NCHRP 9-44

THE NATIONAL ACADEMIES PRESS
This PDF is available at http://nap.edu/23277 SHARE

   
An Experimental Plan for Validation of an Endurance Limit

for HMA Pavements (2009)
DETAILS
265 pages | 8.5 x 11 | PAPERBACK

ISBN 978-0-309-43692-2 | DOI 10.17226/23277
CONTRIBUTORS
GET THIS BOOK
FIND RELATED TITLES
SUGGESTED CITATION
National Academies of Sciences, Engineering, and Medicine 2009. An Experimental

Plan for Validation of an Endurance Limit for HMA Pavements. Washington, DC:
The National Academies Press. https://doi.org/10.17226/23277.

Visit the National Academies Press at NAP.edu and login or register to get:
– Access to free PDF downloads of thousands of scientific reports

– 10% off the price of print titles
– Email or social media notifications of new titles related to your interests
– Special offers and discounts

Distribution, posting, or copying of this PDF is strictly prohibited without written permission of the National Academies Press.
(Request Permission) Unless otherwise indicated, all materials in this PDF are copyrighted by the National Academy of Sciences.
Copyright © National Academy of Sciences. All rights reserved.

An Experimental Plan for Validation of an Endurance Limit for HMA Pavements
ACKNOWLEDGMENT
This work was sponsored by the American Association of State

Highway and Transportation Officials (AASHTO), in cooperation with
the Federal Highway Administration, and was conducted in the
National Cooperative Highway Research Program (NCHRP), which is
administered by the Transportation Research Board (TRB) of the
National Academies.
COPYRIGHT PERMISSION
Authors herein are responsible for the authenticity of their materials

and for obtaining written permissions from publishers or persons who
own the copyright to any previously published or copyrighted material
used herein.
Cooperative Research Programs (CRP) grants permission to

reproduce material in this publication for classroom and not-for-profit
purposes. Permission is given with the understanding that none of the
material will be used to imply TRB, AASHTO, FAA, FHWA,
FMCSA, FTA, Transit Development Corporation, or AOC
endorsement of a particular product, method, or practice. It is expected
that those reproducing the material in this document for educational
and not-for-profit uses will give appropriate acknowledgment of the
source of any reprinted or reproduced material. For other uses of the
material, request permission from CRP.
DISCLAIMER
The opinion and conclusions expressed or implied in the report are

those of the research agency. They are not necessarily those of the
TRB, the National Research Council, AASHTO, or the U.S.
Government.
This report has not been edited by TRB.
Copyright National Academy of Sciences. All rights reserved.


NCHRP Web-Only Document 134: An Experimental Plan for Validation of an Endurance Limit for HMA Pavements
TABLE OF CONTENTS
List of Figures ................................................................................................................................ iv
List of Tables ...................................................................................................................................v
ACKNOWLEDGMENTS ............................................................................................................. vi
ABSTRACT.................................................................................................................................. vii
SUMMARY................................................................................................................................. viii
1. INTRODUCTION AND RESEARCH APPROACH .................................................................1
1.1 Introduction.......................................................................................................................... 1
1.2 Problem Statement and Research Objective ........................................................................ 1
1.3 Research Approach .............................................................................................................. 2
2. FINDINGS................................................................................................................................13
2.1 HMA Endurance Limit ...................................................................................................... 13
2.2 Framework for Incorporating Endurance Limit Behavior in Flexible Pavement Design.. 15
3. INTERPRETATION, APPRAISAL AND APPLICATIONS..................................................30
4. CONCLUSIONS AND RECOMMENDATIONS ...................................................................35
REFERENCES ..............................................................................................................................36
APPENDIX A. HMA ENDURANCE LIMIT WORKSHOP EXECUTIVE SUMMARY ........ A-i
APPENDIX B. HMA ENDURANCE LIMIT VALIDATION STUDY RESEARCH PLAN ... B-i
iii

LIST OF FIGURES
Figure 1. HMA Endurance Limit Workshop Agenda.................................................................... 5

Figure 2. Results of Flexural Fatigue Tests by Carpenter et al., Including Extrapolated
Results at Low Strain Levels . ...................................................................................... 14
Figure 3. Effect of Rest Periods on Plateau Value ...................................................................... 17
Figure 4. Effect of Rest Period on Fatigue Life........................................................................... 18
Figure 5. Plateau Value for Continuous Loading as a Function of Applied Strain Level. .......... 20
Figure 6. Application of Time-Temperature Superposition to Rest Periods. .............................. 23
Figure 7. Comparison of Applied and Allowable Strains............................................................ 27
Figure 8. Example of Minimum Asphalt Thicknesses to Resist Bottom Initiated Fatigue
Cracking With Observed Performance of Four UK Pavement Sections. ..................... 28
Figure 9. Project Flow Chart........................................................................................................ 30
Figure 10. Overall Project Schedule. ........................................................................................... 32
iv

LIST OF TABLES
Table 1. HMA Endurance Limit Workshop Invitees and Attendees. ............................................ 4

Table 2. Summary of Existing Pavement Analysis Approaches Considered. ............................... 7
Table 3. Summary of Proposed Laboratory Experiments.............................................................. 8
Table 4. Matrix for Field Calibration of the Allowable Strain Limit Design Procedure............. 11
Table 5. Effect of Rest Period on Fatigue Life. ........................................................................... 18
Table 6. Approximate Rest Periods for Various Design Traffic Levels...................................... 19
Table 7. Allowable Strains for Various Design Traffic Levels. .................................................. 22
Table 8. Computation of Allowable Strain Strains...................................................................... 25
Table 9. Applied Strains for Design Example. ............................................................................ 26
Table 10. Summary of Man-hour and Cost Estimates................................................................. 31

ACKNOWLEDGMENTS
The research reported herein was performed under NCHRP Project 9-44 by Advanced
Asphalt Technologies, LLC. Ramon Bonaquist, Chief Operating Officer for Advanced Asphalt
Technologies, LLC, served as Principal Investigator for the project and authored this report and
the research plan which was the primary product of NCHRP Project 9-44. Mr. Harold Von
Quintus, Principal Engineer for Applied Research Associates, Inc. co-authored the research plan.
NCHRP Project 9-44 included a facilitated workshop to discuss current asphalt concrete
fatigue research and to identify alternatives for incorporating an endurance limit for asphalt
concrete in mechanistic-empirical pavement design. Mr. Charles Markert, President of Dynamic
Leadership Consulting Group, facilitated the workshop. Mrs. Rane Wagner, President of Rane
Wagner and Associates, provided logistical support of the workshop. Special thanks are extended
to the participants of the HMA Endurance Limit Workshop listed below. These professionals
unselfishly presented ideas to the research team that helped shape the planned research.
Name Affiliation
Dr. David Anderson Consultant
Dr. Samuel Carpenter University of Illinois
Dr. Donald Christensen Advanced Asphalt Technologies, LLC
Dr. Herve Di Benedetto Ecole Nat. des TPE
Mr. Bruce Dietrich Florida Department of Transportation
Mr. Kenneth Fults KWF Pavement Consulting
Mr. Roger Green Ohio Department of Transportation
Dr. Kevin Hall University of Arkansas
Dr. Edward Harrigan National Cooperative Highway Research Program
Dr. Richard Kim North Carolina State University
Dr. Dallas Little Texas A&M University
Dr. Leslie Ann McCarthy Federal Highway Administration
Dr. Andre Molenaar Delft University
Professor Carl Monismith University of California Berkeley
Dr. David Newcomb National Asphalt Pavement Association
Dr. Michael Nunn Lane One Limited
Dr. Brian Prowell Advanced Material Services, LLC
Dr. Rey Roque University of Florida
Ms. Amy Schutzbach Illinois Department of Transportation
Dr. Jacob Uzan Technion University
Dr. Linbing Wang Virginia Polytechnic and State University
Dr. Matthew Witczak Arizona State University
vi

ABSTRACT
This report documents the work completed in National Cooperative Highway Research
Program (NCHRP) Project 9-44. The objective of NCHRP Project 9-44 was to prepare a
research plan and associated cost estimate for a future study to validate the endurance limit for
HMA and to improve mechanistic-empirical pavement design. The primary product of NCHRP
Project 9-44 is the HMA Endurance Limit Validation Study Research Plan.
The planned research is based on the hypothesis that the endurance limit for HMA is the
result of a balance of damage caused by loading and healing or damage recovery that occurs
during rest periods. Under this hypothesis the primary objective in designing a flexible
pavement to resist bottom initiated fatigue cracking will be to make sure that the damage induced
by loading remains small enough so that healing occurs and there is no accumulation of damage
over the life of the pavement. This is a significant departure from current cumulative or
incremental damage models, which assume that no healing occurs and that each load cycle uses
up a portion of the finite fatigue life of the HMA.
This research plan includes the framework for a design procedure that is based on layered
elastic analysis and compatible with the Mechanistic-Empirical Pavement Design Guide
(MEPDG). It uses allowable strains to identify satisfactory conditions for full healing. The
allowable strains are a function of the properties of the HMA, the pavement temperature, and the
duration of rest periods between traffic loads. Five laboratory experiments that are needed to
fully develop the procedure are described. Studies using data from completed accelerated
pavement tests and test roads are proposed to verify critical aspects of the design procedure.
Finally, an experiment to calibrate the design procedure using selected test sections from the
Long Term Pavement Performance Program is presented.
vii

SUMMARY
This report documents the work completed in National Cooperative Highway Research
Program (NCHRP) Project 9-44. The objective of NCHRP Project 9-44 was to prepare a
research plan and associated cost estimate for a future study to validate the endurance limit for
HMA and to improve mechanistic-empirical pavement design. The primary product of NCHRP
Project 9-44 is the HMA Endurance Limit Validation Study Research Plan. The research plan
was formulated from a review of relevant research and the recommendations from a workshop
that included participation by a number of international experts.
Completed laboratory and field investigations clearly show that hot mix asphalt (HMA)
exhibits endurance limit behavior. This endurance limit, however, does not reflect an absence of
load induced damage in the HMA. It is the result of a balance of damage caused by loading and
healing or damage recovery that occurs during rest periods. The endurance limit for HMA is,
therefore, not a single value, but will change depending on the loading and environmental
conditions applied to the HMA. To properly consider this form of an endurance limit in flexible
pavement design requires consideration of the effects of loading, environment and material
properties on both damage accumulation and healing. These findings concerning the endurance
limit for HMA served as the research hypothesis upon which the HMA Endurance Limit
Validation Study Research Plan was formulated.
To effectively design laboratory and field experiments for the HMA Endurance Limit
Validation Study, the framework for a pavement design procedure that incorporates healing and
endurance limit behavior was conceived during NCHRP 9-44. The procedure is based on
layered elastic analysis and is compatible with the Mechanistic-Empirical Pavement Design
Guide (MEPDG). It uses allowable strains to identify satisfactory conditions for full healing.
The allowable strains are a function of the properties of the HMA, the pavement temperature,
and the duration of rest periods between traffic loads. The underlying principal of the design
procedure is to make sure that the damage induced by loading remains small enough so that full
healing occurs between traffic loads and there is no accumulation of damage over the life of the
pavement. This is a significant departure from current cumulative or incremental damage
viii

models, which assume that no healing occurs and that each load cycle uses up a portion of the
finite fatigue life of the HMA.
The HMA Endurance Limit Validation Study is included as a stand-alone appendix to this
report. It is a comprehensive plan for research to rationally incorporate the concept of an
endurance limit for HMA into a mechanistic-empirical algorithm for bottom initiated fatigue
cracking in flexible pavements, and to validate the resulting procedure using performance data
from full-scale pavement sections. The plan presents details of five laboratory experiments that
are needed to fully develop the procedure. Studies using data from completed accelerated
pavement tests and test roads are proposed to verify critical aspects of the design procedure.
Finally, an experiment to calibrate the design procedure using selected test sections from the
Long Term Pavement Performance Program is presented.
ix

1. INTRODUCTION AND RESEARCH APPROACH
1.1 Introduction
This report documents the research completed in National Cooperative Highway Research
Program (NCHRP) Project 9-44, Developing a Plan for Validating an Endurance Limit for HMA
Pavements. For hot mix asphalt (HMA) pavements, the endurance limit has been defined as a
level of strain below which there is no cumulative damage over an indefinite number of load
cycles (1). The endurance limit is an important concept in the design of long life flexible
pavements that is gaining increasing acceptance worldwide. Appropriate application of the
endurance limit in flexible pavement design will lead to more effective pavement sections with
significant benefit and cost savings to the public.
1.2 Problem Statement and Research Objective

1.2.1 Problem Statement
The endurance limit, as applied to HMA and flexible pavement design, is a strain level below
which the fatigue life of the HMA is infinite and the pavement will not experience bottom-up
fatigue cracking. Current mechanistic-empirical fatigue criteria for HMA, including the field
calibrated criterion in the Mechanistic Empirical Pavement Design Guide (MEPDG), assume the
fatigue life of HMA to be a power function of the tensile strain at the bottom of the asphalt layer.
These criteria do not include the provision for an endurance limit. There is mounting evidence
that an endurance limit for asphalt concrete does exist. It has been observed in laboratory studies
of fatigue at low strain levels, and several documented cases studies indicate that bottom-up
fatigue cracking is almost non-existent in properly constructed, thick asphalt concrete
pavements. A concentrated research effort, however, is needed to validate the endurance limit
concept, and to devise effective methods for incorporating it in mechanistic-empirical pavement
design methods.
1.2.2 Objective
The objective of NCHRP Project 9-44 was to prepare a research plan and associated cost
estimate for a future study to validate the endurance limit for HMA and to improve mechanistic-

empirical pavement design. To be successful, the research plan must address the following:
1. Validation of the existence of an endurance limit for HMA in pavements through an

analysis of laboratory and field data;
2. Potential differences in the endurance limit measured in the laboratory and observed
in field performance; and
3. Identification of a recommended methodology for incorporating an asphalt concrete
endurance limit in mechanistic-empirical pavement design.
1.3 Research Approach

The research approach taken in NCHRP Project 9-44 was to synthesize information gathered
from a review of relevant research and a workshop with invited experts to develop a
comprehensive work plan and budget for a future project to validate the endurance limit for
HMA and improve mechanistic-empirical pavement. NCHRP Project 9-44 included six major
tasks, which are briefly described below.
1.3.1 Task 1. Review Relevant Research.

In this task, published research associated with the endurance limit and the design of flexible
pavements and HMA mixtures to resist fatigue cracking was reviewed. Information obtained in
Task 1 was used to select topics for the facilitated workshop that was conducted in Task 2 and to
develop the overall approach for incorporating the endurance limit in mechanistic-empirical
pavement design. This review focused on the following key topics:
• Laboratory endurance limit studies,

• Alternative forms for fatigue testing,
• Approaches for incremental damage analysis,
• Laboratory studies on healing and damage tolerance,
• Field studies of measured strains in thick flexible pavements,
• Case studies of long life flexible pavements.

1.3.2 Task 2. Conduct Facilitated Workshop.

Task 2 included the planning, execution, and documentation of a facilitated workshop
directed at evaluating various methodologies for HMA fatigue characterization, and strategies for
incorporating an endurance limit in mechanistic-empirical design. Recommendations from the
workshop shaped the research plan produced in NCHRP Project 9-44.
The HMA Endurance Limit Workshop was held in Washington, D.C. on August 1 and 2,
2007. Participants included members of the NCHRP Project 9-44 panel and research team, key
researchers and consultants with extensive experience in HMA fatigue analysis, and engineers
from highway agencies who are responsible for designing, constructing, and maintaining flexible
pavements. Thirty-four individuals were invited to attend, including four members of the
research team, the NCHRP Project 9-44 panel members and Project Manager, and 22 invited
experts. There was a high degree of interest in the workshop with 82 percent of the invitees
participating. Table 1 presents summary information about the participants.
The HMA Endurance Limit Workshop was facilitated by Mr. Charles Markert, a Certified
Professional Facilitator. The agenda for the workshop was developed jointly by Mr. Markert and
Dr. Ramon Bonaquist, Principal Investigator for NCHRP Project 9-44. A copy of the agenda is
reproduced as Figure 1. The key element of the workshop was a series of discussion sessions
focusing on four major topics considered important to validating an endurance limit for HMA
pavements:
• Endurance limit and other important fatigue effects,

• Methodologies for HMA fatigue characterization,
• Strategies for incorporating an endurance limit in flexible pavement damage analysis,
and
• Approaches for calibrating and validating pavement analysis methods that include an
endurance limit.
A summary report documenting the HMA Endurance Limit Workshop was prepared. This
report in included as Appendix A.

Table 1. HMA Endurance Limit Workshop Invitees and Attendees.
No Name Affiliation Role Attend

1 Dr. David Anderson Consultant Panel Member Y
2 Dr. Ramon Bonaquist Advanced Asphalt Technologies, LLC Research Team Y
3 Dr. Stephen Brown University of Nottingham Invited Expert N
4 Dr. William Buttlar University of Illinois Invited Expert N
5 Dr. Samuel Carpenter University of Illinois Invited Expert Y
6 Dr. Donald Christensen Advanced Asphalt Technologies, LLC Research Team Y
7 Mr. Danny Dawood Pennsylvania Department of Panel Member N
Transportation
8 Dr. Herve Di Benedetto Ecole Nat. des TPE Invited Expert Y
9 Mr. Bruce Dietrich Florida Department of Transportation Panel Member Y
10 Dr. Jon Epps Granite Construction Company, Inc. Invited Expert N
11 Mr. Kenneth Fults KWF Pavement Consulting Invited Expert Y
12 Mr. Roger Green Ohio Department of Transportation Panel Member Y
13 Dr. Kevin Hall University of Arkansas Invited Expert Y
14 Dr. Edward Harrigan National Cooperative Highway Panel Member Y
Research Program
15 Mr. Frederick Hejl Transportation Research Board TRB Liaison N
16 Dr. Richard Kim North Carolina State University Invited Expert Y
17 Dr. Dallas Little Texas A&M University Invited Expert Y
18 Dr. Robert Lytton Texas A&M University Invited Expert N
19 Dr. Leslie Ann McCarthy Federal Highway Administration Panel Member Y
20 Mr. Charles Markert Dynamic Leadership Consulting Group Research Team Y
21 Dr. Andre Molenaar Delft University Invited Expert Y
22 Professor Carl Monismith University of California Berkeley Invited Expert Y
23 Dr. David Newcomb National Asphalt Pavement Association Invited Expert Y
24 Dr. Michael Nunn Lane One Limited Invited Expert Y
25 Ms. Linda Pierce Washington Department of Invited Expert N
Transportation
26 Dr. Brian Prowell Advanced Material Services, LLC Invited Expert Y
27 Dr. Rey Roque University of Florida Invited Expert Y
28 Ms. Amy Schutzbach Illinois Department of Transportation Panel Member Y
29 Mr. Darin Tedford Nevada Department of Transportation Panel Member N
30 Dr. Jacob Uzan Technion University Invited Expert Y
31 Mr. Harold Von Quintus Applied Research Associates Invited Expert Y
32 Dr. Linbing Wang Virginia Polytechnic and State Panel Member Y
University
33 Ms. Rane Wagner Rane Wagner and Associates Research Team Y
34 Dr. Matthew Witczak Arizona State University Invited Expert Y

AGENDA Item Questions to be answered

Sponsor Welcome, Facilitator Opening, Introductions
Personal Expectations What are your expectations for this session?
Briefing 1 Fatigue in the MEDPG How are fatigue and the endurance limit addressed in
the MEPDG?
Briefing 2 NCHRP 9-38 Laboratory Evaluation of What was found about the endurance limit in NCHRP
Project 9-38?
Endurance Limit
Briefing 3 A Review of UK Pavement Design What approach is taken in the UK?
Purpose Discussion What is the purpose of this session?
Discussion: Existence of Fatigue Endurance Limit Does a Fatigue Endurance- Limit Exist?
Continue Discussion Does a Fatigue Endurance- Limit Exist?
Issues Related to Fatigue -Endurance- Limit “What are the other issues?”
Identify Major Issues Which of these possibilities go on the Short List?
Place your dots – one per card.
What is the Meaning of Each Major Issues? Discussion on the top few.
Discussion: Alternative Methodologies for Characterizing Do we need alternative to Beam Fatigue? Can they
Fatigue address endurance limit? Are they implementable?
Flexible Pavement Damage Analysis How can we improve flexible pavement damage
analysis? What important issues are not currently
addressed?
Plus/Delta
Adjourn
DAY TWO
Review Agenda/Progress/Issues
Strategies for Incorporating Endurance Limit in Flexible What are the possible strategies for incorporating
Pavement Damage Analysis Endurance Limit?
Identify Barriers to Success? What are the barriers to success in this effort?
Identify Countermeasures? What are some countermeasures?
Identify Simplifying Assumptions Can we identify some simplifying assumptions that
will help?
Calibration/Verification Is calibration necessary?
How should it be done? Field sections, accelerated
pavement testing, etc?
Suggest Data Evaluation Approaches What are your suggestions for calibration/verification?
Identify Potential Action for NCHRP 9-44 What should be included in the workplan for future
research developed in NCHRP 9-44?
Silver Bullet Actions Which suggestions from Potential Actions can be
addressed in a 3-year project?
Recommendations, Findings & Conclusions What are your recommendations, findings &
conclusions as a group?
Closing
ADJOURN
Figure 1. HMA Endurance Limit Workshop Agenda.

1.3.3 Task 3. Identify Data Requirements

A major part of Task 3 was the development of a framework for designing pavements to resist
bottom initiated fatigue cracking that considers the effects of an endurance limit. The framework
was developed to identify specific laboratory studies needed to fully develop the design
procedure and the types of pavement test section data needed for the validation.
The framework of the design procedure that was developed is based on the following research
hypothesis that emanated from the HMA Endurance Limit Workshop. The endurance limit for
HMA does not reflect an absence of load induced damage in the HMA. It is the result of a
balance of damage caused by loading and healing or damage recovery that occurs during rest
periods. Under this hypothesis the primary objective in designing a flexible pavement to resist
bottom initiated fatigue cracking is to make sure that the damage induced by loading remains
small enough so that healing occurs and there is no accumulation of damage over the life of the
A number of approaches for designing pavements to resist bottom initiated fatigue cracking
were identified in Task 1. Table 2 briefly summarizes the approaches that were considered.
These range from relatively simple modifications of traditional mechanistic-empirical fatigue
algorithms to sophisticated finite element models based on damage mechanics and fracture
mechanics. The major deficiency of the more practical approaches is that they do not account
for the beneficial effects of healing. In the HMA Endurance Limit Workshop, healing was
identified as a significant factor affecting the endurance limit in HMA (1). The sophisticated
approaches can account for healing, but are not practical at this time for use in routine pavement
design. Since an acceptable existing design procedure could not be identified, the framework for
a new design procedure was developed. It is based on limiting strains at the bottom of the lowest
asphalt bound layer to those that will permit full healing to occur between traffic loads. This
approach results in lower allowable strains for conditions that result in less healing: higher traffic
volumes and colder temperatures. Chapter 2 includes a description of the framework for the new
design procedure.

Table 2. Summary of Existing Pavement Analysis Approaches Considered.
Approach Key Elements Selected Advantages Disadvantages

References
Strain Limit Assume fatigue life is infinite at Timm and Easy to implement in existing M-E Does not consider the beneficial
damage levels below the endurance Young (2) design. effect of rest periods.
limit. Use Miner’s law for strain Witczak (3) Compatible with layered elastic Relies on Miners law for strains
levels above the endurance limit. Thompson and analysis used in MEPDG. above the endurance limit.
Carpenter (4) Above endurance limit fatigue life

of HMA is predefined.
Crack Initiation Limit strain level to that causing Sidess and Easy to implement in existing M-E Does not consider the beneficial
crack initiation in laboratory fatigue Uzan (5) design. effect of rest periods.
tests. Compatible with layered elastic Relies on Miners law.
analysis used in MEPDG. Cycles to crack initiation are
Rational basis for design. predefined.
Strain Limit-Crack Assume fatigue life is infinite at Von Quintus (6, Relatively easy to implement in Does not consider the beneficial
Initiation damage levels below the endurance 7) existing M-E based methods. effect of rest periods.
limit. Use Miner’s law for strain Compatible with layered elastic Relies on Miner’s law for strains
levels above the endurance limit. The analysis used in the MEPDG. above the endurance limit.
endurance limit is estimated from the Value is dependent on the Key property used to estimate
indirect tensile strength test and is temperature (modulus), and endurance limit is highly variable.
dependent on the modulus of the volumetric properties of the
mixture. mixture.
Recursive Miner’s Law Modify fatigue life of HMA to Tsai, et al., (8) Easy to implement in existing M-E Assumes that HMA fatigue life
account for the strength loss of a design. deteriorates with traffic loading.
pavement structure as a function of Compatible with layered elastic Does not consider the beneficial
traffic loading. analysis used in MEPDG. effect of rest periods.
Accounts for changes in fatigue life
of HMA with traffic.
Visco-Elastic Model the evolution of damage in a Mun, et al., (9) Can be used to predict crack Computationally intensive.
Continuum Damage viscoelastic continuum. initiation. Not compatible with layered elastic
Directly accounts for damage analysis used in MEPDG.
accumulation and healing.
Fracture Mechanics Model responses at the crack tip and Roque, et al. Predict crack growth. Requires crack initiation model.
the propagation of cracks. (10) Computationally intensive.
Not compatible with layered elastic
analysis used in MEPDG.
7
Five experiments were identified to full develop the design procedure incorporating an HMA
endurance limit. Table 3 summarizes the laboratory experiments that are needed. The
experiments are briefly described below. Details of these experiments are included in the HMA
Endurance Limit Validation Study Research Plan that is presented in Appendix B.
Table 3. Summary of Proposed Laboratory Experiments.
Experiment Topic Factors

1 Mixture Compositional Factors • Binder Type
Affecting Healing in HMA • Binder Age
• Effective Binder Content
• Air Voids
• Design Compaction
• Gradation
• Filler Content
2 Effect of Applied Strain on • Strain Level
Healing • Healing Rate From Experiment 1
3 Effect of Temperature and Rest • Temperature
Period Duration on Healing • Rest Period Duration
4 Development of Testing and • Healing Rate From Experiment 1
Analysis Procedures to Determine • Mixtures From NCHRP 9-38
Allowable Strain Levels
5 Estimation of Allowable Strain • Mix Compositional Factors Affecting
Levels from Mixture Composition Damage Accumulation
• Significant Factors From Experiment 1
• Temperature
• Rest Period Duration
Experiment 1 is a screening study to identify the mixture compositional factors that affect
healing and therefore, the allowable strain levels in HMA. The results from this experiment will
be used in the remaining experiments. Experiment 2 addresses a major assumption that was
made in developing the allowable strain limit procedure, that is, the healing rate is independent
of the applied strain level. In this experiment healing rates will be determined using different
strain levels. This experiment will be conducted on mixtures from Experiment 1 that have high
and low healing rates. Experiment 3 is a study to verify the applicability of time-temperature
superposition to healing in HMA. This was the second major assumption included in the
development of the allowable strain limit procedure. Experiment 3 will be conducted on a
mixture from Experiment 1 that exhibits a moderate healing rate. Testing and analysis methods

for determining allowable strain limits that result in complete healing will be developed in
Experiment 4. This experiment will include testing and analysis of selected mixtures from
Experiment 1 and mixtures used in the endurance limit testing completed in NCHRP Project 9-
38. This experiment will generate the Level 1 test procedure for use with a future modified
version of the MEPDG. In the last experiment, Experiment 5, a wide range of mixtures will be
tested using the methods developed in Experiment 4 to develop predictive models relating the
allowable strain limits to mixture compositional factors. This last experiment will generate the
relationships between allowable strain and easily measured mixture compositional properties that
will be used in calibrating the procedure and thus verifying the endurance limit for HMA.
These relationships will also provide the Level 2 and 3 analyses for a future modified version of
the MEPDG.
1.3.4 Task 4. Identify Applicable Projects.

Task 4 consisted of assessing the usefulness of various field projects, both accelerated
pavement tests and in-service pavement sections for use in validating an endurance limit for
HMA. Results from accelerated pavement tests can be used to tests critical elements of the
framework developed in Task 3. These include the effects of temperature, applied strain, and
material properties on the allowable strain levels. The accelerated pavement tests recommended
for consideration were:
• Fatigue tests conducted during the Superpave validation study at the FHWA
Pavement Test Facility (11).
• Sections at the NCAT Test Track that have remained in service from the first cycle
through the current cycle (12).
• Sections from the WesTrack experiment containing mixtures with different
composition (13).
• Sections from the structural design experiment performed at the NCAT Test Track
(14, 15).
• Selected sections from the MNRoad project (16).

The allowable strain limit design procedure will be calibrated and validated using in-service
pavement sections. It is important to recognize that the allowable strain limit design procedure is
not intended to be a tool for predicting the extent of bottom initiated cracking with time and
traffic like the MEPDG fatigue model. Its purpose is to identify design features that minimize
the possibility of bottom initiated fatigue cracking. Thus, field calibration of the allowable strain
limit design procedure will be easier and likely more precise than the calibration that was
completed for the MEPDG fatigue model.
Sections from the LTPP program (17) and pavements that have received perpetual pavement
awards from the Asphalt Pavement Alliance (18) were considered for use in the calibration and
validation. The LTPP sections were selected because these sections have received extensive
monitoring over a number of years, and distress, deflection, and material property data are
available from the LTPP database (17). Since sufficient sections for the analysis are available
from the LTPP program, only these sections were included in the research plan. Table 4 presents
the test matrix for using LTPP sections to calibrate and validate the allowable strain limit design
procedure. Since the procedure is not intended for prediction of the extent of cracking in a
pavement section, but rather as a tool to identify design features to minimize the potential for
bottom initiated fatigue cracking, an extremely large data set is not required. The recommended
matrix includes a total of 32 pavement sections: 16 not exhibiting alligator cracking and 16
exhibiting low to moderate amounts of alligator cracking. An equal number of sections from the
four environmental zones are included in the matrix. Only pavements with HMA thicknesses
exceeding 8 inches are included. Subgrade deformation becomes an important consideration in
thinner HMA pavements. Simultaneous calibration and validation can be performed on this data
set using jackknifing as described in Research Results Digest Number 283 (19). Jackknifing
allows the assessment of model accuracy without separating the 32 sections into calibration and
validation subsets. Jackknifing is performed by systematically removing one of the sections,
calibrating the model using the remaining sections, then predicting the value of the section that
was removed. For the section that was removed, the model error is computed as the difference
between the predicted and measured values. The process of withholding, calibrating, and
determining the error is repeated until each section has been removed. This process produces n
values of the error from which the jackknifing goodness of fit statistics can be computed. The
10

advantage of jackknifing is the goodness of fit statistics are based on predictions of

measurements that are not included in the calibration. They are, therefore, better estimates of the
accuracy of future predictions than goodness of fit statistics based on calibration using the full
data set.
Table 4. Matrix for Field Calibration of the Allowable Strain Limit Design Procedure.
No Low
HMA Thickness,
Environment Alligator Alligator
in
Cracking Cracking
8 to 12 2 2
Wet Freeze
>12 2 2
8 to 12 2 2
Wet No Freeze
>12 2 2
8 to 12 2 2
Dry Freeze
>12 2 2
8 to 12 2 2
Dry No Freeze
>12 2 2
1.3.5 Task 5. Prepare Detailed Work Plan.

The HMA Endurance Limit Validation Study Research Plan was prepared in Task 5. The
research plan is a comprehensive document describing the research that must be completed to
successfully incorporate the concept of an endurance limit for HMA into a fatigue algorithm for
bottom initiated fatigue cracking and to validate the resulting procedure using full-scale
pavement sections. It includes four major parts. The first is a summary that briefly describes the
proposed research and presents overall cost estimates and time requirements. The second part is
a description of the required research tasks. This section includes detailed information for each
task and subtask, including (1) a description of the work to be performed, (2) preliminary
experimental designs when appropriate, (3) a list of milestones related to the task, (4) labor hour
estimates, and (5) a listing of pertinent data and reference material that will be needed to
accomplish the task. The third part is a detailed schedule for the project. The schedule addresses
the sequence of the research tasks and the interactions between tasks. Finally, the fourth part
presents the proposed budget for the project. The budget includes detailed estimates of labor and
other costs associated with each task and subtask.
11

1.3.6 Task 6. Prepare Final Report.

The final project task was the preparation of this final report documenting the work
performed in NCHRP Project 9-44. The report was prepared in the format required by NCHRP.
It includes the workshop summary and research plan as stand-alone appendices.
12

2. FINDINGS
2.1 HMA Endurance Limit

Recent laboratory and field studies of HMA fatigue behavior indicate that HMA does exhibit
endurance limit behavior. Although early HMA fatigue research conducted by Monismith and
his colleagues suggested that HMA exhibited an endurance limit at approximately 70 μstrain
(20), only limited HMA fatigue research was conducted at low strain levels until recently when
the Asphalt Pavement Alliance began promoting the concept of perpetual pavement design (21).
A perpetual pavement is an asphalt pavement that provides a very long life without structural
failure and only requires periodic replacement of the surface. A key element of perpetual
pavement design is to eliminate fatigue cracking that initiates at the bottom of the HMA base due
to repeated flexure under traffic loading and to confine distresses to the surface of the pavement,
which can easily be renewed by milling and resurfacing.
In response to increasing interest in perpetual pavements, a substantial amount of laboratory

fatigue testing has recently been performed in the United States in an effort to demonstrate that
HMA does exhibit an endurance limit. Most of this work has been performed at the University
of Illinois (22, 23) and the National Center for Asphalt Technology (NCAT) (24). These studies
provide clear evidence that the fatigue behavior of HMA is much different in low strain level
tests compared to normal strain level tests. Figure 2 shows a consolidated plot of the University
of Illinois fatigue data including low and normal strain level test data. Below approximately 100
μstrain, the fatigue life is significantly longer than estimated from extrapolation of normal strain
level test data. Healing of microdamage has been proposed as the primary reason for the
increased fatigue life at low strain levels (1, 25, 26). For cyclic tests at low strain levels, it
appears that the damage that is caused by loading is offset by healing that occurs during
unloading resulting in essentially infinite fatigue life.
13


Results at Low Strain Levels (22).
Detailed investigation of four heavily trafficked pavements in the United Kingdom support
the perpetual pavement concept and the likelihood of an endurance limit for HMA. This
comprehensive study found no evidence of fatigue damage at the bottom of properly constructed
thick flexible pavements with total HMA thickness ranging from 230 to 350 mm (27). Cracks in
these pavements were found to have initiated at the surface and deflections monitored over a
number of years generally showed steady or decreasing deflection with increasing cumulative
traffic, indicating that fatigue damage to the bottom of the HMA was not occurring. Similar
conclusions concerning the absence of cracking at the bottom of thick HMA pavements have
been reported by others (28, 29, 30).
In summary, there is mounting evidence that an endurance limit for HMA does exist. It has
been observed in laboratory studies of fatigue at low strain levels, and several documented case
studies indicate that bottom initiated fatigue cracking is almost non-existent in properly
constructed, thick HMA pavements. The HMA endurance limit, however, does not reflect an
absence of load induced damage in the HMA. It is the result of a balance of damage caused by
14

loading and healing or damage recovery that occurs during rest periods. The endurance limit for
HMA is, therefore, not a single value, but will change depending on the loading and
environmental conditions applied to the HMA. To properly consider this form of an endurance
limit in flexible pavement design requires consideration of the effects of loading, environment
and material properties on both damage accumulation and healing. These findings concerning
the endurance limit for HMA served as the research hypothesis upon which the HMA Endurance
Limit Validation Study Research Plan was formulated.
2.2 Framework for Incorporating Endurance Limit Behavior in Flexible

Pavement Design
The framework for a pavement design procedure that incorporates healing and endurance
limit behavior was conceived during NCHRP 9-44. The procedure is based on layered elastic
analysis and is compatible with the MEPDG. It uses allowable strains to identify satisfactory
conditions for full healing. The allowable strains are a function of the properties of the HMA,
the pavement temperature, and the duration of rest periods between traffic loads. The underlying
principal of the design procedure is to make sure that the damage induced by loading remains
2.2.1 Effect of Rest Periods

Carpenter and Shen (25) clearly demonstrated the beneficial effects of rest periods on the
fatigue life of HMA. Strain controlled flexural fatigue tests were conducted at 20 °C using a 10
Hz haversine load pulse with a rest period between each pulse to simulate the time between
traffic loads. The rest periods ranged from 0 sec (continuous loading) to 9 seconds. Two 19 mm
mixtures, one with a neat PG 64-22 binder and one with a polymer modified PG 70-22 binder,
were tested. The gradation, binder content and air void content of the two mixtures was the
same. The resulting data were analyzed using the ratio of dissipated energy change (RDEC)
approach developed at the University of Illinois (23). In this approach, the ratio of dissipated
energy change reaches a plateau value (PV) where a constant percentage of the input energy is
15

being converted to damage. The University of Illinois research found a unique relationship
between the plateau value and the traditional definition of failure in flexural fatigue tests, 50
percent stiffness reduction, that holds for a range of mixtures and loading conditions (23).
PV = 0.4429 × N f ( )−1.1102 (1)
where:
PV = plateau value
Nf = number of cycles to 50 percent stiffness reduction
Lower plateau values correspond to longer fatigue lives. Based on the ratio of dissipated energy
change approach, an HMA mixture will exhibit endurance limit behavior when the plateau value
is 6.74x10-9 or less, which based on Equation 1 corresponds to a traditional fatigue life of
1.1x107 cycles or greater.
The effect of rest periods on the plateau value is shown in Figure 3 for the two mixtures that
were tested. Equations 2 and 3 present the relationship between plateau value and the length of
the rest period that were developed for the neat PG 64-22 and the modified PG 70-22 mixtures,
respectively for a strain level of 500 μstrain (25).
PV = 1.018 × 10 −5 (RP + 1)
−0.9069
For neat PG 64-22 (2)
PV = 4.353 ×10 −6 (RP + 1)

−1.352
For modified PG 70-22 (3)
where:
PV = plateau value
RP = duration of intermittent rest period, sec
16

a. Neat PG 64-22
b. Polymer PG 70-22
Figure 3. Effect of Rest Periods on Plateau Value (25).
The decreasing plateau values for tests with rest periods result in increasing fatigue lives.
This can be quantified by substituting plateau values from Equations 2 or 3 into Equation 1. The
results are summarized in Table 5. Figure 4 shows the beneficial effect of the rest periods on the
fatigue lives for the two mixtures. There is a substantial improvement in the fatigue life of both
17

mixtures. The values for the neat PG 64-22 mixture are of similar magnitude to improvements
previously reported by Bonnaure, et al. (31). The effect of rest periods on the modified PG 70-
22 mixture is much more pronounced.
Table 5. Effect of Rest Period on Fatigue Life.
Rest Period, Neat PG 64-22 Modified PG 70-22

sec PV Nf Ratio PV Nf Ratio
0 1.02E-05 1.51E+04 1.00 4.35E-06 3.24E+04 1.00
1 5.43E-06 2.65E+04 1.76 1.71E-06 7.53E+04 2.33
2 3.76E-06 3.70E+04 2.45 9.86E-07 1.23E+05 3.81
3 2.90E-06 4.68E+04 3.10 6.68E-07 1.75E+05 5.41
4 2.37E-06 5.61E+04 3.72 4.94E-07 2.30E+05 7.10
5 2.00E-06 6.51E+04 4.32 3.86E-07 2.87E+05 8.86
6 1.74E-06 7.39E+04 4.90 3.13E-07 3.46E+05 10.69
7 1.54E-06 8.24E+04 5.47 2.62E-07 4.08E+05 12.58
8 1.39E-06 9.07E+04 6.02 2.23E-07 4.70E+05 14.52
9 1.26E-06 9.89E+04 6.56 1.94E-07 5.35E+05 16.51
10 1.16E-06 1.07E+05 7.09 1.70E-07 6.01E+05 18.54
20
18
16
14
Relative Fatigue Life
12
NEAT PG 64-22
Polymer PG 70-22
10
0
0 2 4 6 8 10 12
Duration of Rest Period, sec
Figure 4. Effect of Rest Period on Fatigue Life.
18

An estimate of approximate rest periods can be obtained from the 20 year design traffic level
typically used in mixture design. Table 6 summarizes rest periods for various design traffic
levels. The rest period for a 20 year design traffic level of 100 million ESAL is approximately 6
sec., which results in a factor of 5 improvement in the fatigue life of the mixture with the neat
PG 64-22 binder and a factor of 10 improvement for the polymer modified PG 70-22 mixture.
Table 6. Approximate Rest Periods for Various Design Traffic Levels.
20 Year Rest Period,

Design ESAL ESAL/Day ESAL/sec sec
1.00E+05 13.7 0.0002 6307.2
3.00E+05 41.1 0.0005 2102.4
1.00E+06 137.0 0.0016 630.7
3.00E+06 411.0 0.0048 210.2
1.00E+07 1369.9 0.0159 63.1
3.00E+07 4109.6 0.0476 21.0
1.00E+08 13698.6 0.1585 6.3
3.00E+08 41095.9 0.4756 2.1
2.2.2 Allowable Strains

Continuous loading tests at different strain levels were also conducted by Carpenter and Shen
on the two mixtures and the plateau values are shown in Figure 3 for a rest period of zero
(RP+1=1) (25). From these data relationships between the plateau value for continuous loading
and the applied strain level can be developed as shown in Figure 5. These relationships are given
in Equations 4 and 5 for the neat PG 64-22 mixture and the polymer modified PG 70-22 mixture.
For neat PG 64-22 PV0 = 9.142 ×10 −16 (ε )3.617 (4)
For modified PG 70-22 PV0 = 5.347 ×10 −21 (ε )5.331 (5)
where:
PV0 = plateau value for continuous loading
ε = tensile strain, μstrain
19

1.00E-03
Neat PG 64-22
Polymer PG 70-22
1.00E-04
1.00E-05
3.617E+00
Plateau Value
y = 9.142E-16x
R2 = 9.689E-01
1.00E-06
y = 5.347E-21x5.331E+00
2
R = 9.640E-01
1.00E-07
1.00E-08
1.00E-09
10 100 1000 10000
Strain, μstrain
Figure 5. Plateau Value for Continuous Loading as a Function of Applied Strain Level.
Based on previous work by Bonnaure (31), it is reasonable to assume that the effect of the rest
periods is the same at each strain level. Substituting Equations 4 and 5 for the constants 1.018
x10-5 and 4.353 x10-6 in Equations 2 and 3 respectively, yield the following relationships
between the plateau value, applied strain and rest period for the two mixtures.
PV = 9.142 ×10 −16 (ε ) (RP + 1)−0.9069

3.617
PV = 5.347 ×10 −21 (ε ) (RP + 1)−1.352

5.331
where:
PV = plateau value
20

Equations 6 and 7 can then be substituted into Equation 1 and solved for the allowable strain
level to produce a selected mixture fatigue life.
⎡ (1 + RP )0.2507 ⎤
ε a = 11483.32⎢ ⎥
( )
For neat PG 64-22 0.3069
(8)
⎢⎣ N f ⎥⎦
⎡ (1 + RP )0.2536 ⎤
ε a = 5448.74 ⎢ ⎥
( )
For modified PG 70 -22 0.2082
(9)
⎢⎣ N f ⎥⎦
where:
εa = allowable tensile strain, μstrain
Nf = number of cycles to failure
Recalling that endurance limit behavior occurs when the number of cycles to failure exceeds
1.1 x107, then setting the number of cycles to failure in Equations 8 and 9 to a value above
1.1 x 107 will ensure that full healing occurs at the selected rest period. Conservatively using 2.0
x 107 as the number of cycles to failure yields Equations 10 and 11, which give allowable strain
levels as a function of rest period to ensure that full healing occurs.
For neat PG 64-22 ε af = 66.0(1 + RP )0.2507 (10)
For modified PG 70-22 ε af = 164.5(1 + RP )0.2536 (11)
where:
εaf = allowable tensile strain for full healing, μstrain
If the strains in a pavement at 20 °C are kept below the values given by Equations 10 and 11,
then complete healing will occur during intermittent rest periods, and the pavement will exhibit
endurance limit behavior. Table 7 summarizes these strain levels for various 20 year design
traffic levels.
21

Table 7. Allowable Strains for Various Design Traffic Levels.
20 Year Rest Allowable Strains, μstrain

Design ESAL Period, sec Neat PG 64-22 Modified PG 70-22
1.00E+05 6307.2 592 1513
3.00E+05 2102.4 449 1145
1.00E+06 630.7 332 844
3.00E+06 210.2 253 639
1.00E+07 63.1 187 472
3.00E+07 21.0 143 360
1.00E+08 6.3 109 272
3.00E+08 2.1 88 219
2.2.3 Multiple Temperatures

The allowable strains presented in the previous section were developed from test data
obtained at 20 °C. To be useful in a pavement design procedure, the allowable strains for a wide
range of temperatures must be available. In this procedure the major concern is the effect of
temperature on the healing properties of the mixture. Previous research by Bonnaure, et al. (31)
concluded that the beneficial effect of rest periods increased with increasing temperature. Since
healing can be envisioned as a type of flow phenomenon where the binder flows together to
repair microcracks, it has been hypothesized that the effect of healing at multiple temperatures
can be accounted for using time-temperature superposition. By applying time-temperature
superposition, rest periods at different temperatures can be reduced to an equivalent rest period at
20 °C. The reduced rest period for temperatures above 20 °C will be longer than the actual rest
period, while those for temperatures below 20 °C will be shorter than the actual rest period.
Research conducted in NCHRP Project 9-19 showed that linear, viscoelastic time-temperature
shift factors obtained from dynamic modulus tests could be applied when a high level of
nonlinear damage is present (32). Equation 12 presents the application of time-temperature
superposition to the duration of the rest period.
log(RPR ) = log(RP ) − log( AT ) (12)

where:
RPR = duration of the rest period at the reference temperature, sec
RP = actual duration of the rest period, sec
22

AT = linear viscoelastic time temperature shift factor obtained from dynamic

modulus testing.
Figure 6 illustrates the use of time-temperature superposition for rest periods at temperatures of
40, 20, and 4 °C using 20° C as the reference temperature. In developing Figure 6, typical time-
temperature shift factors were used (log(AT) for 4 °C =2.0 and log(AT) for 40 °C = -2.2).
1.0E+06
1.0E+05
1.0E+04
Reduced Rest Period at 20 C, sec
1.0E+03
4C
1.0E+02 20 C
40 C
1.0E+01
1.0E+00
1.0E-01
1.0E-02
1.0E+04 1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10
20 Year Design Traffic, ESAL
Figure 6. Application of Time-Temperature Superposition to Rest Periods.
2.2.4 Design Example

This section illustrates how the proposed methodology would be used in a mechanistic-
empirical design system such as the MEPDG. To limit the number of computations, a monthly
analysis is illustrated using typical pavement temperatures estimated from mean monthly air
temperature data from Reagan National Airport in Washington, DC. The pavement being
analyzed is 9 in of HMA constructed on a 6 in aggregate subbase base layer and a silty clay
subgrade. The 20 year design traffic level is 1 x 108 ESALs, and the design traffic speed is 45
23

mph. The purpose of the analysis is to determine if the pavement section is sufficiently thick to
resist bottom initiated fatigue cracking assuming the fatigue properties of the neat PG 64-22
mixture discussed in the preceding section.
Material Properties
For this analysis the modulus of the subgrade is assumed to be 4,500 psi and constant
throughout the year. The modulus of the aggregate subbase is assumed to be 25,000 psi and is
also assumed constant throughout the year. Dynamic modulus testing of a typical 19 mm
mixture with PG 64-22 binder using the Simple Performance Test System yielded the master
curve and shift factors given in Equations 13 and 14 for a reference temperature of 20 °C. The
allowable strains for full healing are given in Equation 15.
⎡ 3.259 ⎤
log E * = 0.234 + ⎢
⎣ 1+ e( )
−1.213−0.499 log( f r ) ⎥
⎦
(13)
⎛1 1 ⎞
log f r = log f + 10448.2⎜ − ⎟ (14)
⎝ T 293.2 ⎠
ε af = 66.0(1 + RPr )0.2507 (15)
where:
⎪E*⎪ = dynamic modulus, ksi
f = loading frequency, Hz
fr = reduced frequency, Hz
T = temperature, °K
εaf = allowable tensile strain of full healing, μstrain
RPr = reduced rest period at 20 °C, sec
24

Allowable Strains
Allowable strains at the bottom of the asphalt layer are determined from Equation 15 using
reduced rest periods that depend on the traffic volume and the monthly pavement temperature.
Mean monthly pavement temperatures can be estimated from the mean monthly air temperature
using Equation 16 (33).
⎛ 1 ⎞ 34
M p = M a ⎜1 + ⎟− +6 (16)
⎝ z +4⎠ z +4
where:
Mp = mean monthly pavement temperature at depth z, °F
Ma = mean monthly air temperature, °F
z = depth, in
For a 20 year design traffic of 1 x 108 ESAL, the rest period is 6.3 sec. The reduced rest period
for each month is determined from Equation 12 using the shift factors from the dynamic modulus
master curve and the mean monthly pavement temperature. Table 8 summarizes the computation
of the allowable strains. Because the reduced rest period is much shorter during cold months
compared to warm months, the allowable strain levels for full healing are significantly lower.
Table 8. Computation of Allowable Strain Strains.
Mean Monthly Rest Reduced Allowable

Pavement Period, Rest Strain Level,
Month Temp, C Log (AT) sec Period, sec μstrain
Jan 5.5 1.851 6.3 0.09 67
Feb 7.3 1.611 6.3 0.15 68
Mar 12.2 0.971 6.3 0.67 75
Apr 18.0 0.242 6.3 3.61 97
May 23.7 -0.445 6.3 17.56 137
Jun 29.0 -1.065 6.3 73.20 194
Jul 32.0 -1.397 6.3 157.26 235
Aug 30.9 -1.276 6.3 118.95 219
Sep 26.8 -0.803 6.3 40.04 167
Oct 19.7 0.036 6.3 5.79 107
Nov 13.8 0.773 6.3 1.06 79
Dec 8.4 1.469 6.3 0.21 69
25

Applied Strains
The strains applied by the traffic loading are computed for the design axle load using layered
elastic analysis. In this example an 18 kip single axle load was used for computing applied
strains. For this example the modulus of the subgrade and subbase are constant at 4.5 and 25 ksi,
respectively. The modulus of the asphalt depends on the pavement temperature and the speed of
traffic. Recent research by Al-Qadi, et al., using in-situ instrumentation at the Virginia Smart
Road (34) indicates that loading rates computed by the transformed section analysis in the
MEPDG and other approaches such as that recommended by Barksdale (35) overestimate the
frequency of the load pulse. Based on data presented by Al-Qadi, a loading rate of 16 Hz
appears reasonable for a depth of 9 in under 45 mph traffic. Table 10 summarizes the applied
strains for each month computed using the KENLAYER software (33). The applied strains are
compared to the allowable strains in Figure 7. Since the applied strains in Table 9 are less than
the allowable strains, the proposed section is acceptable with respect to bottom initiated fatigue
cracking. An interesting observation in Figure 7 is that this analysis shows that the critical
condition for bottom initiated fatigue cracking occurs at intermediate to low pavement
temperatures, which is in contrast with traditional cumulative or incremental damage analyses,
which show that the majority of the fatigue damage occurs at high pavement temperatures.
Table 9. Applied Strains for Design Example.
Mean
Monthly AC Subbase Subgrade Applied
Pavement Load Reduced Modulus, Modulus Modulus, Strain,
Month Temp, C Log (AT) Freq, Hz Freq, Hz ksi , ksi ksi μstrain
Jan 5.6 1.841 16 1108.93 1969.7 25 4.5 51
Feb 7.5 1.584 16 614.01 1858.0 25 4.5 54
Mar 12.8 0.900 16 127.08 1535.8 25 4.5 62
Apr 19.0 0.122 16 21.21 1148.4 25 4.5 77
May 25.1 -0.608 16 3.95 801.7 25 4.5 100
Jun 30.8 -1.265 16 0.87 535.6 25 4.5 133
Jul 33.9 -1.616 16 0.39 418.2 25 4.5 157
Aug 32.8 -1.488 16 0.52 458.9 25 4.5 148
Sep 28.4 -0.987 16 1.65 641.1 25 4.5 117
Oct 20.8 -0.096 16 12.83 1041.1 25 4.5 83
Nov 14.4 0.688 16 78.05 1431.1 25 4.5 65
Dec 8.7 1.432 16 432.33 1789.1 25 4.5 55
26

2.0
1.9
Ratio of Allowable Strain to Applied Strain
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Pavement Temperature, C
Figure 7. Comparison of Applied and Allowable Strains.
Traffic Level
The analysis presented above can be performed to determine minimum asphalt thicknesses to
resist bottom initiated fatigue cracking for the given subgrade and subbase conditions as a
function of traffic level. The results are shown in Figure 8 for a 22 kip single axle load. A 22
kip axle load was used to allow comparison with observed data from the analysis of in-service
pavements that was conducted in the United Kingdom (27). Figure 8 also shows the thickness
and accumulated traffic for the four pavements that were analyzed in detail and it was
documented that bottom initiated fatigue cracking had not occurred. This comparison shows the
engineering reasonableness of the proposed approach. It is reasonable to expect that when the
proposed approach is improved to consider the effects of aging and design reliability, the
minimum asphalt thicknesses will increase. It is important to note that at the low traffic levels,
deformation of the subgrade may govern the analysis rather than bottom initiated fatigue
27

cracking. Research in the United Kingdom indicates that for asphalt thicknesses less than about
7 in subgrade deformation governs the performance of the pavement (27). This limit is shown as
the dashed line in Figure 8.
16
UK Pavements With No
Bottom Initiated
14 Fatigue Cracking
Minimum Asphalt Thickness, in
12
Minimum Asphalt
10 Thickness for Structural
Deformation
0
1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10
20 Year Design Traffic, 22 kip ESAL

Cracking With Observed Performance of Four UK Pavement Sections.
2.2.5 Aging
The example presented above does not consider the important effect of aging on either the
applied or allowable strains. As a pavement ages, the modulus of the HMA will increase due to
the increased stiffness of the asphalt binder resulting in lower applied strains. Aging will also
affect the healing rate for the HMA. Although no data is currently available for the effect of
aging on the healing rate, it is reasonable to expect that the healing rate will decrease
significantly on aging resulting in lower allowable strains for full healing. Early research on
healing by Bonnaure, et al. (31) showed that healing rates were much greater in softer binders.
The effect of aging can be incorporated in the procedure outlined above, by computing allowable
28

and applied strains as a function of pavement age. The global aging model currently
incorporated in the MEPDG provides a method for computing aged modulus values (36).
Additional research will be required to develop a model of the effect of aging on HMA healing
and the allowable strains that result in full healing. For perpetual pavement design, it may only
be necessary to perform the analysis for highly aged conditions.
2.2.6 Climate and Mixed Traffic Effects

The MEPDG currently provides excellent capabilities to evaluate the effects of climate and
mixed traffic on the applied strains at the bottom of the asphalt layer. This capability can be
used with the allowable strains described above to determine the HMA thickness needed to resist
bottom initiated fatigue cracking.
2.2.7 Reliability
Because the computations involved in the analysis do not require substantial computer time,
reliability can be included in the analysis using Monte-Carlo simulation. This approach has
already been implemented in the PerRoad program (2). In fact, the allowable strains computed
based on rest periods can be input as the threshold criteria for HMA the in the PerRoad program
and the analysis for a single season can be performed.
29

3. INTERPRETATION, APPRAISAL AND APPLICATIONS
The primary product of NCHRP Project 9-44 is the HMA Endurance Limit Validation Study
Research Plan. This section presents a summary of the research plan. The complete plan is
presented in Appendix B.
The HMA Endurance Limit Validation Study consists of five major tasks: (1) Management
and Reporting, (2) Formulate Design Procedure, (3) Database Management, (4) Laboratory
Studies, and (5) Analysis of Pavement Sections. Figure 9 presents an overall flow chart for the
project with major interactions between tasks identified. Table 10 lists the subtasks for each of
the five major tasks and presents estimated labor hours and costs. The HMA Endurance Limit
Validation Study is estimated to require approximately 12,923 man-hours of effort at a cost of
approximately $1.5 million. Figure 10 presents the overall schedule for the project, which is
estimated to require 48 months to complete.
Task 1
Management and
Reporting
Task 4 Task 2
Laboratory Formulate Design
Studies Procedure
Task 3 Task 5
Database Analysis of
Management Pavement Sections
Figure 9. Project Flow Chart.
30

Table 10. Summary of Man-hour and Cost Estimates.
Estimated Labor Hours

Senior
Task/ Eng./ Eng./ Estimated
Subtask Description Stat Prog. Tech. Admin. Cost
1.0 Management and Reporting
1.1 Project Management 424 0 0 40 $66,000
1.2 Progress Reporting 210 0 0 20 $32,700
1.2 Interim Reports and Presentations 780 0 0 80 $129,780
1.3 Final Report and Presentation 420 0 0 40 $68,400
Task 1 Total 1834 0 0 180 $296,880
2.0 Formulate Design Procedure
2.1 Review Selected Literature 240 160 0 0 $52,000
2.2 Finalize Preliminary Approach 80 160 0 0 $28,000
2.3 Incorporate Findings from Laboratory Studies 80 160 0 0 $28,000
2.4 Modify Approach Based on Analysis of 80 80 0 0 $20,000
Accelerated Pavement Tests
2.5 Prepare Final Design Procedure 120 80 0 0 $26,000
Task 2 Total 600 640 0 0 $154,000
3.0 Database Management
3.1 Develop Plan to Use NCHRP 9-30 Database 120 0 0 0 $18,000
3.2 Develop Needed Tables 80 240 0 0 $36,000
3.3 Input and Manage Data 40 396 0 0 $45,600
Task 3 Total 240 636 0 0 $99,600
4.0 Laboratory Studies
4.1 Experiment 1: Mixture Compositional Factors 42 0 388 0 $39,280
Affecting Healing
4.2 Experiment 2: Effect of Applied Strain on 32 0 214 0 $22,990
Healing
4.3 Experiment 3: Effect of Temperature and Rest 69 0 242 0 $30,920
Period Duration on Healing
4.4 Experiment 4: Testing and Analysis 168 0 392 0 $58,520
Procedures for Allowable Strain Levels
4.5 Experiment 5: Estimation of Allowable Strain 456 0 1890 0 $229,050
Levels from Mixture Composition
Task 4 Total 767 0 3126 0 $380,760
5.0 Analysis of Pavement Sections
5.1 Review Data Sources and Select Sections for 52 320 0 0 $39,800
Analysis
5.2 Obtain Materials and Data for Accelerated 48 280 0 0 $35,200
Pavement Tests
5.3 Perform Testing and Analyze Accelerated 164 512 32 0 $78,520
Pavement Tests
5.4 Obtain Materials and Data for In-Service 120 1280 0 0 $195,600
Pavement Sections
5.5 Perform Testing and Analyze In-Service 300 512 1280 0 $205,000
Pavement Sections
Task 5 Total 684 2904 1312 0 $554,120
Project Total 4,125 4,180 4,438 180 1,485,360
31

Figure 10. Overall Project Schedule.
32
Task 1, Management and Reporting, includes all activities normally associated with
management and reporting for NCHRP Projects. Major management tasks include scheduling,
coordinating, and directing various technical work activities as well as project financial
management. Reporting activities include monthly and quarterly progress reports, the
preparation of several interim reports and presentations, and the preparation of the final report.
Interim reports are required at approximately 6 month intervals and coincide with the completion
of five critical milestones:
(1) Formulation of the preliminary design procedure and selection of the laboratory analysis
approach,
(2) Selection of pavement sections for analysis,
(3) Completion of the laboratory studies,
(4) Modification of the preliminary design procedure to reflect the findings from the
laboratory studies and the analysis of accelerated pavement tests, and
(5) Analysis of the calibration sections and preparation of the final design procedure.
The final report will document the entire study and will be prepared from the interim reports.
Task 2, Formulate Design Procedure, is a critical project task that will be active throughout
the project. This task includes finalizing the preliminary approach that was presented in Chapter
2, modifying the preliminary approach based on the results of the laboratory studies and selected
accelerated pavement tests, and preparation of the final design procedure after analysis of the
calibration pavement sections. It is important to emphasize that the preliminary approach
prepared early in this task will shape the laboratory studies and guide the selection of pavement
sections, both accelerated pavement tests and in-service pavement sections.
Task 3, Database Management, is a support task that will be active throughout the project. A
database will be developed in this task to store and analyze data from the laboratory studies and
the analysis of the pavement sections. It is envisioned that the database will be an adaptation of
the one developed in NCHRP Project 9-30.
33

Task 4, Laboratory Studies, includes the planning and execution of five laboratory studies that
are needed to complete the design procedure that will be formulated in Task 2. The laboratory
studies concentrate on quantifying what affects the healing properties of HMA. The laboratory
studies will be sufficient in breadth to develop models relating mixture and binder properties to
the key engineering properties required for the analysis.
Task 5, Analysis of Pavement Sections, includes several activities associated with the
selection and analysis of full-scale pavements. The preliminary design procedure formulated in
Task 2 will be tested using data from completed accelerated pavement tests, such as the fatigue
studies from the Federal Highway Administration’s (FHWA’s) Pavement Testing Facility or the
structural sections included in the National Center for Asphalt Technology (NCAT) test track.
Calibration of the design procedure will be accomplished through an analysis of in-service
pavements where it has been documented that bottom-up fatigue cracking has occurred or has
not occurred. These analyses will serve to calibrate the design procedure and validate the HMA
endurance limit concept. The predictive models developed in Task 4 will be used in the analysis
of the full-scale pavement sections. This will allow consideration of pavement sections where
original materials are not available since the required data can be obtained from cores taken from
the pavement section.
34

4. CONCLUSIONS AND RECOMMENDATIONS
Completed laboratory and field investigations clearly show that HMA exhibits endurance
limit behavior. This endurance limit, however, does not reflect an absence of load induced
damage in the HMA. It is the result of a balance of damage caused by loading and healing or
damage recovery that occurs during rest periods. The endurance limit for HMA is, therefore, not
a single value, but will change depending on the loading and environmental conditions applied to
the HMA. To properly consider this form of an endurance limit in flexible pavement design
requires consideration of the effects of loading, environment and material properties on both
damage accumulation and healing.
The framework for a pavement design procedure that incorporates healing and endurance
limit behavior was conceived during NCHRP 9-44. The procedure is based on layered elastic
analysis and is compatible with the MEPDG. It uses allowable strains to identify satisfactory
conditions for full healing. The allowable strains are a function of the properties of the HMA,
the pavement temperature, and the duration of rest periods between traffic loads. The underlying
principal of the design procedure is to make sure that the damage induced by loading remains
Additional laboratory and field studies are needed to fully develop and calibrate the design
procedure. The HMA Endurance Limit Validation Study Research Plan prepared in NCHRP
Project 9-44 presents a comprehensive work plan and cost estimate for the needed research. The
research plan includes laboratory experiments to fully develop the new design procedure.
Studies using data from completed accelerated pavement tests and test roads are proposed to
verify critical aspects of the design procedure. Finally, an experiment to calibrate and validate
the new design procedure using selected test sections from the Long Term Pavement
Performance Program is presented.
35

REFERENCES
1. Advanced Asphalt Technologies, LLC, “Hot Mix Asphalt Endurance Limit Workshop:
Executive Summary,” National Cooperative Highway Research Program Project 9-44,
November, 2007.
2. Timm, D., H., and Young, J., B., “Effects of Load Spectra and Variability on Perpetual
Pavement Design,” Proceedings, International Symposium of Design and Construction
of Long Lasting Asphalt Pavements, National Center for Asphalt Technology, Auburn,
AL, 2004.
3. Witczak, M.W., “Use of HMA Fatigue Endurance Limits in the Mechanistic-Empirical

Pavement Design Guide,” Presentation at the Hot Mix Asphalt Endurance Limit Workshop,
August 1, 2007.
4. Thompson, M.R., ad Carpenter, S.H., “Design Principles for Long Lasting Pavement,”
Proceedings, International Symposium of Design and Construction of Long Lasting
Asphalt Pavements, National Center for Asphalt Technology, Auburn, AL, 2004.
5. Sidess, A. and Uzan, J., “A Design Method for Perpetual Flexible Pavement in Israel,”
International Journal of Pavement Engineering, February, 2008.
6. Von Quintus, H. L., “Hot-Mix Asphalt Layer Thickness Design for Longer Life Bituminous
Pavements,” Transportation Research Circular Number 503, Perpetual Bituminous
Pavements, Transportation Research Board, Washington, D.C., December, 2001.
7. Von Quintus, Harold L., “Application of the Endurance Limit in Mechanistic-Empirical

Based Pavement Design Procedures,” paper prepared for presentation and publication at the
International Conference on Perpetual Pavements, Columbus, Ohio, September 13-15, 2006.
8. Tsai, B.W., Harvey, J., Monismith, C., and Bejarano, M., “Calibration of Fatigue Surface
Cracking Using Simplified Recursive Miner’s Law,” Journal of the Association of Asphalt
Paving Technologists, Vol., 76, 2007.
9. Mun, S., Guddati, M. N., and Kim, Y. R., “Fatigue Cracking Mechanisms in Asphalt
Pavements with Viscoelastic Continuum Damage Finite-Element Program,” Transportation
Research Record No. 1896, Transportation Research Board, Washington, D.C., 2004.
10. Roque, R., B. Birgisson, B. Sangpetngam and Z. Zhang, “Hot Mix Asphalt Fracture
Mechanics: A Fundamental Crack Growth Law for Asphalt Mixtures,” Journal of the
Association of Asphalt Paving Technologists, Vol. 71, 2002.
11. Stuart, K.D., Mogawer, W.S., and Romero, P., “Validation of the Superpave Asphalt Binder
Fatigue Cracking Parameter Using an Accelerated Loading Facility,” Report Number
FHWA-RD-01-093, Federal Highway Administration, Washington, D.C., 2002.
12. National Center for Asphalt Technology. http://www.pavetrack.com/ (accessed June, 2008).
36

13. Epps, J. A., Hand, A., Seeds, S., Schultz, T., Alavi, S., Ashmore, C., Monismith, C., Deacon,
J.A., Harvey, J.T., and Leahy, R., “Recommended Performance Related Specifications for
Hot-Mix Asphalt Construction: Results of the WesTrack Project,” NCHRP Report 455,
National Cooperative Highway Research Program, Washington, D.C., 2002.
14. Timm, D. West, R., Priest, A., Powell, B., Selvaraj, I., Zhang, J., and Brwon, R., “Phase II
NCAT Test Track Results,” NCAT Report 06-05, National Center for Asphalt Technology,
Auburn University, Auburn, AL, December, 2006.
15. Timm, D. and Priest, A., “Material Properties of the 2003 NCAT Test Track Structural
Study,” NCAT Report 06-01, National Center for Asphalt Technology, Auburn University,
Auburn, AL, April, 2006.
16. Minnesota Department of Transportation. http://www.dot.state.mn.us/mnroad/ (accessed

June, 2008).
17. Federal Highway Administration. http://www.ltpp-products.com/DataPave/index.asp

(accessed June, 2008).
18. Asphalt Pavement Alliance. http://www.asphaltalliance.com/ (accessed June, 2008).
19. VonQuintus, H.L, Schwartz, C.E., McCuen, R.H., and Andrei, D., “Jackknife Testing – An
Experimental Approach to Refine Model Calibration and Validation,” NCHRP Research
Results Digest Number 283, National Cooperative Highway Research Program,
Washington, D.C, December, 2003.
20. Monismith, C.L., and McLean, D.B., “Structural Design Considerations,” Proceedings of
the Association of Asphalt Paving Technologists, Vol. 41, 1972.
21. Newcomb, D.E., Buncher, M., and Huddleston, I.J., “Concepts of Perpetual Pavements,”
Transportation Research Circular Number 503, Perpetual Bituminous Pavements,
Transportation Research Board, Washington, D.C., December, 2001.
22. Carpenter, S.H., Ghuzlan, K.A., and Shen, S., “Fatigue Endurance Limit for Highway and
Airport Pavements,” Transportation Research Record No. 1832, Transportation Research
Board, Washington, D.C., 2003.
23. Carpenter, S.H., and Shen, S., “Application of the Dissipated Energy Concept in Fatigue
Endurance Limit Testing,” Transportation Research Record No. 1929, Transportation
Research Board, Washington, D.C., 2005.
24. Prowell, B., Brown, E., R., Daniel, J., Bhattacharjee, S., Von Quintus, H., Carpenter, S.,
Shen, S., Anderson, M., Swamy, A. K., and Maghsoodloo, S., “Endurance Limit of Hot Mix
Asphalt Mixtures to Prevent Fatigue Cracking in Flexible Pavements,“ Updated Draft Final
37

Report, NCHRP 9-38, National Cooperative Highway Research Program, Washington, D.C.,
May, 2008.
25. Carpenter, S.H., and Shen, S., “Dissipated Energy Approach to Study Hot-Mix Asphalt
Healing in Fatigue,” Transportation Research Record No. 1970, Transportation Research
26. Kim, B. and Roque, R., “Evaluation of Healing Property of Asphalt Mixtures,”
Transportation Research Record No. 1970, Transportation Research Board, Washington,
D.C., 2006.
27. Nunn, M. and Ferne, B.W., “Design and Assessment of Long-Life Flexible Pavements,”
28. Brown, S.R., Thom, N.H., and Hakim, B.A., “Performance and Rehabilitation of Heavy-duty
Pavements in the UK: Some Case Studies,” Proceedings, International Symposium of
Design and Construction of Long Lasting Asphalt Pavements, National Center for
Asphalt Technology, Auburn, AL, 2004.
29. Wu, Z., Siddiqui, Z.Q., Hossain, M., and Gisi, A.J., “Kansas Turnpike – An Example of
Long Lasting Asphalt Pavement,” Proceedings, International Symposium of Design and
Construction of Long Lasting Asphalt Pavements, National Center for Asphalt
Technology, Auburn, AL, 2004.
30. Uhlmeyer, J.S., Willoughby, K., Pierce, L.M., and Mahoney, J.P., “Top-Down Cracking in
Washington State Asphalt Concrete Wearing Courses,” Transportation Research Record
No. 1730, Transportation Research Board, Washington, D.C., 2000.
31. Bonnaure, F.P, Huibers, A.H.J.J., Boonders, A., “A Laboratory Investigation of the Influence
of Rest Periods on the Fatigue Response of Bituminous Mixes,” Proceedings, Association of
Asphalt Paving Technologists, Vol. 51, 1982.
32. Chehab, G.R., Kim, Y.R., Schapery, R.A., Witczak, M.W., and Bonaquist, R.,
"Characterization of Asphalt Concrete in Uniaxial Tension Using a Viscoelastoplastic
Continuum Damage Model," Journal of the Association of Asphalt Paving Technologists,
Vol. 72, 2003.
33. Huang, Y. H., Pavement Analysis and Design, Englewood Cliffs, N.J.: Prentice-Hall, Inc.,
1993.
34. Al-Qadi, I., Xie, W., and Elseifi, M., “Frequency Determination from Vehicular Loading
Time Pulse to Predict Appropriate Complex Modulus in MEPDG," Paper P:reprint CD,
Association of Asphalt Paving Annual Meeting and Technical Sessions, Philadelphia, PA,
April 27-30, 2008.
38

35. Barksdale, R.G., “Compressive Stress Pulse Times in Flexible Pavements for Use in
Dynamic Testing,” Highway Research Record 345, Highway Research Board, Washington,
D.C., 1971.
36. Mirza, M. W., and M. W. Witczak, “Development of a Global Aging System for Short and
Long Term Aging of Asphalt Cements,” Journal of the Association of Asphalt Paving
Technologists, Vol. 64, 1995.
39

APPENDIX A. HMA ENDURANCE LIMIT WORKSHOP EXECUTIVE

SUMMARY
Table of Contents
List of Tables ..............................................................................................................................A-ii
List of Figures .............................................................................................................................A-ii
Introduction................................................................................................................................. A-1
Participants.................................................................................................................................. A-1
Workshop Overview ................................................................................................................... A-2
Endurance Limit and Other Important Fatigue Effects............................................................... A-6
Methodologies for HMA Fatigue Characterization .................................................................. A-11
Flexible Pavement Damage Analysis ....................................................................................... A-14
Verification and Calibration ..................................................................................................... A-18
Recommendations to the Research Team ................................................................................. A-20
Summary and Conclusion ......................................................................................................... A-24
Attachment. Participant Comments from the Opening and Closing Sessions......................... A-26
A-i

List of Tables
Table 1. HMA Endurance Limit Workshop Invitees and Attendees. ........................................ A-2
Table 2. Issues Identified During Open Brainstorming Discussion. ......................................... A-7
Table 3. Summary of Breakout Group Results.......................................................................... A-8
Table 4. Ideas From the Fatigue Testing Brainstorming Session. ........................................... A-12
Table 5. Barriers and Countermeasures Identified for Incorporating Endurance Limit
in Flexible Pavement Damage Analysis. .................................................................. A-17
Table 6. Ideas From the Verification/Calibration Brainstorming Session............................... A-19
Table 7. Summary of Workshop Recommendations for the NCHRP 9-44 Research Team. .. A-22
List of Figures
Figure 1. Facilitator’s Agenda. .................................................................................................. A-3
Figure 2. Conceptual Representation of Improved Flexible Pavement Fatigue Model........... A-15
A-ii

Introduction
This report documents the Hot Mix Asphalt (HMA) Endurance Limit Workshop held in
Washington, D.C. on August 1 and 2, 2007. The workshop was sponsored by the National
Cooperative Highway Research Program (NCHRP) as part of NCHRP Project 9-44, Developing
a Plan for Validating an Endurance Limit for HMA Pavements. Participants included members
of the NCHRP Project 9-44 panel and research team, key researchers and consultants with
extensive experience in HMA fatigue analysis, and engineers from highway agencies who are
responsible for designing, constructing, and maintaining flexible pavements. The objective of
the workshop was to discuss several topics relevant to an endurance limit for HMA pavements,
and to provide recommendations for consideration by the research team for the work plan that
will be prepared in NCHRP 9-44.
Participants
Participants in the HMA Endurance Limit Workshop included members of the research team,
the NCHRP Project 9-44 panel, and invited experts that were recommended by the research team
and approved by the project panel. Thirty-four individuals were invited to attend, including four
members of the research team, the NCHRP Project 9-44 panel members and Project Manager,
and 22 invited experts. There was a high degree of interest in the workshop with 82 percent of
the invitees participating. Only three invited experts, two panel members, and one liaison to the
panel were not able to participate due to schedule conflicts. Another invited expert was unable
to attend due to an airline delay. Table 1 presents summary information about the participants.
Detailed information on the invitees and participants is included Sections 2 and 9 of the HMA
Endurance Limit Workshop Notebook.
A-1

Table 1. HMA Endurance Limit Workshop Invitees and Attendees.

No Name Affiliation Role Attend
1 Dr. David Anderson Consultant Panel Member Y
2 Dr. Ramon Bonaquist Advanced Asphalt Technologies, LLC Research Team Y
3 Dr. Stephen Brown University of Nottingham Invited Expert N
4 Dr. William Buttlar University of Illinois Invited Expert N
5 Dr. Samuel Carpenter University of Illinois Invited Expert Y
6 Dr. Donald Christensen Advanced Asphalt Technologies, LLC Research Team Y
7 Mr. Danny Dawood Pennsylvania Department of Transportation Panel Member N
8 Dr. Herve Di Benedetto Ecole Nat. des TPE Invited Expert Y
9 Mr. Bruce Dietrich Florida Department of Transportation Panel Member Y
10 Dr. Jon Epps Granite Construction Company, Inc. Invited Expert N
11 Mr. Kenneth Fults KWF Pavement Consulting Invited Expert Y
12 Mr. Roger Green Ohio Department of Transportation Panel Member Y
13 Dr. Kevin Hall University of Arkansas Invited Expert Y
14 Dr. Edward Harrigan National Cooperative Highway Research Program Panel Member Y
15 Mr. Frederick Hejl Transportation Research Board TRB Liaison N
16 Dr. Richard Kim North Carolina State University Invited Expert Y
17 Dr. Dallas Little Texas A&M University Invited Expert Y
18 Dr. Robert Lytton Texas A&M University Invited Expert N
19 Dr. Leslie Ann McCarthy Federal Highway Administration Panel Member Y
20 Mr. Charles Markert Dynamic Leadership Consulting Group Research Team Y
21 Dr. Andre Molenaar Delft University Invited Expert Y
22 Professor Carl Monismith University of California Berkeley Invited Expert Y
23 Dr. David Newcomb National Asphalt Pavement Association Invited Expert Y
24 Dr. Michael Nunn Lane One Limited Invited Expert Y
25 Ms. Linda Pierce Washington Department of Transportation Invited Expert N
26 Dr. Brian Prowell Advanced Material Services, LLC Invited Expert Y
27 Dr. Rey Roque University of Florida Invited Expert Y
28 Ms. Amy Schutzbach Illinois Department of Transportation Panel Member Y
29 Mr. Darin Tedford Nevada Department of Transportation Panel Member N
30 Dr. Jacob Uzan Technion University Invited Expert Y
31 Mr. Harold Von Quintus Applied Research Associates Invited Expert Y
32 Dr. Linbing Wang Virginia Polytechnic and State University Panel Member Y
33 Ms. Rane Wagner Rane Wagner and Associates Research Team Y
34 Dr. Matthew Witczak Arizona State University Invited Expert Y
Workshop Overview
The HMA Endurance Limit Workshop was facilitated by Mr. Charles Markert, a Certified
Professional Facilitator. The agenda for the workshop was developed jointly by Mr. Markert and
Dr. Ramon Bonaquist, Principal Investigator for NCHRP Project 9-44. A copy of the
Facilitator’s Agenda is reproduced as Figure 1.
A-2

AGENDA Item Questions to be answered

Sponsor Welcome, Facilitator Opening, Introductions
Personal Expectations What are your expectations for this session?
Briefing 1 Fatigue in the MEDPG How are fatigue and the endurance limit addressed in
the MEPDG?
Briefing 2 NCHRP 9-38 Laboratory Evaluation of What was found about the endurance limit in NCHRP
Endurance Limit Project 9-38?
Briefing 3 A Review of UK Pavement Design What approach is taken in the UK?
Purpose Discussion What is the purpose of this session?
Discussion: Existence of Fatigue Endurance Limit Does a Fatigue Endurance- Limit Exist?
Continue Discussion Does a Fatigue Endurance- Limit Exist?
Issues Related to Fatigue -Endurance- Limit “What are the other issues?”
Identify Major Issues Which of these possibilities go on the Short List?
Place your dots – one per card.
What is the Meaning of Each Major Issues? Discussion on the top few.
Discussion: Alternative Methodologies for Characterizing Do we need alternative to Beam Fatigue? Can they
Fatigue address endurance limit? Are they implementable?
Flexible Pavement Damage Analysis How can we improve flexible pavement damage
analysis? What important issues are not currently
addressed?
Plus/Delta
ADJOURN
DAY TWO
Review Agenda/Progress/Issues
Strategies for Incorporating Endurance Limit in Flexible What are the possible strategies for incorporating
Pavement Damage Analysis Endurance Limit?
Identify Barriers to Success? What are the barriers to success in this effort?
Identify Countermeasures? What are some countermeasures?
Identify Simplifying Assumptions Can we identify some simplifying assumptions that
will help?
Calibration/Verification Is calibration necessary?
How should it be done? Field sections, accelerated
pavement testing, etc?
Suggest Data Evaluation Approaches What are your suggestions for calibration/verification?
Identify Potential Action for NCHRP 9-44 What should be included in the work plan for future
research developed in NCHRP 9-44?
Silver Bullet Actions Which suggestions from Potential Actions can be
addressed in a 3-year project?
Recommendations, Findings & Conclusions What are your recommendations, findings &
conclusions as a group?
Closing
ADJOURN
Figure 1. Facilitator’s Agenda.
A-3

The HMA Endurance Limit Workshop started with an opening session where the participants
expressed their personal expectations concerning the workshop. Excerpts from this session are
included in the attachment.
The introductory session was followed by three invited presentations on relevant research to
provide background information to the participants. The first presentation, Use of HMA Fatigue
Endurance Limits in the Mechanistic-Empirical Pavement Design Guide, was made by Dr.
Matthew Witczak. Dr. Witczak was responsible for developing the flexible pavement design
methodology contained in the Mechanistic-Empirical Pavement Design Guide (MEPDG)
developed in NCHRP Project 1-37A. Dr. Witczak’s presentation summarized how fatigue
cracking is addressed in the MEPDG, and showed the effect of incorporating endurance limits of
varying magnitude on fatigue damage predicted with the MEPDG. The second presentation,
Endurance Limit of HMA Mixtures to Prevent Fatigue Cracking in Flexible Pavements: NCHRP
9-38, was made by Dr. Brian Prowell. Dr. Prowell is Co-Principal Investigator for NCHRP
Project 9-38. The objectives of this on-going research project are to: (1) confirm the existence of
an HMA endurance limit through laboratory testing, (2) investigate the effect of HMA material
properties on the endurance limit, (3) develop a shortcut method to determine the endurance
limit, and (4) suggest changes to the MEPDG to include an endurance limit. Dr. Prowell’s
presentation updated the participants on the progress that has been made in NCHRP Project 9-38.
The final presentation, Mechanistic-Empirical Design and Fatigue in the United Kingdom, was
made by Dr. Michael Nunn. Dr. Nunn played a key role in the work that led to maximum
asphalt layer thicknesses being included in the flexible pavement design procedure used in the
United Kingdom. His presentation summarized the rationale behind the approach taken in the
United Kingdom. Copies of the presentations and supporting materials that were provided by the
speakers are included in Sections 5, 6, and 7 of the HMA Endurance Limit Workshop Notebook.
The key element of the HMA Endurance Limit Workshop was a series of discussion sessions
focusing on four major topics considered important in the approach for validating an endurance
limit for HMA pavements that was proposed by the research team. This approach involves the
development of an improved mechanistic-empirical fatigue damage analysis that accounts for the
presence of an endurance limit as well as the effects of healing and changes in damage tolerance
A-4

due to temperature and aging. Elements of this improved damage analysis will be developed
through laboratory studies and will require verification using data from accelerated pavement
tests before applying the approach to the analysis of in-service pavements. To validate the
endurance limit, the improved damage analysis will be applied to several in-service pavements
that have documented evidence of no bottom-up fatigue cracking. More detailed information on
this approach was included in Section 4 of the HMA Endurance Limit Workshop Notebook. The
four major topics relevant to this approach that were discussed during the workshop were:

endurance limit.
The discussion of these topics concluded with a session to identify action items. In this session,
the participants outlined their recommendations for items to be included in the work plan that
will be developed in NCHRP Project 9-44.
A variety of techniques were used in these discussion sessions to develop and prioritize ideas,
focus the discussion, and develop recommendations for the research team. These included
brainstorming ideas on a “sticky wall,” “dot polling” to prioritize ideas, small breakout groups to
further develop ideas, and guided discussion. Separate sections discussing each of these sessions
are presented later in this report.
An unplanned discussion session was added during the workshop to develop a definition of
endurance limit for use in NCHRP Projects 9-38 and 9-44. This session was generated by
discussions during the early stages of the workshop. It produced the following definition:
HMA Fatigue Endurance Limit – A level of strain below which there is no cumulative
damage over an indefinite number of load cycles.
A-5

The HMA Endurance Limit Workshop concluded with final comments from each of the
participants. Excerpts of comments contributed by the participants are included in the
attachment.
Endurance Limit and Other Important Fatigue Effects
Objective and Pre-Workshop Position

The objectives of this session were: (1) to discuss whether an endurance limit exists for
asphalt concrete and (2) to identify important asphalt concrete fatigue effects that might
contribute to long fatigue life. The research team’s pre-workshop position was that there is
mounting evidence that an endurance limit for asphalt concrete does exist. It has been observed
in independent laboratory studies of fatigue at low strain levels, and several documented case
studies of in-service pavements indicate that bottom-up fatigue cracking is almost non-existent in
properly constructed, thick asphalt concrete pavements. The research team identified three
components of fatigue performance: (1) rate of damage accumulation; (2) healing rate; and (3)
damage tolerance. Damage accumulation refers to the degradation of material properties during
repeated loading and is what is measured in a standard fatigue test. Healing refers to the
recovery of damage with time. Finally, damage tolerance pertains to the level of damage that
can be sustained before macro-cracking (failure) occurs. Only the first component with a
constant damage tolerance is considered in current flexible pavement design methods. The
endurance limit is the point where the rate of damage accumulation in a laboratory fatigue test
becomes very low and is an important consideration in the design of long-life pavements. The
other two components, however, are also important. A pavement subjected to a strain level
above the endurance limit may accumulate damage fairly quickly, but may still exhibit no
bottom-up fatigue cracking if the healing rate is high, and/or if the damage tolerance is high. All
three components must be considered to validate the concept of an endurance limit and to
effectively improve the fatigue design of flexible pavements.
Overview of Workshop Activities

Issues related to the existence of an endurance limit for HMA and other important fatigue
effects were gathered through an open brainstorming discussion during which the facilitator
A-6

captured the issues on cards that were later prioritized using a “dot poll” technique. The issues
that were identified are summarized in Table 2 along with the number of votes that each issue
received. The issues receiving the highest number of votes were then further developed using
small breakout groups. Eight breakout groups were formed to address the top 10 issues.
Because of their similarities, Issues 7, 8, and 9 were combined. The small breakout groups were
asked to further develop an issue by identifying key points and suggestions for the NCHRP
Project 9-44 research team. The results from the breakout groups are summarized in Table 3.
Table 2. Issues Identified During Open Brainstorming Discussion.

No. Issue From Open Brainstorming Discussion Votes
1 Look at pavements that have and have not cracked. 15
2 Endurance limit may be an algorithm that includes materials, healing, temperature, 15
etc.
3 Is what we do in the lab related to field mechanisms? 13
4 Consider developing confidence/risk level into this approach. 12
5 Determine how the endurance limit is to be used in design. 12
6 Relate the endurance limit to material type and climatic conditions. 11
7 How do we use endurance limit design for long-life pavement design? 10
8 How is the endurance limit to be used in design? 12
9 How do we design long-life pavements? 10
10 Reconcile field, lab, theory. 9
11 Prove existence of endurance limit by looking in the field. 5
12 How thin can I build a pavement without bottom up cracking? 4
13 How extensive is the level of bottom up cracking? 3
14 Define what we mean by “long lasting” pavement. 3
15 What is the mechanism? 2
16 Is endurance limit for crack initiation or some level of crack propagation? 2
17 Define fatigue cracking and permanent deformation 1
18 Definition: A strain level at which no cracking will occur. 1
19 Consider aging of pavement (ability to heal). 1
20 Define what endurance limit means. 1
21 What is the strain limit at which pavement has a long fatigue life? 1
22 Is the endurance limit a material property? 1
23 Endurance may be a change in the slope of the fatigue curve at low strain levels 1
24 There are ways to design for zero fatigue damage without using endurance limit. 0
25 Distinguish between perpetual life vs. endurance limit design. 0
26 Should top down cracking be considered (tradeoff)? 0
27 Is there another name for endurance limit? 0
28 Clearly identify the phenomenon in the road. 0
A-7

Table 3. Summary of Breakout Group Results.

Issue Look at pavements that have Endurance limit may be an Is what we do in the lab related to Consider developing
cracked and have not cracked. algorithm that includes materials, field mechanisms? confidence/risk level into this
healing, temperature, etc. approach.
Key Points • Used to confirm proposed • Probably two approaches: • Stiffness reduction in lab • Complex interaction of
endurance limit values. • Lab tests. mirrored in field variables:
• Must segregate pavements (1) • Field behavior. measurements? • Design.
uncracked, (2) no structural • Concentrate on lab tests: • Europe – no • Thickness.
cracks, (3) structural cracks, (4) • Visco-elastic material. • LTPP? • Properties of each layer.
where cracking initiated. • Temperature – frequency • NCAT Track- yes 7-5” • Load.
• Ensure sample covers range dependent. • Westrack? • Mix Properties
within the design procedure • Healing. • Aging increases stiffness. • Anisotropy.
that will be used. • Lab testing leaves out • Healing? • Volumetrics.
• Most pavements should be in- importance of crack • Fatigue properties of aged • Binder/Mastic.
service roadways as opposed to propagation. material are probably different • External variables
accelerated pavement sections. than unaged materials. • Temperature.
• Construction defects that may
• Rest periods.
have caused cracking must be
• Traffic speed.
identified – segregation, de-
bonding between layers, etc. • Pressing need to weight design
and mix properties.
• Potential to evaluate strain
response magnitude versus • Based on above, consider
mixture composition and local benefit cost ratio approach
condition for pavements with where the cost is a function of
and without structural cracking. tolerable risk.
Suggestions • Make deflection measurements • Use crack initiation, number of • Fatigue tests on material from • Same as key points.
to define structural related cycles ( Ni ) before localization. roads of different ages.
cracking. Check hysteresis • Number of cycles to failure • Measure pavement stiffness as
loop between cracked and un- (Nf ) equals 50% stiffness a function of time.
cracked sections. reduction. • LTPP database for a range
• Measure mixture composition • Use Ni for design of long of pavement thickness.
between sections. lasting pavement. • Indirect Tensile modulus
• Take cores and trenches to on field cores.
confirm direction of cracking
also to ensure construction
defects do not exist or identify
where construction defects
have influenced results.
A-8
Table 3. Summary of Breakout Group Results (continued).

Issue Determine how the endurance Relate the endurance limit to How do we use endurance limit Reconcile field, lab, theory.
limit is to be used in design. material type and climatic design for long-life pavement
conditions. design?
How is the endurance limit to be
used in design?
How do we design long-life
pavements?
Key Points • Design versus analysis: • Evaluate significance of • Variation of endurance limit • In-service pavements.
• Design - prevent fatigue material type and climate on with mix and binder properties, • Good performance and no
from occurring. pavement thicknesses. temperature, aging, healing bottom up cracks.
• Analysis - predict damage • Develop a catalog of endurance (this is not likely to be a routine • Traffic history – large
e.g. MEPDG. limit values based on binder test). commercial volume.
• Use endurance limit as an type, gradation, void content, • Design approach / philosophy • Range of environments.
HMA material property. asphalt content, etc. (risk, load definition). • Performance vs. time.
• Characterize endurance limit • Evidence both in the lab and • Material property relationships • Only with surface maintenance.
versus using a predetermined field indicate endurance limit is (catalog of properties). • Accelerated pavement tests,
value: related to temperature. • Seasonal variation in material HVS, NCAT, MnRoad, ALF,
• Level 1 testing. • Effect of moisture not clear. properties. WESTRACK
• Level 2 catalog of values • Stress state conditions must be • What is the most appropriate • Original construction materials,
based on mix composition. considered. fatigue relationship (or failure)? design.
• Level 3 single default
value.
Suggestions • Recognize need for different • Further lab testing. • From a practical point of view: • Volumetric and binder
levels in material testing. • Use field instrumentation to • Pick very long design life characteristics in-service.
• Recognize that use of define lab testing conditions. – 500 E6 ESALs • FWD and modulus tests on
endurance limit in design is a • Review completed lab testing • Sensitivity analysis to cores.
design policy decision. to potentially establish determine thickness • Analysis:
• Healing effects must be empirical relations. beyond which no • Layered elastic analysis
identified to account for • Stress state conditions may be significant increase in life • Fatigue analysis – MEPDG
overloads / seasonal effects. related to the lab to field shift is observed.
factor. Use computational • Define scope of design
modeling and field approach in terms of:
instrumentation to refine lab • Design model relations
test conditions. (i.e., approach will depend
• Comprehensive analysis to link on model used).
endurance limit to shift factor. • Material properties
available – risk
A-9
Project 9-44 Research Team Assessment

There was not consensus among the participants that an endurance limit exists for asphalt
concrete. Most participants agreed that data from laboratory studies support a change in the
slope of laboratory strain versus number of cycles relationships at low strain levels. However,
there was disagreement over the mechanism causing this change in slope. Healing, lack of crack
propagation, and non-linearity in flexural fatigue testing were three effects that were postulated
as potential causes. Some participants expressed the need for additional laboratory fatigue data
to very high numbers of load cycles to confirm that the apparent change in slope at low strain
levels is not due to variability in the fatigue testing.
Much of the brainstorming portion of the session was devoted to a discussion of terminology.
The terms endurance limit, and long-life or perpetual pavements are often used interchangeably,
in spite of the fact that they have different meanings. Endurance limit implies that no damage
occurs in the asphalt concrete leading to an infinite fatigue life. However, for asphalt concrete at
low strain levels it is possible that damage occurs but it is offset by healing or that only crack
initiation without propagation occurs, both resulting in an apparent endurance limit characterized
by very long fatigue lives. In the interest of time the discussion of terminology was terminated
during this session, but was later addressed in an unplanned session at the end the workshop.
This session led to the definition of endurance limit for HMA that was presented earlier.
Another issue raised during the brainstorming session was whether top-down cracking was
within the scope of NCHRP Project 9-44. The panel members confirmed that NCHRP Project 9-
44 was to develop a work plan to validate an endurance limit for bottom-up fatigue cracking
only.
Many of the issues raised during this session, although relevant to the overall objective of the
workshop, were not directly related the objectives of this particular session. The participants
confirmed the research team’s position that the effects of temperature, aging, healing, and
mixture composition must be considered in any laboratory or field experiment included in the
work plan. One fatigue effect discussed during this session that was not initially identified by
the research team was the concept of crack propagation. At low strain levels cracks that initiate
A-10

in asphalt concrete may not propagate resulting in a change in the slope of laboratory fatigue
relationships. The standard definition of failure in flexural fatigue tests, 50 percent reduction in
stiffness, is accompanied by propagation of cracks from the bottom of the specimen. The
numbers of cycles to crack initiation, which may be substantially less than the number of cycles
to 50 percent stiffness reduction, could potentially be used as a design criterion for long-life
pavements. This approach has been used by Dr. Uzan in a methodology for design of perpetual
pavements developed for Israel.
Methodologies for HMA Fatigue Characterization

The objectives of this session were to identify and discuss the advantages and disadvantages
of using alternatives to the standard flexural fatigue test, AASHTO T321 for laboratory fatigue
characterization. The research team’s pre-workshop position was that consideration should be
given to using an alternative to flexural fatigue testing in the future work that will be required by
the NCHRP Project 9-44 work plan. The primary issue with the flexural fatigue test is that the
stresses and strains vary over the depth of the specimen. To rationally incorporate the effects of
healing, damage tolerance and other fatigue effects that may be stress dependent, a test that has a
uniform stress or strain state is needed. A secondary concern is that a standard method for
fabricating flexural fatigue specimens is not available. Another pre-workshop position held by
the research team is that the NCHRP Project 9-44 work plan should include a study to identify a
less cumbersome surrogate method for determining the endurance limit of asphalt concrete so
that the endurance limit can be evaluated in practice. The shortcut approach proposed in
NCHRP 9-38 involves conducting flexural fatigue tests to a relatively large number of cycles and
extrapolating the data to determine the endurance limit. The research team believes that this
approach is not appropriate for routine mixture characterization. The two approaches that
appeared most promising to the research team based on a review of past work is continuum
damage analysis of cyclic direct tension fatigue tests and the dissipated creep strain energy
approach using the indirect tensile test.
A-11


The participants’ views on fatigue testing were gathered through a brainstorming session
where Dr. Bonaquist first presented the research team’s position, then the facilitator captured the
participants’ ideas on cards. The ideas generated during the brain storming session are
summarized in Table 4. These have been organized into four categories: fatigue test methods,
possible surrogate tests, important fatigue test effects, and other considerations.
Table 4. Ideas From the Fatigue Testing Brainstorming Session.

Category Suggestion
Fatigue Test NCHRP 9-38 shortcut approach.
Methods Torsional fatigue test in dynamic shear rheometer. Methodology well defined
by work at Texas A&M. Can be used on asphalt/mastic/fine aggregate system.
Study healing and effect of fillers.
Standard flexural fatigue tests. Should be used to generate baseline data.
Because it is an AASHTO Standard DOT personnel will have confidence in the
results.
Dissipated creep strain energy from the indirect tension test.
Continuum damage analysis of cyclic direct tension-compression testing.
Continuum damage analysis of constant strain rate direct tension tests.
Possible Linear-viscoelastic limit from pseudo strain analysis.
Surrogate Tests Linear-viscoelastic limit from strain sweep testing.
Threshold value from University of Florida HMA Fracture model.
Crack initiation from two stage Weibull analysis of flexural or tension-
compression fatigue test.
Important Determine what the impact is of healing.
Fatigue Test Stress control not very good. Constant strain is better.
Effects Address healing to either account for it or remove it from test
Consider fracture mechanics & crack propagation.
Other Surrogate test is needed.
Considerations Experimental plan should look at all tests with validation at low strain.
Consider credibility of any other test. (Is it an AASHTO procedure? Will it
become a standard?)
Finite element modeling – with micro damage model.
Tests to determine magnitude of “correction” factors for surrogate tests.
Consider multiple tracks with other test facilities in parallel.
Consider other tests in the work plan.
In addition to the discussion of the various test methods, there was a significant amount of
discussion about healing. Some participants questioned whether healing is an important
consideration for field pavements. Professor Monismith explained that healing does not seem to
A-12

be a significant factor in the Heavy Vehicle Simulator (HVS) tests that have been completed in
California. When loading is stopped, there appears to be a fairly large recovery of damage, but
after reloading the damage quickly returns to close to where it was before the rest period. Others
pointed out that healing is most important when the damage is small and that to observe the
effect the testing must use short alternating cycles of loading and resting. If long cycles are used,
the damage may become too large for healing to occur. It was noted that the torsion fatigue test
developed at Texas A&M University for testing asphalt mastics and sand asphalt mixtures
provided a very good tool for evaluating the effects of healing.
Another issue that was discussed at length was the credibility of the various fatigue test
methods. Only the flexural fatigue test has been standardized. There appeared to be general
consensus that if another method was selected, it would be important to demonstrate how the
approach related to results from the flexural fatigue for both high and low strain levels.

Although there was not consensus among the participants concerning the type of fatigue
testing that should be considered for future laboratory studies that may be required in the work
plan, the recommendations listed below received widespread support.
• If additional long-duration fatigue tests are needed to establish relationships between

the endurance limit and mixture properties, the testing should be based on the flexural
fatigue test or an alternative that provides equivalent results. This will allow the
future researchers to make use of the data developed in the NCHRP 9-38, the
University of Illinois extended fatigue studies, and other endurance limit studies.
Testing in multiple laboratories may be needed to develop the necessary data in a
reasonable time frame.
• Serious consideration should be given to the development of a surrogate test for

determining the endurance limit of asphalt concrete for use in routine design and
analysis. Several potential tests were identified in Table 4. The strain sweep testing
completed at the University of New Hampshire in NCHRP 9-38 appears promising.
A-13

• For studies of healing and other effects that are likely dominated by the binder and
filler portion of the mixture the torsional fatigue test in the dynamic shear rheometer
should be considered.
• When recommending tests in the work plan, priority should be given to those with
standard test methods or where standard test methods are in the process of being
accepted.
Flexible Pavement Damage Analysis

The objective of this session was to identify methods for incorporating an endurance limit in
flexible pavement damage analysis. The research team’s pre-workshop position was that the
endurance limit is one of three factors that must be included in an improved flexible pavement
damage analysis; the other two being healing and changes in damage tolerance. The research
team envisioned that these could be incorporated into an improved mechanistic-empirical
analysis by modifying the fatigue relationship. The modification is shown schematically in
Figure 2, which compares current fatigue damage models with the modified relationship.
Current fatigue models consist of a series of parallel lines that for a given mixture depend on
temperature. Note that these models do not include an endurance limit. The research team
envisions that a comprehensive fatigue model will include an endurance limit that will change
with temperature reflecting the increased damage tolerance of mixtures at higher temperatures.
The change in damage tolerance will also result in fatigue lines that are not parallel. The effect
of healing is shown as the dashed lines. Longer rest periods result in increased life at a given
strain level. This type of fatigue model has the potential to more realistically model the fatigue
behavior of HMA mixtures, resulting in smaller calibration factors. Within this model a
pavement can still exhibit infinite life and be subjected to strains above the 70 to 100 μstrain
level currently considered representative of the endurance limit based on laboratory testing.
A-14

Continuous Loading (Lab)

Long Rest Periods (Field)
Increasing
Temperature
Log εt
Log εt
Increasing
Temperature
Log Nf Log Nf
a. Current Fatigue Model b. Proposed Fatigue Model
Figure 2. Conceptual Representation of Improved Flexible Pavement Fatigue Model.

This session was divided into two parts. First, four breakout groups were formed and asked
to develop strategies for incorporating an endurance limit in flexible pavement damage analysis
and to identify important issues associated with their strategy. After the breakout groups
reported their strategies and issues, a general discussion of barriers to the success of these
strategies and possible countermeasures that could be taken to address the barriers was held. The
strategies developed by the four breakout groups are briefly described below.
Group A
The strategy suggested by Group A is very similar to the pre-workshop position of the
research team and shown in Figure 2b. The endurance limit and fatigue relationship would be a
function of temperature and healing. The primary issues identified for this approach were: (1)
laboratory data to support the changes in endurance limit and fatigue life with temperature and
healing, and (2) field data to calibrate the relationship for use in pavement design models.
Group B
The strategy suggested by Group B is to add a variable or modified endurance limit to current
fatigue models. This endurance limit would be a function of the factors that affect the endurance
limit: mixture properties, binder properties, temperature, rest periods, etc.
A-15

ELmod = EL + C gradation + Cbinder + Ctemperature + Cvolumetrics + C rest period + ..... (1)
Where:
EL = average endurance limit for HMA mixtures from laboratory testing
Ci = modification factors to account for factors affecting the endurance limit
This group identified several important issues associated with this approach including:
• Data to support the modification factors.

• Guidance concerning risks associated with using the endurance limit.
• Clear tie between mixture design and pavement design. Select mixtures to obtain
expected performance.
• Training for designers to fully understand the ramifications of the endurance limit.
Group C
The strategy suggested by Group C is to base pavement design on a comprehensive analysis
of damage and recovery in pavement systems. This analysis would address all forms of
pavement distress (rutting, bottom-up cracking, top-down cracking, thermal cracking, etc) not
just fatigue cracking. For fatigue cracking, an “endurance limit” would occur when there is a
balance between the rate of damage and the rate of healing in the pavement. These rates would
be obtained from appropriate fundamental tests. Endurance limit testing would serve as one
verification of the overall modeling process.
This group identified several issues associated with this approach. First is research to obtain
the damage and healing rates as a function of mixture properties, binder properties, temperature,
age hardening, stress state, etc. Second is the development of an appropriate computational tool
for the analysis. Third is the selection of critical design conditions.
Group D
This group provided an overview of approaches used in France, Israel, and the Netherlands
for fatigue analysis of flexible pavements. This group recommended that to improve flexible
pavement fatigue analysis, consideration should be given to using finite element modeling with
A-16

fracture mechanics because this approach can account for initiation and propagation of cracks in
the pavement. They noted, however, that the approach is not practical at this time. They also
suggested that it may be important to develop a method to account for construction defects.
The discussion of the barriers and countermeasures focused on two general approaches for
incorporating an endurance limit in flexible pavement damage analysis that were suggested by
the breakout groups: (1) add an endurance limit that is a function of design factors to current
fatigue analysis, which is the approach suggested by Groups A and B, and (2) detailed
computation of damage and recovery rates which is the approach suggest by Group C and
supported by the issues raised by Group D. Table 5 summarizes the barriers and
countermeasures that were identified for these approaches.
Table 5. Barriers and Countermeasures Identified for Incorporating Endurance Limit in

Flexible Pavement Damage Analysis.
Add Variable Endurance Limit to Existing Fatigue Detailed Computation of Healing and Damage Rates
Analysis
Barriers Countermeasures Barriers Countermeasures
• Data to support • Additional laboratory • Data to define damage • Additional laboratory
endurance limit as a testing. and healing rates. testing.
function of design
factors.
• Lab to field shift • Test materials from • Appropriate analysis • Finite element
factor. field pavements. tool. analysis, fracture
• Calibration. mechanics, etc.
• There is an interaction • Included in current • There is an interaction • Use mechanistic
of base/foundation. stress/strain analysis. of base/foundation. approach on the entire
structure.
• Effect of construction • Use appropriate • Effect of construction • Use appropriate
defects. material properties. defects. material properties
• Solving bottom-up • Solving bottom-up • Both forms can be
cracking will not cracking will not addressed with this
solve top-down solve top-down approach.
cracking. cracking.
• Cost of • Cost of overbuilding. • Cost of • Cost of overbuilding.
implementation. implementation.
A-17


The discussion during this session led to two general approaches for incorporating an
endurance limit in flexible pavement analysis. The first is modify current mechanistic-empirical
fatigue relationships to include a variable endurance limit that is likely a function of mixture
composition, binder properties, temperature, aging, and duration of rest periods. Additional
laboratory testing may be required to quantify the effect of some of these variables. This
approach is similar to the research team’s pre-workshop position. The second approach is to
perform a comprehensive analysis of damage and damage recovery in the pavement using an
appropriate computational analysis. This approach would require computational analyses that
are not practical for use in routine design at this time. Again, the effect of mixture composition,
binder properties, temperature, and aging on the rate of damage and damage recovery would be
determined through appropriate laboratory study.
There was general agreement among the workshop participants that the work plan to be
developed in NCHRP 9-44 should focus on the first approach: modification of current
mechanistic-empirical fatigue relationships to include a variable endurance limit.
Verification and Calibration

The objective of this session was to gather recommendations for verification and calibration
of an improved mechanistic-empirical fatigue analysis that incorporates an endurance limit. The
research team’s pre-workshop position was that an improved pavement damage analysis would
be assembled from the results of a series of laboratory studies addressing the effect of mixture
composition, binder properties, temperature, and rest periods on fatigue damage, damage
recovery, and damage tolerance. Elements of this pavement analysis would be verified using
accelerated pavement tests before final calibration using data from in-service pavements.

The participants’ views on verification and calibration were gathered through a brainstorming
session where Dr. Bonaquist first described the objective of the session, then the facilitator
captured the participants’ ideas on cards. The ideas generated during the brain storming session
A-18

are summarized in Table 6. These have been organized into three categories: data sources,
analysis approaches, and data collection.
Table 6. Ideas From the Verification/Calibration Brainstorming Session.

Category Suggestion
Data Published case studies.
Sources NAPA Perpetual Pavement Award nominees.
LTPP Sections.
Accelerated pavement tests.
Test Roads.
Rolling Dynamic Deflectometer (RDD).
Lab testing can be useful if it is used properly.
Analysis Find sections with no bottom-up cracking.
Approaches Include some pavements with bottom-up cracks.
Detailed analysis of a small number of sections.
LTPP Database provides for analysis of a large number of sections.
Document assumptions made for each field section (plus forensic data).
Be careful with accelerated pavement tests; overloading, low number of
repetitions, lack of aging, etc.
Simplifying assumptions will need to be made - know which you are
making.
Data Ground penetrating radar (GPR) to help verify crack existence.
Collection Spectral Analysis of Surface Waves (SASW) will pick up cracks you can’t
see.
Random trenching to determine bottom-up cracking is merely a “stab in
the dark.”
States will be willing to gather data (including trenching).

This session generated a number of suggestions concerning data sources, analysis approaches
and data collection that will be helpful in developing the work plan. One or more participants
supported the use of all of the data sources listed in Table 6. There was extended discussion
concerning the advisability of using accelerated pavement tests for calibration because most
accelerated pavement tests are conducted with high load levels resulting in strains outside the
region of the endurance limit. There seemed to be general agreement that accelerated pavement
testing could be used to validate certain effects in the high strain region of the fatigue curve, and
that the overall validation and calibration effort should include both accelerated pavement tests
A-19

and evaluation of in-service pavement sections. Ken Fults noted that the Rolling Dynamic
Deflectometer could be used like an accelerated pavement testing machine to rapidly apply a
large number of load pulses to selected pavements in a short period of time.
There was also an extended discussion of whether it was important to document whether
bottom-up fatigue cracking has occurred. Some participants recommended that it was critical
and the validation/calibration must include sections with and without bottom-up fatigue cracks.
Others argued that the important issue is whether the cracks have propagated to the surface,
noting that is very difficult to propagate bottom initiated cracks to the surface of a thick asphalt
concrete pavement. It was suggested that Ground Penetrating Radar (GPR) or Spectral Analysis
of Surface Waves (SASW) could be used to identify cracks that have not propagated to the
pavement surface.
There was general agreement that for the analysis of in-service pavements it will be very
difficult to determine the composition and construction details of the pavements; therefore, many
assumptions will have to be made and documented. There were mixed opinions concerning the
number of pavements that should be included. Some participants argued for a more in-depth
analysis of a small number of pavements, while others recommended an analysis of a large
number of pavements.
Recommendations to the Research Team

Four major topics relevant to the approach proposed by the research team were discussed
during the workshop: (1) endurance limit and other important fatigue effects, (2) methodologies
for HMA fatigue characterization, (3) strategies for incorporating an endurance limit in flexible
pavement damage analysis, and (4) approaches for calibrating and validating pavement analysis
methods that include an endurance limit. The discussion of these topics concluded with a session
designed to identify recommendations for consideration by the NCHRP Project 9-44 research
team. In this session, the facilitator asked the workshop participants to record their
recommendations on cards, which were then placed on the “sticky wall.” The recommendations
were sorted into groups and each group was named by participant consensus. The group names
A-20

and the related recommendations are summarized in Table 7. The recommendations for each
group are discussed below.
Laboratory Materials Characterization

The workshop participants generally agreed that laboratory testing should be included in the
work plan. They made several recommendations concerning laboratory materials
characterization associated with an endurance limit for asphalt. The individual participant
recommendations can be consolidated into the following four general recommendations.
Prove that Endurance Limit Exists

There was not consensus among the workshop participants that an endurance limit exists for
asphalt concrete. It was recommended that additional fatigue tests be conducted to clearly show
that an endurance limit exists for asphalt concrete. A torsional fatigue test in the dynamic shear
rheometer on sand asphalt mixtures was suggested as method to conduct very long cycle fatigue
testing to definitely show whether an endurance limit exists for HMA.
Review Completed Endurance Limit Research

There was general consensus that before undertaking additional laboratory testing the results
of completed extended cycle fatigue tests should be carefully reviewed. Of particular interest
was the identification of relationships between the endurance limit and mixture and binder
properties. Any additional testing should build upon the results of the completed testing.
Identify Practical Surrogate Test

Many workshop participants were adamant that a practical surrogate test for estimating the
endurance was needed for use in design. The surrogate test should be tied to the results obtained
from flexural fatigue testing.
Additional Laboratory Testing

There was general consensus that additional laboratory testing should be conducted to
establish relationships between the endurance limit and mixture properties, binder properties,
temperature and rest periods. The participants were split over how this additional testing should
be done. Some recommended the use of long cycle fatigue tests while others recommended that
the surrogate test be used.
A-21

Table 7. Summary of Workshop Recommendations for the NCHRP 9-44 Research Team.
Category Recommendation
Laboratory Material Need to go through existing data and literature (material properties).
Characterization Review past work on dissipated energy, rest periods.
Perform experiments to identify relationships between damage and healing properties from
practical short-term tests for endurance limit.
Identification of surrogate test method for sensitivity analysis.
Selection of practical short-term tests for endurance limit.
Lab evaluation of relationships between endurance limit, temperature, rest periods, mix
volumetrics, binder properties, etc.
Additional very long fatigue tests (beam) – evaluate a variety of binders, aggregate types,
volumetric factors, temperature, frequency, etc.
Evaluate surrogate tests (torsion on mastic, push-pull).
Develop plan to tie surrogate tests to beam fatigue.
Prove endurance limit by tests with torsion fatigue on mortar including evaluation of factors
that influence the endurance limit.
Identify material properties that resist bottom-up fatigue and influence the level of the
endurance limit.
Field European LTPP studies particularly Germany (high quality database).
Calibration/Validation Need to consider various accelerated pavement tests (APT) NCAT, Mn/ROAD,
WESTRACK, HVS, ALF.
Need to consider in-service pavements: Perpetual Pavement award, SPS (state forensics),
and non-perpetual pavements.
Select appropriate validation sites and obtain data (in-service and accelerated pavement tests
(APT)).
Determine “shift” or “transfer function” from lab to APT to in-service pavements.
Field verification/calibration using APT/LTPP/forensic cases studies.
Must include both cracked and un-cracked sections in work plan. Analyze each with
reasonable flexible pavement analysis model.
Use LTPP sections include (SPS).
Evaluate how to use APT to measure endurance limit.
Include APT and in-service pavements (inherent assumptions).
Test field samples from wheel path and between wheel paths.
Damage Analysis in Identify and evaluate methods to incorporate endurance limit in design.
Design Modeling (MEPDG, continuum damage mechanics, Models + FEM, crack propagation).
Assess cost/conservatism.
Big Picture Follow-up work should be integrated – suggest single contractor rather then separate
contracts – think system.
Plan should coordinate damage analysis approach with lab/field testing and validation.
Have all pieces fit together.
Field Evaluation Identify nondestructive test methods to detect cracks in HMA layer.
Techniques Identify, evaluate and select methods for identifying macro-cracks.
Technology Transfer Develop training modules.
Cost (FHWA Life Develop a methodology for benefit/cost ratio calculation.
Cycle Cost) Risk analysis.
Definitions Standardize definition terminology (9-38 – 9-44).
Develop glossary/definitions of terms related to endurance limit.
Apparent endurance limit/fatigue endurance limit.
A-22

Field Calibration/Validation
The workshop participants agreed that field calibration or validation should be a component
of the work plan. Although some participants favored use of in-service pavements over
accelerated pavement tests, the participants generally agreed that both should be considered in
the work plan. The participants also agreed that both cracked and un-cracked sections should be
analyzed. There was a high level of support for using sections from the LTPP program.
Damage Analysis in Design

The workshop participants agreed that validation of an endurance limit for asphalt concrete
will require a pavement damage analysis that incorporates the effects of an endurance limit.
There was not, however, agreement on the form for the damage analysis. The participants
recommended that the research team evaluate an number of approaches and develop the work
plan around the most promising approach.
Big Picture
During the workshop, the research team suggested that the future work may require multiple
contracts with agencies having experience in laboratory testing of asphalt concrete and
evaluation and analysis of field pavements. The workshop participants disagreed with this
approach and recommended that the future work be done under a single contract to ensure full
coordination of the work.
Field Evaluation Techniques

Some workshop participants supported including a study in the work plan to identify and
select nondestructive testing methods that could identify subsurface cracking in pavement
sections. It was generally agreed that the ability to identify the presence of subsurface cracking
would be beneficial for the selection and analysis of pavements.
Technology Transfer
The workshop participants recommended that the development of initial training materials be
included in the work plan. Topics for training materials include: defining terminology associated
with the endurance limit, factors affecting the endurance limit in asphalt concrete, methods to
A-23

incorporate an endurance limit in flexible pavement design, and the effect of an endurance limit
of the performance of asphalt concrete pavements.
Cost Analysis
Workshop participants were in general agreement that it is important to use some form of cost
analysis to justify the additional effort that may be required to properly consider an endurance
limit in the design of flexible pavements. It was generally agreed that a more sophisticated
damage analysis with additional material characterization would be required to incorporate an
endurance limit in flexible pavement design. Cost savings associated with minimizing the
occurrence of overdesigned pavements would more than offset the additional design costs.
Definitions
Most workshop participants agreed that it was important to develop standard definitions for
terms associated with the endurance limit. Terms that were mentioned included: endurance
limit, apparent endurance limit, long-life pavement, and perpetual pavement. The workshop
participants took part in a spirited debate to develop a definition of endurance limit that could be
used in NCHRP Project 9-38 and 9-44. This debate produced the following definition:
HMA Fatigue Endurance Limit – A level of strain below which there is no

cumulative damage over an indefinite number of load cycles.
This definition was strongly accepted by only a few participants, but was acceptable to all
participants except one. That participant strongly objected to the word “cumulative” in the
definition maintaining that endurance limit means there is no damage in the asphalt material.
Summary and Conclusion

This report documents the Hot Mix Asphalt (HMA) Endurance Limit Workshop held in
Washington, D.C. on August 1 and 2, 2007. The workshop was sponsored by the National
Cooperative Highway Research Program (NCHRP) as part of NCHRP Project 9-44, Developing
a Plan for Validating an Endurance Limit for HMA Pavements. Participants included members
A-24

of the NCHRP Project 9-44 panel and research team, key researchers and consultants with
extensive experience in HMA fatigue analysis, and engineers from highway agencies who are
responsible for designing, constructing, and maintaining flexible pavements. The objective of
the workshop was to discuss several topics relevant to an endurance limit for HMA pavements,
and to provide recommendations for consideration by the research team for the work plan that
will be prepared in NCHRP 9-44.
The HMA Endurance Limit Workshop focused on four topics relevant to this approach being
considered by the NCHRP Project 9-44 research team:

endurance limit.
Numerous recommendations were made by workshop participants that will be considered by the
research team in preparing the NCHRP 9-44 work plan.
A-25

Attachment. Participant Comments from the Opening and Closing

Sessions
Opening Session
Expectations for the workshop from the participants contributed in Round Robin fashion at the
beginning of the workshop.
• State of the art perspective on endurance limit and related topics so that I can best serve
my role on the panel.
• Take home a few ideas.
• Get an idea of what others are doing in the lab.
• Get a clear definition of what the endurance limit is and what are factors that may affect
it.
• See how we can integrate concepts of an endurance limit in design
• How to field verify the endurance limit concept.
• Latest developments in fatigue measurements in the lab, relationships to field
performance, and incorporation in the design process.
• Develop a process leading to what states can endorse.
• What others are thinking in terms of material and structural level analysis.
• Looking at mechanism of damage recovery during rest periods affecting the endurance
limit.
• Scaling factors from laboratory to field and the testing conditions.
• Practicality.
• What’s now and next generation.
• Getting discussion going on healing and how healing mechanism actually produces an
endurance limit.
• Practical methodology that states can handle and use and how it varies by mix.
• Be convinced that endurance limit exists and why it is important.
• Conclude with a clear plan and not repeat work done on previous projects – Don’t
reinvent the wheel and move towards the next project.
• Definition and material processes.
• Consensus of what endurance limit is exists and what are the conditions.
• Clarify different concepts: fatigue cracking and deformation.
• Concept of failure defined
• Define endurance limit for asphalt concrete.
• How to integrate the concept into design after calibration.
• Concept taken from steel and does it exist in asphalt.
• Integrate endurance limit into Mechanistic-Empirical Pavement Design Guide (MEPDG)
and whether it’s a function of temperature.
• Figure out what needs to go into the work plan.
• Broaden horizons as to what’s going on with endurance limit so that we can better work
with the panel members and the contractors.
A-26

Closing Session
Final thoughts about the workshop from the participants contributed in Round Robin fashion at
the end of the workshop.
• Very helpful exercise – important to get diverse view of different opinions.

• Very productive – Ray has things to work with and all the program was well conducted.
• I appreciate you agreeing on definition.
• We got accomplished what we needed to do. Timing was well set up
• Low strain fatigue testing was good to me. One of the better workshops in terms of
getting different viewpoints aired and discussed.
• Feel from panel’s perspective that we have a better understanding to give good direction
to Ray and Don.
• Constructive ideas, good discussions.
• Productive ideas, valuable.
• Interesting and new experience. Group dynamic was interesting. Concept is complicated
and was only touched upon.
• Learned more than I contributed.
• Lot of people with good opinions.
• Useful just hearing different points of view. I took lots of notes.
• Appreciated being part of this and will take back a lot. Wish we had more of this type of
meeting.
• Appreciate the selection of people (diverse group), voicing viewpoint, liked overseas
visitors.
• Good guidance for Ray and panel.
• Enlightening experience that served purpose of workshop.
• Enjoyable hearing different views
• Thank everyone that participated and learned a lot.
• Thanks for the presentations that helped kick this off – got what I needed to do the work
plan.
A-27

APPENDIX B. HMA ENDURANCE LIMIT VALIDATION STUDY

RESEARCH PLAN
Table of Contents
List of Figures ............................................................................................................................ B-iii

List of Tables ............................................................................................................................. B-iv
Acknowledgements..................................................................................................................... B-v
Abstract ...................................................................................................................................... B-vi
Introduction................................................................................................................................. B-1
Purpose.................................................................................................................................... B-1
Statement of the Problem........................................................................................................ B-1
Objectives and Hypothesis...................................................................................................... B-4
Scope of the Plan .................................................................................................................... B-5
Summary of the Research Plan ................................................................................................... B-7
Task by Task Description of the Research Plan ....................................................................... B-12
Task 1. Management and Reporting .................................................................................... B-12
Subtask 1.1 Project Management..................................................................................... B-12
Subtask 1.2 Progress Reporting ....................................................................................... B-15
Subtask 1.3 Interim Reports and Presentations................................................................ B-15
Subtask 1.4 Final Report and Presentation ...................................................................... B-16
Task 1 Milestones ............................................................................................................. B-16
Task 1 Labor Estimate ...................................................................................................... B-16
Task 1 Sources .................................................................................................................. B-18
Task 2. Formulate Design Procedure................................................................................... B-18
Preliminary Design Procedure .......................................................................................... B-18
Subtask 2.1 Review Selected Literature .......................................................................... B-34
Subtask 2.2 Finalize Preliminary Approach ................................................................... B-36
Subtask 2.3 Incorporate Findings from Laboratory Studies ............................................ B-36
Subtask 2.4 Modify Approach Based on Analysis of Accelerated Pavement Tests........ B-37
Subtask 2.5 Prepare Final Design Procedure................................................................... B-38
Task 2 Sources .................................................................................................................. B-39
Task 3. Database Management ............................................................................................ B-41
Subtask 3.1 Develop a Plan to Use the NCHRP 9-30 Database...................................... B-41
Subtask 3.2 Develop Needed Tables ............................................................................... B-43
Subtask 3.3 Input and Manage Data ................................................................................ B-43
Task 3 Sources .................................................................................................................. B-44
Task 4. Laboratory Studies .................................................................................................. B-44
B-i

Subtask 4.1 Experiment 1: Mixture Compositional Factors Affecting Healing

in HMA......................................................................................................... B-46
Subtask 4.2 Experiment 2: Effect of Applied Strain on Healing..................................... B-55
Subtask 4.3 Experiment 3: Effect of Temperature and Rest Period Duration
on Healing .................................................................................................... B-56
Subtask 4.4 Experiment 4: Development of Testing and Analysis Procedures to Determine
Allowable Strain Levels ............................................................................... B-63
Subtask 4.5 Experiment 5: Estimation of Allowable Strain Levels from Mixture
Composition ................................................................................................. B-72
Task 4 Sources .................................................................................................................. B-77
Task 5. Analysis of Pavement Sections ............................................................................... B-79
Subtask 5.1 Review Data Sources and Select Sections for Analysis............................... B-80
Subtask 5.2 Obtain Materials and Data for Accelerated Pavement Tests
and Test Roads ............................................................................................. B-90
Subtask 5.3 Perform Lab Testing and Analyze Accelerated Pavement Tests
and Test Roads............................................................................................. B-92
Subtask 5.4 Obtain Materials and Data for In-Service Pavement Sections..................... B-92
Subtask 5.5 Perform Lab Testing and Analyze In-Service Pavement Sections .............. B-93
Task 5 Sources .................................................................................................................. B-98
Schedule of Tasks ................................................................................................................... B-101
Budget ..................................................................................................................................... B-104
References............................................................................................................................... B-107
Attachment: Recommended LTPP Test Sections .................................................................. B-113
B-ii

List of Figures
Figure 1. Typical S-N Diagram From Laboratory Fatigue Tests............................................... B-2

Figure 2. Results of Flexural Fatigue Tests by Carpenter et al., Including
Extrapolated Results at Low Strain Levels ............................................................... B-3
Figure 3. Project Flow Chart...................................................................................................... B-7
Figure 4. Overall Project Schedule. ........................................................................................... B-9
Figure 5. Recommended Research Management Structure..................................................... B-14
Figure 6. Effect of Rest Periods on Plateau Value .................................................................. B-22
Figure 7. Effect of Rest Period on Fatigue Life....................................................................... B-23
Figure 8. Plateau Value for Continuous Loading as a Function of Applied Strain Level. ...... B-25
Figure 9. Application of Time-Temperature Superposition to Rest Periods. .......................... B-28
Figure 10. Comparison of Applied and Allowable Strains...................................................... B-32
Figure 11. Example of Minimum Asphalt Thicknesses to Resist Bottom Initiated
Fatigue Cracking With Observed Performance of Four UK Pavement Sections. .. B-33
Figure 12. Schematic of Pulsed, Strain Controlled Fatigue Loading. ..................................... B-53
Figure 13. Expected Results When Healing is Significant. ..................................................... B-54
Figure 14. Schematic of Time-Temperature Superposition Applied to Rest Periods.............. B-62
Figure 15. Typical Plot of Ratio of Dissipated Energy Change Versus Loading Cycles ........ B-64
Figure 16. Typical Damage Ratio Curves From a Cyclic Direct Tension Fatigue Test.......... B-67
Figure 17. Typical Damage Relationship From Continuum Damage Analysis. ..................... B-70
Figure 18. Comparison of Area Fatigue Cracking (Area Alligator Cracking
Based on a Percent of Wheel Path Area) and HMA Layer Thickness ................... B-86
Figure 19. Comparison of the Area Fatigue Cracking for and Maximum Tensile Strain
Computed at the Bottom of the HMA Layer ......................................................... B-86
Figure 20. Comparison of the Maximum Tensile Strain at the Bottom of the HMA
Layer and HMA Thickness..................................................................................... B-88
Figure 21. Schematic of Field Section Data Analysis. ............................................................ B-95
Figure 22. Project Schedule With Critical Path Shown in Black. ......................................... B-102
Figure 23. Project Budget. ..................................................................................................... B-105
Figure 24. Estimated Monthly and Cumulative Expenditures............................................... B-106
B-iii

List of Tables
Table 1. Summary of Man-hour and Cost Estimates................................................................. B-8

Table 2. Major Task 1 Milestones. .......................................................................................... B-17
Table 3. Estimated Labor Hours for Task 1............................................................................. B-17
Table 4. Summary of Existing Pavement Analysis Approaches Considered. ......................... B-20
Table 5. Effect of Rest Period on Fatigue Life. ....................................................................... B-23
Table 6. Approximate Rest Periods for Various Design Traffic Levels.................................. B-24
Table 7. Allowable Strains for Various Design Traffic Levels. .............................................. B-27
Table 8. Computation of Allowable Strain Strains.................................................................. B-30
Table 9. Applied Strains for Design Example. ........................................................................ B-31
Table 10. Major Task 2 Milestones. ........................................................................................ B-38
Table 11. Estimated Labor Hours for Task 2........................................................................... B-39
Table 14. Summary of Proposed Laboratory Experiments...................................................... B-45
Table 15. Design for a Two Level, Seven Factor Plackett-Burman Experiment. ................... B-47
Table 16. Summary of Proposed Experiment 1. ...................................................................... B-52
Table 17. Strain Level Experiment. ......................................................................................... B-56
Table 18. Experimental Design for Experiment 3. .................................................................. B-57
Table 19. Temperature and Frequency Combinations for Dynamic Modulus Tests............... B-58
Table 22. HMA Thicknesses in NCAT Structural Sections .................................................... B-82
Table 23. Summary of MNRoad Mainline HMA Pavement Sections..................................... B-83
Table 24. FHWA Pavement Testing Facility Superpave Fatigue Experiment. ....................... B-83
Table 25. Preliminary Matrix for Field Calibration of the Allowable Strain
Limit Design Procedure. .......................................................................................... B-89
Table 26. LTPP Sections Recommended for Consideration. .................................................. B-90
Table 27. Summary of Required Inputs for Allowable Strain Limit Design........................... B-91
Table 28. Preliminary Testing Plan for Cores From the LTPP Sections. ................................ B-94
Table 31. Project Milestone Summary................................................................................... B-103
Table 32. Labor Costs Used in Budget Preparation............................................................... B-104
Table 33. Travel Cost Estimate.............................................................................................. B-104
Table 34. Estimate of Report Printing Costs. ........................................................................ B-104
B-iv

Acknowledgements
The work reported herein was performed under NCHRP Project 9-44 by Advanced Asphalt
Technologies, LLC, and Applied Research Associates, Inc. Ramon Bonaquist, Chief Operating
Officer for Advanced Asphalt Technologies, LLC, served as Principal Investigator for the project
and was the primary author of this research plan. Mr. Harold Von Quintus, Principal Engineer
for Applied Research Associates, Inc. co-authored this research plan.
Special thanks are extended to the participants of the HMA Endurance Limit Workshop listed
below. These professionals unselfishly presented ideas to the research team that helped shape
the planned research.
Name Affiliation
Dr. David Anderson Consultant
Dr. Samuel Carpenter University of Illinois
Dr. Donald Christensen Advanced Asphalt Technologies, LLC
Dr. Herve Di Benedetto Ecole Nat. des TPE
Mr. Bruce Dietrich Florida Department of Transportation
Mr. Kenneth Fults KWF Pavement Consulting
Mr. Roger Green Ohio Department of Transportation
Dr. Kevin Hall University of Arkansas
Dr. Edward Harrigan National Cooperative Highway Research Program
Dr. Richard Kim North Carolina State University
Dr. Dallas Little Texas A&M University
Dr. Leslie Ann McCarthy Federal Highway Administration
Dr. Andre Molenaar Delft University
Professor Carl Monismith University of California Berkeley
Dr. David Newcomb National Asphalt Pavement Association
Dr. Michael Nunn Lane One Limited
Dr. Brian Prowell Advanced Material Services, LLC
Dr. Rey Roque University of Florida
Ms. Amy Schutzbach Illinois Department of Transportation
Dr. Jacob Uzan Technion University
Dr. Linbing Wang Virginia Polytechnic and State University
Dr. Matthew Witczak Arizona State University
B-v

Abstract
This document presents a plan for research to rationally incorporate the concept of an
endurance limit for hot mix asphalt (HMA) into a mechanistic-empirical algorithm for bottom
initiated fatigue cracking in flexible pavements, and to validate the resulting procedure using
performance data from full-scale pavement sections.
The planned research is based on the hypothesis that the endurance limit for HMA is the
result of a balance of damage caused by loading and healing or damage recovery that occurs
during rest periods. Under this hypothesis the primary objective in designing a flexible
pavement to resist bottom initiated fatigue cracking will be to make sure that the damage induced
by loading remains small enough so that healing occurs and there is no accumulation of damage
over the life of the pavement. This is a significant departure from current cumulative or
incremental damage models, which assume that no healing occurs and that each load cycle uses
up a portion of the finite fatigue life of the HMA.
This research plan includes a preliminary design procedure that is based on layered elastic
analysis and compatible with the Mechanistic-Empirical Pavement Design Guide (MEPDG). It
uses allowable strains to identify satisfactory conditions for full healing. The allowable strains
are a function of the properties of the HMA, the pavement temperature, and the duration of rest
periods between traffic loads. Five laboratory experiments that are needed to fully develop the
procedure are described. Studies using data from completed accelerated pavement tests and test
roads are proposed to verify critical aspects of the design procedure. Finally, an experiment to
calibrate the design procedure using selected test sections from the Long Term Pavement
Performance Program is presented.
The recommended research study has been titled the HMA Endurance Limit Validation
Study. It addresses an important concept in the design of perpetual pavements that is gaining
increasing acceptance worldwide. It is envisioned that application of an endurance limit in
flexible pavement design will lead to more effective pavement sections with significant benefit
and cost savings to the public.
B-vi

Introduction
Purpose
This document presents a plan for research to rationally incorporate the concept of an
endurance limit for hot mix asphalt (HMA) into a mechanistic-empirical algorithm for bottom
initiated fatigue cracking in flexible pavements, and to validate the resulting procedure using
performance data from full-scale pavement sections. For HMA pavements, the endurance limit
has been defined as a level of strain below which there is no cumulative damage over an
indefinite number of load cycles (1).
This research plan is the primary product of National Cooperative Highway Research
Program (NCHRP) Project 9-44, Developing a Plan for Validating an Endurance Limit for HMA
Pavements. The recommended research study has been titled the HMA Endurance Limit
Validation Study. It addresses an important concept in the design of perpetual pavements that is
gaining increasing acceptance worldwide. It is envisioned that application of an endurance limit
in flexible pavement design will lead to more effective pavement sections with significant
benefit and cost savings to the public.
Statement of the Problem

In engineering, fatigue refers to the progressive and localized damage that occurs when a
material is subjected to repeated loading below its ultimate strength. It is an important
consideration in the design of many civil engineering structures including pavements.
The fatigue behavior of materials is evaluated using laboratory fatigue tests, where a sample
is loaded repeatedly using a known stress or strain and the number of load applications are
counted until the sample fails. By performing tests at different stress or strain levels a Wöhler
curve or S-N diagram can be developed. These diagrams are simply plots of the applied stress or
strain and the corresponding number of cycles to failure. Figure 1 shows two typical S-N
diagrams generated from laboratory test data. In curve (a), the fatigue life increases at a
gradually increasing rate with decreasing stress amplitude. In curve (b), on the other hand, the
fatigue life gradually increases until a limit is reached (50 MPa in this case) where the fatigue
B-1

life becomes indefinite. This is called the endurance limit for the material. The endurance limit is
a critical concept in the design of structures that must resist large numbers of repeated loads. If
stresses or strains are kept below the endurance limit, the structure will be able to withstand an
infinite number of load applications.
150
Stress Amplitude, S (MPa)
100
(b) endurance limit = 50 MPa

50
(a) no endurance limit

0
1.E+04 1.E+05 1.E+06 1.E+07 1.E+08
Cycles to Failure, N
Figure 1. Typical S-N Diagram From Laboratory Fatigue Tests: (a) No Endurance Limit;
(b) 50 MPa Endurance Limit.
Many materials do not have the well-defined endurance limit shown schematically in Figure
1. HMA is one of these materials. Although early HMA fatigue research conducted by
Monismith and his colleagues suggested that HMA exhibited an endurance limit at
approximately 70 μstrain (2), only limited HMA fatigue research was conducted at low strain
levels until recently when the Asphalt Pavement Alliance began promoting the concept of
perpetual pavement design (3). A perpetual pavement is an asphalt pavement that provides a
very long life without structural failure and only requires periodic replacement of the surface. A
key element of perpetual pavement design is to eliminate fatigue cracking that initiates at the
bottom of the HMA base due to repeated flexure under traffic loading and to confine distresses to
the surface of the pavement, which can easily be renewed by milling and resurfacing.
B-2

In response to increasing interest in perpetual pavements, a substantial amount of laboratory

fatigue testing has recently been performed in the United States in an effort to demonstrate that
HMA does exhibit an endurance limit. Most of this work has been performed at the University
of Illinois (4,5) and the National Center for Asphalt Technology (NCAT) (6). These studies
provide clear evidence that the fatigue behavior of HMA is much different in low strain level
tests compared to normal strain level tests. Figure 2 shows a consolidated plot of the University
of Illinois fatigue data including low and normal strain level test data. Below approximately 100
μstrain, the fatigue life is significantly longer than estimated from extrapolation of normal strain
level test data. Healing of microdamage has been proposed as the primary reason for the
increased fatigue life at low strain levels (1, 7, 8). For cyclic tests at low strain levels, it appears
that the damage that is caused by loading is offset by healing that occurs during unloading
resulting in essentially infinite fatigue life. Current mechanistic-empirical fatigue criteria for
HMA, including the field calibrated criterion in the Mechanistic Empirical Pavement Design
Guide (MEPDG), are based on results from normal strain level tests and do not include the low
strain level effects shown in Figure 2.

Results at Low Strain Levels (4).
B-3

Detailed investigation of four heavily trafficked pavements in the United Kingdom support
the perpetual pavement concept and the likelihood of an endurance limit for HMA. This
comprehensive study found no evidence of fatigue damage at the bottom of properly constructed
thick flexible pavements with total HMA thickness ranging from 230 to 350 mm (9). Cracks in
these pavements were found to have initiated at the surface and deflections monitored over a
number of years generally showed steady or decreasing deflection with increasing cumulative
traffic, indicating that fatigue damage to the bottom of the HMA was not occurring. Similar
conclusions concerning the absence of cracking at the bottom of thick HMA pavements have
been reported by others (10, 11, 12).
In summary, there is mounting evidence that an endurance limit for HMA does exist. It has
been observed in laboratory studies of fatigue at low strain levels, and several documented case
studies indicate that bottom initiated fatigue cracking is almost non-existent in properly
constructed, thick HMA pavements. A concentrated research effort, however, is needed to
validate the endurance limit concept, and to devise effective methods for incorporating it in
mechanistic-empirical pavement design methods.
Objectives and Hypothesis

The objectives of the HMA Endurance Limit Validation Study are:
1. To incorporate the concept of an endurance limit for HMA into a mechanistic-

empirical algorithm for bottom initiated fatigue cracking in flexible pavements.
2. To validate the methodology using performance data from full-scale pavement

sections.
These objectives could potentially be satisfied using a number of research approaches. The
specific approach presented in this plan is based on the following hypothesis, which was
developed from a review of recent literature concerning the fatigue response of HMA, and
recommendations made during the HMA Endurance Limit Workshop conducted early in
NCHRP Project 9-44 (1):
B-4

HMA does exhibit an endurance limit. This endurance limit, however, does not
reflect an absence of load induced damage in the HMA. It is the result of a
balance of damage caused by loading and healing or damage recovery that
occurs during rest periods. The endurance limit for HMA is, therefore, not a
single value, but will change depending on the loading and environmental
conditions applied to the HMA. To properly consider this form of an endurance
limit in flexible pavement design requires consideration of the effects of loading,
environment and material properties on both damage accumulation and healing.
Under this hypothesis the primary objective in designing a flexible pavement to resist
bottom initiated fatigue cracking will be to make sure that the damage induced by loading
remains small enough so that healing occurs and there is no accumulation of damage over
the life of the pavement. This is a significant departure from current cumulative or
incremental damage models which assume that no healing occurs and that each load
cycle uses up a portion of the finite fatigue life of the HMA. The hypothesis presented
above implies that any flexible pavement structure can be designed to indefinitely resist
bottom initiated fatigue cracking. Thicker pavements will be required for heavier loads,
shorter rest periods (higher traffic volume), and poorer foundation conditions. To
successfully formulate this type of design procedure will require research to quantify the
effects of temperature, aging, and materials properties on damage accumulation and
damage recovery in HMA. Once formulated, the procedure can be validated using
performance data from full-scale pavement sections.
Scope of the Plan

This research plan is a comprehensive document describing the research that must be
completed to successfully incorporate the concept of an endurance limit for HMA into a fatigue
algorithm for bottom initiated fatigue cracking and to validate the resulting procedure using full-
scale pavement sections. It includes four parts in addition to this Introduction. The first is a
summary that briefly describes the proposed research and presents overall cost estimates and
time requirements. The second is a description of the required research tasks. This section
includes detailed information for each task and subtask, including (1) a description of the work to
be performed, (2) preliminary experimental designs when appropriate, (3) a list of milestones
related to the task, (4) labor hour estimates, and (5) a listing of pertinent data and reference
material that will be needed to accomplish the task. The third is a detailed schedule for the
B-5

project. The schedule addresses the sequence of the research tasks and the interactions between
tasks. Finally, the fourth presents the proposed budget for the project. The budget includes
detailed estimates of labor and other costs associated with each task and subtask.
B-6

Summary of the Research Plan

The HMA Endurance Limit Validation Study consist of five major tasks: (1) Management
and Reporting, (2) Formulate Design Procedure, (3) Database Management, (4) Laboratory
Studies, and (5) Analysis of Pavement Sections. Figure 3 presents an overall flow chart for the
project with major interactions between tasks identified. Table 1 lists the subtasks for each of
the five major tasks and presents estimated labor hours and costs. The HMA Endurance Limit
Validation Study is estimated to require approximately 12,923 man-hours of effort at a cost of
approximately $1.5 million. Figure 4 presents the overall schedule for the project, which is
estimated to require 48 months to complete.
Task 1
Management and
Reporting
Task 4 Task 2
Laboratory Formulate Design
Studies Procedure
Task 3 Task 5
Database Analysis of
Management Pavement Sections
Figure 3. Project Flow Chart.
B-7

Table 1. Summary of Man-hour and Cost Estimates.
Estimated Labor Hours

Senior
Task/ Eng./ Eng./ Estimated
Subtask Description Stat Prog. Tech. Admin. Cost
1.0 Management and Reporting
1.1 Project Management 424 0 0 40 $66,000
1.2 Progress Reporting 210 0 0 20 $32,700
1.2 Interim Reports and Presentations 780 0 0 80 $129,780
1.3 Final Report and Presentation 420 0 0 40 $68,400
Task 1 Total 1834 0 0 180 $296,880
2.0 Formulate Design Procedure
2.1 Review Selected Literature 240 160 0 0 $52,000
2.2 Finalize Preliminary Approach 80 160 0 0 $28,000
2.3 Incorporate Findings from Laboratory Studies 80 160 0 0 $28,000
2.4 Modify Approach Based on Analysis of 80 80 0 0 $20,000
Accelerated Pavement Tests
2.5 Prepare Final Design Procedure 120 80 0 0 $26,000
Task 2 Total 600 640 0 0 $154,000
3.0 Database Management
3.1 Develop Plan to Use NCHRP 9-30 Database 120 0 0 0 $18,000
3.2 Develop Needed Tables 80 240 0 0 $36,000
3.3 Input and Manage Data 40 396 0 0 $45,600
Task 3 Total 240 636 0 0 $99,600
4.0 Laboratory Studies
4.1 Experiment 1: Mixture Compositional Factors 42 0 388 0 $39,280
Affecting Healing
4.2 Experiment 2: Effect of Applied Strain on 32 0 214 0 $22,990
Healing
4.3 Experiment 3: Effect of Temperature and Rest 69 0 242 0 $30,920
Period Duration on Healing
4.4 Experiment 4: Testing and Analysis 168 0 392 0 $58,520
Procedures for Allowable Strain Levels
4.5 Experiment 5: Estimation of Allowable Strain 456 0 1890 0 $229,050
Levels from Mixture Composition
Task 4 Total 767 0 3126 0 $380,760
5.0 Analysis of Pavement Sections
5.1 Review Data Sources and Select Sections for 52 320 0 0 $39,800
Analysis
5.2 Obtain Materials and Data for Accelerated 48 280 0 0 $35,200
Pavement Tests
5.3 Perform Testing and Analyze Accelerated 164 512 32 0 $78,520
Pavement Tests
5.4 Obtain Materials and Data for In-Service 120 1280 0 0 $195,600
Pavement Sections
5.5 Perform Testing and Analyze In-Service 300 512 1280 0 $205,000
Pavement Sections
Task 5 Total 684 2904 1312 0 $554,120
Project Total 4,125 4,180 4,438 180 1,485,360
B-8

Figure 4. Overall Project Schedule.
B-9
Task 1, Management and Reporting, includes all activities normally associated with
management and reporting for NCHRP Projects. Major management tasks include scheduling,
coordinating, and directing various technical work activities as well as project financial
management. Reporting activities include monthly and quarterly progress reports, the
preparation of several interim reports and presentations, and the preparation of the final report.
Interim reports are required at approximately 6 month intervals and coincide with the completion
of five critical milestones:
approach,
The final report will document the entire study and will be prepared from the interim reports.
Task 2, Formulate Design Procedure, is a critical project task that will be active throughout
the project. This task includes finalizing the preliminary approach that is presented in this
research plan, modifying the preliminary approach based on the results of the laboratory studies
and selected accelerated pavement tests, and preparation of the final design procedure after
analysis of the calibration pavement sections. It is important to emphasize that the preliminary
approach prepared early in this task will shape the laboratory studies and guide the selection of
pavement sections, both accelerated pavement tests and in-service pavement sections.
Task 3, Database Management, is a support task that will be active throughout the project. A
database will be developed in this task to store and analyze data from the laboratory studies and
the analysis of the pavement sections. It is envisioned that the database will be an adaptation of
the one developed in NCHRP Project 9-30.
B-10

Task 4, Laboratory Studies, includes the planning and execution of five laboratory studies that
are needed to complete the design procedure that will be formulated in Task 2. The laboratory
studies concentrate on quantifying what affects the healing properties of HMA. The laboratory
studies will be sufficient in breadth to develop models relating mixture and binder properties to
the key engineering properties required for the analysis.
Task 5, Analysis of Pavement Sections, includes several activities associated with the
selection and analysis of full-scale pavements. The preliminary design procedure formulated in
Task 2 will be tested using data from completed accelerated pavement tests, such as the fatigue
studies from the Federal Highway Administration’s (FHWA’s) Pavement Testing Facility or the
structural sections included in the National Center for Asphalt Technology (NCAT) test track.
Calibration of the design procedure will be accomplished through an analysis of in-service
pavements where it has been documented that bottom-up fatigue cracking has occurred or has
not occurred. These analyses will serve to calibrate the design procedure and validate the HMA
endurance limit concept. The predictive models developed in Task 4 will be used in the analysis
of the full-scale pavement sections. This will allow consideration of pavement sections where
original materials are not available since the required data can be obtained from cores taken from
the pavement section.
B-11

Task by Task Description of the Research Plan

This section of the research plan presents detailed descriptions of each of the tasks and
subtasks included in the HMA Endurance Limit Validation Study. Each task description
includes a detailed description of the work to be performed including: (1) preliminary
experimental designs when appropriate, (2) a list of milestones related to the task, (3) labor hour
estimates, and (4) a listing of pertinent data and reference material that will be needed to
accomplish the task.
Task 1. Management and Reporting

Task 1 includes all activities normally associated with management and reporting for NCHRP
projects. Task 1 has been divided into four subtasks:
1.1 Project Management,

1.2 Progress Reporting,
1.3 Interim Reports and Presentations, and
1.4 Final Report and Presentation.
Each of these subtasks is described in detail below.
Subtask 1.1 Project Management

Effective project management will be critical to the successful completion of the HMA
Endurance Limit Study. The study requires that the Principal Investigator have in-depth
knowledge of the following technical areas:
• Mechanistic-empirical pavement design and analysis,

• Experimental design,
• Model development,
• Laboratory characterization of HMA,
• Accelerated pavement testing, and
• Pavement evaluation.
B-12

Since the design procedure incorporating an endurance limit for bottom initiated fatigue cracking
will determine the details of the laboratory and field studies, the Principal Investigator should
directly lead Task 2, Formulate Design Procedure. To efficiently manage several tasks that will
be conducted concurrently, the team structure shown in Figure 5 is recommended. In this
structure, the Principal Investigator is supported by three teams: Laboratory, Pavement, and
Data Support, each with a separate team leader. Additionally, it is strongly recommended that a
Statistician be included in the project team to assist the Principal Investigator and team leaders
with detailed experimental design, model formulation, and model calibration. The Principal
Investigator will be responsible for the overall technical content of the project, while the team
leaders will be responsible for the details of the work in their area of expertise. In addition to the
scenario shown in Figure 5 where the management team consists of the Principal Investigator
and three team leaders, other structures are possible depending on the skills and commitment
levels of the senior members of the research team. For example, the Principal Investigator may
also serve as one of the team leaders and one individual may serve as the leader of the remaining
two teams. It is recommended, however, that a single individual not fill more than two
leadership roles.
This research plan as modified during the proposal process will serve as the principal project
management tool. Shortly after contract award, the research management team should meet and
the Principal Investigator should make initial task assignments to the project team. The research
management team should then meet semi-monthly to discuss the progress of the work and
resolve any problems that may develop. These meetings should be scheduled to provide timely
information for the monthly and quarterly progress reports discussed in the next section.
Another important aspect of project management is coordination with other on-going research
efforts. Several studies addressing cracking in flexible pavements are on-going including: (1)
NCHRP 1-41, Models for Predicting Reflection Cracking of Hot-Mix Asphalt Overlays,
(2) NCHRP 1-42A, Models for Predicting Top-Down Cracking of Hot-Mix Asphalt Layers, and
(3) the fatigue studies being conducted in the Asphalt Research Consortium. Although different
approaches are being used in each of these studies, it is important that the research team monitor
and share information with these studies.
B-13

Principal Investigator
• Formulate Design Procedure
• Experimental Design
• Laboratory and Pavement Data Analysis
• Reporting
Statistician
• Experimental Design
• Model Formulation
• Model Calibration
Laboratory Team Leader
• Laboratory Experimental Design
• Oversee Laboratory Testing
• Laboratory Data Analysis
• Model Formulation
• Reporting
Pavement Team Leader

• Select Pavement Sections
• Oversee Pavement Data Collection
• Pavement Data Analysis
• Reporting
Data Support Team Leader

• Design Data Base Structure
• Oversee Data Input
• Assist With Data Analysis
• Reporting
Figure 5. Recommended Research Management Structure.
B-14

Subtask 1.2 Progress Reporting

NCHRP has specific requirements for progress reporting (13). The required reporting
includes brief monthly progress reports and detailed quarterly progress reports. The monthly
progress reports briefly summarize the work that has been completed, planned work, problems
encountered, and expenditures for the project. The detailed quarterly progress reports describe
completed work, planned work, and problems encountered in sufficient detail for review by the
project panel during the course of the project. The quarterly progress reports are the means by
which the project panel provides direction to the research team. Timely progress reporting and
communication with the project panel are essential tools for effective project management.
Subtask 1.3 Interim Reports and Presentations

The HMA Endurance Limit Validation Study includes a series of interim reports that coincide
with the completion of five critical milestones:
approach,
Each interim report should be prepared in accordance with NCHRP requirements (14) and
specifically address the work completed in the relevant tasks. These interim reports will provide
more detailed information than normally contained in the progress reports. The final report will
be compiled from the interim reports.
Presentations to the project panel are included after the second and fourth interim reports.
The purpose of these presentations is to encourage a dialog between the project panel and the
B-15

Principal Investigator on the progress and direction of the work. One full day should be planned
for each of these presentations sessions. Each session should include:
(1) A presentation from the Principal Investigator focusing on the completed interim
reports, planned work, and any changes to the direction of the research.
(2) A discussion period where the project panel discusses critical aspects of the
completed and planned work with the Principal Investigator and other key members
of the research team.
(3) Recommendations concerning the direction of the research.
Subtask 1.4 Final Report and Presentation

The final report will document the entire project and will be compiled from the five interim
reports. This report will be prepared in accordance with NCHRP requirements (14) and revised
as required for publication. Upon completion of the review of the draft of the final report, the
Principal Investigator will meet with the project panel to discuss the outcome of the project and
to jointly develop recommendations concerning implementation and additional research
activities.
Task 1 Milestones
Table 2 summarizes the major milestones for Task 1. This milestone schedule assumes that
this research plan as modified during the proposal process will serve as the work plan for the
project. In addition to the major milestones listed in Table 2, meetings of the research
management team will occur semi-monthly throughout the project, and monthly progress reports
will be submitted as required by NCHRP.
Task 1 Labor Estimate

Table 3 presents the estimated labor required for Task 1. Table 3 presents estimated labor
hours for each of the positions in the research management structure presented in Figure 5.
Project management and reporting is estimated to require at total of 2014 man-hours of effort.
This is approximately 16 percent of the total effort required for the project.
B-16

Table 2. Major Task 1 Milestones.
Milestone Description Months After

Contract Award
1.1 Initial Work Assignments 0.5
1.2 First Quarterly Progress Report 3
1.3 Second Quarterly Progress Report 6
1.4 First Interim Report (Preliminary Design Procedure and Laboratory 7
Analysis Approach)
1.5 Third Quarterly Progress Report 9
1.6 Fourth Quarterly Progress Report 12
1.7 Second Interim Report (Selection of Pavement Sections for Analysis) 13
1.8 First Panel Presentation (Interim Reports 1 and 2) 14
1.9 Fifth Quarterly Progress Report 15
1.10 Sixth Quarterly Progress Report 18
1.11 Seventh Quarterly Progress Report 21
1.12 Third Interim Report (Analysis of Laboratory Studies) 22
1.13 Eighth Quarterly Progress Report 24
1.14 Ninth Quarterly Progress Report 27
1.15 Tenth Quarterly Progress Report 30
1.16 Fourth Interim Report (Design Procedure Incorporating Findings From 30
Laboratory Studies and Analysis of Accelerated Pavement Tests)
1.17 Second Panel Presentation (Interim Reports 3 and 4) 31
1.18 Eleventh Quarterly Progress Report 33
1.19 Twelfth Quarterly Progress Report 36
1.20 Thirteenth Quarterly Progress Report 39
1.21 Fifth Interim Report (Analysis of Validation Sections and Final Design 42
Procedure)
1.22 Fourteenth Quarterly Progress Report 42
1.23 Submit Draft of Final Report 45
1.24 Fifteenth Quarterly Progress Report 45
1.25 Third Panel Presentation (Draft Final Report and Recommendations for 46
Implementation and Additional Research
1.26 Revised Final Report 48
Table 3. Estimated Labor Hours for Task 1.
Laboratory Pavement
Principal Data Support Administrative
Subtask Statistician Team Team
Investigator Team Leader Assistant
Leader Leader
1.1 Project Management 112 0 104 104 104 40
1.2 Progress Reporting 120 0 30 30 30 20
1.3 Interim Reports and Presentations 432 0 116 116 116 80
1.4 Final Report and Presentation 216 0 68 68 68 40
Total 880 0 318 318 318 180
B-17

Task 1 Sources
Procedural Manual for Agencies Conducting Research in the National Cooperative

Highway Research Program, Transportation Research Board, August, 2006.
Preparing Your CRP Final Report, Transportation Research Board, September, 2006
Task 2. Formulate Design Procedure

In Task 2, the procedure for designing pavements to resist bottom initiated fatigue cracking
that considers the effects of an endurance limit for HMA will be developed. Task 2 will build on
the preliminary procedure described in this research plan in four distinct steps:
1. Finalize preliminary procedure,

2. Incorporate findings from laboratory studies,
3. Modify approach based on analysis of accelerated pavement tests, and
4. Prepare final design procedure.
In step 1, the research team will become familiar with the preliminary procedure described in this
research plan, and develop improvements based on their review of the relevant literature and
research in progress. Then in steps 2, 3, and 4 information obtained from Tasks 4 and 5 of the
project will be used to further improve the procedure. The final product will be a procedure for
designing flexible pavements to resist bottom initiated fatigue cracking that accounts for the
effects of an HMA endurance limit. This procedure will be compatible with current mechanistic-
empirical flexible pavement design methods such as the MEPDG.
Preliminary Design Procedure

Background
A major part of the work completed during NCHRP 9-44 was the development of a
preliminary procedure for designing pavements to resist bottom initiated fatigue cracking that
considers the effects of an endurance limit. This preliminary procedure is based on the research
hypothesis that the endurance limit for HMA is the result of a balance of damage caused by
B-18

loading and healing or damage recovery that occurs during rest periods. Under this hypothesis
the primary objective in designing a flexible pavement to resist bottom initiated fatigue cracking
will be to make sure that the damage induced by loading remains small enough so that healing
occurs and there is no accumulation of damage over the life of the pavement. This is a
significant departure from current cumulative or incremental damage models, which assume that
no healing occurs and that each load cycle uses up a portion of the finite fatigue life of the HMA.
A number of approaches for designing pavements to resist bottom initiated fatigue cracking
were reviewed during NCHRP Project 9-44. Table 4 briefly summarizes the approaches that
were considered. These range from relatively simple modifications of traditional mechanistic-
empirical fatigue algorithms to sophisticated finite element models based on damage mechanics
and fracture mechanics. The major deficiency of the more practical approaches is that they do
not account for the beneficial effects of healing. In the HMA Endurance Limit Workshop,
healing was identified as a significant factor affecting the endurance limit in HMA (1). The
sophisticated approaches can account for healing, but are not practical at this time for use in
routine pavement design.
Effect of Rest Periods

An alternative approach was conceived during NCHRP Project 9-44 based on recent
endurance limit research published by Carpenter and Shen (7). In this work, Carpenter and Shen
clearly demonstrated the beneficial effects of rest periods on the fatigue life of HMA. Strain
controlled flexural fatigue tests were conducted at 20 °C using a 10 Hz haversine load pulse with
a rest period between each pulse to simulate the time between traffic loads. The rest periods
ranged from 0 sec (continuous loading) to 9 seconds. Two 19 mm mixtures, one with a neat PG
64-22 binder and one with a polymer modified PG 70-22 binder, were tested. The gradation,
binder content and air void content of the two mixtures was the same.
B-19

Table 4. Summary of Existing Pavement Analysis Approaches Considered.
Approach Key Elements Selected Advantages Disadvantages

References
Strain Limit Assume fatigue life is infinite at Timm and Easy to implement in existing M-E Does not consider the beneficial
damage levels below the endurance Young (15) design. effect of rest periods.
limit. Use Miner’s law for strain Witczak (16) Compatible with layered elastic Relies on Miners law for strains
levels above the endurance limit. Thompson and analysis used in MEPDG. above the endurance limit.
Carpenter (17) Above endurance limit fatigue life

of HMA is predefined.
Crack Initiation Limit strain level to that causing Sidess and Easy to implement in existing M-E Does not consider the beneficial
crack initiation in laboratory fatigue Uzan (18) design. effect of rest periods.
tests. Compatible with layered elastic Relies on Miners law.
analysis used in MEPDG. Cycles to crack initiation are
Rational basis for design. predefined.
Strain Limit-Crack Assume fatigue life is infinite at Von Quintus Relatively easy to implement in Does not consider the beneficial
Initiation damage levels below the endurance (19, 20) existing M-E based methods. effect of rest periods.
limit. Use Miner’s law for strain Compatible with layered elastic Relies on Miner’s law for strains
levels above the endurance limit. The analysis used in the MEPDG. above the endurance limit.
endurance limit is estimated from the Value is dependent on the Key property used to estimate
indirect tensile strength test and is temperature (modulus), and endurance limit is highly variable.
dependent on the modulus of the volumetric properties of the
mixture. mixture.
Recursive Miner’s Law Modify fatigue life of HMA to Tsai, et al., (21) Easy to implement in existing M-E Assumes that HMA fatigue life
account for the strength loss of a design. deteriorates with traffic loading.
pavement structure as a function of Compatible with layered elastic Does not consider the beneficial
traffic loading. analysis used in MEPDG. effect of rest periods.
Accounts for changes in fatigue life
of HMA with traffic.
Visco-Elastic Model the evolution of damage in a Mun, et al., (22) Can be used to predict crack Computationally intensive.
Continuum Damage viscoelastic continuum. initiation. Not compatible with layered elastic
Directly accounts for damage analysis used in MEPDG.
accumulation and healing.
Fracture Mechanics Model responses at the crack tip and Roque, et al. Predict crack growth. Requires crack initiation model.
the propagation of cracks. (23) Computationally intensive.
Not compatible with layered elastic
analysis used in MEPDG.
B-20
The resulting data were analyzed using the ratio of dissipated energy change (RDEC)
approach developed at the University of Illinois (5). In this approach, the ratio of dissipated
energy change reaches a plateau value (PV) where a constant percentage of the input energy is
being converted to damage. The University of Illinois research found a unique relationship
between the plateau value and the traditional definition of failure in flexural fatigue tests, 50
percent stiffness reduction, that holds for a range of mixtures and loading conditions (5).
PV = 0.4429 × N f ( )−1.1102 (1)
where:
PV = plateau value
Lower plateau values correspond to longer fatigue lives. Based on the ratio of dissipated energy
change approach, an HMA mixture will exhibit endurance limit behavior when the plateau value
is 6.74x10-9 or less, which based on Equation 1 corresponds to a traditional fatigue life of
1.1x107 cycles or greater.
The effect of rest periods on the plateau value is shown in Figure 6 for the two mixtures that
were tested. Equations 2 and 3 present the relationship between plateau value and the length of
the rest period that were developed for the neat PG 64-22 and the modified PG 70-22 mixtures,
respectively for a strain level of 500 μstrain (7).
PV = 1.018 × 10 −5 (RP + 1)
−0.9069
PV = 4.353 ×10 −6 (RP + 1)

−1.352
where:
PV = plateau value
B-21

a. Neat PG 64-22
b. Polymer PG 70-22
Figure 6. Effect of Rest Periods on Plateau Value (5).
The decreasing plateau values for tests with rest periods result in increasing fatigue lives.
This can be quantified by substituting plateau values from Equations 2 or 3 into Equation 1. The
results are summarized in Table 5. Figure 7 shows the beneficial effect of the rest periods on the
fatigue lives for the two mixtures. There is a substantial improvement in the fatigue life of both
mixtures. The values for the neat PG 64-22 mixture are of similar magnitude to improvements
previously reported by Bonnaure, et al. (24). The effect of rest periods on the modified PG 70-
22 mixture is much more pronounced.
B-22

Table 5. Effect of Rest Period on Fatigue Life.
Rest Period, Neat PG 64-22 Modified PG 70-22

sec PV Nf Ratio PV Nf Ratio
0 1.02E-05 1.51E+04 1.00 4.35E-06 3.24E+04 1.00
1 5.43E-06 2.65E+04 1.76 1.71E-06 7.53E+04 2.33
2 3.76E-06 3.70E+04 2.45 9.86E-07 1.23E+05 3.81
3 2.90E-06 4.68E+04 3.10 6.68E-07 1.75E+05 5.41
4 2.37E-06 5.61E+04 3.72 4.94E-07 2.30E+05 7.10
5 2.00E-06 6.51E+04 4.32 3.86E-07 2.87E+05 8.86
6 1.74E-06 7.39E+04 4.90 3.13E-07 3.46E+05 10.69
7 1.54E-06 8.24E+04 5.47 2.62E-07 4.08E+05 12.58
8 1.39E-06 9.07E+04 6.02 2.23E-07 4.70E+05 14.52
9 1.26E-06 9.89E+04 6.56 1.94E-07 5.35E+05 16.51
10 1.16E-06 1.07E+05 7.09 1.70E-07 6.01E+05 18.54
20
18
16
14
Relative Fatigue Life
12
NEAT PG 64-22
Polymer PG 70-22
10
0
0 2 4 6 8 10 12
Duration of Rest Period, sec
Figure 7. Effect of Rest Period on Fatigue Life.
An estimate of approximate rest periods can be obtained from the 20 year design traffic level
typically used in mixture design. Table 6 summarizes rest periods for various design traffic
levels. The rest period for a 20 year design traffic level of 100 million ESAL is approximately 6
sec., which results in a factor of 5 improvement in the fatigue life of the mixture with the neat
PG 64-22 binder and a factor of 10 improvement for the polymer modified PG 70-22 mixture.
B-23

Table 6. Approximate Rest Periods for Various Design Traffic Levels.
20 Year Rest Period,

Design ESAL ESAL/Day ESAL/sec sec
1.00E+05 13.7 0.0002 6307.2
3.00E+05 41.1 0.0005 2102.4
1.00E+06 137.0 0.0016 630.7
3.00E+06 411.0 0.0048 210.2
1.00E+07 1369.9 0.0159 63.1
3.00E+07 4109.6 0.0476 21.0
1.00E+08 13698.6 0.1585 6.3
3.00E+08 41095.9 0.4756 2.1
Allowable Strains
Continuous loading tests at different strain levels were also conducted by Carpenter and Shen
on the two mixtures and the plateau values are shown in Figure 6 for a rest period of zero
(RP+1=1) (7). From these data relationships between the plateau value for continuous loading
and the applied strain level can be developed as shown in Figure 8. These relationships are given
in Equations 4 and 5 for the neat PG 64-22 mixture and the polymer modified PG 70-22 mixture.
For neat PG 64-22 PV0 = 9.142 ×10 −16 (ε )3.617 (4)
For modified PG 70-22 PV0 = 5.347 ×10 −21 (ε )5.331 (5)

where:
PV0 = plateau value for continuous loading
B-24

1.00E-03
Neat PG 64-22
Polymer PG 70-22
1.00E-04
1.00E-05
3.617E+00
Plateau Value
y = 9.142E-16x
R2 = 9.689E-01
1.00E-06
y = 5.347E-21x5.331E+00
2
R = 9.640E-01
1.00E-07
1.00E-08
1.00E-09
10 100 1000 10000
Strain, μstrain
Figure 8. Plateau Value for Continuous Loading as a Function of Applied Strain Level.
Based on previous work by Bonnaure (24), it is reasonable to assume that the effect of the rest
periods is the same at each strain level. Substituting Equations 4 and 5 for the constants 1.018
x10-5 and 4.353 x10-6 in Equations 2 and 3 respectively, yield the following relationships
between the plateau value, applied strain and rest period for the two mixtures.
PV = 9.142 ×10 −16 (ε ) (RP + 1)−0.9069

3.617
PV = 5.347 ×10 −21 (ε ) (RP + 1)−1.352

5.331
where:
PV = plateau value
B-25

Equations 6 and 7 can then be substituted into Equation 1 and solved for the allowable strain
level to produce a selected mixture fatigue life.
⎡ (1 + RP )0.2507 ⎤
ε a = 11483.32⎢ ⎥
( )
For neat PG 64-22 0.3069
(8)
⎢⎣ N f ⎥⎦
⎡ (1 + RP )0.2536 ⎤
ε a = 5448.74 ⎢ ⎥
( )
For modified PG 70 -22 0.2082
(9)
⎢⎣ N f ⎥⎦
where:
εa = allowable tensile strain, μstrain
Nf = number of cycles to failure
Recalling that endurance limit behavior occurs when the number of cycles to failure exceeds
1.1 x107, then setting the number of cycles to failure in Equations 8 and 9 to a value above
1.1 x 107 will ensure that full healing occurs at the selected rest period. Conservatively using 2.0
x 107 as the number of cycles to failure yields Equations 10 and 11, which give allowable strain
levels as a function of rest period to ensure that full healing occurs.
For neat PG 64-22 ε af = 66.0(1 + RP )0.2507 (10)
For modified PG 70-22 ε af = 164.5(1 + RP )0.2536 (11)
where:
εaf = allowable tensile strain for full healing, μstrain
If the strains in a pavement at 20 °C are kept below the values given by Equations 10 and 11,
then complete healing will occur during intermittent rest periods, and the pavement will exhibit
endurance limit behavior. Table 7 summarizes these strain levels for various 20 year design
traffic levels.
B-26

Table 7. Allowable Strains for Various Design Traffic Levels.
20 Year Rest Period, Allowable Strains, μstrain

Design ESAL sec Neat PG 64-22 Modified PG 70-22
1.00E+05 6307.2 592 1513
3.00E+05 2102.4 449 1145
1.00E+06 630.7 332 844
3.00E+06 210.2 253 639
1.00E+07 63.1 187 472
3.00E+07 21.0 143 360
1.00E+08 6.3 109 272
3.00E+08 2.1 88 219
Multiple Temperatures
The allowable strains presented in the previous section were developed from test data
obtained at 20 °C. To be useful in a pavement design procedure, the allowable strains for a wide
range of temperatures must be available. In this procedure the major concern is the effect of
temperature on the healing properties of the mixture. Previous research by Bonnaure, et al. (24)
concluded that the beneficial effect of rest periods increased with increasing temperature. Since
healing can be envisioned as a type of flow phenomenon where the binder flows together to
repair microcracks, it has been hypothesized that the effect of healing at multiple temperatures
can be accounted for using time-temperature superposition. By applying time-temperature
superposition, rest periods at different temperatures can be reduced to an equivalent rest period at
20 °C. The reduced rest period for temperatures above 20 °C will be longer than the actual rest
period, while those for temperatures below 20 °C will be shorter than the actual rest period.
Research conducted in NCHRP Project 9-19 showed that linear, viscoelastic time-temperature
shift factors obtained from dynamic modulus tests could be applied when a high level of
nonlinear damage is present (25). Equation 12 presents the application of time-temperature
superposition to the duration of the rest period.

where:
B-27


modulus testing.
Figure 9 illustrates the use of time-temperature superposition for rest periods at temperatures of
40, 20, and 4 °C using 20° C as the reference temperature. In developing Figure 9, typical time-
temperature shift factors were used (log(AT) for 4 °C =2.0 and log(AT) for 40 °C = -2.2).
1.0E+06
1.0E+05
1.0E+04
Reduced Rest Period at 20 C, sec
1.0E+03
4C
1.0E+02 20 C
40 C
1.0E+01
1.0E+00
1.0E-01
1.0E-02
1.0E+04 1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10
20 Year Design Traffic, ESAL
Figure 9. Application of Time-Temperature Superposition to Rest Periods.
Design Example
This section illustrates how the proposed methodology would be used in a mechanistic-
empirical design system such as the MEPDG. To limit the number of computations, a monthly
analysis is illustrated using typical pavement temperatures estimated from mean monthly air
temperature data from Reagan National Airport in Washington, DC. The pavement being
analyzed is 9 in of HMA constructed on a 6 in aggregate subbase base layer and a silty clay
subgrade. The 20 year design traffic level is 1 x 108 ESALs, and the design traffic speed is 45
mph. The purpose of the analysis is to determine if the pavement section is sufficiently thick to
B-28

resist bottom initiated fatigue cracking assuming the fatigue properties of the neat PG 64-22
mixture discussed in the preceding section.
Material Properties
For this analysis the modulus of the subgrade is assumed to be 4,500 psi and constant
throughout the year. The modulus of the aggregate subbase is assumed to be 25,000 psi and is
also assumed constant throughout the year. Dynamic modulus testing of a typical 19 mm
mixture with PG 64-22 binder using the Simple Performance Test System yielded the master
curve and shift factors given in Equations 13 and 14 for a reference temperature of 20 °C. The
allowable strains for full healing are given in Equation 15.
⎡ 3.259 ⎤
log E * = 0.234 + ⎢
⎣ 1+ e( )
−1.213−0.499 log( f r ) ⎥
⎦
(13)
⎛1 1 ⎞
log f r = log f + 10448.2⎜ − ⎟ (14)
⎝ T 293.2 ⎠
ε af = 66.0(1 + RPr )0.2507 (15)
where:
⎪E*⎪ = dynamic modulus, ksi
f = loading frequency, Hz
T = temperature, °K
εaf = allowable tensile strain of full healing, μstrain
RPr = reduced rest period at 20 °C, sec
Allowable Strains
Allowable strains at the bottom of the asphalt layer are determined from Equation 15 using
reduced rest periods that depend on the traffic volume and the monthly pavement temperature.
B-29

Mean monthly pavement temperatures can be estimated from the mean monthly air temperature
using Equation 16 (26).
⎛ 1 ⎞ 34
M p = M a ⎜1 + ⎟− +6 (16)
⎝ z +4⎠ z +4
where:
Mp = mean monthly pavement temperature at depth z, °F
Ma = mean monthly air temperature, °F
z = depth, in
For a 20 year design traffic of 1 x 108 ESAL, the rest period is 6.3 sec. The reduced rest period
for each month is determined from Equation 12 using the shift factors from the dynamic modulus
master curve and the mean monthly pavement temperature. Table 8 summarizes the computation
of the allowable strains. Because the reduced rest period is much shorter during cold months
compared to warm months, the allowable strain levels for full healing are significantly lower.
Table 8. Computation of Allowable Strain Strains.
Mean
Monthly Reduced Allowable
Pavement Rest Rest Strain
Temp, Period, Period, Level,
Month C Log (AT) sec sec μstrain
Jan 5.5 1.851 6.3 0.09 67
Feb 7.3 1.611 6.3 0.15 68
Mar 12.2 0.971 6.3 0.67 75
Apr 18.0 0.242 6.3 3.61 97
May 23.7 -0.445 6.3 17.56 137
Jun 29.0 -1.065 6.3 73.20 194
Jul 32.0 -1.397 6.3 157.26 235
Aug 30.9 -1.276 6.3 118.95 219
Sep 26.8 -0.803 6.3 40.04 167
Oct 19.7 0.036 6.3 5.79 107
Nov 13.8 0.773 6.3 1.06 79
Dec 8.4 1.469 6.3 0.21 69
B-30

Applied Strains
The strains applied by the traffic loading are computed for the design axle load using layered
elastic analysis. In this example an 18 kip single axle load was used for computing applied
strains. For this example the modulus of the subgrade and subbase are constant at 4.5 and 25 ksi,
respectively. The modulus of the asphalt depends on the pavement temperature and the speed of
traffic. Recent research by Al-Qadi, et al, using in-situ instrumentation at the Virginia Smart
Road (27) indicates that loading rates computed by the transformed section analysis in the
MEPDG and other approaches such as that recommended by Barksdale (28) overestimate the
frequency of the load pulse. Based on data presented by Al-Qadi, a loading rate of 16 Hz
appears reasonable for a depth of 9 in under 45 mph traffic. Table 10 summarizes the applied
strains for each month computed using the KENLAYER software (26). The applied strains are
compared to the allowable strains in Figure 10. Since the applied strains in Table 9 are less than
the allowable strains, the proposed section is acceptable with respect to bottom initiated fatigue
cracking. An interesting observation in Figure 10 is that this analysis shows that the critical
condition for bottom initiated fatigue cracking occurs at intermediate to low pavement
temperatures, which is in contrast with traditional cumulative or incremental damage analyses,
which show that the majority of the fatigue damage occurs at high pavement temperatures.
Table 9. Applied Strains for Design Example.
Mean
Monthly
Pavement Load Reduced AC Subbase Subgrade Applied
Temp, Frequency, Frequency, Modulus, Modulus, Modulus, Strain,
Month C Log (AT) Hz Hz ksi ksi ksi μstrain
Jan 5.6 1.841 16 1108.93 1969.7 25 4.5 51
Feb 7.5 1.584 16 614.01 1858.0 25 4.5 54
Mar 12.8 0.900 16 127.08 1535.8 25 4.5 62
Apr 19.0 0.122 16 21.21 1148.4 25 4.5 77
May 25.1 -0.608 16 3.95 801.7 25 4.5 100
Jun 30.8 -1.265 16 0.87 535.6 25 4.5 133
Jul 33.9 -1.616 16 0.39 418.2 25 4.5 157
Aug 32.8 -1.488 16 0.52 458.9 25 4.5 148
Sep 28.4 -0.987 16 1.65 641.1 25 4.5 117
Oct 20.8 -0.096 16 12.83 1041.1 25 4.5 83
Nov 14.4 0.688 16 78.05 1431.1 25 4.5 65
Dec 8.7 1.432 16 432.33 1789.1 25 4.5 55
B-31

2.0
1.9
Ratio of Allowable Strain to Applied Strain
1.8
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Pavement Temperature, C
Figure 10. Comparison of Applied and Allowable Strains.
Traffic Level
The analysis presented above can be performed to determine minimum asphalt thicknesses to
resist bottom initiated fatigue cracking for the given subgrade and subbase conditions as a
function of traffic level. The results are shown in Figure 11 for a 22 kip single axle load. A 22
kip axle load was used to allow comparison with observed data from the analysis of in-service
pavements that was conducted in the United Kingdom (9). Figure 11 also shows the thickness
and accumulated traffic for the four pavements that were analyzed in detail and it was
documented that bottom initiated fatigue cracking had not occurred. This comparison shows the
engineering reasonableness of the proposed approach. It is reasonable to expect that when the
proposed approach is improved to consider the effects of aging and design reliability, the
minimum asphalt thicknesses will increase. It is important to note that at the low traffic levels,
deformation of the subgrade may govern the analysis rather than bottom initiated fatigue
cracking. Research in the United Kingdom indicates that for asphalt thicknesses less than about
7 in subgrade deformation governs the performance of the pavement (9). This limit is shown as
the dashed line in Figure 11.
B-32

16
UK Pavements With No
Bottom Initiated
14 Fatigue Cracking
Minimum Asphalt Thickness, in
12
Minimum Asphalt
10 Thickness for Structural
Deformation
0
1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 1.0E+10
20 Year Design Traffic, 22 kip ESAL
Cracking With Observed Performance of Four UK Pavement Sections.
Aging
The example presented above does not consider the important effect of aging on either the
applied or allowable strains. As a pavement ages, the modulus of the HMA will increase due to
the increased stiffness of the asphalt binder resulting in lower applied strains. Aging will also
affect the healing rate for the HMA. Although no data is currently available for the effect of
aging on the healing rate, it is reasonable to expect that the healing rate will decrease
significantly on aging resulting in lower allowable strains for full healing. Early research on
healing by Bonnaure, et al.(24) showed that healing rates were much greater in softer binders.
The effect of aging can be incorporated in the procedure outlined above, by computing allowable
and applied strains as a function of pavement age. The global aging model currently
incorporated in the MEPDG provides a method for computing aged modulus values (29).
Additional research proposed in the laboratory studies discussed in Task 4 will be required to
develop a model of the effect of aging on HMA healing and the allowable strains that result in
B-33

full healing. For perpetual pavement design, it may only be necessary to perform the analysis for
highly aged conditions.
Climate and Mixed Traffic Effects

The MEPDG currently provides excellent capabilities to evaluate the effects of climate and
mixed traffic on the applied strains at the bottom of the asphalt layer. This capability can be
used with the allowable strains described above to determine the HMA thickness needed to resist
bottom initiated fatigue cracking.
Reliability
Because the computations involved in the analysis do not require substantial computer time,
reliability can be included in the analysis using Monte-Carlo simulation. This approach has
already been implemented in the PerRoad program (15). In fact, the allowable strains computed
based on rest periods can be input as the threshold criteria for HMA the in the PerRoad program
and the analysis for a single season can be performed.
Subtask 2.1 Review Selected Literature

The preceding section presented a rational approach for incorporating an endurance limit for
bottom initiated fatigue cracking in mechanistic-empirical pavement design methods. The
method is based on maintaining tensile strain levels at the bottom of the HMA low enough to
ensure that complete healing occurs between traffic loads and that there is no accumulation of
damage at the bottom of the asphalt concrete. This is accomplished through the use of allowable
strain levels that depend on the damage and healing properties of the HMA, the aging
characteristics of the HMA, the duration of rest periods between traffic loads, and the
temperature of the pavement. Several improvements to this preliminary procedure should be
made based on a detailed review of selected literature. These improvements should be made
before the detailed laboratory testing plans are developed in Task 4. Areas where improvements
should be considered are summarized below:
Duration of Rest Periods. In the preliminary procedure a very simple approach was
used to estimate the duration of rest periods as a function of design traffic level.
B-34

Additional effort should be expended to establish representative rest period durations as a

function of traffic level and roadway classification. Potential sources of information on
the duration of rest periods include: the Highway Capacity Manual (30), data from traffic
studies performed for the Long Term Pavement Performance Program (31), and the
approach used in Strategic Highway Research Program (SHRP) Contract 005 (32).
MEPDG Modifications. It is envisioned that the design procedure will be implemented

in the MEPDG. Because the MEPDG is an AASHTO product, any proposed changes,
including research versions of the software, must be approved by AASHTO. AASHTO
has approved a research version of the software for use in NCHRP 9-30A. It is
envisioned that similar approval will be granted for this project.
A detailed review of the documentation and source code for the MEPDG will be required
to determine specific modifications that will be needed to implement the approach. This
review should concentrate on how the MEDPG addresses the following:
1. Climatic effects,
2. Mixed traffic (Currently hourly traffic distribution factors are not included in
any flexible pavement analysis, they are only considered for the rigid
pavement analysis. It may be necessary to tie daily truck traffic distributions
to temperature distributions to accurately consider the effect of healing),
3. Vehicle speed effects,
4. Vehicle wander (Currently being considered for revision under NCHRP
Project 9-30A),
5. Location of maximum strain at the bottom of the asphalt layer for various axle
configurations,
6. Aging, and
7. Reliability.
B-35

The MEPDG source code should also be reviewed to determine how to remove the
current bottom initiated fatigue cracking algorithm and implement the allowable strain
approach.
Since the major new component of the design procedure is the determination of allowable
strain levels that provide for complete healing between traffic loads, completed research on
healing in HMA should be reviewed before finalizing the laboratory testing program. Several
important publications addressing healing in HMA that should be reviewed are listed at the end
of the Task 2 work description.
Subtask 2.2 Finalize Preliminary Approach

In this Subtask, the preliminary design procedure described in this research plan will be
improved based on the findings from the literature review conducted in Subtask 2.1. The
improved procedure will then be implemented in a research version of the MEPDG software,
designated NCHRP9-44A_Version 0.1. The products of this subtask will be detailed
documentation of the preliminary procedure and a modified research version of the MEPDG
software. The documentation and software will be submitted as part of the first interim report
that is scheduled for delivery during the 7th month of the project.
Subtask 2.3 Incorporate Findings from Laboratory Studies

In Subtask 2.3, the preliminary procedure and software developed in Subtask 2.2 will be
improved by adding the results from the laboratory studies conducted in Task 4. The laboratory
studies are envisioned to result in the following improvements:
1. Verification that time-temperature superposition can be applied to HMA rest periods.

This is an assumption that has been included in the preliminary procedure described
in this research plan.
2. Verification that healing in HMA is not affected by the strain level provided the
strains are low enough that macrocracking does not occur.
3. Testing and data analysis procedures for determining mixture specific allowable
strains levels for HMA. Under the current hierarchical structure of the MEPDG, this
testing and analysis will be used in Level 1 analyses.
B-36

4. A model for estimating allowable strain levels as a function of mixture composition,

binder properties, and age. This model will be used for Level 2 and Level 3 analyses,
and for the analysis of pavement sections to be completed in Task 5. Von Quintus
(19, 20) developed a model to estimate the allowable strain levels at which no
damage is retained in the HMA mixtures. It is estimated from the indirect tensile
strength test, and is dependent on the mixture composition. Healing within this
approach is captured through field calibration factors. A similar type of approach is
expected for this research plan, but using healing directly.
The products of this subtask will be detailed documentation of the improved procedure and a
modified research version of the MEPDG software designated NCHRP9-44A_Version 0.2 that
will be used in Subtask 5.3 for the analysis of selected accelerated pavement test and test road
sections. This documentation and software will be further improved in Subtask 2.4.
Subtask 2.4 Modify Approach Based on Analysis of Accelerated Pavement Tests

In Subtask 2.4, the improved procedure developed in Subtask 2.3 will be further improved
based on the analysis of selected accelerated pavement test and test road sections. The
accelerated pavement test and test road sections will be selected in Subtask 5.1 to exercise
critical aspects of the design procedure. For example, the fatigue tests conducted at the FHWA
Pavement Testing Facility provide data addressing the effect of temperature on HMA fatigue and
healing. These field tests provide the ability to investigate time-temperature superposition as
applied to rest periods. The structural sections from the NCAT test track provide data addressing
the effect of thickness on HMA fatigue, while the WesTrack sections provide data on the effect
of HMA material properties on fatigue. In all three cases, the applied strains should exceed the
allowable strains for full healing. On the other hand, the original sections from the NCAT test
track that are still in-service, should have applied strains that are below the allowable strains for
full healing.
The products of this subtask will be detailed documentation of the improved procedure and a
modified research version of the MEPDG software designated NCHRP9-44A_Version 0.3 that
will be used in Subtask 5.5 for the analysis of selected in-service pavement test sections. The
B-37

documentation and software will be submitted as part of the fourth interim report that is
scheduled for delivery during the 30th month of the project.
Subtask 2.5 Prepare Final Design Procedure

The final subtask in Task 2 is the preparation of the final design procedure. This will be
accomplished after analysis of the in-service pavement calibration sections is completed in
Subtask 5.5. It is envisioned that design reliability will be the primary effort addressed in this
final version of the design procedure.
The products of Subtask 2.5 will be detailed documentation of the final procedure and a
modified research version of the MEPDG software designated NCHRP9-44A_Version 1.0. The
documentation and software will be submitted as part of the draft final report that is scheduled
for delivery during the 45th month of the project.
Task 2 Milestones
Table 10 summarizes the major milestones for Task 2. These are all associated with
improvements to the preliminary design procedure described in this research plan, and the
development of various modified research versions of the MEPDG software.

Contract Award
2.1 Review Selected Literature 3
2.2 Preliminary Approach and NCHRP 944A_Version 0.1 Software 6
2.3 Incorporate Findings from Laboratory Studies into NCHRP 9- 27
44A_Version 0.2 Software
2.4 Modify Approach Based on Analysis of Selected Accelerated 29
Pavement Tests and NCHRP 9-44A_Version 0.3 Software
2.5 Prepare Final Design Procedure and NCHRP 9-44A_Version 1.0 41
Software
B-38


hours for each of the positions in the research management structure presented in Figure 5 and
for programming assistance. Task 2 is estimated to require at total of 1240 man-hours of effort.
This is approximately 10 percent of the total effort required for the project.
Laboratory Pavement
Principal Data Support
Subtask Statistician Team Team
Investigator Team Leader Programmer
Leader Leader
Review Selected Literature 80 0 80 80 0 160
Finalize Preliminary Approach 40 0 20 20 0 160
Incorporate Findings from 40 0 40 0 0 160
Laboratory Studies
Modify Approach Based on Analysis 40 0 0 40 0 80
of Accelerated Pavement Tests
Prepare Final Design Procedure 80 0 20 20 0 80
Total 280 0 160 160 0 640
Task 2 Sources
Endurance Limit Studies

Carpenter, S.H., Ghuzlan, K.A., and Shen, S., “Fatigue Endurance Limit for Highway
and Airport Pavements,” Transportation Research Record No. 1832, Transportation
Carpenter, S.H., and Shen, S., “Application of the Dissipated Energy Concept in Fatigue
Prowell, B., Brown, E., R., Daniel, J., Bhattacharjee, S., Von Quintus, H., Carpenter, S.,
Shen, S., Anderson, M., Swamy, A. K., and Maghsoodloo, S., “Endurance Limit of Hot
Mix Asphalt Mixtures to Prevent Fatigue Cracking in Flexible Pavements,“ Updated
Draft Final Report, NCHRP 9-38, National Cooperative Highway Research Program,
Washington, D.C., May, 2008.
Soltani, A., Solaimanian, M., and Anderson, D.A., “An Investigation of the Endurance
Limit of Hot-Mix Asphalt Concrete Using a New Uniaxial Fatigue Protocol,” Report
Number FHWA-HIF-07-002, Federal Highway Administration, Washington, D.C.,
B-39

HMA Healing Studies

Bonnaure, F.P, Huibers, A.H.J.J., Boonders, A., “A Laboratory Investigation of the
Influence of Rest Periods on the Fatigue Response of Bituminous Mixes,” Proceedings,
Kim, B. and Roque, R., “Evaluation of Healing Property of Asphalt Mixtures,”

Transportation Research Record No. 1970, Transportation Research Board,
Washington, D.C., 2006.
Kim, Y.R., Little, D.N., and Benson, F.C., “Chemical and Mechanical Evaluation of
Healing of Asphalt Concrete, Journal of the Association of Asphalt Paving
Little, D. N., Lytton, R. L., Williams, D., and Chen, C. W., “Microdamage Healing in
Asphalt and Asphalt Concrete, Volume I: Microdamage and Microdamage Healing
Project Summary Report,” Report Number FHWA-RD-98-141, Federal Highway
Administration, Washington, D.C., June 2001.
Pronk, A.C., “Partial Healing, “A New Approach for the Damage Process During Fatigue
Testing of Asphalt Specimens,” Asphalt Concrete Simulation, Modeling, and
Experimental Characterization, Geotechnical Special Publication No. 146, American
Society of Civil Engineers, Reston, VA, 2005.
MEPDG
National Cooperative Highway Research Program, http://www.trb.org/mepdg/guide.htm
(accessed June 30, 2008).
Rest Periods
Hajek, J. J., Selezneva, O., I., Mladenovic, G., and Jiang, Y., J., “Estimating Cumulative
Traffic Loads, Volume II: Traffic Data Assessment and Axle Load Projection for the
Sites With Acceptable Axle Weight Data, Final Report for Phase 2,” Report Number
FHWA-RD-03-094, Federal Highway Administration, Washington, D.C., March, 2005.
Lytton, R L; Uzan, J; Fernando, E G; Roque, R; Hiltunen, D; Stoffels, S M,

“Development And Validation Of Performance Prediction Models And Specifications
For Asphalt Binders And Paving Mixes,” Report Number SHRP-A-357, Strategic
Highway Research Program, Washington, D.C., 1993.
B-40

Transportation Research Board, Highway Capacity Manual, Special Report 209, Third
Edition, Transportation Research Board, Washington, DC 1994.
Task 3. Database Management

Task 3 includes the development and management of a database to store and analyze data
generated in Task 4, Laboratory Testing, and Task 5, Analysis of Pavement Sections. It is
envisioned that the database will be an adaptation of the one developed in NCHRP Project 9-30.
Task 3 has been divided into three subtasks:
3.1 Develop a Plan to Use the NCHRP 9-30 Database

3.2 Develop Needed Tables,
3.3 Input and Manage Data
Each of these subtasks are described in detail below.
Subtask 3.1 Develop a Plan to Use the NCHRP 9-30 Database

In NCRHP Project 9-30 a database called M-E Distress Prediction Models (M-E_DPM) was
developed to provide an appropriate database structure for storing all HMA pavement data
required for the continued improvement of mechanistic-empirical pavement distress prediction
models (33). It was envisioned that this database would serve future mechanistic-empirical
development efforts such as the HMA Endurance Limit Validation Study. Consequently, M-
E_DPM was designed to be flexible to accommodate changes in models and test procedures.
The database was developed in Microsoft Access to take advantage of the standard and custom
features available for entering and storing data, querying data, and generating reports. It consists
of three parts that are briefly described below:
• Descriptive Database. This part of the database includes text files that document
details for the data included in the model inputs portion of the database. This part of
the database provides the flexibility to define the new type of data that will be needed
in the HMA Endurance Limit Validation Study.
• Model Inputs. This part of the database includes the data required to execute the
mechanistic-empirical models. The data are contained in tables that define (1)
B-41

pavement structure, (2) material properties, (3) traffic, and (4) climate. For the HMA
Endurance Limit Validation Study new material properties associated with the
allowable strain levels for full healing will be required.
• Performance Data. This part of the database includes various measures of pavement
distress including (1) area of alligator cracking, (2) longitudinal cracking, (3)
transverse cracking, (4) rutting, (5) smoothness, and (6) other distresses such as
potholes and the extent of patching. Additional detail concerning the performance
data will be required by the Endurance Limit Validation Study to differentiate
bottom-initiated cracking from surface initiated cracking.
In Subtask 3.1, the current version of M-E_DPM and its documentation will be reviewed and
a plan will be developed for modifying this database for use in the analysis of the pavement
sections in Task 5. M-E_DPM is currently being improved and additional data is being added in
NCHRP Project 9-30A. The key HMA property needed for the analysis approach described
earlier in this plan is the allowable strains for full healing, which will be a function of HMA
damage and healing properties, age, and climate. The laboratory experiments in Task 4 will
establish methods for measuring the HMA damage and healing properties and will develop
models for estimating these properties from mixture composition and binder properties that can
be easily measured on field cores. The required material property data tables will have to be
added to the model inputs portion of M-E_DPM. The extent of bottom-initiated fatigue cracking
will be the pavement distress needed for the analysis of the pavement sections. Only the extent
of surface cracking is currently contained in M-E_DPM; therefore, additional tables will be
needed to store this data. The data will be obtained from the crack coring operations described in
Task 5.
A plan for storing the data from the Task 4 laboratory experiments will also be developed in
Subtask 3.1. This will likely be a separate database that can be linked to M-E_DPM upon
completion of the analysis of the laboratory experiments and the development of the models and
procedures for computation of allowable strains for full healing.
B-42

Subtask 3.2 Develop Needed Tables

In Subtask 3.2, the various tables required to use M-E_DPM in the HMA Endurance Limit
Validation Study will be developed. Work in this task will be coordinated with the data
collection and analysis activities in Tasks 4 and 5.
Subtask 3.3 Input and Manage Data

Data from the project will be entered into the database and managed in Subtask 3.3. This
subtask includes entering the data, verifying the entered data, and extracting data in support of
the analyses that will be performed in Tasks 4 and 5 of the project. Subtask 3.3 will be active
during the majority of the project.
Task 3 Milestones
Table 12 summarizes the major milestones for Task 3. These are all associated with the
modification of M-E_DPM for use in this project. In addition to the major milestones listed in
Table 12, data entry and management will occur as needed from month 8 through the completion
of Tasks 3, 4, and 5 in month 41 of the project.

Contract Award
3.1 Database Plan 8
3.2 Tables for Laboratory Data 10
3.3 Tables for Analysis of Accelerated Pavement Tests 15
3.4 Tables for Analysis of In-Service Pavement Sections 23
3.5 Final Database 41

hours for each of the positions in the research management structure presented in Figure 5 and
for programming/engineering assistance. Task 3 is estimated to require at total of 876 man-
hours of effort. This is approximately 6 percent of the total effort required for the project.
B-43

Data
Laboratory Pavement
Subtask Principal Support Programmer/
Statistician Team Team
Investigator Team Engineer
Leader Leader
Leader
Develop Plan to Use NCHRP 9-30 20 0 10 10 80 0
Database
Develop Needed Tables 0 0 0 0 80 240
Input and Manage Data 0 0 0 0 40 396
Total 20 0 10 10 200 636
Task 3 Sources
Von Quintus, H.L., Schwartz, C., McQuen, R., and Andrei, D., “Experimental Plan for
Calibration and Validation of Hot-Mix Asphalt Performance Models for Mix and
Structural Design,” Final Report for National Cooperative Highway Research Program
Project 9-30, National Cooperative Highway Research Program, January, 2004.
Quarterly Reports for NCHRP Project 9-30A.
Task 4. Laboratory Studies

In Task 4 a series of laboratory experiments addressing critical aspects of the allowable strain
limit design procedure described earlier in Task 2 will be designed and executed. Table 14
summarizes the laboratory experiments that are needed.
Experiment 1 is a screening study to identify the mixture compositional factors that affect
healing and therefore, the allowable strain levels in HMA. The results from this experiment will
be used in the remaining experiments. Experiment 2 addresses a major assumption that was
made in developing the allowable strain limit procedure that was described in Task 2. In this
experiment healing rates will be determined using different strain levels. This experiment will
be conducted on mixtures from Experiment 1 that have high and low healing rates. Experiment
3 is a study to verify the applicability of time-temperature superposition to healing in HMA.
This was the second major assumption included in the development of the allowable strain limit
procedure described in Task 2. Experiment 3 will be conducted on a mixture from Experiment 1
that exhibits a moderate healing rate. Testing and analysis methods for determining allowable
strain limits that result in complete healing will be developed in Experiment 4. This experiment
B-44

will include testing and analysis of selected mixtures from Experiment 1 and mixtures used in
the endurance limit testing completed in NCHRP 9-38. This experiment will generate the Level
1 test procedure for use with the modified version of the MEPDG. In the last experiment,
Experiment 5, a wide range of mixtures will be tested using the methods developed in
Experiment 4 to develop predictive models relating the allowable strain limits to mixture
compositional factors. This last experiment will generate the relationships between allowable
strain and easily measured mixture compositional properties that will be used in the analysis of
the pavement sections in Task 5. These relationships will provide the Level 2 and 3 analysis for
the modified version of the MEPDG.
Table 14. Summary of Proposed Laboratory Experiments.
Experiment Topic Factors

1 Mixture Compositional Factors • Binder Type
Affecting Healing in HMA • Binder Age
• Air Voids
• Gradation
• Filler Content
2 Effect of Applied Strain on • Strain Level
Healing • Healing Rate From Experiment 1
3 Effect of Temperature and Rest • Temperature
Period Duration on Healing • Rest Period Duration
4 Development of Testing and • Healing Rate From Experiment 1
Analysis Procedures to Determine • Mixtures From NCHRP 9-38
Allowable Strain Levels
5 Estimation of Allowable Strain • Mix Compositional Factors Affecting
Levels from Mixture Composition Damage Accumulation
• Significant Factors From Experiment 1
• Temperature
• Rest Period Duration
For each experiment, detailed laboratory work plans will be prepared based on the experiment
descriptions and preliminary designs in this research plan and the results from completed
experiments. The experiments will then be executed and the resulting data analyzed. Pertinent
interim findings from the laboratory studies will be included in the quarterly progress reports.
The laboratory testing and analysis will be fully documented in the third interim report that will
B-45

be submitted at the end of the 22nd month of the project. The five experiments are described in
greater detail below.
Subtask 4.1 Experiment 1: Mixture Compositional Factors Affecting Healing in HMA

Experimental Design
Past studies of healing in HMA have assumed that only the properties of the binder affect the
healing characteristics of the mixture (7, 24, 34). Experiment 1 is a screening study that will use
an appropriate statistical design to verify or refute this assumption and to identify mixture
compositional factors affecting healing in HMA that should be included in Experiment 5.
Experiment 1 is based on a Plackett-Burman experimental design. This is a specific type of

partial factorial experiment that can simultaneously assess the effect of multiple factors with a
limited amount of testing. It is routinely used in ruggedness testing to quickly assess the effect
of a number of controllable test factors. ASTM E 1169 presents detailed information on the
design and analysis of Plackett-Burman experiments. Inherent to this type of statistical design is
the assumption that the effect of each of the factors on the result is independent. Therefore, the
observed effect resulting from simultaneous variation of several factors is simply the sum of the
individual effects. Since screening experiments are concerned with identifying significant
effects and not necessarily the form of the effect, each factor is evaluated at only two levels.
Replication is included in the experiment to estimate the variance of a single measurement.
A Plackett-Burman design with replication to simultaneously evaluate 7 factors requires only

16 tests, two for each of the specific combinations shown in Table 15. The seven factors are
designated by letters A through G. A “+” indicates high levels for the factors while a “-“
indicates low levels. Thus, determination 1 will be made with factors A, B, C and E at high
levels, and factors D, F, and G at low levels. The order of the tests should be randomized within
each replication of the experiment. ASTM E1169 describes designs for other numbers of factors.
B-46

Table 15. Design for a Two Level, Seven Factor Plackett-Burman Experiment.
Factor
Determination A B C D E F G
1 + + + - + - -
2 - + + + - + -
3 - - + + + - +
4 + - - + + + -
5 - + - - + + +
6 + - + - - + +
7 + + - + - - +
8 - - - - - - -
Selection of Factors
The selection of factors for Experiment 1 was based on a review of literature concerning
fatigue damage and healing in HMA. The factors are discussed individually below.
Binder Type. Several studies of fatigue and healing in HMA have shown that binder
properties affect the fatigue response of the mixture (35). The Shell fatigue equation is,
perhaps, the earliest example (36). It included the penetration index, which was an early
measure of the rheology of the binder. Research into healing that has been conducted at
the Texas Transportation Institute has shown that the properties of the binder affect
healing (34).
Less information is available on the effect of polymer modification on the fatigue and
healing characteristics HMA. Using continuum damage analysis, Lee, et al.demonstrated
better fatigue resistance for mixtures incorporating SBS modified binders (37). Recent
research on healing conducted at the University of Illinois using one neat and one
polymer modified binder showed the mixture with the polymer modified binder had
improved healing characteristics compared to the mixture with the neat binder (7). In
both of these studies, the neat and polymer modified binders were different grades.
Clearly, Experiment 5 will have to include a wide variety of binders, both neat and
modified, from different sources. In an attempt to better quantify the effect of polymer
B-47

modification on healing, Experiment 1 will use two binders from the same source having
the same performance grade, one neat and one modified with styrene butadiene styrene
(SBS). The recommended binders are neat PG 70-22 and a modified PG 70-22 produced
by adding SBS polymer to neat PG 58-28 binder obtained from the same refinery as the
neat PG 70-22.
Binder Aging. It is generally assumed by pavement and materials engineers that binder
aging has a detrimental effect on the fatigue life of asphalt mixtures. With this in mind, it
is interesting that only one study was identified where the effect of binder aging on
laboratory fatigue results was directly evaluated (38). In most laboratory fatigue studies,
unaged or short-term aged binders of different consistencies were used, and the results
generalized to describe the effect of mixture stiffness on fatigue life. The general
conclusions drawn from these studies that used relatively unaged binders are (35):
1. For continuous, controlled stress flexural testing, which is typically associated

with thick asphalt pavements, laboratory fatigue life increases with increasing
mixture stiffness.
2. For continuous, controlled strain flexural testing, which is typically associated
with thin asphalt pavements, laboratory fatigue life decreases with increasing
mixture stiffness.
3. When the results from either controlled stress or controlled strain flexural tests are
used in a mechanistic-empirical analysis of pavements with 6 or more inches of
asphalt, the predicted fatigue life increases with increasing mixture stiffness.
These conclusions imply that binder aging improves the fatigue life of pavements with
relatively thick asphalt layers. Because unaged and short-term aged binders were used in
these studies, the important effect of binder embittlement was not included in the
analysis. As asphalt binders age, they become, not only stiffer, but also more brittle due
to oxidation.
B-48

Recently, researchers at the Texas Transportation Institute performed controlled strain

flexural fatigue testing on compacted specimens from two mixtures that were aged for 0,
3, and 6 months at 60 °C (38). Three months of aging at 60 °C simulates 3 to 6 years of
field service for Texas conditions while 6 months of aging simulates 6 to 12 years of field
service (38). The loose mix for all specimens was short-term oven aged for 4 hours at
135 °C prior to compaction. Fatigue lives were 25 percent shorter for specimens aged for
three months, and 50 percent shorter for specimens aged for six months (38). The study
also included direct tension strength tests. In these tests strength increased while the
strain at failure decreased with increased aging, confirming that the mixtures become
stiffer and more brittle on aging (38). Aged, brittle mixtures would be expected to have
significantly poorer healing characteristics compared to unaged, ductile mixtures.
Short-term and long-term aged mixtures will be included in Experiment 1. The short-
term aging will be done for 4 hours at 135 °C as specified in AASHTO R30 for mixture
performance testing. The long-term aged specimens will be oven aged for 120 hours at
85 °C in accordance with AASHTO R30. Since the effects of aging are binder specific,
preliminary dynamic modulus and tensile strength tests should be conducted to ensure
that the selected binders exhibit significant stiffening and embrittlement as a result of the
laboratory, long-term aging process.
Effective Binder Content. Models for predicting the fatigue life of asphalt concrete
based on the results of continuous laboratory fatigue tests all indicate that fatigue life
increases as the mixture becomes increasingly rich in asphalt binder (39). These models
use either the effective volumetric binder content of the mixture, VBE, or voids filled
with asphalt, VFA, to indicate the richness of the mixture. Binder content effects have
not been included in past studies of healing in asphalt concrete.
It is reasonable to expect that richer mixtures may have improved healing characteristics,
resulting in improved fatigue lives, and higher allowable strains for complete healing.
Binder content will, therefore, be one of the factors included in Experiment 1.
Volumetric design procedures for asphalt mixtures set minimum limits for the effective
binder content of the mixture. These limits depend on the nominal maximum aggregate
B-49

size; increasing with decreasing nominal maximum aggregate size. Since this project is
concerned with fatigue cracking that initiates at the bottom of the asphalt layer, a typical
25 mm base course mixture will be used. The minimum effective binder content for 25
mm mixtures in AASHTO M323 is 8.0 percent by volume. The recommended
production tolerance for asphalt content in ASTM D 3535 is ± 0.5 percent by weight,
which is approximately ± 1 percent for the effective binder content by volume. These are
reasonable ranges for use in Experiment 1.
Air Voids. Nearly all laboratory fatigue studies have found the air void content of the
mixture to be a significant factor affecting mixture fatigue life (35, 39). Fatigue life
decreases with increasing air voids. It is reasonable to expect that air voids will also have
a significant effect on healing in asphalt concrete mixtures. Based on typical compaction
specifications, specimen air void contents of 4 and 8 percent will be included in
Experiment 1.
Design Compaction. An interesting finding in NCHRP Report 567, Volumetric

Requirements for Superpave Mix Design, is that the fatigue life of asphalt concrete
mixtures is significantly affected by the design compaction level; increasing as the design
gyration level increases (39).
Design compaction level was included in Experiment 1 to determine if healing properties

of asphalt mixtures are affected by the design compaction level. Considering the current
design compaction levels in AASHTO R35, the recommendations in NCHRP Report
573, Superpave Mix Design: Verifying the Gyration Levels in the Ndesign Table (40) and
approximate equivalencies between Marshall and gyratory compaction (39), design
gyration levels of 65 and 100 will be used in Experiment 1.
Gradation. The WesTrack project demonstrated that there is a difference in the fatigue
life of coarse-graded mixtures compared to fine-graded mixtures. Significantly more
cracking was observed in the coarse-graded mixture sections (41). Mixture gradation has
B-50

not been found to be a significant factor in fatigue models based on analysis of laboratory
test data.
As a result of the WesTrack experience, gradation was included in Experiment 1 to

determine if healing is different in coarse-graded compared to fine-graded mixtures. The
primary control sieve designation in AASHTO M323 will be used to distinguish between
coarse-graded and fine-graded mixtures. For 25 mm mixtures, the 4.75 mm sieve is the
primary control sieve and mixtures with less than 40 percent passing the 4.75 mm sieve
are considered coarse-graded.
Filler Content. Like aging, the effect of filler on the fatigue life of asphalt concrete has
not been systematically investigated. Currently, the influence of mineral filler on HMA
properties is being studied in NCHRP Project 9-45. The dust to binder ratio, defined as
the percent by weight passing the 0.075 mm sieve divided by the effective binder content
by weight of total mixture, is used in AASHTO M323 to control the filler content of
mixtures. A reasonable median value for the dust to binder ratio for design is 1.0. The
recommended production tolerance for the percent passing the 0.075 mm sieve in ASTM
D 3535 is ± 3.0 percent. This range is considered reasonable for Experiment 1.
Table 16 summarizes the factors and factor levels to be included in Experiment 1.

Experiment 1 requires the selection of a neat and polymer modified binder from the same source
and having the same performance grade, and the design of four 25 mm mixtures.
• 100 gyration coarse-graded

• 100 gyration fine-graded
• 65 gyration coarse-graded
• 65 gyration fine-graded
B-51

Table 16. Summary of Proposed Experiment 1.
Factor
%
Binder Air Passing
Determination Binder Aging Content Voids Ndesign 4.75 mm Filler
1 Polymer LTOA + 0.5 4.0 100 Coarse Low
2 Neat LTOA + 0.5 8.0 65 Fine Low
3 Neat STOA + 0.5 8.0 100 Coarse High
4 Polymer STOA - 0.5 8.0 100 Fine Low
5 Neat LTOA - 0.5 4.0 100 Fine High
6 Polymer STOA + 0.5 4.0 65 Fine High
7 Polymer LTOA - 0.5 8.0 65 Coarse High
8 Neat STOA - 0.5 4.0 65 Coarse Low
In designing these mixtures, the target effective binder content for all mixtures should be kept
constant at approximately 8.5 percent by volume, which will result in design voids in the mineral
aggregate (VMA) of 12.5 percent. The design dust to binder ratio should also be kept constant
for the four mixtures at approximately 1.0. These binder selection and mixture design
requirements will eliminate major interactions between the factors. During binder selection,
preliminary dynamic modulus and tensile strength tests should be conducted on specimens after
short- and long-term aging to ensure that the selected binders exhibit significant stiffening and
embrittlement as a result of the long-term aging.
The factor levels for binder content and filler will be obtained by making the appropriate
adjustment to the design mixture during batching. The factor levels for aging will be obtained by
appropriately aging the loose mixture and, for long-term aging, the test specimen. Finally, the
factor levels for air voids will be obtained by compacting specimens to the height needed to
achieve the target air voids.
Replicate tests for each determination in Table 16 will be made. This results in a total of 16
healing tests for Experiment 1.
B-52

Test Procedure
The objective of Experiment 1 is to identify the mixture compositional factors that affect
healing in asphalt concrete. To evaluate healing, a pulsed, strain controlled fatigue test must be
used. Either direct tension or flexural beam fatigue tests may be used, but the loading must be
such that a rest period is included after each load pulse. Figure 12 presents a schematic of the
required loading. The amount of healing that occurs will be evaluated by conducting fatigue
tests at 20 °C using two rest periods: 0 sec (continuous loading), and 3 sec. The modulus of the
specimen will be recorded for each load pulse. For each test, the accumulated damage in the
specimen will be determined from the ratio of the current modulus to the initial modulus. Figure
13 presents a schematic of the expected results when significant healing occurs.
Target Strain
Strain
Rest Period
0.1 sec Pulse
Time
Figure 12. Schematic of Pulsed, Strain Controlled Fatigue Loading.
B-53

1.0
0.9
Continuous
0.8 3 sec Rest Period
0.7
Damage (1-E*/E*I)
0.6
0.5
0.4
Healing
0.3
0.2
0.1
0.0
0 2000 4000 6000 8000 10000 12000
Cycles
Figure 13. Expected Results When Healing is Significant.
The same strain level must be used for all specimens tested in Experiment 1. The strain level
should be selected to produce a high degree of damage, approximately 30 to 40 percent, in the
specimens after 10,000 cycles when tested with continuous loading. Fifty percent damage is
typically used as the failure criterion for controlled strain tests. A maximum of 10,000 cycles
was selected because tests using the 3 sec rest period will require approximately 8.6 hours to
complete. Selection of an appropriate strain level will require some initial trial and error testing
with selected combinations. For example, the combination of factors used in Determination 6 in
Table 16 (polymer modified binder, short-term aging, high binder content, low air voids) would
be expected to give low amounts of damage during the testing. On the other had, the
combination of factors used in Determination 7 (polymer modified binder, long-term aging, low
binder content, high air voids) would be expected to give high amounts of damage during the
testing. Initial testing with these combinations at various strain levels will be needed to select an
appropriate strain level for the testing.
B-54

Data Analysis
For Experiment 1, healing is defined as the difference in damage between continuous loading
and loading with 3 sec rest period at 30,000 cycles. Linear regression is an efficient method for
analyzing the resulting healing data. The healing can be fit to a linear model of the form:
Y = B0 + B1 X 1 + B2 X 2 + B3 X 3 + B4 X 4 + B5 X 5 + B6 X 6 + B7 X 7 + Error (17)
where:
Y = healing
Xi = seven factors included in the experiment
Bi = model coefficients
Error = model error
From this analysis, the statistical significance of the model coefficients can be used to determine
which factors affect healing in HMA. For statistically significant factors, the model coefficients
can be used to select appropriate factor levels to be used in other experiments. Combinations
yielding low, moderate, and high levels of healing in Experiment 1 will be used in Experiments
2, 3, and 4. Significant factors identified in Experiment 1 will be included in Experiment 5.
Subtask 4.2 Experiment 2: Effect of Applied Strain on Healing

Experimental Design
One of the major assumptions that was made in developing the allowable strain limit design
approach described in Task 2 is that healing in HMA is independent of the applied strain level.
Early healing research provided some data supporting this assumption, but the testing was not
specifically designed to evaluate the effect of strain level (24).
In Experiment 2, the healing tests described for Experiment 1 will be conducted using three
different strain levels. Two different mixtures from Experiment 1 will be used: one exhibiting a
high amount of healing and one exhibiting a low amount of healing. All tests will be conducted
at 20 °C. The strain level used in Experiment 1 will be the medium strain level for Experiment
2. Tests at higher and lower strain levels will be added to complete the factorial. In selecting the
high strain level, it is important that the strain be such that macro-cracking does not occur during
B-55

the tests. Three replicates will be tested for each mixture. The experimental design is
summarized in Table 17.
Table 17. Strain Level Experiment.
Mixture Strain Level Replicates

Low 3
Low Healing Medium 3
High 3
Low 3
High Healing Medium 3
High 3
Data Analysis
Analysis of variance will be used to analyze the data from Experiment 2. For each mixture a
one-way analysis of variance will be conducted. It is anticipated that this analysis will confirm
that healing in HMA is not significantly affected by the applied strain level, provided the strains
are low enough that macro-cracking does not occur.
Subtask 4.3 Experiment 3: Effect of Temperature and Rest Period Duration on Healing
Experimental Design
The second major assumption that was made in developing the allowable strain limit design
approach described in Task 2 is that time-temperature superposition can be applied to the rest
periods to account for the effect of varying temperatures. The objective of Experiment 3 is to
confirm that this assumption is valid. Previous research on healing clearly showed that healing
effects were greater at higher temperatures (24). It is reasonable to expect that time-temperature
superposition will apply to rest period effects as it does for many other aspects of asphalt
material response. It is well known that time-temperature superposition is valid for measures of
binder and mixture stiffness. Time-temperature superposition is also an integral part of the
continuum damage approach to fatigue analysis that has become popular with a number of
researchers (42, 43, 44).
B-56

In Experiment 3, the healing tests described for Experiment 1 will be conducted using a
factorial of temperatures and rest period duration. A single mixture from Experiment 1, one
exhibiting a moderate amount of healing, will be used. Two replicates will be tested for each
mixture. The experimental design is summarized in Table 18. In addition to the healing tests
outlined in Table 18, dynamic modulus tests will be performed on replicate specimens at the
temperatures and frequencies listed in Table 19 to determine time-temperature shift factors for
the mixture. The dynamic modulus testing will be performed in accordance with NCHRP 9-29:
PT1, Determining the Dynamic Modulus and Flow Number for Hot Mix Asphalt (HMA) Using
the Simple Performance Test System (45).
Table 18. Experimental Design for Experiment 3.
Mixture Temperature, C Rest Period, sec Replicates

4 0 2
4 0.1 2
4 1 2
4 10 2
10 0 2
10 0.1 2
10 1 2
10 10 2
20 0 2
Moderate 20 0.1 2
Healing 20 1 2
20 10 2
30 0 2
30 0.1 2
30 1 2
30 10 2
40 0 2
40 0.1 2
40 1 2
40 10 2
B-57

Table 19. Temperature and Frequency Combinations for Dynamic Modulus Tests.
Temperature, C Frequency, Hz
4 10
4 1
4 0.1
4 0.01
10 10
10 1
10 0.1
10 0.01
20 10
20 1
20 0.1
20 0.01
30 10
30 1
30 0.1
30 0.01
40 10
40 1
40 0.1
40 0.01
Data Analysis
The data analysis for Experiment 3 is somewhat more complicated than that for Experiments
1 and 2. First, time-temperature shift factors must be determined from the dynamic modulus
measurements. Then the time-temperature shift factors will be applied to the rest periods to shift
the measured healing data. If time-temperature superposition applies to the rest periods, then the
healing results will form a continuous function after shifting.
Dynamic Modulus Master Curve and Shift Factors

Equation 18 presents a modified version of the dynamic modulus master curve equation
included in the MEPDG that is appropriate for this analysis (46).
log E * = δ +
(log E * max
−δ ) (18)
β +γ log f r
1+ e
B-58

where:
⎮E*⎮ = dynamic modulus
⎮E*⎮max = limiting maximum modulus
δ, β, and γ = fitting parameters
A second order polynomial can be used to describe the time-temperature shift factors:
log[A(T )] = a1 (TR − T ) + a2 (TR − T ) 2 (19)
where:
A(T) = time-temperature factor
T = test temperature
TR = reference temperature (normally 20 °C)
a1, a2 = fitting coefficients
The reduced frequency in Equation 18 is given by:
log f r = log f + a1 (TR − T ) + a2 (TR − T ) 2 (20)
where:
fr = reduced frequency at the reference temperature
f = loading frequency at the test temperature
Substituting Equation 20 into Equation 18 yields the final form of the dynamic modulus master
curve equation.
log E * = δ +
(log E * max
−δ )
1+ e [
β +γ log f + a1 (TR −T ) + a2 (TR −T ) 2 ] (21)
B-59

The limiting maximum modulus, ⎮E*⎮max, in Equation 21 is estimated from mixture volumetric
properties using the Hirsch model (47) and a limiting binder modulus of 1 GPa (145,000 psi),
using Equations 22 and 23. Chirstensen and Anderson recommended 1 GPa as a reasonable
estimate of the glassy modulus for all asphalt binders (48).
⎡ ⎛ VMA ⎞ ⎛ VFA x VMA ⎞⎤ 1 − Pc (22)

| E* | max = Pc ⎢4,200,000⎜1 − ⎟ + 435,000⎜ ⎟⎥ +
⎣ ⎝ 100 ⎠ ⎝ 10,000 ⎠⎦ ⎡ ⎛ VMA ⎞ ⎤
⎢ ⎜1 − 100 ⎟ ⎥
⎢⎝ ⎠+ VMA
⎥
⎢ 4,200,000 435,000( VFA) ⎥
⎢ ⎥
⎣ ⎦
where
0.58
⎛ 435,000(VFA ) ⎞
⎜ 20 + ⎟
Pc = ⎝ VMA ⎠ (23)
0.58
⎛ 435,000(VFA) ⎞
650 + ⎜ ⎟
⎝ VMA ⎠
⏐E*⏐max = limiting maximum mixture dynamic modulus, psi

VMA = Voids in mineral aggregates, %
VFA = Voids filled with asphalt, %
Using the limiting maximum modulus estimated from the volumetric properties of the test
specimens, the fitting coefficients (δ, β, γ, a1, and a2) are determined by numerical optimization
of Equation 21 using the measured modulus data. The optimization can be performed using the
Solver function in Microsoft EXCEL®. This is done by setting up a spreadsheet to compute the
sum of the squared errors between the logarithm of the average measured dynamic moduli at
each temperature/frequency combination and the values predicted by Equation 21.
∑ error
2 n
1
(
= ∑ log Eˆ * − log E * i
i
)
2
(24)
where:
2
∑ error = sum of squared errors
B-60

n = number of temperature/frequency combinations used in the testing
log Ê * i = value predicted by Equation 20 for each
temperature/frequency combination
log E * i = logarithm of the average measured dynamic modulus for each
temperature/frequency combination.
The time-temperature shift factors are then determined from Equation 18 using the fitting
coefficients, a1 and a2, obtained from the numerical optimization.
Reduced Rest Period

Knowing the time-temperature shift factors from the dynamic modulus testing, the results of
the healing tests will be shifted according to Equation 25. If time-temperature superposition
applies to the rest periods, then the healing results will form a continuous function after shifting.
This is shown schematically in Figure 14.

where:
modulus testing.
B-61

100
90
80
70 4C
Healing, %
10 C
60
20 C
50
30 C
40 40 C
30
20
10
0
0.001 0.01 0.1 1 10 100 1000
Reduced Rest Period, sec
a. Original Data
100
90
80
70 4C
Healing, %
60 10 C
20 C
50
30 C
40
40 C
30
20
10
0
0.001 0.01 0.1 1 10 100 1000
Reduced Rest Period, sec
a. Shifted Data
Figure 14. Schematic of Time-Temperature Superposition Applied to Rest Periods.
B-62

Subtask 4.4 Experiment 4: Development of Testing and Analysis Procedures to

Determine Allowable Strain Levels
Possible Approaches
In Experiment 4, testing and analysis procedures for determining the allowable strain levels
will be developed. One approach, using flexural fatigue testing and the ratio of dissipated energy
change (RDEC) method was illustrated in the description of Task 2. A second approach based
on cyclic direct tension testing and continuum damage analysis is also possible. Brief
descriptions of these two approaches are presented below.
Ratio of Dissipated Energy Change

Recently a substantial amount of HMA fatigue research has been performed at the University
of Illinois (4, 5, 7). This research has concentrated on using the ratio of dissipated energy change
to describe the fatigue response of HMA. The basic premise of this research is that the change in
dissipated energy per cycle of loading is related to the growth of damage that occurs in HMA.
The dissipated energy for each cycle of loading is the area within the stress-strain hysteresis loop
generated for that cycle of loading. The ratio of dissipated energy change is defined as the
average change in dissipated energy between two cycles divided by the dissipated energy from
the first of the two cycles:
RDECa =
(DEa − DEb ) (26)
(b − a ) × DEa
where:
RDECa = ratio of dissipated energy change for cycle a
DEa = dissipated energy for cycle a
DEb = dissipated energy for cycle b
For a given mixture a plot of the ratio of dissipated energy change as a function of loading cycles
forms a broad “U” shape as shown in Figure 15. The ratio of dissipated energy change initially
decreases, then reaches a broad plateau, where a constant percentage of the input energy is being
converted to damage, then finally increases as the sample begins to fail. Because of the high
variability of the cyclic dissipated energy measurements due to the small amount of energy
B-63

dissipated in each cycle, statistical methods were developed to determine the plateau value (5).
Lower plateau values imply lower damage per cycle. The plateau value for a given mixture
depends on the mixture properties, the applied strain level, and the duration of rest periods.
Plateau values decrease with decreasing applied strain and increasing rest period duration (7).
The effect of mixture properties on the plateau value is not clear from the research that has been
completed to date. From tests on a number of mixtures, the University of Illinois researchers
also found a unique relationship between the plateau value and number of cycles to 50 percent
reduction in stiffness, the traditional definition of failure in constant strain fatigue tests (7).
Figure 15. Typical Plot of Ratio of Dissipated Energy Change Versus Loading Cycles (6).
PV = 0.4429 × N f ( )−1.1102 (27)
where:
PV = plateau value
B-64

The University of Illinois research further found that an HMA mixture will exhibit endurance
limit behavior when the plateau value is 6.74x10-9 or less, which based on Equation 27
corresponds to a traditional fatigue life of 1.1x107 cycles or greater.
The testing and analysis required to use the ratio of dissipated energy change to establish
allowable strain limits for complete healing is summarized below:
1. Conduct dynamic modulus tests on the mixture and develop a dynamic modulus
master curve and associated time-temperature shift factors.
2. Conduct continuous loading, controlled strain flexural fatigue tests at 20 °C using

different strain levels to develop a relationship between the plateau value and the
applied strain (Equations 4 and 5 in Task 2).
3. Conduct pulsed, controlled strain flexural fatigue tests at a moderate strain level using
various temperatures and rest periods to determine a relationship between the plateau
value and reduced rest period (Equations 2 and 3 in Task 2).
4. Combine the relationships from Steps 2 and 3 to form a relationship for the plateau
value as a function of applied strain level and reduced rest period (Equations 6 and 7
in Task 2).
5. Substitute the relationship from Step 4 into the unique plateau value – number of
cycles to 50 percent stiffness reduction relationship (Equation 27) established by the
University of Illinois research (Equations 8 and 9 in Task 2).
6. Solve the equation developed in Step 5 for the allowable strain level for full healing
by substituting a value greater than 1.1 x107 for the number of cycles to 50 percent
stiffness reduction (Equations 10 and 11 in Task 2).
B-65

Continuum Damage Analysis

Continuum damage analysis has recently been introduced as a rapid method for characterizing
fatigue properties of HMA (44). Pioneering work in the application of continuum damage
analysis to HMA was performed at the North Carolina State University (42). Since its
introduction, continuum damage analysis has been used by several researchers in the United
States and abroad. The analysis is usually applied to the results of direct tension cyclic fatigue
tests or monotonic direct tension tests, although an approximate solution has been developed for
use with flexural fatigue tests (44).
Continuum damage analysis models the decay of the modulus of the mixture with increasing
load cycles. Figure 16 shows typical cyclic direct tension data. In traditional continuum damage
analysis, the curves for different strain levels and temperatures are collapsed into a unique
relationship by introducing an internal state variable, S, to represent the current damage in the
material. The internal state variable is difficult for many practicing engineers to understand and
can only be computed using approximate, numerical integration. Additionally, traditional
continuum damage analysis assumes that even very small levels of strain induce damage in the
material, implying that asphalt concrete does not exhibit endurance limit behavior. Recently
Christensen and Bonaquist, simplified continuum damage analysis and included the direct
consideration of the endurance limit (49). This improved analysis uses the concept of reduced
cycles defined by Equation 28 to collapse the data shown in Figure 16 into a unique relationship.
The endurance limit of asphalt concrete is accounted for using the concept of effective strain.
Effective strain is defined as applied strain minus the endurance limit. This innovation in
continuum damage analysis allows for the calculation of endurance limits from relatively limited
fatigue data.
2α 2α
⎛ f ⎞⎛ E * LVE ⎞ ⎛ε E ⎞ ⎡ 1 ⎤
N R = N R −ini + N ⎜⎜ 0 ⎟⎟⎜ ⎟ ⎜ E
⎜ε ⎟ ⎢
⎟ a(T / T ) ⎥ (28)
⎝ f ⎠⎜⎝ E * LVE / 0 ⎟
⎠ ⎝ 0 ⎠ ⎣ 0 ⎦
Where
NR = reduced cycles
NR-ini = initial value of reduced cycles, prior to the selected loading period
B-66

N = actual loading cycles

F0 = reference frequency (10 Hz suggested)
f = actual test frequency
|E*|LVE = undamaged (linear viscoealstic or LVE) dynamic modulus under given
conditions, lb/in2
|E*|LVE/0 = reference initial (LVE) dynamic modulus, lb/in2 (the LVE modulus at
20°C is suggested)
α = continuum damage material constant with a typical value of about 2.0
ε E = effective applied strain level = applied strain minus the endurance limit strain
ε 0E = reference effective strain level (0.0002 suggested)
a(T/T0) = shift factor at test temperature T relative to reference temperature T0
1.0
0.9
0.8
0.7
C(N) =E*/E*Initial
0.6
0.5
0.4 20 C, High Strain

20 C Low Strain
0.3 40 C, High Strain
40 C, Low Strain
0.2
0.1
0.0
0 20000 40000 60000 80000 100000 120000
Cycles
Figure 16. Typical Damage Ratio Curves From a Cyclic Direct Tension Fatigue Test.
Analysis of uniaxial fatigue data using the reduced cycles approach is done using the
following procedure.
B-67

1. Select the reference conditions. The suggested reference strain is 0.000200, peak-to-peak.
The recommended reference temperature is 68 °F (20°C). The reference modulus should
be the undamaged dynamic modulus or linear viscoelastic LVE modulus at 68 °F (20°C).
The reference frequency should be 10 Hz—the same as the most commonly used test
frequency for modulus and fatigue testing of asphalt concrete mixtures.
2. Perform dynamic modulus master curve testing on two samples to determine time-
temperature shift factors for the mixture.
3. Test a total of four to eight specimens, two to four at both 39.2 °F (4°C) and 68 °F
(20°C). Other temperatures may be used if desired, but temperatures much higher or
lower than these might prove difficult to test using the procedures given here. At each
test temperature, the specimens should be tested at different strain levels for each test.
4. Set up a spreadsheet to compute the damage ratio, C, and the reduced cycles for each test.
The damage ratio is given by Equation 29:
E*n
C= (29)
E * LVE
where:
C = damage ratio
|E*|n = damaged modulus at cycle n
|E*|LVE = undamaged (linear viscoealstic or LVE) dynamic modulus
Reduced cycles are calculated using Equation 28 and value of 2.00 for the continuum
damage constant α and an endurance limit strain of zero. Variation in the applied strain
during the test can be accounted for by splitting the data up into a number of segments,
calculating reduced cycles for each segment, and adding this value to the initial value
calculated at the end of the previous segment.
B-68

The LVE modulus can be estimated by visual examination of a plot of |E*| as a function
of loading cycles at the lowest strain level tested. The LVE modulus should be within a
few percent of the maximum observed value.
In some tests, macro damage (“localization”) might occur, which means that data beyond
this point is not valid for analysis using continuum damage methods. Macro damage is
indicated when there is a sudden drop in the modulus, or if modulus values suddenly
become erratic, rather than decreasing smoothly. Data after macro damage has occurred
should be eliminated from the analysis.
5. Fit Equation 30 to the C versus NR data.
1
C= (30)
1 + (N R K1 )K 2
where
K1 = cycles to 50 % damage = the fatigue half-life
K2 = fitting parameter
Linear regression can be used for the fitting by performing a logarithmic transformation
of Equation 30 to produce:
⎛1 ⎞
ln⎜ − 1⎟ = A + B ln N R (31)
⎝C ⎠
where:
A = -K2(lnK1)
B = K2
A problem in practical application of this approach is that because of noise in the

experimental data at low strains, the measured modulus can approach the LVE, resulting
in very noisy data when it is transformed using Equation 31. For this reason, a weighted
least squares approach to linear regression should be used, with a weight of N R 0.5 . This
B-69

approach gives very little weight to data points representing little or no damage, while
giving relatively more weight to data points associated with more heavily damaged states.
This prevents noisy data collected at low temperatures and/or low strains from skewing
the function relating C and NR, and also results in a more ideal distribution of the
residuals.
6. Keeping the value of α at 2.00, adjust the endurance limit strain for the data at 68 °F
(20°C) until the R2 value for the regression is maximized. Then adjust the endurance limit
strain value for the data at 39.2 °F (4°C), again, until the R2 value for the regression is
maximized.
Although it is possible to vary the value of α, it has been found that excellent
convergence of the data is generally possible while keeping α at 2.00 for all asphalt
concrete mixtures tested to date using this procedure. However, if the steps above do not
result in complete convergence, it might be necessary to vary the assumed value of α.
Figure 17 presents a typical fatigue damage curve developed using the procedure described
above.
1.0
|E*|/|E*|-LVE
0.8 20 C low strain

20 C high strain
0.6
4 C low strain
0.4 4 C high strain
Fit
0.2
1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
Reduced Cycles
Figure 17. Typical Damage Relationship From Continuum Damage Analysis.
B-70

Continuum damage analysis has not been applied to pulsed fatigue tests where intermittent
healing is permitted to occur. It is expected that the endurance limit will increase as the duration
of rest period increases. The testing and analysis required to use continuum damage analysis to
establish allowable strain limits for complete healing is summarized below:
1. Conduct dynamic modulus tests on the mixture and develop a dynamic modulus
master curve and associated time-temperature shift factors.
2. Conduct cyclic direct tension controlled strain fatigue tests using various
temperatures, strain levels, and rest periods.
3. Perform continuum damage analysis and determine the endurance limit for each of
the test conditions.
4. Develop a relationship of the endurance limit as a function of temperature and rest

period using time-temperature superposition if appropriate.
5. The endurance limit relationship developed in Step 4 is the allowable strain for full
healing.
It should be noted that the allowable strains from the continuum damage analysis will likely
be lower than the allowable strains developed using flexural fatigue testing and the RDEC
method. The reason is the endurance limit in the continuum damage analysis is defined as the
strain below which no measurable damage occurs in the mixture. The endurance limit in the
RDEC approach is defined as the strain that results in less than a 50 percent reduction in the
modulus of the material after an infinite number of loading cycles.
Experimental Design and Data Analysis

The two approaches are very similar. In both cases the rate of damage accumulation should
depend on the HMA properties, the applied strain level, the temperature, and the duration of rest
periods. The allowable strain limit for design is the strain level for specific temperatures and rest
B-71

periods where no damage accumulates in the HMA. The primary issue for both approaches is
determining the testing conditions that provide for an efficient and robust analysis. This
includes:
• Strain levels,
• Test temperatures,
• Duration of rest periods,
• Number of replicates.
The results of Experiments 1, 2, and 3 will provide initial estimates for the testing conditions.
Data will then be collected on two mixtures from Experiment 1, one exhibiting a low healing rate
and one exhibiting a high healing rate using a wider than estimated range and more intervals for
each of the testing conditions. The analysis will then be repeated using a reduced data set to
determine the optimum testing conditions. Tests using the optimum testing conditions will then
be conducted on selected mixtures from NCHRP 9-38 and the results will be compared to the
endurance limit strain levels determined in NCHRP 9-38.
Subtask 4.5 Experiment 5: Estimation of Allowable Strain Levels from Mixture

Composition
The final experiment that will be conducted is one to establish a predictive model to estimate
allowable strain levels from mixture composition. This is an extremely important experiment for
two reasons. First, it is unlikely that original materials or appropriate size field specimens will
be available from the calibration pavement sections; therefore, estimates of allowable strain
levels will be needed for the Task 5 analyses. Using the models developed in Experiment 5,
estimates of allowable strain levels can be made using test data from standard tests on a small
number of cores removed from the pavement sections. Second, a method of estimating
allowable strain levels will also be needed for use in Level 2 and 3 design with the modified
version of the MEPDG. The testing and analysis procedure developed in Experiment 4 will
provide methods for Level 1 analysis. The predictive model developed in Experiment 5 will
provide relationships for Level 2 and Level 3 analyses.
B-72

Experiment Design
Regression analysis will be used to develop a predictive model to estimate allowable strain
levels from mixture composition. In Experiment 5 a database of allowable strains and mixture
properties will be assembled by performing the analysis developed in Experiment 4 on a
representative sample of HMA base course mixtures. Since it is envisioned that the model will
be used for both analysis of existing pavements and the design of future pavements, the mixtures
tested should include past, current, and likely future features that affect HMA fatigue response
and healing. For example the base course of many existing pavements was designed using
Marshall compaction resulting in somewhat richer mixtures than designed today using gyratory
compaction. If healing is found to be much greater in modified binders, then it may be likely
that modified binders will be considered for base courses in the future, an uncommon practice
today.
Guidance on the factors and their ranges to be included in Experiment 5 will be obtained from
Experiment 1. As discussed previously, the following factors have been identified as potentially
affecting the allowable strain levels:
• Binder grade
• Binder modification
• Aging
• Air Voids
• Gradation
• Filler Content
The purpose of Experiment 1 is to narrow this list to the factors that significantly affect the
fatigue damage and healing characteristics of HMA. The results of Experiment 1 and a review
of past and current mixture design and mixture production specifications will be used to
determine the specific factors and the ranges that must be included in Experiment 5. It is
envisioned that approximately 30 mixtures will be tested in Experiment 5. It is not necessary
B-73

that Experiment 5 be a full or partial factorial design. The major experimental design
requirements are that (1) the mixtures that are selected to be representative of base courses (2)
they span the desired range of each important factor, and (3) at least three levels are included for
each factor so that non-linear analyses can be made.
Data Analysis
The database of allowable strains and associated mixture compositional properties will be
analyzed using graphical and regression techniques. First scatter plots will be prepared for each
of the factors included in the experiment to determine appropriate mathematical functions for the
model. At this point consideration will be given to using a more general factor that combines
some of the individual factors. For example, the effects of binder grade and aging could both be
addressed using the rheological index obtained from a binder master curve. Or the effects of air
voids and effective binder content could both be addressed using the voids in the mineral
aggregate or voids filled with asphalt. Additionally, consideration will be given to using easily
measured or estimated mechanical properties such as indirect tensile strength or modulus.
Once appropriate model forms have been identified using graphical analysis, a regression
analysis will be performed to determine the model coefficients. Most likely the relationships
will be non-linear resulting in the need to use numerical optimization. Several statistical
packages are available for performing non-linear regression analyses.
The final step in the process, which is often overlooked, is to evaluate the appropriateness of
the model. There are several analyses that must be performed to evaluate the model including:
1. Goodness of Fit. Two measures of the goodness of fit of the model should be evaluated.
The first is the square of the correlation coefficient, R2, which is the percentage of the
variance of the criterion variable explained by the predictor variables. The second
measure of the goodness of fit of the model is the standard error of estimate, Se, which is
the standard deviation of the errors. The standard error of estimate has the same units as
the criterion, and its magnitude is a direct indicator of the model errors. If the model
B-74

provides a good prediction, the standard error of estimate should be much lower than the
standard deviation of the data used to fit the model.
2. Statistical Significance of the Predictor Variables. Only statistically significant

predictor variables should be included in the model. If predictor variables that are not
statistically significant are included, then irrational effects may be predicted for important
predictor variables. The standard error of the parameter estimates should be used in a t-
test to determine if each of the model parameters is significantly different from zero.
3. Residual Analysis. An analysis of the residuals or errors should always be performed to

ensure that the underlying assumptions of regression analysis are not violated by the
model. The model errors should (1) be independent, (2) have zero mean, (3) have a
constant variance across all predictor variables, and (4) be normally distributed. Plots of
the residuals as a function of the predictor variables should be used to identify bias in the
model and to identify potential violations of the underlying regression assumptions.
4. Reliability of the Model. Confidence intervals should be constructed to assess the

reliability of the model. Since the model will be used to predict properties for design and
analysis, the width of prediction intervals for the model are of primary concern. The
prediction interval is the confidence interval associated with the prediction of a future
value.
Task 4 Milestones
Table 20 summarizes the major milestones for Task 4. These are all associated with the
design, execution, and analysis of the five laboratory experiments.
B-75


Contract Award
4.1 Select Analysis Approach and Prepare Detailed Work Plan for 5
Experiment 1
4.2 Complete Experiment 1 8
4.3 Detailed Work Plan for Experiments 2 8
4.5 Detailed Work Plan for Experiment 3 and Experiment 4 10
4.8 Detailed Work Plan for Experiment 5 13

hours for each of the positions in the research management structure presented in Table 5 and for
laboratory technicians. Task 4 is estimated to require a total of 3,893 man-hours of effort. This
is approximately 30 percent of the total effort required for the project.
Data
Laboratory Pavement
Subtask Principal Support
Statistician Team Team
Investigator Team Technicians
Leader Leader
Leader
Experiment 1: Mixture 4 4 34 0 0 388
Compositional Factors Affecting
Healing
Experiment 2: Effect of Applied 4 4 24 0 0 214
Strain on Healing
Experiment 3: Effect of Temperature 8 4 57 0 0 242
and Rest Period Duration on Healing
Experiment 4: Testing and Analysis 54 16 98 0 0 392
Procedures for Allowable Strain
Levels
Experiment 5: Estimation of 146 40 270 0 0 1890
Allowable Strain Levels from
Mixture Composition
Total 216 68 483 0 0 3126
B-76

Task 4 Sources
HMA Fatigue
Christensen, D.W., and Bonaquist, R.F., “Volumetric Requirements for Superpave Mix
Design,” NCHRP Report 567, National Cooperative Highway Research Program,
Rao Tangella, S.C.S., Craus, J., Deacon, J.A., and Monismith, C.L., “Summary Report on
Fatigue Response of Asphalt Mixtures,” Report Number SHRP-A/IR-90-011, Strategic
Shook, J. F., F. N. Finn, M. W. Witczak, and C. L. Monismith, “Thickness Design of

Asphalt Pavements—The Asphalt Institute Method,” Proceedings, Fifth International
Conference on the Structural Design of Asphalt Pavements, Vol. 1, The University of
Michigan and The Delft University of Technology, August 1982
University of California, Berkeley, Asphalt Research Program, Institute of Transportation

Studies, Fatigue Response of Asphalt-Aggregate Mixes, Report SHRP-404, Washington,
D.C.: National Research Council, 1994.
Walubita, L.F., Epps-Martin, A., Jung, S. H., Glover, C. J., Park, E.S., Chowdhury, A.,
and Lytton, R. L., “Comparison of Fatigue Analysis Approaches for Two Hot Mix
Asphalt Concrete (HMAC) Mixtures, Report Number FHWA/TX-05/0-4468-2, Texas
Department of Transportation, Austin, TX, August, 2005.
Healing in HMA
Influence of Rest Periods on the Fatigue Response of Bituminous Mixes,” Proceedings,
Kim, B. and Roque, R., “Evaluation of Healing Property of Asphalt Mixtures,”

Transportation Research Record No. 1970, Transportation Research Board,
Kim, Y.R., Little, D.N., and Benson, F.C., “Chemical and Mechanical Evaluation of
Healing of Asphalt Concrete, Journal of the Association of Asphalt Paving
Little, D. N., Lytton, R. L., Williams, D., and Chen, C. W., “Microdamage Healing in
Asphalt and Asphalt Concrete, Volume I: Microdamage and Microdamage Healing
Project Summary Report,” Report Number FHWA-RD-98-141, Federal Highway
Administration, Washington, D.C., June 2001.
B-77

Pronk, A.C., “Partial Healing, “A New Approach for the Damage Process During Fatigue
Testing of Asphalt Specimens,” Asphalt Concrete Simulation, Modeling, and
Experimental Characterization, Geotechnical Special Publication No. 146, American
Society of Civil Engineers, Reston, VA, 2005.
Ratio of Dissipated Energy Change

Carpenter, S.H., Ghuzlan, K.A., and Shen, S., “Fatigue Endurance Limit for Highway
and Airport Pavements,” Transportation Research Record No. 1832, Transportation
Carpenter, S.H., and Shen, S., “Dissipated Energy Approach to Study Hot-Mix Asphalt
Healing in Fatigue,” Transportation Research Record No. 1970, Transportation
Prowell, B., Brown, E., R., Daniel, J., Bhattacharjee, S., Von Quintus, H., Carpenter, S.,
Shen, S., Anderson, M., Swamy, A. K., and Maghsoodloo, S., “Endurance Limit of Hot
Mix Asphalt Mixtures to Prevent Fatigue Cracking in Flexible Pavements,“ Updated
Draft Final Report, NCHRP 9-38, National Cooperative Highway Research Program,
Washington, D.C., May, 2008.
Continuum Damage Analysis

Christensen, D.W., and Bonaquist, R.F., “Practical Application of Continuum Damage
Theory to Fatigue Phenomena in Asphalt Concrete Mixtures,” Journal of the
Association of Asphalt Paving Technologists, Vol. 74, 2005
Kim, Y. R. and Little, D. N. “One Dimensional Constitutive Modeling of Asphalt

Concrete,” ASCE Journal of Engineering Mechanics, Vol. 116, No. 4, 1990
Kim, Y. R., Little, D. N., and Lytton, R. L., “Use of Dynamic Mechanical Analysis
(DMA) to Evaluate the Fatigue and Healing Potential of Asphalt Binders in Sand Asphalt
Mixtures,” Journal of the Association of Asphalt Paving Technologists, Vol. 71, 2002
Kim, Y.R., Lee, H. J., and Little, D. N ., “Fatigue Characterization of Asphalt Concrete
Using Viscoelasticity and Continuum Damage Theory,” Journal of the Association of
Lee, H. J., and Kim, Y. R. “A Uniaxial Viscoelastic Constitutive Model for Asphalt
Concrete Under cyclic Loading,” ASCE Journal of Engineering Mechanics, Vol. 124,
1998, No. 11.
B-78

Statistical Analysis
American Society for Testing Materials, “ASTM Designation E 1169-02: Standard Guide
for Conducting Ruggedness Tests,” Annual Book of ASTM Standards, Vol. 14.02.
Statistica for Windows. [Computer software]. StatSoft, Inc., Tulas, Oklahoma, 1994.
Task 5. Analysis of Pavement Sections

The final task in the HMA Endurance Limit Validation Study is an analysis of full-scale
pavement sections using the allowable strain limit design procedure formulated in Task 2 and
improved through the laboratory experiments in Task 4. Two types of full-scale pavement
sections will be analyzed. First data from selected accelerated pavement tests and test roads will
be used to test critical elements of the procedure. These include the effects of temperature,
applied strain, and material properties on the allowable strain levels. Results from these analyses
will be used to further improve the allowable strain limit design procedure for use in analysis of
the second type of full-scale pavement: in-service pavement sections. For the in-service
pavements, both cracked and uncracked pavements will be analyzed. These analyses will be
used to calibrate the procedure and serve as validation of the concept of an endurance limit for
flexible pavement design. It is important to recognize that the allowable strain limit design
procedure is not intended to be a tool for predicting the extent of bottom initiated cracking with
time and traffic like the MEPDG fatigue model. Its purpose is to identify design features that
minimize the possibility of bottom initiated fatigue cracking. Thus field calibration of the
allowable strain limit design procedure will be easier and likely more precise than the calibration
that was completed for the MEPDG fatigue model. Task 5 has been divided into five subtasks:
5.1 Review Data Sources and Select Sections for Analysis

5.2 Obtain Materials and Data for Accelerated Pavement Tests and Test Roads
5.3 Perform Testing and Analyze Accelerated Pavement Tests and Test Roads
5.4 Obtain Materials and Data for In-Service Pavement Sections
5.5 Perform Testing and Analyze In-Service Pavement Sections
Each of these subtasks are described in detail below.
B-79

Subtask 5.1 Review Data Sources and Select Sections for Analysis
In this subtask the sources identified in this research plan will be reviewed considering the
final preliminary approach developed in Subtask 2.2 and specific pavement sections will be
selected for subsequent analysis. Subtask 5.1 will begin immediately after the preliminary
approach is finalized in Subtask 2.2. Initial selection of sections for analysis will be documented
in the second interim report that will be submitted at the end of the 13th month of the project.
This initial selection will be reviewed as results from the laboratory experiments become
available and adjusted as needed. Two types of full-scale pavements: accelerated pavement tests
and test roads, and in-service pavements will be selected for analysis. The sections that follow
describe specific pavement sections that are recommended for consideration in Task 5.
Accelerated Pavement Tests and Test Roads

Selected, well documented accelerated pavements tests and test roads will serve the important
role of verifying critical aspects of the allowable strain limit design procedure. Specific elements
of the procedure that can be verified include:
1. The overall engineering reasonableness of the approach,

2. Applicability of time-temperature superposition to healing and allowable strains,
3. Independence of healing on applied strain, and
4. Effect of material properties on allowable strains.
Although there are now a number of accelerated pavement testing devices and test road
facilities in the United States, few of the testing programs have addressed fatigue of HMA in a
structured manner. For flexible pavements, accelerated pavement testing has mostly been used
to investigate rutting in HMA surfaces, or to evaluate specific materials or design features. Only
four projects were identified where structured, full-scale testing was conducted that is useful in
verifying the above aspects of the allowable strain limit design procedure. The following
projects are recommended for analysis:
B-80

• Fatigue tests conducted during the Superpave validation study at the FHWA
Pavement Test Facility (50).
• Sections at the NCAT Test Track that have remained in service from the first cycle
through the current cycle (51).
• Sections from the WesTrack experiment containing mixtures with different
composition (41).
• Sections from the structural design experiment performed at the NCAT Test Track
(52, 53).
• Selected sections from the MNRoad project (54).
Although the MNRoad sections are actually in-service pavements loaded with normal traffic,
they are included in the verification studies because there are a number of sections that can be
analyzed and all of the sections are exposed to the same environmental conditions. If MNRoad
sections are included in the calibration, then only a limited number of sections can be used,
otherwise the analysis will be biased toward the environmental and construction conditions at
MNRoad. The sections that follow describe analyses that should be conducted considering the
preliminary design approach described in Task 2.
Overall Engineering Reasonableness

All of the accelerated pavement tests will be used to judge the engineering reasonableness of
the allowable strain limit design procedure. An analysis of each section using the procedure
should provide the correct conclusion concerning cracking in the pavement. For sections that
have cracked, the analysis should show that the allowable strain levels were exceeded. For
sections that have not yet cracked, such as the first cycle sections at the NCAT Test Track that
remain in-service, the analysis should show that the allowable strain levels were not exceeded.
It should be noted that the allowable strain limit design procedure developed in Task 2 does
not require the pavement to exhibit endurance limit behavior. Equations 8 and 9 in Task 2 can
be solved for the allowable strains for any number of loading cycles. Endurance limit behavior
occurs when the number of cycles to failure exceeds 1.1x107. This will be very useful for
analysis of the structural sections at the NCAT test track. Table 22 presents the HMA
B-81

thicknesses in the NCAT structural sections (51). For the materials used in the base course of
these sections, analysis can be done assuming endurance limit behavior, then the analysis can be
repeated using the observed load cycles to failure and the allowable and actual strains can be
compared.
Table 22. HMA Thicknesses in NCAT Structural Sections

2003 Construction 2006 Construction
Section HMA Thickness, HMA Base Binder HMA Thickness, HMA Base Binder
in in
N1 5 Polymer 76-22 7 Neat PG 67-22
N2 5 Neat 67-22 7 Polymer 76-22
N3 9 Neat 67-22 NA NA
N4 9 Polymer 76-22 NA NA
N5 7 Polymer 76-22 7 Neat PG 67-22
N8 7 (rich bottom) Neat 67-22 10 Polymer 76-28
N9 NA NA 14 Polymer PG 76-28
N10 NA NA 8 Polymer PG 70-22
The MNRoad sections also provide the opportunity to perform a systematic analysis of the
overall reasonableness of the approach for pavements of different thickness and composition
exposed to the same traffic and environment. At MNRoad, sections were constructed using
different thicknesses, design compaction levels, and binders. Table 23 summarizes the main line
HMA cells at MNRoad that could be used in the verification analyses (54). Although bottom
initiated fatigue cracking was not reported as a distress for any of the HMA sections in the last
condition report (55), the pavements have received seven years of additional traffic and selected
sections will remain in service after reconstruction is completed in 2008 and 2009 (54).
B-82

Table 23. Summary of MNRoad Mainline HMA Pavement Sections.
Section HMA Thickness, in Design Compaction Binder

1 6.0 75 Blow Marshall PG 58-28
4 9.1 Gyratory PG 58-28
16 8.0 Gyratory PG 64-22
Applicability of Time-Temperature Superposition to Rest Periods

The fatigue experiment that was conducted during the Superpave validation study at the
FHWA Pavement Testing Facility provides an excellent opportunity to validate that application
of time-temperature superposition to rest periods. In this study, accelerated pavement tests were
conducted with the FHWA Accelerated Loading Facility on two pavements at three different
pavement temperatures. The tests were performed when ambient air temperatures were low. An
infrared heating system was used to maintain the pavement temperatures (50). Table 24
summarizes the tests that were performed. Analysis of these tests at different temperatures using
the allowable strain limit design procedure will provide validation of the use of time-temperature
superposition to model HMA healing effects.
Table 24. FHWA Pavement Testing Facility Superpave Fatigue Experiment.
HMA Thickness, mm Binder Load, kN 10 °C 19 °C 28 °C

AC-5 53 X X X
100
AC-20 53 X X X
AC-5 53 X X X
200
AC-20 53 X X X
B-83

The instrumented structural sections at the NCAT Test Track can be used to evaluate the
effect of damage and healing during different temperature conditions. Measured strains and
deflections in these sections can be used to determine the effects of rest periods on healing at
different temperatures. Within the current loading experiment, four of the structural test sections
are instrumented.
Independence of Healing on Strain Level

The FHWA Superpave validation study fatigue experiments also provide the opportunity to
verify that healing is independent of strain level. Since the same mixtures were tested at the
same temperature and load in two different pavement structures, the effect of strain level on
healing can be evaluated. The thicker pavement has significantly lower tensile strains at the
bottom of the HMA compared to the thinner pavement. The structural sections at the NCAT
Test Track and sections at MNRoad where the same base course material was used in pavements
of different thicknesses can also be used to verify that healing is independent of strain level.
Effect of Material Properties on Allowable Strains

All four recommended projects can be used to assess how well the allowable strain limit
design procedure addresses the effect of changes in mixture composition. The WesTrack
experiment included variations in gradation, filler content, binder content, and in-place density
(41). A single asphalt binder and aggregate source were used in the original sections. In the
replacement sections a different aggregate was used (41). As shown in Table 22, the structural
sections at the NCAT Test Track includes pavements of the same thickness made with a polymer
modified PG 76-22 binder and a neat PG 67-22 binder. The FHWA experiment included two
neat binders, AC-5 and AC-20. Finally as shown in Table 23, the MNRoad project includes
sections of the same thickness designed with different compaction and two different binders.
The predictive model developed in Experiment 5 of Task 4 addresses the effect of material
properties on allowable strains. The effects predicted by this model can be compared to the
observed effects within each of the experiments.
B-84

In-Service Pavement Sections

Calibration of the allowable strain limit design procedure will be performed using in-service
pavements. Analyses will be conducted for a number of sections, both cracked and uncracked,
using the procedure. Sections from the LTPP program (56) and pavements that have received
perpetual pavement awards from the Asphalt Pavement Alliance (57) were considered for use in
the calibration. The LTPP sections were selected because these sections have received extensive
monitoring over a number of years, and distress, deflection, and material property data are
available from the LTPP database (56). Since sufficient sections for the analysis are available
from the LTPP program, only these sections are included in this research plan.
LTPP Sections
In NCHRP Project 9-38, analyses were conducted using data from the LTPP database to
determine if an endurance limit for HMA could be identified from field data (6). The following
assumptions were made in these analyses:
1. Alligator cracking reported in the LTPP database initiated at the bottom of the section.
2. Wheel path longitudinal cracking reported in the LTPP database initiated at the surface.
3. The endurance limit can be defined by a single value of strain that is independent of
temperature, mixture modulus, and type of mixture.
From these analyses, an endurance limit could not be definitively identified. The NCHRP 9-38
research team hypothesized that one of the reasons why an endurance limit could not be defined
is that the endurance limit is mixture composition dependent and it varies with temperature.
Figures 18 and 19 compare the amount of fatigue cracking (percent of wheel path area) from
the most recent LTPP distress survey with HMA thickness and maximum tensile strain at the
bottom of the HMA, respectively. As shown and expected, the test sections with thinner HMA
layers and higher tensile strains generally exhibit more fatigue cracking.
B-85

Series1 Log. (Series1)
100
Area Fatigue Cracking, %
90
80
70
60
50
40
30
20
10
0
0 5 10 15 20 25 30
HMA Thickness, inches
Figure 18. Comparison of Area Fatigue Cracking (Area Alligator Cracking Based on a
Percent of Wheel Path Area) and HMA Layer Thickness (6).
Fatigue Cracking Log. (Fatigue Cracking)
100
90
Fatigue Cracking, %
80
70
60
50
40
30
20
10
0
10 100 1000 10000
Tensile Strain, micro-inches
Figure 19. Comparison of the Area Fatigue Cracking for and Maximum Tensile Strain
Computed at the Bottom of the HMA Layer (6).
A number of test sections with thick HMA layers and low tensile strains, however, have
levels of fatigue cracking exceeding 5 percent. Reasons given for the cracking in these sections
included (6):
B-86

• Misclassification of longitudinal cracking as alligator cracking.

• The presence of construction defects, such as high air voids, debonding of layers, etc.
• Moisture damage in the section,
• The endurance limit is dependent on the quality of the HMA base; therefore, sections
with poor HMA base quality require lower strains to exhibit endurance limit
behavior.
Forensic evaluation of the thick HMA sections with reported alligator cracking was
recommended for future endurance limit validation studies.
An observation of the data in Figures 18 and 19 that was not made by the NCHRP Project 9-
38 research team is the pavements in the LTPP database are generally properly designed to resist
fatigue cracking for the level of traffic that they have received. This is indicated by the large
number of sections having zero alligator cracking. This is particularly true for pavements having
maximum tensile strains at the bottom of the asphalt layer below about 100 microstrain when
calculated using the equivalent annual layer moduli for each pavement layer. Figure 20 presents
a plot of tensile strain at the bottom of the asphalt layer versus HMA layer thickness that was
used to develop Figure 19. From Figure 20 tensile strains of 100 microstrain correspond to
approximately 10 inches of HMA, which is similar to the thicknesses reported for the heavily
trafficked pavements in the United Kingdom having no evidence of bottom initiated fatigue
cracking (9). This observation suggests that the thick sections with high levels of alligator
cracking likely contain construction defects and should not be included in the calibration of the
allowable strain limit design procedure. Forensic evaluation of these sections should definitely
be conducted, but not as part of the HMA Endurance Limit Validation Study.
B-87

Series1 Power (Series1)

Tensile Strain, Micro-inches
1000
100
10
0 5 10 15 20 25 30
HMA Thickness, inches
Figure 20. Comparison of the Maximum Tensile Strain at the Bottom of the HMA Layer
and HMA Thickness (6).
Table 25 presents the preliminary test matrix for using LTPP sections to calibrate the
allowable strain limit design procedure. Since the procedure is not intended for prediction of the
extent of cracking in a pavement section, but rather as a tool to identify design features to
minimize the potential for bottom initiated fatigue cracking, an extremely large data set is not
required. The recommended matrix includes a total of 32 pavement sections: 16 not exhibiting
alligator cracking and 16 exhibiting low to moderate amounts of alligator cracking. An equal
number of sections from the four environmental zones are included in the matrix. Only
pavements with HMA thicknesses exceeding 8 inches are included. Subgrade deformation
becomes an important consideration in thinner HMA pavements.
B-88

Table 25. Preliminary Matrix for Field Calibration of the Allowable Strain Limit Design
Procedure.
No Low
HMA Thickness,
Environment Alligator Alligator
in
Cracking Cracking
8 to 12 2 2
Wet Freeze
>12 2 2
8 to 12 2 2
Wet No Freeze
>12 2 2
8 to 12 2 2
Dry Freeze
>12 2 2
8 to 12 2 2
Dry No Freeze
>12 2 2
Table 26 presents a summary of applicable LTPP sections for each of the cells in the
experimental matrix. Information from the LTPP database on these sections and others that may
be considered is presented in the attachment. Specific sections to be included in the calibration
effort will be selected in Subtask 5.1. Items that should be considered in the final selection
include:
• Current status of the section (active or out of service).

• Willingness of the state agency to assist with providing traffic control for distress
verification and seismic testing, and to provide limited coring to investigate cracking
and obtain samples for laboratory testing.
• Consistency of time series distress data for the section in the LTPP database.
• Consistency of time series deflection data for uncracked sections.
• Availability of traffic information or an estimate of traffic for the section.
B-89

Table 26. LTPP Sections Recommended for Consideration.
HMA Total Fatigue or Alligator Cracking

Climate
Thickness, in.
None Appreciable
12-0101; 12-0103; 22-0114; 01-0101; 05-0114; 12-0107;
8 to 12
40-0160 40-0114
Wet-No
Freeze 05-3071; 12-0106; 12-0104;
>12 01-0111; 05-0115; 05-0116
13-4113; 22-0116; 40-0115
8 to 12 35-0111; 35-0103; 35-0107 04-0162; 48-1070

Dry-No Freeze
>12 04-1065; 35-0106; 48-0116 04-1062; 04-0116
8 to 12 32-0101; 32-0105 16-9034; 30-0114; 32-0103

Dry-Freeze
31-0115; 31-0116; 32-0106;
>12 30-0116; 30-0115; 30-0124
32-0104
19-0101; 19-0105; 55-01114;
8 to 12 19-0103; 55-C960
55-C901
Wet-Freeze
19-0112; 26-0115; 39-0902;
>12 39-0106; 39-0112; 39-0903
55-0116
Subtask 5.2 Obtain Materials and Data for Accelerated Pavement Tests and Test
Roads
The primary activity required in Subtask 5.2 is extracting the data required for analysis of the
accelerated pavement tests and test road sections from various research reports. This includes
information on the pavement structure, loading, environmental conditions, material properties,
and distress for each section that will be analyzed. The data will be entered into the database and
managed in Subtask 3.2.
The inputs needed to apply the allowable strain limit design procedure to accelerated
pavement tests and test roads are similar to those required for current mechanistic-empirical
design, such as the MEPDG. Table 27 summarizes the required inputs. The elements in bold in
Table 27 are ones required by the allowable strain limit design procedure that are not included in
current mechanistic-empirical analysis. Since mechanistic-empirical pavement analyses were
included in the recommended projects, most of the information needed for the analyses are in
published reports for the projects or available from the project websites (41, 50, 51, 52, 53, 54,
58).
B-90

Table 27. Summary of Required Inputs for Allowable Strain Limit Design.
Category Required Input

Pavement Structure Layer thicknesses
Layer moduli
Layer Poisson’s ratios
Mixture composition and binder properties for HMA base
Traffic Axle configuration
Tire configuration
Tire loads
Tire pressure
Speed
Wander
Rest Period
Environmental Pavement temperature history
Base modulus history
Subgrade modulus history
It is envisioned that the model for predicting allowable strains in HMA developed in
Experiment 5 of Task 4 will relate allowable strains to mixture composition and binder
properties. The required mixture composition data are available in the published research
reports; however, it is expected that binder properties in addition to the performance grade of the
binder will be required. Extensive testing of the binders used in the FHWA Superpave validation
study, WesTrack, and MNRoad was completed during NCHRP Project 9-19 (59, 60, 61).
Therefore, the only material sampling and testing that will be needed for analysis of the
accelerated pavement tests and test roads will be characterization of the binders used in the
structural sections at the NCAT Test Track. One quart samples of these binders will be
requested from NCAT or the test section sponsors.
The required performance data for the recommended projects are included in published
reports. Updated information on performance of the MNRoad test sections is available by
request through the MNRoad website (54). Traffic loading for the 2006 sections included in the
structural sections at the NCAT Test Track is scheduled for completion in the Fall of 2008 (51).
B-91

Subtask 5.3 Perform Lab Testing and Analyze Accelerated Pavement Tests and Test
Roads
The only laboratory testing envisioned in Subtask 5.3 is further characterization of the binders
used in the structural test sections from the NCAT test track. It is unlikely that master curves
characterizing the flow characteristics of the binders over a wide temperature range and for
various aging conditions are available; therefore, they will have to be developed. Master curves
are developed by testing the binder at multiple temperatures and frequencies using the dynamic
shear rheometer, AASHTO T315, and conducting bending beam rheometer tests, AASHTO
T313, at multiple temperatures.
For each accelerated pavement test and test road section, an analysis will be performed with
the research version of the MEPDG software, NCHRP9-44A_Version 0.2, using section specific
material properties, loading, and environment. Two analyses will be performed. For all sections
an analysis will be conducted to determine the allowable strains that will produce endurance
limit behavior (full healing). Then, for those sections that have exhibited cracking an analysis
will be performed using the observed cycles to first cracking. Comparisons will be made within
projects and between projects to verify the following aspects of the allowable strain limit design
procedure:
• The overall engineering reasonableness of the approach,

• Applicability of time-temperature superposition to healing and allowable strains,
• Independence of healing on applied strain, and
• Effect of material properties on allowable strains.
Pertinent interim results from these analyses will be discussed in the quarterly progress
reports. The analyses will be thoroughly documented in the fourth interim report submitted at
the end of the 30th month of the project.
Subtask 5.4 Obtain Materials and Data for In-Service Pavement Sections
In this Subtask, data and materials needed to analyze each of the LTPP sections included in
the final matrix of in-service pavements will be obtained. First, the most recent data for the test
B-92

section will be retrieved from the LTPP database (56). This data will be entered into the project
database and managed under Subtask 3.2. The relevant data for the analyses include:
• Traffic.
• Time-series deflection data.
• Time-series fatigue cracking.
• Time-series longitudinal cracking.
• Layer material properties.
A site visit to each of the selected pavement sections is required. The site visit will include:
1. A visual condition survey to confirm the distresses obtained from the LTPP database,
2. Non-destructive testing at various locations in the section using the Portable Seismic
Pavement Analyzer (PSPA) (62, 63) to identify damage in the base layers that is not
apparent from surface distress measurements.
3. Coring to obtain 3 to 5 full depth samples for laboratory testing, and
4. Additional coring to confirm the distress survey and seismic testing. If cracks are
present, cores will be taken through selected cracks to confirm where the cracks
initiated and confirm the cause of cracking.
Each site visit will require two full days. It is envisioned that the necessary traffic control and
coring will be provided by the state highway agencies. Their willingness to participate in the
field testing is an important consideration in the final selection of pavements for analysis.
Subtask 5.5 Perform Lab Testing and Analyze In-Service Pavement Sections
Laboratory Testing
The pavement section cores will be used to determine modulus values for analysis of the
seismic test data and to obtain the properties of the HMA base for use in the predictive model
developed in Experiment 5 of Task 4. This model will relate allowable strains for full healing to
easily measured volumetric properties of the mixture and flow characteristics of the binder.
Mixture properties will be obtained from normal volumetric analysis of the cores. The binder
B-93

will be recovered to determine the required binder properties. A preliminary testing plan is
presented in Table 28 assuming that an indirect tensile strength will be used in the model and a
binder master curve will be required to characterize the flow properties of the binder in the
predictive model developed in Experiment 5 of Task 4.
Table 28. Preliminary Testing Plan for Cores From the LTPP Sections.
Test Method Number Reason

Bulk specific gravity AASHTO T169 3 Volumetric properties
Indirect Tensile Modified 3 Analysis of seismic data
Modulus AASHTO T322
Indirect Tensile AASHTO T322 3 Mixture strength
Strength
Asphalt content AASHTO T164 3 Volumetric properties
Sieve analysis AASHTO T30 3 Gradation
Aggregate bulk specific AASHTO T84 1
Volumetric properties
gravity AASHTO T85 1
Binder Recovery AASHTO T170 3 Obtain binder for rheological testing
Dynamic Shear AASHTO T315 Frequency sweep at Binder master curve
Rheometer 6 temperatures
Bending Beam AASHTO T313 3 temperatures Binder master curve
Rheometer
Analysis
Analysis of the LTPP sections will be performed using the research version of the MEPDG
software, NCHRP9-44A_Version 0.3, developed in Subtask 2.4. The analysis will involve
performing simulations for each of the 32 pavement sections to determine the frequency at which
the allowable strains for full healing (endurance limit behavior) are exceeded. For all of the
simulations, the best available information on the traffic and unbound layers will be used.
Since the field data consists of cracked and uncracked sections, the analysis will produce
binary data (either cracked or uncracked) as shown schematically in Figure 21. From this data a
model for the probability that bottom initiated cracking will occur can be developed using the
logistic function given in Equation 32.
e[b0 + b1 ( PE ) ]
p= (32)
1 + e[b0 + b1 ( PE )]
B-94

where:
p = probability of bottom initiated fatigue cracking
PE = percent of axle loads with strains exceeding the endurance limit
b0 and b1 = fitting parameters
1.0
0.9
Probability of Bottom Inititated Cracking
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0 10 20 30 40 50 60 70 80 90 100
Percent of Axle Loads Exceeding Endurance Limit Strain, %
Figure 21. Schematic of Field Section Data Analysis.
Jackknifing as described in Research Results Digest Number 283 (64) can be used to assess
the accuracy of the of the model coefficients without having to separate the 32 sections into
calibration and validation subsets. Jackknifing is performed by systematically removing one of
the sections, calibrating the model using the remaining sections, then predicting the value of the
section that was removed. For the section that was removed, the model error, ei, is computed as
the difference between the predicted and measured values. The process of withholding,
B-95

calibrating, and determining the error is repeated until each section has been removed. This
process produces n values of the error from which the following jackknifing goodness of fit
statistics can be computed.
0.5 0.5
⎡1 n ⎤ ⎡1 n ⎤
S e = ⎢ ∑ ei 2 ⎥ = ⎢ ∑ (Yî − Yi ) 2 ⎥
⎣ν i =1 ⎦ ⎣ν i =1 ⎦
(33)
where
Se = standard error
ei = errors computed from jackknifing
n = number of measurements taken
ν = degrees of freedom = n minus number of unknowns
Ŷi = predicted value for the ith jackknifing set
Yi = measured value for the ith jackknifing set
⎡⎛ 2 ⎞ ⎤
2 ⎢⎜ S e ⎟⎛ n − p ⎞⎥
R = 1− ⎜ ⎟ (34)
⎢⎜ S y 2 ⎟⎝ n − 1 ⎠⎥
⎣⎝ ⎠ ⎦
where
R2 = explained variance
Se = standard error
Sy = standard deviation of the measured data
p = number of unknowns
n
bias = ∑ ei (35)
i =1
where
ei = errors computed from jackknifing
B-96

The advantage of jackknifing is the goodness of fit statistics are based on predictions of
measurements that are not included in the calibration. They are, therefore, better estimates of the
accuracy of future predictions than goodness of fit statistics based on calibration using the full
data set. The stability of the model can also be assessed by performing the jackknifing again by
withholding two sets of measurements and calibrating using the remaining n-2 measurements.
For n-2 jackknifing, two errors are computed for each set of two measurements that are withheld.
The change in the jackknifing goodness of fit statistics between n-1 and n-2 jackknifing is an
indicator of the stability of the statistics. Stable goodness of fit statistics indicate a model with
reliable prediction accuracy.
Pertinent interim results from these analyses will be discussed in the quarterly progress
reports. The analyses will be thoroughly documented in the fifth interim report submitted at the
end of the 42nd month of the project.
Task 5 Milestones
Table 29 summarizes the major milestones for Task 5. Initially the emphasis of the project
will be on the formulation of the design procedure and the laboratory testing and analysis. This
provides substantial time for compiling the accelerated pavement test and test road data and for
final selection of the LTPP sections. After the laboratory testing and analysis are complete, the
emphasis of the project shifts to collection and analysis of the data from the LTPP sections.

Contract Award
5.1 Initial Selection of Sections for Analysis 12
5.2 Final Selection of LTPP Sections for Analysis 20
5.3 Compile Data From Accelerated Pavement Tests and Test Roads 24
5.4 Complete Analysis of Accelerated Pavement Tests and Test 27
Roads
5.5 Complete Data Collection for LTPP Sections 32
5.6 Complete Testing and Analysis of LTPP Sections 35
B-97


hours for each of the positions in the research management structure presented in Figure 5,
engineering support for collection and analysis of the pavement sections, and technician support
for laboratory testing. Task 5 is estimated to require a total of 4,900 man-hours of effort. This is
approximately 38 percent of the total effort required for the project.
Data
Laboratory Pavement
Principal Statistician Support
Team Team
Subtask Investigator Team Engineers Technicians
Leader Leader
Leader
5.1 Review Data Sources and Select 16 8 0 28 0 320 0
Sections for Analysis
5.2 Obtain Materials and Data for 20 0 0 28 0 280 0
Accelerated Pavement Tests and Test
Roads
5.3 Perform Lab Testing and 36 16 4 108 0 512 32
Analyze Accelerated Pavement Tests
and Test Roads
5.4 Obtain Materials and Data for 20 0 0 100 0 1280 0
In-Service Pavement Sections
5.5 Perform Lab Testing and 90 30 90 90 0 512 1280
Analyze In-Service Pavement
Sections
Total 182 54 94 354 0 2904 1312
Task 5 Sources
FHWA Superpave Validation Study

Stuart, K.D., Mogawer, W.S., and Romero, P., “Validation of the Superpave Asphalt
Binder Fatigue Cracking Parameter Using an Accelerated Loading Facility,” Report
Number FHWA-RD-01-093, Federal Highway Administration, Washington, D.C.,
2002.
NCHRP Project 9-19, “Binder Characterization, Team Report BC-4, “Rheological

Characterization of the FHWA-ALF Binders,” PDF File No, 04-E, CRP-CD-46, National
Cooperative Highway Research Program, 2005.
B-98

MnRoad
Minnesota Department of Transportation. http://www.dot.state.mn.us/mnroad/ (accessed
June, 2008).
Palmquist, D., Worel, B., Zerfas, W., “2002 Hot-Mix Asphalt Mainline Test Cell
Condition Report,” Minnesota Department of Transportation, September 6, 2002.
Worel, B., “MnRoad Database Guide,” Minnesota Department of Transportation,

January, 2006.

Characterization of the MnRoad Binders,” PDF File No. 04-D, CRP-CD-46, National
LTPP Sections
Federal Highway Administration. http//www.ltpp-products.com/DataPave/index.asp
National Cooperative Highway Research Program, http://www.trb.org/mepdg/guide.htm

(accessed June 30, 2008).
NCAT Test Track

National Center for Asphalt Technology. http://www.pavetrack.com/ (accessed June,
2008).
Timm, D. West, R., Priest, A., Powell, B., Selvaraj, I., Zhang, J., and Brown, R., “Phase
II NCAT Test Track Results,” NCAT Report 06-05, National Center for Asphalt
Technology, Auburn University, Auburn, AL, December, 2006.
Timm, D. and Priest, A., “Material Properties of the 2003 NCAT Test Track Structural
Study,” NCAT Report 06-01, National Center for Asphalt Technology, Auburn
University, Auburn, AL, April, 2006.
Willis, J., and Timm, D., “Forensic Investigation of a Rich Bottom Pavement”, NCAT
Report 06-04, National Center for Asphalt Technology, Auburn University, Auburn, AL,
December, 2006.
Timm, D.H., Priest, A.L., and McEwen, T. V., “Design and Instrumentation of the
Structural Pavement Experiment at the NCAT Test Track,” NCAT Report 04-01,
National Center for Asphalt Technology, Auburn University, Auburn, AL, April, 2004.
Brown, E.R., Cooley, L.A., Hanson, D., Lynn, C., Powell, B., Prowell, B., and Watson,
D., “NCAT Test Track Design, Construction, and Performance,” NCAT Report 02-12,
B-99

National Center for Asphalt Technology, Auburn University, Auburn, AL, November,
2002.
Westrack
Epps, J. A., Hand, A., Seeds, S., Schultz, T., Alavi, S., Ashmore, C., Monismith, C.,
Deacon, J.A., Harvey, J.T., and Leahy, R., “Recommended Performance Related
Specifications for Hot-Mix Asphalt Construction: Results of the WesTrack Project,”
NCHRP Report 455, National Cooperative Highway Research Program, Washington,
D.C., 2002.
NCHRP Web-Only Document 111: “Recommended Performance Related Specifications

for Hot-Mix Asphalt Construction: Results of the WesTrack Project,”
http://onlinepugs.trb.org/onlinepubs/nchrp/nchrp_w111.pdf (accessed June, 2008).

Characterization of the WesTrack Binders,” PDF File No, 04-F, CRP-CD-46, National
Seismic Pavement Analyzer

Geomedia Research and Development, http://geomedia.us/ (accessed June, 2008)
McDaniel, M., Yuan. D., Chen, D., and Nazarian, S., “Use of Seismic Pavement
Analyzer in Forensic Studies in Texas,” Nondestructive Testing of Pavements and
Backcalculation of Moduli, Third Volume, ASTM STP 1735, American Society for
Testing and Materials, West Conshohoken, PA, 2000.
Nazarian, Soheil, et al., Development of Structural Field Testing of Flexible Pavement

Layers, Project No. 1735, Final Report, University of Texas at El Paso, Center for
Materials and Research, El Paso, Texas, 2002.
Nazarian, S., Baker, M. R., and Crain, K., “Development and Testing of a Seismic
Pavement Analyzer,” Report Number SHRP-H-377 Strategic Highway Research
Program, Washington, D.C., 1993.
Steyn, W., and Sadzik, E., “Application of the Portable Seismic Pavement Analyzer
(PSPA) for Pavement Analysis,” http://hdl.handle.net/1024/1246, 2007.
Von Quintus, et al., Nondestructive Testing Technology for Quality Control and
Acceptance of Flexible Pavement Construction; Volume I - Procedure Manual, Final
Report, NCHRP Project 10-65, June 2008.
B-100

Schedule of Tasks
The HMA Endurance Limit Validation Study will require 48 months to complete. Figure 22
presents a Gantt Chart for the project with the critical path identified. Table 31 presents a
complete listing of milestones for the project.
Perhaps the most critical task in the project is Task 2.2, Finalize Preliminary Approach,
because the procedure assembled in this task will shape the final design of the laboratory
experiments and the final selection of in-service pavements for analysis. Once the preliminary
design procedure is finalized, then the critical path shifts to the laboratory studies in Task 4.
When the laboratory studies are completed, the critical path splits. The development of
NCHRP944A_Version 0.2 of the research MEPDG software in Task 2.3 becomes critical. This
version of the software will be used to analyze the accelerated pavement and test road data in
Task 5.3. Then based on the finding from these analyses, NCHRP944A_Version 0.3 will be
developed for the calibration studies using data from the LTPP sections. The collection of data
from the LTPP sections in Task 5.4 also becomes critical. The site visits required in this task
can not begin until the form of the model for predicting allowable strains from mixture
composition is determined. The final field coring and laboratory testing plans will depend on the
form of the model developed in Task 4.5. The schedule provides 12 months to perform the 32
site visits. This is a compressed schedule for the site visits and likely will require at least two
field engineers to complete the work as scheduled.
Analysis of the LTPP sections can begin as soon as the NCHRP944A_Version 0.3 is
completed in Task 2.4. Laboratory testing of the field cores will lag the site visits by
approximately 1 month; therefore, the data required to analyze most of the LTPP sections will be
available when NCHRP944A_Version 0.3 is completed.
The final tasks of the project begin after the calibration analyses are completed in Task 5.5.
This includes development of the final design procedure, NCHRP9-44A_Version 1.0 of the
software, and the preparation of the final report for the project.
B-101

Figure 22. Project Schedule With Critical Path Shown in Black.
B-102
Table 31. Project Milestone Summary.
B-103

Budget
The budget for the project is based on the labor hour estimates provided in the Task by Task
Description of the Research Plan and the loaded hourly rates presented in Table 32 for various
categories of labor. Travel costs were included for the panel meetings in Task 1.3 and for the
LTPP site visits in Task 5.4. Printing costs were also included in Task 1.3 for each of the Interim
Reports and the Final Report. The overall budget is presented in Figure 23. Details of the travel
and printing estimates are provided in Tables 33 and 34, respectively.
Table 32. Labor Costs Used in Budget Preparation.
Labor Category Loaded Hourly Rate

Senior Engineers and Statistician $150.00
Engineers and Programmers $100.00
Technicians $85.00
Administrative Support $60.00
Table 33. Travel Cost Estimate.
Task Item Detail Estimate

1.3 Transportation 3 presentations × 2 people × $800 per trip $4,800
1.3 Lodging & Per Diem 3 presentations × 2 people × 2 days × $265/day $3,180
Task 1.3 Total $7,980
5.4 Airfare 2 person × 16 projects × $800 per site $25,600
5.4 Rental Car 1 car × 4 days × 16 sites × $75.00/day $4,800
5.4 Lodging & Per Diem 2 × 5 days × 16 sites × $120.00/ day $19,200
Task 5.4 Total $49,600
Table 34. Estimate of Report Printing Costs.
Report Pages Copies Cost /Page Cost

Interim 1 300 20 $0.05 $300
Interim 2 300 20 $0.05 $300
Interim 3 300 20 $0.05 $300
Interim 4 300 20 $0.05 $300
Interim 5 300 20 $0.05 $300
Revised Final 300 100 $0.05 $1,500
Total $3,000
B-104

Figure 23. Project Budget.
B-105
The total cost of the the project is $1,485,360. Figure 24 provides a estimate of monthly
expenditures for the project. Monthly expenditures reach approximately $52,000 per month
when the laboratory experiments are being conducted.
$60,000
Montly
Cumulative $1,400,000
$50,000
$1,200,000
Estimate Cummulative Expenditure

Estimated Monthly Expenditure
$40,000 $1,000,000
$800,000
$30,000
$600,000
$20,000
$400,000
$10,000
$200,000
$0 $0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48
Month After Contract Award
Figure 24. Estimated Monthly and Cumulative Expenditures.
B-106

References
1. Advanced Asphalt Technologies, LLC, “Hot Mix Asphalt Endurance Limit Workshop:
Executive Summary,” National Cooperative Highway Research Program Project 9-44,
November, 2007.
2. Monismith, C.L., and McLean, D.B., “Structural Design Considerations,” Proceedings of

the Association of Asphalt Paving Technologists, Vol. 41, 1972.
3. Newcomb, D.E., Buncher, M., and Huddleston, I.J., “Concepts of Perpetual Pavements,”
4. Carpenter, S.H., Ghuzlan, K.A., and Shen, S., “Fatigue Endurance Limit for Highway and
Airport Pavements,” Transportation Research Record No. 1832, Transportation Research
5. Carpenter, S.H., and Shen, S., “Application of the Dissipated Energy Concept in Fatigue
6. Prowell, B., Brown, E., R., Daniel, J., Bhattacharjee, S., Von Quintus, H., Carpenter, S.,
Shen, S., Anderson, M., Swamy, A. K., and Maghsoodloo, S., “Endurance Limit of Hot Mix
Asphalt Mixtures to Prevent Fatigue Cracking in Flexible Pavements,“ Updated Draft Final
Report, NCHRP 9-38, National Cooperative Highway Research Program, Washington, D.C.,
May, 2008.
7. Carpenter, S.H., and Shen, S., “Dissipated Energy Approach to Study Hot-Mix Asphalt
Healing in Fatigue,” Transportation Research Record No. 1970, Transportation Research
8. Kim, B. and Roque, R., “Evaluation of Healing Property of Asphalt Mixtures,”

Transportation Research Record No. 1970, Transportation Research Board, Washington,
D.C., 2006.
9. Nunn, M. and Ferne, B.W., “Design and Assessment of Long-Life Flexible Pavements,”
10. Brown, S.R., Thom, N.H., and Hakim, B.A., “Performance and Rehabilitation of Heavy-duty
Pavements in the UK: Some Case Studies,” Proceedings, International Symposium of
Design and Construction of Long Lasting Asphalt Pavements, National Center for
Asphalt Technology, Auburn, AL, 2004.
11. Wu, Z., Siddiqui, Z.Q., Hossain, M., and Gisi, A.J., “Kansas Turnpike – An Example of
Long Lasting Asphalt Pavement,” Proceedings, International Symposium of Design and
B-107

Construction of Long Lasting Asphalt Pavements, National Center for Asphalt

Technology, Auburn, AL, 2004.
12. Uhlmeyer, J.S., Willoughby, K., Pierce, L.M., and Mahoney, J.P., “Top-Down Cracking in
Washington State Asphalt Concrete Wearing Courses,” Transportation Research Record
No. 1730, Transportation Research Board, Washington, D.C., 2000.
13. Procedural Manual for Agencies Conducting Research in the National Cooperative Highway
Research Program, Transportation Research Board, August, 2006.
14. Preparing Your CRP Final Report, Transportation Research Board, September, 2006.
15. Timm, D., H., and Young, J., B., “Effects of Load Spectra and Variability on Perpetual
Pavement Design,” Proceedings, International Symposium of Design and Construction
of Long Lasting Asphalt Pavements, National Center for Asphalt Technology, Auburn,
AL, 2004.
16. Witczak, M.W., “Use of HMA Fatigue Endurance Limits in the Mechanistic-Empirical
Pavement Design Guide,” Presentation at the Hot Mix Asphalt Endurance Limit Workshop,
August 1, 2007.
17. Thompson, M.R., ad Carpenter, S.H., “Design Principles for Long Lasting Pavement,”
Proceedings, International Symposium of Design and Construction of Long Lasting
Asphalt Pavements, National Center for Asphalt Technology, Auburn, AL, 2004.
18. Sidess, A. and Uzan, J., “A Design Method for Perpetual Flexible Pavement in Israel,”
International Journal of Pavement Engineering, February, 2008.
19. Von Quintus, H. L., “Hot-Mix Asphalt Layer Thickness Design for Longer Life Bituminous
Pavements,” Transportation Research Circular Number 503, Perpetual Bituminous
Pavements, Transportation Research Board, Washington, D.C., December, 2001.
20. Von Quintus, Harold L., “Application of the Endurance Limit in Mechanistic-Empirical
Based Pavement Design Procedures,” paper prepared for presentation and publication at the
International Conference on Perpetual Pavements, Columbus, Ohio, September 13-15, 2006.
21. Tsai, B.W., Harvey, J., Monismith, C., and Bejarano, M., “Calibration of Fatigue Surface
Cracking Using Simplified Recursive Miner’s Law,” Journal of the Association of Asphalt
Paving Technologists, Vol., 76, 2007.
22. Mun, S., Guddati, M. N., and Kim, Y. R., “Fatigue Cracking Mechanisms in Asphalt
Pavements with Viscoelastic Continuum Damage Finite-Element Program,” Transportation
Research Record No. 1896, Transportation Research Board, Washington, D.C., 2004.
B-108

23. Roque, R., B. Birgisson, B. Sangpetngam and Z. Zhang, “Hot Mix Asphalt Fracture
Mechanics: A Fundamental Crack Growth Law for Asphalt Mixtures,” Journal of the
24. Bonnaure, F.P, Huibers, A.H.J.J., Boonders, A., “A Laboratory Investigation of the Influence
of Rest Periods on the Fatigue Response of Bituminous Mixes,” Proceedings, Association of
25. Chehab, G.R., Kim, Y.R., Schapery, R.A., Witczak, M.W., and Bonaquist, R.,
"Characterization of Asphalt Concrete in Uniaxial Tension Using a Viscoelastoplastic
Continuum Damage Model," Journal of the Association of Asphalt Paving Technologists,
Vol. 72, 2003.
26. Huang, Y. H., Pavement Analysis and Design, Englewood Cliffs, N.J.: Prentice-Hall, Inc.,
1993.
27. Al-Qadi, I., Xie, W., and Elseifi, M., “Frequency Determination from Vehicular Loading
Time Pulse to Predict Appropriate Complex Modulus in MEPDG," Paper P:reprint CD,
Association of Asphalt Paving Annual Meeting and Technical Sessions, Philadelphia, PA,
April 27-30, 2008.
28. Barksdale, R.G., “Compressive Stress Pulse Times in Flexible Pavements for Use in
Dynamic Testing,” Highway Research Record 345, Highway Research Board, Washington,
D.C., 1971.
29. Mirza, M. W., and M. W. Witczak, “Development of a Global Aging System for Short and
Long Term Aging of Asphalt Cements,” Journal of the Association of Asphalt Paving
30. Transportation Research Board, Highway Capacity Manual, Special Report 209, Third
Edition, Transportation Research Board, Washington, DC 1994.
31. Hajek, J. J., Selezneva, O., I., Mladenovic, G., and Jiang, Y., J., “Estimating Cumulative
Traffic Loads, Volume II: Traffic Data Assessment and Axle Load Projection for the Sites
With Acceptable Axle Weight Data, Final Report for Phase 2,” Report Number FHWA-
RD-03-094, Federal Highway Administration, Washington, D.C., March, 2005.
32. Lytton, R L; Uzan, J; Fernando, E G; Roque, R; Hiltunen, D; Stoffels, S M, “Development

And Validation Of Performance Prediction Models And Specifications For Asphalt Binders
And Paving Mixes,” Report Number SHRP-A-357, Strategic Highway Research Program,
33. Von Quintus, H.L., Schwartz, C., McQuen, R., and Andrei, D., “Experimental Plan for
Calibration and Validation of Hot-Mix Asphalt Performance Models for Mix and Structural
Design,” Final Report for National Cooperative Highway Research Program Project 9-30,
National Cooperative Highway Research Program, January, 2004.
B-109

34. Little, D. N., Lytton, R. L., Williams, D., and Chen, C. W., “Microdamage Healing in
Asphalt and Asphalt Concrete, Volume I: Microdamage and Microdamage Healing Project
Summary Report,” Report Number FHWA-RD-98-141, Federal Highway Administration,
Washington, D.C., June 2001.
35. Rao Tangella, S.C.S., Craus, J., Deacon, J.A., and Monismith, C.L., “Summary Report on
Fatigue Response of Asphalt Mixtures,” Report Number SHRP-A/IR-90-011, Strategic
36. Bonnaure, R., Gravois, A., and Udron, J., “A New Method for Predicting the Fatigue Life of
Bituminous Mixes,” Proceedings of the Association of Asphalt Paving Technologists,
Vol. 49, 1980.
37. Lee, H., J., Daniel, J.S., and Kim, Y.R., “Laboratory Performance Evaluation of Modified
Asphalt Mixtures for Inchon Airport Pavements,” International Journal of Pavement
Engineering, Volume 1, Number 2, 2000.
38. Walubita, L.F., Epps-Martin, A., Jung, S. H., Glover, C. J., Park, E.S., Chowdhury, A., and
Lytton, R. L., “Comparison of Fatigue Analysis Approaches for Two Hot Mix Asphalt
Concrete (HMAC) Mixtures, Report Number FHWA/TX-05/0-4468-2, Texas Department
of Transportation, Austin, TX, August, 2005.
39. Christensen, D.W., and Bonaquist, R.F., “Volumetric Requirements for Superpave Mix
Design,” NCHRP Report 567, National Cooperative Highway Research Program,
40. Prowell, B. D., and Brown, E. R., “Superpave Mix Design: Verifying the Gyration Levels in
the Ndesign Table,” NCHRP Report 573, National Cooperative Highway Research Program,
41. Epps, J. A., Hand, A., Seeds, S., Schultz, T., Alavi, S., Ashmore, C., Monismith, C., Deacon,
J.A., Harvey, J.T., and Leahy, R., “Recommended Performance Related Specifications for
Hot-Mix Asphalt Construction: Results of the WesTrack Project,” NCHRP Report 455,
National Cooperative Highway Research Program, Washington, D.C., 2002.
42. Kim, Y.R., and Little, D.N., “One Dimensional Constitutive Modeling of Asphalt Concrete,”
ASCE Journal of Engineering Mechanics, Vol. 116, No. 4, 1990.
43. Lundstrom, R., and Isacsson, U., “Characterization of Asphalt Concrete Deterioration Using
Monotonic and Cyclic Tests,” International Journal of Pavement Engineering, Vol. 4, No.
3, September, 2003.
44. Christensen, D.W., and Bonaquist, R.F., “Practical Application of Continuum Damage
Theory to Fatigue Phenomena in Asphalt Concrete Mixtures,” Journal of the Association of
B-110

45. Advanced Asphalt Technologies, LLC, “NCHRP 9-29 Phase V Draft Final Report,” National
Cooperative Highway Research Program, July 30, 2007.
46. Advanced Asphalt Technologies, LLC, “NCHRP 9-29 Phase IV Draft Final Report,”
National Cooperative Highway Research Program, June, 2006.
47. Christensen, D.W., Pellinen, T, and Bonaquist, R.F., “Hirsch Model for Estimating the
Modulus of Asphalt Concrete,” Journal of the Association of Asphalt Paving
48. Christensen, D.W., and Anderson, D.A., “Interpretation of Dynamic Mechanical Test Data
for Paving Grade Asphalt Cements,” Proceedings of the Association of Asphalt Paving
49. Christensen, D.W., and Bonaquist, R., “Analysis of HMA Fatigue Data Using The Concepts
of Reduced Loading Cycles and Endurance Limit,” Submitted for Presentation and
Publication at the 2009 Annual Meeting of the Associations of Asphalt Paving
Technologists.
50. Stuart, K.D., Mogawer, W.S., and Romero, P., “Validation of the Superpave Asphalt Binder
Fatigue Cracking Parameter Using an Accelerated Loading Facility,” Report Number
FHWA-RD-01-093, Federal Highway Administration, Washington, D.C., 2002.
51. National Center for Asphalt Technology. http://www.pavetrack.com/ (accessed June, 2008).
52. Timm, D. West, R., Priest, A., Powell, B., Selvaraj, I., Zhang, J., and Brwon, R., “Phase II
NCAT Test Track Results,” NCAT Report 06-05, National Center for Asphalt Technology,
Auburn University, Auburn, AL, December, 2006.
53. Timm, D. and Priest, A., “Material Properties of the 2003 NCAT Test Track Structural
Study,” NCAT Report 06-01, National Center for Asphalt Technology, Auburn University,
Auburn, AL, April, 2006.
54. Minnesota Department of Transportation. http://www.dot.state.mn.us/mnroad/ (accessed

June, 2008).
55. Palmquist, D., Worel, B., Zerfas, W., “2002 Hot-Mix Asphalt Mainline Test Cell Condition
Report,” Minnesota Department of Transportation, September 6, 2002.
56. Federal Highway Administration. http://www.ltpp-products.com/DataPave/index.asp

57. Asphalt Pavement Alliance. http://www.asphaltalliance.com/ (accessed June, 2008).
B-111

58. Worel, B., “MnRoad Database Guide,” Minnesota Department of Transportation, January,
2006
59. NCHRP Project 9-19, “Binder Characterization, Team Report BC-4, “Rheological
Characterization of the FHWA-ALF Binders,” PDF File No, 04-E, CRP-CD-46, National
Characterization of the WesTrack Binders,” PDF File No, 04-F, CRP-CD-46, National
Characterization of the MnRoad Binders,” PDF File No. 04-D, CRP-CD-46, National
62. Von Quintus, et al., Nondestructive Testing Technology for Quality Control and Acceptance
of Flexible Pavement Construction; Volume I - Procedure Manual, Final Report, NCHRP
Project 10-65, June 2008
63. Geomedia Research and Development, http://geomedia.us/ (accessed June, 2008)
64. VonQuintus, H.L, Schwartz, C.E., McCuen, R.H., and Andrei, D., “Jackknife Testing – An
Experimental Approach to Refine Model Calibration and Validation,” NCHRP Research
Results Digest Number 283, National Cooperative Highway Research Program,
Washington, D.C, December, 2003.
B-112

Attachment: Recommended LTPP Test Sections

The following provides location and summary information for the LTPP test sections that are
applicable for use in confirming the endurance limit and values. The distress information listed
for each test sections are the values included within the LTPP database. Specifically, the
longitudinal cracking and transverse cracking values are in meters, while the block and alligator
cracking values are in square meters.
B-113

LTPP Site Identification Number: 0103 State: Alabama (01)

Roadway or Route No.: US-280
Date of Construction: April 1991 Status: In Service
Location:
Longitude: 85.25 Latitude: 32.62 Elevation: 151
Pavement Cross Section:

Layer Material Type Thickness, inches
4 HMA Hot Mixed, Hot Laid HMA, Dense Graded (1) 1.5
2 ATB Asphalt Treated Base (319) 7.4
1 Subgrade Soil Sandy Lean Clay (114) ---
Traffic Data:
Number of Years with Data:
AADTT (One-way): Year:
KESALS per year:
Unbound Layers Resilient Modulus:

Layer Material/Soil Type Equivalent Resilient Modulus, ksi
1 Sandy Lean Clay 6
Hot Mix Asphalt Mixtures:

Equivalent Annual Modulus of HMA: 200
Type of Asphalt: PG64-22
HMA HMA ATB

Asphalt Content, % 5.1 4.0 4.3
Air Voids, % 3.3 5.1 11.6
Gradation; percent passing:

#3/4 #3/8 #4 #40 #200
HMA 100 83 61 24 6.7
HMA 90 66 48 21 7.2
ATB 90 65 42 18 5.6
Tensile Strain at Bottom of HMA Layer:
Total Amount of Fatigue Cracking:

Year Alligator Block Long. Transverse Year Alligator Block Long. Transverse
1994 0 0 0 0 2001 22.9 0 9.7 0
1995 0 0 0 0 2002 28.7 0 0 1.5
1996 0 0 0 0 2003 34.4 0 0 7.9
2000 34.9 0 30.6 0 2004 40.1 0 0 8.7
2005 41.4 0 0 8.0
B-114


Roadway or Route No.: US-280
Location:

2 GB Crushed Stone, Granular Base (303) 7.9
1 Subgrade Soil Sandy Silt (145) ---
Traffic Data:
KESALS per year:

1 Sandy Lean Clay 6

HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1995 0 0 0 0 2001 25.3 0 13.7 0.8
1996 1.2 0 4.5 0 2002 31.1 0 0 3.5
1997 0 0 37.5 0 2003 64.9 0 0 14.6
1998 3.0 0 0.2 0.3 2004 68.0 0 0 15.7
2000 38.6 0 16.2 0.6 2005 70.4 0 0 14.6
B-115


Roadway or Route No.: US 280
Location:

4,5 HMA Hot Mixed, Hot Laid AC, Dense Graded (1) 4.0
3 ATB Hot Mixed, Hot Laid AC Dense Graded Mix (319) 7.9
2 PATB Open-Graded Hot Laid Mix (325) 3.7
1 Subgrade Soil Silt with Sand (143) ---
Traffic Data:
KESALS per year:

1 Silt with Sand 5

HMA HMA ATB PATB

Asphalt Content, % 5.2 4.0 4.3 2.2
Air Voids, % 3.3 5.1 11.6 ---

#3/4 #3/8 #4 #40 #200
HMA 100 83 61 24 6.7
HMA 90 66 48 21 7.2
ATB 90 65 42 18 5.6
PATB 71 19 10 7.0 4.3
Tensile Strain at Bottom of HMA Layer: 144.29

1994 0 0 0 0 2001 92.7 0 0.5 0
1995 0 0 0 0 2001 67.8 0 0 1.2
1996 0 0 5.5 0 2003 86.4 0 0 2.4
2000 60.2 0 23.2 0 2004 86.9 0 0 3.3
2005 89.0 0 0 3.5
B-116

LTPP Site Identification Number: 0116 State: Arizona (04)

Date of Construction: Jan. 1993 Status: In Service
Location:

4 AC Surface Seal (72), 2003 4.0
3 HMA Hot Mixed, Hot Laid AC, Dense Graded (1)
2 ATB Hot Mix, Hot Laid, AC, Dense Graded (1) 12.1
1 Subgrade Soil Silty Sand with Gravel (215) 132
Traffic Data:
Number of Years with Data: 3
AADTT (One-way): 1190 Year: 1995
KESALS per year: 300

Layer Material/Soil Type Equivalent Resilient Modulus. ksi
1 Silty Sand with Gravel 4
Last 5

Type of Asphalt:
HMA ATB
Asphalt Content, % 4.7 4.5
Air Voids, % 10.3 6.1

#3/4 #3/8 #4 #40 #200
HMA 99 82 64 18 3.9
ATB 88 72 56 16 4.0

1995 0 0 0 0 2002 69.9 0 16.2 1.5
1998 0 0 0 0 2003 5.8 0 19.2 1.1
1999 1.4 0 2.1 0 2004 11.2 0 17.3 2.8
2000 4 0 3.3 0 2005 11.6 0 28.5 5.3
2001 9.9 0 22 0
B-117

LTPP Site Identification Number: 04-0162 State: Arizona (04)

Location:

3 AC Seal Coat (72), 2003 9.0

2 HMA Hot Mixed, Hot Laid AC, Dense Graded (1)
1 Subgrade Soil Well Graded Gravel with Sand & Silt (261) ---
Traffic Data:

Layer Material/Soil Type Equivalent Resilient Modulus. ksi

Type of Asphalt:
HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA

1995 0 0 0 0 2002 1.7 0 3.8 2.0
1998 0 0 0 0 2005 2.8 0 4.5 10.4
1999 0 0 0 0 2006 4.4 0 1.4 20.4
2000 0 0 0 0
2001 0 0 1.0 0
B-118


Roadway or Route No.: I 40
Date of Construction: 10-1-1977 Status: Milled/Overlay; Friction Course
Location:

6 AC Open-Graded Friction Course (82); 9-1999 0.3
5 AC Open Graded, Sand Seal (2) 4.6 – after milling
4 HMA Hot Mixed, Hot Laid AC, Dense Graded (1) (5.8 – Original)
3 ATB Asphalt Treated Mixture (321) 11.2
2 TS Lime Treated Mixture (338) 6
1 Subgrade Soil Clayey Gravel with Sand (267) 54
Traffic Data:

2 Lime Treated Mixture 28
1 Clayey Gravel with Sand 15

Type of Asphalt:
HMA
Asphalt Content, % 5.3
Air Voids, % 5.5

#3/4 #3/8 #4 #40 #200
HMA 80.5 57 48.5 14.5 7.4

1996 0 0 0 6.4 2005 7.0 0 1.4 0.6
1998 0 0 0 8.7
2000 0 0 0 0
2003 0 0 0 0
B-119


Roadway or Route No.: I 40
Date of Construction: 10-1-1977 Status: Milled/Overlay; Friction Course
Location:

5 AC Open-Graded Friction Course (82); 9-1999 0.3
4 AC Hot Mixed, Hot Laid AC, Dense Graded (1) 5.3
3 TB Asphalt Treated Mixture (319) 13.7
2 TS Lime Treated Soil (338) 5
1 Subgrade Soil Clayey Gravel with Sand (A-2-6) ---
Traffic Data:

2 Lime Treated Soil 28
1 Clayey Gravel with Sand 15

Type of Asphalt:
HMA
Air Voids, % 4.2

#3/4 #3/8 #4 #40 #200
HMA 77 54.5 47.5 14.5 7.2

1992 0 0 0 0 2005 0 0 2.7 0.7
1998 0 0 0 12.2
2000 0 0 0 0
2003 0 0 0 0.3
B-120

LTPP Site Identification Number: 0115 State: Arkansas (05)

Location:

4 HMA Hot Mixed, Hot Laid AC, Dense Graded (1) 1.8
2 ATB Hot Mixed, Hot Laid, Dense Graded (319) 7.4
1 Subgrade Soil Silty Sand (214) ---
Traffic Data:

1 Silty Sand ---

Equivalent Annual Modulus of HMA: ---
Type of Asphalt:
HMA HMA ATB

Air Voids, % 9.9 9.9 6.7

#3/4 #3/8 #4 #40 #200
HMA 98 78 53 12 5.0
HMA 91 61 43 20 6.0
ATB 77 47 35 16 4.5

1997 0 0 0 0 2003 21 0 0.7 7.7
2000 6.7 0 1.7 0.5 2004 22 0 0.8 8.9
2001 10.2 0 0 1.7 2005 25.4 0 0 12.6
2002 17.3 0 0 4.3
B-121


Location:

2 GB Crushed Stone, Granular Base (303) 11.3
Traffic Data:

1 Silty Sand ---

Type of Asphalt:
HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1997 0 0 0 0 2003 39.8 0 0 18.6
2000 6.2 0 2.3 0.4 2004 59.1 0 1.2 37.3
2001 10.7 0 1.0 2.4 2005 81.6 0 0 79.3
2002 31.2 0 0 9.7
B-122


Location:

1 Subgrade Soil Poorly Graded Sand with Silt (204) ---
Traffic Data:

1 Silty Sand ---

Type of Asphalt:
HMA HMA ATB

Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA
ATB

1997 0 0 0 0 2003 15.5 0 0 11.7
2000 2.7 0 0.7 1.4 2004 15.5 0 0 16.9
2001 5.4 0 0 0.8 2005 19.9 0 0 21.1
2002 10.8 0 0 6.3
B-123


Date of Construction: 7-1-1987 Status: In Service
Location:

6 AC Seal Coat (72); Placed after construction 0.4
5 AC Seal Coat (71) 0.5
2 ATB Hot Mix, Hot Laid, Dense Graded (319) 10.5
1 Subgrade Soil Lean Clay (214) ---
Traffic Data:

1 Subgrade Soil/Lean Clay 5

Type of Asphalt:
HMA HMA
Air Voids, % 3.7 6.0

#3/4 #3/8 #4 #40 #200
HMA 100 80 64 30 8.5
HMA 84 58 45 21 7.4

1991 0 0 0 0 1999 0 0 0 0
1994 0 0 1 0 2000 0 0 0 2.2
1995 0 0 0 0 2003 0.7 0 0 60.7
1997 0 0 0 0 2004 0.6 0 0 95.9
B-124

LTPP Site Identification Number: 0106 State: Florida (12)

Location:

3 ATB Hot Mixed, Hot Laid AC, Dense Graded (319) 8.4
2 GB Crushed Stone (303) 4
1 Subgrade Soil Silty Sand with Gravel (215) 87.6
Traffic Data:
KESALS per year:

2 Crushed Stone 25

Type of Asphalt:
HMA HMA ATB

Air Voids, % 8.1 5.5 4.6

#3/4 #3/8 #4 #40 #200
HMA 100 90 68 28 3.1
HMA 99 77 60 25 3.2
ATB

1996 0 0 0 0 2003 0 0 0 0
2000 0 0 0 0 2004 0 0 1.7 1.1
2001 0 0 0 0 2005 0 0 2.3 1.3
2002 0 0 0 0
B-125


Location:

1 Subgrade Soil Poorly Graded Sand with Silt and Gravel (205) 87.6
Traffic Data:
KESALS per year:

1 Poorly Graded Sand with Silt and Gravel 5

Type of Asphalt:
HMA HMA ATB

Air Voids, % 4.2 3.9 5.9

#3/4 #3/8 #4 #40 #200
HMA 100 90 68 28 3.1
HMA 99 77 60 25 3.2
ATB

1996 0 0 0 0 2003 0 0 0 0
2000 0 0 0 0 2004 0.2 0 0 0
2001 0 0 0 0 2005 0.5 0 1.2 0
2002 0 0 0 0
B-126


Location:

2 GB Crushed Stone (303) 8.1
Traffic Data:
KESALS per year:

2 Crushed Stone 25

Equivalent Annual Modulus of HMA:
Type of Asphalt:
HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1996 0 0 0 0 2003 0 0 0 0
2000 0 0 0 0 2004 0.1 0 0 0.3
2001 0 0 0 0 2005 0.2 0 0 0.8
2002 0 0 0 0
B-127


Location:

1 Subgrade Soil Poorly Graded Sand with Gravel & Silt (205) 87.6
Traffic Data:
KESALS per year:


Type of Asphalt:
HMA HMA ATB

Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA
ATB

1996 0 0 0 0 2003 0.6 0 0 0
2000 0 0 0 0 2004 2.6 0 0.6 0
2001 0 0 0 0 2005 3.4 0 7.8 0
2002 0 0 0.4 0
B-128


Location:

1 Subgrade Soil Poorly Graded Sand with Gravel & Silt (205) 105.6
Traffic Data:
KESALS per year:

2 Crushed Stone 25

Type of Asphalt:
HMA HMA ATB

Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA
ATB

1996 0 0 0 0 2003 1.6 0 5.9 0
2000 0 0 0 0 2004 11.6 0 54.3 0.3
2001 0 0 0 0 2005 12.1 0 52.8 0.6
2002 0 0 0 0
B-129


Location:

3 ATB HMAC, Hot Laid, Dense Graded (319) 8.2
2 PATB Open Graded Hot Mix, Hot Laid (325) 4.0
Traffic Data:
KESALS per year:


Type of Asphalt:
HMA HMA ATB

Air Voids, % 7.9 6.8 5.3

#3/4 #3/8 #4 #40 #200
HMA 98 78 59 29 2.5
HMA 99 77 60 25 3.2
ATB

1996 0 0 0 0 2003 0 0 0 0
2000 0 0 0 0 2004 0.3 0 8.0 1.2
2001 0 0 0 0 2005 0.5 0 13.1 1.6
2002 0 0 0 0
B-130

LTPP Site Identification Number: 4113 State: Georgia (13)

Roadway or Route No.: IH 95
Location:

AC Seal Coat (71) 0.1
2 ATB Asphalt Treated Mixture (321) 11.5
1 Subgrade Soil Poorly Graded Sand with Silt (204) ---
Traffic Data:

1 Poorly Graded Sand with Silt 12

Type of Asphalt:
HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 99 63 46 27 3.7
ATB 74 47 40 22 3.9

1991 0 0 0 0 1999 0 0 0 0
1994 0 0 0 1 2000 0 0 0 0
1997 4.4 0 9.5 2
1998 4.1 0 11.3 2.1
B-131


Location:

5 AC Friction Course (2) 0.8
4 HMAA Hot Mixed, Hot Laid AC, Dense Graded (1) 2.0
2 GS Soil Agg. Mix (308) 16.4
1 Subgrade Soil Sandy Silt (145) ---
Traffic Data:

Layer Material/Soil Type Equivalent Resilient Modulus ksi
1 Soil Agg. Mix 14
2 Sandy Silt 9

Type of Asphalt:
HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 96 75 61 17 9.5
ATB 72 63 52 15 8.5

1992 0 0 0 0
1994 0 0 0 0
1995 0 0 0 1.7
B-132


Location:

1 Subgrade Soil Poorly Graded Sand (202) ---
Traffic Data:

1 Poorly Graded Sand 12

Type of Asphalt:
HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 99 59 44 31 3.5
ATB 84 51 44 25 3.5

1991 0 0 0 0 1998 0 0 0 0
1994 0 0 0 0 1999 0 0 0 0
1997 0 0 0 0 2000 0 0 0 0
B-133

LTPP Site Identification Number: 9034 State: Idaho (16)

Roadway or Route No.: 95
Location:

5 AC Seal Coat (71)
2 GB Crushed Stone Base (303) 18.8
1 Subgrade Soil Silty Gravel with Sand (265) ---
Traffic Data:
KESALS per year:

1 Poorly Graded Sand

Type of Asphalt:
HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1994 0 0 44.5 0.7 2004 17.3 0 0 4.8
1997 2.3 0 66.2 5.3
1998 2.3 0 68.8 3.5
2001 0 0 0 0
B-134

LTPP Site Identification Number: 2009 State: Indiana (18)

Roadway or Route No.: ST 37
Date of Construction: Jan. 1981 Status: Out of Service; 4-1999
Location:
Longitude: 86 Latitude: 40.03 Elevation: 785

6 AC Seal Coat, Slurry Seal (72) 0.5
4 ATB Dense Graded, Cold Laid, Plant Mix (326) 6.5
3 PATB Open Graded, Hot Mix, Hot Laid (323) 3.3
2 GB Gravel, Uncrushed (302) 9.5
1 Subgrade Soil Sandy Lean Clay (114) ---
Traffic Data:

2 Gravel (uncrushed) 100
1 Sandy Lean Clay 7

Type of Asphalt:
HMA ATB
Air Voids, % 6.6

#3/4 #3/8 #4 #40 #200
HMA 78 45 32 8.0 3.7
ATB

1993 3.7 0 0 44.7
1995 5.9 0 12 51.7
B-135

LTPP Site Identification Number: 0112 State: Iowa (19)

Date of Construction: May 1992 Status: In Service
Location:

6 HMA Hot Mixed, Hot Laid AC, Dense Graded 2.5
4 ATB HMAC, Hot Laid, Dense Graded 12.4
3 PATB Open-Graded Hot-Mix, Hot Laid; 4.1
2 GS Lean Clay with Sand 24
1 Subgrade Soil Clay ---
Traffic Data:
KESALS per year:

2 Lean Clay with Sand 6
1 Clay 5

HMA ATB
Air Voids, % 10.0 8.5

#3/4 #3/8 #4 #40 #200
HMA 92 57 39 5.3
ATB 96 68 47 6.2

1995 0 0 0 0 2005 0.3 0 16.4 12.3
1999 0 0 1.2 10.4
2001 0 0 5 17.3
2002 0 0 0 6.3
B-136


Location:


2 GS Embankment Soil; Clay with Gravel (104) 24
1 Subgrade Soil Clay with Gravel (104) ---
Traffic Data:
KESALS per year:

1 Clay 5

HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1995 15.0 0 0 0 2005 0.8 0 0 14.8
1999 7.5 0 7.3 15.3
2001 12.1 0 15.3 32.0
2002 1.3 0 0 9.9
B-137


Location:

3 ATB Dense Graded Asphalt Treated Base (319) 8.4
1 Subgrade Soil Clay with Sand (107) ---
Traffic Data:
KESALS per year:

1 Clay 5

HMA HMA ATB

Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA
ATB

1995 0 0 0 0 2005 9.8 0 0 12.5
1999 5.3 0 17.1 16.4
2001 5.4 0 21.4 34.6
2002 2.2 0 6.8 12.8
B-138


Location:

3 ATB Dense Graded Asphalt Treated Base (319) 4.7
1 Subgrade Soil Clay with Gravel (104) ---
Traffic Data:
KESALS per year:

1 Clay 5

HMA HMA ATB

Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA
ATB

1995 0 0 18.5 1.0 2005 1.1 0 0 23.0
1999 0 0 105.0 25.7
2001 2.1 0 103.4 26.7
2002 0.7 0 0 8.9
B-139

LTPP Site Identification Number: 0115 State: Louisiana (22)

Date of Construction: Nov. 1992 Status: In Service
Location:

2 GS Crushed Stone (131) 12.0
1 Subgrade Soil Lean Inorganic Clay (102) ---
Traffic Data:
KESALS per year:

2 Crushed Stone 8
1 Lean Inorganic Clay 5

Type of Asphalt:
HMA HMA ATB

Air Voids, % 5.4 2.0 4.8

#3/4 #3/8 #4 #40 #200
HMA 100 82 58 24 6.0
HMA 100 83 57 25 5.3
ATB

1998 0 0 0 0
1999 0 0 0 0
2004 0 0 0 0
B-140


Location:

2 GS Crushed Stone (131) 18.0
Traffic Data:
KESALS per year:

Layer Material/Soil Type Equivalent Resilient Modulus
No.
2 Crushed Stone 8

Type of Asphalt:
HMA HMA ATB

Air Voids, % 2.1 3.3 6.1

#3/4 #3/8 #4 #40 #200
HMA 100 82 58 24 6.0
HMA 100 83 57 25 5.3
ATB

1998 0 0 0 0
1999 0 0 0 0
2004 0 0 0 0
B-141


Location:

2 GS Embankment, Silty Clay with Sand (133) 12.0
Traffic Data:
KESALS per year:

No.
3 Crushed Stone 8

Type of Asphalt:
HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1998 0 0 0 0
1999 0 0 0 0
2004 0 0 0 0
B-142


Location:

4 ATB Hot Mixed, Hot Laid, Plant Mix (319) 10.6
3 PATB Open Graded, Hot Mix, Hot Laid (325) 3.6
2 GS Embankment; Silt (141) 30
Traffic Data:
KESALS per year:

2 Silt 8

Type of Asphalt:
HMA HMA ATB

Air Voids, % 5.3 2.9 6.8

#3/4 #3/8 #4 #40 #200
HMA 100 82 58 24 6.0
HMA 100 83 57 25 5.3
ATB

1998 0 0 0 0
1999 0 0 0 0
2004 0 0 0 0
B-143

LTPP Site Identification Number: 0116 State: Michigan (26)

Date of Construction: Aug. 1995 Status: Out of Service; 10-2002
Location:

2 ATB HMAC 12.0
1 Subgrade Soil Sandy Clay ---
Traffic Data:
KESALS per year:

1 Sandy Clay 4

Type of Asphalt:
HMA HMA ATB

Air Voids, % 5.0 2.7 4.8

#3/4 #3/8 #4 #40 #200
HMA 100 78 47
HMA 86 58 42 5.5
ATB 4.8

1995 0 0 0 0 2000 0 0 0 0
1996 0 0 0 0 2001 0 0 0 0
1998 0 0 0 0 2002 30.5 0 0 0
1999 0 0 0 0 2003 0 0 0 0
B-144

LTPP Site Identification Number: 0115 State: Michigan (26)

Date of Construction: Aug. 1995 Status: Out of Service; 10-2002
Location:

2 ATB HMAC 9.6
1 Subgrade Soil Sandy Clay ---
Traffic Data:
KESALS per year:

1 Sandy Clay 4

Type of Asphalt:
HMA HMA ATB

Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA
ATB

1995 0 0 0 0 2000 0 0 0 0
1996 0 0 0 0 2001 0 0 0 0
1998 0 0 0 0 2002 105.7 0 0 0
1999 0 0 0 0 2003 0 0 0 0
B-145

LTPP Site Identification Number: 0116 State: Montana (30)

Roadway or Route No.: Interstate 15
Date of Construction: Oct. 1997 Status: In Service
Location:

3 ATB Hot Mixed, Hot Laid AC, Dense Graded 12.6
2 SS Embankment Soil; A2-4, (SP-SM) 24
1 Subgrade Soil Poorly Graded Sand with Silt; A-2-6 ---
Traffic Data:
KESALS per year:

1 Subgrade Soil 12

Type of Asphalt: AC-10
HMA HMA ATB

Air Voids, % 7.5 6.0 5.5

#3/4 #3/8 #4 #40 #200
HMA 100 85 45 6.0
HMA 88 57 39 4.2
ATB 84 48 32 4.5

1998 0 0 0 0 2002 3 0 0 0
1999 0 0 0 0 2003 8.1 0 0 0
2000 0 0 0 0 2004 55.4 0 0 0
2001 0.4 0 0 0 2005 0 0 0 0
B-146


Location:

5
4 AC Seal Coat 0.2
2 GB Granular Base, Crushed Stone (303) 12.4
Traffic Data:
KESALS per year:

1 Subgrade Soil 12

HMA
Air Voids, % 7.5

#3/4 #3/8 #4 #40 #200
HMA

1998 0 0 0 0 2002 20.0 0 1.5 0
1999 0 0 0 0 2003 46.6 0 0 0
2000 5.9 0 0 0 2004 47.8 0 2.7 3.8
2001 9.0 0 5.1 0 2005 1.4 0 0 7.0
B-147


Location:

5
4 AC Seal Coat 0.2
Traffic Data:
KESALS per year:

1 Subgrade Soil 12

HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1998 0 0 0 0 2002 26.5 0 0 0
1999 0 0 0 0 2003 64.5 0 11.2 8.6
2000 0 0 0 0 2004 48.5 0 10.0 7.0
2001 23.1 0 0 0 2005 1.2 0 0 0
B-148


Location:

3 ATB HMAC, Hot Laid, Central Plant Mix (319) 13.7
2 PATB Open Graded, Hot Mixed, Hot Laid (323) 4.2
1 Subgrade Soil Poorly Graded Sand with Silt ---
Traffic Data:
KESALS per year:


HMA HMA ATB

Air Voids, % 7.5 6.0 5.5

#3/4 #3/8 #4 #40 #200
HMA 100 85 45 6.0
HMA 88 57 39 4.2
ATB 84 48 32 4.5

1998 0 0 0 0 2002 8 0 1.4 0
1999 0 0 0 0 2003 14.8 0 0 0
2000 0.6 0 3.3 0 2004 29.0 0 0 2.3
2001 4.3 0 5.4 0 2005 0.2 0 0 3.7
B-149

LTPP Site Identification Number: 0124 State: Nebraska (31)

Date of Construction: July 1995 Status: Out of Service; 9-2002
Location:

4 ATB HMAC, Hot Laid, Central Plant Mix 10.5
3 PATB Open Graded, Hot Mixed, Hot Laid 3.4
2 GS Lean Inorganic Clay 24
1 Subgrade Soil Lean Inorganic Clay ---
Traffic Data:
KESALS per year:

No.

HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 100 87 74 6.9
ATB 97 72 53 3.9

1995 0 0 0 0
1999 0 0 0 0
2002 0 0 0 0
B-150


Date of Construction: July 1995 Status: In Service
Location:

3 ATB HMAC 8.6
2 GS Lean Inorganic Clay; A-6 24
1 Subgrade Soil Lean Inorganic Clay; A-7-5 ---
Traffic Data:
KESALS per year:


Type of Asphalt:
HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 100 87 74 6.9
ATB 97 72 53 3.9

1995 0 0 0 0
1999 0 0 0 0
2002 0 0 0 0
B-151


Location:

4 HMA Hot Mixed, Hot Laid AC, Dense Graded; (1) 4.1
2 GS Lean Inorganic Clay; A-6 24.0
1 Subgrade Soil Lean Inorganic Clay; A-7-6 ---
Traffic Data:
KESALS per year:

No.

HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 100 87 74 6.9
ATB 95 68 50 2.2

1995 0 0 0 0
1999 0 0 0 0
2002 0 0 0 0
B-152

LTPP Site Identification Number: 0104 State: Nevada (32)

Date of Construction: Aug. 1995 Status: In Service
Location:

4 ATB Hot Mixed, Hot Laid, Plant Mix 12.4
3 GS Soil Agg. Mix. 18.4
2 TS Lime Treated Soil 12.0
1 Subgrade Soil Silty Sand ---
Traffic Data:

2 Soil Agg. Mix. 28
1 Silty Sand 10

HMA ATB
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA 96 73 62 21 5.6
ATB

1996 0 0 0 0 2002 0 0 0 0
1998 0 0 18.8 0 2003 0 0 0 0.9
1999 5 0 0 0 2004 0 0 0 1.2
2000 0 0 0 0 2005 0 0 0 3.1
2001 0 0 0 0 2006 0 0 0 5.6
B-153


Location:

4 GB Crushed Gravel (304) 8.5
3 GS Soil Agg. Mix, predominately coarse grained (308) 22.8
2 TS Lime Treated Soil (338) 12.0
Traffic Data:

3 Soil Agg. Mix. 28
1 Silty Sand 10

HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA

1997 0 0 0 0 2002 1.2 0 10.6 0
1998 0 0 7.3 0 2003 0 0 12.0 0
1999 0 0 0 0 2004 0.9 0 0 1.3
2000 0.5 0 0.3 0 2005 1.6 0 0 2.5
2001 1.4 0 0 0 2006 2.0 0 0 6.5
B-154


Location:

3 GS Soil Agg. Mix.; coarse grained (308) 24.5
1 Subgrade Soil Clayey Sand (216) ---
Traffic Data:

3 Soil Agg. Mix. 28
1 Clayey Sand 10

HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA

1997 0 0 0 0 2002 28.8 0 0 0
1998 0 0 0 0 2003 32.9 0 0 0
1999 0 0 0 0 2004 34.1 0 0 0
2000 33.8 0 0 0
2001 16.5 0 0 0
B-155


Location:

Traffic Data:

3 Soil Agg. Mix. 28
1 Silty Sand 10

HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1997 0 0 0 0 2002 0 0 1.5 0
1998 0 0 0 0 2003 1.3 0 0 0
1999 0.4 0 0 0 2004 1.5 0 0 0
2000 16.4 0 0 0
2001 10.4 0 0 0
B-156


Location:

Traffic Data:

3 Soil Agg. Mix. 28
1 Clayey Sand 10

HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1997 0 0 0 0 2002 0 0 1.3 1.0
1998 0 0 0 0 2003 0.2 0 0 1.3
1999 0 0 0 0 2004 0.2 0 0 1.7
2000 0 0 0 0 2005 0.3 0 0 2.3
2001 0 0 0 1.2 2006 0.5 0 0 3.9
B-157


Location:

4 PATB Open Graded, Hot Mixed, Hot Laid 4.2
3 GS Soil Agg. Mix. 15.1
1 Subgrade Soil Clayey Sand ---
Traffic Data:

2 Soil Agg. Mix. 28
1 Clayey Sand 10

HMA ATB
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA 95 70 54 16 5.0
ATB

1996 0 0 0 0 2002 0 0 0 1.1
1998 0 0 0 0 2003 0 0 0 1.1
1999 0 0 0 0 2004 0 0 0 1.1
2000 0 0 0 0 2005 0.3 0 0 4.8
2001 0 0 0 1.1 2006 0.3 0 0 4.6
B-158

LTPP Site Identification Number: 0106 State: New Mexico (35)

Location:

5 AC Friction Course 0.6
3 ATB Hot Mixed, Hot Laid AC, Dense Graded 8
2 GB Crushed Stone 2.9
1 Subgrade Soil Sandy Fat Clay ---
Traffic Data:

No.
2 Crushed Stone 25
1 Sandy Fat Clay 5.5

Type of Asphalt:
HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 95 71 53 17 5.2
ATB 97 78 57 19 4.5

1997 0 0 0 0 2003 0 0 3.5 0
1999 0 0 0 0 2004 0 0 3.6 0
2000 0 0 0 0 2005 0 0 3.6 0
2001 0 0 0 0 2006 0.6 0 5.1 0
2002 0 0 3.5 0
B-159


Location:

5
1 Subgrade Soil Fat Inorganic Clay (103) ---
Traffic Data:

No.
1 Fat Inorganic Clay 5.5

Type of Asphalt:
HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1997 0 0 0 0 2003 0.1 0 12.5 0
1999 0 0 0 0 2004 0.2 0 20.4 0
2000 0 0 0 0 2005 0.3 0 26.7 0
2001 0 0 0 0 2006 0.3 0 36.2 0.4
2002 0.1 0 11.0 0
B-160


Location:

3 ATB Open Graded, Hot Mixed, Hot Laid AC (325) 3.7
1 Subgrade Soil Sandy Fat Clay with Sand (109) ---
Traffic Data:

No.
1 Fat Clay with Sand 5.5

Type of Asphalt:
HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1997 0 0 0 0 2003 0.5 0 33.0 0
1999 0 0 0 0 2004 1.0 0 46.6 1.7
2000 0 0 0 0 2005 1.0 0 48.1 1.8
2001 0 0 0 0 2006 1.2 0 56.4 6.0
2002 0 0 22.6 0
B-161


Location:

5 AC Friction Course 0.6
2 PATB Open Graded Mix 3.7
1 Subgrade Soil Clayey Sand ---
Traffic Data:

1 Clayey Sand 10

Type of Asphalt:
HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 95 73 51 16 4.0
ATB 97 78 57 19 4.5

1997 0 0 0 0 2003 0 0 0 0
1999 0 0 0 0 2004 0 0 0 0
2000 0 0 0 0 2005 0 0 0 0
2001 0 0 0 0 2006 0 0 0 0
2002 0 0 0 0
B-162

LTPP Site Identification Number: 0106 State: Ohio (39)

Location:

3 ATB Hot Mixed, Hot Laid AC, Plant Mix 7.9
2 GB Crushed Stone 3.9
1 Subgrade Soil Silty Clay ---
Traffic Data:
KESALS per year:

2 Crushed Stone 10
1 Silty Clay 5

Type of Asphalt:
HMA HMA ATB

Air Voids, % 10.4 6.8 14.6

#3/4 #3/8 #4 #40 #200
HMA 100 88 52 13 5.9
HMA 89 61 44 10 5.0
ATB 67 54 37 12 7.0

1996 0 0 0 0 2002 62.8 0 223.4 0
1997 0 0 0 0 2004 204.1 0 0 0
1999 0 0 9.5 0 2005 274.1 290 0 0
2001 17.7 0 201.6 0
B-163


Location:

3 ATB Hot Mixed, Hot Laid AC, Plant Mix 11.8
2 PATB Open Graded, Plant Mix 4.0
Traffic Data:
KESALS per year:

1 Silty Clay 5

Type of Asphalt:
HMA HMA ATB

Air Voids, % 11.3 7.6 5.0

#3/4 #3/8 #4 #40 #200
HMA 100 89 53 12 5.7
HMA 94 74 54 12 6.1
ATB 62 49 33 12 7.3

1996 0 0 0 0 2004 138.3 0 0 0
1999 0 0 3.5 0 2005 244.0 320 0 0
2001 0 0 107.5 0
2002 20.5 0 37.8 0
B-164


Location:

4 ATB Asphalt Treated Mixture, Plant Mix 12.0
3 PATB Open Graded, Hot Laid, Central Plant Mix 3.7
2 GS Crushed Stone 6
Traffic Data:
KESALS per year:

3 Open Graded, Hot Laid, Central Plant Mix 25
2 Crushed Stone 80
1 Silty Clay 8

Type of Asphalt:
HMA ATB

#3/4 #3/8 #4 #40 #200
HMA 100 88 61 4.6
ATB 100 78 58 4.8

1996 0 0 0 0 2004 1.3 0 0 0
1999 0 0 0 0
2001 0 0 4.9 0
B-165


Location:

4 ATB Asphalt Treated Mixture, Plant Mix 12.0
3 PATB Open Graded, Hot Laid, Plant Mix 3.7
2 GS Crushed Stone 6.0
Traffic Data:
KESALS per year:

No.
3 Open Graded, Hot Laid, Central Plant Mix 25
2 Crushed Stone 80
1 Silty Clay 8

Type of Asphalt:
HMA ATB
Air Voids, % 12.8 11.4

#3/4 #3/8 #4 #40 #200
HMA 100 86 51 4.7
ATB 100 67 49 7.0

1996 0 0 0 0 2004 154.1 0 0 0
1999 0 0 0 0
2001 0 0 123.8 0
B-166

LTPP Site Identification Number: 0115 State: Oklahoma (40)

Location:
Longitude: 98.66 Latitude: 34.64 Elevation:

3 ATB Asphalt Treated Base, Plant Mix 9.0
1 Subgrade Soil Clayey Sand 144
Traffic Data:
KESALS per year:

1 Clayey Sand 10

Type of Asphalt: PG-6422
HMA HMA ATB

Air Voids, % 4.5 3.0 4.5

#3/4 #3/8 #4 #40 #200
HMA 86 63 46 5.3
HMA 82 55 41 7.6
ATB 79 52 38 8.9

1997 0 0 0 0 2001 0 0 0 3.8
1997 0 0 0 0 2002 0 0 0 3.8
1999 0 0 0 0 2003 0 0 0 3.8
2000 0 0 0 2.2
B-167


Location:

6
3 GB Granular, Crushed Stone (303) 11.3
1 Subgrade Soil Clayey Sand (216) 6
Traffic Data:
KESALS per year:

3 Crushed Gravel 25
1 Clayey Sand 10

HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1997 0 0 0 0 2002 3.9 0 0 0
1999 0 0 0 0 2003 12.6 0 0 0
2000 0 0 0 0 2004 22.5 0 0 0.6
2001 0 0 0 0 2006 32.3 0 0 5.0
B-168


Location:

3 GB Granular, Crushed Gravel 3.6
1 Subgrade Soil Clayey Sand 72
Traffic Data:
KESALS per year:

3 Crushed Gravel 25
1 Clayey Sand 10

HMA HMA ATB

Air Voids, % 8.0 5.0 3.0

#3/4 #3/8 #4 #40 #200
HMA 100 82 63 4.1
HMA 85 50 47 4.8
ATB 82 55 41 8.3

1997 0 0 0 0 2002 0 0 0 0
1999 0 0 0 0 2003 0.5 0 0 0
2000 0 0 0 0
2001 0 0 0 0
B-169


Location:

4 ATB Asphalt Treated Base, Plant Mix (319) 4.0
3 GB Granular, Crushed Stone (303) 5.4
Traffic Data:
KESALS per year:

3 Crushed Gravel 25
1 Clayey Sand 10

HMA HMA ATB

Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA
ATB

1997 0 0 0 0 2002 0 0 0 0
1999 0 0 0 0 2003 0 0 0 0
2000 0 0 0 0 2004 0 0 0 0
2001 0 0 0 0 2006 0 0 0 0
B-170

LTPP Site Identification Number: 0116 State: Texas (48)

Date of Construction: April 1997 Status: Out of Service; 4-2002
Location:

1 Subgrade Soil Poorly Graded Sand With Silt ---
Traffic Data:
KESALS per year:

1 Poorly Graded Sand With Silt 12

Type of Asphalt:
HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1997 0 0 0 0 2002 0 0 0 0
1998 0 0 0 0 2002 0 0 0 0
1999 0 0 0 0 2003 0 0 0 0
2000 0 0 0 0 2004 0 0 0 0
2001 0 0 0 0 2005 0.4 0 0 0
B-171


Date of Construction: April 1997 Status: Out of Service; 4-2002
Location:

1 Subgrade Soil Poorly Graded Sand with Silt ---
Traffic Data:
KESALS per year:


Type of Asphalt:
HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1997 0 0 0 0 2002 0 0 0 0
1998 0 0 0 0 2003 0 0 0 0
1999 0 0 0 0 2004 0 0 0 0
2000 0 0 0 0 2005 0 0 0 0
2001 0 0 0 0
B-172


Roadway or Route No.: SH 175
Date of Construction: 7-1-1977 Status: Out of Service; 7-2003
Location:

3 ATB Other; Treated Layer 13.5
1 Subgrade Soil Fat Inorganic Clay ---
Traffic Data:

1 Fat Inorganic Clay 6

Type of Asphalt:
HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1991 0 0 46 82 2000 1.8 216.7 25.6 215.2
1993 0 0 53.2 85.7 2002 1.5 220.8 23.6 215.0
1995 0 0 77.9 147.6 2003 1.7 211.6 16.1 245.9
1998 3.2 320.3 3 30.9
B-173


Date of Construction: 11-1-1974 Status: Out of Service; 8-1996
Location:

3 HMA Hot Mixed, Hot Laid, Dense Graded Mix (1) 1.2
2 ATB Open Graded, Plant Mix (319) 9.8
1 Subgrade Soil Coarse Grained Soil (215) ---
Traffic Data:
KESALS per year: 20

No.
1 215 ---

Type of Asphalt:
HMA ATB
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
ATB

1991 0 0 0 56.4
1993 0 0 35.1 61.1
1995 0.7 0 46.2 63.5
1996 7 7.5 7 51.9
B-174

LTPP Site Identification Number: 0124 State: Wisconsin (55)

Date of Construction: Nov. 1997 Status: Out of Service in 2008
Location:

2 GS Soil Agg. Mix.; A-1-b 8.0
1 Subgrade Soil Silty Sand; A-1-b ---
Traffic Data:
KESALS per year:

2 Soil Agg. Mix. 14
1 Silty Sand 10

HMA HMA ATB

Air Voids, % 7.5 6.2 5.9

#3/4 #3/8 #4 #40 #200
HMA 99 86 66 3.5
HMA 99 86 66 3.5
ATB 92 53 37 3.0

1998 0 0 0 0 2005 0 0 0 0
2000 0 0 0 0
2002 0 0 0 0
2004 0 0 0 0
B-175


Location:

2 GB Granular Base; A-1-a 10.8
Traffic Data:
KESALS per year:

2 Granular Base, Soil Agg. Mix. 14
1 Silty Sand 10

HMA HMA ATB

Air Voids, % 5.1 7.3 6.6

#3/4 #3/8 #4 #40 #200
HMA 100 88 69 6.5
HMA 99 80 57 3.9
ATB 92 53 37 3.0

1998 0 0 0 0 2005 0 0 0 0
2000 0 0 0 0
2002 0 0 0 0
2004 1.61 0 0 0
B-176


Location:

2 GB Granular Base; A-1-b 14.2
1 Subgrade Soil Silty Sand; A-1-a ---
Traffic Data:
KESALS per year:

1 Silty Sand 10

HMA HMA ATB

Air Voids, % 6.6 7.2 6.1

#3/4 #3/8 #4 #40 #200
HMA 99 86 66 3.5
HMA 99 86 66 3.5
ATB 88 68 58 4.1

1998 0 0 0 0 2005 0 0 0 0
2000 0 0 0 0
2002 0 0 0 0
2004 0 0 0 0
B-177


Location:

2 GS Soil Agg. Mix.; A-1-b (308) 10.0
1 Subgrade Soil Poorly Graded Sand with Gravel & Silt (205) ---
Traffic Data:
KESALS per year:

2 Soil Agg. Mix. 14
1 Poorly Graded Sand 10

HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1998 0 0 0 0 2005 0 0 0 0
2000 0 0 0 0
2002 0 0 0 0
2005
B-178

LTPP Site Identification Number: C903 State: Wisconsin (55)

Location:

3 GB Crushed Stone Base; A-1-a 13.0
2 GS Embankment; Coarse-Fine soil; A-3 5.0
Traffic Data:
KESALS per year:

1 Silty Sand 10

HMA HMA

#3/4 #3/8 #4 #40 #200
HMA 100 73 42 3.7
HMA 99 69 51 3.4

1998 0 0 0 0 2005 5.2 0 0 1.0
2000 0 0 0 0
2002 0 0 0 0
2004 2.9 0 0 0
B-179


Location:

3 GB Crushed Stone Base; A-1-a (303) 13.0
2 GS Embankment; Coarse-Fine soil; A-3 (210) 24.0
1 Subgrade Soil Well Graded Sand with Silt & Gravel (211) ---
Traffic Data:
KESALS per year:

1 Well Graded Sand 10

HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1998 0 0 0 0
2000 0 0 0 0
2002 0 0 0 0
2004 0 0 0 0
B-180


Location:

3 GB Crushed Stone Base; A-1-a (303) 13.0
2 GS Embankment; Coarse-Fine soil; A-3 (210) 5.0
1 Subgrade Soil Silty Sand; A-1-b (214) ---
Traffic Data:
KESALS per year:

1 Silty Sand 10

HMA HMA
Asphalt Content, %
Air Voids, %

#3/4 #3/8 #4 #40 #200
HMA
HMA

1998 0 0 0 0
2000 52.4 0 0 0
2004 262.6 0 0 0
B-181

NCHRP 9-44

Uploaded by

Copyright:

Available Formats

NCHRP 9-44

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

NCHRP 9-44

Uploaded by

Copyright:

Available Formats

THE NATIONAL ACADEMIES PRESS

This PDF is available at http://nap.edu/23277 SHARE

An Experimental Plan for Validation of an Endurance Limit

265 pages | 8.5 x 11 | PAPERBACK

GET THIS BOOK

FIND RELATED TITLES

National Academies of Sciences, Engineering, and Medicine 2009. An Experimental

– Access to free PDF downloads of thousands of scientiﬁc reports

Copyright © National Academy of Sciences. All rights reserved.

This work was sponsored by the American Association of State

Authors herein are responsible for the authenticity of their materials

Cooperative Research Programs (CRP) grants permission to

The opinion and conclusions expressed or implied in the report are

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Figure 1. HMA Endurance Limit Workshop Agenda.................................................................... 5

Copyright National Academy of Sciences. All rights reserved.

Table 1. HMA Endurance Limit Workshop Invitees and Attendees. ............................................ 4

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

1. INTRODUCTION AND RESEARCH APPROACH

1.2 Problem Statement and Research Objective

Copyright National Academy of Sciences. All rights reserved.

1. Validation of the existence of an endurance limit for HMA in pavements through an

1.3 Research Approach

1.3.1 Task 1. Review Relevant Research.

• Laboratory endurance limit studies,

Copyright National Academy of Sciences. All rights reserved.

1.3.2 Task 2. Conduct Facilitated Workshop.

• Endurance limit and other important fatigue effects,

Copyright National Academy of Sciences. All rights reserved.

Table 1. HMA Endurance Limit Workshop Invitees and Attendees.

No Name Affiliation Role Attend

Copyright National Academy of Sciences. All rights reserved.

AGENDA Item Questions to be answered

Copyright National Academy of Sciences. All rights reserved.

1.3.3 Task 3. Identify Data Requirements

Copyright National Academy of Sciences. All rights reserved.

Table 2. Summary of Existing Pavement Analysis Approaches Considered.

Approach Key Elements Selected Advantages Disadvantages

Carpenter (4) Above endurance limit fatigue life

Table 3. Summary of Proposed Laboratory Experiments.

Experiment Topic Factors

Copyright National Academy of Sciences. All rights reserved.

1.3.4 Task 4. Identify Applicable Projects.

Copyright National Academy of Sciences. All rights reserved.

Copyright National Academy of Sciences. All rights reserved.

advantage of jackknifing is the goodness of fit statistics are based on predictions of

1.3.5 Task 5. Prepare Detailed Work Plan.

Copyright National Academy of Sciences. All rights reserved.

1.3.6 Task 6. Prepare Final Report.

Copyright National Academy of Sciences. All rights reserved.

2.1 HMA Endurance Limit

In response to increasing interest in perpetual pavements, a substantial amount of laboratory

Copyright National Academy of Sciences. All rights reserved.

Figure 2. Results of Flexural Fatigue Tests by Carpenter et al., Including Extrapolated