You are on page 1of 49

Research Rep ort

UKTRP-87-29

PAVEMENT DESIGNS BASED ON WORK

by
Herbert F. Southgate , P.E.
Chief Research Engineer
and
Robert c. Deen , P . E.
D irector

Kentucky Transpor tation Research Program


College of Engineering
University of Kentucky
Lexington, Kentucky 40506-0043

in cooperation with
K entucky Transpor tation Cabinet
and
Federal Highway Administration
U S Department of Transpor tation
T he contents of this report r e f l e c t the views of the
authors who are responsible for the facts and accuracy of the data
p r e s en t e d here in. The c o n t e n t s do not nec e s s ar i ly r e f l e c t the
official views or policies of the Univer s i ty of Kentucky , of the
K e n t u c k y T r a n s p o r t a t i o n C a b i n e t , or o f t h e F e d e r a l H i g hway
Adm i n i s t r a t i on.
This report does not cons t i tu t e a s t andar d ,
specification, or regulation.
The inclusion of manufacturer names
and t r adenam es are for ident i f ic a t i on pur p o s e s and ar e not to
considered as endorsementse
October 1987

REPORTS ISSUED
Research Study KHYPR-84-96
Development of a Composite Pavement Design Methodology

==============================================================================

UKTRP
REPORT NO.

84-3

84-11

84-12

85-13

86-21

87-28

87-29

TITLE

DATE

"Thickness Design Curves for Portland


Cement Concrete Pavements"

Feb 1984

"Variations of Fatigue due to Unevenly


Loaded Axles within Tridem Groups''

Apr 1984

"Variable Serviceability Concent for


Pavement Design Confirmed by AASHO

Apr 1984

"Effects of Load Distributions and


Axle and Tire Configurations on
Pavement Fatigue"

May 1985

"Distribution of Strain Components


and Work within Flexible Pavement
Structures"

Sep 1986

"Modi fication to Chevron N-layer


Computer Program"

Oct 1987

"Pavement Designs Based on Work"


(Final Report)

Oct 1987

Technical Report Documentation Poge


1. Report No.

3. Recipient's Catalog No.

2. Government Accession No.

UKTRP-87-29
4. Title oncf Subtitle

5. Report Dote

October 1987
Pavement Designs Based on Work

6. Performing Organization Code


B. Performing Organization Report No.

7. Author's)

H. F. Southgate and R.

c.

Deen

UKTRP-87-29

10. WorSe Unit No. (TRAIS)

9. Performing Orgonlzoflon Nome oncl Address

K entucky Transportation Research Program


College of Engineering
University of K entucky
Lexington, K entucky 40506-0043

11. Contract or Grant No.


13.

KYHPR-84-96
Type of Report and Period Covered

1 2. Spcnsoring Agency Nome and Address


.

Kentucky Transportation Cabinet


State Of fice Building
Frankfort, K entucky 40622

Final
u. Spons oring Agency Code

15. Supplementary Notes

Prepared in cooperation with the US Department of Transportation, Federal Highway


Administration
Study Title: Development of a Composite Pavement Design Methodologv

16. Abstract

Thickness design curves presented in the report provide a sy stematic


methodology for the selection of equivalent pavement designs for a broad
range of layered systems. This is a unified sy s t em s1noe the failure criteria
are founded on the same concept of work s train an work . The analyses of
stress-strain fields in the layered systems are based on elastic layered
theory. This theory is represented by the Chevron N-layered compu ter program.
The report also sunnn arizes the historical development and evolution of
pavement design in K entucky.

18. Distribution Statement

17. Key Word

Pavement Design
Failure Criteria
Work Concepts
Thickness
Portland Cement Conc rete
Asphaltic Conc rete
19. Security Claul f. (af thh report)

Unclassified
Form DOT F 1700.7

Unlimited with approval of K entucky


Transportation Cabinet

20. Security Claulf. (of this po;e)

Unclassified
(B-72)

Reproduction of completed p 1lge lluthotized


1

21. No. of Pages

50

22. Price

INTRODUCTION

The 1949 (1)

and 1959 (2) Kentucky flexible pavement

Revisions made in 1968 ( 3 ) , 1971

curves were based on empirical experience.


(4 ) ,

and 1981 (5) combined elastic

thickness design

theory with

empirical

experience.

The

lat ter three revisions ut ilized a failure criterion of tensile strain at the
bottom of

the asphaltic concrete based on laboratory testing as well as a

criterion developed from theoretical analyses of ver tical compressive strains


at the top of the subgrade.

BACKGROUND

1959 KENTUCKY THICKNESS DESIGN CURVES


The 1949 Kentucky thickness design curves (1) were updated in 1959 (2 )
after a testing program consisting of pavement deflect ion tests and opening of
t es t p i ts on s e l e c t e d pavements to perform p l a t e bearing and in situ CBR
tests,

sampling of materials

from

each

and/or distortion of layer boundaries .

layer ,

and measuring surface ruts

Traffic analyses provided estimates of

a ccum u l a t e d f a t i gue, and the pavements were ass igned a s a t i s f a c t ory or


unsatisfactory performance rating.

Pavements were sorted into groups having

essentially the same accumulated traffic fatigue ( equivalent 18-kip axleloads )


and a curve was drawn through the thickness-CBR plots to separate satisfactory
from unsatisfactory performances.
sufficient

data

to permit

Only two groups of "traffic fatigue" had

complete

analysis.

Those

designated

Traffic

Curves IV and VI (3-6 million and 10-20 million EWL , respectively) represented
two levels of equivalent wheelloads ( EWLs ) .

O ther traffic curves ( accumulated

fatigue) were obtained by interpolation and extrapolation.

Further analyses

indicated that pavement thicknesses associated with Traffic Curve X


m illion EWL) should have been greater.

(160-320

The 1968 curves ( 3 ) did provide such

adjustments , making pavements for Curve X (160-320 million EWL) thicker than
before .

The 1959 Traffic Curves were based on a 5-kip EWL and were converted

to an equivalent 18-kip EAL for the 1968 Kentucky design curves.

Pavements

associated with a given traffic curve were thought at that time to manifest
approximately the same surface deflections.

ELASTIC

THEORY

In 196 8 ,

APPLIED

TO

1959 CURVES

the Chevron N-1ayer computer program (based on elas tic theory)

( 6 ) was obtained and used to analyze

the 1959 Kentucky design curves

(2 ) .

Those analyses indicated that pavements associated with the 1959 Traf fic Curve
X r e s u l t e d in the same ver tical compr e s sive s t r ains rather t h an t h e same
surface deflections.

ELASTIC

MODULI

The elas tic modulus for the asphaltic concrete pavements on the AASHO
Road Te s t was determined to be approximately 600 ksi (7) . This corresponded
a l s o t o a m e an annu a l t em p e r a t u r e o f 6 0 d e gr e e s F.

T h e m e a n annu a l

t emperature for Kentucky is 70 degrees F , resulting in an equivalent modulus


of approximately 480 ksi ( 3 ) .
average CBR o f 7.

Subgrades of pavements tested in 1957 had an

Mi t c h e l l and Shen ( 8 ) had de t ermined t h a t the elas tic

modulus of clay could be estimated by multiplying the CBR by 1500 .

CRUSHED STONE BASE


In the 1968 analyses, a value of 25 ksi was assigned as the modulus for
crushed stone base.

Subsequent analyses showed that this created a weak layer

between two stronger

layers for

subgrade

CBRs

greater

than 17,

leading

to

unrealistic results.
Ma tching theory to Tr a f fic Curve X a l s o permi t t ed deve lopm ent of a
r e l a tions hip of moduli for t h e cru s h e d s tone layer as a function of t h e
m o du l u s o f t h e asphaltic concr e t e and t h e CBR o f the su b gr ade .

Ana l y s e s

indicated that the modulus o f the crushed stone layer increased a s the CBR o f
t h e subgrade increased.

The modulus o f the crushed stone base was 2 . 8 times

the modulus of the subgrade at CBR 7.

A theoretical Bousinesq solution could

be estimated as the CBR equivalent to the asphaltic modulus (480 ksi) divided
by 1500 -- a CBR of 320.
log(F)

i n which F

These two data points were used to define

0 . 674797 - 0.269364 * log ( CBR)

fac tor u s ed
modulus

CBR

to multiply 1500

for

California

crushed stone bas e , and

Bearing Ratio.

CBR

to

o b t ai n

the

FATIGUE CRITERIA
Dorman and Metcalf (9) determined from laboratory tests that fatigue of
asphaltic concrete could be described by

log( a )

-2. 69897 - 0.163382

tens i1e strain at

in which ea
concrete.

log(EAL)

the bo t tom of

t h e a s p h a 1 t ic

The above relationship ( Figure 1 ) was used in the development of the 1971 (4 )
and 1981 (5) thickness design methods.
While develop ing the 1 9 6 8 Kentu cky thi ckne s s de s ign curves (3 ) , a
vertical compres s ive s train of 2.4 x 10 4 at the top of a CBR 7 subgrade was
determined to correspond to a fatigue of 8 x 106 18-kip EALs under the center
o f one 9-kip circularly loaded area.

Damage factors ,

sometimes called load

equivalency factors , were calculated using


DF

in which

(1 .2504 ) (P-1B)

P
DF

axleload in kips and


damage factor
(EAL assigned to e a caused by one 18-kip axleload)

--------------------------------------------------

(EAL ass igned to e a caused by a given axleload)


S e veral pavement s tructures rang ing from r e l a t ively thin to

thick were

subjected to a range of loads and analyzed using the Chevron N-layer program.
An arithmetic mean of the vert ical compressive strains ( for different pavement
t hicknesses) was calculated for each of the various loads.

Each mean strain

value was plotted versus the associated EAL calculated using Equation 2 to
produce Figure 2.

This procedure resulted in a prominent downward bend of the

criterion relat ionship for high EALs that was not an extension ( extrapolation)
of trends for EALs less than 8 x 1 06 ,
resulting in thicker designs for large
EALs.

1971

KENTUCKY

THICKNESS DESIGN METHOD

A series of nomographs (4 ) were developed that accurately described the


theoretical behavioral of pavements.

Use of those nomographs by a group of

engineering students resulted in such a range of pavement thickness designs


that the nomographs were assessed as inappropriate for general use.
Several attributes and characteristics of the nomographs, however, became
an integral part of subsequent research efforts.

Results from the 1959 test

pits ( 2 ) indicated that distortion at the top of the subgrade diminished as


the pavement thickness increased, which in turn was related to an increase in
design traffic.

Thus, for designs greater than 4 million EALs,

should be fully protected from distortion.

the subgrade

In general, geometries of farm-to

market roads would preclude speeds necessary for hydroplaning because of water
s t anding in the ruts on the pavement surface.

Thus,

the criterion for such

r o ads would a l low the su bgrade to dis t o r t ( ru t ) provided the aspha l t i c


concrete is protected from fatigue cracking during the design life.
of

the 1959 Kentucky

thickness design method was

Curve IA

associated with farm-to

m arket roads subj e c t e d to t h e equivalent of one application of an 18-kip


axleload per day for 20 years.

Criteria between 7, 300 and 4 million EALs were

established so the allowable distortion of the top of the subgrade decreased


as EALs inc r e as ed.

That was accomplished by de t e rmining

t h e required

thicknesses associated with both the asphaltic concrete and subgrade strain
c ri t eria and adding a proportion of the di f f erence in thicknesses to t h e
t hickness required by the

asphaltic concrete criterion ( Figure 3 ) .

A n a t t e m p t was made to assess the compatibility o f bo th a s t rain


controlled and stress-controlled criterion for a range of moduli of asphaltic
concrete

(3 ).

For a strain-control criterion, results of laboratory tests

reported in the li terature indicated that strain-fatigue relationships for a


range of asphaltic concrete moduli

could be

expressed as parallel lines by

mos t investigators or as a series of lines converging at very low tensile


s trains and a large number of repetitions by others.

Analysis of Kentucky

experience indicated that the most appropriate relationship could be expressed


a s a family of s t r ain-fatigue lines asso ciated wi t h a range of mo duli of
asphaltic concrete that converge at one repetition and a large ( catastrophic)
v a l u e of tensile s train.

Thus,

for a given level of fa tigue, values of

t ensile s t rain co u l d be d e t e rmined for any desired mo dulus of asp h a l t i c


concrete.

However, the same approach was not applicable for the development

of a single stress-controlled cri terion.

One repetition of a catastrophic

load resulted in a different critical stress value

and a separate stress-

fatigue relationship for each modulus of asphaltic concrete.


possible

to

develop

design method

incorporating

both

Thu s , it was not

stress and

strain

c riteriae

REVISIONS TO CHEVRON N-LAYER COMPUTER PROGRAM


The original version of the Chevron N-layer computer program ( 6 ) utilized
a single circular loaded

area to obtain stresses, strains, and deflec tions at

radii from the center of that load.

The program was revised to input multiple

loads at spec ified X-Y coord inates on


s trains ,

and

deflections

us ing

the surface and to obtain stresse s ,

superposition principles

at

any

designated

location within the X-Y-Z coordinate system , the Z coordinate being the depth
below the surface .
Reference 1 1 .

Documentat ion and example p r o b lems are conta ined in

This development made it poss ible to inves t igate various loads ,

tire spacing s , axle spacing s ,


loads

within

the

same

group

number of axles/ tires w i thin a group,


of

tires/ axles ,

tire

contact

uneven

pressure s ,

etc.

Another revision was the addition of an equation ( 1 0 , 1 1 ) to calculate strain


energy dens ity .
W ith the revised program , all 100 combinat ions of layer thicknesses used
a t the AASHO Road Te s t , of which 67 were constructed ( 12 ) , were analyzed .
matrix

of

t i re

loads

ranging f r om 2 , 00 0

t o 8 , 00 0 p ounds

on 500-pound

increments per tire were applied to each pavement for a two-tired single axle
w i th a spacing to represent a steering axle ,

a four- tired

eight-tired tandem axle , and a twelve- tired tridem (5).

single axle ,

an

Additional studies

were initia ted to quantify the effects of different axleloads within the same
g r oup

of axles

( tandem and

tridem)

(13 ) ,

ad d i t ional ax les

addi tional tires per axle , and various tire pressures ( 14 ).

in a g r oup ,

These analyses

permit ted development of load equivalency factor s , or damage fac tor s , for axle
and t i re arrangemen t s u sed at

the AASHO Road Te s t

( 1 5 ) a s w e l l a s new

arrangements now seen on highways.


Results for the four- tired single axle were in ve ry close agreement with
results us ing the AASHTO Equation C-16 ( 1 6 ) at a level of serviceability of
2. 5 and within the range of axleloads employed at the Road Tes t .
factors for tandem axles were different.

Howeve r , the

Ad jus tment factors were developed to

account for uneven axleloads within a tandem or tridem group for axles weighed
o n s ta t ic weigh scales ( 1 4 ) .

Ad j u s tment fac t o r s a l s o were developed to

account f o r inc reased t i re contact pressures curre n t ly being used.


5

The

effects of

increased

tire

contact pressures vary widely,

thickness of the asphaltic concrete layer

( 14 ) .

depending on

Results of

the

those analyses

w e r e app l i ed to a t r a f f i c s t ream on a p a r t i cu l a r i n t e r s t a t e pavement and


comparisons made .

Accounting for tire/axle arrangements, uneven loading among

axles in the same group, and increased tire contact pressures,


load

equivalencies

resulted

in an accumulated

fatigue

that

the Kentucky

was

1. 27

times

greater than that obtained using AASHTO factors ( 1 6, 17 ).


A compu t e r program was w r i t ten to ca lcu l a t e an average EAL for each
vehicle type based on truck traffic weighed at a particular weigh stat ion and
r e corded on a lo adome t e r compu t e r tape.

The EAL w a s calcula ted for each

vehicle, summed for that truck classification, and an average value calculated
for each truck classification.
U s ing the Kentucky load equivalency r e l a t ions hip s to calcu l a t e the
average load equivalency per trip for each vehicle classification results in a
c a lcul a t ed 1 8-kip EAL t h a t is appro x i m a t e l y 7 3 p e rcent of that u s ing the
AASHTO load equivalency r e l a t ionships.

Combining the e f f e c t s of uneven

loading and the Kentucky load equivalency relationships results in an 18-kip


EAL t h a t i s app rox i m a t e l y 88 p e rcent of the AASHTO EAL.

Combining the

K entucky load equivalency relationships, uneven loading, and effects of tire


contact pressure results in an 18-kip EAL that is approximately 127 percent of
the AASHTO EAL.

Comparing results within the Kentucky procedures only, uneven

loading between the axles within that group of axles increases the calculated
EAL by approximately 20 percent.

Increased tire contact pressure resulted in

appro x i m a t e l y an ad d i t ional 55 p e rcent i n c r e a s e in EAL.


additional

effects

of uneven load

dis tribut ion and

Combining the

increased

tire

contact

pressure increased the calculated EAL approximately 75 percent compared to the


EAL u s ing only Kentucky load equivalency relationships.

1981

KENTUCKY THICKNESS DESIGN CURVES


The same matrix of pavement layer thicknesses used in the 1968 ( 3 ) and

1971

( 4 ) studies was used in this revision.

The modulus of

the asphaltic

concrete was selected to be 480 ksi, based on results discussed previously.


The modulus of the crushed stone base was varied according to Equation 1.

An

18-kip four- tired single axleload for tire spacings in use on current trucks
w a s app l i e d to the s u r f ace o f each pavement.
p r e s s u r e w a s u s e d i n a l l ca s e s.

The 8 0- p s i tire co n t a c t

Po i s s on's r a t io w a s 0.40 f o r aspha l t i c

concrete and dense-graded aggregate bases and 0. 45 for the subgrade.


6

For a given subgrade modulus , the vertical compressive strain at the top
o f the subgrade was plotted as a function of thickness of asphaltic concrete
by v i sual f i t .
d rawn.

Curves of equal thickne s s of d e n s e-graded aggr egate were

The same me thodology was u s ed to pl o t rela t ionships of sur face

deflection and tensile strain at the bottom of the asphaltic concrete.


The same criteria used in the 1971 and 1981 Kentucky thickness design
methods ( 4 ,

5) were used in the development of

pavem e n t s where

the

thickne s s

of

p ercentage of the total thickness.


f ou r p e r c e n t a g e s ( 3 3 ,
Subsequently ,

SO ,

thickness design curves for

the a s p ha l t ic conc r e t e wa s

the

same

Thicknes s des ign curves were developed for

7 5 , and 1 0 0 )

curves were developed

o f a s p ha l t ic c o nc r e t e

(5) .

for pavements consisting of asphaltic

concrete on 4 inches of dense-graded aggregate base and asphalt ic concrete on


s tabilized base materials

(18)

such as pozzolanic materials ,

cement- treated

material s , fly-ash mixtures , etc.

RUTTING INVESTIGATIONS
W i th the influx of heavy trucks into the coal fields of Kentucky came
increased ru t ting of heavy duty pavements.

Most severe rutt ing was located on

fairly steep upgrades and at intersections having traffic control devices.

To

i nve s tiga t e this phenom enon, 4 - f o o t wide trenches were dug to a depth of
approximately 3 feet so various layers could be observed carefully.
t h o s e pavem e n t s c o n s i s ted of 1 7 and 18
c oncrete.

For both those pavements ,

inc h e s of

Two of

full- d ep th asphal t i c

there was no dis tortion from the bottom

of the asphaltic concrete up to 6 inches below the surface of the pavement.


From that depth to the surface ,

the distort ion increased.

Distortion in the

top 6 inches was noted in three ways:

1 . In the wheelp a t hs , the

cons truc t i on i n t er face for each

succeeding lift of pavement was displaced vertically more severely


than the interface below.
2.

The top layer

wheelpaths and much

o f sur face m ix was much thinner in the

thicker

between wheelpaths ,

indicating

shear

flow within the surface mix.


3 . Random orienta t i on of aggr ega te par t ic l e s in the lower
layers changed gradually from the 6 - inch d e p t h to a parallel
o r i e n t a t i on at the surface.

Ag g r eg a t e par t ic l e s very near the

surface appeared to be separated by,

and sliding

on ,

a layer of

asphalt cement.

To verify shear flow was occurring , two 0.25-inch wide by 0. 25-inch deep cuts
were made in the surface of the pavement.

The first cut was made at an angle

t o the centerline of the road , and the second cut was perpendicular to the
centerline.

These cuts were filled with glass beads used in traffic paint to

p revent closure and to provide an identification of the cuts later.

After

three weeks , the cuts in the wheelpaths were displaced downhill approximately
0 . 6 inch.

No me asurements were made to de termine if the cu ts had been

displaced laterally.

FUNDAMENTAL RELATIONSHIPS

WORK
The following equation ( 1 0 ,

11,

19)

was

added

to the Chevron N-layer

c omputer p r o g r am to calcu l a te the s t rain energy density at a given point


within the pavement structure:
2
2
2
w = (1/2 ) * ).. + )l* (
1 1 + e 22 + e 3 3
+ 2*2 1 2 + 2*8 223 + 2*6 2 13 )
in which W = strain energy density,

4
or energy

volume , or the volume density

eij

of

of

deformation per unit

strain energy,

i , jth component of the strain tensor,

)l = E/ ( 2 ( 1 + ) ) , the modulus of rigidity or the shear modulus ,


E = Young's modulus ,
(J"

= Poisson's ratio

).. = E/ ( ( 1 + rr)(l - 2) , and


v = 8 1 1 + 8 22 + 633
S train energy density takes into account all nine components of strain,
o r s t ress , four of which wi l l have no resul t ant value because one shear
c omponent balances another

component

for

c ompone n t s are c a l c u l a te d and printed.

two situati ons. Howeve r ,


W o r k is

the

all

three- dim ensional

summation of strain energy density for the volume of material involved .

Thus ,

i t was assum e d tha t , for a unit volume a t a given point in the pavement

s tructure, work also was equal to the calculated strain energy density ( Work =
in. 3 x psi = in.-lb).
WORK STRAIN
The equation

"Work strain" as used in this study is not a pure strain.


t o c a l c u l a t e s tr ain energy density contains
involving Poisson's ratio.

two

di f f erent

terms,

each

Each term involves either the square of the sum of

t he princip l e strains or the sum of the square of each s train component.


Dividing the s train energy density by 0 . 5

E and taking the square root

yields a numerical value of the same order of magnitude as


s train component.

any individual

This value has been called work strain and is taken as the

net effect of all strain components.

Similar calculations may be made using

s tress components and yielding work stress,

equal to the product of Young's

modulus of elasticity and work strain.


Laboratory fatigue tes ting typically utilizes strain gages placed at the
bot tom of rectangular beams of asphaltic concrete.

Dimensions of the sample

beam genera l ly are such that the s train gage is fu nc tional only in the
direction of the longest dimension.

Thus, the tensile strain at the bottom of

the asphaltic concrete has been used to develop theoretical thickness design
curveso
I t should be noted that, direc t l y under the c e n t er of the load, the
radial strain equals the tangential tensile strain.

However,

for any other

location, the tangential tensile strain is the larger of the two, but in most
c ases is only slightly larger than the radial component.

Wby, then, should

t h e r a dial component be ignor ed, and how shou l d the radial component be
included?

Likewise, the shear component also may be significant, particularly

at any location other than under the center of the loaded area.

The 1981

K entucky thickness design method, The Asphalt Institute design method

(20 ) ,

and the Shell design method (2 1 ) have utilized the tensile strain component at
the bottom of the asphaltic concrete and the vertical compressive strain at
the top of the subgrade and ignored all other components.
design procedure incorporates
cri teria.

all

strain,

or

stress,

No known thickness
components

into

its

Work strain incorp orates the effects of all strain components and

r esolves the conflict discussed above.

WORK VERSUS WORK STRAIN


The TOTAL amount of work done by the pavement structure due to an applied
load involves

a three-dimensional summation

of

strain

energy

becomes extremely expensive in terms of money and compu ter time.

density

that

The area of

g r e a t e s t range of w o r k would be wi t hin a volume tric slice of unit width


p assing through the center of the loaded areas along a given axle that is
either a single axle or one of a group of axles .

Locations of greatest unit

work were identified in a previous study ( 2 2 ) , and maximum unit work was noted
t o be under the e d ge of the tire closest to
s tress/ strain component

the o t h e r dual tire.

having the greatest variation from point

under the tire contact area was shear.

The

to point

Shear has its greatest influence at

t h e o u t e r e d ge o f t h e t i re c o n t a c t a r e a.

T h e ve r t i c a l c o m p o ne n t o f

s tress/strain is greatest at the center of the tire contact area.


Theoretically, the total amount of work caused by a given axleload is the
The work calculated at specific

same without regard to pavement structure .

locations within a pavement structure varies according


and subgrade moduli.

to layer thicknesses

This variation is relatively smal l , and a small change

in work requires a large change in thickness.

Howeve r , work strain provides a

much wider range of values and permits easier usage in developing thickness
design curves .

Thus , for convenience and ease of use , work st rain has been

chosen as the variable upon which to base the thickness design curves.

LOCATION OF GREATEST REACTION


A n inves tigation was made to de te rmine the l o ca tion of t h e greatest
reaction to applied loads within two- and three-layered pavement structures.
For normally spaced dual tires , this location was beneath the edge of the tire
c l oses t t o the adjacent tire ( 5 , 1 3 , 1 4 , 2 2 ) .

Figure 4 i l lu s t r a tes the

relationship between work at the bottom of the asphaltic concrete with respect
t o location along the axis of the axle.

The numeri c a l value of work was

slightly larger beneath the edge of the inside tire than under the edge of the
outside tire; but the difference was negligible , particularly when compared to
t he value of work under the center of the loaded area.
For two-layered pavement structures ( 2 2 ) , the primary contributor to work
was the shear component -- stress or strain.

Als o , the distribution of work

from the center to the edge of the tire varied greatly, and the variation was
predominantly due to the shear component.

Maximum shear was reached at

d ep th f r om the s u r f a c e equal to approxim a t e ly 3 5 t o 4 0 pe rcent o f t h e

10

thickness of the asphaltic concrete layer.


asphaltic concrete,
below the surface.

Thus , for an 7 . 5-inch thickness of

the maximum shear would be at approximately 2 . 5 inches


Considering that Kentucky often uses a l-inch surface over

a 1 . 5-inch binder cours e , the maximum shear is very close to a construction


i nt e r f a c e -- the point of least aggregate in t e r l o ck.

A t an interf ace ,

compaction equipment tends to orient the longest axis of an aggregate particle


p ar a l l e l to the s u r f ace.

Thus , resis tance to shearing is minim a l ,

and

p ar ticles may move l a t e r a lly if fric tion between a g g r e g a t e p a r t i c les is


overcome by loading forces and/or the tensile properties of the asphalt cement
Lateral

is exceede d , whi ch may o c cur at higher pavement temperatu res.


m o vement o f lower p a r t i c les allows upper p a r t i c l es

t o m o v e vertically

downward , resul ting in surface rutting within the wheelpath.

CHEVRON N-LAYER ANALYSES


A matrix of layer thicknesses,

moduli of

asphaltic

concrete,

sub grade

moduli , and crushed stone moduli ( as a function of subgrade modulus)

were

analyzed using the Chevron N-layer computer program as modified by Kentucky


(11 ) .

The LaGr ange i n t e rp o l a tion m e t hod was used t o f i t

p o lynomial

equations

for

t hicknesses and then used

various

combinations

to interp o l a t e

for

within
other

third degree

that

m a trix

of

required pavement

thicknesses.

COMPARISON OF WORK STRAIN AND STRAIN COMPONENTS


Figu res 5 and 6 i l lu s t r a t e the relati onship between wo rk s t r ain and
t ensile s t r ain

at

the

bot t om

of

the

asphaltic

concrete

compressive strain at the top of the subgrade, respectively.

and

vertical

The correlation

in Figure 5 is not as g o o d as shown in Figure 6 because the rela tionship


between radial st rain and layer thicknesses results in a larger scatter of
data than for

the relationship between tensile strain and layer thickness.

Since work strain is the net effect of all components , the correlation between
work strain and radial strain is not as good as for work strain and tensile
s t rain .

The radial s t r ai n at the top of the subgrade is not nearly as

inf luential as at the b o t t om of the asphaltic concr e t e layer.

Thus ,

the

relationship between work strain and vertical compressive strain at the top of
the subgrade has a very narrow band of data scatter .
Many laboratory fa tigue tes ts measure the tensile st rain component.
Thus , the relatively close correlation between work strain and tensile strain

11

of the bottom fiber has

the advantage of utilizing all previous

test results and permits utilization of previous experience .

laboratory

Confidence in

the newer approach is increased because previous experience is utilized.

DAMAGE FACTORS
Fatigue is defined as

EAL associated wi th the wo rk strain caused by 18-kip


single axleload
DF

----------------------------------------------------

EAL associated wi th wo rk strain caused by axleload P

The 1959 Kentucky design curves were based on the following fatigue equation
for single and tandem axles (2 , 3 ) :
6

in which DF5 9
A

damage factor used in the 1959 Kentucky design system;

co n s t ant

that

is

the

slope

of

the

s e m i l o g a r i thm i c

r e l a t io n s hip b e tween axleload and r ep e t i t ions


and has

a value o f 1.2504 for

f o u r t i r e s a n d a v a l u e of

single axles with

1 . 1254 for

t andem

axleloads;
B

(P - 1 0 )

to correspond to the EWL s y s tem u s e d in the

1 9 59 Kentucky design system ,


=

(P

18 )

(P - 3 2 )

for four- tired single axles (1968 ) , and


f o r e i g h t- t i r e d t a n d e m a x l e i n t h e 1 9 6 8

Kentucky des ign system; and


P

axleload in kips.

The 1981 Kentucky thicknes s design curves were based on the general equation
expressed as
log(DF8 1 )
i n which DF 8 1

C + D(P ) + E(P) 2

damage f a c t o r u s e d a s a p a r t of the 1 9 8 1 K e n tucky d e s ign


curves

and

12

C , D, E

coefficients

of correlation depending upon tire and axle

c onfiguration ( see Table 1 of Reference 14) .

COMPARISON OF 1959 AND 1981 KENTUCKY THICKNESS DESIGN CURVES


Pavement thicknesses represented by the 1959 Kentucky design curves were
t o be comprised of 1/3 asphaltic concrete and 2/3 crushed stone base .
of layer design thicknesses had been prepared for convenience .

A table

A s a resu l t ,

t h e 1 / 3 - 2/3 r a t i o o f a s p h a l t i c concrete t o cr u s hed s t one b a s e w a s no t


maintained .

For example ,

the thicknes s of crushed stone base might vary as

much as 3 to 4 inches for the same thickness of asphaltic concrete ,


i n per centage s of asphaltic concrete thickne s se s

resulting

ranging from 2 8 t o 5 2 ,

r ather than the design value of 3 3 .


Thickness designs were determined for given values of CBR and 1 8-kip EALs
from both the 1959 and 1981 curve s .

Values of work strain at the bottom of

the asphaltic concrete and at the top of the subgrade were


those designs .

determined

for

Figure 7 shows the relationship between EAL and work strain at

the bottom of the asphaltic concrete .

The correlation is not good .

Figure 8

s hows the relationship between EAL and work strain at the top of the subgrade ,
and the correlation is much better .

The larger scatter for the 1959 design

curves is a s s ociated with the range in the per centages of the asphaltic
concrete of the total thickness described above.

Some of the scatter for the

1981 curves might be attributable to the use of one component of strain rather
than work strain.

However , best fit curves through their respective sets of

data intersect each other as shown in Figure 8 .

For relatively high values of

work strain, the 1959 curve is associated with a higher design EALs ( thicker
pavement) than that for the 1981 curve .
s train ,

Conversely, for lower values of work

the 1 9 5 9 curve is a s s ociated with a le s ser design EAL ( thinner

pavement) than for the 1981 curve .

FATIGUE RELATIONSHIPS
The 1959 Kentucky thickness

design curves were empirically based;

the

1981 Kentucky thickness design curves established a theoretical basis to merge


with those empirical dat a .

However , Figure 8 displays two intersecting curves

( one for 1959 design curves and one for 1981 curves) for data that had been
a s sumed to be essentially the same .

Kentucky W-4 tables for 1959 through 1984

were used to describe a traffic stream .

Using the count and weight data, it

was possible to es timate the 20-year EAL based on each year's traffic data as
13

shown in Figure 9 .

Thu s , a relationship was obtained that permitted adjusting

an EAL calculated using the 1959 damage factors to an equivalent value based
upon the 1981 damage factors given by

DF

in which

2
(
a+
b*P
+c*P
)
10

a, b, c

=
=

axleload in kip s and


constants given in Table 1 for respective axle configurations .

These correlations show that the empirical experience matches the theoretical
work strain at the top of the subgrade while the limiting tensile capacity of
the asphaltic concrete is no t exceede d .

STRESS- OR STRAIN-CONTROLLED FATIGUE RELATIONSHIPS


Reference has been made earlier to the procedure used to proportion the
design thicknes s according to the difference in thicknesses required by the
t en s i le s t rain a t the b o t tom o f the aspha l t i c concrete and the ve rtical
compressive s train at the top of the subgrade.
the 1981 Kentucky curves (5) for EALs

less

Final design thicknesses from


than 4 x 10 6 has an equivalent

vertical compressive strain that may be equated to an equivalent work strain


and/or work stre s s .

Therefore , one design procedure is possible

for both

s tress-controlled and strain-controlled fatigue criteria as shown in Figure


1 0.
The "hooks" at either end of the 1981 Kentucky design curves (5) were the
result of averaging large variances beyond
encountered on the highways .

the

range in axleloads normally

In the range of typical 1 8-kip design EALs , the

relationship is almos t a straight line on a log-log plo t , as shown in Figure


11.
A regression equation was fitted to the relationship between work st rain
a t the top of the subgrade and 1 8-kip EALs
positioning thicknes s design curves .

to determine the criterion for

One equation produced close agreement

with the 1981 Kentucky thicknes s designs for the range of low EALs.

However ,

large discrepancies occurred for low CBR values in the range of high EALs.
Thu s ,

regression equations

for criteria were obtained for

CBRs: 3 to 5 , 5 to 7 , and 7 and greater .

three

ranges of

These three curves became the basis

for the development of all thicknes s design curves presented in this report .
14

The "hook" at each end of the criterion line (5 ) has been straightened
f or each of the three criterion lines in Figure 11 .

The re s u l t i s that

designs for low volume roads are slightly thicker than those required by the
7 EAL s ) , design
1981 Kentucky design curve s .
For very high 18-kip EALs (10
thicknesses are slightly thinner than those required by the 1981 curves ( 5 ).

1987 KENTUCKY THICKNESS DESIGN CURVES


CONVENTIONAL ASPHALTIC CONCRETE/DENSE-GRADED AGGREGATE DESIGNS
Total thicknes s design curves have been constructed for structures where
the thicknesses of the asphaltic concrete are 3 3 , 5 0 , and 7 5 percent of the
total pavement thicknes s as shown in Figures 12-14 .
s tructures

having

asphaltic concrete.

4-inch

thickness

of

Figure 15 is for pavement

dense-graded

aggregate

below

the

These design curves were based on work strain at the top

of the subgrade as the failure criterion.

Design thicknesses

shown on the

vertical axes in Figures 12-15 are the total thickness of all layers above the
subgrade .

ASPHALTIC CONCRETE OVER BROKEN PORTLAND CEMENT CONCRETE


A n a ly s e s o f dynamic d e f l e c t ions indicate the modul u s o f fragmented
p o r t land cement concre te i s

a function of

the size

of broken piece s .

Preliminary analyses indicated the modulus is approximately 10 ksi for totally


crushed material and increased to approximately 1 , 000 ksi for 30- to 3 6-inch
pieces of portland cement

Overlay thicknes s design curves using

concrete.

a sphaltic concrete are presen ted for three moduli of broken and sea ted
portland cement concrete pavements ( Figures 16-18 ) ,
s hown

in

Figures 16-18 may

be calculated

Overlay thicknes s designs

using the equation presented in

Table 2.

A SPHALTIC

CONCRETE OVER PORTLAND CEMENT CONCRETE

Analyses ( 2 4 ) of temperature gradients in asphaltic and port land cement


concrete pavement s indicated

that

at

least 5

inches

of

asphaltic

concrete

overlay is required so the temperature at the aspha l t i c-port land cement


concrete interface wi ll be no higher than if the port land cement concrete were
exposed directly to the sun.
materials is

almo s t identical ,

The coefficient of heat transfer for the


but

the

15

coefficient of heat

two

absorption for

b l ack m a t e r i a l s i s almo s t twi c e that for wh i t e m a t er i al s .

Ther e f or e , a

s ignificant thicknes s of black (asphaltic) material is required to dissipate


the absorbed hea t .

Without the proper thickness of asphaltic concrete,

portland cement concrete will tend to expand with rising temperatures .

the
If

expansion i s not possible , compressive forces will increase and may result in
crushed concrete at the faces of sawed joint s , or blowups may occur .
Any overlay over a joint

that has differential ver tical movement will

exhib i t a r e f l e c t ion crack and no thickn e s s of over lay wi l l pr event the


reflection of

If

that joint .

there is no vertical movement ,

the required

overlay thickness is not a function of fatigue but is a function of the annual


r ange of tempera tur e s .

Thi ckne s s es in exc e s s of 5 inches ( for K entu cky

conditions) will minimize horizontal (expansion/contraction) movements of the


port land cement slab .
I t does not appear that fatigue is the controlling factor for this type
o f pavement

With heavier

structure .

axle loads

and increased

tire

contact

pressures , rutting has been observed in other states and provinces as a major
problem . This phenomenon may be a result of shear flow within the asphaltic
concrete layer .
Nei ther a cr itical magnitude of shear stress/ strain nor a fatigue-shear
relationship has been identified in literatur e .

However,

current research

involves subjecting an asphaltic concrete tubular specimen to tors ional shear


(25 ) .

T e s t r e s u l t s i n d i c a t e that the cr i t ical tor s ional shear s t r e s s i s

relatively low.

Efforts are underway to compare torsional shear results for

hollow and solid specimens.

Therefor e , it is premature to develop thicknes s

des ign curves based o n shear cr iterion .

PORTLAND CEMENT CONCRETE OVER ASPHALTIC CONCRETE


An exis t ing asphaltic concrete pavement on a crushed stone base has a
total thickness such than even a 4-inch portland cement concrete pavement on
asphaltic concrete results in a very low value of work strain at the top of
the subgrade .

Thu s , the work strain-fatigue relationship presented in Figure

8 is not appropriat e .

The required fatigue criterion must be appropriate to

portland cement concrete .


D evelopment of thicknes s design curves for port land cement concrete over
a crushed stone base has been reported previously ( 2 6 ) .

The design cr iterion

depended almos t entirely upon the work strain at the bottom of the portland
c em ent concr e t e s lab .

The f a t igue cr i t er ion was a merger o f emp i r i c a l

16

c ri t eria u s ed by the Po r t land Cem ent A s so ci a tion and

by

A s sociation of S tate Highway and Transportation Officials .


system was based on empirical data

the Ameri c an
The PCA design

from highways having relatively low truck

volumes and lesser gross loads in the 1940's as compared to much higher truck
The concept of work

traffic with larger gross loads at the AASHO Road Tes t .


s t rain provided

the n e c e s s a r y key to merge the two cri t e ria into one .

Thi ckne s s d e sign curves p r e s e n t ed in Figure 1 9 were


that

a s sumption

the

existing

asphaltic

developed

concrete pavement

has

on

the

deteriorated

until the modulus is 200 ksi instead of the 480 ksi of new material .

The same

fatigue criteria for port land cement concrete (2 6 ) is used herein .

PORTLAND CEMENT CONCRETE OVER PORTLAND CEMENT CONCRETE


K entu cky's thickness d e sign cu rves (2 6 ) should be u s e d f o r port land
cement concrete overlays that are to be fully bonded
c ement concr e t e .

to

exis ting portland

Unhanded p o r t land cement concr e t e overlays

r ecommended at this time .

Warping and curling stresses

are no t

cause the unhanded

s lab to be seated or unseated, depending on the time of day and traffic .

It

i s conceivable that corners might break off and cracks develop at mid- slab
p rematurely.

Locating new sawed joints directly over old joints and cracks is

a difficult and tedious task.

Some pavements overlaid with portland cement

concrete were seen in Iowa where the sawed joint was only 0 . 25 inch from a
crack reflected from the underlying slab.

Reflection cracks are a severe

p rob lem for ei ther bonded or unhanded conc r e t e overlay s .

Joint s e a ling

becomes more expensive when two closely spaced joints occur because the new
s awed joint does not coincide with the old joint .

SUMMARY

A matching of empirical experience with elastic theory adequately defines


f a tigue.

F a tigue failure is d e s cribed by wo rk s t rain at the top of the

subgrade.
Shear is a material problem ,
However,

at least partially related to mix design.

the location of maximum shear is much nearer the surface than had

been thought and the maximum tolerable value of shear strain, or stre s s , is
much less than had been thought .

Maximum shear occurs at depths that very

o f t en coincide with con s t ru c tion i n t e r f a c e s between layers of asphaltic

17

c oncrete ..

Consider ation should be given t o changing

the

cons truction

thicknesses of layers near the surface of the pavement .


C alculated strains and stresses for multiple-tired loads usually are less
t han those due to a single loaded area .
The point of maximum work strain is located beneath the edge of the tire
closest to the adjacent tire .
C oncepts of work and work strain combine all components of strain into
one net value .

The work strain at the top of the subgrade appears to closely

c orrespond wi th over 40 years of empirical experience in Kentucky.


Figure 10 shows that the concept of work strain and work stress provide
the means to base a thickness design method on both strain-controlled and
s tress-controlled criteri a .
The fatigue criterion for portland cement concrete pavements also applies
t o portland cement concrete overlays .

Fatigue criteria for asphaltic concrete

o verlays are applicable

asphaltic

to

existing

concrete

broken and seated portland cement concrete pavements .

pavements

and

for

Shear may overshadow

f a tigue criteria for asphaltic concrete overlays on existing non-broken


por tland cement concrete pavement s .
The following is a listing of charts appropriate for various pavement
designs:
FIGURE NO .

DESCRIPTION
New pavements, 33 percent asphaltic concret e ,
67 percent dense-graded crushed stone base

12

New pavements , 50 percent asphaltic concrete,


50 percent dense-graded crushed stone base

13

New pavements, 75 percent asphaltic concrete ,


2 5 percent dense-graded crushed stone base

14

New pavements , asphaltic concrete on 4-inch layer


of dense-graded crushed stone base

15

Asphaltic concrete over broken and seated portland


cement concrete having a modulus of 25 ksi
9-inch concrete pavement

16a

10-inch concrete pavement

16b
18

A sphaltic concrete over broken and seated portland


cement concrete having a modulus of 100 ksi
9-inch concrete pavement

17a

1 0-inch concrete pavement

17b

A sphaltic concrete over broken and seated portland


c ement concrete having a modulus of 200 ksi
9-inch concrete pavement

18 a

1 0-inch concrete pavement

18b

Port land cement concrete over asphaltic concrete


having a modulus of 200 ksi

19

FUTURE RESEARCH

The theme of the final sess ion of the Sixth International Conference on
S tructural Design of Asphalt Pavements held on July 17 , 1987 , was "Paving the
Gap".

Several speakers stated that results of research should be in a format

t h a t i s p r a c t i c a l and easy to imp l ement and u s e.

P r o f e s s o r P e t er Pe l l ,

University of No t tingham , stated that research should be practical; yet , there


a lways will be a need to have fundamental research so frontiers will continue
to be advanced.

Such is the case with the concep ts of work and work strain.

Preliminary analyses of the observed behavior of pavements at the AASHO Road


T e s t us ing principles of work st rain show promise.
the analyses is needed.

Additional refinement of

Combining the behavior of the AASHO Road Test with 45

years of Kentucky's empirical and theoretical designs would provide a sound


b as i s for the development of a mechanis t ic thickness design system covering a
w i d e range o f inp u t values for required parameters.

Such a s y s tem has

potential for analyzing data collected for the Long Term Pavement Performance
port ion of the Strategic Highway Research Program.
A

cursory analysis

indicated

significant

than

f a tigue in asphaltic concrete pavements and overlays as axleloads and

tire

contact pressures increase.

that

shear may be more

Also , the amount of shear and the amount of work

m ay be grea t e s t in the top 3 to 4 inches ( top 2 5


thickness of asphaltic concrete thicknes s ) .

19

to 40 p e rcent of

the

IMPLEMENTATION

Thickness design curves have been prepared for new asphaltic

concrete

p avements and asphaltic concrete overlays on broken and seated portland cement
concrete pavement .

All sets of curves are based upon the principal of work,

provide equivalent designs , and may be implemented as presented in the report.

20

REFERENCES
1.

R. F. Baker and w. B. Drake , "Investigation of Field and


Methods for Evaluating Subgrade Support in the Design
F l exible Pavement s ," Bulletin No. 1 , Vol 4 , Engineering
S t ation, University of Kentucky, Lexington, KY, September

2.

w. B. Drake and J. H. Havens , "Kentucky Flexible Pavement Design


S tudies , " Bulletin No. 4 , Vol 1 3 , Engineering Experiment Station,
University of Kentucky , Lexington, KY , June 1959.

3.

H. F. Southgat e , R. c. Deen, and J. H. Havens , "Rational Analysis of


K entucky Flexible Pavement Design Criterion," Research Report 270,
Division of Research, Kentucky Department of Highways , Lexing ton,
K Y , November 1968.

4.

R. c. Deen, H. F. Southgate , and J. H. Havens , "Structural Analys is


of Bituminous Concrete Pavement s , " Research Report 305 , Division of
Research, Kentucky Department of Highway s , Lexington, KY, May 1971.

5.

J. 11. Havens , R. c. Deen, and II, F. Southgat e , "Design Guide for


B ituminous Concrete Pavement Structures , " Research Report UKTRP-811 7 , K e n t u c ky Tr a n s p o r t a t i on Re s e a r c h P r o g r am , C o l l e g e o f
Engineering , University of Kentucky, Lexington, KY, August 1981.

6.

J. Michelow, "Analysis of Stresses and Di splacements in an rr-Layered


Elas tic System under a Load Uniformily Distributed on a Circular
Area," unpub lished, California Research Corporation, Chevron Oil
C ompany, Richmond, CA, September 2 4 , 1963.

7.

AASIIO Road Test: S t. Louis Conference P roceeding s ,


D.C. , Special Report 7 3 , Highway Research Board , 1962.

8.

J. K. Mitchell and c. K. Shen , "Soil-Cement Properties Determined by


R e p e a t e d L o a ding in Relat ion to B a s e s f o r F l e x i b l e Pavement s , "
P roceedings , Second International Conferenc e , Structural Design of
A sphalt Pavement s , University of Michigan, Ann Arbor, MI, 1968.

9.

G. M. D o rmon , and c. T. Metcal f , " D e s i gn Curves for F l e x i b l e


Pavements Based on Layered Systems Theory , " Record No. 7 1 , Highway
Research Board, Washington, D. C. , 1965.

10.

R. c. Deen, H. F. Southgate, and J. G. Mayes , "The Effect of Truck


D e s i g n on P a v e m e n t P e r f o rm a nc e , " P r o c e e d i n g s , V o l 4 9 , T h e
A s sociation of Asphalt Paving Technologis t s , St. Pau l , MN, 1980.

11.

H. F . S o u t hg a t e , R. c. D e e n , D . 11. C a i n , a n d J. G. M a y e s ,
"Modification to Chevron N-Layer Computer Program ," Research Report
UKTRP-87-28 , Kentucky Transportation Research Program , College of
Engineering , University of Kentucky , Lexington, KY , Oc tober 1987.

12.

AASIIO Road Test: Report 2


Materials and Construction,
Research Board, Washington, D. c., 1962.

21

Laboratory
of Highway
Experiment
1949.

Washington,

Highway

13.

H . F. Sou thgate , R . c. Deen , and J, G . Maye s , "S train Energy


A n a l y s i s of Pavement De s i gns f o r Heavy Trucks , " Re cord 9 4 9 ,
Transportation Research Board, Washington, D . c., 1982 .

14.

H . F . Southgate and R . c. Deen, "Effects of Load Distributions and


Axle and Tire Configurations on Pavement Fatigue , " P roceeding s ,
S ix t h In ternat ional Conference , S t ru c t u r a l De s ign of Asphalt
P a vemen t s , Unive r s i ty of Michi g a n , Ann Arb o r , MI , 1 9 8 7 ; a l s o
P roceeding s , Second National Weigh- in-Motion Conference , Vo l One ,
A t lanta, GA , 1985 .

15.

AASHO Road T e s t: Repo r t 5--P avement R e s e a r c h , H i ghway Re search


Board , Washington, D . c., 1962 .

16.

1 9 7 2 AASHTO I nterim G u i de f o r D e s ig n o f P avement S t ructure s ,


American Association of State Highway and Transportation Official s ,
Washington, D . C . , 1972 (1982 Printing) .

17.

Guide for Design of P avement S tructure s , American Association of


S tate Highway and Transportation Official s , Washington, D . c. , 1986 .

18.

H . F . Southgate , "Thicknes s Design Curves for Asphaltic Concrete on


a 4-inch Layer of Dense-Graded Aggregate , or on 6 , 9 , or 12 Inches
o f S t a b i lized So i l , or for Maximum U t i l i z a t ion of Dense-Graded
Aggregate ," Research Report UKTRP-86-14 , Kentucky Transportation
Research Program , College of Engineering , University of Kentucky,
Lexington, KY, May 1986 .

19.

I . S . S o ko lniko f f , Mathem a t i c a l T heory o f E la s t i c i ty , Second


Edition, McGraw-Hill Book Company , New York, 1956 .

20.

Thickness De sign
Asphalt Pavements for Highways and Stree t s ,
Manual Series No . 1 (MS-1 ) , The Asphalt Institute, College Park, MD,
Sep tember 198 1 .

21.

Shell 1 9 6 3 Design Charts for Flexible Pavement s , Shell International


Pe troleum Company Limite d , London, UK , 1963; also , Addendum to the
Shell Pavement Design Manual, London, UK , 1985 .

22.

H . F. Southgate and R . c. Deen, "Distributions of Strain Components


a nd Wo rk within Flexible Pavement S t ru c ture s , " Re search Rep o r t
UKTRP-86-2 1 , Kentucky Transportation Research Program , College of
Engineering, University of Kentucky, Lexington, KY, September 1986.

23.

K . Atkinson, Elementary Numerical Analy s i s , John Wiley


York, NY, 1985 .

24.

H . F. Southgate , D . c. Newberry, Jr . , R. c. Deen , and J, H . Havens ,


" Re su r f ac in g , Re s t o ra t io n , and Reha b i l i t a t i on of I n t e r s t a t e
H ighways: Cri teria and Logic Used t o Determine January 3 , 1 9 7 7 Needs
and E s t im a t e s of Cos t s , " Re search Rep o r t
4 7 5 , D i v i s ion of
Research, Kentucky Department of Highways , Lexington , KY, July 1977,

22

&

Sons , New

2 5.

J. M . B. de Sou s a , "Dynam i c Proper t i e s of Pavement Ma t er i a l s , "


Dissertation Series UCB-ITS-DS-86-3 , Institute of Transportation
S tudies , University of California, Berkeley, December 1986 .

2 6.

H . F. S o u t h g a t e and R. c. D e e n , "Thickn e s s D e s i gn Curves for


Portland Cement Concrete Pavements , " Research Report UKTRP-84-3,
K entucky Transportation Research Program , College of Engineering ,
Univers ity of Kentucky, Lexington , KY, February 1984 .

23

LIST OF FIGURES
F igure

l.

Relationship between Tensile Strain at Bottom of


A sp ha l t i c Conc r e t e and Rep e t i t i ons of 18-kip
Single Axleload.

F igure 2.

Relationship between Vert ical Compressive Strain


a t Top of Subgrade and Rep e t i t ions of 18-kip
Single Axleload.

F i gure 3.

Adjustment of Design
Thickness for Rut ting as a
Function of Repetitions of 18-kip EAL .

F igure

4.

Work at
the Bottom of Asphaltic Concrete versus
Location along Axis of Axle.

F igure

5.

Re l a t i onship
b e tween W o r k S t rain and Tens i l e
S train a t Bottom
of
Asphaltic Concrete.

F igure

6.

Relationship
between Work Strain
and Vertical
Compressive Strain at Top of Subgrade .

F igure

7.

Relationship
between
Work S train at Bottom of
Asphaltic Concrete and Repetitions of 18-kip EAL .

F igure

8.

Re l a t i o n s h i p b e twe e n W o r k S t r a i n a t T o p o f
Subgrade and Repetitions of 18-kip EAL.

F igure

9.

Estimated Annual 18-kip EAL versus Calendar Year.

F igure 10.

Pavement
D e s ign P r o cedure
Control or Strain Control .

F igure 11.

Work S t r a i n at Top o f S u b g r ade vs 18-kip


Criterion for Three Ranges of CBRs .

F igure 12.

Thickness
Design Curves for Pavement Structures
Comprised of 33 Percent Asphaltic Concrete and 67
Percent Crushed Stone Aggregate Bas e .

F igure 13.

Thicknes s Design Curves for Pavement Structures


Comprised of 50 Percent Asphaltic Concrete and 50
Percent Crushed Stone Aggregate Bas e .

F igure 14.

Thicknes s Design Curves


for Pavement Structures
Comprised of 75 Percent Asphaltic Concrete and 2 5
Percent Crushed S tone Aggregate Base.

F i gure 15 .

Thicknes s Design Curves


for Pavement Structures
C omp r i s e d of A s p h a l t i c C o n c r e t e on 4 Inches o f
Crushed Stone Aggregate Base.

24

for E i t h e r S t r e s s

EAL

Figure 16.

Thickness
Design Curves
for Asphaltic Concrete
on Broken and S e a t e d Po r t land Cement Conc r e t e
Having a Young's Modulus o f 2 5 ksi .

Figure 17.

Thickness
Design
Curves for Asphaltic Concrete
on Broken and S e a t e d Po r t land Cement Conc r e t e
Having a Young's Modulus o f 100 ksi .

Figure 18.

Thickness
Design
Curves for Asphaltic Concrete
on Broken and S e a t e d Po r t l and Cement Conc r e t e
Having a Young's Modulus o f 200 ksi .

Figure 19.

Thickn e s s Design Curves for Port land Cement


Concrete on Asphaltic Concrete Having a Young's
Modulus of Elasticity of 200 ksi.

25

TABLE 1 .

REGRESSION COEFFICIENTS TO CALCULATE DAMAGE


FACTORS FOR VARIOUS AXLE CONFIGURATIONS

log(Damage Factor)

a + b( log(Load ) ) + c ( log( load) ) 2

================================================================

COEFFICIENTS
AXLE
CONFIGURATION

Two-Tired Single
Front Axle

-3.540112

2 . 728860

o. 289 133

F our-Tired Single
Rear Axle

-3 .439501

0.423747

1. 846657

E ight-Tired
Tandem Axle

-2.979479

-1 . 26 5 144

2. 007989

Twelve-Tired
Tridem Axle

-2 . 740987

-1. 873428

1. 964442

S ixteen-T ired
Quad Axle

-2. 589482

-2. 224981

1. 923512

26

TABLE 2 .

VALUES FOR COEFFICIENTS TO CALCULATE OVERLAY


THICICNESS OF ASPHALTIC CONCRETE ON BROKEN PCC

==========================--=====--===========
===-======
--

OLT
a + b*log(EAL) + c*( log(EAL) ) 2
where OLT = overlay thickness and
=

a , b, and c

regression coefficients

f(CBR)
d + e*log(CBR) + f*(log(CBR))2
=

MODULUS OF
BROKEN PCC
(ksi)

THICKNESS OF
BROKEN PCC
( inches)

25

a
b
c

-39 . 55341825
12 . 27682107
-{) . 612631087

7 . 559521621
-2 . 774707235
0 . 210962531

-o. 181394144
0 . 0884178858
-o. 0077519796

10

a
b
c

-70. 28424248
20. 54802239
-1 . 169270477

21 . 272191482
-6. 5107715697
0 . 4633902917

-1 . 578324527
0 . 46901845188
-o.o335234984

a
b
c

-51 . 97485309
15 . 181766663
-{) .819240853

10. 138201032
-3 . 5133660621
0 . 2696910591

-{) . 494600481
0 . 1729875084
-{). 0137252471

10

a
b
c

-72 . 03822884
20. 439819522
-1 . 171268401

15 . 174210016
-4 . 8767616024
0. 3613069921

-o.837114874
0 . 2654316355
-{). 0199286739

a
b
c

-58. 32810531
16 . 282821001
-o.898751519

11 . 980305722
-3. 9776753781
0. 3016042334

-o . 661994556
0 . 2156855133
-o. 0165359824

10

a
b
c

13 . 616843885
-5 . 016295968
0 . 6511910044

-{) . 538274742
-o. 3930013094
0.0458919219

0 . 0051893437
0 . 0298301792
-{). 0036225523

100

200

COEFFICIENT
d

COEFFICIENT

27

10

21
6(

18pc

.....
......

... ...

r-.. ...

...

r-.. ...

......

.....

......
1--...

-...
..._

.....
....

--o

......

-...

:"""

-...

...

-5
10
3
10
F i gure

1.

4
10
Re l a t i onsh i p
Aspha l t i c

10
1 8-KIP EAL

10

between

7
10

Tens i l e Stra i n

Concrete and Repet i t i o ns o f

A x l e l oad .

28

at

8
10
Bottom

1 8-k i p

of

S i ng l e

....

....

....

"

....

'\
'!\

10

-4
10

10

10

10

10

10

18-KIP EAL
F i gure

2.

R e l a t i onsh i p between Ver t i c a l Compressive


Str a i n
at
Top
of
Subgrade and Repet i t i ons
of
1 8-k i p
S i ng l e
Ax l e l oad .

29

1 00

II

80

60

40

20

0
10

10

10

10

10

10

1 8 KIP EAL
F i gure 3 .

Ad j us tment

of

Des i g n Th i c: l:ness

Func: t i on of Repet i t i ons of

30

for

18-k i p EAL .

R u tt i.ng

as

il

4!100 LB

4!100 LB

c:L c:L
l

I
.c.
u

.5

AXLE

SUM OF WORK
THROUGH 8 INCHES
O
---- F ASPHALT I C
CONC R E T E AND 8
INCHES OF SUBGRADE
....____ 8 1NCHES OF

--

ASPHALTIC CONCRET
ONLY

8 INCHES OF SUB
GRADE ONLY

30

20

50

40

60

70

80

Y D i stanc e , Inches
F i gure 4 .

Work

at

the Bo t tom

of

Aspha l t i c

Loca t ion a l ong Ax i s o f A x l e .

31

Concrete

versus

_,

'!l'

IV

/
/

/
10 "4

TENSILE STRAIN AT BOTTOM OF ASPHALT


F i gure

5.

10

Re l a t i onh i p
between Wor k S tr a i n and Tens i l e
a t Bo t tom o f A5pha l t i c Concrete .

32

-3

Str a i n

w
0
<(
a:
(!)
m
:::>
C/)
LL

2
1 0..

1/

3
1 0-

1 4 Jlliflr

a..

lei:

<

4
1 0-

a:
0

-5 v

1 0 -5
10

4
1 0VERTICAL STRAIN AT TOP OF SUBGRADE

Fi gure

6.

R e l a t i onsh i p

between

Work

Str a i n

Comp r es s i ve Str a i n a t Top o f Subgrade .

33

and

Ver t i c a l

,1959

NS

()

""' I
.....

'

191 IKtf
10

IGNS I
10

r-

10

10

10

1 8-KIP EAL
Fi gure

7.

Re l a t i onsh i p
Aspha l t i c

between

Conc r e t e

Work

S tr a i n

and Repet i t i ons

34

of

at

Bottom

1 8-k i p EAL .

of

N.

t.

1 95

m;

DE Sl l.jI"

m. 1 ,K'l II

10

...,

,_
I

IG s
10

10

1"":00

10

18-KIP EAL
Fi gure

B.

R e l a t i onsh i p be tween Wor k Str a i n at


and Repe t i t i ons of 1 8- k i p EAL.

35

Top

of

Subgrade

:.-

,.

'

;...-

103

60

62

l.-'

64

/.

,.,..
loo""'

,.,..

..... ,...

--

_,

66

'

68

70

72

74

76

78

80

82

84

86

CALENDAR YEAR
Fi gure

9.

Est i mated

Annua l

1 8-k i p

36

EAL versus C a l endar

Year .

10

10

10
10
18-KIP EAL

WORK STRESS VS WORK STRAIN

FATIGUE AT BOTTOM OF
ASPHALTIC CONCRETE
At:, MODUWS
480 KSI

10

10

10

106

10
cb

... 10

10
10

F i gure

10.

Pavement
or

Des i g n P r o c edure f o r

Str a i n Cont r o l .

37

-4

10
10
10
WORK STRAIN

E i ther

S t r ess

Con t r o l

10

-2

(!)
LL

.....

r--im

>7 v
106
1 8-KIP . EAL
F i gure

11.

Wor k

Str a i n

Cr i ter i on for

at

Top

of

Sub g r ad e

Three Ranges of CBRs .

38

vs

1 8-k i p

EAL

40

f3
25z

38
36
34

. ,""'

1"\\[J

32

30

...-,,
;,r"J[

' .

28

i=

26

,-,.. n /IV/0
,_

,,,

0\

1\.P

..

I l

24

....

22

20

18

16
14
12

10

"'
v
/ 1/
r7
.L; :..17r::;
4

/
/
1-L- 2

...

_/

...

...Iii'

.....
1./
'".:.;;
r.,...o"

'"

""'
""'

s:
e:;

-7
/

3 -7"
/
/
,
/
....

....

7
""

_/
I;
li"'
_/ _.. ""'
" 1.-v
7 17' "' '""
_/' 1/
.... _/' v.... 1:0...... ....

_/

_ii

""

_/ 5 17

/ r,r
/1 6 'I'

'7,v

..

_/ 1/ IV
_/ _/ ""'
_/' _.. ..
/.-- "/

_/[h 1':....-"'./. K 1':;..-"'_/' "\..

10

1 8-KIP EAL
F i gure

12.

Th i c kness
Comp r i sed

Des i gn
of

Curves

for

Pavement

Structures

33 Percent Aspha l t i c Concrete

Percent Crushed S_tone Aggregate Base .

39

and

67

CBR
36
34

ffi

32

:::c
0

30
28

26
24

22
20
18

I
"" hMl'J I I

16

ffi

14

12

lfJ

""'

'-

'"''"' -.

.;'

./

./
/

,.

!:;;;

""'

.;'

li'
-:;..

.;'

GO'

,..,.

....

....

/ I;' I;'

11

/ I;'
V :/ I;'
._, ""- !:"
":;..

""'
/ .,. ......
/ ....

10 8
10

5
10

10

10

10

1 8-KIP EAL
F i gure

13.

T h i c k ness
Compr i sed
F'er c:ent

Des i g n
of

Crushed

50

Curves
F'er c:ent

fcr

Pavement

Aspha l t i c:

S t one A g g r e g a t e Base .

40

Struc tures

Conc r e t e

and 50

CBR
26

f3
:I:

PL l HI

24

22

20
en

f3
z

1-

12

f3
c

I'

10

8
10

I;'

I/
I;'
/ / I;'
/ /l/
V"/....::
l'_h

/ l/
v v l/
/ I/:

//

. J::

/
V/::...
.-:
/"h
/

"'

Q 16
14

..........

_"l

nut Jl

18

FE

-2

ss I = ]!"

11

10

10

10

10

1 8-KIP EAL
F i gure

14.

T h i c kness
Comp r i sed
Pel-cent

Des i g n Cul-ves
cf

Crushed

for

Pavement

S t r uc tures

75 Percent Aspha 1 t i c Ccncr e t e


S t o ne Aggregate B ase .

41

and

25

24

4-"":..
Rl

'A

"''"

liAnr II
cR

lEI

1 1 Jr

:J I\

v ..;I-'
v....

1-".,.

./ /

I/

v
/ ILI'
/

CBR
-2
-3
r-5

l -7

-1 1

810

:,....:

'I

10

1 8-KIP EAL
F i gure

15.

Th i c: kness

Des i g n

Comp r i sed

of

Curves

Aspha l t i c:

for

Crushed Stone A g g r e g a te Base .

42

Pavement

Conc:rete

on

S t r Lic: tures
Inc:hes

of

18
16
14
I 12

N Pi

PCC MC o

su

II

1 lx

10

....

1.......
!/

/ v ,.

//.......... iii'

Ki

25 Sl
..us

10

16
UJ

"

PCc
suBG M
HiI

14

,.

/ /
v/ :.....

1I1C.I 25

16.

EAL

CBR

be

KS
v

BR

Th i c kness

1.;
...

/ 1./
"/ 1/ ""
v 1/'!.; I.v....""
.. ....
..
y
!::"

10

-9
-1 1

/v

4 5
10

VA....

I/
1::7 v

1/ 1.lh r., 1-.#.)...-

bKEfl

12

F i gure

I/
v
v ., I/
v .,v I;
v..,...._,.. ......1.-'
V
/v....

v
/

/ v :...-

-3

,.
/
......
...... ..,

18-KIP

18

,.
"""

i:

Si

NC1-1

9 -

CBR

"""

......

v 1/
/v

3
....
1.-'

1.-'

....

/ v l.-'
/ ......
V/

!::?'

10

18-KIP EAl..

Des i g n Curves

f o r Aspha l t i c

-5
-7
9

-1 1

10

Concrete

Broken
and Seated Pcr t l and Cement Concrete Having
Young ' s Mod u l us o f E5 k s i .

43

on
a

CBR
16

9 - I" c 4

PCC M 'r
SUB 3R I[) =
-

"'

ws

!3R

5(

/ "' ..-

v .-v/ /17 .--

7
-9

-1 1

V.h v
f9
w

v/ jI/ .;;V_/v,

v
/ v/
I'
/v/
"
/ 0 /
V/v.: .I

100 KSI

2
3

KEN p be

//

/ /.I
6

10

10

10

10

18-KIP EAL

16

--

1 0 IN<
PCC

su_

M< D Ll
u

EN c

00 Kl l

liS - 1 50b

Cl

v ""
I/ ,;

'/!/ ::;
.
1/ ....,
1/J..- .-

II'"'

K:;BR

/
-./I/!.-

./

'...--'

%
%;;

"/

'

18-KIP EAL
Figure

17.

T h i c kness
Des i gn
Curves for Asph a l t i c
Concrete
on
a
Bro ken
and Seated P o r t l and Cement Ccncr e t e Hav i ng
Young ' s Modu l us of

1 00 k s i .
44

PC

uhr
-

EN

1M
I

9 ln "

. Ill

200
ln 1 L

11 0

F c:c

,KSI

ICBR

-2

. ...r
7
9

V' ""

./

v
./"I-" ......
v
V/ / ......
V//v ""

4 6
10

18

16

PCC

14

10

10

l '"

Ul 1n

18-KIP

200
In 1 1

II' l

EAL

18.

""'

Th i c kness
Brc k en

and

10

10

<SI

10

10

CBR

6
0

_......
/ /""
.....
./ / .......

18-KIP

Des i g n
Seated

PO :;

If

F i gure

/ /L,...
""'/v ...... ,.....
h
//.t:'l ::::
v

EAL

....
/ ......
/ / /./
"'/ ,....,...... .. //'/
/..,- %
v../.:: ,.
0

Young ' s Mod u l us o f 200

45

ksi .

-7

l- 9
.:....1 1

7
10

10

Curves fcr Asph a l t i c:


Por t l and

-5

Cement

Cone: rete
cn
a
Concrete Hav i ng

5-INCI

14

,_.,
ft

ov

....

/
/

1-4

1/

f
10
111-KP EAI.

10

18

1
10

CBR
6.5-INCH
IC

CRETE

Sl

-6
I-'

1/

/
//
V/
v

18

CBR

IC

It;.

ll
-6
8
-

/
:,....-'

/ 1.;'
/
/

1
10

F i gure

19.

T h i c kness Des i gn CLirves for

Po r t l and Cement

on
Aspha l t i c
Concrete Havi ng
E l ast i c i ty o f 200 k s i .

46

a Young ' s

Concrete

Mod u l us

of