DOT/FAAIRD-91/15, I

DOT-VNTSC-FAIJ1 -91-7 Unified

PavementMethodology forDesign
Analysis and Airport

Research and Vol. I-State of the Art

Development Service
Washington, D.C. 20591

AD-A238 812

John Zaniewski

Arizona State University

Temple, AZ 85287

June 1991
S D I
JUL29GETE
1981 -

This document is available to the public

through the N'itional Technical Information
Service, Springfield, Virginia 221bl.

I UTION 3TATKI A
Dbstzutihm UUimbted w
U.S DepartmenI of Transportaton
rc .1u Aviutfon .dm!n'stra! ,,

91-05839
91l!lllilllll8
 NOTICE

This document is disseminated under the sponsorship of the

Department of Transportation in the interest of information
exchange. The United States Government assumes no
liability for its contents or use thereof.

NOTICE
The United States Government does not endorse
products or manufacturers. Trade or manufacturers'
names appear herein solely because they are considered
essential to the object of this report

• n~um mn m ml no n Im nnM MM nm iMMMMM

 Technical Report Documentation Page
1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.

DOT/FAA/RD-91/15, I

4. Title and Subtitle 5. Report Date

June 1991
Unified Methodology for Airport Pavement Analysis and
6. Performing Organization Code
Design
Vol. I State of the Art DTS-77
8. Performing Organization Report No.

7. Author s) DOT-VNTSC-FAlJl-91-7
John P. Zaniewski*
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
U.S. Department of Transportation FAIJl/A1063
,John A. Volpe National Transportation Systems Center 1. ContractorGrantNo.
Research and Special Programs Administration
Cambridge, MA 02142 13. Type of Report and Period Covered
I.. Spornboring Ae..cy N.,me on; ',dress Final Report
U.S. Department of Transportation January 1989 - January 1990
Federal Aviation Administration
_ Research and Development Service 14. Sponsoring Agency Code
Washington, DC 20591 ARD-240
15. Supplementary Notes

*Arizona State University

Temple, AZ 85287
16. Abstract
This report presents an assessment of the state of the art of airport pavement analysis
and design. The objective is to identify those areas in current airport pavement
analysis methodology that need to be substantially improved from the perspective of
airport pavement design and management needs. The foundations of current design
practice are examined with emphasis on the last Lwenty years of research and
advancement in the area of pavement response prediction and cross section/layer
thickness design. The author presents a review of empirical methods of design for
rigid and flexible pavement design, and quasi-mechanistic analyses that are based on
layercd-elastic theory or finite element methods. Weakness of current methods as
applied to airport operational needs are discussed. The report presents a rational
argument for developing a unified pavement analysis and design procedure that can be
used for pavements of any material type and that are based on mathematical formulations
of the actual stress/strain response processes in airport pavement mate-ials.

17. Key Words 18. Ostrbution Statement

Airport Pavetnlents, Pavement Analysis I DOCUMENT IS AVAILABLE T3 THE PjdLtL
P"Vt-ineret. Design THROUGH THE NATIONAL TECHNICAL
INFORMATION SERVICE. SPRINGFIELD.
VIRGINIA 22161

19. Se ry Closf.(of rh-Isreport) 20, Security Clossil. (of this page) 21. No. of Pages 22. Pr..c

Uo c i7ssifi
it-d Unclassified 14)

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

 PREFACE

Airport pavements are designed for a nominal life of twenty

years; but most need major rehabilitation, reconstruction, or
replacement before twelve to fourteen years have elapsed. Part
of the problem stems from design procedures that are empirically
based, not representing well the response of airport pavement to
load or appropriate failure mechanisms. It is not possible with
the existing pavement models to assess accurately the effect of
incorporating new materials into pavement designs or the impact
of changes in traffic - either increased number or weight of
aircraft. Nearly one billion dollars is spent yearly in
building, improving or maintaining airport runways, taxiways, and
aprons. Producing a more accurate model of pavement response and
design procedures in order that pavements achieve their design
lives can result in significant savings.

This report reviews the state of the art of airport pavement

analysis and design. The specific objective is to determine what
areas need to be improved to meet the FAA's goals of improving
the reliability of pavement analysis and design, and developing a
tool that can be used to realistically study the effect of new
pavement materials and changes in traffic.

A general framework for the design of airport pavements is

portrayed. A more detailed examination of recent literature and
research in analytical models is presented including a summary of
elastic layer theory, viscoelastic layer analysis, thin plate
thecry, and numerical methods (including finite element
representations). Available pavement distress models are also
described and assessed. Finally, airport pavement design methods
are reviewed.

It is concluded that accurate estimates of pavement response

and subsequently of pavement distress and performance will
require a new approach. This approach should be founded on
three-dimensional finite element analysis, formulated with the

iii
specific characteristics of pavement materials, i.e., the
conqtitutive equations of the materials, with respect to elastic,
plastic, and viscous behavior, properly represented. The task is
a formidable one; but with recent advances in fracture mechanics
and numerical analysis, and particularly with the advances in
computational capabilities of personal computers, it is a task
that can be successfully performed over the next several years.

This report was prepared for the Infrastructure Systems and

Technology Division, John A. Volpe National Transportation
Systems Center, Research and Special Programs Administration.
The effort was sponsored by the Federal Aviation Administration,
System Technology Division, Research and Development Service and
the technical monitor was Dr. Aston McLaughlin.

Accession For

NTIS GRA&l
DTIC TAB
Unannounced 0
justificat0io

By
- ~ DI!,tribuItiouf

" st siaf1c o or

iv
 METRIC/ ENGLISH CONVERSION FACTORS

ENGLISH TO METRIC METRIC TO ENGLISH

LENGTH tAPPROXIMATE) LENGTH CAPPRXIMATE)

I inch (in) = 2.5 centimeters (cm) 1 millimeter (mm) = 0.04 inch (in)
1 foot (ft) = 30 centimeters (cm) I centimeter (cm) 0 inch (in)
0.4
1 yard (yd) = 0.9 meter (m) 1 meter (m) = 3.3 feet ft)
I mile (mi) = 1.6 kiiom 2ters (km) I meter (m) = 1.1 yards (yd)
1 kilometer (kin) = 0.6 mile (ni)

AREA (APPPCXWMATE AREA (APXROXWATEJ

2
1 square inch (sq ii, irs) - 6.5 square centimeters (cm,) I square centimeter (cm ) = 0.16 square inch (sq in, in-)
2
I square foot (sq ,t, f-t ) = 0.09 square meter (m2) I square meter (m) = 1.2 square yards (sq yd, ydz)
-
square yard (sq yd, yd -) = 0.8 square meter (m ) 1 square kilometer (km) = 0.4 square mile (sq mi, mi2)
2
1 square mile (sq mi, mi) = 2.6 square kilometers (km ) 1 hectare (he)= 10,000 sq-zare meters (m2) = 2.5 acres
1 acre = 0.4 hectares (he) = 4,000 square meters (m)

MASS - WEIGHT (APPROX:ATE) MASS - WEIGHT (APPAOXIMAE)

1 ounce (oz) = 28 grams (gr) 1 gram (gr) = 0.036 ounce (cz)
1 pound (Ib) = .45 kilogram (kg) i kilogram (kg) = 2.2 pounds (lb)
1 shon ton = 2,000 pounds (Ib) = 0.5 tonne (t) I tonne (t) = 1,000 kilograms (kg) = 1.1 shcrt tons

VOLUME (-0 xiMA7c) VOLUME ,.P'c;'CX A7'.E

I teaspoon (tsp) = m.illiliters (ml) 1 milliliter (ml) = 0,03 fluid ource fcz)
1 tablespoon (tbsp) 15 milliliters (ml) 1 liter (1) = 2.1 pints (pt)
1 fluid ounce ri cz) = 30 milliliters (ml) 1 liter (I) = 106 quar-ts(qI)
1 cup (c) = 0.24 liter (1) 1 liter C) = 0.26 gallon (ga)
1 pint (pt) = 1.47 liter (I) 1 cubic meter (ins) = 35 cubc feet (cu f-t, h-)
1 quart (qt) = l.6liter (I) 1 cubic meter (m?) = 1.3 cubic yards (cu yd, yd')
1 Sallon (ga!) = 3.8 liters (I)
1 cubic foot (cu ft, ft!) = 0.03 -ibic meter (m )
-
I cubic yard (cu yd, yd ) = 0.76 cubic meter (m)

TEMPERATURE ExAc-7 TEP",PERATURE (ExAci

[(x-32)(5'9)]PF y'C [(9'5) y+32]'C = x'F

QUICK INCH-CENTiMETER LENGTH CONVERSION

INChE S 0 1 2 3 4 5 6 7 8 9 .0

CENT. £'-,S 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
25 1

QUICK FAHRENHEIT-CELCIUS TEMPERATUFE CONVERSION

* °
22 ,' -4 141 32' 50 683 l:S' 104' 122 1.20' 15 176' 154' 2'2'

... -,'.. ,'-01*0. 0 104'", -1 ]0 + 60" ,

70* , 0 ,'J*
1031

Fcr- ,e e) l a J(4 " .cC, .o n f42 0 s, ssEc ,ES e aneous Pub lCzt;cn 2S6. U rts cf V ' g ,s and
ce'sc
%'tI SL" ;-:,e S2 . SD C2- No. C 13 1 2 6.

17
 TABLE OF CONTENTS

Section Page

1. THE AIRPORT PAVEMENT PROBLEM................1

1.1 Airport Pavement Types.................3

1.1.1 Rigid Pavements.................3

1.1.2 Flexible Pavements...............6

1.2 Pavement Materials..................6

1.2.1 Fundamental Material Characterization . . . 7

1.2.2 Steel .................................. 12
1.2.3 Portland Cement Concrete............12
1.2.4 Asphalt Concrete................14
1.2.5 Stabilized Mateials.............19
1.2.6 Soils and Granular Materials ......... 19

1.2.6.1 Cohesive Soils............21

1.2.6.2 Cohesionless Soils...........21
1.2.6.3 Untreated Granular Materials . . . . 23

1.3 Aircraft Characteristics...............26

1.4 Functional Areas of Airport Pavements ......... 29
1.5 Environmental Effects.................33
1.6 Interaction Effects.................34

2. A GENERAL FRAMEWORK FOR THE DESIGN OF AIRPORT

PAVEMENTS........................37

3. ANALYSIS MODELS......................47

3.1 Elastic Layer Theory - Static Loads ......... 49

3.2 Viscoelastic Layer Analysis.............55
3.3 Layer Analysis - Dynamic Loads............59
3.4 Thin Plate Theory...................63
3.5 Numerical Methods...................64
3.6 Environmental Models................ 4

4. DISTRESS MODELS......................79

4.1 Fracture.........................79

4.1.1 Phenomcnological Model.............0

4.1.2 Power Law ................... 82
4.2 Deformations....................83

Vii
 TABLE OF CONTENTS (cont.)

Section Page

5. AIRPORT PAVEMENT DESIGN METHODS .... ............. . 87

5.1 Federal Aviation Administration .. .......... . 87

5.1.1 FAA Flexible Pavement Design .. ........ 89

5.1.2 FAA Rigid Pavement Design . . ......... 90

5.2 Asphalt Institute ...... ................. 91

5.3 Portland Cement Association .... ............ 93
5.4 Department of Defense Pavement Design Methods . 96

5.4.1 Triservice Manual for Flexible Pavement

Design ....... ................... 96
5.4.2 Army-Air Force Rigid Pavement Design . . . 101

5.4.2.1 Plain Concrete Airfield Pavements . 102

5.4.2.2 Reinforced Concrete Pavements . . .. 102
5.4.2.3 Fibrous Concrete Pavements ..... .103
5.4.2.4 Prestressed Pavement Design ...... .. 104

5.4.3 Navy Rigid Pavement Design ... ......... .104

5.4.4 Army-Air Force Flexible Pavement Design -
Elastic Layer Theory Method ... ........ 105

6. DISCUSSION AND RECOMMENDATIONS ...... ............. ill

6.1 Discussion of the State Of The Art in Airport

Pavement Analysis ...... ................. 112
6.2 Recommendations for Improving the State Of The Art
in Airport Pavement Design and Analysis ....... .. 115

LIST OF SYMBOLS .......... ....................... 119

REFERENCES ........... ......................... R-1

viii
 LIST OF ILLUSTRATIONS

Figure Page

1-1 PAVEMENT TYPES: (A) FLEXIBLF AND (B) RIGID ... ...... 4

1-2 DEPENDENCE OF STIFFNESS ON ASPHALT CONCRETE TEMPERATURE

AND TIME OF LOADING ....... ................. 17

1-3 RANGE OF STIFFNESS UNDER MOVING TRAFFIC FOR ASPHALT

CONCRETE DEPENDING ON ENVIRONMENT, 12-INCH THICK
ASPHALT BOUND LAYER ....... ................. 17

1-4 TYPICAL PLOT OF STRESS AND STRAIN VERSUS TIME DURING

THE COMPLEX (DYNAMIC) MODULUS TEST .... ........... 20

1-5 TYPICAL PLOT OF LOAD AND DEFORMATION VERSUS TIME DURING

RESILIENT MODULUS TEST ....... ................. 20

1-6 TYPES OF GEAR ASSEMBLIES ON CIVILIAN AIRCRAFT . ... 28

1-7 AIRCRAFT LANDING GEAR-TRUCK TYPES ... .......... 28

1-8 TYPE IMPRINT AREA ........ .................. 30

1-9 AIRCRAFT GROUND OPERATION CYCLE .... ........... 30

1-10 A STEREOGRAM OF CALCULATED MONTHLY AND HOURLY

CHANGE OF FATIGUE DAMAGE OF ASPHALT PAVEMENT AT
BOTTOM OF ASPHALT MIX LAYER ..... ............. 36

1-11 A STEREOGRAM OF CALCULATED MONTHLY AND HOURLY

CHANGES OF 'ATIGUE DAMAGE OF ASPHALT PAVEMENT AT TOP OF
ASPHALT MIX LAYER ........ .................. 36

2-1 MAJOR CLASSES OF ACTIVITIES IN A PAVEMENT

MANAGEMENT SYSTEM ........ .................. 38

2-2 MAJOR PAVEMENT DESIGN COMPONENTS ... .......... 40

2-3 BLOCK DIAGRAM OF PAVEMENT DESIGN SYSTEM . ....... 41

2-4 COMPONENTS OF A MECHANISTIC PAVEMENT DESIGN

PROCEDURE .......... ...................... 42

2-5 FLOW CHART FOR ESTIMATING FATIGUE DAMAGE OF

ASPHALT PAVEMENTS ........ .................. 43

Jx
 LIST OF ILLUSTRATIONS

Figure Page

3-1 GENERALIZED MULTILAYER ELASTIC SYSTEM .. ........ 50

3-2 SPRING-DASHPOT SINGLE DEGREE OF FREEDOM PAVEMENT

MODEL ........... ........................ 61

3-3 COUPLED MECHANICAL MODEL WITH SEVEN PARAMETERS . . .. 61

3-4 AXISYMMETRIC FINITE ELEMENT MODEL OF TWO LAYER SYSTEM 69

3-5 DEFLECTIONS OF A PAVEMENT SLAB FOR THE (A) WINKLER

AND (B) ELASTIC CONTINUUM SUPPORT CONDITIONS ..... 72

3-6 FINITE ELEMENT MODEL OF A THREE SLAB PAVEMENT SYSTEM . 73

3-7 STRESSES RESULTING FROM CONTRACTION DUE TO TEMPERATURE:

(A) FORCES ACTING ON CONTRACTING SLAB; (B) VARIATION OF
SUBGRADE RESISTANCE WITH LENGTH .... .......... 76

5-1 FLOW OF ASPHALT INSTITUTE DESIGN PROCEDURE ....... 92

5-2 AIR FORCE TRAFFIC AREAS, LIGHT-LOAD AIRFIELDS . . . . 97

5-3 AIR FORCE TRAFFIC AREAS, MEDIUM AND MODIFIED HEAVY

LOAD AIRFIELDS ......... .................... 98

5-4 AIR FORCE TRAFFIC AREAS, HEAVY LOAD AIRFIELDS . . . . 99

5-5 ELASTIC LAYER THEORY DESIGN !aBTHOD, CONVENTIONAL

PAVEMENTS .......... ...................... 106

5-6 ELASTIC LAYER THEORY DESIGN METHOD, PAVEMENTS WITH

STABILIZED BASES ........ ................... 106

5-7 ELASTIC LAYER THEORY DESIGN METHOD, PAVEMENTS WITH

STABILIZED SUBBASES ........ .................. 107

x
 LIST OF TABLES

Table Page

1-1 EFFECT OF INCREASING VARIABLES ON THE STIFFNESS

OF COHESIVE SOILS ........ .................. 22

1-2 EFFECT OF INCREASING VARIABLES ON THE STIFFNESS

OF COHESIONLESS SOILS ....... ................ 24

1-3 EFFECT OF INCREASING VARIABLES ON THE STIFFNESS

OF UNTREATED GRANULAR AGGREGATES .... ............ 25

1-4 ACTIVITIES OF AN AIRCRAFT IN ITS AIRPORT OPERATIONAL

CYCLE ............ ........................ 31

3-1 COMPARISON OF ELASTIC LAYER THEORY PROGRAMS ..... 56

5-1 DESIGN WEIGHTS AND PASS LEVELS FOR AIRFIELD PAVEM7NTS,

AIR FORCE, (WEIGHTS IN 1000 POUNDS) .. ........ 100

xi
 1. THE AIRPORT PAVEMENT PROBLEM

Air transportation is a v-tal component of interreg-onal,

interstate and international commerce and recreation of nrie United
States. In the past two decades there has been a tremendous
growth in air traffic. The construction of airport facilities
during this time, however, has been virtually stagnant. As a
resuli, the airport system has been stretched near its limits of
capacity. Closure of a pavement, especially a runway at a major
airport, can affect the operations of the entire airport system.
Pavement reliability, therefore, is critical to air transportation
convenience and dependability.

The need for reliable pavement design has long been

recognized as integral to the smooth functioning of airports.
However, the airport pavements that serve their original design
life without extensive maintenan-ce are the exception rather than
the rule. One of the most prevalent problems in airport pavement
design has been underestimating the rate of air traffic growth ani
hence the underdesign of the pavements.

Paivements are one of the most difficul ' design problems faced
by civ4l engineers. Pavements are constructed with low-cost
materials whose properties are highly variable and dependent on
environmental and load conditions. Traffic loadings are difficrnl
to f, 2cast as air traffic growth frequently exceeds expectatic.1s
and as new aircraft are introduced. Environmental conditions can
be evaluated on a probabilistic basis f:om historical tiends;
however, the specific environment at a particular point in time
can have a dramatic effect on the perfoimanc- of the pavement.

One of the most difficult aspects of pavement analysis is t

definition of pavemcnt failure. Excess stress of the pavement
structure resul-ts in fracture of the material. However, cracking
alone docs not signify a failure of the pavement surface.
Aircraft can traverse distressed pavements. In the highway, fieli
pavement failure is cietined in terms of the fur-ctional

I
characteristics of the pavement surface, primarily with respect to
ride quality. There is not a comparable definition of failure for
airport pavements. In airport pavements, the accumulated effect
of different types of distress are of major concern to the
pavement engineer. Toward this end, the pavement evaluation
procedures and subsequent computation of a pavement condition
index developed for the PAVER pavement management system " give
guidance for determining when pavement rehabilitation and
reconstruction is needed. The need for these treatments is
indicative of pavement failure. The PAVER method is widely used
for the evaluation of military airfields and civilian airports
and, therefore, limits of the PAVER defined pavement condition
index may become a de facto standard for the definition of airport
pavement failure.

Due to the difficulty of the airport pavement analysis

process, design methods have evolved in an empirical manner.
While these methods produced workable designs, they have several
shortcomings. There have been significant developments in the
areas of engineering mechanics and materials evaluation that can
provide the foundation for the development of improved airport
pavement design procedures. The purpose of this report is to
summarize the state of the art in airport pavement analysis
models. This task does not have a clear boundary. There are
models that have been used for many years for pavement design.
There are also models that are applied only by engineers that are
on the leading edge of technology for the design of pavement
s-tructures. Other models have been proposed by researchers but
have not been used extensively for airport pavement analysis.
Finally, there are models that have been developed in other
engineering fields that can be applied to the analysis of airport
pavements. This report attempts to span across all these levels
of development.

* Numbers refer to reference list on pg. R-1.

2
 This chapter presents a review of the factors affecting
airport pavement performance as a msans of establishing the
foundation for the review of pdvement analysis models. Included
is a discussion of airport pavement types, pavement material
characteristics, aircraft characteristics, functional areas of
airport pavements, and environmental effects. While the
complexity of these individual factors makes the development of
pavement analysis models a difficult task, the interaction of
these factors makes the modeling task even more challenging.

1.1 AIRPORT PAVEMENT TYPES

Pavements are generally classified as either rigid or
flexible based on the manner they distribute the load over the
subgrade 2 . Rigid pavements use a stiff surface that carries the
load in flexure, a major portion of the structural capacity is
supplied by the surface. Flexible pavements use a lower modulus
surface on base and subbase materials that distributes the load
throughout the pavement structure such that the stresses on the
subgrade do not exceed the strength of the subgrade. In general,
rigid pavements have a portland cement concrete surface while
flexible pavements have an asphalt concrete surface, as shown in
Figure 1-1. Some authors have introduced a third type of pavement
that has an asphalt concrete surface over a portland cement slab.
This has been termed a composite pavement. Functionally, the
composite pavement behaves as a rigid pavement.

1.1.1 Rigid Pavements

Since rigid pavements carry most of the load in the pavement
surface, they may not need a base layer. The base layer of a
rigid pavement can serve to control pumping, frost action,
drainage, shrink and swell of the subgrade, and expedite
construction2 . Depending on the function of the base, it may be
stabilized with portland cement or asphalt cement.

Due to stresses developed during curing and temperature

changes, large portland cement concrete slabs will crack. These

3
 Binder Surface I Seal coat
coat
cors coreTc
coure~\coure i i /Tac ,-Prime

I Base course
Subbase course
Compacted subgrade
L----------------------------------------------
Natural subgrade
(a)I

Prlnd-cement concrete
IBas coure mayor may not be used
------------- L-------------------

(b)I

FIGURE 1-1. PAVEMENT TYPES: (A) FLEXIBLE AND (B) RIGID

(YODER AND WITCZAK)

4
cracks can be controlled through the use of joints, reinforcement,
or combinations of such.

In general, concrete pavement types are classified based on

the type of reinforcing:

1. Jointed concrete pavements (JCP) are designed without

reinforcement in the slab. The spacings of the joints are
selected to keep the curing and temperature stresses below
the working strength of the concrete. Due to the
discontinuity in the slab at the joints, load transfer
devices, such as dowel bars, are frequently but not always
used with JCP.

2. Jointed reinforced concrete pavements (JRCP) are

designed to have a greater spacing between the joints to
reduce the discontinuities caused by the joints. However,
due to the greater spacing between the joints, the slabs
will crack. To combat this, reinforcing steel is used to
hold these cracks tightly together such that aggregate
interlock is maintained across the crack to provide load
transfer. The amount of steel is designed for the control
of the crack width rather than as a traffic load carrying
element of the slab. Due to the distance between the
joints, the joint movement of JRCP is greater than with
JCP and therefore mechanical load transfer devices, such
as dowel bars, are required.

3. Continuously reinforced concrete pavements (CRCP) are

desiged to eliminate the need for the joints. As with
JRCP, the amount of steel is selected to control crack
movement and maintain aggregate interlock across the
cracks.

4. Prestressed concrete pavements (PCP) are designed to

effectively use the high compressive strength of concrete
to offset the inherent low flexural strength. Compressive

5
 stresses in the pavement due to prestressing are
cumulative, with the flexural strength to produce an
increase in the stress range in the flexural zone3 . Due
to the greater range of allowable stresses in the flexural
zone, PCP can be thinner than other concrete pavements.
The prestressing is generally accomplished by
post-tensioning steel strands. The size of slabs of
prestressed pavements is limited by the ability to
post-tension the steel strands.

1.1.2 Flexible Pavements

Traditional flexible pavements consist of a series of layers
with the highest quality material at or near the pavement
surface 2 . For airport pavements, the surface layer of a flexible
pavement is asphalt concrete. The base course can be aggregates
or aggregates stabilized with cement, asphalt, or lime. Subbases
are generally granular materials with a better quality than the
subgrade. The Asphalt Institute promotes the use of full-depth
asphalt concrete pavements without bases or subbases for some
situations.

In certain situations, such as with full-depth asphalt

concrete and pavements with stabilized bases, the pavement can
behave like a rigid pavement and the classical methods for
designing flexible pavements do not apply. In these cases, the
concepts of rigid pavement design may apply 2

1.2 PAVEMENT MATERIALS

As noted above, the primary materials used fnr pavement
construction are the asphalt concrete or portland cement concrete
surface, granular bases and subbases, asphalt, cement, and lime
stabilized bases, and steel used for reinforcement and dowel bars
in concrete pavements. In addition, due to the importance of the
subgrade in the performance of pavements, a discussion of pavement
materials needs to address subgrade properties.

6
 There have been several reviews of the literature on material
properties and the use of these characteristics for pavement
design. Research specifically on airport pavement design has been
performed at the Waterways Experiment Station under contract to
4
the Federal Aviation Administration. In particular, Chou
compiled material characteristics data for bituminous concrete,
portland cement concrete, granular materials, stabilized soil and
cohesive subgrade soils.

1.2.1 Fundamental Material Characterization

Many pavement design methods use the empirical tests of
material quality as opposed to methods that quantify the
engineering properties of the materials. Examples of empirical
tests include the California Bearing Ratio (CBR) and the Marshall
stability test. Empirical material characterization has many
drawbacks, the primary one being the inability to extrapolate
historical knowledge to changing conditions. Since pavement
design is a dynamic process with new materials being introduced
and changing traffic loads, there is increasing interest in the
use of mechanistic analysis procedures. The purpose of the
following discussion is to present background information
regarding the engineering characteristics of materials to
establish a foundation for the discussion of material
characteristics.

Engineering evaluation of materials is generally concerned

with the deformation response of a material subjected to loads.
The deformation of the material is normalized with respect to the
sample dimensions in order to define the strain tensor. The load
is normalized with respect to the area of the sample to defi.c
stress. The accuracy attainable in modeling material properties
depends on the complexity of the material response. Materials can
be categorized by their primary response characteristics. The
primary responses include linear or nonlinear and rate-dependent
or rate-independent 5 behavior.

7
 Linear response indicates that the deformation, or strain, is
proportional to the load or stress level (e.g., Houbling the
stress doubles the strain). Conversely, a nonlinear material
would not demonstrate a proportional relationship between stress
and strain.

If the response of the material is not affected by the

loading rate, then the material is rate-independent. In other
words, for a given magnitude of load the response of the material
does not change if a test is performed rapidly or over an extended
period of time.

There are three basic terms used to describe material

behavior: elastic, plastic and viscous. Elastic response
indicates deformation response is instantaneous and all of the
deformation is recovered when the load or stress is removed.
viscous material has time-dependent response to load and does not
demonstrate any recovery of the deformation when the load is
removed. In general, viscous materials are rate-dependent and
elastic materials are rate-independent. A viscoelastic material
displays a rate-independent initial deformation when the load is
first applied and a time-dependent deformation under sustained
load; the rate-independent deformation recovers when the load is
removed. If an elastic material is stressed beyond the yield
point, plastic deformation will occur if the material is ductile,
otherwise the material will fracture in a brittle mode. The
plastic deformation is an instantaneous response to the load and
is therefore rate-independent. However, the deformation of the
material is not proportional to the ioad level and therefore
plastic deformation is a nonlinear response. There is no recovery
of plastic deformation when the load is removed.

An isotropic material demonstrates the same material

properties regardless of the orientation of the material during
testing. Finally, a homogeneous material has the same properties
throughout the volume. Considering these definitions, the
simplest material to model is linear, rate-independent, elastic,
isotropic and homogeneous. (As will be seen later, this set of
material characteristics is frequently applied to the analysis of
materials in pavement structures.)

No material is linear or rate-independent for all magnitudes

and frequencies of loads5 . Generally, tests for the
characterization of material properties should be performed in a
manner that reflects the in-service conditions in order to measure
the properties of the materials in a specific application.
However, in-service conditions can never be fully duplicated in
the laboratory. Sentler 6 states:

The strength characteristics of materials is to

a large extent based on the results obtained in
standardized tests. Such information is
valuable because it is often the only
information available. But very few structural
members, if any, fail in a way which resembles a
standardized test. Instead other types of
failures like fatigue play a much more important
role in practice. It is also obvioas that the
environmental influences have to be considered
in a more appropriate manner.

For mechanistic analysis, the stress-strain characteristics

of materials are defined by constitutive laws that describe the
primary response behavior of the material when subjected to loads.
The form of the constitutive relationship depends on the type of
the material. A constitutive model formulates a mathematical
functional form of the material behavior. The constant terms in
the functional forms of the constitutive equations are the
material properties that must be defined experimentally. Linear
rate-independent materials are the easiest to quantify, while
nonlinear rate-dependent are the most complex.

For linear-elastic isotropic materials, the generalized

Hooke's law applies and the deformation response of the material
is completely defined by the modulus of elasticity and Poisson's
ratio. The modulus of elasticity, E, is the ratio of the
stress-strain measurements and Poisson's ratio is the absolute

9
value of the ratio of transverse to longitudinal strains for a
sample measured in uniaxial test condition:

E = ale

p = let/le

As the isotropic assumption is relaxed, the number of

constants required to quantify the material properties rapidly
increases. In the case where the material has three perpendicular
planes of symmetry, nine material constants are required. When
the symmetry condition is completely relaxed (i.e., an anisotropic
material), 21 material constants are required5 .

For linear rate-dependent materials, the behavior varics

depending on the stress state of the test. Uniaxial tension or
compression tests are common in some engineering fields. The
tests are performed at either constant stress or constant strain.
In the constant stress test, a stress is instantaneously applied
to the sample and held constant over time. The resulting strain,
E(t) is measured and the normalized result forms the creep
function:

J(t) = e(t)/a 0

For constant strain or relaxation tests, an instantaneous

strain is applied and maintained on the sample. The resulting
uniaxial stress in the sample relaxes with respect to time. The
normalized function defines the relaxation function:

G(t) = a(t)/E 0

For linear isotropic materials, the volume and shear

responses are independent 5 . This fact allows the three-
dimensional characterization of these materials by performing
creep tests in hydrostatic compression to define the volume creep
function, Jv(t), and creep or relaxation tests in pure shear to

10
define the shear creep function, J.(t). Superposition can be used
to define the cumulative effect of these two creep functions,
i.e., the total strain in the three-dimensional case is the
addition of the strain due to volume and shear strains.

Nonlinear rate-independent materials can be modeled with

either deformation or incremental laws 5 . The deformation law uses
the results of experimental tests to quantify a functional form
for the shape of the load-deformation or stress-strain curve.
This may be thought of as fitting a nonlinear regression equation
to the experimental results. The incremental law separates the
measured strain into elastic and inelastic components. Elastic
strains are estimated in the same manner as linear
rate-independent materials, except the constitutive equations are
developed for incremental form. The plastic strain is estimated
based on the concept of a yield surface which expands as the
stresses are incrementally increased in the model in the region of
the yield surface. The amount of deformation depends on the flow
rules that are used to formulate the model. There are many flow
rules that capture the different plastic behavior of materials,
such as strain-hardening, during the plastic deformation.

Characterization of the load-deformation behavior of

nonlinear rate-dependent materials requires the most sophisticated
constitutive laws. Whatever forms are ultimately selected must be
reducible, in some sense, to each of the three previous
classifications5 . Since the exact form of these constitutive
models is dependent on the modeling approach and since there are
so many modeling approaches, the various forms of constitutive
models for nonlinear, rate-dependent materials that exist in the
literature are not presented at this point.

Also pertinent to a discussion of material characteristics

are the effects of aging and temperature. As materials age, their
material properties change due to ongoing chemical changes and
environmental effects. Aging and temperature effects are
difficult if not impossible to address in a constitutive model of

11
the material behavior. Generally, the aging and temperature
effects are considered in an incremental manner in mechanistic
analysis. The following discussion will address these
characteristics for each of the materials. Discussion of the
mechanistic models in subsequent chapters will address how these
characteristics can be included in an incremental manner for
estimating the performance of the pavements.

1.2.2 Steel
Steel is used for pavement reinforcement and load transfer
devices. Steel has a lattice structure composed of iron and
carbon atoms with other alloying agents used to impart special
properties. Compared to other materials used in pavements, the
properties of steel are relatively easy to quantify. For the
conditions encountered in pavement performance, steel is a linear
elastic material. Structural design of the steel in pavements is
relatively straightforward as Hooke's law applies and the modulus
of elasticity and Poisson's ratio are well-defined.

Generally, corrosion is the limiting factor in the

performance of the steel used for reinforcing and load transfer
devices in pavements. Epoxy coatings can be used to limit the
corrosion of reinforcing steel. Load transfer devices can either
be epoxy-coated or made with stainless steel.

1.2.3 Portland Cement Concrete

Although there are many types of cement, and therefore many
types of concrete, the generic term concrete almost always refers
to a material made with portland cement, aggregates, and water.
In many cases, admixtures, including fly ash, are used with these
basic ingredients to augment the properties of the concrete. The
material characteristics of the concrete are strongly dependent on
the proportions of the ingredients or the mix design. Simply
altering the amount of water in the mix without changing the
quantities of the other ingredients will have a major effect on
the properties of the hardened concrete. In fact if all other

12
factors are equal, the strength of the concrete will be determined
by the water-to-cement ratio.

Concrete is classified as a ceramic material. The hydration

process, or the chemical reaction between the cement and water,
causes the growth of calcium silicate hydrate crystals. The
amount of water required for hydration is approximately 30% of the
weight of the cement. Excess water that cannot react with the
cement will result in capillary cavities in the matrix of the
concrete8 and therefore reduces the strength of the concrete,
increasing permeability. Unfortunately, the workability and flow
requirements for construction dictate the need for water in excess
of the amount required for hydration. Recognition of the presence
of the capillary cavities in the concrete structure is important
for understanding the failure mechanisms.

Concrete can be considered a three-component composite

material consisting of the matrix, the aggregates, and the
matrix-aggregate interface (halo interface). Due to the nature cl
the crystal growth, the halo interface is the weakest portion ot
the concrete9 . According to Derucher and Korfiatis, bond
microcracks form during the hardening of the concrete in the halo
interface due to shrinkage. Under load, the bond microcracks
widen at stresses in the range of 15 to 20% of the ultimate
strength of the concrete, f'c" As the stress increases, the boni
cracks bridge each other at 20% of f 1. At 75% of f' the
microcracks in the matrix begin to bridge together. At 90% ot f'
microcracks form in the aggregates.

This progression of the microcracks explains, to some extent,

the load deformation response of concrete. In general, concrete
behaves as a nonlinear elastic material with a brittle failure
when the compressive strength is exceeded.

The failure mode of concrete depends on the stress state of

the material. Uniaxial compressive tests are generally used in
concrete design and quality control tests to determine the

13
ultimate compressive strength. (Shear stress in the sample as a
result of the compression loading is actually responsible for the
failure of the material at the conclusion of the test.) However,
concrete is much weaker in tension than in compressicn. Since
pavements carry traftic loads in a flexural mode, the
tensile-flexural test is usually used for the design and quality
control of concrete used in pavements.

In addition to mechanical failures, many concrete pavements

fail due to detericration of the concrete as a result of
environmental and chemical attack. Freezing water in the pores
and cavities of the concrete expands about 9% producing
hydrostatic pressure. Generally, an air entraining admixture is
added to the concrete to provide protection against freeze-thaw
deterioration. Deicing salts and chemicals can inc:ease the water
retention of the concrete and contribute to recrystallization and
weathering. These effects will cause the properties of the
concrete to vary with time.

1.2.4 Asphalt Concrete

Asphalt cement is blended with aggregates to make asphalt
concrete. (It should be noted that tar and asphalt are different
materials and have very different properties. Tar is used in
Europe for some pavements, but the use in the United States is
limited to some specific applications. Tar will not be considered
in this report.) The quality of the asphalt concrete depends on
both the quality of the asphalt cement and the aggregate
characteristics.

Asphalt cement is a high molecular weight hydroca-bon.

Naturally occurring asphalt cement deposits exist, but they are
primarily used for specialty asphalt products. By a wide margin,
most of the asphalt us-i for pavement construction is the by-
product of the reduction of crude oil. Due to the variety of
crude oil sources and refining processes. the chemical composition
and the distribution of the molecular weights of hydrocarbon
chains are highly variable as compared to steel and portland

14
cement whose chemical compositions are carefully controlled. In
fact, ASTM specifications for asphalt cement are written around
physical tests of characteristics rather than chemical
composition. The gradc of the asphalt is evaluated with viscosity
tests.

Aggregates have a ma]or effect on the quality of the asphalt

concrete. In asphalt concrete, the asphalt cement acts as a
binder to hold the aggregates together. Stresses generated by
traffic loads are transmitted th.ough the aggregates by aggregate
interlock and friction. Hence, the gradation, shape, and texture
of the aggregates are important for the stability of the mix.

Asphalt concrete mix design consists o? selecting the

aggregate gradation and asphalt content required to meet design
criteria such as the stability, flow, and void content. Stability
and flow are mEasured with the Marshall apparatus. This is an
empirical test method in that the results of the methcd are only
meaningful relative to the experience of the agency using the
test. Monismith, Epps, and Finn'0 have proposed a new mix design
procedure based on engineering measures of the asphalt concrete
properties. One of the unique features of the methods proposed by
Monismith, Epps, and Finn is simultaneous consideration of the mix
design and pavement design. In other words, greater consideration
of the application of the material needs to be included in the mix
design process. The mixture properties that should be considered
during mix design include:

1. mixture stiffness;

2. resistance to permanent deformation;

3. durability;

4. fatigue resistance;

5. low temperature response (including stiffness at long

loading times and fracture characteristics); and

6. permeability.

15
 With the exception of durability and permeability, Monismith,
Epps and Finn recommend measuring the properties of the mixes in a
form which permits mechanistic analysis. For example, they
recommend measuring the stiffness as:

Smix(tT) = o/E

where:

Smix(tT) = mixture stiffness at a particular time of

loading, t, and temperature, T

G,= applied stress and resulting strain

Figure 1-2 shows the dependence of modulus on time and

loading temperature and Figure 1-3 shows the ranges of stiffness
under highway traffic loadings for three environments in the
United States. These curves would be flattened out for runway
operations where the duration of the load varies from slow roll to
takeoff speed. These figures demonstrate that asphalt concrete is
a nonlinear viscoelastic material. In other words, this is an
example of the most complex material type discussed in the section
on material characterization.

In addition to the effect of temperature on the stiffness,

there are a variety of environmental effects on the properties of
asphalt. Solar radiation evaporates off the lighter molecular
weight molecules increasing the stiffness of the asphalt and
reducing the flexibility. Oxygen from the atmosphere can bond
with the asphalt at the surface of the pavement altering the
qualities of the asphalt. Water can penetrate into the asphalt
concrete and can lead to debonding of the asphalt and stripping of
the aggregates.

Since asphalt is a petroleum product, fuel spills can soften

or even wash away the asphalt. Obviously this will ruin the
structural capacity of the material and is a primary reason for
the use of portland cement concrete in the fueling areas.

16
 10'

101

-30' C
d 10' .20* C

0 ,

10 10
* 10-aC
20* C
0 250 C
2 10-2 30

~.10-3 400 C
Estimated using shell procedure;
10-4
- ~ Asphalt property:
Pen. @ 25*C =50 dmm 0C
50ayC
1mmin 1hr. lSk
1da 0C
TR &B= T80pn.=54C

*~
10-5

16 1
~lweeks
10'
10-7 10- 10-6 10-4 10-3 10-2 10-1 10 10, 102 1023 10 4 10 3 10'6 O 101
Time, sec.

FIGURE 1-2. DEPENDENCE OF STIFFNESS ON ASPHALT CONCRETE

TEMPERATURE AND TIME OF LOADING (MONISMITH, EPPS AND FINN)

-E 100
(D
*0
> =c

so

'U0
S60

_
040__

00

2 10i0516 0
Asphalt Concrete Stiffness, psi

FIGURE 1-3. RANGE OF STIF'FNESS UNDER MOVING TRAFFIC FOR ASPHALT

CONCRETE DEPENDING ON ENV!.RONMENT, 12-INCH THICK ASPHALT BOUND
LAYER (MONTSMTTH, EPPS AND FTN:N)

17
 The behavior of asphalt concrete is very complex and
difficult to quantify. In the conclusion to an in-depth
literature survey of asphalt material properties, Deacon 11 states:

This discussion indicates that one may characterize

the behavior of pavement materials in numerous ways
depending in part on the nature of the problem and in
part on personal preferences. It must be
emphasized, however, that in most cases pavement
materials do not possess idealized properties and
that the measured properties are often significantly
influenced by the test procedures and equipment. It
is important, therefore, for laboratory procedures to
simulate to as great a degree as possible actual
field loading conditions. Test procedures that
result in nearly homogeneous stress and strain states
are necessary to investigate the properties of a
small volume element.

Although this statement was prepared almost 20 years ago, there is

no evidence in the literature reviewed for this report indicating
the situation has significantly changed.

Mamlouk and Sarofim 12 reviewed the state of the art in the

measurement of asphalt concrete modulus for the design of pavement
structures. There are two ASTM methods for determining the
modulus of asphalt concrete: the complex modulus and the resilient
modulus. The complex modulus uses measures of the strain response
to a sinusoidal wave under uniaxial loading, as shown in Figure
1-1. The complex modulus is defined as:

E* = osin(tat)/{Eosin(t -

where:

E = Complex modulus

)= angular frequency of the vibration

= phase difference between stress and strain

The resilient modulus is the other procedure for determining a

modulus for asphalt. When asphalt is loaded and the load is

18
removed, most of the strain will be recovered (after sufficient
conditioning), as shown on Figure 1-5. The resilient modulus is
the ratio of repeated stress to the corresponding recoverable or
resilient strain. The ASTM procedure for measuring resilient
modulus uses a diametral test mode. Researchers have also used
uniaxial and triaxial test modes.

1.2.5 Stabilized Materials

Granular materials can be stabilized with either portland
cement, lime, or asphalt cement. Generally, the amount of cement
used for stabilization and the specifications for the aggregate
gradation are relaxed for the production of the stabilized
materials. Theretore, the quality of the stabilized materials
will not bc equal to asphalt concrete or portland cement concrete.
The functional behavior of these materials, however, is similar
(e.g., cement stabilized bases are nonlinear elastic and asphalt
stabilized bases are nonlinear viscoelastic).

1.2.6 Soils and Granular Materials

Aggregate bases are conglomerates of individual granular
materials that meet specific gradation requirements. The
properties of the granular materials depend on gradation, moisture
content, density, stress state, and the aggregate shape and
texture. Depending on the amount of fine material in the
aggregates, they may be classified as either cohesive or
cohesionless. Frcr a general mechanistic response viewpoint,
soils and aggregate matcrials can be combined into these two
classifications.

The characteristics of aggregates and soils are strongly

influenced by the moisture content. During construction, care is
taken to compact these materials near the optimum moisture
content. However, during the life of the pavement various
mechanisms such as percolation of water through cracks in the
pavement surface and capillary action tend to allow water to enter
the base and the subgrade. As a result, many airport pavement
foundations, even in arid regions, are near saturation for a major

19
 U) 0 C/= CFOSIN ((o t)

U)

TIME

z c£0 E £f, SIN (0t -)

IE'I=r o o

FIGURE 1-4. TYPICAL PLOT OF STRESS AND STRAIN VERSUS TIME DURING
THE COMPLEX (DYNAMIC) MODULUS TEST (MAMLOUK AND SAROFIM)

TO7ALI

INSTANTANEOUS

LOAD
it I

DEFORMATION

FIGURE 1-5. TYPICAL PLOT OF LOAD AND DEFORMATION VERSUS TIME

DURING RESILIENT MODULUS TEST (MAMLOUK AND SAROFIM)

20
portion of the time. Many pavement design procedures recognize
the problem with saturated bases, subbases and subgrades, and
require material characterization be performed on saturated
samples. Cedergren13 developed a damage factor for comparing the
damage to pavements with saturated bases and subgrades to well-
drained pavements. The damage factors ranged from 10 to 70,000.
Cedergren also demonstrated that pavements can remain saturated
for up to 20 days following a rainstorm.

1.2.6.1 Cohesive Soils - Deacon states that most investigators

report the behavior of cohesive soils is highly nonlinear. Clays
show immediate and time-dependent recoverable and permanent
strains, the immediate strains being predominant under short
duration loads and the permanent strain per cycle decreasing to an
insignificant amount after many cycles of stress. Stress history
may have a significant effect on the nonlinear response to load in
two ways:

1. The stiffness of these materials is dependent on the

initial stress state and increases as the effective
mean principal stress increases.

2. The stiffness decreases with an increase in the

incremental stress amplitude (increase of the deviator
stress while maintaining the confining stress in the
triaxial tests).

The effects of load, mixture features and environment on the

stiffness of cohesive soils are summarized in Table 1-1. As
this table indicates, the stiffness or load deformation
characteristics of cohesive soils are very complex. In addition,
cohesive soils can be cross isotropic (e.g., the horizontal
stiffness can excecd the vertical stiffness).

1.2.6.2 Cohesionless Soils - The stiffness of cohesionless soils

(sand) is affected by many of the same factors as cohesive soils.

21
 TABLE 1-1. EFFECT OF INCREASING VARIABLES ON THE STIFFNESS
OF COHESIVE SOILS (DEACON)

Variable Effect on Remarks

Stiffness

Loading
Number of cycles Decrease
Minimum Minimum at 1 to 5000
cycles
Incremental strain Decrease Rate of decrease
amplitude depends on maximum
stiffness and shear
stress
Effective mean Increase Effect depends on stress
initial principal or strain amplitude
stress
Transverse stress no effect
Initial octahedral effect negligible after
shear stress 10 cycles
Frequency of loading Increase Effect minor above
10 cps
Strain rate Increase
Overconsolidation Increase Any effect can be
ratio explained on basis of
effective pressure and
void ratio
Stress path Large dependency

Mixture
Soil disturbance Decrease
Void ratio Decrease Maximum effect at low
confining pressure
Dispersion Decrease At small strains
Structure Little effect on max.
shear modulus
Degree of saturation Decrease Modulus of resilient
at compaction deformation
Plasticity Decrease
Compaction Energy Maximum Impact compaction

Environmental
Aging Increase
Degree of saturation Decrease
Time (thixotropy) Increase Recovery after high amp-
litude cyclic loading
or many load cycles
Densification Increase
Time (during sec- Increase Bentonite
ondary compression)

22
However, the response to an increase in the number of cycles is
different for cohesive and cohesionless soils. Table 1-2 is a
summary of how the various factors affect the stiffness of
cohesionless soils11 . Many investigators relate the stiffness of
cohesionless soils to the mean effective stress c o as:

S = K aon

Where K and n are experimentally determined constants. Cohesionless

soils, by definition, do not have tensile strength. However, they
are probably more isotropic than other paving materials 11

1.2.6.3 Untreated Granular Materials - The effects of various

factors on the stiffness of granular materials are summarized in
Table 1-311. The major effect is the initial confining pressure
on the sample. The relationship between the modulus of resilient
deformation, MR, and the initial stress state is:

MR = Ka 3n

and

MR = Kaon

These equations are similar to the equations for cohesionless

soil, except that a3 is the initial confining pressure in a triaxial
test (rather than a).

Deacon reported that one researcher, Hicks 14 identified

factors influencing Poisson's ratio of aggregate materials.
Poisson's ratio increases with:

1. decreasing confining pressure;

2. increasing incremental stresses;

3. decreasing degree of saturation; and

4. decreasing fines.

23
TABLE 1-2. EFFECT OF INCREASING VARIABLES ON THE STIFFNESS
OF COHESIONLESS SOILS (DEACON)

Variable Effect on Remarks

Stiffness

Loading

Number of cycles Increase Approaches a

maximum

Incremental strain Decrease Rapid decrease

amplitude

Incremental stress Decrease

amplitude

Load duration Decrease Pulsating loads

Load rate or Constant No effect after the

frequency first few cycles

Initial effective Increase

Mean principal stress

Initial octahedral Decrease Very small effect

shear stress after 10 load
cycles

Mixture

Void ratio Decrease

Environmental

Degree of saturation Constant Effective stresses

must be used

24
TABLE 1-3. EFFECT OF INCREASING VARIABLES ON THE STIFFNESS
OF UNTREATED GRANULAR AGGREGATES (DEACON)

Variable Effect on Remarks

Stiffness

Loading
Number of cycles Constant After 50 to 100 cycles
Initial confining Increase Triaxial compression
pressure
Initial effective Increase
mean principal stress
Incremental stress Constant to Differences in liter-
level Increase ature, large effect if
shear failure
Load duration Constant 0.1 to 0.25 sec.
Load rate or Increase Small increase
frequency
Drainage Constant

Mixture
Void ratio Decrease At low moisture
contents
Increase At high moisture
contents
Angularity and Increase
surface roughness
Fines Decrease Minor effect
Compaction water Decrease
content

Environmental
Degree of saturation Decrease

25
1.3 AIRCRAFT CHARACTERISTICS
One of the primary differences between highway and airport
pavements is the nature of the traffic loadings. Essentially,
commercial airport pavements are designed for a fewer number of
repetitions of heavier loads than highway pavements. However, the
concept of fewer number of repetitions is a relative term. The
1974 FAA advisotv circuLar for pavement design and evaluation (AC
150/5320-6B) had design curves for pavcments receiving up to
2!,000 annual departure,. These curves were extrapolated in the
1978 report to cover up to 200,000 annual departures based on
accelerated test data performed by the U.S. Corps of Fngineers.
However, Kohri ajQ Bentsen 15 report these test data were based on
the equivalent of 17,400 passes of a dual-tandem gear. The Dalla.-
- Ft. Worth Airport averaged 314,000 annual operations from 1974
to 1983 and Atlanta's Hartsfield Airport averaaed 577,000 annual
operations from 1984 to 1986.

Traffic loading factors that affect pavement analysis are:

1. total aircraft weight;

2. distribution of the aircraft load to the wheel

assemblies;

3. geometry of the wheel assemblies and the distribution ot

the loads to the individual wheels;

4. characteristics of the tires, including inflation

pressure;

5. lateral distribution of the load across the pa'ement

structure;

6. duration of the load;

dynamic nature of the wheel loads; and

8. numler of repetitions.

The first four parameters are defined by the design

characteristics of aircraft. The other parameters are defined hy
the ope itional characteristics ot the airport.

26
 Aircraft characteristics are available in several references
and are only briefly summ.arized here. As shown in Figure 1-6,
there are three basic types of gear assemblies for civilian
aircraft: single tricycle, single bicycle-tricycle combination,
and double tricycle. The Boeinlg 727, McDonnell-Douglas DC-10 anl
the Boeing 747 respectively, are examples of each of these types
of assemblies. There are several types of tire assemblies as
shown in Figuie 1-7. Nose gea- assemblies are predominantly twin
tire assemblies. The main truck assembly of the heavier aircraft
are predominantly twin-twin tandems, e.g., B 747, B 707, DC 8 and
some models of the L 1011 (model 1). The DC 10 has a twin-twin
assembly under the wings and a twin gear in the center. Th ain
truck of the lighter commercial jets, B 727, B 737 and DC 9 are
twin tire assemblies.

The introduction of the wide-body aircraft greatly increased

the maximum gross weight of aircraft during the 1970s. The Boeing
747-F is the heaviest civilian aircraft at 778,000 pounds. Boeinq
is in the process of developing a model of the B 747 estimated to
have a maximum gross weight of 987,000 pounds.

Due to the load spreading effect of multiple gears and tires,

the weight on the truck assembly and spacing of tires is o
greater concern than the maximum gross weight of the aircraft.
The DC-10-30 and 20 models have the highest maximum weight per
tire of any civtilian aircraft. These tires have a spacing of 54
inches whereas the B 727, also with a high tire pressure, has a
tire spacing of 34 inches. This narrower spacing can result in a
greater concentration of the stresses and in some cases the B 722
can actually cause greater damage to the pavement than the neavier
aircraft.

The construction and pressure of the tire affect the contact

stress on the pveraent. In general, the contact pressure of

27
 / \
/ \

/
/\

TMTM' TT
TNN
Single Tricycle

Tm T

/I' / I '
/ I \\/ I ' '

a Tb b, T M/l
T"O/ TO \T

Single Bicycle - Tricycle Combination Double - Tricycle

a a , c\

FIGURE 1-6. TYPES OF GEAR ASSEMBLIES ON CIVILIAN AIRCRAFT

( SARGIOUS )

Single Twin Twin-Twin

Single: Tandem Twin: Twin Tandem Twin: Triangular

•0 0 00 0
Twin: Triple Tandem Twin - Twin: Trlp!e Tandem

FIGURE 1-7. AIRCiZAFT LANDING GEAR-TRUCK TYPES (SARGIOUS)

28
aircraft ti -es has the shape of an ellipse, as shown jn Figure
i-a. The total contact area is determined as a function of the
total weiqht on the tLre and the tire pressure. cenerally, the
conac2 pressure between the tire and pavement is assumed to be
,niforn. Research h:s shown that contact pressure between truck
:_ires and pavements is not uniform ' 1 71 8 "' '. The tire pressure of
commercial ajr' ;aft tires is in the range of 100 to 180 psi.

1.4 FUNCTIONAL AREAS OF AIRPORT PAVEMENTS

Wignot et. al. 20 , as reported by Highter and Harr21 , developed
T'aole 1-4 and Figure 1-9 to de vonstrate the functional areas of
airo.-ift activities. The majority of loads o i an airport pavement
are djnami . Static loads onl, exist when the aircraft is parked.
Horizontal lo-is are generated during turning and breaking
maneuvers. Turning maneuvers can be critical when the aircraft
uses high exit. taxiuays as there is a transfer of load to the
outer undercarriage of the aircraft gear. Dynamic loads are
generated during the landing and takeoff operations.

Aircraf'. )perations on runways, taxiways and aprons are

highly channelized. Instruments on modern aircraft allow., the
pilots to accurately place the aircraft during landing. Due to
lower speeds on taxiways, traffic is more highly concentrated. On
aprons, care is taken to repeatedly place the aircraft at the same
location to permit the concourse hookup. Ho!ang 22 ,23 reported on
the results of an FAA--sponsored study on the lateral distribution
of aircraft at nine airports. Data were collected both day and
night and for :i rar je of winds and rain. HoSang reported the
average offset to the nose gear from the center of the facil ity
and the standa-rd deviation:
Type of Pavement Average Cffset Standard
feature width _m)_ _ M Deviat ion
Runway - landing 45.7 0.27 to 0.48 2.1 to 3.1
61.0 0.27 to 0.70 2.7 to 3.4
takeoff 45.7 0.15 to 0.37 1.8 to 2.5
61.0 0.70 to 0.76 2.3 to 2-5
Taxiway 22.R 0.64 0.76 to 1.2
30. 5 (.97 1.8
High speed exit varies 2.4 to 3.2

C
29
 03Li 0L 0.31L
l

..............

L.

FIGURE 1-8. TYPE IMPRINT AREA (SARGIOUS)

150 MPH 100 MPH 50 MPH

Landing Decelerating High-Speed Taxi, 41j, o
Impact, Braking, Dynamic Oscillation v
Spinup Accelerating High-Speed Taxi,

Oscillation Takeoff -
elDynamic
50 MPH 100 MPH 150 MPH o
Rejected Takeoff
Ree(Emergency)

Low-Speed Taxi Low-Speed Taxi

20 MPH Low- 20 MPH
Speed
zTaxi

] U'l

Parking
(Static)

FIGURE 1-9. AIRCRAFT GROUND OPERATION CYCLE (WIGNOT)

30
 TABLE 1-4. ACTIVITIES OF AN AIRCRAFT
IN ITS AIRPORT OPERATIONAL CYCLE (WIGNOT)

Departure Arrival

Static(parked) Landing impact

Low-speed taxiing High-speed braking

Turning Deceleration roll-out

Low-speed braking Low-speed braking

Acceleration takeoff roll Turning

Aborted takeoff roll Low-speed taxi

(emergency operation
infrequent use)

Takeoff rotation Static (parked)

31
 Due to the width of airport pavements, edge loading is of
less concern than it is on highway pavements. The outer wheels
are usually more than 15 ft. from the pavement edge 3

The load duration of aircraft on pavements is highly variable

depending on the functional class of the pavement. In the apron
and maintenance areas, the aircraft are static for extended
periods of time. On the runway, the load duration can be
extremely short depending on the speed of the aircraft and the
length of the tire contact area.

There is a complex interaction between the aircraft and

pavement roughness. The aircraft characteristics that enter this
interaction are weight, center of gravity, aerodynamic lift,
landing gear configuration size and tire pressure (Wignot et al.
cited by Highter and Harr). The aircraft response to roughness
continues for some time after its initiation causing the dynamic
loads to vary. According to Highter and Harr, the actual loads
imposed on pavements defy anything more than qualitative
description. These authors argue that pavement analysis methods
that require precise knowledge of the induced loads have a small
probability of success in predicting pavement performance.
Ledbetter 24 reported on an experimental study of the effects of
dynamic loads on airport pavements. Based on instrumentation of
aircraft and pavements, Ledbetter reports:

1. None of the basic ground operations induced pavement

responses greater than that for static loads, even when
the dynamic load was 1.2 times greater than the static
load.

2. Rough surfaces and stiff pavements could increase the

importance of dynamic loads.

3. Elastic and inelastic responses of the pavement decrease

at higher speeds.

32
 4. High horizontal loads during turns produce responses that
are temperature and rate of loading dependent.

1.5 ENVIRONMENTAL EFFECTS

As noted in the section on material characteristics,
environmental conditions have a major influence on the
characteristics of materials and the performance of pavements.
The purpose of this section is to summarize the direct influences
of the environment on pavements. The environmental factors that
influence the pavement are rainfall or moisture, temperature and
solar radiation.

Moisture in the pavement structure has the following effects:

1. concrete swell;

2. transport of the contaminants into cracks and joints;

3. reduction of the strength and stability of base,

subbase and subgrade;

4. corrosion of reinforcing steel; and

5. promotes stripping of asphalt concrete.

Both low and high temperatures can have detrimental effects

on the performance of pavements. High temperatuze produce6 the
following effects:

1. expansion of concrete producing high compressive

stresses at the joints and curling and warping
stresses in the slab;

2. rapid curing of concrete;

3. softening of asphalt cement, reduction in the

stiffness; and

4. reduction in the viscosity of asphalt cement

contributing to bleeding.

33
Low temperatures promote the following effects:

1. widening of joints and cracks in concrete pavements;

2. thermal contraction stresses in both asphalt and

concrete pavement;

3. increase on the modulus of asphalt concrete resulting

in loss of flexibility; and

4. expansion of frozen moisture in concrete generating

internal hydrostatic stresses in the pavement.

One of the most critical conditions that can develop in a

pavement structure is the spring thaw process. Low temperatures
freeze the moisture in the pavement structure. As the temperature
rises, the pavement thaws from the top down. The ice in the lower
areas of the pavement traps the water in the pavement structure
greatly reducing its strength. Traffic on the pavement will then
result in an excess amount of destruction to the pavement. The
freeze-thaw problem is frequently the most damaging in moderate
temperature areas where several cycles can occur annually.

The primary effect of solar radiation to pavement materials

is the hardening of asphalt concrete surfaces. This is the result
of volatilization of the light molecular weights portion of the
asphalt cement, reducing the flexibility.

1.6 INTERACTION EFFECTS

Obviously the performance of pavements is a complex process
affected by the interactions of the material properties, traffic
loadings and environment. Static loadings on pavements softened
by high temperatures can result in shoving and permanent
deformation of the surface. Wheels crossing joints opened by low
temperatures and subgrades saturated by moisture can cause
thousands of times more damage than ideal conditions. Development
of a uniform model of the performance of pavements must
incorporate the combined effects of traffic, environment and
materials, and their variation over the life of the pavement.

34
 Methods are being developed that can be applied to the
analysis of these complex interactions. Figures 1-10 and
i-1i demonstrate the results of pavement fatigue analysis that was
25
developed at the University of Tokyo

35
 -8.8
8, -8.9

6 -9.1
E -9.2
u! -9.3
0 c
,, -9.4

FIGURE 1-10. A STEREOGRAM OF CALCULATED MONTHLY AND HOURLY

CHANGE OF FATIGUE DAMAGE OF ASPHALT PAVEMENT AT BOTTOM OF
ASPHALT MIX LAYER (HIMENO ET AL.)
0A

._,..me. -10. ' -

g -9.5

0E

-l -11.0

Mon th oft YearT

CT -. CD

FIGURE 1-11. A STEREOGRAM OF CALCULATED MONTHLY AND HOURLY

CHANGES OF FATIGUE DAMAGE OF ASPHALT PAVEMENT AT TOP OF
ASPHALT MIX LAYER (HIMENO ET AL.)

36
 2. A GENERAL FRAMEWORK FOR THE DESIGN OF AIRPORT
PAVEMENTS

Traditionally, pavement design consists of the selection of

layer materials and thicknesses. This straightforward task can be
accomplished simply with any one of a number of pavement design
procedures such as those reviewed in this paper. However, the
straightforward application of these procedures have several
drawbacks. There are vast differences between the actual
performance of materials and the complexity of the traffic and
environmental loading conditions, versus the simplifying
assumptions used in the development of many airport pavement
design procedures. Due to the complexity of the real-world
conditions that affect pavement performance, the development of a
pavement design method is a difficult task requiring a series of
compromises between the true field conditions and the ability of
engineers to develop accurate models and test methods to quantify
their inputs. However, the development and verification of the
models and test methods is a prerequisite for predicting pavement
performance in the field.

It must be understood that the pavement design process is

only one part of the universe of factors that affect pavement
performance. Haas and Hudson 26 have defined six major classes of
activities required for the management of pavements: 1) planning;
2) design; 3) construction; 4) maintenance; 5) pavement
evaluation; and 6) research. As shown on Figure 2-1, these
activities are interrelated and a failure to understand these
relationships will affect the performance of the entire pavement
system. For example, clearly the work of the pavement design
engineer will be meaningless if the contractor fails to follow the
design and specifications. Hence, a review of the pavement design
methods, as a foundation for the development of a new method, must
recognize the importance of the pavement construction, maintenance
and evaluation process.

37
 PLANNING D N V
ACTIVITIES DESIGN ACTIVITIES
* Assess Network Input Information on Materials
Deficiencies Traffic, Climate, Costs, etc.
" Establish Priorities Alternative Design Strategies
" Program and 0
Budget Analysis
Economic Evaluation
Optimization

Il .
r(CONSTRUCTIO ACTIVITIES
(1MAINTENANCE ACTIVITES
_S

RESEARCH-DATAPAVEMENT EVALUATION
AcTIITIES BAN

FIGURE 2-1. MAJOR CLASSES OF ACTIVITIES IN A PAVEMENT

MANAGEMENT SYSTEM (HAAS AND HUDSON)

38
 The main focus of this project is a review of pavement design
and analysis methods. There are several pavement design methods
available ranging from empirical procedures and guidelines to
sophisticated analytical models. However, as shown on Figure 2-1,
the design activities can basically be divided into three levels:

1) information needs relating to inputs, objectives,

constraints, and so on;

2) the generation of alternative design strategies; and

3) the structural analysis, economic evaluation and

optimization of the design strategies.

The three levels of pavement design activities are expanded

in Figure 2-2 26 . This figure is a generalized diagram of the
activities required for many pavement design methods that are
currently available, although not all of the design methods
include all of these activities. Figure 2-3 presens a more
specific formulation of the overall pavement design process
27
formulated by Hudson and McCullough . The overall sequence of
the pavement design process defined by McCullough and Hudson has
been followed by several other researchers. Figure 2-4 defines
the steps used by Gomez-Achecar and Thompson 28 at the University
of Illinois, while Figure 2-5 presents a more detailed flow of the
24
pavement analysis steps defined by Himeno et al.

The pavement design process consists of all of the activities

shown in Figures 2-2 and 2-3. The top row of Figure 2-2 and the
upper left portion of Figure 2-3 represent the information and
analysis methods that should be available or acquired for the
design process. Data on the available materials, expected traffic
and climatic factors are often the first information items
acquired. However, the specific data requirements are a function
of the analysis model used in the design procedure. For example,
traditional pavement design procedures use California Bearing
Ratio (CBR) or plate bearing test results while mechanistic design
methods require measures of the "elastic constants" of the
materials.
39
 Expected el. Available Available O~bjectives Constraints: Expected
Available Traffic Climatic Structural Variance Related to: Costs: -
Mec ( Mtds
andI Data Selected
Design 2.Models
Available Data on:
a)Construction a)Performance a)Costs
b)Structure / a)Construction 0
Volumes) Period Mo esC
Economic b)Maintenance b)Economy c)Overlays, etc. b)Maintenance
PLAN NG
(ProjectI (PoetModels c)Safety c) User. etc>
Investment 0
Z
Decision)

Characterize Materials over Simulated

Range of Loads and Environment

Generate Alternative Pavement W

Design Strategies
a)Layer Types and
d) Maintenance I
- Thiicknesses Policies Expected /M 0)
b)Materials Types e)Overlays and
X
and Sources Seal Coats
c) Construction f) Performance
PoliciesExpected Evaluation Policies

(Combined Subjective
0 0/Analytical
00 0 0 l Link)
O 4l O 0 O 0 tm

O O 0
O0

Optimize and <

Predict Outputs Assign Costs
* Predict Outputs CZ
Apply Available :
* in Terms of Distress:
a)Fatigue ina)Terms
RidingofQuality
Performance:
I vs. Apply and
Available Recommend
Best Strategy Z r
-
(n
Structural b) Permanent Relate orfor
Analysis Deformation vs. R
c) Skid Resistance to
Costs and 0 Imlementation
mpModels 0
Models c) Shrink. Cr. Time Transform Determine '100
e) Disintegration
-40

Z
IMPLEMENTATION

FIGURE 2-2. MAJOR PAVEMENT DESIGN COMPONENTS (HAAS AND

HUDSON)

40
 INPUTS Feedback

-.. aintenance Fund Available k__- Maintenance

Variables Deig Peid
Min' Thickness
... ucraMin. TimesOUPT

[i.Pavement
onstruc on
atriaRESPONSE
C FrMdlDeft

Structure
~ion
StanDistortio
Permanent
(Distress)

Disint rain
Weighting
Functions

W~
WOu
SYSTEM OUTPUT FUNCTION

Performance

nivironmentps
Atualy Iransporled--------- ----- -~*FLoad Apps

....... ' Rogaeet eeae io- VARIABLES Compare Solutions and

Depleted Resources Opimze Pavement

FIGUR 2-3. BLOCK DIAp A OFgh AVEMENilty DESINaSYTELih D

Constuc AND lynRoUGH)
McCULt ecsonCle

01 ECSIN ainennc F ndsNose ec Usaisa41r

 INPUTS
MATERIALS CHARACTERI7ATION
PAVING MATERIALS

SUBGRADE SOILS STMOUEL

TRAFFIC ch
z
CLIMATE PAVEME RESPONSES P

I--
z
FUNCTIONS
TRANSFER a

DISTRESS /
PERFORMANCE

FINAL
DESIGN

FIGURE 2-4. COMPONENTS OF A MECHANISTIC PAVEMENT DESIGN

PROCEDURE (GOMEZ-ACHECAR AND THOMPSON)

42
 Solar Wind Air Thermal
Radiation Velocity T/ature Prepert

,oe l i Traffic l Pavement Binder Van der Poels

Volume Temperature Property Nomograph
IVehicle
Speed

Transverse
i Pavement IMix Stiffness Multi-Layered
Wheel Position Structure Modulus Elastic Theory

I I
Stress, Strain ,
Phase Angle I
Dissipated
Energy

Estimation of Unit
Fatigue Damage

Accumulation of
Unit Fatigue Damage

Final ]
Fatigue Life

FIGURE 2-5. FLOW CHART FOR ESTIMATING FATIGUE DAMAGE OF

ASPHALT PAVEMENTS (HIMENO ET AL.)

43
 While empirical pavement design methods were useful in their
time, the dynamics of modern pavement design, including new
materials, changing load conditions, and the need for greater
reliability in the pavement design, have limited the utility of
empirical methods. Thus, the focus of this project is on the
development of mechanistic design methods. The primary
distinction between the empirical and mechanistic methods is the
use of traditional engineering analysis methods for estimating the
performance of the pavements. The key feature of the mechanistic
approach, as shown in the upper left portion of Figure 2-3, is the
use of structural models of the pavement to predict the response
of the pavement to traffic and environmental loading.

The central feature of a mechanistic analysis procedure is

the structural model. Material characteristics and traffic and
environmental loadings must be expressed in terms of the
parameters required in the structural model. The structural model
uses the input parameters to predict the primary response of the
model in terms of stresses, strains and deflections of the
pavement. However, these structural responses are not sufficient
to predict the life of the pavement. Limiting response criteria
are required for estimating distresses in the pavement as a
function of the primary response of the pavement structure. As
opposed to the majority of structures designed by civil engineers,
pavements must be designed to withstand multiple load applications
over an extended period of time. In other words, the limiting
response or distress functions used for pavement design must
capture the cumulative damage that results from the traffic and
environmental loading and their interaction.

The basic idea of structural design, as stated by Sentler, is

to ensure the load carrying capacity of the structure is larger
than the anticipated loads that affect the structure2 .
Although
this concept is simple to state, it is difficult to quantify with
theory and more difficult to translate into practice. Material
characterization is largely performed with standardized test
methods, but very few structural members tail in a way which

44
resembles a standardized test29 . Of particular concern is the
long-term interaction between traffic and environmental loads and
the characteristics of the materials. While the challenge is
great, the application of advanced mechanistic models and test
methods can be used to improve the state of the art in pavement
design and yield greater reliability in the analysis.

45
 3. ANALYSIS MODELS

The central feature of a mechanistic design method is the

structural analysis model used for computing the response of the
pavement with respect to the load and environmental inputs or
stimulants. There are several excellent reviews of the various
structural analysis models, such as those by Yoder and Witczak2 .
In general these models can be separated into two groups: models
for predicting traffic stresses and models for predicting
environmental stresses.

The review of structural models poses a dilemma as to ,:hethe-

to strive foi completeness of the theoretical aspects of the
models or to address in broad terms the features and abilitIes of
the models. The latter approach was selected for this revie'.-.. A
relatively brief review of the mathematical models would not be
sufficient for those interested in the formulation of the mnodels.
Those interested in the models at this level of detail will be
capable of retrieving the required information from the
literature. Furthermore, the computer codes are available for all
of the models described in this paper. This permits the
application of the models by users without the mathematical
sophistication required to understand some of the complex
mathematical formulations. However, prior to application the
users should understand the premise of the theories, the
underlying assumptions, and be able to interpret the output of th,
models.

According to Nair 30 , the formulation of mechanistic models of

pavement response involves idealizing the real physical em
and casting it into mathematical form. The general mather -tical
form of pavement response models consists of a set of viutia
differential equations subjected to various initiiJ Ird bounii.rv
conditions. The essential components of these models are thr
governing equations, constitutive equations, and boundary cr1I

47
initial conditions. Nair describes the relationship of these
components as:

For the analysis of pavements, the governing equations

are the equations of equilibrium, motion (for dynamic
problems) and compatibility. These equations are
derived from the basic laws of classical physics and
from continuity considerations in the material. Various
approximations can be introduced at this level (e.g.,
small strains to obtain linearity and symmetry of the
stress tensor). It should be recognized that the
governing equations are independent of any material
properties.

Constitutive equations are representations of the

properties of the particular materials under
consideration and represent idealizations of actual
material behavior.

Boundary conditions may consist of prescribed

displacements and stresses on various boundaries. (For
thermal and hydro stresses it is necessary to define the
temperature and moisture contents as functions of space
and time.) For static problems this is sufficient: for
dynamic problems it is necessary to specify the
conditions at some arbitrary time, generally at t = 0,
when they are called initial conditions. The governing
and constitutive equations can only be solved in general
terms; it is boundary and initial conditions that make
the general solution specific for the problem under
consideration. The boundary and initial conditions also
represent various levels of idealization. For example,
the actual time variation of load might be approximated
by a simple analytic function (e.g., sine) or non-
axisymmetric loads might be approximated by axisymmetric
load distributions.

There are two basic approaches to the solution of the

boundary value problems; analytical or classical methods and
numerical or approximation techniques. The analytical techniques
carry the development of the mathematical formulation of the
problem as far as possible before resorting to numerical
calculations. On the other hand, numerical techniques use a
problem formulation directed toward a computational procedure
from the outset30 . The elastic layer theory developed by
32
Burmister 31 and the thin plate solutions developed by Westergaard
are examples of the analytical approach. Finite differences and
finite element models are examples of the numerical approach.

48
 Methods for the analysis of the pavement due to load will be
presented first followed by a discussion of the models for
analyzing environmentally induced stresses. The analytical
solutions are the first load response models addressed including
layer theories for elastic and viscoelastic materials and plate
theory. This will be followed by a discussion of the finite
differences and finite element numerical analysis methods. Models
of environmentally induced stresses are then presented. The
theories discussed in this chapter are for predicting the primary
response (i.e , stresses, strains and displacements), of the
pavements. Con;epts for the analysis of the limiting response of
the pavements (e.g., cracking) are discussed in the following
chapter.

3.1 ELASTIC LAYER THEORY - STATIC LOADS

Burmister's solution of the elastic two-layer problem laid
the foundation for the extension of the theory to multiple layers.
The equations for the two-layer case are relatively simple and can
be solved on a pocket calculator. However, the extension of the
theory to multiple layers greatly complicates the problem and
practical application of the theory requires computer analysis.
Fortunately, there are several computer programs available for
performing this analysis.

The general concept of elastic layer theory (ELT) is shown in

Figure 3-1. Yoder and Witczak 2 state that the assumptions used
for model development are:

1. homogeneous material properties;

2. finite layer thickness except for the bottom layer which
is assumed to be infinite;

3. isotropic material properties;

4. full friction between the layers;

49
 P

Interface 1

h2 , E2 , /.L2 A1 Bi

Interface 2

Interface n - 1

'Itt

FIGURE 3-1. GENERALIZED MULTILAYER ELASTIC SYSTEM (VO11LR AND

WITCZAK)

50
 5. no shear stresses at the surface; and

6. materials are linear-elastic and obey Hooke's law, i.e.,

the constitutive behavior of the material is defined by
the elastic modulus and Poisson's ratio.

In addition, the load is assumed to be uniformly distributed

over a circular area and static. Some of these assumptions have
been relaxed with the development of various elastic layer theory
computer codes.

Strictly speaking, elastic layer theory is not always an

accurate model of a pavement structure. Comparison of the
material characteristics, pavement geometry, and traffic loading
conditions described in Chapter 1 indicates that real pavement
structures do not conform with the assumptions specified for the
theory development. When the theory was introduced, engineers
recognized the potential of the model, if properly applied, to
improve the state of the art in pavement design. Numerous
researchers investigated the effects of the differences between
the theory and reality on the utility of elastic layer theory for
pavement analysis.

Avramesco 33 concluded from a theoretical study that for

elastic materials, if the speed of the load is a fraction of the
Rayleigh wave velocity of the subgrade, the distribution of
stresses and strains is equal to the static case. Comparison
stresses and strains for a static load to values corresponding to
vehicles with a speed of 270 mph resulted in a difference of less
than 10%. Other studies reported at the Second International
Conference on the Structural Design of Asphalt Pavement show that
as the speed of the load increases, strains and deflections
decrease and stresses increase. This is attributed to an increase
in the modulus of asphalt concrete with decreasing load duration
and stress relaxation with time. These studies suggest the
assumption of static load can be compensated for by characterizing
the properties of the asphalt concrete at a load duration equal to
the field conditions.

51
 Lister and Jones 34 studied the effect of nonuniform,
noncircular loads and concluded that the net effect of these two
assumptions resulted in an error of less than two percent under a
whole range of realistic tire and load conditions. Gross
overloading of a tire results in an error of about seven percent.
Saraf et al. 35 concluded elastic layer theory overestimates the
tensile strain at the bottom of the asphalt layer compared to a
finite element analysis.

The principle of superposition is commonly used for the

analysis of structures. Ahlvin et al. 36 presented experimental
data supporting the use of this theory for pavement analysis.

McCullough 37 and Brown and Pell 38 concluded the assumption of

continuity across the interface of the layers is valid for
pavement analysis.

The assumption of infinite horizontal dimensionis of the

layers is a major drawback to the use of the model for pavement
structures. Edges, joints and cracks in pavements increase the
stresses generated by wheel loads. Since the interior loading
condition is a more realistic assumption for flexible pavements
than for '-ncrpte pavements ELT has been more widely applied to
flexible pavements.

Several investigators have demonstrated that the stresses and

strains in a pavement are sensitive to the thickness of the
subgrade. While determining the thickness of the subgrade is a
concern, it is not a limitation of the model as a rigid layer can
be simulated by using high modulus values to simulate the presence
of a bed rock layer.

The material assumptions used in ELT are vastly different

from the behavior of the material characteristics, especially for
asphalt pavement. Therefore, there has been considerable research
into the effects of these assumptions on the reliability ot the

52
ELT analysis. Engineers that employ ELT for pavement analysis
rely on measuring the material properties under simulated field
conditions. Different test procedures will produce different
measures of the material properties. This will result in
different primary responses estimated by ELT and, therefore, when
the primary responses are used in the limiting response functions,
the prediction of distress may be in error due to the formulation
of the analysis problem.

There are several computer codes for the solution of elastic

layer theory equations. Three widely used programs in the United
States are:

1. CHEV5L or LAYER developed by California Research

Company39 , a division of Chevron Oil;

2. BISAR developed by Shell Oil Co.40 ; and

3. ELSYM5 developed by Ahlborn41 at the University of

California.

The Chevron program is capable of analyzing five layers and a

single load. Input and output are in the radial coordinate system
which complicates the interpretation of the data. The ELSYM5
program uses the same basic algorithm as CHEV5L for the solution
of layer theory equations. However, an input and output processor
are used to allow the use of rectangular coordinates. In addition
to being more convenient to the user, the rectangular coordinate
system permits the use of superposition to permit the analysis of
multiple loads. ELSYM5 can analyze up to ten identical loads. As
with CHEV5, ELSYM5 can analyze up to five layers. In addition,
ELSYM5 permits the definition of a rigid layer under the five
conventional layers. The interface at the rigid layer can be full
friction or full slip.

The accuracy of these programs has been tested by several

researchers with the conclusion that the )rograms faithfully
perform the required calculations in most cases. However, there

53
are some problems near the pavement surface directly under the
load. With respect to ELSYM5 Ahlborn states,

The program uses a truncated series for the integration

process that leads to some approximations for the
results at and near the surface and at some points out
at some distance from the load.

The CHEV5L and ELSYM5 programs are widely available and are
in the public domain. The Federal Highway Administration
sponsored the modification of the program to operate on a
microcomputer with a full screen editor for inputting the data.
This program is available from the McTRANS 42 Center at the
University of Florida.

BISAR is the most powerful of the ELT programs. It can

handle up to 10 layers and 10 different loads. Burmister's theory
has been modified in this program to permit the analysis of shear
loads at the pavement surface and varying interface continuity
between the pavement layers ranginG from full continuity to full
slip. The mathematical techniques used in the BISAR programs are
reported to be more sophisticated than the CHEV5L program44
Shahin, Krichner, and Blackmon 43 demonstrated the application of
the capability of the BISAR program for the analysis of slip
between an overlay and the original pavement surface. Parker et
al. 44
selected the BISAR program for use in a pavement design
procedure developed by the Corps of Engineers. Reasons for this
selection included the mathematical sophistication of the solution
process and the ability to analyze varying interface conditions.

The capabilities of the CHEV5L program have been expanded by

several researchers. Shahin 45 developed an iteration method for
introducing stress sensitivity into the elastic layer theory
analysis. Other modifications permit the analysis of up to 15
layers.

54
 Zaniewski 46 compared the output of the three programs. As
shown in Table 3-1, the surface deflection and horizontal strain
at the bottom of the surface are equal for ELSYM5 and CHEV5L and
are similar to the BISAR output. There are considerable
differences in the computed strains at the top of the subgrade for
the three different programs. Parker et al. also found
significant differences in the computed deflections near the load
although the stresses and strains were not very different between
the BISAR and CHEV5L programs.

3.2 VISCOELASTIC LAYER ANALYSIS

There have been several theories proposed for the analysis of
pavement structures. Aston and Moavenzadeh 47 reviewed the
approaches to viscoelastic modeling, then continued with the
development of the VESYS program for the Federal Highway
Administration. The following discussion is based on information
from Aston and Moavenzadeh. Viscoelastic models may be placed in
two broad classes, rheological models and creep/relaxation
functions.

The rheological approach uses discrete models of springs and

dashpots in series and/or parallel to characterize the visco-
elastic material behavior. Mathematical complexities arise when a
large number of elements are used. Thus, models are limited to
two to five elements that limit the ability of this approach to
model the behavior of real materials.

Literature citations and comments by Aston and Moavenzadeh

concerning the use of the rheological approach identified the
state of the art in these models as of 1967 as:

Lee48 illustrated the basic idea in his paper in 1955

with the solution for a fixed and moving point load on a
viscoelastic half-space which was assumed to behave as a
Voight model in shear, and to behave elastically in
hydrostatic tension or compression. In 1961, Pister 9
presented the solution for a viscoelastic foundation
under a uniform circular load where both the plate and
the foundation were assumed to behave as incompressible

55
 Q. W) I r'.Q ~ .
Vu 4 -7i J-4C .1 0 *

(n r w r' ~) Y. N-
I
-4-4
I I
P
41~

P.4

0 0 4 Az C
Cd) N -4 -4 000u~ O 1 E 2 ~ 3%A T
:3 0 0i ~ 0 0 C% 0 N VIA .7
M. '.I .4 61

'A 0

(n -4-

7,' ;- Go- 0 0% 0

o4

04 a o 0 0
n - n.. 4- VI N -

>4 - -443

>4 -4 C 04

-4 0 0 0 13 N - VI

E-4 0'' 'L

N -7- V) w Z'

U)w -44 0 /10

Q rj

IV ?A r 0j
M-

F,- . -4tv w 0
E-4- 1) C67 V. .0 0 >7

>' F3
.11~~ z -w4-

oo UO -4

O 56 '0
 Maxwell materials. In 1962, Dister and W.entman 0 used a
three-element model to characterize the hhavior of a
beam on a Winkler foundation and analyzed this for a
moving point load. Ishihara presented a solution in
1962 for a two-layer viscoolastic system in which he
assumed that the layers were characterized as
incompressible Voight and Maxwell models. However, he
examined the behavior only at zero and infinite times.
Kraft presented an analysis of the deflection of a
two-layered system in 1965 in which the layers were each
composed of three-element models, and the volumetric
.behavior was assumed to be elastic.

Based on this review of the state of the art, Aston and

Moavenzadeh rejected the use of the discrete rheological model
approach in favor of the creep/relaxation method. Some of the
limitations of the rheological models were attributed to limited
computer capabilities at the time of the review. However, it
appears the selection of the creep/relaxation method was selected
based on the fundamental capabilities of the models rather than
the computer capabilities.

Aston and Moavenzadeh attribute the development of the

approach to Lee and Rogers who used numerical techniques
suggested by Hopkins and Hamming53 . The model introduced by Aston
and Moavenzadeh was refined and reported on by Moavenzadeh and
Elliot54 . The viscoelastic model starts with the Burmister
formulation of the primary response of the pavement system.
Modifications are then introduced to model a limited time duration
of the load and the viscoelastic form of the constitutive equation
_for the material characteristics. The creep compliance function
used in this development was:

iD- EGI(e)t6i , j = 1, 2, 3,
where:
Dj = the creep compliance function
e = natural base
t = time interval
G1 = coefficient of the Dirichlet series
5, - exponent of the Dirichlet series

b/*
 G i and 5i require experimental determination of the complex
modulus and the time temperature shift function of the asphalt
concrete. Rauhut et al. 55 demonstrated that rutting predictions
with the VESYS program developed for the Federal Highway
Administration were very sensitive to these parameters. There
have been several revisions to the VEoiS model to refine
capabilities. Hufferd and Lai 56 expanded the capabilities of the
original program to include "N-layers" and reformulated the rut
prediction procedure to reduce the complexity of the mater-Lal
characterization.

Khosla 57 described the inputs to the VESYS IIIA model as:

1. Geometry of the pavement system: the thickness of the

first (N-i) layers with the thickness of the N layer
being infinite.

2. Traffic loadings: numb)er of 18,000 pound equivalent axle

loads per day, intensity, and duration of loads.

3. Temperature: average seasonal temperature and winter

design temperature.

4. Material Response Properties: modulus of resilience and

Poisson's ratio of every layer for every season; these
properties are needed in o-der to calculate the stresses,
strains, and deflection response in a pavement system
under the application of external loadings.

5. Material Damage Properties: fatigue coefficients of the

surface layer and permanent deformation parameters of
every layer; these coefficients and parameters together
with the stress, strain, and deflection response in the
pavement system are used to estimate the pavement damage
in terms of cracking, rutting, and roughness under various
stages of its life.

This list of input requirements demonstrates the VESYS model

considers fatigue cracking as well as the rutting of the pavement
trom viscoelastic strains. The fatigue model will be addressed
later.

58
 Khosla concluded the structural subsystem of VESYS IIIA
predicts pavement performance accurately and that the triaxial
compression test can be used for easuring the material
properties. In another paper presented at the same conference as
the Khosla paper, Beckedahl et al.5 8 criticized the ability of the
VESYS model, particularly its characterization of the material
properties. Beckedahl et al. prop)sed several improvements to the
model including the development of procedures for capturing
fluctuations in the material properties over the life of the
pavement.

3.3 LAYER ANALYSIS - DYNAMIC LOADS

Research into the development of models that capture the
dynamic nature of traffic loadings has been performed for many
years. Many researchers have included dynamic analysis in models
of viscoelastic behavior due to the need to include the duration
and rate of loading for estimating viscoelastic deformations.
Lattes, Lions, and Bonitzer 59 and Bastiani 60 presented analytical
approaches for the calculation of the response of a pavement to
dynamic loads. In both papers, extensive equations are presented
for the dynamic analysis but numerical results are not available
due to the complexity of the calculations and the lack of "large
high capacity computers." Similarly, several authors at the
Second International Conference on the Structural Design of
Asphalt Concrete Pavements presented formulations for the dynamic
analysis of pavements, including papers by Ishihara and Kimura6 ,
and Perloff and Moavenzadeh62 , and Avramesco3 3 .

Mamlouk 63 reported on the use of dynamic analysis of

pavements, which is the source for the following discussion.
Dynamic m)dels capture the inertia components of the deflection of
a pavement in response to moving loads. The simplest models use i
single degree of freedom to represent the motion of the
pavementwhile multiple degree of freedom models are more complete
and necessarily more complex.

59
 Single degree of freedom models developed by Weiss6,65 use a
combination of masses, springs and dashpots to represent the
pavement, as shown in Figure 3-2. The spring represents the
stiffness of the pavement, the dashpot represents damping of the
pavement materials and the mass represents the weight of the
pavement. When a force is applied to the model, the mass deflects
and may oscillate before coming to rest. Oscillations are a
function of the relative magnitudes of the mass, spring stiffness,
and dashpot dampening coefficient. If there was no mass in the
system, there would be no vibration and, therefore, no dynamic
response and the problem could be analyzed with static models.
Since pavements have mass, there is an expectation of a difference
in the dynamic and static response of the pavement. Although the
single degree of freedom model considers the dynamics of pavement
response, it is limited to vertical loads and responses. Thus,
there is no consideration of the propagation of the response
laterally through the pavement. Deflections at points away from
the load cannot be modeled.

Earlier work by Szendrei and Freeme presented a seven-

parameter model of a pavement, as shown in Figure 3-3. This model
uses a resistance component (Z, on Figure 3-3) to couple two
masses, each with a spring and dashpot component. The researchers
used a vibratory pavement deflection device to quantify the
parameters of the model before developing equations for computing
pavement response to an impulse traffic load. Data presented in
the paper demonstrated good correlation between the pavement model
and measured response.

A muitidegree of freedom model can capture the inertial

effects in three dimensions. Dynamic wheel loads are represented
by a series of hal 4 sine waves. The wave (transient) mode of
loading is represented by a series of harmonic loadings with
different frequencies and magnitudes using Fourier
transformations. Once the pavement response to harmonic loading
as a tunction of the frequency and magnitude is evaluated, the
response to any wave can be obtained.

60
 MASS

SPRINGI DASHPOT
FIGURE 3-2. SPRING-DASHPOT SINGLE DEGREE OF FREEDOM PAVEMENT
MODEL (MAMLOUK)

APPLIED FORCE

S2 ! R2

FIGURE 3-3. COUPLED MECHANICAL MODEL WITH SEVEN PARAMETERS

(SZENDREI AND FREEME)

61
 The single degree of freedom model considers the dynamics of
pavement response, it is limited to vertical loads and responses.
Thus, there is no consideration of the propagation of the response
laterally through the pavement. Deflections at points away from
the load cannot be modeled.

The governing equation for steady-state elastodynamics is the

Helmholtz function 67 written in tensor form as:

2
(A)uil1 + (E + W)u 1 ij +P& Ui = 0

where:

£,g = Lame's constants

P= mass density

C) = angular frequency of excitation and

ui = ith Cartesian component of the displacement vector

In the equation, the Cartesian indicial notation is assumed

in which the subscripts range from 1 to 3, addition is implied
over repeated subscripts and a comma denotes differentiation with
respect to the space variable, i.e., Ui j = dui/dx i . Thus, this
tensor form differential equation is a short representation of a
number of regular differential equations. The time displacement
is also assumed to be a time harmonic.

Analytical or closed form solutions are not available for the

solution of the displacement equation for layered systems. Kausel
and Peek developed a numerical solution, in the form of a
computer model, based on the assumption that the displacement
field is linear in the direction of layering between adjacent
interfaces. This requires use of sufficiently thin layers
toensure the validity of this representation. This may require
subdividing the pavement layers for the purposes of analysis.

62
 The Kausel and Peek program computes displacements. This
program was modified by Sebaaly 69 to include the calculation of
stresses and strains.

3.4 THIN PLATE THEORY

As opposed to flexible pavements which distribute the load
gradually through the pavement structure, rigid pavement slabs act
as a structural element (a plate) resting on an elastic
foundation. Since the deflection of rigid pavements is small
3
relative to their thickness, they can be analyzed as thin plates
The following approximations are required for the development of
thin plate models:

1. There is no deformation in the middle plane of the slab;

this plane remains neutral.

2. The planes in the slab initially normal to the middle

plane of the slab remain normal after bending.

3. The normal stresses in the direction transverse to the

plane of the slab can be ignored.

Sargious reports that the differential equation describing

the deflected surface of a slab subjected to a uniform load was
developed by LaGrange in 1811, and identifies Westergaard07 ,71 as
the first to develop a theoretical solution for rigid pavement
design. Sargious identifies the assumptions used in the
developmen' of the Westergaard equations as:

1. The concrete slab acts as a homogeneous elastic solid

in equilibrium.

2. The reaction of the subgrade is solely vertical,

and proportional to the deflection of the slab.

3. The reaction of the subgrade per unit area at any

given point is equal to a constant, K (modulus of
subgrade reaction), multiplied by the deflection at
that point.

4. The thickness of the slab is uniform.

63
 5. The load at the interior and at the corner of the
slab is distributed uniformly over a circular contact
area; for the corner loading the circumference of
this circular area is tangential to the edge of the
slab.

6. The load at the interior edge of the slab is

distributed uniformly over a semicircular contact
area, the diameter of the semicircle being along the
edge of the slab.

Although not commonly stated, it should be noted that

Westergaard also assumed a static load condition. Based on these
assumptions, Westergaard developed equations for computing the
stresses in the slab for the following cases:

1. Maximum tensile stress at the bottom of the slab due to

central loading.

2. Maximum tensile stress at the bottom of the slab for an

interior edge loading in a direction parallel to the edge.

3. Maximum tensile stress at the top of the slab in a

direction parallel to the bisector of the corner angle for
corner loading.

Subsequently, Westergaard modified the equations specifically for

the analysis of airport pavements assuming elliptical load areas
and load transfer across the joint or edge of the pavement. In
1951, Pickett presented equations for "protected" and
"unprotected" corners. The Westergaard equations are widely used
for the design of concrete pavements. Pickett and Ray 72 developed
influence charts for the solution of these equations and
Packard737 4 incorporated them into a generalized program for the
design of portland cement concrete pavements.

3. 5 NUMERICAL METHObS
There are two basic numerical techniques that can be applied
to the analysis of pavement structures: finite differences and
finite elements. The application of finite difference methods for
the analysis of rigid pavementz was dcveloped at the University of
Texas in the 1960s and several successful computer programs were

64
produced . However, advances in the finite element method (FEM),
along with the development of more powerful computers, have led
researchers to concentrate on the development and application of
the finite element method in preference to the finite difference
method. This fulfills the prediction of Nair who stated in 1971,

The finite difference technique has been used

fairly extensively in the analysis of plate
problems. However, because of the difficulties in
handling corners and because of the physically
motivated formulation of the finite element method,
most of the new developments in the analysis of
plate problems are likely to be with the use of
finite element techniques.

Several researchers have developed finite element computer

code for a wide variety of pavement analysis probipms. Due to the
flexibility of the method, the applications will probably continue
to grow. The flexibility of the FEM stems from the basic
definition of the method as "a computer-aided mathematical
technique for obtaining approximate numerical solutions to the
abstract equations of calculus that predict the response of
physical systems subjected to external influences 76 .

This is a very broad definition of the finite element method

and is not necessarily common to all engineers that work in the
77
area of numerical analysis methods. For example, Ioannides
developed a numerical analysis method for evaluation of slabs on
grade that meets the Burnet's FEM definition, yet Ioannides claims
the method is not a finite element solution.

Under Burnet's definition, FEM is not so much a model per se

but rather an analytical technique for solving a problem once the
proper equations have been formulated. However, use of the FEM
required formulating the problem in a specific manner to promote
the development of an accurate solution. Burnet points out that
most people's contact with FEM will be as users rather than as
developers of computer programs. Users do not become directly
involved with the underlying mathematics. However, experience has

65
shown it is difficult to be an effective user of an FEM program
without understanding some of the basic concepts and mathematical
techniques employed by the method. In this vein, the Burnet's
description of the salient features of every FEM is presented
followed by a description of various applications to pavement
analysis.

Burnet defines the system as the subject of the model;

generally, but not always, the system is a physical object, such
as a specific section of pavement. The domain is the region of
space that is occupied by the system. It may also be an interval
of time during which there are changes in the system. The
governing equations describe a conservation or balance of a
physical property such as mass, momentum, or energy. They may be
differential or integral equations or constitutive equations that
describe material behavior; these equations contain experimentally
determined physical properties of the materials that constitute
the system. Loading conditions are externally originating forces,
temperature, etc., that interact with the system causing the state
of the system to change. Loads acting in the interior of the
domain (interior loads) are included in the governing equations.
Loads acting on the boundary of the domain (boundary loads) appear
in separate equations called boundary conditions.

The domain of the problem is divided into smaller regions

called elements. The shapes of the elements are simplified as
much as possible. The entire mosaic-like pattern is called a
mesh. Mesh generation is generally performed by a preprocessor to
the finite element analysis program based on the geometry of the
domain and the accuracy of the required solution. There is a
direct tradeoff between mesh size, accuracy of the solution, and
the amount of computer time required for the solution.

In each element the governing equations, uzually differential

or variational form, are transformed into algebraic equations,
called element equations, which are an approximation of the

66
governing equations. Algebraic equations are much easier to work
w~th than calculus equations. Derivation of the element equations
is a theoretical procedure performed by the analyst or program
developer. The element equations are algebraically identical for
all elements of the same type. Consequently, element equations
usually need to be derived for only one or two typical eiements,
not every element in the mesh. In addition, since the element
geometry is simple, derivation of the equations is usually
straightforward. The analytical effort for the entire problem has
been reduced to deriving a few algebraic equations for usually
only one or two small elements.

The terms in the element equations are numerically evaluated

for each element in the mesh (internally by the computer program)
The results are assembled into a set or system of algebraic
equations. This system of equations characterizes the response o-
the entire system and generally constitutes a very large nunher of
equations. However, the solution of the equations is econorical
because the matrix of coefficients is "sparse."

The boundary condition-, including the external loading, ar,-

then imposed by modifying the system equations. This involvc;
adding values to existing terrs and/or shifting terms from on'.
side of the equations to the other.

Numerical analysis techniques are then applied to the

sciution of the equations. These techniques have been rerined t:
t.... advantage of the formulation of the FE system equations to
provide efficient solutions.

Post processing displays the solution ot ',he Peations in

tabular, graphical or pi-torial form. Post process i can ilsn
include the derivation of other meaningful quantities from the
r;o lut ion.

67
 Much of the work described above needs to be performed only
once, when developing the computer program. Application of an
existing program requires supplying specific data on the
constituent constants, mesh generation commands, and output
specifications. Due to the popularity of FEM, there are many
commercially available programs that provide a wide variety of
capabilities.

Duncan, Monismith and Wilson78 demonstrated the use of the

linear elastic finite element model SAP for the analysis of
pavements in 1968. Pichumani9 compared two finite element
programs with elastic layer theory models and concluded the models
were almost identical with regard to stresses and strains, but
there were slight differences in the computed deflections
attributed to the differences in the boundary conditions. The FEM
applications by these authors did not offer any advantage over the
analytical solutions to elastic layer theory. However, they did
demonstrate the applicability of this approach to the analysis of
pavements.

Subsequent applications in the use of FEM for the analysis of

pavements have permitted modeling of more complex material
characteristics and pavement geometries than is possible with the
analytical solution methods developed by Burmister and
Westergaard. The FEM procedure has become very popular and there
are numerous publications regarding its application. The
following paragraphs discuss a number of these publications.

Raad and Figueroa presented an FEM, ILLI-PAVE, for the

28
analysis of flexible pavements. Gomez-Achecar and Thompson
summarize the characteristics of this model as:

1. an axisymmetric solid of revolution, e.g., Figure 3-4;

2. nonlinear, stcess-Ud-pendent resilient modulus of layer

materials; and

68
 I 200"
PRigid Plate (10" X 10") P/ 1500 b

*Granular Base-- -- - 20

Subgrade ----- -

180".

FIGUJRE 3-4. AXISYMNETRIC FINITE ELEM4ENT MODEL, OF TJWO TAYER SYSTEN

(GOMEZ-ACHECAR AND THOMPSON)

69
 3. limits on the principal stresses in granular and
fine-grain soils so that they do not exceed the
Mohr-Coulomb theory of failure.

Gomez-Achecar and Thompson report that several authors have tested

the validity of the ILLI-PAVE program with favorable results.

MICH-PAVE, a nonlinear-anisotropic FEM for the analysis of

flexible pavements was developed at Michigan State University81
The axisymmetrical formulation is capable of calculating stresses
and strains and the surface deflections developed in a pavement
section due to a wheel load.

Smith and Yandell82 present a discussion of a "mecho-lattice"

FEM of elasto-plastic behavior applied to flexible pavements.
This model considers different modulus values for the soil support
in the loading and unloading modes. A dampening factor is used to
capture the plastic behavior of the material. Transient wheel
loadings are modeled by transferring the load across the surface
nodes. Unbound material behavior can be modeled by permitting
cracks to form when tensile stresses are computed in the material.
The boundary conditions are described as follows:

All the joints on the top surface are free to move. Since
the sides and approached end cross section are often within
the deflection bowl, the theory of linear elasticity is used
to designate boundary deflections. Provision is made for
lateral and longitudinal continuity for each layer with
plastic hystereses on the sides. The main conditions to be
satisfied for the passed end cross section is that the
residual longitudinal strain must equal zero whilst still
allowing flow; also the vertical and lateral strain must tend
not to vary at points that vary in the longitudinal
coordinate only.

Lim and Yandell83 used the mecho-lattice analysis procedure

to reevaluate the Shell pavement design criteria. In this work,
residual stresses were considered for the determination of ruttinq
and fatigue cracking.

70
 Sargious3 and Wang, Sargious, and Cheung 8 presented a general
description of FEM for the analysis of concrete pavements. The
foundation for the slab is modeled as either an elastic continuum
or as a Winkler foundation. The differences in these
foundation models are shown in Figure 3-5. The Winkler
foundation, as also used in the Westergaard solution, consists of
a series of springs having a constant modulus of reaction, K. The
reaction of subgrade per unit area at any point is proportional tc
the vertical deflection at that point, but independent of the
vertical deflection at any other point. The elastic continuum
foundation is considered as an idealized half space. The
flexibility matrix for the foundation is obtained by determining
the deflections at all points for each location of a unit vertic7i
point load.

The stiffness matrix of the foundation is obtained by

inverting the flexibility matrix. It is then combined with the
slab matrix to obtain the complete stiffness matrix.

Chou85 describes two programs developed by the Corps of

Engineers with similar formulation and capabilities. These are
known as the WESLIQID and WESLAYER programs. Use of these

programs to investigate the contact pressures under rigid

pavements led Chou8 6 to conclude subgrade support has a greater
influence on pavement life than is indicated with the Westergaard
formula.

The ILLI-SLAB and FEACONS are FEMs developed specifically

for the analysis of rigid pavements. FEACONS uses a three-slab
model as shown in Figure 3-6. The middle slab is of interest in
the analysis; the outer slabs establish boundary conditions. A
concrete slab is modeled as an assemblage of rectangular plate
bending elements with three degrees of freedom at each node. The
slab can be homogeneous or a composite slab consisting of two
layers bound together. The subgrade is modeled with a Winkler
foundation. Load transfer across the joints are modeled by shear
and torsional springs. Frictional effects at the edges are

71
 pm

177-7777 7 7 7 7 -Z 77-7 -

r P

(a) (b)

FIGURE 3-5. DEFLECTIONS OF PAVEMENT SLAB FOR THE (A) ;'!::LER

AND (9) ELASTIC CONTINUUM SUPPORT CONDITIONS (SARGIJUS)

72
 V

FIGURE 3-6. FINITE ELEMENT MODEL OF A THREE SLAB PAVEMENT SYSTEM

(TIA)

73
modeled by shear springs along the edges. The concrete is modeled
as linearly elastic and isotropic. Force vectors due to the
weight of the slab, thermal gradient and loads are applied
incrementally. The stiffness matrix is adjusted at the end on
each increment according to the new subgrade support conditions.

Tirado-Crovetti et al. 89 demonstrated the use of FEM for the

analysis of the stress intensity factor used in a fracture
mechanics approach for the prediction of reflection cracking of
asphalt concrete surfaces. A similar approach was applied by
Majidzadeh et al.9" for the analysis of reflection cracking of
airfield pavements.

This brief review of some of the applications of FEM for the

analysis of both flexible and rigid pavements demonstrates the
flexibility of the analysis method. The fact that elemental
models are available in the literature means that further
application of the FEM for pavement analysis is a promising tool.
It appears the main constraints to the application of this method
will rest in the development of more constitutive equations to
further relax the number of assumptions of material behavior.
Development of these equations must be supported by the laboratory
procedures to quantify the material behavior described by the
equations.

Traditionally, the other constraint to the use of FEM has

been computer costs. Numerical methods are more computationally
intensive than analytical methods. However, consideration of the
consequence on engineering decisions based on the output of the
models would always favor the use of the most applicable theory
for the analysis of pavements.

3.6 ENVIRONMENTAL MODELS

Materials reduce in volume as their temperature drops.
Pavements resting on a base material are constrained by the weight
of the pavement surface causing forces at the interface. When the
temperature of the pavement drops, the constraint to movement

74
causes the development of stresses in the surface. When these
stresses exceed the strength of the material, transverse cracks
develop. This basic mechanism is the reason conventional concrete
pavements are constructed with joints. The need for design models
for the selection of joint spacings in concrete pavements has led
to the development of models of the environmental stresses in
concrete slabs. Although the mechanism also affects flexible
pavements, there has been relatively little research into
quantifying the effect.

The model developed for computing the stresses in a concrete

slab due to friction is shown in Figure 3-7. Balancing the forces
defined in the figure yields:

cc = WLf/(24h)

where:

Uc = "friction stress"

W weight of Uie slab (psf)

L = length of slab, ft

f = average coefficient of subgrade resistance

h depth of the slab, in.

The f term is sometimes called the friction between the slab

and the subgrade. However, this is incorrect since the
contraction of the slab results in shear forces that are
transmitted into the subgrade2 .

For jointed concrete pave.nents, the stress due to friction

can be computed and compared to the concrete strength. If the
stress is excessive, a shorter slab is designed. For the design
of continuously reinforced concrete pavements, there is no length
of slab and distance between the cracks is unknown. McCullough et
al. 91 de-,eloped a numerical tecnnique for the analysis of crack
spacing in reinforced concrete pavements. This model estimates

75
 L/2
I IA
00

.Friction

(a)

Free end e Center of Slab

Fully mobilied
resistance f

!= 1000

At =temperature drop (degrees F)

L/2 -

(b)

FIGURE 3-7. STRESSES RESULTING FROM CONTRACTION DUE TO

TEMPERATURE: (A) FORCES ACTING ON CONTRACTING SLAB; (B) VARIATION
OF SUBGRADE RESISTANCE WITH LENGTH (YODER AND WITCZAK)

76
crack spacing as a function of temperature drop, dryinv-, shrinkage,
moisture change, wheel load, and coefficient of subgrade
resistance.

Warping of a concrete slab is developed by a thermaJ gradient

in the slab. Yoder and Witczak 2 presented equations formulated by
Westergaard7 0 71
along with the solution developed by Bradbury 92 as:

edge stress

a = CEat/2
interior stress

= (Eat/2) ((C 1 +MC2 ) / (I-M 2 ))

where:

C1,C? = coefficients defined as functions of the

relative stiffness of the slab

a = coefficient of thermal expansion of concrete

L = Po;-son's ratio

E = elastic modulus

t = temparature differential between the top and

bottom of the slab

Traditionally, the principle of superposition is used to add

the curlinq and traffic stresses to obtain the total stres- in the
Iivement. However, Ioannides and Salsilli-Murua 93 report this
principle does not apply due to the loss of support of the slab
during curl ig. They resort to FEM for the analysis of the
combined temperature and loading stlesses du( to the lack of
-losed-form model.

77
 4. DISTRESS MODELS

The mechanistic models described in the preceding chapter

compute the response of the pavement to traffic and environmental
loads. Relationships are now needed for using these responses to
estimate the life of the pavement. Since pavements are designed
for multiple applications of repeated loads, the distress models
should capture the accumulated damage to the pavement materials.
However, damage is a vague term. Pavements display many types of
distress and, in general, damage functions must be developed for
the prediction of specific distress types. In the related
literature, there is a preponderance of information on two primary
distress types:

1. fracture of the pavement surface due to repeated a :le load

applications, fatigue; and

2. uistortion of the surface to repeated accumulation of

plastic or viscous strains in the wheel path, rutting.

Several other distress types have also been examined in the

literature including reflection cracking of asphalt overlays of
jointed pavements and stripping of asphalt concrete.

4.1 FRACTURE
Fatigue is generally considered to be the fracture of the
pavement surface due to the repeated application of traffic loads.
Both asphalt and concrete pavements are subject to fatigue
failure. Two basic approaches have been taken to the modeling of
fatigue behavior: phenomenological and power law. In the
phenomenological approacn, the number of applications a pavement
can carry prior to failure is estimated directly as a function o!
the stress or strain levels generated by the traffic loadings.
The power law approach uses concepts developed in fracture
mechanics to estimate the growth of a crack through the pavement
layer.

79
 The phenomenological approach has been widely appiied in the
analysis of pavement fatigue life. Although both asphalt and
concrete pavements will fail in fatigue, the behavior of these
materials with respect to repeated loads is very different in that
concrete appears to have a "fatigue limit." It is generally
assumed that concrete will not fail in fatigue if the stresses in
the slab are kept below 50% of the modulus of rupture, as measured
with the flexural test.

Conversely, asphalt concrete does not appear to have a

fatigue limit. In other words, repeated application of even the
smallest level of strain will eventually result in fatigue
failure.

4.1.1 Phenomenological Model

The fatigue life of the asphalt is generally related to the
strain in the pavement as:

N = a(l/E)b

where:

N = the number of applications to failure

a,b = fatigue-life coefficients

E = strain in the asphalt concrete

Some authors have modified this equation to include the stiffness

of the asphalt and others substitute stress for the strain term.
Laboratory tests have demonstrated that a and b coefficients are
sensitive to mix design parameters, such as asphalt content, air
voids, aggregate gradation, etc., and to the mode of testing.
Attempts to compare the results of laboratory testing with field
performance have demonstrated poor correlation. Generally,
pavements last longer than laboratory fatigue testing and analysis
predict. The difference between field performance and laboratory

8o
estimates are expressed in terms of a shift factor. This shift
factor is frequently in the range of 20 to 25, but factors as
large as 1000 have been reported.

In addition to the problem of the correlation of laboratory

and field testing, the variability of fatigue testing should be
considered in the development of a design procedure based on the
phenomenological approach. Navarro and Kennedy9 4 reported the
coefficient of variation of laboratory fatigue testing ranges from
53 to 73%.

There are many examples of the use of the phenomological

approach for the prediction of the fatigue life of highway
pavements. However, there are relatively few examples of the use
of this approach for the design and analysis of airport pavements.
Kelly and Thompson95 used the ILLIPAVE program and a fatigue
equation for the analysis of airfield pavements for F15 aircraft.
Strain at the bottom of the asphalt concrete surface was computed
with ILLIPAVE and the fatigue life was computed with the equation:

N = 259(1/6)3.16(l/E) 1.4

where:

N = number of estimated applications of strain

E = strain at the bottom of the asphalt concrete layer

E = modulus of the asphalt concrete.

This equation was developed through an ILLIPAVE analysis of the

AASHO Road Test96 .

The Waterways Experiment Station has developed an elastic

layer theory method for the design of flexible pavements that use ;
the phenomenological approach for estimating the fatigue life of
airfield pavements 97 . This design procedure uses the equation:

' 5
N = 479(1/E) (I/E) -

81
This equation was developed from an analysis of experimental
pavement sections designed and tested to simulate airfield
pavements 98

While the WES and the Kelly-Thompson equations have the same
form, the coefficients are considerably different. For a strain
of 0.0005 in/in and a modulus of 500,000 psi, the WES equation
estimates approximately 10,000 repetitions can be applied to the
pavement while the Kelly-Thomas equation estimates in excess of
73,000 applications. This demonstrates the problem with the
phenomological approach to fatigue analysis. Prediction models
developed by different researchers appear to be more a function of
the analysis procedure used in the development than determined
from the basic properties of the materials.

4.1.2 Power Law

A power law approach to the estimation of the fatigue life of
a pavement was developed by Majidzadeh et al.9 '100
'1 01 and
Ramsamooj 1 2 based on the application of fracture mechanics.
Fatigue is considered to be a process of accumulative damage
where, under a given stress state, damage grows according to a
crack propagation law from an initial state to a critical and
final level. The form of the crack propagation law is:

n
dc/dN = Ak

where:

dc/dN = rate of crack propagation

A,n = material constants

k = stress intensity factor

This equation has the same basic form as the phenomenological

equation. Thus, the primary differences between the two
approaches are the calculation of the stress intensity factor at
the crack tip and relating that to the local failure of the
material, and the method for quantifying the material constants.

82
Determination of the stress intensity factor requires analysis of
the discontinuity of the pavement material in the area of the
crack. Currently, this requires the application of FEM to model
both the material behavior and the stress intensity.

The material constants in the power law equation are related

to the ability of the material to absorb energy before fracture.
Jayawickrama and Lytton I1 3 have shown that for viscoelastic
materials, n is inversely proportional to the slope of the
stiffness-load time curve on a log-log scale and a linear
relationship exists between n and log(A). This extension of
fracture mechanics from linear elastic materials to viscoelastic
materials is based on the work of Schaprey I .

It should be noted that the power law predicts the growth of

a crack rather than an instantaneous failure when a fixed number
of strain applications have been applied. Actually modeling the
growth of the crack is superior to the phenomenological approach
in that the method is not limited to the analysis of the repeated
load applications. Thus, the power law is applicable to a broad
category of problems. Examples of applications of the fracture
mechanics and power law relationship to pavement analysis include:

1. George 1 5 for the analysis of soil cement pavement layers.

2. Crockford and Little1 06 also for the analysis of soil

cement bases in pavements.

3. Majidzadeh et al. 10 for the analysis of reflection

cracking of asphalt overlays on concrete pavements.

4.2 DEFORMATIONS
Plastic and viscous deformations of the pavement materials
result in permanent deformations of the surface. Channelized
traffic generates an accumulation of deformations in the wheel

83
paths of the vehicle or rutting. Three approaches have been
defined for relating the primary response of pavements to rutting:

1. limiting the magnitude of the maximum vertical strain in

the subgrade;

2. relating permanent strains to stresses or strains computed

with elastic theories; and

3. direct estimates of permanent strains with viscoelastic

models.

1 08
The limiting strain approach was first presented by Dorman
in 1962 and Klomp and Dorman1 0 in 1964. It is currently used in
the Asphalt Institute airfield pavement design method110 and the
Joint Department of Army and Air Force elastic layer theory method
for the design of flexible pavements. The basic hypothesis of
this approach is that if the maximum compressive strain at the
surface of the subgrade is less than a critical value, then
excessive rutting will not occur for a specified number of
repetitions. These relationships were developed based on analysis
of the Corps of Engineers pavement design procedures.

The Army and Air Force elastic layer theory procedure for the
design of flexible pavements uses limiting subgrade criteria for
estimating the number of applications a pavement can withstand
before excessive permanent deformation occurs. The criteria are
specified by the equation:

N = 10,000(A/Ss)

where:

N = number of applications the pavement can sustain at a

given strain level

Ss = vertical compressive strain at the top of the subgrade

A = 0.000 2 7 4 +0.000 245

logM R

MR = Modulus of the subgrade

559
B = 0.0 658 (MR) 0°

84
 Relating permanent deformation to elastic stresses has been
proposed by several researchers. Some have assumed a fundamental
deformation jaw exists through which permanent strains can be
predicted based on the stress state of the material. Other
authors have used statistically formulated equations for relating
the permanent and elastic strains. Neither method has been
particularly successful since the concept, from a mechanistic
aspect, is not fundamentally sound and there has not been an
adequate database for developing the statistical models.

Calculation of permanent strains is incorporated in the

constitutive equations of viscoelastic theory. By definition,
rutting can be estimated directly from the application of this
theory and thus several authors have dismissed the need for
further model development in this area. To some extent this claim
is justified by the verification studies of the VESYS IIIA program
by Khosla111 and Sneddon1 2 . This is not to say, however, that
VESYS IIIA is the final answer to the prediction of permanent
deformations since the response model is based on elastic layer
theory.

85
 5. AIRPORT PAVEMENT DESIGN METHODS

As described in the preceding chapters, pavements are

complicated structures, composed of materials whose behavior is
difficult to characterize. They are continually subjected to
diverse traffic loadings and environmental conditions. The
spectrum of factors affecting pavement performance has led
engineers and researchers into the development of multiple
theories for the analysis of pavements. This has also led to the
development of multiple methods for airport pavement design.
Methods have been developed by the Federal Aviation
Administration, the Department of Defense, trade associations, and
consultants and researchers.

The purpose of this chapter is to review several pavement

design methods with respect to the theories and concepts used fcr
the establishment of the design curves. For rigid pavements, on
the pavement thickness design methods are reviewed. A- o-f the
rigid pavement design procedures also include the desiqn of joint
details and, in some cases, reinforcement design. It is nct the
intention of this chapter to present a detailed reproductic-r o.
the design methods since they are readily availabie in tho
li teiature.

. FEDERAl AV7IATION ADMINISTRATION

The Federal Aviation Administration pavement design"

-rcedJi*
presents methods for the design ot flexible and rigid pa'eements
tor light and heavy aircraft 3 . C:i} the procedure for ....
aircraft, weighing mote than 30,000 pounds is reviewed _ .
Flexyible pavement design procedure is base:d on tle Ci] _fo-n..
B,, -ring Rati. (CBR), L-,st. The rigid pavement design prc :-ciu-e
based on the Westergaard stress equation for ,i :it edge stress.

All paoerent designs are based on a 20-year 3esijg, Iife ar.

the traffic vclume is expressed in terms of tht annual nee of
departures. A!! departures are assumed to be at 95% of the
maxinum aircr.aft weijhtr. "The design urves are ased on the
P7
number of coverages of aircraft based on the statistical
distribution of aircraft wander across the pavement rather than
the actual number of departures. This rodification to the number
of applications is transparent to the designer and is documented
in an appendix to the design procedure.

The traffic analysis procedure is common to the design of

both flexible and rigid pavements. All classes of aircraft that
will use the facility are reduced to the number of equipilent
annual departures of a design aircraft. The design airzraft is
defined as the aircraft that would require the greatest pavement
thickness if it were the only aircraft using the facility. -,.ce
the pavement design charts consider both the number of
applications and the aircraft weight, the design aircraft will not
necessarily be the heaviest aircraft. In essence, determining the
equivalent annual applications requires:

1. Convertin, all aircraft types to the gear type of the

design aircraft by multiplying the number of annual
applications for each aircraft type by equivalency
factors.

2. Converting the number of adjusted annual applications to

the equivalent number of design aircraft applications by
using the equation:

log(R) = (w2 /wl) 1og(R 2 )

where:

R, = equivalent number of annual departures for

the design aircraft

R 2 = annual departures expressed in design

aircraft landing gea-

w I = wheel load of design aircraft

W = wheel load of the aircraft being analyzed

(the wheel load for wide bodied aircraft is
computed on the ba is of 300,000 lbs maximum
aircraft weight rather than the actual gross
maximum weight)

88
 3. The number of equivalent departures is summed to determine
the total number of departures for use in design.

5.1.1 FAA Flexible Pavement Design

The essence of the CBR design procedure is the protection of
the subgrade from overstressing by using layers of successively
stronger materials. Separate design curves are available for
single, dual., and dual tandem gears and for each of the wide body
aircraft.

In this procedure, it is assumed the asphalt concrete will

meet specified criteria in terms of mix design and cons'ruction
quality. There is no method for adjusting the thicknesses based
on the quality of the surface material. It is assumed the base
will be a granular material with a CBR of 80.

Required layer thickness is determined by the following

sequence of functional relationships:
=f(SGBR,WDNE) 1

Ts = Specified on the design charts

T = f\ZBcBR, W NE)

TBmin = f(SGBR,TP)

Ts = max(TrinTsb)

TSB =Tp - T s - TB 6

Design charts are used to quantify equations I, 3, and 4. I1

some cases, the thickness requirement for the subbase is adjusted
if the upper portion of the subgrade has a thin layer of material
near the surface.

The minimum thickness of the surface is 5 inches lor wide

bodied aircraft, 4 inches for all other aircraTt in the critical
areas, and 1 inch less for al] ctY. i .eas. The thickness of the
surface is determined based on whether or rot "-.er ar, wide
bodied aircraft and rnt on the cJesqn ajrcraft.
89
 Stabilized bases are required for all airports serving jet
aircraft weighing more than 100,000 pounds, unless there is a
history of satisfactory performance of the locally available
granular materials. Stabilized bases and subbases can also be
used if there is an economic advantage in using these materials.
A reduction in the thickness of the pavement is allowed for
stabilized materials through the use of material equivalency
factors. However, the design manual only provides ranges for the
adjustment factors based on the type of the material. Selection
of a specific material equivalency factor is left to the
experience of the designer.

The design curves are limited to 25,000 applications. For a

greater number of departures, the design is increased on a simple
extrapolation. For high traffic volumes, the total pavement
thickness is increased by a multiplier and the surface thickness
is increased by 1 inch.

5.1.2 FAA Rigid Pavement Design

The slab thickness for a concrete pavement is determined from
design charts based on the flexural strength of the concrete, the
modulus of subgrade reaction, weight of the design aircraft, and
equivalent number of applications. Subbase is required to be a
minimum of 4 inches of granular material. If the airport serves
aircraft with weights in excess of 100,000 pounds, then a
stabilized base is required. The strength of the stabilized base
can be used to increase the modulus of subgrade reaction and
thereby reduces the thickness of the slab.

As with flexible pavements, the design curves are limited to

25,000 load applications and extrapolated; multiplying factors are
used to increase the thickness of the slab. These factors are
relatively insensitive to the number of applications. Increasing
the number of applications from 25,000 to 50,000 only increases
the thickness of the slab by 4%.

90
5.2 ASPHALT INSTITUTE
The pavement design charts for the Asphalt Institute110 design
procedure for full-depth asphalt concrete pavements were developed
based on elastic layer theory analysis and two distress types,
fatigue and permanent deformation. The fatigue criteria are based
on limiting the horizontal tensile strain at the bottom of the
surface layer. The permanent deformation criteria is based on
limiting the vertical compressive stress at the top of the
subgrade. All aircraft classes are converted to the "standard
aircraft"; the DC-8-63F was "arbitrarily" selected to be the
standard. Taxiways are considered to be the critical portion of
the airfield pavement. The design process is summarized in Figure
5-1.

The inputs to the design process are the mean annual air
temperature, the design subgrade modulus, and the expected number
of repetitions of each aircraft type. The design subgrade modulus
requires performing a series of soils tests and selecting the 85
percentile value, e.g., only 15% of the subgrade tests have a
lower value than the design value. The manual recommends
laboratory evaluation of the modulus using the triaxial test
procedure specified in the manual. Approximations are available
for estimating the modulus from the CBR test, plate bearing test
and the FAA Soil Classification. The manual provides
specifications for the asphalt concrete but the properties of the
asphalt concrete are not a direct input to the design process.

The procedure requires the development of curves for the

allowable number of applications and the estimation of equivalent
applications for both fatigue and plastic deformation. Design
charts are used for selecting the allowable number of applications
as a function of the subgrade modulus and the mean annual
temperature. For a given subgrade modulus and temperature, the
required thickness is determined for several different assumed
traffic loadings and the results are plotted on a design graph for
both the fatigue and plastic analysis. This establishes the
allowable traffic curve.

91
 EE
LL.)

CNL-
4) =

06 H

2 W2
- uJi

.= 51 C
0 1A c H

C#2=
W=Z.

co U.2
(U C.5 TO)

Eo m31 W E-4

, %- - L ----

-cm

921
 The traffic equivalency curves are then used to convert the
number of aircraft loadings into equivalent loadings.
Equivalencies are determined for four wheel paths and four
different pavement thicknesses. For each assumed thickness, the
critical wheel path is determined as the one with the greatest
number of total equivalencies. The number of equivalencies and
corresponding thicknesses are plotted on the design graph to
establish the predicted traffic curve.

The intersection of the predicted and allowable traffic

curves defines the required pavement thickness for each of the
design modes. The final pavement thickness is the greater of the
thicknesses required for either the fatigue or plastic deformation
load.

That equivalency factors for each of the aircraft types is a

function of the pavement thickness is a major difference between
the Asphalt Institute method and the FAA procedure. Theory would
favor the Asphalt Institute method since the relative damage
caused by an aircraft is a function of the stiffness of the
pavement and, therefore, the equivalency factors should consider
the thickness of the pavements.

5.3 PORTLAND CEMENT ASSOCIATION

The Portland Cement Association design manua174 for rigid
pavements presents two design methods: one based on the critical
aircraft that will use the facility and the other based on the
fatigue life of the pavement. Both methods use the Westergaard
theories for determining the stresses in the pavement structure.

The steps in the critical aircraft design method are:

1. Determine the k value of subgrade support with plate

bearing tests or correlation to subgrade soil test data.

2. Select a safety factor based on the estimated operating

and load conditions. The safety factor is the ratio of

93
 the design modulus of rupture of the concrete to the
working stress that will be used for design. The safety
factor ranges from 1.7 to 2.0 for the critical areas,
and 1.4 to 1.7 for noncritical areas. The critical
areas are aprons, taxiways, hard stands, runway ends and
hangar floors. The noncritical areas are the central
portion of the runway and some high-speed exit taxiways.
The selection of a specific safety factor for an
aircraft depends on the expected number of loadings.

3. The working stress for the design is determined by

dividing the modulus of rupture of the concrete by the
chosen safety factor.

4. The required pavement thickness is determined from the

design charts as a function of the working stress, gear
load, and k value.

5. The process is repeated for other aircraft that can have

critical loads. The safety factor is adjusted for each
of the types of critical aircraft.

The PCA states that the fatigue method of design applies to:

1. Design for specific volumes of mixed traffi .;

2. Evaluation of future traffic capacity of existing
pavements or of an existing pavement's capacity to carry
a limited number of overloads;

3. Evaluation of the fatigue effects of future aircraft

with complex gear arrangements; and

4. More precise definition of the comparative thicknesses

of runways, taxiways, and other pavement areas depending
on the operational characteristics.

The fatigue method introduces three additional design parameters:

1. Traffic width for taxiways, runways and ramps;

2. Variability of concrete strength; and

3. Downgrading of service life where a good subbase support

is not provided.

94
The basic st'eps in the fatigue analysis method are:

1. Fstimate the number of operations of each type of

a, ocraft.

2. Use .nfluence charts or the PCA computer prograr to

estimate the stress each aircraft will cause in the
pavement. For design, this requires assuming a
thickness.

3. Estimate the design modulus of rupture, DMR, for the

concrete as a function of the variability of the
concrete:

DMR = MR9 0 (1-V/100)M

where:
MR9 0 = average modulus of rupture at 90 days
V = Coefficient of variation of modulus of
rupture in percent (range 10 to 18%)

M = factor for the averagc modulus of rupture

during design life, recognizing that
concrete strength increases with age
(typically 1.10).

4. Compute the stress ratio of the estimated stress for

each aircraft to the design modulus of rupture.

5. Determine the load repetition factor (LRF) tor each

aircraft. LRF is determined from stresses, the fatigue
curve and the probability distribution of aircraft
wander. The design manual provides tabular values for
LRF.

6. Determine the number of fatigue repetitions for each

aircraft by multiplying the expected number of
departures by the LRF.

7. Determine the number of allowable repetitions for each

aircraft as a function uf the stress ratio and the
fatigue curve.

8. Determine the percert of structural capacity used by

each aircraft as the ratio of the fatigue repetitions
the allowable repetitions. Sum the percent of
structural capacity used by all of the aircraft.
Adjust the percent of fatigue life used when the k valw:_
is less than 200 pci. This adjustment ranges from S fo-
a k value of 50 to I for a k value of 200 pci. The
adjusted percent c-f fatigue life should be close to but
not exceed 100 percent.

95
 The PCA manual addresses the design of continuously
reinforced concrete pavements, but recommends that the thickness
design be the same as for plain concrete pavements. This manual
states that reducing the thickness can increase deflection and
promote spalling of the joints.

5.4 DEPARTMENT OF DEFENSE PAVEMENT DESIGN METHODS

The Army and Air Force share a common procedure for the
design of rigid pavements, while the Navy has a separate design
procedure. A triservice manual is used for the design of flexible
pavements for all services. In addition, the Army and Air Force
have a manual for the design of flexible pavements using elastic
layer theory.

Each of the pavement design manuals requires designing

different pavement sections based on the airfield class and
traffic areas. For example, the Air Force defines four pavement
area types ranging from highly channelized traffic such as on
primary taxiways, type A, to low volume and low weight traffic
areas, type D. In addition, each airfield is designated as a
light, medium and modified-heavy, heavy load, or short-field
facility. The traffic area types for each facility are shown in
Figures 5-2, 5-3 and 5-4114. The aircraft design loads and number
of applications for Air Force pavement design are given in Table
5-1. The Army and Navy use a similar concept with different
terminology for the airfield designations and traffic areas.

5.4.1 Triservice Manual for Flexible Pavement Design

The design of conventional flexible pavements and flexible
pavements with stabilized bases and/or subbases are covered in
115
reference' . The design process uses the CBR procedure for the
distribution of stresses through the layered pavement section.

96
 WARM-UP APRON TYPE B

LEGEND

TYPE A TRAFFIC AREA

~ TYPE BTRAFFIC AREA

~] TYPE CTRAFFIC AREA

POWER CHECK PAO

WASHRACK

APRON,

FIGURE 5-2. AIR FORCE TRAFFIC AREAS, LIGHT-LOAD AIRFIELDS

97
 LEGEND_

TYPE ATRAFFIC AREA

TYPE BTRAFFIC AREA

TYPE CTRAFFIC AREA

__ TYPF D TRAFFIC AREA

WARM-ipPAPRNn TYPE STRAFFIC AREA
FOR LIGHT-LOAD PAVEMENT

X POWER CHECK PAD

LADDER TAXIWAY /

WASH RACK

FIGURE 5-3. AIR FORCE F TAF'F'( A RARES Mi:DITUM

, AND MODI F I D HEAVY
WOAD A IRFl El PS;

98
 LEGEND

______TYPEABTRAFFICAREA

TYPE B TRAFFIC AREA

TYPEPDDTRAFFIC AREA

WARMUP ARON YPE CTRAFFIC AREA

~PRONFOR
WAR UP LIGHT-LOAD PiAVEMEN1 0!DSIGN

F I7 TYPE B TRAFFIC AREA

FO IH-LOAD PAVEN1E ! SIGN

POWER CHECK PAD

Ur

- LDDE TAXIWAY

~U E. -~.~1WASH ACK
 zz

~~t-~e r-
~~~ tr4r--C.~ m-o
0ot~

0~0

< - C-',
C <

t, < C)- r-r--

E-4 -a z w -M

m -n

Cl l C Cl6

U) () Qn ) k o
U) 0~ cz 1)CI Zq

0T 0
oc 0 '0T : 0 r00
zn en It0 IT0 000 I
U ClO 0 Cl~~c~Cl~~e~"~ ~z
U) C

HC
kn kn

Hn OCtm -

- U

I~ :D

100
Design of full-depth bituminous concrete pavements is beyond the
scope of this report. The design sequence is:

1. Determine design CBR of the subgrade, depending on the

variability of the subgrade. (Either distinct pavement
design areas can be used, or the 85 percentile of the
subgrade tests can be used.

2. Determine the total pavement thickness required based on

the type of facility, subgrade CBR, gross aircraft
weight and number of passes.

3. Determine the design CBR of the subbase.

4. Determine the thickness of surface and base required by

entering the design curves with the subbase CBR.

5. Determine the minimum thickness of the surface from the

appropriate table. Minimum base thicknesses are given
for base materials with CBR values of 80 and 100
percent. There is no incentive in the design procedure
for using a surface thickness greater than the minimum,
so practical design would use the minimum surface
thickness. The thickness of the base is the maximum of
either the minimum required base thickness or the
difference between the total pavement thickness and the
minimum surface thickness.

6. The thickness of the subbase is equal to the difference

between the total required thickness and the combined
thickness of the surface and the base. If this produces
a subbase thickness less than 6 inches, consideration
should be given to increasing the thickness of the base
and eliminating the subbase.

Design of pavements with stabilized base and subbases

requires the design of a conventional pavement and then a
reduction of the required thicknesses based on equivalency
factors. There are a variety of rules that address the
appropriate application of stabilization materials.

5.4.2 Army-Air Force Rigid Pavement Design

The Army-Air Force rigid pavement design manual. presents
procedures for the design of plain, jointed-reinforced,
steel-fibrous, continuously reinforced and prestressed concrete
pavements1 '

10]
5.4.2.1 Plain Concrete Airfield Pavements - The edge stresses are
reduced by 25% to account for load transfer afforded by the joint
designs required. The flexural modulus of elasticity of the
concrete is assumed to be 4,000,000 psi and Poisson's ratio is
assumed to be 0.15. The design curves are available for light,
medium, heavy and modified-heavy load pavements. Select pavement
thickness is a function of the flexural strength of the concrete,
modulus of subgrade reaction, and type of traffic area. The
design curve for short-field pavements uses the gross weight of
the aircraft and number of aircraft passes instead of the type of
traffic area. The design thickness is rounded to the nearest half
inch.

When the base or subgrade is stabilized, or the base is lean

or existing concrete, the pavement slab is designed as an overlay
using the equation:
(E,/EC) 1/3(hb))4)1/14
h = ( (hd) 1.4 -

where:

ho = required thickness of the plain concrete slab on a

stabilized subgrade

hd = thickness of plain slab that would be required if the

slab was placed directly on the subgrade

EI = modulus of elasticity of the base

E, = modulus of elasticity of the concrete

hb = thickness of stabilized layer or lean concrete base

5.4.2.2 Reinforced Concrete Pavements - Thickness design for both

continuously and jointed reinforced concrete pavements is the
same. Different procedures are used for designing the amount of
steel. Design of a reinforced concrete pavement requires first
selecting the thickness for a plain concrete pavement. A
n'mograph is then used for the selection of the reduced thickness
of the reinforced pavement based on the thickness of slab required
for a plain concrete pavement, the area of steel reinforcement,
the percent of steel reinforcement and the length of the slab.

102
There is an interdependence between the amount of steel required,
the size of the slab and the thickness, so designing the pavement
requires assuming either a percent steel or reduced slab thickness
and solving for the other value.

5.4.2.3 Fibrous Concrete Pavements - The design of Fibrous

concrete pavements is based on limiting the ratio of the flexural
strength and maximum tensile stress at the joint, with the load
either parallel or normal to the pavement edge. The limiting
criteria for the stress ratio is based on field experiments.
These experiments were performed with steel fibers which limit the
application ol- the design method. In addition, the design
procedure limits the vertical deflection to prevent potential
pumping, densitication and/or shear failures of the subgrade.

Design curves are presented for each of the classes and types
of Army and Air Force airfields. The Army design curves consider:

flexural strength of the Fibrous concrete;

modulus of subgrade reaction;

aircraft gross weight;

n)Amber of passes; and

type of traffic area.

The Air Forc(- lesign curves consider:

flexuial strength of the Fibrous concrete;

modulus of subqrade reaction; and

type of traffic area.

In the Air Forcie procedure, the aircraft weight and number of

passes are defined by the traffic area, as shown in Table 5-1.

103
5.4.2.4 Prestressed Pavement Design - The design of prestressed
concrete pavements requires balancing the level of prestressing
with the thickness of the slab to obtain an economical design.
The design equation is:

ds= (6PNB/(w(hp)P)) - R + r. +

where:

ds = design prestress required in the concrete

P = aircraft gear load

N = load repetition factor

B = load moment factor

w = ratio of multiple wheel gear load to single wheel gear

load

hp= design thickness of prestressed concrete

R = design flexural strength of concrete

rs = foundation restraint stress

ts = temperature warping stress

The design manual presents curves and equations for

quantifying each of the design factors except for the design
prestress and the thickness of the slab. The manual suggests the
design prestress should be in the range of 100 to 400 psi and the
minimum thickness of the slab is 6 inches.

5.4.3 Navy Rigid Pavement DesiQn

The Navy design manual for rigid pavements 11 is similar to
the Portland Cement Association procedures. Westergaard's theory
is used for computing the maximum stress in the pavement. The
thickness is selected to keep the computed stress less than the
working stress. The working stress is the flexural strength
divided by a safety factor of 1.4 and 1.2 for primary and
secondary traffic areas respectively. Design charts are presented
for single, dual and dual-tandem gear types. The manual endorses
the use of the PCA pavement design computer program when designing

104
for other gear types. The design curves are limited to modulus of
subgrade reactions of 100 and 500 pci and interpolation is used
for other k values.

5.4.4 Army-Air Force Flexible Pavement Design - Elastic Layer

Theory Method
The Corps of Engineers has developed a mechanistic design
procedure for flexible pavements97 . The analysis is performed
with either CHEV5L or BISAR. As described in Chapter 4, two
design criteria are used for the selection of the pavement
thickness, fatigue of the asphalt surface or stabilized base
layer, and subgrade strain criteria. The pavement designs are
performed for the critical aircraft at the airfield rather than
for a mix of aircraft.

The design procedure considers three design situations:

1. granular base and subbase;

2. stabilized base and granular subbase; and

3. stabilized base and subbase.

Several steps are required for the design process, as shown in

Figures 5-5, 5-6, and 5-7. The variables used in these figures
are defined as:

EBc modulus of the base course

CALL allowable strain, may be for either the subgrade strain

criteria or the fatigue criteria depending on the step
in the flow chart
Eh horizontal tensile strain at the bottom of the surface
layer or stabilized base or subbase

Ev vertical compressive strain at the top of the subgrade

n, expected number of strain repetitions of traffic for the

fatigue or subgrade strain analysis

Ni allowable number of strain repetitions estimated from

either the fatigue or subgrade strain criteria

105
 SITUMIROUSNRET AR OT IONETTOOLSFO
PARAMETERS. SRAFAIC %

LIHITIN
STRACOMPTERE
ETRI
lAECUREMDUU EIN DLETERMINLE ~
RUTODLUORS INGfRMESIALN
UETR EI.LRCE NOCTS HIKEST
MODULU
SURRAS BONMAT

-a ROUITEDCKCNEES
I

START5 NEUS
USDNoH EISO 05TP PPVMN
TIF DRIINED

E0 ONE AYeST
STRAINS CONST

FR0SABIIZE BASESIMM

cONIDRAIO
EG N AF
FO SEEC Rlil106NA
STARTWITH:
BITUMINOUS CONCRETEMODULUS
FOREACHMONTH BASE COURSE DETERMINED
MODULUS INFLEXURE
EIf
SUBGRADE MODULUS FOR ANDTHAW
NORMAL COURSE
PERIODSBASE MODULUS FROM
DETERMINED EQUIVALENTCRACKED EBc
SECTION
TRAFFIC PARAMETERS COURSE
SUBBASE MODULUS INFLEXURE
DETERMINED ES
LIMITING STRAIN
CRITERIA SUBBASE
COURSE DETERMINED
MODULUS FROMEQUIVALENTCRACKED
SECTION
Ek
ESTIMATED INITIAL
THdICKNESS

INCEASE
THI
KNESS

I OPTs h sN

FIGUR 5-7 ELASTI LAYER THEORY DE I N M TH D A EM N S W T

STABILIZED VUBBAUE

107KES T
 As demonstrated with the design procedure flow charts, many
steps are required for designing a flexible pavement with this
procedure. However, the procedure can be summarized in five
steps.

1. Determine material properties.

2. Determine a trial pavement section.

3. Computation of the critical strains.

4. Determine expected number of applied strain repetitionc.

5. Computation of damage factors and the cumulative damage.

Material characterization requires measurement of the modulus

of elasticity for each layer in the pavement. In general, dynamic
testing of the materials is required with tae xcption of
granular materials where the manual states "an empirical based
procedure was judged a better approach for obtaining usable
material parameters." The subgrade modulus is determined with
respect to the anticipated deviator stress on the subgrade. The
modulus of the asphalt concrete must be determined for a variety
of temperatures. In the design process, two temperatures are used
for determining the asphalt concrete modulus for each month. For
the fatigue analysis, a modulus is selected corresponding to the
average daily maximum temperature. For the subgrade strain
analysis, the average of the average daily mean temperature and
the average daily maximum temperature is used to select the
asphalt concrete modulus. These air temperatures are corrected to
a design temperature of the pavement. If the temperature
variations between months are relatively small, months can be
grouped to reduce the calculation requirements.

To determine trial pavement design for analysis the procedure

in TM 5-825-2/AFM 88-6, Ch. 2115 should be used. The mechanistic
analysis procedure is then used to check the design of a thicker
and thinner pavement structure to determine the optimum pavement
design.

108
 Strains in the pavement structure due to the design aircraft
wheel loading are computed with an elastic layer theory program.
These strains are input to the criteria equations for fatigue and
subgrade strain to determine the allowable number of strain
repetitions.

The expe'ted number of passes on the pavement for the design

aircraft is then reduced to the number of strain repetitions based
on the configuration of the landing gear, tire imprint area and
wander of the aircraft, and the thickness of the pavement. Design
curves are presented for determining the percent of passes of an
aircraft that produce strain repetitions based on the aircraft
type and the thickness of the pavement. For aircraft with tandem
tires, the number of strain repetitions can actually exceed the
number of operations.

The final step in the design process is to compute the

cumulative damage. since only critical aircraft are considered in
the design, the computation is required to account for the
different strains resulting from changes in temperature.

As a final note, it should be emphasized that the design of

conventional flexible pavements only considers the subgrade straj-
criteria. Fatigue criteria are not considered in the design. The
stated reason for this limitation is:

Conventional pavements consist of relatively thick

aggregate layers with a thin (3 to 5 inch) wearing
course of bituminous concrete. In this type of
pavement, the bituminous concrete structure is a minor
structural element of the pavement and thus the
temperature effects on the stiffness properties of the
bituminous concrete may be neglected. Also, it must be
assumed that if the minimum thickness of bituminous
concrete is used as specified in TM 5-825-2/AFM 88-6,
Ch. 2, then the fatigue cracking will not be considered.
Thus, for a conventional pavement, the design problem is
one of determining the thickness of pavement required to
protect the subgrade.

109
 6. DISCUSSION AND RECOMMENDATIONS

From the preceding chapters it is clear that pavements

behave in an extremely complex manner. Material behaviors
can be difficult to predict with respect to their response to
load and environmental conditions. In addition, material
characteristics change with time, environmental conditions
and stress-strain history, further complicating the task of
capturing the mechanistic response of the pavement. The
process is further complicated when attempting to relate
pavement response to pavement performance. Finally, there is
no consensus definition of airport pavement failure.

Design procedures based on empirical methods or

relationships, such as the CBR method, prescribe a pavement
thickness for protecting the subgrade (and therefore, the
pavement structure) from excessive deformation. However,
there is no statement in these design procedures for pavement
failure. Mechanistic-empirical methods that are based on
fatigue cracking do not generally specify a level of cracking
associated with pavement failure. Fatigue cracking of
pavements is a relatively common occurrence. Pavement
cracking alone, however, is not necessarily an indication of
pavement failure. Similarly, other pavement distresses do
not necessarily indicate that a pavement has "failed."

There is a general consensus that highway pavement

failure is related to the quality of the service provided to
its users. Under the concepts developed by Carey and Iric at
the AASHO Road Test, pavement failure is defined with respect
to the serviceability level of the pavement; a concept
largely related to the roughness of the traveling surface.
No comparable definition exists, however, for airport
pavements.

111
 Generally, maintenance can be divided into two
categories: preventive and responsive. Since preventive
maintenance is performed to reduce the occurrence of
distress, it is not an indicator of pavement failure.
Responsive maintenance, on the other hand, corrects distress
conditions. Taken to an extreme, responsive maintenance can
become cost prohibitive and rehabilitation or reconstruction
may be required. If airport pavement maintenance is used as
a failure criteria and if maintenance of the pavement is
performed to correct the occurrence of distresses, then the
prediction of pavement life is directly related to the
prediction of pavement distress.

Conceptually, the use of pavement distress as the

limiting criteria for determining the life of the pavement
requires defining type, extent, and severity of distresses
requiring maintenance. Prediction of type, severity, and
extent of distress in a mechanistic manner requires
predicting pavement response to load for a set of
environmental conditions, followed by a relationship between
pavement response and distress. The prediction of pavement
response requires constitutive relationships defining the
deformation response of the pavement to loading conditions
for a set of environmental conditions. Use of the
constitutive models requires quantification of the material
behavior with respect to the material constants required by
the models. In addition, the model must be capable of
handling mixed traffic distributions both with respect to the
type and frequency of aircraft.

6.1 DISCUSSION OF THE STATE OF THE ART IN AIRPORT PAVEMENT

ANALYSIS
The state of the art in practical airport pavement
analysis is limited to elastic theories of pavement response.
Some of these models can account for the stress-sensitivity
pavement materials. Furthermore, the prediction of distress
from pavement response is accomplished with empirical

112
relationships developed primarily from laboratory and highway
experience with limited calibration from airport pavement
performance data. Thus, the state of the art of airport
pavement design methodology falls far short of a true
mechanistic analysis of pavement behavior and performance.

There are many models for pavement response that have

been proposed by researchers and by advanced pavement
engineers. Notably, the application of finite element models
in conjunction with the fracture mechanics of crack growth
and viscoelastic-plastic analysis for permanent deformation
hold promise for improving the state of the art in the
prediction of pavement performance. Conceptually, the finite
element technique can be used for the solution of a broad
class of material behavior, traffic loads and environmental
conditions. As the flexibility of the finite element
analysis process is increased, the need for computer storage
and speed increases by a disproportionate amount. In
addition, laboratory determination of the required material
characteristics becomes increasingly complex as the number of
constants required by the models increases. Furthermore,
models have yet to be developed to define completely the
response of pavement materials to all types of load
conditions, material types and pavement geometries.

The types and capabilities of pavement models range

widely from linear elastic response to viscoelastic models
that include an element of fracture mechanics for the
prediction of rracking. Many of the advanced models were
developed for other engineering fields and their application
to pavement design has been relatively limited. There is no
uniform pavement analysis model that can be used for
analyzing a pavement structure for all possible environmental
and traffic load conditions. All pavement response models
require a series of simplifying assumptions that consequently
limit their universal applicability. Simplifying assumptions
cover material behavior, traffic loadings, environmental

113
conditions, pavement geometry and the interaction of these
factors.

Existing mechanistic models that are applied to the

design of pavement structures are limited with respect to
several considerations regarding the design and analysis of
these structures. The models used for pavement analysis
generally assume linear elastic behavior, whereas pavement
structures are composed of materials that also display a
viscous and plastic behavior. The models generally assume a
uniform distribution of contact stress between the tire and
pavement and do not account for the side wall stiffness of
the tires.

Generally, the pavement structure is modeled as a

uniform structure with homogeneous characteristics in all but
the vertical direction, where distinctions in the properties
of the layers are modeled. In actuality, material properties
of pavement structures are not uniform nor homogeneous.
Construction variability affects material characteristics in
a random manner. Cracking and joints provide geometric
discontinuities in the pavement structure.

One of the major areas of concern in the analysis of

airport pavements is the treatment of mixed traffic effects.
Current analytical procedures for the effect of mixed traffic
is dependent on the application of Minor's hypothesis based
on an accumulation of the incremental damage caused by each
type of traffic loading. While this hypothesis has been
verified on a statistical basis for laboratory conditions,
there has not been a field experiment for either highway or
airport pavements that verifies -he applicability of Minor's
hypothesis for pavement damage.

The above discussion has focused primarily on the design

of new pavements. In the aging U.S. airport system, however,
the greatest demand for pavement engineering is in the area

114
of preservation and restoration of existing pavements. In
many ways, pavement preservation and restoration is more
difficult than the design of new pavements. For the design
of pavement preservation and restoration, the structural
capacity and condition of the existing pavements must be
evaluated. Deflection testing is the current practice fet
evaluating existing pavements. However, analysis of
deflection data suffers from the same variances between
pavement behavior and mechanistic models as is encountered
for the design of new pavements (e.g., the discrepancies
between the true behavior of the materials and the
assumptions used in the analytical models, and the effects of
environmental conditions on the behavior of the pavement).

Pavements fail in many ways as noted by the various

types of distresses that are identified in pavement condition
evaluation procedures. Some of these procedures define up to
nineteen distress types for both asphalt and concrete
pavements. Current mechanistic models are limited in the
number of distress types that are simulated. Fatigue
cracking is the most common distress type modeled with
current mechanistic models. Other models estimate the
development of rutting. Some model the development of
reflection cracking. Arguably, these are the predominate
distress types. However, the discrepancy between the number
of distresses that exist on the pavements versus the
distresses that are modeled emphasizes the shortcomings of
the existing analysis procedures.

6 2 RECOMMFMI",XV'IONS FOR Ii'HOVN. i{E STATP OF THE ART IN

AIRPORT PAVEMENT DESIGN AND ANALYSIS
The state of the art in pavement analysis has r-aI
cansiderable advance, particularly since the advent of
powerful mainframe and microcomputers. However, the state of
the practice in airport pavement design and analysis has not
kept abreast of other advances in the air transportation
industry. There are no constitutive models that can analyze

115
all pavement materials under all types of traffic loadings
and environmental conditions. Even the most precise models
fail to recognize the full range of behavior with respect to
pavement conditions, material properties and environmental
conditions.

Pavement response to load, in terms of stresses, strains

and deformaticn is estimated with models that contain a great
many assumptions regarding material behavior, environmental
conditions, and load characteristics. Hence, pavement
response estimatEs are approximations at best. Estimating
the pavement response, however, is only a part of the
requirement for a mechanistic analysis. Translation of
pavement response into distress is an equally challenging
task. The phenomenological approach that is widely used for
the analysis of fatigue failure often does not capture the
differences between specific materials at various airports.
As demonstrated, different phenomenological equations can
produce vastly different estimates of pavement life even when
the same level of strain is used in the equation. In the
past, pavement engineers have been able to use imprecise
models due to the noncatastrophic nature of pavement
failures. Generally, pavement failure does not result in
loss of life or aircraft accidents. But, premature pavement
failure generates tremendous economic and operational
hardships for both the individual airport and the air
transport industry in general.

Accurate estimates of pavement response and subsequently

of pavement distress and performance will require a new
approach to the design and analysis of airport pavements.
Improving the sophistication of pavement analysis will likely
require the use of three-dimensional finite element models
that are formulated with specific characteristics of pavement
materials with respect to elastic, viscous and plastic
behavior. In addition, the constitutive equations in the
three-dimensional analysis should capture the effect of

116
environmental conditions of the stress state of the materials
and their properties. While three-dimensional finite element
analysis is firmly entrenched in other areas of engineering,
due to its complexity this technology has not been extended
to the analysis of pavements. The intricacy of pavement
materials, loading conditions and environmental effects makes
pavements among the most complex structures designed and
analyzed by engineers.

Development of an improved model of pavement behavior

should follow the traditional approach to the development of
new engineering methods. This would include the formal
development of the problem statement, encompassing a
definition of pavement performance and failure. At this
point, analytical methods should be developed for modeling
the pavement response to both environmental and traffic
loading conditions. Laboratory testing would be necessary
for defining the required material characteristics and
failure or distress criteria. Finally, field verification
and calibration of the models would be required.

To develop an improved model for pavement performance,

attention must focus on the most critical elements of airport
pavements. As earlier mentioned, the majority of U.S.
airports are already in place and therefore, with regard to
pavements, the area needing attention is preservation and
restoration. Research in this area would include field
evaluation of existing and future structural capacity of
airport pavements. Currently, deflection testing is used for
quantifying pavement structural capacity. While this form of
testing is arguably superior to laboratory testing of the
materials, it is limited by the inability of existing
theories to describe the results of the deflection testing
and the extrapolation of those results to the future
performance of the pavements.

117
 In addition, the existing shortage of quality virgin
materials will lead to a greater emphasis being placed on the
use of new materials, recycling and the use of marginal
materials. The existing models and laboratory procedures are
not adequate to determine the future response of new
materials. This is a critical area demanding the development
of the unified pavement theory. While field verification of
material performance is always desirable, it can also impede
the development and introduction of new materials. Normally,
a pavement should perform for 20 years. Thus, it would take
a minimum of 20 years to prove the value of a new material.
A unified theory for the analysis of airport pavements,
supported by appropriate laboratory testing, can provide a
vehicle for the proof and acceptance of new materials in a
timely manner.

The Federal Aviation Administration has essentially

defined the steps required for the development of a unified
theory for the design and analysis of airport pavement
structures. Due to the importance of pavements to the air
transportation industry and the importance of this industry
to the economic development of the United States, an
investment in pavement research that can lead to improved
pavement performance is an investment that will deliver a
large benefit-to-cost ratio.

118
 LIST OF SYMBOLS

Symbol Definition(s) as used in Report

a,b,n,A,k constants
C1, C 2 relative stiffness coefficients
C/S design prestress in concrete
DMR design modulus of rupture
Dj creep compliance function
E Young's modulus
ER resilient modulus
E* complex modulus
f average subgrade coefficient
f' eultimate strength of concrete
G(t) relaxation function for a constant strain
test
Gi coefficient of the Dirichlet series
hi thickness of the ith layer
J(t) creep function under constant stress
J,(t) volumetric creep function
k stress intensity factor
L length
Mi mass of i
MR modulus of rupture
MR90 90 day average modulus of ruputurer
N number of applications to failure
NE number of equivalent departures of design
aircraft
P vertical point load or aircraft gear load
v (as subscript) radial direction
rs foundation restraint stress
R design flexural strength of concrete
Ri equivalent number of design aircraft
annual departures
R2 annual departures expressed in design
aircraft landing gear
S stiffness of cohesionless soils

119
Smix(tT) asphalt mixture stiffness at time, t, and
temperature, T
Ss vertical compressive strain at top of
subgrade
SGcBR subgrade California Bearing Ratio (CBR)
t time, (as subscript) time, tangential or
transverse
TpTB, Ts,TsB thickness of, respectively, pavement,
base, surface, and subbase. Total
pavement thickness Tp = T s + T B + TSB
Ai ith component of the displacement vector
in Cartesian coordinates
w ratio of multiple wheel gear load to
single wheel gear load
wi wheel load of the ith whee
W, W D weight, weight of design aircraft
a coefficient of thermal expansion
6i exponent of the Dirichlet series
6 strain
Ce longitudinal strain
6o initial and constant strains
E(t) strain as a function of time
£ Lamds constant
IPoisson's ratio, Lames constant
phase difference between stress and
strain
mass density
a stress
cc "frictional" stress
o hydrostatic stress
a3 transverse principal stress
T shear stress

(angular frequency

120
 REFERENCES

1. Shahin, M.Y., and Kohn, S.D.. (19 i) "Pavement Management

System for Roads and Parking Lots," Technical Report M-294,
U.S. Army Corps of Enginecrs Construction Engineering
Research Laboratory, Champaign, IL.

2. Yoder, E.J., and Witczak, M.W. (1975) Principles of

Pavement Design, John Wiley & Sons, New York

3. Sargious, M., (1975) Pavements and Surfacings for Highway

and Airports, John Wiley and Sons, New York

4. Chou, Y.I. 1977, "Engineering Behavior of Pavement

Materials; State of the Art," Federal Aviation
Administration Report No. FAA-RD-77-37, Washington, DC

5. Westmann, R.A., (1971) "Fundamentals of Material

Characterization," Structural Design of Asphalt Concrete
Pave'-ent Systems, Highway Research Board Special Report
126, Washington, DC

6. Sentler, L., (1987) "A Strength Theory for Viscoelastic

Materials," Document D9:1987, Swedish Council for Building
Research, Stockholm

7. Derucher, K.N., and Korfiatis, G.P. (1988) Materials for

Civil and Highway Engineers, Prentice Hall, New Jersey

8. Cordon, W.A. (1979) Properties, Evaluation and Control of

Engineering Materials, McGraw-Hill, Inc., New York

9. Mazaras, J., and Pijandier-Cabot, G. (1989) Continuum

Damae-Theory - Application to Concrete, Journal of
Engineering Mechanics, Vol. 115(2) ASCE

10. Monismith, C., Epps, J.A., and Finn, F.N. (1985) "Improved
Asphalt Concrete Mix Design," American Association of
Asphalt Paving Technologist, Proceedings, Volume 54

11. Deacon, J.A. (1971) "Material Characterization -

Experimental Behavior," Structural Design of Asphalt
Concrete Pavement Systems, Highway Research Board Special
Report 126, Washington, DC

12. Ma,,ilouk, M., and Sarofim, R.T. (1988) "Modulus of Asphalt

Mixtures - An Unresolved Dilemma," Transportation Research
Board, Research Record 1171, p. 193, Washington, DC

13. Cedergren, H.R. (1988) "Why All Important Pavements Should

Be Well Drained," Transportation Research Board, Research
Record 1188, p. 56, Washington, DC

R-1
14. Hicks, R.G., (19 Factot nfluencing the Resilient
Properties of Grai.ular Mat <iais, Dissertation Series,
University of California, Berkeley

15. Kohn, S.D., and Bentsen, R.A. (1987) "Performance of

Flexible Airfield Pavements Subjected to High Traffic
Volumes," p. 638 Proceedinqs, The Sixth International
Conference Structural Design of Asphalt Pavements, Ann
Arbor, Michigan

16. Saraf, C., Marshek, K., Chen, H., and Hudson, W.R., (1987)
"The Effect of Truck Tire Contact Pressure Distribution of
the Design of Flexible Pavenents," Proceedings, rhe Sixth
International Conference Structural Design of Asphalt
Pavements, Ann Arbor, Michigan

17. Bonaquist, R., Surdahl, R., and Mogawer, W. (1989) "Effect

of Tire Pressure on Flexible Pavement Response and
Performance," Transportation Research Board, Research
Record 1227, p. 97, Washington, DC

18. Huhatla, M., Pihlajamaki, J., and Pieninaki, M. (1989),

"Effect of Tires and Tire Pressures on Road Pavements,"
Transportation Research Board, Research Record 1227,
p. 107, Washington, DC

19. Sebaaly, P., and Tabatabaee, N. (1989), "Effect of Tire

Pressure and Type on the Response of Flexible Pavements,"
Transportation Research Board, Research Record 1227,
p. 115, Washington, DC

20. Wignot, J.E., Durup, P.C., Wittlin, G., Scott, R.B., and
Gamom, M.A. (1970) "Aircraft Dynamic Wheel Load Effects on
Airport Pavements," Federal Aviation Administration, Report
No. FAA-RD-70-19, Washington, DC

21. Highter, W.H., and Harr, M.E. (1973) "Application of Energy

Concepts to the Performance of Airfield Pavements," Air
Force Weapons Laboratory, Technical Report Number AFWL-TR-
72-225, Kirtland Air Force Base, New Mexico

22. HoSang, V.A. (1975) "Field Survey and Analysis of Aircraft

Distribution on Airport Pavements," Federal Aviation
Administration, Report FAA-RD-74-36, Washington, DC

23. HoSang, V.A. (1978) "Field Survey and Analysis of Aircraft

Distribution on Airport Pavements," Transportation Research
Board, Special Report 175, Washington, DC

24. Ledbetter, R.H. (1978) "Effects of Dynamic Loads on Airpor:

Pavements," Transportation Research Board, Special Rleport
175, Washington, DC

R-2
25. Himeno, K., Watanabe, T., and Maruyama, T. (1987)
"Estimation of Fatigue Life of Asphalt Pavement," p. 272,
Proceedings, The Sixth International Conference Structural
Design of Asphalt Pavements, Ann Arbor, Michigan

26. Haas, R.C.G., and Hudson, W.R. (1978) Pavement Management

Systems, McGraw-Hill, Inc., New York

27. Hudson, W.R., and McCullough, B.F., (1973) "Flexible

Pavement Design and Management: Systems Formulation,"
National Cooperative Highway Research Program Report 139

28. Gomez-Achecar, M. and Thompson, M.R. (1986) "ILLI-PAVE-

Based Response Algorithms for Full-Depth Asphalt Concrete
Flexible Pavements," p. 11, Transportation Research Board,
Transportation Research Record 1095, Washington, DC

29. Sentler, L., (1987) "A Strength Theory for Viscoelastic

Materials," Report D9:1987, Swedish Council for Building
Research, Stockholm, Sweden

30. Nair, K. (1971), "Solutions and Solution Techniques for

Boundary Value Problems," p. 103, Structural Design of
Asphalt Concrete Pavement Systems, Highway Research Board,
Special Report 126, Washington, DC

31. Burmister, D.M. (1945) "The General Theory of Stresses and

Displacements in Layered Systems," Journal of Applied
Physics, Vol. 16(2) p. 69, Vol. 16(3) p. 126, Vol. 16(5),
p. 296

32. Westergaard, H.M. (1927), "Theory of Concrete Pavement

Design," Proceedings, Highway Research Board

33. Avramesco, A., (1967) "Dynamic Phenomena in Pavements

Considered as Elastic Layered Structures," Proceedings, The
Second International Conference on the Structural Design of
Asphalt Pavements, Ann Arbor, Michigan

34. Lister, N.W., and Jones, R. (1967), "The Behavior of

Flexible Pavements Under Moving Wheel Loads," Proceedings,
The Second International Conference on the Structural
Design of Asphalt Pavements, Ann Arbor, Michigan

35. Saraf, C., Marshek, K., Chen, H., and Hudson, W.R., (1987)
"The Effect of Truck Tire Contact Pressure Distribution of
the Design of Flexible Pavements," Proceedings, The Sixth
International Conference Structural Design of Asphalt
Pavements, Ann Arbor, Michigan

36. Ahlvin, R.G., Chou, Y.T., and Hutchinson, R.L. (1973), "The
Principle of Superposition in Pavement Analysis," Highway
Research Board, Highway Research Record 466, p. 153,
Washington, DC

R-3
37. McCullough, B.F. (1969) A Pavement Overlay Design System
Considering Wheel Loads, Temperature Changes and
Performance, Doctoral Dissertation, University of
California, Berkeley

38. Brown, S.F., and Pell, P.S. (1967), "An Experimental

Investigation of the Stresses, Strains and Deflections in a
Layered Pavement Subjected to Dynamic Loads," Proceedings,
The Second International Conference on the Structural
Design of Asphalt Pavements, Ann Arbor, Michigan

39. Chevron Oil Company, (1963) development of the elastic

layer theory program

40. De Jong, D.L., Peutz, M.G.F., and Korswagen, (1973)

"Computer Program BISAR," External Report Koninklyke/Shell
Laboratorium, Amsterdam, Holland

41. Ahlborn, G., (1972) "Elastic Layer System with Normal

Loads," Institute for Transportation and Traffic
Engineering, University of California, Berkeley

42. McTrans, The Center for Microcomputers in Transportation,

The University of Florida, 512 Weil Hall, Gainsville, FL,
32611, (904) 392-0378

43. Shahin, M.Y., Kirchner, K., and Blackmon, E. (1987),

"Analysis of Asphalt Concrete Layer Slippage and Its Effect
on Pavement Performance," p. 958, Proceedings, Sixth
International Conference on the Structural Design of
Asphalt Concrete Pavements, Ann Arbor, Michigan

44. Parker, F., Barker, W.R., Gumkel, R.C., and Odom E.C.
(1979) "Development of a Structural Design Procedure for
Rigid Airport Pavements," Federal Aviation Administration
Report No. FAA-RD-77-81, Washington, DC

45. Shahin, M.Y., (1971), "A Computer Program for an Iterative

Method of Pavement Structural Analysis," Unpublished
technical memorandum, Center for Highway Research,
University of Texas, Austin

46. Zaniewski, J.P., (1978) Design Procedure for Asphalt

Concrete Overlays of Flexible Pavements, Doctoral
Dissertation, University of Texas, Austin

47. Aston, J.E., and Moavenzadeh, F. (1967), "Analysis of

Stresses and Displacements in a Three-Layered Viscoelastic
System," p. 149, Proceedings, Second International
Conference on the Structural Design of Asphalt Concrete
Pavements, Ann Arbor, Michigan

48. Lee, E.H. (1961), "Stress Analysis in Viscoelastic

Materials," Journal of Applied Physics, Vol. 13(2) p. 183

R-4
49. Pister, K.S. (1961), "Viscoelastic Plates on Viscoelastie
Foundations," Journal of the Engineering Mechanics
Division, ASCE, p. 43

50. Pister, K.S., and Westman, R.A. (1962), "Analysis of

Viscoelastic Pavements," p. 20, Proceedings, International
Conference on the Structural Design of Asphalt Pavements,
Ann Arbor, Michigan

51. Ishihara, K., (1962), "The General Theory of Stresses and

Displacements in Two-Layer Viscoelastic Systems," The
Japanese Society of Soil Mechanics and Foundations
Engineering, Soils and Foundations, Vol. 2(2), Tokyo, Japan

52. Lee, E.H. and Rogers, T.G. (1961) "Solution of Viscoelastic

Stress Analysis Problems Using Measured Creep or Relaxation
Functions," Division of Applied Mathematics, Brown
University

53. Hopkins, I.L. and Hamming, R.W. (1957), "On Creep and
Relaxation," p. 906, Journal of Applied Physics, Vol. 28

54. Moavenzadeh, F., and Elliot, J.F. (1972), "A Stochastic

Approach to Analysis and Design of Highway Pavements,"
Proceedings, Third International Conference on the
Structural Design of Asphalt Pavements, London, England

55. Rauhut, J.B., et al. (1975), Sensitivity Analysis of VESYS,

Austin Research Engineers, Final Report to the Federal
Highway Administration

56. Hufferd, W.L. and Lai, J.S. (1978) "Analysis of N-Layered

Viscoelastic Pavement Systems," Federal Highway
Administration Report No. FHWA-RD-78-22, Washington, DC

57. Khosla, N.P. (1987), "A Field Verification of VESYS IIIA

Structural Subsystems," p. 486, Proceedings, Sixth
International Conference Structural Design of Asphalt
Pavements, Ann Arbor, Michigan

58. Beckedahl, H., Gerlach, A., LOcke, H., and Schwaderer, W.

(1987), "On Improvements of the Existing VESYS Concepts,"
p. 677, Proceedings, Sixth International Conference
Structural Design of Asphalt Pavements, Ann Arbor, Michigan

59. Lattes, R., Lions, J.L., and Bonitzer, J. (1962), "Use of

Galerkin's Method for the Study of Static and Dynamic
Behavior of Road Structures," p. 238, Proceedings,
International Conference on the Structural Design of
Asphalt Pavements, Ann Arbor, Michigan

R-5
60. Bastiani, A. (1962), "The Explicit Solution Cf the
Eq]atioiis of the Elastic Deformdtion= for d Stratified Road
Under Given Stresses in the Dynamic Case," p. 351,
Proceedings, International Conference on the Structural
Design of Asphalt Pavements, Ann Arbor, Michigan

61. Ishihara, K., and Kimura, T. (1967), "The Theory of

Viscoelastic Two-Layer Systems and Conception of Its
Application to the Pavement Design," p. 189, Second
International Conference on the Structural Design of
Asphalt Concrete Pavements, Ann Arbor, Michigan

62. Perloff, W.H., and Moavenzadeh, F. (1967), "Deflection of

Viscoelastic Medium Due to a Moving Load," p. 212, Second
International Conference on the Structural Design of
Asphalt Concrete Pavements, Ann Arbor, Michigan

63. Mamlouk, M.S. (1987), "Dynamic Analysis of Multilayered

Pavement Structures - Theory, Significance and
Verification," p. 466, Proceedings, Sixth International
Conference on the Structural Design of Asphalt Concrete
Pavements, Ann Arbor, Michigan

64. Weiss, R.A. (1977), "Subgrade Elastic Moduli Determined

from Vibratory Testing of Pavements," U.S. Army Engineer,
Waterways Experiment Station, Vicksburg, Miss.

65. Weiss, R.A. (1979), "Pavement Evaluation and Overlay

Design: Method that Combines Layered-Elastic Theory and
Vibratory Nondestructive Testing," Transportation Research
Board, Research Record 700, p. 20, Washington, DC

66. Szendrei, M.E. and Freeme, C.R. (1967), "The Computation of

Road Deflections Under Impulsive Loads from the Results of
Vibration Measurements," p. 89, International Coiiferencc )n
the Structural Design of Asphalt Concrete Pavements, Ann
Arbor, Michigan
67. Eringen, A.C. and Suhubi, E.S. (1975), Elastodynamics,
Vol. 2, Linear Theory, Academic Press, New York

68. Kausel, E. and Peek, R. (1982) "Dynamic Loads in the

Interior of a Layered Stratum: An Explicit Solution,"
p. 1459, Bulletin of the Seismological Society of America,
Vol. 72(5)

69. Sebaaly, B.E. (1987), Dynamic Models for Pavement

Analysis," Doctoral Dissertation, Arizona State University,
Tempe

70. Westergaard, H.M., (1927) "Theory of Concrete Pavement

Design," Proceedings, Highway Research Board

71. Westergaard, H.M., (1948) "New Formulae for Stresses in

Concrete Pavements of Airfields," Transactions, ASCE

R-6
72. Pickett, G., and Ray, G.K. (1951), "Influence Charts for
Concrete Pavements," Transactions, ASCE

73. Packard, R.G., "Computer Program for Airport Pavement

Design," Portland Cement Association, Skokie, IL.

74. Packard, R.G., (1973) "Design of Concrete Airport

Pavement," Portland Cement Association, Skokie, IL.

75. Stelzer, C.F., and Hudson, W.R. (1967), "A Direct Computer
Solution for Plates and Pavement Slabs," Research Report
569, Center for Highway Research, University of Texas,
Austin

76. Burnet, D. (1987), Finite Element Analysis from Concepts to

Applications, Addison-Wesley Publishing Co., Mass

77. Ioaniides, A.M. (1987) "Axisymmetric Slabs of Finite Extent

Solid," American Society of Civil Engineers, Journal of
Transportation Engineering, Vol. 113, pp. 277-290

78. Duncan, J.M., Monismith, C.L., and Wilson, E.L. (1968),

"Finite Element Analysis of Pavements," Highway Research
Board, Highway Research Record 228, Washington, DC

79. Pichumani, R. (1972), "Application of Computer Codes to the

Analysis of Flexible Pavement," Proceedings, Third
International Conference on the Structural Design of
Asphalt Pavements, London, England

80. Raad, L., and Figueroa, J.L. (1980), "Load Response of

Transportation Support Systems," p. 111, Transportation
Engineering Journal, ASCE, Vol. 106, No. TEl

81. Baladi, G. (1989), "Fatigue Life and Permanent Deformati-n

Characteristics of Asphalt Concrete Mixes," p. 75,
Transportation Research Board, Transportation Research
Record 1227, Washington, DC

82. Smith, R.B., and Yandell, W.O. (1987), "Predicted and Field
Performance of a Then Full-Depth Asphalt Pavement Placed
Over a Weak Subgrade," p. 443, Proceedings, Sixth
International Conference on the Structural Design of
Asphalt Concrete Pavements, Ann Arbor, Michigan

83. Lim, P.N., and Yandell, W.O. (1988) "Reevaluation of Shell

Life by Mechano-Lattice Analysis," ASCE Journal of
Transportation Engineering, Vol. 114, pp. 435-449

84. Wang, S.K., Sargious, M.A., and Cheung, Y.K., (1972)

"Advanced Analysis of Rigid Pavements," ASCE,
Transportation Engineering Journal

R-7
85. Chou, Y.T., (1981) "Structural Analysis Computer Programs
for Rigid Multicomponent Pavement Structures with
Discontinuities, WESLIQID and WESLAYER," U.S. Army
Engincering Waterways Experiment Station, Technical Report,
GL-81-6, Vicksburg, Mississippi

86. Chou, Y.T., (1983) "Subgrade Contact Pressures Under Rigid

Pavements," ASCE Journal of Transportation Engineering,
Vol. 109, pp. 363-379

87. Tabatabaie, A.M., and Barenberg, E.J. (1980), "Structural

Analysis of Concrete Pavement Systems," p. 493,
Transportation Engineering Journal, ASCE, Vol. 106, No. TE5

88. Tia, M., Arnaghani, J.M., Wu, C.L., Lei, S., and Toye, K.L.
(1987), "FEACONS III Computer Program for an Analysis of
Jointed Concrete Pavements," p. 12, Transportation Research
Board, Transportation Research Record 1136, Washington, DC

89. Triado-Crovetti, M.R., Darter M.I., Jayawickrama, P.W.,

Smith, R.E., and Lytton, R.L. (1988), "ODE Computer
Program: Mechanistic-Empirical Asphalt Concrete Overlay
Design," p. 30, Transportation Research Board, Research
Record 1207, Washington, DC

90. Majidzadeh, K., Abdulshafi, A., Ilves, G.J., and

McLaughlin, A. (1987), "A Mechanistic Approach for
Thermally Induced Reflection Cracking of Portland Cement
Concrete Pavement with Reinforced Asphalt Overlay," p. 83,
Transportation Research Board, Transportation Research
Record 1117, Washington, DC

91. McCullough, B.F., Abou-Ayash, A., Hudson, W.R., and

Randall, J.P. (1975', "Design of Continuously Reinforced
Concrete Pavements for Highways," Final Report, NCHRP
Project 1-15

92. Bradburu, R.D., (1938) "Reinforced Concrete Pavement," Wire

Reinforced Institute

93. Ioannides, A.M., and Salsilli-Murua, R.A. (1989),

"Temperature Curling in Rigid Pavements: An Application of
Dimensional Analysis," p. 1, Transportation Research Board,
Transportation Research Record 1227, Washington, DC

94. Navarro, D., and Kennedy, T.W. (1975), "Fatigue and

Repeated Load Elastic Characteristics of Inservice Asphalt-
Treated Materials," Center for Highway Research, University
of Texas, Austin

95. Kelly, H.F., and Thompson, M.R. (1988) "Mechanistic Design

Concepts for Heavyweight F15 Aircraft on Flexible
Pavements," ASCE, Journal of Transportation Engineering,
Vol. 114, pp. 323-340

R -8
96. "The AA:YK Read Test. Report 5 - Pavement Researt;<,'
Highway Research Board, Special Report 61E, 1962

97. "Flexible Pavement Design Method for Airfields (Elastic

Layer Method)," Joint Departments of the Army and Air
Force, U.S.A., Report No. rM 5-825-2-11, AFM 88-6, Ch. 2,
1989

98. "Comparatsve Performance of Structural Layers in Pavement

Systems, Vol. II, Analysis of Test Section Pata and
Presentation of Design and Construrtion Procedures,"
Federal Aviation Administration Report No. FAA-RD-73-198II,
1977

q9. Majidzadeh, K., and Kauffmann, E.M. (1971) "Analysis of

Fatigue and Fracture of Bituminous Pavement mixtures,"
Final Report, Phases I and II, Prnject R 2845, Ohio
Research Foundation

100. Majidzadeh, K., Kauffmann, E.M., and Chang. C.W. (1973)

"Verification of Fracture Mechanics Concepts to Predict
Cracking of Flexibie Pavement," FHWA Final Report No. FHWA-
RD-73-91

!01. Ma-idzadeh, K., Kauffmann, E.M., and Ramsamooj, D.C. (1971)

"Application of Fracture Mechanics in the Analysis of
Pavement Fatigue," Association of Asphalt Paving
Technologist, Vol. 40

102. Ramsamooj, D.V. (1970) The Design and Analysis of Flexible

Pavements, PHD Dissertation, Ohio State University

103. Jayawickrama, P.W., and Lytton, R.L. (1987), "Method for

Predicting Asphalt Concrete Overlay Life Against Reflectior
Cracking," p. 913, Proceedings, Sixth International
Conference Structural Design of Asphalt Concrete Pavements,
Ann Arbor, Michigan

104. Schapery, R.A. (1975), "A Theory of Crack Initiation and

Growth in *iscoelastic Media, II Approximate Methods of
Analysis," p. 369, International Journal o Fracture,
Netherlands

105. George, K.P. (1970) "Theory of Brittle Fracture Applied ta

Soil Cement," p. 991, Journal of Soil Mechdnics and
Foundation Engineering, ASCE Vol. 96(3)

L06. Crockford, W.W., and Little, D.N. (1987), "Tensile Fracture

and Fatigue of Cement-Stabilized Soil," p. 520, Journal of
Transportation Engineering, ASCE, Vol. 11,,K)
107. Majidzadeh, K., Abdulshafi, A., Ilves, G.J., and
McLaughlin, A. (1987), "A Mechanistic Approach for
Thermally Induced Reflection Cracking of Portland Cement
Concrete Pavement with Reinforced Asphalt Overlay," p. 83,
Transportation Research Board, Transportation Research
Record 1117, Washington, DC

108. Dormon, G.M. (1962) "The Extension to Practice of a

Fundamental Procedure for the Design of Flexible
Pavements," International Conference on the Structural
Design of Asphalt Pavements, University of Michigan, Ann
Arbor

109. Klomp, A.J.G., and Dormon, G.M. (1964) "Stress

Distributions and Dynamic Testing in Relation to Road
Design," Australian Road Research Board

110. Asphalt Institute, (1973) "Full-Depth Asphalt Pavements for

Air Carrier Airports," The Asphalt Institute, Manual Series
No. 11

111. Khosla, N.P. (1987), "A Field Verification of VESYS IIIA

Structural Subsystems," p. 486, Proceedings, Sixth
International Conference Structural Design of Asphalt
Pavements, Ann Arbor, Michigan

112. Sneddon, R.V. (1986) "Evaluation and Verification of the

VESYS IIIA Structural Design System for Two Test Sites in
Nebraska," p. 18, Transportation Research Board,
Transportation Research Record 1095, Washington, DC

113. Federal Aviation Administration, (1978) "Airport Pavement

Design and Evaluation," AC 150/5320-6C, with change 1.

114. General Provisions for Airfield Pavement Design, Department

of Defense, TM 5-824-1/AFM 88-1, Ch. 1 (1987)

115. Flexible Pavement Design for Airfields (1978), Triservice

manual, Navy DM 21.3, Army TM 5-825.2, Air Force AFM 88-6,
Ch. 2

116. Rigid Pavements for Airfields (1988) Army TM 5-825-3, Air

Force AFM 88-6, Ch. 3

117. Airfield Pavements, (1973) NAVFAC DM-21, with changes

1 and 2

R-10