An evaluation of design highwater clearances for pavements


Material Information

An evaluation of design highwater clearances for pavements
Physical Description:
xiii, 149 leaves : ill. ; 28 cm.
Elfino, Mohamed K., 1945-
Publication Date:


Subjects / Keywords:
Soil moisture -- Measurement   ( lcsh )
Pavements -- Subgrades   ( lcsh )
Road drainage   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1986.
Includes bibliographical references (leaves 144-147).
Statement of Responsibility:
by Mohamed K. Elfino.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000935232
notis - AEP6307
oclc - 16396402
System ID:

Full Text






19. R6







Due to the broad scope of this study, many people provided valuable

assistance. Knowing that it is not possible to cite by name all of the

individuals involved, the author would like to express his sincere

appreciation to every person who has contributed toward the completion

of this study.

The author wishes to acknowledge Dr. John L. Davidson, chairman of

the supervisory committee, for his valued counsel, guidance, and

assistance throughout the entire course of this study. The author also

thanks Dr. Frank C. Townsend, Professor Walter H. Zimpfer, Dr. James L.

Fades, and Dr. Luther C. Hammond for their assistance and for serving on

the Supervisory Committee.

The author wishes to acknowledge the valuable contributions of

Dr. James H. Schaub, Chairman of the Department of Civil Engineering.

The guidance and interest of Dr. Byron E. Ruth are also appreciated.

The author is grateful to the Florida Oepartment of Transportation

for their financial support through the Bureau of Materials and

Research. His thanks go to the administration and the technical staff

of the Soil Section of the Bureau.

Gratitude is expressed to Professor rMichael T. Talbot and Mr. Art

Taylor of the Agricultural Engineering Department for their friendship

and support.

Special recognition is extended to Dr. Robert S. Mcnsell,

Dr. Victor W. Carlisle and Mr. David L. Cantlin of the Soil Science

Department for their assistance and interest.

Special thanks go to Dr. David G. Bloomquist and Mr. Michael S.

Kinne of the Department of Civil Engineering for their technical

assistance and friendship.

The author wishes to acknowledge the assistance of Mr. Kwasi Badu-

Tweneboah in runnina the BISAR computer program and Mr. Andre M. Gallet

for collecting the data for column study as part of the same project and

his thesis.

The author wishes to acknowledge the invaluable consultation

received from Or. Rarry J. nempsey, and Dr. Marshall M. Thompson, both

of the University of Illinois, and Dr. Yu T. Chou of the U.S. Army

Engineer Waterways Experiment Station at Vicksburg, Mississiooi, and

Dr. Mathew M. Witczak of the University of Maryland.

The author is indebted to Ms. Annette M. Davidson for her editing

and proofing of the final copy of this dissertation.

The author wishes to commend Ms. Candace J. Leggett for excellence

and dedication in typing and preparing this manuscript.

The author wishes to recognize the moral support of his wife

Nariman, daughter Nancy, and of his entire family in Alexandria, Egypt,

and the special prayers of his father in California.



ACKNOWLEnGMENTS ..................................................... iii

LIST OF TABLES ...................................................... vii

LIST OF FIGURES ..................................................... ix

ABSTRACT ............................................................ xii


I INTRODUCTION .................................................. 1

1.1 Background ............................................. 1
1.2 Purpose of Study......................................... 4
1.3 Scope of Work ........................................... 4

II REVIEW OF LITERATURE ........ ................................ 6
2.1 Water Retention in Subgrades ............................. 6
2.2 Deformation Characteristics of Subgrades Under
Repetitive Loading .................................... 15

III QUESTIONNAIRE SURVEY.......................................... 18

3.1 Background ............................................. 18
3.2 Analysis of Responses.................................... 19

IV MATERIALS DESCRIPTION ......................................... 22

4.1 Introduction........... ................................ 22
4.2 Classification of Materials ........................... 22
4.3 Soil Density-Water Content Relationships................. 25
4.4 Soil Mineralogy ......................................... 29
4.5 Permeability Test ..................................... 29
4.6 Soil-Water Retention Characteristics .................... 29
4.7 Static Triaxial Compression Tests ....................... 45

V EQUIPMENT AND PROCEDURES.................................... 49

5.1 Column Study ........................................... 49
5.2 Repetitive Load Testing.................................. 56

VI RESULTS AND DISCUSSION ........................................ 71

6.1 Introduction ............................................. 71
6.? information Characteristics at Optimum Water Content..... 71
6.3 DeFormation Characteristics at Varied Water
Retention Conditions................................... 81
6.4 Deformation Characteristics at Selected Water
Retention Conditions.................................... 85
6.5 Deformation Characteristics During Dynamic
Conditioning ........................................... 88
6.6 Application of Test Results.............................. 98

VII CONCLUSIONS AND RECOMMENDATIONS............................... 109

7.1 Conclusions .............................................. 109
7.2 Recommendations ........................................ Ill


A QUESTIONNAIRE AND COVER LETTER................................ 114

TESTS ....................................................... 119


PEFERENCES .......................................................... 144

RIOGRAPHICAL SKETCH ................................................. 148



Table Page

2.1 Typical Values of Height of Capillary Rise .................. 13

4.1 Visual Description and Identification of Droject Soils
in Accordance with ASTM 0-2488-84.......................... 23

4.? Summary of Classification Results for All Project Soils...... 26

4.3 Summary of Soil 9ensity-Water Content Relationships.......... 28R

A.4 Mineralogy Analysis of Clay Fraction ........................ 34

4.5 Summary of Permeability Test Results........................ 34

4.6 Triaxial Static Compression Test Results.................... 48p

6.1 Test Program of Project Soils Under Varied Water Content..... 72

6.2 Summary of the Prediction Equations and Their
Coefficients for All Project Soils at Optimum .............. 82

6.3 Summary of Prediction Equations and MR Values for All
Project Soils at Different Water Retention Conditions...... 91

6.4 Subgrade Resilient Moduli Used as Input in BISAR
Programs ................................................... 104

6.5 Summary of Strains at Top of Subgrade From Laboratory
Tests and from BISAR's Output at 18 and 30 kips............ 105

6.6 The Shell Criteria Limiting Subgrade Compressive Strain
Values Corresponding to Different Load Applications........ 106

B.I Soil-Water Retention Test Results, A-3 Soil,
Standard Proctor ......................................... 119

B.2 Soil-Water Retention Test Results, 1-2-4 Soil,
Standard Proctor ......................................... 120

B.3 Soil-Water Retention Test Results, A-2-5 Soil,
Standard Proctor ......................................... 121

8.4 Soil-Water Retention Test Results, A-5 Soil,
Standard Proctor ......................................... 122

B.5 Soil-Water Retention Test Results, A-3 Soil,
Hand Tools ................................................. 123

B.6 Soil-Water Retention Test Results, A-2-4 Soil,
Hand Tools ................................................. 124

B.7 Soil-Water Retention Test Results, A-2-6 Soil,
Hand Tools ................................................. 125

B.8 Soil-Water Retention Test Results, A-5 Soil,
Hand Tools ................................................. 1?6

C.1 Repetitive Load Testing Results, A-3 Soil,
w = 14.5 Percent ........................................... 128

C.2 Repetitive Load Testing Results, A-2-4 Soil,
w = 11.8 Dercent ........................................... 129

C.3 Repetitive Load Testing Results, A-2-5 Soil,
w = 11.6 Percent ........................................... 130

C.4 Repetitive Load Testing Results, A-5 Soil,
w = 37.84 Percent .......................................... 131

C.5 Repetitive Load Testing Results, A-3 Soil,
w = 13.25 Percent.......................................... 132

C.6 Repetitive Load Testing Results, A-3 Soil,
w = 10.53 Percent.......................................... 133

C.7 Repetitive Load Testing Results, A-3 Soil,
w = 8.10 Percent ........................................... 134

C.8 Repetitive Load Testing Results, A-2-4 Soil,
w = 11.70 Percent .......................................... 135

C.9 Repetitive Load Testing Results, A-2-4 Soil,
w = 10.50 Percent .......................................... 136

C.10 Repetitive Load Testing Results, A-2-4 Soil,
w = 8.7 Percent ............................................ 137

C.11 Repetitive Load Testing Results, A-2-6 Soil,
w = 12.20 Percent........................................ 138

C.12 Repetitive Load Testing Results, A-5 Soil,
w = 44.4 Percent ......................................... 139

C.13 Repetitive Load Conditioning Results, A-3 Soil .............. 140

C.14 Repetitive Load Conditioning Results, A-2-4 Soil............. 141

C.15 Repetitive Load Conditioning Results, A-2-6 Soil............. 142

C.16 Repetitive Load Conditioning Results, A-5 Soil............... 143



Figure Page

1.1 Ways in Which Water Can Enter a Highway Subgrade............. 2

4.1 Map of the State of Florida and the Location of the
Six FOOT Districts...................... ................... 24

4.2 Grain Size Distribution Curves for All Project Soils......... 27

4.3 Soil Density-Water Content Relationships for A-3 Soil........ 30

4.4 Soil Density-Water Content Relationships for A-2-d Soil...... 31

4.5 Soil Density-Water Content Relationships for A-?-6 Soil...... 32

4.6 Soil lensity-Water Content Relationships for A-5 Soil........ 33

4.7 Different Size Tempe Cells ................................... 36

4.8 The Tempe Cell Setup.......................................... 37

4.9 Gravimetric Water Content Versus Height Above Water
Table for A-3 Soil ......................................... 39

4.10 Gravimetric Water Content Versus Height Above Water
Table for A-2-4 Soil ..................................... 40

4.11 Gravimetric Water Content Versus Height Above Water
Table for A-2-6 Soil ....................................... 41

4.12 Gravimetric Water Content Versus Height Above Water
Table for A-5 Soil ......................................... 42

4.13 Retention Characteristic Curves for all Project Soils
Compacted to Standard Proctor ............................... 46

5.1 Lucite Column Section and Tensiometer ....................... 50

5.2 Inverted Flask to Maintain Fixed Water Level................. 51

5.3 Column Setup During Testing ................................. 52

5.4 Mercury Manometers for Measuring Soil Suction in
the Column ................................................. 53

5.5 Hand Compaction Tools........................................ 54

5.6 Lucite Mold and Compaction Equipment......................... 57

5.7 Mold Assembly in the Deep Sink During Saturation............. 59

5.8 Mold Assembly During Free Drainage........................... 60

5.9 Components of the Conditioning Cup.......................... 6?

5.10 Specimen During Conditioning................................ 63

5.11 Clamps With LVDT's for Measuring Lateral and Vertical
Deformations ............................................... 65

5.12 An Overall View of the Repetitive Loading Test Equipment..... 66

5.13 The MTS System, Load Cell and Triaxial Chamber............... 67

5.14 LVDT Clamps Placed Over the Specimen ......................... 9

6.1 Axial Permanent Strain Versus Number of Stress
Applications for All Project Soils......................... 73

6.2 Axial Strain Versus Number of Stress Applications
for A-3 Soil ............................................... 74

6.3 Axial Strain Versus Number of Stress Applications
for A-2-4 Soil ............................................. 75

6.4 Axial Strain Versus Number of Stress Applications
for A-2-6 Soil ............................................. 76

6.5 Axial Strain Versus Number of Stress Applications
for A-5 Soil ............................................... 77

6.6 Gravimetric Water Content Versus Height Above Water Table
for A-3 Soil From 3oth Tempe Cells and Column Study........ 84

6.7 Axial Permanent Strain Versus Number of Stress Applications
for A-3 Soil at Varied Water Retention Conditions.......... 86

6.8 Axial Permanent Strain Versus Number of Stress Applications
for A-2-4 Soil at Varied Water Retention Conditions........ 87

6.Q Axial Permanent Strain Versus Mumber of Stress Aoplications
for A-2-6 Soil at Varied Water Retention Conditions........ 89

6.10 Axial Permanent Strain Versus Number of Stress Applications
for A-5 Soil at Varied Water Retention Conditions.......... 90

6.11 Axial Permanent Strain Versus Number of Stress Applications
for A-3 Soil during Conditioning........................... 92

6.12 Axial Permanent Strain Versus Number of Stress Applications
for A-2-4 Soil During Conditioning........................ 94

6.13 Axial Permanent Strain Versus Number of Stress Applications
for A-2-6 Soil During Conditioning........................ 95

6.14 Axial Permanent Strain Versus Number of Stress Applications
for A-5 Soil During Conditioning.......................... 97

6.15 Recommended Subgrade Clearance Heights for the Project
Soils Based on Capillary Fringe (Retention Case)........... 102

6.16 Recommended Subgrade Clearance Heights Based on a
30 kip Axle Load and 10 Load Applications for
the Project Soils ........................................ 108

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Mohamed K. Elfino

May 1986

Chairman: Or. John L. Davidson
Major Department: Civil Engineering

This research was carried out to investigate the effect of water

presence on the permanent deformation of four subgrade soils. First,

the physical and engineering properties were determined, with emphasis

on developing the soil-water retention characteristic of the four

soils. Second, repetitive load testing was performed at two different

water conditions. These conditions were (1) at optimum, to represent

the as-built condition, and (2) at varied water retention conditions to

represent the subgrade condition in-service, i.e. in equilibrium with

the designated water table. The deformation characteristics of the

suhgrade fill, at different water conditions, were related to pavement

rutting in accordance with the Shell Criteria. When a tolerable

deformation was obtained from a specimen at a specific water condition,

its location on the soil-water retention curve was determined and the

height of the subgrade fill fixed.

Based both on the height of the capillary fringe and on the Shell

Criteria, the following subgrade fill heights were found to be adequate;

24 inches for the A-3 soil, 36 inches for the A-2-4 soil, and 12 inches

for the A-2-6 soil. If the A-5 soil must he used, further careful

analysis should be undertaken. A fill thickness in the range of 4 to 6

feet is probably adequate.

Because of the possible variations in soil within any classifica-

tion grouping, the results of this research are limited to the parti-

cular soils investigated.



1.1 Background

The presence of water in highway pavement systems accelerates the

deterioration and destruction of the pavement. This has long been

recognized. Great road builders of the past, such as Pierre M.

Tresaquet of France and John L. Mc4dam of Great Britain, knew of the

importance of keeping roadbeds dry and protecting them against water


In 1820, John McAdam stated:

The roads can never be rendered thus perfectly se-
cure until the following principles be fully under-
stood, admitted and acted upon: namely, that it is
the native soil which really supports the weight of
the traffic; that whilst it is preserved in dry
state it will carry any weight without sinking
that. if water pass through a road and fill
the native soil, the road whatever may be its
thickness, loses support and goes to pieces.

Pavement systems must he designed such that either water is

prevented from entering places where it can cause damage, or

alternatively, any water which does enter can he quickly and safely


Figure 1.1 shows six ways in which the moisture content of a

highway suhgrade can be changed. These are

1. Downward flow through joints, cracks, and porous surfaces.

2. Lateral flow from water ponded on high medians.

3. Upward flow from high groundwater, springs, or rivers.

I .







I I '-







4. Capillary suction from underlying water table.

5. Transfer of moisture either to or from the soil in the verges
as a result of differences in moisture content.

6. Condensation of water vapor as a result of fluctuations in
temperature and other atmospheric conditions.

The research described in this dissertatioy- relates specifically to

the fourth pause listed above, capillary suction. To prevent water rise

into critical areas of the pavement system or to reduce its effects,

some State Departments of Transportation have developed highwater

clearance guidelines. In these guidelines a minimum height, the

clearance, between a groundwater level and a particular elevation within

the pavement system is specified.

The Florida Department of Transportation (FDOT) has such guide-

lines. They state that the bottom of the pavement base should he

located a specific height above a designated highwater level. The

details are as follows:

4-lane Primary and Interstate 3 feet above 50-year highwater

2-lane Primary 2 feet above 50 year highwater

Secondary Roads 1 foot above 25 year highwater

The policy has been applied uniformly statewide, and is intended to

satisfy two concerns:

1. Achieving the required compaction and stability during

construction operations.

2. That adequate pavement performance is provided.

The guidelines, however, do not address a most critical factor,

that of subgrade soil type. As a result, there have been locations

where the specified clearance appeared excessive for the conditions

encountered. However, with no other guidelines to follow, the policy

was adhered to. The resulting high fill costs may not be justified. At

the other extreme, situations may exist where the specified clearance

may be inadequate. This also leads to eventual waste of funds due to

construction problems or poor pavement performance.

To take subgrade soil type into account in setting such guidelines

requires a determination of which physical and mechanical soil proper-

ties influence behavior during construction and under the expected

dynamic service loadings. It is also very important that the water

present in the subgrade be correctly modeled in any laboratory testing,

i.e., correct water content and water pressure. This research evolved

from a recognition of the need to include subgrade soil type in high-

water clearance guidelines and the problems involved in doing this.

1.? Purpose of Study

The purpose of the study was to investigate the soil-water reten-

tion behavior of four different Florida soils, to determine their defor-

mation characteristics under repetitive loading at different water

retention conditions, and to evaluate the suitability of the soils as

subgrade material for a major highway.

1.3 Scope of Work

This study was planned in six phases, as follows:

First Phase

This phase consisted oftwo parts. First, a questionnaire survey

was sent to all State De6artments of Transportation to ascertain if they

used highwater clearance specifications and, if so, what were the bases

for their criteria. A literature review was then made in the areas of

soil water retention in subgrades and of repeated load testing.

Second Phase

The physical and engineering properties were determined for four

different materials commonly used as fill by FDOT. These properties

included grain size distribution, Atterberg limits, specific gravity,

permeability, soil mineralogy, soil-water retention characteristic,

density-water content relationships, cohesion, and internal friction


Third Phase

This phase consisted of column studies to simulate field conditions

for compacted and stabilized subgrade in order to obtain the soil-water

characteristic curves.

Fourth Phase

The deformation characteristics of the four subgrade materials

compacted at optimum water content were determined. This represents the

pavement condition as built.

Fifth Phase

The deformation characteristics of the four subgrade materials at

varied soil-water retention conditions were determined. This represents

the subgrade condition in-service, i.e., in equilibrium with the desig-

nated water table. Repetitive axial loading was used in both the fourth

and fifth phases.

Sixth Dhase

The deformation characteristics of the subgrade fill, at different

water conditions, were related to pavement rutting potential in accor-

dance with the Shell Criteria. When a tolerable deformation was

obtained from a specimen at a specific water condition, its location on

the soil-water retention curve was determined and the height of the most

economical subgrade fill then fixed.


2.1 Water Retention in Subgrades

In 1897 Briggs developed a classification describing soil water in

the following manner:

a. Hygroscopic water which is held in thin films on the soil

particles by adsorption.

b. Capillary water which is held in soil voids by surface tension.

c. Gravitational water which moves in and through the voids under

the influence of gravity and which drains from the soil when it

is held at a higher elevation than the water table.

Soil moisture content is commonly expressed as the percentage of

water weight per unit weight of oven dry soil. Experience has shown

that moisture percentages are of little significance in predicting soil

behavior, unless something is known of the texture and mineralogical

nature of the soil itself. A more reliable approach in characterizing

soil moisture conditions is to base it on the security or tenacity with

which the water is held by the soil.

In 1907, Buckingham introduced the energy concept which provided a

means of describing soil moisture conditions in terms of a potential

function which is a property of the state of stress of the soil water

itself. Buckingham reasoned that the flow of moisture through

unsaturated soils was similar in many respects to the flow of

electricity through a wire or the flow of heat through a rod. The flow

in each case being from regions of high potential to regions of lower

potential. In the case of moist soil in which the capillary potentials

are balanced with the gravitational potential, the capillary soil

moisture will approach a state of static eauilibrium with a water table

below and no upward or downward flow will take place.

For texturally homogenous soils in which the moisture is in

equilibrium, the soil will be saturated at the water table and will

decrease in moisture content with distance above the water table.

In 1920 Gardner, making use of Buckingham's energy concept, stated

that capillary potential was equivalent to the pressure potential in

Bernouilli's equation, and consequently could be determined by measuring

the soil water pressure. He also noted that the capillary potential was

equal to the negative pressure (tension). In 1941, Russell and Spangler

brought to the attention of highway engineers the Buckingham capillary

potential relationships in soil moisture. They believed this concept

would be a useful tool in the study of subgrade moisture. In 1944,

Kersten reported on such a study. He investigated the field moisture of

the upper six inches of the subgrade near the interior of flexible

pavement systems. His work is summarized as follows:

a. The degree of saturation existing in the subgrades of numerous

projects in 6 states averaged 73 percent, the range being from 60 to 81


b. The suhgrade soils of projects in which a high average percent

of saturation occurred were in mot instances either clay or silty clay.

c. Saturation percentages varied with soil texture; in general,

( they were high for the clays and became orogressively lower for the clay

loams, the loams, and the sandy loams.

> d. Soils of groups A-6 and A-7 showed higher average percentages of

saturation than those of groups A-1, A-2, and A-4.

e. Subgrade moistures expressed as percentages of the plastic limit

for a large variety of soils in six states averaged 77 percent.

Averages for individual states varied from 64 to 82 percent.

Approximately 17 percent of the determinations disclosed moisture

contents in excess of the plastic limit.

f. The fine-textured soils, such as clays, exhibited a marked

tendency to attain moisture contents in excess of their plastic limit.

Sandy loams rarely had moisture contents as great as their plastic

limit. Loessial silty soils tended to attain moistures close to their

plastic limit.

g. The optimum moisture contents of the soils were exceeded by the

field moistures in about one-third of the determinations reported. Clay

soils exceeded the optimum most commonly, but soils of all textures,

including the sandy loams, had moistures greater than optimum in a

substantial proportion of the tests.

h. Only slight changes in moisture content for periods of from 1 to

5 years were indicated in tests on several projects. Soils on most of

these projects, at the time of the initial tests, already had moisture

contents approaching the plastic limits.

i. Clay soils with high percentages of saturation were encountered

in areas with annual precipitations as low as 14 inches. Most tests in

such regions, however, give relatively low saturation values.

The presence of water in a pavement system may decrease its

strength in several ways, as summarized by Barber and Sawyer (1952):

1. It reduces the cohesion by lowering the capillary forces.

2. It reduces the friction by reducing the effective weight of the

material helow the water table.

3. For quickly applied loads, it may reduce the strength by the

development of excess pore pressure.

4. Bearing capacity of sand is decreased more than 50 percent due

to complete submergence as compared to dry sand. Capillary

saturation gives somewhat less reduction.

5. For loose saturated sand under dynamic loading, the tendency to

become denser causes pore pressures to increase. This reduces

the effective normal stresses and thereby reduces the strength.

6. While high permeabilities may be obtained by using coarse

aggregates, care must be taken to prevent a reduction in their

permeability and stability by the intrusion of adjacent finer


Darter et al. (1982) gave a list of several of the more common

distress types, for flexible and rigid oavements, which are caused by

water presence in the pavement system. The list included surface |

deformation in the form of distortion, corrugation, rutting, waves,

depression, and potholes for the flexible pavement, while for rigid

pavement the surface deformation is in the form of pumping and faulting.

Roque and Ruth (1983) in their analysis of pavement cracking on the '

Florida Turnpike, found that the in situ water content of the embankment

soil of the uncracked pavement was six percent less than the cracked
pavement sections. This increase in water content would substantially /

reduce foundation strength, making it susceptible to localized shear

failure or high deformation under stress. Haynes and Yoder (1963) found

that the degree of saturation has substantial effect on the repeated

load deformation properties of the AASHTO Road Test crushed stone and

gravel materials. In these tests, the total deformation after 1,000

load cycles increased markedly once the moulding saturation exceeded

about 85 percent. In contrast with this, the total deformation of

specimens at saturation between 70 and 85 percent showed only a slight

dependence on the moulding saturation.

Monismith and Finn (1977) reported that water presence in the

pavement system is one of the most important environmental factors,

particularly because it influences the response of the materials in the

pavement section to load, and because it may cause undesirable volume

change. They also stated that for design purposes, the influence of

water may be considered by measuring the properties of materials at

water contents, which are assumed representative of those that may

develop at some time subsequent to construction.

Chu et al. (1977) indicated that the results of investigations on

the variation of subgrade moisture, which were conducted in the United

States and abroad, showed that, after a certain period of time following

construction, subgrade soils below impervious pavements remain at a

fairly stable moisture condition, except for a zone close to the

pavement edge. Under idealized conditions and in areas where frost

penetration does not extend to the subgrade, the variation in subgrade

moisture depends mostly on the relative elevation of the groundwater

table if it is within a certain depth below the pavement. In this

respect, the critical_depth of the groundwater table is dependent

primarily upon soil type. Russam (1970) summarized the finding from the

previous investigations and stated that, for highway and airport

pavements, this depth would be approximately 20 feet in clays, 10 feet

in sandy clays or silts, and 3 feet in sands. / The moisture content of

subgrade soils may then be estimated on the basis of the depth of the

water table below the pavement together with soil suction data, provided

that a moisture equilibrium condition has been reached in the

subgrade. It could then be concluded from the previous studies that \

subgrade soils below pavements are seldom saturated, as is often assumed '

in formulating laboratory test procedures for the evaluation of subgrade

soils. For this reason, the common practice of soaking soil specimens

in water for a number of days before using them for laboratory tests,

such as in the California bearing ratio (CBR) test, may result in an

extremely severe moisture condition which usually does not occur under

the pavement. If pavement structures are to be designed by rational

procedures, simulation of anticipated field moisture is necessary.

Burland (1965), and Aitchison and Richards (1965) emphasized the. need of

simulating negative pore pressure and utilizing equilibrium suction in

determining the behavior characteristics of partially saturated soils.

Christensen (1940) reported that if a column of soil is in

equilibrium with a water table and there is no tendency for the water in

the soil to move either upward or downward, then for each foot of

increase in elevation above the water table the tension in the water

should increase by an amount equal to one foot of water. This relation

between the tension and the elevation in the column can be represented

by a straight-line extension upward from the water table with unit slope

(450 for the scale being the same for the abscissa and the ordinate).

Similar conclusions, to those of Christensen, were reported by Croney et

al. (1958).

Capillarity in soils is similar in many respects to the rise and

retention of water in a capillary tube, although there are also

important differences between the two cases. It is convenient to

introduce the subject of capillary water in soil by reviewing the action

of water in a capillary tube. It is well known that if a clean glass

tube having a fine bore is placed vertically in a container of water,

the water in the tube rises above the level in the container. Two

phenomena are responsible for this rise:

a. Forces of attraction between water molecules which at an air-

water interface give rise to surface tension.

b. Attraction of water to the material of the tube which causes


Capillary rise is related to surface tension, radius of capillary tube,

and angle of contact as follows:
h = 2- cos e
c rY
in which

hc = height of capillary rise

r = radius of circular capillary tube

T = surface tension of water

o = contact angle between the surface of the liquid
and the surface of the tube

Yw = unit weight of water

Some typical values of capillary rise are given in Table 2.1.

Capillary water is at a pressure less than atmospheric, which

creates capillary tension in the pore water and a counteracting

effective stress known as capillary pressure. The effect of capillary

stress is evident in the shearing strength of the damp, fine sand. The

Table 2.1 Typical Values of Height of Capillary Rise

Soil Type Height of capillary rise hc, cm

Coarse sand 2-5

Sand 12-35

Fine Sand 35-70

Silt 70-150

Clay 200-400 and greater

Source: A. I. Silin-Bekchurin (1958)

wheel-load stability of damp sand on some beaches reportedly results

from such pressures. When the sand is completely submerged or

completely dry, this stability is largely lost.

Spangler and Handy (1982) stated that capillary water may be

expressed quantitatively by a stress property called capillary

potential, matric potential, or soil suction and they defined it as the

work required to pull a unit mass of water away from a unit mass of soil

exclusive of osmotic and other influences. Several factors affecting

the relationship between soil suction and water content were discussed

in detail by Spangler and Handy. These included temperature, dissolved

salts, grain size, state of oacking, angle of contact, and mineralogy of


A soil-water characteristic curve is a form of expressing the

relationships between water content and pressure potential. In an

unsaturated soil column, when it is at equilibrium with free water, the

water content decreases with height above the free water. Physically,

the curve tells (at any given water content) how much energy (per unit

quantity of water moved) is required to move a small quantity of water

from the soil. It indicates how tightly water is held in the soil. The

area of soil physics lends itself to such topics, where references are

available from such researchers as Hillel, Taylor, Ashcroft, Kirkham,

Powers, and Childs just to name a few.

Soil water characteristic curves are divided into two types,


a. Sorption curves representing the water rise case.

b. Desorption curves representing the drying or the draining case.

The difference between them is known as hysteresis. Hillel (1980)

gave the following causes which may contribute to hysteresis:

1. The geometric nonuniformity of the individual pores which are

generally irregular in shape voids and are interconnected by small

passages. (This results in the ink bottle effect.)

2. The contact-angle effect, i.e., the greater the contact angle

the larger the radius of curvature in an advancing meniscus than in a

receding one. (A given water content will tend therefore to exhibit

greater suction in desorption than in sorption.)

3. Entrapped air, which further decreases the water content of

newly wetted soil. (Failure to attain true equilibrium can accentuate

the hysteresis effect.)

4. Swelling and shrinking, or aging phenomena, which result in

differential changes of soil structure, depending on the wetting and

drying history of the sample.

Janssen and Dempsey (1981) stated that generally the desorption

(the drying curve) is sufficient for most civil engineering uses and it

is the most critical.

The soil-water characteristics curve can be determined using one or

more of the following methods depending on the soil encountered.

a. tensiometer method

h. direct suction method

c. pressure plate method

d. centrifuge method

2.2 Deformation Characteristics of Subgrades Under Repetitive Loading

Monismith et al. (1975) developed a constitutive relationship

between plastic strain and number of stress application. The

relationship was represented by a power function as follows:

p A NI


E = permanent strain

N = number of stress applications

A & B = experimentally determined coefficients

This equation was verified up to 100,000 stress applications. The

authors also indicated that the subgrades of well-designed pavements are

subjected to comparatively small stresses from conventional traffic

loads (9,000-lb. wheel load). At stiffnesses in the asphalt bound layer

larger than 200,000 psi, the vertical compressive stresses in the

subgrade were less than 5 psi. At these stress levels, measurements of

permanent deformation are not as precise as at higher stresses. Dempsey

(1976) indicated that strength and stiffness parameters for aggregate

materials are effected by moisture. However, the relative effect is

influenced by the material's gradation, percent of fines and the degree

of saturation.

Hicks and Monismith (1971) showed that increased saturation lead to

reduced resilient moduli values for granular materials.

Barksdale (1972) found that the plastic strain in granular

specimens increased an average of 68 percent when they were soaked.

Seed et al. (1962) showed that pavements containing a saturated granular

base layer displayed higher magnitudes of deflection than those

pavements that were "dry."

Dempsey (1976) noted that unsoaked and soaked CBR values (a static

type test) for granular materials were normally not significantly

different, indicating the significance of repeated, dynamic loading


Shackel (1973) concluded that for a particular molding saturation,

the resilient axial strains decreased linearly with increasing

suctions. He also concluded that the cumulative, nonrecoverable

(residual) axial strains decreased rapidly as the suctions increased.

Monismith et al. (1975) reported on two aoproaches which are avail-

able to consider rutting (permanent deformation) as a result of repeated

traffic loading. One of the approaches involves limiting the vertical

compressive strain at the subgrade surface to some tolerable amount

associated with a specific number of load repetitions (Shell

Criteria). By controlling the characteristics of the material in the

pavement section through materials design and proper construction

procedures (unit weight or relative compaction requirements) and by

ensuring that materials of adequate stiffness and sufficient thickness

are used so that the strain level is not exceeded, permanent deformation

equal or less than some prescribed amount is thus assured. The other

approach involves an estimation of the actual amount of ruttinq which

might occur using materials characterization data developed from

laboratory tests. Chou (1977) stated that the major advantage of the

first approach is that it could be used as a workable tool for the

pavement design, and several agencies have introduced procedures based

upon it.

Only the Criteria developed hy Shell Oil Company will be considered

in this study. Such Criteria can be utilized to ensure that permanent

deformation in the subgrade will not lead to excessive rutting at the

pavement surface. These Criteria may be thought to he associated with

ultimate rut depths of the order of 3/4 inch.

The Shell Criteria are based on California Bearing Ratio (CBR)

procedures and emperical correlation with results from the AASHTO Road

Tests. These tests included single axle loads ranged from 2,000 to

30,000 lbs, including the standard 18,000-lb. The pavement section

consisted of a surface course (Bituminous concrete 1 to 6 inches), a

base course (a well-graded crushed limestone 0 to 9 inches), and a

subbase layer (a uniformly graded sand-gravel mixture 0 to 16 inches).

The testing covered a wide range of combinations of the pavement

components. These combinations are still applicable today, which makes

use of the Shell Criteria a reasonable one. To use the Shell Criteria,

the asphalt concrete stiffness should range between 100,000 and 200,000

psi, and the Poisson's ratios of the materials in the pavement section

should be in the range of 0.35 to 0.40.

In this study, the simulation of the soil-water retention in the

subgrade soils will be used to characterize the permanent deformation

due to repetitive loading.


3.1 Background

A twenty item questionnaire was sent to the Departments of

Transportation of all fifty states, to the District of Columbia, and to

three Canadian provinces, to determine if they made use of highwater

clearance specifications and if so, the bases for their criteria. A

copy of the questionnaire and the cover letter are included in Appendix


Questions were designed to provide information on the existence of

similar highwater clearance guidelines at the other Departments of

Transportation, the development of such guidelines, consideration of

soil type, duration of highwater level, determination of designated

highwater level, common materials used for subgrade, capillary

phenomena, measures taken to reduce the capillary saturation problem,

equipment used for evaluating pavement deformation, criteria used for

evaluating pavement deformation, and any current research or previous

studies in the area of highwater clearance.

A total of 52 out of 54 departments (or 96 percent) responded.

Most of the surveys were completed by pavement designers or drainage

engineers. Their practical experience and sound judgment was evident

from the responses.

3.? Analysis of Responses

Since many of the survey questions elicited descriptive responses,

the survey is not amenable to any type of computer analysis or even to

any useful tabular listing of results. The responses have instead been

summarized in the following ten paragraphs.

1. Fifteen departments have guidelines similar to those of the

Florida Department of Transportation. Ten of the fifteen do not

consider soil type. This suggests that these departments apply their

guidelines uniformly to attain the same goals as the FOOT. In areas

where frost is a problem, clearances of 4 to 6 feet from the finished

grade .to designated highwater are required, especially where the

subgrade soil is silt or clay. Of those departments which do not have

guidelines similar to those of the FDOOT, some considered each project

individually, some keep the highwater level below the subbase, while

still others take measures to break the caoillary rise. In all cases

reported, experience and engineering judgment were used to develop the

clearance guidelines. No testing procedures are followed in considering

soil type.

2. Only six departments consider the duration of the highwater

level. These departments are in the north central states. The spring

season, with snow melt, is considered the critical time. For short

highwater durations, they believe pavement damage will not result. For

longer durations, they take protective action such as closing lanes or

limiting vehicle loads.

3. Thirty-four of the departments responding place great

importance on the determination of the designated highwater level. Some

of the procedures mentioned make use of

Nearby rivers or lakes.


Interviews with maintenance personnel, mail carriers, local
residents, etc.

Reviewing previous records.

Near bridge crossings use water surface profile program.

Routing the 100 year flow through the bridge.

Using National Flood insurance maps.

Corps of Engineers as a source of information and records.
Computer programs from Corps of Engineers, Soil Conservation
Service, and USGS.

Soil conservation runoff methodology.

4. The subgrade materials used by the different departments ranged

from gravel and sand to silt and clay. The material depended on what

was locally available with economic feasibility as the deciding factor.

5. Capillary rise was acknowledged by thirty-eight of the

departments as one of the factors leading to a reduction of subgrade

strength. Departments which use only gravel and coarse sand as subgrade

material did not consider the effect of capillary rise on subgrade

strength, simply because these soils exhibit very little capillary rise.

6. Measures taken to reduce capillary effects were

Use of free draining materials with controlled percent of fines
(maximum 10 percent passing #200).

Use of lime treated subbase.

Use of under-drains and adequate surface drainage.

Increased subbase depth where suspect material was encountered.

7. Methods used for evaluating pavement deformation were


Road rater and profilometer.

Benkleman Beam.

Field measurement of rutting, and experimental sections.

The most commonly employed methods were the Dynaflect and the Renkleman


8. Thirty-one of the departments responding do not have specific

criteria for evaluating pavement deformation. Twenty-one departments

use the Asphalt Institute or the Shell Oil Company criteria. American

Association of State Highway Transportation Officials (AASHTO) road data

and Chevron research are also used.

9. No departments were involved in, or were aware of, any ongoing

or previous studies on highwater clearance effects.

10. Some departments sent chapters of their design manuals and

expressed their interest in the study by requesting copies of the final



4.1 Introduction

The Bureau of Materials and Research, FOOT, at Gainesville,

provided four soils for the research. These represent the most commonly

used fill soils in the State of Florida.

Ten bags of each material, weighing a total of approximately 500

pounds, were received. Visual identification was made in accordance

with ASTM 0 2488-R4. A summary of this identification, the as-received

water content, and the source of the materials by districts, are listed

in Table 4.1. Figure 4.1 shows a map of the State of Florida and the

location of the six FOOT districts.

Following the inspection and identification, the materials were

air-dried by spreading on the laboratory floor and passing air over them

with a fan. Each was then placed in a 30-gallon rubber container with

two plastic liners for sealing. This maintained the material in the

air-dry condition throughout the research.

4.2 Classification of Materials

The four soils were classified according to the AASHTO and the

Unified Systems, using grain-size analysis (AASHTO T-11, T-27-82) and

liquid and plastic limits (AASHTO T-89, T-90-81). Specific gravities

(AASHTO T-100-75) were also determined.

>, >. *-
.C -C C
-4-' 4-~ <
S- I- C
L 5,.
LL) '0 lO

F ,

C .-4

S.- L a
C C 3



0 4



.> <




I Q.
-C o
I .c




1 .
* r- *r.-


I C3

o (

*r- X

o =



U. 0L 0,


3 C

Vd) I

4- >-,

i- Iv
r-- toi
*r q--

-4i C-si *

-4 cc l-






c 5- C
c cc:


C .0
U. ~


Figure 4.1 Map of the State of Florida and the Location of the
Six FDOOT Districts

The results from the classification and specific gravity tests are

presented in Table 4.2. Based on AASHTO classification, the following

identification will be used throughout the text:

Soil #1 A-3

Soil #2 A-2-4

Soil #3 A-2-6

Soil #4 A-5

Figure 4.2 shows the grain size distribution curves for all four soils.

4.3 Soil Density-Water Content Relationships

The soil density-water content relationships were determined by

performing "Proctor" compaction tests. The Standard Proctor test

(AASHTO-99-81) was performed on the four FDOOT subgrade soils.

Stabilized subgrade materials, made by mixing three parts subgrade soil

and one part lime rock by weight, were tested using the Modified Proctor

test (AASHTO-180-74). A summary of the compaction test results is

included in Table 4.3, along with calculated values of the void ratio

and the percent saturation at optimum water content and maximum

density. The following equations were used to calculate the void ratio

(e) and the degree of saturation (S).
Gs w G s w
e = 1 and S -
Yd e


e = void ratio
Gs = specific gravity
Y = water unit weight
Yd = soil dry unit weight
w = moisture content (percentage)
S = degree of saturation (percentage)

Table 4.2 Summary of Classification Results for All Project Soils

Soil Number

Sieve Diameter 1 2 3 4
Number (mm) % Passing

4 4.76 99.67 99.90 100 100

10 2.00 99.51 99.88 99.38 99.95

Sieve 40 0.42 94.80 99.70 77.95 98.95

Analysis 60 0.25 70.87 99.05 59.27 97.99

100 0.149 16.31 63.24 42.31 96.31

200 0.074 4.10 10.46 28.85a 94.50b

Specific Gravity 2.61 2.62 2.68 2.55

Liquid Limit 31 52
Atterberg Limits NP NP
Plasticity Index 11 8

AASHTO A-3 A-2-4 A-2-6 A-5
Unified SP SP-SM SC MHc


Hydrometer test results showed 6.95% silt size and 21.9% clay size.
Hydrometer test results showed 73.5% silt size and 21.0% clay size.
Hydrochloric acid added to the soil resulted in a violent reaction
indicating the presence of calcium carbonate rather than organic in
the soil.




~ n
-- ---- -- __ __ i


__ I I
-Z__ := = = : := = LL

^ =====^==s "

0 -- --- -- --- *^ ^ ^ -- _
0 ^ -- --- -- ---^-- --- 0 0 0

f~~~~~~N. %-'^ <<^ --'^

4-- -

4i/' (

0 -

4-_ a--

S- xo c-

c El-
0 "



x V) o

1 0
>1 CT

0l- -


*r- -F -



*3* 0- *- 'C

a, U: C\ C

a.i .' a-

- *-4

LO Lc C- N-
0 o N
0! C'" l

trc a N-

4*F Lr 4C

O C L;

N- Co

(- -




-4 -4



a) 0

01 O
>1 r-
(V .0

4), 0

.0 SC
Lt I

Q. E

S.- .-


I\ .Q
*> C

(1 **--'

E n)


ir :3

a) V
LV --
U* to

0 N

S *4r-

"O r
-o 4-

4 4-


4- 0

0 > 0
3. .4

-. CI
4- r- -4

Cc 0)-
H- r- H- V
"CI Er~-
C- I
4-a V)--O
1-0 3 --

0> 0

1 L'u
2 *r *'

MN e-

a) 0a
I- E L- S

-4 U)
(V .0

Figures 4.3 through 4.6 show the test plots for the individual

soils 1 through 4, respectively, for the Standard (T-Q9) and modified

(T-1SO) proctor tests.

4.4 Soil Mineralogy

The purpose of this analysis was to establish the nature of the

clay minerals present in the project soils. Particular emphasis was

placed on the presence of montmorillonite in the A-5 material, which

exists in South Florida and may cause problems if used as fill.

Representative samples of the four soils were analyzed at the FOOT

Bureau of Research and Materials in Gainesville, using X-ray diffraction

techniques. Table 4.4 summarizes the results.

4.5 Permeability Test

Permeability tests were performed on all soils. Each specimen was

compacted at optimum water content (T-99) in a 6-inch diameter

compaction mold. The mold was then fitted in a permeameter and the soil

saturated. The constant head procedure was used for the A-3 soil, while

falling head tests were performed on the other three soils. Table 4.5

summarizes the permeability test results.

4.6 Soil-Water Retention Characteristics

The soil-water retention characteristics for the project soils were

determined, using commercially available Tempe pressure cells. Two

different size cells were used, one with a 3 3/8-inch diameter, the

other with a 2 1/4-inch diameter. For the larger diameter apparatus,

specimens were hand compacted directly in the cell. For the smaller

S* *


\ ro






a /


0) 0.-'




\) ., CM



\) N 0

(od) i HI3A IN O


F- F-

-- M

U* Z





(ad) iiI3M iNn ko
/ \ *(A
y \ 2
,/ V* LU
/ ^s.'"" C-

f x ^ Srt| -' 40

f a)
( C-

\ ed HI3 ..a.f A




I- i- \ 0


3 c

Q .-

i oC
c a

(4 d) 3


0 a
(us LvIB Il A

Table 4.4 Mineralogy Analysis of Clay Fraction

Soil No. Soil Type Clay Minerals Present

1 A-3

2 A-2-4 Kaolinite, Chlorite

3 A-2-6 Chlorite

4 A-5 Kaolinite, Montmorillonite

Table 4.5 Summary of Permeability Test Results

Coeff. of Permeability
Soil No. Soil Type Type of Test cm/sec

1 A-3 Constant head 7.3 x 10-4

2 A-2-4 Falling head 1.1 x 10-s

3 A-2-6 Falling head 4.R x in"7

4 A-5 Falling head 6.4 x 10-6

cell, specimens were obtained using a thin wall tube and sampling from a

soil mechanically compacted in a standard 5-inch mold. Densities

obtained were within 2 percent of the T-99 maximum dry density.

The testing was performed at the Soil Physics laboratory of the

Soil Science Department, University of Florida. After preparing the

soil specimen as described above, air was purged from under the plate by

injecting water through the bottom nipple. The lid was then set on and

the cell placed in a pan of water to saturate. The cell inlet tube was

then connected to an air pressure source and the cell allowed to drain,

initially under zero pressure. The cell was weighed daily. When the

change in weight was less than 0.02 g for two consecutive days,

equilibrium was assumed to exist, and a pressure of 3.5 cm of water was

applied. This process was then repeated for pressures of 20, 30, 45,

60, 80, 100, 150, 200 and 345 cm of water, and at each stage the weight

of the cell and sample was determined. The air pressure was measured by

means of a water column. After reaching equilibrium under the 345 cm

pressure, the apparatus was disassembled and the dry weight of the

specimen determined after drying in an oven for 24 hours at 1050 C.

Knowing this weight, the water contents at the different pressures

applied during the test can be computed. The system and procedure used

are similar to those reported by Janssen and Dempsey, 1980. The

photograph in Figure 4.7 shows the two sizes of Tempe cells. Figure 4.8

shows the setup during testing, including the water columns to the

left. The Tempe cell utilizes pressure rather than suction to drive the

water out of the soil pores and reach static equilibrium at the specific

air pressure applied.

-iutje 4.7 liferent Size Tezre Cells

cioure A.8

The Temopll t1 JO

Appendix B contains a summary of the data collected from the soil-

water retention tests on both hand compacted and mechanically compacted

specimens. Hand compacted specimens were tested because their fabric

and density would match those obtained in the column study, which will

be described later. The tests using mechanically compacted specimens

will be the references for the subsequent repetitive loading testing at

varied water conditions. Figures 4.9 through 4.12 show plots of

gravimetric water content (w) versus distance above the water table for

the four project soils compacted both mechanically and using the hand


It is also common practice to plot the above relationships using

the volumetric water content O. This is defined as the fraction of the

soil volume occupied by water (Yong and Warkentin 1975).

Vw Vw
S V Vt

9 = volumetric water content

Vw = volume of water

Vs = volume of solids

Vy = volume of voids

Vt = total volume

The relationship between w and 8 is
w 1
TwTT x Yd x-I

w = gravimetric water content
Yd = dry density = -
Yw = unit weight of water

Ws = weight of solids



A- 3


200 -

> 150-




0 II I I
0 5 10 15 20 25


Figure 4.9 Gravimetric Water Content Versus Height Above Water
Table for A-3 Soil



A- 2-4



I- 200-


0 I I I
0 5 I0 15 20


Figure 4.10 Gravimetric Water Content Versus Height Above Water
Table for A-2-4 Soil
Table for A-2-4 Soil

A 2 -6

2 250 -

r 200 -
i I

o 150


I 100 -

50 -


0 II 12 13 14 15


Figure 4.11 Gravimetric Water Content Versus Height Above Water
Table for A-2-6 Soil




Figure 4.12 Gravimetric Water Content Versus Height Above Water
Table for A-5 Soil







In the testing only a limited range of soil suction (to 1/3 bar,

about 11.3? feet of water) was necessary to represent the soil-water at

static equilibrium with a shallow water table. Because soil suction is

given in units of water head, it is eaual to the distance above the

water table at equilibrium condition.

The A-3 soil-water retention curve in Figure 4.9, shows a distinct

decrease in water content at a small height above the water table.

Between 30 and 45 cm, the specimen water content decreased from 20.27 to

10.63 percent. On the other hand, an almost constant water content was

reached at a height between 150 and 345 cm where the variation in water

content was only from 4.20 to 3.56 percent. This characteristic shape

is due to the greater percentage of the larger pore sizes in sand, which

would empty at low suction values. The hand compacted specimen and the

mechanically compacted specimen showed similar trends, but with a small

difference at low height above the water table. The difference could be

due to variation in the structure of the soil, resulting from the method

of compaction and differences in density. The term structure is used to

cover the size, shape, and arrangement of the particles and voids.

The A-2-4 soil-water retention curves, in Figure 4.10, show a trend

similar to that found in the A-3 soil, except that they have distinct

capillary saturations of 60 and 45 cm for the mechanically and hand

compacted specimens, respectively. The overall fine, uniform gradation,

and the presence of 11 percent fines (< 0.074 mm) in the A-2-4 soil

could be responsible for the development of the capillary saturation.

Again, the differences between the two curves are probably due to

variations in structure. Between 150 and 345 cm there was, again, no

significant change in water content.

The A-2-6 material, which is a well-graded soil, shows in Figure

4.11 a gradual and uniform decrease in water content with increasing

height above the water table. The difference between the mechanically

and hand compacted specimens was very pronounced, although the

difference in densities was the same as in the A-3 and A-2-4 soils. The

reason for the difference is the large number of layers in the hand

compaction (twelve) versus (three) in the mechanical compaction,

resulting in differences in the developed structure. The mechanically

compacted specimen did not reach the optimum water content (11.8

percent) within the testing range of 345 cm (11.32 ft). It would

require much higher suction to reach this value. This observation is in

agreement with what has been reported by Kersten (1944), and Janssen and

Oempsey (1980). They found in field studies that a significant number

of subgrade soils, which had clay contents, had water contents which

were above the optimum water content. The number of such cases

increased directly with increasing clay content. The A-2-6 soil tested

contained 21.9 percent clay.

The characteristic curve for the A-5 soil, which is predominantly

silt, is shown in Figure 4.12. The water contents remained near

saturation up to a height of 200 cm. Then a relatively rapid reduction

in water content took place. This specimen also did not reach optimum

water content (36.4 percent) within the range of the test (1/3 bar,

11.32 ft). The A-5 soil also has 21 percent clay content. hue to an

experimental problem with the last reading at 345 cm for the A-5 hand

compacted specimen, the water content was not considered in the


For both the A-2-6 and A-5 soils it seems that the method of

compaction and the number layers, used to obtain the desired density,

had great effect on the shape of the characteristic curve and the amount

of water retained at any height above the water table.

Figure 4.13 shows a cumulative graphical presentation of the soil-

water characteristic curves for all project soils mechanically compacted

to the Standard Proctor. The distinction in shape for each soil-water

characteristic curve is evident in Figure 4.13. The A-3 curve is

located to the far left followed by the A-2-4, then the A-2-6, and

finally to the far right, the A-5 curve. It was noted that the amount

of fines (< 0.074 mm) increased from left to right. These curves can be

helpful in recognizing the soils that are susceptible to large changes

in water content as a result of changing the position of the water


The mechanical behavior of subgrade soils is affected by the

presence of water and the development of capillary potential above the

water table. It is therefore essential in any study to simulate the

field condition for the subgrade soils as it exists, in-service, under

the pavement, and in equilibrium with the designated water table. The

soil-water characteristic curves provide the necessary parameters to

achieve this simulation.

4.7 Static Triaxial Compression Tests

Standard static triaxial compression tests were performed on all

project soils in order to define failure envelooes. Tests were per-

formed consolidated drained and under strain controlled conditions.

Soil specimens, which were 4 inches in diameter and 8 inches high, were

0 A-3
E0 A- 2-4
A 2-6
A -5

2 5 10 15 20 23 43 45 50



Figure 4.13 Retention Characteristic Curves for all Project Soils
Compacted to Standard Proctor









prepared hy mechanical compaction to the Standard Proctor at optimum

water content.

The triaxial tests provide the angle of internal friction, 4, and

the cohesion, c. Table 4.6 summarizes the testing results.

Table 4.6 Triaxial Static Compression Test Results

Confining Deviator Stress Angle of Internal
Pressure (peak) Cohesion Friction
a 01 a c C
Soil Type psi psi psi degrees

A-3 2 11.50
10 42.50 1.5 40.0
15 58.50

A-2-4 2 21.00
5 32.50
10 44.70 4 36.0
15 58.50

A-2-6 2 26.40
10 46.80 6 34.0
15 60.90

A-5 2 40.90
5 49.50
10 61.00 10 33.0
15 72.20


5.1 Column Study

To study capillary effects in the subgrade soil, full height soil

columns were prepared. Each column consisted of sections of lucite

cylinder, 5 5/8 inches 1.0. and 6 inches high, stacked on top of one

other to provide the desired height. Holes 7/3-inch in diameter were

drilled in each section for tensiometers which measure soil suction,

Figure 5.1. Each column rested on a base to which a tall cylinder could

be connected to provide upward flow saturation of the soil. A simple

inverted flask was used to keep the water table at a fixed level, as

shown in Figure 5.2. Figure 5.3 is a photograph showing the setup

during testing. Mercury manometers, Figure 5.4, connected to the

tensiometers measured the developed suctions.

The test setup was intended to represent the subgrade and

stabilized subgrade layers in a pavement system. The subgrade soil was

compacted to AASHTO T-99 and the stabilized subgrade to AASHTO T-180.

This was achieved by hand compacting a determined weight of soil into a

known volume using the tools shown in FiQure 5.5. The tensiometers were

inserted during compaction to ensure intimate contact between the porous

cup and the surrounding material. The primed tensiometers were then

connected to the mercury manometers. Each column was saturated from the

bottom up by connecting the bottom of the column to a tall cylinder.

The water level in the tall cylinder was maintained at the same level as

Criure !.

Lucite 'ol'n Sect


aure ^i.7 Inverted clask'K to '^aintain FixeH ater Level



niaure ;.3 Column Setup Durino Testinq

4f L "

'Aercurv Manometers -or 'Rasjrin. Soi Suction in tf)c. 'W ;Jlmn

Figure -.I

FiQure ;5. Hand Comrpaction Tools

the top of the soil in the column. The upward infiltration of water

into the soil simulates a rising water table in a pavement system due to

a heavy rain storm. This represents the worst condition expected in the

field. When the tensiometers in the column indicated zero tension, the

column was considered to be saturated. Free drainage was then allowed

to occur until the soil reached a condition of static equilibrium with

the water table at the bottom of the column. The drained water was

allowed to overflow, while the water table was maintained at a fixed

level using a simple inverted flask. The water distribution in the soil

column at equilibrium is near saturation at the water table level and

decreases with height above the water table. When the soil is at some

water content less than saturation, it draws water through the porous

walls of the tensiometer cup. As a result, a negative pressure is

produced in the cup water and the mercury rises in the closed leg of the

manometer. This process continues until the soil and the cup water have

the same pressure deficiency or tension. The pressure deficiency, which

is numerically equal to the capillary potential, may be determined

directly by reading the height of mercury rise, above the mercury sump,

in the manometer leg.

In this study only the retention case, which is the more critical

condition, was considered. In it the soil at a certain height above the

water table would have a higher water content than in the wetting up

case. The higher the water content the less stable the subgrade. The

relationships developed between suction and water content will be

compared later to the results obtained using the Tempe cells. The

column study is described in detail in a thesis by Gallet, 1986, as part

of the same project.

5.2 Repetitive Load Testing

Specimen Preparation and Conditioning

The air dried material was sieved through a #4 sieve (4.76 mm) and

mixed with enough water to provide the optimum water content previously

determined in accordance with AASHTO T-99. The mix was then compacted

(5 equal layers, 26 blows/layer, 5.5-lb hammer, 12-inch drop) to give a

specimen size of 4-inch diameter and 8-inch height, Figure 5.6. A

lucite cylindrical mold was used so that the specimen could be observed

throughout the testing. These specimens, at optimum water content,

represent the subgrade as built. Repetitive loading triaxial tests were

performed on such specimens of all four subgrade materials to obtain the

as built deformation characteristics.

The water content at any location in the subgrade soil will not

remain at the optimum compacted value, but will change until it comes

into equilibrium with groundwater conditions. In this research only the

more critical case of a draining soil after saturation was considered.

Section 5.1 described this for the column testing. To prepare samples

for repetitive loading triaxial testing, specimens were prepared as

described above for the optimum condition, then conditioned to bring

them to the desired water content and soil suction state.

The lucite mold containing the compacted specimen was first fitted

with a perforated base, and filter paper at the bottom and a collar at

the top. Threaded rods held the collar and base together. The mold

assembly was then placed in a deep sink in which water was allowed to

rise to the top of the mold and was maintained at that level. Water

could enter the specimen only from the bottom. The mold top was covered

with filter paper and a perforated light weight plate. When water drops

Figure G.6

Lucite Mold and Comoaction Eauioment

were observed on the perforated top, the cover and filter paper were

removed and the specimen visually inspected. If a glaze of water was

observed on top of the specimen, it was considered to be saturated. The

A-3 (sand) took only a few hours to saturate. On the other hand, the A-

2-6 (Clayey sand) took almost P days. Figure 5.7 shows a mold assembly

in the sink during saturation. Following saturation, the water in the

sink was allowed to drain slowly, simulating water drawdown in the

field. The mold assembly was then lifted out of the sink and the excess

water allowed to drip for at least 30 minutes. Figure 5.8 shows a mold

assembly during this phase. When no more dripping was observed, the

assembly was taken apart and the mold, filled with the saturated soil,

was weighed. Knowing the weight and volume of the empty mold, the unit

weight Y of the specimen could be determined. Since the dry unit

weight Yd was also known, from the initial compaction, the water content

could be determined based on

d 1 + w

Knowing the water content (w), the void ratio (e), and the specific

gravity (Gs) of the soil, the degree of saturation (S) could then be


Gsw = Se

This was the assurance that the specimen was saturated. When

satisfactory saturation was achieved (S greater than 98 percent) the mold

was fitted with a pressure plate (rated at 1 bar) a lucite cup with two

drains, a small reservoir and a water manometer. 0-rings and silicon

grease were used to seal any gaps between the mold and the lucite cup.

The upper end of the mold was covered with a perforated lucite plate to

allow air to enter the specimen. Connecting rods were used to hold the

-101 -

figure 5.7 o01 Asse '1v in the "neen Sink nurinn Saturanion

Fiure 5. ''nold Assembly 1urrine ree 'rainaqe

cuo, mold, and the cover together as one unit. The arrangement described

above is essentially a large Tempe cell, except it uses a negative column

rather than air pressure to condition the specimen. Figure 5.9 shows the

components of the conditioning cup. The idea behind the conditioning cup

is that the saturated specimen is subjected to a negative pressure

through the pressure plate and the hanging column. When equilibrium is

reached, the water content of the specimen is representative of the water

content of a similar specimen in the subgrade at a height above the water

table, equal to the length of the negative column. Figure 5.10 shows a

specimen during conditioning.

When the specimen, under the effect of the applied, negative column,

ceased to drain water for 48 hours, it was considered to be at


The mold assembly was then taken apart and the wet unit weight and

water content determined as before. At this point, a decision was made

on whether the specimen had actually reached the target water content.

The target water content was obtained from the soil-water characteristic

curve, that is, the corresponding water content at a certain height above

the water table. If the water content was not satisfactory, the specimen

was refitted with the conditioning cup and the process repeated until the

target water content was obtained. At that point the specimen was ready

for the repetitive load testing.

Due to the size of the specimen, 4-inch diameter and 8-inch height,

and the limited distance above the water table to be investigated as a

fill height, attaining the target water content throughout the entire

specimen was impractical. For this reason, it was decided that if the

middle 4 inches of the specimen was within 1 percent of the target value,

eiqure 5. Components of the uConditonifn CuD



iiure 5.10 Soecirmen Durina Conditioning

the specimen was considered acceptable. Achieving this criterion

required considerable time (up to 45 days in some cases).

Repetitive Loading Equipment

The repetitive loading equipment consisted of a triaxial chamber

capable of accommodating a 4-inch by 8-inch specimen, vertical and

lateral deformation measuring equipment and a vertical loading source.

The external loading source was a closed loop electro-hydraulic system

manufactured by MTS Systems Corporation. An electronic load cell

measured the axial forces. The axial and radial deformation measurements

were made using two pairs of linear variable differential transformers

(LVOT's). These were mounted on a pair of expandable clamps which

contacted the specimen through four wide feet on each clamp (see Figure

5.11). Air was used as the chamber fluid and was monitored with

conventional pressure gauges. Signal excitation, conditioning, and

recording equipment provided for simultaneous recording of the axial

loads and deformations. The LVDT's were wired so that the average signal

from each pair was recorded. Figure 5.12 shows an overall view of the

equipment used. Figure 5.13 shows a close up to the MTS System, load

cell and the triaxial chamber.

Testing Procedure

The objective of the repetitive load testing was to characterize the

permanent deformation of the project soils at both optimum and varied

water conditions. The testing procedures and equipment were the same in

both cases. The specimen was extruded using a hydraulic jack, then

transferred to the triaxial chamber base plate. Porous stones were used

Figure %.11

CliamDs With L',"T's for "easurinn Lateral and Vertical

ioure 5.12 An Overall View of the enoetitive Loadin4n Test

- "1

Figure 5.13 The MTS System, Load Cel and Trixial Charter

on top and at the bottom of the specimen. A rubber membrane was placed

over the specimen, using a membrane stretcher, and secured to the top and

bottom platens using 0-rings. The membrane was marked with a felt-tip

pen at the 2- and 6-inch heights on opposite sides (1800 apart) of the

specimen. This was to aid in the placement of the clamps at the ends of

the middle 4 inches of the specimen. The diameter of the specimen was

measured, using a long-jaws caliper, at the 2-, 4- and 6-inch heights and

the specimen tilt was checked using an air bubble level. The LVDT clamps

were then placed at the 2- and 6-inch heights (see Figure 5.14). The

vertical and horizontal LVDT's were zeroed and the chamber cylinder

placed and connected to the chamber base plate using 6 tie rods. The

loading piston was inserted through the top of the chamber cylinder and

the load cell lowered until full contact with the loading piston was made

through a steel ball. A confining pressure of 2 psi was applied for 30

minutes under drained condition.

The MTS controls were set to give a 0.1 sec load on, and 0.9 sec

load off. The pulse wave was haversine. Knowing the specimen diameter

and the stress desired, the load could be calculated and the controls set

accordingly. The confining pressure was maintained at 2 psi in all

tests. Dynamic conditioning of the specimen was performed to eliminate

the end imperfection of the specimen, to allow for better seating of the

porous stones, and to eliminate the effects of the interval between

compaction and loading. This dynamic conditioning consisted of first

applying 200 load repetitions of a 2 psi deviator stress (aI = 4 psi

and o3 = 2 psi). The axial load was then incremented by 2 psi after each

200 repetitions until a total axial load of 8 psi was applied, deviatorr

stress ad = 6 psi). Specimen testing was started at the end of the 600

LVr)T lamps MaceI

ver the Secmen

Figure 5.14

repetitions of conditioning with same stress as the last 200

repetitions. The stress level of ad = 6 psi and 3 = ?2 psi was

considered representative of the stresses encountered in the field by the

subgrade layer. Stresses due to the surface load were conservatively

calculated using the Boussinesq equations at a depth of 25 inches below

the pavement surface. An 18 kip axle loading and 100 psi tire pressure

were assumed. The resulting stress increments were A. = 6.5 psi and

Auh = 0.13 psi. To these were added geostatic stresses, and the totals

rounded off to give 8 psi vertical and 2 psi horizontal. All specimens

were tested for 10,000 repetitions. Test data were recorded at or near

the following repetition levels; 1, 10, 100, 200, 400, 600, 800, 1,000,

then every 1,000 thereon up to the termination of the test. Test data

monitored included axial load, axial deformation and radial

deformation. These were recorded on a strip chart recorder manufactured

by the Gould Brush Company.


6.1 Introduction

This chapter discusses the results obtained during the course of the

testing program. The principal objective of the test program was to

characterize the permanent deformation of the four subgrade soils under a

variety of water conditions. All four soils were triaxial repetitive

load tested at optimum water contents. In addition, the A-3 and A-2-4

soils were tested under three different water retention conditions,

corresponding to three different heights above the water table. These

heights and water contents were selected from the respective soil-water

retention curves. The A-2-6 and A-5 soils were tested at only one

additional water retention condition. In these soils, because of their

high capillary fringes, there was no significant change in water content

up to a height of 11 feet above the water table level. The selected

water retention conditions were chosen, based on the range of economical

fill height above the water table. Table 6.1 summarizes the test program


6.2 Deformation Characteristics at Optimum Water Content

The relationships between axial permanent strain, a and number of

stress applications, N, were established for all the project soils.

Tables C.1 through C.4 in Appendix C provide the test results. These

include the permanent and resilient axial and radial strains, resilient

modulus, and resilient Poisson's ratio.

Table 6.1 Test Program of Project Soils Under Varied Water Content

Height Above Soil Type

Water Table A-3 A-2-4 A-2-6 A-5

in inches Water Content Percentage

15 13.24
18 10.53
24 8.10
30 11.70
36 10.50
48 8.70
0 12.20
36 44.40

The resilient modulus was defined as the deviator stress divided by

the axial resilient strain, while the resilient Poisson's ratio was

defined as the radial resilient strain divided by the axial resilient

strain. Figure 6.1 shows an arithmetic plot of the number of stress

applications versus the accumulated axial permanent strain for all four

project soils. Figures 6.2 through 6.5 provide semilogarithmic plots of

the number of stress applications versus both the total and accumulated

permanent axial strains, for the A-3, A-2-4, A-2-6, and A-5 soils,


The A-3 soil was compacted according to Standard Proctor to provide

optimum conditions of dry density equal to 103.1 pcf and water content

equal to 14.8 percent. These were therefore the initial conditions prior

to the repetitive load testing. The water mixed with the dry material is

essentially free water for a granular free draining material such as the

A-3 soil. When the confining pressure was applied to the specimen as

part of the conditioning procedure, the specimen drained 113 grams of

water. The end of conditioning water content was therefore 10.4 per-

cent. By the end of the actual testing, an additional 54 grams of water

had drained, resulting in a final overall water content of 8.6 percent.



1 )


l < -c


CI) z:






\\_0 x 'N IdN I^ L3V/E~ VX


0 !




U* ) 0
\- <.n

Oc =


I -

o I -o

0-01 )Nl 83d NI NIVUIS "IVIXV



a- --

M 4



SOD w to IT to ..

01 x 83d -N NIV1S 1VIXV
\\~ r 5
\ \ ^ <

\\ y
\\o /
\\ ft- "
\\ *Q
\\, ::3

\.- \ <
I \ -


.^ I



_- s-


oW E



\0 4-

a E

S41 L)


01 XNI N3d "NI: NIVIJS -r-xv






,0 0.


C E) "

0 n
-'- '

- (
< ,-,

t,O01 x NI d3d 'N1 = NIV8IS -IVIXV

< @ 0

o 0

The plot for the A-3 soil in Figure 6.1, showing the relationship between

accumulated axial permanent strain and number of stress applications, is

typical of such plots reported in the literature. Two stages can be

recognized in this relationship. The first stage shows a high rate of

deformation and can be represented by a power function of the form
:a = AN3

ea = accumulated axial permanent strain

N = number of stress applications

A,B = constants (which would represent intercept and slope
terms, respectively, on a log-log plot)

This high rate of deformation then decreases, leading to the second

stage where the deformation becomes almost constant and can be

represented by a semilogarithmic function of the form

Ea = A + 3 log N

Data from both stages were fitted using least squares procedures. There

was a transition ooint between the stages at 1,000 stress applications.

Figure 6.2 shows plots of the total accumulated axial strain, which is

the sum of the accumulated axial permanent strain and the axial resilient

strain, and accumulated axial permanent strain versus number of stress

applications. The total and permanent strains followed almost the same

trend, indicating the small variation in resilient strain. The range of

values for the resilient modulus (MR) was between 24,615 and 43,63A psi

with a value of 27,429 psi at 1,000 stress applications. The wide range

of values is due to the water draining during the repetitive load

testing. The A-3 specimen became stiffer toward the end of the test and

the values of MR stabilized. It is to be noted that all strain values

were read from the strip chart recorder by eye using a magnifying glass,

therefore, they were subject to a small variation considering that each

division on the chart represented 0.0005 inch.

Early in the testing program, it was realized that some

inconsistencies existed in the measuring of the radial strain. This was

believed due to the clamp feet digging into the specimen. Two reasons

for this were, the small contact area between the clamp's feet and the

specimen, and the elasticity of the rubber bands used to pull the two

halves of the LVDT's clamps together. An effort was made to reduce these

problems by increasing the width of the feet and using rubber bands with

appropriate elasticity. Radial strain measurements, however, were still

not satisfactory all of the time. This resulted in resilient Poisson's

ratios beina higher than 0.5 in some cases. In all tests, physical

inspection was made to assure that there was no indication of the feet

digging into the specimen. It was finally decided that only the axial

strain could be considered in this study. Similar testing problems have

been reported by Monismith et al., 1975, and by Brabston, 1982.

The A-2-4 soil was compacted according to Standard Proctor with a

dry density of 107.2 pcf and a water content of 11.8 percent. In this

soil there was no drainage of water out of the specimen during the entire

test. This was to be expected because of the fine gradation and the

presence of 11 percent Fines (passing #200 sieve). Figure 6.1 shows the

almost identical axial permanent strain number of stress applications

relationships of the A-2-4 and A-3 soils. The MR values for the A-2-4

soil ranged between 19,750 and 25,622 psi with a value of 24,947 psi at

1,000 stress applications. These values are a consequence of the higher

resilient strains, which can be observed by comparing Figures 6.2 and


The A-2-6 soil was also compacted according to Standard Proctor and

resulted in a dry density of 120.3 pcf and a water content of 11.6

percent. No water drained from the specimen during the entire test. The

A-2-6 plot in Figure 6.1 shows two distinct stages. The first stage

could be represented by a power function as before, except that there

existed a very shallow slope at the start up to 1,000 stress

applications. This then increased from that point up to 7,000 stress

application. In the second stage, between 7,000 10,000 applications,

the specimen response was a plateau with no increase in the axial

permanent strain. Figure 6.4, at a much enlarged scale, shows that in

the beginning of the test the axial resilient strain was much larger than

the axial permanent strain. However, by 1,000 stress applications,

almost equal values were obtained. The resilient strain continued to

decrease until it reached a constant value between 7,000 and 10,000

applications. This A-2-6 soil is customarily used as a base course in

Chipley, West Florida, compacted to Modified Proctor. It is not

therefore surprising it performed so well even when compacted only to

Standard Proctor. The well graded distribution of the grain sizes

results in high dry density and low void ratio, which in turn contribute

to less permanent deformation. The MR values ranged between 28,235 and

60,000 psi with a value of 38,400 psi at 1,000 stress applications. The

reason for the wide range of MR values could be attributed to re-

orientation of particles from the beginning of the test up to 5,000

repetitions. The MR values from 6,000 to 10,000 repetitions were

constant, as if the specimen had reached a stage of threshold. The MR

value of 60,000 psi is attributed to the low stress values used in the

repetitive loading compared to the peak deviator stress at static

triaxial failure. This ratio was approximately 0.25 (repetitive deviator

stress/peak deviator stress at failure, o/24-4, see Table 4.6). Both

terms in the above ratio are based on confining pressures, 03, of 2 psi.

The reason for the wide range of MR values is that the stress level

applied is too small for such material. It appears that it will take

larger number of repetitions to reach a constant value. Again, the

values between 3000 and 6000 repetitions were read off the strip chart

and in this range it was very difficult to precisely estimate the

difference in resilient strain values.

The A-5 soil was compacted to Standard Proctor at a dry density of

72.0 pcf and a water content of 37.84 percent. This soil showed similar

behavior to the A-2-6 soil, except that the first stage had a much

steeper slope, as shown in Fioure 6.1. Figure 6.5 shows that initially

the axial resilient strain was much higher than the axial permanent

strain. By 3,000 stress applications, equal values were observed. A

plateau was then reached between 7,000 and 10,000 applications. MR

values ranged between 10,213 and 12,000 psi, with a value of 10,909 psi

at 1,000 stress applications.

Table 6.2 summarizes the prediction equations, their coefficients,

and the R square (R2) values for the four project soils tested at optimum

moisture content conditions.

6.3 Deformation Characteristics at Varied Water Retention Conditions

To determine the soil water-pressure conditions for the second phase

repetitive triaxial tests, it was necessary to model the water retention

conditions as they would exist in the subgrade soil, in-service, at

equilibrium with the designated water table. To accomplish this, two

E 0a C C Ce C c\J 0a
c'3 J c C c C C' C 0% C"

S* * *



i w r. 0a r- .-4 CC
a- Q- I.r Ci' C, Lr )M
*- cC O C0
o ,- C ', c (') C c
a0 ) V) M 0


0 4-' I I

r- S 0 C- 0 S 4 S


ua I

LL 0 C C 0
-- 4- 'L '--

c -

C *r -J

a) C) C
u ca
I-.,-- 3 ) .



, I I I I I I
c- -4 4 *-


L J1
a) C C0

m .,,o F..-c -.


studies were made, one using soil columns and the second using Tempe

pressure cells. The soil columns, described in Chapter 5, represent a

one-to-one simulation of the pavement substructure. Subgrade, stabilized

subgrade and even the base course can be prepared as they would occur in

the field. A profile of water content and soil suction can then be found

experimentally for a particular pavement subsurface. Soil columns up to

4 feet in height have been employed. The Tempe pressure cells on the

other hand provide a single point value, and have been used only for the

subgrade materials. Since small volumes of soil are used, these tests

are much more rapid than column testing. Conditions representing up to

11 feet of subgrade above a water table have been modeled. Results from

the two procedures compare reasonably well, as shown in Figure 6.6 for

the A-3 soil.

Testing with the Tempe cell had the following advantages:

1. Obtaining full saturation was much easier and faster in the
Tempe cell than in the column experiments.

2. Reaching equilibrium was also much faster due to the smaller
size specimen.

3. The water content obtained from the Tempe cell was representa-
tive of the entire specimen. That obtained from the column
study was representative only of a localized area around the

4. Due to its small size, the Tempe cell does not experience any
effect due to temperature. This could pose a problem in the
4-foot column. Temperature variations leading to a scattering
of test results have been reported by Spangler and Handy (1982).

5. A large number of soils can be studied in a smaller space and in
a shorter time.

6. Specimens can be retrieved from actual pavement profiles and
studied using the Tempe cells.

The testing of the project soils, under varied water retention

conditions, was limited to a range between I and 4 feet above the water



A- 3
300 0- HAND TOOL


2 250-


a= 200-

o 150


I 100


O 1 I I A I
0 5 LO 15 20 25


Figure 6.6 Gravimetric Water Content Versus Height Above Water
Table for A-3 Soil From Both Tempe Cells and Column

The A-3 soil was tested by repetitive loading under three different

water retention conditions. These had water content values of 13.50,

10.53, and 8.10 percent representing 15, 18, and 24 inches above the

water table, respectively. Tables C.5 through C.7 contain the deforma-

tion characteristics data collected from these tests, respectively.

Figure 6.7 shows a cumulative plot of axial permanent strain ea versus
number of stress applications N for the three water retention condi-

tions. The curves stack, with the one representing 24 inches at the

bottom, that representing 18 inches in the middle and, the one

representing 15 inches at the top. For a certain number of stress

applications, the bottom curve yields the lowest axial permanent

deformation and the top curve the highest. A power function, as

presented before, was found to fit all the data in all three tests.

The A-2-4 soil was also tested under three different levels of water

retention. Water contents chosen were 11.70, 10.50, and 8.7 percent,

representing 30, 36, and 48 inches above the water table. Tables C.8

through C.10 contain the deformation characteristics data collected from

these tests, respectively. Figure 6.8 provides the plots of a versus

N. Again, as expected, the plot representing the greatest distance, 48

inches, was located at the bottom of the figure, that representing 36

inches was in the middle and the 30 inches plot was on top. The A-2-4

soil showed behavior very similar to that of the A-3 soil. Dower

functions again were used to fit the data.

6.4 Deformation Characteristics at Selected Water Retention Conditions

The A-2-6 and A-5 soils were each tested at only one selected water

retention level. The A-2-6 specimen was prepared by soaking in water for



S- C

\\ 0




4C ) -

S S-0 r
Q- 4-


u c
10 4-' '

t.INI 83d 'NI NIVIS N3NV 3d 7VX
(0 E 0


0 00


R 0

mo o

o o

-- ,





r- 0


_ o a c
S d 4-)

\0 \ *

\0 \' 0
\0 1 i--