FINITE ELEMENT ANALYSIS OF

ARIZONA'S THREE-LAYER OVERLAY SYSTEM OF RIGID PAVEMENTS

TO PREVENT REFLECTIVE CRACKING

BY

NAN JIM CHEN, P.E., Ph.D.

JOSEPH A. DI VITO, P.E.

GENE R. MORRIS, P.E.

OF

ARIZONA TRANSPORTATION RESEARCH CENTER

ARIZONA DEPARTMENT OF TRANSPORTATION

FOR PRESENTATION AT THE 1982 ANNUAL MEETING OF

ASSOCIATION OF ASPHALT PAVING TECHNOLOGISTS

FEBRUARY, 1982

ABSTRACT

Asphalt-rubber has been used as a Stress Absorbing Membrane Interlayer

(SAMI) for overlay of flexible and rigid pavement for more than ten years in

Arizona to retard reflective cracking with satisfactory performance.

This paper discusses several field experiments of asphalt-rubber

overlays on rigid pavement with a three-layer system which consists of an

asphalt concrete leveling course, a SAMI and a wearing course. This

three-layer overlay system has provided a very smooth ride on an interstate

highway for two years without a single visible reflective crack. The

Average Daily Traffic (ADT) is approximately 100,000 with nine percent

commercial vehicles.

Analytical studies utilizing the Finite Element Method (FEM) indicate this

3 layer system is far superior to simple overlays without interlayers.

St~esses in an overlay above an existing joint are reduced signficantly (up

to 95% or more) due to the presence of a SAMI, and overlay life is extended.

As a result of the analysis, the three-layer system is recommended for

overlay of a rigid pavement, especially when the rideability of the existing

pavement needs to be improved. The leveling couse not only improves rideability

of the final overlay, but also the performance of SAMI.

I NTRODUCTI ON

Reflection cracking has been defined as the cracking of a resurface

or overlay pavement above underlying cracks or joints. Reflection cracks

may be induced by environmental conditions or traffic loads or a combina­

tion of the two. It is a worldwide problem occurring in virtually all

types of pavements unless steps are taken to prevent it. The problem is

especially critical in overlays of rigid pavements and this paper is

directed towards that problem.

Several treatments have been used to reduce reflective cracking in

asphalt concrete overlays with varying degrees of success. Generally

they fall into one of the following categories:

1. Treatments to existing concrete pavements (sucn as breaking into

small sections) prior to any rehabilitative work,

2. Stress or strain relieving interlayers,

3. Cushion courses or special overlay treatments,

4. Increased overlay thickness.

Arizona's three-layer asphalt-rubber overlay system has been used on

several experimental projects where Portland Cement Concrete Pavements

(PCCP) were overlaid. The asphalt-rubber materials have been discussed tho­

roughly in other reports (1 through 9) and will not be discussed in this paper.

The 3-layer system incorporates a thin flexible overlay placed in two

layers with a low modulus asphalt-rubber interlayer between. It consists

of an open-graded mix 1/2" to 3/4" thick placed as a leveling course

directly on the cracked, rigid pavement followed by an asphalt-rubber

flush-in with cover material. The final course is another open-graded

mix 1/2" to 3/4" thick. The result is a thin, flexible overlay that

resists reflective cracking for a much longer time than conventional

overlays. Figure 1 is a simplified cross-section of this system.

-1-

ASPHALT - RUBBER'FLtjSHi!lN i"\lO.rvu'(1 !'--No;;..

BLOTTED WITH COVER MATERIAL

!~'ACFC OR LEVELING COURSE o ~ (:;.1>..\ G. fl I'. () I::: u \l <'l-.;:: - r.t l "-

PORTLAND CEMENT CONCRETE PAVEMENT

BASE MATERIAL

FII'URE 1/.1 TYPICA(rnoss S~G1'iON OF A..RH\lNA'S/TtI~E-L.A.Y,f;R_ V/ \.AsPHALT-~EIt'OVml,AY....sySTErl~ ~ .

-2-

As with most of the previously mentioned treatments, trial and error

applications preceded mathematical modeling. The same is true for asphalt­

rubber and the need currently exists to prepare analytical studies to

verify field performance. The intent of this research was to establish

an appropriate mathematical model to verify empirical observations. The

analytical study consisted of a finite element analysis of stresses and

strains with different layer configurations and is explained in detail

in Part II. Part I will address the design and performance of the various

experimental projects that received this specialized overlay treatment.

PART I - PROJECT PERFORMANCE

All pavements overlaid with the three-layer system were plain,

undowe1ed pavements exhibiting varying degrees of structural cracking

and load failure. Three of the projects are in the Flagstaff area at

an approximate elevation of 7000 feet, while the three other projects

are in the Phoenix valley area. Comparison of the two climate extremes

has been informative. Flagstaff winter temperatures have been as low as

-30oF while Phoenix summer temperatures can be as high as +1150 F.

Table 1 provides a description of each of the projects and the dates

overlaid with the three-layer system. A detailed description of each

project follows.

1. 1-17 at Bell Road - This section of Interstate 17 in north

Phoenix was not originally scheduled to receive the three-layer

overlay system. In June, 1974, an open-graded Asphalt Concrete

Friction Course (ACFC) was placed and asphalt-rubber was flushed

into the mat with cover material applied. The treatment served

very well until another adjacent resurfacing project in 1980

called for an ACFC to be placed over this experimental section,

-3-

TABLE 1. SUMMARY OF FIELD PROJECTS WITH THE THREE-LAYER ASPHALT-RUBBER OVERLAY SYSTEM

Three-Layer Project System Placement Length and Width Traffic ADT and %

Date Commercial Vehicles

l. Southbound 1-17 at Bell Road Summer, 1974 1500 feet 38' wide 50,000+, 9%

(Phoenix area)

2. Eastbound and Westbound 1-40 Summer, 1974 4 miles 38' wide 6,000, 20%

Riordan to 1-17 (Flagstaff area)

3. Westbound 1-40 at Butler T.1. Summer, 1974 3000 feet 38' wide 7,500,28% , ..,. (Flagstaff area) ,

4. Mesa Country Club Drive Fall, 1976 4000 feet 68' wide 25,000, 5%

S.R. 87 (Phoenix area)

5. Westbound 1-40 at East Fall, 1977 1000 feet 38' wide 7,500, 28%

Flagstaff T.1. (Flagstaff area)

6. Southbound 1-17 at Durango Fall, 1979 1500 feet 48' wide 95,000+, 9%

Curve (Phoenix area)

making it a three-layer overlay. So far, it has not exhibited any

reflective cracks. Before 1980, the original treatment controlled

90% of the reflection of all joints and cracks.

2. 1-40, Riordan to 1-17 - A three-layer system was placed on this

four-mile divided interstate highway project in 1974 in an

attempt to control severe structural failure related to frost

heave and thaw weakening. Cracking and joint faulting had

progressed to the point where the highway had to be inspected

daily for broken chunks of concrete. Structural failure (cracking)

had occurred in a very large percentage of the travel lane. A

typical view of cracking and skewed joints prior to overlay is

shown in Figure 2. In addition, severe faulting was evident

throughout the section. Prior to overlay, the integrity of the

structural section was restored with sub-slab grouting. Struc­

tural cracks were grouted and sealed with asphalt-rubber as was

the longitudinal edge joint. The PCCP was originally built 9"

thick, unreinforced, with centerline tie bars. Shoulders were

constructed of asphalt concrete. Since the three-layer layer has

been in place, approximately 25% of the travel lane transverse

joints in the PCCP have reflected through as cracks in the overlay.

The cracks that have reflected are narrow, have not spal1ed and

have not required any major maintenance. There is little

reflective cracking in the passing lane. Approximately 10%

of the longitudinal edge joint has cracked through the overlay.

However, neither the tied centerline joint nor the structural

cracks have appeared on the surface. All surface cracks have

remained narrow and have not spal1ed, very likely due to the

-5-

TYPICAL VIEW OF CRACKING AND SKEWED JOINTS PRIOR TO OVERLAY

SHOULDER

131 lSi 111 12'

PASSIN~ LANE - - - -

12' TRAVEL LANE , ~ -r- ... ..

} ..............

1

10' SHOULDER

FIGURE 2. TYPICAL VWI OF CRACKING AND SKEl-IED JOINTS PRIOR TO OVERLI\Y

-6-

1.0

extended flexibility of the asphalt-rubber at cold temperatures.

While not completely controlling joints and cracks, the three­

layer system has performed well in rehabilitating this section of

highway that really should have been reconstructed.

3. Butler Traffic Interchange, 1-40 - Except for sub-slab grouting,

this section was exactly like the Riordan experiment and was

placed at the same time, summer 1974. Performance has also been

similar to the Riordan project indicating that magnitude of

vertical deflection may not be critical.

4. Mesa Underpass on S.R. 87 - This experiment was originally

established to test different application rates of the asphalt­

rubber flush-in on the three-layer overlay system. It was

constructed in the fall 1976. The lower ACFC layer was placed

using two different gradations to test the effect of voids in

the mat versus reflective crack prevention. After close

observation for four years, the entire project had no reflective

cracks. Since no changes were apparent in traffic patterns or

loading, or any other factor, it is believed that asphalt­

rubber stiffness properties may have changed or the life of

fatigue crack propagation has been reached. Some surface

instability was observed in the areas where 0.8 to 1.2 gallons

per-square-yard (gpsyd) of asphalt-rubber was placed. Applica­

tion rate was originally specified as 0.4, 0.6 and 0.8 but some

areas received as much as 1.2 gallons per-square-yard. At date

of this report, the surface has approximately 10% reflective

cracking. Again, these cracks are hairline and have required no

maintenance. Application rate (in the range of 0.4 to 0.8 gpsyd)

-7-

seems to have no effect on reflective cracking prevention nor

did the variable gradation of the first-layer. Figures 3 and

4 were typical pavement conditions before overlay and five years

after overlay.

5. East Flagstaff, 1-40 - The three-layer overlay system was

placed as one of six different overlay test sections and is the

only one of the six to perform well. All others have already

been dismissed as unacceptable. The three-layer system has

prevented 90% of all cracks and joints from reflecting through

to the surface. It was placed in late summer, 1977 and to date,

has shown very good results.

6. Durango Curve on 1-17 - The overlay of this urban Phoenix

freeway section of rigid pavement took place in November, 1979.

The three-layer system was to be compared with the very expensive

alternative of grinding and grooving to restore ride and skid

number. At the time of design, it was estimated to be 1/3 the

cost of grinding. In addition, it could be constructed in a

much shorter time period. After two years of service there are

no reflective cracks of joints whatsoever. (An 18" transverse

hairline crack was observed during January, 1982 inspection.)

The only surface defects noticeable are two small areas (less

than three square yards each) exhibiting rutting and shoving.

They are the result of inadvertent excessive applications of

asphalt-rubber during construction, a problem not attributed to

design nor one that should occur during normal project construc­

tion. Figures 5 and 6 were pavement conditions before overlay

and two years after overlay.

-8-

FIGURE 3. MESA UNDERPASS ON S.R. 87 BEFORE OVERLAY IN 1976

rIGURE 4. MESA UNDERPASS ON S.R. 87 FIVE YEARS AFTER OVERLAY (JANUARY, 1982)

-9-

FIGURE 5. DURANGO CURVE ON 1-17 BEFORE OVERLAY IN 1979

FIGURE 5. DURANGO CURVE ON 1-17 TWO YEARS AFTER OVERLAY (JANUARY, 1982)

-10-

The three projects in the Flagstaff vicinity have all demonstrated

similar performance. Some reflective cracking is evident but it is essen­

tially confined to the transverse joints in the travel lane. This indicates

reflective cracking is related to the combination of thermal and load

associated stresses. The fact that structural cracks did not reflect

through demonstrates loading alone is insufficient to cause reflective

cracking of the three-layer overlay system.

Also, the passing lane of the three projects is virtually crack­

free which indicates thermal stresses alone are insufficient to cause

cracking. The tied centerline joint has not cracked through, further

supporting these conclusions.

The Phoenix area experiments demonstrate the three-layer overlay

system will prevent reflective cracking of joints and cracks for up to

four years in mild climatic environments. After this length of

service, approximately 10% cracking is apparent. The Durango curve

project has proven the system is very stable under the extreme conditions

of high traffic volume with heavy loadings combined with extreme high

summer temperatures.

In all cases, the system has maintained good surface friction

qualities (high skid numbers) and good ride quality. The system is far

less expensive than grinding to smooth PCCP and much less disruptive

to traffic during construction.

-11-

PART II - ANALYTICAL STUDIES

To explore the basic reason why asphalt-rubber prevents reflective

cracking of rigid pavement and to determine the structural adequacy of

the system, an analysis was conducted of the theoretical behavior of the

overlay structure. Several Finite Element Method (FEM) computer programs

were utilized for the analysis of stresses and strains with different

layer configurations. The primary computer program used for this study

was a slightly modified static analysis program for solid structures -

SOLID SAP by Wilson (10). The slight modification of this program was

in the calculation of the effective stresses (11), which is essentially

the "root mean square" shear stresses, and defined by using the normal

and shear stresses in an orthogonal Cartesian coordinate system as

They were considered to be a realistic determinant for fracture (cracking)

under the triaxial stress state existing in the overlay pavement. In

order to reduce the cost of computer time, a linear elastic plane strain

analysis was assumed 11ith 685 nodes and 620 elements. Among the field

experiments, there are several overlay configurations with different

existing conditions. The finite element analysis 11as simplified by

studying a nine-inch concrete pavement (PCep) with a O.3-inch jOint

overlayed by different thicknesses of asphalt concrete and a SAMI. The

general configuration of the pavement overlay system is shown in Figure?

Seven different overlay designs were analyzed ranging from a very thin

(1 5/8") three-layer design to a seven-inch overlay. Information on input

thickness and layer arrangement is given in Table 2. Mechanical properties

-12-

I ..... W I

TL

15CM 15CM +TL iL SURFACE (S-) (611

)

COURSE 0.69 MPa (100 PSI) 'V'v

MOVING TRAFFIC

\ ~ /SAMI

'AI'

~LEVELING COURSE ...

• ~O. 76 CM JOINT 22.8 CM PCeP-"",

(0.3 11 ) (9111

)

"''/~'' ... ~~, ''<1/ SUSGRAOE

lOADING HI G1 FI EI 0 1 C' SI AI CASE A S C o E F G H I

TL (CM) 40 35 30 25 20 15 10 5 o -5 -10 -15 -20 -2.5"30 -35-40

UN) 16 14 '2 10 8 6 4 2 o .. 2 .. 4 ·6 -8 -to -12 -14 -16

FIGURE 7. GENERAL CONFIGURATION OF THE OVERLAY SYSTEr~ ON RIGID PAVEMENT

I I-' ... I

OVERLAY DESIGN THICKNESS

CASE I

CASE 2

CASE :3

CASE 4

CASE 5

CASE 6

CASE 7

SURFACE SAMI _EVE LING COURSE COURSE CM (IN.) CM UN.) CM (IN.)

I

0.95 :0.315 ,

1.59 10.625 1.59 '0.625 I I , I I 3.18 11.25 0.95 ,0.315 0.0 K>.O

I

I I I

4.13 ~ 1.625 0.0 10.0 0.0 10.0 ,

I I i 9.21 0.95 '0.315 0.0 13 .625 1°.0 I

I I

:0.0 10.16 :4.0 0.0 10.0 0.0 I

I I I 16.83 16.625 0.95 ,0.315 0.0 ,0.0

I

11.18 :1.0 0.0 I I 10.0 0.0 10.0

I I I

TABLE 2. ARRANGEr'IENT OF OVERLAY THI CK~,ESS

TOTAL

CM ( IN.)

I

4.13 11.625 I

I 4.13 ,1.625

I

I i

4.13 11.625 I

I

10.16 J4.0

10.16 14.0

I 11.18 11.0

I 11.18 11.0

I

of different pavement layers used for analysis are listed in Table 3.

From the results of very limited tensile and shear tests on viscoelastic

properties of asphalt-rubber, the ranges of creep compliance values can

be roughly determined as shown in Figure 8. The elastic moduli of SAM!

were selected accordingly for finite element analysis. For the study of

moving traffic, 2000 psi was selected and 10 psi chosen for the thermal

study.

The thermal effect on an overlay was studied by using two different

temperature profiles from the surface of the overlay down to the subgrade,

with the higher one used as a stress free temperature. A profile of

temperature differences is given in Figure 9. A 50-degree (oF) tempera­

ture change on the surface of the overlay was used for analysis.

Moving traffic on the overlay was represented by a 12" long, 100 psi

load for the FEM computer analysis. The location of this loading was

moved every two inches from one side of the joint to the other (see

Figure 7) and influence lines of stresses above the existing joint were

plotted and evaluated.

The stress~s which were investigated in this study are located appro­

ximately 0.09" right above the existing joint (center of the lowest layer

of mesh within overlay used for finite element analysis) and those stresses

are not the average stresses in different layers. A higher stress indi­

cates that reflective cracking may start to propagate sooner than that

with a lower stress.

The major findings of this study are as follows:

(a) Three-layer overlay system - The three-layer overlay analyzed,

consisted of a thin (5/8") asphalt concrete leveling course, a

3/8" SAM! and a thin (5/8") asphalt concrete surface wearing

-15-

..... en ,

ELASTIC SHEAR POISSOO MODULUS MODULUS RATIO

E G J.I M Po KSI MPa KSI

I 1 SURFACE AND

2070:300 828 1120 0.3 LEVELING COURSE I I

SAMI i 1

(TRAFFIC) 13.8 12 5.52 10.8 0.35 I i

i I

SAMI 0.027610.004 0.06910.01 0.35

(THERMAL) I I I

I I

11040 :1600 PCCP 27SOO14OOO 0.2 I I

I I

1 SUBGRADE 69 110 27.6 14 0.48

I I

TABLE 3. ELASTIC PROPERTIES OF OVERLAY cmiPON~;m FOR ANALYSIS

THERMAL COEF.

a of

0.0000125

0.0000150

0.0000150

0.0000039 I

0.0000100

10 0 .....

• 0 .....

~ rI)

0 "" ·0 L'.t ...... co

4 co :::>

"" ~

I-

~ -' ~

~ 0 a..

~ V1 ...... -<

~ 0 C": C>

't w u..

en w .... L>

~ 0 Z < - ~ -- -' "(' a..

~ ~O W C5 u

~ 0 a..

:> ~ '" 0 LeI

~ 0:: ..... U ......

...... LL

't 0

{;J

~ '"' TO

z < 0::

"(' ...... ~

CO

w

~ "" OJ :::>

'"' ·0 ~

LL .....

--.... o o .... ~ If) "'" o '0 '0 ·0 '0 _ ..... _- ..... --

-17-

CHANGE OF TEMPERATURE (OF) 10 20 30 40 50

-:2 0 --• z ..... - ..... (.)

10 ~ 0:: :;) fI)

~ 0

20 ...J ..... II)

..... 0 z

..... i:!

30~ c c

15

40

FIGURE 9. THlPERATURE PROFILE USED FOR THERtlAL ANALYSIS

-18-

course. The major advantage of this system over a 2-layer (SAM!

and AC) system is that the leveling course will provide a very

smooth surface for the SAM! to behave as a stress absorbing

layer. Without this leveling course, very often, the asphalt­

rubber will flow into joints or large cracks and a continuous

stress absorbing layer can not be achieved. Results of the analy­

tical study indicate stresses in the leveling course above the

joint are very high but due to the existence of the SAM!,

stresses in the surface course are very low. See Figures 10

and 11. Both shear stress (S12) and effective stress (Seff) are

at a maximum when the simulated traffic loading is completely

on one side of the concrete slab. The very low stress level in

the surface course indicates a very low stress intensity factor

which in turn will provide a longer service life for the overlay

versus reflective cracking.

(b) Overlay Thickness - Overlay thickness from 1 5/8" to 7" were

studied, with and without a SAM!. The typical influence lines of

shear and effective stresses within the overlay are indicated in

Figures 12 and 13. The stresses in the overlay reduced signifi­

cantly because of the existence of the SAM!. For an overlay

without a SAM!, stresses reduce significantly due to increasing

thickness (refer to Figures 14 and 15). However, for an overlay

with a SAM!, stresses were only slightly reduced due to the

increasing overlay thickness (Figure 16). This indicates, from

an economical point of view, a thick overlay may not be justified

for a reflective overlay design when a SAM! is utilized.

-19-

MPQ PSI 800

5

4 600

0.. 400 ..... ..... ~ 2 () « 200 0: I ()

I.IJ > 0 0 m « (/) (/) IJJ 0: t; 0: -« -3 I.IJ :J: (/) -4

-600 •

~ 2 -5

-800

LEVELING COURSE (5/8 11

)

SURFACE COURSE (51 II)

HI G F' E' 0' C' Sl do

f-!A~SrlC ~O~E ~F~G~H~~ TRAFFIC LOADI NG CASE

FIGURE 10. INFLUE~CE LINES OF SHEAR STRESS ABOVE THE CRACK TIP DUE TO f10VING LOAD (THREE-LAYER SYSTEt1)

-20-

I

'" ..... I

MP01

PSI !!: 1800 ...... 12

~ ~ IISOO o

IAJ 10 ~

= 1400

fn m ~ 811200

iii ~ 1000 --...... o ~

LEVELING COURSE {5/8 111

}

.h 0.-.... SURFACE COURSE

SAMI (5/8 11

,

~ ~ .6::V--I&.. LuJ

• O' ~ • i •

X

~ A S C D E F G H I HI G' P E' DD C' SI A'

TRAFFIC LOADING CASE FIGURE 11. INFLUENCE LINES OF EFFECTIVE STRESS ABOVE THE CRACK TIP DUE TO MOVING LOAD

(THREE-LAYER SYSTEM)

I N N I

0.. .... .... t§ c 3 L&.I >-0 m C en en L&.I a:: .... en a:: c L&.I l: en

MPOII PSI 4 600

3\ 400

2l I I 2001

I 5/881 OVERLAY

I I WITH SAMI (IN AC)

/ WITij SAMHIN SAMI)'

o~ I ~ 0==8:$ Q ~, H' G' F' E' 0' c i S' AD 0

I A S C 0 E F G

-'r ~ TRAFFIC LOADING -200 CASE

1-400 -3

i -4r-soo

FIGURE 12. TYPICAL INFLUE~ICE LINES OF SH~AR STR~SS ABOV': HIE C~,1\CK TIP DUE TO nOVI!~G LOAD

f1. MAli PSI I 1= I SIS'" OVERLAY ~ 9 (,)

~ 11200 (,) 8 1&1 > 1

11000 0 m 4 U) 6 U)

1800 1&1 1 0:: N 5 w l;; 1

2L'V· I ,WITH SAMI (IN SAMI) 1&1 > ....

1200l ~/ /WITH SAMI (IN AC) t5 1&1 I I&. I&. 1&1 ot 0

• ABCOEFGH I HI G' F' F OB C' fJ A' X 4

TRAffiC LOADING CASE :I FIGURE 13. TYPICAL I~lFLUErlCE LINES OF EFFECTIVE STRESS ABOVE THE CRACK TIP DUE TO :lOVING LOAD

, GO .;0-,

MPal PSI

4lsoo

Q" -D- 3. 400 ~ (.) c( 2 a:: (.) 200

hi I

~

I 5/S 11 OVERLAY

7- OVERLAY HI G' F' E' 01 C' BI AI m

ct 01 i , ". ",

m en _I w a:: 1-200 D-en -2 0: ~ W ~ m

•

~ :2

AB CDE FG

TRAFFIC LOADING CASES

H I

FIGJRE ;4. I\[FLUr:~CE LINES OF SHEIIR STRr::S5 iIROV':: THe S!I'\CK TIP D:':E TO noV!"[" [.DAO (WITHOUT SAI·n)

MPOI PSI

ft1. .1400

r:: 9 ~

11200 I I \ I 5/8111 OVERLAY

'0 ~ 8 « 0 hi :> 0

~ 6 U)

I 800 m I

~ N U"I Q:': 5 I

t; kI 41 600 :> ..." ..... 0 liJ iI.. I!.L 1&1

• .., .., A 8 C 0 E F G H H' G' F' E' 0' C' 8' tJ: X I

c( TRAFFIC LOADING CASE :2

FIGURE 15. INFLUENCE LINES OF EFFECTIVE STRESS ABOVE THE CRACK TIP D~~ TO nOVIrlG LOAD (WITHOUT SA;,I)

MPo P51 1400

9 5EFF

8 1200

Q" ..... .... 1 1000 ~ 0 6 <r 800 It: 0 5 5 12 lIJ 4 ~

3 m 400 <r en 2 y;ITH SAM I 1&1 200 5EFF :-en en I ~ lIJ 5'2 It: 0 t; 0 I 2 3 4 5 6 1 (IN)

• • • • • X 5 10 15 (eM) <r ::I OVERLAY TH ICKNESS

FIGURE 16. EFFECTS OF SAt1l ON SHEAR liND EFFECTIVE STRESSES

-26~

(c) Thermal Effects - The elastic moduli of asphal t concrete (AC)

and asphalt concrete friction course (ACFC) for thermal study

were as'sumed to be the same as that for traffic study, i. e. ,

300 ksi, this value may be too high for slow loading such as

thermal changes. Unfortunately, no effect was being made to

study the visco-elastic behavior of AC and ACFC. Results of

this study indicate the horizontal tensile stress near an

existing joint is reduced significantly with the low modulus

interlayer, especially for thin overlays. The SAMI does not

reduce the overall horizontal stress in an asphalt concrete

(AC) or asphalt concrete friction course (ACFC) due to thermal

expansion or contraction. However, it will significantly reduce

stress concentration above the existing joint, minimizing

reflection cracking. For an overlay without a SAMI, horizontal

tensile stress reduces significantly due to increasing thickness

as shown in Figure 17. For an overlay with a SAM!, horizontal

tensile stress is only slightly reduced due to the increasing

overlay thickness. !n a three-layer overlay system, horizontal

tensile stresses are very high in the leveling course and may

cause cracks. Stresses in the surface course are very uniform

and the influence of existing cracks can hardly be detected.

This indicates a SAM! can effectively retard or eliminate

reflection cracking due to thermal expansion or contraction of

a jointed rigid pavement.

(d) Effects of Movi ng Tra ffi c - Throughout the ana.lytica 1 study,

it was assumed there was no load transfer capability across the

joint. This condition would exist when vertical differential

-27-

-~M :2 0: W ::r: ..... -~ ..J 0: W > 0 Z .... en en W 0: ..... UJ

laJ ..J ..... UJ Z laJ .....

• N .... 0: 0 ::r:

• X « :2

PSI

3000

15 2000

10

1000

5

500

250

0 0 0 •

I 3 4 5 6 7 (IN ) • 5

OVERLAY

, ,

10 15 (eM)

THICKNESS

FIGURE 17. EFFECTS OF SAm ON THERHAL HORIZONTAL TENSILE STRESS

-28-

movement of a joint is relatively small under loading. Sufficient

vertical movement can occur that aggregate interlock and dowels

may not have an effect in reduction of shear stress. This

study has revealed that both shear and effective stresses are

at a maximum when traffic loading is completely on one side of

the concrete slab. Results also indicate that a SAMI can reduce

horizontal thermal stress, effective stress and shear stress

in the overlay which many researchers believe to be the major

factors causing reflective cracking.

SUMMARY AND CONCLUSIONS

Based on the field performance of several projects and analytical

studies, the following conclusions can be made:

1. The three-layer system can be used for reducing reflective

cracking when placed over a rigid pavement. The leveling course

provides a smooth surface for a SAMI to behave as a continuous

stress absorbing layer. The stresses in the surface course are

extremely low and insure a much longer service life.

2. When a SAMI is used in the three-layer overlay system, asphalt­

rubber application rate is important because excessive rates

promote instability of the system. A specified rate of appro­

ximately 0.60 gallon-per-square-yard yields good results.

3. Field performance indicates that load stress alone or thermal

stress alone are insufficient to cauSe major reflective

cracking. The combination of these stresses in cold climatic

regions may be sufficient to cause reflective cracks.

4. Field performance indicates that those cracks that do reflect

through are narrow, do not spall and require little if any

maintenance.

-29-

5. A three-dimensional finite element analysis with improved time

and temperature dependent material properties will provide

better results.

10 to 20 times.

However, computer time required may increase

The primary purpose of this study is to report

a successful application of asphalt rubber as a SAM! in rigid

pavement overlay and attempt to provide analytical explanations

for the apparent success of this new approach in pavement

rehabilitation.

6. "Shearing" action is more inducive to reflective cracking of

overlays than "Bending" action when a traffic loading is moving

from one side of a joint to the other.

7. SAM!s can not only reduce stresses caused by thermal changes

but also vertical shear stresses due to moving traffic loading.

8. It is intuitively obvious that reduced stress level will result

in increased service life of overlay. The prediction of service

life with or without SAM! could be accomplished by analyzing

the fatigue behavior of the systems.

9. When a SAM! is utilized, the thickness of overlay becomes less

critical. This may result in very economical approaches to

overlay design.

10. A better understanding of SAM! properties is needed through

continued laboratory testing.

-30-

REFERENCES

1. C. H. McDonald, "Asphalt-Rubber Compounds and Their Applications for

Pavement" 21 st Ca 1 i forni a Streets and Hi ghway Conference, 1969.

2. G. R. Morris and C. H. McDonald, "Asphalt-Rubber Stress-Absorbing

r~embranes: Field Performance and State of the Art" TRB, Transpor­

tation Research Record 595, 1976, pp. 52-58.

3. G. Cooper and G. R. Morris, "Reclaimed Rubber in Seal Coats and

Resea rch in Rubber/ Aspha lt Products" Proceedi ngs of Ari zona Confer­

ence on Roads and Streets, Tucson, University of Arizona, 1974.

4. B. A. Vallerga, G. R. Morris, J. E. Huffman and B. J. Huff "Applica­

bility of Asphalt-Rubber r~embranes in Reducing Reflection Cracking"

Symposium - Prevention and Control of Reflective Cracking, Proceedings

AAPT, Volume 49, 1980, pp. 330-353.

5. G. B. Way, "Prevention of Reflective Cracking in Arizona" presented

at 59th Annual Meeting, Transportation Research Board, 19BO.

6. E. L. Green, "The Chemical and Physical Properties of Asphalt­

Rubber Mixtures" Arizona Department of Transportation Report

ADOT-RS-14(162), July 1977.

7. R. A. Jimenez, "Testing r1ethod for Asphalt-Rubber" Arizona Department

of Transportation Report ADOT-RS-15(164), January 1978.

B. R. A. Jimenez, "Testing of Asphalt-Rubber and Aggregate Mixtures"

Arizona Department of Transportation Report FHWA/AZ-79/lll, 1979.

9. R. A. Jimenez, "Testing Asphalt-Rubber with the Schweyer Rheometer"

Report on Grant Number ENG-7B-10395, National Science Foundation, 1980.

-31-

10. E. L. Wilson, Solid SAP - A Static Analysis Program for Three

Dimensional Solid Structures, Structural Engineering Laboratory

U. C. SESM 71-19. University of California, Berkeley, September,

1971, revised December, 1972.

11. N. F. Coetzee and D. L. Monismith, "Analytical Study of Minimization

of Reflection Cracking in Asphalt Concrete Overlays by Use of a

Rubber-Asphalt Interlayer" TRB, Transportation Research Record 700,

1979, pp. 100-108.

-32-