<%BANNER%>

Tensile Properties of Open Graded Friction Course (OGFC) Mixture to Evaluate Top-Down Cracking Performance

Permanent Link: http://ufdc.ufl.edu/UFE0024746/00001

Material Information

Title: Tensile Properties of Open Graded Friction Course (OGFC) Mixture to Evaluate Top-Down Cracking Performance
Physical Description: 1 online resource (168 p.)
Language: english
Creator: Koh, Chul
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: asphalt, dog, open, superpave, tensile
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: Since its introduction, open graded friction course (OGFC) has provided unique functions as a surface layer for pavements. However, the tensile properties of OGFC, which influence the top-down cracking performance of asphalt pavement, have not been properly evaluated. Stress development at the top of the asphalt structural layer may be significantly affected by the material properties/characteristics of OGFC mixture. Therefore, it is necessary to accurately and reliably determine the fracture properties of OGFC mixture to determine its overall contribution to top-down cracking resistance of the pavement structure. A dog-bone direct tension test (DBDT) to accurately determine tensile properties of asphalt concrete, including OGFC, was conceived, developed and validated. Proper data reduction, analysis methods, and correction factors were developed based on three dimensional finite element analysis to account for non-uniform stress, strain, and rotation effects. The newly developed DBDT and existing Superpave IDT were used to perform resilient modulus, creep, and strength tests at multiple temperatures on dense graded and OGFC mixture. Tensile properties of dense graded and OGFC mixture were successfully obtained and both testing systems provided reasonable and consistent test results with respect to temperature and aging. Excellent correspondence was observed between properties determined from each type of test, indicating that fundamental properties can be accurately determined using either test. Differences in strain rate between the two tests resulted in expected differences in strength and failure strain. Creep compliance was highly correlated between the two tests but was lower for IDT than for DBDT, an effect that was attributed to the higher confinement in IDT. It was concluded that DBDT compliance is more appropriate for uniaxial stress states, while IDT compliance is more appropriate for biaxial stress states. This effect also helps to explain why conditions at the surface are more conducive to top-down cracking. Analysis with HMA fracture mechanics model indicated that OGFC mixture probably accelerates development of cracking relative to pavements with no OGFC. Continued use of Superpave IDT was recommended because it is much more practical than DBDT.
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 Chul Koh.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Roque, Reynaldo.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024746:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024746/00001

Material Information

Title: Tensile Properties of Open Graded Friction Course (OGFC) Mixture to Evaluate Top-Down Cracking Performance
Physical Description: 1 online resource (168 p.)
Language: english
Creator: Koh, Chul
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: asphalt, dog, open, superpave, tensile
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: Since its introduction, open graded friction course (OGFC) has provided unique functions as a surface layer for pavements. However, the tensile properties of OGFC, which influence the top-down cracking performance of asphalt pavement, have not been properly evaluated. Stress development at the top of the asphalt structural layer may be significantly affected by the material properties/characteristics of OGFC mixture. Therefore, it is necessary to accurately and reliably determine the fracture properties of OGFC mixture to determine its overall contribution to top-down cracking resistance of the pavement structure. A dog-bone direct tension test (DBDT) to accurately determine tensile properties of asphalt concrete, including OGFC, was conceived, developed and validated. Proper data reduction, analysis methods, and correction factors were developed based on three dimensional finite element analysis to account for non-uniform stress, strain, and rotation effects. The newly developed DBDT and existing Superpave IDT were used to perform resilient modulus, creep, and strength tests at multiple temperatures on dense graded and OGFC mixture. Tensile properties of dense graded and OGFC mixture were successfully obtained and both testing systems provided reasonable and consistent test results with respect to temperature and aging. Excellent correspondence was observed between properties determined from each type of test, indicating that fundamental properties can be accurately determined using either test. Differences in strain rate between the two tests resulted in expected differences in strength and failure strain. Creep compliance was highly correlated between the two tests but was lower for IDT than for DBDT, an effect that was attributed to the higher confinement in IDT. It was concluded that DBDT compliance is more appropriate for uniaxial stress states, while IDT compliance is more appropriate for biaxial stress states. This effect also helps to explain why conditions at the surface are more conducive to top-down cracking. Analysis with HMA fracture mechanics model indicated that OGFC mixture probably accelerates development of cracking relative to pavements with no OGFC. Continued use of Superpave IDT was recommended because it is much more practical than DBDT.
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 Chul Koh.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Roque, Reynaldo.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024746:00001


This item has the following downloads:


Full Text

PAGE 1

TENSILE PROPERTIES OF OPEN GRADED FRICTION COURSE (OGFC) MIXTURE TO EVALUATE TOP-DOWN CRACKING PERFORMANCE By CHULSEUNG KOH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009 1

PAGE 2

2009 Chulseung Koh 2

PAGE 3

To my wife, Kyoungnam Lee and my baby, Amy Jungwon Koh 3

PAGE 4

ACKNOWLEDGMENTS Now, I am finishing the long journey, the so-called Ph.D., with what you have read herein. I could not have completed my Ph.D. without support from some people. I would like to express my appreciation to them here with these limited words even though words cannot express my entire appreciation. I have no idea how to express my appreciation to my wife, Kyoungnam Lee. Without her smile, belief, patience, and sacrifice, there would not been Dr. Koh. She has always encouraged me to focus on my research and study without complaining to her terrible husband. Kyoungnam, you have done half of this work and I promise you a better future. You deserve it. Thank you and I love you. I also thank my baby, Amy Jungwon Koh who invites me to a beautiful and pleasant world. I cannot imagine how my wife and I would have lived without YOU. I come to know a true love because of you. You always make me very happy which allows my hard work to be enjoyable. It is my turn to make you happy. I love my baby. I would like to express my sincere appreciation to my adviser and chairman of my supervisory committee, Dr. Reynaldo Roque. He always listened to and respected my opinion. Furthermore, he provided me with many chances to acquire valuable experience during my graduate education. All tasks in this dissertation were accomplished under his support and guidance. I would like to offer a heartfelt gratefulness and have a great respect for him. I will never forget his help. One more thing that I thank him for is giving a name to my baby, Amy. That is a beautiful name. I also thank my committee members, Dr. Mang Tia, Dr. Dennis R. Hiltunen, and Dr. Bhavani V. Sankar, for the generous contribution of discussion, support, encouragement, and precious guidance. I must extend my gratitude to Dr. Namho Kim, a professor at the Korea 4

PAGE 5

University of Technology and Education. He was always willing to share his knowledge and experience with me while doing my research. I would like to thank the Florida Department of Transportation (FDOT) for providing technical and financial support and materials for this research. Special thanks go to the engineers and technicians of the Bituminous Section of the State Materials Office (SMO) for their contributions in terms of their expert knowledge, experience, and constructive advice throughout the course of this work. Their efforts are sincerely appreciated and clearly made a positive impact on the quality of the research. Also, a special thank must go to Mr. George Lopp for his support in the laboratory and his valuable advice. I would like to acknowledge the former graduate students, Dr. Jaeseung Kim and Dr. Sungho Kim, for their generous help and advice. I also would like to thank the Korean students in the Department of Civil and Coastal Engineering including Minwoo Son, Byoungil Kim, Sanghyun Chun and Yongwan Kwon for their friendship and encouragement. I appreciate the friendship and help of all the students in the materials division including Alvaro Guarine, Aditya Ayedia, Weitao Li, Guangming Wang, and Jian Zou. I have really enjoyed working and talking with them. Lastly, I would like to thank my family and wifes family for their endless trust, encouragement, support and love. I do not want to specify more detail since it will make this acknowledge turned into tragedy. I just would like to say thank you and I love them. Everyone who has supported me is an author of this dissertation. I would like to dedicate this dissertation to them with my love and respect. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 TABLE OF CONTENTS .................................................................................................................6 LIST OF TABLES ...........................................................................................................................9 LIST OF FIGURES .......................................................................................................................10 ABSTRACT ...................................................................................................................................15 CHAPTER 1 INTRODUCTION..................................................................................................................17 1.1 Background..................................................................................................................17 1.2 Hypothesis....................................................................................................................19 1.3 Objectives.....................................................................................................................19 1.4 Scope............................................................................................................................20 1.5 Research Approach......................................................................................................21 2 LITERATURE REVIEW.......................................................................................................23 2.1 Evaluation of Open Graded Friction Courses..............................................................23 2.1.1 Introduction........................................................................................................23 2.1.2 Functions and Durability....................................................................................24 2.1.3 Structural Capacity.............................................................................................25 2.1.4 Closure................................................................................................................28 2.2 Top-Down Cracking Mechanism.................................................................................28 2.2.1 Introduction........................................................................................................28 2.2.2 Influencing Factors on Top-Down Cracking......................................................29 2.2.3 Characteristics of Top-Down Cracking..............................................................32 2.2.4 Closure................................................................................................................33 2.3 Asphalt Mixture Tests in Tension Mode......................................................................33 2.3.1 Introduction........................................................................................................33 2.3.2 Superpave Indirect Tension Test (Superpve IDT)..............................................35 2.3.3 Hollow Cylinder Tension Test (HCT)................................................................36 2.3.4 Disk-Shaped Compact Tension Test (DCT).......................................................37 2.3.5 Semi-Circular Bending Test (SCB)....................................................................38 2.3.6 Uniaxial Direct Tension Test..............................................................................39 2.3.7 Closure................................................................................................................41 3 DEVELOPMENT OF DOG-BONE DIRECT TENSION TEST (DBDT)............................42 6

PAGE 7

3.1 Introduction..................................................................................................................42 3.2 Optimization of Specimen Geometries for DBDT.......................................................42 3.2.1 Two Dimensional Finite Element Analysis........................................................43 3.3 Dog-Bone Direct Tension Testing System..................................................................46 3.3.1 Coring Process....................................................................................................47 3.3.2 Bonding Specimen to Loading Heads................................................................49 3.3.3 Strain Gage Mounting System............................................................................53 3.3.4 Load Equalization System..................................................................................57 4 TEST MATERIALS AND SPECIMEN PREPARATION....................................................61 4.1 Materials.......................................................................................................................61 4.1.1 Dense Graded Asphalt Mixtures........................................................................61 4.1.2 Open Graded Asphalt Mixtures..........................................................................62 4.2 Asphalt Mixture Design...............................................................................................64 4.2.1 Dense Graded Asphalt Mixtures........................................................................64 4.2.1.1 Selection of traffic level............................................................................64 4.2.1.2 Batching and mixing.................................................................................65 4.2.1.3 Compaction...............................................................................................66 4.2.1.4 Long term oven aging (LTOA).................................................................67 4.2.2 Open Graded Asphalt Mixtures..........................................................................68 4.2.2.1 Measuring air void for open graded asphalt mixtures..............................69 4.3 Test Specimen Preparation...........................................................................................73 4.3.1 Dog-Bone Direct Tension Test Specimen..........................................................73 5 DBDT TEST PROCEDURE AND DATA INTERPRETATION METHODS.....................80 5.1 Introduction..................................................................................................................80 5.2 Dog-Bone Direct Tension Test Procedures..................................................................81 5.2.1 Resilient Modulus Test.......................................................................................81 5.2.2 Creep Test...........................................................................................................83 5.2.3 Tensile Strength Test..........................................................................................84 5.3 Data Interpretation Methods for Dog-Bone Direct Tension Test................................85 5.3.1 Three Dimensional Finite Element Analysis......................................................86 5.3.1.1 Stress analysis...........................................................................................87 5.3.1.2 Strain analysis...........................................................................................88 5.3.1.3 Rotation effect on edge measurements.....................................................90 5.3.2 Resilient Modulus Test.......................................................................................92 5.3.3 Creep Test...........................................................................................................99 5.3.4 Tensile Strength Test........................................................................................101 5.4 Verification of DBDT Correction Factors.................................................................106 5.4.1 Calibration Sample (Delrin Specimen)............................................................106 5.4.2 Asphalt Mixtures..............................................................................................109 6 MIXTURE PERFORMANCE EVALUATION...................................................................112 6.1 Introduction................................................................................................................112 7

PAGE 8

6.2 Use of Non-Uniform Stress States of Tests for Tensile Failure Limits.....................112 6.2.1 Uniform Stress State of Uniaxial Direct Tension Test.....................................114 6.2.2 Non-Uniform Stress State of Superpave IDT...................................................116 6.3 Superpave IDT Test Results......................................................................................119 6.4 Dog-Bone Direct Tension Test Results......................................................................126 6.5 Comparison between Superpave IDT and DBDT Test Results.................................133 7 CLOSURE............................................................................................................................142 7.1 Summary and Findings..............................................................................................142 7.2 Conclusion.................................................................................................................143 7.3 Recommendations......................................................................................................143 APPENDIX A LABORATORY MIXTURES INFORMATION.................................................................145 B SUPERPAVE IDT TEST RESULTS...................................................................................149 C DBDT TEST RESULTS.......................................................................................................151 D CREEP PARAMETERS FROM SUPERPAVE IDT AND DBDT.....................................153 LIST OF REFERENCES.............................................................................................................157 BIOGRAPHICAL SKETCH.......................................................................................................167 8

PAGE 9

LIST OF TABLES Table page 2-1 Terminologies for open graded surface friction course (after Suresha et al., 2009)..........23 3-1 Analysis results for determining DBDT sample geometry................................................45 3-2 Characteristics of LOCTITE Hysol Product E-20HP........................................................51 4-1 Aggregate source for dense graded asphalt mixture..........................................................62 4-2 Aggregate source for open graded asphalt mixture...........................................................63 4-3 Traffic levels and gyratory compaction efforts..................................................................65 4-4 Volumetric information for dense graded asphalt mixtures..............................................67 4-5 Air void contents for open graded asphalt mixture............................................................72 A-1 JMF for Georgia granite dense gradation........................................................................146 A-2 Batch weight for Georgia granite dense gradation..........................................................146 A-3 JMF for Florida limestone open gradation.......................................................................147 A-4 Batch weight for Florida limestone open gradation.........................................................147 A-5 JMF for Nova Scotia granite open gradation...................................................................148 A-6 Batch weight for Nova Scotia granite open gradation.....................................................148 B-1 Superpave IDT test results for dense graded asphalt mixtures........................................150 B-2 Superpave IDT test results for open graded asphalt mixtures.........................................150 C-1 DBDT test results for dense graded asphalt mixtures......................................................152 C-2 DBDT test results for open graded asphalt mixtures.......................................................152 9

PAGE 10

LIST OF FIGURES Figure page 1-1 Research Approach............................................................................................................22 2-1 Stress state in asphalt layer (after Roque and Buttlar, 1992).............................................34 2-2 Superpave IDT...................................................................................................................36 2-3 Hollow cylinder tension test sample..................................................................................37 2-4 Disk-shaped compact tension test......................................................................................38 2-5 Semi-circular bending test.................................................................................................39 2-6 Uniaxial direct tension test.................................................................................................40 3-1 Schematic drawing of load application in Superpave IDT and DBDT.............................43 3-2 Coring radius (r) and coring overlap (x)............................................................................44 3-3 2-D FEM mesh and coordinates........................................................................................45 3-4 Stress distribution on centerline and headline (3 in coring radius and 2 in coring overlap)..............................................................................................................................46 3-5 Dog-bone direct tension test prototype..............................................................................47 3-6 Coring jig...........................................................................................................................48 3-7 Loading head with grooved contact surface......................................................................49 3-8 Loading heads with alignment bars and shim stocks.........................................................50 3-9 Loading heads with a series of shim stocks.......................................................................51 3-10 Bonding agent gun with mixing nozzle.............................................................................52 3-11 Machinist granite block......................................................................................................52 3-12 Gage point positioning on edge.........................................................................................54 3-13 Modification of gage point for edge..................................................................................54 3-14 Template to attach gage point on faces..............................................................................55 3-15 Template to attach gage point on edges.............................................................................55 10

PAGE 11

3-16 Gage point attachment.......................................................................................................56 3-17 Schematic illustration of knife edge modification.............................................................56 3-18 Modification of knife edge.................................................................................................57 3-19 Dual cylinder loading assembly.........................................................................................58 3-20 Dual cylinder loading assembly attached to loading rod...................................................59 3-21 Environmental chamber.....................................................................................................59 4-1 Dense graded asphalt mixture gradation............................................................................62 4-2 Open graded asphalt mixture gradation.............................................................................64 4-3 Asphalt mixer.....................................................................................................................66 4-4 Superpave Servopac gyratory compactor..........................................................................67 4-5 Long term oven aging (LTOA) conditioning set-up..........................................................68 4-6 Corelok equipment to measure air void.............................................................................70 4-7 Vacuum sealing in corelok.................................................................................................70 4-8 Air voids of open graded asphalt mixture..........................................................................72 4-9 Slicing asphalt mixture......................................................................................................74 4-10 Coring asphalt mixture sliced............................................................................................74 4-11 Coring the other side of asphalt mixture sliced.................................................................75 4-12 Specimen after coring........................................................................................................75 4-13 Sanding asphalt mixture to eliminate asphalt film.............................................................76 4-14 Bonding between a specimen and loading heads...............................................................77 4-15 Gage points attachment on faces........................................................................................78 4-16 Gage points attachment on edges.......................................................................................78 4-17 Set-up of knife edge on gage points...................................................................................79 5-1 Test plan.............................................................................................................................80 5-2 Typical resilient load..........................................................................................................82 11

PAGE 12

5-3 Typical creep load..............................................................................................................83 5-4 Determination of strain rate for strength test.....................................................................85 5-5 3-D FEM mesh and coordinates........................................................................................86 5-6 Stress distributions at the center of DBDT specimen (2-in thickness specimen with 2-in width).............................................................................................................................88 5-7 Strain distributions on faces (2-in thickness specimen with 2-in width)...........................89 5-8 Strain distributions on edges (2-in thickness specimen with 2-in width)..........................90 5-9 Rotational effect on edges..................................................................................................91 5-10 Normalized displacement along axis of gage point on edge (2-in thickness specimen with 2-in width)..................................................................................................................91 5-11 Determination of beginning of load cycle.........................................................................93 5-12 Determination of load amplitude.......................................................................................94 5-13 Definition of instantaneous and total recoverable deformation.........................................95 5-14 Determination of instantaneous and total recoverable deformation..................................96 5-15 Trimmed mean method for obtaining the creep compliance...........................................100 5-16 Determination of the beginning of loading......................................................................102 5-17 Instant of fracture for DBDT system...............................................................................103 5-18 Calculation of fracture energy.........................................................................................105 5-19 Test set-up on Delrin calibration sample.........................................................................107 5-20 Resilient deformation on Delrin calibration sample........................................................108 5-21 Resilient modulus on Delrin calibration sample..............................................................108 5-22 Resilient modulus results on face and edge.....................................................................109 5-23 Creep compliance of Dense (Unmod) STOA mixtures on face and edge.......................110 5-24 Creep compliance of FC-5 (FLime) LTOA mixtures on face and edge..........................111 5-25 Creep compliance results on face and edge.....................................................................111 6-1 Uniform stress condition under direct tension.................................................................115 12

PAGE 13

6-2 Non-uniform stress state of Superpave IDT and DBDT..................................................116 6-3 Non-uniform stress state of Superpave IDT from 3D FEM analysis...............................118 6-4 Detection of the fracture instant of Superpave IDT.........................................................119 6-5 Resilient modulus from Superpave IDT test....................................................................120 6-6 Creep compliance from Superpave IDT test....................................................................121 6-7 Creep rate from Superpave IDT test................................................................................121 6-8 Strength from Superpave IDT test...................................................................................122 6-9 Failure strain from Superpave IDT test...........................................................................123 6-10 Fracture energy from Superpave IDT test.......................................................................124 6-11 Dissipated creep strain energy from Superpave IDT test................................................125 6-12 Energy ratio from Superpave IDT test.............................................................................125 6-13 Resilient modulus from DBDT test.................................................................................126 6-14 Creep compliance from DBDT test.................................................................................127 6-15 Creep rate from DBDT test..............................................................................................128 6-16 Strength from DBDT test.................................................................................................128 6-17 Failure strain from DBDT test.........................................................................................129 6-18 Fracture energy from DBDT test.....................................................................................130 6-19 Dissipated creep strain energy from DBDT test..............................................................130 6-20 Specimen after strength test.............................................................................................131 6-21 Crack length with load cycles..........................................................................................132 6-22 No. of load cycles at 1 inch crack length.........................................................................132 6-23 Resilient modulus from Superpave IDT and DBDT tests...............................................133 6-24 Creep compliance from Superpave IDT and DBDT tests...............................................135 6-25 Creep rate from Superpave IDT and DBDT tests............................................................135 6-26 Strength from Superpave IDT and DBDT tests...............................................................136 13

PAGE 14

6-27 Initial tangent modulus from Superpave IDT and DBDT tests.......................................136 6-28 Failure strain from Superpave IDT and DBDT tests.......................................................137 6-29 Fracture energy Superpave IDT and DBDT tests............................................................138 6-30 Dissipated creep strain energy from Superpave IDT and DBDT tests............................139 6-31 Schematic illustration of stress-strain relation between DBDT and Superpave IDT......139 6-32 Loading rate correction with tangent modulus................................................................141 6-33 Corrected dissipated creep strain energy from Superpave IDT and DBDT tests............141 D-1 m-value from Superpave IDT test....................................................................................154 D-2 D 1 from Superpave IDT test............................................................................................154 D-3 m-value from DBDT test.................................................................................................155 D-4 D 1 from DBDT test..........................................................................................................155 D-5 m-value from Superpave IDT and DBDT tests...............................................................156 D-6 D 1 from Superpave IDT and DBDT tests........................................................................156 14

PAGE 15

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TENSILE PROPERTIES OF OPEN GRADED FRICTION COURSE (OGFC) MIXTURE TO EVALUATE TOP-DOWN CRACKING PERFORMANCE By Chulseung Koh August 2009 Chair: Reynaldo Roque Major: Civil Engineering Since its introduction, open graded friction course (OGFC) has provided unique functions as a surface layer for pavements. However, the tensile properties of OGFC, which influence the top-down cracking performance of asphalt pavement, have not been properly evaluated. Stress development at the top of the asphalt structural layer may be significantly affected by the material properties/characteristics of OGFC mixture. Therefore, it is necessary to accurately and reliably determine the fracture properties of OGFC mixture to determine its overall contribution to top-down cracking resistance of the pavement structure. A dog-bone direct tension test (DBDT) to accurately determine tensile properties of asphalt concrete, including OGFC, was conceived, developed and validated. Proper data reduction, analysis methods, and correction factors were developed based on three dimensional finite element analysis to account for non-uniform stress, strain, and rotation effects. The newly developed DBDT and existing Superpave IDT were used to perform resilient modulus, creep, and strength tests at multiple temperatures on dense graded and OGFC mixture. Tensile properties of dense graded and OGFC mixture were successfully obtained and both testing systems provided reasonable and consistent test results with respect to temperature and aging. 15

PAGE 16

Excellent correspondence was observed between properties determined from each type of test, indicating that fundamental properties can be accurately determined using either test. Differences in strain rate between the two tests resulted in expected differences in strength and failure strain. Creep compliance was highly correlated between the two tests but was lower for IDT than for DBDT, an effect that was attributed to the higher confinement in IDT. It was concluded that DBDT compliance is more appropriate for uniaxial stress states, while IDT compliance is more appropriate for biaxial stress states. This effect also helps to explain why conditions at the surface are more conducive to top-down cracking. Analysis with HMA fracture mechanics model indicated that OGFC mixture probably accelerates development of cracking relative to pavements with no OGFC. Continued use of Superpave IDT was recommended because it is much more practical than DBDT. 16

PAGE 17

CHAPTER 1 INTRODUCTION 1.1 Background Since its introduction to the United States in the 1940s, open graded friction course (OGFC) or porous/permeable friction course (PFC) has been used primarily as a functional layer that does not increase structural load bearing capacity. The function of this layer includes increasing skid resistance of pavement surface through increased macro and micro texture, quick water drainage with high air void contents, and reduction in pavement noise. With these functional advantages, open graded friction course has attained global attention. In 1978, NCHRP reported that 15 states in the US were using OGFC extensively and several additional states were considering the use of OGFC (Halstead, 1978). Since that time, several state DOTs have initiated an implementation program for OGFC that included work related to its design and construction. Kandhal and Mallick (1998) reported that 19 of 50 highway agencies contacted, including the state of Florida, indicated they were using OGFC. However, several problems were reported with OGFC mixtures, including premature oxidation, early raveling, stripping of the underlying layers, reduction of the OGFC benefits by time through plugging of the surface pores, difficulties during snow and ice removal, and construction difficulties. Furthermore, most if not all open graded friction course mixture design guidelines are empirical in nature. Asphalt content is determined to produce a minimum VMA (Voids in the Mineral Aggregate) at a sufficiently high effective asphalt film thickness to ensure adequate durability of the mixture. Evaluation of this kind mixture has primarily focused on its function such as skid resistance, permeability and durability. Existing design processes and evaluation methods do not accommodate pavement performance-based criteria (e.g. top-down cracking). However, open graded friction courses may play a key role in top-down cracking performance of pavement. 17

PAGE 18

Surface initiated longitudinal cracking, which is often called top-down cracking, is now recognized as a common distress mode in asphalt pavement. Previously, it was generally thought that load-associated cracking initiated only at the bottom of asphalt layer, and then propagated to the surface (bottom-up cracking). Top-down cracking is predominantly parallel to the asphalt pavement centerline and located in the vicinity of the wheel paths. For years, many researchers all over the world have endeavored to define the underlying mechanisms of top-down cracking. Prior research has indicated that top-down cracking is initiated by critical tensile stresses in the surface of the asphalt pavement due to non-uniform contact stress by modern radial truck tires (Myers et al., 1998; Baladi et al., 2002). These cracks then propagate downward by the combined effects of both load and temperature (Dauzats and Rampal, 1987; Kim, 2005). Many other factors have been associated with this mode of failure as will be discussed in section 2.2. However, the effects of open graded friction course mixtures on the development of this mode of failure have not yet been fully identified or verified. As mentioned earlier, it seems fairly evident that top-down cracking can be affected by the properties and characteristics of open graded friction course mixtures because they are directly exposed to surface tensile stress induced by traffic loading, thermal stresses, and environmental effects. Thus, it is necessary to properly evaluate fracture resistance of OGFC in order to acquire more insight regarding its effect on top-down cracking of pavement. To date, very little work has been done to evaluate and characterize fracture resistance of open graded asphalt mixture. There are currently several laboratory test methods available to evaluate fracture resistance of asphalt mixture as will be described in section 2.3. Superpave IDT has been widely used to evaluate fracture resistance of dense graded asphalt mixtures. However, there is a concern that the open graded nature of OGFC mixtures may require a modified testing 18

PAGE 19

protocol from that used for dense graded mixtures. Previous research by Varadhan (2004) has shown a region of high damage under the loading strips during Superpave IDT testing of open graded asphalt mixtures. In order to overcome this effect, specimen thickness and loading rate during strength testing was increased. However, a study is needed to optimize the test conditions needed to ensure that consistent fracture properties of open graded asphalt mixtures can be obtained with the Superpave IDT. In particular, the effects of temperature need to be evaluated. Based on the experience of the researchers, a lower test temperature will likely minimize the potential problem with localized damage under the loading strips in the Superpave IDT test. Therefore, there is a clear need to identify evaluation methods that allow for the characterization and optimization of fracture resistance of open graded asphalt mixture. Also, it is necessary to accurately and reliably determine the fracture properties of open graded asphalt mixture to identify the overall contribution to cracking resistance of the total pavement structure. 1.2 Hypothesis In Florida, it is estimated that top-down cracking accounts for more than 90% of cracking in pavements. According to Florida specification, a friction course is required on most state roads and highways, so it appears that the effect of friction course mixtures may play a significant role in the development of top-down cracking. Dense-graded friction courses can be handled as a dense graded structural mix but open graded friction courses are another matter, as indicated in the previous discussion. Hence, it was hypothesized for this study that properties of OGFC mixtures necessary to evaluate top-down cracking performance can be reliably determined from the existing Superpave IDT. 1.3 Objectives The primary objectives of this research study were as follows: 19

PAGE 20

Identify an appropriate testing system to accurately and reliably measure relevant tensile properties of open graded asphalt mixtures and dense graded asphalt mixtures in general. Develop a dog-bone direct tension test (DBDT), associated components, and data interpretation methods for evaluating tensile properties of open graded asphalt mixtures and describe potential benefits of the dog-bone direct tension test system. Evaluate tensile properties of open graded asphalt mixture, as well as dense graded asphalt mixture with both Superpave IDT and DBDT. Compare the tensile properties from both proposed direct tension test and Superpave IDT to investigate the fundamental tensile properties of asphalt mixtures and to optimize the use of Superpave IDT for open graded asphalt mixtures. Identify key properties and characteristics of thin surface courses that may affect top-down cracking performance to evaluate the contribution of the OGFC to the cracking resistance of the HMA pavement structure. 1.4 Scope This study was initiated because the effect of OGFC on the top-down cracking performance has not been extensively investigated. A detailed literature review also revealed that no appropriate test method has been available to determine tensile properties of open graded asphalt mixture. Accurate determination of the tensile properties of open graded asphalt mixture may allow for clearer understanding of the top-down cracking mechanism. In order to assess its relative contribution to the cracking resistance on overall fracture performance, it is necessary to obtain fracture properties of dense graded and open graded asphalt mixture individually. Evaluation of the effect of friction course on top-down cracking performance may be complicated by combination of aggregates, binders and additives, aging, traffic loading, thermal stress, and many other environmental factors. In order to reduce the level of complication and increase relevancy, this study is primarily focused on well-controlled samples produced in the laboratory. 20

PAGE 21

Two mixture types were evaluated for each gradation. Resilient modulus, creep compliance, and strength tests were performed using Superpave IDT and the dog-bone direct tension test developed in this study. Tests were performed at the following temperatures: -10C, 0C, 5C and 10C. Two aging conditions were used: short term oven aging (STOA) and long term oven aging (LTOA). HMA fracture mechanics model was used to analyze test results and to make it possible to evaluate fracture resistance of the asphalt mixtures tested. 1.5 Research Approach This study focused primarily on identifying the effect of open graded friction course mixture on top-down cracking performance of asphalt pavement. For that purpose, an appropriate test method to accurately and reliably obtain tensile properties of open graded asphalt mixtures was conceived and identified. The approach used involved the following steps: Review previous research on open graded friction course, bonded interface in pavement systems, top-down cracking mechanism, and existing methods to obtain tensile properties of asphalt mixture for assessing top-down cracking performance. Develop and validate a dog-bone direct tension test (DBDT) and associated data interpretation method and investigate its feasibility and accuracy for determining tensile properties of asphalt mixtures. Conduct laboratory tests using Superpave IDT and DBDT on dense and open graded asphalt mixtures subjected to various conditions to evaluate tensile properties. The entire research program was accomplished as shown in Figure 1-1, starting with problem identification and a comprehensive worldwide literature review. Most of the research effort was focused on two main tasks: development of dog-bone direct tension test (DBDT) and implementation of Superpave IDT and DBDT. During development of DBDT, specimen geometries were optimized and adequate correction factors were established. In laboratory tests, dense graded and open graded asphalt mixtures were tested and evaluated individually using both Superpave IDT and the DBDT developed as part of this study. 21

PAGE 22

Figure 1-1. Research Approach 22

PAGE 23

CHAPTER 2 LITERATURE REVIEW 2.1 Evaluation of Open Graded Friction Courses 2.1.1 Introduction Various agencies around the world have used open graded friction course. Different terminologies have been used to describe this type of mixtures as shown in Table 2-1 (Suresha et al., 2009). Although there are slight difference between open graded friction course (OGFC) in the US and porous asphalt (PA) in other countries, their functions are the same. Therefore, OGFC, PA and open graded asphalt mixture (OGAM) will be referred to interchangeably in this dissertation. Table 2-1. Terminologies for open graded surface friction course (after Suresha et al., 2009) Country Agency Terminology References American Society for Testing and Materials (ASTM) Open-Graded Friction Course (OGFC) ASTM (2004) United States of America Federal Aviation Administration (FAA) Porous Friction Course (PFC) FAA (2005) Australia Australian Asphalt Pavement Association (AAPA) Open-Graded Asphalt (OGA) AAPA (2004) New Zealand Transit New Zealand (TNZ) Porous Asphalt (PA) TNZ (2007) South Africa Southern African Bitumen Association (Sabita) Porous Asphalt (PA) Sabita (1995) Japan Japan Highway Public Corporation (JHPC) Porous Asphalt (PA) Asahi and Kawamura (2000) Open graded asphalt mixture is distinctly different than dense graded asphalt mixture. As expected from its name, it is designed to contain about 15-25% air voids, preferably interconnected. In order to gain relatively higher air voids, open graded asphalt mixture uses a uniform gradation. The aggregate gradation is made up of mostly a single coarse aggregate size 23

PAGE 24

with relatively low fine aggregate and filler. Asphalt contents for this mixture are slightly higher than dense graded asphalt mixture (Cooley et al., 2000). 2.1.2 Functions and Durability Alvarez et al. (2006) summarized advantages and disadvantages of OGFC. They reported that advantages of OGFC could be divided into three areas: safety, environment, and economy. In order for this layer to achieve its functions, it should have proper air void content (15-25%), consequently, surface texture evaluation might be needed to clarify the distinction between OGFC and other pavement types. McDaniel et al. (2004) used the Circular Texture Meter (CTM) that is equipped with a charge coupled device (CCD) laser displacement sensor to measure the surface profile. They showed that OGFC has significantly greater texture depth than dense graded conventional surface (more than four times), indicating that OGFC has more surface air void contents. Other researchers also addressed higher macro surface texture depth of OGFC (Perez-Jimenez and Gordillo, 1990; Wang and Flintsch, 2007). These characteristics essentially allow for quick water drainage from the pavement during and shortly after rainfall. Thus, OGFC can improve traffic safety even when pavement is wet (Ruiz et al., 1990; van der Zwan et al., 1990). Kandhal and Mallick (1998) also mentioned from their survey that 85% of states that use or have used OGFC in the United States gave OGFC a higher than good rating in terms of safety. In addition, fast drainage of water keeps the pavement from accumulating water on the surface which would cause hydroplaning, prevents splash and spray behind vehicles (van Heystraeten and Moraux, 1990), and decreases glare from the pavement by diffusing reflection of light (van der Zwan et al., 1990). On the other hand, dense graded asphalt layer is much less permeable and thus water during rainfall tends to drain over the surface, which may cause reduction in skid resistance in wet weather and increase a potential for hydroplaning. From the environmental point of view, it has been widely reported that OGFC can reduce traffic noise. 24

PAGE 25

Perez-Jimenez and Gordillo (1990) found that OGFC has a beneficial effect on noise reduction both outside and inside of a vehicle. Recently, McDaniel et al. (2004) have also observed that OGFC reduced pavement noise compared with dense graded asphalt mixture by 3.6 dB and 4.2 dB with pass-by and close proximity noise measurements, respectively. Moreover, noise reduction capability in the pavement itself provides economic savings since the use of other provisions such as noise barriers might be decreased or not needed. However, disadvantages of this layer, however, have also been reported by many researchers. Raveling, which is defined as progressive disintegration of HMA layer due to dislodgement of aggregate particles, is one of the main problems (Swart, 1997; Huber, 2000). During the 1990s, improvements in binder (modified binder) and gradation improved OGFC performance in terms of raveling (Huber, 2000; Mallick et al., 2000; McDaniel and Thornton, 2005). Perez-Jimenez and Cordillo (1990) reported that open graded asphalt mixture with polymeric binder exhibited higher resistance to disintegration. Another problem associated with open graded surface is clogging of pores with time. Clogging pores gradually reduce the functional effectiveness of OGFC. Based on 20-years trials, Bowskill and Colwill (1997) pointed out that the lifetime of open graded asphalt is limited by clogging and by accelerated hardening of the binder. Based on multi-year monitoring in the US, Kowalski et al. (2009) stated that an increase in noise with time could be explained by clogging of the surface pores. However, Kraemer (1997) stated that it is important not to regard the reduction of hydraulic and acoustic surface properties as a failure when it would simply mean that such asphalt gradually turns into non-porous surfacing and still affords good surface characteristics. 2.1.3 Structural Capacity OGFC is a relatively thin asphalt layer placed on the structural dense graded layer. Open graded friction courses are generally considered to have no or little structural contribution in 25

PAGE 26

current structural design systems because it is unclear whether the relatively low stiffness and open graded mixture of OGFC significantly contributes to structural capacity. Therefore, it is important to evaluate structural capacity of OGFC, especially as it affects top-down cracking. In general, it has been observed that the stiffness of open graded asphalt mixture is about 50 to 80% less than conventional dense graded asphalt mixture. Therefore, OGFC has less ability to distribute load-induced stresses than dense graded asphalt mixture. Argentina adopted a 50% structural capacity for OGFC in their pavement design system based on the resilient modulus test that showed roughly 60% of the conventional mixture (Bolzan et al., 2001). van Heystraeten and Moraux (1990) summarized the results of research conducted by the Belgian Road Research Center (BRRC). Based on modulus tests, they reported the structural contribution of OGFC lies between 73 and 79% of that of a wearing course in conventional dense asphalt concrete. In the Netherlands, van der Zwan et al. (1990) showed that the initial dynamic modulus of open graded asphalt mixture is approximately 80% of that of dense asphalt concrete. Consequently the effective structural contribution is about 80 to 90% with multilayer elastic analysis. However, they also stated aging, stripping and temperature retaining of OGFC affects structural integrity. Considering these combined effects, they concluded that depending on the thickness of the structure, OGFC can be expected to contribute about 50% of the equivalent bearing capacity achievable with dense asphalt concrete. Verhaeghe et al. (1994) concluded that open graded asphalt mixture is not suitable for use as strengthening layers because of its poorer mechanical properties compared with conventional dense asphalt. They also observed that the use of polymer modified binder results in a significant increase in the fatigue resistance of OGFC. Khalid and Walsh (1996) also reported the lower modulus value of open graded asphalt than conventional dense graded asphalt mixtures. McDaniel et al. (2004) conducted the 26

PAGE 27

frequency sweep test using the Superpave Shear Test (SST). They reported lower stiffness values in open graded mixtures tested in the laboratory, indicating that there is very little mastic to stiffen the mixture. However, other researchers have pointed out that reduction in modulus should not always mean corresponding reduction in structural capacity. Poulikakos et al. (2006) developed a fatigue evolution model using elastic layer theory. On the basis of results of their model, they concluded that the modulus value of the open graded asphalt does not have a strong effect on the horizontal strain at the bottom of the asphalt layer and thus does not have a strong influence on the design life of the structure if underlying pavement structure remains the same. Similarly, Toppeiner (1993) mentioned that the use of 60 to 75% structural layer coefficients for OGFC compared with for dense graded asphalt is conservative, since similar structural contributions can be obtained from well designed OGFC. Moreover, he added that structural layer coefficient based on the resilient modulus can underestimate the structural contribution of open graded asphalt mixture. In Spain, based on the analysis of the reinforcement capacity and the reduction in deflection induced by OGFC and dense graded asphalt layer, they have determined that both dense graded layer and OGFC have similar structural capacity (Alvarez et al., 2006). Oregon Department of Transportation also applies a similar structural coefficient for both layers on the basis of the deflection measurements (Kandhal, 2002). According to a research in New Zealand, average lifetime of OGFC is 10.5 years, which is relatively short compared with that of dense graded asphalt layer, 16.2 years (Bartley consultants, 1999). Similarly, van der Zwan et al. (1990) reported the service life of OGFC is 10 years in the Netherlands. They also emphasized that OGFC is sensitive to mechanical damage in the first year after installation. Voskuilen et al. (2004) mentioned damage in the initial stage is one of the causes of shorter service life in OGFC. However, the use of modified binder provided 27

PAGE 28

the benefits to enhance the initial strength, but they indicated that modified binder does not necessarily guarantee the longer service life. Nevertheless, it is believed that modified binders increase the service life of OGFC by increasing the cohesion and adhesion in the asphalt mixture and by increasing the binder film thickness without the binder drainage during construction. 2.1.4 Closure A review of literature revealed that most research on the evaluation of OGFC has focused on its functionality, durability and structural integrity based on conventional fatigue (bottom-up cracking). Relatively little research to date had been carried out on top-down cracking resistance of open graded asphalt mixture. Given its position top of the pavement system, open graded friction course may play an important role to either resist or accelerate top-down cracking, which is induced by surface tensile stress. More comprehensive evaluation on the tensile properties of OGFC is necessary to properly assess its effect on fracture resistance and top-down cracking in flexible pavement. 2.2 Top-Down Cracking Mechanism 2.2.1 Introduction It is now well recognized that load-related top-down cracking, which initiates at the surface of the pavement and propagates downward, commonly occurs in hot mix asphalt pavements. This phenomenon has been reported to occur in many parts of the United States (Roque and Ruth, 1990; Myers et al., 1998; Myers, 2000; Uhlmeyer et al., 2000; Myers et al., 2001; Svasdisant et al., 2002; Schorsch et al., 2003; Kim, 2005) as well as in Europe (Molenaar, 1984; Dauzats and Rampal., 1987; Gerritsen et al., 1987; Nunn, 1998; De Freitas et al., 2005), Japan (Matsuno and Nishizawa, 1992; Himeno et al., 1997; Komoriya et al., 2001; Uchida et al., 2002) and other countries (Wambura et al., 1999; Raju et al., 2008). 28

PAGE 29

Previously, research on asphalt pavement cracking performance has generally focused on the traditional fatigue mechanism that considers failure which initiates at the bottom of the asphalt surface layer. Conventional multi-layer analysis of pavements subject to a uniformly distributed load would always yield maximum tensile stresses and strains to occur at the bottom of the asphalt concrete layer. Also, traditional fatigue approach assumes there is an average condition over the life of the pavement, where an equal amount of damage is done by each wheel load applied for that condition. Interestingly enough, however, contradictions associated with this approach have been raised by many researchers. Francken (1979) found that rest periods significantly increase the fatigue life of asphalt mixtures. This appears to indicate that asphalt concrete has the potential to heal, or that the actual failure mode is not a true fatigue phenomenon. Both of these ideas negate the validity of conventional fatigue mechanism. Roque and Ruth (1990) also contradicted this approach by indicating that fatigue parameters generally do not consider the effects of temperature changes, rest periods, and age-hardening when they explained the critical condition concept. Top-down cracking clearly cannot be explained by this traditional fatigue mechanism. As indicated above, pavement modeling as a linear-elastic multilayer system with homogeneous materials and a circular, uniformly distributed vertical loading will not predict maximum tensile stresses and/or strains to occur near the pavement surface. 2.2.2 Influencing Factors on Top-Down Cracking For years, many researchers have made efforts to identify fundamental mechanisms that may lead to top-down cracking initiation and propagation. A detailed review of this work is presented below. It is well accepted that near surface tensile stress and/or strain induced by non-uniform contact stress between the tire and asphalt surface layer may cause top-down cracking. Based on 29

PAGE 30

multiple computer models using measured tire-pavement interface stresses, Myers et al. (1998) indicated that this mode of distress may be initiated by high surface tensile stresses due to non-uniform contact stress between the ribs of radial truck tires and the surface of the asphalt layer. One of their interesting findings is that the radial truck tires behave very differently than bias-ply truck tire. Consequently, the location where the tensile stresses are acting is under the treads of the tire rather than at the tire edges. Researches presented by Jacobs (1995) and De Beer et al. (1997) indicate that most significant tension is found under the widest tread. Nunn (1998) also concluded that surface initiated cracking (top-down cracking) in the UK was caused by horizontal tensile stresses at the surface generated by truck tires. Wide based tires generated the highest tensile stress. In fact, much work on top-down cracking resulted in similar conclusion (Himeno et al., 1997; Baladi et al., 2002). Thermal effects also induce surface tensile stress. Dauzats and Rampal (1987) reported that thermal stress in the asphalt surface course may initiate top-down cracking, which is then propagated by traffic loads. Myers et al. (1998) also indicated that significant thermal stresses that can contribute to the initiation and propagation of surface cracks can develop near the surface of the pavement. More recent work by Kim (2005) showed that although top-down cracking performance in Florida was most strongly affected by traffic loading, thermal effects also appeared to affect performance. In the study presented herein, most efforts were concentrated on effects of open graded friction course mixtures and interface condition on top-down cracking. In Florida, the top pavement layer often consists of a thin Open Graded Friction Course (OGFC) designed to quickly remove the surface water during rain events. The performance of these porous surface mixtures are certainly more severely affected by environmental exposure issues than underlying 30

PAGE 31

dense graded mixtures. They are directly exposed to UV radiation, intense heat and due to their inherent higher air void content, they are prone to oxidation, environmental leaching and are more susceptible to sudden dramatic temperature changes from rainfall. These combined effects most likely cause accelerated asphalt binder hardening. Through the analysis of field data and laboratory test results, Svasdisant et al. (2002) concluded that pavements with higher asphalt surface modulus due to aging would have higher potential for top-down cracking. They reported that because of increased exposure to sunlight and oxygen, age hardening mainly affects the surface course while the lower courses are less affected. Aging increases the stiffness of binder with corresponding effect on the asphalt mixtures (Bell et al., 1991) and higher air voids generally accelerates aging (Martin et al., 1990). This seems to imply that higher air void contents of open graded asphalt mixtures might severely accelerate age-hardening. Some researchers have stated that this aging issue might be one of cause of top-down cracking (Wambura et al., 1999; Uchida et al., 2002; De Freitas et al., 2003). Another interesting issue associated with environmental condition was presented in Japan. Matsuno and Nishizawa (1992) reported that top-down cracking is absent in shadowy areas such as under bridges and appears mostly in areas exposed to extensive solar radiation. Based on analysis using finite element method, they concluded that traffic loads cause high tensile strains in hot pavements. Similarly, based on multi year survey, Komoriya et al. (2001) reported that surface initiated cracking grew faster with longer exposure to the sunshine and that high rainfall is prone to the generation of surface longitudinal cracks. They also emphasized the surface material in stating that fine graded mixture is more resistant to surface longitudinal cracks than coarse graded mixture, which supports the finding reported by De Freitas et al. (2005). 31

PAGE 32

It is also noted that stiffness gradients might make contribution to top-down cracking. It is well known that stiffness in the asphalt concrete layer is non-uniform because of induced temperature gradients, age-hardening and pavement cooling rates (Roque et al., 1988). Stiffness gradients in the asphalt concrete layer had significant effects on near-surface tensile response, which are not considered in traditional fatigue approaches (Myers, 2000; Myers et al., 2001). Secondary effects for cracking at the asphalt surface might be related to construction quality, including effects of segregation and poor compaction (Gerritsen et al., 1987; Chang et al., 2002; Schorsch et al., 2003; De Freitas et al. 2005). 2.2.3 Characteristics of Top-Down Cracking Gerritsen et al. (1987) reported that pavements in the Netherlands were experiencing premature cracking in the wearing courses, and cracks did not extend into the lower asphalt layer. Uhlmeyer et al. (2000) similarly observed that surface initiated cracking often stops at the interface between the wearing course and the underlying layer (around 50 mm). They also reported that the crack width is around 3 to 4 mm at the surface and closed with depth. Myers et al. (1998) made similar observations and explained that tensile stresses dissipate quickly with depth. Literature from many countries concurs that top-down cracking is predominantly parallel to the asphalt pavement centerline and located in the vicinity of the wheel paths (Gerristen et al., 1987; Matsuno and Nishzawa, 1992; Schorsch et al., 2003). However, researchers from different countries reported different ages at which this mode of failure initiated: the Netherlands (Gerristen et al., 1987), soon after construction; Japan (Matsuno and Nishzawa, 1992), from 1 to 5 years; France (Dauzats and Rampal, 1987), from 3 to 5 years; the state of Washington in US (Uhlmeyer et al., 2000) from 3 to 8 years; the state of Florida in US (Myers et al,. 1998), from 5 to 10 years; UK (Nunn, 1998), even after 10 years. 32

PAGE 33

Matsuno and Nishzawa (1992) conducted finite element analysis on two typical pavement cross sections and concluded the pavement cross section had little effect on surface tensile strain. However, Nunn (1998) and Uhlmeyer et al. (2000) found top-down cracking was observed in thicker asphalt pavement while Bensalem et al. (2000) observed top-down cracking in mostly thin asphalt layer. 2.2.4 Closure It seems obvious that there is a fundamentally close relationship between top-down cracking and properties of surface materials since surface mixtures are directly exposed to the surface tensile stress induced by both traffic loading and environmental conditions. However, studies on top-down cracking have not included the effect of open graded friction course mixtures. Recently, there has been a growing recognition that these open graded mixtures may be the first front in resisting top-down cracking. Myers et al. (1998) concluded that this type of cracking is in fact primarily related to mixture composition. Specifically, more fracture resistant asphalt mixtures are needed to mitigate top-down cracking. This seems to indicate the need to better characterize and determine properties of open graded asphalt mixture and to evaluate their effect on top-down cracking performance. 2.3 Asphalt Mixture Tests in Tension Mode 2.3.1 Introduction Very little work to date has been reported on the testing and characterization of fracture resistance of open graded asphalt mixtures. The laboratory mixture evaluation tests required are of course dependent on the mechanisms determined to drive top-down cracking performance (Figure 2-1). In addition, it was reported that the only way to determine asphalt mixture properties is to measure them directly since they cannot be obtained reliably from the relationship with current asphalt binder properties (Roque et al., 1997). Asphalt mixture tests 33

PAGE 34

must provide accurate information on the damage and fracture properties of the HMA mixture. In the case of top-down cracking, it is extremely important that these properties be obtained using a tensile mode of loading, as it has been shown that damage and fracture primarily develop in the presence of tensile stresses. Figure 2-1. Stress state in asphalt layer (after Roque and Buttlar, 1992) Determination of tensile properties are critical for evaluating top-down cracking mechanisms, since the cracks appear to develop mostly in opening mode, fracture mode (Myers, 2000), indicating that tension is at least partially, if not predominantly responsible for the development of cracks. There are currently several tests which are being used to measure the tensile properties of asphalt mixture in the laboratory: Superpave IDT (Roque and Buttlar, 1992; Buttlar and Roque, 1994; Roque et al., 1997), hollow cylinder test (Buttlar et al., 1999; Buttlar et al., 2004), disk-shaped compact tension test (Wagoner et al., 2005; Wagoner et al., 2006), semi-circular beam test (Li and Marastreanu, 2004; Huang and Shu, 2005), and uniaxial direct tension test (Haas, 1973; Bolzan and Huber, 1993; Kim et al., 2002). Each of these testing modes offers 34

PAGE 35

advantages and disadvantages from the standpoint of practicality as well as their ability to provide accurate damage and fracture properties. A complete review was performed to identify an appropriate measurement system for determining fracture properties of open graded asphalt mixtures and/or asphalt mixture in general. 2.3.2 Superpave Indirect Tension Test (Superpve IDT) Currently, Superpave IDT test is widely used to evaluate the tensile properties of dense graded asphalt mixtures. Superpave IDT best represents a biaxial state of stress at the bottom of the asphalt layer (Figure 2-1). The Superpave IDT loading configuration results in fairly uniform tensile stresses perpendicular to the direction of the applied load. A key advantage of the Superpave IDT over the other testing systems is that the failure plane is known a priori; it is almost always along the vertical diametral plane (Roque and Buttlar, 1992). Thus, measurements can be obtained on the failure plane at the time of failure. This results in the potential for very accurate determination of mixture failure limits. Roque et al. (1997) developed the measurement and data analysis systems employed in the Superpave IDT, which overcame the interpretation problems typically associated with this test. From a practical point of view, the test is relatively simple in nature using compressive loads through loading strips. One disadvantage of the Superpave IDT is that test results are more difficult to analyze because of the more complex stress states that develop within the test specimen. Also, tensile response depends strongly on Poissons ratio. In addition, local failure can occur due to shear fracture under the loading strips at higher temperatures, particularly in open-graded mixtures. These problems are exacerbated for thin specimens (e.g., thin wearing courses). Based on the authors experience using the Superpave IDT test on open-graded mixtures, it appears that surface air void content and aggregate structure may have an effect on the test results at relatively high test temperatures. Similar problems can be expected with SMA mixtures. These effects can be expressed more 35

PAGE 36

generally as localized damage under the loading strips. Figure 2-2 shows Superpave IDT in the environmental chamber. Figure 2-2. Superpave IDT 2.3.3 Hollow Cylinder Tension Test (HCT) Recently, hollow cylinder testing has been proposed for the determination of tensile properties of asphalt mixtures (Buttlar, 1999). Figure 2-3 presents a sample for hollow cylinder tension test. Hollow cylinder tests offer near uniform stress fields around the entire circumference of the cylinder without the alignment problems presented by the direct tension test. As a consequence, there is no significant stress concentration. However, the test requires relatively thick specimens (approximately 150 mm), which may make it a problem when properties need to be obtained from field cores from thin surface layers associated with top-down cracking problems on open graded asphalt mixtures. In addition, the plane of failure is arbitrary in this test, which makes it difficult or impossible to obtain 36

PAGE 37

measurements on the failure plane at the instant of failure. As Buttlar et al. (1999) mentioned in their paper, this test has innate problems such as density gradient problems and wall thickness determination with respect to aggregate particle size. Coring the gyratory compacted sample to produce hollow test sample makes this test less practical. Figure 2-3. Hollow cylinder tension test sample 2.3.4 Disk-Shaped Compact Tension Test (DCT) As a fracture toughness test, disk-shaped compact tension test has been developed to be able to test cylindrical cores obtained from the field as well as compacted in the laboratory (Wagoner et al., 2005). Figure 2-4 shows disk-shaped compact tension test. This test is a common fracture test specified in ASTM E 399 for metallic materials. However, application to asphalt mixtures still has not been established. Also the test is primarily suitable for determination of fracture toughness, and not other properties. Compact 37

PAGE 38

tension test could produce erroneous results if the crack front deviates from a straight (pure Mode 1) crack path. It could happen in the asphalt mixture due to aggregate particles deflecting the crack around an aggregate and/or eccentric loading. The stress distribution around the holes used for loading is unknown, and premature rupture could occur at the loading holes. Sample preparation must go through several stages including slicing, drilling holes for loading, slicing flat edge at the notch mouth, and producing a suitable notch. These steps also make this test less practical. Figure 2-4. Disk-shaped compact tension test 2.3.5 Semi-Circular Bending Test (SCB) Semi-circular bending test has been used to measure the fracture resistance of asphalt mixtures (Li and Marasteanu, 2004; Huang and Shu, 2005). Figure 2-5 shows semi-circular bending test. Originally, single-edge notched beam test (SEB) was utilized to obtain fracture toughness in fracture mechanics (Anderson, 2005). In the asphalt community, this test was 38

PAGE 39

modified to apply to asphalt mixture because it is not easy to obtain the specimen for SEB from the field and in the laboratory. Figure 2-5. Semi-circular bending test However, use of the semi-circular bending test resulted in other problems. The SCB test arrangement for asphalt mixture is a 150 mm in diameter semi-circle supported by two rollers with a span of 120 mm which leads to a relatively short fracture ligament. The initial ligament length should be as large as possible to produce reliable fracture properties. The fundamental measurement of tensile strength is a difficulty in beam testing due to neutral axis shifting and stress redistribution after the initiation of tensile fracture. 2.3.6 Uniaxial Direct Tension Test Haas (1973) stated that the stiffness modulus of an asphalt mixture should be obtained using a direct tension approach. The typical uniaxial direct tension test for asphalt mixtures uses 39

PAGE 40

a cylindrical specimen that is bonded to end caps of the same or slightly larger diameter as shown in Figure 2-6. Figure 2-6. Uniaxial direct tension test In theory, the uniaxial direct tension test benefits from the uniform stress and strain fields in the middle of the specimen, but in practice, evaluation of asphalt mixtures is difficult due to the many detrimental effects associated with this method. Bolzan and Huber (1993) summarized the following disadvantages: Stress concentration near the ends of the samples has been observed in many experiments. Sample failure can occur due to misalignment. Sample preparation requires a long time and a skilled operator/technician. The failure plane is presumed to occur at the center of the specimen perpendicular to the vertical axis, but in practice, it may occur at any locations over the specimens. This test has had some difficulty in obtaining repeatability 40

PAGE 41

Furthermore, it might not be possible to do this test on field cores since it would consist of several layers. That may not be thick enough to make cylindrical specimens for this test. 2.3.7 Closure In summary, none of existing test methods, including Superpave IDT, provides a suitable method for obtaining tensile properties of open graded asphalt mixture. Thus, it is necessary to develop a test method and system which mitigates the disadvantages of existing tests and at the same time, meets the following requirements: a direct tension mode test; can be performed on gyratory compacted specimens or field cores; can test samples of various thickness (thin or thick); and can be used on all mixture types (dense-graded, open-graded). 41

PAGE 42

CHAPTER 3 DEVELOPMENT OF DOG-BONE DIRECT TENSION TEST (DBDT) 3.1 Introduction A mechanical test to evaluate the tensile properties of asphalt mixture is necessary in mixture design, quality control, and thus pavement design. As mentioned in the literature review, because the existing direct tension test is unsuitable for evaluating the tensile properties of asphalt mixtures, a Dog-Bone Direct Tension test (DBDT) was conceived to overcome its disadvantages and to provide appropriate tensile properties of asphalt mixture. The DBDT provides some potential advantages including the fact that the failure plane is known a priori, which means failure limits can be measured directly on the failure plane. Due to DBDT specimen geometry, stress concentrations near the ends of specimen are less critical, and the location where failure is likely to occur is maximized. The DBDT specimens can be produced by simply coring opposing sides from slices or disks obtained from cylindrical laboratory samples or field cores. Development and evaluation efforts for the DBDT are described in the following sections. The intent behind the development of this direct tension test is for it to provide a comparison to the more easily performed Superpave IDT test. Thus, the new direct tension test should be viewed as a research tool, rather than a production test. Figure 3-1 shows a schematic drawing of load application in Superpave IDT and DBDT. Tensile stress is induced by applied compression load in Superpave IDT while DBDT is directly applied with tension load. 3.2 Optimization of Specimen Geometries for DBDT The primary objective of this section is to describe the design, development and evaluation of a new direct tension test system that was conceived for determining tensile properties of asphalt mixtures. 42

PAGE 43

Figure 3-1. Schematic drawing of load application in Superpave IDT and DBDT 3.2.1 Two Dimensional Finite Element Analysis Two dimensional finite element method (FEM) analyses were conducted to determine optimum width of the specimen, coring radius, and coring overlap considering stress distributions at both the centerline of the specimen and the area near the loading heads. Stress distributions of the centerline and headline of the specimen were checked to ensure higher enough stress concentration at edge of a specimen, thereby increasing potential for failure at the center of the specimen and not near the loading heads (problem with traditional specimens). Simultaneously, the stress distributions across the centerline should be as uniform as possible to guarantee stress concentration over a large enough area to capture a representative portion of the mixture. Coring radii (r) and coring overlaps (x) were variables used for analysis to optimize the specimen geometry as presented in Figure 3-2. 43

PAGE 44

CL rr xxw/2w/2 Asphalt SampleCoring Bit Variables: coring radius (r, in) 1.5, 2.0, 2.5, 3.0coring overlap (x, in) 0.5, 1.0, 1.5, 2.0 CL rr xxw/2w/2 Asphalt SampleCoring Bit CL rr xxw/2w/2 Asphalt SampleCoring Bit Variables: coring radius (r, in) 1.5, 2.0, 2.5, 3.0coring overlap (x, in) 0.5, 1.0, 1.5, 2.0 Figure 3-2. Coring radius (r) and coring overlap (x) Figure 3-3 shows two dimensional finite element mesh and its coordinate. Based on the predicted stress distributions for two-inch wide specimens, a coring radius of 3 inches and a coring overlap of 2 inches produced centerline stresses that were reasonably uniform and considerably greater than stresses near the loading heads (headline). Analysis results for this case are summarized in Table 3-1. This final specimen geometry results in a large enough cross section for testing without sacrificing the integrity of the mixture. As indicated in this table, a significant stress difference between the centerline and the headline of the specimen was observed. The stress concentration at the centerline should cause the specimen to break in this region, thereby compensating for any density gradients in gyratory compacted specimens as previously exposed by researchers (Harvey et al., 1991; Shashidhar, 1999, Chehab et al, 2000). As also shown in Figure 3-4, the percent difference in the computed stress along the centerline is around 30%. However, this 30% increase in stress between the outer edge and center of the 44

PAGE 45

specimen is about the same as that observed for Superpave IDT specimens from which material properties have been determined successfully (Roque and Buttlar, 1992). The key to success is to have the gages directly on the planes of maximum stress. Therefore, measurements are obtained in the center of edges of the specimen as well as on the specimen faces. Figure 3-3. 2-D FEM mesh and coordinates Table 3-1. Analysis results for determining DBDT sample geometry Various Coring Overlaps with Fixed Coring Radius (3 in) Coring Overlap (in) 0.5 1.0 1.5 2.0 % Difference of Stress on Center Line (%) 20.39 27.04 28.02 31.31 Stress Difference between Center Line and Head Line (%) 20.14 20.26 26.62 43.17 Various Coring Radii with Fixed Coring Overlap (2 in) Coring Radius (in) 1.5 2.0 2.5 3.0 % Difference of Stress on Center Line (%) 61.17 46.42 37.39 31.31 Stress Difference between Center Line and Head Line (%) 50.59 48.54 45.91 43.17 45

PAGE 46

Figure 3-4. Stress distribution on centerline and headline (3 in coring radius and 2 in coring overlap) 3.3 Dog-Bone Direct Tension Testing System After thorough analysis and discussion, a DBDT prototype system was proposed, designed and built. The complete system is composed of several pieces, including a specimen coring jig, specimen loading heads, strain gage sensors and attachment kits, a dual cylinder tensile load equalizer, and a PC controlled servo-hydraulic load frame with an integrally mounted environmental chamber. All shop drawings of components and subassemblies were generated during prototype development. The various components of the DBDT system that were designed, developed, and/or procured are described in this section. An asphalt concrete specimen instrumented with the new DBDT system and ready for testing in the loading frame is shown in Figure 3-5. 46

PAGE 47

The proposed system was designed to properly perform three types of mixture tests: resilient modulus, creep compliance, and tensile strength tests, which provide the properties needed to fully define fracture behavior using the HMA fracture model developed at the University of Florida (Zhang et al., 2001; Roque et al., 2004a). Figure 3-5. Dog-bone direct tension test prototype 3.3.1 Coring Process A coring jig was designed and built to make dog-bone shaped asphalt specimens. The opposing sides of the specimen were cored with the coring jig shown in Figure 3-6. 47

PAGE 48

Figure 3-6. Coring jig The coring barrel of the coring machine pushes down onto an asphalt specimen on the coring fixture, which has grooves at its base to accommodate the coring barrel after it goes through the base of the specimens. The coring jig has three bearings to prevent the coring barrel from chattering and wavering as well as hold it during the coring process. The coring fixture was fitted with an adjustable round shaped steel block, which restrains the specimen laterally to accommodate different diameter specimens. A hex key causes the block to move the specimen laterally such that the specimen is exactly centered on the coring fixture. The steel block is reversible so that both sides of the specimen can be cored without readjustment. 48

PAGE 49

3.3.2 Bonding Specimen to Loading Heads Suitable loading heads were designed to prevent eccentricity due to bending or misalignment. The system can be used with any independent loading frame including the MTS system. Loading heads were machined from 6061-T6 aluminum with V-grooved specimen contact surfaces to increase the bonding strength between the loading heads and the specimens as shown in Figure 3-7. Also, the upper loading head has two holes through which the load is applied with loading pins. Figure 3-7. Loading head with grooved contact surface To make the loading head usable for specimen of various thicknesses, a series of shim stocks were employed as shown in Figures 3-8 and 3-9. If the thickness of a specimen is not the same as that of loading heads, a series of shim stocks can be placed beneath either the specimen or the loading head in order to eliminate eccentricity due to bending or poor alignment. 49

PAGE 50

The upper and lower loading heads were connected with distance-adjustable two alignment bars as shown in Figures 3-8 and 3-9. These also help to ensure good alignment during bonding process. An area of concern was the bond between a specimen and the loading heads. Similar to the uniaxial direct tension test, this interface can make or break the test. Several epoxies and adhesives from different manufactures were procured and evaluated to determine which chemistry provided sufficient bond strength and practicality. The final selection was based on working time, curing time, and ultimate tensile strength. Detailed properties of the bond agent selected are shown in Table 3-2. This bond agent consists of two materials: resin and hardener. Figure 3-10 shows the bond agent and mixing nozzle. The ratio of resin to hardener is two to one by volume. The entire bonding process was performed on a machinist granite block to provide a flat reference surface as shown in Figure 3-11. Figure 3-8. Loading heads with alignment bars and shim stocks 50

PAGE 51

Figure 3-9. Loading heads with a series of shim stocks Table 3-2. Characteristics of LOCTITE Hysol Product E-20HP Physical Properties (25C) Dielectric Strength (Volts/Mil) 500 Tensile Strength, ASTM D 638 (psi) 5,700 Tensile Elongation, ASTM D 638 (%) 8 Hardness, ASTM D 1706 (Shore D) 80 Glass Transition Temperature, T g (C) 60 Curing Properties (25C) Working Life (min) 20 Tack Free Time (min) 40 51

PAGE 52

Figure 3-10. Bonding agent gun with mixing nozzle Figure 3-11. Machinist granite block 52

PAGE 53

3.3.3 Strain Gage Mounting System Four extensometers were used to measure on-specimen deformations. Two of the sensors are placed at the center of the specimens flat faces and the other two are placed at the center of curved edges. The extensometers placed on the faces are the most appropriate for determination of resilient modulus and creep compliance. Since the edge measures deformation in the immediate vicinity of maximum tensile stress, they can detect the instant when the material fractures and are most appropriate to determine failure limits (i.e., strength, failure strain, fracture energy etc.). Detailed stress analysis based on finite element method will be discussed in the section 5.5. At first, it was difficult to place and attach the aluminum gage points at the exact position planned, especially at the curved edges of the specimen. Longer gage points with an angle were conceived and prepared for both edges as shown in Figures 3-12 and 3-13. The same gage points used for IDT specimens were used for the flat faces. Also, simple templates shown in Figures 3-14 and 3-15 were designed and machined to mount gage points precisely at the center of test specimens face and edge considering the curvature of a specimen. Placement of the gage points with these templates guarantees that the gage points on one face or edge of the specimen are perfectly aligned with gage points on the other face or edge. Gage point spacing of 1.5 in was used. The longer angled gage points were formed to be more effective for use on the specimens curved edges. Figure 3-16 shows a specimen after attaching gage points on faces and edges. Knife edges are affixed on the gage points using a set screw. The extensometers are then clipped onto these knife edges to measure both face and edge deformations. Extensometers are designed to rely on a minimum spring force to maintain contact with knife edges mounted to gage points glued on the specimen. This is the same system currently in use for Superpave IDT. Knife edges were also 53

PAGE 54

modified to simplify their placement onto the gage points for edges as shown in Figures 3-17 and 3-18. Figure 3-12. Gage point positioning on edge Figure 3-13. Modification of gage point for edge 54

PAGE 55

Figure 3-14. Template to attach gage point on faces Figure 3-15. Template to attach gage point on edges 55

PAGE 56

Figure 3-16. Gage point attachment Figure 3-17. Schematic illustration of knife edge modification 56

PAGE 57

Figure 3-18. Modification of knife edge 3.3.4 Load Equalization System A load equalization system consisting of two interactive hydraulic cylinders was conceived and designed to assure uniform loading across the specimen, that is, to minimize or eliminate eccentric loading that can induce premature failure near the loading heads. This load frame system attaches to the loading heads with two pins to pull on the specimen in a uniform fashion. The upper chambers of both cylinders are filled and plumbed together, and the lower piston chambers are similarly plumbed. Therefore, as one piston goes up, the other piston comes down. In essence, if one side of the specimen starts to fail, the complete load is not transferred to the other side of the specimen. The piston rises moving hydraulic fluid to the other cylinder thereby reducing the load on the non-failing side until the load equalizes and then continues to equalize until failure. This takes place very quickly and extra low friction (ELF) seals were secured and 57

PAGE 58

incorporated inside each cylinder to reduce the effects of seal friction on load uniformity. Figure 3-19 shows the actual cylinder loading assembly and Figure 3-20 shows the loading assembly attached to loading rod. The load frame has a capacity of 22,000 pounds, which far exceeds the needs for most asphalt materials characterization, at even the coldest testing temperatures, and includes an environmental chamber manufactured by ESPEC. The temperature range of the environmental chamber is -20 C to 100 C, with an accuracy of 0.2 C, as required by Superpave specifications. Temperature control systems are also available at the University of Florida as shown in Figure 3-21. This type of environmental chamber is useful to maintain the specimen at test temperature during placement of the extensometers through two access holes. Since behavior of asphalt mixtures are highly dependent on temperature, careful control of chamber temperature is necessary to obtain accurate material properties. Figure 3-19. Dual cylinder loading assembly 58

PAGE 59

Figure 3-20. Dual cylinder loading assembly attached to loading rod Figure 3-21. Environmental chamber 59

PAGE 60

Hydraulic power is provided to the MTS load frame via an MTS SilentFlo hydraulic power supply. This hydraulic power supply provides the actuator, or system, high-pressure hydraulic fluid at 3000 psi, at a flow rate of approximately 5.5 gallons per minute. Heat removal from the hydraulic power supply is accomplished via a portable chiller manufactured by BUDZAR Industries, and is properly sized to the hydraulic power supply. A temperature control chamber is required to be rapid and stable such that a specimen should be quickly stabilized at the testing temperature. 60

PAGE 61

CHAPTER 4 TEST MATERIALS AND SPECIMEN PREPARATION 4.1 Materials This section provides information regarding the materials used for production of asphalt mixture specimens in the laboratory for this study. Two gradations of asphalt mixtures were used: one dense graded asphalt mixture and one open graded asphalt mixture. Usually, dense graded asphalt mixture is used for surface structural layer while open graded asphalt mixture is used as wearing course. 4.1.1 Dense Graded Asphalt Mixtures Georgia granite was used for dense graded asphalt mixtures in this study with both unmodified asphalt binder (PG 67-22) and modified asphalt binder (PG 76-22). Aggregate sources for dense graded asphalt mixtures are shown in Table 4-1. All mixtures were 12.5 mm nominal maximum aggregate size gradations according to Superpave system. Figure 4-1 shows the gradation chart of the dense graded asphalt mixture including the restricted zone and control points. Detailed gradation information and blend proportions are available in Appendix A. Two binders were used in this study: a control binder and an SBS modified asphalt binder. The control asphalt binder graded as a PG 67-22 or AC 30, while the modified asphalt binder graded as a PG 76-22. SBS polymer (3%) was blended with the control asphalt in the process to produce the SBS modified asphalt. Both of asphalt binders were provided by CITGO Asphalt Refining Company. It is of particular interest to compare an SBS modified (PG 76-22) mixture an unmodified mixture (PG 67-22), with all other mixture characteristics being the same. Previous research by Roque et al. (2004b) has shown that SBS modification results in slower rate of fracture damage in mixtures, as compared to unmodified mixtures. 61

PAGE 62

Table 4-1. Aggregate source for dense graded asphalt mixture Type of Material 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 Figure 4-1. Dense graded asphalt mixture gradation 4.1.2 Open Graded Asphalt Mixtures Two aggregates were used for open graded asphalt mixtures: Florida oolitic limestone and Nova Scotia granite. Both were used to produce mixtures conforming to FDOTs specifications for FC-5 mixture FC-5 is one kind of open graded friction course gradation designated in specifications of Florida Department of Transportation (FDOT, 2007). Table 4-2 shows aggregate sources and 62

PAGE 63

Figure 4-2 shows the gradation chart for the open graded asphalt mixtures. In this particular case, gradations follow that used in the test sections on U. S. Highway 27 Highlands County test project (Varadhan, 2004; Thai, 2005). These are aggregates extensively used in the state of Florida and are approved by the FDOT for road construction and rehabilitation projects. Table 4-2. Aggregate source for open graded asphalt 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 Florida Limestone Filler # 7 Stone 44 Martine Marietta Aggregates TM-322 NS-315 # 789 Stone 51 Martine Marietta Aggregates TM-322 NS-315 Screenings 22 Martine Marietta Aggregates TM-322 NS-315 Nova Scotia Granite Hydrated Lime Asphalt rubber binder (ARB-12) was used, along with 1% lime pretreatment for the granite mixture. Base binder for ARB-12 is unmodified binder, which is PG67-22 or AC-30 which is the same unmodified binder used for dense graded asphalt mixtures. Asphalt rubber binder (ARB-12) containing 12% ground tire rubber was blended prior to use. Based on multi-year research project on state road 16 in Florida, this pre-blended asphalt rubber binder (wet process) improved the cracking resistance of open graded asphalt mixture compared to the dry mix process (Choubane et al., 1998). They also suggested that the amount of rubber used for FC-5 be within 10 to 15%. 63

PAGE 64

Figure 4-2. Open graded asphalt mixture gradation 4.2 Asphalt Mixture Design 4.2.1 Dense Graded Asphalt Mixtures The dense graded asphalt mixtures produced for testing and evaluation were designed with the Superpave volumetric mix design procedure, which bases its selection for design asphalt content on a set of criteria on the volumetric properties of the mixture (VMA, VFA, density) at 4% air voids. Also, the aggregates need to meet criteria for the consensus and source properties that aim to prevent the use of substandard aggregates in producing asphalt mixture. 4.2.1.1 Selection of traffic level The Superpave design method for compacted asphalt mixtures specifies the number of gyrations to which a sample must be compacted with the Superpave gyratory compactor. The number of gyrations for these traffic levels as specified by the Florida Department of 64

PAGE 65

Transportation is presented in Table 4-3. Per FDOTs request, dense graded asphalt mixtures were designed for traffic level C, which is more than 3 million and less than 10 million ESALs. Table 4-3. Traffic levels and gyratory compaction efforts Traffic Level (Millions of ESALs) N ini N des N max A (<0.3) 6 50 75 B (0.3-3) 7 75 115 C (3-10) 7 75 115 D(10-30) 8 100 160 E ( 30) 8 100 160 4.2.1.2 Batching and mixing Aggregate batching sheets, attached in Appendix A, were prepared for 4500g samples based on the JMF for the aggregates. The batched 4500g samples were heated in oven at the mixing temperature for approximately 3 hours. The mixing tools and asphalt used were also heated at the mixing temperature. The mixing temperature of the unmodified mixture and modified mixture are 300-315F and 320-335F, respectively. The aggregates were then removed from the oven and mixed until the aggregates were well coated (approximately 3-5 minutes) with asphalt binder. Figure 4-3 shows a picture of laboratory mixer used. The mixed samples were spread out in pans and heated in an oven for 2 hours at the same temperature as mixing for short-term oven aging. Each of the mixtures was stirred after 1 hour to obtain a uniformly aged sample. 65

PAGE 66

Figure 4-3. Asphalt mixer 4.2.1.3 Compaction After short-term oven aging, the 4500g samples were then removed and quickly compacted using the Superpave Gyratory Compactor. Figure 4-4 shows the Superpave Gyratory Compactor used. The samples were compacted with a ram pressure of 600kPa at a gyratory angle of 1.25. The compaction data from the samples were used in determining the design asphalt content. That is, volumetric properties of the mixture such as air voids (AV), voids in mineral aggregates (VMA), and voids filled with asphalt (VFA) were calculated at these asphalt contents and then each was plotted as a function of asphalt content at N des The design asphalt content was obtained by interpolating the air void versus asphalt content curve to obtain asphalt content at 4% air voids. 66

PAGE 67

Figure 4-4. Superpave Servopac gyratory compactor The other volumetric properties were then obtained at this design asphalt content as presented in Table 4-4. Design asphalt content was determined to be 4.8% for dense graded asphalt mixtures. Table 4-4. Volumetric information for dense graded asphalt mixtures AC (%) Gmm Gsb VMA (%) VFA (%) 4.8 2.579 2.770 14.9 73.1 4.2.1.4 Long term oven aging (LTOA) Laboratory aging is one of the most influential factors that affect asphalt mixture properties and predicted performance. Aging effects were evaluated by comparing long term oven aging tension test results to short term oven aging tension test results. The mixes were subjected to long term aging according to AASHTO PP2 (2001). A new procedure was developed by 67

PAGE 68

Varadhan (2004) to contain the open graded compacted specimens from falling apart during aging. A wire mesh with openings of 1/8 and steel hose clamps were used. The mesh size was chosen to ensure good circulation of air within the sample for oxidation and at the same time, to prevent the smaller aggregate particles from falling through the mesh. The samples were placed on porous plates in an oven at 185F for 5 days. The specimen was turned over twice during long term oven aging. The test samples with the wired mesh and porous plate for LTOA are shown in Figure 4-5. Figure 4-5. Long term oven aging (LTOA) conditioning set-up 4.2.2 Open Graded Asphalt Mixtures Asphalt content for open graded asphalt mixture was determined to produce a minimum voids in the mineral aggregate (VMA) at a sufficient effective asphalt film thickness to ensure adequate durability of the mixture. The processes of batching and mixing are the same as those of dense graded asphalt mixtures. Mixing temperature was kept around 320-335F to obtain 68

PAGE 69

proper workability. A Superpave Gyratory Compaction (SGC) set to 50 gyrations was used for the open graded mixture compaction as suggested by prior research (Mallick et al., 2000; Varadhan, 2004) to meet specified air void content. The sample was allowed to cool down for about 2 hours after compaction before ejecting from the mold. This was done to prevent collapse of these high air void content mixtures. Once the specimen was ejected from the mold, it was allowed to cool for 5 minutes before removing from the ejector. All compacted specimens were then allowed to cool at room temperature for at least 24 hours before further handling. The optimum asphalt content was selected as the asphalt content resulting in the lowest VMA. Optimum asphalt contents were determined as 6.4% for Florida limestone and 6.0% for Nova Scotia Granite. Open graded asphalt mixtures were also subjected to LTOA. Since these mixtures are very coarse and open, special care must be taken to prevent specimen damage during aging process. 4.2.2.1 Measuring air void for open graded asphalt mixtures For many years, bulk specific gravity for asphalt mixture have been measured by weighing compacted specimens in air and water, commonly referred to as water displacement. Test procedures for this method are provided in AASHTO T 166 (AASHTO, 2001) or ASTM D 2726 (ASTM, 2002). However, it has been found that this method results in erroneous air void contents when testing coarser gradations such as open graded mixture and stone matrix asphalt (SMA) (Buchanan, 1999; Cooley et al., 2002). Therefore, determination of G mb with the Corelok device has been recommended for coarse graded asphalt mixtures. Crouch et al. (2002) stated that the Corelok device performed well with a variety of sample types and was the most widely applicable method of G mb determination. Figure 4-6 shows the Corelok equipment and Figure 4-7 shows the vacuum sealing process for asphalt mixture. 69

PAGE 70

Figure 4-6. Corelok equipment to measure air void Figure 4-7. Vacuum sealing in corelok 70

PAGE 71

This method is specified in ASTM D 6752 (2002) as a vacuum sealing method. Bulk specific gravity (G mb ) was calculated using following equation (InstroTek, 2003): FWWWWWGcbsdbdmb21 (4-1) Where, W d1 = dry sample weight before sealing (g), W d2 = dry sample weight after water submersion (g), W b = bag weight (g), W s = sealed sample weight in water (g), F c = bag volume correction factor provided by manufacturer. Dimensional analysis for measuring bulk specific gravity specified in AASHTO T 269 (2001) was also performed on open graded asphalt mixture. Alvarez et al. (2008) recommended utilizing dimensional analysis for determining Gmb of porous friction course mixtures rather than the vacuum sealing method. They also stated that this method is relatively simpler, fast, and less expensive and additional testing supplies are not needed. The dimensional analysis method directly calculates volume assuming that the specimen is a regular cylinder with smooth faces. The dimensional G mb was calculated as: wmmmbVWG (4-2) Where, W m = total dry weight of asphalt mixture (g), V m = total volume of asphalt mixture (cm 3 ), w = density of water (g/cm 3 ) Table 4-5 and Figure 4-8 show the air void contents from both Corelok and dimensional method. 71

PAGE 72

Table 4-5. Air void contents for open graded asphalt mixture Dimensional Method Corelok Method Type of Aggregate Test G mm G mb Air Void (%) G mb Air Void (%) IDT 1.918 16.9 1.963 15.0 DBDT 1.927 16.6 1.973 14.5 Florida Limestone Ave. 2.309 1.922 16.7 1.968 14.8 IDT 1.948 20.2 1.982 18.8 DBDT 1.953 20.0 1.997 18.2 Nova Scotia Granite Ave. 2.441 1.950 20.1 1.990 18.5 Figure 4-8. Air voids of open graded asphalt mixture These results imply that the air voids determined from dimensional analysis are higher than those obtained from the vacuum method. This discrepancy between two methods resulted from the surface air voids. Dimensional analysis includes all surface voids, whereas the vacuum method partially includes them because the bag partially follows the surface voids with depth once the vacuum sealing process is applied. 72

PAGE 73

The specimens were then measured for thickness at four equally spaced locations approximately 90 apart, and an average thickness was calculated. These specimens were laboratory produced and compacted in a Superpave Gyratory Compactor (SGC), therefore, their diameter was fixed at 150 mm. 4.3 Test Specimen Preparation 4.3.1 Dog-Bone Direct Tension Test Specimen Dog-Bone Direct Tension Test (DBDT) specimens were prepared with the materials for dense and open graded asphalt mixtures mentioned section 4.1. 150mm diameter specimens were compacted by the Superpave gyratory compactor. After waiting for the specimen to fully cool down, for dense graded asphalt mixture, the bulk specific gravity of each compacted specimen was measured, and the air void content was determined. The target air voids of the gyratory pill was 7.5%, because the air void of the sliced specimen taken from the middle of the pill is generally 0.5% less than that of the compacted specimen. The compacted specimens were sliced with a masonry cutting saw with a special attachment to hold the pills (Figure 4-9) to obtain the desired thickness of specimen for each gradation (1.5 inch thickness for dense graded asphalt mixtures 2 inch for open graded asphalt mixtures). Based on prior research done by Varadhan (2004), this thickness was determined to minimize end effect during testing. Also, this thickness provides for representative sample of OGFC. Figure 4-10 shows that after cutting, each side of the test specimens was cored to produce the dog-bone shape specimen using a specially designed coring fixture. As stated in section 3.3.1, coring fixture was designed to core each opposing side using round shape steel block as shown in Figure 4-11. Figure 4-12 shows the specimen after coring each side. Because the coring saw uses water to keep the blade wet, the specimens were dried one day at room temperature prior to further testing. 73

PAGE 74

Figure 4-9. Slicing asphalt mixture Figure 4-10. Coring asphalt mixture sliced 74

PAGE 75

Figure 4-11. Coring the other side of asphalt mixture sliced Figure 4-12. Specimen after coring 75

PAGE 76

The bulk specific gravity was taken to obtain air void contents for each dense graded specimen. Specimens had to be in the range of 7 0.5% air voids to be considered for testing. Specimens were placed in the humidity chamber for at least two days to negate moisture effects. To increase the effectiveness of the bonding interface, the top and bottom of the test specimens were coarsely sanded to remove the asphalt film as shown in Figure 4-13. The test samples were then bonded to the loading heads using simple alignment bars. Figure 4-14 shows a specimen being bonded to the loading heads. All bonding was performed on a machinist granite block to provide a flat reference surface. After bonding, the specimens were allowed to cure for one day at room temperature. This ensures the bonding agent has revealed full strength. Gage points were attached onto the face and edge of specimens using a simple steel template, adhesive, and activator for rapid cure. Figure 4-13. Sanding asphalt mixture to eliminate asphalt film 76

PAGE 77

Figure 4-14. Bonding between a specimen and loading heads Figures 4-15 and 4-16 illustrate the attachment of gage points. Four pairs of gage points were placed at distance of 1.5 in for each pair: two on the faces and two on the edge along the vertical axes on the center of the specimen. Figure 4-17 shows how knife edges are placed over the gage points. Prior to testing, the prepared DBDT specimen was kept in the environmental chamber at the desired test temperature for at least six hors to ensure uniform temperature within the specimen. Mounting of extensometers is completed in the environmental chamber through access holes. 77

PAGE 78

Figure 4-15. Gage points attachment on faces Figure 4-16. Gage points attachment on edges 78

PAGE 79

Figure 4-17. Set-up of knife edge on gage points 79

PAGE 80

CHAPTER 5 DBDT TEST PROCEDURE AND DATA INTERPRETATION METHODS 5.1 Introduction The need for asphalt mixtures tests that satisfactorily provide tensile fracture properties of mixture is well recognized since it was observed that the asphalt mixture properties cannot be reliably determined through the current empirical binder-to-mixture relationships (Roque et al., 1997). Superpave IDT and a newly developed tension test, DBDT, were conducted to evaluate tensile properties of dense graded and open graded asphalt mixtures. Three types of mixture tests were performed: resilient modulus test, creep compliance test, and tensile strength test. Figure 5-1 shows test plan with Superpave IDT and DBDT test for this study. Figure 5-1. Test plan 80

PAGE 81

5.2 Dog-Bone Direct Tension Test Procedures Three kinds of tests (i.e., resilient modulus, creep compliance, and strength test) were performed on all mixtures to determine resilient modulus, creep compliance, m-value, tensile strength, failure strain, fracture energy, and dissipated creep strain energy to failure. As mentioned previously, these tests provide the properties needed to define cracking performance using the HMA fracture model developed at the University of Florida (Zhang et al., 2001; Roque et al., 2004a). 5.2.1 Resilient Modulus Test The resilient modulus is defined as the ratio of the applied stress to the recoverable strain when repeated loads are applied. It is one of the most common methods for measuring the stiffness of the asphalt mixtures. The resilient modulus test was performed in load control mode by applying a repeated haversine waveform load to the specimen for 0.1 second along with a rest period of 0.9 second. The load was selected to keep on-face strain between 100 and 300 micro-strains. A typical load versus time plot is shown in Figure 5-2. The procedures for resilient modulus test are as follows: (1) Dog-bone test samples are stored in a humidity chamber at a constant relative humidity of 60% for at least 2 days prior to testing to ensure uniform moisture conditions. Specimens are there cooled at the test temperature for at least 6 hours before testing. (2) Extensometers are mounted and centered on the specimen to the gage points for the measurements of the face and edge deformations. (3) A constant preloading of approximately 5 pounds is applied to the test specimens to ensure proper contact between loading pins and holes in upper loading head through which the load is applied. 81

PAGE 82

Figure 5-2. Typical resilient load (4) The specimen is then tested by applying a repeated haversine waveform load for seven seconds to obtain on-face strain between 100 to 300 micro-strains. If the horizontal strains are higher than 300 micro-strains, the load is immediately removed from the specimen, and the specimen is allowed to recover for a minimum 3 minutes before reloading at a lower load. (5) When the applied load is determined, the data acquisition program begins recording test data. Data are acquired at a rate of 500 points per second. (6) The resilient modulus is calculated based on three dimensional finite element analysis described in section 5.3. 82

PAGE 83

5.2.2 Creep Test Creep compliance is a function of time dependent strain over constant stress. Creep compliance curve is useful not only for predicting thermally induced stress in an asphalt pavement but also for evaluating the rate of damage accumulation of asphalt mixture. The creep test requires the application of a constant load, applied as a step function, for 1000 seconds. Figure 5-3 shows typical creep load measurements. The magnitude of load is adjusted so that the horizontal deformations meet certain criteria. These criteria are limits set on the horizontal deformations at 100 seconds, and as a not-to-exceed value at the end of the test as mentioned below. These limits can vary with test temperature or with the specimen type, but most importantly, with heavily aged specimens. The load was selected to keep the maximum on-face strain below 750 micro-strain. Figure 5-3. Typical creep load 83

PAGE 84

The test procedures for creep test with DBDT system consist of the following steps: (1) The test sample preparation, mounting the extensometers, and preloading procedures are the same as those for resilient modulus test. (2) The static loading is applying and held for 1000 seconds. If the on-face strains are not between 100 and 180 micro-strains at 100 seconds or are higher that 750 micro-strains at 1000 seconds, the load is immediately removed from the specimen, and specimen is allowed to recover until deformations are relatively constant before reloading at a different level. (3) When the applied load is determined, the data acquisition program records the loads and deformations at a rate of 10 Hz for the first 10 seconds, 1 Hz for the next 290 seconds, and 0.2 Hz for the remaining 700 seconds of the creep test. The creep parameters, which are m-value, D 1 creep compliance, and creep rate, are calculated based on three dimensional finite element analysis. 5.2.3 Tensile Strength Test Failure limits including tensile strength, failure strain, and fracture energy were determined from strength test. These properties are used for estimating the cracking resistance of the asphalt mixtures. The strength test was conducted by applying a constant rate of displacement of loading ram until the specimen failed. After careful observation of displacement rate from two testing modes, displacement rate of DBDT test was determined based on the comparison between horizontal deformation of Superpave IDT and on-face deformation of DBDT test as shown in Figure 5-4. It should be noted that a given rate of ram displacement will result in a strain rate in the Superpave IDT that is approximately twice as high as in the DBDT. Therefore, in order to achieve approximately the same strain rate in both tests, a deformation rate of 1in/min for dense graded asphalt mixture was used in DBDT test compared to the standard 2in/min used in the 84

PAGE 85

Superpave IDT. For open graded asphalt mixture, a deformation rate of 2in/min was used in DBDT test. It should be noted that the non-uniform stress distribution on the DBDT cross-section results in non-uniform strain rates in the cross-section. Therefore, the actual strain rate at the edge of the DBDT specimen will be greater than at its center of face and greater than the Superpave IDT. The on-face and on-edge deformation, and the applied load are recorded. Figure 5-4. Determination of strain rate for strength test 5.3 Data Interpretation Methods for Dog-Bone Direct Tension Test The development of appropriate data interpretation methods to accurately determine material properties based on measurements obtained from the DBDT is described in this section. Accurate determination of properties requires accurate prediction of stresses within the specimen, an approach that can accurately identify the instant of fracture from load and deformation measurements, and procedures to properly determine strain accurately from deformation measurements. Developments in each of these areas are described in the sub-sections below. 85

PAGE 86

5.3.1 Three Dimensional Finite Element Analysis Three dimensional finite element analysis was conducted to analyze stress and strain of a 6-inch diameter specimen with the proposed system. The commercial computer program ADINA was used for analysis (ADINA, 2005). To simplify analysis and for time savings, symmetry was taken advantage of in all three dimensions. Therefore, only one eighth of the specimen was analyzed without sacrificing any accuracy or information. Roller supports were applied at each node along planes of symmetry. The program ADINA provides proper supports to meet given conditions at the lines and points where the roller supports meet. Full friction was assumed at the interface between steel loading head and the specimen. Figure 5-5 shows the 3-D FEM mesh used to represent the specimen and the coordinate system used. The analysis was conducted assuming an asphalt modulus of 400,000 psi, while varying Poissons ratio, the center width, and thickness of the specimen. Figure 5-5. 3-D FEM mesh and coordinates 86

PAGE 87

5.3.1.1 Stress analysis Stress at the center of a specimen of uniform cross-section can simply be calculated as applied load divided by cross-sectional area. However, results of the 3-D FEM analyses presented in Figure 5-6 indicate that the stress distribution of the DBDT specimen having a 2-inch thickness and a 2-inch width varies from the center of the face to the center of the edge. The tensile stresses vary along centerline of the specimen (xand yaxis in Figure 5-6) and reach a maximum at the edge of the specimen. According to this analysis, tensile failure will initiate at the edge of the specimen. An important finding is that stress distributions are nearly independent of the Poissons ratio. Based on multiple 3-D finite element analyses performed with a range of variables, stress correction factors were developed to determine the point stress on the face and edge of the proposed DBDT system based on average cross-sectional stress (P/A). The corrected point stresses on the face and edge can be calculated using the following equations 5-1 to 5-4. CAPfcfc (5-1) CAPecec (5-2) 960700158001980.w.t.Cf (5-3) 034610901001770.w.t.Ce (5-4) Where, fc = stress on face corrected, ec = stress on edge corrected, P = applied load (lbs), A c = area at the center of a specimen (in 2 ), C f = stress correction factor on face, C e = stress correction factor on edge, t = thickness of a specimen (in), w = width at the center of a specimen (in). 87

PAGE 88

Figure 5-6. Stress distributions at the center of DBDT specimen (2-in thickness specimen with 2-in width) 5.3.1.2 Strain analysis The strain values obtained from the DBDT test are based on the point-to-point displacements measured by the finite length extensometers and are better represented by the average of the strain distributions between gage points as expressed by equations 5-5 and 5-6. 2211/GL/GLfadxxfGL (5-5) 2221/GL/GLeadxxfGL (5-6) Where, fa = average strain between two gage points on face, ea = average strain between two gage points on edge, GL = gage length (in) = distance over which vertical deformations are obtained. Since only the average strain can be measured with a finite length gage, strain correction factors were developed to convert the average strain determined from the extensometer to a 88

PAGE 89

vertical point strain at the center of a specimens face and edge. Figures 5-7 and 5-8 present the strain distributions between gage points on the face and edge. It was noted that these values were nearly independent of Poissons ratio. With respect to strain on the face (Figure 3-27), small differences with Poissons ratio were observed. However, it was determined that the maximum error in strain resulting from assuming an incorrect value of Poissons ratio was only 2.6%. The strain is corrected using following equations 5-7 to 5-10. Cffafcfp (5-7) Ceeaep (5-8) 054510178000280.w.t.Cf (5-9) (5-10) 07701009000150003902.w.t.t.Ce Where, fp = fc =corrected vertical point strain at the center of a specimens face, ep = corrected vertical point strain at the center of a specimens edge, C f = strain correction factors on face, C e = strain correction factors on edge, t and w as previously defined. Figure 5-7. Strain distributions on faces (2-in thickness specimen with 2-in width) 89

PAGE 90

Figure 5-8. Strain distributions on edges (2-in thickness specimen with 2-in width) 5.3.1.3 Rotation effect on edge measurements Figure 5-9 shows how rotational effects of the gage points can influence the deformation measurements at the edges. According to the analysis, it was found that the rotational effects are significant, but not affected by Poissons ratio, thickness or width of specimen. Figure 5-10 presents the effect of Poissons ratio for a 2-inch thick specimen with a 2-inch width. This shows a normalized deformation along axis of the gage point, where the contact surface between specimen and the gage point is given a value of 1. A single correction factor can be applied in all cases as long as gage point length is known. Displacement correction factors can be obtained with knowing that the actual measurement is comprised of the expected measurement plus two times the measurement error as shown in Figure 5-9. 90

PAGE 91

Figure 5-9. Rotational effect on edges Figure 5-10. Normalized displacement along axis of gage point on edge (2-in thickness specimen with 2-in width) 91

PAGE 92

Cdeeaed (5-11) Where, = corrected strain for rotation effect on the edge, Cde= displacement correction factoran be obtained using a combination of equat ed s for the edge = 0.81, ea as previously defined. Therefore, a corrected point strain on the edge c ions 5-8 and 5-11. CCdeeeaec (5-12) Where, = corrected strain for rotation effect on the edge, All other variables as previoulus Test n extensively used in the pavement engineering community to evalu a rest r measured face defortotal unloading and rest-period portion of each cycle. ec usly defined. 5.3.2 Resilient Mod Resilient modulus has bee ate the relative quality of mixtures and to predict load response of pavements subjected to wheel loads. The resilient modulus test is performed in load control mode by applying a repeated haversine waveform load to the specimen for a period of 0.1 second followed byperiod of 0.9 seconds. The load for the resilient modulus test was selected to keep the on-surface strains below linear viscoelastic range. Resilient modulus is a function of the load, specimen dimensions, and resilient deformation. For the DBDT specimen, stress and strains required focalculation of resilient modulus can be determined using the correction factor equations which were developed based on 3-D finite element analysis as described above. The resilient modulus of asphalt mixture was calculated by using the mation. The instantaneous resilient modulus is calculated by using the recoverable deformation that occurs instantaneously during the unloading portion of each cycle. The resilient modulus is calculated by using the total recoverable deformation, which includes boththe instantaneous recoverable and the time-dependent recoverable deformation during the 92

PAGE 93

Analysis procedure is summarized as followings: (1) Determine the beginning of each load cycle. e was determined as the first data point in a s thss of how small), and for which the total o le. s shown in Figure 5-12, the maximum load is determined as the peak load and the seating load iermded immediately prior to the beginning of the load cycle. For each load cycle, the beginning of the load cycl eriesat has three sequential load increases (regardle f the three load increases is greater than 10 lbs. Figure 5-11 shows the load start point usedfor analysis based on measured load data. The noise in the load measurements can be no greaterthan 10 lbs for this approach to work. Figure 5-11. Determination of beginning of load cycle (2) Determine the maximum load for each cyc A s detined by averaging the 30 loading points recor 93

PAGE 94

(3) Calculate the stress for each cycle. Load amplitude for each cycle is simply maximum load minus the seating load priocycle: r to the )i_(seati1 PPmax_i P (5-13) Stress for each cycle can be simply calculated using following equations with correction factors. C Afcfci Pi(5-14) CAPecieci (5-15) Where, P = load amplitude for each cycle, Pmax_i = maximum load for each cycle, Pi-1) = seating load prior to each cycle, i = number of cycles (1 ~ 7), fci = corrected stress on the specimen facd stress on the specimen edge for each cyc Figure 5-12. Determination of load amplitude i seat_( e for each cycle, eci = correctele. 94

PAGE 95

(4) Calculate the instantaneous and total recoverable deformation for each cycle The maximum deformation is determined as the peak deformation for each cycle (see Figure 5-13), and determining the value of the deformation is made by performing the regression is of recovery portion of each deformation wave. The definitions of the instantaneous and total recoverable deformations for resilient modulus test are presented in Figure 5-13. From the maximum deformation point of each cycle, three data points are skipped and the next seven data points are selected to perform the regression on the unloading portion of the wave (regression #1, Figure 5-14). Starting at the end of the loading cycle, the previous 100 data points were used to performe linear regression analysis (regression #2, Figure 5-14). analys Figure 5-13. Definition of instantaneous and total recoverable deformation 95

PAGE 96

The intersection of the two best fit lines was used to compute the instantaneous recoverable as shown in Figures 5-14. The total recoverable deformation is based on the ation at the end of the load cycle where the deformation is defined as the intersection ofregression line #2 and the beginning of the following cycle. deform Figure 5-14. Determination of instantaneous and total recoverable deformation (5) Calculate the instantaneous and total recoverable strains for each cycle 2211/GL/GLIi_fadxxfGL (5-16) 2211/GL/GLTi_fadxxfGL (5-17) 2221/GL/GLIi_eadxxfGL (5-18) 2221/GL/GLTi_eadxxfGL (5-19) 96

PAGE 97

Where, fa_Ii = instantaneous average strain between two gage points on face for each fa_Ti = total average strain between two gage points on face for each cycle, ea__Ii = taneous average strain between two gage points on edge for each cycle, ea__Ti = total average strain between two gage points on edge for each cycle. Corrected point strains are determined as follows: cycle, instan CfIi_faIi_fc (5-20) CfTi_faTi_fc (5-21) CCdeeIi_eaIi_ec (5-22) CCdeeTi_eaTi_ec (5-23) Where, fc_Ii = instantaneous corrected strain at the center of a specimens face for each le, ec_Ii =instanneous corrected strain at the center of a specimens edge for each cycle, ec_Ti = total corrected stra(6) Obtain trimmed mean modulus for each replicate Even though seven cycles are recorded, only five are analyzed. It was determined that seven load cyssure that five completeoad cycles are recordeseven resilient mare ranked. The highest and lowest values are deleted and the remaining five moduli are averaged to obtain the trimmed mean moduli. cycle, fc_Ti = total corrected strain at the center of a specimens face for each cyc ta in at the center of a specimens edge for each cycle. cles must be obtained to a ld. The oduli from the seven cycles 27iMinMaxMfcifcifcin_RfI (5-24 1iIi_fcIi_fcIi_fc) 21iMinMaxMiTi_fcTi_fcTi_fcn_RfT (5-25 7fcifcifci) 2iMinMaxMn_IRe (5-26) 7eciecieci 1iIi_ecIi_ecIi_ec 97

PAGE 98

2 7eci 1iTi_ecn_TRe iMinMaxMTi_ececiTi_ececi (5-27) Whereent modulus on face, MRfT = total resilient modulus on face, MReI = ulus on edge, MReT = total resilient modulus on face, n = numb M RfI = instantaneous resili instantaneous resilient mod er of gage on faces or edges from three replicates (3 replicates 2 faces/edges) (7) Obtain trimmed mean modulus from three replicates To more reliably obtain resilient deformation, three replicate specimens were tested and evaluated. Six moduli with on-face measurements and six moduli with on -edge measurements were obtainend lowest moduli were deleted and the rem d. The six moduli from each case were ranked. The highest a aining four moduli were averaged to obtain the trimmed mean modulus. 21nMMinMMaxMMn_RfIn_RfInn_RfIRfI (5-28 6) 26nMMinMMaxMMn_RfTn_RfTn_RfTRfT (51n29) 261nMMinMMaxMMn_efIn_IRenn_IReIRe (5-30) 261nMMinMMaxMMn_efTn_TRenn_TReTRe (5-31) According to Roque et al. (1997), however, the method used to determine the instantaneous recoverable deformation should be reviewed. This meod may not be suitable for mixtures that exhibit significant delayed elastic recovery, where the recovery portion of the unloading deformation curve is highly nonlinear since this fits a straight line though the last 100 pointsf th of each loading cycle regardless of mixture stiffness. At this present time, thus, the used ototal resilient modulus is recommended because it is less affected by this effect. 98

PAGE 99

5.3.3 n ed from creep compliance tests. D1 and m-value have been used to determine the rate of damage accumulation in asphalt mixture. Analysis procedure is summarized as follows: (1) Calculate stresses and strains with correction factors The stresses and strains for face and edge were corrected with correction factor equations. Creep Test Creep compliance is a function of time-dependent strain over stress. Creep compliance cabe used to predict thermally induced stress in asphalt pavement, as well as to evaluate the rate of damage accumulation of asphalt mixture. D 0 D 1 and m-value are mixture parameters obtain CPfffc00 Ac(5-32) C Apeceec00 (5-33) Cttffafc (5-34) Ctteeaec (5-35) fcecorrected strain ecimens edge with time. Where, fc = f0 = constant stress with time on face, ec = e0 = constant stress with time on face, (t) =corrected strain at the center of a specimens face with time, (t) =c at the center of a sp (2) Calculate creep compliance For uniaxial stress conditions, creep compliance is defined as time-dependent over constant stress (equations 5-36 and 5-37). Six creep compliances on faces from three replicates can be fitted with the generalized power. Similarly, six compliances are obtained on edges. n_fcn_fct (5-36) n_ftD 99

PAGE 100

n_ecn_eWhere with time on face, D(t)e = creep compliance wiedge, n = number of gage on faces or edges from three replicates (3 replicates 2 faces/edge) (3) pliance The triance was obtained by plotting creep compliance (Figure 5-15) and ranking each of the face and edge compliance. The data in the middle of test were used e compliance were ranked numeiddle four edge compliances. included in the calculation. Results from the Superpave IDT introduced by Buttlar and Roque n_ecttD (5-37) D(t) f = creep complianceth time on Obtain trimmed mean creep com mmed mean of creep compli for ranking. That is, ranking was performed using sum of compliance between 450 seconds and 550 seconds (equation 5-38). These sums of the face and edg rically to identify the middle four face and m Figure 5-15. Trimmed mean method for obtaining the creep compliance The upper and lower values from the three replicates were discarded (trimmed) and not 100

PAGE 101

(1994) showed that more reliable estimates of creep compliance were obtained using thetrimmed mean approach. 550450dttDtDtestofmiddle (5-38) Where, D(t)middle of test = sum of creep compliance in the middle of test. After trimming the upper and lower values, the remaining four compliances were averaged to create representative creep parameters for face and edge. DDttDfmftrimmed_fctrimmed_fctrimmed_ff10 (5-39) DDttDemetrimmed_ectrimmed_ectrimmed_ee10 (5-40) Where, D(t)f_trimmed, D0f, D1f, mf = representative creep parameters on face at certain condition, D(t)e_trimmed, D0e, D1e, me = representative creep parameters on edge at certain condition. is possible to determine a measure of the mixtures stiffness by computing its tangent modulus from stress-strain results. The true tensile strength of an asphalt mixture can be determined by determining the stress level at the edge of the specimen at the instant of failure. To accomplish this experimentally, a measurement system must be able to detect the instant when fracture occurs at the edge of the specimen. According to 3-D finite element analysis, fracture should occur on either edge first or 5.3.4 Tensile Strength Test F ailure limits of asphalt mixtures provide insight into their resistance to fracture. In addition, failure limits are required to determine whether load and temperature induced stresses and strains will cause failure of asphalt pavement. Also, since stress-strain information can be obtained from the strength test, it 101

PAGE 102

on both edges simultaneously in the DBDT testing system, since tensile stress is highest at the edges. Analysis procedure for strength test is summarized as follows: quential ses, and the total load increase is greater than 30 lbs as shown in Figure 5-16. This t data points are part of the loading cycle and not just initial load fluctu Figur6. g (1) Determine the beginning of the loading The start of the load is determined as the first data point in a series that has three seload increa guarantees that the subsequen ation or noise. e 5-1 Determination of the beginning of loadin 102

PAGE 103

(2) Determine the end of the loading It was determine d that the instant of failure can be detected by examining the difference between the average face deformations and each edge deformation, since failure first occurs at the location where the rate of deformation increases as shown in Figure 5-17. Figure 5-17. Instant of fracture for DBDT system ) Calculate the strength of each specimen ensile strength of each specimen is calculated as follows: (3 T CPSefracturet (5-41) Where, St = tensile strength, Pfracture = applied load at the instant of fracture, A Ac) Calculate stress and strain at loading cycle time c = initial cross section area of the specimen. (4 103

PAGE 104

Stresses and strains are calculated from the start of the load to the instant of specimen failure. When calculating stress, the initial seating load is subtracted from the applied loads to simulate a zero stress state at the beginning of the loading cycle. Failure strain was determined strain on edge at failure. CAPtPtecseating (5-42) CCttdeeeaec (5-43) Where, (t) = corrected stress during loading time (t), P(t) = tensile load during loading time (t), Pseating = initial seating load applied to specimen, ec (t) = corrected strain effect on the edge during loading time (t). modulus o obtain fracture energy and tangent modulus, the stress-strain curve was fitted with a polynomtion that forces the fitted curve through the origin. (5) Calculate fracture energy and tangent T ial function. The three order polynomial func t at(5-44) tata33221 data at the end of the test. Sinceres elastic, it was decided to reduce the influence that this data ha r the curve, which presents the fracture energy as shown in Figure 5-18. It was noticed that when the function was not forced through the origin, the fit of the curve through the initial stress-strain points was strongly influenced by the the ponse at the end of the test is probably not purely s on the initial tangent modulus. By applying this constraint, the basic law that strain equals zero when stress equals zero is satisfied. By taking the derivative ofthe function, a tangent modulus can be computed at any strain level. Also, the function can be integrated to obtain the area unde 104

PAGE 105

The fitrom stress-strain data that does not have the seati includr, ted curve was obtained fng stress ed. As mentioned earlier, this simulates a zero stress state at time equal to zero. Howevewhen calculating fracture energy, the initial seating load must be included. failt@seatingfailt@Ptatata033221 (5-45) eremed mean method which eliminates the highest and lowest values was also used for analyzing the data. Tensile strengy (FE), and tangent modulus (MT) were obtained from strength test. t@FE Wh, @t=fail = failure strain. Figure 5-18. Calculation of fracture energy (6) Obtain trimmed mean values from three replicates Like resilient modulus and creep compliance tests, three replicate specimens were tested and evaluated to reliably obtain the tensile properties from strength test. Trim th (S t ), failure strain ( failure ), fracture energ 105

PAGE 106

261nSMinSMaxSSn_tn_tnn_tt (5-46) 261nMinMaxn_failuren_failurenn_failurefailure (5-47) 261nFEMinFEMaxFEFEn_n_nn_ (5-48) 261nMMMaxMMn_Tn_Tnn_TT (5-49) All the variables were previously defined. 5.4 Verification of DBDT Correction Factors 5.4.1 Calibration Sample (Delrin Specimen) Preliminary tests were performed using the proposed DBDT prototype to investigate the feasibility and accuracy of this system for determining the tensile properties of asphalt mxtures addition, preliminary tests are to evaluate all the sub-systeere any needs for modification. The se oise aluated at this time. Delrin plastic specimen was fabricated for calibration purposes. It has the same dimension as asphalt mixture tested in the prototype of DBDT test system. Only difference is that calibration specimen is one single unit as shown in Figure 5-19, not separad with loaheads. Published material specifications for Delrin report a modus of 3.1 Gpa (450 kresilient modulus test was performed on the Delrin DBDT specimen to determine the material modulus. i and to validate the correction factors identified. In ms of the testing unit and determine whether there w ub-systems to be tested include: the on-specimen measuring system, the loading system, thdata acquisition and control systems. Modifications to the subsystems were made as necessary, until the entire unit was functioning properly. Preliminary testing of the prototype system was performed on material of known mechanical and physical properties, Delrin. The system nand data quality were also ev teding lusi). A 106

PAGE 107

The specimen was loaded using a repeated haversine waveform lasting 0.1 seconds, followed by a 0.9 second rest period. Figure 5-20 shows resilient deformation results on face and on edg Figure 5-19. Test set-up on Delrin calibration sample e. Results from the test presented in Figure 5-21 clearly show that before applying the correction factors, moduli on the edges and on the faces are significantly different and do not agree with the modulus reported by the manufacturer. However, after stress, strain and rotation correction factors are applied, modulus values as calculated from the edge and face deformations approach or equal the manufacturer reported value of 3.1 GPa. This result appears to indicate that these correction factors allow stress and strain to be accurately determined from the proposed DBDT system. The verification of correction factors with asphalt mixtures is also shown in the following section. 107

PAGE 108

Figure 5-20. Resilient deformation on Delrin calibration sample 5 00.51.522.53.54M (GP) 134.51234Ra Edge Face Before After Strain & tion A fter Stres s Correction Correction Rotation Correc Manufacturer ReportedValue : 3.1GPa Figure 5-21. Resilient modulus on Delrin calibration sample 108

PAGE 109

1095.4.2 Asphalt Mixtures It is imperative to validate correction factors based on three dimensional finite element analysis for asphalt mixtures even though they were verified with a Delrin plastic specimen. The results from on-face and on-edge measurements in two testing modes, resilient modulus and creep compliance, were evaluated for this purpose. Figure 5-22 shows that resilient modulus results from face and edge measurements were in excellent agreement, regardless of mixture type, once correction factors were applied. This indicates correction factors work well for short loading times or small strain tests. Figure 5-22. Resilient modulus results on face and edge Examination of creep compliance test results which involve longer loading times (1000 nmodified short term oven aged dense graded asphalt mixture at multiple temperatures are presented in Figure 5-23. As can be seen in this Figure, creep compliance curves from both face and edge measurements were in seconds) yielded similar results. Creep compliance results of u

PAGE 110

excellent agreement. This was also observed for open graded asphalt mixture as shown in Figure 5-24,. ogmixtures on face and edge which shows creep compliance curves for long term oven aged Florida limestone OGFCTo corroborate this further, creep compliance values for all the mixtures evaluated were compared in Figure 5-25. Again, values from both measurement positions were in excellent agreement. Based on these comparisons, it appears that correction factors developed result in accurate predictions of stresses and strains in asphalt mixtures subjected to tension using the dbone direct tension test system proposed. In addition, correction factors were found to be independent of loading time. Figure 5-23. Creep compliance of Dense (Unmod) STOA 110

PAGE 111

Creep compliance of FC-5 (FLime) LTOA mixtures on face and edge Figure 5-24. Figure 5-25. Creep compliance results on face and edge 111

PAGE 112

CHAPTER 6 MIXTURE PERFORMANCE EVALUATION 6.1 Introduction A mixture tension test is necessary to evaluate and control cracking performance of the asphalt mixtures for use in the pavements. Both Superpave IDT and a newly developed DBDT test were performed on all mixtures at various conditions. The use of both tests might enhance reliability of the asphalt mixture properties. However, routine use of a direct tension test may not be feasible due to the multiple complications involved with this test such as specimen preparation and test interpretation. Therefore, it would be advantageous to adopt the Superpave IDT to the testing of open graded asphalt mixtures. Also, there is a concern that the open graded nature of these mixtures may require a modified IDT testing protocol from that used for dense graded asphalt mixtures. A study is essential to optimize the test conditions needed to obtain consistent fracture properties of open graded asphalt mixtures with the Superpave IDT. In particular, the effects of temperature need to be evaluated. Thus, experiments were performed with asphalt mixtures at four temperatures to evaluate the precision and accuracy of the new DBDT testing system. At the same time, cracking performance of the asphalt mixture was evaluated with three different loading modes. 6.2 Use of Non-Uniform Stress States of Tests for Tensile Failure Limits Superpave IDT and DBDT share a common advantage. Both tests result in non-uniform tensile stress states within a mixture, which allows for identification of the failure plane a priori. Cracking (fatigue, thermal, and top-down) in flexible pavement is one of the most common and crucial distresses that clearly affects the service quality and life of flexible pavement. ciency that should be addressed pavement design and maintenance planning. In order to evaluate and predict asphalt pavement Therefore, cracking is an important structural and functional defi in 112

PAGE 113

performance and develop appropriate models with respect to cracking resistance, it is necessary to measure the tensile failure limits of asphalt mixtures such as tensile strength, failure strain, and fracture energy. For many years, various testing methods have been used for evaluating tensile failure limits of asphalt mixture that are related to cracking performance in the field. However, if the properties are not fundamental, the results may be dependent on the testing systems employed. Fracture energy is believed to be one of the most important failure limits for describing and modeling the fracture behavior of asphalt mixtures. It is defined as the energy required to initiate fracture in a mixture. Previous studies have determined that fracture energy is a reliable indicator of the crack resistance of a mixture when other conditions such as pavement structure and traffic are similar (Sedwick, 1998). Research conducted by Wen and Kim (2002) confirms that fracture energy is an excellent indicator of the resistance of mixture to fatigue cracking through the use of the indirect tension test. In the HMA fracture mechanics model developed at the University of Florida, fracture energy is one of the key properties governing a mixtures resistance to fracture (Zhang, 2000; Zhang et al., 2001). Tensile strength and failure strain have also been commonly used for evaluating cracking performance of the flexible pavement. However, these properties are rate-dependent, so they are not fundamental (i.e., independent of geometry and mode of loading). It is noted that once fracture occurs in asphalt mixture, post behavior has no meaning in terms of cracking performance and is not easily interpretable because of localizing effects of the crack. Thus, for an accurate determination of the tensile failure limits, the testing system should capture the exact instant of fracture. According to Roque et al. (1997), first fracture in the indirect tension test sample occurs prior to the peak load. They developed an approach to identify 113

PAGE 114

the instant of first fracture in the Superpavusing vertical and horizontal measurements. Tensile strength, failure strain, and fracture energy were obtained using load and deformt. 6.2.1 Uniform Stress State of Uniaxial Direct Tension Test For many years, it has been assumed that uniform normal stresses distribute over a cross-section of a cylindrical specimen, providing satisfactory results unless sudden changes in cross-section along the specimen were involved. The stress and strain can be simply calculated with following equations 6-1 and 6-2. e IDT sample ation measurements corresponding to the instant of first fracture. The same phenomenon was observed in the dog-bone direct tension test in this study. First fracture occurs either at the center of one edge or the center of both edges at the same time where tensile stress is the highes A Papplied (6-1) L G (6-2) Where, = stress over the cross-section, Papplied = applied load, A = cross-section area, G = global strain, = global deformation, L = height of a specimen. However, it has also been recognized that uniform stress distribution changes once non-uniform (non-homogeneous) damage and/or the first fracture occurs within the specimen, most frequently near the loading platens. Consequently, the stress and/or strain distribution cannot be predicted or determined after non-homogeneous damage or first fracture since one does not know the exact location and geometry of damage zones and cracks, as shown in Figure 6-1. Therefore, failure limits including fracture energy calculated from global stress-strain measurements from uniform specimens may not be accurate or reliable. Measurements obtained from direct tension tests can only provide an average failure strain and energy. It must be acknowledged that microcrack and localized damage occur throughout the loading within a 114

PAGE 115

whole specimen. The key assumption that must be made in calculating and/or determining fracture energy from a non-uniformly loaded specimen is the surface area of the first crack areover which the fairly uniform maximum tensile stress was acting. a that most microdamage also occurs there. If the measurement systems could be well positith the deformation and gage length, respectively. Figure 6-1. Uniform stress condition under direct tension Using a non-uniform stress state test is one possible solution (Figure 6-2). Non-uniform stress state or a stress concentration would induce the first fracture at a known location, indicating oned in that region, predicted stress and measured strain through first fracture would be more accurate. Strains, which can be measured at the location of first fracture, will be less affected by microdamage that occurs throughout the rest of the specimen, whereas global strains are affected by all damage throughout the specimen. The local strains can be calculated wisame equation 6-3, substituting global deformation and height of specimen with local 115

PAGE 116

ll Where, = local strain, = local deformation, l = gage length. (6-3) l Figure 6-2. Non-uniform stress state of Superpave IDT and DBDT 6.2.2 occurs at a lower load than that res f ading phalt Non-Uniform Stress State of Superpave IDT As mentioned previously, fracture in asphalt mixture usually quired to break the whole specimen. In other words, the load to initiate fracture is alwayless than the maximum load measured in the system (P fail < P measured = P max ). Thus, it is necessaryto measure the load at the instant of fracture (P fail = P measured < P max ) to accurately determine failure limits. Therefore, identification of first fracture is very important. Furthermore, it is oconsiderable interest that comparative analysis of failure limits at first fracture obtained fromindirect and direct loading conditions would be made to verify their independence of loconditions. Superpave IDT has been extensively used to characterize the tensile properties of asmixtures among pavement engineers and researchers since it was introduced to the asphalt 116

PAGE 117

community. This test was developed and improved during the Strategic Highway ResearchProgram (Roque and Buttlar, 1992; Buttlar and Roque, 1994). Roque and Buttlar (1992) showedthat the Superpave IDT overcame many of the problems that have been typically associated with the existing indirect tensile test. One of the major improvements they made was the introduction of on-face measurements at the center of the specimen, where maximum compressive and tensile stresses take place. Figure 6-3 represents the tensile stress distribution on the horizontal axis, which indicates that first fracture will develop on the vertical axis at the center of the specimen. Therefore, measurements can be obtained on the failure plane, which allows accurate determination of failure limits. That is, more accurate measurements can be obtained on the failure plane by placing a horizontal sensor at the center of the specimen. The on-face measurement system also eliminates effects of specimen rotation. Roque et al. (1997) also developed a method for identifying first fracture using the vertical and horizontal measurements on the center area of the indirect tension specimen. This idea was initiated with the fact that the tal deformation increased. According to Romeo (2008), this method to detect fracture initiation matches very well with the obsers ttlar failure occurred first at the location where the rate of horizon vation using Digital Image Correlation (DIC) system. Test procedure and analysis methodfor Superpave IDT were well documented in the references (Roque and Buttlar, 1992; Buand Roque, 1994; Roque et al, 1997). For convenience, the method developed to consistently identify the instant of fracture from deformation from Superpave IDT is described below. 117

PAGE 118

Figure 6-3. Non-uniform stress state of Superpave IDT from 3D FEM analysis In the Superpave IDT system, the extensometers measure deformations in the immediate vicinity of maximum tensile stress, and were found to be suitable in identifying the instant when cking initiates at the faces eformation in the vicinity of the crack can be seehe f se of entified as the time when the difference between the vertical and horizontal defor fracture develops at the face of the specimen. At the instant when cra of the specimen, an increase in the rate of horizontal d n due to the weakening of the specimen near the crack. If this is the case, the instant of specimen fracture can be identified by analyzing the rate of deformation of the horizontal strain gage during strength test. Figure 6-4 presents the method to detect the instant of fracture from deformation measurements. The change in horizontal response can be identified by plotting tdifference between vertical and horizontal measurements. When first fracture occurs, the rate ohorizontal deformation will increase relative to the rate of vertical deformations. This will cauthe difference between horizontal and vertical deformations to reach a maximum. The instant fracture is id mations reaches a peak. If both sets of vertical and horizontal gages reach a peak, the instant of fracture is taken as the time when the first peak is reached. 118

PAGE 119

Figure 6-4. Detection of the fracture instant of Superpave IDT 6.3 Superpave IDT Test Results The test results obtained from Superpave IDT were analyzed with the ITLT program developed at the University of Florida. Detailed analysis and interpretation methods for IDT test results were presented by Roque, Buttlar and their associates (1992; 1994; 1997). Superpave IDT results presented in Appendix B, appear to follow the expected trend with respect to temperature, aging condition, and binder modification. For example, as test temperature increased, resilient modulus and strength exhibited lower values while the absolute value and rate of creep compliance and failure strain showed higher values. Also, results indicated that aging conditioning made an asphalt specimen stiffer. With respect to binder modification, SBS decreased creep response and increased fracture energy. However, SBS did not affect resilient modulus or strength. 119

PAGE 120

The resilient modulus (MR) is a measure of a materials elastic stiffness. Resilient modulus results are shown in Figure 6-5. As expected, the dense graded asphalt mixture had greater resilient modulus than open graded asphalt mixture. This clearly indicates that mixture structure affects resilient response. Figure 6-5. Resilient modulus from Superpave IDT test Creep compliance is related to the ability of a mixture to relax stresses. Generally speaking, mixtures with higher creep compliances can relax stresses faster than mixtures with low creep compliance, which is primarily of significance for thermal stress evaluation. As shownin Figures 6-6 and 6-7, open graded asphalt mixture and unmodified dense graded asphalt mixture shows higher creep response than modified dense graded asphalt mixture. 120

PAGE 121

Figure 6-6. Creep compliance from Superpave IDT test Figure 6-7. Creep rate from Superpave IDT test 121

PAGE 122

The SBS polymer appears to have much greater influence on the time-dependent response of dense graded asphalt mixture than on the short loading time response. As shown in Figur6 and 6-7, it was also noted that reduction in absolute value and rate of creep compliance of unmodified dense graded m es 6-ixture with aging is remarkable compared to modified dense graded asphalt mixture. This implies that modified binder reduces the rate of oxidative aging. fore fracture. As odulus. ixtures. Figure 6-8. Strength from Superpave IDT test dense graded asphalt mixtures than for open graded asphalt mixtures at both aging conditions as presented in Figure 6-9. With respect to open graded asphalt mixture, age conditioning did not have significant influence on the failure strain for either aggregate type. This might be due to binder Tensile strength is the maximum tensile stress the mixture can tolerate beshown in Figure 6-8, results of tensile strength had a similar trend as that of resilient mDense graded asphalt mixtures have higher tensile strength than open graded asphalt m Failure strain, which characterizes the brittleness of a mixture, was higher for 122

PAGE 123

contents. The binder contents of open graded asphalt mixtures are higher than that of dense graded asphalt mixture and it might take open graded asphalt mixtures more time to be aged enough to influence failure strain. Similar values of fracture energy and dissipated creep strain energy were observed regardless of age conditioning for open graded asphalt mixtures. Figure 6-9. Failure strain from Superpave IDT test Fracture energy (FE) results are shown in Figure 6-10. Fracture energy is calculated from the indirect tensile strength test by computing the area under the stress-strain curve up to the point of first fracture. The fracture energy of dense graded asphalt mixture is much higher than that of open graded asphalt mixture at all test temperatures and aging conditions. As mentioned above, there is no significant difference in fracture energy between two aging conditions for difference was observed between the two aggregate types. open grade asphalt mixture. Also, no considerable 123

PAGE 124

Figure 6-10. Fracture energy from Superpave IDT test The dissipated creep strain energy at failure (DCSEf) is defined as the fracture energynus the elastic energy (Zhang, 2000). Since DCSEf is a function of the fracture energy, it generally followed a similar trend as fracture energy (Figure 6-11). The energy ratio (ER) was developed by Roque et al. (2004a) to represent the cracking ance of asphalt mixtures. Energy ratios at 10C with STOA conditioned asphalt mishown in Figure 6-12. This parameter allows the evaluation of cracking performance for different pavement structures by incorporating the effects of mixture properties and pavem miresistxture are ent nsidered as a criterion for fatigue cracking. Higher ER values typically result for mixture with higher DCSEf and lower creep rate. The results indicate the significantly higher energy ratio of the dense graded asphalt mixtures compared to open graded asphalt mixtures. Even though the energy ratio was developed for dense graded mixtures and it has not been truly calibrated for friction course mixtures, it may at structural characteristics. Energy ratio can be co 124

PAGE 125

least provide a relative evaluation for cracking performance, which indicates that open gradeasphalt mixtures would exhibit worse cracking performance than dense graded asphalt mixtur d e. Figure 6-12. Energy ratio from Superpave IDT test Figure 6-11. Dissipated creep strain energy from Superpave IDT test 125

PAGE 126

6.4 Dog-Bone Direct Tension Test Results The newly developed dog-bone direct tension test (DBDT) was also performed with the three loading modes as for Superpave IDT (i.e., resilient modulus, creep compliance, and strength test). Test procedure and data analysis were already discussed in the chapter 5. All DBDT test results are presented in Appendix C. Generally, the DBDT test results shown in Figures 6-13 through 6-19 followed the similar trend as results from Superpave IDT. Resilient modulus results are presented in Figure 6-13. Dense graded asphalt mixture showed higher resilient modulus than open graded asphalt mixture, and the resilient modulus of estone open graded asphalt mixture was higher than granite open graded asphalt mixture. These observations clearly indicate that elastic response of asphalt mixture is dependent on samelim. characteristics of aggregate gradation and aggregate type Figure 6-13. Resilient modulus from DBDT test 126

PAGE 127

Absolute value and rate of creep compliance are shown in Figures 6-14 and 6-15. Clearlymodified dense graded asphalt mixture exhibited lower creep compliance at 1000 seconds and lower creep rate than open graded asphalt mixture. It is well known that creep behavior of asphalt mcombasphalt msignificanill be presen Figure 6-14. Creep compliance from DBDT test ixture is strongly related to both aggregate and binder characteristics. Thus, the ination of dense gradation and binder modification plays a beneficial role in reducing the absolute value and rate of creep compliance. Strength results shown in Figure 6-16 are very intriguing. As expected, dense graded ixture showed much higher strength than open graded asphalt mixture. However, no t difference with temperature or aging condition was observed. This was unexpected and different from Superpave IDT test results. Potential explanations for these observations wted later in this chapter. 127

PAGE 128

Figure 6-15. Creep rate from DBDT test Figure 6-16. Strength from DBDT test 128

PAGE 129

Figure 6-17 shows the failure strain from DBDT. Failure strain seems to be strongly related to mixture gradation. Dense graded asphalt mixture showed higher values of failure strain. This can be observed more clearly for the short term oven aging condition. As for thSuperpave IDT results, the change of failure strain due to aging was relatively small for opgraded asphalt mixture. As mentioned previously, this might be due to higher binder contents inopen graded asphalt mixture. e en asphalt mpresents dissmiwhich clearly shows fracture occurred at the center of the specimen, as expected. As indicated Figure 6-17. Failure strain from DBDT test Fracture energy of dense graded asphalt mixture was also higher than that of open graded ixture as shown in Figure 6-18 because of higher strength and strain. Figure 6-19 ipated creep strain energy at failure (DCSEf). Higher DCSEf of dense graded asphalt xture is attributed to higher fracture energy. Figure 6-20 shows a specimen after strength test, 129

PAGE 130

previously, it is one of advantages of DBDT test system that failure limits can be measured directly on the failure plane because failure plane is known a priori. Figure 6-18. Fracture energy from DBDT test rgy from DBDT test Figure 6-19. Dissipated creep strain ene 130

PAGE 131

131 Figure 6-20. Specimen after strength test Because energy ratio has not been fully calibrated for DBDT, it is not certain whether energy ratio determined from DBDT provides reliable information for evaluating top-down ing performance. Therefore, HMA fracture mechanics model, which is based on more crack growth was used to assess DBDT data. HMA fractu fundamental prediction of crack initiation and re mechanics model developed at the University of Florida (Zhang et al., 2001) was used toevaluate top-down cracking performance by calculating the number of load cycles needed tocreate a certain length of crack. STOA conditioned asphalt mixtures at 10C were used for this analysis. Figure 6-21 shows crack depth as a function of load cycles and Figure 6-22 illustrates the number of load cycles needed to create a 1 inch crack. Results clearly indicate that modified

PAGE 132

dense graded asphalt mixture exhibits the best performance while two types of open graded asphalt mixtures show the poorest cracking performance. Figure 6-21. Crack length with load cycles Figure 6-22. No. of load cycles at 1 inch crack length 132

PAGE 133

1336.5 Comparison between Superpave IDT and DBDT Test Results This section provides comparisons of results from Superpave IDT and DBDT tests. It is generally thought that Superpave IDT test is the most convenient test to evaluate tensile properties of asphalt mixture while direct tension mode is more of a research tool for the asphalt testing community. Therefore, comparative analysis of two tests might be of great interest. Figure 6-23 presents resilient modulus comparison for all tests performed. The comparison clearly shows that resilient modulus from the two tests were in excellent agreement, which indicates that elastic response due to rapid and repeated loading could be reliably obtained regardless of testing mode, i.e. indirect and direct mode. respectively. These two properties show almost the same tendency between the two tests. Figure 6-23. Resilient modulus from Superpave IDT and DBDT tests Absolute value and rate of creep compliance are shown in Figures 6-24 and 6-25, 133

PAGE 134

Absolute value and rate of creep compliance from DBDT at 1000 seconds are higher than thofrom Superpave IDT. Christensen and Bonaquist (2004) ob se served a similar trend when they studied Superpave IDT and uniaxial tension test. They showed significant discrepancy between two tests, having higher compliance in uniaxial tension test. They reported that this might result from the anisotropy and differences in air void, air void distribution, or both. However, there may be another potential reason. That is related to the fundamental behavior of asphalt mixture. Superpave IDT induces tension stresses in the presence of significant confinement induced by vertical compressive stresses. Lanaro et al. (2009) stated that compressive confinement is induced perpendicularly to the direction of the tensile stress in IDT test system. Behavior of particulate materials, such as granular soils and asphalt concrete at high and intermediate temperatures, is known to be strongly influenced by confinement. Greater confinement results in higher stiffness or lower compliance, which is the effect reflected in the T results in lower compliance. data: for a given level of stress, Superpave ID 134

PAGE 135

Figure 6-24. Creep compliance from Superpave IDT and DBDT tests ion-(wa relativetermined that this was due to the difference in strain rate between Superpave IDT and DBDT. Because of stress concentrations at the specimens edge, the DBDT tension test results in higher strain rates than Figure 6-25. Creep rate from Superpave IDT and DBDT tests It indicates that uniaxial stresses in the surface course (no confinement) accelerate tensinduced damage accumulation relative to uniaxial stresses at the bottom of the structural layer ith confinement). These results also indicate that open graded asphalt mixture exhibited greater confining effect than dense graded asphalt mixture. Open graded asphalt mixture that has ely larger dominant aggregate size, which has a higher friction angle that results in a greater confinement effect in the IDT system. Figure 6-26 presents strength comparisons between the two tests. As can be seen, strength results from DBDT are higher than those from Superpave IDT. It was d 135

PAGE 136

Superpave IDT, even when the same rate of displacement is used for both tests. This effect was also observed in differences in initial tangent modulus presented in Figure 6-27. Figure 6-26. Strength from Superpave IDT and DBDT tests Figure 6-27. Initial tangent modulus from Superpave IDT and DBDT tests 136

PAGE 137

As shown in Figure 6-28, failure strain obtained from DBDT is g enerally less than that from Superpave IDT. This is also explained by the difference in strain rate between the two tests. It is wgure 6-28 showkes sense, sinost Figur ell known that higher strain rate results in lower failure strain and higher strength. The higher failure strain, the greater the difference in failure strain between the two tests. Fis that the difference between the tests is negligible for very brittle mixtures. This mace more brittle materials also exhibit more elastic, less strain-rate dependent behavior. One of the more interesting observations from this study is that fracture energy was almexactly the same for the two tests (Figure 6-29), even though the tensile strength and failure strain between Superpave IDT and DBDT were different. e 6-28. Failure strain from Superpave IDT and DBDT tests 137

PAGE 138

Figure 6-29. Fracture energy Superpave IDT and DBDT tests act a fundamental property that is independenportant models. ure 6-e DBDT warences in the -31 shows schematic This provides further support that fracture energy is in f t of stress state and loading conditions, and is thus the most appropriate and necessary property to evaluate mixture fracture potential. Significantly different stress and strain states are induced in DBDT than in Superpave IDT and both strength and failure strain are known to be stress-state and load-rate dependent. However, fracture energy, which accounts for both stress and strain, was found not to be dependent on these factors. This result is in itself a very imvalidation involving the used of fracture energy as the basis for the HMA fracture mechanics Dissipated creep strain energy at failure (DCSEf) was also compared as presented Fig30. This was calculated based on fracture energy and resilient modulus. DCSEf from ths slightly lower than that from Superpave IDT. This might result from slight diffe stress-strain curves from the strength tests. For illustrative purpose, Figure 6 138

PAGE 139

relatio Figure 6-31. Schematic illustration of stress-strain relation between DBDT and Superpave IDT nship of strength, failure strain, fracture energy, and DCSE f from two tests. DCSE is inherently dependent on loading rate. This makes sense since the faster loading rate, the more elastic response, while the slower loading rate, the less elastic response. Figure 6-30. Dissipated creep strain energy from Superpave IDT and DBDT tests 139

PAGE 140

Since the stress-strain relationship is different for each test while both fracture energy and resilient modulus are almost the same between two tests, DCSEf from DBDT shows slightly lower values. This was corrected by using the tangent modulus to compute DCSEf, since elastic energy is dependent on loading rate. Equation 6-4 shows rate correction to the MR from DBDT test, which is illustrated in Figure 6-32. This equation does not mean that tangent modulus should be used for distinguishing elastic energy (EE) from fracture energy (FE). Resilient modulus is still recommended since it is more consistent and repeatable than tangent modulus. However, it was thought that the simplest and easiest way to incorporate the effect of loading rate was to use the ratio of tangent modulus between two loading rates. i.e., this was done for comparative purposes only. MIDT_TWhere, M MMMDBDT_TDBDT_Rcorrected_DBDT_R (6-4) R_DBDT_corrected = corrected resilient modulus from DBDT used in obtaining energy ratio, M R_DBDT = resilient modulus from DBDT, M T_DBDT = initial tangent modulus from DBDT, MT_IDT = initial tangent modulus from Superpave IDT. Corrected DCSEf was re-plotted as shown in Figure 6-33. As can be seen, DCSEf from both tests were in excellent agreement. 140

PAGE 141

Figure 6-33. Corrected dissipated creep strain energy from Superpave IDT and DBDT tests Figure 6-32. Loading rate correction with tangent modulus 141

PAGE 142

CHAPTER 7 ures g g may be summarized as follows: Both Superpave IDT and DBDT provided reasonable and consistent test results with respect to test temperature, aging and binder modification. Observed differences were well explained by satisfactorily known phenomena including effects of stress state and strain rate on mixture properties. Both Superpave IDT and DBDT tests clearly showed that open graded asphalt mixture had energy) than dense graded asphalt mixture. Resilient modulus values obtained from both Superpave IDT and DBDT were almost y obtained regardless of testing mode, i.e. indirect or direct. Absolute value and rate of creep compliance were highly correlated between Superpave T and DBDT but were lower for Superpave IDT than for DBDT, which was attributed to the higher confinement in Superpave IDT. This shows that rate of damage is in fact dependent on stress state. It also indicates that IDT may underpredict the rate of damage for open graded friction course, which is generally subjected to uniaxial tension at the pavement surface. Because of its lower fracture energy and greater rate of damage relative to conventional dense graded asphalt mixture, it appears that OGFC accelerates development of cracking in pavement relative to pavement without OGFC. Non-uniform stress states induced in both Superpave IDT and DBDT result in the unique advantage that the failure plane is known a priori, so that tensile failure limits can be CLOSURE 7.1 Summary and Findings This study was conducted to develope a system to determine tensile properties of OGFC sothat its effect on top-down cracking resistance of asphalt pavement could be evaluated. A new dog-bone direct tension test (DBDT) was conceived, developed and validated for asphalt mixtures. Resilient modulus, creep, and strength tests were performed at multiple temperaton dense graded and open graded asphalt mixtures with the newly developed DBDT and existinSuperpave IDT. The tensile properties of dense and open graded asphalt mixtures were successfully obtained with the DBDT developed, as well as with the Superpave IDT. Findingsassociated with tension testin much lower resilient modulus and failure limits (strength, failure strain, and fracture identical, indicating that elastic response due to rapid and repeated loading can be reliabl ID 142

PAGE 143

measured directly on the failure plane. This advantage mitigates the effects of non-homogeneous microdamage and/or marcrocrack that might be induced in the asphalt mixture during testing using a uniform stress distribution. It appears that although tensile strength and failure strain between Superpave IDT and DBDT are different due to strain rate, the resulting fracture energy from the two tests is the same. This supports previous work indicating that fracture energy is a fundamental property of asphalt mixture. 7.2 Conclusion A comprehensive evaluation of the fracture resistance of open graded asphalt mixture was completed. The following conclusions were based on the above findings and results: Open graded friction course mixture appears to reduce the cracking performance of asphalt pavement. One of potential reason is that the thresholds of fracture energy and dissipated creep strain energy for open graded asphalt mixture shows considerably lower values than dense graded asphalt mixture. This results in lower energy ratio and less number of load MA fracture mechanics model. oth Superpave IDT and the new DBDT test developed in this study provide suitable and accurate tensile properties of asphalt mixture, including open graded asphalt mixture. Thus, continued use of Superpave IDT is recommended because it provides reasonable properties and is much more practical. However, relationships to account for stress state effects need to be considered for mechanisms involving uniaxial stress states. DBDT creep compliance is more appropriate for uniaxial stress states, while Superpave IDT creep compliance is more appropriate for stress states involving confinement. Excellent agreement in fracture energy was observed between Superpave IDT and DBDT, indicating that fundamental properties can be accurately determined using either test. This further supports that fracture energy is in fact independent of stress state, loading conditions, and specimen geometry. It is thus the most appropriate and necessary property to evaluate mixture fracture potential. 7.3 Recommendations Based on extensive evaluation throughout this study, the following section will present recommendations for further investigation of effect of the open graded friction course and interface condition on the top-down cracking resistance of flexible pavement system. The new DBDT test system developed in this study were evaluated with only laboratory produced specimens. Laboratory specimens were also used for evaluation of fracture cycles to create a certain length of crack obtained by H B 143

PAGE 144

resistance of open graded asphalt mire more insight into the crack behavior of asphalt pavement, there is need to evaluate the fracture properties of field cores gathered from various sections. Aging conditioning procedures that may capture the characteristics of OGFC and bonded certainly has an influence on cracking performance of asphalt pavement and need to be e of most critical issues that need further investigation when evaluating the cracking behavior different healing characteristics than that of dense graded asphalt mixture. Current characteristics are fully integrated when evaluating asphalt mixture. One of the primary functions of open graded friction course is to drain the pavement affected by damage induced by water and should have suitable moisture resistance. To oisture conditioning. rates s racture resistance and performance characteristics. ixture. To acqu interface need to be developed for more accurate and reasonable evaluation. Age hardening properly taken into account when assessing the tensile properties of asphalt mixtures. Healing characteristics were also not evaluated in this study. Healing is probably onof asphalt mixtures. The physical difference of open graded asphalt mixture might result indeficiencies in evaluating asphalt mixture may be compensated for if healing quickly during rainfall events. Thus, these mixtures and bonded interface could be more evaluate the moisture resistance, there is need to develop proper m Finally, a rational design specification of open graded asphalt mixtures which incorpotensile properties needs to be developed. The design specification should be practical and feasible. It should also be integrated with pavement design and analysis process as well aaddress and optimize their f 144

PAGE 145

APPENDIX A LABORATORY MIXTURES INFORMATION 145

PAGE 146

A-1. JMF for Georgia Tablegranite dense gradation W-10 Local # 78 Stone # 89 Stone Screenings Sand JMF Sieve (mm) Sieve Size 33% 7% 50% 10% 100% 19.0 3/4 100.0 100.0 100.0 100.0 100.0 12.5 1/2 97.0 100.0 100.0 100.0 99.0 9.5 3/8 59.0 99.7 100.0 100.0 86.5 4.75 # 4 9.0 30.0 100.0 100.0 65.1 2.36 # 8 4.0 4.0 70.0 100.0 46.6 1.18 # 16 2.0 2.0 42.0 100.0 31.80.600 # 30 2.0 1.0 25.0 94.0 22.6 0.300 # 50 1.0 1.0 16.0 53.0 13.7 Gsb2.809 2.799 2.77 2.626 2.770 0.150 # 100 1.0 1.0 10.0 11.0 6.5 0.075 # 200 1.0 1.0 7.0 3.0 4.2 0 Pan 0.0 0.0 0.0 0.0 0.0 Table A-2. Batch weight for Georgia granite dense gradation Retained Weight, g Sieve (mm) Sieve Size # 78 Stone # 89 Stone W-10 Screenings Local Sand 19.0 3/4" 0.0 1,485.0 1,800.0 4,050.0 12.5 1/2" 44.6 1,485.0 1,800.0 4,050.0 9.5 3/8" 608.9 1,485.9 1,800.0 4,050.0 4.75 # 4 1,351.4 1,705.5 1,800.0 4,050.0 2.36 # 8 1,425.6 1,787.4 2,475.0 4,050.0 1.18 # 16 1,455.3 1,793.7 3,105.0 4,050.0 0.600 # 30 1,455.3 1,796.9 3,487.5 4,077.0 0.300 # 50 1,470.2 1,796.9 3,690.0 4,261.5 0.150 # 100 1,470.2 1,796.9 3,825.0 4,450.5 0.075 # 200 1,470.2 1,796.9 3,892.5 4,486.5 0 Pan 1,485.0 1,800.0 4,050.0 4,500.0 Sum 1,485 315 2,250 450 146

PAGE 147

Table A-3. JMF for Florida limestone opiller JMF en gradation S1A S1B Screenings F Sieve (mm) Sieve Size 44.7% 49.4% 3.2% 2.7% 100% 19.0 3/4 100.0 100.0 100.0 100.0 100.0 12.5 1/2 79.0 100.0 100.0 100.0 90.6 9.5 3/8 36.0 92.0 100.0 100.0 67.4 4.75 # 4 7.0 26.0 100.0 100.0 21.9 2.36 # 8 3.0 7.0 68.0 100.0 9.7 1.18 # 16 3.0 3.0 67.0 100.0 7.7 0.600 # 30 3.0 3.0 55.0 100.0 7.3 0.300 # 50 3.0 2.0 35.0 100.0 6.1 0.150 # 100 2.0 2.0 14.0 100.0 5.0 0.075 # 200 1.0 1.0 3.0 100.0 3.7 0 Pan 0.0 0.0 0.0 0.0 0.0 Gsb2.425 2.451 2.527 2.600 2.445 Table A-4. Batch weight for Florida limestone open gradation Retained Weight, g Sieve (mm) Sieve Size S1A S1B Screenings Filler 19.0 3/4" 0.0 2,011.5 4,234.5 4,378.5 12.5 1/2" 422.4 2,011.5 4,234.5 4,378.5 9.5 3/8" 1,287.4 2,189.3 4,234.5 4,378.5 4.75 # 4 1,870.7 3,656.5 4,234.5 4,378.5 2.36 # 8 1,951.2 4,078.9 4,280.6 4,378.5 1.18 # 16 1,951.2 4,167.8 4,282.0 4,378.5 0.600 # 30 1,951.2 4,167.8 4,299.3 4,378.5 0.300 # 50 1,951.2 4,190.0 4,328.1 4,378.5 0.150 # 100 1,971.3 4,190.0 4,358.3 4,378.5 0.075 # 200 1,991.4 4,212.3 4,374.2 4,378.5 0 Pan 2,011.5 4,234.5 4,378.5 4,500.0 Sum 2,011.5 2,223.0 144.0 121.5 147

PAGE 148

Table A-5. JMF for Nova Scotia granite open gradatio n Hydrated # 7 # 789 Screenings Lime JMF Sieve (mm) Sieve Size 76.7% 11.7% 10.6% 1.0% 100% 19.0 3/4 100.0 100.0 100.0 100.0 100.0 12.5 1/2 95.0 100.0 100.0 100.0 96.2 9.5 3/8 64.0 92.0 100.0 100.0 71.5 4.75 # 4 11.0 20.0 97.0 100.0 22.1 2.36 # 8 3.0 5.0 68.0 100.0 11.1 1.18 # 16 2.0 3.0 43.0 100.0 7.4 0.600 # 30 2.0 3.0 28.0 100.0 5.9 0.300 # 50 2.0 3.0 18.0 100.0 4.8 0.150 # 100 2.0 3.0 11.0 100.0 4.1 0.075 # 200 1.1 2.5 8.0 100.0 3.0 0 Pan 0.0 0.0 0.0 0.0 0.0 G sb 2.627 2.633 2.580 2.600 2.622 Table A-6. Batch weight for Nova Scotia granite open gradatio n Retained Weight, g Sieve (mm) Sieve Size # 7 # 789 Screenings Hydrated Lime 19.0 3/4" 0.0 3,451.5 3,978.0 4,455.0 12.5 1/2" 172.6 3,451.5 3,978.0 4,455.0 9.5 3/8" 135 ,242.5 3,493.6 3,978.0 4,455.0 4.75 # 4 3,071.8 3,872.7 3,992.3 4,455.0 2.36 # 8 3,348.0 3,951.7 4,130.6 4,455.0 1.18 # 16 3,382.5 3,962.2 4,249.9 4,455.0 0.600 # 30 3,382.5 3,962.2 4,321.4 4,455.0 0.300 # 50 3,382.5 3,962.2 4,369.1 4,455.0 0.150 # 100 3,382.5 3,962.2 4,402.5 4,455.0 0.075 # 200 3,413.5 3,964.8 4,416.8 4,455.0 0 Pan 3,451.5 3,978.0 4,455.0 4,500.0 Sum ,451.5 26.5 477.0 45.0 148

PAGE 149

APPENDIX B PERPAV TEST SU E IDT RESULTS 149

PAGE 150

Table B-1. Superpave IDT test results for dense graded asphalt mixtures Types/ m-D Aging Con. p. (1/p(Mpa) (kJ/DCSEf Creep D(t) (1/GFailure Strain () C 3.53.51 0.80 0.52 6.33E-10 0.225 1 St M R FE Tem value si) (Gpa) m 3 ) (kJ/m 3 ) Rate (1/psi-sec) Pa) -10 0.514 4E-08 22.21 398.41 0 d) C 2.25E-07 3.05 1.80 1.50 4.59E-09 1.330 C 2.27E-07 2.79 2.70 2.41 1.13E-08 2.673 Dense(UnmoST 0C 4.77E-07 2.14 4.20 3.99 3.20E-08 7.055 C 7.19E-08 3.95 1.00 0.65 4.18E-10 0.204 404.23 C 9 1.66E-07 3.56 1.70 1.36 2.18E-09 0.735 751.07 C 2.36E-07 3.02 2.40 2.11 3.41E-09 1.082 Dense(UnmoLT C 4.48E-07 2.25 2.20 1.99 9.43E-09 2.619 1336.78 C 1.57E-07 3.72 1.10 0.78 5.24E-10 0.282 515.24 C 3.57E-07 3.20 2.30 2.01 2.81E-09 1.013 1038.17 C 5.89E-07 2.81 4.50 4.17 7.31E-09 2.304 2180.68 Dens(ModSTC 7.54E-07 2.23 5.50 5.26 1.61E-08 4.414 3326.20 5.89E-08 3.84 0.80 0.49 1.49E-10 0.117 348.31 0C 0.360 2.27E-07 3.55 2.40 2.04 9.79E-10 0.443 5C 0.380 3.66E-07 2.99 15.97 3.00 2.72 1.93E-09 0.788 1434.74 Dense (ModLT 10C 0.413 5.43E-07 2.59 11.37 3.50 3.21 3.88E-09 1.414 1824.64 0.529 15.28 870.23 5 0.632 13.55 1289.29 1 0.6681 10.85 2566.05 -10 0.39 22.26 0d) 0.47 18.43 5 0.490 15.66 1120.48 10-10 0.5320.334 11.9921.40 0e 0.423 17.40 5) 0.473 11.98 10 0.534 10.55 -10C 0.306 23.88 17.40 974.00 ) xtures Table B-2. Superpave IDT test results for open graded asphalt mi Types/ Aging Con. m-vD1 St (Mpa) MR (Gpa) Creep Rate D(t) Failure Strain () -10C 0.229 3.84E-07 1.97 11.47 -10 315.03 Temp. alue (1/psi) FE (kJ/m 3 ) DCSE f (kJ/m 3 ) (1/psi-sec) 4.28E (1/GPa) 0.40 0.23 0.317 0C 0.400E-07 1.67 9.84 0.36E-09 448.59 0.497E-07 8.16 0.78E-09 881.90 FC-5 (FLime) ST 0.533E-07 9.10 1.05E-08 1058.80 0.189 E-07 14.71 0.15E-10 285.21 0.324 E-07 11.35 0.35E-09 446.92 0.421E-07 10.35 0.72E-09 773.89 FC-5 (FLime) LT 0.427E-07 10.16 0.93E-09 1013.60 0.292E-07 11.06 0.13E-10 259.18 0.43-07 9.39 0.35E-09 447.62 0.50-07 8.84 0.74E-08 838.63 FC-5 (NGran) ST C 0.557E-07 6.93 1.29E-08 1499.10 11.69 0.15E-10 251.72 0C 0.355 5.25E-07 9.25 0.2017E-09 400.95 5C 0.442 7.13E-07 1.38 8.83 0.90 0.79 6.69E-09 2.275 805.92 FC-5 (NGran) LT 10C 0.555 7.47E-07 1.27 7.29 1.10 0.99 1.91E-08 5.118 1215.67 4.80 0.50 6 3.0 1.185 5C 6.20 1.36 0.90 9 9.5 2.879 10C 5.87 1.45 1.20 8 1.2 3.500 -10C 2.65 2.07 0.30 5 1.8 0.180 0C 4.46 1.93 0.50 4 1.3 0.670 5C 4.69 1.72 0.90 6 3.6 1.324 10C 6.26 1.57 1.10 8 5.1 1.824 -10C 3.46 1.44 0.20 1 7.6 0.433 0C 1 3.72E 1.39 0.40 0 3.1 1.135 5C 8 6.69E 1.36 0.80 0 1.1 3.387 10-10C 0.223 8.384.65E-07 1.17 1.30 0 2.118 5.828 1.44 1.34 0.200.30 4. 2. 0.351 0.943 150

PAGE 151

APPENDIX C DBDT TEST RESULTS 151

PAGE 152

Table C-1. DBDT test results for dense graded asphalt mixtures Types/ Aging Con. Temp. m-value D 1 (1/psi) St (Mpa) M R (Gpa) FE (kJ/m 3 ) DCSE f (kJ/m 3 ) Creep Rate (1/psi-sec) D(t) (1/GPa) Failure Strain () -10C 0.435 7.73E-08 4.81 21.37 0.75 0.41 6.82E-10 0.275 302.11 0C 0.545 2.66E-07 4.98 15.12 1.67 1.18 6.23E-09 1.708 612.43 5C 0.568 4.39E-07 4.67 13.17 3.24 2.53 1.26E-08 3.261 1051.09 Dense (Unmod) ST 10C 0.659 7.14E-07 4.56 10.25 4.38 3.85 4.45E-08 9.850 1449.85 -10C 0.381 1.11E-07 5.09 24.80 0.86 0.47 5.88E-10 0.272 303.17 0C 0.497 1.88E-07 4.97 18.32 1.52 1.07 2.89E-09 0.892 527.24 5C 0.524 2.91E-07 4.54 15.52 2.03 1.56 5.69E-09 1.623 756.12 Dense (Unmod) LT 10C 0.543 5.04E-07 4.46 11.71 2.30 1.82 1.17E-08 3.162 872.13 -10C 0.326 2.16E-07 6.04 20.88 1.39 0.89 6.70E-10 0.346 403.75 0C 0.481 2.36E-07 5.97 16.79 2.76 2.12 3.14E-09 0.996 784.61 5C 0.562 3.29E-07 5.86 12.83 4.92 4.25 9.00E-09 2.371 1344.09 Dense (Mod) ST 10C 0.599 4.96E-07 5.48 10.69 7.14 6.57 1.86E-08 4.549 1961.02 -10C 0.322 1.29E-07 5.54 20.61 1.23 0.68 3.83E-10 0.221 386.82 0C 0.449 1.87E-07 5.71 17.29 2.28 1.67 1.87E-09 0.651 672.70 5C 0.458 2.78E-07 5.45 18.55 3.12 2.56 3.02E-09 1.004 922.83 Dense (Mod) LT 10C 0.466 3.67E-07 5.28 12.15 3.38 2.87 4.26E-09 1.376 1035.28 Table C-2. DBDT test results for open graded asphalt mixtures Types/ Aging Con. Temp. m-value D 1 (1/psi) St (Mpa) M R (Gpa) FE (kJ/m 3 ) DCSE f (kJ/m 3 ) Creep Rate (1/psi-sec) D(t) (1/GPa) Failure Strain () -10C 0.230 3.26E-07 3.22 12.78 0.48 0.20 3.69E-10 0.280 266.80 0C 0.408 4.21E-07 3.14 9.97 0.82 0.46 2.87E-09 1.069 444.88 5C 0.582 4.31E-07 3.04 9.56 1.11 0.84 1.40E-08 3.539 586.09 FC-5 (FLime) ST 10C 0.616 6.28E-07 2.97 7.29 1.39 1.02 2.72E-08 6.455 764.92 -10C 0.268 2.09E-07 2.59 13.97 0.39 0.19 3.55E-10 0.241 271.73 0C 0.366 4.00E-07 2.77 11.20 0.50 0.23 1.83E-09 0.775 316.37 5C 0.456 4.58E-07 2.77 9.37 0.74 0.51 4.86E-09 1.595 440.72 FC-5 (FLime) LT 10C 0.471 6.40E-07 2.62 7.78 1.14 0.91 7.77E-09 2.443 598.35 -10C 0.330 2.92E-07 2.41 12.17 0.38 0.19 9.38E-10 0.461 277.24 0C 0.513 2.74E-07 2.3 8.85 0.39 0.22 4.89E-09 1.429 301.30 5C 0.530 6.28E-07 2.22 7.71 0.72 0.53 1.30E-08 3.598 537.53 FC-5 (NGran) ST 10C 0.614 8.80E-07 2.22 6.03 1.23 0.96 3.77E-08 8.941 905.09 -10C 0.245 4.45E-07 2.21 10.83 0.28 0.09 5.94E-10 0.400 228.00 0C 0.431 3.94E-07 2.34 9.64 0.41 0.21 3.33E-09 1.169 315.00 5C 0.475 7.80E-07 2.65 7.61 0.83 0.55 9.88E-09 3.064 531.66 FC-5 (NGran) LT 10C 0.609 7.51E-07 2.60 7.23 1.20 0.97 3.08E-08 7.384 742.24 152

PAGE 153

APPENDIX D CREEP PARAMETERS FROM SUPERPAVE IDT AND DBDT 153

PAGE 154

Figure D-1. m-value from Superpave IDT test Fie D-1 IDst gur 2. D from S uperpave T te 154

PAGE 155

Figure D-3. m-value from DBDT test Figure D-4. D1 from DBDT test 155

PAGE 156

Figure D-5. m-value from Superpave IDT and D BDT tests BDT tests Figure D-6. D 1 from Superpave IDT and D 156

PAGE 157

LIST OF REFERENCES AAPA (2004). National Asphalt Specifications, Second Edition, Kew Victoria, Australia. AASHTO (2001). Standard Method of Test for Bulk Specific Gravity of Compacted Bituminous Mixtures Using Saturated Surface-Dry Specimens, AASHTO T 166, Washington, D. C. AASHTO (2001). Standard Method of Test for Percent Air Voids in Compacted Dense and Open Bituminous Paving Mixtures, AASHTO T 269, Washington, D. C. AASHTO (2001). Standard Practice for Mixture Conditioning of Hot Mix Asphalt, AASHTO PP2, Washington, D. C. ADINA (2005). ADINA Users Manual-Version 8.3, ADINA R&D, INC., Watertown, MA. Alvarez, A. E., Martin, A. E., Estakhri, C. K., Button, J. W., Glover, C. J., and Jung, S. H. (2006). Synthesis of Current Practrice on the Design, Construction, and Maintenance of Porous Friction Courses, FHWA/TX-06/0-5262-1, Texas Department of Transportation, Texas Transportation Institute/Texas A&M University, College Station, TX. Alvarez, A. E., Martin, A. E., Estakhri, C. K., Button, J. W., Kraus, Z., Prapaitrakul, N., and Glover, C. J. (2008). Evaluation and Recommeded Improvements for Mix Design of Permeable Friction Courses, FHWA/TX-08/0-5262-3, Texas Department of Transportation, Texas Transportation Institute/Texas A&M University, College Station, TX. Asahi, M. and Kawamura, K. (2000). Activities of Porous Asphalt on Expressway, Proceeding of Road Engineering Association of Asia and Australasia, Tokyo, Japan. ASTM (2004). Standard Practice for Open-Graded Friction Course (OGFC) Mix Design, ASTM D 7064, West Conshohocken, PA. ASTM (2002). Standard Test Method for Bulk Specific Gravity and Density of non-absorptive compacted bituminous Mixtures, ASTM D 2726, West Conshohocken, PA. ASTM (2002). Standard Test Method for Bulk Specific Gravity and Density of Compacted Bituminous Mixtures Using Automatic Vacuum Sealing Method, ASTM D 6752, West Conshohocken, PA. Baladi, G. Y., Schorsch, M., and Svasdisant, T. (2002). Determination of Top-Down Cracks in Bituminous Pavements, MDOT-PRCE-MSU-2003-110, Michigan Department of Transportation, MI. Bartley Consultants Ltd. (1999). Survey of New Zealand Roading Authorities Regarding Pavement Engineering Issues, Transfund Research Report No. 143, Australia. 157

PAGE 158

Bell, C. A., AbWahab, Y., and Cristi, M. E. (1991). Investigation of Laboratory Aging Procedures for Asphalt-Aggregate Mixtures, In Transportation Research Record: Journal of the Transportation Research Board, No. 1323, National Research Council, National Academy Press, Washington, D. C., pp. 32-46. Bensalem, A., Brown, A. J., Nunn, M. E., Merrill, D. B., and Lloyd, W. G. (2000). Finite Element Modeling of Fully Flexible Pavement: Surface Cracking and Wheel Interaction, 2nd International Symposium on 3D Finite Element for Pavement Analysis, Design, and Research, Charleston, WV, pp. 103-113. Birgisson, B., Roque, R., Varadhan, A., Thai, T., and Jaiswal, L. (2006). Evaluation of Thick Open Graded and Bonded Friction Courses for Florida, Final Report of Florida Department of Transportation, University of Florida, Gainesville, FL. Bolzan, P. E. and Huber, G. (1993). Direct Tension Test Experiments, SHRP-A-641, Strategic Highway Research Program, Transportation Research Board, National Research Council, Washington, D.C. Bolzan, P. E., Nicholls, J. C., and Huber, G. A. (2001). Searching for Superior Performing Porous Asphalt Wearing Courses, Transportation Research Board, National Research Council, Washington, D. C. 7). Experience with porous asphalt in the United ropean Conference on Porous Asphalt, Madrid, Spain. Buchanan, M. S. (1999). An Evaluation of Selected Methods to Determine the Bulk Specific Gravity of Compacted Hot Mix Asphalt (HMA), Journal of the Association of Asphalt Paving Technologists, Vol. 68, pp. 608-634. Buttlar, W. G., Al-Khateeb, G. G., and Bozkurt, D. (1999). Development of a Hollow Cylinder Tensile Tester to Obtain Mechanical Properties of Bituminous Paving Mixtures, Journal of the Association of Asphalt Paving Technologists, Vol. 68, pp. 369-403. 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, In Transportation Research Record: Journal of the Transportation Research Board, No. 1454, National Research Council, National Academy Press, Washington, D. C., pp. 163-171. Buttlar, W. G., Wagoner, M. P., You, Z., and Brovold, S. T. (2004). Simplifying the Hollow Cylinder Tensile Test Procedure through Volume-Based Strain, Journal of the Association of Asphalt Paving Technologists, Vol. 73, pp. 367-400. Chang, C., Baladi, G. Y. and Wolff, T. F. (2002). Detecting Segregation in Bituminous Pavements, In Transportation Research Record: Journal of the Transportation Research Research Council, National Academy Press, Washington, D. Bowskill, G. J. and Colwill, D. M. (199Kingdom In Proceedings of the Eu Board, No. 1813, NationalC., pp. 77-86. 158

PAGE 159

Chehab, G. R., OQuinn, E., and Kim, Y. R. (2000). Specimen Geometry Study for Direct Tension Test Based on Mechanical Tests and Air Void Variation in Asphalt Concrete Specimens Compacted by Superpave Gyratory Compactor, In Transportation Research Record: Journal of the Transportation Research Board, No. 1723, National Research Council, National Academy Press, Washington, D. C., pp. 125-132. Cooley, L. A., Brwon, E. R., and Watson, D. E. (2000). Evaluation of OGFC Mixtures Containing Cellulose Fibers, NCAT Report 00-05, National Center for Asphalt Technology, Auburn, AL, 2000 Cooley, L. A., Prowell, B. D., Hainin, M. R., Buchanan, M. S., and Harrington, J. (2002). Bulk Specific Gravity Round Robin Using the Corelok Vacuum Sealing Device, NCAT Report 02-11, National Center for Asphalt Technology, Auburn, AL. Christensen, D. W. and Bonaquist, R. F. (2004). Evaluation of Indirect Tensile Test (IDT) Procedures for Low-Temperature Performance of Hot Mix Asphalt, National Cooperative Highway Research Program (NCHRP) Synthesis 530, Transportation Research Board, National Research Council, Washington, D. C. Crouch, L. K., Copeland, A. R., Walker, C. T., Maxwell, R. A., Duncan, G. M., Goodwin, W. A., Badoe, D. A., Leimer, H. W. (2002). Determining Air Void Content of Compacted Hot-Mix Asphalt Mixtures, In Transportation Research Record: Journal of the Transportation Research Board, No. 1813, National Research Council, National Academy Press, Washington, D. C., pp. 39-46. Choubane, B., Sholar, G. A., Musselman, J. A. (1998). and Page, G. C., Long Term Performance Evaluation of Asphalt-Rubber Surface Mixes, Research Report of Florida Department of Transportation 98-431, Gainesville, FL. Dauzats, M. and Rampal, A. (1987). Mechanism of Surface Cracking in Wearing Courses, 6th International Conference Structural Design of Asphalt Pavements, The University of Michigan, Ann Arbor, MI, pp. 232-247. De Beer, M., Fisher, C., and Jootse, F. J. (1997). Determination of Pneumatic Tyre/Pavement Interface Contact Stresses under Moving Loads and Some Effects on Pavement with Thin Asphalt Surfacing Layers, 8th International Conference on Asphalt Pavements, University of Washington, Seattle, WA, pp. 179-227. De Freitas, E. F., Pereira, P., Picado-Santos, L. and Papagiannakis, A. T. (2005). Effect of Construction Quality, Temperature and Rutting on Initiation of Top-Down Cracking, In Transportation Research Record: Journal of the Transportation Research Board, No. 1929, National Research Council, National Academy Press, Washington, D. C., pp. 174-182. De Freitas, E. F., Pereira, P., and Picado-Santos, L. (2003). Assessement of Top-Down Cracking Causes in Asphalt Pavements, 3th International Symposium on Maintenance and Rehabilitation of Pavements and Technical Control, Guimaraes, Portugal. 159

PAGE 160

FAA (2005). Standards for Specifying Construction of Airports, Advisory Circular No. 150/5370-10B, US Department of Transportation, Washington, D. C. FDOT (2007). Standard Specifications for Road and Bridge Construction, Florida Department of Transportation, Tallahassee, FL. Francken, L. (1979). Fatigue Performance of a Bituminous Road Mix under Realistic Test Conditions, In Transportation Research Record: Journal of the Transportation Research Board, No. 712, National Research Council, National Academy Press, Washington, D. C., pp. 30-36. Gerritsen, A. H., van Gurp, C. A. P. M., van der Heide, J. P. J., Molenaar, A. A. A., and Pronk, A.C. (1987). Prediction and Prevention of Surface Cracking in Asphaltic Pavements, 6th International Conference Structural Design of Asphalt Pavements, The University of Michigan, Ann Arbor, MI, pp. 378-391. Haas, R. C. G. (1973). A Method for Designing Asphalt Pavements to Minimize Low -Temperature Shrinkage Cracking, The Asphalt Institute Research Report 73-1 (RRR-73-1). Halstead, W. J. (1978). Open-Graded Friction Courses for Highways, National Cooperative Highway Research Program (NCHRP) Synthesis 49, Transportation Research Board, Harvey, J., Eriksen, K., Sousa, J., and Monismith, C. L. (1994). Effects of Laboratory Specimen Preparation on Aggregate-Asphalt Structure, Air-Void Content Measurement, and Repetitive Simple Shear Test Results, In Transportation Research Record: Journal of the Transportation Research Board, No. 1454, National Research Council, National Academy Press, Washington, D. C., pp. 113-122. Himeno, K., Ikeda, T., Kamijima, T. (1997). and Abe, T., Distribution of Tire Contact Pressure of Vehicles and Its Influence on Pavement Distress, 8th International Conference on Asphalt Pavements, University of Washington, Seattle, WA, pp. 129-139. Huber, G. (2000), Performance Survey on Open-Graded Friction Course Mixes, Synthesis of Highway Practice 284, Transportation Research Board, National Research Council, Washington, D. C. Huang, B. and Shu, X. (2005). Laboratory Evaluation of Semi-Circular Bending Tensile Strength Test for HMA Mixtures, Transportation Research Board, National Research Council, Washington, D. C. InstroTek (2003). Corelok Operators Guide, InstroTek, Raleigh, NC. Jacobs, M. M. (1995). Crack Growth in Asphaltic Mixes, Ph.D. Dissertation, Delft University of Technology, The Netherlands. National Research Council, Washington, D. C. 160

PAGE 161

Kandhal, P. S., and Mallick, R. B. (lt Friction Course: State of the Practice, NCAT Report 98-7, National Center for Asphalt Technology, Auburn, AL. Kandhal, P. S. (2002). Design, Construction, and Maintenance of Open-Graded Asphalt Friction nduced Cracking of Asphalt Pavement, Ph.D. Dissertation, Kim, Y. R., Daniel, J. S., and Wen, H. (2002). Fatigue Performance Evaluation of WesTrack al Kowaliction Properties of PFC, SMA, and DGA Pavements, Transportation Research BoardKraemeropean Li, X. ahe Low Temperature Fracture Resistance of Mallick, R. B., Kandhal, P. S., Cooley, L. earch gton, D. C., pp. 9-19. 1998). Open Graded Aspha Courses, NAPA Information Series 115, National Asphalt Pavement Association, Lanham, MD. Khalid, H. and Walsh, C. (1996). A Rational Mix Design Method for Porous Asphalt, 24th European Transport Forum, London, UK. Kim, J. (2005). Accurate Determination of Dissipated Creep Strain Energy and Its Effect on Loadand Temperature-I University of Florida, Gainesville, FL. Asphalt Mixtures Using Viscoelastic Continuum Damage Approach, Final Report of North Carolina Department of Transportation, North Carolina State University, Raleigh, NC. Komoriya, K., Yoshida T. and Nitta, H. (2001). WA-DA-CHI-WA-RE Surface LongitudinCracks on Asphalt Concrete Pavement, Transportation Research Board, National Research Council, Washington, D. C. ski, K. J., McDaniel, R. S., Shah, A., and Olek, J. (2009). Long Term Monitoring of the Noise and Fr National Research Council, Washington, D. C. r, C. (1997). Porous asphalt: Past and Present, In Proceedings of the Eu Conference on Porous Asphalt, Madrid, Spain. nd Marasteanu, M. (2004). Evaluation of t Asphalt Mixtures Using the Semi Circular Bend Test, Journal of the Association of Asphalt Paving Technologists, Vol. 73, pp. 401-426. A. Jr., and Watson, D. E. (2004). Design, Construction, and Performance of New Generation Open Graded Friction Courses, Journal of the Association of Asphalt Paving Technologists, Vol. 69, pp. 391-422. Mallick, R., Kandhal, P., Cooley, A., and Watson, D. (2000). Design, Construction and Performance of New Generation Open-Graded Friction Courses, National Center for Asphalt Technology Report 00-01, NCAT, Auburn, AL. Martin, K. L., Davison, C. J., and Bullin, J. A., Asphalt Aging in Texas Roads and Test Sections, In Transportation Research Record: Journal of the Transportation ResBoard, No. 1269, National Research Council, National Academy Press, Washin 161

PAGE 162

Matsuno, S. and Nishizawa, T. (1992). Mechanism of Longitudinal Surface Cracking inPavement, 7th International Conference on Asphalt Pavements, Vol. 2, University Nottingham, UK, pp. 277-291. Asphalt of McDaniel, R. S. and Thornton W. D. (2005). Field Evaluation of a Porous Friction Course for McDanz, J. G. (2004). Field Evaluation of Porous Asphalt Pavement, SQDH 2004-3, Final Report of The Institute for Safe, Quiet, and MolenaJournal of the Association of Asphalt Paving Technologists, Vol. 53, pp. 440-474. Myers,ible Highway Pavements, Ph.D. Dissertation, University of Florida, Gainesville, FL. Myers,nt Management, Journal of the Association of Asphalt Paving Technologists, Vol. 71, pp. 651-670. Myers,iated dinal Wheel Path Cracks, In Transportation Research Record: Journal of the Transportation Research Board, No. 1778, National Research Council, National Myers,echanisms of Surface-Initiated Longitudinal Wheel Path Cracks in High-Type Bituminous Pavements, Journal of the Association of Myers,ontact Stresses for Different Truck Types to Evaluate Their Influence on Near-Surface Cracking Nunn, oads for Heavy Traffic, Industry Conference, Australian Asphalt Pavement Association, Surfers Paradise, Queensland, Australia. Perez-Je Use of Transportation Research Board, No. 1265, National Research Council, National Academy Press, Noise Control, Transportation Research Board, National Research Council, Washington, D. C. iel, R. S., Thornton, W. D., and Domingue Durable Highways at the Purdue University, West Lafayette, IN. ar, A. A. A. (1984). Fatigue and Reflective Cracking due to Traffic, L. A. (2000). Development and Propagation of Surface-Initiated Longitudinal Wheel Path Cracks in Flex L. A. and Roque, R. (2002). Top-Down Crack Propagation in Bituminous Pavements and Implications for Paveme L. A., Roque, R. and Birgisson B. (2001). Propagation Mechanisms for Surface-InitLongitu Academy Press, Washington, D. C., pp. 113-122. L. A., Roque, R., and Ruth, B. E. (1998). M Asphalt Paving Technologists, Vol. 67, pp. 401-432. L. A., Roque, R., Ruth, B. E. and Drakos, C. A. (1999). Measurement of C and Rutting, In Transportation Research Record: Journal of the Transportation Research Board, No. 1655, National Research Council, National Academy Press, Washington, D. C., pp. 175-184. M. (1998). Design of Long-Life R imenez, F. E. and Gordillo, J. (1990). Optimization of Porous Mixes through thSpecial Binders, In Transportation Research Record: Journal of the Washington, D. C., pp. 59-68. 162

PAGE 163

Poulikakos, L. D., Pittet, M., M., Arnaud, Junod, A., Gubler, R., Simond, E., Partl, L., and Dumont, A. (2006). Mechanical Properties of Porous Asphalt, Recommendations for Standardization, Swiss Federal Laboratory for Materials Testing and Research (EMPA) and Ecole Polytechnique Federale de Lausanne (EPFL)-Laboratoire des voies de circulation (LAVOC). Raju, S. Kumar, S. S., Reddy, K. S., Bose, S., and Pandey, B. B. (2008). Analysis of TCracking Behavior of Asphalt Pavements, Transportation Research Board, National Research Council, Washington, D op-Down C. Romeo, E.(2008). Measurement and Prediction of Fundamental Tensile Failure Limits of Hot Roque,wn Cracking Performance of Hot Mix Asphalt, Journal of the Association of Asphalt Paving Technologists, Vol. 73, pp. 229-Roque,f Modifiers iversity ing Technologists, Vol. 61, pp. 304-332. ment of Transportation, University of Florida, Gainesville, FL. Roque, pp. alt Ruiz, A., Alberola, R., PTransportation Research Record: Journal of the Transportation Research Board, No. p. 87Romanoschi, S. A. (1999). Characterization of Pavement Layer Interfaces, Ph.D. Dissertation, Louisiana State University, Baton Rouge, LA. Mix Asphalt (HMA), Ph.D. Dissertation, University of Florida, Gainesville, FL. R., Birgisson, B., Drakos, C. A., and Dietrich, B. (2004a). Development and Field Evaluation of Energy-Based Criteria for Top-Do 260. R., Birgisson, B., Tia, M., Kim, B., and Cui, Z. (2004b). Guidelines for Use oin Superpave Mixtures, Final Report of Florida Department of Transportation, Unof Florida, Gainesville, FL. Roque, R. and Buttlar, W. G. (1992). The Development of a Measurement and Analysis System to Accurately Determine Asphalt Concrete Properties Using the Indirect Tensile Mode, Journal of the Association of Asphalt Pav Roque, R., Buttlar, W. G., Ruth, B. E., Tia, M., Dickson, S. W., and Reid, B. (1997). Evaluation of SHRP Indirect Tension Tester to Mitigate Cracking in Asphalt Pavements and Overlays, Final Report of Florida Depart R., Hardee, H., and Ruth, B. E. (1988). Thermal Rippling of Asphalt Concrete Pavements, Journal of the Association of Asphalt Paving Technologists, Vol. 57,464-483. Roque, R. and Ruth, B. E. (1990). Mechanisms and Modeling of Surface Cracking in AsphPavements, Journal of the Association of Asphalt Paving Technologists, Vol. 59, pp. 396-421. erez, F., and Sanchez, B. (1990). Porous Asphalt Mixtures in Spain, In 1265, National Research Council, National Academy Press, Washington, D. C., p94. 163

PAGE 164

Sabita (1995). The Design and Use of Porous Asphalt Mixes, Manual 17, Roggebaai, SoAfrica. uth Transportation Research Board, National Research Council, Washington D.C. Sedwicce-da, Suresha, S. N., Varghese, G., and Shankar, A. U. R. (2009). A Comparative Study on Properties Svasdisant, T., Schorsch, M., Baladi G. Y., and Pinyosunun, S. (2002). Mechanistic Analysis of nal Tappei, MD. TNZ (2007). SP/SP11 070704. Uhlmeyg No. 1730, National Research 16. Schorsch, M., Chang, C. and Baladi G. Y. (2003). Effects of Segregation on the Initiation and Propagation of Top-Down Cracks k, S. C. (1998). Effect of Asphalt Mixture Properties and Characteristics on SurfaInitiated Longitudinal Wheel Path Cracking, Masters Thesis, University of FloriGainesville, FL. Shashidhar, N. (1999). X-Ray Tomography of Asphalt Concrete, In Transportation Research Record: Journal of the Transportation Research Board, No. 1681, National Research Council, National Academy Press, Washington, D. C., pp. 186-192. of Porous Friction Course Mixes with Neat Bitumen and Modified Binders, International Journal of Construction and Building Materials, Vol. 23, pp. 1211-1217. Top-Down Cracking in Asphalt Pavement, In Transportation Research Record: Jourof the Transportation Research Board, No. 1809, National Research Council, NationalAcademy Press, Washington, D. C., pp. 126-136 Swart, J. H. (1997). Experiences with Porous Asphalt in the Netherlands, In Proceedings of the European Conference on Porous Asphalt, Madrid, Spain. ner, W. (1993). Open-Graded Asphalt Friction Course, NAPA Information Series 115National Asphalt Pavement Association, Lanham, Thai, T. T. (2005). Approaches for using Open-Graded Friction Courses for Paving Floridas Highways, Master Thesis, University of Florida, Gainesville, FL. Specification for Open Graded Porous Asphalt Uchida K., Kurokawa, T., Himeno, K. and Nishizawa, T. (2002). Healing Characteristics of Asphalt Mixture under High Temperature Conditions, 9th International Conference on Asphalt Pavements, Copenhagen. er, J. S., Willoughby, K., Pierce, L. M. and Mahoney, J. P. (2000). Top-Down Crackinin Washington State Asphalt Concrete Wearing Courses, In Transportation Research Record: Journal of the Transportation Research Board Council, National Academy Press, Washington, D. C., pp. 110-1 164

PAGE 165

van der Zwan, J. Th., Goeman, Th., Gruis, H. J. A. J., Swart, J. H., and Oldenburger, R. H. (1990). Porous Asphalt Wearing Courses in the Netherlands: State of the Art Review In Transportation Research Record: Journal of the Transportation Research Board, No. van Heystraeten, G. and Moraux, C. (1990). Ten Years Experience of Porous Asphalt in h il, National Academy Press, Washington, D. VaradhGainesville, FL. Voskuilen, J. L. M., Tolman, F., and Rutten, E. (2004). Do Modified Porous Asphalt Mixture Wagoner, M. P., Buttlar, W. G., Paulino, G. H., and Blankenship, P. (2005). Investigation of the Academies, Wagonnkenship, P. (2006). Laboratory Testing Wambu Bituminous Materials Study, In No. 29-Wang, H. and Flintsch, G. W. (2007). Investigation of Short and Long-Term Variations of search e 1265, National Research Council, National Academy Press, Washington, D. C., pp. 95-110. Belgium, In Transportation Research Record: Journal of the Transportation ResearcBoard, No. 1265, National Research Counc C., pp. 34-40. an, A. (2004). Evaluation of Open Graded and Bonded Friction Course for Florida, Masters Thesis, University of Florida, Verhaeghe, B. M. J. A., Rust, F. C., Vos, R. M., and Visser, A. T. (1994). Properties of Polymerand Fibre-Modified Porous Asphalt Mixes, 6th Conference on Asphalt Pavements for Southern Africa, Cape Town, South Africa. Have a Longer Service Life? 3rd Eurasphalt and Eurobitume Congress, Vienna, Austria. Fracture Resistance of Hot-Mix Asphalt Concrete Using a Disk-Shaped Compact TensionTest, In Transportation Research Record: Journal of the Transportation Research Board, No. 1929, Transportation Research Board of the National Washington, D.C., pp. 183-192. er, M. P., Buttlar, W. G., Paulino, G. H., and Bla Suite for Characterization of Asphalt Concrete Mixtures Obtained from Field Cores, Journal of the Association of Asphalt Paving Technologists, Vol. 75, pp. 815-851. ra, J. H. Mania, J., and Smith, H. R. (1999). Kenya Transportation Research Record: Journal of the Transportation Research Board1681, National Research Council, National Academy Press, Washington, D.C., pp. 1137. Pavement Surface Characteristics at the Virginia Smart Road, Transportation ReBoard, National Research Council, Washington, D. C. Wen, H. and Kim, Y. R. (2002). A Simple Performance Test for Fatigue Cracking and Validation with WesTrack Mixtures, In Transportation Research Record: Journal of thTransportation Research Board, No. 1789, National Research Council, National Academy Press, Washington, D. C., pp. 66-72. 165

PAGE 166

Zhang, Z. (2000). Identification of Suitable Crack Growth Law for Asphalt Mixtures Using the Superpave Indirect Tensile Test (IDT), Ph.D. Dissertation, University of Florida, Gainesville, FL. Zhang, Z., Roque, R., Birgisson, B., and Sangpetngam, B. (2001). Identification and Verification of a Suitable Crack Growth Law, Journal of the Association of APaving Technologis sphalt ts, Vol. 70, pp. 206-241. 166

PAGE 167

BIOGRAPHICAL SKETCH hulseung Koh was born in Incheon, South Korea in 1976. After graduating from Deain chool in February 1995, CHigh S he was accepted to attend the Department of Civil Engineering at Incheonials, specifically, asphalt material. After master aterial thesis ethe fallHe wasthe pub have more professional knowledge, however relief after such tension made him stronger. He acquiregraduation education and did an outstanding job in the research project. He also worked as an instrucresponsibility of preparing lectures, homework assignments, and examinations as well as created class cuDuring his stay at the University of Florida, he also received the Master of Science in the Department of Civil and Coastal Engineering in December 2007. He received his Ph.D. with his the University of Incheon, Incheon, South Korea. When he was enrolling the University of he was interested in construction mater receiving a Bachelor of Engineering in February 2001, that led him master program at the University of Incheon to learn more about asphalt material. While he was spending two years in program, although work was exhausting, he became more familiar with asphalt m and deeply interested in asphalt pavement. In February 2003, Koh graduated with his masters ntitled Correlation Analysis of Mechanical Strengths for Hot Mix Asphalt. His academic curiosity led him to attend the Ph.D. program at the University of Florida in of 2005. Before he came to US, he got married to a wonderful woman, Kyoungnam Lee. He has worked as a graduate research assistant with his doctoral advisor, Dr. Reynaldo Roque. involved in many projects related to the experimental testing and analytical modeling of pavement materials and pavement. One of the experiences he had was presenting his research to lic in asphalt community. It was definitely not easy to present the research to people who d extensive experimental and analytical skills, gained broad experience during his tor for a class, Superpave Hot Mix Asphalt Technology. As an instructor, he had full rriculums combined with class lectures, laboratory tests, and field trips. 167

PAGE 168

dissertation entitled Tensile Properties of Open Graded Friction Course (OGFC) Mixture to Evaluate Top-Down Cracking Performance in Department of Civil and Coastal Engineering from thhaving e University of Florida in the summer of 2009. Challenges faced in graduate school helped him to gain more confidence and he ended up mind that he is ready to face any challenge ahead. 168