Effects of asphalt mixture characteristics on dynamic modulus and fatigue performance

GEORGIA DOT RESEARCH PROJECT 16-19 FINAL REPORT
EFFECTS OF ASPHALT MIXTURE CHARACTERISTICS ON DYNAMIC MODULUS
AND FATIGUE PERFORMANCE
OFFICE OF PERFORMANCE-BASED MANAGEMENT AND RESEARCH 15 KENNEDY DRIVE FOREST PARK, GA 30297-2534

1. Report No.:

2. Government Accession No.:

FHWA-GA-19-1619

4. Title and Subtitle:

Development of Geosynthetic Design and Construction

Guidelines for Pavement Embankment Construction in North

Georgia

7. Author(s): S. Sonny Kim, Ph.D., P.E., F.ASCE, J. Robert A. Etheridge, M.S., Mi G. Chorzepa, Ph.D., P.E., Y. Richard Kim, Ph.D., P.E., F.ASCE,
9. Performing Organization Name and Address: University of Georgia, College of Engineering Driftmier Engineering Center, Athens, GA 30602 Phone: (706) 542-9804, Email: kims@uga.edu
12. Sponsoring Agency Name and Address: Georgia Department of Transportation Office of Performance-Based Management and Research 15 Kennedy Drive, Forest Park, GA 30297-2534

3. Recipient's Catalog No.:
5. Report Date: April 2019
6. Performing Organization Code:
8. Performing Organization Report No.: 16-19
10. Work Unit No.:
11. Contract or Grant No.: PI# 0015117
13. Type of Report and Period Covered:
Final; July 2016 April 2019 14. Sponsoring Agency Code:

15. Supplementary Notes: Prepared in cooperation with the U.S. Department of Transportation, Federal Highway Administration.
16. Abstract: Fatigue cracking is one of the most critical distresses involved in flexible pavement failure. For state
departments of transportation, the accurate prediction of flexible pavement service life in terms of potential fatigue cracking is crucial for pavement design, maintenance, and rehabilitation. Therefore, it is imperative to understand the asphalt mixture characteristics that contribute to fatigue cracking.
The Georgia Department of Transportation (GDOT) is interested in learning more about the material characteristics of Georgia-sourced asphalt material and how different mixtures can affect the predicted fatigue life. This interest includes investigating different asphalt mixtures to see how the growth of damage and fatigue cracking performance change over time.
This research study investigates predictions of the fatigue cracking performance of Georgia asphalt mixtures using three fatigue test methods: (1) overlay test, (2) Illinois flexibility index test (I-FIT), and (3) cyclic direct tension based on the simplified viscoelastic continuum damage (S-VECD) model. The results from these fatigue test methods were further used to examine the material characteristics that make up the asphalt material influence on the performance of the pavement. This report summarizes the effects of asphalt material characteristics such as asphalt binder type, nominal maximum aggregate size (NMAS), reclaimed asphalt pavement (RAP) percentage, and asphalt binder content on asphalt fatigue cracking resisting performance.
Further, this report provides a database of dynamic modulus (|E*|) for 19 different asphalt mixtures from across the state of Georgia that can be used in the MechanisticEmpirical Pavement Design Guide (MEPDG) for pavement design and performance analysis. Long-term aging and dynamic shear rheometer (DSR) test results of two binders (PG 64-22 and PG 76-22) from the state of Georgia are also present in this report for MEPDG Level 2 inputs.

17. Keywords: Dynamic Modulus, S-VECD, SCB, Overlay Test Fracture, Fatigue Cracking

18. Distribution Statement:

19. Security Classif. (of this report): Unclassified

20. Security Classification (of this page): Unclassified

Form DOT 1700.7 (8-69)

21. No. of Pages: 22. Price: 124

GDOT Research Project No. 16-19
Draft Final Report
EFFECTS OF ASPHALT MIXTURE CHARACTERISTICS ON DYNAMIC MODULUS AND FATIGUE PERFORMANCE
By
S. Sonny Kim, Ph.D., P.E., F.ASCE Associate Professor
Robert A. Etheridge Graduate Research Assistant
Mi G. Chorzepa, Ph.D., P.E. Associate Professor
Civil Engineering, College of Engineering University of Georgia
Y. Richard Kim, Ph.D., P.E., F.ASCE Jimmy D. Clark Distinguished University Professor Alumni Association Distinguished Graduate Professor Department of Civil, Construction, and Environmental Engineering
North Carolina State University
Contract with
Georgia Department of Transportation In cooperation with
U.S. Department of Transportation Federal Highway Administration
April 2019
The contents of this report reflect the views of the authors, who are solely responsible for the facts and accuracy of the data, the opinions, and the conclusions presented herein. The contents do not necessarily reflect the official view or policies of the Georgia Department of Transportation (GDOT) or the Federal Highway Administration (FHWA). This report does not constitute a standard, specification, or regulation, and its contents are not intended for construction, bidding, or permit purposes. The use of names or specific products or manufacturers listed herein does not imply endorsement of those products or manufacturers.

TABLE OF CONTENTS
Page
LIST OF TABLES ........................................................................................................... vi LIST OF FIGURES ........................................................................................................ vii EXECUTIVE SUMMARY ............................................................................................. ix ACKNOWLEDGMENTS ............................................................................................. xiii SI CONVERSION FACTORS ........................................................................................... 1. INTRODUCTION ......................................................................................................1
1.1 PROBLEM STATEMENT ...........................................................................................1 1.2 STUDY OBJECTIVES................................................................................................2 2. LITERATURE REVIEW ..........................................................................................3 2.1 MIXTURE CHARACTERISTICS .................................................................................3 2.2 FRACTURE MECHANICS..........................................................................................7 2.3 DYNAMIC MODULUS ............................................................................................10 2.4 FATIGUE CRACKING TESTS FOR ASPHALT MIXTURES ..........................................18 2.5 SUMMARY OF TEST AND COMPARISON CHART ....................................................30 3. MATERIAL AND QUALITY CONTROL............................................................32 4. DYNAMIC MODULUS ...........................................................................................35 4.1 SPECIMEN FABRICATION ......................................................................................35 4.2 EXPERIMENTAL PROCEDURE ................................................................................36 4.3 RESULTS AND ANALYSIS ......................................................................................37 4.4 NMAS .................................................................................................................38 4.5 BINDER TYPE .......................................................................................................39 4.6 RAP CONTENT .....................................................................................................40 5. CYCLIC FATGUE TEST FOR S-VECD ..............................................................41 5.1 SPECIMEN FABRICATION ......................................................................................41 5.2 EXPERIMENTAL PROCEDURE ................................................................................42 5.3 DAMAGE CHARACTERISTIC CURVES ....................................................................45 5.4 DR FAILURE CRITERION........................................................................................46 5.5 SAPP CRACKING INDEX ..........................................................................................47 5.6 FATIGUE PERFORMANCE SIMULATION .................................................................51 5.7 SECTION DESIGN ..................................................................................................52 5.8 RESULTS FROM SIMULATION................................................................................53
iv

6. SEMICIRCLE BEND TEST ...................................................................................59 6.1 SPECIMEN FABRICATION ......................................................................................59 6.2 EXPERIMENTAL PROCEDURE ................................................................................61 6.3 RESULTS AND ANALYSIS ......................................................................................62
7. MODIFIED OVERLAY TEST ...............................................................................67 7.1 SPECIMEN FABRICATION ......................................................................................67 7.2 EXPERIMENTAL PROCEDURE ................................................................................67 7.3 RESULTS AND ANALYSIS ......................................................................................68
8. CONCLUSIONS .......................................................................................................72 8.1 DYNAMIC MODULUS ............................................................................................72 8.2 DIRECT TENSION CYCLIC FATIGUE TEST USING S-VECD MODEL ......................73 8.3 SEMICIRCLE BEND TEST.......................................................................................74 8.4 MODIFIED OVERLAY TEST ...................................................................................74 8.5 FATIGUE TEST METHOD COMPARISONS ...............................................................75
9. RECOMMENDATIONS AND FUTURE WORK ................................................76 10. REFERENCES .........................................................................................................77
Catalog of MEPDG Design Inputs for Asphalt Mixtures...................83 Aging and DSR Testing of Georgia Binders ......................................103 Proposed Standard Operating Procedure (SOP) ..............................111
v

LIST OF TABLES

Table

Page

2.1. Comparison of Test Summary................................................................................ 31
3.1. Plant Produced Mixture Properties......................................................................... 32
4.1 Dynamic Modulus Comparison Against NMAS.................................................... 38
4.2. Dynamic Modulus Comparison Against Binder Type ........................................... 39
4.3. Dynamic Modulus Comparison Against RAP Content.......................................... 40 5.1 DR and Sapp Values with the Corresponding R2 and Standard Deviation
Values ..................................................................................................................... 47
5.2. Recommended Threshold Values for the Sapp Fatigue Index Parameter................ 49
5.3. Recommended Threshold Values for the Sapp Based on GDOT Mixture Selection Criteria (Assuming 5% Truck Traffic and 1.17 ESAL Factor) .............. 50
6.1. Mixtures Used for SCB Testing ............................................................................. 61
6.2. Analysis of Variance of VMA, VFA, and AV for All Mixtures at 95% Confidence Interval ................................................................................................ 64
6.3. Analysis of Variance of Binder Type, RAP Content, Aggregate Source, and NMAS with a 95% Confidence Interval ................................................................ 65
6.4. I-FIT Results for All Mixtures (a) Fracture Energy and (b) Flexible Index .......... 66
7.1. Crack Progression Rate and Cycles to Failure for Various AC Mixtures.............. 69

vi

LIST OF FIGURES

Figure

Page

2.1. 0.45 Power Curve for 12.5, 19, and 25 mm NMAS Asphalt Mixtures .................... 7
2.2. Stress and Strain Curves from Sinusoidal Loading................................................ 11
2.3. Results of Dynamic Modulus Test Before Shifting ............................................... 12
2.4. Master Curve with Shifted Dynamic Modulus Values .......................................... 14
2.5. Results from a Cyclic Direct Tension Fatigue Test ............................................... 21
2.6. Load vs. Load Line Displacement Curve for a Typical SCB Test......................... 23
2.7. Overlay Test Result, Zhou and Scullion (2005)..................................................... 27
2.8. Typical Load Reduction Curve for Overlay Test ................................................... 29
3.1. Regional Separation of Plant Produced Mixtures .................................................. 34
4.1. Dynamic Modulus Test Setup ................................................................................ 36
4.2. Dynamic Modulus |E*| Results for Mixture with Different NMAS, Binder Types, RAP, and Aggregate Source ....................................................................... 37
5.1. SGC Compacted Specimen with Four Cores Taken from the Center .................... 41
5.2. Small Specimen Cut from 178 mm to 110 mm...................................................... 42
5.3. Small Specimen with Glued End Plates ................................................................. 43
5.4. Secured Test Specimen with Feeler Gauges .......................................................... 44
5.5. (a) C vs. S Curves for 12.5 mm Mixtures with PG 76-22 and PG 64-22 Binders, and (b) C vs. S Curves for 9.5 mm and 12.5 mm Mixtures with PG 64-22 Binder and 30% RAP ................................................................................... 45
5.6. (a) DR Failure Criterion Used to Compare Polymer-Modified PG 76-22 Binder with Unmodified PG 67-22 and PG 64-22 Binders, and (b) DR Failure Criterion Used to Compare 9.5 mm, 12.5 mm, 19 mm, and 25 mm Mixtures ...... 46
5.7. (a) Sapp Values for Mixtures with Different NMAS Values, and (b) Sapp Values for 12.5 mm Mixtures with PG 64-22, PG 67-22, and PG 76-22 Binders.................................................................................................................... 48
5.8. Sapp Values for Mixtures with Different Binder Contents for the Same NMAS and Binder Type......................................................................................... 49

vii

5.9. Damage Area Used in FlexPAVETM Separated into Top and Bottom for Cracking Reference ................................................................................................ 52
5.10. Damage Contours Based on Pavement Section with 12.5 mm NMAS and PG 64-22 Binder..................................................................................................... 54
5.11. FlexPAVETM Predicted Fatigue Cracking for 4 Inch Pavement Section: (a) (c) Total Percent Damage, (d)(f) Top Down Percent Damage, and (g)(i) Bottom Up Percent Damage................................................................................... 56
5.12. FlexPAVETM Predicted Fatigue Cracking for 6 Inch Two-Layer Pavement Section: (a)(c) Total Percent Damage, (d)(f) Top Down Percent Damage, and (g)(i) Bottom Up Percent Damage................................................................. 57
5.13. Pavement ME Predicted Fatigue Cracking: (a) 4 Inch Layer Top Down Cracking, (b) 4 Inch Layer Bottom Up Cracking, (c) Two 3 Inch Layers Top Down Cracking, and (d) Two 3 Inch Layers Bottom Up Cracking ....................... 58
6.1. SGC Specimen Prepared to be Cut into Two 50 mm Disks................................... 59 6.2. 50 mm Disk Cut from an SGC Specimen .............................................................. 60 6.3. Semicircle with 15 mm Notch in Center ................................................................ 60 6.4. Semicircle with 15 mm Notch in Center ................................................................ 63 7.1. Test Specimens for Modified Overlay Test ........................................................... 67 7.2. Tightening Pattern for Bolts Tex-248-F ................................................................. 68 7.3. Normalized Load Reduction Curve with Fitted Power Curve ............................... 69 7.4. Crack Progression Rate Vs. NMAS ....................................................................... 70 7.5. Crack Progression Rate Vs. Binder Type............................................................... 70 7.6. Cycles to Failure Vs. Crack Progression Rate ....................................................... 71
viii

EXECUTIVE SUMMARY
Fatigue cracking is one of the critical distresses in asphalt concrete pavement. This distress is a result of repeated loading from traffic that reduces the pavement performance and structural life. Pavement failure that results from these distresses is costly to state departments of transportation that maintain and repair these issues. Letting these cracks go unrepaired will only quicken deterioration through the presence of moisture and freeze/thaw cycles, leading to more costly repairs. To mitigate fatigue cracking, the underlying properties of asphalt concrete (AC) mixtures that contribute to crack propagation must be well understood.
One fundamental property of asphalt concrete is dynamic modulus. It defines the stiffness characteristics as a function of loading frequency and temperature. In order for Georgia's Department of Transportation (GDOT) to implement the MechanisticEmpirical Pavement Design Guide (MEPDG), level 1 inputs of dynamic modulus consisting of laboratory testing must be conducted. This will allow for GDOT to use the predictive capabilities of MEPDG to predict pavement performance and the amount of fatigue cracking over the pavement's design life.
This research project built a database of dynamic modulus values for 19 different asphalt mixtures from across the state of Georgia to be used in the MEPDG for pavement design and performance analysis. It also investigated three different fatigue tests that could possibly be implemented by GDOT as a way to rank asphalt mixtures in their ability to resist fatigue cracking. These different tests were further used to examine the material characteristics that make up the asphalt material's influence on the performance of the pavement. A number of characteristics were investigated, such as asphalt binder type,
ix

nominal maximum aggregate size (NMAS), reclaimed asphalt pavement (RAP) percentage, and asphalt binder content. Quality control was conducted on all asphalt materials that the lab received, including theoretical maximum specific gravity, bulk specific gravity, gradation, and air voids.
After conducting numerous dynamic modulus and fatigue tests, the following conclusions and recommendations were made:
Dynamic Modulus Generally, Superpave (Superior Performing Asphalt Pavements) mixtures with higher PG binder and increased RAP content (up to 30% RAP) result in higher dynamic modulus. Dynamic modulus values for NMAS between 25 mm and 19 mm were not significantly different. The same was true for 12.5 mm and 9.5 mm mixtures. Values for dynamic modulus were significantly different between 12.5 mm and 19 mm, as well as 12.5 mm and 25 mm. Binder type influenced dynamic modulus values, with the stiffer PG 76-22 binder being significantly different from both PG 64-22 and PG 67-22. However, there was not a notable difference between PG 64-22 and PG 67-22. RAP content had a great effect on dynamic modulus between 15% and 30% RAP contents.
Fatigue Test Method Comparisons For the fatigue tests, the semicircular bed (SCB) test and cyclic direct tension test with simplified viscoelastic continuum damage (S-VECD) model provide
x

consistent test results that could be used in identifying AC cracking potential. The advantage of the SCB test over the S-VECD test is simple sample fabrication, ease of operation, and quick testing time. The cyclic direct tension test with the S-VECD model provides more theoretically sound in-depth information to better understand AC mixture behavior. On the other hand, the cyclic direct tension test with the S-VECD model requires intensive training to complete a successful test compared to the SCB test. This concern could be overcome through lab training and a workshop at the University of Georgia upon GDOT's request. It is apparent from the results that the overlay test (OT) is the least favorable method to predict fatigue performance of asphalt mixtures. The issues of reliability and repeatability give concern for its use. With a larger database of dynamic modulus values created, the MEPDG can be implemented for design of flexible roadways. The implementation of the MEPDG would be most successful with training of staff and personnel regarding the inputs needed for the MEPDG. Having a firm background about these inputs and their significance will help GDOT use the MEPDG successfully in their designbuild projects. For successful MEPDG implementation, calibration of AASHTOWare Pavement ME and an accurate calibration coefficient for AC pavement is essential. Future studies to predict AC fatigue cracking should focus on the investigation of cracking performance using field-cored specimens and comparison of pavement condition surveys. The cyclic direct tension test method has a capability to estimate the calibration coefficients for fatigue cracking in Pavement ME based on S-VECD analyses.
xi

Based on laboratory test results using field-cored specimens with different NMAS (i.e., 12.5 mm, 19 mm, and 25 mm), the calibration coefficients in Pavement ME can be obtained. Also, Sapp criteria in the S-VECD model can be developed to select the appropriate mixture for field construction according to the design traffic. Finally, it is recommended that the flexibility index (FI) criteria are developed based on the SCB test method using field-cored specimens to accurately assess cracking performance.
xii

ACKNOWLEDGMENTS This project was conducted in cooperation with the Georgia Department of Transportation. The authors gratefully acknowledge the contributions of many individuals to the successful completion of this research project. This especially includes Mr. Binh Bui, Ms. Sheila Hines, Dr. Peter Wu, and Mr. Ian Rish, who have helped and advised the research team toward successful completion of the study.
xiii

SI CONVERSION FACTORS

1. INTRODUCTION
1.1 Problem Statement Developed under National Cooperative Highway Research Program (NCHRP)
1-37A, the MechanisticEmpirical Pavement Design Guide (MEPDG) provides three hierarchical levels of design inputs (i.e., levels 1, 2, and 3) to allow the designer to select the quality and the level of details of design inputs according to the level of importance of the project. Typically, level 1 offers the highest design reliability but requires the highest level of accuracy and laboratory dynamic modulus (|E*|) testing to run the MEPDG software for flexible pavement designs. The |E*| is considered one of the fundamental asphalt mix properties and is obtained from a series of complex modulus tests at different temperature and loading frequency conditions. Several State highway agencies (SHAs) have already created or are in the process of creating an |E*| database for the calibration and implementation of the MEPDG.
Georgia Department of Transportation (GDOT) has made a continued commitment to the performance enhancement of pavement and has proactively calibrated and implemented the MEPDG methodology for the design of flexible pavement structures. There already exists an |E*| database for some hot mix asphalt (HMA) mixes that are conventionally used in the state of Georgia through GDOT RP 12-07 and GDOT RP 14-12. Although the GDOT material input library includes |E*| for 25 mm, 19 mm, and 12.5 mm Superpave mixes with PG 64-22 and PG 67-22, the library is based on only two sources of aggregate. Further, an |E*| library for polymer-modified asphalt (PMA) mixtures has not been developed yet even though the PMA mixtures are being used for high-volume traffic roads in Georgia.
1

1.2 Study Objectives The primary objectives of the proposed research are to: (1) extend the |E*| database
for different aggregate sources with PG 64-22, PG 67-22, and PG 76-22 PMA; (2) recommend to GDOT a fatigue test method that provides better fatigue cracking prediction; and (3) identify the effects of nominal maximum aggregate size (NMAS), aggregate source, binder type, and other mix characteristics on the |E*|, and the long-term pavement performance to propose guidelines for the choice of input data.
This study uses the asphalt mixture performance tester (AMPT) to expand the GDOT material input database. Three fatigue test methods--the cyclic direct tension test based on the simplified-viscoelastic continuum damage (S-VECD) model, the semicircular bend (SCB) test, and the modified overlay test (OT)--are used to the determine cracking potential of asphalt concrete (AC) mixtures.
This study presents the fatigue index parameters from those fatigue test methods for different asphalt mixtures that are commonly used in Georgia. The relations among fatigue index parameters from each test method and AC mixture properties such as NMAS, reclaimed asphalt pavement (RAP) content, asphalt binder type, and asphalt binder content were also investigated to: (1) determine how these properties affect fatigue cracking potential of AC mixtures; and (2) compare fatigue test methods for usefulness as a test method and better fatigue cracking prediction.
Finally, this report presents the pavement performance analyses using AASHTOWare Pavement ME and FlexPAVETM to rank the mixes on their ability to resist cracking.
2

2. LITERATURE REVIEW
2.1 Mixture Characteristics In order to determine how mixture characteristics impact asphalt pavement behavior,
the research team performed tests to determine those properties. This section details those laboratory tests that were performed to obtain the physical properties of the different asphalt mixtures.

2.1.1 Bulk Specific Gravity and Theoretical Maximum Specific Gravity The bulk specific gravity (Gmb) test determines the specific gravity of compacted hot
mix asphalt by determining the ratio of a specimen's weight to the weight of an equal volume of water (PI, 2011). The test is performed according to AASHTO Standard T 166 "Bulk Specific Gravity of Compacted Asphalt Mixtures using Saturated Surface-Dry Specimens" (AASHTO T 166, 2015). The test measures a specimen's weight under three different conditions: dry, saturated surface dry (SSD), and submerged in water. After a specimen's dry weight is recorded, it is placed in a water bath for 4 minutes. At the end of the 4 minutes, the submerged weight is recorded and then the specimen is rolled on top of damp towels to remove any excess water on the surface while leaving the voids saturated. The SSD weight is recorded and the three different masses are used to calculate the bulk specific gravity using Eq. (1):

Bulk

Specific

Gravity

(Gmb)=

A B - C

(1)

Where, A = mass of specimen in air (g)

3

B = mass of SSD in air (g) C = mass of specimen in water (g)

The theoretical maximum specific gravity (Gmm) of an HMA mixture is the specific gravity of a sample excluding air voids (PI 2011). This test is performed on a sample of loose HMA by weighing the sample and then determining the volume by calculating the volume of water the sample displaces. The test is performed according to AASHTO T209 "Theoretical Maximum Specific Gravity and Density of Hot Mix Asphalt (HMA)" (AASHTO T209, 2016). A loose mixture, which is a broken-up sample with fine aggregates separated into particles smaller than 0.25 inch, is weight and dry mass recorded. The sample is then placed into a rigid container and filled with water enough to cover the sample by about 1 inch. The container is sealed and a vacuum of 2530 mm Hg is applied for 15 minutes. The container is periodically struck with a hammer to release trapped air bubbles. After 15 minutes, the vacuum is released and the container is submerged in water for 10 minutes and then the submerged weight is recorded. The recorded weights are used to determine Gmm using Eq. (2):

Theoretical

Maximum

Specific

Gravity

()=

A A-C

(2)

Where, A = mass of dry sample in air C = mass of water displaced by the sample

4

2.1.2 Air Voids Air voids in HMA pavement have a significant effect on its long-term performance.
Studies on the effects high percentages of air voids have on HMA have concluded that tensile strength, static and resilient moduli, stability, and fatigue life are reduced. (Kennedy et al. 1984, Pell and Taylor 1969, Epps and Monismith 1969, Linden et al. 1989, Finn et al. 1973). For these reasons, it is important in this study to create specimens with consistent air voids of 7 0.5% for super gyratory compacted (SGC) specimens. The percent air voids was chosen based on the target air voids described in the procedures for each test, and 7 0.5% satisfied all four tests. Percent air voids of a compacted HMA specimen can be determined from Gmb and Gmm using Eq. (3):

Air

Voids

(Va)

=

(1

-

)


100

(3)

2.1.3 Binder Content Binder content affects asphalt mixture performance related to stiffness, strength,
durability, fatigue life, raveling, rutting, and moisture damage (PI 2011). In order to determine binder content of HMA, the ignition test is commonly used. The ignition test is performed in accordance with AASHTO T 308, "Determining the Asphalt Binder Content of HMA by the Ignition Method." A sample of loose mix asphalt is placed into a mesh basket and into a forced air furnace. For this study, an NCAT Asphalt Content Furnace was used to determine binder content. The furnace heats the basket containing the loose mix to a temperature of 1000F. The internal scale measures the weight of the asphalt as the binder burns off. The weight before ignition and after ignition is used to determine the binder content, and a correction factor is applied to account for the loss of aggregate mass.

5

2.1.4 Gradation Performance of pavement is greatly influenced by particle size distribution or
gradation of aggregate. Gradation is an important aggregate characteristic that determines properties such as stiffness, stability, durability, permeability, workability, fatigue resistance, frictional resistance, and moisture susceptibility (Roberts et al. 1996). A maximum density gradation is a common reference in determining the desired gradation. To determine the maximum density gradation, a standard gradation graph known as the 0.45 power curve was introduced by the Federal Highway Administration (FHWA), which plots sieve sizes raised to the 0.45 power to percent passing of a sieve analysis. The maximum density line in this graph is a straight diagonal line from zero to the maximum aggregate size of the mixture being considered. The maximum aggregate size is considered to be one sieve larger than the nominal maximum aggregate size, while the NMAS is one sieve size larger than the first sieve to retain more than 10 percent of the material (Roberts et al. 1996). Typical HMA mix designs are considered dense graded which have a gradation near the 0.45 power curve but not exactly on it because there needs to be adequate volume for the binder to occupy. Figure 2.1 shows a 0.45 power curve.
6

Percent Passing (100%)

120% 100%
80% 60% 40% 20%
0% 0.075

Gradation Verification
2.36 4.75 9.512.5 19 25 37.5

12.5mm NMAS 19 mm NMAS 25 mm NMAS

Sieve Size (mm)
FIGURE 2.1 0.45 Power Curve for 12.5, 19, and 25 mm NMAS Asphalt Mixtures

2.2 Fracture Mechanics Fatigue cracking is one of the critical distresses in asphalt concrete pavement. This
distress is a result of repeated loading from traffic that reduces the performance and life cycle of roads. Pavement failure that results from these distresses are costly to state departments of transportation that maintain and repair these issues. Letting these cracks go unrepaired will quicken deterioration through the presence of moisture and freeze/thaw cycles, leading to more costly repairs. To mitigate fatigue cracking, the underlying properties of AC mixtures that contribute to crack propagation must be well understood. Fracture mechanics can be used to determine the properties of AC mixes that provide better cracking resistance.
To address the issues related to fatigue cracking, fracture mechanicsbased tests were developed. These include a variety of tests, such as the single-edge notch beam (SEB)

7

test, disk-shaped compaction test (DCT), semicircular bend test, and modified overlay test. While reliable, the SEB test suffers due to fabrication of a rectangular specimen (Wagoner et al. 2005). DCT, initially more favorable than the SCB test due to its potential crack surface being larger than SCB, can result in erroneous results if the crack propagation deviates from a straight path. The geometry is also much harder to create than an SCB specimen. The SCB test was chosen due to the research that indicates its success with identifying mixes that have fracture resistance properties and its repeatability (Wu et al. 2005, Li and Marasteanu 2009, Im et al. 2014)
Linear elastic fracture mechanics (LEFM) was developed to describe crack growth and fracture within a material under essentially linear elastic conditions (Irwin 1948). LEFM was first used in the fracture mechanics of metals. With the introduction of fracture mechanics to geological materials, the research and influence of rock mechanics fracture covered a huge field of studies. This led to the development of the SCB test by Chong and Kuruppa (1984) for rock fracture tests. The concept was then applied to asphalt, which acts at a quasi-brittle material especially at low temperatures. The underlying principle is that if energy stored near the crack tip exceeds the crack resistance of the material, then cracking initiates in the vicinity of the crack. Plastic deformation in the material creates a crack inelastic zone around the crack tip. If the inelastic region at the crack tip grows too large, then the elastic stress analysis will be inaccurate. The stress field at the crack tip is defined by the stress intensity factor, K. This factor depends on the mode of loading. The three principle modes are tensile mode, sliding mode, and tearing mode, which are called Mode I, Mode II, and Mode III, respectively. A combination of one or more modes is called mixed mode. This study focuses on Mode I. In Mode I, the crack initiation occurs when the stress
8

intensity factor reaches its critical value, KIC, which is known as fracture toughness. The stress intensity factor can be written as Eq. (4) (Lim et al. 1993):

K1 = Y1o

(4)

Where, = notch depth o = applied stress Y1 = normalized Mode I stress intensity factor
Stress is given by Eq. (5):

o

=

2

(5)

Where, = applied load = radius = thickness

Lim et al. (1993) developed expressions for Y1 for different specimen geometry

based

on

span

length

divided

by

the

radius.

For

example,

when



=

0.8,

Y1

can

be

expressed

as Eq. (6):

Y1{0.8} = 4.782 1.219(a/r) + 0.063exp(7.045(a/r))

(6)

Stress intensity factor has been used to study fracture behavior on asphalt mixtures below subzero temperatures (Biligiri 2012, Khalid and Monney 2009). However, LEFM
9

may not be applicable to mixtures that are above subzero temperatures. At higher temperatures, asphalt exhibits a viscoelastic response that creates a large inelastic zone around the crack tip, which leads to inaccuracies. An alternate to the LEFM approach is elastic plastic fracture mechanics (EPFM), which has been used to measure fracture resistance based on the energy of fracture.

2.3 Dynamic Modulus Dynamic modulus, |E*|, is a fundamental property of asphalt concrete that defines
the stiffness characteristics as a function of loading frequency and temperature. Dynamic modulus is used as a property input in the MechanisticEmpirical Pavement Design Guide developed by NCHRP Project 1-37A (ARA 2004). Using the principles of time temperature superposition, a master curve can be created to predict the behavior of asphalt under a given loading condition and temperature. |E*| is an important linear viscoelastic property that can be used in pavement models based on viscoelasticity. The dynamic modulus tests were performed according to AASHTO T 342 at three temperatures of 4oC, 20oC, and 40oC (39.2oF, 68oF, and 104oF) and six frequencies 25, 10, 5, 1, 0.5, and 0.1 Hz.

2.3.1 Complex Modulus Complex modulus, E*, is a stressstrain ratio of linear viscoelastic materials under
sinusoidal loading. The complex modulus contains a storage or elastic component (E) and a loss or viscous component (E) and can be written as Eq. (7):

E* = E + iE

(7)

10

Taking the absolute value of this complex number gives the dynamic modulus. It is equal to the amplitude of the sinusoidal stress, 0, divided by the maximum recoverable strain, 0, as in Eq. (8):

|E*| = 0

(8)

0

Due to the responses being time-dependent, the strain occurs after the load is applied in a time lag, which is defined as the phase angle, , and determined from Eq. (9):

= 2ft

(9)

Where, f = loading frequency in Hz t = time delay between the stress and strain cycles
For perfectly elastic materials the phase angle would be equal to 0, and it would be equal to 1 for perfectly viscous materials. Figure 2.2 shows the time lag between the stress and strain for a uniaxial sinusoidal compressive stress test.

FIGURE 2.2 Stress and Strain Curves from Sinusoidal Loading
11

2.3.2 Master Curve Development In asphalt mixtures, dynamic modulus values vary with temperature and loading
frequency. Due to this variation, it is difficult to compare test results. The master curve was introduced to give better comparison between test results (AASHTO R84-17). The master curve is based upon the thermorheological attributes of asphalt, which allow the time temperature superposition (tTs) principle to be applied. This principle allows the same modulus value to be inferred at either low temperatures and long loading times or high temperatures and short loading times. Using a timetemperature shift factor, dynamic modulus results from each temperature at each loading frequency can be shifted graphically along the frequency domain to create the master curve. Figure 2.3 shows the results of a dynamic modulus test and how the data can be shifted.

|E*| (ksi)

10000

1000

4C

100

20C

40C

Fit 10

1 1.E-06

1.E-04

1.E-02

1.E+00 1.E+02 1.E+04 1.E+06

Reduced Frequency, Hz
FIGURE 2.3 Results of Dynamic Modulus Test Before Shifting

The master curve can be defined mathematically with a sigmoidal function as Eq. (10):

12

Log

|E*|

=

+

1+ +()

(10)

Where, = reduced time of loading at reference temperature = minimum value of E* + = maximum value of E* and = parameters describing the shape of the sigmoidal function

The shift factor can be shown as Eq. (11).

a(T) =


(11)

Where, a(T) = shift factor as a function of temperature = time of loading at desired temperature = reduced time of loading at reference temperature T = temperature of interest

While the shift factor as a function of temperature is defined by a linear relationship, a second order polynomial fit is more accurate leading a(T) to be commonly described by the quadratic Eq. (12):

Log a(T) = aT2 + bT + c

(12)

Where, T = temperature of interest

13

a, b, c = regression coefficients
The resulting master curve from Figure 2.4 can be used to estimate |E*| at temperatures and frequencies that available equipment cannot mechanically test.

|E*| (ksi)

10000

1000

4C

20C

100

40C

Fit

10

1 1.E-07

1.E-05

1.E-03 1.E-01 1.E+01 1.E+03 1.E+05 Reduced Frequency, Hz

FIGURE 2.4 Master Curve with Shifted Dynamic Modulus Values

2.3.3 Predictive Models The MEPDG provides three levels of input design. Level 1 is the highest level and
requires regional material characterization of |E*| from laboratory testing. Levels 2 and 3 determine |E*| through predictive models. These models are based on simpler material properties and volumetric properties. The predictive models are briefly described in the following subsections.
2.3.4 Original Witczak Equation The original Witczak equation, developed as part of NCHRP 1-37A, used data from
205 mixtures to create Eq. (13):

14

Log10 | E*|= -1.249937 + 0.02923 200 - 0.001767(200)2 - 0.002841 4 - 0.05809Va

- 0.082208 +
+

3.871977-0.00214+0.0039583/80.0000173/82+0.00547 1+(-0.603313-0.313351 log -0.393532 log )

(13)

Where, | E*| = dynamic modulus 200 = percent passing #200 sieve 4 = percent retained on #4 sieve 3/8 = percent retained on inch sieve 3/4 = percent retained on inch sieve = percent of air voids = percent of effective asphalt content f = loading frequency (Hz) = binder viscosity at temperature of interest (106 poise)
This Witczak equation based on nonlinear regression analysis is an option for Level 2 analysis. There are limitations to the Witczak equation (Bari 2005). It relies on other models to change the binder shear modulus |G*| into binder viscosity. There is also a need for improved sensitivity to volumetric properties such as voids in mineral aggregate (VMA), voids filled with asphalt (VFA), asphalt content, and air void (AV).
2.3.5 Modified Witczak Equation Under NCHRP Project 1-40D, Witczak reformulated Eq. (13) to include binder |G*|b
in the model as Eq. (14):

15

Log10 | E*|= -0.349 + 0.754( |G*|b-0.0052)(6.65 - 0.032200 - 0.027(200)2 + 0.0114

+

0.0063/8

0.00014(3/8)2

-

0.0014(3/8)2

-

0.08Va



1.06

( ))
+

+

2.558

+

0.032

+

0.713
+

+

0.01243/8

-

0.0001(3/8)2

+0.00983/4

1+exp(-0.7814 - 0.5785log|G*|b + 0.8834logb))

(14)

Where, |G*|b = dynamic shear modulus of asphalt binder b = binder phase angle associated with |G*|b
The NCHRP 1-40D model is based on nonlinear regression similar to NCHRP 1-37A, and it used 346 mixtures. This G*-based model is used in Level 2 analysis in the MEPDG.

2.3.6 Hirsch Model Another model used to estimate |E*| is the Hirsch model suggested by Christensen
et al. (2003) that incorporates the binder modulus, VMA, and VFA as Eq. (15), Eq. (16), and Eq. (17):

|E*|

=

Pc(4,200,00(1-100)

+

3|G*|b

()
10,000

+

(1-)

(41,2-001,00000)+

3||()

(15)

= -21(logPc)2 55logPc

(16)

P = c

(20+3||().58 650+3||().58

(17)

16

Where, Pc = the aggregate contact volume This model lacks a strong dependence on volumetric parameters, but the use of the
empirical phase angle equation is beneficial when converting |E*| to the relaxation modulus or creep compliance.
2.3.7 Effects of AC Materials Characteristics on |E*| The continued effort by GDOT to update the |E*| database relies on laboratory testing
to provide the highest design input level. The dynamic testing done in this study to extend the database for GDOT investigated the properties of AC mixtures and the effect on the resulting dynamic modulus. A number of studies have been conducted to evaluate dynamic modulus test results and the effect of different factors on dynamic modulus and phase angle. Factors including aggregate, asphalt content, and RAP percentage had influence on the response (Flintsch et al. 2007, Cross and Jakatimath 2007). Binder was also determined to affect dynamic modulus, with a softer binder providing a lower modulus (Clyne et al. 2003). Studies completed show that a stiff asphalt binder, low asphalt content, and air voids contribute significantly to increase dynamic modulus values (Shu and Huang 2008). Polymer-modified binders have been shown to increase dynamic modulus values (Zhu et al. 2011).
17

2.4 Fatigue Cracking Tests for Asphalt Mixtures
2.4.1 Cyclic Direct Tension Test with Simplified Viscoelastic Continuum Damage Model (S-VECD)
Materials that show time-dependent behavior, such as viscoelastic materials, are affected by current input and past input history. The stressstrain relationship of linear viscoelastic materials is expressed with two convolution integrals, Eq. (18) and Eq. (19):



=

0

(

-

)





(18)



=

0

(

-

)





(19)

Where, () = relaxation modulus D() = creep compliance = integration variable

Complex modulus that is composed of two parts--the storage modulus and the loss modulus--is a parameter of linear viscoelastic behavior. AC stiffness is dependent on loading rate and temperature, which would cause the need to test stiffness over a large range of frequencies and temperatures. Due to the impracticality of performing a large number of tests, researchers are able to take advantage of the timetemperature superposition principle, which greatly reduces the amount of testing needed. Tests at different frequencies and temperatures can be shifted to a reference temperature, with the resulting shifted frequency called reduced frequency. These shifted frequencies form a master curve over which a dynamic modulus value can be obtained over a range of

18

temperatures and frequencies. Materials that are capable of forming a master curve are called thermorheologically simple materials. Linear viscoelastic theory allows for relaxation modulus and creep compliance to be converted from the complex modulus in the frequency domain. All three functions are considered unit response functions. Relaxation modulus is a stress response due to a unit step strain input, and creep compliance is a strain response due to a unit step stress input. Schapery (1984) suggested that the stress and strain terms in viscoelastic materials are defined as pseudo variables in the form of convolution integrals. Physical stress or strain in elastic solutions can be replaced by pseudo stress or strain. Eq. (20) presents the pseudo strain, R, that can be calculated based on the correspondence principle:

R

=

1

0

(

-

)





(20)

Where, R = pseudo strain = the measured strain E(t) = the linear viscoelastic relaxation modulus ER = the reference modulus (typically taken as 1)
Pseudo strain equals the stress response of linear viscoelastic material due to a certain strain input. This property allows for the time effect in a stresspseudo strain plot that forms nonlinear behavior to be removed. Removing the time effect proves that damage does not actually occur. Continuous damage ignores microscale behaviors and characterizes material using macroscale observations. To assess the structural integrity, an

19

instantaneous secant modulus can be employed; however, damage is more difficult to quantify. A theory to address this is Schapery's work potential theory based on thermodynamics principles. The theory quantifies damage by an internal state variable, S, that accounts for microstructural changes in the material. Eq. (21), Eq. (22), and Eq. (23) summarize the damage evolution law:

WR = f (R,S);

(21)



=





=

()

(22) (23)

Where, WR = pseudo strain energy density function = the damage growth rate
Based on this theory, cyclic fatigue tests are conducted in accordance with AASHTO TP107 entitled, "Standard Method of Test for Determining the Damage Characteristic Curve and Failure Criterion Using the Asphalt Mixture Performance Tester (AMPT) Cyclic Fatigue Test". By applying a cyclic fatigue test using the AMPT shown in Figure 2.5, the S-VECD model shows the fatigue damage growth as the modulus changes based on the pseudo strain energy input history (Kim and Little 1990, Daniel and Kim 2002, Chehab et al. 2003, Underwood et al. 2010).

20

Phase Angle () Dynamic Modulus (kPa)

40

9,000

35 30 25 20 15 10
5 0
0

Phase Angle |E*|
20,000 40,000 60,000 Number of Cycles

8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 80,000

FIGURE 2.5

Results from a Cyclic Direct Tension Fatigue Test

The relationship between the internal state variable representing damage, S, and the pseudo stiffness, C, is the primary interest of the model. A damage characteristic curve can be formed between C and S where the pseudo stiffness value starts at 1, indicating the material is intact and decreases as damage accumulates. The function can be fitted as a power function represented by Eq. (24), where C11 and C12 are model coefficients:

C = 1 C11SC12

(24)

This relationship is independent of mode of loading, temperature, and load amplitude, making it a fundamental material property. A study was performed to investigate how air voids and binder content affect damage characteristic curves, and it concluded that higher air voids increase damage accumulation and higher binder content decreases damage accumulation (Zeiada et al. 2013). Another study looked at RAP content and showed that an increase of RAP decreases fatigue performance (Sabouri et al. 2015).

21

In this study, this model is compared to other fatigue tests to determine which provides the best information about cracking resistance. The damage characteristic curve is useful to represent how damage grows, but failure criterion is needed to determine failure.
The DR failure criterion is used to predict material failure in the S-VECD model. The DR failure criterion is based on observation that the average reduction in pseudo stiffness up to failure is independent of the mode of loading, temperature, and load amplitude (Wang and Kim 2017). DR is defined by the slope of the linear relationship between the sum of 1-C to failure and the number of cycles to failure, Nf , in Eq. (25):

DR = 0(1-) = (1-)





(25)

Wang and Kim (2017) reported that DR changes with mixture characteristics, such as a higher RAP content lowering DR and polymer-modified binders having a higher DR value. While good trends have been recognized with DR, this value alone cannot compare the fatigue performance of different asphalt mixtures. To index mixtures for fatigue performance, a cracking index property was developed, called the apparent damage capacity (Sapp). The Sapp was developed by Wang and Kim (2017) based on concepts of the S-VECD model and the DR failure criterion. It can be defined as the corresponding S value on the damage characteristic curve when C is equal to (1-DR). Sapp is expressed in Eq. (26):

1

Sapp

=

1 10000



1
C11



DRC12

(26)

The value of 110000 is a normalization factor used to make the Sapp value in the range of

040 if the unit of stress is kPa. Using the Sapp parameter, the predictions of fatigue

performances on Georgia-sourced asphalt mixtures were investigated.

22

2.4.2 Semicircle Bend Test The AAHSTO TP105 standard for the SCB test is based on the calculation of
fracture energy, Gf, which is the energy required to create a unit surface area of a crack and is less dependent on linear elasticity and homogeneity (Marasteanu 2004). Figure 2.6 shows the results from a typical SCB test.

FIGURE 2.6 Load vs. Load Line Displacement Curve for a Typical SCB Test

Fracture energy is obtained using the work of fracture, Wf, which is the total area under the load line displacement curve as in Eq. (27):

G =
f

(27)

In Eq. (28), Wf is the work of fracture and calculated by finding the area underneath the load line displacement curve. Arealig is the ligament area calculated as in Eq. (28):

Arealig = t(r-a)

(28)

23

Where, t =thickness r = radius a = notch length
This test has been used to characterize the low temperature cracking resistance in AC mixtures. More recently, a new procedure has been developed by the University of Illinois to characterize fracture cracking at intermediate temperature. This test procedure (AASHTO TP124) is able to screen AC mixtures for their ability to resist fatigue cracking by creating a new index parameter to better discriminate AC mixture performance, called the flexibility index (FI). This procedure uses fracture mechanic principles to distinguish between AC mix characteristics. The Illinois Flexibility Index Test (I-FIT) uses fracture energy and post-peak slope to determine the FI. This new parameter is meant to identify brittle mixtures that are prone to cracking. I-FIT has shown consistent and repeatable trends and is able to determine with greater distinction between fracture properties than fracture energy alone (Ling et al. 2017; Ozer, Al-Qadi, Lambros et al. 2016; Ozer, Al-Qadi, Singhvi et al. 2016). The FI is more sensitive to changes within mix designs compared to fracture energy (Ling et al. 2017). Mixture characteristics such as reclaimed asphalt pavement have been shown to impact fracture resistance with great significance and that the use of RAP can increase the amount of cracking in pavement (Ling et al. 2017; Ozer, Al-Qadi, Lambros et al. 2016; Ozer, Al Qadi, Singhvi et al. 2016; Norouzi et al. 2017; Cascione et al. 2015). The hardened and stiffer binders within RAP are prone to cracking and are shown to decrease FI (Ling et al. 2017, Ozer et al. 2016). In Georgia, 25%30% RAP is widely used in surface mixes. This increase in RAP usage as a sustainable practice could have future
24

impacts on the cost to GDOT to maintain roads. Aging of asphalt mixtures was shown to greatly affect FI and its resistance to cracking (Ling et al. 2017, Kim et al. 2012). Since this study was not investigating the effect of aging on asphalt, precautions were taken to ensure the material was not aged.
Performance grade (PG) of binder has also been shown to greatly affect FI, as a stiffer binder has significant variation to a softer binder (Ling et al. 2017, Ozer et al. 2016). The use of modified binders is to provide better resistance to rutting, thermal cracking, and fatigue damage. Previous studies have proven that polymer-modified asphalt binder can improve permanent deformation from rutting and fatigue cracking resistance (Sargand and Kim 2001, Bahia 2011). The expected trend from the I-FIT test resulting from the modified binder would be an increase in fracture energy and FI. However, in one study on the sensitivity of the I-FIT procedure, a peculiar result occurred during the testing of polymermodified asphalt binder mixtures. The study results suggested that the modified binder affected FI negatively, resulting in lower FI for the asphalt with modified binder (Ling et al. 2017). This trend is the opposite of what was expected and was noted as such in the study. This could be an important finding to improve the test method because the study suggests that the procedure cannot adequately characterize polymer- modified asphalt binders.
Several studies have focused on the sensitivity analyses of testing variables, including loading rate, specimen thickness, and testing temperature (Ling et al. 2017). These study showed sensitivities of the testing variables (Ling et al. 2017). Further, an investigation was done on the variables of the SCB test to determine its reliability and repeatability (Nsengiyumva 2016). The study evaluated the number of specimens needed
25

for sufficient sample representation, specimen thickness, notch length, loading rate, and testing temperature. Their key findings are summarized as follows:
A reasonable number of SCB tests is six to evaluate representative fracture behavior of asphalt concrete with a 95% level of confidence.
A thickness of SCB specimen in the range of 40 mm to 60 mm provides consistent fracture energies.
Notch depth from 5 mm to 40 mm presents consistent fracture energy. Although it is known that loading rate affects fracture energy, the loading rates in
their study did not find significant difference, as the loading rate was low (0.1 to 10 mm/min). Mixtures with 25% and 30% RAP content are not significant for fracture energy or FI since this was due in part to the different binder types controlling the response, as the change in RAP content was only 5%.
2.4.3 Modified Overlay Test The overlay test (TxDOT Tex-248-F) was first designed in the 1970s by Germann
and Lytton (1979) and consisted of two steel plates with one fixed and the other moveable in the horizontal direction. The test was developed to simulate cracks that formed in the old pavement beneath an overlay and the reflective cracking caused to the overlay pavement. The original overlay test was upgraded by Zhou and Scullion (2005) to ease the fabrication of specimens and be compatible with field cores in order to evaluate the reflective cracking resistance. Testing is performed at room temperature with a loading rate of one cycle per 10 seconds with a maximum displacement of 0.025 inch. These loading rates do not actually represent field conditions; however, the purpose of the test is to be an
26

accelerated crack resistance test. The results of the test can be interpreted in two ways: (1) the reflective cracking life of an asphalt mixture, and (2) the fracture parameters of the mixture. The reflective cracking life of an asphalt mixture is defined as the number of cycles needed to propagate a crack through a specimen under a defined test condition. The plot formed by the load and displacement versus time results in three distinct phases, as shown in Figure 2.7.
FIGURE 2.7 Overlay Test Result, Zhou and Scullion (2005) Phase I is the crack initiation and steady propagation. The load reaches a maximum value before the displacement reaches a maximum, indicating a crack initiation at the bottom. The load then decreases rapidly as the crack propagates, but the load and displacement reach maximums at the same time, indicating a steady and slow propagating crack at the top surface. Phase II is the late-stage crack propagation represented as a saddleshaped load, which indicates the crack has partially gone through the entire cross section of the specimen. The first peak load is associated with minor adhesion, which rapidly decreases after breaking the weak adhesion bonds. Continuing the cyclic loading will break
27

the specimen and starts the beginning of phase III. Phase III has the crack propagated completely through the specimen. The reflective life cracking of the asphalt mixture can then be defined as the onset of phase II.
While the OT has been validated for reflective cracking, which is driven by crack propagation, the interest of this study is on fatigue cracking which centers on crack initiation and crack propagation. Work presented by Zhou and Scullion (2005) summarizes how crack initiation is related to crack propagation, which gives theory and then validation for the usefulness of the OT to be used for fatigue cracking. While the OT mainly characterizes crack propagation, the validated results from that study conclude that the OT can be used as a performance test for fatigue cracking. Previous studies show factors that affect the reflective cracking life on the OT include RAP percent, NMAS, binder content, and polymer-modified binders. An increase in asphalt content improved reflective cracking life of asphalt (Zhou and Scullion 2005). In the same study, binder grade PG 64-22 and PG 76-22 were investigated, and it was determined that the polymer-modified binder decreased reflective cracking life. Several studies investigated the effect of RAP on the OT and concluded that an increase in RAP decreases the reflective cracking life. Finally, a study looked at the NMAS and concluded that the smaller 9.5 mm NMAS compared to 12.5 mm NMAS increased reflective cracking life.
While some studies have shown good results from the OT, others have had challenges due to repeatability and variability issues (Walubita et al. 2012, Walubita et al. 2013). Issues with the OT came from a number of different sources. Walubita et al. (2012) discussed that one of the reasons for the large variability was not adhering to test specifications and procedures. Other factors the study found that contributed to the
28

variability of results was a function of sample fabrication and test setup. A consistent gluing method was also found to be crucial to improving variability (Garcia and Miramontes 2015). Recently, researchers have suggested an alternative way to interpret data from the OT, which resulted in the crack progression rate during the crack propagation phase (Garcia et al. 2017). The crack propagation can be quantified by fitting a power equation (29) to the load reduction curve. Figure 2.8 shows a typical load reduction curve.

y = ax

(29)

Where, a = 1 = crack progression rate

Normalized Peak Tensile Load (kN)

1.2 1
0.8 0.6 0.4 0.2
0 0

Normalized Load Reduction Curve

20

40

60

80 100 120 140

Cycles

FIGURE 2.8

Typical Load Reduction Curve for Overlay Test

29

The crack progression rate was shown to follow a trend in which a lower value for indicated better cracking resistance. Even with the studies that have been performed to help improve the repeatability and variability of the OT, concerns still remain regarding it reliably predicting fatigue cracking.
2.5 Summary of Test and Comparison Chart Table 2.1 compares the test procedure, loading mode, outcomes, and factors
affecting test from literature reviews.
30

31

Test Procedure SVECD (AASHTO TP107)

Loading Mode
Fingerprint Cyclical compression Fatigue Cyclical tension

SCB (AASHTO TP124)

Monotonic

TXOverlay Cyclical tension (TxDOT Tex-248-F)

DM (AASHTO T342)

Cyclical compression

Load Control
Fingerprint Strain 5075 micro strain Fatigue Strain Low micro strain Mid micro strain High micro strain Displacement 50 mm/min (1.97 in/min)
Displacement 0.025 in (0.06 cm)
Strain 5075 micro strain

TABLE 2.1 Comparison of Test Summary

Dimension, AV, Test Temperature 38 mm diam., 110 mm height
AV: 7.0 0.5%
Based on PG T= ((Hi Temp+Low Temp)/2) - 3

Outcome
Relationship between damage and the pseudo secant modulus to create the Damage Characteristic Curve

Factors Affecting Test from Literature Review Increased RAP decreases fatigue
resistance Increased binder content
increases fatigue resistance Increased air voids decreases
fatigue resistance

150 mm 1 mm diam., 50 mm 1 mm thick, cut in half (form two semi-circles) AV: 7.0 0.5% 25C (77F) 150 mm 2 mm diam., 38 mm height, 76 0.5 mm width AV: 7.0 0.5% 25C (77F)
38 mm diam., 110 mm height AV: 7.0 0.5% 4C, 20C, 40C (39.2F, 68F, 104F)

Fracture Energy (G), Flexibility Index (FI) for damage resistance
Susceptibility to fatigue or reflective cracking
Dynamic modulus (E*) and phase angle, master curve

Increased RAP decreases fracture energy and FI
Aging decreases FI Polymer-modified binder
decreases FI
Increased RAP percent decreases reflective cracking life (RCL)
NMAS of 9.5 mm increases RCL compared to 12.5 mm
Increased binder content increases RCL
PM binder decreases RCL
Increased RAP percent increases DM
PM binder increases DM Increased air voids decreases
DM

3. MATERIAL AND QUALITY CONTROL
The materials for this study were obtained from four hot mix asphalt production plants from different aggregates sources within the state of Georgia. Table 3.1 summarizes the materials used in this study.

TABLE 3.1 Plant Produced Mixture Properties

Specimen_ID A 19_64_N1 A 25_64_N1 A 12.5_67_N

NMAS (mm)
19
25
12.5

Binder Grade PG 64-22
PG 64-22
PG 67-22

A 12.5_76_N 12.5 PG 76-22

A 19_64_N2

19 PG 64-22

A 25_64_N2

25 PG 64-22

B 9.5_64_M1 9.5 PG 64-22

B 9.5_64_M2

9.5

PG 64-22

C 9.5_67_M

9.5 PG 67-22

A 12.5_64_M2 12.5 PG 64-22

A 12.5_64_M1 12.5 PG 64-22
B 12.5_64_M 12.5 PG 64-22

C 12.5_67_M 12.5 PG 67-22

C 12.5_76_M 12.5 PG 76-22

B 19_64_M

19 PG 64-22

B 25_64_M

25 PG 64-22

B 9.5_67_S

9.5 PG 67-22

B 12.5_67_S

12.5 PG 67-22

D 12.5_76_S 12.5 PG 76-22

RAP (%) 25 25 30
30 30 30 30 30
30
30
30 30
30 15 30 30
25
25
25

Binder (%)
4.6

Gmm 2.545

4.3 2.542

5.52 2.466

5.41 2.549 5.25 2.501 5.20 2.513 5.90 2.447 5.60 2.498

5.63 2.494

5.40 2.468

5.50 2.459 5.50 2.463

5.68 2.526 5.10 2.477 4.70 2.529 4.40 2.554

5.84 2.454

5.40 2.468

5.37 2.483

Air Void (%) 5.5 5.5 6.3
5.7 5.5 5.5 6.5 6.4
5.5
5.6
5.5 5.6
5.8 5.5 5.5 5.9
5.6
6.0
5.6

VMA (%) 14.7 15.0 18.0
18.4 17.1 16.7 19.3 18.1
17.8
17.7
17.7 18.0
17.3 16.8 15.8 15.3
18.4
18.1
17.5

VFA (%) 68.8 65.1 65.3
68.7 68.0 67.3 65.2 64.3
72.9
68.7
70.7 69.2
66.3 68.6 66.3 61.4
69.4
66.8
68.1

Effective Binder
(%) 10.1 9.8 11.8
12.6 11.6 11.2 12.6 11.6
12.9
12.2
12.5 12.5
11.5 11.5 10.5 9.4
12.8
12.1
11.9

Test Performed
|E*| |E*| |E*|, OT |E*|, SVECD, SCB |E*|, SVECD |E*| |E*|, OT |E*|, SVECD |E*|, SCB, OT |E*|, SVECD, SCB, OT |E*|, SVECD, SCB |E*| |E*|, SVECD, OT |E*|, OT |E*|, SVECD |E*| |E*|, SCB, OT |E*|, SVECD, SCB |E*|. SVECD, SCB OT

Note: Specimen ID labeled as X ##_##_X denotes Plant Source, NMAS, Binder Type, and Location.

32

For this study and the procurement of materials, the state was divided into three separate regions (i.e., North Georgia, Middle Georgia, and South Georgia) to differentiate the aggregate sources shown in Figure 3.1. The North Georgia region was subdivided into 1A and 1B due to 1A having limestone aggregates. The regions were divided based on soil support value (SSV) and climate differences. Plant A had material sourced from North and Middle Georgia. Plant B was sourced from Middle and South Georgia. Plant C was sourced from Middle Georgia, and Plant D was sourced from South Georgia. All of the mixtures except for A 19_64_N1 and A 25_64_N1 had granite aggregates, while those two had limestone aggregates. Three different binder types were used, which are PG 64-22, PG 67-22, and PG 76-22. All three binder types were used to create 12.5 mm NMAS mixtures. PG 64-22 and PG 67-22 were used to create three 9.5 mm NMAS mixtures. PG 64-22 was used for a 19 mm mixture and a 25 mmm mixture. Five mixtures had a RAP content of 25% and thirteen mixtures had a RAP content of 30% and one had a RAP content of 15%. Testing took place on the mixture available at the time. This is the reason that not all the mixtures presented in Table 3.1 were used in each test. More information can be found about relating selected mixtures and binders to MEPDG inputs in Appendix A and Appendix B, respectively.
33

FIGURE 3.1 Regional Separation of Plant Produced Mixtures
34

4. DYNAMIC MODULUS
4.1 Specimen Fabrication Developed under NCHRP Project 9-29, the asphalt mixture performance tester is
used in this study to measure dynamic modulus. AASHTO TP 79 (now AASHTO T378) was developed from the research performed in NCHRP Project 9-19 specifically for measuring dynamic modulus in an AMPT. The procedure for dynamic modulus testing in AASHTP TP 79 requires a specimen size of 100 mm in diameter by 150 mm in height. However, these dimensions are often impossible to obtain from field cores. In order to solve this problem, research was done to develop a test procedure and determine dynamic modulus from specimen sizes of 38 mm by 150 mm using an indirect tension test (Kim et al. 2004). Further research was performed to develop small-specimen dynamic modulus testing through uniaxial compression, and the results concluded that small and large specimens provided equivalent results (Kutay et al. 2009). Li and Gibson (2013) and Bowers et al. (2015) concluded that a specimen height of 110 mm provided the most consistent data. Lee et al. (2017) performed S-VECD tests using mixtures with an NMAS range from 9.5 mm to 25 mm, different binder types and gradations, and concluded the equivalence of the small specimen results with the large specimens. This study used small specimen geometry in accordance with NCHRP IDEA Project 181 (Castorena 2017).
All mixtures were compacted in a super gyratory compactor to a height of 178 mm and 150 mm in diameter. Small specimens were cored vertically from the inner 100 mm diameter of the large specimens. This provided 4 small specimens of 38 mm diameters from each 150 mm diameter specimen. The ends of the 38 mm specimens were sawed off to a height of 110 mm. The target air void of the compacted specimen was 7 0.5%. Cores
35

taken from the compacted specimen had their air voids measured, and the three with the most similar air voids were used.
4.2 Experimental Procedure The uniaxial compression test was performed in the AMPT at temperatures of 4C,
20C, and 40C. The loading frequency at each temperature was 25, 10, 5, 1, 0.5, and 0.1 Hz. Each specimen had mounting studs glued to the specimen with a gage length of 70 mm. Linear variable differential transformers (LVDT) were mounted to the specimen and a polytetrafluoroethylene sheet was used to reduce friction between the specimen and the loading platen shown in Figure 4.1. The allowed strain range was between 50 and 75 microstrains.
FIGURE 4.1 Dynamic Modulus Test Setup
36

4.3 Results and Analysis

Results from the test were used to create dynamic modulus master curves of each

mixture. Data collected from dynamic modulus testing were used to form the master

curves plotted in Figure 4.2. The data presented are the average for the three replicates.

1.0E+08

ALL PG 64-22

2.5E+07

ALL PG 64-22

|E*| (kPa)

1.0E+07 1.0E+06 1.0E+05 1.0E+04

A 25_64_N2 B 25_64_M A 19_64_N2 A 12.5_64_M1 B 12.5_64_M B 9.5_64_M1 B 9.5_64_M2

|E*| (kPa)

2.0E+07 1.5E+07 1.0E+07 5.0E+06

A 25_64_N2 B 25_64_M A 19_64_N2 A 12.5_64_M1 B 12.5_64_M B 9.5_64_M1 B 9.5_64_M2

1.0E+03 1.0E-04 1.0E-02 1.0E+00 1.0E+02 1.0E+04

1.0E+03 1.0E-04 1.0E-02 1.0E+001.0E+021.0E+04

(a)

Reduced Frequency (Hz)

(b)

Reduced Frequency (Hz)

1.0E+08

ALL PG 67-22

2.0E+07

ALL PG 67-22

1.0E+07 1.0E+06

1.6E+07 1.2E+07

A 12.5_67_N C 12.5_67_M C 9.5_67_M

|E'| (kPa)

|E'| (kPa)

1.0E+05 1.0E+04

A 12.5_67_N C 12.5_67_M C 9.5_67_M

8.0E+06 4.0E+06

1.0E+03 1.0E-04 1.0E-02 1.0E+00 1.0E+02 1.0E+04

1.0E+03 1.0E-04 1.0E-02 1.0E+00 1.0E+02 1.0E+04

Reduced Frequency (Hz)

Reduced Frequency (Hz)

(c)

(d)

|E* (kPa)

1.0E+08 1.0E+07 1.0E+06 1.0E+05 1.0E+04

ALL PG 76-22
A 12.5_76_N C 12.5_76_M D 12.5_76_S

|E* (kPa)

2.0E+07 1.6E+07 1.2E+07 8.0E+06

ALL PG 76-22
A 12.5_76_N C 12.5_76_M D 12.5_76_S

4.0E+06

1.0E+03 1.0E-04 1.0E-02 1.0E+00 1.0E+02 1.0E+04
Reduced Frequency (Hz)
(e)

1.0E+03 1.0E-04 1.0E-02 1.0E+00 1.0E+02 1.0E+04
Reduced Frequency (Hz)
(f)

FIGURE 4.2

Dynamic Modulus |E*| Results for Mixture with Different NMAS, Binder Types, RAP, and Aggregate Source

37

Statistical analysis was performed using an unequal variance t-test for the mixtures. The t-test was performed for all three temperatures at three different frequencies (i.e., 25, 5, and 0.5 Hz) to compare the difference in NMAS, binder type, and RAP content. The null hypothesis is that the dynamic moduli for different NMAS, binder grade, or RAP content are the same. The P-value was calculated and compared to the critical value of 0.05 to reject or accept the null hypothesis. A value greater than 0.05 indicates that the dynamic values are statistically the same.

4.4 NMAS P-values are summarized in Table 4.1 for different NMAS. In this comparison, the
only difference between the mixtures is the NMAS. The binder type and RAP content are the same for each mixture.

TABLE 4.1 Dynamic Modulus Comparison Against NMAS

25 mm vs 19 mm

Temperature 4C 20C

P-Value < 0.05 50% 50%

P-Value > 0.05 50% 50%

25 mm vs 12.5 mm

Temperature

4C 20C

P-Value < 0.05 50% 75%

P-Value > 0.05 50% 75%

19 mm vs 12.5 mm

Temperature

4C 20C

P-Value < 0.05 83% 83%

P-Value > 0.05 17% 17%

12.5 mm vs 9.5 mm

Temperature

4C 20C

P-Value < 0.05 17% 58%

P-Value > 0.05 83% 42%

40C 50% 50%
40C 75% 25%
40C 50% 50%
40C 0% 100%

Total 50% 50%
Total 67% 33%
Total 72% 28%
Total 25% 75%

38

For NMAS of 25 mm and 19 mm, this table shows that there was not statistical difference between dynamic modulus values. The same is true for the 12.5 mm and 9.5 mm NMAS. However, the NMAS of 25 mm and 19 mm compared to the 12.5 mm showed that the larger NMAS influenced the dynamic modulus.

4.5 Binder Type Table 4.2 summarizes the P-values between different binder types. These binder
types were compared against mixtures that had the same NMAS and RAP content.

TABLE 4.2 Dynamic Modulus Comparison Against Binder Type

PG 64-22 vs PG 76-22

Temperature 4C 20C 40C

P-Value < 0.05 58% 92% 50%

P-Value > 0.05 42% 8% 50%

PG 64-22 vs PG 67-22

Temperature

4C 20C 40C

P-Value < 0.05 42% 0% 25%

P-Value > 0.05 58% 100% 75%

PG 67-22 vs PG 76-22

Temperature

4C 20C 40C

P-Value < 0.05 50% 92% 100%

P-Value > 0.05 50% 8%

0%

Total 67% 33%
Total 22% 78%
Total 81% 19%

For binder types PG 64-22 and PG 67-22, the P-value was greater than 0.05 for 78% of the frequencies at all temperatures, meaning that there is no statistical difference between the PG 64-22 and PG 67-22 mixtures. However, the total amount of P-values that are below 0.05 when comparing PG 64-22 and PG 76-22 would suggest that there is a significant difference between the two binders. The same is true for PG 67-22 and PG 76-22. PG 76-22 is a stiffer binder, which is the reason for the higher moduli values.

39

4.6 RAP Content RAP content was compared between a mixture that had 15% RAP and mixtures that
had 30% RAP but the same NMAS and binder type. The results are shown in Table 4.3.

TABLE 4.3 Dynamic Modulus Comparison Against RAP Content

15% RAP vs 30% RAP Temperature 4C 20C 40C P-Value < 0.05 89% 100% 78% P-Value > 0.05 11% 0% 22%

Total 89% 11%

The results from the RAP comparison can conclude that RAP content greatly influences the dynamic modulus. This can be attributed to the RAP adding aged binder, which has hardened over time contributing to a stiffer AC mixture.

40

5. CYCLIC FATGUE TEST FOR S-VECD
5.1 Specimen Fabrication Loose plant-produced HMA mix was obtained and short-term aged at 20C below
compaction temperature to separate the mixture that arrived at the lab in covered metal buckets. The mixture was then heated to compaction temperature and then poured into molds to be compacted in a gyrator compactor. To obtain small specimen geometry, specimens were vertically cored from the gyratory-compacted specimen with a height of 178 mm and 150 mm in diameter in accordance with AASHTO TP107. The cored specimens have 38 mm diameters and four cores are obtained from each 150 mm diameter specimen shown in Figure 5.1. The ends of the 38 mm specimens were sawed off to a height of 110 mm, as shown in Figure 5.2.
FIGURE 5.1 SGC Compacted Specimen with Four Cores Taken from the Center
41

FIGURE 5.2 Small Specimen Cut from 178 mm to 110 mm The target air void of the compacted specimen was 7 0.5%, and the cored specimens had an air void of 6 0.5%. Four cores taken from the compacted specimen had their air voids measured, and the three with the most similar air voids were used. 5.2 Experimental Procedure With the prepared specimens, mounting studs were glued to the specimen at a gage length of 70 mm, and the end plates were glued to the specimen as shown in Figure 5.3.
42

FIGURE 5.3 Small Specimen with Glued End Plates Then, the specimen was inserted into the AMPT machine and the bottom support was tightened. The actuator applied a seating of 0.01 kN for the purpose of securing the upper loading platen with screws. Feeler gauges were used to ensure proper leveling of the specimen. The feeler gauges and the way the screws hold the test specimen in place is shown in Figure 5.4.
43

FIGURE 5.4 Secured Test Specimen with Feeler Gauges

The load was reduced to zero, which was the starting load for the test, and LVDTs were mounted to the specimen. The test temperature was based on LTPPBind Version 3.1 (AASHTO TP107, 2014) and used Eq. (30).

Test Temperature (C) = (if + - 3 21, + - 3 otherwise, 21C) (30)

2

2

Where, TH = high-temperature PG Grade from LTPPBind (C) TL = low-temperature PG Grade from LTPPBind (C), generally a negative number
Once the specimen reached the target temperature, a dynamic modulus (|E*|) fingerprint test was performed at a frequency of 10 Hz and a target strain range of 5075 microstrains. Following the fingerprint test, the specimen rested for 20 minutes. The cyclic fatigue test was started with the peak-to-peak specimen strain amplitude of 300, 500, or

44

800 microstrains based on the |E*|fingerprint ranges. The test was terminated when the phase angle began to drop.

5.3 Damage Characteristic Curves C versus S curves were constructed with the aid of FlexMATTM software. Three
replicates were used to construct the curves and a single model was fitted via the curves. Figure 5.5(a) shows the accumulated damage between polymer-modified binder PG 76-22 and unmodified binder PG 64-22. The C versus S curves for the mixtures with polymermodified PG 76-22 binder are higher than the curves for the mixtures with PG 64-22 binder. This outcome is expected because the modulus heavily influences these curves. The polymer-modified binder is stiffer than the unmodified binder, thus leading to a higher dynamic modulus value and a higher C versus S curve. Figure 5.5(b) shows the effect of the NMAS on the C versus S curves; the smaller 9.5 mm mixture with PG 64-22 binder has a higher curve than the 12.5 mm mixture with PG 64-22 binder.

1.0

1.0

D

A

0.8

12.5_76_N A

0.8

12.5_64_M1 B 9.5_64_M

0.6

12.5_76_N 0.6

C C

0.4

0.4

0.2

0.2

0.0 0
(a)

0.0

100000 S 200000

300000
(b)

0

100000 S 200000 300000

FIGURE 5.5

(a) C vs. S Curves for 12.5 mm Mixtures with PG 76-22 and PG 64-22 Binders, and (b) C vs. S Curves for 9.5 mm and 12.5 mm Mixtures with PG 64-22 Binder and 30% RAP

45

5.4 DR Failure Criterion
The following figures were arranged based on expected results of binder type and NMAS. The expected trend would be a decrease in DR from high to low binder grade and a decrease with a larger NMAS. Figure 5.6(a) shows the trends for the mixtures with polymer-modified PG 76-22 binder that have a higher DR value than the mixtures with unmodified PG binders. The surface mixtures also have a higher DR value than the base mixtures. The DR values for the 9.5 mm and 12.5 mm surface mixtures are both higher than the larger 19 mm and 25 mm mixtures, as seen in Figure 5.6(b). These mixtures contained the same PG 64-22 binder. The 9.5 mm mixtures generally have the same DR values as the 12.5 mm mixtures with the same performance grade binder, i.e., PG 64-22. Both of these results show expected trends; i.e., the polymer-modified binder and smaller NMAS present higher DR values that indicate better cracking resistance. Table 5.1 includes the R2 values for DR and the standard deviation for Sapp for all mixtures.

DR vs. PG
0.6 0.5 0.4 0.3 0.2 0.1
0

DR vs. NMAS
0.5 0.4 0.3 0.2 0.1
0

DR DR

(a)

(b)

FIGURE 5.6
(a) DR Failure Criterion Used to Compare Polymer-Modified PG 76-22 Binder with Unmodified PG 67-22 and PG 64-22 Binders, and (b) DR Failure Criterion Used to Compare 9.5 mm,
12.5 mm, 19 mm, and 25 mm Mixtures

46

Table 5.1

DR and Sapp Values with the Corresponding R2 and Standard Deviation Values

NMAS Binder Specimen_ID (mm) Grade DR R2 Sapp SD

B 9.5_64_M 9.5 PG 64-22 0.48 0.99 10.36 0.62

B 9.5_67_S

9.5 PG 67-22 0.50 1

9.80 1.04

A 12.5_64_M1 12.5 PG 64-22 0.48 1 9.37 0.23

A 12.5_64_M2 12.5 PG 64-22 0.49 1 8.90 0.53

B 12.5_67_S 12.5 PG 67-22 0.47 1 9.02 0.17

C 12.5_67_M 12.5 PG 67-22 0.45 1 9.94 0.61

A 12.5_76_N 12.5 PG 76-22 0.52 1 11.81 0.69

D 12.5_76_S 12.5 PG 76-22 0.55 0.99 15.40 0.49

A 19_64_S

19 PG 64-22 0.41 1 8.72 0.95

B 25_64_M

25 PG 64-22 0.42 1

8.57 0.87

5.5 Sapp Cracking Index Binder type is an important mixture characteristic that a fatigue index should
accurately reflect. Mixtures in Georgia typically use PG 64-22 or PG 67-22 binder for normal traffic loading conditions and polymer-modified PG 76-22 binder for heavy traffic. The 12.5 mm mixtures were used to investigate the effect of binder type on Sapp. Figure 5.7(a) shows the differences between the three binder types. The Sapp values for PG 67-22 are slightly higher overall compared to those for PG 64-22. However, the Sapp values for the PG 76-22 binder are significantly higher than for PG 64-22 or PG 67-22 binder. This finding implies that Sapp can accurately rank fatigue resistance based on binder type. This finding also agrees with the results shown in Figure 5.6(a), where polymermodified binders have higher DR values. Mixtures with different NMAS values were investigated to determine the effects of NMAS on Sapp and fatigue cracking resistance. Figure 5.7(b) presents the mixtures with different NMAS values. All mixtures consist of 30% RAP and PG 64-22 binder with differences in NMAS for 9.5 mm, 12.5 mm, 19 mm,
47

and 25 mm. This figure shows that an increase in the NMAS corresponds to a decrease in Sapp value. This outcome again agrees with Figure 5.6(b), where DR is shown to decrease with an increase in NMAS. DR shows similar trends to Sapp, but it is unable to be used by itself to rank mixtures because it only measures toughness, while Sapp combines toughness and moduli.

Sapp Sapp

Sapp vs PG
16 14 12 10 8 6 4 2 0

Sapp vs NMAS
10 8 6 4 2 0

(a)

(b)

FIGURE 5.7

(a) Sapp Values for Mixtures with Different NMAS Values, and (b) Sapp Values for 12.5 mm Mixtures with PG 64-22, PG 67-22, and PG 76-22 Binders

The effect of binder content on fatigue cracking/resistance also was investigated by observing how different binder contents would affect Sapp values. Mixtures with the same NMAS and binder type were chosen for comparison. Figure 5.8 compares 12.5 mm mixtures with different binder types, i.e., PG 67-22 and PG 64-22. For both PG 67-22 and PG 64-22 binders, the Sapp value increased as the binder content increased. This result is intuitive, as an increase in the binder percentage would generally result in a softer AC mixture. No significant difference in Sapp values due to the half-grade difference of binder was observed.

48

Sapp

10 9.8 9.6 9.4 9.2
9 8.8
5.3

Sapp vs. Binder Content

PG 64-22 PG 67-22

5.4

5.5

5.6

5.7

5.8

Binder Content %

FIGURE 5.8
Sapp Values for Mixtures with Different Binder Contents for the Same NMAS and Binder Type

Table 5.2 suggests the threshold values for Sapp along with different traffic levels (Wang and Kim 2017). The Sapp values from Figure 5.8 indicate that all mixes should satisfy S designation and D 12.5_76_S satisfies H designation, which has traffic level between 3 and 30 million ESALs (equivalent single axle loads).

TABLE 5.2

Recommended Threshold Values for the Sapp Fatigue Index Parameter

Traffic Level (million ESALs)

Sapp

Tier

Designation

3

8

Light

L

>3 and 10

>8 and 15

Standard

S

>10 and 30

>15 and 20

Heavy

H

>30

>20 and 25

Very Heavy

V

>30 and slow traffic

>25

Extremely Heavy

E

49

To compare the traffic level shown in Table 5.2 against one suggested by GDOT's practical guideline for specific mixtures, Table 5.3 was developed. Table 5.3 presents the calculated ESALs and measured Sapp values for the 12.5 mm NMAS mixture in Georgia. GDOTallows 12.5 mm Superpave mixtures with PG 64 or PG 67 binders for the two-way average daily traffic (ADT) between 10,000 to 25,000 while 12.5 mm mixtures with polymer-modified binder is allowed for the two-way ADT between 25,000 and 50,000. Since GDOT uses two-way average daily traffic to select mixture type, the two-way ADT was converted into ESALs assuming 5% truck traffic, 1.17 for ESAL factor, and 1.0 for lane distribution factor.

TABLE 5.3
Recommended Threshold Values for the Sapp Based on GDOT Mixture Selection Criteria (Assuming 5% Truck Traffic and 1.17 ESAL Factor)

Two-way ADT
10,000 25,000
25,000 50,000
>50,000

Traffic Level (million ESALs)
>4 and 10
>10 and 20
>20

Sapp (from Test Results) >12
>15.5
N/A

Mix Type

Remarks

12.5 mm Superpave with
PG 64-22 or PG 67-22
12.5 mm Superpave with
polymer modified binder
12.5 mm Stone Matrix Asphalt

For State Routes and for shoulders of Interstate Routes
For high ADT State Routes, Interstate Routes when recommended by GDOT, all flexible pavement Interstate Ramps, and all flexible pavement roundabouts For Interstate Routes and for State Routes when recommended by GDOT

As shown in Table 5.3, all mixes used in this study are adequate for the traffic level between 4 and 20 million ESALs. Although the recommended threshold values at different traffic level in Table 5.2 and Table 5.3 show reasonable agreement, it is still based on the

50

calculated ESALs from two-way ADT with assumption of percent truck. Therefore, it is suggested that each state agency develop their own Sapp threshold criteria reflecting the state agency's mixture selection criteria and practical guideline.
5.6 Fatigue Performance Simulation In this study, the goal of fatigue cracking performance simulation is to determine the
practical application of available pavement evaluation programs in ranking AC mixture performance using a GDOT-approved pavement section design. This study used AASHTOWare Pavement ME (Ver. 2.3.1) and FlexPAVETM to simulate fatigue cracking performance.
Pavement ME was originally developed under National Cooperative Highway Research Program (NCHRP) Project 1-37A and uses layered elastic theory and empirical models to determine fatigue damage and permanent deformation (Advanced Research Associates, Inc. 2004). Pavement ME has a number of global calibration factors that can be adjusted to meet local calibration factors. GDOT developed local calibration factors for rutting, bottom-up fatigue cracking, and thermal cracking.
FlexPAVETM, developed by North Carolina State University researchers, uses the finite element method to predict pavement distresses. The S-VECD model is used in FlexPAVETM to predict fatigue damage throughout the pavement design life. Damage is calculated in FlexPAVETM through two overlapping triangles that form the reference cross section area (Wang and Kim 2018). The top inverted triangle has a base of 170 cm, while the bottom triangle has a base of 120 cm. Each triangle calculates damage, which allows for top down cracking and bottom up cracking to be separated. Figure 5.9 shows the overlapping triangles and the separation into the top and bottom for cracking purposes. The
51

program outputs damage contours that show the location of damage and the degree of damage. Recent research has shown that FlexPAVETM along with DR can reasonably predict and rank pavement performance (Wang and Kim 2018).
FIGURE 5.9 Damage Area Used in FlexPAVETM Separated into Top and Bottom for Cracking Reference 5.7 Section Design Two different pavement sections were considered in this study to determine pavement performance based on DR and Sapp values. The first was a single-layer pavement section 4 inch (10.16 cm) thick. A 12 inch (305 mm) unbound aggregate base was used under the asphalt layer. Underneath the aggregate base was a subgrade that was considered to be a semi-infinite layer, and modulus values were used that are specific to the typical soil in Georgia. To investigate the effect of the surface mixtures on the fatigue cracking performance of a pavement, this single layer was changed between the various 9.5 mm and 12.5 mm
52

NMAS mixtures. The results from the cyclic uniaxial fatigue test and the dynamic modulus test were used in FlexPAVETM for pavement performance evaluation. Pavement ME requires mixture dynamic modulus data along with binder complex shear modulus (G*) data for Level 1 analysis. Traffic inputs were based on actual traffic data from an approved GDOT design. FlexPAVETM uses the daily equivalent single-axle load from the approved design report and the vehicle speed. Pavement ME currently offers additional options for traffic inputs. A two-way average annual daily truck traffic count based on the design report was input for the number of lanes, truck percentage in the design lane, operational speed, and traffic load distribution. Climate data specific to the state of Georgia, which were available for both programs, also were used.
The second design that was used to compare the programs was a two-layer pavement section. The design consisted of a surface layer (12.5 mm NMAS) and a bottom layer (19 mm NMAS) with thicknesses of 3 inch (7.62 mm) for each, providing a total thickness of 6 inch (15.24 mm). The aggregate base and subgrade were left the same at 12 inch (305 mm) and the subgrade layer was considered to be infinite in the depth direction. The traffic conditions were applied in the programs the same way as for the single-layer section.
5.8 Results from Simulation Figure 5.10 shows an example of the damage contours for one of the 4 inch
pavement sections, to illustrate the growth of damage predicted by the FlexPAVETM software. This figure was created using mixture A 12.5_64_M2 and shows the damage due to top down and bottom up cracking.
53

FIGURE 5.10 Damage Contours Based on Pavement Section with 12.5 mm NMAS
and PG 64-22 Binder
Figure 5.11(a) through (i) show the fatigue cracking predicted by FlexPAVETM of the 4 inch pavement section over the design life of 20 years and their correlations with Sapp and DR. Figure 5.12(a) through (i) show the same for the 6 inch two-layer section. Figure 5.13(a) through (d) show the top down and bottom up cracking predicted by Pavement ME over the design life of 20 years for the 4 inch section and the two 3 inch layer section. The 4 inch pavement section in FlexPAVETM showed very good trends for a decrease in damage with an increase in binder type. It also showed that the mixtures with lowest Sapp values had the most damage and those with the highest Sapp values had the least. The correlation for Sapp also shows why DR alone is not sufficient to rank asphalt mixtures. Sapp has a much higher R2 value than DR when compared against percent damage. When separated into top down and bottom up damage, it can be seen that the polymer-modified mixtures reduced cracking for both of these instances. In the case of the 6 inch two-layers, the total percent damage were very similar to each other according to FlexPAVETM; however, once separated into top down and bottom up, it is seen that the higher binder
54

grades were able to reduce top down cracking. Pavement ME had similar results for both the 4 inch section and the 6 inch two-layer section. For both pavement sections, Pavement ME showed that the polymer-modified binders had the least amount of top down and bottom up cracking. Pavement ME shows that the most cracking would occur in the 9.5 mm NMAS mixture, while FlexPAVETM has the 12.5 mm NMAS PG 64-22 as experiencing the most cracking.
55

56

FIGURE 5.11 FlexPAVETM Predicted Fatigue Cracking for 4 Inch Pavement Section: (a)(c) Total Percent Damage,
(d)(f) Top Down Percent Damage, and (g)(i) Bottom Up Percent Damage

57

FIGURE 5.12 FlexPAVETM Predicted Fatigue Cracking for 6 Inch Two-Layer Pavement Section: (a)(c) Total Percent Damage,
(d)(f) Top Down Percent Damage, and (g)(i) Bottom Up Percent Damage

FIGURE 5.13 Pavement ME Predicted Fatigue Cracking: (a) 4 Inch Layer Top Down Cracking, (b) 4 Inch Layer Bottom Up Cracking, (c) Two 3 Inch Layers Top Down Cracking, and (d) Two 3 Inch
Layers Bottom Up Cracking
58

6. SEMICIRCLE BEND TEST
6.1 Specimen Fabrication Specimens were made from plant-produced loose mix asphalt in a super gyratory
compactor in accordance with AASHTO T312, "Standard Method of Test for Preparing and Determining the Density of Asphalt Mixture Specimens by Means of the Superpave Gyratory Compactor (SGC)" (AASHTO T312, 2015). Specimens were made to a height of 178 mm at a target air void of 7 0.5%. These SGC specimens were then cut as shown in Figure 6.1 to obtain two 50 mm thick discs from the middle of the specimen shown in Figure 6.2. Each disk was then cut in half to create four semicircle shapes, and a 15 mm notch was made at the center of the specimen as shown in Figure 6.3.
FIGURE 6.1 SGC Specimen Prepared to be Cut into Two 50 mm Disks
59

FIGURE 6.2 50 mm Disk Cut from an SGC Specimen
FIGURE 6.3 Semicircle with 15 mm Notch in Center
60

6.2 Experimental Procedure Using the newly developed I-FIT test (AASHTO TP124), the materials were tested
to determine how their mixture properties lend to their fracture resistance. The test was specifically looking at evaluating different percentages of RAP, binder types, aggregate sources, and NMAS. Prior to the I-FIT test, each mixture was subjected to quality control tests including theoretical maximum specific gravity (Gmm), bulk specific gravity (Gmb), asphalt content by ignition oven, and sieve analysis. The measurements were compared to the job mix formulas (JMFs) supplied by the producers since it is important for the state agency to understand how its approved mixes, which come from different regions of the state and are made of different binder types, perform comparative to each other. This will benefit those who create mix designs to select the optimum mix for the particular project. For this study, eight different GDOT-approved plant mixes were obtained and fabricated into specimens as shown in Table 6.1.

NMAS Specimen_ID (mm)

Binder Grade

TABLE 6.1 Mixtures Used for SCB Testing

RAP Binder Gmm (%) (%)

Air Void (%)

VMA (%)

VFA (%)

A 12.5_76_N

12.5 PG 76-22 30 5.41 2.549 5.7 18.4

68.7

C 9.5_67_M

9.5 PG 67-22 30 5.63 2.494 5.5 17.8

72.9

A 12.5_64_M2 12.5 PG 64-22 30 5.40 2.468 5.5 17.7

68.7

A 12.5_64_M1

12.5 PG 64-22 30 5.50 2.459 5.5 17.7

70.7

B 9.5_67_S

9.5 PG 67-22 25 5.84 2.454 5.5 18.2

70.3

B 12.5_67_S

12.5 PG 67-22 25 5.40 2.468 6.0 18.1

66.8

D 12.5_76_S

12.5 PG 76-22 25 5.37 2.483 5.5 17.4

68.6

Effective Binder
(%)
12.6 12.9
12.2
12.5 12.8
12.1
11.9

Test Performed
|E*|, SVECD, SCB
|E*|, SCB, OT |E*|, SVECD,
SCB, OT |E*|, SVECD,
SCB |E*|, SCB, OT |E*|, SVECD,
SCB |E*|. SVECD,
SCB OT

61

The test was run in an asphalt mixture performance tester (AMPT) in accordance with the I-FIT procedure at 25C with four (4) replicates for each mix type. The test had an initial contact load of 0.1 kN and used line load displacement control at a rate of 50 mm/min. The test terminated when the load dropped below 0.1 kN.
6.3 Results and Analysis Figure 6.4 shows the results of the I-FIT tests. As shown in Figure 6.4(a), the fracture
energy values ranged from 1333 to 2521 J/m2. The FI had values from 1.5 to 5.2, as shown in Figure 6.4(b). The coefficient of variation (CV) for FI was between 5% and 26% with an average of 15%. The fracture energy of the mixtures had a CV between 4% and 20% with an average of 11%. Interestingly, the mixtures with polymer-modified binder (PG 76-22) show lowered fracture energy and FI compared to others, which was unexpected.
The eight mixtures varied in four distinct categories: RAP content, binder type, aggregate source, and NMAS. Comparisons between mixtures needed to be appropriate in order to draw conclusions about the local fracture characteristics of plant-produced AC mixtures in Georgia. Due to the mixtures being plant-produced as opposed to laboratoryproduced, the mixtures had variability in volumetric properties. Before analysis was performed on the mixture characteristics of interest, analysis of variance (ANOVA) was performed on voids in the mineral aggregate, voids filled with asphalt, and percent of air voids. The results of this analysis are presented Table 6.2.
62

Fracture Energy (J/m2)

2200 2000 1800 1600 1400 1200 1000 800 600 400 200
0
(a) Fracture Energy 5
4
3
2
1
0
(b) Flexible Index FIGURE 6.4
I-FIT Results for All Mixtures (a) Fracture Energy and (b) Flexible Index
63

Flexibility Index

TABLE 6.2 Analysis of Variance of VMA, VFA, and AV for All Mixtures at 95% Confidence Interval

Variable
VMA VFA AV

FI P-Value
0.370 0.389 0.912

Responses Fracture Energy P-Value 0.108 0.788 0.866

Based on the ANOVA of VMA, VFA, and AV, it was determined that these three volumetric properties did not have any significant influence on fracture energy or FI for this study. Knowing these three variables did not create differences within the responses, further analysis was possible on the mixture characteristics of interest. ANOVA was performed on the results from the different mixtures, presented in Table 6.3, with the variables being binder type, RAP content, aggregate source, NMAS, and the response fracture energy and FI. The order of analysis was important in this study because of the variable mixture designs. The study was limited to the material being produced by plants at the time. This led to analysis of the results in a systematic manner in order to filter out mixture characteristics that did not influence FI or fracture energy. NMAS was the first variable selected because it provided the most mixtures to analyze that had consistent variables. Aggregate source was analyzed second for similar reasons as NMAS. These two analyses provided the following findings:
Aggregate sources had no significant difference in FI or fracture energy. NMAS was not found to be significant in fracture energy or FI between the 9.5 mm
and 12.5 mm surface mixes.

64

Since these two mixture characteristics did not have any influence on FI or fracture energy, further analysis was possible on binder type and RAP content. In review of the literature, numerous researchers showed that RAP content has a significant influence on fracture energy and FI (Norouzi et al. 2017, Cascione et al. 2015, Kim, Mohammad, and Elseifi 2012). Based on those studies, RAP content was analyzed across data groups after it was determined aggregate source and NMAS were inconsequential. Binder type was then compared between mixtures that had the same RAP content. The analysis of those two mixture characteristics shows as follows:
PG was seen to have significant impact on FI, but fracture energy failed to discriminate between mixtures.
Mixtures with 25% and 30% RAP content are not significant for fracture energy or FI, since this was due in part to the different binder types controlling the response, as the change in RAP content was only 5%.

TABLE 6.3
Analysis of Variance of Binder Type, RAP Content, Aggregate Source, and NMAS with a 95% Confidence Interval

Variable
NMAS Aggregate Source

FI P-Value
0.312 0.075

Responses Fracture Energy P-Value 0.911 0.087

RAP Content

0.404

0.819

Binder Type

0.004

0.110

Because ANOVA does not show where the difference in the data lies, paired t-tests were performed. A paired t-test was performed for the difference in RAP content, as well,
65

in order to determine if the nonresponse found in the ANOVA was due to the binder variable. The results of the t-test are shown in Table 6.4.

TABLE 6.4 Results of Paired t-test with a 95% Confidence Interval

Variable
RAP 25% and 30% PG 76-22 and PG 64-22

Response

FI

Fracture

T-Value P-Value T-Value P-Value

4.50

0.0102 0.983

0.199

4.27

0.003

2.38

0.019

PG 76-22 and PG 67-22

4.21

<0.001

2.90

0.007

PG 67-22 and PG 64-22

0.367

0.359

0.237

0.408

From the t-test, the following observations were made: Fracture energy failed to discriminate the fatigue-resisting performance between 25% and 30% RAP mixes, while FI is statistically significant to differentiate fatigue-resisting performance of mixtures. This t-test allowed for the exclusion of binder type, which resulted in nonresponses in the ANOVA. FI values between mixtures with PG 64-22 and PG 76-22 had significant difference, as well as PG 67-22 and PG 76-22. There was no significant difference between PG 67-22 and PG 64-22 in FI or fracture energy. The results of lowered FI for the mixtures with polymer-modified binder (PG 76-22) compared to other mixtures with softer binders (PG 64-22 and PG 67-22) was unexpected.

66

7. MODIFIED OVERLAY TEST
7.1 Specimen Fabrication Specimen fabrication for the modified overlay test was created from a super gyratory
compacted specimen at a height of 178 mm and a diameter of 150 mm. Two specimens were cut from a single SGC specimen to a height of 38 mm and a diameter of 150 mm in accordance with Tex-248-F). A template was used to obtain an accurately cut specimen perpendicular to the top surface, resulting in a width of 76 mm. The specimen was glued to the base plates with weights placed on top of the specimen while the glue cured. Figure 7.1 shows the geometry of the cut specimens.
FIGURE 7.1 Test Specimens for Modified Overlay Test 7.2 Experimental Procedure The specimen was placed into the AMPT after the glue was cured, and the target temperature of 25C had been reached. The specimen was secured using a torque wrench by applying 1.7 Newton-meter (15 lb-in) to each bolt in a specific tightening pattern, as
67

shown in Figure 7.2. The numbers indicate the order that the screws should the tightened. The test was started and ran until 93% reduction of the maximum load occurred or the test completed 1000 cycles.

3

5

FIGURE 7.2 Tightening Pattern for Bolts Tex-248-F

7.3 Results and Analysis The load reduction curves for each mixture were normalized by the maximum load
of the first cycle shown in Figure 7.3. The power curve in Eq. (30) was fitted to the load reduction curves to obtain the value . The value for , or the crack progression rate, was calculated for each specimen that was tested and presented in Table 7.1, along with the cycles to failure. The crack progression rate is presented as an absolute value.

68

Normalized Peak Tensile Load (kN)

1.2

1

0.8

0.6

A 12.5_64_M1

0.4

Beta Power Fit

0.2

0

0

20

40

60

80 100 120 140

Cycles

FIGURE 7.3

Normalized Load Reduction Curve with Fitted Power Curve

TABLE 7.1 Crack Progression Rate and Cycles to Failure for Various AC Mixtures

Mix ID
B 9.5_64_M1 B 9.5_67_S C 9.5_67_S A 12.5_64_M A 12.5_67_N B 12.5_67_M C 12.5_67_M C 12.5_76_M D 12.5_76_S

Crack Progression Rate

Mean

SD

COV

0.37

0.03

7%

0.37

0.00

0%

0.46

0.09

20%

0.56

0.11

19%

0.37

0.03

7%

0.74

0.30

41%

0.39

0.03

8%

0.34

0.03

7%

0.71

0.24

34%

Cycles to Failure

Mean SD COV

226

10

4%

172

12

7%

213

29 13%

110

23 21%

1000

0

0%

316 285 90%

152

58 38%

594 406 68%

67

56 83%

As shown in Table 7.1, the coefficient of variation is as high as 41% for the crack progression rate and 90% for the cycles to failure. These high COVs suggest the test results are not reliable for its repeatability. Figure 7.4 and Figure 7.5 show factors influencing
69

Crack Progression Rate

crack progression. There appears to be a clear trend for a lower NMAS having a lower , which indicates better crack resistance. However, Figure 7.4 shows an opposite trend to what would be expected for different binder types. The PG 64-22 binder type performed better than the PG 67-22 and the polymer-modified PG 76-22. The large stand error bars suggest the repeatability is unreliable.
1.20 1.00 0.80 0.60 0.40 0.20 0.00
B_9.5_64_M1 B 9.5_67_S A 12.5_64_M B 12.5_67_M FIGURE 7.4
Crack Progression Rate Vs. NMAS

Crack Progression Rate

1.20

1.00

0.80

0.60

0.40

0.20

0.00 A 12.5_64_M

B 12.5_67_M

D 12.5_76_S

FIGURE 7.5 Crack Progression Rate Vs. Binder Type

70

The cycles to failure have a slight correlation to the crack progression rate, as shown in Figure 7.6. In general, as the cycles to failure increase, the crack progression rate decreases, which would be the expected trend. The crack progression rate does offer more insight to the OT, but the issues of reliability and repeatability damper the OT results from offering valuable data.

Crack Progression Rate

0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00
0

R = 0.2635
200 400 600 800 1000 1200 Cycles to Failure

FIGURE 7.6 Cycles to Failure Vs. Crack Progression Rate

71

8. CONCLUSIONS
In this study, the effects of NMAS, RAP content, binder type, and binder content on mixture characteristics and fatigue cracking resistance performance were investigated. Categorical mixtures used in Georgia were fabricated and tested for dynamic modulus (|E*|) and fatigue cracking potential measurements using direct cyclic tension, semicircular bending, and modified overlay test methods. Based on laboratory tests and analyses, the following conclusions were made.
8.1 Dynamic Modulus Generally, Superpave mixtures with higher PG binder and increased RAP content
(up to 30% RAP) result in a higher dynamic modulus master curve. Dynamic modulus values for NMAS between 25 mm and 19 mm were not
significantly different. The same conclusion is made for 12.5 mm and 9.5 mm mixtures. Values for dynamic modulus were significantly different between 12.5 mm and 19 mm, as well as 12.5 mm and 25 mm, with P-values less than 0.05 for 70% of the master curves. Binder type influenced dynamic modulus values, with the stiffer PG 76-22 binder being significantly different with P-values less than 0.05 for 74% of the master curves from both PG 64-22 and PG 67-22. However, there was not a notable difference between PG 64-22 and PG 67-22. RAP content had a great effect on the dynamic modulus between 15% and 30% RAP contents with P-values less than 0.05 for 89% of the master curve.
72

8.2 Direct Tension Cyclic Fatigue Test Using S-VECD Model Controlled crosshead tension cyclic fatigue tests were performed to investigate the
fatigue performance of mixtures with different mixture properties. Cracking parameter, Sapp, was used to rank fatigue cracking performance of asphalt
mixtures commonly used in Georgia and investigate correlations between mixture properties and fatigue cracking performances. Sapp is a theoretically sound index parameter that can be used to predict and rank the fatigue cracking performance of asphalt mixtures. The results show that Sapp adequately reflects the effect of mixture properties on fatigue cracking performance. Sapp is significantly influenced by NMAS and binder type. Sapp values decreased with larger aggregate and increased for polymer-modified asphalt mixtures. Binder content also was shown to affect Sapp, with a higher binder content leading to a higher Sapp value. The DR failure criterion also is influenced by NMAS and binder type. The mixtures with larger aggregate had lower DR values, indicating that cracking resistance would be lower, too. The mixtures with modified binders showed the highest DR values in this study. The trends of Sapp as a function of binder type and NMAS are clearer than those of DR. Therefore, Sapp is recommended as the cracking index property. Both FlexPAVETM and Pavement ME showed that the polymer-modified mixtures performed the best in terms of fatigue cracking resistance. Reasonable correlations were found between Pavement ME and FlexPAVETM analyses for the top-down and bottom-up cracking predictions for a 10.2 cm (4-
73

inch) thick single-layer pavement, whereas the correlations were poor for a 15.2 cm (6-inch) two-layer pavement. The poor correlations for the two-layer pavement are attributable to the fact that the cracking prediction in Pavement ME depends solely on the modulus of the top layer, whereas FlexPAVETM uses layer-specific modulus and fatigue properties throughout the asphalt layers to estimate fatigue cracking resistance.
8.3 Semicircle Bend Test Fracture energy alone is not enough to discriminate fatigue resisting mixture
performance, while the newly developed FI is able to determine significant differences between mixture performances. Between two surface mixes (i.e., 9.5 mm and 12.5 mm NMAS mixtures), there was no significant difference in FI or fracture energy. AC surface mix from different locations in Georgia showed to have no significant difference between their fracture resistant properties. Notch width significantly affects FI. With an increased notch width from 1.5 mm to 3.5 mm, it was observed that FI was reduced up to 75%. The test results for mixtures with polymer-modified binder (i.e., PG 76-22) lowered FI, which was unexpected.
8.4 Modified Overlay Test This study was conducted to determine if the OT could give reliable and repeatable
results to rank AC mixtures for cracking resistance. The materials used were sourced from the state of Georgia and were composed of different mixture characteristics in order to
74

determine if the OT accurately captures the effect on pavement performance. Based on laboratory tests and analyses, the following conclusion were made:
The crack progression rate provided more useful information than the cycles to failure and gave a trend of a decreasing crack progression for smaller NMAS mixtures, indicating less cracking.
With high variability of the OT results, it was challenging to identify the relationship between OT results and AC mixture properties, although the variability of the test results may be attributed to the gluing method during the test setup stage. This variability of the OT results could make this test method less favorable to predict fatigue cracking potential of AC mixtures.
8.5 Fatigue Test Method Comparisons For the fatigue tests, the SCB and cyclic direct tension tests with S-VECD model
provide consistent test results that could be used in identifying AC cracking potential. The advantages of the SCB test over S-VECD are simple sample fabrication, ease of operation, and quick testing time. The cyclic direct tension test with S-VECD model provides more theoretically sound in-depth information to better understand AC mixture behavior. On the other hand, the cyclic direct tension test with S-VECD model requires intensive training to complete a successful test compared to the SCB test. This concern could be overcome through lab training and a workshop at the University of Georgia upon GDOT's request.
75

9. RECOMMENDATIONS AND FUTURE WORK
Future study to predict AC fatigue cracking should focus on the investigation of cracking performance using field-cored specimens to compare against fatigue index rankings and pavement performance. For this task, a draft standard operating procedure (SOP) was developed to evaluate fatigue cracking resistance performance of asphalt mixture. The SOP is provided in Appendix C.
With a larger database of dynamic modulus values created, the MEPDG can be implemented for design of flexible roadways. Its implementation would be most successful with training of staff and personnel about the inputs needed for the MEPDG. Having a firm background about these inputs and their significance will help GDOT use the MEPDG successfully in their designbuild projects. For successful MEPDG implementation, Pavement ME needs accurately calibrated coefficients of AC mixtures.
Future studies should focus on investigating cracking performance using fieldcored specimens and comparing pavement condition surveys. Based on the field evaluations, index parameter criteria to select appropriate mixtures for field construction could be refined for design traffic.
76

10. REFERENCES
American Association of State Highway Transportation Officials (AASHTO) R83 (2018), "Preparation of Cylindrical Performance Test Specimens Using the Superpave Gyratory Compactor (SGC)." Washington, D.C.: AASHTO
AASHTO R84 (2017), "Practice For Developing Dynamic Modulus Master Curves For Asphalt Mixtures Using The Asphalt Mixture Performance Tester (AMPT)." Washington, D.C.: AASHTO.
AASHTO TP107 (2018), "Standard Method of Damage Characteristic Curve of Asphalt Concrete from Direct Tension Cyclic Fatigue Tests." Washington, D.C.: AASHTO.
AASHTO TP124 (2018). "Standard Method of Test for Determining the Fracture Potential of Asphalt Mixtures Using the Flexibility Index Test (FIT)." Washington, D.C.: AASHTO.
AASHTO T166 (2015). "Bulk Specific Gravity of Compacted Asphalt Mixtures using Saturated Surface-Dry Specimens." Washington, D.C.: AASHTO.
AASHTO T209 (2015). "Standard Method of Test for Theoretical Maximum Specific Gravity (Gmm) and Density of Hot Mix Asphalt (HMA)." Washington, D.C.:AASHTO
AASHTO, T342 (2015). "Standard Method of Test for Determining Dynamic Modulus of Hot Mix Asphalt (HMA)." Washington, D.C.: AASHTO.
Advanced Research Associates. (2004). 2002 Design Guide: Design of New and Rehabilitated Pavement Structures. NCHRP 1-37A Project, National Cooperative Highway Research Program. National Research Council. Washington, D.C.
Bahia, H.U. (2001). Characterization of Modified Asphalt Binders in Superpave Mix Design. Washington, D.C.: National Academy Press.
Bari, J. (2005). Development of a New Revised Version of the Witczak E* Predictive Models for Hot Mix Asphalt Mixtures. Ph.D. Dissertation. Arizona State University.
Biligiri K.P., S.F. Said, H. Hakim. (2012). "Asphalt Mixtures Crack Propagation Assessment Using Semi-Circular Bending Test." Int J Pavement Res Technol, Vol. 5, No. 4, pp. 20917.
Bowers, B.F., B.K. Diefenderfer, and S.D. Diefenderfer. (2015). "Evaluation of Dynamic Modulus in Asphalt Paving Mixtures Utilizing Small-Scale Specimen Geometries." Electronic Journal of the Association of Asphalt Paving Technologists, Vol. 84, pp. 497526.
Cascione, A.A., R.C. Williams, and J. Yu. (2015). "Performance Testing of Asphalt Pavements with Recycled Asphalt Shingles from Multiple Field Trials." Construction
77

and

Building

Materials,

Vol.

101,

pp.

628642.

doi:10.1016/j.conbuildmat.2015.09.027

Castorena, C., et al. (2017, September). Development of Small Specimen Geometry for Asphalt Mixture Performance Testing. National Cooperative Highway Research Program Innovations Deserving Exploratory Analysis Project 181. Transportation Research Board, National Research Council. Washington, D.C.

Chehab, G.R., Y.R. Kim, R.A. Schapery, M. Witczack, and R. Bonaquist. (2003). "Characterization of Asphalt Concrete in Uniaxial Tension Using a Viscoelastoplastic 23 Model." Journal of the Association of Asphalt Paving Technology, Vol. 72, pp. 315355.

Christensen, Jr., D.W., T.K. Pellinen, and R.F. Bonaquist. (2003). "Hirsch Model for Estimating the Modulus of Asphalt Concrete." Journal of the Association of Asphalt Paving Technologists (AAPT), Vol. 72.

Chong, K.P., and Kuruppu, M.D. (1984). "New Specimen for Fracture Toughness Determination for Rock and Other Materials." International Journal of Fracture, 26(2). doi:10.1007/bf01157555

Clyne, T.R., X. Li, M.O. Marasteanu, and K. Engene. (2003). Dynamic Modulus and Resilient Modulus of Mn/DOT Asphalt Mixtures. Report: MN/RC2003-09: Minnesota Department of Transportation, St. Paul.

Cross S.A., and Y. Jakatimath (2007). Determination of Dynamic Modulus Master Curves for Oklahoma HMA Mixtures. Final Report: Oklahoma Department of Transportation, Oklahoma, USA.

Daniel, J.S.. and Y.R. Kim. (2002). "Development of a Simplified Fatigue Test and Analysis Procedure Using a Viscoelastic Continuum Damage Model. Journal of the Association of Asphalt Paving Technologists, Vol. 71, pp. 619650.

Epps, J.A. and C.L. Monismith. (1969). "Influences of Mixture Variables on the Flexural Fatigue Properties of Asphalt Concrete." Proceedings of the Association of Asphalt Paving Technologists, Vol. 38. pp. 423464

Finn, F.N., K. Nair, and J. Hilliard. (1973, February). Minimizing Premature Cracking of Asphalt Concrete Pavements. National Cooperative Highway Research Program Project 9-4. Transportation Research Board, National Research Council. Washington, D.C.

Flintsch G.W., A. Loulizi, and B.K. Diefenderfer. (2007). Asphalt Material Characterization in Support of Implementation of Mechanistic Empirical Pavement Design Guide. Virginia Transportation Research Council: Final Report, Report No. VTRC 07-CR10

Garcia, V., et al. (2017). Improved Overlay Tester for Fatigue Cracking Resistance of Asphalt Mixtures. Texas Department of Transportation, Research Project 0-6815,

78

Center for Transportation Infrastructure Systems, The University of Texas at El Paso, Tex.
Garcia, V., and A. Miramontes. (2015). "Understanding Sources of Variability of Overlay Test Procedure." Transportation Research Record: Journal of the Transportation Research Board, No. 2507, Transportation Research Board, Washington, D.C., pp. 1018. DOI: 10.3141/2507-02
Germann, F.P., and R.L. Lytton. (1979, March). Methodology for Predicting the Reflection Cracking Life of Asphalt Concrete Overlays. Research Report FHWA/TX79/09+207-5, College Station, Texas.
Im, S., H. Ban, and Y.-R. Kim. (2014). "Mode-Dependent Fracture Behavior of Asphalt Mixtures with Semicircular Bend Test." Transportation Research Record: Journal of the Transportation Research Board, No. 2447, pp. 2331.
Irwin, G.R. (1948). "Fracture Dynamics," Trans. Am. Soc. Met. Vol. 40A, pp. 147 166.
Kennedy, T.W., F.L. Roberts, and R.B. McGennis. (1984). "Effects of Compaction Temperature and Effort on the Engineering Properties of Asphalt Concrete Mixtures." Placement and Compaction of Asphalt Mixtures, F.T. Wagner, Ed. ASTM Special Technical Publication 829. American Society for Testing and Materials, Philadelphia, Penn. pp. 4866.
Khalid, H.A., and O.K. Monney. (2009). "Moisture Damage Potential of Cold Asphalt," International Journal of Pavement Engineering, Vol. 10, No. 5, pp. 311 318. http://dx.doi.org/10.1080/10298430802169838
Kim, Y.R., and D.N. Little. "One-dimensional Constitutive Modeling of Asphalt Concrete." ASCE Journal of Engineering Mechanics, Vol. 116, No. 4, 1990, pp. 751 772.
Kim, M., L. Mohammad, and M. Elseifi. (2012). "Characterization of Fracture Properties of Asphalt Mixtures as Measured by Semicircular Bend Test and Indirect Tension Test." Transportation Research Record: Journal of the Transportation Research Board, 2296, pp. 115124. doi:10.3141/2296-12
Kim, Y.R., Y. Seo, M. King, and M. Momen. (2004). "Dynamic Modulus Testing of Asphalt Concrete in Indirect Tension Mode." Transportation Research Record: Journal of the Transportation Research Board, No. 1891, pp. 163173. http://dx.doi.org/10.3141/1891-19
Kutay, M.E., N. Gibson, J. Youtcheff, and R. Dongre. "Use of Small Samples to Predict Fatigue Lives of Field Cores: Newly Developed Formulation Based on Viscoelastic Continuum Damage Theory." Transportation Research Record: Journal of the Transportation Research Board, No. 2127, 2009, pp. 9097. http://dx.doi.org/10.3141/2127-11
79

Lee, K., S. Pape, C. Castorena, and Y.-R. Kim. (2017). "Evaluation of Small Specimen Geometries for Asphalt Mixture Performance Testing and Pavement Performance Prediction." Transportation Research Record: Journal of the Transportation Research Board, No. 2631, pp. 7482. http://dx.doi.org/10.3141/2631-09
Li, X.J., and M.O. Marasteanu (2009). "Using Semi Circular Bending Test to Evaluate Low Temperature Fracture Resistance for Asphalt Concrete." Experimental Mechanics, Vol. 50, No. 7, pp. 867876.
Li, X., and N.H. Gibson. (2013). "Using Small Scale Specimens for AMPT Dynamic Modulus and Fatigue Tests." Electronic Journal of the Association of Asphalt Paving Technologists, Vol. 82, pp. 579616.
Lim, I.L., I.W. Johnson, and S.K. Choi. (1993). "Stress Intensity Factors for Semicircular Specimens Under Three-Point Bending." Engineering Fracture Mechanics Vol. 44, No. 3, pp. 363382.
Linden, R.N., J.P. Mahoney, and N.C. Jackson. (1989). The Effect of Compaction on Asphalt Concrete Performance. 1989 Annual Meeting of the Transportation Research Board, Washington, D.C.
Ling, C., D. Swiertz, T. Mandal, P. Teymourpour, and H. Bahia. (2017). "Sensitivity of the Illinois Flexibility Index Test to Mixture Design Factors." Transportation Research Record: Journal of the Transportation Research Board, 2631, pp. 153159. doi:10.3141/2631-17
Marasteanu, M.O., et al. (2004). "Low Temperature Cracking of Asphalt Concrete Pavement."
Norouzi, A., Y.-R. Kim, S. S. Kim, and J. Yang. (2017). "Effect of Reclaimed Asphalt Pavement Content and Binder Grade on Fatigue-Resisting Performance of Asphalt Mixtures in Georgia." Journal of Materials in Civil Engineering, Vol. 29, No. 9, pp. 04017115. doi:10.1061/(asce)mt.1943-5533.0001960
Nsengiyumva, G., T. You, and Y. Kim. (2016). "Experimental-Statistical Investigation of Testing Variables of a Semicircular Bending (SCB) Fracture Test Repeatability for Bituminous Mixtures." Journal of Testing and Evaluation, Vol. 45, No. 5, 20160103. doi:10.1520/jte20160103
Ozer, H., I.L. Al-Qadi, J. Lambros, A. El-Khatib, P. Singhvi, and B. Doll. (2016). "Development of the Fracture-based Flexibility Index for Asphalt Concrete Cracking Potential Using Modified Semi-circle Bending Test Parameters. Construction and Building Materials, Vol. 115, pp. 390401. doi:10.1016/j.conbuildmat.2016.03.144
Ozer, H., I.L. Al-Qadi, P. Singhvi, T. Khan, J. Rivera-Perez, and A. El-Khatib. (2016). "Fracture Characterization of Asphalt Mixtures with High Recycled Content Using Illinois Semicircular Bending Test Method and Flexibility Index." Transportation Research Record: Journal of the Transportation Research Board, No. 2575, pp. 130 137. doi:10.3141/2575-14
80

Pell, P.S., and I.F. Taylor. (1969). "Asphalt Road Materials in Fatigue." Proceedings of the Association of Asphalt Paving Technologists, Vol. 38. pp. 371464.

PI (Pavement Interactive). (2011). "Bulk Specific http://www.pavementinteractive.org/article/bulk-specific-gravity/

Gravity."

PI (Pavement Interactive). (2011). "Theoretical Maximum Specific Gravity." http://www.pavementinteractive.org/theoretical-maximum-specific-gravity/

PI

(Pavement

Interactive).

(2011).

http://www.pavementinteractive.org/binder-content/

"Binder

Content."

Roberts, F.L., P.S. Kandhal, E.R. Brown, D.Y. Lee, and T.W. Kennedy. (1996). "Hot Mix Asphalt Materials, Mixture Design, and Construction." National Asphalt Paving Association Education Foundation. Lanham, MD.

Sabouri, M., T. Bennert, J.S. Daniel and Y.R. Kim (2015). "A Comprehensive Evaluation of the Fatigue Behaviour of Plant-produced RAP Mixtures." Road Materials and Pavement Design, Vol. 16, Sup. 2, pp. 2954, DOI: 10.1080/14680629.2015.1076997

Sargand, S.M. and S.S. Kim. (2001). "Performance Evaluation of Polymer Modified and Unmodified Superpave Mixes. Second International Symposium on Maintenance and Rehabilitation of Pavements and Technological Control, Auburn, AL.

Schapery, R.A. (1984). "Correspondence Principles and a Generalized J-integral for Large Deformation and Fracture Analysis of Viscoelastic Media." Int. J. Fract., Vol. 25, pp: 195

Shu, X., and B. Huang. (2008). "Micromechanics-based Dynamic Modulus Prediction of Polymeric Asphalt Concrete Mixtures." Composites B,Vol. 39, No. 44. pp. 704 713.

Underwood, B.S. Y.R. Kim, and M.N. Guddati. (2010). "Improved Calculation Method of Damage Parameter in Viscoelastic Continuum Damage Model." Inter. J. Pave. Eng., Vol.. 11, No. 6, pp. 459476.

Wagoner, M.P., W.G. Buttlar, and G.H. Paulino. (2005). "Development of a Singleedge Notched Beam Test for Asphalt Concrete Mixtures." Journal of Testing and Evaluation, Vol. 33, No. 6, pp. 452460.

Walubita, L.F., A.N.M. Faruk, G. Das, H.A. Tanvir, J. Zhang, and T. Scullion. (2012). The Overlay Tester: A Sensitivity Study to Improve OT Repeatability and Minimize Variability in the OT Test Results. Report FHWA/TX-12/0-6607-1. Texas A&M Transportation Institute, Texas A&M University System, College Station.

Walubita, L.F., A.N.M. Faruk, Y. Koohi, R. Luo, T. Scullion, and R.L. Lytton. (2013). The Overlay Tester (OT): Comparison with Other Crack Test Methods and Recommendations for Surrogate Crack Test. Report FHWA/ TX-13/0-6607-2. Texas A&M Transportation Institute, Texas A&M University System, College Station.

81

Wang, Y. and Y.R. Kim (2017). "Development of a Pseudo Strain Energy-based Fatigue Failure Criterion for Asphalt Mixtures." International Journal of Pavement Engineering, 2017, DOI: 10.1080/10298436.2017.1394100. Wang, Y. and Y.R. Kim. (2018). "Development of a Simple Fatigue Index Parameter for Asphalt Concrete based on the Viscoelastic Continuum Damage Theory." Journal of Association of Asphalt Paving Technologies, accepted, 2018. Wu, Z., et al. (2005). "Fracture Resistance Characterization of Superpave Mixtures Using the Semi-circular Bending Test." Journal of ASTM International, Vol. 2, No. 3, pp. 115. Zeiada, W.A., K.E. Kaloush, B.S. Underwood, and M.S. Mamlouk. (2013). "Effect of Air Voids and Asphalt Content on Fatigue Damage Using the Viscoelastic Continuum Damage Analysis." Airfield and Highway Pavement 2013. doi:10.1061/9780784413005.094 Zhou, F., and T. Scullion. Overlay Tester: A Rapid Performance Related Crack Resistance Test. FHWA/TX-05/0-4467-2. Texas Transportation Institute, College Station, 2005. Zhu, H., Sun, L., Yang, J., Z. Chen, and W. Gu. (2011). "Developing Master Curves and Predicting Dynamic Modulus of Polymer-modified Asphalt Mixtures. J Mater Civ Eng, Vol. 23 No. 2, pp. 131137.
82

Catalog of MEPDG Design Inputs for Asphalt Mixtures 83

TABLE A.1 Mixture ID A_12.5_67_N

Mixture Type: A 12.5_67_N

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1196709

1468848 1600929 1911746 2046825

2218405

68

368348

546358

635750

895078 1022953

1200915

104

72843

120391

147741

254010 318939

426653

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

Effective Binder Content (%)

11.8

Air Voids (%)

6.3

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

13

25

6.3

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.8

Air Voids (%)

6.3

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Level 3

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

11.8

Cumulative % Retained on 3/8 Inch Sieve

13

Air Voids (%)

6.3

Cumulative % Retained on #4 Sieve

25

Total Unit Weight (pcf)

145

% Passing #200 Sieve

6.3

Asphalt Binder: Superpave Binder Grading:

PG 67-22

Note: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

84

TABLE A.2 Mixture ID A _12.5_76_N

Mixture Type: A 12.5_76_N

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1236304 1525751 1657929 1981267 2122679 2303977

68

356213

535287 624437 887874 1013864 1199077

104

68424

107193 129408

227033 288819

420707

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

12.6

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.7

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

10

27

6.3

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

12.6

Air Voids (%)

5.7

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

12.6

Cumulative % Retained on 3/8 Inch Sieve

10

Air Voids (%)

5.7

Cumulative % Retained on #4 Sieve

27

Total Unit Weight (pcf)

145

% Passing #200 Sieve

6.3

Asphalt Binder: Superpave Binder Grading:

PG 76-22

Note: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

85

TABLE A.3 Mixture ID A 19_64_N1

Mixture Type: A 19_64_N1

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1470782 1766998 1902608 2232473 2373885 2555425

68

495643

703531 805154 1093393 1229197 1418133

104

97615

162491 198267 333056 410506 537414

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

11.6

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

5

11

27

5.8

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.6

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Level 3

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

5

Effective Binder Content (%)

11.6

Cumulative % Retained on 3/8 Inch Sieve

11

Air Voids (%)

5.5

Cumulative % Retained on #4 Sieve

27

Total Unit Weight (pcf)

145

% Passing #200 Sieve

5.8

Asphalt Binder: Superpave Binder Grading:

PG 64-22

Note: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

86

TABLE A.4 Mixture ID A 25_64_N1

Mixture Type: A 25_64_N1

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1491958 1756555 1875438 2161550 2283430 2438814

68

518414

718035 814243 1085754 1212228 1390576

104

112825

181733 218814 359743 438885 573577

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

11.2

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

12

9

20

5.7

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.2

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

12

Effective Binder Content (%)

11.2

Cumulative % Retained on 3/8 Inch Sieve

9

Air Voids (%)

5.5

Cumulative % Retained on #4 Sieve

20

Total Unit Weight (pcf)

145

% Passing #200 Sieve

5.7

Asphalt Binder: Superpave Binder Grading:

PG 64-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

87

TABLE A.5 Mixture ID B 9.5_64_M1

Mixture Type: B 9.5_64_M1

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1112490

1419100 1557998 1895453 2039379 2221354

68

294717

466152 555012 829472 963391 1161368

104

51861

87236 109857 205132 267015 376422

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

12.6

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

6.5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

1

28

6

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

12.6

Air Voids (%)

6.5

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

12.6

Cumulative % Retained on 3/8 Inch Sieve

1

Air Voids (%)

6.5

Cumulative % Retained on #4 Sieve

28

Total Unit Weight (pcf)

145

% Passing #200 Sieve

6

Asphalt Binder: Superpave Binder Grading:

PG 64-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

88

TABLE A.6 Mixture ID B 9.5_64_M2

Mixture Type: B 9.5_64_M2

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1144592 1448011 1586619 1929102 2077428 2269265

68

320631

484185 569129 840592 972480 1161706

104

55898

95773 121218 221715 286305 401223

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

11.6

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

6.5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

6

27

6.5

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.6

Air Voids (%)

6.5

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

11.6

Cumulative % Retained on 3/8 Inch Sieve

6

Air Voids (%)

6.5

Cumulative % Retained on #4 Sieve

27

Total Unit Weight (pcf)

145

% Passing #200 Sieve

6.5

Asphalt Binder: Superpave Binder Grading:

PG 64-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

89

TABLE A.7 Mixture ID C 9.5_67_M

Mixture Type: C 9.5_67_M

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1463361 1726339 1852135 2156522 2288071 2452061

68

476933

676989 773053 1053556 1184912 1368772

104

93071

138676 172015 292107 362692 482300

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

12.9

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

5

32

5.5

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

12.9

Air Voids (%)

5

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

12.9

Cumulative % Retained on 3/8 Inch Sieve

5

Air Voids (%)

5

Cumulative % Retained on #4 Sieve

32

Total Unit Weight (pcf)

145

% Passing #200 Sieve

5.5

Asphalt Binder: Superpave Binder Grading:

PG 67-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

90

TABLE A.8 Mixture ID A 12_64_M2

Mixture Type: A 12.5_64_M2

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

Mixture |E*|, psi

0.5 Hz

1 Hz

5 Hz

10 Hz 25 Hz

39.2

1323423

1568103 1683069 1996206 2129448 2302478

68

406687

590256 681679 962279 1097889 1284602

104

73500

139754 178784 316715 396631 528228

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

12.2

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve Cumulative % Retained on 3/8 Inch Sieve Cumulative % Retained on #4 Sieve % Passing #200 Sieve

Level 2
0 12 27 5.9

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

12.2

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve

Asphalt General: Volumetric Properties as Built

Effective Binder Content

0

(%)

12.2

Cumulative % Retained on 3/8 Inch Sieve

12

Air Voids (%)

5.5

Cumulative % Retained on #4 Sieve

27

Total Unit Weight (pcf)

145

% Passing #200 Sieve

5.9

Asphalt Binder: Superpave Binder Grading:

PG 64-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

91

TABLE A.9 Mixture ID A 12.5_64_M1

A Mixture Type: 12.5_64_M1

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1179401 1480935 1618672 1977206 2130270 2327183

68

346061

545004 644597 941345 1085126 1296881

104

47684

84325

114638 225679 299020 452422

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

12.5

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

14

26

5.8

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

12.5

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

12.5

Cumulative % Retained on 3/8 Inch Sieve

14

Air Voids (%)

5.5

Cumulative % Retained on #4 Sieve

26

Total Unit Weight (pcf)

145

% Passing #200 Sieve

5.8

Asphalt Binder: Superpave Binder Grading:

PG 64-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

92

TABLE A.10 Mixture ID B 12.5_64_M

B Mixture Type: 12.5_64_M

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1210826 1490991 1616013 1923010 2057219 2298901

68

357905

536254 625984 890388 1017006 1194050

104

82169

139865 172499 295732 368928 485152

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

12.5

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.6

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

13

25

6

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

12.5

Air Voids (%)

5.6

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

12.5

Cumulative % Retained on 3/8 Inch Sieve

13

Air Voids (%)

5.6

Cumulative % Retained on #4 Sieve

25

Total Unit Weight (pcf)

145

% Passing #200 Sieve

6

Asphalt Binder: Superpave Binder Grading:

PG 64-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

93

TABLE A.11 Mixture ID B 25_64_M

Mixture Type: B 25_64_M

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1633901 1958448 2101746 2442295 2582160 2761910

68

603116

861767 982874 1334978 1492103 1710868

104

133923

224422 274267 466104 571788 736406

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

9.4

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.9

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

10

8

17

5

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

9.4

Air Voids (%)

5.9

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

10

Effective Binder Content (%)

9.4

Cumulative % Retained on 3/8 Inch Sieve

8

Air Voids (%)

5.9

Cumulative % Retained on #4 Sieve

17

Total Unit Weight (pcf)

145

% Passing #200 Sieve

5

Asphalt Binder: Superpave Binder Grading:

PG 64-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

94

TABLE A.12 Mixture ID B 9.5_67_S

Mixture Type: B 9.5_67_S

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1168813 1443853 1568731 1874955 2005150 2167254

68

321549

489407 574060 834984 959330 1146235

104

56396

96736

121416 227903 290221 383335

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

12.8

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

3

28

5.3

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

12.8

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

12.8

Cumulative % Retained on 3/8 Inch Sieve

3

Air Voids (%)

5.5

Cumulative % Retained on #4 Sieve

28

Total Unit Weight (pcf)

145

% Passing #200 Sieve

5.3

Asphalt Binder: Superpave Binder Grading:

PG 67-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

95

TABLE A.13 Mixture ID B 12.5_67_S

Mixture Type: B 12.5_67_S

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1210826 1490991 1616013 1923010 2057219 2298901

68

357905

536254 625984 890388 1017006 1194050

104

82169

139865 172499 295732 368928 485152

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

12.1

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

6

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

14

25

5

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

12.1

Air Voids (%)

6

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

12.1

Cumulative % Retained on 3/8 Inch Sieve

14

Air Voids (%)

6

Cumulative % Retained on #4 Sieve

25

Total Unit Weight (pcf)

145

% Passing #200 Sieve

5

Asphalt Binder: Superpave Binder Grading:

PG 67-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

96

TABLE A.14 Mixture ID D 12.5_76_S

Mixture Type: D 12.5_76_S

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1389174

1643619 1759069 2039331 2166433 2309682

68

519574

721854 811778 1080291 1209569 1381294

104

105805

173756 210257 343160 417564 537994

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

11.9

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

12

28

4.9

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.9

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

Level 3

Asphalt Mix: Aggregate Gradation

Asphalt General: Volumetric Properties as Built

Cumulative % Retained on 3/4 Inch Sieve

0

Effective Binder Content (%)

11.9

Cumulative % Retained on 3/8 Inch Sieve

12

Air Voids (%)

5.5

Cumulative % Retained on #4 Sieve
% Passing #200 Sieve

28

Total Unit Weight (pcf)

145

4.9

Asphalt Binder: Superpave Binder Grading:

PG 76-22

Notes: The table summarizes the test data using extracted asphalt binder from asphalt plant mix.

97

TABLE A.15 Mixture ID A_19_64_N2

Mixture Type: A 19_64_N2

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1930214 2251667 2390420 2712936 2844002 3043236

68

755068

1051187 1146139 1477986 1610840 1825110

104

168897

280358 343208 597073 702661 882556

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

10.1

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

1

9

19

5.3

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

10.1

Air Voids (%)

5

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve Cumulative % Retained on 3/8 Inch Sieve Cumulative % Retained on #4 Sieve % Passing #200 Sieve Asphalt Binder: Superpave Binder Grading:

Level 3
1 9 19 5.3 PG 64-22

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

10.1

Air Voids (%)

5

Total Unit Weight (pcf)

145

98

TABLE A.16 Mixture ID A_25_64_N2

Mixture Type: A 25_64_N2

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1956490 2270932 2411402 2723306 2852245 3012222

68

727075

1000134 1128444 1479194 1631436 1842418

104

139570

208758 265661

451310

557671

718615

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

9.8

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.2

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve Cumulative % Retained on 3/8 Inch Sieve Cumulative % Retained on #4 Sieve % Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

9 7 15 5.5

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

9.8

Air Voids (%)

5.2

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve Cumulative % Retained on 3/8 Inch Sieve Cumulative % Retained on #4 Sieve % Passing #200 Sieve Asphalt Binder: Superpave Binder Grading:

Level 3
9 7 15 5.5 PG 64-22

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

9.8

Air Voids (%)

5.2

Total Unit Weight (pcf)

145

99

TABLE A.17 Mixture ID C_12.5_67_M

Mixture Type: C 12.5_67_M

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

1279574

1578594 1720199 2067855 2214972 2402313

68

384399

572417

664709

940910 1071976 1265166

104

68453

116446

147049

265903

339389

451552

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

11.5

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.8

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

12

27

6.1

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.5

Air Voids (%)

5.8

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve Cumulative % Retained on 3/8 Inch Sieve Cumulative % Retained on #4 Sieve % Passing #200 Sieve Asphalt Binder: Superpave Binder Grading:

Level 3
0 12 27 6.1 PG 67-22

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.5

Air Voids (%)

5.8

Total Unit Weight (pcf)

145

100

TABLE A.18 Mixture ID C_12.5_76_M

Mixture Type: C 12.5_76_M

Level 1

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

Mixture |E*|, psi

1 Hz

5 Hz

10 Hz

25 Hz

39.2

958049

1208384 1337299 1634965 1766756 1937031

68

273203

424236 502073 743997 859882 1030978

104

55376

87076 107410 193819 250529 351234

Asphalt Binder: Superpave Binder Test Data

Asphalt General: Volumetric Properties as Built

Temperature Angular Freq. = 10 rad/sec

Effective Binder Content (%)

11.5

(F)

G* (Pa)

Delta (degree)

Air Voids (%)

5.8

See Appendix B

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation

Cumulative % Retained on 3/4 Inch Sieve

Cumulative % Retained on 3/8 Inch Sieve

Cumulative % Retained on #4 Sieve

% Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

0

12

27

6.1

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.5

Air Voids (%)

5.8

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve Cumulative % Retained on 3/8 Inch Sieve Cumulative % Retained on #4 Sieve % Passing #200 Sieve Asphalt Binder: Superpave Binder Grading:

Level 3
0 12 27 6.1 PG 76-22

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

11.5

Air Voids (%)

5.8

Total Unit Weight (pcf)

145

101

TABLE A.19 Mixture ID B_19_67_M

Mixture Type: B 19_64_M

Asphalt Mix: Dynamic Modulus Table

Temperature (F)

0.1 Hz

0.5 Hz

39.2

1404621

1732479

68

410167

619119

104

58218

114875

Level 1

Mixture |E*|, psi

1 Hz

5 Hz

1884914 2249032

720307 1036345

149650 283163

10 Hz 2401467 1186508 369557

25 Hz 2601329 1403871 539590

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

10.5

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve Cumulative % Retained on 3/8 Inch Sieve Cumulative % Retained on #4 Sieve % Passing #200 Sieve

Asphalt Binder: Superpave Binder Test Data

Temperature (F)

Angular Freq. = 10 rad/sec

G* (Pa)

Delta (degree)

See Appendix B

Level 2

1 14 25
6

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

10.5

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

Asphalt Mix: Aggregate Gradation Cumulative % Retained on 3/4 Inch Sieve Cumulative % Retained on 3/8 Inch Sieve Cumulative % Retained on #4 Sieve % Passing #200 Sieve Asphalt Binder: Superpave Binder Grading:

Level 3
1 14 25
6 PG 64-22

Asphalt General: Volumetric Properties as Built

Effective Binder Content (%)

10.5

Air Voids (%)

5.5

Total Unit Weight (pcf)

145

102

Aging and DSR Testing of Georgia Binders 103

INTRODUCTION
The accurate characterization of in situ aging of asphalt pavement materials over the service life of the pavement is of utmost importance to the implementation of mechanistic empirical (ME) pavement design and analysis methods. The key product of NCHRP 09-54 is a laboratory aging procedure that prescribes a set of laboratory aging conditions to represent the long-term aged state of asphalt mixtures in a pavement as a function of climate and depth. The results of this project will also yield a pavement aging model that can serve as a basis for the future development of a methodology that integrates the effects of long-term aging in Pavement ME Design.
The development of the pavement aging model is ongoing. This model will predict the evolution of asphalt mixture performance with long-term aging. Implementation of the model will require quantifying asphalt binder kinetics using relatively simple and efficient test methods. The universal simple aging test (USAT) or the rolling thin film oven (RTFO) and pressure aging vessel (PAV) can be used to obtain various levels of binder aging, which can then be characterized using the dynamic shear rheometer (DSR) to obtain the binder kinetics.
The binder kinetics and properties can then be coupled with the mixture properties at shortterm aged condition to predict the aged mixture properties at any duration and depth in the field using the pavement aging model that will be developed under NCHRP 09-54.
In this report, both USAT and RTFO/PAV aging methods were used to age two different asphalt binder sources obtained from Georgia. The kinetics, linear viscoelastic properties, as well as damage properties of both these binder sources were characterized at various age levels using the DSR. When the NCHRP 9-54 pavement aging model is complete and implemented in the Pavement ME Design and FlexPAVETM programs, the experimental data obtained from this project can be used to evaluate the pavement performance with long-term aging.
MATERIALS AND TEST METHODS
The two asphalt binder sources obtained from Georgia are PG64-22 and PG76-22 binders.
Aging Methods
Universal Simple Aging Test (USAT) Farrar et al. (2014) proposed the USAT for the efficient simulation of asphalt binder aging in the laboratory. The USAT uses thin binder films to induce a kinetics-controlled reaction. The binder is placed in grooved plates to achieve a film thickness of 300 micrometers. The USAT plates are placed in an oven at 135C for four hours to simulate short-term aging and best mimic the short-term aging of loose mixtures. After this binder short-term aging
104

process, the USAT plates are placed in an oven at 95C for 2 days, 4 days, and 8 days to simulate various levels of long-term aging. Rolling Thin Film Oven (RTFO) RTFO aging was conducted using selected original asphalt binder samples according to AASHTO T240 to simulate short-term aging. Pressure Aging Vessel (PAV) Asphalt binder residue obtained from the RTFO aging was subjected to PAV aging based on AASHTO R28 at 100C for 20 hours and 40 hours to simulate two levels of long-term aging. Testing Method: Dynamic Shear Rheometer (DSR) TemperatureFrequency Sweep Test Temperaturefrequency sweep testing is conducted at frequencies ranging from 0.1 Hz to 30 Hz and multiple temperatures (i.e., 5C, 20C, 35C, 50C, and 64C) using asphalt binders in the DSR with 8 mm parallel plate geometry. The strain amplitude used is chosen such that the linear viscoelastic limit is maintained. The rheological properties obtained are the dynamic shear modulus (G*) and phase angle. Linear Amplitude Sweep (LAS) Test The LAS test (AASHTO TP101) consists of oscillatory shear in strain-controlled mode in the DSR using an 8 mm parallel plate geometry at a frequency of 10 Hz. The strain amplitude is increased linearly from 0.1% to 30% to induce fatigue damage at an accelerated rate. The simplified viscoelastic continuum damage (S-VECD) modeling can be applied to LAS test results to predict the fatigue life at any loading history of interest.
RESULTS
Both binders were tested after short-term aging (STA), and after 2 days, 4 days, and 8 days of long-term aging (LTA). Figure B.1 and Figure B.2 show the evolution of |G*| with aging using USAT and RTFO/PAV, respectively. Figure B.3 and Figure B.4 show the evolution of the phase angle for both binders with aging using USAT and RTFO/PAV, respectively.
105

|G*| (Pa)

1.E+09
1.E+08 (a)

1.E+07

1.E+06

1.E+05

1.E+04

1.E+03

1.E+02

0.0000001 0.0001

0.1

STA 2D 4D 8D
100

Reduced Frequency (Hz)

|G*| (Pa)

1.E+09
1.E+08 (b)

1.E+07

1.E+06

1.E+05

STA

1.E+04

2D

1.E+03

1.E+02

0.0000001 0.0001

0.1

4D 8D
100

Reduced Frequency (Hz)

FIGURE B.1
Evolution of the Dynamic Shear Modulus as Aging Advances Using USAT for: (a) PG 64-22, and (b) PG 76-22

|G*| (Pa)

1.E+09
(a)
1.E+08

1.E+07

1.E+06 1.E+05 1.E+04 1.E+03

RTFO 20hr PAV 40hr PAV

1.E+02

0.0000001 0.0001

0.1

100

Reduced Frequency (Hz)

|G*| (Pa)

1.E+09
(b)
1.E+08

1.E+07

1.E+06 1.E+05 1.E+04 1.E+03

RTFO 20hr PAV 40hr PAV

1.E+02

0.0000001 0.0001

0.1

100

Reduced Frequency (Hz)

FIGURE B.2
Evolution of the Dynamic Shear Modulus as Aging Advances Using RTFO/PAV for: (a) PG 64-22, and (b) PG 76-22

Phase Angle ()

90

80

(a)

70

60

50

40

30

STA

20

2D

10

4D

0

8D

0.0000001 0.0001

0.1

100

Reduced Frequency (Hz)

Phase Angle ()

90

80

(b)

70

60

50

40

30

STA

20

2D

10

4D

0

8D

0.0000001 0.0001

0.1

100

Reduced Frequency (Hz)

FIGURE B.3
Evolution of Phase Angle as Aging Advances Using USAT for: (a) PG 64-22, and (b) PG 76-22

106

Phase Angle () Phase Angle ()

90

80

(a)

70

60

50

40

30

RTFO

20

20hr PAV

10

40hr PAV

0

0.0000001 0.0001

0.1

100

Reduced Frequency (Hz)

90

80

(b)

70

60

50

40

30

RTFO

20

20hr PAV

10

40hr PAV

0

0.0000001 0.0001

0.1

100

Reduced Frequency (Hz)

FIGURE B.4
Evolution of Phase Angle as Aging Advances Using RTFO/PAV for: (a) PG 64-22, and (b) PG 76-22

A significant increase is evident in the dynamic shear modulus values when the aging duration is increased. The phase angle, on the other hand, drops when age level increases. Both of these trends are expected since the binder is stiffened with aging. Table B.1 and Table B.2 show the values of |G*| at 10 rad/s and the phase angle at different temperatures for both binders aged using both USAT and RTFO/PAV.

107

TABLE B.1 |G*| at 10 rad/s and Phase Angle as a Function of Temperature (USAT Aging)

Temperature (F)

|G*| (Pa)

Phase Angle ()

PG 64-22 STA

147.2

15,136.4

158.0

7,455.4

168.8

3,963.9

179.6

2,289.2

190.4

1,442.9

70.64 72.39 73.80 74.92 75.80

147.2 158.0 168.8 179.6

PG 64-22 2D 43,059.3 20,327.6 10,129.8 5,369.4

64.28 66.48 68.33 69.88

190.4

3,046.2

71.16

147.2 158.0 168.8 179.6 190.4

PG 64-22 4D 76,013.9 35,750.7 17,541.1 9,051.7 4,946.2

60.41 62.82 64.90 66.67 68.17

147.2 158.0 168.8 179.6 190.4

PG 64-22 8D 293,245.2 140,595.3 68,221.4 33,768.0 17,172.9

50.72 53.51 56.05 58.35 60.40

Temperature (F)

|G*| (Pa)

Phase Angle ()

PG 76-22 STA

147.2

23,305.8

158.0

11,916.7

168.8

6,526.4

179.6

3,853.2

190.4

2,465.0

67.39 69.27 70.82 72.09 73.09

PG 76-22 2D

147.2

48,644.9

158.0

23,767.8

168.8

12,209.7

179.6

6,645.0

63.06 65.38 67.37 69.06

190.4

3,855.9 70.47

PG 76-22 4D

147.2

77,727.2

158.0

37,750.4

168.8

19,097.8

179.6

10,144.4

190.4

5,696.4

59.87 62.36 64.53 66.41 68.01

PG 76-22 8D

147.2

253,791.8

158.0

123,718.0

168.8

61,163.5

179.6

30,904.1

190.4

16,072.2

51.33 54.13 56.70 59.02 61.10

108

TABLE B.2 |G*| at 10 rad/s and Phase Angle as a Function of Temperature (RTFO/PAV Aging)

Temperature (F)

|G*| (Pa)

Phase Angle ()

PG 64-22 RTFO

147.2

4,711.0

158.0

2,418.1

168.8

1,376.7

179.6

873.6

190.4

620.0

79.82 81.16 82.19 82.96 83.50

PG 64-22 20hr PAV

147.2

22,717.5

68.63

158.0

10,716.1

70.63

168.8

5,398.0

72.29

179.6

2,924.4

73.64

190.4

1,713.6

74.73

PG 64-22 40hr PAV

147.2

100,935.9 57.74

158.0

47,170.8

60.30

168.8

22,745.9

62.54

179.6

11,409.2

64.49

190.4

5,994.7

66.18

Temperature (F)

|G*| (Pa)

Phase Angle ()

PG 76-22 RTFO

147.2

9,309.8 73.71

158.0

4,994.6 75.13

168.8

2,935.1 76.25

179.6

1,898.7 77.10

190.4

1,357.3 77.72

PG 76-22 20hr PAV

147.2

28,113.1 66.86

158.0

13,920.4 68.96

168.8

7,340.9 70.72

179.6

4,152.5 72.17

190.4

2,534.2 73.35

PG 76-22 40hr PAV

147.2

67,055.8 60.53

158.0

32,275.0 63.02

168.8

16,176.6 65.19

179.6

8,510.1 67.07

190.4

4,730.9 68.67

The damage characteristic curves, which indicate the level of accumulated damage under fatigue loading, show an upward shift with aging as shown in Figure B.5 and Figure B.6 for USAT and RTFO/PAV aging, respectively. For the same value of C (pseudo stiffness or material integrity), the STA curve shows lower accumulated damage values than the LTA curves. These trends are expected since the damage characteristic curves of stiffer materials are generally higher than the curves of softer materials. The fatigue resistance is expected to decrease with prolonged aging due to the embrittlement imposed by oxidation.

109

C

1.0 0.8 0.6 0.4 0.2 0.0
0

1

STA

2D

STA

0.8

4D

8D

4D

0.6

C

0.4

5 S 10

(a)
15

0.2
0 0

5

10

S

FIGURE B.5

Damage Characteristic Curves as Aging Advances Using USAT for:

(a) PG 64-22, and (b) PG 76-22

2D 8D
(b)
15

C

1

1

RTFO

RTFO

0.8

20hr PAV

0.8

20hr PAV

40hr PAV

40hr PAV

0.6

0.6

C

0.4

0.4

0.2
0 0

0.2

(a)

(b)

0

5

10

0

2

4

6

S

S

FIGURE B.6

Damage Characteristic Curves as Aging Advances Using RTFO/PAV for: (a) PG 64-22, and (b) PG 76-22

REFERENCES
Farrar, M.J., J.P. Planche, R.W. Grimes, and Q. Qin. (2014). "The Universal Simple Aging Test (USAT): Simulating Short- and Long-Term Hot and Warm Mix Oxidative Aging in the Laboratory." Asphalt Pavements, Kim, Y.R., Ed., London: CRC Press, Taylor & Francis Group. pp. 7987.

110

Proposed Standard Operating Procedure (SOP) 111

Laboratory SOP DRAFT
Georgia Department of Transportation Office of Materials and Testing
Proposed Standard Operating Procedure (SOP) DRAFT Measurements of Dynamic Modulus (|E*|) and Development of Mastercurve of Asphalt
Concrete Mixture
I. General The purpose of this Standard Operating Procedure is to outline the methodology for measuring Dynamic Modulus (|E*|) of Asphaltic Concrete Mixture. This test is designed to be performed with an asphalt mixture performance tester (AMPT) in accordance with AASHTO TP107 for small specimen (1.5-in. diameter). The measurements of dynamic modulus of asphaltic concrete mixtures is a very technical process requiring highly skilled testing personnel, precision testing equipment, and close adherence to design guidelines and test procedures to assure high quality mix designs. It is a requirement for lab certification that the design equipment must meet all requirements and tolerances stated in the test procedures. Equipment calibration records shall be furnished to OMAT for review prior to initial certification and shall be available for inspection at all times. In case commercial laboratories that satisfy the requirements, research universities in Georgia that have extensive experience to measure dynamic modulus with small specimen (38 mm diameter) of asphaltic concrete mix should conduct this test with plant mix or field cores.
II. Specimen Fabrication
This procedure governs the sampling procedure to fabricate hot mix asphaltic concrete for dynamic modulus and fatigue cracking tests.
A. Sampling
Randomly select plant mix (verification mix) is collected in accordance with below references. The sampling testing, and inspection duties are to be performed by a GDOT Certified Contractor QCT and/or University of Georgia (UGA) Pavement Research Lab:
References: GDOT Specifications
GSP 15 (Sampling Procedures For Asphalt Concrete Mixtures) GDT 73 (Method of Random Selection And Acceptance Testing of Asphaltic
Concrete). DOT 162 (Asphaltic Concrete Plant Sampling Report).
This procedure also utilize 6-inch diameter field cores to run dynamic modulus test. The 6-inch diameter field cores will be used to prepare 38-mm-diameter by 110-mm-height for dynamic modulus test utilizing small cylindrical performance test specimens. This practice
112

is intended for dense-graded asphalt mixtures with nominal maximum aggregate sizes up to 25.0 mm.
B. Procedure for Specimen Fabrication from Gyratory Specimens
1. Asphalt Mixture Preparation:
a. Prepare asphalt mixture for each Superpave Gyratory Compactor (SGC) specimen in accordance with T312 and prepare a companion test specimen for maximum specific gravity (Gmm) in accordance with T 209.
b. The mass of asphalt mixture needed for each specimen will depend on the SGC specimen height, the Gmm of the mixture, the nominal maximum aggregate size, gradation (coarse or fine), and target air void content of the test specimens.
c. Perform conditioning on the asphalt mixture for the test specimens and companion Gmm sample in accordance with SOP 2.
d. SGC Specimen Compaction:
i. Compact the SGC specimens to a height of 180 mm or higher, in accordance with T312, carefully following the exceptions noted.
ii. Pour the mixture into the center of the mold to minimize air void variation between samples. Pouring material down the sides of the mold will result in lower air voids on that side of the mold.
iii. Charge the mold in two equal lifts, and rod the sample 20 times after each lift, to minimize vertical air void variance.
2. SGC Specimen Density and Air Voids:
a. Determine the Gmm of the asphalt mixture in accordance with SOP 2 and Section 828.
b. Determine Gmb of the SGC specimen in accordance with SOP 2 and Section 828. Record the Gmb of the SGC specimen.
c. Compute the air void content of the SGC specimen in accordance with SOP 2 and Section 828. Record the air void content of the SGC specimen.
3. Test Specimen Preparation:
a. Prepare the gyratory specimen by marking the location(s) where the cores will be taken. All cores must be taken within the inner 100 mm of the gyratory specimen. As many as four 38-mm diameter cores can be extracted from one gyratory specimen, as shown by the gray circles in Figure 1. The optimal lines to mark to extract four gyratory specimens are shown in white in Figure 1.
b. Drill a core of nominal diameter of 38 mm from the SGC specimen. Both the SGC specimen and the drill shall be adequately supported to ensure
113

that the resulting core is cylindrical with sides that are smooth, parallel, and meet the tolerances on specimen diameter given in Table 1.
c. Saw the ends of the core to obtain a test specimen of a nominal height of 110 mm. Both the core and the saw shall be adequately supported to ensure that the resulting test specimen meets the tolerances given in Table 1 for height, end flatness, and end perpendicularity.
d. With most equipment, it is better to perform the coring before the sawing. However, these operations may be performed in either order as long as the dimensional tolerances in Table 1 are satisfied.
e. Test specimens shall meet the dimensional tolerances given in Table 1.

Figure 1-Graphic of a marked gyratory specimen

Table 1-- Test Specimen Dimensional Tolerances

Item

Specification

Method Reference

Average diameter

36 to 40 mm

9.5.4.1

Standard deviation of diameter

0.5 mm

9.5.4.1

Height

107.5 to 112.5 mm

9.5.4.2

End flatness

0.5 mm

9.5.4.3

End perpendicularity

1.0 mm

9.5.4.4

114

4. Test Specimen Density and Air Voids: a. Determine the Gmm of the asphalt mixture in accordance with SOP 2. b. Determine Gmb of the test specimen in accordance with SOP 2. Record the Gmb of the SGC specimen. c. Compute the air void content of the SGC specimen in accordance with T 269. Record the air void content of the SGC specimen.
References: AASHTO Standards:
R 30, Mixture Conditioning of Hot Mix Asphalt T 166, Bulk Specific Gravity (Gmb) of Compacted Hot Mix Asphalt (HMA) Using
Saturated Surface Dry Specimens T 209, Theoretical Maximum Specific Gravity (Gmm) and Density of Hot Mix
Asphalt (HMA) T 269, Percent Air Voids in Compacted Dense and Open Asphalt Mixtures T 312, Preparing and Determining the Density of Asphalt Mixture Specimens by
Means of the Superpave Gyratory Compactor T 342, Determining the Dynamic Modulus of Hot Mix Asphalt (HMA) TP 107, Determining the Damage Characteristic Curve and Analysis Parameters
Using Small Specimens in the Asphalt Mixture Performance Tester (AMPT) Cyclic Fatigue Test
ASTM Standard:
D3549/D3549M, Standard Test Method for Thickness or Height of Compacted Bituminous Paving Mixture Specimens
III. Test Procedure
Dynamic Modulus (|E*|) is the absolute value of the complex modulus calculated by dividing the peak-to-peak stress by the peak-to-peak strain for a material subjected to a sinusoidal loading.
Phase Angle () is the angle in degrees between a sinusoidally applied stress and the resulting strain in a controlled stress test.
1. Dynamic Modulus Test
This test method describes the procedure for measuring the dynamic modulus of asphaltic concrete mixture using 38-mm diameter small specimen. A test specimen at a specific test temperature is subjected to a controlled sinusoidal (haversine) compressive stress of various frequencies. The applied stresses and resulting axial strains are measured as a function of time and used to calculate the dynamic modulus and phase angle.
115

a. Place the specimens to be tested in the environmental chamber with the "dummy" specimen and monitor the temperature of the "dummy" specimen to determine when testing can begin.
b. Place platens and friction reducers inside the testing chamber. Turn on the AMPT, set the temperature control to the desired testing temperature, and allow the testing chamber to equilibrate at the testing temperature for at least 1 h.
c. When the "dummy" specimen and the testing chamber reach the target temperature, open the testing chamber. Remove a test specimen from the conditioning chamber and quickly place it in the testing chamber.
d. Assemble the specimen to be tested with platens in the following order from bottom to top: bottom loading platen, bottom friction reducer, specimen, top friction reducer, and top loading platen.
e. Install the specimen-mounted deformation-measuring system on the gauge points per the manufacturer's instructions. Ensure that the deformationmeasuring system is within its calibrated range. Ensure that the top loading platen is free to rotate during loading.
f. Close the testing chamber and allow the chamber temperature to return to the testing temperature.
g. Procedures in Step (c) through Step (i), including the return of the test chamber to the target temperature, shall be completed in 5 min.
h. Enter the required identification and control information into the dynamic modulus software.
i. Follow the software prompts to begin the test. The AMPT will automatically unload when the test is complete and will display the test data and data quality indicators.
j. Review the data quality indicators as discussed in Step (e). Retest specimens with data quality indicators above the values specified in Step (e).
k. Once acceptable data have been collected, open the test chamber and remove the tested specimen. Repeat procedures in Step (c) through Step (k) for the remaining test specimens.
2. Computations and Data Quality
a. The calculation of dynamic modulus, phase angle, and the data quality indicators is performed automatically by the AMPT software.
b. Accept only test data meeting the data quality statistics given in Table 2. Table 3 summarizes actions that can be taken to improve the data quality statistic. Repeat tests as necessary to obtain test data meeting the data quality statistics requirements.
116

Table 2--Data Quality Statistics Requirements

Data Quality Statistic

Limit

Deformation drift

In direction of applied load

Peak-to-peak strain

50 to 75 strain

Load standard error

10%

Deformation standard error 10%

Deformation uniformity 30%

Phase uniformity

3

Table 3--Data Quality Statistics Requirements

Item

Cause

Possible Solutions

Deformation drift not in direction Gauge points are moving apart of applied load

Peak-to-peak strain too high Peak-to-peak strain too low

Load level too high Load level too low

Reduce LVDT spring force. Add compensation springs. Reduce test temperature. Reduce load level. Increase load level.

Load standard error >10% Deformation standard error >10%
Deformation uniformity >30%

Applied load not sinusoidal
1. Deformation not sinusoidal
2. Loose gauge point 3. Excessive noise on deformation signals 4. Damaged LVDT 1. Eccentric loading
2. Loose gauge point 3. Sample ends not parallel 4. Poor gauge point placement 5. Non-uniform air void distribution

Adjust tuning of hydraulics.
1. Adjust tuning of hydraulics. 2. Check gauge points. Reinstall if loose. 3. Check wiring of deformation sensors. 4. Replace LVDT.
1. Ensure specimen is properly aligned. 2. Check gauge points. Reinstall if loose. 3. Check parallelism of sample ends. Mill ends if out of tolerance. 4. Check for specimen non-uniformity (segregation, air voids). Move gauge points. 5. Ensure test specimens are cored from the middle of the gyratory specimen.

Phase uniformity >3

1. Eccentric loading
2. Loose gauge point 3. Poor gauge point placement 4. Damaged LVDT

1. Ensure specimen is properly aligned. 2. Check gauge points. Reinstall if loose. 3. Check for specimen non-uniformity (segregation, air voids). Move gauge points. 4. Replace LVDT.

3. Development of Mastercurve
Using Excel spreadsheet, mastercurve of dynamic modulus is generated in accordance with AASHTO R84.
Step 1: Copy and paste frequency, dynamic modulus, phase angle, and test temperature data into the pink cells within the measured data table. Each block of test data should correspond to a single replicate and temperature of

117

testing. Include data for each test specimen. Do not average data prior to entry into the spreadsheet. Step 2: Enter the desired reference temperature into the Tref cell of the shift factor table. Step 3: Run Solver
IV. Maintenance
This list should serve as a guideline for when the dynamic modulus test should be conducted re re-evaluated.
1. Per GDOT SOP 2, all mix designs shall be subjected to one or more field verifications during production at the discretion of the State Bituminous Construction Engineer. When field verification tests are conducted in accordance with SOP 2, dynamic modulus test should be conducted.
2. Field cores taken for performance evaluation and remaining design life prediction. 3. Asphalt mixtures with binder grades outside of PG 76-22, PG 67-22, and PG 64-22. 4. Asphalt mixtures with a reclaimed asphalt pavement (RAP) above 30% or below
25%. 5. Asphalt mixtures that use limestone as an aggregate source. 6. Asphalt mixtures in pavements that failed to reach the design life. 7. The dynamic modulus test is conducted in accordance with AASHTO T342 (4-in.
diameter specimen) or TP107 (1.5-in. diameter specimen) at the standard test temperatures and frequencies from the dynamic modulus mastercurve of each replicate specimen. This SOP introduce to use small specimen to measure dynamic modulus. The dynamic modulus test results of the small and large specimens generally differ significantly at 54C whereas the majority of the mixtures evaluated demonstrated statistically equivalent dynamic modulus results at low and intermediate temperatures. Additionally, at low and intermediate temperatures, COV values are less than 15%, indicating that specimen-to-specimen variability is within the generally accepted range. Therefore, it is recommended to limit small specimen testing to the temperatures outlined in AASHTO PP61, which specifies three test temperatures with the highest temperature selected as a function of the Performance Grade (PG). The highest temperature specified by AASHTO PP61 ranges between 35C and 45C for different asphalt binder PG grades. In Georgia, the recommended three temperatures for dynamic modulus test are 4oC, 20oC, and 40oC. The observed difference in the mastercurve at high temperature does not significantly affect pavement fatigue performance predictions (Castorena et al., 2017).
118

Laboratory SOP DRAFT
Georgia Department of Transportation
Office of Materials and Testing
Proposed Standard Operating Procedure (SOP) DRAFT Asphalt Mixture Test to Evaluate Fatigue Cracking Resistance Performance
I. General The purpose of this Standard Operating Procedure is to outline the methodology for maintaining, managing, and updating the materials library as it relates to the Asphalt Mixture Database. All tests are designed to be performed with an asphalt mixture performance tester (AMPT).
II. Specimen Fabrication
A. Semicircular Bend (SCB)
This test procedure is for asphaltic concrete mixtures composed of aggregates with an NMAS of 19 mm or less. The following instructions detail how to prepare a specimen for testing.
1. Compact an asphalt mixture into a cylinder with a height of 178 mm, a diameter of 150 mm, and a target air void percentage of 7 0.5 %, per AASHTO T 312.
2. Cut two 50 mm thick disks from the compacted cylinder. 3. Cut each of the two disks in half, creating a total of four semicircle shapes with a
thickness of 50 mm. 4. Cut a notch in the middle of the semicircle perpendicular to the flat surface with a
length of 15 1 mm. 5. Measure the bulk specific gravity of each cut specimen according to AASHTO T
166 to determine the air voids of the specimens. 6. Condition the test specimens in the environmental chamber at a temperature of 25
0.5 C for 2 0.5 hr.
B. Direct Tension Test for S-VECD
This test procedure is for asphaltic concrete mixtures composed of aggregates with an NMAS of 25 mm or less. The following instructions detail how to prepare a specimen for testing.
1. Compact an asphalt mixture into a cylinder with a height of 178 mm, a diameter of 150 mm, and a target air void percentage of 7 0.5 % per AASHTO T 312.
2. Core vertically, from the inner 100 mm diameter, four cylindrical specimens with a diameter of 38 mm.
119

3. Cut the top and bottom of the 100 mm specimens to form a cylinder with a height of 110 1 mm.
4. Measure the bulk specific gravity of each cut specimen according to AASHTO T 166 to determine the air voids of the specimens.
5. Glue the mounting studs to the 100 mm cylindrical specimen using a gage length of 70 1 mm center to center. Note: For gluing, use Devcon 10240 plastic steel putty, or another putty with equivalent properties and a spacing fixture to ensure an accurate gage length.
6. Clean the end plates to ensure the grooves are free of any debris or old glue by heating them in an oven or by soaking in acetone.
7. Glue the end plates to the cylindrical specimen using approximately 7 grams of Devcon 10240 plastic steel putty.
8. Divide the glue into quarters using each to spread over the four contact areas, ensuring the glue fills all the grooves in the end plates.
9. Use a gluing jig to center the specimen on top of the plates to ensure when the load is not eccentrically applied and that screw holes align with the AMPT.
10. Lower the gluing jig's weight onto the specimen and allow for initial set to occur before moving the specimen.
11. Move the specimen after initial set to an environmental chamber for conditioning. Note: Hold the specimen from the bottom so that tension is not applied to the adhesive.
12. Condition the specimens at one of the following temperatures based on the asphalt binder grade: a. PG 64-22: 18C b. PG 67-22: 19.5C c. PG 76-22: 21C
13. Allow the glue to fully cure based on the manufacturer's curing time before testing.
III. Test Procedure
A. Semicircular Bend (SCB)
This method covers the determination of fracture energy (Gf) and the post-peak slope, using semicircular asphalt specimens at 25C. These parameters are used to calculate the flexibility index (FI), which ranks asphalt mixtures based on their resistance to asphalt cracking. A mixture with a higher FI indicates better pavement performance than a mixture with a lower FI.
1. Place the SCB testing apparatus inside the AMPT test chamber. 2. Turn on the AMPT and set the climate-controlled chamber to the test temperature
of 25C. 3. Allow the chamber to reach the test temperature before proceeding. 4. Remove one of the cut specimens from its conditioning chamber and quickly place
it inside the AMPT to keep the specimen temperature constant. 5. Place the specimen with the flat side on the rollers of the testing apparatus and the
loading head centered above the notch.
120

6. Ensure that the specimen is centered in both the x and y directions on the testing apparatus.
7. Place the steel ball bearing on top of the loading head. 8. Lower the AMPT chamber. Note: Steps 48 should be completed in less than 5
minutes to help maintain testing temperature. 9. Apply an initial contact load of 0.1 0.01 kN at a loading rate of 0.05 kN/s. 10. Set the linear load displacement (LLD) control to a rate of 50 mm/min. 11. Set the test termination to when load drops below 0.1 kN and start the test.
Fracture Energy (Gf) Calculate by dividing the work of fracture (Wf), which is the total area under the load line displacement curve in Figure C.1, by the ligament of the area (Alig) using Eq. (C.1) and Eq. (C.2)

Wf

FIGURE C.1 Load vs Load Line Displacement Curve

G =
f

Alig = t(ra)

Where,

t =thickness of specimen (mm) r = radius of specimen (mm) a = notch length of specimen (mm)

121

(C.1) (C.2)

Determine the inflection point of the curve after the peak load. Calculate the slope of the tangential line passing through the inflection point. Designate this slope as m. Calculate FI using Eq. (C.3).

FI

=

||





(C.3)

Where, A = 0.01 (unit-less conversion factor)
B. Direct Tension for S-VECD
This test method covers the procedure for testing asphalt concrete mixtures to determine the damage characteristic curve and fatigue analysis parameters (i.e., DR, Sapp) via the direct tension cyclic fatigue test using the AMPT. The failure criterion DR when used with the mixture's modulus can determine the mixture's Sapp value. A higher Sapp value indicates better pavement performance than a lower value.
1. Turn on the AMPT and set the climate-controlled chamber to the correct testing temperature based on the asphalt binder grade.
2. Attach spacers to the machine to compensate for the reduced height of the specimen. 3. Remove the specimen from the conditioning chamber and insert it into the AMPT. 4. Tighten the bottom platen to the bottom support. Note: A torque wrench is useful
to use here with the torque set to 12 N*m. 5. Raise the actuator by applying a seating load of 0.01 kN. 6. Check the top platen to determine if it is level. 7. Proceed to Step 10 if the top platen is level; otherwise, continue with Step 8. 8. Loosen the screws on the bottom spacer until the top platen sits level with the
machine surface. 9. Insert feeler gauges as needed under the bottom spacer and retighten the screws,
ensuring the top platen does not move. 10. Insert screws through the top platen into the ring at the top of the machine, but do
not tighten the screws. 11. Use feeler gauges on each side of the screws to fill all the gaps before tightening
the screws. 12. Tighten the top platen to the ring using the torque wrench in increments starting at
4 N*m of torque and going up to 12 N*m. Note: This increased incremental tightening is to ensure that the specimen does not break. 13. Reduce the load to 0 kN after the top platen has been filly secured. 14. Attach the three strain gauges to the specimen and adjust the stain gauges so that the displacement during testing will not exceed the gauges' limits. 15. Lower the cell and allow for adequate time for the specimen to reach its testing temperature. 16. Input a test frequency of 10 Hz with a target strain range of 5075 microstrain in the tension compression mode of loading for the dynamic modulus fingerprint test.

122

17. Start the fingerprint test and ensure the strain does not exceed 150 microstrain and reaches a minimum of 50 data points per cycle.
18. Allow the specimen to rest for 20 minutes after the fingerprint test. 19. Set the cyclic fatigue test to run a pull-pull actuator displacement. 20. Set the target peak to peak on-strain amplitude for the first test specimen to 300,
500, or 800 microstrain (os1) based on the |E*|fingerprint ranges in Table C.1. Note: These are suggested values, but after running numerous tests, operators may have a more educated guess of the expected cycle count for an asphalt mixture at a strain level. A good test is a cycle count above 2000.

TABLE C.1 Target on Specimen Strain Levels for the First Specimen

Case (units in MPa)
|E*|fingerprint > 8,800 4,400 < |E*|fingerprint < 8,800
|E*|fingerprint < 4,400

os1
300 500 800

21. Manually terminate the test when the phase angle beings to drop. 22. Export the test results to FlexMAT for analysis. 23. Repeat the necessary steps to complete a total of three tests. 24. Adjust the cyclic fatigue test target strain level so that each test uses a different
target strain level. Note: It is important to try to have tests with at least one cycle count above 10,000. The lower the target strain, the more likely a test will have higher cycle counts.
Figure C.2 shows the results from a successful test. The test was terminated as soon as the phase angle began to drop.

123

Phase Angle () Dynamic Modulus (kPa)

40 35 30 25 20 15 10
5 0
0

9,000

8,000

7,000

6,000

5,000

4,000

3,000

Phase Angle |E*|
20,000 40,000 60,000 Number of Cycles

2,000
1,000
0 80,000

FIGURE C.2

Expected Results of a Cyclic Fatigue Test

IV. Maintenance
This list should serve as a guideline for when the asphalt mixtures should be added and re-evaluated.
1. Field cores taken for performance evaluation. 2. Asphalt mixtures with binder grades outside of PG 76-22, PG 67-22, and PG 64-22. 3. Asphalt mixtures with a reclaimed asphalt pavement (RAP) above 30% or below
25%. 4. Asphalt mixtures that use limestone as an aggregate source. 5. Asphalt mixtures in pavements that failed to reach the design life.

124