage: The portion of resilient modulus not rec ove red during the healing phase 5) Development of s t eric hardening: The portion of re silient modulus not immediately recovered. This remaining portion is time dependent and is completely reversible These five phenomena are illustrated in Figure 7 1.

PAGE 98

98 Figu re 7 1 Effects responsible for changes in resilient modulus during the HPT

PAGE 99

99 CHAPTER 8 CLOSURE 8.1 Summary and Findings In this study, healable and non healable micro damage were characterized by changes in resilient modulus during repeate d load of asphalt mixture. To allow for both qualitative and quantitative ev aluation of the effects of micro damage and micro damage healing, a test was conceived, developed, and evaluated for a range of asphalt mixture. Furthermore, testing protocol and d ata reduction and interpretation methods were established to obtain healing rate in a standardized manner. The healing potential test consists of a damage phase during which micro damage is accumulated, indicated by a reduction in resilient modulus, and a healing phase during which micro damage is recovered, indicated by an increase in resilient modulus. In addition, a mechanism involving differential thermal contraction was proposed to explain non healable micro damage development in asphalt mixture. For t he mixtures and conditioning procedures used in this study, HPT results were found to be inconclusive when assessing non healable micro damage development through temperature cycling of asphalt mixture Results did, however, indicate the presence of load a ssociated non healable micro damage in asphalt mixture. Whereas load induced permanent micro damage was not found to exist in previous studies, it was found to exist in this study which may suggest the existence of a micro damage threshold. Findings assoc iated with this study are summarized as follows: Accumulated delayed elasticity during repeated load may cause error in determination of resilient modulus. This error is greater with shorter rest periods which allow less time for recovery of delayed elasti city and less for longer rest periods which allow more time for recovery of de layed elasticity However, shorter rest periods allow for greater number of load cycles and hence, greater rate of damage for a given period of time. An intermediate rest period (0.4 second) was found to result in a n acceptable balance between accumulated delayed elasticity and micro damage accumulation while providing a controllable damage development and a longer steady state damage range.

PAGE 100

100 Modulus at facture, which was defined a s strength divided by failure strain, was found to successfully s ort relative asphalt mixture brittleness in terms of temperature and degree of oxidative aging. The natural logarithmic relationship established between modulus at fracture and load level was found to successfully select appropriate load levels for use during the damage phase of the HPT. Healing rate where healing was defined as recovery of resilient modulus, was found to not be constant, but rather varied with time. Healing rates determined with the HPT were found to be in agreement with expected trends thereby validating the developed test. 8.2 Conclusions After comprehensive evaluation of results from the healing potential test s performed on all asphalt mixtures tested in this study, the foll owing conclusions can be drawn: The developed healing potential test can successfully evaluate healing characteristics of asphalt mixture as well as quantify healing rates of asphalt mixture. Healing rate parameter, a, can be used to evaluate asphalt mixtu re healing potential. The parameter modulus at fracture introduced in this study can be used to determine appropriate load levels for use in inducing micro damage in asphalt mixture Both healable and non healable micro damage exists in asphalt mixture both of which are load associated. 8.3 Recommendations Based on extensive evaluation throughout this study the following items are recommended for further investigation of the effects of micro damage and micro damage healing in asphalt mixture : For the dense grad ed mixture used in this study, permanent micro damage induced by differential thermal contraction of the asphalt mixture was inconclusive for conditions evaluated Thermal conditioning of a different asphalt mixture one that has marginal adhesive propert ies should be evaluated in an attempt to successfully induce permanent micro damage through temperature cycling.

PAGE 101

101 A combination of oxidative aging, thermal co nditioning, and cyclic pore pressure conditioning should be used to age the asphalt mixture in an attempt to induce permanent micro damage. A strength based approach as opposed to a repeated load approach should be used to assess possible permanent micro damage induced by differential thermal contraction of the asphalt mixture X ray computed tomograp hy should be explored to exam ine its potential to identify permanent micro damage in asphalt mixture. This would allow for better understanding of the nature of micro damage and micro damage healing. Results indicate not all load induced micro damage may b e fully healable. An alternate and more direct measure of damage and healing by way of failure limits (strength, failure strain, fracture energy dissipated creep strain energy ) should be explored to further evaluate the findings of this study based on res ilient modulus. Effects of steric hardening heat generation, and delayed elasticity should be studied in more detail A broader range of factors/mixture parameters ( e.g. binder type dominant aggregate size range porosity, disruption factor, effective fil m thickness, ratio between coarse portion of fine aggregate and fine portion of fine aggregate, etc ) should be evaluated using the developed healing potential test.

PAGE 102

102 APPENDIX A ASPHALT MIXTURE INFORMATION

PAGE 103

103 Table A 1 Dense gradation jo b mix formula (JMF) Sieve Size 33% 7% 50% 10% 100% # 78 Stone # 89 Stone W 10 Screenings Local Sand JMF 3/4" 100.0 100.0 100.0 100.0 100.0 1/2" 97.0 100.0 100.0 100.0 99.0 3/8" 59.0 99.7 100.0 100.0 86.5 # 4 9.0 30.0 100.0 100.0 65.1 # 8 4.0 4.0 70. 0 100.0 46.6 # 16 2.0 2.0 42.0 100.0 31.8 # 30 2.0 1.0 25.0 94.0 22.6 # 50 1.0 1.0 16.0 53.0 13.7 # 100 1.0 1.0 10.0 11.0 6.5 # 200 1.0 1.0 7.0 3.0 4.2 G sb 2.809 2.799 2.770 2.626 2.770 Table A 2 Open gradation job mix formula (JMF) Sieve Size 44.7% 49.4% 3.2% 2.7% 100% S1A Stone S1B Stone Screenings Filler JMF 3/4" 100.0 100.0 100.0 100.0 100.0 1/2" 79.0 100.0 100.0 100.0 90.6 3/8" 36.0 92.0 100.0 100.0 67.4 # 4 7.0 26.0 100.0 100.0 21.9 # 8 3.0 7.0 68.0 100 .0 9.7 # 16 3.0 3.0 67.0 100.0 7.7 # 30 3.0 3.0 55.0 100.0 7.3 # 50 3.0 2.0 35.0 100.0 6.1 # 100 2.0 2.0 14.0 100.0 5.0 # 200 1.0 1.0 3.0 100.0 3.7 G sb 2.425 2.451 2.527 2.600 2.445

PAGE 104

104 Table A 3 Dense gradation batch wei ghts (cumulative) Sieve Size Retained Weight, g # 78 Stone # 89 Stone W 10 Screenings Local Sand 3/4" 0.0 1485.0 1800.0 4050.0 1/2" 44.6 1485.0 1800.0 4050.0 3/8" 608.9 1485.9 1800.0 4050.0 # 4 1351.4 1705.5 1800.0 4050.0 # 8 1425.6 1787.4 2475.0 40 50.0 # 16 1455.3 1793.7 3105.0 4050.0 # 30 1455.3 1796.9 3487.5 4077.0 # 50 1470.2 1796.9 3690.0 4261.5 # 100 1470.2 1796.9 3825.0 4450.5 # 200 1470.2 1796.9 3892.5 4486.5 Pan 1485.0 1800.0 4050.0 4500.0 Sum 1485.0 315.0 2250.0 450.0 Table A 4 Open gradation batch weights (cumulative) Sieve Size Retained Weight, g S1A Stone S1B Stone Screenings Filler 3/4" 0.0 2011.5 4234.5 4378.5 1/2" 422.4 2011.5 4234.5 4378.5 3/8" 1287.4 2189.3 4234.5 4378.5 # 4 1870.7 3656 .5 4234.5 4378.5 # 8 1951.2 4078.9 4280.6 4378.5 # 16 1951.2 4167.8 4282.0 4378.5 # 30 1951.2 4167.8 4299.3 4378.5 # 50 1951.2 4190.0 4328.1 4378.5 # 100 1971.3 4190.0 4358.3 4378.5 # 200 1991.4 4212.3 4374.2 4378.5 Pan 2011.5 4234.5 4378.5 4500.0 Sum 2011.5 2223.0 144.0 121.5

PAGE 105

105 APPENDIX B BULK SPECIFIC GRAVITY RESULTS

PAGE 106

106 Table B 1 Bulk specific gravity test results for open graded mixture with SBS modified binder Sample ID A B C D E F G H I J K L M Bag Weight (g) Weight before Sealing (g) Sealed Sample Weight in Water (g) Weight after Water Submersion (g) Ratio B/A Bag Volume Correction from Table Total Volume (A+D) C Volume of Bag A/F Volume of Sample G H G mb (g/cm 3 ) Maximum Specific Gravity Air Voids (%) Average G mb OMS_A 46.3 4752.9 2360.0 4752.9 102.654 0.689 2439.2 67.18 2372.02 2.004 2.309 13.22 1.98 OMS_B 46.3 4752.9 2345.0 4752.9 102.654 0.689 2454.2 67.18 2387.02 1.991 2.309 13.77 OMS_C 46.3 4765.3 2347.0 4765.3 102.922 0.689 2464.6 67.2 2 2397.38 1.988 2.309 13.91 OMS_D 47.3 4772.7 2350.1 4772.7 100.903 0.692 2469.9 68.34 2401.56 1.987 2.309 13.93 OMS_E 47.2 4789.9 2341.3 4789.9 101.481 0.691 2495.8 68.29 2427.51 1.973 2.309 14.54 OMS_F 46.3 4769.5 2345.2 4769.5 103.013 0.689 2470.6 67.24 2403.36 1.985 2.309 14.05 OMS_G 46.3 4766.9 2309.0 4766.9 102.957 0.689 2504.2 67.23 2436.97 1.956 2.309 15.28 OML_A 46.3 4770.7 2319.2 4770.7 103.039 0.689 2497.8 67.24 2430.56 1.963 2.309 14.99 1.95 OML_B 46.3 4757.4 2281.0 4757.4 102.752 0.6 89 2522.7 67.20 2455.50 1.937 2.309 16.09 OML_C 46.3 4757.4 2251.6 4757.4 102.752 0.689 2552.1 67.20 2484.90 1.915 2.309 17.08 OML_D 47 4770.0 2323.1 4770.0 101.489 0.691 2493.9 68.00 2425.90 1.966 2.309 14.84 OML_E 47.2 4782.3 2324.3 4782.3 101.320 0.691 2505.2 68.27 2436.93 1.962 2.309 15.01 OML_F 46.3 4764.7 2306.8 4764.7 102.909 0.689 2504.2 67.22 2436.98 1.955 2.309 15.32 OML_G 46.3 4771.8 2326.5 4771.8 103.063 0.689 2491.6 67.25 2424.35 1.968 2.309 14.76

PAGE 107

107 Table B 2 Bulk specific gravity test results for open graded mixture with asphalt rubber binder Sample I D A B C D E F G H I J K L M Bag Weight (g) Weight before Sealing (g) Sealed Sample Weight in Water (g) Weight after Water Submersion (g) Ratio B/A Bag Volu me Correction from Table Total Volume (A+D) C Volume of Bag A/F Volume of Sample G H G mb (g/cm 3 ) Maximum Specific Gravity Air Voids (%) Average G mb ORS_A 49.85 4770.9 2320.0 4770.9 95.705 0.701 2500.75 71.14 2429.61 1.964 2.309 14.96 1.97 ORS_B 49.85 4770.9 2324.7 4770.9 95.705 0.701 2496.05 71.14 2424.91 1.967 2.309 14.79 ORS_C 49.85 4804.8 2335.9 4804.8 96.385 0.700 2518.75 71.25 2447.50 1.963 2.309 14.98 ORS_D 49.85 4804.8 2320.3 4804.8 96.385 0.700 2534.35 71.25 2463.10 1.951 2.309 15.52 ORS_E 49.85 4794.9 2337.0 4794.9 96.187 0.700 2507.75 71.22 2436.53 1.968 2.309 14.77 ORS_F 49.85 4794.9 2337.7 4794.9 96.187 0.700 2507.05 71.22 2435.83 1.968 2.309 14.75 ORS_G 47.5 4790.4 2390.0 4790.4 100.851 0.692 2447.9 68.62 2379.28 2.013 2.309 12.80 ORS_H 47.5 4790.4 2390.0 4790.4 100.851 0.692 2447.9 68.62 2379.28 2.013 2.309 12.80 ORS_I 46.3 4785.5 2299.5 4785.5 103.359 0.688 2532.3 67.29 2465.01 1.941 2.309 15.92 ORS_J 46.3 4782.9 2321.0 4782.9 103.302 0.688 2508.2 67.28 2440.92 1.959 2.309 15.14 ORL_A 49.85 4791.3 2305.8 4791.3 96.114 0.700 2535.35 71.21 2464.14 1.944 2.309 15.79 1.96 ORL_B 49.85 4791.3 2312.7 4791.3 96.114 0.700 2528.45 71.21 2457.24 1.950 2.309 15.55 ORL_C 49.85 4784.0 2323.6 4784.0 95.968 0.700 2510.25 71 .18 2439.07 1.961 2.309 15.05 ORL_D 49.85 4784.0 2323.6 4784.0 95.968 0.700 2510.25 71.18 2439.07 1.961 2.309 15.05 ORL_E 49.85 4756.6 2311.7 4756.6 95.418 0.701 2494.75 71.09 2423.66 1.963 2.309 15.00 ORL_F 49.85 4756.6 2318.0 4756.6 95.418 0.701 24 88.45 71.09 2417.36 1.968 2.309 14.78 ORL_G 49.85 4810.9 2353.4 4810.9 96.508 0.699 2507.35 71.28 2436.07 1.975 2.309 14.47 ORL_H 49.85 4810.9 2358.0 4810.9 96.508 0.699 2502.75 71.28 2431.47 1.979 2.309 14.31 ORL_I 46.3 4780.7 2319.5 4780.7 103.255 0.688 2507.5 67.28 2440.22 1.959 2.309 15.15 ORL_J 46.3 4785.2 2298.0 4785.2 103.352 0.688 2533.5 67.29 2466.21 1.940 2.309 15.97

PAGE 108

108 APPENDIX C SUPERPAVE IDT TEST RESULTS

PAGE 109

109 Table C 1 Superpave IDT test results Mixture Aging Temp. m value D 1 (1/psi) St (MPa) M R (GPa) FE (kJ/m3) DCSE f (kJ/m3) Creep Rate (1/psi sec) D(t) (1/GPa) Failure Strain Dense Graded (PG 67 22) STOA 0C 0.529 2.25E 07 3.05 15.28 1.80 1.50 4.59E 09 1.330 870.23 10C 0.668 4.77E 07 2.14 10.85 4.20 3.99 3.20E 08 7.055 2566.05 20C 0.740 2.00E 06 1.16 6.34 2.90 2.79 2.46E 07 47.994 3556. 95 LTOA 0C 0.479 1.66E 07 3.56 18.43 1.70 1.36 2.18E 09 0.735 751.07 10C 0.532 4.48E 07 2.25 11.99 2.20 1.99 9.43E 09 2.619 1336.78 20C 0.680 1.07E 06 1.19 6.37 2.50 2.39 7.99E 08 17.081 2838.78 Dense Graded (PG 76 22) STOA 0C 0.423 3.57E 07 3 .20 17.40 2.30 2.01 2.81E 09 1.013 1038.17 10C 0.534 7.54E 07 2.23 10.55 5.50 5.26 1.61E 08 4.414 3326.20 20C 0.623 1.70E 06 1.32 5.72 3.90 3.75 7.81E 08 18.189 3988.09 LTOA 0C 0.360 2.27E 07 3.55 17.40 2.40 2.04 9.79E 10 0.443 974.00 10C 0. 413 5.43E 07 2.59 11.37 3.50 3.21 3.88E 09 1.414 1824.64 20C 0.551 1.08E 06 1.38 6.21 2.90 2.75 2.68E 08 7.114 2919.52 Open Graded (PG 76 22) STOA 0C 0.393 2.98E 07 1.87 10.48 0.60 0.43 1.76E 09 0.708 491.42 10C 0.434 8.83E 07 1.58 7.83 1.20 1.00 7.65E 09 2.657 1107.59 20C 0.646 1.51E 06 1.09 4.29 2.50 2.36 8.49E 08 19.049 3050.47 LTOA 0C 0.291 3.02E 07 1.98 10.18 0.50 0.31 6.57E 10 0.386 457.41 10C 0.365 9.02E 07 1.50 8.53 0.70 0.60 4.11E 09 1.741 732.86 20C 0.520 1.31E 06 1.03 5.1 5 1.10 1.00 2.47E 08 7.011 1409.67 Open Graded (ARB 12) STOA 0C 0.400 4.80E 07 1.67 9.84 0.50 0.36 3.06E 09 1.185 448.59 10C 0.533 5.87E 07 1.45 9.10 1.20 1.08 1.25E 08 3.500 1058.80 20C 0.717 1.24E 06 1.07 4.52 2.00 1.87 1.25E 07 25.304 2364.51 LTOA 0C 0.324 4.46E 07 1.93 11.35 0.50 0.34 1.35E 09 0.670 446.92 10C 0.427 6.26E 07 1.57 10.16 1.10 0.98 5.13E 09 1.824 1013.60 20C 0.585 9.62E 07 1.10 4.92 1.50 1.38 3.19E 08 7.988 1772.16

PAGE 110

110 APPENDIX D HEALING POTENTIAL TEST RESULTS

PAGE 111

111 Figure D 1 HPT results for dense graded unmodified mixture (STOA) Figure D 2 HPT results for dense graded unmodified mixture (LTOA) 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 10 20 30 40 50 60 70 Normalized MR Time(min) Dense_U ST 0C 10C 20C 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 10 20 30 40 50 60 70 Normalized MR Time(min) Dense_U LT 0C 10C 20C

PAGE 112

112 Figure D 3 HPT results for dense graded m odified mixture (STOA) Figure D 4 HPT results for dense graded modified mixture (LTOA) 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 10 20 30 40 50 60 70 Normalized MR Time(min) Dense_M ST 0C 10C 20C 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 10 20 30 40 50 60 70 Normalized MR Time(min) Dense_M LT 0C 10C 20C

PAGE 113

113 Figure D 5 HPT results for open graded SBS modified mixture (STOA) Figure D 6 HPT results for open graded SBS modified mixture (LTOA) 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 10 20 30 40 50 60 70 Normalized MR Time(min ) OGFC_M ST 0C 10C 20C 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 10 20 30 40 50 60 70 Normalized MR Time(min ) OGFC_M LT 0C 10C 20C

PAGE 114

114 Figure D 7 HPT results for open graded ARB modified mixture (STOA) Figure D 8 HPT results for open graded ARB modified mixt ure (LTOA) 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 10 20 30 40 50 60 70 Normalized MR Time(min) OGFC_ARB ST 0C 10C 20C 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 0 10 20 30 40 50 60 70 Normalized MR Time(min ) OGFC_ARB LT 0C 10C 20C

PAGE 115

115 APPENDIX E HEALING RATE PARAMETER

PAGE 116

116 Figure E 1 Healing rate parameter for dense graded mixture Figure E 2 Healing rate parameter for open graded mixture 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 DUS DUL DMS DML Healing Rate Parameter 0c 10C 20C 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 OMS OML ORS ORL Healing Rate Parameter 0C 10C 20C

PAGE 117

117 LIST OF REFERENCES AASHTO. (2012 a ). Standard Method of Test for Bulk Specific Gravity of Compacted Bituminous Mixtures using Saturated Surface Dry Specimens T 166 Washington, D.C. AASHTO. (2012 b ). Standard Practice for Mixture Conditioning of Hot Mix Asphalt PP2 Washington, D.C. AASHTO. (20 07 ). Standard Method of Test for Resistance of Compacted Hot Mix Asphalt (HMA) to Moisture Induced Damage T 283 Washington, D.C. Asphalt Institute. (2001). Superpave Mix Design: Superpave Series no.2 (SP 2), 3rd Ed., Asphalt Institute, Le xington, KY. ASTM. (2011 a ). Standard Test Method for Bulk Specific Gravity and Density of Compacted Bituminous Mixtures using Automatic Vacuum Sealing Method D6752 West Conshohocken, PA. ASTM. (2011 b ). Standard Test Method for Bulk Specific Gravity and Density of Non Absorptiv D2726 West Conshohocken, PA. Bhasin, A., Palvadi, S., Little, D. N. (2011). "Influence of Aging and Temperature on Intrinsic Healing of Asphalt Binders." Transportation Research Record 2207, T ransportation Research Board, National Research Council 70 78. Bonnaure, F.P., Huibers, A.H.J.J., and Boonders, A. (1982). "A Laboratory Investigation of the Influence of Rest Periods on the Fatigue Characteristics of Bituminous Mixes." Journal of the As sociation of Asphalt Paving Technologists, 51, 104 128. Buchanan, M. S., and White, T. D. (2005). "Hot Mix Asphalt Mix Design Evaluation using the Corelok Vacuum Sealing Device." Journal of Materials in Civil Engineering, 17(2), 137 142. Buttlar, W. G., and Roque, R. (1994). "Development and Evaluation of the Strategic Highway Research Program Measurement and Analysis System for Indirect Tensile Testing at Low Temperatures." Transportation Research Record 1454, Transportation Research Board, National Rese arch Council 163 171. Carpenter, S. H., and Shen, S. (2006). "Dissipated Energy Approach to Study HMA Healing in Fatigue." Transportation Research Record 1970, Transportation Research Board, National Research Council, 178 185. Carpenter, S. H., Ghuzlan, K. A., Shen, S. (2003). "Fatigue Endurance Limit for Highway and Airport Pavements." Transportation Research Record 1832, Transportation Research Board, National Research Council, 131 138.

PAGE 118

118 Coates, D. F., and Parsons, R. C. (1966). "Experimental Criteria for Classification of Rock Substances." International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 3(3), 181 189. Cooley, L. A., Prowell, B. D., Hainin, M. R., Buchanan, M. S., Harrington, J. (2002). Bulk Specific Gravity Round Robin using the Corelok Vacuum Sealing Device, Report 02 11 National Center for Asphalt Technology, Auburn, AL. Daniel, J. S., and Kim, Y. R. (2001). "Laboratory Evaluation of Fatigue Damage and Healing of Asphalt Mixtures." Journal of Materials in Civil Engineering, 13(6), 434 440. Deme, I. J., and Young, F. D. (1987). "Ste. Anne Test Road Revisited Twenty Years Later." Proc. Canadian Technical Asphalt Association, 32, 254 283. Di Benedetto, H., Nguyen, Q. T., Sauzeat, C. (2011). "Nonlinearity, Heatin g, Fatigue and Thixotropy during Cyclic Loading of Asphalt Mixtures." Road Materials and Pavement Design, 12(1), 129 158. El Hussein, H. M., and Abd El Halim, A. O. (1993). "Differential Thermal Expansion and Contraction: A M echanistic Approach to Adhesio n in Asphalt Concrete." Canadian Journal of Civil Engineering, 20(3), 366 373. FDOT. (2010 a ). Standard Specifications for Road and Bridge Construction, Florida Department of Transportation, Tallahassee, FL. FDOT. (2010 b ). Florida Test Method for Resista nce of Compacted Bituminous Mixture to Moisture FM 1 T 283, Florida Department of Tr ansportation, Tallahassee, FL. Finn, F., Saraf, C., Kulkarni, K., Smith, W., and Abdullah, A. (1977). "The U se of P rediction S ubsystems for the D esign of P avement S tructures." Proc., 4th International Conference on Structural Design of Asphalt Pavements University of Michigan, Ann Arbor, MI 3 38. Francken, L., and Clauwaert, C. (1987). Characterization and S tructural A ssessment of B ound M aterials for F lexible R oad S tructures." Proc., 6th International Conference on the Structural Design of Asphalt Pavements, University of Michigan, Ann Arbor, M I 130 144. Ghuzlan, K. A., and Carpenter, S. H.,. (2000). "Energy Derived, Damage Based Failure Criterion for Fatigue Testing." Transportation Research Record 1723, Transportation Research Board, 141 149. Grant, T. P. (2001). Determination of Asphalt Mixture Healing Rate using the Superpave Indirect Tensile Test, Master's Thesis University of Florida, Gainesvil le, FL

PAGE 119

119 Groenendijk, J., Vogelzang, C.H., Miradi, A., Molenaar, A.A.A ., and Dohmen, L.J.M. (1997). Linear Tracking Performance Tests on Full Depth Asphalt Pavement." Transportation Research Record 1570, Transportation Research Board, National Research Co uncil, 39 54. Hucka, V., and Das, B. (1974). "Brittleness Determination of Rocks by Different Methods." International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 11(10), 389 392. eal Transportation Research Record 1970 Transportation Research Board, National Research Council 84 91. Kim, S., and Coree, B. J. (2005). Evaluation of Hot Mix Asphalt Moisture Sensitivity using the Nottingham Asphalt Tes t Equipment, CTRE Project 02 117 I owa State University, Ames, IA Kim, Y., and Kim, Y. R. (1997). "In Situ Evaluation of Fatigue Damage Growth and Healing of Asphalt Concrete Pavements using Stress Wave Method." Transportation Research Record 1568, Transportation Research Board, National Research Council, 106 113. Koh, C., and Roque, R. (2010 a ). "Characterization of the Tensile Properties of Open Graded Friction Course Mixtures Based on Direct and Indirect Tension Tests." Journal of Testing and Eval uation, 38(4), 1 12. Koh, C., and Roque, R. (2010 b ). "Use of Nonuniform Stress State Tests to Determine Fracture Energy of Asphalt Mixtures Accurately." Transportation Research Record 2181, Transportation Research Board, National Research Council, 55 66. Lamond, J. F., and Pielert, J. H. (2006). Significance of Tests and Properties of Concrete and Concrete Making Materials, ASTM International, West Conshohocken PA. Little, D. N., and Bhasin, A. (2007). "Exploring Mechanisms of Healing in Asphalt Mixture s and Quantifying its Impact." Self Healing Mechanisms: An Alternative Approach to 20 Centuries of Materials Science, Springer, the Netherlands, 205 218. Liu, Q., Garca, ., Schlangen, E., Ven, M. (2011). "Induction Healing of Asphalt Mastic and Porous A sphalt Concrete." Construction and Building Materials, 25(9), 3746 3752. Lytton, R. L., Uzan, J., Fernando, E. G., Roque, R., Hiltunen, D., Stoffels, S. M. (1993). Development and Validation of Performance Prediction Models and Specifications for Asphalt Binders and Paving Mixes, SHRP A 357 National Research Council, Washington, D.C. Monismith, C.L., Epps, J.A., and Finn, F.N. (1985). "Improved Asphalt Mix Design." Journal of the Association of Asphalt Paving Technologists, 54, 347 406.

PAGE 120

120 Monismith, C.L., Epps, J.A., Kasianchuk, D.A., and McLean, D.B. (1970). "Asphalt Mixture Behavior in Repeated Flexure." Institute of Transportation and Traffic Engineering, Report No. TE 70 5. Mouillet, V., De, l. R., Chailleux, E., Coussot, P. (2012). "Thixotropic Behav ior of Paving Grade Bitumens under Dynamic Shear." Journal of Materials in Civil Engineering 24(1), 23 31. Myre, J. (1992). "Fatigue of A sphalt M aterials for N orwegian C onditions." Proc., 7th International Conference on Asphalt Pavements, Nottingham, UK 238 251. Read, J., Whiteoak, D., Hunter, R. (2003). Shell Bitumen Handbook, 5th Ed., Thomas Telford, London, UK. Roque, R., Isola, M., Chun, S., Zou, J., Koh, C., Lopp, G. (2012 a ). Effects of Laboratory Heating, Cyclic Pore Pressure, and Cyclic Loading on Fracture Properties of Asphalt Mixture, Final Report of Florida Department of Transportation University of Florida, Gainesville, FL Roque, R., Simms, R., Chen, Y., Koh, C., Lopp, G. (2012 b ). Development of a Test Method that Will Allow Evaluation and Quantification of the Effects of Healing on Asphalt Mixture, Final Report of Florida Department of Transportation University of Florida, Gainesville, FL Roque, R., Chun, S., Zou, J., Lopp, G., Villiers, C. (2011). Continuation of Superpave Projects Mon itoring, Final Report of Florida Department of Transportation University of Florida, Gainesville FL Roque, R., Birgisson, B., Sangpetngam, B., Zhang, Z. (2002). "Hot Mix Asphalt Fracture Mechanics: A Fundamental Crack Growth Law for Asphalt Mixtures." Journal of the Association of Asphalt Paving Technologists, 71, 816 827. Roque, R., Buttlar, W. G., Ruth, B. E., Dickson, S. W. (1997 a ). "Short Loading Time Stiffness from Creep, Resilient Modulus, and Strength Tests using Superpave Indirect Tension Test. Transportation Research Record 1630, Transportation Research Board, National Research Council, 10 20. Roque, R., Buttlar, W. G., Ruth, B. E., Tia, M., Dickson, S. W., Reid, B. (1997 b ). Evaluation of SHRP Indirect Tension Tester to Mitigate Cracking in A sphalt Pavements and Overlays, Final Report of Florida Department of Transportation University of Florida, Gainesville, FL. Roque, R., and Buttlar, W. G. (1992). "The Development of a Measurement and Analysis System to Accurately Determine Asphalt Concre te Properties using the Indirect Tensile Mode." Journal of the Association of Asphalt Paving Technologists, 61, 304 332.

PAGE 121

121 Santagata, E., Baglieri, O., Tsantilis, L., Dalmazzo, D. (2013). "Evaluation of Self Healing Properties of Bituminous Binders Taking i nto Account Steric Hardening Effects." Construction and Building Materials 41 60 67. Seed, H. B., Chan, C. K., Monismith, C. L. (1955). "Effects of R epeated L oad on the S trength and D eformation of C ompacted C lay." Proc., Highway Research Board Proceedin gs, Highway Research Board 541 558. Shen, S., an d Carpenter, S. H. (2005). "Application of the Dissipated Energy Concept in Fatigue Endurance Limit Testing." Transportation Research Record 1929, Transportation Research Board, National Research Council, 1 65 173. Si, Z., Little, D. N., Lytton, R. L. (2002). "Characterization of Microdamage and Healing of Asphalt Concrete Mixtures." Journal of Materials in Civil Engineering, 14(6), 461 470. Tabatabaee, H. A., Velasquez, R., Bahia, H. U. (2012). "Predicting Low Temperature Physical Hardening in Asphalt Binders." Construction and Building Materials 34 162 169. Thompson, M.R. and Carpenter, S.H. (2006). "Considering Hot Mix Asphalt Fatigue Endurance Limit in Full Depth Mechanistic Empirical Pavement Design. Pro c., International Conference on Perpetual Pavements, Ohio University, Colum bus, OH. Varadhan, A. (2004). Evaluation of Open Graded and Bond ed Friction Course for Florida, Master' s Thesis, University of Florida, Gainesville, FL Wool, R. P., and O'Conn or, K. M. (1981). "A Theory of Crack Healing in Polymers." Journal of Applied Physics, 52(10), 5953 5963. Zhang, Z., Roque, R., Birgisson, B., Sangpetngam, B. (2001). Journal of the Association of Asphalt Paving Technologists, 70, 206 241. Zou, J., and R oque, R. (2012). "Effect of HMA Ageing and Potential Healing on Top Down Cracking using HVS." Road Materials and Pavement Design, 13(3), 518 533. Zou, J., and Roque, R. (2011). "Top Down Cracking: Enhanced Performance Model and Improved Understanding of M echanisms." Journal of the Association of Asphalt Paving Technologists, 80, 255 288.

PAGE 122

122 BIOGRAPHICAL SKETCH Reebie Simms was born in Sarasota, Florida in 1986. After graduating from Booker High School in 2004, she atten ded the University of Florida where she received a Bachelor of Science with honors in c ivil e ngineering in December of 2008. Upon graduation f rom the University of Florida, s he worked briefly in the private sector before entering into graduate school at th e University of Florida as a Ph.D. student. During her time at the University of Florida, she also managed to rece ive a Master of Engineering in civil e ngineering in August of 2011. After completing her Ph.D., she plans to continue to work in the field of civil engineering outside of academia.