<%BANNER%>

Determination of Asphalt Mixture Healing Rate Using the Superpave Indirect Tensile Test


PAGE 1

DETERMINATION OF ASPHALT MIXTURE HEALING RATE USING THE SUPERPAVE INDIRECT TENSILE TEST By THOMAS PAUL GRANT A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR TH E DEGREE OF MASTER OF ENGINEERING UNIVERSITY OF FLORIDA 2001

PAGE 2

ii ACKNOWLEDGMENTS The two years of work shown in this thesis could not have been accomplished without the support of many individuals. I have been immersed in the complexities of asphalt research, and the many classes, discussions, laboratory testing and a nalysis have widened my scope of knowledge. I gratefully acknowledge my advisor and chairman of my supervisory committee, Professor Reynaldo Roque, for the time and effort he gave to make this project successful. Over many discussions, he guided my testi ng and molded my results into a thesis that lays the groundwork for much future research. Thanks go to Professor Mang Tia and Assistant Professor Bjorn Birgisson for being part of my supervisory committee. My fellow graduate students were of great assist ance. These include Shirley Zhang, Karina Honeycutt, Boonchai Sangpengam, Booil Kim, Christos Drakos, Bensa Nukumya, D.J. Swan, Mike Wagoner, Jeff Frank, Oscar Garcia and Paola Ariza. Our lab guru, George Lopp, was invaluable in keeping equipment operati onal and a big all around help. I thank my parents for their support over the years and helping with my schooling until grad school could help with the tab. And finally, a special thank you goes to my wonderful wife, Jamie, for supporting and standing be hind me in my pursuit of a graduate degree.

PAGE 3

iii TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. ii LIST OF TABLES ................................ ................................ ................................ ............. vi LIST OF FIGURES ................................ ................................ ................................ ......... viii ABSTRACT ................................ ................................ ................................ ........................ xi CHAPTERS 1 INTRODUCTION ................................ ................................ ................................ ........... 1 1.1 Background ................................ ................................ ................................ ............... 1 1.2 Objectives ................................ ................................ ................................ .................. 2 1.3 Scope ................................ ................................ ................................ ......................... 3 1.4 Research Approach ................................ ................................ ................................ ... 3 2 LITERATURE REVIEW ................................ ................................ ................................ 5 2.1 Introduction ................................ ................................ ................................ ............... 5 2.2 Traditional Fatigue Approach ................................ ................................ ................... 5 2.2.1 Strain Based ................................ ................................ ................................ ....... 5 2.2.2 Energy Based ................................ ................................ ................................ ..... 6 2.3 Fracture Mechanics Approach ................................ ................................ .................. 6 2.4 Crack Growth Model ................................ ................................ ................................ 8 2.4.1 Threshold Concept ................................ ................................ ............................. 8 2.4.2 Diss ipated Creep Strain Energy ................................ ................................ ......... 8 2.5 Healing ................................ ................................ ................................ .................... 10 3 MATERIALS AND METHODS ................................ ................................ ................... 12 3.1 Introduction ................................ ................................ ................................ ............. 12 3.2 Aggregates ................................ ................................ ................................ .............. 12 3.3 Material Handling ................................ ................................ ................................ ... 13 3.3.1 Gradations ................................ ................................ ................................ ........ 13 3.3.2 Preparation ................................ ................................ ................................ ....... 14 3.4 Mixture Production ................................ ................................ ................................ 15 3.4.1 Preparation ................................ ................................ ................................ ....... 15

PAGE 4

iv 3.4.2 Compaction ................................ ................................ ................................ ...... 15 3.5 Volumetric Properties ................................ ................................ ............................. 16 3.6 Testing Preparation ................................ ................................ ................................ 16 3.6.1 Slicing ................................ ................................ ................................ .............. 16 3.6.2 Preparation ................................ ................................ ................................ ....... 17 3.7 Mixture Testing ................................ ................................ ................................ ....... 19 3.7.1 Resilient Modulus, Creep Compliance and Tensile Strength Tests ................. 21 3.7.2 Fracture Tests ................................ ................................ ................................ ... 22 3.7.2.1 Two temperature tes ts ................................ ................................ ............... 23 3.7.2.2 Healing tests ................................ ................................ .............................. 23 3.8 Data Reduction ................................ ................................ ................................ ........ 23 4 FINDINGS AND ANALYSIS ................................ ................................ ....................... 25 4.1 Introduction ................................ ................................ ................................ ............. 25 4.2 Fracture Testing and Analysis ................................ ................................ ................. 25 4.2.1 Determination of Temperatures and Loadings ................................ ................. 25 4.2.2 Determination of Initial Deformation ................................ .............................. 26 4.2.3 Determination of Failure Limi t ................................ ................................ ........ 26 4.2.4 Explanation of Expectations ................................ ................................ ............ 27 4.3 Determination of Healing Rate ................................ ................................ ............... 28 4.3.1 Determination of Dissipated Creep Strain Energy ................................ ........... 34 4.3.2 DCSE/DCSE applied vs. d H / d 0 ................................ ................................ ............. 35 4.3.3 Steric Hardening Effects ................................ ................................ .................. 36 4.4 Damage Accumulation and Healing Across Temperatures ................................ .... 37 4.4.1 Stress Values ................................ ................................ ................................ .... 38 4.4.2 Damage Translation Process ................................ ................................ ............ 41 4.5 Permanent Deformation ................................ ................................ .......................... 43 4.6 Summary of Findings and Analyses ................................ ................................ ....... 44 5 SUMMARY, CONCLUSIONS AND RECOMMENDATIONS ................................ .. 46 5.1 Summary of Findings ................................ ................................ .............................. 46 5.2 Conclusions ................................ ................................ ................................ ............. 46 5.3 Recommendations ................................ ................................ ................................ ... 47 APPENDICES A DATA FROM TESTED SPECIMENS ................................ ................................ ........ 48 B SUMMARY OF MIXTURES TEST RESULTS ................................ ......................... 55 C FRACTURE TEST FIGURES ................................ ................................ ...................... 58 D HEALING TEST DATA ................................ ................................ .............................. 69

PAGE 5

v LIST OF REFERENCES ................................ ................................ ................................ ... 80 BIOGRAPHIC AL SKETCH ................................ ................................ ............................. 82

PAGE 6

vi LIST OF TABLES Table Page 3.1: Blend Proportions ................................ ................................ ................................ ........... 14 3.2: Job Mix Formulas ................................ ................................ ................................ ........... 14 3.3: Fracture Tests Loadings ................................ ................................ ................................ .. 23 4.1: DCSE t o Failure for Each Specimen ................................ ................................ .............. 35 4.2: Recovered d H / d 0 ................................ ................................ ................................ .............. 42 4.3: Comparison of Calculated and Tested d H / d 0 ................................ ................................ ... 42 A1: Healing Fracture Tests for C1 at 15 o C ................................ ................................ ............ 49 A2: Healing Fracture Tests for F1 at 15 o C ................................ ................................ ............ 49 A3: Healing Fracture Tests for C1 at 10 o C ................................ ................................ ............ 50 A4: Healing Fracture Tests for F1 at 10 o C ................................ ................................ ............ 50 A5: Fracture Test s to Failure for C1 at 15 o C ................................ ................................ ......... 51 A6: Fracture Tests to Failure for F1 at 15 o C ................................ ................................ .......... 51 A7: Fracture Tests to Failure for C1 at 10 o C ................................ ................................ ......... 52 A8: Fracture Tests to Failure for F1 at 10 o C ................................ ................................ .......... 52 A9: Two Temperature Fracture Tests to Failure for C1 ................................ ........................ 53 A10: Two Temperature Fracture Tests to Failure for F1 ................................ ...................... 53 A11: M R Creep and Strength Tests for C1 at 15 o C ................................ ............................... 54 A12: M R Creep and Strength Tests for F1 at 15 o C ................................ ............................... 54 B1: Resilient Modulus Test Data ................................ ................................ ........................... 56

PAGE 7

vii B2: Tensile Strength Test Data ................................ ................................ .............................. 56 B3: Cre ep Test Data ................................ ................................ ................................ ............... 57 B4: Dissipated Creep Strain Energy Calculation ................................ ................................ ... 57 D1: Coarse 1 at 10 o C Healing Specimens ................................ ................................ .............. 70 D2: Coarse 1 at 10 o C Healing Calculations ................................ ................................ ........... 71 D3: Fine 1 at 10 o C Healing Specimens ................................ ................................ .................. 72 D4: Fine 1 at 10 o C Healing Specimens continued ................................ ............................... 73 D5: Fine 1 at 10 o C Healing Calculations ................................ ................................ ............... 74 D6: Coarse 1 a t 15 o C Healing Specimens ................................ ................................ .............. 75 D7: Coarse 1 at 15 o C Healing Calculations ................................ ................................ ........... 76 D8: Fine 1 at 15 o C Healing Specimens ................................ ................................ .................. 77 D9: Fine 1 at 15 o C Healing Specimens continued ................................ ............................... 78 D10: Fine 1 at 15 o C Healing Calculations ................................ ................................ ............. 79

PAGE 8

viii LIST OF FIGURES Figure Page 2.1: Fatigue Crack Growth Behavior ................................ ................................ ..................... 7 2.2: Dissipated Creep Strain Energy ................................ ................................ ...................... 9 3.1: C1 and F1 Gradations (12.5 mm Nominal Size) ................................ ............................. 13 3.2: Pine Model Superpave Gyratory Compactor ................................ ................................ .. 16 3.3: Diamond Pacific Cutting Saw with Specimen Holder ................................ .................... 17 3.4: Aluminum Template for Drilling 8 mm Hole ................................ ................................ 18 3.5: Gage Point Placement Template with Vacuum Pump ................................ .................... 19 3.6: LVDT Setup ................................ ................................ ................................ .................... 20 3.7: Testing Setup ................................ ................................ ................................ .................. 20 4.1: Resilient Horizontal Deformation vs. Number o f Load Replications ............................. 27 4.2: Resilient Deformation vs. Time ................................ ................................ ...................... 29 4.3: d H / d 0 vs. Time ................................ ................................ ................................ ................. 30 4.4: Average d H / d 0 (Modified for Steric Hardening) vs. Time ................................ .............. 31 4.5: DCSE/DCSE applied vs. d H / d 0 ................................ ................................ ............................ 31 4.6: Healing vs. Time after 1000 Cycles of Loading ................................ ............................. 32 4.7: Comparison of Healing vs. Time after 1000 Cycles of Loading ................................ .... 33 4.8: Comparison of Time to Full Healing ................................ ................................ .............. 33 4.9: Comparisons of Healing Rates ................................ ................................ ........................ 34 4.10: Comparison of DCSE/DCSE applied vs. d H / d 0 ................................ ................................ 36

PAGE 9

ix 4.11: d H / d 0 vs. Cycles for Initia l Loading at 15 o C ................................ ................................ 39 4.12: d H / d 0 vs. Number of Load Replications for C1 ................................ ............................. 39 4.13: d H / d 0 vs. Number of Load Replications for F1 ................................ ............................. 40 4.14: DCSE applied(using s 0 )/DCSE failu re vs. Mixture Type and Temperature .............................. 40 4.15: d H / d 0 vs. DCSE/DCSE failure ................................ ................................ ........................... 41 4.16: Permanent Deformation vs. Time ................................ ................................ ................. 44 C1: Resilient Deformation vs. Load Replications (C1 at 15 o C to failure) ............................. 59 C2: Resilient Deformation vs. Load Replications (F1 at 15 o C to failure) ............................. 59 C3: Resilient Deformation vs. Load Replications (C1 at 10 o C to failure) ............................. 60 C4: Resilient Deformation vs. Load Replications (F1 at 10 o C to failure) ............................. 60 C5: Resilient Deformation vs. Time (C1 at 15 o C healing test) ................................ .............. 61 C6: Resilient Deformation vs. Time (F1 at 15 o C he aling test) ................................ .............. 61 C7: Resilient Deformation vs. Time (C1 at 10 o C healing test) ................................ .............. 62 C8: Resilient Deformation vs. Time (F1 at 10 o C healing test) ................................ .............. 62 C9: Average d H / d 0 vs. T ime (C1 at 15 o C healing test) ................................ .......................... 63 C10: Average d H / d 0 vs. Time (F1 at 15 o C healing test) ................................ ......................... 63 C11: Average d H / d 0 vs. Time (C1 at 10 o C healing test) ................................ ........................ 64 C12: Average d H / d 0 vs. Time (F1 at 10 o C healing test) ................................ ......................... 64 C13: Average d H / d 0 without Steric Hardening vs. Time (C1 at 15 o C healing) ...................... 65 C14: Average d H / d 0 without Steric Hardening vs. Time (F1 at 15 o C healing) ...................... 65 C15: Average d H / d 0 without Steric Hardening vs. Time (C1 at 10 o C healing) ...................... 66 C16: Average d H / d 0 without Steric Hardening vs. Time (F1 at 10 o C healing) ...................... 66 C17: Permanent D eformation vs. Time (C1 at 15 o C healing test) ................................ ......... 67 C18: Permanent Deformation vs. Time (F1 at 15 o C healing test) ................................ ......... 67

PAGE 10

x C19: Permanent Deformation vs. Time (C1 at 10 o C healing test) ................................ ......... 68 C20: Permanent Deformation vs. Time (F1 at 10 o C healing test) ................................ ......... 68

PAGE 11

xi Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Engineering DETERMINATION OF ASPHALT MIXTURE HEALING RATE USING THE SUPERPAVE INDIRECT TENSILE TEST By Thomas Paul Grant December 2001 Chairman: Dr. Reynaldo Roque Cochair: Dr. Bjorn Birgisson Major Department: Civil and Coastal Engineering Cracking is one of the major failure modes in asphalt concrete. Microcracks form within the as phalt concrete during repetitive loading. Eventually these microcracks will join together to form macrocracks, which result in permanent damage to the pavement. Healing occurs due to rest periods and temperature increases while in the microcracking range Healing increases the useful life of the pavement. Asphalt concrete specimens were prepared in the laboratory and tested using the Superpave Indirect Tensile Test. Two Superpave mixtures were prepared: a fine blend and a coarse blend. Tests were condu cted at 15 C and 10 C in an environmental chamber using hydraulic loading equipment. Repetitive loading tests applied 0.1 second haversine loads followed by 0.9 second rest periods. Damage accumulation was measured using the normalized resilient modulus.

PAGE 12

xii A method was developed to determine healing rate in terms of the recovered dissipated creep strain energy by converting changes in normalized resilient modulus to changes in energy. A crack growth law and model developed at the University of Florida us es dissipated creep strain energy as a measure of damage. With the work from this thesis, healing can now be incorporated into the model. Tests were also conducted at two temperatures using one specimen to analyze damage accumulation and translation acros s temperatures. Specimens were subjected to 1000 load repetitions at 15 C, cooled to 10 C for one hour and tested again. A procedure was devised to relate damage incurred at one temperature to equivalent damage at other temperatures. A number of findings and conclusions were drawn from this research. The relationship between the normalized resilient deformation and dissipated creep strain energy provided the ability to measure healing in terms of energy. Steric hardening appeared to play a significant r ole in a mixtures response during the loading and healing portions of laboratory tests. Healing occurred much faster at 15 C than 10 C and the coarse mixture healed much faster than the fine mixture at 15 C. Recovered strains do not appear to be an adeq uate measure of healing. Healing appears to play an important role in field performance.

PAGE 13

1 CHAPTER 1 INTRODUCTION 1.1 Background There are more than four million miles of roadways in the United States of America. Ninety six percent of the 2.42 million miles of paved public roadways are constructed of asphalt concrete. Currently, more than 130 million cars and 7.7 million heavy trucks compete for space on the roads. Functional roadways are vitally important to the economy to allow for speedy transport of people and freight. When an asphalt roadway is in disrepair, it results in a drain on resources to repair automobiles and the roadway. It also results in delays due to defensive driving and construction. In 1998, only 43.4% of urban interstate highways had a rating of GOOD or VERY GOOD (NTS, 2000). If better roadways are created that are more resista nt to failure, the resulting time and cost savings would be tremendous. The major failure modes of asphalt concrete (AC) are cracking, distortion, and disintegration. Cracking occurs for a number of reasons, but this thesis focuses on fatigue cracking, wh ich is cracking due to repetitive loading until the fatigue life of the pavement is reached. When loads are applied to a flexible pavement, tensile stresses arise at the bottom of the pavement. Numerous microcracks form within the asphalt concrete. If a n AC pavement is loaded until fatigue failure begins, the microcracks will join together to form macrocracks, which result in permanent damage to the pavement. According to Kim et al.,

PAGE 14

2 When an asphalt concrete pavement is subjected to repetitive applicat ions of multi level vehicular loads and various durations of rest periods, three major mechanisms take place: fatigue, which can be regarded as damage accumulation during loading; time dependent behavior related to the viscoelastic nature of asphalt concr ete; and chemical healing across microcrack and macrocrack faces during rest periods. (1994, pg. 89) The healing that occurs is due to rest periods and temperature increases. This thesis analyzes the healing that occurs after a period of continuous loading and explains the process for determination of a healing rate. Another aspect covered is damage accumulation across temperatures and a way to compare equivalent damage at different temperatures. A recently developed crack growth model at the University of Florida uses dissipated creep strain energy (DCSE) to measure damage accumulation in gyratory specimens. The healing rates developed in this thesis are in terms of DCSE and can improve the crack growth model with incorporation into the model. 1.2 Object ives The main objectives of this research are summarized below: Determine the necessary loadings to promote macrocracking in the specimens using the Superpave Indirect Tensile Test (IDT) in a certain time period at 10 o C and 15 o C. Develop a procedure to id entify and quantify microcrack healing. Evaluate and if possible quantify how microdamage is translated across temperatures. Develop a formula of healing with respect to time for each mixture at each temperature that can be compared and used. Identify a relationship to express time dependent microcrack healing in terms of recovered dissipated creep strain energy.

PAGE 15

3 1.3 Scope This study focuses on the microcrack healing of asphalt concrete during rest periods and the translation of accumulated damage acros s temperatures. Healing was monitored on two Florida Department of Transportation (FDOT) Superpave mixtures: a fine blend and a coarse blend. Tests were performed at 10 o C and 15 o C. Gyratory specimens were prepared with short term aging and compacted to 7% air voids. The asphalt content was controlled to achieve 4% air voids at N design. Specimens accumulated damage due to repeated loadings using the Superpave Indirect Tensile Test (IDT). Tests were performed to microdamage the specimens, but the loads were stopped before macrocracking began. The healing of the specimen was measured after the loading was stopped. Additional tests were performed to determine mixture properties, including resilient modulus, tensile strength, failure strain, fracture ene rgy m value, and D 1 value. 1.4 Research Approach The research was performed in five phases: 1. Literature Review: Examination of existing research on fracture mechanisms, healing, and measurement of crack growth in asphalt pavements. 2. Laboratory Mixtures: Preparation of six inch diameter specimens, including sieving, batching, mixing, compacting, aging, cutting to one inch and two inch thicknesses, determining air voids, dehumidifying, and placement of gage points. 3. Testing: Cooling of specimens to requisi te temperature and completion of resilient modulus, creep compliance, tensile strength, and repetitive loading tests. 4. Data Analysis: Evaluation of each test to produce figures showing healing effects, damage accumulation, and crack initiation, propagatio n, and disintegration.

PAGE 16

4 5. Development of a healing rate based on dissipated creep strain energy. Use healing rate for comparison and later incorporation into HMA crack growth model.

PAGE 17

5 CHAPTER 2 LITERATURE REVIEW 2.1 Introduction The purpose for this chapter is to examine published literature concerning healing, fatigue damage and crack growth in asphalt pavements. Fatigue is the wearing down of an asphalt pavement caused by repeated loads belo w the tensile strength of the pavement. Without healing, fatigue eventually results in failure, or permanent cracking, of the pavement. This section reviews the literature addressing the occurrence and quantification of healing, work performed to describ e the mechanisms of crack growth, and ways of measuring fatigue damage. 2.2 Traditional Fatigue Approach The Traditional Fatigue Approach assumes that damage occurs in a specimen from repetitive loading that leads to eventual fatigue failure of the specim en. The method can be based on stress, strain or energy. 2.2.1 Strain Based An oft used strain based equation developed by Monismith et al. (1985) N f = K (1/ e t ) a (1/S mix ) b provides the number of cycles to failure (N f ) where K is a mix related factor based on asphalt content and degree of compaction, e t is the tensile strain, a and b are coefficients based on beam fatigue tests and S mix is the stiffness of th e mixture.

PAGE 18

6 Many other equations have been developed using various parameters dependent on the asphalt mixture composition and temperature. None of the fatigue relationships that have been used are all encompassing. Many factors affect fatigue and the va rious formulae try to predict the number or load replications until cracking begins. Francken (1979) showed that fatigue relationships are dependent on rest periods and that longer rest periods result in longer fatigue life. From this, one can deduce tha t healing of the specimen occurs with longer rest periods. 2.2.2 Energy Based Fatigue has also been predicted using the cumulative dissipated energy from repetitive loading. A common energy based equation by van Dijk (1969) is W f = B f N f z where W f is th e cumulative dissipated strain energy per volume (J/m 3 ), B f and z are mixture coefficients, and N f is the number of cycles to fatigue failure. As loading occurs, energy within the viscoelastic material is dissipated until the failure limit is reached. Mi ners Rule simply states that damage is the ratio of number of cycles to number of cycles to failure. This does not account for healing effects from rest periods. The Traditional Fatigue Approach can add damage but not subtract it. The results are not f undamental because the approach is dependent on mode of loading. Can this approach explain the complexities of asphalt pavements? How does damage translate across temperatures? 2.3 Fracture Mechanics Approach The Fracture Mechanics Approach assumes ther e are inherent flaws in the material. When loading occurs, there are higher stress concentrations around the flaws

PAGE 19

7 because load is distributed over a smaller area, which means the material no longer has a uniform stress distribution. Fracture mechanics d escribes the propagation of cracks through materials. Paris and Erdogan (1963) developed a crack rate law for use in linear elastic homogeneous materials. Paris Law is defined as da/dN = A ( D K) n where a is the crack length, N is the number of load repetitions, K is the stress intensity factor, and A and n are material parameters. The stages of crack growth are shown in Figure 2.1. Paris law is only applicable in the propagation phase. log D D K log da/dN D K threshold D K critical 1 n Initiation Phase Propagation Phase Disintegration Phase Figure 2.1: Fatigue Crack Growth Behavior (After Jacobs, 1995) Asphalt concrete is a heterogeneous viscoelastic material, so Paris Law may not completely explain how asphalt mixtures behave. For example, cracks are known to grow di scontinuously in asphalt concrete mixtures (Zhang, 2000). Paris Law does not explain crack initiation because it assumes cracks are already present. The parameters A and n are a function of loading condition. When looking at da/dN from laboratory

PAGE 20

8 testi ng, results do not correspond to actual field performance of different mixtures (Zhang, 2000). Despite the Fracture Mechanics Approachs flaws, crack length is a measurable interpretation of damage, which provides a sound base from which to proceed. 2.4 Crack Growth Model The two approaches to measuring fatigue damage thus far presented are unable by themselves to accurately explain fatigue in asphalt concrete. Zhang et al. (2001) created a crack growth law based on viscoelastic fracture mechanics with t he addition of a threshold concept. Zhangs paper further explains the use of this crack growth law to create a pavement cracking prediction model. 2.4.1 Threshold Concept Paris Law assumes the crack advances when any stress is applied. Asphalt pavemen ts can perform well in the field for many years without developing visible macrocracks. Zhang et al. (2001) proposed that if a certain threshold was not exceeded for a given mixture during repeated loadings, that no macrocrack would form. The microcracks could heal themselves with the help of temperature increases and/or rest periods. From laboratory testing by Zhang and Honeycutt, it was shown that if a test specimen was tested and loading was stopped before the threshold was reached, full healing occur red with rest and temperature increases. It was also shown that once the threshold was exceeded (i.e., once macrocracks developed), no healing was achieved with rest and temperature increases (Zhang, 2000). 2.4.2 Dissipated Creep Strain Energy Zhang et al (2001) determined that dissipated creep strain energy (DCSE) to failure appears to be a fundamental parameter that serves as an appropriate threshold

PAGE 21

9 between microcracks and macrocracks. DCSE is defined as the fracture energy, FE, minus the elastic ener gy, EE (see Figure 2.2). From the strength test, failure strain ( e f ), tensile strength (St) and fracture energy can be determined. From the resilient modulus test, the resilient modulus (M R ) value can be found (see Figure 2.2). e 0 = (M R e f S t )/ M R EE = S t ( e f e 0 ) DCSE = FE EE Strain Stress e f M R M R S t Dissipated Creep Strain Energy (DCSE) Elastic Energy (EE) e 0 0 Fracture Energy, FE = DCSE + EE Figur e 2.2: Dissipated Creep Strain Energy (After Zhang et al., 2001) Zhang also showed that dissipated creep strain energy appears to be independent of mode of loading. This means the strength test results can be used to evaluate the repetitive loading test r esults. If the DCSE limit is not reached, macrodamage will not occur.

PAGE 22

10 In addition, Zhang developed a method to determine DCSE/cycle, so that the DCSE at any given time during loading could be calculated. The formula for a 0.1 second haversine load is DCSE/cycle = 1/20 s 2 D 1 m (100) m 1 where s is the stress and D 1 and m are determined from creep tests. This process uses an energy based approach to fracture mechanics with a threshold boundary. A pavement cracking prediction model based on Zhan gs crack growth law is currently being developed at the University of Florida. To make the model complete, healing needs to be incorporated in terms of recovered DCSE. 2.5 Healing The occurrence of microcrack healing in polymers is well documented. Asp halt binder contains hydrocarbon polymers, but only in recent years has any attempt been made to quantify healing in asphalt pavements. Little et al. (1997) found that long chain aliphatic molecules in binders are directly proportional to healing effects. The Western Research Institute has developed a microstructural asphalt cement model that exhibits the molecular structure as continuously reforming, thus incorporating healing effects. For many years, it has been known that identical mixtures perform bet ter in the field than in laboratory testing. The accelerated process of laboratory testing does not adequately mimic field performance. Because of this, a shift factor (SF) was developed in 1977 using AASHTO Road Test data. A shift factor developed by L ytton et al. (1993) is SF = 1 + n ri /N 0 a (t i /t 0 ) h

PAGE 23

11 where n ri is the number of rest periods, N 0 is the number of loading cycles to failure without rest periods in cyclic laboratory tests, t i is the length of the period between field load applications, t 0 is the length of the rest period between loading cycles in laboratory tests, and a and h are regression values. Kim et al. (1994) measured healing in a laboratory setting and in the field. For the laboratory testing, Kim et al. used Schaperys (1984) non linear viscoelastic correspondence principle (CP). For the correspondence principle case, a cyclic loading with rest periods was administered. A pseudo strain was used to differentiate between the relaxation effect and healing. The appropriate pseudo st rain formula as developed by Schapery (1984) is e R = 1/E R 0 t E (t t ) d e /d t d t where e is uniaxial strain, e R is pseudo strain, E R is reference modulus, and E(t) is the uniaxial relaxation modulus. Kim et al. transformed a time dependent issue into a time independent issue and calculated the re covered pseudostrain energy from rest periods. For the field testing, Kim et al. (1994) used stress wave testing. A section of roadway was closed to traffic and temperature sensors were placed in the pavement. From hourly stress wave tests for 24 hours, wavespeeds were obtained and the change in elastic modulus was calculated as a function of temperature. The elastic modulus of asphalt pavement increased at a given temperature over the 24 hour rest period. This increase is conjectured to be microcrack h ealing (Kim et al., 1994).

PAGE 24

12 CHAPTER 3 MATERIALS AND METHODS 3.1 Introduction This chapter explains the materials and processes used to create the gyratory AC specimens suitable for testing. The testing methods and data reduction are described. Relevant figures are included, with most of th e test results in the appendices. The same asphalt binder type AC 30 (Coastal) was used for each mixture. The appendices provide detailed information on material properties and fracture test results. Honeycutt (2000) produced identical mixtures to the o nes used in this research project and additional information on the aggregate properties, gradation and mixture design are provided in Honeycutts thesis. 3.2 Aggregates The aggregate used for this study was Miami oolite limestone produced by Whiterock. Limestone is a relatively soft, porous aggregate that is commonly used in asphalt paving in Florida because of its abundance. The aggregate components provided were a coarse aggregate, S1A; a fine aggregate, S1B; and screenings. Granite mineral filler f rom Georgia was also used in the mixtures.

PAGE 25

13 3.3 Material Handling 3.3.1 Gradations Two mixture designs provided by the FDOT Materials Office were used for this research. Both mix designs passed all Superpave criteria. The coarse blend, C1, and the fine b lend, F1, passed below and above the Superpave restricted zone, respectively and passed between the control points. Figure 3.1 shows the gradation chart of the mixtures and the restricted zone and control points. The blend proportions of the aggregates i n each mixture are shown in Table 3.1. The Job Mix Formulas are provided in Table 3.2. 0 20 40 60 80 100 Sieve Size^0.45 (mm) Percent Passing (%) Control Points Restricted Zone Coarse 1 Fine 1 19.5 12.5 9.5 4.75 2.36 Actual Sieve Size 1.18 0.6 0.3 0.15 0.075 Figure 3.1: C1 and F1 Gradations (12.5 mm Nominal Size)

PAGE 26

14 Table 3.1: Blend Proportions S1A S1B Screenings Filler (%) (%) (%) (%) C1 10. 20 63.27 25.51 1.02 F1 20.30 25.37 53.29 1.03 Bulk Specific Gravity 2.43 2.45 2.53 2.69 Table 3.2: Job Mix Formulas Sieve Size C1 F1 (mm) 25 (1") 100 100 19 (3/4") 100 100 12.5 (1/2") 97.4 95. 5 9.5 (3/8") 90 85.1 4.75 (#4) 60.2 69.3 2.36 (#8) 33.1 52.7 1.18 (#16) 20.3 34 0.6 (#30) 14.7 22.9 0.3 (#50) 10.8 15.3 0.15 (#100) 7.6 9.6 0.075 (#200) 4.8 4.8 3.3.2 Preparation The aggregates were dried and sieved to each individual sieve size This was done to assure that the samples represent the true gradation of the mixtures, which may not be representative in small quantities. The aggregates were then batched out in the appropriate quantities to produce 4500 g samples.

PAGE 27

15 3.4 Mixture Prod uction 3.4.1 Preparation The 4500 gram aggregate batches, asphalt binder and mixing equipment were heated for three hours at 150 o C (300 o F) to achieve appropriate uniform mixing temperature. The batches were then mixed with the proper amount of asphalt bi nder and heated for another two hours at 135 o C (275 o F) for short term aging. This aging represents the aging that occurs in the field between mixing and placement and allows for absorption of the asphalt binder into the aggregate pores. After one hour of aging, the mixture is stirred to prevent the outside of the mixture from aging more than the inner because of increased air exposure. During the aging, the compaction equipment was also heated to assure uniform temperature for compaction. 3.4.2 Compactio n The mixtures were compacted into 150 mm diameter specimens using the Pine model Superpave Gyratory Compactor (see Figure 3.2). The specimens were compacted to 7% air voids for testing. This is typical of the air void percentage in mixtures when they a re placed in the field. After compaction by vehicular traffic over time, the roadway reaches its optimum 4% air void level. Lower air void levels result in bleeding, while higher air void levels may result in raveling. The mixtures can be compacted by number of gyrations or by height. Compacting based on number of gyrations to control density results in large variances in air void percentages. Because of the different sizes of aggregate in each sieve size, each replicated mixture has a different total volume of aggregate, while the total mass is the same. Compacting based on height results in more uniform air void percentages for the test specimens, so this method was used for compaction.

PAGE 28

16 Figure 3.2: Pine Model Superpave Gyratory Compactor 3.5 Vol umetric Properties After compaction and cooling of the specimens, bulk specific gravities of the specimens were taken. Rice tests were performed in accordance with ASTM D 2041 to attain the theoretical maximum densities on 1000 gram uncompacted test mixt ures (Honeycutt, 2000). From these two tests, the air void percentage was determined. 3.6 Testing Preparation 3.6.1 Slicing The 150 mm diameter gyratory specimens are tested in one inch and two inch thicknesses using the Indirect Tensile Tests (IDT). A Diamond Pacific cutting saw (see Figure 3.3) with a special attachment to hold the gyratory specimens was used to achieve the desired thicknesses. Two two inch test specimens or three one inch test specimens

PAGE 29

17 were attained from each gyratory specimen. Be cause the saw uses water to keep the blade wet, the test specimens were dried for two days at room temperature to achieve the natural moisture content after cutting. Figure 3.3: Diamond Pacific Cutting Saw with Specimen Holder 3.6.2 Preparation The bu lk specific gravity and air void content were determined for each test specimen. Gyratory specimens had to be in the range of 7 0.5 % air voids to be considered for testing. Air void percentages for the test specimens used are provided in Appendix A. Test specimens were placed in a low humidity chamber for two days to negate moisture effects in testing. The two inch thick test specimens were used for the Resilient Modulus, Creep Compliance and Indirect Tensile Strength tests. The one inch thick test specimens were used for the fracture tests. The fracture test specimens required the drilling of an eight mm diameter hole in the center to create a focal point from which

PAGE 30

18 the crack would propagate. An aluminum template (see Figure 3.4) was used to align the specimen and hold it in place, while a drill press with a concrete drill bit was used to drill the hole. Figure 3.4: Aluminum Template for Drilling 8 mm Hole Brass gage points were attached to the test specimens with a strong adhesive using a stee l template and vacuum pump setup (see Figure 3.5). Four gage points were placed on each side of the test specimens 19 mm (0.75 in.) from the center along the vertical and horizontal axes. A steel plate that fit over the attached gage points was used to m ark the loading axis with a marker. This helped with placement of the test specimen in the testing chamber and assured proper loading of the specimen.

PAGE 31

19 Figure 3.5: Gage Point Placement Template with Vacuum Pump 3.7 Mixture Testing Roque et al. develope d the IDT testing procedure and data reduction process used for this research (1997, 1999). Thorough explanation of the procedures and the reasoning behind them are given in those papers. Tests were performed using an MTS hydraulic loading system with th e Teststar IIs data acquisition system. An environmental chamber kept the temperature constant 0.1 o C. Tests were performed at 10 o C and 15 o C. The deformation in the test specimens was measured using linear voltage differential transducers (LVDT) attache d to the gage points (see Figure 3.6). The LVDTs took voltage readings, which were converted to micro inches using a signal conditioning unit manufactured by MGC. The testing equipment is shown in Figure 3.7.

PAGE 32

20 Figure 3.6: LVDT Setup Figure 3.7: Test ing Setup

PAGE 33

21 After the LVDTs were in place, the test specimen was carefully placed in the testing chamber so that the loading occurred along the vertical plane of the test specimen. The test specimen was then cooled for eight hours to assure constant temper ature throughout the specimen. 3.7.1 Resilient Modulus, Creep Compliance and Tensile Strength Tests The resilient Modulus, creep compliance and tensile strength tests can all be performed on a single specimen. A 10 lb seating load is applied to each test specimen immediately before testing. The resilient modulus (M R ) test applies 0.1 second haversine loads followed by 0.9 second rest periods for five cycles. For this test, different loadings are tried, starting with a small loading and slowly increasing the load, to produce a maximum horizontal deformation between 200 300 micro inches. Five hundred and twelve data points are recorded during each second of testing. Creep tests applying a constant stress to the test specimen were performed for 100 secon ds. As with the M R test, small loads were used to check the amount of creep occurring. Horizontal deformations, d H were kept within the approximate range of 100 150 micro inches at 30 seconds to prevent excessive deformation of the specimen. A maximum horizontal deformation of 1000 min was used. Acquisition rates for the creep tests varied from 10 points per second at the beginning to one point every five seconds at the end of the test. The strength test loads the specimen to failure by applying a c onstant stroke of 50 mm per minute. An acquisition rate of 20 points per second was used. Through data reduction, these tests provide the resilient modulus, m value, D 0 and D 1 values, tensile

PAGE 34

22 strength, failure strain, fracture energy, and Poissons ratio for the tested specimen at a given temperature. The test results used for this research project at 10 o C were performed by Honeycutt (2000). This includes all results and mixture properties derived from M R creep, and strength tests for F1 and C1 mixtures at 10 o C. One difference of note is that Honeycutt ran 1000 second creep tests, although this should have no bearing on material properties, such as m value. 3.7.2 Fracture Tests The fracture tests are repetitive loading tests performed on the on e inch gyratory specimens. A 10 lb seating load is applied to the specimen before testing begins. Much like the M R tests, the fracture test applies 0.1 second haversine loads followed by a rest period of 0.9 seconds. This test continues until it is manu ally stopped. This allows the user to test a specimen for a certain amount of time. On screen, a constantly updated graph of the deformations shows the user the current state of the gyratory specimen and a rough estimate of when macrocrack propagation be gins. The load is similar to that used for the M R test, but was varied to achieve complete disintegration of the test specimen in about one hour. Standard loads were determined for both the C1 and F1 specimens at 10 o C and 15 o C (see Table 3.3). The user manually controlled the data collection. This allowed more data points to be recorded when the deformation slope began changing. Every time the data collection button was pushed during the test, the program took an M R reading over five cycles of loading at an arbitrary time. When the test was terminated, the testing and seating loads were removed from the test specimen.

PAGE 35

23 Table 3.3: Fracture Tests Loadings Temperature Loading Mixture ( o C) (lb) C1 10 1300 C1 15 700 F1 10 1300 F1 15 925 3.7.2.1 T wo temperature tests The two temperature fracture tests were performed after standard loadings were determined from standard fracture tests. The tests began at 15 o C for 1000 cycles of loading. The test was then stopped and the test specimen was cooled t o 10 o C. After an hour, the test was continued with the new appropriate loading. The test was continued until the specimen failed. 3.7.2.2 Healing tests The healing fracture tests loaded the specimens for 1000 cycles and then the test was terminated. Dur ing the following hour, five M R tests were performed on the specimen to monitor the healing. These tests were performed at 3 min., 6 min., 10 min., 30 min. and one hour after repetitive loading had stopped. The same load was used for the healing portion M R tests as with the fracture tests. 3.8 Data Reduction The tests output the data files in various formats. Some alterations have to be performed to reduce the data using a spreadsheet and program. A set of three test specimens is required to correctly calculate the appropriate values for the M R creep and strength tests. An Excel spreadsheet and Fortran code have been developed by the

PAGE 36

24 University of Floridas Civil Engineering Materials Department and thoroughly explained in a report to the Florida Depa rtment of Transportation (Roque et al., 1997). Several outputs are provided for the various tests. The M R test shows the horizontal and vertical instantaneous and total deformations of the middle three cycles. These cycles are used for the calculation o f the resilient modulus, which is also shown on the printout. The fracture test outputs the horizontal and vertical deformations for each data set collected during the test. With the M R creep, and strength tests, outputs include creep compliance over ti me, Poissons ratio, average tensile strength, m value, D 0 value, D 1 value, failure strain, and fracture energy. The results of these outputs are shown in Appendix B along with pertinent information about tested samples in Appendix A.

PAGE 37

25 CHAPTER 4 FINDINGS AND ANALYSIS 4.1 Introduction The repetitive loading fracture test mimics a wheel rolling over a section of asphalt pavement every second. This method of loading will eventually fail a pavement because the only time available for healing is bet ween each load. By loading the test specimen for a time and then removing the load, healing can be monitored and better understood. This chapter explains the testing results and analysis procedures used in the determination of a healing rate. The damage accumulation across temperatures is explained. Findings are explained in detail. 4.2 Fracture Testing and Analysis The coarse and fine mixtures that were chosen for this research are approved FDOT Superpave mixtures using native Florida aggregates. Th e two mixtures perform differently because of interlocking effects, but are produced using the same aggregates and binder. This prevents other factors from contributing to differences in performance and healing. Chapter 3 details the process by which gyr atory specimens were created and tested. 4.2.1 Determination of Temperatures and Loadings Historically at the University of Florida, repetitive loading tests were performed at 10 o C. After preliminary testing, 15 o C was chosen as the other temperature for this

PAGE 38

26 project. Loads were determined that would provide an acceptable linear portion of microcrack growth while still totally failing within an hour. Even with the narrow range of air voids and constant loads, each test specimen performs differently, resulting in a range of failure limits and damage rates. 4.2.2 Determination of Initial Deformation After analysis of a fracture test, the resulting resilient deformations can be plotted versus time. Figure 4.1 shows a representative test taken to fai lure. At the beginning of the test, the microdamage is not linear. In the past, this effect had been attributed to local elevation of temperature in the specimen during cycling that stabilizes after about two minutes of loading. However, based on data ob tained in this investigation, the author maintains that steric hardening may be the main force behind this effect. This will be explained in Section 4.3. The modified initial resilient deformation, d 0 is determined by interpolating the linear portion bac k to the beginning of cycling. 4.2.3 Determination of Failure Limit As explained in the literature review, Zhangs threshold value denotes the boundary between microcracking and macrocracking. Figure 4.1 shows that macrocracking on that test specimen occ urred at about 1300 cycles of loading. A conservative cutoff point was to halt loading at 1000 cycles to insure macrocracking had not developed. Thus, after 1000 cycles of loading, the specimens were still in the microdamage range and the damage was heal able. For both the damage accumulation across temperatures tests and the healing tests, 1000 cycles of loading was used as the standard. This cutoff point was used for both mixtures at both temperatures.

PAGE 39

27 0.000000 0.000200 0.000400 0.000600 0.000800 0.001000 0.001200 0 500 1000 1500 2000 2500 Number of Load Replications (cycles) Resilient Deformation (inches) d i d o Microdamage Macrodamage Figure 4.1: Resilient H orizontal Deformation vs. Number of Load Replications 4.2.4 Explanation of Expectations Initially, the testing procedure was developed to translate damage effects across temperatures. Tests were performed that induced a certain amount of damage at one tem perature (15 o C) before quickly cooling the specimen to a second temperature (10 o C). The intent was to lock the damage in, to evaluate the effect of damage accumulated at one temperature on the failure limit at a second temperature. An adequate way to measure damage was needed. A change in resilient horizontal deformation is a measure of damage. Earlier work (Honeycutt, 2000) has shown that by normalizing the change in horizontal deformation by d 0 it could be directly related to the theoretical crack length of the specimen. From the crack length, the da/dN, or crack growth rate, could be determined. Both the crack length and da/dN are ways of monitoring damage. Zhang (2000) noted, Crack grow th rates obtained from

PAGE 40

28 the laboratory tests do not correlate well with the field performance. One of the major reasons for this is the absence of healing due to the lack of rest periods during laboratory tests. After testing and analysis of damage accumu lation fracture tests, it was determined that the damage was not locked in and partial recovery of the resilient deformation of the specimen was seen. This recovery of resilient deformation is an increase in stiffness and represents healing. This healing needed to be quantified, but with the LVDTs temperature susceptibility, testing could not be performed when the temperature in the environmental chamber was not steady. As previously mentioned, full healing can occur with rest periods and temperature in creases. It was decided that healing tests should be done at both temperatures to look at the difference in the healing rate of the specimens at each temperature. 4.3 Determination of Healing Rate Development of a healing rate was an iterative process The goal was to find a way to measure healing in terms of DCSE. The processes explained in Section 4.2 had already been carried out before this analysis began. Explanation of the healing testing methods is provided in Chapter 3. One cycle of loading lasts one second and because the loading is stopped at 1000 cycles, the following figures are shown in terms of time. The example figures used in this section are the C1 tests at 15 o C. Appendix C shows figures generated from the healing tests. Refer to Appendix D for healing test data and further explanation of the healing rate calculations. The following step by step guide details the process used for each mixture at both temperatures, with explanation provided afterward: 1. Plot Horizontal Deformation vs Time for each specimen. See Figure 4.2.

PAGE 41

29 2. Determine d 0 for each specimen. See Section 4.2.2 for explanation. 3. Normalize the specimens by thickness and diameter. 4. Determine d H / d 0 at each time of acquisition. 5. Average the times and d H / d 0 of the specimens. 6. Determine DCSE at each time during loading. See Section 4.3.1. 7. Determine DCSE/DCSE applied at each time. DCSE applied is the value of DCSE at 1000 cycles, when the loading was stopped. 0.0003 0.00035 0.0004 0.00045 0.0005 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (seconds) Resilient Deformation (inches) Loading Healing Figure 4.2: Resilient Deformation vs. Time 8. Plot average d H / d 0 vs. time. Place a linear trendline th rough the data points collected during the loading portion. Place a logarithmic trendline through healing data points. See Figure 4.3. 9. Backwards forecast the logarithmic trendline to ending point of loading. 10. Determine the difference between d H / d 0 at 1000 cycles and backwards forecasted value. This represents the initial steric hardening effect (see Section 4.3.3). The value shown for the example in Figure 4.3 is 0.0265.

PAGE 42

30 y = 0.00010x + 1.00207 R 2 = 0.98307 y = -3.439427E-02Ln(x) + 1.313134E+00 R 2 = 9.512166E-01 0.9 0.95 1 1.05 1.1 1.15 1.2 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (seconds) d d H H /d /d 0 0 Average loading Average healing Logarithmic Trendline 0.0265 Linear Trendline 1000 Cycles of Loading Healing for 1 hour Figure 4.3: d d H / d d 0 vs. Time 11. Subtract initial steric harden ing effect from each averaged d H / d 0 healing data point by adding the value determined in Step 10 to each point. 12. Re plot average d H / d 0 vs. time using new values for healing. See Figure 4.4. 13. Plot DCSE/DCSE applied vs. d H / d 0 during the loading portion. See Figure 4.5.

PAGE 43

31 y = -3.439427E-02Ln(x) + 1.339657E+00 R 2 = 9.512166E-01 0.9 0.95 1 1.05 1.1 1.15 1.2 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (seconds) d d H H /d /d 0 0 Average loading Average healing Logarithmic Trendline Figure 4.4: Average d d H / d d 0 (Modified for Steric Hardening) vs. Time y = 9.7249x 9.7368 R 2 = 0.9831 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 0.9 0.95 1 1.05 1.1 1.15 1.2 d d H H /d /d 0 0 DCSE/DCSE(applied) Linear Trendline Figure 4.5: DCSE/DCSE applied vs. d d H / d d 0

PAGE 44

32 14. Find slope of DCSE/DCSE applied vs. d H / d 0 line. Ignore the intercept value. 15. (Relativ e absolute healing) time, t = [( d H / d 0 ) 1000 ( d H / d 0 ) t ] (slope in Step 14.) 16. Plot relative absolute healing vs. log time. See Figure 4.6. 17. Add a logarithmic trendline. Show equation. Healing rate is slope of the line. The relative healing begins at 1000 c ycles when the loading is halted. Figure 4.7 shows the comparison of healing versus time for each specimen at each temperature. When relative absolute healing equal 1.0, the specimen is fully healed. By substitution, one can determine the time to full h ealing. Figure 4.8 shows the comparison of time to full healing. Figure 4.9 shows the comparison of the healing rates. Appendix C shows the healing test figures for both mixtures at both tested temperatures. Appendix D shows the data used to generate the figures and healing rates. y = 0.3345Ln(x) 2.3105 R 2 = 0.9512 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1000 10000 100000 1000000 log Time (seconds) Relative Healing from DCSE/DCSE(applied) Logarithmic Trendline Figure 4.6: Healing vs. Time after 1000 Cycles of Loading

PAGE 45

33 y = 0.1673Ln(x) 1.1556 R 2 = 0.9862 y = 0.3345Ln(x) 2.3105 R 2 = 0.9512 y = 0.1803Ln(x) 1.2451 R 2 = 0.9849 y = 0.2321Ln(x) 1.6122 R 2 = 0.9423 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1000 10000 100000 1000000 log Time (seconds) Relative Healing from DCSE/DCSE(applied) Coarse 1 @ 10 Coarse 1 @ 15 Fine 1 @ 10 Fine 1 @ 15 Coarse 1 @ 15 Fine 1 @ 15 Fine 1 @ 10 Coarse 1 @ 10 Figure 4.7: Comparison of Healing vs. Time after 1000 Cycles of Loading 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 Mixture Type and Temperature Time to Full Healing (days) Coarse 1 Fine 1 10 o C 15 o C Figure 4.8: Compa rison of Time to Full Healing

PAGE 46

34 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Mixture Type and Temperature Healing Rate ((DCSE/DCSE(applied))/(ln(time))) Coarse 1 Fine 1 10 o C 15 o C 0.1673 0.1803 0.3345 0.2321 Figure 4.9: Comparisons of Healing Rates 4.3.1 Determination of Dissipated Creep Strain Energy As explained in the literature review, Zhang et al. (2001) developed a method to determine DCSE/cycle dur ing the linear portion of microcracking up to the threshold boundary. The formula for a 0.1 second haversine load is DCSE/cycle = 1/20 s 2 D 1 m (100) m 1 where s is the stress and D 1 and m are determined from creep tests. A variation of this formu la, showing the DCSE at a given time during testing is DCSE N = 1/20 s 2 D 1 m (100) m 1 N where N is the number of cycles. For the stress, either the tensile strength or the far away stress can be used, depending on the situation. The far away str ess equation is s FA = 2 P/( p t D) where P is the load (psi), t is the thickness (in.) and D is the diameter (in.).

PAGE 47

35 Zhang also determined a method for determining DCSE failure This process is detailed in Chapter 2, Section 2.4, with results depende nt on M R creep and strength IDT tests. Table 4.1 shows the DCSE to failure for both mixtures at both temperatures. Appendix B shows the applicable values in tabular form for determining DCSE failure and DCSE N Table 4.1: DCSE to Failure for Each Specimen Mixture Temperature DCSE failure ( o C) (psi) C1 10 1.049 C1 15 0.836 F1 10 0.75 F1 15 0.775 4.3.2 DCSE/DCSE applied vs. d H / d 0 The DCSE equation is only applicable during the loading portion of testing because DCSE is a function of cycles of loading and was developed for repetitive loading. This prevents it from being directly measured during the healing portion of the test For this study, it was desirable to relate the healing in terms of recovered energy. The relationship between DCSE and d H / d 0 is linear during the loading portion of the test. Because d H can be measured during the healing portion from M R tests and d 0 ca n be measured from the beginning of loading, d H / d 0 can be used to determine the recovered DCSE. The DCSE equation is dependent on the level of stress and different stress levels were used for each specimen to standardize time to failure. Also, each speci men has different DCSE failure limits. This prevents DCSE by itself from being used to determine a healing rate, since it cannot be related to other specimens or temperatures.

PAGE 48

36 The DCSE/DCSE applied value ratios the dissipated energy to the total dissipate d energy when loading is halted and was deemed the most appropriate representation of damage for comparison of healing. Figure 4.10 shows the comparison of DCSE/DCSE applied vs. d H / d 0 for each mixture at both temperatures. y = 3.0312x 3.0332 R 2 = 0.9961 y = 9.7249x 9.7368 R 2 = 0.9831 y = 4.9261x 4.9585 R 2 = 0.9875 y = 5.4389x 5.4617 R 2 = 0.9906 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.9 1 1.1 1.2 1.3 1.4 1.5 d d H H /d /d 0 0 DCSE/DCSE(applied) Coarse 1 @ 10 Coarse 1 @ 15 Fine 1 @ 10 Fine 1 @ 15 Coarse 1 @ 15 Fine 1 @ 15 Fine 1 @ 10 Coarse 1 @ 10 Figure 4.10: Comparison of DCSE/DCSE applied vs. d d H / d d 0 4.3.3 Steric Hardening Effects Steric hardening is a structural hardening associated with asphalt materials. Over time, an asphal t pavement will creep to a more stable structure and develop a higher stiffness. This stiffness is reversible and mechanical action will return the asphalt concrete to its original state. E. Barth (1962) stated that Age hardening or steric hardening in an asphalt occurs rapidly at first but appears to approach a limiting degree of hardness on prolonged standing. Tests supporting this statement found that 36% of ultimate steric hardening occurred within 10 hours of testing near freezing at 2.5 o C.

PAGE 49

37 Steric hardening explains the sudden decrease in stiffness at the beginning of loading and increase in stiffness after the repeated loading is stopped. Reversal of steric hardening occurs during the first couple of hundred cycles of loading. After loading stops the asphalt specimen experiences an immediate restructuring that is a partial recovery of steric hardening. By ignoring the initial stiffness in the determination of the d 0 value, the d 0 is associated with the response of the material with no steric hardening. While loading occurs, steric hardening is gone. When loading stops, the material starts to restructure and begins to heal. By bringing the backwards forecasted tr endline up, initial steric hardening effects are eliminated from the response of the mixture. Because initial steric hardening is accounted for, when d 0 equals 1.0, the specimen is fully healed. By looking at Figure 4.2, one can see the large jump between the first two recorded data points before the specimen is damaged in a linear fashion. If temperature effects are the sole cause for the jump, then the difference between d 0 and d i should be able to be subtracted out at the end of the loading portion of the test. Because the temperature stabilizes after two minutes of loading, then it should return to its previous level within two minutes of unloading. However, this would show that no healing occurred and all healing activity was temperature restabiliza tion. Steric hardening provides a rational explanation to this inconsistency. 4.4 Damage Accumulation and Healing Across Temperatures Development of a healing rate provided a way to analyze damage accumulation that translated across temperatures. The tw o temperature fracture tests involved loading the specimen for 1000 cycles at 15 o C, cooling to 10 o C for an hour and then continuing

PAGE 50

38 the test until failure of the specimen. Figure 4.11 shows the loading portion at 15 o C. Figures 4.12 and 4.13 show the cont inuation of the tests at 10 o C. The average original d 0 from virgin 10 o C tests was used to accurately portray the d H / d 0 values. Damage translation can be seen by the fact that d H / d 0 at the beginning of the secondary test is not 1.0. 4.4.1 Stress Values The stress used in DCSE and reference values for compar ison are important for correctly explaining what is occurring during healing. The correct stress to use for DCSE/cycle calculations varies based on the situation. On a local level, to determine the percentage to failure of a specimen at the end of loadin g, the DCSE should be calculated with tensile strength as the stress. At the crack tip, tensile strength dominates and shows how close the specimen is to ultimate failure. Figure 4.14 shows the damage to each specimen at each temperature after 1000 cycles of loading. For purposes of comparison on a global scale, the far away stress is appropriate. As a whole, the stress in a specimen during loading is much closer to the far away stress. The far away stress accounts for the load applied. The damage lev el is related to the applied stress and different loads were used for each mixture and temperature. To correctly compare a mixture at two temperatures, the global approach is required. Much of the specimen that influences deformation is being subjected t o stresses that are closer to the s FA which dominates the response.

PAGE 51

39 0.9 0.95 1 1.05 1.1 1.15 1.2 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Number of Load Replications (cycles) d d H / d d 0 Average loading Figure 4.11: d d H / d d 0 vs. Cycles for Initial Loading at 15 o C 1 1.4 1.8 2.2 2.6 3 0 500 1000 1500 2000 2500 3000 3500 Number of Load Replications (cycles) d d H / d d 0 t10b-c1 linear t10b-c1 curve t10c-c1 linear t10c-c1 curve t13c-c1 linear t13c-c1 curve 1.082 At 10 o C after 1000 cycles of loading at 15 o C Figure 4.12: d d H / d d 0 vs. Number of Load Replications for C1

PAGE 52

40 1 1.4 1.8 2.2 2.6 3 0 500 1000 1500 2000 2500 3000 Number of Load Replications (cycles) d d H / d d 0 t8f1c linear t8f1c curve t10f1b linear t10f1b curve At 10 o C after 1000 cycles of loading at 15 o C 1.236 Figure 4.13: d d H / d d 0 vs. Number of Load Replications for F1 0 0.1 0.2 0.3 0.4 0.5 0.6 Mixture Type and Temperature % to Failure at End of Loading Coarse 1 Fine 1 10 o C 15 o C Figure 4.14: DCSE applied(using s s 0 )/DCSE failure vs. Mixture Type and Temperature

PAGE 53

41 4.4.2 Damage Translation Process The process to translate damage across temperature s is shown in Tables 4.2 and 4.3. Because the temperature is cooled from 15 o C to 10 o C, healing is occurring at both temperatures. While the temperature will be at 10 o C for most of the hour, it has been shown that healing occurs very quickly at first, so the effect of the higher temperature present at the beginning of healing influences the healing rate. Because of this, healing has been analyzed at both temperatures to understand what is occurring. y = 3.0815x + 1.0013 R 2 = 0.9961 y = 1.1709x + 1.0021 R 2 = 0.9831 y = 2.1082x + 1.0078 R 2 = 0.9875 y = 2.1413x + 1.005 R 2 = 0.9906 0.9 1 1.1 1.2 1.3 1.4 1.5 0 0.05 0.1 0.15 0.2 0.25 0.3 DCSE/DCSE(failure) d d H H /d /d 0 0 Coarse 1 @ 10 Coarse 1 @ 15 Fine 1 @ 10 Fine 1 @ 15 Coarse 1 @ 15 Fine 1 @ 15 Fine 1 @ 10 Coarse 1 @ 10 Figure 4.15: d d H / d d 0 vs. DCSE/DCSE failure First, the relative healing at one hour must be found using the healing rate equations in Figure 4.7. Table 4.2 shows the steps to determining recovered d H / d 0 Because the loading stopped at 1000 seconds, 4600 seconds represent s one hour of healing. The relative healing values are then inserted into Figure 4.10 to determine the equivalent d H / d 0 The recovered d H / d 0 during the hour of healing is determined by

PAGE 54

42 subtracting the value 1.0 from the equivalent d H / d 0 Figure 4.15 sho ws d H / d 0 vs. DCSE/DCSE failure The DCSE values calculated using the far away stress as discussed in Section 4.2.1. As shown by the arrows, the DCSE/DCSE failure value at the end of loading at 15 o C is converted to the equivalent d H / d 0 at 10 o C. Table 4.3 s hows the remaining steps for damage translation across temperatures. The damage translated d H / d 0 value at the end of the hour of healing is then determined by subtracting the recovered d H / d 0 from the d H / d 0 at 10 o C. Table 4.2: Recovered d d H / d d 0 Mixture Temp erature Relative Healing in One Hour Equivalent d H / d 0 Recovered d H / d 0 ( o C) (Using Figure 4.7) (Using Figure 4.10) (1 Equivalent d H / d 0 ) C1 15 0.510609974 1.169111235 0.169111235 10 0.255376678 1.084909171 0.084909171 F1 15 0.345287668 1 .076670727 0.076670727 10 0.275516228 1.062507101 0.062507101 Table 4.3: Comparison of Calculated and Tested d d H / d d 0 Mixture Temp. d H / d 0 @ 10 o C d H / d 0 @ end of hour d H / d 0 from at End of Loading Two Temperature Test ( o C) (Using Figure 4.15) (d H / d 0 @ 10 o C Recovered d H / d 0 ) (From Fig.4.12 and 4.13) C1 15 1.098232954 10 1.267344188 1.182435018 1.082 F1 15 1.11044478 10 1.187115507 1.124608406 1.236

PAGE 55

43 In comparing with the tested d H / d 0 value at the beginning of the secondar y test, it can be seen that the C1 healed almost completely at 15 o C in the hour. That corresponds with the C1 having a very high healing rate at 15 o C and a low healing rate at 10 o C. The F1 does not match up as well. The two temperature tested d H / d 0 is h igher than either of the calculated healed values. It can be noted that F1 heals almost as fast at 10 o C as at 15 o C in the short term. Clearly everything did not heal, but the process is there and makes sense. Some damage translated across temperatures and some healing occurred. 4.5 Permanent Deformation Permanent deformation plots were generated for each healing test. Figure 4.16 shows a sample plot. Permanent deformation, or creep strain, is normally what is associated with damage. The plot shows t hat very little permanent deformation is recovered during the healing portion of the test. Delayed elasticity, or time dependent elastic behavior, effects are almost entirely eliminated by the 0.9 second rest period within each cycle, which indicates the test is being run properly. By observing the permanent deformation plots, one can see that a specimen does not have to have much strain (creep) recovery to have healing. Also, healing cannot be measured or interpreted from recovered strain. The percent of strain recovered is much less than the percent of DCSE recovered. The permanent deformation plot also supports the steric hardening idea. There is a drastic change in the rate of change of resilient deformation at the beginning of loading. A proportio nal increase in rate of permanent deformation, or creep strain, is not seen. If

PAGE 56

44 temperature were the controlling factor for the initial jump, it would be evident in the creep strain as well. 0.000000 0.001000 0.002000 0.003000 0.004000 0.005000 0.006000 0.007000 0.008000 0.009000 0.010000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (seconds) Permanent Deformation (inches) Figure 4.16: Permanent Deformation vs. Time 4.6 Summary of Findings and Analyses Relative healing is the recovered energy relative to the applied energy as a function of time. At any given point in time, the relative healing can be determined using the Healing vs. Time relationship. Just sh owing the DCSE healed is ineffectual as a comparative tool because each specimen has different failure limits and different stress levels during loading. An assumption made for this healing rate determination is that the rate of healing is directly relate d to the amount of damage. The healing rate will be faster with a larger load and is primarily dependent on damage. If there is more energy, there will be more healing.

PAGE 57

45 There are three identified possible mechanisms associated with recovery of the effe ctive stiffness of mixtures: actual healing (microcrack), reduction (restabilization) of temperature, and steric hardening. Steric hardening appears to be the initial mechanism when loading is halted. By considering steric hardening to be the primary eff ect, one can subtract out the initial steric hardening recovery and assume the specimen will eventually recover all steric hardening over time with long enough rest periods. The remainder can be considered as healing. The healing rate probably overestim ates true healing due to residual steric hardening occurring during the monitored healing portion of the test. Because an M R test cannot be run for about two minutes after the fracture test is halted, it is difficult to accurately determine what is occurr ing in the early stages of healing. The healing rate converted a normalized horizontal deformation into dissipated creep strain energy. This overcame the barrier of measuring DCSE without repetitive loading, as previously developed formulae are only appli cable during loading. The relationship between d H / d 0 and DCSE during loading was used to convert measured d H values from M R tests into DCSE. With healing computed in terms of DCSE, the healing rates can be incorporated into the HMA Crack Growth Model for further refinement of that model. Damage was sh own to accumulate across temperatures and a process was developed to translate damage and recovery across temperatures. At higher temperatures, the healing rates are much higher and mixture C1 heals much faster than F1 at 15 o C. Even over the small temper ature range of 5 o C, a difference of multiple days separates the amount of healing that occurs.

PAGE 58

46 CHAPTER 5 SUMMARY, CONCLUSIONS AND RECOMMENDATIONS 5.1 Summary of Findings Healing is a quantifiable measurement in asphalt pavements that can be compared between mixtures and at different temperatures by using healing rate process and comparison procedures develo ped in this thesis. Findings of note: Healing rate was measured as a recovery of the effective stiffness of a mixture. An approach was developed to describe healing rate in terms of recovered DCSE. Creep strain recovery did not appear to be necessary for healing. Healing was determined to be a non linear process that occurred quickly at first, but decreased over time. A logarithmic relationship appeared to provide the best fit. Steric hardening appeared to be the primary reason for initial rapid reductio n in stiffness during testing. A procedure was developed to separate the steric hardening effect from the healing effect so that healing rates could be determined more accurately. The coarse graded mixture was found to heal much faster than the fine grad ed mixture at 15 o C, even though the same binder was used. This implies that below a certain temperature, there is almost no healing. It also implies that above a certain temperature, the sample incurs almost no damage because of the immediate healing. 5.2 Conclusions The process to describe healing in terms of DCSE involves many steps and many possible interpretations. Some conclusions of note:

PAGE 59

47 The recovered strains do not appear to be an adequate measure of healing. The fact that healing can be expressed in terms of DCSE indicates that it can be incorporated into the HMA Crack Growth Model developed by Zhang (2000). Steric hardening appears to play a significant role in a mixtures response during both the damage and the healing portions of labo ratory tests and therefore must be considered in their interpretation. At 10 o C, healing is much slower than at 15 o C. This shows that healing is very dependent on temperature and implies that at high enough temperatures and viscosities, healing is almost immediate. Healing is an important aspect of field performance that needs to be considered. 5.3 Recommendations This research project has set the stage for future work. With the developed healing rate, a greater understanding of healing can be achieve d. Some recommendations of note: Run tests at higher temperatures, such as 20 o C and 25 o C. This will provide a more complete look of healing effects due to temperature. Specimen internal temperatures should be measured with thermocouples, so that a mea sure of the temperature as a function of time can be seen. Steric hardening needs to be looked at carefully over time. Tests should be performed that include healing for some time and reloading; in essence, a healing test on top of one. This could be u sed to see if the healing rate changes. One of the assumptions made in the determination of the healing rate is that the amount of induced damage determines the rate of healing. Run tests to two different damage levels and see if they converge at same healing point. Incorporate the healing rate into the HMA Fracture Mechanics Law and Crack Growth Model.

PAGE 60

APPENDIX A DATA FROM TESTED SPECIMENS

PAGE 61

49 Table A1: Healing Fracture Tests for C1 at 15 o C Specimen T18a c1 T18c c1 Asphalt Content (AC%) 6.5 6.5 Theoretical Maximum Specific Gravity (Gmm) 2.327 9 2.3279 Apparent Specific Gravity (Gsa) 1.035 1.035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.159 2.17 Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 Asphalt Absorption (Pba ) 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 Percent VMA in Compacted Mix (VMA) 18.24 17.82 Percent Air Voids in Compacted Mix (AV) 7.26 6.78 Percent VFA in Compacted Mix (VFA) 60.22 61.94 (Honeycutt, 2000) Table A2: Healing Fracture Tests for F1 at 15 o C Specimen T11f1b T11f1c T12f1b Asphalt Content (AC%) 6.3 6.3 6.3 Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378 Apparent Specific Gravity (Gsa) 1.035 1.035 1.035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.183 2.178 2.171 Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 2.488 Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effecti ve Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273 Percent VMA in Compacted Mix (VMA) 17.79 17.97 18.24 Percent Air Voids in Compacted Mix (AV) 6.62 6.84 7.13 Percent VFA in Compacted Mix (VFA) 62.77 61.97 60.88 (Honeycutt, 2000)

PAGE 62

50 Table A3: Healing Fracture Tests for C1 at 10 o C Specimen T17a c1 T19b c1 Asphalt Content (AC%) 6.5 6.5 Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 Apparent Specific Gravity (Gsa) 1.035 1. 035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.153 2.155 Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 Asphalt Absorption (Pba) 1.313 1.313 Effective Asphalt Content of Mixtu re (Pbe) 5.273 5.273 Percent VMA in Compacted Mix (VMA) 18.47 18.39 Percent Air Voids in Compacted Mix (AV) 7.51 7.43 Percent VFA in Compacted Mix (VFA) 59.32 59.61 (Honeycutt, 2000) Table A4: Healing Frac ture Tests for F1 at 10 o C Specimen T13f1a T13f1b T12f1b Asphalt Content (AC%) 6.3 6.3 6.3 Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378 Apparent Specific Gravity (Gsa) 1.035 1.035 1.035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.177 2.182 2.171 Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 2.488 Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.2 73 5.273 Percent VMA in Compacted Mix (VMA) 18.01 17.82 18.24 Percent Air Voids in Compacted Mix (AV) 6.88 6.66 7.13 Percent VFA in Compacted Mix (VFA) 61.81 62.61 60.88 (Honeycutt, 2000)

PAGE 63

51 Table A5: Fracture Tests to Failure for C1 at 15 o C Specimen T9a c1 T9b c1 T10a c1 Asphalt Content (AC%) 6.5 6.5 6.5 Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 2.3279 Apparent Specific Gravity (Gsa) 1.035 1.035 1.035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.153 2.156 2.162 Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 2.469 Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 2.549 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.2 73 5.273 Percent VMA in Compacted Mix (VMA) 18.47 18.35 18.13 Percent Air Voids in Compacted Mix (AV) 7.51 7.38 7.13 Percent VFA in Compacted Mix (VFA) 59.32 59.77 60.68 (Honeycutt, 2000) Table A6: Fracture Tests to Failur e for F1 at 15 o C Specimen T4f1b T7f1c T8f1b Asphalt Content (AC%) 6.3 6.3 6.3 Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378 Apparent Specific Gravity (Gsa) 1.035 1.035 1.035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.17 6 2.185 2.165 Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 2.488 Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273 Pe rcent VMA in Compacted Mix (VMA) 18.05 17.71 18.46 Percent Air Voids in Compacted Mix (AV) 6.92 6.54 7.39 Percent VFA in Compacted Mix (VFA) 61.66 63.10 59.97 (Honeycutt, 2000)

PAGE 64

52 Table A7: Fracture Tests to Failure for C1 at 10 o C Specimen T8a c1 T11c c1 c1 18 a Asphalt Content (AC%) 6.5 6.5 6.5 Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 2.3279 Apparent Specific Gravity (Gsa) 1.035 1.035 1.035 Bulk Specific Gr avity of Compacted Mix (Gmb) 2.148 2.154 2.159 Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 2.469 Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 2.549 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effective Asphalt Content of Mi xture (Pbe) 5.273 5.273 5.273 Percent VMA in Compacted Mix (VMA) 18.66 18.43 18.24 Percent Air Voids in Compacted Mix (AV) 7.73 7.47 7.26 Percent VFA in Compacted Mix (VFA) 58.58 59.46 60.22 (Honeycutt, 2000) Table A8: Fracture Tests to Failure for F1 at 10 o C Specimen T10f1c T7f1a f1 10 d Asphalt Content (AC%) 6.3 6.3 6.3 Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378 Apparent Specific Gravity (Gsa) 1.035 1.035 1.035 Bulk Spec ific Gravity of Compacted Mix (Gmb) 2.163 2.18 2.172 Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 2.488 Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273 Percent VMA in Compacted Mix (VMA) 18.54 17.90 18.20 Percent Air Voids in Compacted Mix (AV) 7.48 6.75 7.09 Percent VFA in Compacted Mix (VFA) 59.67 62.29 61.03 (Honeycutt, 2000)

PAGE 65

53 Table A9: Two Temperature Fracture Tests to Failure for C1 Specimen T10b c1 T10c c1 T13c c1 Asphalt Content (AC%) 6.5 6.5 6.5 Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 2.3279 Apparent Specific Gravity (Gsa) 1.035 1.035 1. 035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.163 2.158 2.155 Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 2.469 Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 2.549 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273 Percent VMA in Compacted Mix (VMA) 18.09 18.28 18.39 Percent Air Voids in Compacted Mix (AV) 7.08 7.30 7.43 Percent VFA in Compacted Mix (VFA) 60.84 60.07 59.61 (Honeycutt, 2000) Table A10: Two Temperature Fracture Tests to Failure for F1 Specimen T8f1c T10f1b Asphalt Content (AC%) 6.3 6.3 Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 Apparent Specific Gravity (Gsa) 1.035 1.035 Bulk Specific Gravity o f Compacted Mix (Gmb) 2.164 2.162 Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 Asphalt Absorption (Pba) 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 Perc ent VMA in Compacted Mix (VMA) 18.50 18.58 Percent Air Voids in Compacted Mix (AV) 7.43 7.52 Percent VFA in Compacted Mix (VFA) 59.82 59.52 (Honeycutt, 2000)

PAGE 66

54 Table A11: M R Creep and Strength Tests for C1 at 15 o C Specimen T15a c1 T15b c1 T16a c1 Asphalt Content (AC%) 6.5 6.5 6.5 Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 2.3279 Apparent Specific Gravity (Gsa) 1.035 1.035 1.035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.154 2.155 2.159 Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 2.469 Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 2.549 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273 Percent VMA in Compact ed Mix (VMA) 18.43 18.39 18.24 Percent Air Voids in Compacted Mix (AV) 7.47 7.43 7.26 Percent VFA in Compacted Mix (VFA) 59.46 59.61 60.22 (Honeycutt, 2000) Table A12: M R Creep and Strength Tests for F1 at 15 o C Specimen T1f1a T1f1b T2f1a Asphalt Content (AC%) 6.3 6.3 6.3 Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378 Apparent Specific Gravity (Gsa) 1.035 1.035 1.035 Bulk Specific Gravity of Compacted Mix (Gmb) 2.178 2.175 2.169 Bulk Specifi c Gravity of Aggregate (Gsb) 2.488 2.488 2.488 Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554 Asphalt Absorption (Pba) 1.313 1.313 1.313 Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273 Percent VMA in Compacted Mix (VMA) 17.97 18.09 18.31 Percent Air Voids in Compacted Mix (AV) 6.84 6.96 7.22 Percent VFA in Compacted Mix (VFA) 61.97 61.50 60.57 (Honeycutt, 2000)

PAGE 67

APPENDIX B SUMMARY OF MIXTURES TEST RESULTS

PAGE 68

56 Table B1: Resilient Modulus Test Data Resilient Modulus for STOA Mixture Type Temp. Trimmed Mean Values Average Value Poisson Ratio ( o C) (GPa) (GPa) 7.85 0.42 Coarse 1 10 7.82 7.92 0.42 8.08 0.39 5.82 0.37 Coarse 1 15 5.68 5.75 0.38 5.75 0.36 9.4 0.27 Fine 1 10 9.39 9.49 0.28 9.68 0.25 5.85 0.25 Fine 1 15 5.9 5.87 0.22 5.87 0.23 (Properties from 10 C tests from Honeycutt, 2000) Table B2: Tensile Strength Test Data Tensile Strength for STOA Mixture Type Temp. Trimmed Mean Values Average Value Poisson Ratio ( o C) (MPa) (MPa) 1.57 Coarse 1 10 1.75 1.64 0.5 1.5 1.25 Coarse 1 15 1.22 1.24 0.38 1.24 2.05 Fine 1 10 2.28 2.08 0.46 1.9 1.37 Fine 1 15 1.37 1.36 0.31 1.35 (Properties from 10 C tests from Hon eycutt, 2000)

PAGE 69

57 Table B3: Creep Test Data Mixture Type Temp. m value D1 value Failure Strain Fracture Energy Poisson Ratio ( o C) (1/psi) ( me ) (kJ/m 3 ) (from creep) Coarse 1 10 0.796 3.81E 07 4629.75 7.4 0.5 Coarse 1 15 0.8 8.37E 07 5862.07 5.9 0.38 Fine 1 10 0.656 5.73E 07 2919.62 5.4 0.46 Fine 1 15 0.6027 1.57E 06 5045.68 5.5 0.31 (Properties from 10 C tests from Honeycutt, 2000) Table B4: Dissipated Creep Strain Energy Calculation Mixture Type Coarse 1 Fine 1 Temp. ( o C) 10 15 10 15 Resilient Modulus (GPa) 7.92 5.75 9.49 5.87 Failure Strai n ( me ) 4629.75 5862.07 2919.62 5045.68 Tensile Strength (MPa) 1.64 1.24 2.08 1.36 Initial Strain ( me ) 4422.68 5646.42 2700.44 4813.99 Elastic Energy (kJ/m 3 ) 0.1698 0.1337 0.2279 0.1575 Fracture Energy (kJ/m 3 ) 7.4 5.9 5.4 5.5 Dissipated Energy to Failu re (kJ/m 3 ) 7.2302 5.7663 5.1721 5.3425 Dissipated Energy to Failure (psi) 1.0486 0.8363 0.7501 0.7748 (Properties from 10 C tests from Zhang, 2000)

PAGE 70

APPENDIX C FRACTURE TEST FIGURES

PAGE 71

59 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0.0009 0.001 0.0011 0.0012 0.0013 0.0014 0 500 1000 1500 2000 2500 3000 3500 4000 Number of Load Replications (cycles) Resilient Deformation (inches) T9a linear T9a curve T9b linear T9b curve T10a linear T10a curve Figure C1: Resilient Deformation vs. Load Replications (C1 at 15 o C to failure) 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0.0009 0.001 0.0011 0.0012 0.0013 0.0014 0 500 1000 1500 2000 2500 3000 3500 4000 Number of Load Replications (cycles) Resilient Deformation (inches) T4f1b linear T4f1b curve T7f1c linear T7f1c curve T8f1b linear T8f1b curve Figure C2: Resilient Deformation vs. Load Replications (F1 a t 15 o C to failure)

PAGE 72

60 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0.0009 0.001 0.0011 0.0012 0.0013 0.0014 0 500 1000 1500 2000 2500 3000 3500 4000 Number of Load Replications (cycles) Resilient Deformation (inches) T8a linear T8a curve T11c linear T11c curve c1-18-a linear c1-18-a curve Figure C3: Resilient Deformation vs. Load Replications (C1 at 10 o C to failure) 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0.0009 0.001 0.0011 0.0012 0.0013 0.0014 0 500 1000 1500 2000 2500 3000 3500 4000 Number of Load Replications (cycles) Resilient Deformation (inches) f1-10-d linear f1-10-d curve t10f1c linear t10f1c curve t7f1a linear t7f1a curve Figure C4: Resilient Deformation vs. Load Replications (F1 at 10 o C to failure)

PAGE 73

61 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) Resilient Deformation (inches) t18a-c1 t18c-c1 Figure C5: Resilient Deformation vs. Time (C1 at 15 o C healing test) 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) Resilient Deformation (inches) t11f1b t11f1c t12f1b Figure C6: Resilient Deformation vs. Time (F1 at 15 o C healing test)

PAGE 74

62 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) Resilient Deformation (inches) t17a-c1 t19b-c1 Figure C7: Resilient Deformation vs. Time (C1 at 10 o C healing test ) 0.0002 0.0003 0.0004 0.0005 0.0006 0.0007 0.0008 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) Resilient Deformation (inches) t13f1a t13f1b t12f1b Figure C8: Resilient Deformation vs. Time (F1 at 10 o C healing test)

PAGE 75

63 y = 0.00010x + 1.00207 R 2 = 0.98307 y = -3.439427E-02Ln(x) + 1.313134E+00 R 2 = 9.512166E-01 0.9 1 1.1 1.2 1.3 1.4 1.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) d d H / d d 0 Average loading Average healing Logarithmic Trendline 0.0265 Figure C9: Average d d H / d d 0 vs. Time (C1 at 15 o C healing test) y = 0.00018x + 1.00503 R 2 = 0.99059 y = -4.267977E-02Ln(x) + 1.481454E+00 R 2 = 9.423406E-01 0.9 1 1.1 1.2 1.3 1.4 1.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) d d H / d d 0 Average loading Average healing Logarithmic Trendline 0.000 Figure C10: Average d d H / d d 0 vs. Time (F1 at 15 o C healing test)

PAGE 76

64 y = 0.00033x + 1.00130 R 2 = 0.99606 y = -5.518856E-02Ln(x) + 1.677661E+00 R 2 = 9.862454E-01 0.9 1 1.1 1.2 1.3 1.4 1.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) d d H / d d 0 Average loading Average healing Logarithmic Trendline 0.0349 Figure C11: Average d d H / d d 0 vs. Time (C1 at 10 o C healing test) y = 0.00020x + 1.00779 R 2 = 0.98753 y = -3.659121E-02Ln(x) + 1.427938E+00 R 2 = 9.849328E-01 0.9 1 1.1 1.2 1.3 1.4 1.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) d d H / d d 0 Average loading Average healing Logarithmic Trendline 0.0326 Figure C12: Av erage d d H / d d 0 vs. Time (F1 at 10 o C healing test)

PAGE 77

65 y = 0.00010x + 1.00207 R 2 = 0.98307 y = -3.439427E-02Ln(x) + 1.339657E+00 R 2 = 9.512166E-01 0.9 1 1.1 1.2 1.3 1.4 1.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) d d H / d d 0 Average loading Average healing Logarithmic Trendline Figure C13: Average d d H / d d 0 without Steric Hardening vs. Time (C1 at 15 o C healing) y = 0.00018x + 1.00503 R 2 = 0.99059 y = -4.267977E-02Ln(x) + 1.481454E+00 R 2 = 9.423406E-01 0.9 1 1.1 1.2 1.3 1.4 1.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) d d H / d d 0 Average loading Average healing Logarithmic Trendline Figure C14: Average d d H / d d 0 without Steric Hardening vs. Time (F1 at 15 o C hea ling)

PAGE 78

66 y = 0.00033x + 1.00130 R 2 = 0.99606 y = -5.518856E-02Ln(x) + 1.712529E+00 R 2 = 9.862454E-01 0.9 1 1.1 1.2 1.3 1.4 1.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) d d H / d d 0 Average loading Average healing Logarithmic Trendline Figure C15: Average d d H / d d 0 without Steric Hardening vs. Time (C1 at 10 o C healing) y = 0.00020x + 1.00779 R 2 = 0.98753 y = -3.659121E-02Ln(x) + 1.460553E+00 R 2 = 9.849328E-01 0.9 1 1.1 1.2 1.3 1.4 1.5 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) d d H / d d 0 Average loading Average healing Logarithmic Trendline Figure C16: Average d d H / d d 0 without Steric Hardening vs. Time (F1 at 10 o C healing)

PAGE 79

67 0.000000 0.001000 0.002000 0.003000 0.004000 0.005000 0.006000 0.007000 0.008000 0.009000 0.010000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) Permanent Deformation (inches) t18a-c1 t18c-c1 Figure C 17: Permanent Deformation vs. Time (C1 at 15 o C healing test) 0.000000 0.001000 0.002000 0.003000 0.004000 0.005000 0.006000 0.007000 0.008000 0.009000 0.010000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) Permanent Deformation (inches) t11f1b t11f1c t12f1b Figure C18: Permanent Deformation vs. Time (F1 at 15 o C healing test)

PAGE 80

68 0.000000 0.001000 0.002000 0.003000 0.004000 0.005000 0.006000 0.007000 0.008000 0.009000 0.010000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) Permanent Deformation (inches) t17a-c1 t19b-c1 Figure C19: Permanent Deformation vs. Time (C1 at 10 o C healing test) 0.000000 0.001000 0.002000 0.003000 0.004000 0.005000 0.006000 0.007000 0.008000 0.009000 0.010000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Time (second) Permanent Deformation (inches) t13f1a t13f1b t12f1b Figure C20: Permanent Deformation vs. Time (F1 at 10 o C healing test)

PAGE 81

APPENDIX D HEALING TEST DATA

PAGE 82

70 Table D1: Coarse 1 at 10 o C Healing Specimens Specimen: t17a c1 Time d H d H norm d H /d 0 (seconds) (inches) (inches) 14 0.000495 1. 0.000485 1.0000 120 0. 000511 0.000501 1.0316 250 0.000545 0.000535 1.1017 360 0.000565 0.000554 1.1418 480 0.000590 0.000578 1.1916 600 0.000616 0.000604 1.2451 730 0.000629 0.000617 1.2710 840 0.000650 0.000637 1.3135 960 0.000675 0.000662 1.3643 1187 0 .000661 0.000648 1.3347 1375 0.000655 0.000642 1.3229 1633 0.000652 0.000639 1.3168 2830 0.000632 0.000620 1.2774 4599 0.000620 0.000607 1.2515 1. d 0 = 0.000495 in., Actual d H = 0.000439 in. Specimen: t19b c1 Time d H d H norm d H /d 0 (seconds) (inches) (inches) 8 0.000570 2. 0.000581 1.0000 120 0.000585 0.000597 1.0266 240 0.000611 0.000623 1.0716 360 0.000628 0.000641 1.10 23 480 0.000650 0.000662 1.1398 600 0.000668 0.000681 1.1725 720 0.000686 0.000700 1.2041 840 0.000704 0.000717 1.2345 960 0.000717 0.000731 1.2570 1145 0.000704 0.000717 1.2345 1359 0.000709 0.000723 1.2442 1629 0.000698 0.000712 1 .2249 2834 0.000679 0.000693 1.1918 4620 0.000671 0.000684 1.1766 2. d 0 = 0.000570 in., Actual d H = 0.000509 in. Average thickness = 1.02 in. Load = 1300 lb T17a c1 thickness = 1.00 in. T19b c1 thickness = 1.04 in. Normalizing formula for thickness = d H t i /t avgi P avg /P i

PAGE 83

71 Table D2: Coarse 1 at 10 o C Healing Calculations Average Average Average d H /d 0 DCSE/ DCSE/ Relative Abs. Time d H /d 0 w/o st. hard. DCSE DCSEfail DCSEapplied Healing 0 1.000 0.0000 0.0000 0.0000 120 1.029 0.0134 0.0 128 0.1200 245 1.087 0.0274 0.0261 0.2450 360 1.122 0.0403 0.0384 0.3600 480 1.166 0.0537 0.0512 0.4800 600 1.209 0.0671 0.0640 0.6000 725 1.238 0.0811 0.0773 0.7250 840 1.274 0.0939 0.0896 0.8400 960 1.311 0.1073 0.1024 0.9600 1000 1.331 0.1118 0.1066 1.0000 1166 1.285 1.31946 0.03589 1367 1.284 1.31839 0.03913 1631 1.271 1.30571 0.07756 2832 1.235 1.26950 0.18734 4609.5 1.214 1.24893 0.24968 Linear trendline equation for x = 1000, y=0.00033x+1.00130 = 1.331 Log trendline equation for x = 1000, y = 0.055189*Ln(x)+1.6777 = 1.296 Diff. between backcalculated dh/do from log. at 1000s. and linear at 1000s = 0.035 Slope of DCSE/DCSEapplied = 3.0312 m value = 0.796 D 1 value = 3.81E 07 s FA = 137.359 psi = 2*1300/(3.14*1.02*5.91) s 0 = 233.019 psi = (1.57+1.75+1.5)/3*1000/6.895 DCSEfail = 1.04862 psi

PAGE 84

72 Table D3: Fine 1 at 10 o C Healing Specimens Specimen: t13f1a Time d H d H norm d H /d 0 (seconds) (inches) (inches) 8 0.000430 1. 0.000437 1.0000 120 0.000438 0.000445 1.0178 240 0.000449 0.000457 1.0450 360 0.000459 0.000467 1.0678 480 0.000472 0.000479 1.0969 640 0.000477 0.000485 1.1101 720 0. 000488 0.000496 1.1345 840 0.000493 0.000500 1.1453 960 0.000498 0.000506 1.1589 1165 0.000489 0.000496 1.1360 1367 0.000488 0.000496 1.1353 1602 0.000483 0.000491 1.1229 2898 0.000479 0.000487 1.1136 4621 0.000469 0.000477 1.0915 1 d 0 = 0.000430 in., Actual d H = 0.000386 in. Specimen: t13f1b Time d H d H norm d H /d 0 (seconds) (inches) (inches) 9 0.000410 2. 0.000385 1.0000 120 0.000419 0.000393 1.0224 240 0.000436 0.000409 1.0626 365 0.000445 0.000418 1.085 8 480 0.000447 0.000420 1.0907 600 0.000457 0.000429 1.1154 720 0.000464 0.000436 1.1325 840 0.000470 0.000441 1.1467 960 0.000477 0.000447 1.1622 1173 0.000472 0.000443 1.1508 1370 0.000465 0.000436 1.1333 1596 0.000464 0.000436 1. 1325 2889 0.000450 0.000423 1.0980 4734 0.000443 0.000416 1.0809 2. d 0 = 0.000410 in., Actual d H = 0.000371 in.

PAGE 85

73 Table D4: Fine 1 at 10 o C Healing Specimens continued Specimen: t12f1b Time d H d H norm d H /d 0 (seconds) (inches) (inches) 17 0.000400 3. 0.000418 1.0000 120 0.0004 12 0.000430 1.0292 240 0.000430 0.000450 1.0754 360 0.000447 0.000467 1.1179 480 0.000459 0.000479 1.1463 605 0.000471 0.000492 1.1779 720 0.000479 0.000501 1.1983 840 0.000491 0.000513 1.2279 960 0.000500 0.000523 1.2508 1217 0.000 492 0.000514 1.2296 1439 0.000484 0.000506 1.2096 1622 0.000488 0.000510 1.2200 2885 0.000478 0.000499 1.1938 4678 0.000475 0.000496 1.1871 1. d 0 = 0.000400 in., Actual d H = 0.000368 in. Average thickness = 1.03 in. Load = 1300 lb T13f1a thickness = 1.05 in. T13f1b thickness = 0.97 in. T12f1b thickness = 1.08 in. Normalizing formula for thickness = d H t i /t avgi P avg /P i

PAGE 86

74 Table D5: Fine 1 at 10 o C Healing Calculations Average Average Average di/do DCSE/ DCSE/ Relative Abs. Time di/do w/o st. hard. DCSE DCSEfail DCSEapplied Healing 0 1.000 0 .0000 0.0000 0.0000 120 1.023 0.0086 0.0114 0.1200 240 1.061 0.0171 0.0228 0.2400 361.66667 1.091 0.0258 0.0344 0.3617 480 1.111 0.0342 0.0456 0.4800 615 1.134 0.0439 0.0585 0.6150 720 1.155 0.0514 0.0685 0.7200 840 1.173 0.0599 0.0799 0.8400 960 1.191 0.0685 0.0913 0.9600 1000 1.208 0.0713 0.0951 1.0000 0.00000 1185 1.172 1.20476 0.01491 1392 1.159 1.19201 0.07773 1606.6667 1.158 1.19108 0.08233 2890.6667 1.135 1.16771 0.19744 4677.6 667 1.120 1.15243 0.27270 Linear trendline equation for x = 1000, y=0.0002x+1.00779 = 1.208 Log trendline equation for x = 1000, y = 0.03659*Ln(x)+1.4279 = 1.175 Diff. between backcalculated dh/do from log. at 1000s. and linea r at 1000s = 0.033 Slope of DCSE/DCSEapplied = 4.9261 m value = 0.656 D 1 value = 5.73E 07 s FA = 136.025 psi = 2*1300/(3.14*1.03*5.91) s 0 = 301.184 psi = (2.05+2.28+1.9)/3*1000/6.895 DCSEfail = 0.75012 psi

PAGE 87

75 Table D6: Coar se 1 at 15 o C Healing Specimens Specimen: t18a c1 Time d H d H norm d H /d 0 (seconds) (inches) (inches) 8 0.000373 1. 0.000378 1.0000 120 0.000379 0.000384 1.0152 240 0.000383 0.000388 1.0264 360 0.000387 0.000392 1.0366 480 0.000390 0 .000395 1.0442 600 0.000396 0.000402 1.0621 720 0.000400 0.000406 1.0728 840 0.000407 0.000413 1.0912 960 0.000410 0.000416 1.0992 1188 0.000395 0.000401 1.0599 1372 0.000396 0.000401 1.0608 1634 0.000394 0.000400 1.0567 2764 0.000386 0.000392 1.0357 4775 0.000377 0.000383 1.0112 1. d 0 = 0.000373 in., Actual d H = 0.000336 in. Specimen: t18c c1 Time d H d H norm d H /d 0 (seconds) (inches) (inches) 9 0.000360 2. 0.000355 1.0000 120 0.000367 0.000361 1.0185 240 0.000371 0.000366 1.0310 360 0.000372 0.000366 1.03 19 480 0.000377 0.000372 1.0481 600 0.000385 0.000379 1.0685 720 0.000391 0.000385 1.0861 845 0.000394 0.000388 1.0935 965 0.000391 0.000385 1.0861 1169 0.000384 0.000378 1.0662 1348 0.000389 0.000383 1.0796 1575 0.000384 0.000379 1 .0671 3000 0.000376 0.000370 1.0431 4800 0.000371 0.000365 1.0296 2. d 0 = 0.000360 in., Actual d H = 0.000333 in. Average thickness = 1.025 in. Load = 700 lb T18a c1 thickness = 1.04 in. T18c c1 thickness = 1.01 in. Normalizing formula for thickness = d H t i /t avgi P avg /P i

PAGE 88

76 Table D7: Coarse 1 at 15 o C Healing Calculations Average Average Average di/do DCSE/ DCSE/ Relative Abs. Time di/do w/o st. hard. DCSE DCSEfail DCSEapplied Healing 0 1.000 0.0000 0.0000 0.0000 120 1.017 0.0087 0.0104 0.1200 240 1.029 0.0173 0.0207 0.2400 360 1.034 0.0260 0.0311 0.3600 480 1.046 0.0347 0.0414 0.4800 600 1.065 0.0433 0.0518 0.6000 720 1.079 0.0520 0.0622 0.7200 842.5 1.092 0.0608 0.0727 0.8425 962.5 1.093 0.0695 0.0831 0.9625 1000 1.102 0.0722 0.0863 1.0000 0.00000 1178.5 1.063 1.08956 0.12163 1360 1.070 1.09672 0.05201 1604.5 1.062 1.08846 0.13234 2882 1.039 1.06592 0.35152 4787.5 1.020 1.04692 0.53630 Linear trendline equation for x = 1000, y=0.00010x+1.00207 = 1.102 Log trendline equation for x = 1000, y = 0.03439*Ln(x)+1.3131 = 1.076 Diff. between backcalculated dh/do from log. at 1000s. and linear at 1000s = 0.027 Slope of DCSE/DCSEa pplied = 9.7249 m value = 0.8 D 1 value = 8.37E 07 s FA = 73.602 psi = 2*700/(3.14*1.025*5.91) s 0 = 179.357 psi = (1.25+1.22+1.24)/3*1000/6.895 DCSEfail = 0.8363 psi

PAGE 89

77 Table D8: Fine 1 at 15 o C Healing Specimens Specimen: t11f1b Time d H d H norm d H /d 0 (seconds) (inches) (inches) 18 0.000470 1. 0.000464 1.0000 120 0.000479 0.000473 1.0195 240 0.000487 0.000481 1.0355 360 0.000493 0.000487 1.0496 480 0.000511 0.000504 1.0869 600 0.000513 0.000507 1.0922 720 0.000523 0.00 0517 1.1128 840 0.000534 0.000527 1.1355 960 0.000540 0.000533 1.1482 1187 0.000545 0.000538 1.1589 1380 0.000543 0.000536 1.1546 1740 0.000535 0.000528 1.1379 2228 0.000533 0.000526 1.1330 4628 0.000521 0.000514 1.1082 1. d 0 = 0.0004 70 in., Actual d H = 0.000413 in. Specimen: t11f1c Time d H d H norm d H /d 0 (seconds) (inches) (inches) 10 0.000373 2. 0.000375 1.0000 120 0.000381 0.000383 1.0206 240 0.000396 0.000398 1.0617 360 0.000407 0.000410 1.0916 495 0.0 00414 0.000416 1.1090 621 0.000424 0.000427 1.1367 720 0.000425 0.000427 1.1385 840 0.000441 0.000443 1.1814 960 0.000446 0.000449 1.1953 1161 0.000445 0.000448 1.1926 1575 0.000444 0.000447 1.1903 2252 0.000431 0.000433 1.1542 3405 0.000429 0.000432 1.1506 4586 0.000424 0.000427 1.1376 2. d 0 = 0.000373 in., Actual d H = 0.000323 in.

PAGE 90

78 Table D9: Fine 1 at 15 o C Healing Specimens continued Specimen: t12f1b Time d H d H norm d H /d 0 (seconds) (inches) (inches) 8 0.000370 3. 0.000372 1.0000 120 0.000379 0.000381 1.0234 240 0.000391 0.000393 1.0568 360 0.000404 0.000407 1.0923 480 0.000408 0.000411 1.1027 600 0.000420 0.000423 1.1360 720 0.000420 0.000423 1.1351 840 0.000429 0.000432 1.1604 960 0.000437 0.000440 1.1811 1240 0.000444 0.000447 1.2005 1447 0.000434 0.000436 1.1716 1676 0.000429 0.000431 1.1581 2855 0.000422 0.000424 1.1401 4690 0.000418 0.000420 1.1284 3. d 0 = 0.000370 in., Actual d H = 0.000348 in. Average thickness = 1.07 in. Load = 925 lb T11f1b thickness = 1.06 in. T11f1c thickness = 1.08 in. T12f1b thickness = 1.08 in. Normalizing formula for thickness = d H t i /t avgi P avg /P i

PAGE 91

79 Table D10: Fine 1 at 15 o C Healing Calculations Average Average Average di/do DCSE/ DCSE/ Relative Abs. Time di/do w/o st. hard. DCSE DCSEfail DCSEapplied Healing 0 1 .000 0.0000 0.0000 0.0000 120 1.021 0.0079 0.0102 0.1200 240 1.051 0.0158 0.0204 0.2400 360 1.078 0.0237 0.0306 0.3600 485 1.100 0.0320 0.0413 0.4850 607 1.122 0.0400 0.0516 0.6070 720 1.129 0.0475 0.0612 0.7200 840 1.159 0.0554 0.0714 0.8400 960 1.175 0.0633 0.0817 0.9600 1000 1.185 0.0659 0.0851 1.0000 0.00000 1196 1.184 1.18397 0.00579 1467.3 1.172 1.17219 0.06982 1889.3 1.150 1.15007 0.19015 2829.3 1.141 1.14122 0.23830 4634.7 1 .125 1.12472 0.32802 Linear trendline equation for x = 1000, y=0.00018x+1.00503 = 1.185 Log trendline equation for x = 1000, y = 0.042680*Ln(x)+1.48145 = 1.187 Diff. between backcalculated dh/do from log. at 1000s. and linear at 1000s = 0 Slope of DCSE/DCSEapplied = 5.4389 m value = 0.6027 D 1 value = 1.57E 06 s FA = 93.169 psi =2*925/(3.14*1.07*5.91) s 0 = 197.728 psi = (1.37+1.37+1.35)/3*1000/6.895 DCSEfail = 0.77483 psi

PAGE 92

80 LIST OF REFERENCES Barth, E. J., Asphalt Science and Technology New York, Science Publishers, Inc., 699 pp., 1962. Bureau of Transportation Statistics, National Transportation Statistics (NTS) 2000, Washington, DC, Bureau of Transportation Statistics, 2000. Francken, L., Fatigue Performance or a Bituminous Road Mix Under Realistic Test Conditions, T ransportation Research Record No. 712, pp. 30 36, 1979. Honeycutt, K. E., Effect of Gradation and Other Mixture Properties on the Cracking Resistance of Asphalt Mixtures, Masters Thesis, University of Florida, 2000. Jacobs, M.M.J., Crack Growth in Asphaltic Mixes, Ph.D. Dissertation, Delft, The Netherlands, Delft Univ ersity of Technology, 1995. Kim, Y.R., S.L. Whitmoyer, and D.N. Little, Healing in Asphalt Concrete Pavements: Is it Real?, Transportation Research Record No. 1454, pp. 89 96, 1994. Little, D.N., R.L. Lytton, D. Williams, and Y.R. Kim, Propagation and Healing of Microcracks in Asphalt Concrete and Their Contributions to Fatigue, Asphalt Science and Technology ed. by Arthur M. Usmani, New York, Marcel Dekker, pp. 149 195, 1997. Lytton, R.L., J. Uzan, E.G. Fernando, R. Roque, D. Hiltunen, and S.M. Stoffels, Development and V alidation of Performance Prediction Models and Specifications for Asphalt Binders and Paving Mixes, Report SHRP A 357, Federal Highway Adminstration, Washington, DC, 1993. Monismith, C.L., J.A. Epps, and F.N. Finn, Improved Asphalt Mix Design, Proceedings, Association of Asphalt Paving Technologist s, Vol. 55, pp. 347 406, 1985. Paris, P.C. and F. Erdogan, A Critical Analysis of Crack Propagation Laws, Transactions of the ASME, Journal of Basic Engineering, Vol. 85, pp. 528 534, 1963. Roque, R., W.G. Buttlar, B.E. Ruth, M. Tia, S.W. Dickison, and B. Reid, Evaluation of SHRP Indirect Tension Tester to Mitigate Cracking in Asphalt Pavements and Overlays, Final Report to the Florida Department of Transportati on, University of Florida, Gainesville, 1997.

PAGE 93

81 Roque, R., B. Sankar, and Z. Zhang, Determination of Crack Growth Rate Parameters of Asphalt mixtures Using the Superpave IDT, Proceedings of the Association of Asphalt Paving Technologists, Vol. 68, pp. 404 433, 1999. Schapery, R.A., Correspondence Principles and a Generalized J integral for Large Deformation and Fracture Analysis of Viscoelastic Media, International Journal of Fracture, Vol. 25, pp. 195 223, 1984. van Dijk, W., Practical Fatigue Charact erization of Bituminous Mixtures, Proceedings of the Association of Asphalt Paving Technologists, Vol. 38, pp. 423 456, 1969. Zhang, Z., Identification of Suitable Crack Growth Law for Asphalt Mixtures Using the Superpave Indirect Tensile Test (IDT), P h.D. Dissertation, University of Florida, Gainesville, 2000. Zhang, Z., R. Roque, B. Birgisson, and B. Sangpetngam, Identification and Verification of a Suitable Crack Growth Law, Proceedings of the Association of Asphalt Paving Technologists, Vol. 70, 2001 (in press).

PAGE 94

82 BIOGRAPHICAL SKETCH Thomas Paul Grant was born in Biloxi, Mississippi, on January 29, 1976, to Steven Bain and Catherine Ware Grant. He moved around the world due to his fathers assignments in the United States Air Force. He completed his secondary edu cation at London Central High School in the United Kingdom as valedictorian. Thomas enrolled at the University of Florida in 1994 and received a Bachelor of Science degree in c ivil e ngineering in December 1999, graduating summa cum laude. During that time he also completed a three semester co op with Robert Bates & Associates, Inc., in Jacksonville, Florida. Thomas passed the Engineer Intern exam in 1999. Thomas started working on his Master of Engineering in civil engineering in January 2000 at the Univ ersity of Florida. He is now working for Kimley Horn and Associates, Inc., in Ocala, Florida, as a traffic engineer.


Permanent Link: http://ufdc.ufl.edu/UFE0000321/00001

Material Information

Title: Determination of Asphalt Mixture Healing Rate Using the Superpave Indirect Tensile Test
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000321:00001

Permanent Link: http://ufdc.ufl.edu/UFE0000321/00001

Material Information

Title: Determination of Asphalt Mixture Healing Rate Using the Superpave Indirect Tensile Test
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000321:00001


This item has the following downloads:


Full Text











DETERMINATION OF ASPHALT MIXTURE HEALING RATE USING THE
SUPERPAVE INDIRECT TENSILE TEST

















By

THOMAS PAUL GRANT


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF ENGINEERING

UNIVERSITY OF FLORIDA


2001















ACKNOWLEDGMENTS

The two years of work shown in this thesis could not have been accomplished

without the support of many individuals. I have been immersed in the complexities of

asphalt research, and the many classes, discussions, laboratory testing and analysis have

widened my scope of knowledge. I gratefully acknowledge my advisor and chairman of

my supervisory committee, Professor Reynaldo Roque, for the time and effort he gave to

make this project successful. Over many discussions, he guided my testing and molded

my results into a thesis that lays the groundwork for much future research. Thanks go to

Professor Mang Tia and Assistant Professor Bjom Birgisson for being part of my

supervisory committee. My fellow graduate students were of great assistance. These

include Shirley Zhang, Karina Honeycutt, Boonchai Sangpengam, Booil Kim, Christos

Drakos, Bensa Nukumya, D.J. Swan, Mike Wagoner, Jeff Frank, Oscar Garcia and Paola

Ariza. Our lab guru, George Lopp, was invaluable in keeping equipment operational and

a big all around help. I thank my parents for their support over the years and helping

with my schooling until grad school could help with the tab. And finally, a special thank

you goes to my wonderful wife, Jamie, for supporting and standing behind me in my

pursuit of a graduate degree.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ......... ..................................................................................... ii

L IST O F TA B L E S .... ...... .................................................. .. .. .... .. ............ vi

LIST OF FIGURES ................................. .. .. ................. .......... viii

A B STR A C T ................................................... ....... ............ ................. xi

CHAPTERS

1 IN TR O D U C T IO N ....................... ........................... .......... ..............

1.1 B background ........................................ .................... ....................... ......... 1
1.2 O bjectiv es............................... .............. ...... 2
1 .3 S c o p e ............... ............................ ............................................................................ 3
1.4 R research A approach ................................................... 3

2 LITERA TU RE REV IEW ....................................................... 5

2 .1 Introdu action ........................................... ................................................. 5
2.2 Traditional Fatigue Approach....................................... .............. 5
2.2.1 Strain-B asked ....................... ......... .......................... 5
2.2.2 Energy-B asked ................ .. .................................... .............. 6
2.3 Fracture Mechanics Approach ................................... ........ 6
2.4 Crack Growth M odel ........................................ 8
2.4.1 Threshold C oncept........................................ 8
2.4.2 D issipated Creep Strain Energy ........................................... ................... 8
2 .5 H e a lin g ................................................................................................... 1 0

3 MATERIALS AND METHODS .....................................................................12

3.1 Introduction ............................ .............. ...... 12
3.2 A ggregates ........................................ 12
3.3 M material H handling ....................... ................................. 13
3 .3 .1 G rad atio n s ........................................................................ 13
3.3.2 Preparation ................................................................................................. 14
3.4 M mixture Production ........................................... .......... ............................ 15
3.4.1 Preparation ................................................................................................. 15









3 .4 .2 C om p action .............................. .................. .. ........ ............ .. ................ 15
3 .5 V olu m etric P rop erties ............................................................................. ...... ...... 16
3.6 T testing Preparation ............................................................ ..... .............. 16
3 .6 .1 S licin g ........................................ 16
3 .6 .2 P rep aratio n ..................................................................... 17
3.7 M mixture Testing................. .................. ............ .. ....... ........ .............. 19
3.7.1 Resilient Modulus, Creep Compliance and Tensile Strength Tests ............... 21
3.7.2 F fracture T ests .......... ....................................................... .............. .......... .. 22
3.7 .2 .1 T w o-tem perature tests........................................................................ ... 23
3.7.2.2 H dealing tests .................. ................................. .... .. .......... 23
3.8 D ata R education .................. ................................ ...... ............ .. 23

4 FIND IN G S AN D AN ALY SIS ............................................... ............................ 25

4.1 Introduction..................................... 25
4.2 Fracture Testing and Analysis.................................................. 25
4.2.1 Determination of Temperatures and Loadings.............................................. 25
4.2.2 Determination of Initial Deformation.................................................... 26
4.2.3 Determ nation of Failure Lim it................................ ............................... 26
4.2.4 Explanation of Expectations ........................................ ....................... 27
4.3 D eterm nation of H dealing R ate .................................... ...................................... 28
4.3.1 Determination of Dissipated Creep Strain Energy........................................ 34
4.3.2 D CSE/D CSEapplied vs. 6H/60 ............................................. ................... 35
4.3.3 Steric H gardening Effects .............................. .. ............ .............. 36
4.4 Damage Accumulation and Healing Across Temperatures .............. ................. 37
4 .4 .1 Stress V alu es ................. ....................................................... ......... 3 8
4.4.2 D am age Translation Process................................. ............... ... ................. 41
4.5 Perm anent D eform ation.................................................... .......................... 43
4.6 Summary of Findings and Analyses ................. ......................... 44

5 SUMMARY, CONCLUSIONS AND RECOMMENDATIONS..............................46

5 .1 Su m m ary of F in ding s ............................................................................................ 4 6
5.2 C onclu sions .................................................................................. ............... 46
5.3 Recom m endations .............. .. .............. .. .. .............. ...... .. .......... .. 47

APPENDICES

A DATA FROM TESTED SPECIMENS ............................................. ...............48

B SUMMARY OF MIXTURES' TEST RESULTS..................................................55

C FR A C TU R E TE ST FIG U R E S ....................................... ........... ......... .....................58

D HEALING TEST DATA ..................... ........... ........................ ...............69










L IST O F R E FE R E N C E S ........................................................................ ......................80

BIOGRAPHICAL SKETCH .............................................................. ..................82

























































V
















LIST OF TABLES



Table Page

3.1: Blend Proportions ............. ................. .. ........ ........... ........... ........14

3.2: Job-M ix-Form ulas...... .............................................. .......... .... ..... .. ........... 14

3 .3 : F fracture T ests L oadings ......................................................................... ...................23

4.1: D CSE to Failure for Each Specim en ........................................ ......................... 35

4.2: Recovered 6H/60.... ............. ...................................... ...............42

4.3: Comparison of Calculated and Tested 6H/60 ............................................ 42

Al: Healing Fracture Tests for Cl at 15C ...... ............ ......... .......... ....... 49

A2: H dealing Fracture Tests for Fl at 15 C ........................................ ........................ 49

A3: H dealing Fracture Tests for C at 10 C ..................................................................50

A4: H dealing Fracture Tests for Fl at 10 C ............................... .................................. 50

A5: Fracture Tests to Failure for Cl at 150C ............ .....................................51

A6: Fracture Tests to Failure for Fl at 15C ........... ............ ............... 51

A7: Fracture Tests to Failure for Cl at 100C .................. ......... ...................52

A8: Fracture Tests to Failure for Fl at 100C .................. ......... ..................52

A9: Two-Temperature Fracture Tests to Failure for C .................. .............................. 53

A10: Two-Temperature Fracture Tests to Failure for Fl ................................................53

Al1: MR, Creep and Strength Tests for Cl at 150C ........... ..............................54

A12: M R, Creep and Strength Tests for Fl at 15C .................................... ............... 54

B : Resilient M odulus Test Data .............. ............................. ................. ...56









B 2 : T ensile Strength T est D ata ...................................................................... .................. 56

B 3 : C re ep T e st D ata ......................................................................................................... 5 7

B4: Dissipated Creep Strain Energy Calculation........................................ ............... 57

D l: Coarse 1 at 10C H dealing Specim ens......... ........................................ ............... 70

D 2: Coarse 1 at 10C H dealing Calculations ........................................ ....................... 71

D3: Fine 1 at 10C Healing Specimens............................ ................. ..................72

D4: Fine 1 at 10C Healing Specimens continued........... ...................................... 73

D 5: Fine 1 at 10C H dealing Calculations ........................................ .......................... 74

D6: Coarse 1 at 150C Healing Specimens .................................... ............... 75

D 7: Coarse 1 at 150C H dealing Calculations ........................................ ....................... 76

D 8: Fine 1 at 150C Healing Specim ens ................................................... ..................77

D9: Fine 1 at 150C Healing Specimens continued........... ...................................... 78

D 10: Fine 1 at 150C H dealing Calculations ........................................ ......................... 79
















LIST OF FIGURES



Figure Page

2.1: Fatigue Crack G row th B ehavior.......................................................... ............... 7

2.2: Dissipated Creep Strain Energy ........... ............. ..... ...............9

3.1: Cl and Fl Gradations (12.5 mm Nominal Size).................................. ............... 13

3.2: Pine Model Superpave Gyratory Compactor.......................... ....... ... ............ 16

3.3: Diamond Pacific Cutting Saw with Specimen Holder ..............................................17

3.4: Aluminum Template for Drilling 8 mm Hole.................................... ............... 18

3.5: Gage Point Placement Template with Vacuum Pump ............................... ...............19

3 .6 : L V D T Setup .............................................................................20

3 .7 : T testing Setup ............................................................................ 20

4.1: Resilient Horizontal Deformation vs. Number of Load Replications.............................27

4.2: R esilient D eform ation vs. Tim e ............................................. ............................. 29

4 .3 : 6 H/6 0 v s. T im e ......................................................................... 3 0

4.4: Average 6H/60 (Modified for Steric Hardening) vs. Time ...........................................31

4.5: D C SE /D C SE applied vs. 6H/60 ............................................................... .................. ......31

4.6: Healing vs. Time after 1000 Cycles of Loading......................................................32

4.7: Comparison of Healing vs. Time after 1000 Cycles of Loading .............. ............... 33

4.8: Comparison of Time to Full Healing .................................................................. 33

4.9: Com prisons of H dealing R ates............................................... ............................. 34

4.10: Comparison of DCSE/DCSEapplied vs. H/0 .............. ........................ ...............36









4.11: 6H/60 vs. Cycles for Initial Loading at 15 C ...................................... ............... 39

4.12: 6H/60 vs. Number of Load Replications for C1 ......................................................39

4.13: 6H/60 vs. Number of Load Replications for F1 ......................................................40

4.14: DCSEapplied(using oo)/DCSEfailure vs. Mixture Type and Temperature.............................40

4.15: 6H/60 vs. D C SE /D C SE failure ................................................................ ............... 41

4.16: Perm anent D eform ation vs. Tim e ........................................ ........................... 44

Cl: Resilient Deformation vs. Load Replications (Cl at 150C to failure).............................59

C2: Resilient Deformation vs. Load Replications (Fl at 150C to failure)............................59

C3: Resilient Deformation vs. Load Replications (C1 at 10C to failure)..........................60

C4: Resilient Deformation vs. Load Replications (Fl at 10C to failure)..........................60

C5: Resilient Deformation vs. Time (Cl at 150C healing test)........................................61

C6: Resilient Deformation vs. Time (Fl at 150C healing test) ........................................61

C7: Resilient Deformation vs. Time (Cl at 10C healing test)....................... ...........62

C8: Resilient Deformation vs. Time (Fl at 10C healing test) ........................................62

C9: Average 6H/60 vs. Time (Cl at 150C healing test) .................................. ............... 63

C10: Average 6H/60 vs. Tim e (Fl at 150C healing test) ......................... ............... ......63

C11: Average 6H/60 vs. Time (Cl at 10C healing test) ................................ ............... 64

C12: Average 6H/60 vs. Time (Fl at 10C healing test) ....... ....... ............ 64

C13: Average 6H/60 without Steric Hardening vs. Time (C at 150C healing)...................65

C14: Average 6H/60 without Steric Hardening vs. Time (Fl at 15C healing) ....................65

C 15: Average 6H/60 without Steric Hardening vs. Time (C at 10C healing)...................66

C16: Average 6H/60 without Steric Hardening vs. Time (Fl at 10C healing) ....................66

C17: Permanent Deformation vs. Time (Cl at 150C healing test)......................................67

C18: Permanent Deformation vs. Time (Fl at 150C healing test) ......................................67









C19: Permanent Deformation vs. Time (Cl at 10C healing test)......................................68

C20: Permanent Deformation vs. Time (Fl at 10C healing test) ......................................68















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Engineering

DETERMINATION OF ASPHALT MIXTURE HEALING RATE USING THE
SUPERPAVE INDIRECT TENSILE TEST

By

Thomas Paul Grant

December 2001


Chairman: Dr. Reynaldo Roque
Cochair: Dr. Bjorn Birgisson
Major Department: Civil and Coastal Engineering

Cracking is one of the major failure modes in asphalt concrete. Microcracks form

within the asphalt concrete during repetitive loading. Eventually these microcracks will

join together to form macrocracks, which result in permanent damage to the pavement.

Healing occurs due to rest periods and temperature increases while in the microcracking

range. Healing increases the useful life of the pavement.

Asphalt concrete specimens were prepared in the laboratory and tested using the

Superpave Indirect Tensile Test. Two Superpave mixtures were prepared: a fine blend

and a coarse blend. Tests were conducted at 150C and 100C in an environmental

chamber using hydraulic loading equipment. Repetitive loading tests applied 0.1-second

haversine loads followed by 0.9-second rest periods. Damage accumulation was

measured using the normalized resilient modulus.









A method was developed to determine healing rate in terms of the recovered

dissipated creep strain energy by converting changes in normalized resilient modulus to

changes in energy. A crack growth law and model developed at the University of Florida

uses dissipated creep strain energy as a measure of damage. With the work from this

thesis, healing can now be incorporated into the model.

Tests were also conducted at two temperatures using one specimen to analyze

damage accumulation and translation across temperatures. Specimens were subjected to

1000 load repetitions at 150C, cooled to 100C for one hour and tested again. A procedure

was devised to relate damage incurred at one temperature to equivalent damage at other

temperatures.

A number of findings and conclusions were drawn from this research. The

relationship between the normalized resilient deformation and dissipated creep strain

energy provided the ability to measure healing in terms of energy. Steric hardening

appeared to play a significant role in a mixture's response during the loading and healing

portions of laboratory tests. Healing occurred much faster at 150C than 100C and the

coarse mixture healed much faster than the fine mixture at 150C. Recovered strains do

not appear to be an adequate measure of healing. Healing appears to play an important

role in field performance.














CHAPTER 1
INTRODUCTION



1.1 Background

There are more than four million miles of roadways in the United States of

America. Ninety-six percent of the 2.42 million miles of paved public roadways are

constructed of asphalt concrete. Currently, more than 130 million cars and 7.7 million

heavy trucks compete for space on the roads. Functional roadways are vitally important

to the economy to allow for speedy transport of people and freight. When an asphalt

roadway is in disrepair, it results in a drain on resources to repair automobiles and the

roadway. It also results in delays due to defensive driving and construction. In 1998,

only 43.4% of urban interstate highways had a rating of GOOD or VERY GOOD (NTS,

2000). If better roadways are created that are more resistant to failure, the resulting time

and cost savings would be tremendous.

The major failure modes of asphalt concrete (AC) are cracking, distortion, and

disintegration. Cracking occurs for a number of reasons, but this thesis focuses on

fatigue cracking, which is cracking due to repetitive loading until the fatigue life of the

pavement is reached. When loads are applied to a flexible pavement, tensile stresses

arise at the bottom of the pavement. Numerous microcracks form within the asphalt

concrete. If an AC pavement is loaded until fatigue failure begins, the microcracks will

join together to form macrocracks, which result in permanent damage to the pavement.

According to Kim et al.,









When an asphalt concrete pavement is subjected to repetitive applications
of multi-level vehicular loads and various durations of rest periods, three
major mechanisms take place: fatigue, which can be regarded as damage-
accumulation during loading; time-dependent behavior related to the
viscoelastic nature of asphalt concrete; and chemical healing across
microcrack and macrocrack faces during rest periods. (1994, pg. 89)

The healing that occurs is due to rest periods and temperature increases.

This thesis analyzes the healing that occurs after a period of continuous loading

and explains the process for determination of a healing rate. Another aspect covered is

damage accumulation across temperatures and a way to compare equivalent damage at

different temperatures.

A recently developed crack growth model at the University of Florida uses

dissipated creep strain energy (DCSE) to measure damage accumulation in gyratory

specimens. The healing rates developed in this thesis are in terms of DCSE and can

improve the crack growth model with incorporation into the model.



1.2 Objectives

The main objectives of this research are summarized below:

Determine the necessary loadings to promote macrocracking in the specimens
using the Superpave Indirect Tensile Test (IDT) in a certain time period at 10C
and 15C.

Develop a procedure to identify and quantify microcrack healing.

Evaluate and if possible quantify how microdamage is translated across
temperatures.

Develop a formula of healing with respect to time for each mixture at each
temperature that can be compared and used.


Identify a relationship to express time-dependent microcrack healing in terms of
recovered dissipated creep strain energy.









1.3 Scope

This study focuses on the microcrack healing of asphalt concrete during rest

periods and the translation of accumulated damage across temperatures. Healing was

monitored on two Florida Department of Transportation (FDOT) Superpave mixtures: a

fine blend and a coarse blend. Tests were performed at 10C and 15C. Gyratory

specimens were prepared with short-term aging and compacted to 7% air voids. The

asphalt content was controlled to achieve 4% air voids at N design. Specimens

accumulated damage due to repeated loadings using the Superpave Indirect Tensile Test

(IDT). Tests were performed to microdamage the specimens, but the loads were stopped

before macrocracking began. The healing of the specimen was measured after the

loading was stopped. Additional tests were performed to determine mixture properties,

including resilient modulus, tensile strength, failure strain, fracture energy m-value, and

Di-value.



1.4 Research Approach

The research was performed in five phases:

1. Literature Review: Examination of existing research on fracture mechanisms,
healing, and measurement of crack growth in asphalt pavements.

2. Laboratory Mixtures: Preparation of six-inch diameter specimens, including
sieving, watching, mixing, compacting, aging, cutting to one-inch and two-inch
thicknesses, determining air voids, dehumidifying, and placement of gage points.

3. Testing: Cooling of specimens to requisite temperature and completion of
resilient modulus, creep compliance, tensile strength, and repetitive loading tests.

4. Data Analysis: Evaluation of each test to produce figures showing healing
effects, damage accumulation, and crack initiation, propagation, and
disintegration.






4


5. Development of a healing rate based on dissipated creep strain energy. Use
healing rate for comparison and later incorporation into HMA crack growth
model.














CHAPTER 2
LITERATURE REVIEW



2.1 Introduction

The purpose for this chapter is to examine published literature concerning healing,

fatigue damage and crack growth in asphalt pavements. Fatigue is the wearing down of

an asphalt pavement caused by repeated loads below the tensile strength of the pavement.

Without healing, fatigue eventually results in failure, or permanent cracking, of the

pavement. This section reviews the literature addressing the occurrence and

quantification of healing, work performed to describe the mechanisms of crack growth,

and ways of measuring fatigue damage.



2.2 Traditional Fatigue Approach

The Traditional Fatigue Approach assumes that damage occurs in a specimen

from repetitive loading that leads to eventual fatigue failure of the specimen. The method

can be based on stress, strain or energy.

2.2.1 Strain-Based

An oft-used strain-based equation developed by Monismith et al. (1985)

Nf = K (1/st)a (1/Smix)b

provides the number of cycles to failure (Nf) where K is a mix-related factor based on

asphalt content and degree of compaction, ;t is the tensile strain, a and b are coefficients

based on beam fatigue tests and Smix is the stiffness of the mixture.









Many other equations have been developed using various parameters dependent

on the asphalt mixture composition and temperature. None of the fatigue relationships

that have been used are all encompassing. Many factors affect fatigue and the various

formulae try to predict the number or load replications until cracking begins. Francken

(1979) showed that fatigue relationships are dependent on rest periods and that longer

rest periods result in longer fatigue life. From this, one can deduce that healing of the

specimen occurs with longer rest periods.

2.2.2 Energy-Based

Fatigue has also been predicted using the cumulative dissipated energy from

repetitive loading. A common energy-based equation by van Dijk (1969) is

Wf = BfNf

where Wf is the cumulative dissipated strain energy per volume (J/m3), Bf and z are

mixture coefficients, and Nf is the number of cycles to fatigue failure. As loading occurs,

energy within the viscoelastic material is dissipated until the failure limit is reached.

Miner's Rule simply states that damage is the ratio of number of cycles to number

of cycles to failure. This does not account for healing effects from rest periods. The

Traditional Fatigue Approach can add damage but not subtract it. The results are not

fundamental because the approach is dependent on mode of loading. Can this approach

explain the complexities of asphalt pavements? How does damage translate across

temperatures?



2.3 Fracture Mechanics Approach

The Fracture Mechanics Approach assumes there are inherent flaws in the

material. When loading occurs, there are higher stress concentrations around the flaws










because load is distributed over a smaller area, which means the material no longer has a

uniform stress distribution. Fracture mechanics describes the propagation of cracks

through materials. Paris and Erdogan (1963) developed a crack rate law for use in linear

elastic homogeneous materials. Paris' Law is defined as

da/dN = A (AK)n

where a is the crack length, N is the number of load repetitions, K is the stress intensity

factor, and A and n are material parameters. The stages of crack growth are shown in

Figure 2.1. Paris' law is only applicable in the propagation phase.





A Kcntical



z Initiation Propagation Disintegration
Phase Phase Phase








AKthreshold
log AK


Figure 2.1: Fatigue Crack Growth Behavior (After Jacobs, 1995)



Asphalt concrete is a heterogeneous viscoelastic material, so Paris' Law may not

completely explain how asphalt mixtures behave. For example, cracks are known to

grow discontinuously in asphalt concrete mixtures (Zhang, 2000). Paris' Law does not

explain crack initiation because it assumes cracks are already present. The parameters A

and n are a function of loading condition. When looking at da/dN from laboratory









testing, results do not correspond to actual field performance of different mixtures

(Zhang, 2000). Despite the Fracture Mechanics Approach's flaws, crack length is a

measurable interpretation of damage, which provides a sound base from which to

proceed.



2.4 Crack Growth Model

The two approaches to measuring fatigue damage thus far presented are unable by

themselves to accurately explain fatigue in asphalt concrete. Zhang et al. (2001) created

a crack growth law based on viscoelastic fracture mechanics with the addition of a

threshold concept. Zhang's paper further explains the use of this crack growth law to

create a pavement cracking prediction model.

2.4.1 Threshold Concept

Paris' Law assumes the crack advances when any stress is applied. Asphalt

pavements can perform well in the field for many years without developing visible

macrocracks. Zhang et al. (2001) proposed that if a certain threshold was not exceeded

for a given mixture during repeated loadings, that no macrocrack would form. The

microcracks could heal themselves with the help of temperature increases and/or rest

periods. From laboratory testing by Zhang and Honeycutt, it was shown that if a test

specimen was tested and loading was stopped before the threshold was reached, full

healing occurred with rest and temperature increases. It was also shown that once the

threshold was exceeded (i.e., once macrocracks developed), no healing was achieved

with rest and temperature increases (Zhang, 2000).

2.4.2 Dissipated Creep Strain Energy

Zhang et al. (2001) determined that dissipated creep strain energy (DCSE) to

failure appears to be a fundamental parameter that serves as an appropriate threshold










between microcracks and macrocracks. DCSE is defined as the fracture energy, FE,

minus the elastic energy, EE (see Figure 2.2). From the strength test, failure strain (sf),

tensile strength (St) and fracture energy can be determined. From the resilient modulus

test, the resilient modulus (MR) value can be found (see Figure 2.2).

0s = (MR* Sf St)/ MR

EE = /2 St (gf- go)

DCSE = FE EE


Fracture Energy, FE = DCSE + EE


Elastic
Energy (EE)


Strain o Ef


Figure 2.2: Dissipated Creep Strain Energy (After Zhang et al., 2001)


Zhang also showed that dissipated creep strain energy appears to be independent

of mode of loading. This means the strength test results can be used to evaluate the

repetitive loading test results. If the DCSE limit is not reached, macrodamage will not

occur.









In addition, Zhang developed a method to determine DCSE/cycle, so that the

DCSE at any given time during loading could be calculated. The formula for a 0.1-

second haversine load is

DCSE/cycle = 1/20 02 D1 m (100)-1

where o is the stress and D1 and m are determined from creep tests. This process uses an

energy-based approach to fracture mechanics with a threshold boundary.

A pavement cracking prediction model based on Zhang's crack growth law is

currently being developed at the University of Florida. To make the model complete,

healing needs to be incorporated in terms of recovered DCSE.



2.5 Healing

The occurrence of microcrack healing in polymers is well documented. Asphalt

binder contains hydrocarbon polymers, but only in recent years has any attempt been

made to quantify healing in asphalt pavements. Little et al. (1997) found that long-chain

aliphatic molecules in binders are directly proportional to healing effects. The Western

Research Institute has developed a microstructural asphalt cement model that exhibits the

molecular structure as continuously reforming, thus incorporating healing effects.

For many years, it has been known that identical mixtures perform better in the

field than in laboratory testing. The accelerated process of laboratory testing does not

adequately mimic field performance. Because of this, a shift factor (SF) was developed

in 1977 using AASHTO Road Test data. A shift factor developed by Lytton et al. (1993)

is


SF = 1 + ni/No a (t/to)h









where ri is the number of rest periods, N is the number of loading cycles to failure

without rest periods in cyclic laboratory tests, 1 is the length of the period between field

load applications, to is the length of the rest period between loading cycles in laboratory

tests, and a and h are regression values.

Kim et al. (1994) measured healing in a laboratory setting and in the field. For

the laboratory testing, Kim et al. used Schapery's (1984) nonlinear-viscoelastic

correspondence principle (CP). For the correspondence principle case, a cyclic loading

with rest periods was administered. A pseudo strain was used to differentiate between

the relaxation effect and healing. The appropriate pseudo strain formula as developed by

Schapery (1984) is

SR= 1/ER ft E (t-'c) ds/dr' dr

where s is uniaxial strain, sR is pseudo strain, ER is reference modulus, and E(t) is the

uniaxial relaxation modulus. Kim et al. transformed a time-dependent issue into a time-

independent issue and calculated the recovered pseudostrain energy from rest periods.

For the field-testing, Kim et al. (1994) used stress-wave testing. A section of

roadway was closed to traffic and temperature sensors were placed in the pavement.

From hourly stress-wave tests for 24 hours, wavespeeds were obtained and the change in

elastic modulus was calculated as a function of temperature. The elastic modulus of

asphalt pavement increased at a given temperature over the 24-hour rest period. This

increase is conjectured to be microcrack healing (Kim et al., 1994).














CHAPTER 3
MATERIALS AND METHODS



3.1 Introduction

This chapter explains the materials and processes used to create the gyratory AC

specimens suitable for testing. The testing methods and data reduction are described.

Relevant figures are included, with most of the test results in the appendices. The same

asphalt binder type AC 30 (Coastal) was used for each mixture. The appendices provide

detailed information on material properties and fracture test results. Honeycutt (2000)

produced identical mixtures to the ones used in this research project and additional

information on the aggregate properties, gradation and mixture design are provided in

Honeycutt's thesis.



3.2 Aggregates

The aggregate used for this study was Miami oolite limestone produced by

Whiterock. Limestone is a relatively soft, porous aggregate that is commonly used in

asphalt paving in Florida because of its abundance. The aggregate components provided

were a coarse aggregate, S1A; a fine aggregate, S1B; and screenings. Granite mineral

filler from Georgia was also used in the mixtures.














3.3 Material Handling

3.3.1 Gradations

Two mixture designs provided by the FDOT Materials Office were used for this


research. Both mix designs passed all Superpave criteria. The coarse blend, C1, and the

fine blend, Fl, passed below and above the Superpave restricted zone, respectively and


passed between the control points. Figure 3.1 shows the gradation chart of the mixtures

and the restricted zone and control points. The blend proportions of the aggregates in

each mixture are shown in Table 3.1. The Job-Mix-Formulas are provided in Table 3.2.





100
Control Points
- -Restricted Zone
80- 1- Coarse 1
A Fine 1

F60

(-



0.15
20 4
0.075 0 Actual Sieve Size

0.3 0.6 1.18 2.36 4.75 9.5 12.5 19.5

Sieve Size^0.45 (mm)


Figure 3.1: C1 and Fl Gradations (12.5 mm Nominal Size)














Blend Proportions


S1A S1B Screenings Filler
(%) (%) (%) (%)
C1 10.20 63.27 25.51 1.02
F1 20.30 25.37 53.29 1.03

Bulk Specific Gravity 2.43 2.45 2.53 2.69


Table 3.2: Job-Mix-Formulas

Sieve Size C1 F1
(mm)
25 (1") 100 100
19 (3/4") 100 100
12.5 (1/2") 97.4 95.5
9.5 (3/8") 90 85.1
4.75 (#4) 60.2 69.3
2.36 (#8) 33.1 52.7
1.18 (#16) 20.3 34
0.6 (#30) 14.7 22.9
0.3 (#50) 10.8 15.3
0.15 (#100) 7.6 9.6
0.075 (#200) 4.8 4.8


3.3.2 Preparation

The aggregates were dried and sieved to each individual sieve size. This was

done to assure that the samples represent the true gradation of the mixtures, which may

not be representative in small quantities. The aggregates were then batched out in the

appropriate quantities to produce 4500 g samples.


Table 3.1:









3.4 Mixture Production

3.4.1 Preparation

The 4500-gram aggregate batches, asphalt binder and mixing equipment were

heated for three hours at 1500C (300F) to achieve appropriate uniform mixing

temperature. The batches were then mixed with the proper amount of asphalt binder and

heated for another two hours at 1350C (275F) for short-term aging. This aging

represents the aging that occurs in the field between mixing and placement and allows for

absorption of the asphalt binder into the aggregate pores. After one hour of aging, the

mixture is stirred to prevent the outside of the mixture from aging more than the inner

because of increased air exposure. During the aging, the compaction equipment was also

heated to assure uniform temperature for compaction.

3.4.2 Compaction

The mixtures were compacted into 150 mm diameter specimens using the Pine

model Superpave Gyratory Compactor (see Figure 3.2). The specimens were compacted

to 7% air voids for testing. This is typical of the air void percentage in mixtures when

they are placed in the field. After compaction by vehicular traffic over time, the roadway

reaches its optimum 4% air void level. Lower air void levels result in bleeding, while

higher air void levels may result in traveling.

The mixtures can be compacted by number of gyrations or by height.

Compacting based on number of gyrations to control density results in large variances in

air void percentages. Because of the different sizes of aggregate in each sieve size, each

replicated mixture has a different total volume of aggregate, while the total mass is the

same. Compacting based on height results in more uniform air void percentages for the

test specimens, so this method was used for compaction.






























Figure 3.2: Pine Model Superpave Gyratory Compactor


3.5 Volumetric Properties

After compaction and cooling of the specimens, bulk specific gravities of the

specimens were taken. Rice tests were performed in accordance with ASTM D 2041 to

attain the theoretical maximum densities on 1000-gram uncompacted test mixtures

(Honeycutt, 2000). From these two tests, the air void percentage was determined.



3.6 Testing Preparation

3.6.1 Slicing

The 150 mm diameter gyratory specimens are tested in one-inch and two-inch

thicknesses using the Indirect Tensile Tests (IDT). A Diamond Pacific cutting saw (see

Figure 3.3) with a special attachment to hold the gyratory specimens was used to achieve

the desired thicknesses. Two two-inch test specimens or three one-inch test specimens









were attained from each gyratory specimen. Because the saw uses water to keep the

blade wet, the test specimens were dried for two days at room temperature to achieve the

natural moisture content after cutting.


Figure 3.3: Diamond Pacific Cutting Saw with Specimen Holder


3.6.2 Preparation

The bulk specific gravity and air void content were determined for each test

specimen. Gyratory specimens had to be in the range of 7 + 0.5 % air voids to be

considered for testing. Air void percentages for the test specimens used are provided in

Appendix A. Test specimens were placed in a low humidity chamber for two days to

negate moisture effects in testing. The two-inch thick test specimens were used for the

Resilient Modulus, Creep Compliance and Indirect Tensile Strength tests. The one-inch

thick test specimens were used for the fracture tests. The fracture test specimens required

the drilling of an eight-mm diameter hole in the center to create a focal point from which









the crack would propagate. An aluminum template (see Figure 3.4) was used to align the

specimen and hold it in place, while a drill press with a concrete drill bit was used to drill

the hole.


Figure 3.4: Aluminum Template for Drilling 8 mm Hole


Brass gage points were attached to the test specimens with a strong adhesive

using a steel template and vacuum pump setup (see Figure 3.5). Four gage points were

placed on each side of the test specimens 19 mm (0.75 in.) from the center along the

vertical and horizontal axes. A steel plate that fit over the attached gage points was used

to nmrk the loading axis with a marker. This helped with placement of the test specimen

in the testing chamber and assured proper loading of the specimen.






19























Figure 3.5: Gage Point Placement Template with Vacuum Pump




3.7 Mixture Testing

Roque et al. developed the IDT testing procedure and data reduction process used

for this research (1997, 1999). Thorough explanation of the procedures and the reasoning

behind them are given in those papers. Tests were performed using an MTS hydraulic

loading system with the Teststar IIs data acquisition system. An environmental chamber

kept the temperature constant + 0. 1C. Tests were performed at 10C and 15C.

The deformation in the test specimens was measured using linear voltage

differential transducers (LVDT) attached to the gage points (see Figure 3.6). The

LVDT's took voltage readings, which were converted to micro-inches using a signal-

conditioning unit manufactured by MGC. The testing equipment is shown in Figure 3.7.































Figure 3.6: LVDT Setup


Figure 3.7: Testing Setup









After the LVDT's were in place, the test specimen was carefully placed in the

testing chamber so that the loading occurred along the vertical plane of the test specimen.

The test specimen was then cooled for eight hours to assure constant temperature

throughout the specimen.

3.7.1 Resilient Modulus, Creep Compliance and Tensile Strength Tests

The resilient Modulus, creep compliance and tensile strength tests can all be

performed on a single specimen. A 10-lb seating load is applied to each test specimen

immediately before testing. The resilient modulus (MR) test applies 0.1-second haversine

loads followed by 0.9-second rest periods for five cycles. For this test, different loadings

are tried, starting with a small loading and slowly increasing the load, to produce a

maximum horizontal deformation between 200-300 micro-inches. Five hundred and

twelve data points are recorded during each second of testing.

Creep tests applying a constant stress to the test specimen were performed for 100

seconds. As with the MR test, small loads were used to check the amount of creep

occurring. Horizontal deformations, 6H, were kept within the approximate range of 100-

150 micro-inches at 30 seconds to prevent excessive deformation of the specimen. A

maximum horizontal deformation of 1000 [t-in was used. Acquisition rates for the creep

tests varied from 10 points per second at the beginning to one point every five seconds at

the end of the test.

The strength test loads the specimen to failure by applying a constant stroke of 50

mm per minute. An acquisition rate of 20 points per second was used. Through data

reduction, these tests provide the resilient modulus, m-value, Do and DI-values, tensile









strength, failure strain, fracture energy, and Poisson's ratio for the tested specimen at a

given temperature.

The test results used for this research project at 100C were performed by

Honeycutt (2000). This includes all results and mixture properties derived from MR,

creep, and strength tests for Fl and Cl mixtures at 10C. One difference of note is that

Honeycutt ran 1000-second creep tests, although this should have no bearing on material

properties, such as m-value.

3.7.2 Fracture Tests

The fracture tests are repetitive loading tests performed on the one-inch gyratory

specimens. A 10-lb seating load is applied to the specimen before testing begins. Much

like the MR tests, the fracture test applies 0.1-second haversine loads followed by a rest

period of 0.9 seconds. This test continues until it is manually stopped. This allows the

user to test a specimen for a certain amount of time. On screen, a constantly updated

graph of the deformations shows the user the current state of the gyratory specimen and a

rough estimate of when macrocrack propagation begins.

The load is similar to that used for the MR test, but was varied to achieve

complete disintegration of the test specimen in about one hour. Standard loads were

determined for both the Cl and Fl specimens at 10C and 15C (see Table 3.3). The user

manually controlled the data collection. This allowed more data points to be recorded

when the deformation slope began changing. Every time the data collection button was

pushed during the test, the program took an MR reading over five cycles of loading at an

arbitrary time. When the test was terminated, the testing and seating loads were removed

from the test specimen.









Table 3.3: Fracture Tests Loadings

Temperature Loading
Mixture (C) (Ib)
C1 10 1300
C1 15 700
F1 10 1300
F1 15 925


3.7.2.1 Two-temperature tests

The two-temperature fracture tests were performed after standard loadings were

determined from standard fracture tests. The tests began at 150C for 1000 cycles of

loading. The test was then stopped and the test specimen was cooled to 100C. After an

hour, the test was continued with the new appropriate loading. The test was continued

until the specimen failed.

3.7.2.2 Healing tests

The healing fracture tests loaded the specimens for 1000 cycles and then the test

was terminated. During the following hour, five MR tests were performed on the

specimen to monitor the healing. These tests were performed at 3 min., 6 min., 10 min.,

30 min. and one hour after repetitive loading had stopped. The same load was used for

the healing portion MR tests as with the fracture tests.



3.8 Data Reduction

The tests output the data files in various formats. Some alterations have to be

performed to reduce the data using a spreadsheet and program. A set of three test

specimens is required to correctly calculate the appropriate values for the MR, creep and

strength tests. An Excel spreadsheet and Fortran code have been developed by the









University of Florida's Civil Engineering Materials Department and thoroughly

explained in a report to the Florida Department of Transportation (Roque et al., 1997).

Several outputs are provided for the various tests. The MR test shows the horizontal and

vertical instantaneous and total deformations of the middle three cycles. These cycles are

used for the calculation of the resilient modulus, which is also shown on the printout.

The fracture test outputs the horizontal and vertical deformations for each data set

collected during the test. With the IR, creep, and strength tests, outputs include creep

compliance over time, Poisson's ratio, average tensile strength, m-value, Db-value, Di-

value, failure strain, and fracture energy. The results of these outputs are shown in

Appendix B along with pertinent information about tested samples in Appendix A.














CHAPTER 4
FINDINGS AND ANALYSIS



4.1 Introduction

The repetitive loading fracture test mimics a wheel rolling over a section of

asphalt pavement every second. This method of loading will eventually fail a pavement

because the only time available for healing is between each load. By loading the test

specimen for a time and then removing the load, healing can be monitored and better

understood. This chapter explains the testing results and analysis procedures used in the

determination of a healing rate. The damage accumulation across temperatures is

explained. Findings are explained in detail.



4.2 Fracture Testing and Analysis

The coarse and fine mixtures that were chosen for this research are approved

FDOT Superpave mixtures using native Florida aggregates. The two mixtures perform

differently because of interlocking effects, but are produced using the same aggregates

and binder. This prevents other factors from contributing to differences in performance

and healing. Chapter 3 details the process by which gyratory specimens were created and

tested.

4.2.1 Determination of Temperatures and Loadings

Historically at the University of Florida, repetitive loading tests were performed

at 10C. After preliminary testing, 15C was chosen as the other temperature for this









project. Loads were determined that would provide an acceptable linear portion of

microcrack growth while still totally failing within an hour. Even with the narrow range

of air voids and constant loads, each test specimen performs differently, resulting in a

range of failure limits and damage rates.

4.2.2 Determination of Initial Deformation

After analysis of a fracture test, the resulting resilient deformations can be plotted

versus time. Figure 4.1 shows a representative test taken to failure. At the beginning of

the test, the microdamage is not linear. In the past, this effect had been attributed to local

elevation of temperature in the specimen during cycling that stabilizes after about two

minutes of loading. However, based on data obtained in this investigation, the author

maintains that steric hardening may be the main force behind this effect. This will be

explained in Section 4.3. The modified initial resilient deformation, 6o, is determined by

interpolating the linear portion back to the beginning of cycling.

4.2.3 Determination of Failure Limit

As explained in the literature review, Zhang's threshold value denotes the

boundary between microcracking and macrocracking. Figure 4.1 shows that

macrocracking on that test specimen occurred at about 1300 cycles of loading. A

conservative cutoff point was to halt loading at 1000 cycles to insure macrocracking had

not developed. Thus, after 1000 cycles of loading, the specimens were still in the

microdamage range and the damage was healable. For both the damage accumulation

across temperatures tests and the healing tests, 1000 cycles of loading was used as the

standard. This cutoff point was used for both mixtures at both temperatures.

















v-
U

0
0
.0


0.001200

0.001000

0.000800

0.000600

0.00040

n nnnn --


r Microdamage Macrodamage -
0.000000
0 500 1000 1500 2000 2500

Number of Load Replications (cycles)



Figure 4.1: Resilient Horizontal Deformation vs. Number of Load Replications




4.2.4 Explanation of Expectations

Initially, the testing procedure was developed to translate damage effects across


temperatures. Tests were performed that induced a certain amount of damage at one


temperature (150C) before quickly cooling the specimen to a second temperature (10C).


The intent was to "lock" the damage in, to evaluate the effect of damage accumulated at


one temperature on the failure limit at a second temperature.


An adequate way to measure damage was needed. A change in resilient


horizontal deformation is a measure of damage. Earlier work (Honeycutt, 2000) has


shown that by normalizing the change in horizontal deformation by 60, it could be


directly related to the theoretical crack length of the specimen. From the crack length,


the da/dN, or crack growth rate, could be determined. Both the crack length and da/dN


are ways of monitoring damage. Zhang (2000) noted, "Crack growth rates obtained from









the laboratory tests do not correlate well with the field performance." One of the major

reasons for this is the absence of healing due to the lack of rest periods during laboratory

tests.

After testing and analysis of damage accumulation fracture tests, it was

determined that the damage was not locked in and partial recovery of the resilient

deformation of the specimen was seen. This recovery of resilient deformation is an

increase in stiffness and represents healing. This healing needed to be quantified, but

with the LVDT's temperature susceptibility, testing could not be performed when the

temperature in the environmental chamber was not steady. As previously mentioned, full

healing can occur with rest periods and temperature increases. It was decided that

healing tests should be done at both temperatures to look at the difference in the healing

rate of the specimens at each temperature.



4.3 Determination of Healing Rate

Development of a healing rate was an iterative process. The goal was to find a

way to measure healing in terms of DCSE. The processes explained in Section 4.2 had

already been carried out before this analysis began. Explanation of the healing testing

methods is provided in Chapter 3. One cycle of loading lasts one second and because the

loading is stopped at 1000 cycles, the following figures are shown in terms of time. The

example figures used in this section are the Cl tests at 150C. Appendix C shows figures

generated from the healing tests. Refer to Appendix D for healing test data and further

explanation of the healing rate calculations. The following step-by-step guide details the

process used for each mixture at both temperatures, with explanation provided afterward:

1. Plot Horizontal Deformation vs. Time for each specimen. See Figure 4.2.











2. Determine 60 for each specimen. See Section 4.2.2 for explanation.

3. Normalize the specimens by thickness and diameter.


4. Determine 6H/60 at each time of acquisition.


5. Average the times and 6H/60 of the specimens.

6. Determine DCSE at each time during loading. See Section 4.3.1.

7. Determine DCSE/DCSEapplied at each time. DCSEapplied is the value of DCSE
at 1000 cycles, when the loading was stopped.






0.0005
I Loading
I Healing
S0.00045
U
t-

0
0.0004 I-
00I

.2 0.00035



0.0003
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

Time (seconds)



Figure 4.2: Resilient Deformation vs. Time




8. Plot average 6H/ 0 vs. time. Place a linear trendline through the data points
collected during the loading portion. Place a logarithmic trendline through
healing data points. See Figure 4.3.
9. Backwards forecast the logarithmic trendline to ending point of loading.


10. Determine the difference between 6H/60 at 1000 cycles and backwards-
forecasted value. This represents the initial steric hardening effect (see
Section 4.3.3). The value shown for the example in Figure 4.3 is 0.0265.














1.2

I Average loading

0.0265 I Average healing

1.1
Logarithmic Trendline
y = -3.439427E-02Ln(x) + 1.313134E+00
t o 1.05= l191RrF-n
1.0 Linear Trendline
y = 0.00010x + 1.00207
1 R2 = 0.98307
1000 Cycles Healing for 1
of Loading hour
0.95


0.9
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (seconds)




Figure 4.3: 8/Ho vs. Time




11. Subtract initial steric hardening effect from each averaged 6H/60 healing data
point by adding the value determined in Step 10 to each point.


12. Re-plot average 6H/60 VS. time using new values for healing. See Figure 4.4.


13. Plot DCSE/DCSEapplied vs. 6H/60 during the loading portion. See Figure 4.5.















1.2


1.15

Logarithmic Trendline
1.1
1.1 y = -3.439427E-02Ln(x) + 1.339657E+00
R2 = 9.512166E-01

1.05
ta




SAverage loading
0.95
*Average healing

0.9-
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (seconds)




Figure 4.4: Average SH/80 (Modified for Steric Hardening) vs. Time








1.00

0.90

0.80

S0.70
S-Linear Trendline
R. 0.60
S y = 9.7249x 9.7368
(V
S0.50-
0.50 R 0.9831

0.40

i 0.30

0.20

0.10

0.00
0.9 0.95 1 1.05 1.1 1.15 1.2
B/d0


Figure 4.5: DCSE/DCSEapplied VS. 8H80











14. Find slope of DCSE/DCSEapplied vs. 6H/60 line. Ignore the intercept value.


15. (Relative absolute healing)time, t = [(H/60)1000ooo (6H/60)t] (slope in Step 14.)

16. Plot relative absolute healing vs. log time. See Figure 4.6.

17. Add a logarithmic trendline. Show equation. Healing rate is slope of the line.

The relative healing begins at 1000 cycles when the loading is halted. Figure 4.7

shows the comparison of healing versus time for each specimen at each temperature.

When relative absolute healing equal 1.0, the specimen is fully healed. By substitution,

one can determine the time to full healing. Figure 4.8 shows the comparison of time to

full healing. Figure 4.9 shows the comparison of the healing rates. Appendix C shows

the healing test figures for both mixtures at both tested temperatures. Appendix D shows

the data used to generate the figures and healing rates.




1

2D 0 .9 ------------ -----------------
0.9
E0.8
,LI
0.7
0.6
U Logarithmic Trendline
0.5
S0.4- = 0.3345Ln(x)- 2.3105

0.3
I



0
1000 10000 100000 1000000
log Time (seconds)


Figure 4.6: Healing vs. Time after 1000 Cycles of Loading



































1000 10000 100000
log Time (seconds)


1000000


Figure 4.7: Comparison of Healing vs. Time after 1000 Cycles of Loading


0 Coarse 1
*Fine 1


100C


150C
Mixture Type and Temperature


Figure 4.8: Comparison of Time to Full Healing


11,
















0.45__ OnCoarse 1
S.4 Fine 1
S 0.4
0.3345
0.35

a 0.3
"U 0.2321
S 0.25
0170.1803
LU 0.1673
/ 0.2

0.15

0.1
c)
S0.05


100C 150C
Mixture Type and Temperature


Figure 4.9: Comparisons of Healing Rates




4.3.1 Determination of Dissipated Creep Strain Energy

As explained in the literature review, Zhang et al. (2001) developed a method to

determine DCSE/cycle during the linear portion of microcracking up to the threshold

boundary. The formula for a 0.1-second haversine load is


DCSE/cycle = 1/20 02 Di m (100)"1


where o is the stress and D1 and m are determined from creep tests. A variation of this


formula, showing the DCSE at a given time during testing is


DCSEN= 1/20 *2 D1 m (100)" N


where N is the number of cycles. For the stress, either the tensile strength or the far-away

stress can be used, depending on the situation. The far-away stress equation is


CFA = 2 P/(27 t D)

where P is the load (psi), t is the thickness (in.) and D is the diameter (in.).









Zhang also determined a method for determining DCSEfailure. This process is

detailed in Chapter 2, Section 2.4, with results dependent on MR, creep and strength IDT

tests. Table 4.1 shows the DCSE to failure for both mixtures at both temperatures.

Appendix B shows the applicable values in tabular form for determining DCSEfailure and

DCSEN.




Table 4.1: DCSE to Failure for Each Specimen

Mixture Temperature DCSEfalure
(C) (psi)
C1 10 1.049
C1 15 0.836
Fl 10 0.75
Fl 15 0.775


4.3.2 DCSE/DCSEmpid vs. 6/6

The DCSE equation is only applicable during the loading portion of testing

because DCSE is a function of cycles of loading and was developed for repetitive

loading. This prevents it from being directly measured during the healing portion of the

test. For this study, it was desirable to relate the healing in terms of recovered energy.

The relationship between DCSE and 6H/60 is linear during the loading portion of the test.

Because 6H can be measured during the healing portion from MR tests and 60 can be

measured from the beginning of loading, 6H/60 can be used to determine the recovered

DCSE. The DCSE equation is dependent on the level of stress and different stress levels

were used for each specimen to standardize time to failure. Also, each specimen has

different DCSE failure limits. This prevents DCSE by itself from being used to

determine a healing rate, since it cannot be related to other specimens or temperatures.











The DCSE/DCSEapplied value ratios the dissipated energy to the total dissipated energy

when loading is halted and was deemed the most appropriate representation of damage

for comparison of healing. Figure 4.10 shows the comparison of DCSE/DCSEapplied


vs. 6H/60 for each mixture at both temperatures.




1
Fine 1 @ 15
0.9
y = 5.4389x- 5.4617 Fine 1@ 10
0.8 2 =n
y = 4.9261x 4.9585
0.7 -2 =n a-q--7
Coarse 1 @ 15
y = 9.7249x 9.7368 C 10
M Coarse 1 @ 10
L 0 R2=0.9831
S0.5 y = 3.0312x 3.0332
0.4 R2 = 0.9961

o 0.3 I Coarse 1 @ 10 -
0.2- iCoarse 1 @ 15
AFine 1 @10
0.1 Y" Fine 1 @15
0
0.9 1 1.1 1.2 1.3 1.4 1.5




Figure 4.10: Comparison of DCSE/DCSEapplied vs. 6J 0




4.3.3 Steric Hardening Effects

Steric hardening is a structural hardening associated with asphalt materials. Over

time, an asphalt pavement will creep to a more stable structure and develop a higher

stiffness. This stiffness is reversible and mechanical action will return the asphalt

concrete to its original state. E. Barth (1962) stated that "Age hardening or steric

hardening in an asphalt occurs rapidly at first but appears to approach a limiting degree of

hardness on prolonged standing." Tests supporting this statement found that 36% of

ultimate steric hardening occurred within 10 hours of testing near freezing at 2.50C.









Steric hardening explains the sudden decrease in stiffness at the beginning of

loading and increase in stiffness after the repeated loading is stopped. Reversal of steric

hardening occurs during the first couple of hundred cycles of loading. After loading

stops, the asphalt specimen experiences an immediate restructuring that is a partial

recovery of steric hardening. By ignoring the initial stiffness in the determination of the

60 value, the 60 is associated with the response of the material with no steric hardening.

While loading occurs, steric hardening is gone. When loading stops, the material starts to

restructure and begins to heal. By bringing the backwards-forecasted trendline up, initial

steric hardening effects are eliminated from the response of the mixture. Because initial

steric hardening is accounted for, when 60 equals 1.0, the specimen is fully healed.

By looking at Figure 4.2, one can see the large jump between the first two

recorded data points before the specimen is damaged in a linear fashion. If temperature

effects are the sole cause for the jump, then the difference between 60 and 6i should be

able to be subtracted out at the end of the loading portion of the test. Because the

temperature stabilizes after two minutes of loading, then it should return to its previous

level within two minutes of unloading. However, this would show that no healing

occurred and all healing activity was temperature restabilization. Steric hardening

provides a rational explanation to this inconsistency.



4.4 Damage Accumulation and Healing Across Temperatures

Development of a healing rate provided a way to analyze damage accumulation

that translated across temperatures. The two-temperature fracture tests involved loading

the specimen for 1000 cycles at 150C, cooling to 10C for an hour and then continuing









the test until failure of the specimen. Figure 4.11 shows the loading portion at 150C.

Figures 4.12 and 4.13 show the continuation of the tests at 10C. The average original 60

from virgin 10C tests was used to accurately portray the 6H/80 values. Damage

translation can be seen by the fact that 6H/80 at the beginning of the secondary test is not

1.0.

4.4.1 Stress Values

The stress used in DCSE and reference values for comparison are important for

correctly explaining what is occurring during healing. The correct stress to use for

DCSE/cycle calculations varies based on the situation. On a local level, to determine the

percentage to failure of a specimen at the end of loading, the DCSE should be calculated

with tensile strength as the stress. At the crack tip, tensile strength dominates and shows

how close the specimen is to ultimate failure. Figure 4.14 shows the damage to each

specimen at each temperature after 1000 cycles of loading.

For purposes of comparison on a global scale, the far-away stress is appropriate.

As a whole, the stress in a specimen during loading is much closer to the far-away stress.

The far-away stress accounts for the load applied. The damage level is related to the

applied stress and different loads were used for each mixture and temperature. To

correctly compare a mixture at two temperatures, the global approach is required. Much

of the specimen that influences deformation is being subjected Io stresses that are closer

to the OFA, which dominates the response.


















1.2


1.15


1.1

to
1.05
to

1


0.95

1


0 500 1000 1500 2000 2500 3000 3500
Number of Load Replications (cycles)


4000 4500 5000


Figure 4.11: 8 /8o vs. Cycles for Initial Loading at 15C


0 500 1000 1500 2000 2500 3000
Number of Load Replications (cycles)





Figure 4.12: 8H/8o vs. Number of Load Replications for C1


I Average loading








40






3-

At 100C after 1000 cycles of
loading at 150C
2.6




I I
2.2-



1.8

Ot8flc linear
1.4 t8fc curve











Figure 4.13: 8Ho0 vs. Number of Load Replications for F1








0.6

OCoarse 1
c 0.5 WrFine 1
0.5


0 0.4

"c
W 0.3


' 0.2
Ll


Mixture Type and Temperature


Figure 4.14: DCSEapplied(using o)/DCSEfaure vs. Mixture Type and Temperature







41


4.4.2 Damage Translation Process

The process to translate damage across temperatures is shown in Tables 4.2 and

4.3. Because the temperature is cooled from 15C to 10C, healing is occurring at both

temperatures. While the temperature will be at 10C for most of the hour, it has been

shown that healing occurs very quickly at first, so the effect of the higher temperature

present at the beginning of healing influences the healing rate. Because of this, healing

has been analyzed at both temperatures to understand what is occurring.




1.5-


1.4
Coarse 1 @ 10
y=3.0815x+ 1.0013
1.3 R2= 0.9961
14
1.2
to Fine 1 @ 10 Fine 1 @ 15
y= 2.1082x+ 1.0078 y= 2.1413x + 1.005
1.1 R2= n 7R2= 0.9906 I Coarse 1 @ 10
SCoarse 1 @ 15
1 Coarse 1 @15 AFine 1 @10
y= 1.1709x + 1.0021Fine 1 @ 15
R2 = 0.9831 *Fine 1 @ 15
0.9
0 0.05 0.1 0.15 0.2 0.25 0.3
DCSE/DCSE(failure)



Figure 4.15: BE/go vs. DCSE/DCSEfaiure




First, the relative healing at one hour must be found using the healing rate


equations in Figure 4.7. Table 4.2 shows the steps to determining recovered 6H/80.

Because the loading stopped at 1000 seconds, 4600 seconds represents one hour of

healing. The relative healing values are then inserted into Figure 4.10 to determine the


equivalent 6H/60. The recovered 6H/80 during the hour of healing is determined by










subtracting the value 1.0 from the equivalent 6H/60. Figure 4.15 shows 6H/60 vs.

DCSE/DCSEfailure. The DCSE values calculated using the far-away stress as discussed in

Section 4.2.1. As shown by the arrows, the DCSE/DCSEfailure value at the end of loading

at 150C is converted to the equivalent 6H/60 at 100C. Table 4.3 shows the remaining steps

for damage translation across temperatures. The damage translated 6H/60 value at the end

of the hour of healing is then determined by subtracting the recovered 6H/60 from the

6H/60 at 10C.


Table 4.2: Recovered 88Ho


Mixture Temperature Relative Healing in One Hour Equivalent 6H/80 Recovered 6H/80
(C) (Using Figure 4.7) (Using Figure 4.10) (1 Equivalent H/60)
C1 15 0.510609974 1.169111235 0.169111235
10 0.255376678 1.084909171 0.084909171

F1 15 0.345287668 1.076670727 0.076670727
10 0.275516228 1.062507101 0.062507101





Table 4.3: Comparison of Calculated and Tested 8o80

Mixture Temp. 6H/80 @ 10C H/80( @ end of hour 6H/80 from
at End of Loading Two-Temperature Test
(C) (Using Figure 4.15) (SH/6o @ 10C Recovered 6,, .;.. (From Fig.4.12 and 4.13)
C 15 1.267344188 1.098232954 1.082
10 1.182435018

F 15 1.187115507 1.11044478 1.236
10 1.124608406









In comparing with the tested 6H/60 value at the beginning of the secondary test, it

can be seen that the Cl healed almost completely at 15C in the hour. That corresponds

with the Cl having a very high healing rate at 150C and a low healing rate at 100C. The

Fl does not match up as well. The two-temperature tested 6H/60 is higher than either of

the calculated healed values. It can be noted that Fl heals almost as fast at 100C as at

150C in the short term.

Clearly everything did not heal, but the process is there and makes sense. Some

damage translated across temperatures and some healing occurred.



4.5 Permanent Deformation

Permanent deformation plots were generated for each healing test. Figure 4.16

shows a sample plot. Permanent deformation, or creep strain, is normally what is

associated with damage. The plot shows that very little permanent deformation is

recovered during the healing portion of the test. Delayed elasticity, or time-dependent

elastic behavior, effects are almost entirely eliminated by the 0.9-second rest period

within each cycle, which indicates the test is being run properly. By observing the

permanent deformation plots, one can see that a specimen does not have to have much

strain (creep) recovery to have healing. Also, healing cannot be measured or interpreted

from recovered strain. The percent of strain recovered is much less than the percent of

DCSE recovered.

The permanent deformation plot also supports the steric hardening idea. There is

a drastic change in the rate of change of resilient deformation at the beginning of loading.

A proportional increase in rate of permanent deformation, or creep strain, is not seen. If










temperature were the controlling factor for the initial jump, it would be evident in the

creep strain as well.




0.010000
0.009000
w 0.008000
0.007000
o 0.006000
E 0.005000
/
S0.004000
0.003000
S0.002000
0.001000
0.000000
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

Time (seconds)



Figure 4.16: Permanent Deformation vs. Time






4.6 Summary of Findings and Analyses

Relative healing is the recovered energy relative to the applied energy as a

function of time. At any given point in time, the relative healing can be determined using

the Healing vs. Time relationship. Just showing the DCSE healed is ineffectual as a

comparative tool because each specimen has different failure limits and different stress

levels during loading. An assumption made for this healing rate determination is that the

rate of healing is directly related to the amount of damage. The healing rate will be faster

with a larger load and is primarily dependent on damage. If there is more energy, there

will be more healing.









There are three identified possible mechanisms associated with recovery of the

effective stiffness of mixtures: actual healing (microcrack), reduction (restabilization) of

temperature, and steric hardening. Steric hardening appears to be the initial mechanism

when loading is halted. By considering steric hardening to be the primary effect, one can

subtract out the initial steric hardening recovery and assume the specimen will eventually

recover all steric hardening over time with long enough rest periods. The remainder can

be considered as healing.

The healing rate probably overestimates true healing due to residual steric

hardening occurring during the monitored healing portion of the test. Because an MR test

cannot be run for about two minutes after the fracture test is halted, it is difficult to

accurately determine what is occurring in the early stages of healing.

The healing rate converted a normalized horizontal deformation into dissipated

creep strain energy. This overcame the barrier of measuring DCSE without repetitive

loading, as previously developed formulae are only applicable during loading. The

relationship between 6H/60 and DCSE during loading was used to convert measured 6H

values from MR tests into DCSE. With healing computed in terms of DCSE, the healing

rates can be incorporated into the HMA Crack Growth Model for further refinement of

that model.

Damage was shown to accumulate across temperatures and a process was

developed to translate damage and recovery across temperatures. At higher temperatures,

the healing rates are much higher and mixture Cl heals much faster than Fl at 150C.

Even over the small temperature range of 5C, a difference of multiple days separates the

amount of healing that occurs.














CHAPTER 5
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS



5.1 Summary of Findings

Healing is a quantifiable measurement in asphalt pavements that can be compared

between mixtures and at different temperatures by using healing rate process and

comparison procedures developed in this thesis. Findings of note:

Healing rate was measured as a recovery of the effective stiffness of a mixture.

An approach was developed to describe healing rate in terms of recovered DCSE.

Creep strain recovery did not appear to be necessary for healing.

Healing was determined to be a non-linear process that occurred quickly at first,
but decreased over time. A logarithmic relationship appeared to provide the best
fit.

Steric hardening appeared to be the primary reason for initial rapid reduction in
stiffness during testing. A procedure was developed to separate the steric
hardening effect from the healing effect so that healing rates could be determined
more accurately.

The coarse-graded mixture was found to heal much faster than the fine-graded
mixture at 15C, even though the same binder was used. This implies that below
a certain temperature, there is almost no healing. It also implies that above a
certain temperature, the sample incurs almost no damage because of the
immediate healing.



5.2 Conclusions

The process to describe healing in terms of DCSE involves many steps and many

possible interpretations. Some conclusions of note:









The recovered strains do not appear to be an adequate measure of healing.

The fact that healing can be expressed in terms of DCSE indicates that it can be
incorporated into the HMA Crack Growth Model developed by Zhang (2000).

Steric hardening appears to play a significant role in a mixture's response during
both the damage and the healing portions of laboratory tests and therefore must be
considered in their interpretation.

At 10C, healing is much slower than at 15C. This shows that healing is very
dependent on temperature and implies that at high enough temperatures and
viscosities, healing is almost immediate.

Healing is an important aspect of field performance that needs to be considered.



5.3 Recommendations

This research project has set the stage for future work. With the developed

healing rate, a greater understanding of healing can be achieved. Some recommendations

of note:

Run tests at higher temperatures, such as 20C and 25C. This will provide a
more complete look of healing effects due to temperature.

Specimen internal temperatures should be measured with thermocouples, so that a
measure of the temperature as a function of time can be seen.

Steric hardening needs to be looked at carefully over time.

Tests should be performed that include healing for some time and reloading; in
essence, a healing test on top of one. This could be used to see if the healing rate
changes.

One of the assumptions made in the determination of the healing rate is that the
amount of induced damage determines the rate of healing. Run tests to two
different damage levels and see if they converge at same healing point.

Incorporate the healing rate into the HMA Fracture Mechanics Law and Crack
Growth Model.















APPENDIX A
DATA FROM TESTED SPECIMENS













Table Al: Healing Fracture Tests for C1 at 15C


Specimen T18a-c1 T18c-cl
Asphalt Content (AC%) 6.5 6.5
Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279
Apparent Specific Gravity (Gsa) 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.159 2.17
Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469
Effective Specific Gravity of Aggregate (Gse) 2.549 2.549
Asphalt Absorption (Pba) 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.24 17.82
Percent Air Voids in Compacted Mix (AV) 7.26 6.78
Percent VFA in Compacted Mix (VFA) 60.22 61.94
* (Honeycutt, 2000)


Table A2: Healing Fracture Tests for Fl at 15C


Specimen T11flb T11f1c T12flb
Asphalt Content (AC%) 6.3 6.3 6.3
Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378
Apparent Specific Gravity (Gsa) 1.035 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.183 2.178 2.171
Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 2.488
Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554
Asphalt Absorption (Pba) 1.313 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273
Percent VMA in Compacted Mix (VMA) 17.79 17.97 18.24
Percent Air Voids in Compacted Mix (AV) 6.62 6.84 7.13
Percent VFA in Compacted Mix (VFA) 62.77 61.97 60.88
* (Honeycutt, 2000)














Table A3: Healing Fracture Tests for C1 at 100C


Specimen T17a-c1 T19b-c1
Asphalt Content (AC%) 6.5 6.5
Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279
Apparent Specific Gravity (Gsa) 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.153 2.155
Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469
Effective Specific Gravity of Aggregate (Gse) 2.549 2.549
Asphalt Absorption (Pba) 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.47 18.39
Percent Air Voids in Compacted Mix (AV) 7.51 7.43
Percent VFA in Compacted Mix (VFA) 59.32 59.61
* (Honeycutt, 2000)


Table A4: Healing Fracture Tests for Fl at 10C


Specimen T13fla T13flb T12flb
Asphalt Content (AC%) 6.3 6.3 6.3
Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378
Apparent Specific Gravity (Gsa) 1.035 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.177 2.182 2.171
Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 2.488
Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554
Asphalt Absorption (Pba) 1.313 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.01 17.82 18.24
Percent Air Voids in Compacted Mix (AV) 6.88 6.66 7.13
Percent VFA in Compacted Mix (VFA) 61.81 62.61 60.88
* (Honeycutt, 2000)














Table A5: Fracture Tests to Failure for C1 at 150C


Specimen T9a-cl T9b-cl T10a-cl
Asphalt Content (AC%) 6.5 6.5 6.5
Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 2.3279
Apparent Specific Gravity (Gsa) 1.035 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.153 2.156 2.162
Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 2.469
Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 2.549
Asphalt Absorption (Pba) 1.313 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.47 18.35 18.13
Percent Air Voids in Compacted Mix (AV) 7.51 7.38 7.13
Percent VFA in Compacted Mix (VFA) 59.32 59.77 60.68
* (Honeycutt, 2000)





Table A6: Fracture Tests to Failure for Fl at 15C

Specimen T4flb T7f1c T8f1b
Asphalt Content (AC%) 6. 6. 6.
Theoretical Maximum Specific Gravity (Gmm) 2.337E 2.337E 2.337E
Apparent Specific Gravity (Gsa) 1.03 1.03E 1.03E
Bulk Specific Gravity of Compacted Mix (Gmb) 2.176 2.18q 2.16q
Bulk Specific Gravity of Aggregate (Gsb) 2.48 2.48 2.48
Effective Specific Gravity of Aggregate (Gse) 2.55 2.554 2.55
Asphalt Absorption (Pba) 1.31 1.313 1.31
Effective Asphalt Content of Mixture (Pbe) 5.27 5.273 5.27
Percent VMA in Compacted Mix (VMA) 18.05 17.71 18.4
Percent Air Voids in Compacted Mix (AV) 6.92 6.54 7.3
Percent VFA in Compacted Mix (VFA) 61.66 63.1C 59.97
* (Honeycutt, 2000)


















Table A7: Fracture Tests to Failure for C1 at 10C


Specimen T8a-cl Tllc-cl c1-18-a*
Asphalt Content (AC%) 6.5 6.5 6.5
Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 2.3279
Apparent Specific Gravity (Gsa) 1.035 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.148 2.154 2.159
Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 2.469
Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 2.549
Asphalt Absorption (Pba) 1.313 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.66 18.43 18.24
Percent Air Voids in Compacted Mix (AV) 7.73 7.47 7.26
Percent VFA in Compacted Mix (VFA) 58.58 59.46 60.22
* (Honeycutt, 2000)








Table A8: Fracture Tests to Failure for Fl at 10C

Specimen T10f1c T7fla fl-10-d*
Asphalt Content (AC%) 6.3 6.3 6.3
Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378
Apparent Specific Gravity (Gsa) 1.035 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.163 2.18 2.172
Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 2.488
Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554
Asphalt Absorption (Pba) 1.313 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.54 17.90 18.20
Percent Air Voids in Compacted Mix (AV) 7.48 6.75 7.09
Percent VFA in Compacted Mix (VFA) 59.67 62.29 61.03
* (Honeycutt, 2000)














Table A9: Two-Temperature Fracture Tests to Failure for C1

Specimen T10b-c1 T10c-cl T13c-cl
Asphalt Content (AC%) 6.5 6.5 6.5
Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 2.3279
Apparent Specific Gravity (Gsa) 1.035 1.035 1.03
Bulk Specific Gravity of Compacted Mix (Gmb) 2.163 2.158 2.155
Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 2.469
Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 2.549
Asphalt Absorption (Pba) 1.313 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.09 18.28 18.39
Percent Air Voids in Compacted Mix (AV) 7.08 7.30 7.43
Percent VFA in Compacted Mix (VFA) 60.84 60.07 59.61
* (Honeycutt, 2000)


Table A10: Two-Temperature Fracture Tests to Failure for Fl

Specimen T8flc T10flb
Asphalt Content (AC%) 6.3 6.3
Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378
Apparent Specific Gravity (Gsa) 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.164 2.162
Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488
Effective Specific Gravity of Aggregate (Gse) 2.554 2.554
Asphalt Absorption (Pba) 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.50 18.58
Percent Air Voids in Compacted Mix (AV) 7.43 7.52
Percent VFA in Compacted Mix (VFA) 59.82 59.52
* (Honeycutt, 2000)










Table All: MR, Creep and Strength Tests for C1 at 15C


Specimen T15a-c1 T15b-c1 T16a-c1
Asphalt Content (AC%) 6.5 6.5 6.5
Theoretical Maximum Specific Gravity (Gmm) 2.3279 2.3279 2.3279
Apparent Specific Gravity (Gsa) 1.035 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.154 2.155 2.159
Bulk Specific Gravity of Aggregate (Gsb) 2.469 2.469 2.469
Effective Specific Gravity of Aggregate (Gse) 2.549 2.549 2.549
Asphalt Absorption (Pba) 1.313 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273
Percent VMA in Compacted Mix (VMA) 18.43 18.39 18.24
Percent Air Voids in Compacted Mix (AV) 7.47 7.43 7.26
Percent VFA in Compacted Mix (VFA) 59.46 59.61 60.22
* (Honeycutt, 2000)





Table A12: MR, Creep and Strength Tests for Fl at 15C

Specimen T1fla T1flb T2fla
Asphalt Content (AC%) 6.3 6.3 6.3
Theoretical Maximum Specific Gravity (Gmm) 2.3378 2.3378 2.3378
Apparent Specific Gravity (Gsa) 1.035 1.035 1.035
Bulk Specific Gravity of Compacted Mix (Gmb) 2.178 2.175 2.169
Bulk Specific Gravity of Aggregate (Gsb) 2.488 2.488 2.488
Effective Specific Gravity of Aggregate (Gse) 2.554 2.554 2.554
Asphalt Absorption (Pba) 1.313 1.313 1.313
Effective Asphalt Content of Mixture (Pbe) 5.273 5.273 5.273
Percent VMA in Compacted Mix (VMA) 17.97 18.09 18.31
Percent Air Voids in Compacted Mix (AV) 6.84 6.96 7.22
Percent VFA in Compacted Mix (VFA) 61.97 61.50 60.57
* (Honeycutt, 2000)















APPENDIX B
SUMMARY OF MIXTURES' TEST RESULTS










Table Bl: Resilient Modulus Test Data


Resilient Modulus for STOA
Mixture Trimmed Average Poisson
Type Temp. Mean Values Value Ratio
(C) (GPa) (GPa)
7.85 0.42
Coarse 1 10 7.82 7.92 0.42
8.08 0.39
5.82 0.37
Coarse 1 15 5.68 5.75 0.38
5.75 0.36
9.4 0.27
Fine 1 10 9.39 9.49 0.28
9.68 0.25
5.85 0.25
Fine 1 15 5.9 5.87 0.22
5.87 0.23
(Properties from 100C tests from Honeycutt, 2000)




Table B2: Tensile Strength Test Data

Tensile Strength for STOA
Mixture Trimmed Mean Average Poisson
Type Temp. Values Value Ratio
(C) (MPa) (MPa)
1.57
Coarse 1 10 1.75 1.64 0.5
1.5
1.25
Coarse 1 15 1.22 1.24 0.38
1.24
2.05
Fine 1 10 2.28 2.08 0.46
1.9
1.37
Fine 1 15 1.37 1.36 0.31
1.35
(Properties from 100C tests from Honeycutt, 2000)













Table B3: Creep Test Data


Table B4: Dissipated Creep Strain Energy Calculation

Mixture Type
Coarse 1 Fine 1
Temp. (OC) 10 15 10 15
Resilient
Modulus (GPa) 7.92 5.75 9.49 5.87
Failure Strain (E) 4629.75 5862.07 2919.62 5045.68
Tensile Strength (MPa) 1.64 1.24 2.08 1.36
Initial Strain (.w) 4422.68 5646.42 2700.44 4813.99
Elastic Energy (kJ/m3) 0.1698 0.1337 0.2279 0.1575
Fracture Energy (kJ/m3) 7.4 5.9 5.4 5.5
Dissipated
Energy to Failure (kJ/m3) 7.2302 5.7663 5.1721 5.3425
Dissipated
Energy to Failure (psi) 1.0486 0.8363 0.7501 0.7748
* (Properties from 100C tests from Zhang, 2000)


Failure Fracture
Mixture Type Temp. m-value D1-value Strain Energy Poisson Ratio
(C) (1/psi) (Gw) (kJ/m3) (from creep)

Coarse 1 10 0.796 3.81E-07 4629.75 7.4 0.5


Coarse 1 15 0.8 8.37E-07 5862.07 5.9 0.38


Fine 1 10 0.656 5.73E-07 2919.62 5.4 0.46


Fine 1 15 0.6027 1.57E-06 5045.68 5.5 0.31

(Properties from 100C tests from Honeycutt, 2000)















APPENDIX C
FRACTURE TEST FIGURES

















0.0014
0.0013
0.0012
i 0.0011
S 0.001
. 0.0009
E 0.0008
0 0.0007
0.0007

0.0005
0.0004
0.0003
0.0002


0 500 1000 1500 2000


2500 3000 3500 4000


Number of Load Replications (cycles)





Figure Cl: Resilient Deformation vs. Load Replications (C1 at 150C to failure)


0.0014
0.0013 --AT4fl b linear
S 0.0012 -_ TA4f1 b curve
M 0.0011
. 0.0011 T7f1clinear ,
0 o.01 "T7f1c curve
OT8fl b linear
S 0.0009-
0.0009 T8fl b curve
k 0.0008
o 0.0007
S 0.0006
0.0005
0.0004
0.0003
0.0002


0 500 1000 1500 2000


2500 3000 3500 4000


Number of Load Replications (cycles)




Figure C2: Resilient Deformation vs. Load Replications (F1 at 150C to failure)


- T9a linear
S* T9a curve
A T9b linear
A T9b curve
O TO1a linear
T10a curve


















0.0014
0.0013--- A T8a linear
0.0012 -- AT8a curve
0.0011 --XT11c linear
0XT1cclinear_____________________________________
0.001 T1 1 c curve
Scl-18-a linear
c0.000 c1-18-a curve Al MAW_

0.0007-
0.0006-
0.0005
n nnnd


0
c


E
0

0


0.0003-
0.0002-


0 500 1000 1500 2000 2500 3000 3500 4000


Number of Load Replications (cycles)


Figure C3: Resilient Deformation vs. Load Replications (Cl at 10C to failure)


0.0014
0.0013- Ofl-10-d linear
- 0.0012-- *fl-10-d curve
w Ot10f1c linear
.c 0.0011 --
U* t10f1c curve
o 0.0009
0.0008 fl curve


o 0.0007 -
oi 0.0006 -- -
S 0.0005

,' 0.0004
S 0.0003

0.0002
0 500 1000 1500 2000 2500 3000 3500 4000


Number of Load Replications (cycles)





Figure C4: Resilient Deformation vs. Load Replications (F1 at 100C to failure)



















0.0007


U.UUUB


0.0005





0.0003 1t18a-cl
*-t18c-c1
0.0002
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000


Time (second)




Figure C5: Resilient Deformation vs. Time (Cl at 150C healing test)








0.0008


0.0007


0.0006


0.0005-

0.0004



0.0003 tfc


0.0002
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000


Time (second)




Figure C6: Resilient Deformation vs. Time (F1 at 150C healing test)




































0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000


Time (second)





Figure C7: Resilient Deformation vs. Time (Cl at 100C healing test)


n nnn


,. 0.0007
-,
U
u
. 0.0006

0
" 0.0005

4


Figure C8: Resilient Deformation vs. Time (F1 at 100C healing test)


L'
3
E


0.0004

0 t1l3fla
0.0003- -t13flb

-At12flb
0.0002 1
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000


Time (second)


i


- -Y -


I


491~25~


























S1.2


1.1
1 1-


0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (second)





Figure C9: Average SH/S0 VS. Time (Cl at 150C healing test)


SLogarithmic Trendline
M 1.2
r 12y = -4.267977E-02Ln(x) + 1.481454E+00
2= 9.423406E-01
1.1 0.000
y= 0.00018x + 1.00503
S R2 = 0.99059
1


0.9
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (second)


Figure C10: Average 8H/80 vs. Time (F1 at 150C healing test)


*Average loading

*Average healing


= 0.00010x + 1.00207
R2 = 0.98307


Logarithmic Trendline


y = -3.439427E-02Ln(x) + 1.313134E+00
6- R2 = 9.512166E-01
0.0265


SAverage loading

*Average healing








64






1.5
I Average loading
1.4 I Average healing

Logarithmic Trendline






1.1
y = -.00033x + 1.00130
R2 = 0.99606
862 01





0.9
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (second)




Figure C11: Average 8H/go vs. Time (Cl at 100C healing test)







1.5
OAverage loading

1.4 *Average healing


1.3
y = 0.00020x + 1.00779
R2 = 0.98753 .
S= 0Logarithmic Trendline
lo 2 y= -3.659121E-02Ln(x) + 1.427938E+00


0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (second)


Figure C12: Average SH/SO vs. Time (F1 at 100C healing test)















1.5

OAverage loading
1.4
I Average healing

1.3


j1.2
o y = 0.00010x + 1.00207
R2 = 0.98307
1.1

Logarithmic Trendline
1 y = -3.439427E-02Ln(x) + 1.339657E+00
R2= 9.512166E-01
0.9
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (second)





Figure C13: Average SH/80 without Steric Hardening vs. Time (Cl at 150C healing)








1.5

OAverage loading
1.4
*Average healing

1.3
y= 0.00018x + 1.00503
R2 = 0.99059
1.2


1.1
SLogarithmic Trendline
y = -4.267977E-02Ln(x) + 1.481454E+00
R2 = 9.423406E-01



0.9
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (second)


Figure C14: Average SH/80 without Steric Hardening vs. Time (F1 at 150C healing)



















y= 0.00033x+ 1.00130
R2 = 0.99606


Logarithmic Trendline
y = -5.518856E-02Ln(x) + 1.712529E+00
2 i


1.3 .


S1.2
Co

1.1


1. Average loading

*Average healing

0.9
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (second)





Figure C15: Average SH/&0 without Steric Hardening vs. Time (Cl at 100C healing)


0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Time (second)


Figure C16: Average SH/80 without Steric Hardening vs. Time (F1 at 100C healing)








67






0.010000

0.009000-

w 0.008000
U
0.007000

. 0.006000

0.005000
' 0.004000

0.003000

S 0.002000

S 0.001000t18-c
0.0000001
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000


Time (second)




Figure C17: Permanent Deformation vs. Time (Cl at 150C healing test)








0.010000

0.009000

S 0.008000

0.007000

S 0.006000

E 0.005000 -

I 0.004000

E 0.003000

S 0.002000 I- tlt1f1b

0. 0.001000
0.000000 11--- 2
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000


Time (second)




Figure C18: Permanent Deformation vs. Time (F1 at 150C healing test)








68






0.010000

0.009000

w 0.008000
U
S 0.007000

. 0.006000

S 0.005000
' 0.004000

0.003000

0.002000
[ tl 7a-c1
o. 0.001000 -t19b-c1

0.000000
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000


Time (second)




Figure C19: Permanent Deformation vs. Time (Cl at 100C healing test)








0.010000

0.009000- tl3fla
t13flb
w 0.008000
U
,. 0.007000
o 0.006000

0.005000
w 0.004000

E 0.003000

0.002000

C. 0.001000
0.000000
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000


Time (second)




Figure C20: Permanent Deformation vs. Time (F1 at 100C healing test)















APPENDIX D
HEALING TEST DATA










Table Dl: Coarse 1 at 10C Healing Specimens
Specimen: tl7a-c1
Time 6H 6H norm SH80
(seconds) (inches) (inches)
14 0.000495 1. 0.000485 1.0000
120 0.000511 0.000501 1.0316
250 0.000545 0.000535 1.1017
360 0.000565 0.000554 1.1418
480 0.000590 0.000578 1.1916
600 0.000616 0.000604 1.2451
730 0.000629 0.000617 1.2710
840 0.000650 0.000637 1.3135
960 0.000675 0.000662 1.3643
1187 0.000661 0.000648 1.3347
1375 0.000655 0.000642 1.3229
1633 0.000652 0.000639 1.3168
2830 0.000632 0.000620 1.2774
4599 0.000620 0.000607 1.2515
1. o5 = 0.000495 in., Actual 8H = 0.000439 in.

Specimen: tl9b-c1
Time 6H H norm 8H80
(seconds) (inches) (inches)
8 0.000570 2 0.000581 1.0000
120 0.000585 0.000597 1.0266
240 0.000611 0.000623 1.0716
360 0.000628 0.000641 1.1023
480 0.000650 0.000662 1.1398
600 0.000668 0.000681 1.1725
720 0.000686 0.000700 1.2041
840 0.000704 0.000717 1.2345
960 0.000717 0.000731 1.2570
1145 0.000704 0.000717 1.2345
1359 0.000709 0.000723 1.2442
1629 0.000698 0.000712 1.2249
2834 0.000679 0.000693 1.1918
4620 0.000671 0.000684 1.1766
2. o0 = 0.000570 in., Actual SH = 0.000509 in.

Average thickness = 1.02in.
Load = 13001b
T17a-c1 thickness = 1.00 in.
T19b-c1 thickness = 1.04in.

Normalizing formula for thickness =

6H* t/tavgi Pavg/Pi
















Table D2: Coarse 1 at 100C Healing Calculations
Average
Average Average 8660 DCSE/ DCSE/ Relative Abs.
Time 6H80 w/o st. hard. DCSE DCSEfail DCSEapplied Healing
0 1.000 0.0000 0.0000 0.0000
120 1.029 0.0134 0.0128 0.1200
245 1.087 0.0274 0.0261 0.2450
360 1.122 0.0403 0.0384 0.3600
480 1.166 0.0537 0.0512 0.4800
600 1.209 0.0671 0.0640 0.6000
725 1.238 0.0811 0.0773 0.7250
840 1.274 0.0939 0.0896 0.8400
960 1.311 0.1073 0.1024 0.9600
1000 1.331 0.1118 0.1066 1.0000
1166 1.285 1.31946 0.03589
1367 1.284 1.31839 0.03913
1631 1.271 1.30571 0.07756
2832 1.235 1.26950 0.18734
4609.5 1.214 1.24893 0.24968


Linear trendline equation for x = 1000, y=0.00033x+1.00130 = 1.331
Log trendline equation for x = 1000, y = -0.055189*Ln(x)+ 1.6777 = 1.296
Diff. between backcalculated dh/do from log. at 1000s. and linear at 1000s = 0.035


Slope of DCSE/DCSEapplied = 3.0312


m-value = 0.796
Di-value = 3.81E-07
GFA = 137.359psi = 2*1300/(3.14*1.02*5.91)
co = 233.019psi =(1.57+1.75+1.5)/3*1000/6.895
DCSEfail = 1.04862psi

















Table D3: Fine 1 at 100C Healing Specimens
Specimen: tl3fla
Time 6H 6H norm 6H8o
(seconds) (inches) (inches)
8 0.000430 1. 0.000437 1.0000
120 0.000438 0.000445 1.0178
240 0.000449 0.000457 1.0450
360 0.000459 0.000467 1.0678
480 0.000472 0.000479 1.0969
640 0.000477 0.000485 1.1101
720 0.000488 0.000496 1.1345
840 0.000493 0.000500 1.1453
960 0.000498 0.000506 1.1589
1165 0.000489 0.000496 1.1360
1367 0.000488 0.000496 1.1353
1602 0.000483 0.000491 1.1229
2898 0.000479 0.000487 1.1136
4621 0.000469 0.000477 1.0915
1. Jo = 0.000430 in., Actual 8H = 0.000386 in.

Specimen: tl3flb
Time 6H 6H norm 6H8o
(seconds) (inches) (inches)
9 0.000410 2. 0.000385 1.0000
120 0.000419 0.000393 1.0224
240 0.000436 0.000409 1.0626
365 0.000445 0.000418 1.0858
480 0.000447 0.000420 1.0907
600 0.000457 0.000429 1.1154
720 0.000464 0.000436 1.1325
840 0.000470 0.000441 1.1467
960 0.000477 0.000447 1.1622
1173 0.000472 0.000443 1.1508
1370 0.000465 0.000436 1.1333
1596 0.000464 0.000436 1.1325
2889 0.000450 0.000423 1.0980
4734 0.000443 0.000416 1.0809
2. 5o = 0.000410 in., Actual SH = 0.000371 in.

















Table D4: Fine 1 at 100C Healing Specimens continued
Specimen: tl2f1b
Time 6H 6H norm 6H/8o
(seconds) (inches) (inches)
17 0.000400 3. 0.000418 1.0000
120 0.000412 0.000430 1.0292
240 0.000430 0.000450 1.0754
360 0.000447 0.000467 1.1179
480 0.000459 0.000479 1.1463
605 0.000471 0.000492 1.1779
720 0.000479 0.000501 1.1983
840 0.000491 0.000513 1.2279
960 0.000500 0.000523 1.2508
1217 0.000492 0.000514 1.2296
1439 0.000484 0.000506 1.2096
1622 0.000488 0.000510 1.2200
2885 0.000478 0.000499 1.1938
4678 0.000475 0.000496 1.1871
1. 5o = 0.000400 in., Actual 8H = 0.000368 in.

Average thickness = 1.03in.
Load = 13001b
T13f1a thickness = 1.05 in.
T13flb thickness = 0.97 in.
T12f1b thickness = 1.08 in.

Normalizing formula for thickness =

H* t/tavgi Pavg/Pi
















Table D5: Fine 1 at 100C Healing Calculations
Average
Average Average 8&/80 DCSE/ DCSE/ Relative Abs.
Time 8t/80 w/o st. hard. DCSE DCSEfail DCSEapplied Healing
0 1.000 0.0000 0.0000 0.0000
120 1.023 0.0086 0.0114 0.1200
240 1.061 0.0171 0.0228 0.2400
361.66667 1.091 0.0258 0.0344 0.3617
480 1.111 0.0342 0.0456 0.4800
615 1.134 0.0439 0.0585 0.6150
720 1.155 0.0514 0.0685 0.7200
840 1.173 0.0599 0.0799 0.8400
960 1.191 0.0685 0.0913 0.9600
1000 1.208 0.0713 0.0951 1.0000 0.00000
1185 1.172 1.20476 0.01491
1392 1.159 1.19201 0.07773
1606.6667 1.158 1.19108 0.08233
2890.6667 1.135 1.16771 0.19744
4677.6667 1.120 1.15243 0.27270


Linear trendline equation for x = 1000, y=0.0002x+1.00779 = 1.208
Log trendline equation forx = 1000, y= -0.03659*Ln(x)+1.4279 = 1.175
Diff. between backcalculated dh/do from log. at 1000s. and linear at 1000s = 0.033


Slope of DCSE/DCSEapplied = 4.9261


m-value = 0.656
Di-value = 5.73E-07
GFA = 136.025psi =2*1300/(3.14*1.03*5.91)
co = 301.184psi = (2.05+2.28+1.9)/3*1000/6.895
DCSEfail = 0.75012psi










Table D6: Coarse 1 at 150C Healing Specimens
Specimen: t18a-c1
Time 6H 6H norm 6H8o
(seconds) (inches) (inches)
8 0.000373 1. 0.000378 1.0000
120 0.000379 0.000384 1.0152
240 0.000383 0.000388 1.0264
360 0.000387 0.000392 1.0366
480 0.000390 0.000395 1.0442
600 0.000396 0.000402 1.0621
720 0.000400 0.000406 1.0728
840 0.000407 0.000413 1.0912
960 0.000410 0.000416 1.0992
1188 0.000395 0.000401 1.0599
1372 0.000396 0.000401 1.0608
1634 0.000394 0.000400 1.0567
2764 0.000386 0.000392 1.0357
4775 0.000377 0.000383 1.0112
1. o5 = 0.000373 in., Actual 6H = 0.000336 in.

Specimen: t18c-cl
Time 6H H norm 8H80
(seconds) (inches) (inches)
9 0.000360 2. 0.000355 1.0000
120 0.000367 0.000361 1.0185
240 0.000371 0.000366 1.0310
360 0.000372 0.000366 1.0319
480 0.000377 0.000372 1.0481
600 0.000385 0.000379 1.0685
720 0.000391 0.000385 1.0861
845 0.000394 0.000388 1.0935
965 0.000391 0.000385 1.0861
1169 0.000384 0.000378 1.0662
1348 0.000389 0.000383 1.0796
1575 0.000384 0.000379 1.0671
3000 0.000376 0.000370 1.0431
4800 0.000371 0.000365 1.0296
2.0 = 0.000360 in., Actual H = 0.000333 in.

Average thickness = 1.025in.
Load = 7001b
T18a-c1 thickness = 1.04in.
T18c-cl thickness = 1.01 in.

Normalizing formula for thickness =

H* t/tavgi Pavg/Pi













Table D7: Coarse 1 at 150C Healing Calculations
Average
Average Average 8&/80 DCSE/ DCSE/ Relative Abs.
Time 8&/80 w/o st. hard. DCSE DCSEfail DCSEapplied Healing
0 1.000 0.0000 0.0000 0.0000
120 1.017 0.0087 0.0104 0.1200
240 1.029 0.0173 0.0207 0.2400
360 1.034 0.0260 0.0311 0.3600
480 1.046 0.0347 0.0414 0.4800
600 1.065 0.0433 0.0518 0.6000
720 1.079 0.0520 0.0622 0.7200
842.5 1.092 0.0608 0.0727 0.8425
962.5 1.093 0.0695 0.0831 0.9625
1000 1.102 0.0722 0.0863 1.0000 0.00000
1178.5 1.063 1.08956 0.12163
1360 1.070 1.09672 0.05201
1604.5 1.062 1.08846 0.13234
2882 1.039 1.06592 0.35152
4787.5 1.020 1.04692 0.53630


Linear trendline equation for x= 1000, y=0.00010x+1.00207 = 1.102
Log trendline equation forx = 1000, y= -0.03439*Ln(x)+1.3131 = 1.076
Diff. between backcalculated dh/do from log. at 1000s. and linear at 1000s = 0.027


Slope of DCSE/DCSEapplied = 9.7249


m-value = 0.8
D,-value = 8.37E-07
GFA = 73.602psi = 2*700/(3.14*1.025*5.91)
co = 179.357psi = (1.25+1.22+1.24)/3*1000/6.895
DCSEfail = 0.8363psi
















Table D8: Fine 1 at 150C Healing Specimens
Specimen: t11flb
Time 6H H norm 6H8o
(seconds) (inches) (inches)
18 0.000470 1. 0.000464 1.0000
120 0.000479 0.000473 1.0195
240 0.000487 0.000481 1.0355
360 0.000493 0.000487 1.0496
480 0.000511 0.000504 1.0869
600 0.000513 0.000507 1.0922
720 0.000523 0.000517 1.1128
840 0.000534 0.000527 1.1355
960 0.000540 0.000533 1.1482
1187 0.000545 0.000538 1.1589
1380 0.000543 0.000536 1.1546
1740 0.000535 0.000528 1.1379
2228 0.000533 0.000526 1.1330
4628 0.000521 0.000514 1.1082
1. 5o = 0.000470 in., Actual SH = 0.000413 in.

Specimen: tlfic
Time 8H 6H norm 6H8o
(seconds) (inches) (inches)
10 0.000373 2. 0.000375 1.0000
120 0.000381 0.000383 1.0206
240 0.000396 0.000398 1.0617
360 0.000407 0.000410 1.0916
495 0.000414 0.000416 1.1090
621 0.000424 0.000427 1.1367
720 0.000425 0.000427 1.1385
840 0.000441 0.000443 1.1814
960 0.000446 0.000449 1.1953
1161 0.000445 0.000448 1.1926
1575 0.000444 0.000447 1.1903
2252 0.000431 0.000433 1.1542
3405 0.000429 0.000432 1.1506
4586 0.000424 0.000427 1.1376
2. 0o = 0.000373 in., Actual SH = 0.000323 in.














Table D9: Fine 1 at 150C Healing Specimens continued
Specimen: tl2f1b
Time 6H 6H norm 6H/8o
(seconds) (inches) (inches)
8 0.000370 3. 0.000372 1.0000
120 0.000379 0.000381 1.0234
240 0.000391 0.000393 1.0568
360 0.000404 0.000407 1.0923
480 0.000408 0.000411 1.1027
600 0.000420 0.000423 1.1360
720 0.000420 0.000423 1.1351
840 0.000429 0.000432 1.1604
960 0.000437 0.000440 1.1811
1240 0.000444 0.000447 1.2005
1447 0.000434 0.000436 1.1716
1676 0.000429 0.000431 1.1581
2855 0.000422 0.000424 1.1401
4690 0.000418 0.000420 1.1284
3. 5o = 0.000370 in., Actual 8H = 0.000348 in.

Average thickness = 1.07in.
Load = 925 Ib
T1lflb thickness = 1.06in.
T1lflcthickness = 1.08in.
T12f1b thickness = 1.08 in.

Normalizing formula for thickness =

H* t/tavgi Pavg/Pi


















Table D10: Fine 1 at 150C Healing Calculations
Average
Average Average 86/80 DCSE/ DCSE/ Relative Abs.
Time 8&/80 w/o st. hard. DCSE DCSEfail DCSEapplied Healing
0 1.000 0.0000 0.0000 0.0000
120 1.021 0.0079 0.0102 0.1200
240 1.051 0.0158 0.0204 0.2400
360 1.078 0.0237 0.0306 0.3600
485 1.100 0.0320 0.0413 0.4850
607 1.122 0.0400 0.0516 0.6070
720 1.129 0.0475 0.0612 0.7200
840 1.159 0.0554 0.0714 0.8400
960 1.175 0.0633 0.0817 0.9600
1000 1.185 0.0659 0.0851 1.0000 0.00000
1196 1.184 1.18397 0.00579
1467.3 1.172 1.17219 0.06982
1889.3 1.150 1.15007 0.19015
2829.3 1.141 1.14122 0.23830
4634.7 1.125 1.12472 0.32802


Linear trendline equation for x = 1000, y=0.00018x+1.00503 = 1.185
Log trendline equation for x = 1000, y = -0.042680*Ln(x)+1.48145 = 1.187
Diff. between backcalculated dh/do from log. at 1000s. and linear at 1000s = 0


Slope of DCSE/DCSEapplied = 5.4389


m-value = 0.6027
Di-value = 1.57E-06
GFA 93.169psi =2*925/(3.14*1.07*5.91)
cO = 197.728psi = (1.37+1.37+1.35)/3*1000/6.895
DCSEfail = 0.77483psi
















LIST OF REFERENCES


Barth, E. J., Asphalt Science and Technology, New York, Science Publishers, Inc., 699
pp., 1962.

Bureau of Transportation Statistics, National Transportation Statistics (NTS) 2000,
Washington, DC, Bureau of Transportation Statistics, 2000.

Francken, L., "Fatigue Performance or a Bituminous Road Mix Under Realistic Test
Conditions," Transportation Research Record No. 712, pp. 30-36, 1979.

Honeycutt, K. E., "Effect of Gradation and Other Mixture Properties on the Cracking
Resistance of Asphalt Mixtures," Master's Thesis, University of Florida, 2000.

Jacobs, M.M.J., "Crack Growth in Asphaltic Mixes," Ph.D. Dissertation, Delft, The
Netherlands, Delft University of Technology, 1995.

Kim, Y.R., S.L. Whitmoyer, and D.N. Little, "Healing in Asphalt Concrete Pavements: Is
it Real?," Transportation Research Record No. 1454, pp. 89-96, 1994.

Little, D.N., R.L. Lytton, D. Williams, and Y.R. Kim, "Propagation and Healing of
Microcracks in Asphalt Concrete and Their Contributions to Fatigue," Asphalt Science
and Technology, ed. by Arthur M. Usmani, New York, Marcel Dekker, pp. 149-195,
1997.

Lytton, R.L., J. Uzan, E.G. Fernando, R. Roque, D. Hiltunen, and S.M. Stoffels,
"Development and Validation of Performance Prediction Models and Specifications for
Asphalt Binders and Paving Mixes," Report SHRP-A-357, Federal Highway
Administration, Washington, DC, 1993.

Monismith, C.L., J.A. Epps, and F.N. Finn, "Improved Asphalt Mix Design,"
Proceedings, Association of Asphalt Paving Technologists, Vol. 55, pp. 347-406, 1985.


Paris, P.C. and F. Erdogan, "A Critical Analysis of Crack Propagation Laws,"
Transactions of the ASME, Journal of Basic Engineering, Vol. 85, pp. 528-534, 1963.

Roque, R., W.G. Buttlar, B.E. Ruth, M. Tia, S.W. Dickison, and B. Reid, "Evaluation of
SHRP Indirect Tension Tester to Mitigate Cracking in Asphalt Pavements and Overlays,"
Final Report to the Florida Department of Transportation, University of Florida,
Gainesville, 1997.










Roque, R., B. Sankar, and Z. Zhang, "Determination of Crack Growth Rate Parameters of
Asphalt mixtures Using the Superpave IDT," Proceedings of the Association of Asphalt
Paving Technologists, Vol. 68, pp. 404-433, 1999.

Schapery, R.A., "Correspondence Principles and a Generalized J-integral for Large
Deformation and Fracture Analysis of Viscoelastic Media," International Journal of
Fracture, Vol. 25, pp. 195-223, 1984.

van Dijk, W., "Practical Fatigue Characterization of Bituminous Mixtures," Proceedings
of the Association of Asphalt Paving Technologists, Vol. 38, pp. 423-456, 1969.

Zhang, Z., "Identification of Suitable Crack Growth Law for Asphalt Mixtures Using the
Superpave Indirect Tensile Test (IDT)," Ph.D. Dissertation, University of Florida,
Gainesville, 2000.

Zhang, Z., R. Roque, B. Birgisson, and B. Sangpetngam, "Identification and Verification
of a Suitable Crack Growth Law," Proceedings of the Association of Asphalt Paving
Technologists, Vol. 70, 2001 (in press).















BIOGRAPHICAL SKETCH

Thomas Paul Grant was born in Biloxi, Mississippi, on January 29, 1976, to

Steven Bain and Catherine Ware Grant. He moved around the world due to his father's

assignments in the United States Air Force. He completed his secondary education at

London Central High School in the United Kingdom as valedictorian.

Thomas enrolled at the University of Florida in 1994 and received a Bachelor of

Science degree in dvil engineering in December 1999, graduating summa cum laude.

During that time, he also completed a three-semester co-op with Robert Bates &

Associates, Inc., in Jacksonville, Florida. Thomas passed the Engineer-Intern exam in

1999.

Thomas started working on his Master of Engineering in civil engineering in

January 2000 at the University of Florida. He is now working for Kimley-Horn and

Associates, Inc., in Ocala, Florida, as a traffic engineer.