PERFORMANCE AND DESIGN OF EXPANSIVE SOILS

AS ROAD SUBGRADE

BY

MAGDI MOHAMED ELTAYEB ZUMRAWI

A thesis submitted for the degree of

Doctor of Philosophy

Highway Engineering Institute

Chang'an University

Xi'an

August 2000

1

ABSTRACT

The construction of roads on naturally occurring

expansive soils has been generally avoided due to their

high potential to swell that will produce significant

volume changes and uplift forces on the pavement layers

and have relatively low strength values. The prediction

of volume changes, uplift forces and strength of these

soils is also complicated by the fact that, these soils

are affected by the boundary conditions.

The aim of this research was to investigate the swelling

behaviour and strength characteristics of expansive soils

by means of various laboratory tests such as an oedometer

cell test, CBR test and triaxial compression test.

Three types of soils with different plasticity have been

compacted at a wide range of placement conditions (i.e.

water content and dry density) then using laboratory

equipment to enable the measurement of the swell percent

and swelling pressure as well as CBR and shear strength

of these soils.

The results show that the swelling behaviour of these

soils is controlled by the surcharge pressure imposed on

the soil as well as the initial state of the soil as

described by the dry density and water content. On the

other hand, the strength of these soils is greatly

influenced by the initial dry density and water content

of the soil as well as the testing conditions.

Analysis of the experimental data indicated that it is

possible to combine the initial dry density and water

content in a way reflecting the influence of each of them

on the swell percent and swelling pressure as well as CBR

and shear strength. Therefore a new concept was

developed; this is called the placement condition factor.

On basis of this concept, a chart of swell percent and

swelling pressure – placement conditions behaviour as

well as CBR and shear strength – placement conditions

behaviour was developed. The use of these charts in

conjunction with conventional methods for pavement

2

design, pavement evaluation and compaction control on

expansive areas was demonstrated.

16

CHAPTER ONE

INTRODUCTION

1.1 INTRODUCTION

Expansive or swelling soil is a kind of high plasticity clay

that contains mainly clay minerals of kaolinite, illite and

montmorillonite. This soil has clay content (i.e. soil

fraction less than 2m) of more than 30% and known to have high liquid limit and low plastic limit, therefore the

plasticity index is high. Expansive soil has considerable

ability of swelling and shrinking deformation; swelling when

absorbs water and shrinking when losing its water. As the

moisture content of this soil increases, the volume will

expand producing uplift forces. On the other hand,

decreasing in moisture content, the volume shrinks and

cracks will appear. A dry mass of this soil is stiff, has

high strength and wide cracks that appear in the vertical

and horizontal directions of the soil mass, all these

changes will reduce in wetting state.

Expansive soils are widely distributed all over the world in

more than forty countries in the six continents. In these

areas expansive soils are known by different names such as

"black cotton soils" of central and eastern Africa and

India, "black earth" of Australia and northern Africa,

"swell-shrinking soils or cracking soils" of China, "problem

soils" of Japan, "London clay" of England and "hidden hazard

and grumusols" of USA.

Expansive soils occur extensively in tropical climates and

known locally as black cotton soils. In many areas where

expansive soils occur, there are no suitable natural gravels

or aggregates. Most of the tropical black clay deposits are

formed residually by weathering of the basic rock basalt and

rivers alluvial deposits which cover such large areas that

avoiding or by-passing them is not feasible.

Recently there is a wide interest in studying expansive

soils; this is due to the problems caused by these soils in

many countries. Expansive soils when used as a foundation

17

material for engineering structures such as roads, railways,

airfields, light buildings and others will suffer damages

due to excessive volume changes caused by variation of water

content. According to records, the economic losses caused by

expansive soils as a foundation material exceed five

billions dollars annually in the world.

In road construction, specifications and standards normally

do not permit expansive soils to be used in embankments or

fill below formation level, moreover excavation below the

embankment of expansive soil and replacing it by better

quality materials may also be required. This is because of

the large volume changes that can occur as a result of

wetting and drying with seasonal changes of climate. Few

roads constructed through these areas have proved

satisfactory, but the majority especially in Africa and

southern Asia has failed completely.

In Sudan black cotton soils cover a wide strip along the

river Nile and its tributaries (see map 1.1), in these areas

are located most of the main economical and agricultural

schemes and production factories. Most of our express

highways are constructed on black cotton clays. These major

engineering structures consequently suffer from series

distresses and damages. This due to the fact that expansive

soils exist as road subgrade when absorb water will expand

producing high uplift forces on the above road layers, on

the other hand when losing water expansive soils will shrink

and crack, as a result cracks usually appear on road

surface. These severe cracks subsequently permit water to

enter the road’s structure and create a weakness, which eventually causes road failure. In Sudan, according to

records, the annual estimated economical losses caused by

expansive soils exceed a million-dollar.

1.2 OBJECTIVES OF RESEARCH

Construction of roads often requires the use of granular

soil materials as road subgrade. It is a common practice to

remove away expansive soils and replace them by granular

materials of high strength and low potential to swelling or

using technical methods developed by researchers such as

18

soil stabilization method to reduce the swelling potential

of expansive soils. However, the practice of removing

expansive soils or using technical methods is becoming more

costly especially in undeveloped countries. When considering

expansive soils as road subgrade, one of the main concerns

of the design engineer should be concentrated on the

behaviour of the subgrade soil during both the construction

and design life of the road. Expansive soils have high

potential to swelling that will produce significant volume

changes and up lift forces. Therefore it is of great

important in design of roads on expansive soils to consider

both the strength and swelling behaviour of this soil as

main factors of design.

The main objectives of the research were the followings:-

To study the swelling potential including the swell

percent and the swelling pressure of expansive soils and

the main influence factors such as the soil composition,

initial state parameters (i.e. water content, dry density

and void ratio) and the applied surcharge pressure. Among

these factors, the soil initial state factor is the most

important one and is the main objective of the study.

To study the strength characteristics (i.e. CBR and

shear stress) of soils and the main influencing factors

including the soil composition and plasticity factor, the

initial state factor and the soaking state factor. The

initial and soaking state factors and the are the most

important factors and are considered as the main

objective of the study.

To develop the soil states factors considering both the

initial state and soaking state. These factors are formed

mainly by a combination of water content and dry density

in the different states, depending on the study analysis

results.

To relate both the swelling potential and strength

characteristics of compacted expansive soils to the soil

states factors.

19

To investigate the dependence of the swelling potential

and strength characteristics on the soil composition and

plasticity index of the soil as well as the testing

conditions such as applied surcharge pressure, cell

pressure and soaking conditions.

To develop on basis of the above relationships,

empirical chart that can be used in design, evaluation

and field compaction control of expansive soil subgrade.

The experiments have been conducted on a high plasticity

expansive soil of Southern Shaanxi clay in the north west of

China. The apparatus used for the experiments were:-

An oedometer cell: In this apparatus, the sample was

restrained laterally by the confining ring and allowed to

swell and compressed vertically by a weight on the hanger

system. This instrument was used for measuring the swell

percent at three different surcharge loads as well as

measuring the swelling pressure of the soil tested. After

swelling completed, loads were gradually placed. The

final pressure that retained the sample to its original

volume was taken as the swelling pressure.

A conventional triaxial cell: In this project the

unconsolidated undrained compression tests of compacted

expansive samples were performed. The test was relatively

easy to perform and used to give some indication of the

soil tested shear stress.

A standard CBR: This is a well known test in road

engineering and is used for measuring the CBR as soaked

or unsoaked. The soil samples were compacted at a variety

of water contents and dry densities and tested at

different boundary conditions.

1.3 STRUCTURE OF THESIS

The first part of this thesis considers those topics, which

are relevant to the study and form the theory review of the

research project. In the second part the experimental work

performed is described with an emphasis on the procedures

20

adopted in the sample preparation and testing methods. The

presentation of the results, analysis and discussion

critically examine the validity of the developed soil states

factors and the empirical chart that can be utilised in

design of expansive soils as road subgrade were clearly

demonstrated.

The following is a brief preview of the topics considered in

the various chapters of the thesis: -

Chapter 2 provides the theory review of the thesis. The

topic of swelling potential is introduced and the major

factors influencing it are described. This chapter is also

described the soil strength and the factors influencing it.

Finally a brief review of some technical methods adopted in

design of roads on expansive soils and the conventional

flexible pavement design methods are considered.

Chapter 3 covers the experimental work performed in the

laboratory both in terms of laboratory apparatus and testing

procedures; in order to measure the swelling potential

including the swell percent and swelling pressure as well as

measuring the soil strength such as the CBR and shear

stress.

Chapter 4 contains the results and analysis of the swelling

and strength data obtained from the laboratory experiments

of the soil samples studied and those reported by previous

researchers. These analysed data are used to develop new

factors, which are called the soil states factors, then

using the previous researchers’ data to verify the validity of these developed factors. Finally this chapter presents

the analysis results obtained.

Chapter 5 presents the design chart developed on basis of

the soil state factor concept. The validity of this chart is

examined by the experimental results obtained. The final

section of this chapter describes the practical applications

of the developed chart in design and construction of road

subgrade on expansive soils.

21

Chapter 6 presents the conclusions drawn from the whole

research and a summery of the study findings. Presented

Proposes recommendations for the current specifications used

for construction and design of roads are given. Also

recommendations for the future researches works are given.

22

CHAPTER TWO

THEORY REVIEW

2.1 INTRODUCTION

The research undertaken uses expansive soils as compacted

and investigates the influence of the soil initial state

parameters such as water content and dry density on the

swelling and strength properties of the soil. Several topics

of geotechnical and highway engineering, which are relevant

to the objectives of this project, will be reviewed in this

chapter. The following is a brief description of these

topics:

Mineral composition of expansive clays, i.e. the

different clay mineral compositions and the related basic

engineering properties.

Compaction, i.e. the different aspects of soil compaction

and their influence on the mechanical behaviour of the

soil mass formed.

The swelling potential, i.e. the definitions of the main

swelling components "swell percent" and "swelling

pressure" and the factors influence them.

California Bearing Ratio (CBR), i.e. the development and

applications of CBR on road design and the main influence

factors.

The shear stress, i.e. the definition of the soil shear

stress and the affecting factors.

Road design on expansive soils, i.e. the conventional

methods of pavement design as well as the various

technical methods developed by researchers to be used as

practical solutions for the problem of expansive clays as

road subgrade.

23

2.2 THE COMPOSION OF EXPANSIVE CLAY

The clay minerals in general are very small (less than 2m) and very electrochemical active particles which can be seen

using electron microscope only with difficulty. Clay

particles consist of platelets whose surfaces are negatively

charged. To maintain chemical equilibrium cations in the

soil water such as sodium (Na+), calcium (Ca2+) or magnesium

(Mg2+) are attracted to the surface of the platelets. The

concentration of these cations is higher near the surface of

the platelet than in the body of the pore fluid. Any

diffusion of the cations is prevented by the electrostatic

attraction of the clay surface. The surface of the platelet

where the positively charged cations are held is called the

diffuse double layer. The outer surface of the diffuse

double layer has a positive charge and maintains an

electrical equilibrium by attracting water molecules (which

are electro-statically polarised). As more water molecules

are attracted to the diffuse double layer, the clay

platelets are forced further apart until a new state of

equilibrium is reached with the ambient stress levels. This

increased distance between the clay platelets is observed on

a macroscopic scale as an increase in the volume occupied by

the soil.

The same process is reversible so that if water leaves the

pore fluid by, for example, evaporation an unbalance in the

electrostatic equilibrium is created. Water molecules that

were once attracted to the diffuse double layer become free

and leave in the form of vapour. Reduction in the water

molecules attracted to the diffuse double layer allows the

platelets to move closer together resulting in a decrease in

volume. The diffuse double layer then remains with a

positive charge (due to the cations) until other water

molecules (or anions) are attracted to it.

The various clay mineral types have different degrees of

expansion, which are dependent on the constituent atoms in

the clay crystal lattice. The kaolinite group of minerals

generally have low expansive properties, illites show some

expansion characteristics and montmorillonites (smectites)

are highly expansive. Figure 2.1 shows a comparative study

of the expansive properties of montmorillonite and kaolinite

24

(Grim (1962)). All of the samples were restrained laterally

and allowed to expand vertically under a constant stress of

7kpa (1 psi). It can be seen that different proportions of

the mineral can effect the amount of volume change that the

soil undergoes on inundation.

The pore water chemistry itself can also have a large effect

on the magnitude of volume change and swelling behaviour of

a soil. The type and concentration of the cations present in

the soil water affects the adsorption of water and

consequent swelling. For example the presence of calcium or

magnesium ions in the pore water produces less swelling of

the soil than sodium ions when present.

2.3 CLAY MINERALOGY RELATED TO ENGINEERING PROPERTIES

The engineering properties of soils are dependent upon the

particle size, shape, surface area, stress history as well

as the mineral composition. It is not possible to classify

all soils or clays in such a way that their mechanical

behaviour can be predicted solely on the basis of the clay

mineral composition. However the knowledge of the clay

minerals and their effects will become more important as

soils with higher clay contents are used more widely in

engineering.

2.3.1 Index tests

The index tests as proposed by Atterberg were adapted for

use in soil mechanics by Casagrande (1947) and are still

used as a convenient way of expressing the properties of

soils containing clay minerals. The test provides a

measurement of the water content of the soil at two

arbitrary strengths. The definitions and methods used for

the determination of the index properties are fully defined

in the current specification (BS 1377 part 4).

The plastic limit (PL) is defined as the minimum water

content at which the soil can be deformed plastically by

rolling into a 3 mm thick thread. The definition of the

liquid limit (LL) is the water content of the soil at which

a standard 80 gm 30 cone will penetrate a sample of soil

25

for a displacement of 20 mm when dropped under its own

weight. Plasticity index (PI) is the difference between the

liquid limit (LL) and plastic limit (PL) and is the range of

moisture contents over which the soil remains in a plastic

state.

1.2PLLLPI

The liquidity index (LI) of a soil is defined as the ratio

of the water content (w) minus the plastic limit (PL) all

divided by the plastic index (PI). The expression for

liquidity index is:

2.2PI

PLwLI

It is a way of expressing the natural water content of a

soil or clay in relation to its plastic and liquid limit.

The plasticity characteristics of a soil are related to the

amount of clay sized particles present. A relevant parameter

that reflects the amount of clay minerals present in the

soil is known as Activity (A) (Skempton (1953)). In general,

a soil with a high activity number has relatively high water

holding capacity and a low permeability; the converse is

true for a soil with a low activity.

3.22%

)(mclayofweightby

IndexPlasticityAActivity

It is possible to interpret empirically the Atterberg limits

in terms of the interaction of the water molecules with the

clay particles. At low moisture contents water molecules are

adsorbed to the surface of the clay particles and held in a

tight, well-orientated pattern. Insufficient water is

available and the molecules are not sufficiently free to

lubricate the soil particles. As the water content of the

soil increases the layers of water adsorbed to the diffuse

double layer increase. The outermost layers are less

attracted and are able to move more freely and lubricate the

movement of the clay particles under the application of a

small load. In this case the plastic limit can be

26

interpreted as the water content at which the surface of the

particles can adsorb just slightly more water than can be

held in a rigid condition. In a similar way the addition of

further water to obtain the liquid limit can be interpreted

as the water content at which the clay particles are still

able to retain a sufficiently strong attraction to the water

molecules. This prevents the soil from loosing rigidity and

becoming dispersed within the water.

Holtz and Gibbs (1956) used index tests for classifying

expansive soils as shown in table (2.1). Most expansive

soils are known by their high liquid limit, plasticity

index, shrinkage ratio and clay content.

Although the liquid and plastic limit values are based on

empirical tests they may be related to more fundamental

properties such as shear stress (Skempton and Northey

(1953)), angle of internal friction (Mitchell (1976)) and

compressibility (Wroth and Wood (1978)). For example, Figure

2.2 shows the relationship between liquidity index and shear

stress for several remoulded clays (Skempton and Northey

(1953)). As the water content of the soil moves from the

Plastic Limit to the Liquid Limit the strength and stiffness

of the soil decreases.

2.4 COMPACTION: DEFINITIONS AND INFLUENCES

A definition of compaction was introduced by the Road

Research laboratory in a widely used reference book on soil

mechanics for road engineers (Road Research Laboratory

(1952)):

"Soil compaction is the process whereby soil particles

are constrained to pack more closely together through a

reduction in air voids, generally by mechanical means."

The compaction of soil produces a material which has greater

shear stress and at the same time it reduces the propensity

for settlement and deformation as well as its permeability

to water. To use soil as an engineering material it is

necessary to understand the factors which can effect its

27

integrity both in the short and long term. The factors that

are of the greatest influence in the compaction of soil are:

The soil characteristics – grading, plasticity etc. The moisture content of the soil

The volume of the soil compacted

The amount of energy used to compact the soil.

2.4.1 Definitions

The measurements used to define quantitatively the

compaction of soil are: bulk density (), dry density (d), saturated density (sat), void ratio (e) and degree of

saturation (Sr). The relationships define the proportions of

solid, water and air within the soil. Useful definitions are

as follows:

Bulk density ()- is the ratio of the mass of a given volume of soil to the volume that the soil

occupies. Units are Mg/m3.

Dry density (d)- is the ratio of the mass of solid

particles in the given volume of soil to

the volume that the soil occupies. Units

are Mg/m3.

Saturated density (sat)- is the ratio of the mass of a given volume of saturated soil to the volume

that the soil occupies. Units are Mg/m3.

Void ratio (e) - is the ratio of the volume of voids in the

sample (water and air) to the volume of

solids. This term is dimensionless.

Degree of Saturation (Sr)- is the percentage of the volume

of water in the soil to that occupied by

the total voids.

28

2.4.2 Influence of soil type

For the input of a standard compactive effort the highest

dry densities are achieved by soils having low or no

plasticity, for example, sands and gravels. For these types

of soil greater densities are achievable when the soil is

well graded, rather than poorly graded. The soils which

exhibit plasticity due to the presence of silt and clay

particles are able to retain more water the lower the

achievable dry density for the same compactive effort.

Figure 2.3 shows the compaction curves (using the standard

2.5 kg rammer test) for a wide variety of soils ranging from

crushed limestone to high plasticity clay. It can be seen

that different soils subjected to the same amount and type

of compactive effort achieve quite different degrees of

compaction.

2.4.3 Influence of moisture content

The presence of water in soils has a large effect on the

strength of the soil. As water is added to a sand soil the

achievable dry density for a standard compaction will

increase until an optimum water content is reached after

which the achievable dry density reduces (Figure 2.4). This

peak in the dry density results from the way in which the

water forms menisci between the particles of sand creating a

negative pore water pressure (suction). At water contents

drier than the optimum the higher suctions result in an

increased effective stress within the soil, making its shear

stress greater and compaction harder. As more water is added

the suction within the soil decreases and higher densities

are achieved. At water contents greater than the optimum the

high suctions are reduced and during compaction the pore

water develops positive pore water pressures which, unless

they dissipate quickly, prevent the soil from becoming

denser. The same effects are observed in soils containing

silt or clay, however because the particles are smaller

(with larger surface areas) the suctions are much higher so

for the same water content these soils will have lower dry

densities following compaction. Increasing the water content

above the optimum produces a reduction in the dry density.

This reduction is due to the generation of excess water

29

pressures in the soil water which are unable to dissipate

sufficiently fast to allow further compaction.

2.4.4 Influence of volume compacted

As the thickness of the layer which is being compacted

increases there is a variation of shear stresses exerted by

the compactor throughout the layer this results in a

variation of the density of the soil with depth. In order to

achieve a more uniform distribution of dry density the soil

mass can be compacted in smaller layers (Figure 2.5). The

effect of layer thickness was demonstrated by Parsons (1992)

who interpreted results from a previous study, where it was

shown that an increase in the depth of fill produced a lower

dry density for the same compactive effort (Figure 2.6). The

soil used in this experiment was high plasticity clay

compacted in a standard way by an 8 tone roller.

2.4.5 Influences of compactive energy

The energy used in compacting a soil is directly related to

the achievable dry density. Generally as the compaction

energy input into a soil mass increases there will be a

greater reduction in the air voids of the soil and a higher

dry density will be achieved. The required amount of energy

can be applied by successive applications of a load such as

a weight dropping from a fixed distance or a roller passing

over the surface of the soil. After each application of the

load the dry density of the soil mass increases. During the

compaction process the stresses imposed by the roller can be

seen to increase as the successive passes reduce the

thickness of the top layer of soil and reducing the area of

contact that the roller has with the soil (Figure 2.7). At

the surface of the compacted layer the shear stress applied

can exceed the maximum shear stress of the soil and over

stressing can occur resulting in a decrease in density.

Figure 2.8 shows the change in the density observed when

compacting a granular soil with standard weight machine and

an excessively heavy machine for the same granular soil.

30

2.4.6 Effects of fabric in compacted soils

Lambe (1958) put forward a theory that the fabric of a

compacted clay changed with moisture content. When soil was

compacted at water content dry of the optimum the clay

particles would form a "flocculated" fabric (without

orientation of the particles), whereas if the clay was

compacted wetter than the optimum moisture content a

"dispersed" fabric (particles having an orientation) would

result. The idea, illustrated in Figure 2.9, formed the

basis of a thorough investigation into the fabric and

strength characteristics of compacted clays performed by

Seed and Chan (1959). A simplified model proposed by

Brackley (1975) considered unsaturated clay soils existing

as packets of soil particles, with each packet being

completely saturated and the inter-packet voids being filled

with air (Figure 2.10). This meant that the soil mass was

unsaturated whereas the individual soil packets were

saturated. By assuming that the packets were saturated,

Brackley (1975) developed the idea that the total volume

change of the soil mass would be due to the summation of the

effects of swelling or compression of the packets and their

shear behaviour.

From observations of electron micrographs of natural soils

McGowan and Collins (1975) proposed a classification system

for identifying the different types of collapsible and

expansive soils. They observed that soil microfabric could

be classified using three basic forms: "elementary particle

arrangements" – where groups of clay platelets are joined together to form elementary particles, "particle

assemblages" – where the elementary particles are arranged in aggregations or matrices with other larger sand or silt

grains and finally "pore spaces". The arrangement and

proportions of these three forms were indicative of the type

of behaviour of the soil. Figure 2.11 illustrates the types

of microfabric observed.

31

2.5 SWELLING POTENTIAL

The swelling potential of a given expansive soil can be

considered in terms of two main components "swell percent"

and "swelling pressure".

The swell percent can be defined as the ratio of the linear

(or volumetric) increase in the soil sample dimensions

relative to the initial dimensions, due to increase in water

content. The swell percent is usually measured under a small

vertical surcharge pressure (1 to 20 kPa).

The swelling pressure is equivalent to the pressure, which

must be applied to prevent swelling (or volume change) of

the soil sample when water is fed into it.

The factors that have the greatest influence in the swell

percent and swelling pressure of a given expansive soil are

summarized in the followings:-

The soil physical properties

The applied surcharge pressure

The soil initial state

2.5.1 Swell percent

The swell percent or volume change has been defined by many

researchers in different ways, such as:

Holtz (1959) defined the swell percent as the percentage of

total volume change of a soil when tested in an oedometer

such that its moisture content varies from the air-dry to

the saturated condition under a vertical surcharge pressure

of 1 psi (≈7kpa).

Seed et al (1962) stated that swell percent could be defined

as the percentage of volume change of laterally defined

sample compacted at optimum moisture content in the standard

AASHTO compaction test and a surcharge pressure of 1psi.

Snethen (1984) defined the swell percent as the equilibrium

vertical volume change or deformation from an oedometer-type

32

test, expressed as a percent of original height of an

undisturbed sample from its natural moisture content and dry

density to a state of saturation under an applied load

equivalent to the in-situ overburden pressure.

The swelling of a given expansive soil can be measured in

three different ways such as free swell, swelling under a

small load and swelling without loading the soil sample.

Among these methods the swelling under a load seems to be

the most popular used method for measuring the swell percent

of the soil.

2.5.2 Swelling Pressure

The term swelling pressure has been used by many researchers

in many different ways, and the definition is often

dependent in some way on the test method. Brackley (1973)

stated that swelling pressure could be defined as the

pressure required to hold the soil at constant volume when

water was added. He gave three different methods for

determining the swelling pressure but they produced

considerably different values.

The three methods are the consolidation test, the

equilibrium of void ratios and the constant volume test.

Each method will be described briefly:

Method 1: consolidation test – The specimen with a known initial thickness and dry density is placed in a

consolidometer and allowed under a seating pressure to swell

on the addition of water until equilibrium is reached.

Subsequently incremental loads are added and the specimen is

allowed to consolidate. The percentage volume change is

plotted against the log of the pressure on the specimen. The

point at which the curve intersects the zero volume change

line is the swelling pressure. This test results in an upper

bound for the swelling pressure.

Method 2: equilibrium of void ratios – Four to five samples with identical thickness, initial moisture content and dry

density are placed in oedometers under the same seating

pressure. The loads on the different specimens are increased

33

to different values and allowed to equilibrate. Water is

then added to the samples which are allowed to swell or

compress until equilibrium is again established. These

equilibrium positions are used to obtain (by interpolation)

the load under which the sample does not undergo volume

change on saturation. Brackley (1975) suggested that this

method followed the probable stress path that the soil may

undergo in the field where after construction the clay may

not be exposed to water for several months and after wetting

a long time may elapse before equilibrium is reached. This

test gives a lower bound to the swelling pressure.

Method 3: constant volume method – After placing in the

consolidometer and adding water, the swelling of the

specimen was controlled by the addition or subtraction of

loads so that there was neither swelling or compression

whilst maintaining a constant volume. The manual control of

this test proved difficult; the loading path was such that

there was always some swelling and small variations in

volume that could not be avoided. At near equilibrium the

addition of a small weight made the sample compress and

cross the zero volume change line. This intersection

represents the swelling pressure and gave a median value

compared to the other two methods.

Figure 2.12 shows the results from tests performed on black

cotton soil (PL = 60 %, LL = 98 %) by Sridharan et al (1986)

using the three methods discussed by Brackley (1975). The

three methods produced very different results. Method 1 was

allowed to swell fully prior to consolidation and provided

an upper bound for the swelling pressure of the soil. Method

2 had the merit of following the probable stress path that

the soil may undergo in the field, as discussed above. The

method produced the lowest value of swelling pressure but

three separate samples were requied. The last method (method

3) produced an intermediate swelling pressure value by

maintaining a small volume change during the inundation.

Johnson (1989) observed that the magnitude of the swelling

pressure depended on the degree of confinement of the soil

with the greater degree of confinement leading to increased

34

swelling pressure. Table 2.2 shows the various definitions

of swelling pressure in decreasing order of confinement.

2.5.3 Influence of soil physical properties

The soil physical properties include the mineralogical

composition of clay which has a great influence on the

swelling potential as explained by Li Bing et al (1992) and

shown in Figure 2.13. The various clay minerals types have

different degree of expansion (expansiveness) such as the

montmorillonite group of minerals have high degree of

expansion, illities are of moderate expansion and kaolinites

show low expansion. Expansive soils can be identified by the

determination of the constituent clay minerals in the soil

as indicated in table 2.1.

2.5.4 Influence of soil initial state

The initial state or the placement condition of a given

expansive soil as defined by dry density, water content and

void ratio has a great influence on the swell percent and

swelling pressure.

Influence of dry density

The initial dry density of a soil is strongly influencing

the swell percent and swelling pressure. The data of Chen

(1975) and Kassif et al (1965) as shown in Figures 2.14

and 2.15 indicated that, the higher the dry density the

higher the swell percent.

The swelling pressure is also affected by the variation

of dry density. As reported by Seed and Chan (1959) and

shown in Figure 2.16, the swelling pressure increases

with increasing in dry density.

Influence of moisture content

The presence of water in expansive soils has a large

effect on the swell percent and swelling pressure. Chen

(1975) reported that, the increase in the moisture

35

content of a soil produces a reduction in the swell

percent as shown in Figure 2.18.

The swell percent versus initial moisture content is

approximately linear if the initial dry density is fixed,

as done by Zein (1985) Figure 2.17.

O’Connor (1994) Figure 2.19 shows a higher plasticity soil with low water content produces a higher value of

swelling pressure.

Influence of void ratio

Soil is an aggregate of mineral grains in which the void

spaces are filled with water and/or air. A state of

equilibrium between the water and air phase exists which

governs the amount of air that can be dissolved in water.

The general mechanical properties of any particular

sample of soil are dependent on the proportions of the

soil, water and air present as well as the soil itself.

Figure 2.20 shows how the percentage of the three phases

in a soil change, as a sample that is saturated with air

(dry) becomes a sample saturated with water. The soil

that exists between these two extremes of saturation is

known as unsaturated.

The swelling is strongly influenced by the void ratio of

a given expansive soil sample. As reported by Brackley

(1975) and shown in Figure 2.21 that at a constant water

content the swell percent increases with decreasing in

void ratio, as can be seen the soil sample will have high

swelling at low water content and low void ratio values.

O’Connor (1994) Figure 2.19 shows the influence of three types of clay at different void ratios on the swelling

pressure. It is clear that, at a certain moisture content

value as the void ratio increases the swelling pressure

decreases.

36

2.5.5 Influence of applied surcharge pressure

The surcharge pressure under which the soil is tested

affects the measured swell percent value. Results reported

by Brackley (1975) as shown in Figure 2.22 and Rengmark and

Eriksson (1953) Figure 2.23 indicated that, other conditions

being the same, increasing surcharge pressure reduces the

measured swell percent value. In practice, the amount of

swell is reduced by the overburden pressure, the higher the

foundation pressure the smaller the swell percent due to

wetting up of the supporting soil as reported by Kassif and

Zeitlen (1960) Figure 2.24. When the overburden pressure is

just sufficient to prevent swelling, this pressure is termed

the swelling pressure.

2.6 SOIL STRENGTH

The strength of a soil as defined by Atkinson (1993) is the

largest stress that the soil can sustain and it is this

which governs the stability or collapse of structures. The

term strength usually refers to shear stress.

Yoder (1975) reported that, many tests have been devised for

measuring the strength properties of soils such as

California Bearing Ratio (CBR) test, triaxial compression

test and plate loading tests. The test results can be

correlated with field performance. As a result, many of the

procedures for testing have been standardized, which must be

strictly followed in order to obtain more reasonable

results.

In the present study the tests performed in the laboratory

to measure the soil strength are the CBR test and the

triaxial compression test. Therefore the soil strength will

be considered in terms of CBR and shear stress.

2.6.1 California Bearing Ratio

The California Bearing Ratio (CBR) test was originated by

Porter (1938) of the California State Highway Department,

since then has been modified and revised by various states

and federal agencies. The CBR test is the most widespread

37

method of determining the bearing strength of the pavement

materials and is fundamental to pavement design practice in

most countries. To use the CBR method in designing a

pavement, it is necessary to carry out a standard CBR test

and then, using values obtained from the test, an empirical

design chart is entered and the pavement layer thickness

required is read off the proper curve. The original test

procedure and design curves developed by the California

investigators have since been modified by various agencies

for their own purposes.

The CBR test is relatively simple and can be performed both

in the laboratory and field. It is essential that the

standard test procedure should be strictly followed (BS 1377

[110]). The CBR test may be conducted on remoulded or

undisturbed soil samples or on the soil in place. The

samples may be tested at their natural or as moulded

moisture content (unsoaked CBR), or they may be soaked by

immersing in water or flood for a specified period of time

(four days) in order to simulate highly unfavorable moisture

conditions of the soil type.

The CBR value of a soil tested is a ratio between the load

required to force a piston of 5 cm diameter into the soil

sample 2.50 mm penetration depth, and that required to force

the piston the same depth into a standard sample of crushed

stone (13.24 KN/m2). Therefore, the CBR may be considered as

the strength of the soil relative to that of crushed stone.

Yoder (1975) reported that, the CBR value in general

corresponding to 2.50 mm penetration is greater than at 5.00

mm penetration, i.e. the CBR will decrease as the

penetration value increases. In some cases, however, the

value at 5.00 mm penetration may be higher than at 2.50 mm

penetration if this happens the value at 5.00 mm penetration

is used.

The CBR test is probably the most widely used test to

provide the relative bearing value of subgrade, subbase and

base materials. The specifications and standards of pavement

design recommend CBR values of greater than 30% for subbase

and greater than 80% for road-base are generally specified.

38

The main factors affecting the CBR value of a soil as

reported by Murphy (1966), Yoder (1975) and Glanville (1951)

are grouped as follows:

The soil composition and Atterberg limits

The soil initial state

The soaking condition.

2.6.1.1 Influence of soil composition and Atterberg limits

The soil composition affects the CBR value. A soil contains

crushed stones or gravels has higher CBR while that contains

large quantities of clay minerals (i.e. expansive soil) has

very low CBR, Yoder (1975).

The basic engineering soil indices such as Atterburg limits

(i.e. liquid limit, plastic limit and plasticity index) much

affect the CBR value. Soils have relatively high liquid

limits and plasticity indices such as expansive soils have

very low CBR values, while those have relatively low liquid

limits and plasticity indices such as crushed stones or

gravels have high CBR values.

2.6.1.2 Influence of soil initial state

The soil initial state as indicted by moisture content, dry

density and void ratio have great influence on unsoaked CBR

value. The available data of the previous researchers

suggests the followings:

The unsoaked CBR is strongly influenced by initial

moisture content; the higher the moisture content the

lower the CBR. This was proved by Glanville (1951) and

TRL team (1981) and indicated in Figures 2.25 and 2.26.

The CBR and moisture content relationship is

approximately linear, if the dry density is fixed as

shown in Figures 2.25 and 2.26.

The unsoaked CBR is greatly affected by the initial dry

density; the unsoaked CBR value increases as the dry

density is increased, the moisture content being kept

constant (TRL team (1981) Figure 2.26 and the US Corps of

39

Engineers (1945) Figure 2.27). As shown in Figure 2.27,

the unsoaked CBR – dry density relationship is close to linear, if the moisture content is fixed.

As reported by Yoder (1975) and TRL team (1981) that, the

void ratio affected the unsoaked CBR value; the CBR value

is reduced with increasing in initial void ratio and

increased with decreasing in void ratio. In general,

dense soils (i.e. soils with low void ratios) have high

unsoaked CBR, whereas loose soils (i.e. soils with high

void ratios) have low unsoaked CBR.

2.6.1.3 Influence of soaking condition

It has been found that soaking for 4 days leads to realistic

design of roads constructed on expansive clay soils in

regions where the climate is almost wet throughout the year.

The effect of soaking on the CBR value of expansive clay

soils was found to reduce the CBR value. The considerable

drop of CBR value in a soaked clay sample was attributed to

excessive moisture absorption accompanied by considerable

swell, and reduced density after soaking. In contrast,

granular soils, which are not affected by swelling, the CBR

values have no significant changes after the soaking period.

As reported by Bissada (1970) and shown in Figure (2.28)

that, the soaking of compacted expansive soil samples for 4

days will influence the CBR values. For soaked samples

compacted at moisture contents dry of optimum moisture

content tend to give the lowest CBR values. The drop of CBR

values as the moisture content decreased is due to water

absorption and swelling during the soaking period as

indicated in Figure 2.29. Expansive soils compacted at

relatively low moisture contents will swell more than those

compacted at higher moisture contents. Swell decreases as

moisture content increases to about the optimum value and

then becomes relatively constant for moisture contents

greater than optimum (Yoder (1975)). Whereas soil samples

compacted at moisture contents wet of optimum moisture

content, the unsoaked CBR is slightly greater than that of

the soaked CBR values due to swelling of soil, and the

difference become greater as the swelling increases with

40

decrease in moisture content up to the optimum moisture

content as explained by Yoder (1975) and shown in Figure

2.29.

As previously explained, it is clear that saturation due to

soaking has great influence on the CBR values of expansive

soils. In this case the soil final state such as water

content and dry density will be saturated. The saturated dry

density (satd ) and the saturated water content ( sat ) can be

calculated using the following equations:

4.201.01 w

sat

dsat

5.2r

satS

w

Where:

sat is the saturated density of soil(Mg/m3)

w is the water content(%)

rS is the degree of saturation(%).

The influence of the saturated dry density and saturated

water content as well as the swelling occurs during the

soaking period on the soaked CBR value will be explained in

the following paragraphs.

The saturated dry density as determined by equation 2.4

is mainly affected by the saturated density and water

content. As explained by Bissada (1970) and Yoder (1975)

that at water contents less than the optimum water

content (i.e. dry of optimum side) there is a reduction

of density due to excessive moisture absorption during

the soaking period. In this case, as the water content

decreases the reduction in density will increase which

reduces the value of saturated dry density. Therefore

decreasing in water content from optimum decreases the

saturated dry density that leading to reduced the soaked

41

CBR value as shown in Figures 2.28 and 2.29. On the other

hand, at high water contents (wet of optimum side) where

the soil is almost saturated, the soaking has less effect

on the density, accordingly the saturated dry density to

some extent is similar to the initial dry density. As

indicated in Figures 2.28 and 2.29 and discussed in

Section (2.6.1.2), it is clear in that increasing in

saturated dry density will increase the soaked CBR value.

The saturated water content value depends on the initial

water content and degree of saturation as indicated by

equation 2.5. On the wet of optimum moisture contents

side where the soil is almost saturated (i.e the degree

of saturation is approximately equal hundred percent),

the saturated water content value seems to be equal to

the water content value. As explained by Bissada (1970)

and Yoder (1975) and indicted in Figures 2.28 and 2.29

that, the soaked CBR is much affected by the saturated

water content (i.e. water content). It is clear that the

soaked CBR values are decreased with increasing in the

saturated water contents. Whereas on the dry of optimum

moisture contents side, the saturated water content value

is much affected by the degree of saturation. As water

content decreases from optimum water content, the degree

of saturation value becoming very low that will increase

the saturated water content value as indicated by

equation 2.5. The drop of the soaked CBR value at low

water contents as shown in Figures 2.28 and 2.29, is due

to decreasing in degree of saturation that will give high

saturated water content values. Therefore the soaked CBR

decreases with increasing in saturated water content.

It was found that the swelling of expansive soil during

the soaking period has much effect on reducing the CBR

value (Yoder, 1975). As can be seen in figure 2.29 that

increasing in the amount of swelling reduces the soaked

CBR value. Whereas in granular soils, swelling has no

influence on the soaked CBR values.

42

2.6.2 Shear stress

A load placed on a soil mass will always produce stresses of

varying intensity in a bulb shaped zone beneath the load.

The shear stress applied to a soil must be carried by inter-

particle forces in the soil skeleton, but in general only a

proportion of the normal stresses will be carried by the

soil skeleton.

The behaviour of a soil mass is controlled by the external

total stresses applied to a soil element, and also by the

water and air pressures developed in the pores of the soil.

Terzaghi (1925) was the first to recognize that the

controlling feature was the resultant inter-particle forces,

and expressed this fact for a saturated soil by what he

termed the effective stress law. He postulated that the

strength of a soil depended on the difference of the total

stress and pore water pressure (termed the effective stress)

not on the absolute values. It followed that there would be

strains induced by a change in total stresses only if the

effective stress also change, that is, if the total stress

and the pore water pressure increments differ. Another

fundamental concept is that of the "state" of drainage. Two

"limiting" conditions are recognized as the undrained state

in which the pore water (and air, if present) is prevented

from draining, which would otherwise occur as a result of

the pressure induced in the pore water; and the drained

state in which the pore pressures are controlled to specific

values, commonly atmospheric pressure. These states model

the conditions of rapid loading and long term loading.

The soil failure is defined as the inability of the soil

element to withstand the applied stress state. Failure is

associated with large strains and a rapid decrease in the

stress state which can be resisted by the soil. The Mohr-

Coulomb failure theory considers a sample of soil which is

just about to fail when the normal and shear stresses ( and ) reach some limiting values on a failure plane. It is

found that the combinations of and which cause failure on the plane can usually be represented by the linear

equation

6.2tan c

43

The soil constants c and are termed the cohesion and the angle of internal friction, respectively. The values of c

and will be different for different soil types, while

different values of c and will be relevant, depending on whether the normal stress is taken as the total or effective

stress. In case of undrained triaxial test when no

measurement of pore water pressure is taken, the above

equation can be written in terms of total stresses for

undrained loading as follows:

7.2tan uuc

The degree of importance of either the cohesion or the angle

of internal friction depends on the type of soil. In fine

grained soils such as clay, the cohesion component is the

major contributor to the shear stress. In fact, it is

usually assumed that the angle of internal friction of

saturated clays is equal zero, cohesion C.

The failure criteria of equations (2.6 and 2.7) are known as

the Mohr-Coulomb failure criteria. The Mohr-Coulomb

relationship says that neither shear stress () nor normal stress () by themselves cause failure, but a critical

combination is required. To determine whether a particular

, combination indicates failure, that stress state has

merely to be plotted; stresses on the failure envelope imply

failure. If 1 and 3 are the major and minor principal

stresses at which failure has occurred, and the

corresponding Mohr’s circle is constructed, then one of the points on the circle must satisfy the failure criterion

equation (2.7). Furthermore the failure envelope must be a

tangent to the circle at that point, with the angle of the

failure plane equal to /4 + /2.

In actual fact the shear resistance of a soil is a function

of more than the parameters indicated in equations (2.6 and

2.7). It is also likely to influence by the followings:

Type and shape of soil particles

Drainage conditions

Soil initial states

44

2.6.2.1 Influence of soil particles type and shape

In coarse grained soils such as sand and gravel, the shear

stress is achieved mainly through the internal resistance to

sliding as the particles roll over each other. The angle of

internal friction is therefore an important shear stress

parameter. The value of the angle of internal friction

depends on the shape of the individual soil particles and

the surface texture. Thus Soils with rough particles such as

angular sand grains will have higher shear stress (larger

angle of internal friction) than rounded particles of

crushed gravel and bank run gravel. On the other hand, soils

with fine grained particles (clay soils) will have a low

angle of internal friction and a high cohesion.

2.6.2.2 Influence of Drainage conditions

The shear stress of a soil may be obtained in the laboratory

by conducting the triaxial test, the unconfined compression

test, or the direct shear test. There are three main types

of test conditions are employed in the laboratory. These

briefly described thus:-

A. Undrained, or "quick" test, in which no drainage is

allowed as testing proceeds to failure. Normally pore

water pressures are measured during the test, which

enables the results to be expressed in terms of either

total or effective stress, when the samples are not fully

saturated.

B. Consolidated – undrained or "Consolidated – quick" test, in which the test sample is loaded and allowed to

consolidate under a system of applied lateral stresses.

During this period drainage is allowed until the pore

water pressure is reduced to zero. Once this

consolidation is complete, a normal stress is increased

rapidly without allowing drainage, until shear failure

results. Providing pore water pressure measurements are

taken during the shearing phase, the resulting can be

expressed in terms of total or effective stress.

C. Drained or "slow" test in which the sample is loaded and

consolidated as in (B) above. When consolidation is

45

complete under these conditions, drainage continues to be

allowed while a normal stress is increased at a rate such

that no pore water pressure can develop. The resulting

shear stress is expressed in terms of effective stress

only.

2.6.2.3 influence of soil initial state

The shear stress value of a given expansive clay soil is

influenced by the soil initial state as described by the dry

density, moisture content and void ratio.

The shear stress is greatly influenced by the initial

moisture content; increasing moisture content will reduce

the shear stress as explained by Lewis (1959) and

O’Connor (1994) in Figures 2.30 and 2.31. It is clear

that, as shown in Figure 2.30, the shear stress and water

content relationship is approximately linear, if the

initial dry density is fixed.

The initial dry density is strongly influencing the shear

stress. When the initial dry density is high, the shear

stress will have a high value. This is verified by

O’Connor (1994) and Lewis (1959) and shown in Figures 2.30 and 2.32, as it can be seen this relationship is

almost linear.

Mitchell (1976) reported that, the initial void ratio has

much influence on the internal angle of friction;

increasing in the soil initial void ratio decreases the

internal angle of friction as indicated in Figure 2.33.

It is known in practice, soils with low void ratios (i.e.

dense soils) have high shear stress than those with high

void ratios (i.e. loose soils).

2.7 CONVENTIONAL METHODS OF PAVEMENT DESIGN

Methods of pavement design can be subdivided into two main

groups namely:

46

(a) Methods derived from empirical studies of pavement

performance

(b) Methods which either use or are derived from

theoretical studies of the mechanical behaviour of the

pavement.

Methods which are based on empirical studies of pavement

performance usually include field and laboratory testing of

materials for identification and classification. In the more

sophisticated methods materials are required to satisfy

quite stringent specifications. These usually include

strength tests giving quantitative information about the

mechanical properties of the materials. Empirically based

methods have proved to be satisfactory provided the

materials, environment and conditions of loading do not

differ significantly from those which applied during the

original empirical studies on which the design methods were

based.

The essential features of an analytical or mechanistic

structural design system for flexible pavements are:-

1) The stresses and strains in the system have to be

expressed in terms of load stresses and the mechanical

properties of the materials.

2) The characterisation of the mechanical properties of the

materials in the layers under the appropriate climatic

and loading conditions.

3) The solution of the equations for the stresses, strains

and displacements in the model.

4) The definition of criteria for design and the selection

of limiting values of stress, strain and deformation.

5) The presentation of the design system in a form that is

convenient to use.

In this project the main conventional methods of design are

briefly reviewed such as the CBR, AASHTO and SHELL method.

2.7.1 CBR Design Method

The California Bearing Ratio (CBR) method is one of the most

important design methods developed by the California State

Highway Department during the 1930’s. An example of this is

47

shown in Figure 2.34. It can be seen that the depth of

pavement construction is related to the wheel load and to

the soaked CBR value of the subgrade; in this, chart it is

not related to the volume of traffic, the environmental

conditions or the properties of the pavement materials

themselves except in as far as the pavement materials are

required to satisfy some minimum specifications. The CBR

method was initially adopted by the US Corps of Engineers

for the design of airfield pavements but has subsequently

been refined and used by many highway authorities. It is the

basis for United Kingdom method of pavement design as

described by Transport and Road Laboratory (TRL) in Road

Note No.29 and also the design method for tropical and sub-

tropical countries written by the TRL and described in Road

Note No.31.

In the United Kingdom method, the thickness of the subbase

depends on the traffic loading expressed in terms of

equivalent standard axles and on the CBR of the subgrade

measured at the equilibrium moisture content and density

expected under the road after construction. Figure 2.35

shows the Iso-CBR chart which provides the strength of the

subgrade soil at different moisture contents and densities.

The thickness of the base course also depends on the traffic

loading in addition, also depends on the type of material.

Even such a simple design method as the CBR method is open

to misuse. In the original formulation of the method the CBR

of the subgrade material is obtained after four days of

soaking – a condition which is very unlikely to apply under a properly build road. Thus the measured CBR is merely used

as an index of strength to classify the material. On the

other hand the pavement performance data on which the design

thicknesses are based were obtained from in-service

pavements. Under such pavements the in-situ CBR of the

subgrade (which is related to the actual supporting value of

the subgrade) would have been much higher than the soaked

CBR value. The result of using a soaked CBR value for design

purposes will be a thicker pavement than necessary.

2.7.2 AASHTO Design Method

48

The AASHTO design method (or modifications of it) is

probably the most widely used method in tropical countries.

This method was developed from the results of the AASHO Road

Test conducted during 1959 and 1960. In the AASHO road test,

Four large loops of road approximately two miles long were

constructed together with two smaller loops. Each loop

comprised a two lane highway. Within each loop test sections

of pavement with a minimum length of 100ft were constructed.

A total of 468 test sections of flexible pavement were

included together with test sections of rigid pavement. Each

lane of each loop was trafficked almost continuously for two

years by specially selected and loaded vehicles with

particular axle loads and wheel configurations. Each lane of

each loop always carried the same type of vehicle with the

same axle load throughout the two year study period but no

two lanes carried vehicle of the same type and axle load.

The Road Test was clearly a massive undertaking and has

provided an enormous amount of data, some of which are still

being analysed or reanalysed in the light of theoretical

advances.

AASHTO initially published an interim guide for the design

of pavement structures in 1961, which was revised in 1972. A

further revision was published in 1986, incorporating new

developments and specifically address-ing pavement

rehabilitation.

2.7.2.1 The Pavement Design Factors

The factors considered in the AASHTO procedure for the

design of flexible pavement as presented in the 1986 guide

are pavement performance, traffic, subgrade strength,

materials properties, drainage and reliability.

Pavement Performance: The primary factors considered

under pavement performance are the structural and

functional performance of the pavement. Structural

performance is related to the physical condition of the

pavement with respect to factors that have a negative

impact on the capability of the pavement to carry the

traffic load. These factors include cracking, faulting,

raveling, and so forth. Functional performance is an

49

indication of how effectively the pavement serves the

user. The main factor considered under functional per-

formance is riding comfort.

To quantify pavement performance, a concept known as the

serviceability performance was developed. Under this

concept, a procedure was developed to determine the

present serviceability index (PSI) of the pavement, based

on its roughness and distress, which were measured in

terms of extend of cracking, patching and rut depth for

flexible pavements. The mean of the ratings was used to

relate the PSI with the factors considered. The scale

ranges for 0 to 5, where 0 is the lowest PSI and 5 is the

highest.

The serviceability indices are used in the design

procedure include the initial serviceability index (Pi),

which is the serviceability index immediately after the

construction of the pavement, and the terminal

serviceability index (Pt), which is the minimum acceptable

value before resurfacing or reconstruction is necessary.

In the AASHTO road test, a value of 4.2 was used for Pi

for flexible pavements. Recommended values for Pt are 2.5

or 3.0 for major highways and 2.0 for highways with a

lower classification.

Traffic: The traffic load application is given in terms

of the number of repetitions of an 18000 lb (80 KN)

single axle loads, tandem axle is treated as two single

axles. This is usually referred to as the Equivalent

Single Axle Load (ESAL). The procedure used to determine

the design ESAL is fully described in the AASHTO manual.

Subgrade strength: The AASHTO uses the resilient modulus

(Mr) of the soil to define the subgrade material strength.

However, the method allows for the conversion of the CBR

or R value of the soil to an equivalent Mr value using the

following conversion factors.

8.2)/(15002

inlbCBRM r

9.2)/(55510002

inlbvalueRM r

50

The AASHTO recommended that the above two equations to be

used for fine grain soils with soaked CBR of 10% or less

and R value 20.

Material properties: unlike most empirically based design

methods the AASHTO method takes account of variations in

material properties and allows the overall thickness to

be reduced as the strength of the materials increase

above the minimum values required by the specifications.

The Layers coefficients are related to the standard

strength tests for the materials in question, such as for

crushed stone the relationship has been derived between

a2, the layer coefficient for crushed stone base course,

and the CBR value.

Drainage: The effect of drainage on the performance of

flexible pavements is in the 1986 guide with respect to

the effect of water has on the strength of the base

material and subgrade soil. The approach used is to

provide for the rapid drainage of the free water from the

pavement structure by providing a suitable drainage layer

and by modifying the structural layer coefficient. The

modification is carried out by incorporating a factor mi

for the base and subbase layer coefficients (a2 and a3).

The mi factors are based on the percentage of time during

which the pavement structure will be nearly saturated,

and the quality of drainage, which is dependent on the

time it takes to drain the base layer to 50 percent of

saturation.

Reliability: The AASHTO proposes the use of a reliability

factor that considers the possible uncertainties in

traffic prediction and performance. Reliability design

levels (R%), which determine assurance levels that the

pavement section designed using the procedure will

survive for its design period, have been developed for

different types of highways.

2.7.2.2 The Structural Design

The objective of the design using the AASHTO method is to

determine a flexible pavement SN adequate to carry the

51

projected design Equivalent Single Axile Load (ESAL) by

using the chart in figure 2.36. The AASHTO gives the

expression for SN as:

10.233322211 mDamDaDaSN

Where

mi: drainage coefficient for layer i.

a1,a2,a3: layer coefficients representative of surface,

base, and subbase course, respectively.

D1,D2,D3: actual thickness of surface, base, and subbase

courses, respectively.

2.7.3 Shell Method

The Shell method of pavement design is possibly the most

widley known semi-theoretical procedure. With this design

process, which was originally published in 1963, the

pavement is regarded as a three layer system in which the

lowest of the layers (assumed infinite in the vertical

direction) represents the subgrade; the middle layer

represents the combined unbound roadbase and subbase, whilst

the uppermost layer includes all bituminous bound materials

above the roadbase. A "full-depth" bituminous bound pavement

resting directly on the subgrade is dealt with by assuming

zero thickness for the unbound layers.

The Shell design approach involves estimating the bituminous

bound and unbound layer thickness required to satisfy

governing strain criteria. These criteria are:

1. The compressive strain in the surface of the subgrade,

i.e. if this is excessive, permanent deformation will

occur at the top of the subgrade and this will cause

deformation at the pavement surface.

2. The horizontal tensile strain in the bituminous bound

layer, generally at its bottom, i.e. if this is

excessive, cracking of the layer will occur.

The permissible value for compressive subgrade strain was

derived from the analysis of data from AASHO Road Test

pavements which conformed to CBR design. The permissible

52

strain in the bituminous layer was determined from extensive

laboratory measurements for various bituminous mix types at

different stiffness moduli.

Other criteria taken into account include the permissi-ble

tensile stress or strain in any cementitious base in the

middle layer and the integrated permanent deformation at the

pavement surface due to deformations in the individual

layers.

A complex computer program BISAR has been developed by Shell

which enables all stresses, strains and displacements at any

point in the pavement system to be determined under any

number of vertical and/or horizontal surface loads. However,

as it may be difficult to run this sophisticated computer

program, a series of 296 design charts has been prepared for

the Shell Manual; these charts obviate the need to carry out

complex calculations. Instead, the design engineer is able

to enter these charts with input data reflecting the

subgrade modulus, bituminous mix design, traffic volume, and

mean annual air temperature, and derive various combinations

of pavement structures which satisfy the critical strain

criteria, i.e. pavements that will not crack and will not

result in excessive strain in the subgrade which could lead

to excessive surface deformation.

However, bituminous layers, being partially viscous

materials, will deform in themselves under traffic. Thus,

when some alternative structural designs have been selected,

the next stage in the design process involves determining

the magnitude of the permanent deformation, i.e. rut depth,

anticipated in the surface layer in each candidate cross

section during its design life. The amount of permanent

deformation directly attributable to the bituminous layer is

estimated from the product of the thickness of that layer,

the average stress in the layer, the reciprocal of the

stiffness of the layer, and a correction factor for dynamic

effects. Permanent deformation in the unbound road base and

subbase layers can also contribute to the ultimate surface

deformation, so an estimate of this deformation is also made

and added to the estimated bituminous layer deformation. To

determine the ultimate surface deformation, the bituminous

53

layer is subdivided into a number of sub-layers. Then for

each sub-layer, calculation the effective viscosity of the

bitumen, determination of the mix stiffness and the average

stress is to done.

The estimated total surface deformation in each candidate

pavement design is next compared with the allowable depth of

rut, and either judged acceptable or not acceptable. The

final choice of pavement design is then made from the

acceptable candidates, normally on the basis of economics.

2.8 DESIGN OF ROADS ON EXPANSIVE SOILS

Many methods are used as practical solutions for the problem

of expansive soils when exist as road subgrade. These can be

subdivided into two categories:

Technical design methods: These are the methods that

recommended by the previous researchers and still under

study.

African specifications of pavement design: certain

African countries which have Code of Practice for design

and construction of roads on expansive soil.

2.8.1 Technical Design Methods

A wide range of design methods which recommended by previous

researchers for minimizing the damaging effects of moisture

changes and related volume changes in expansive soils

beneath roads. Many of these methods are embodied in the

specifications of certain African countries. This section

will briefly describe the methods and their intended effect

as well as the experimental works done by previous

researchers.

2.8.1.1 Soil stabilization

It is any process by which a soil may be improved and made

more stable i.e. increase strength. This may increase or

decrease the permeability, reduce compressibility, improve

stability or decrease heave due to swelling. The various

54

forms of soil stabilization include mechanical

stabilization, chemical stabilization and thermal method.

Mechanical stabilization rearranges, adds, or removes soil

particles. The objective is to modify density, water

content, or gradation. If a soil cannot be made stable

simply by compaction or consolidation, then additional soil

or other aggregate materials may be admixed to produce a

mixture having the required stability characteristics.

The properties of the soil related to the interaction with

water can be altered by chemical means to reduce the

expansiveness of the soil. Addition of lime or cement is

often used in road construction to increase the strength of

the compacted soil. It is also effective in reducing the

expansiveness of a soil. The change in the soil is most

apparent in the reduction in the liquid limit, although the

plastic limit may be raised to a lesser extent.

In lime stabilization there is a physical and a chemical

component to the reaction of lime with clay. The physical

reaction is one of cation absorption, calcium first

replacing any other ion present as a base exchange ion. This

is followed by the flocculation into groups of coarse

particles which produce an immediate increase in strength.

The addition of lime to a soil causes an immediate increase

in the pH of the molding water, due to the partial

dissociation of the calcium hydroxide. The calcium ions in

turn combine with the reactive silica or alumina or both,

present at soil surfaces to form insoluble calcium silicates

or aluminates or both, which harden on curing to stabilize

the soil. This process continues for some months.

Lime treatment levels of two to eight percent by weight of

dry soil are typical. The unconfined compressive strength

increases with increasing lime content to an upper limit

generally proportional to the clay content (Herzog and

Croft (1964) Figure 2.37). Hoskins et al (1972) found a

reduction in swell percent under a surcharge pressure of 2

psi (13 kPa) from 4.9% to 0.4% on the addition of 5% lime.

The samples were compacted to 95% Proctor maximum density at

optimum moisture content. Mukhtabant and Bunng (1973) found

55

a reduction in swell under 1 psi (6.7 kPa) from 7% to 0.5%

and 0.3% for the addition of 2 and 5% lime respectively at

93% of Proctor maximum dry density. Samples compacted to

100% of maximum density showed a smaller improvement,

reducing from 13.5% swell to 9.5% and 3% for the same lime

contents. There is a significant decrease in the swelling

pressure of active clays by the addition of lime (Mitchell

(1976) Figure 2.38). Gray et al (1980) found an increase in

CBR to the range 20 to 60% using 3% lime. A typical CBR for

untreated black clay would fall in the range 1 to 5%.

Logavinayagam (1980) measured an increase from 1 to 4% for

untreated soil to 20 to 54% for the addition of 3% lime and

15% fly ash.

Cement stabilization is one of the commonest methods of soil

stabilization. Cement hydrates when water is added,

producing cementitious compounds independently of the soil.

These products are calcium silicate hydrates, calcium

aluminate hydrates and hydrated lime. The first two products

constitute the major cementitious components, whereas the

lime is deposited as a separate crystalline solid phase. The

increase in strength is due to the development of

cementitious linkages between these hydration products and

soil particles. The lime released during the hydration of

the cement may react with any pozzolanic material, for

example clay, present in the soil to form a secondary

cementitious materials which also contributes to inter-

particle bonding.

Addition of cement generally increases the strength of most

soils. The major problem is to achieve adequate mixing.

Indeed the percentages of cement required for adequate

stabilization of the lateritic soils vary from 3% to about

7%. Unconfined compressive strength increases approximately

linearly with cement content (Ingles and Metcalf (1972)

Figure 2.39). The use of cement appears to be more effective

than the use of lime in increasing the strength of an

expansive soil. Quigley and di Nardo (1978) compared the

effect of equal amounts of lime and cement and found a

greater unconfined compressive strength using cement. The

strengths found using cement were about three times the

strength using lime. They note that cement is more effective

56

in stabilizing the whole soil, while lime is more effective

on the clay fraction. A soil with a high clay content could

thus show a higher strength when lime stabilized than when

cement stabilized.

Heating will alter the clay minerals, ultimately to ceramics

such as china, but at lower temperatures to earthenware.

This has been evaluated in the past and found to be too

expensive and slow. The heating methods have been crude and

inefficient, and it may be that modern techniques could

develop a more effective treatment. Temperatures of 400 to

600 C are typically required for permanent alteration of montmorillonites to non-swelling material.

2.8.1.2 Impermeable Membranes or Barriers

Impermeable membranes have seen widespread testing in

Southern Africa, Australia and the USA. The membranes may

either be used to minimize moisture content change by

preventing moisture movement through the soil or be used to

move the zone of moisture movement away from the road

pavement. Vertical barriers fall in the first category.

These may be effective where lateral movements of water

either causes edge heave or where evaporation at the edge

defines the lateral extent of a center heave profile. To be

fully effective, the barrier should penetrate through the

zone of moisture movement, but may be effective at a

shallower depth in reducing moisture transfer to an adequate

rate. They may be used either to keep the soil at a low

moisture content where moisture would otherwise penetrate

from the edge, or to maintain a high but uniform moisture

content where moisture movement is from the center toward

the edge. The barrier is generally formed using a plastic

sheet unrolled into a narrow trench, but could also be

formed using a bituminous material poured into a narrow

trench or sprayed onto a vertical face. Care must be taken

to ensure that the trench does not act as a water trap

supplying moisture to expansive soil at the toe of the

membrane. The membranes are usually placed at the edge of

and tied into the road pavement, thus forming a three-sided

box.

57

Vertical moisture barriers have been used for both roads and

houses. Lee and Kocherhaus (1973) used 1 to 1.5 m vertical

plastic membranes or concrete barriers together with paved

sidewalks around houses as remedial works to damaged houses

in California. The area is subjected to 2 m of potential

evaporation but only receives 0.75 m of annual rainfall.

Edge heave was being caused by watering of gardens. Further

damage was prevented after installation of the barriers.

Sterinberg (1980 and 1981) used a 2.4 m deep vertical

membrane beneath the edges of a road in Texas and

successfully reduced the variations in moisture content

within the protected zone compared with the soil outside the

protected zone. The soil outside the protected zone

experienced significant drying, while the soil within the

protected zone maintained a high moisture content. This

experiment might have been less successful had the soil

initially been dry as the membrane trench was backfilled

with sand which would have acted as a water source at the

toe of the barrier. Similar results were obtained by

Picornell et al (1983 and 1984) on another road project

which was also in San Antonio, Texas. The suction range

expected in this project was in the range 0 to 1 bar, which

corresponds to a high moisture content.

Horizontal membranes, such as the sealed shoulder, fall in

the second category. These could be used for both of the

moisture movements for which vertical barriers may be

effective. However, where a vertical barrier will tend to

minimize soil movements, the horizontal membrane will tend

to increase the soil movements by concentrating runoff from

a greater area at the edge of the sealed area or by reducing

moisture content losses by evaporation. The lateral extent

of the barrier must be sufficient to reduce differential

movement of the road pavement to an acceptable amount.

Horizontal membranes have been placed at various positions

beneath roads. Strongman (1963) and Gordon and Waters (1984)

describe cases where the membranes were placed immediately

beneath the road shoulder. The former case also used a

blanket of compacted weathered phonolite lava of low

permeability on the embankment slopes with expansive soil

forming the core of the embankment. He reported no

58

significant change in moisture content over 4 years, and the

road was reported to have performed well for many years

thereafter (Jones 1985). The latter case was coupled with

the use of dry compaction, and was not as effective in

preventing damage. They found that the sealed shoulders were

only effective in delaying the deformation of the roads and

longitudinal cracking was not prevented. Merlen and Brakey

(1968) used a horizontal membrane placed 0.6 m below ground

level to prevent penetration of water due to hydrogenesis

(Brakey 1968) into the underlying expansive soil subgrade.

They found an accumulation of water in the fine sand fill

above the membrane and no increase in the moisture content

of the expansive soil beneath the membrane. Where the

membrane was omitted there was an increase in the moisture

content of the expansive soil.

Inclined barriers are a compromise between the two extremes,

but are also of benefit in that any drainage ditch that is

required may be lined with the membrane. They are generally

of greatest value on embankments, where the membrane will

also reduce erosion. Luttrel and Reeves (1984) has used an

inclined membrane in Australia. The membrane extended from

beneath the pavement, across the shoulder, down the

embankment and 30 cm below natural ground level. No distress

was noted to the road over a period of 5 years where erosion

was expected to occur, a reinforced concrete cover was

placed over the membrane on the embankment slope and

penetrating below the ground surface.

2.8.1.3 Injected Lime Barriers

In this technique a lime slurry is injected under high

pressure along the edge of the road. The depth to which

injection is performed may be estimated as for the vertical

membranes. The lime slurry penetrates into the soil by

hydraulic fracture or along existing fissures. The soil

between the sheets of lime placed in this way is virtually

unaffected by the lime. Reduction of movements is therefore

by inhibiting moisture movement due to the horizontal sheets

of impermeable lime.

59

This method is extensively used in North Texas. Wright

(1973) describes the lime injection performed at Dallas–Fort Worth Airport where it was used as a pre-construction

treatment for most of buildings. Injection was to 2.1 m

below ground at 1.5 m centers to form an overlapping network

of horizontal sheets. Apart from reducing moisture movement

because of the horizontal sheets of lime, the injection

process also caused pre-swelling of the soil. Post-

construction injection was used by Saragunam (1983) on 22

buildings in Madras. Injection was most effective when

performed in spring as the damage was caused by soil

shrinkage during the summer. Twelve millimeter diameter

perforated steel pipes were used at 45 cm spacing around the

buildings. The 1.5 meter depth of treatment was effective in

preventing further damage in 90 percent of cases. Injection

pressures range from 200 to 600 kPa. Sand and silt lenses

were present in the soil profile and would have facilitated

the lateral penetration of the lime. Remedial work to a road

was reported by Cothren (1984). Failure of the road had been

by longitudinal cracking and it was resurfaced after the

lime injection treatment. Injection points were at 1.2 meter

centers along both sides of the pavement. Up to 220 liters

of slurry was injected at 0.75, 1.5 and 2 m depth at

pressures up to 1380 kPa. The slurry comprised 0.3 kg of

lime per liter of water, and a wetting agent was used to aid

penetration of the slurry into the soil. After two wet and

dry seasonal cycles there was no indication of movement and

no cracking of the asphalt surface.

2.8.1.4 Granular Blankets

Seasonal or cyclic heave in particular is dependent on loss

of moisture by evapotranspiration adjacent to the covered

area. Removal of vegetation will reduce the rate of moisture

loss, but where the clayey soil extends to the surface,

moisture will be moved to the zone of maximum suction at the

ground surface and will evaporate from there. The maximum

capillary rise of moisture through a sand is very much less

than the potential rise due to soil suction in clays. A

clean sand or gravel blanket of 150 to 250 mm thickness will

be effective in preventing almost all evaporation from the

ground surface.

60

The effectiveness of a granular surface cover in preventing

evaporation and maintaining high and stable moisture

contents has been demonstrated by the early work of de

Bruijn and Donaldson (1965). They measured moisture changes

using both psychrometers and nuclear meters and found less

variation beneath the sand blanket than was observed in the

open field.

Gogoll (1970) used 0.6 and 1.8 m thick blankets over a

prewet clay to minimize heave. The foundations were placed

several months after the blanket had been placed, and no

movements were observed during a six month period. A control

peg outside of the blanketed area indicated a subsidence of

60 mm, indicating that the blanket was effective in

minimizing moisture changes.

2.8.1.5 Prewetting and Ponding

These methods attempt to cause the majority of heave to

occur before construction is commenced. Water is supplied by

irrigation or by forming a pond covering at least the area

of subsequent construction. Rainfall may be utilized

together with removal of vegetation and a sand blanket to

prevent evaporation, but the rate of heave is likely to be

much slower using rainfall than using artificially provided

water. Difficulty may be experienced in obtaining adequate

penetration through relatively impermeable soils,

particularly where the potential heaving layer is thick.

This may be alleviated by drilling a series of boreholes

through the potential heaving layer and supplying water by

filling the boreholes.

This method will only be effective where the equilibrium

moisture content is high. If prewetting is used in areas

where the equilibrium moisture content is low, then drying

and shrinkage will occur over a period of years and may

cause severe damage. Similar effects may occur at the edge

of a road if the heave profile caused by the prewetting does

not reasonably match the long term moisture distribution

beneath the road.

61

Artificially increasing the moisture content of a desiccated

soil before construction has employed for many years.

Steinberg (1977 and 1980) has used ponding on road projects

in the USA. The depth of wetting varies, in one project

being less than 1.4 m in one month, in the other being 3 m.

In the former case the ponding caused about half the final

total heave to occur, and the remainder which took place

after construction, occurred at depths greater than 1.4 m.

Movement of the control sections which were not prewet did

not commence for 2 years but were similar in final magnitude

to the total movement in the ponded sections. In both cases

damage to the road was less in the ponded sections than in

the control sections.

Weston (1980) and Netterberg and Bam (1984) compared

prewetting with other movement control techniques. The

observed movements were inadequate to reach firm

conclusions, but the latter authors suggest that prewetting

with an irrigation layer of sand would have been adequate.

The road was constructed with prewetting and membranes as

standard measures for a 25 km section over an expansive soil

subgrade.

2.8.1.6 Placement at Equilibrium Moisture Conditions

This technique also attempts to obtain the equilibrium

moisture conditions in the soil during construction. Where

the above method is primarily applicable to in situ soils,

this technique is applicable to compacted soils. The

standard practice for compaction of soils is to use a

moisture content close to optimum in order to minimise the

effort required to obtain the desired density. Where the

equilibrium conditions are significantly drier than the

compaction moisture content drying will take place and may

cause severe damage. If the compaction moisture content is

near to the equilibrium moisture content then volume change

due to moisture content change will be minimised.

Although this technique could be of great value under

appropriate ground and climatic conditions, very little

effort has been directed toward its evaluation in practice.

Ellis (1980) has employed dry compaction in Sudan. This was

62

intended to evaluate the use of soil expansive, at its

natural moisture content in order to avoid the use of

additional water. The moisture content at compaction ranged

from 7 to 14%, compared with the optimum moisture content 20

to 30%. In the 18 months following construction, the

moisture contents increased slightly toward the 24% moisture

content measured at about 2 m depth at the site before

construction. It appears that the compaction moisture

contents, which reflect the natural moisture content at

shallow depth may be lower than the equilibrium conditions,

and further wetting up is to be expected. No damage had been

observed after two wet seasons, and it is suggested by Ellis

(1980) that the uniform wetting up is the reason.

In the above case the moisture condition was governed by

climatic conditions during construction. Very careful

control of the moisture content will be required to maintain

the desired conditions during construction. However, it

appears that wetting up from dry conditions is less damaging

than drying after construction which is held to be

responsible for some cases of longitudinal cracking (Gordon

and Water (1984)).

2.8.2 African Specifications for Pavement Design

Certain African countries have manuals or reports for design

and construction of roads on naturally occurring expansive

soils. In this section a brief description of the

recommendations given by these specifications design manuals

or reports.

Kenya Road Design Manual (MOTC, 1981) considers soil

stabilization, using 4 to 6% of lime will reduce the

swelling and increase the soaked CBR of expansive soils. In

order to control the moisture content of expansive soils,

the manual recommends that soil should be placed at the

Equilibrium Moisture Content (EMC) which is usually the

optimum moisture content for standard compaction, except in

arid areas or where the water table is near to the surface.

Extended sealed shoulders using impermeable materials are

required by the manual where removal or stabilization are

not employed, in order to minimize the effect of seasonal

63

changes on the pavement. The manual recommends that side

drains should be kept as far from the road as possible, or

avoided entirely, as they have acted as water sources and

promoted swelling (Mitchell and Mckechnie, 1972). In order

to minimize the swelling due to moisture changes, the manual

recommends surcharging the natural expansive soil using 1 to

3 m of non-expansive fill. Also compaction to a lower dry

density than is normally required to minimize swelling due

to wetting.

Zimbabwe Rural Road Report (Mitchel et al, 1975) recommends

removal of 60 cm of the naturally occurring expansive soil.

In Zimbabwe the equilibrium moisture content is close to

optimum moisture content and that construction should not

take place where the soil is excessively wet or dry. Side

drains should be kept as far from the road as possible or

avoided entirely, as they have acted as water sources and

promoted swelling.

TRH9 from South Africa (NITRR, 1978) considers soil

stabilization. Using 4 to 6% of lime will reduce swell to

negligible values and increasing the soaked CBR of the soil,

and suggested that two or three partial applications of the

lime will be required to obtain uniform mix. The mount of

volume change that occurs on wetting may be reduced by

surcharging the expansive soil (see section 2.5.5). This is

utilized by the NITRR in recommending that using 3 m

embankment of non-expansive soil placed on it to reduce the

amount of swelling that occur on wetting.

Botswana Road Design Manual (MOWC, 1982) recommends

prewetting of the natural soil as was done by Netterberg and

Bam (1984). Extended sealed shoulders are required using

embankment slopes where removal or stabilization is not

employed.

2.9 DISCUSSION

The technical methods of design and construction of roads on

expansive soils which recommended by the previous

researchers and adopted by certain African countries as

practical solutions of swelling problems, as described in

64

the previous section, have some disadvantages and may be

unsuitable options. This due to the reasons will be

discussed in the following paragraphs.

Soil stabilization using chemicals such as lime or cement is

an expensive option in undeveloped countries. The mixing of

lime or cement with expansive soils was found to be one of

the problems met by the engineers working in expansive soil

areas (e.g. TRL teams and Howard Humphreys working in East

Africa, 1980~1990).

Impermeable membranes which are used to control moisture

content of the expansive soil, may crack due to the soil

movements. Where the membrane crack, water may penetrate

through the cracks and cause a progressive inward

penetration of the zone of soil movement leading to soil

expansion and ultimately failure of the pavement. Using

plastic sheets, bituminous or concrete materials as vertical

barriers in road construction on expansive soils as well as

using injected lime barriers, these will be very expensive

techniques especially in undeveloped countries.

The granular surface cover technique is not often effective.

Shiming (1984) used a 1.5 m thick sand blanket beneath an

industrial building. Differential heave and damage resulted

from leakage of water from pipes serving the building.

The prewetting and ponding methods of naturally occurring

expansive soils will not be effective for arid climates,

where the equilibrium moisture content is low. This is due

to the fact that, these soils have low permeability, also

drying and shrinkage could occur where Equilibrium Moisture

Content (EMC) is less than that achieved by prewetting and

may cause sever damage (Blight and de Wet, 1965). Similar

effects may occur at the edge of a road if the heave profile

caused by the prewetting does not reasonably match the long

term moisture distribution beneath the road. The Botswana

manual recommends pre-wetting of the natural soil. This may

be unsuitable option for arid climates.

In most arid regions where Equilibrium Moisture Content

(EMC) is less than the compaction optimum moisture content,

65

drying will take place and may cause deformation and

longitudinal cracking appear on road’s surface. With the

climatic and geological conditions so careful control of the

moisture content is recommended to maintain the desired

conditions during construction.

Some manuals consider the removal of expansive soil and

replacement with non-expansive soils to be the most

effective option where avoidance of expansive soil is not

feasible. But this option is only possible in areas where

suitable non-expansive fill material is available, and where

the depth of expansive soil is small.

Jones (1985) found that the sealed shoulders were only

effective in delaying the deformation of the roads and

longitudinal cracking was not prevented. The extended sealed

shoulders which are recommended by the Kenya manual where

removal or stabilization are not employed, in order to

minimize the effect of seasonal changes on the pavement.

This will not prevent seasonal movements, but such movements

will initially occur beneath the shoulders and only

penetrate toward the pavement if longitudinal cracking

permits water entry through the shoulders.

The conventional methods of pavement design which included

analytical design methods and empirically based design

methods (see section 2.7) are often criticized because of

the deficiencies that will be pointed out in the following

paragraphs.

Analytical design methods require a considerable amount of

material testing and computational effort before they can be

properly used. It is therefore unlikely that in the

immediate future the full technique will be used for the

design of individual roads in most countries. It is more

likely that the analytical technique will be used in the

preparation of simplified design manuals.

Recently, traffic volumes and traffic loadings have

increased quite dramatically and it is unlikely that

existing empirical methods can be simply extrapolated to

accommodate these changes. Further more, the use of new or

66

unusual materials may be inhibited if it is not possible to

determine the thicknesses and conditions under which they

should be used. The extension of empirical methods to

different loadings, different materials and different

environmental conditions can be achieved only by carrying

out expensive and time consuming full-scale pavement

experiments. The extension of empirical design methods is a

serious problem and difficult even for most developing

countries. Indeed many developing countries do not yet have

a satisfactory design method of their own based on empirical

studies. In such countries pavement design relies on methods

borrowed from elsewhere with results that are often far from

satisfactory.

Finally, it is to be noted that all the technical design

methods previously reviewed (in Section 2.8) do not consider

the strength characteristics of the subgrade material,

although it is the main factor of pavement design. On the

other hand, conventional methods of flexible pavement design

consider the strength (CBR and shear stress) alone and not

swelling. The work in this thesis hopefully considers both

the swelling behaviour and the strength characteristics of

the soil.

67

CHAPTER THREE

LABORATORY TESTS AND PROCEDURES

3.1 INTRODUCTION

The laboratory tests were carried out in this research

project for number of purposes, the most important being:

For description and classification of the soil used.

To investigate the swelling and strength behaviour of

soil samples prepared at a wide range of water contents

and compacted to different dry densities and subjected to

different testing conditions in order to develop a new

factor concept for the different soil states.

To determine design parameters (i.e. numerical values for

swelling, CBR and shear stress) for later analysis.

The tests were performed using different types of apparatus.

In this chapter a detailed description of the tests and the

fundamental components of the apparatus used will be given,

followed by the procedures adopted during the tests.

The main piece of apparatus used was a conventional

oedometer cell which was used to measure the swell percent

of compacted soil samples tested under different surcharge

loads. The oedometer cell was also used for measuring the

swelling pressure of compacted samples. The second piece of

apparatus used was a standard 38 mm triaxial cell in which

different stresses could be applied to samples of 38 mm in

diameter and 76 mm in height. This apparatus was used to

measure the shear stress of the sample in a quick triaxial

test. Finally a standard CBR equipment was used to measure

the CBR of compacted samples. Some of these samples were

tested as remoulded in order to measure their unsoaked CBR,

whereas others were subjected to soaking before testing so

as to measure the soaked CBR.

The samples which were prepared and compacted at different

water contents and dry densities, were then used to perform

standard soil tests such as swelling tests, undrained

triaxial test, CBR tests as well as index tests were also

68

carried out to classify the soil used. The measurements from

these tests were able to give a useful indication of the

relationship between the combination of soil initial state

parameters (water content, dry density and void ratio) and

the swelling or strength properties. Since the aim of these

tests was to investigate this relationship it was possible

to perform the tests with one type of soil. The existence of

a relationship would then support the possibility of testing

other soils. This helped with the analysis of the results

reported by previous researchers.

3.2 SOIL USED

The experiments of this study had been conducted on

expansive soil which was brought from Ankang City in

Southern Shaanxi province located in the western part of

china. In Shaanxi province expansive soils are widely spread

in Hanzhong basin, Xixiang basin and Ankang basin of Han

River basin to the south of Qinling Mountains. This area is

known to have the most potentially expansive soil in China.

Ankang clay soil which was selected for this study is famous

for being harmful to roads, railways and light structures.

Serious damages, landslides and cracks are clearly visible

in this area and they reflect the seasonal variations and

changes in moisture content causing upheaval of soil.

Expansive soils appear to be one of the main problems to

agricultural and industrial development in this area and

cause enormous economic losses to the Shaanxi province and

its people. Therefore, the study soil samples were taken

from this area.

Disturbed soils were collected from excavated bore pits at

depths 0.5 to 3 m. The samples were packed in big bags and

the send to Chang’An university laboratory for study.

3.3 SAMPLE PREPARATION

Sample preparation was a very important part of the testing

procedure. It was critical that the method of soil

preparation and compaction method chosen would result in

consistent, uniform samples. In practice clayey backfill

69

material is excavated from a borrow pit and recompacted in

location at a water content similar to its natural in situ

water content. The study of the compaction of soil contained

in chapter two highlighted the large influence that the

compaction process has on the behaviour of the final mass of

soil. For the experiments performed in this thesis the soil

preparation and compaction method chosen to simulate similar

techniques used in the field. The process was a well

controlled standardized procedure that could produce

consistent samples on a small scale. The general method of

preparation was to produce small lumps of soil at a known

and controlled moisture content and to remould these into a

container by applying a load. For this project, soil samples

were prepared with different initial water contents ranging

from 7% to about 35%. To ensure the repeatability of the

samples a careful study of each process in the preparation

was made.

Soil was initially air dried and crushed using a jack

hammer. This was then sieved and the fraction passed 5 mm

sieve was kept. It was possible to produce samples of

different water contents by controlling the ratio of water

to dry soil grains. To achieve the desired moisture content,

a desired quantity of water was sprayed onto a known amount

of dried soil and they were thoroughly mixed for about five

minutes. Then the wet soil was put in plastic bags and

stored for two days to allow a uniform moisture distribution

throughout the sample. After two days of storage, the moist

soil sample was remixed and then compacted into a U8O tube,

120 mm in length with a ring of 20 x 76 mm diameter

positioned inside the tube from which samples could be

prepared for the swelling tests. On the other hand, The

samples prepared for the triaxial test were compacted into a

U38 tube of 76 mm length, while those tested for CBR were

compacted into a standard CBR mould of 2177 cm3 volume.

The compactive effort for remoulding the soil samples in the

CBR mould was provided manually by a rammer of mass 2.5 Kg

(in low compaction) or 4.5 Kg (in heavy compaction). The

rammer was dropped free from a height of 30 cm (or 45 cm)

above the elevation of the soil. The sample was compacted in

three (or five) layers. By adding to the mould a known mass

70

of soil and applying 59 (or 98) uniformly distributed blows

from the rammer it was possible to achieve a consistent

layer thickness. The samples remoulded in the U38 tube were

compressed in eight layers to avoid the presence of a large

density gradient in the sample (Section 2.4.4), whereas

those remoulded in the U80 tube were compressed in one

layer. The compactive effort for compressing the soils in

the preparation tubes was provided by a hydraulic jack.

Once the soil samples had been compacted in the tubes it

could be extruded and the following samples prepared:

20 x 76 mm diameter samples for the swelling tests in the

oedometer cell.

76 x 38 mm diameter samples for the shear stress test in

the triaxial cell.

The prepared compacted soil samples were then weighed for

determining the bulk density and the moisture content of

each sample was also determined. The samples were then ready

for swelling or shear stress tests.

The samples compacted in the CBR mould, after compaction was

completed, the extension collar was removed and the

compacted soil was carefully trimmed with the top of the

mould using a straightedge. Then for each prepared sample

the density and moisture content were determined. The

prepared compacted samples were ready for unsoaked CBR test.

The samples prepared for soaked CBR test were subjected to

soaking process before testing. The compacted sample with

the mould were immersed in a water tank for about four days

to obtain a saturation condition similar to what may occur

in the field. During the soaking period, the sample was

loaded with a 5 kg surcharge load placed on top of it and

the expansion of the soil due to saturation was measured

using a dial gauge positioned over the sample. After the

four days of soaking, the sample was removed from the water

tank and allowed to drain for about 15 minutes. The soaked

sample was then ready for performing the CBR test.

71

3.4 PRELIMINARY TESTS

The engineering properties of soils are controlled by a

number of factors such as soil mineral composition, organic

material, particle size distribution and geological history.

However for the majority of soils the mechanical properties

of soils are largely determined by the finest 20 % of the

constituent grains for the soils used, that is the clay and

fine silt fractions. The preliminary tests aimed to classify

and described the soils within the existing standard methods

and quantify their properties.

The preliminary tests or the routine soil tests performed in

this project were including the index tests, clay content,

specific gravity determination and compaction tests. These

tests are very carefully and precisely specified in a number

of national standards and codes of practice. The tests

provided standard data for the soil which could be used in

subsequent analysis. The tests also provided an opportunity

to ensure that methods of mixing and drying employed in the

sample preparation did not affect the properties of the

soil.

3.4.1 Index tests

The properties of fine grained clay soils depend largely on

the type of clay. The basic behaviour of plastic clay soils

can be assessed from the index tests (i.e. liquid limit and

plastic limit). The liquid and plastic limits are the water

contents at which the soil changes its mechanical behaviour.

The definitions and methods of testing used are fully

defined in the standard specifications.

In this research the Liquid Limit was determined using a

cone penetrometer apparatus. The penetration of a standard

cone into a soil sample was measured at a variety of

moisture contents and the moisture content corresponding to

a penetration of 20 mm was taken as the liquid limit of the

soil tested.

The Plastic Limit test was performed on the soil used in

accordance with BS 1377 part 4. The Plastic Limit was

72

measured as the water content at which the sample deformed

plastically when rolled into a 3 mm thick thread.

3.4.2 Compaction tests

The soil used was subjected to both standard and heavy

compaction tests. These tests aimed to achieve the optimum

moisture content and the maximum dry density (i.e the

compaction characteristics) of the study soil. The

compaction tests were carried out in accordance with the

standard specifications.

In the standard compaction tests the sample was compacted in

three layers, each layer was subjected to 27 uniformly

distributed blows from a standard rammer of 2.5 kg mass,

dropped from a height 30 cm. The mould was then trimmed and

weighed, from which the bulk density of the sample was

determined. A small soil sample was taken from the mould to

determine the moisture content of the compacted soil sample.

Seven soil samples with different moisture contents were

tested.

The procedure followed in the heavy compaction test was the

same as that used for the standard compaction test except in

that the soil was placed in five layers, each of which was

subjected to 56 blows from a 4.5 kg rammer having a free

fall of 45 cm. The moisture content and density were

determined in the manner described above. Seven soil samples

with different moisture contents were also tested.

3.5 SWELLING TESTS

These tests were carried out using an oedometer cell to

measure both the swell percent and swelling pressure. The

swell percent of compacted samples was measured at different

surcharge loads. In this research the oedometer cell was

also used to provide a simple measure of swelling pressure

of compacted soil samples by adopting the pre-swell method

of testing.

73

3.5.1 General arrangement

These swelling tests were performed in a conventional

oedometer apparatus. This piece of apparatus is a standard

equipment in soil testing laboratories. The frame consists

of two steel rods each 10 mm diameter and 300 mm long. The

rods are separated a distance of 200 mm by another steel bar

of 10 mm thick. The bars are arranged in a triangular

pattern, sufficiently spaced to allow easy access for the

oedometer cell. The base plate of the frame has a spigot

fixed so that the oedometer cell can always be located in

the same position. A load hanger system is attached through

the top plate by a reverse screw threaded shaft which allows

it to be raised or lowered on to the sample. At the front

edge of the shaft hanged a weight hanger and at the back

edge fixed a counter balance weight. The surcharge weight

under which the sample was tested positioned on a weight

hanger.

The oedometer cell as shown in Figure 3.1 consists of a

stiff ring, two porous discs, the soil sample and the top

and bottom platens. Normally the dimension of the ring is 76

mm inner diameter by 20 mm height and 1 mm thick. The top

and bottom platens consist of porous discs which are placed

in contact with the sample faces and used saturated by

boiling in de-aired water. It is important that the two

surfaces of the soil sample which in contact with the porous

stones should be plane.

The vertical displacement (axial strain) of the sample was

measured by observing the vertical movement of the top

platen using a dial gauge positioned on top of it. The test

was performed to provide a simple measure of the swell

percent and swelling pressure of the soil tested.

3.5.2 Testing procedure

Each sample was prepared and then placed in a standard

oedometer cell. Two types of experiment were under taken.

The first series of experiments measured the swell percent

at different surcharge loads. A number of samples with

different moisture contents were compacted to different dry

densities. These samples were tested under three different

74

surcharge pressures 7, 25 and 50 Kpa. This series of

experiments investigated the effects of different surcharge

pressures on the volumetric changes of the sample. A

detailed description of the procedure adopted in testing

will be given in the following section.

The second series of experiments used soil samples that were

compacted, again at a wide range of moisture content and

different dry densities. These samples were initially

allowed to swell under a seated pressure of 1 psi (7 kpa) and then consolidated until retained back to its initial

volume. This series of experiments measured the swelling

pressure.

3.5.2.1 Swelling under loading

In this project the swell percent was measured as defined by

Holtz (1959) and described in section 2.5.1. The soil

samples were prepared and compacted at a variety of water

contents and dry densities and were tested under three

different surcharge pressures of 7, 25 and 50 KPa.

For each test a standard oedometer sample was trimmed, the

trimmed part of the sample was used for determining the

sample moisture content. The lower saturated porous stone

was placed on the base of the unit. The prepared sample

weighed and then positioned on the lower porous stone, the

saturated upper stone and loading platen were then added.

The whole unit was placed on the pedestal in the loading

arrangement, the lever arm was supported and the dial gauge

which seated on the ball bearing was securely fastened and

checked for sufficient vertical travel. After placing the

oedometer cell in the frame a surcharge load was applied on

the sample by placing a certain weight on the weight hanger

and the level of the liver arm was adjusted by a water bulb

tube. Then the dial gauge that used for measuring the

swelling was set to zero or any starting reading. After 5

minutes distilled water was added to the cell until the

sample was covered. The sample was let to swell to a period

of two to three days until the swelling value was stable.

During the testing period, water was added to ensure that

the sample was in direct contact with water. The final dial

75

gauge reading was recorded when no further changes in the

reading. The sample was quickly removed from the oemdometer

cell.

3.5.2.2 Swelling pressure

The swelling pressure of a compacted soil sample was

measured using a pre-swelled method which is also called a

consolidation method (see section 2.5.2). The swelling

pressure in this method is defined as the external pressure

required to consolidate a pre-swelled sample until it goes

back to its initial volume. In this study a compacted soil

sample was first allowed to swell under a light pressure

equal to 1 psi (7 kpa) until reaching the peak swelling, then it was consolidated by the increase of applied

surcharge loads until the sample retained back to its

initial volume. The pressure required for this is the

swelling pressure. The procedure adopted for measuring the

swelling pressure of a sample will be described in the

following paragraphs.

The compacted soil sample after prepared was then placed in

the oedometer cell with dry porous discs above and below.

The oedometer cell positioned in the frame and then the

loading hanger system was lowered until the sample was under

a small seating pressure 1 psi( 7 Kpa). The seating

pressure was used to ensure that the soil, porous stone and

load system were in full contact prior to wetting. This also

removed any looseness in the fit of the threaded ring

system. Distilled water was then added to the cell. The

sample was allowed under the seating pressure to swell on

the addition of water until when there was no changes in the

dial gauge readings (i.e. reaching the peak swelling).

After swelling under 1 psi surcharge pressure had been

completed, the sample was then subjected to consolidation

stages in order to retain back the sample to its initial

volume by subsequently increment of loads. The surcharge

load was increased each day by an amount approximately equal

to the load under which the sample had consolidated. The

sample was allowed to consolidate until the sample retained

to its original volume. The final pressure was taken to be

76

the swelling pressure of the sample tested. After

consolidation was completed, the sample was unloaded and the

cell was dismantled from the frame.

3.6 CBR TESTS

The California Bearing Ratio (CBR) penetration test, as

described earlier (Section 2.6.1) is a widely used test in

highway engineering for measuring the strength of the road

subgrade material. The tests performed on compacted samples

for measuring both the soaked and unsoaked CBR. Some of the

samples were tested directly as remoulded while others were

subjected to a soaking process prior to the penetration

test. The boundary conditions of testing were chosen to

simulate the boundary conditions that a subgrade soil would

experience below a road pavement.

The CBR measurements from these tests were able to give a

useful indication of the soil strength at different

conditions of testing as well as allowed an investigation of

the relationship between the combination of the soil initial

state parameters (or soaking state parameters) and the

unsoaked (or soaked) CBR.

3.6.1 General arrangement

The fundamental equipment of a standard testing machine is

the penetration piston, which consists of a cylindrical

plunger having an end diameter of exactly 49.6 mm. The load

transmitted to the sample via the plunger fitted into the

upper part of the loading machine where a proving ring is

fixed. The application of the load was performed manually at

a constant rate of 1 mm per minute to force the plunger to

penetrate the soil sample. A sensitive dial gauge fixed into

a proving ring of the loading frame measures the applied

load. The penetration of the plunger in the sample is

measured externally using two resistive displacement dial

gauges with a nominal stroke of 25 mm. As the plunger

penetrates axially into the soil sample the dial gauge

measures this penetration depth. Figure 3.2 shows the

arrangement of a standard CBR machine.

77

The compacted samples and those subjected to soaking process

after completion of compaction were then transferred to the

penetration machine for measuring the CBR. For soaked or

remoulded (unsoaked) sample the testing procedure adopted

was the same in measuring the CBR value.

3.6.2 Testing procedure

The soil samples were prepared to water contents between 7%

and 35% and were allowed to equilibrate as described in

Section 3.3. For these water contents samples were compacted

to different dry densities using the same method as in

Section 3.3. The procedure adopted in testing the samples

can be subdivided into two stages namely the soaking stage

and the penetration testing stage. In the soaked CBR test

the samples were subjected to a soaking process prior to the

penetration test. On the other hand, the unsoaked CBR tests

were performed on samples not subjected to soaking but

directly tested as remoulded.

3.6.2.1 Soaking process

This process was performed on compacted sample before the

penetration testing stage. A coarse filter paper was placed

on a perforated base plate and then the mould contained

compacted sample was inverted and placed on the filter paper

so that the compacted sample was in contact with it. A swell

plate with adjustable stem was placed over the sample and a

5 kg surcharge weight was imposed upon it. The complete

assembly was then immersed in a tank filled with water. A

tripod and a dial indicator attached to the stem for

measuring the sample expansion and an initial dial reading

was set to zero or a starting reading. The assembly was left

immersed in water for 96 hours and during this period the

dial indicator reading was recorded at the ends of 1, 2, 4,

8, 24, 48, 72 and 96 hours. At the end of the soaking

period, the mould was removed from the water tank, the water

poured off the top and the surcharge weights were removed.

The sample was ready for penetration test.

78

3.6.2.2 Penetration test

For each prepared sample a surcharge weight of 5 kg was

placed on top of it and then the combination was seated

under the penetration piston of the loading machine. The

penetration piston was placed in contact with the center of

the sample. The initial reading of the proving ring (the

load indicator) was recorded or might set to zero. The

penetration dial gauges were mounted on the top of the mould

and were set to zero. The vertical load, which was applied

manually, was increased gradually to force the plunger to

penetrate the sample at a constant rate of penetration 1 mm

per minute. The applied load was recorded at intervals of

penetration of 0.5 mm to a total penetration not exceeding

12.5 mm. The test stopped at 12.5 mm penetration or when the

load indicated a peak before 12.5 mm penetration arrived.

After recording the final reading, the surcharge weight was

removed from the top of the sample and then the mould was

removed from the penetration mechanism. The soil was then

removed from the mould by an extruder and soil pieces were

taken for determining the moisture content. For unsoaked

samples, soil pieces taken from top, middle and bottom part

of the mould for determining the moisture content, whereas

the soaked samples moisture contents were determined before

the soaking process by taken trimmed soil pieces from the

compacted soil.

3.7 TRIAXIAL COMPRESSION TEST

The triaxial test is by far the most common and versatile

test for measuring the shear stress of fine grained soils.

The tests performed using a standard triaxial apparatus were

conventional undrained triaxial compression tests and were

used to determine the value of shear stress of the samples

tested. This type of triaxial test is known to be quick and

simple test. The tests were performed at three different

cell pressures of 50, 100 and 200 kPa.

The samples were compacted at different moisture contents

and dry densities. A cylindrical sample of soil was prepared

and then enclosed in a pressurized chamber which subjected

the sample to compressive stresses in three mutually

79

perpendicular directions. The vertical compressive stress

was then increased in excess of the horizontal stresses

until eventually the sample failed in shear or strained to

such a point that excessive deformation could result.

3.7.1 General arrangement

The triaxial apparatus used in this project was manufactured

by Nanjin Company Ltd in China and is frequently found in

most commercial soil testing laboratories. The conventional

triaxial apparatus has been described in detail by Bishop

and Henkel (1962) in their standard text on triaxial testing

of soils.

The basic features of the conventional triaxial tests are

shown in Figure 3.3. The soil sample is a cylinder with

height about twice the diameter; the size used in this study

is 38 mm diameter. The sample is enclosed in a thin rubber

sleeve sealed to the top platen and to the base pedestal by

rubber O-rings. The sample sits on a pedestal at the central

axis of the apparatus. The cell was formed from a Perspex

cylinder surround by six tension bars, which spaced at an

angle of 60. Their purpose is to connect the base of the main pressure vessel to a top plate to which a load cell is

mounted. The pedestal sits on a column that passes through

the base of the pressure vessel into a rubber rolling

bellofram which seals the cell. Linear bearings placed

around the column allow free, low friction, vertical

movement whilst maintaining the centrality of the pedestal.

The base of the column passes through another bellofram seal

into a lower sealed pressure chamber. By controlling the

pressure in the lower axial pressure chamber it is possible

to control the stress transmitted to the sample in the upper

chamber. The internal load cell is fixed onto a metal rod

that passes through two O-ring seals. The rod then screws

internally into a thread ring which allows the internal load

cell to be raised or lowered. A ball seated at the upper

face of the top Perspex platen to allow the axial load to be

in contact with the sample by means of a plunger fitted in a

rotating bush in the top plate of the cell.

80

The apparatus is provided with two pressure control units

that vary the pressure in the cell and the pressure in the

axial ram applied to the sample. The input air pressure is

supplied by a centralized air compressor that can deliver a

maximum air pressure of 1000 kPa. The air/water interfaces

consist of sealed cylinders in which water is pressurized by

the air. Air pressure enters at the top of the cylinder

which then forces water out through a pipe at the base of

the cylinder.

In the base plate of the cell there are three different

valves. From the base pedestal a connection via a valve

connected the pore water pressure of the sample to a pore

water pressure measuring device, from the top Perspex platen

is a connection via a valve to a burette system to allow

drainage and volume change observations. A further valve is

connected to the base plate of the cell, allowing water in

the cell and surrounding the externally sealed sample to be

pressurized.

3.7.2 Instrumentation

The axial load is measured using a dial gauge fixed into a

proving ring supplied by Nanjin Company Limited in China.

Because the load indicator is mounted inside the cell the

deviator force measurements are not affected by the

frictional errors that external load indicators suffer. The

pressure of the cell water is measured by means of a Master

pressure gauge inserted directly into the base of the cell.

This Master gauge works in a pressure range from 0 to 1000

kPa. The axial strain is measured externally using a

resistive displacement dial gauge with a nominal stroke of

10 mm. As the sample deforms axially the dial gauge measures

the relative displacement between the top of the cell that

is firmly connected to the top of the sample and the axial

ram which acts at the base of the sample. The compliance of

the triaxial cell and measuring devices (displacement and

load) may cause significant errors in the measurements

taken.

81

3.7.3 Testing procedure

After compaction a sample tube of 38 mm in diameter and 76

mm in length was taken from the preparation tube. The

prepared sample was then accurately weighed to determine the

bulk density. Three small soil samples were used to measure

the sample moisture content. The sample was enclosed in a

thin rubber membrane in order to be effectively sealed from

interaction with water. The main function of the surrounding

membrane was to prevent moisture movement across it and any

potential moisture seepage between the membrane and the

upper platen and base pedestal was prevented by the champing

effect of two O-ring seals. The sample was then placed

inside a Perspex cell and seated on the base pedestal. The

top drainage leads were connected to the top cap and then

the outside of the cell was bolted into position. After

readings from the load cell and strain dial gauge indicators

were zeroed and the cell filled with water and a cell

pressure applied by pressurizing water the test started.

The sample was loaded axially through the reaction with a

ram bearing on the top cap as the cell was driven upwards by

a motor and gear box. The axial load applied on the sample

was shown by a proving ring fitted externally to the

pressurized cell. This load was recorded at predetermined

strain intervals and since this apparatus was not

instrumented so the measurements of the axial load and axial

strain were recorded manually. The test stopped when shear

failure of the sample occurred which indicated by a rapid

increasing in strain without any change in axial load or

when the recorded strain reached the probable failure strain

which is normally around twenty percent rate of strain.

This quick undrained compression test was performed at three

different cell pressures. No pore pressure measurements were

taken since the samples were unsaturated. This simple fairly

rapid compression test provided an easy and quick method for

determining a value of shear stress for the remoulded

samples.

82

3.8 ERRORS AND ACCURACY

The function of a measuring device is to sense change in its

physical environment such as load or displacement. The

measuring device must be capable of faithfully and

accurately detecting any changes that occur in the measured

quantity. To obtain the best performance from any instrument

it is very important to understand the principles by which

the instrument functions, its basic characteristics and the

errors which influence the measurement of the reading.

Without an understanding or appreciation of these factors

the reliability of the data obtained using such instruments

should be questioned. It is also important to understand the

relative magnitudes and hence importance of the various

factors influencing the behaviour of the measurement system.

3.8.1 Error

The algebraic difference between the measured value and the

true value is termed the error of the measurement. The error

of a measurement results from the combination of a number of

individual errors. By understanding how the individual

errors arise, corrections to the final data may be made

there by increasing the overall accuracy of the measurement.

The major causes of errors in the measurements are:

Interference (noise) – electrical, electromagnetic and

electrostatic pick-up in the measuring system which can

superimpose large variations in the output signal. It is

important to isolate the measuring system from any

external influences such as generators and shield the

components from interfering with each other.

Zero Drift – this represents a variation in the output with time under constant conditions input. It is caused

by changes in the ambient conditions of pressure and

temperature or aging of the electrical components. The

zero drift of the load or displacement indicators can be

assessed by comparing the variation between the readings

corresponding to the zero conditions at the beginning and

end of the test.

83

3.8.2 Accuracy and Precision

The accuracy of a measurement can be defined as the

closeness with which the reading approaches an accepted

standard value. Accuracy is a relative term influenced by

static error, drift, reproducibility and non-linearity. In

any experiment the accuracy is numerically equal to the

referred error value, i.e. the degree of error in the final

result. The accuracy is determined by calibrating under

certain operation conditions and is expressed as a

percentage at a certain point of the scale. For a complete

system the accuracy is dependent upon the individual

accuracy of the sensing element and the manipulating device.

The overall accuracy can be determined by summing up the

accuracy limits of the individual components i.e. if a1, a2 and a3 are the accuracy limits then overall accuracy (A) is expressed as A = (a1 + a2 + a3). In practice it is not probable that all the elements of the system will have the

greatest static error at the same time so to account for

this the root mean square accuracy is often specified, this

is express as A = (a12 + a22 + a32).

The precision of an instrument is the closeness with which

individual measurements are distributed about their mean

value. It is a measure of the scatter of the set of readings

among themselves. It includes the uncertainty in the reading

due to random errors, however it gives no information

relating to accuracy. The precision of an instrument is

found by calculating the mean value of the scatter of the

individual measurements.

3.8.3 Linearity

To enable more accurate data gathering and reduction most

measuring devices are designed to produce a linear

relationship between the input and output. The closeness of

a calibration curve to a specified straight line is known as

the linearity of the measuring device. The linearity of the

instrument is expressed as the maximum deviation of the

output curve from the best-fit straight line; the value of

linearity is given as a percentage of the full scale. In

order to avoid misleading statements the range to which the

linearity refers to should be stated.

84

CHAPTER FOUR

RESULTS AND ANALYSIS

4.1 INTRODUCTION

Presented in this chapter is the results and analysis of the

experimental work. The results of the tests carried out

during the research project are analysed. Some data reported

by previous investigators which related to the objective of

the research will also be analysed.

In the first section, The results of the preliminary tests

such as index testing and compaction tests are described.

The main results are then presented as follows: for each

type of testing, the results and their analysis are shown.

Summaries of the experimental results are given in Tables

4.1 to 4.9.

The analysis of results will be carried out and presented in

this chapter. As the main objective of this project is to

study the swelling behaviour and strength characteristics of

compacted soils in order to develop a new factor concept.

Therefore the results from the experimental work and those

reported by previous researchers are analysed and used to

develop this concept.

4.2 PRELIMINARY TESTS RESULTS

The preliminary tests performed in this project were index

tests, clay content, specific gravity determination and

compaction. These tests provided standard data for the soil

which could be used in subsequent analysis. The preliminary

tests were aimed to classify the soils within the existing

standard methods and quantify their properties. The clay

content and specific gravity were determined as presented in

Table 4.1.

85

4.2.1 Index tests

These tests provide a measurement of the liquid limit and

plastic limit. The methods used for testing are the British

standard specifications for the determination of the Liquid

Limit and Plastic Limit. As discussed in Section 2.3.1 the

index tests, although based on simple empirical tests, can

be correlated with more fundamental properties such as shear

stress (Skempton and Northey (1953)) and compressibility

(Wroth and Wood (1978)). Previous research in to the use of

index tests (Sherwood (1970)) has shown that these tests are

subjected to variability. Sherwood (1970) sent samples of

the same soil to different laboratories for the

determination of the liquid limit. The imprecision in the

results were attributed to defects in the apparatus such as

damaged cones as well as operator induced errors. Since in

this research later analysis would use the index tests to

develop empirical equations, the liquid limit and plastic

limit were performed using carefully checked apparatus and

different but experienced operators.

The liquid limit determined the water content at which the

soil had weakened so much that it started to flow like a

liquid. On the other hand, the plastic limit determined the

water content at which the soil had became so brittle that

it crumbled. The index tests results are given in Table 4.1.

4.2.2 Compaction tests

These tests were performed to study the moisture content/dry

density behaviour of samples compacted at two different

compactive efforts. The experimental results of compaction

tests are listed in Table 4.2 and indicated in Figure 4.1.

The curves produced give a measure of the variation of the

achievable dry density for the light and heavy compactive

efforts as the moisture content of the soil changes. Studies

into the compaction of soils using different compactive

efforts (Parsons (1992)) indicate that the maximum

achievable dry density as determined by the heavy (modified)

compaction test is higher than that determined by the

standard compaction test, this is demonstrated by the

compaction curves of the soil used in the experiments.

86

The method of compaction used in the preparation of the

samples for the swelling and triaxial tests differed from

the Standard Proctor compaction method (Section 3.3). These

samples were prepared by static compaction methods which can

easily be compacted and may have higher densities than

similar samples which have been compacted by dynamic

compaction.

4.3 MAIN TESTS RESULTS

The main laboratory tests of this research project include

the swelling, CBR and triaxial tests. These tests were

carried out using in a variety of apparatus. In excess of

104 individual tests were performed under different

conditions of initial water content and dry density. The

results of different types of tests are fully described in

the following sections.

4.3.1 Swelling tests

The experiments performed in a conventional oedometer cell

were to measure the swell percent and the swelling pressure

of compacted samples. Radial movement was prevented by the

presence of the stiff oedometer ring and axial movement was

allowed against a surcharge load positioned on a weight

hanger.

Thirty nine tests were performed for measuring the swell

percent of samples compacted at a wider range of water

content and various dry densities and tested under three

different surcharge pressures (7kPa, 25kPa and 50kPa). When

the oedometer cell was flooded with water there was a rapid

increase in the dial gauge reading as the soil absorbed

water and started to expand against the top platen of the

cell. Because the sample was prevented by a stiff proving

ring from expanding radially, only axial strains were

allowed. During the testing period, it was observed that

most of the swelling took place within the first day hours,

but in some tests the swelling increased to a peak value

after two days of testing. This is probably due to the fact

that, the rate of water absorption in dense soils being

slower than that in loose soils. The measured swell percent

87

of the soil observed to be influenced by the initial dry

density and water content as well as the surcharge load

under which the sample was tested. As the surcharge load was

increased the resulting amount of swell percent was reduced.

The highest swelling values were observed in samples that

compacted at lower water contents and higher dry densities

and tested under light surcharge loads. The experimental

results of the swell percent measured at different pressures

and the initial state of the tested samples (i.e. water

content, dry density, void ratio and swell percent) are

given in Tables 4.3 to 4.5.

The oedometer cell was also used to measure the swelling

pressure of the compacted soil samples. The test samples

were compacted at different initial water contents and dry

densities. The sample was initially allowed to swell under a

seating pressure of 1 psi and after reaching a peak swelling

value, it was then compressed by adding weights. The weights

were added each day by a certain amount as described in

Section 3.5.2.2. By increasing the weights sufficiently it

was possible to retain back to the started dial gauge

reading. The pressure compressed the expanded sample to its

original volume was considered as the swelling pressure. It

was observed that both the initial water content and dry

density of the soil sample influenced the amount of swelling

pressure: low initial water contents and high dry densities

have greater swelling pressure values than drier denser

soils. The test results of these experiments are presented

in Table 4.6.

4.3.2 CBR tests

These tests were aimed to give a useful indication of the

soil strength at different testing conditions. The tests in

this project were performed on remoulded as well as soaked

samples in accordance with BS 1377 part 110. In the first

series of tests samples which prepared at different water

contents and dry densities were directly subjected to CBR

penetration test. The soil was compressed by axial load

which applied at a constant rate to penetrate the plunger

into the soil. The measured penetration force found to be

influenced by the soil initial state (water content and dry

88

density). For all samples tested the highest penetration

forces were measured in samples prepared at high dry

densities and low water contents. As the dry density was

decreased and water content increased the resulting amount

of force was reduced. This demonstrate that samples with

lower water contents and compacted to higher densities have

high CBR values (Section 2.6.1). The results from these

tests are listed in Table 4.7.

The second series of experiments were performed on samples

which prepared at different water contents and dry densities

but subjected to soaking prior to penetration testing. The

measured soaked CBR values of the samples as shown in Table

4.8 were clearly seen to be much less than those measured

unsoaked CBR. The reduction in the CBR value can be

attributed to the softening of the soil in the soaking

process. That is, as the soil became wetter a general

softening of the sample occurred and the ability of the

sample to resist the penetration of the plunger tended to be

very low.

4.3.3 Triaxial compression test

This test was used to measure the shear stress of compacted

samples. The sample was compressed to vertical stress

increased at a constant rate while maintaining the side cell

pressure constant. The test results for soil samples when

subjected to undrained compression are given in Tables 4.9

to 4.11. It was seen that after approximately 20% axial

strain the peak shear stress was reached and a rupture plane

formed in the sample. An analysis of the failure plane

revealed that its surface was not smooth but had some

undulations, possibly a remnant of the clods of soil that

had not been totally remolded during compaction.

4.4 ANALYSIS OF SWELLING RESULTS

In this section analysis of swelling results is presented.

It will cover the analysis of swell percent and swelling

pressure data which obtained from the experiments carried

out in this project and others reported by investigators.

Based on the empirical results a factor, which is formed by

89

a combination of easy measured soil parameters, will be

developed. A linear relationship of this factor with the

swell percent as well as the swelling pressure is

investigated. By the analysis of the experimental swelling

data this relationship will be verified.

4.4.1 Development of the initial state factor

When performing laboratory experiments on compacted

expansive soils it is difficult to control and measure the

swell percent and swelling pressure without using sensitive

and well-controlled apparatus and following restrictedly the

standard procedures of testing. So in order to by-pass this

problem, an alternative indirect method to assist in

measuring the swelling was considered. The basic idea

consisted of trying to relate both the swell percent and

swelling pressure of compacted soils to a factor which is

formed by a combination of easy measured soil parameters

such as dry density, water content and void ratio and is

termed the soil initial state factor. The soil parameters

are combined in such a way to reflect the influence of each

of them in the measured swelling (swell percent and swelling

pressure). The factor can be applied for soils tested as

compacted or undisturbed and is formed of initial state soil

parameters such as water content, dry density and void ratio

and is termed as the initial state factor.

4.4.1.1 Definition of the factor

The initial state factor of compacted expansive soil is

defined as a combination of the soil initial state

parameters such as dry density, water content and void ratio

and can be expressed thus:

1.41

eF d

i

Where:

d is the initial dry density of soil

is the density of water

90

is the initial water content

e is the initial void ratio.

In fact, among these variables e is supernumerary since it

dependents on dry density according to the following

relationship:

2.41d

sGe

Where sG is the specific gravity of soil.

4.4.1.2 Derivation of the factor

This factor was empirically formulated on the basis of the

following reasons:

The results of swell percent data reported by Chen (1975)

Figure 2.14 and Kassif et al (1965) Figure 2.15 clearly

indicate that a linear relationship can be drawn between

the swell percent and the initial dry density of the

samples of the same soil, having the same moisture

content and subjected to the same surcharge pressure.

Hence, the swell percent (S) can be assumed to be

directly proportional to the dry density (d) i.e.

3.4tan; tconsisS d

The results of swell percent data reported by Zein (1985)

Figure 2.17 and Chen (1975) Figure 2.18 suggest that an

inverse linear relationship exists between the swell

percent and the moisture content of the samples of the

same soil, having the same dry density and subjected to

the same surcharge pressure. Empirical examination of

these results, however, suggests that an equally good fit

can be obtained by assuming a hyperbolic relationship

instead of an inverse linear one. Hence, it can be

assumed that the swell percent (S) is inversely

proportional to the initial moisture content() i.e.

4.4tan;1

tconsisS d

91

As reported by Brackley (1975) and shown in Figure 2.21

that at a constant water content the swell percent

increases with decreasing in void ratio. As can be seen

from Figure 2.21, an inverse linear relationship exists

between the swell percent and the void ratio of the

samples of the same moisture content. So it can be

assumed that the swell percent (S) is inversely

proportional to the void ratio (e) i.e.

5.4tan;1

tconsiswe

S

Hence from the above equations and using

d instead of d

to make the term dimensionless results in:

6.41

eS d

From the above equation it can be concluded that the swell

percent (S) is directly proportional to the initial state

factor (Fi). This relationship can be expressed as:

7.4iFS

4.4.2 Swell percent and factor Fi relationship

This section analyses the swell percent results in order to

investigate its relationship with the developed initial

state factor. The dependence of the swell percent on

surcharge pressure, soil plasticity and clay content will be

studied to develop an empirical equation.

As previously verified that a direct linear proportional

relationship exists between the swell percent (S) and the

developed initial state factor (Fi) as indicated by equation

4.7. Hence, this relationship can be assumed as a linear

relationship and expressed thus:

8.4bFaS i

Where a and b are constants of the linear equation.

92

To investigate the validity of this linear relationship

between the swell percent and the developed factor Fi, the

results obtained in this study and others reported by

previous investigators were analysed. These results were for

undisturbed and compacted soils of different origins.

Different methods of compaction, testing procedures and

applied surcharge pressures had been used by the

investigators are listed in Table 4.12(a).

The plots of swell percent versus initial state factor for

these soils are shown in Figures 4.2 to 4.15. The straight

line shown in each of the plots is the linear regression of

the swell percent to the factor. In each of Figures 4.2 to

4.15, the data from which the values of the swell percent

(S) and the factor (Fi) were obtained are given on the right

hand side of the figure. For each of the data analysed the

magnitude of the correlation coefficient (R) which indicates

the degree of closeness of the linear relationship between

the swell percent and the factor is given in Table 4.12(a).

The plots in Figures 4.2 to 4.15 have clearly demonstrated

that a direct linear relationship exists between the swell

percent and the factor for all the data analysed. The

magnitude of the correlation coefficient (R) as presented in

Table 4.12(a) is in the range 0.80 to 0.99 which indicates a

very strong linear relationship between swell percent and

initial state factor. This linear relationship can be

expressed as:

9.4)( 0FFMS i

Where:

S is the swell percent

iF is the initial state factor

0F is the value of F at which no swelling occurs

M is the gradient of the straight line.

93

The values of the constants Fo and M for each of the plots

in Figures 4.2 to 4.15 are listed in Table 4.12(a) and

collectively illustrated in Figures 4.16 to 4.19. As can be

seen, differences in soils’ composition, surcharge pressures and methods of testing result in different Fo and M values.

The results of the studied samples as shown in Figure 4.17

and Table 4.12(a) suggest that, as the surcharge pressure

increases from 7 to 50 kPa, the value of M decreases from

0.90 to 0.53 and the value of Fo increases from 1.72 to

3.73. This confirms with the results of Yevnin and Zaslavsky

(1970) as shown in Figure 4.18 and listed in Table 4.12(a).

From Figure 4.19 and Table 4.12(a), it can be noticed that,

although the soils were tested under similar surcharge

pressures, the differences in soil types (i.e. different

composition or expansiveness) and methods of testing seem to

result in different Fo and M values. It can, however, be

noticed from Figure 4.19, the soils have relatively high

clay contents, liquid limits and plastic limit (such as

soils(5)&(12)) seem to have higher M and lower Fo values

than soils have relatively low clay contents, liquid limits

and plastic limits (such as soil(13)). For expansive soils,

as indicated in Section (2.3.1), the higher the clay

content, liquid limit and plastic limit the higher the

expansiveness of the soil. It can be said from Figure 4.19

that the increase in soil expansiveness increases the value

of M and decreases the value of Fo.

The values of Fo and M as listed in Table 4.12(a) and the

plasticity index (PI) and clay content (C) were normalised

with respect to the surcharge pressure (P). The results of

this are shown in Figures 4.20 and 4.21. A power regression

has been used to provide best fit lines. When normalising

the Fo values as plotted in Figure 4.20, the equations of

the best fit lines are given by:

10.4165.0100

58.00

P

PILog

P

FLog

11.4063.0100

56.00

P

CLog

P

FLog

94

By adding equations 4.10 & 4.11 and rearranged to represent

Fo in the following way:

12.4)()(0474.056.058.043.0

0 CPIPF

And when normalising the values of M, plasticity index (PI)

and clay content (C) with respect to the surcharge pressure

(P). The normalised relationships were plotted in Figure

4.21. The equations of the best fit lines are given by:

13.4021.1100

33.1

P

PILog

P

MLog

14.4859.0100

32.1

P

CLog

P

MLog

By adding equations 4.13 & 4.14 and rearranged to express M

in the following way:

15.4)()(01.032.133.132.0

CPIPM

Then by substituting the expressions of Fo and M (as given

by equations 4.12 and 4.15) in equation 4.9 allows the

definition of swell percent as a function of initial state

factor as well as surcharge pressure, plasticity index and

clay content. This result is expressed by the following

equation:

16.4)()()(0474.0)()(01.056.058.043.032.133.132.0

CPIPFCPIPS i

Where:

S is the swell percent(%)

iF is the initial state factor

P is the surcharge pressure (kPa)

PI is the plasticity index (%)

C is the clay content (%)

95

The calculated values of the constants Fo and M using above

developed equations (4.12)&(4.15) and the ratio of the

calculated and measured swell percent values for each of the

data analysed are listed in Table 4.12(b). As can be seen in

table that the ratio between calculated swell percent using

equation (4.16) and measured swell percent within the range

0.70 1.40, which clearly demonstrated the validity of the developed equation. This result allows the definition and

measurement of swell percent as a function of initial state

factor, surcharge pressure and soil expansiveness (as

indicated by equation (4.16)).

4.4.3 Swelling pressure and factor Fi relationship

A linear relationship between the swelling pressure and the

initial state factor exists as previously discussed in case

of the swell percent. To prove the validity of this

relationship the present study results and data reported by

O’Connor, 1994 were analysed.

The plots of the swelling pressure (SP) versus initial state

factor (Fi) for the analysed samples are shown in Figures

4.22 to 4.25. The straight lines drawn in these plots and

the values of the correlation coefficient (R) as given in

Table 4.13(a) seem to indicate that for all the data

analysed a very goad linear relationship exists between the

swelling pressure (SP) and the factor (Fi). The straight

line shown in each of the plots is the linear regression of

the swelling pressure and the factor relationship and can be

expressed by the following equation:

17.4)( 0FFMSP i

For each plot shown in Figures 4.22 to 4.25, the values of

the straight line intercept (Fo) and gradient (M) and the

magnitude of the correlation coefficient (R) are given in

Table 4.13(a). The plots of the swelling pressure analysis

results are collectively illustrated in Figure 4.26. From

this figure it can be noticed that differences in soils

expansiveness (clay contents, liquid limit and plasticity

index) and methods of testing result in different Fo and M

values. The soils with high clay contents, liquid limits and

96

plasticity indices (e.g. soil (4)) seem to have higher M and

lower Fo values than those with low clay contents, liquid

limits and plasticity indices (e.g. soil (2)). It can be

noticed that for expansive soils increasing in soil

expansiveness will increase the M value and decrease the Fo

value. This was demonstrated by Figures 4.27 & 4.28. The

equations of the best fit lines of Fo value relations are

given by:

18.41366.0100

10.20

PILogFLog

19.4084.0100

54.10

CLogFLog

By adding equations 4.18 & 4.19 and rearranged to represent

Fo in the following way:

20.4)(719)(570754.11.2

0

CPIF

Similarly the relations of M value with plasticity index

(PI) and clay content (C) were plotted in Figure 4.28. The

equations of the best fit lines are given by:

21.4567.2100

81.2

PILogMLog

22.435.2100

21.2

CLogMLog

By adding equations 4.21 & 4.22 and rearranged to express M

in the following way:

23.4)(0042.0)(0004.021.281.2

CPIM

The definitions of Fo and M as expressed by equations 4.20

and 4.23 were substituted in equation 4.17. This allows the

definition of swelling pressure as a function of initial

state factor as well as plasticity index and clay content.

This result is expressed thus:

97

24.4)(719)(5707)(0042.0)(0004.054.11.221.281.2 CPIFCPISP i

Where:

SP is the swelling pressure (kPa)

iF is the initial state factor

PI is the plasticity index (%)

C is the clay content (%)

The results obtained when using the above equation (4.24)

for each of the data analysed are listed in Table 4.13(b).

It is clear that the ratio of calculated swelling pressure

using equation (4.24) and measured swelling pressure within

the range 0.80 1.20, which strongly verified the validity of the relationship as expressed by equation (4.24) between

the swelling pressure, the initial state factor and the soil

expansiveness for all the data analysed. This result allows

the definition and measurement of swelling pressure (as

expressed by equation (4.24)) as a function of initial state

factor and soil expansiveness.

4.5 ANALYSIS OF CBR RESULTS

The experimental data of soaked and unsoaked CBR are

analysed in this section. The data analysed of unsoaked CBR

will be used to investigate the validity of the linear

relationship of the CBR and the initial state factor. A

soaking state factor which is formed of easy measured soil

parameters in the saturated state will be developed. The

Soaked CBR data are analysed and used to investigate its

relationship with the soaking state factor. The previous

researchers’ data are also analysed and used with the experimental results to demonstrate the validity of the CBR

relationship with the new factor.

4.5.1 Unsoaked CBR and factor Fi relationship

To investigate the validity of the linear relationship

between unsoaked CBR and the initial state factor (Fi) which

was developed in Section (4.4.1.2), the results obtained in

this study and others reported by El-Tayeb (1991), Murphy

(1966) and Glanville (1951) were analysed.

98

The unsoaked CBR and the initial state factor Fi

relationship of these analysed data are shown in Figures

4.29 to 4.33. The plots in these figures and the values of

the correlation coefficient (R) as listed in Table 4.14(a)

have clearly demonstrated that a direct linear relationship

exists between unsoaked CBR and the initial state factor for

all the data analysed.

The straight line shown in the plots of Figures 4.29 to 4.33

can be expressed as:

25.4)( 0FFMCBRunsoaked i

The summary of results of all the data analysed is given in

Table 4.14(a) and drawn in Figure 4.34. As can be seen, the

magnitudes of M and Fo are different for different types of

soils. The soils have relatively low liquid limits and

plasticity indices such as soil (4) and (3) seem to have

higher values of M and Fo than soils with relatively high

liquid limits and plasticity indices (expansive soils) such

as soil (2). It can be noted that in expansive soils,

increasing in degree of expansiveness will decrease M and Fo

values. The relationship of Fo and M with plasticity index

was plotted in Figure 4.35. The equations of the best fit

curve and line are expressed thus:

26.43022.2035.02

0 PIPIF

27.4)(161582.1 PIM

By substituting the above two equations in the general

equation 4.17 and rearranged to express unsoaked CBR as:

28.43022.2)(035.0)(1615282.1

PIPIFPICBRunsoaked i

Where:

iF is the initial state factor

PI is the plasticity index (%).

99

The analysis results when using the above equation (4.28)

are presented in Table 4.14(b). As seen in table the ratio

of calculated CBR using equation (4.28) and measured CBR

(0.90 1.30) for all the data analysed proved the validity of the developed equation (4.28). This result suggested that

unsoaked CBR could be defined as a function of soil initial

state factor and the plasticity index.

4.5.2 Development of the soaking state factor

When a given soil subjected to soaking by immersing in water

for about four days as in the soaked CBR test, in this case

the soil initial state parameters will be affected by the

soaking conditions. It was known that saturation due to

soaking has great influence on the CBR values of expansive

soils (Section (2.6.1.3)). Therefore the factor in the

soaking state is suggested to be formed by a combination of

saturated dry density, saturated water content and the

amount of swelling during the soaking period. This factor is

called the soaking state factor and is expressed as:

29.4)1(

1.

SF

sat

d

s

sat

Where:

satd is the saturated dry density

sat is the saturated water content

is the density of water

S is the amount of swelling due to soaking.

The soaking state factor Fs was developed on basis of the

following reasons:

As previously discussed in Section (2.6.1.3) and the

soaked CBR test results of this study as presented in

Figure 4.36 and those reported by El-Tayeb (1991) in

Figure 4.37. It is clear that the soaked CBR values are

decreased with increasing in the calculated saturated

100

water contents of samples have the same saturated dry

density. hence, it can be assumed that the soaked CBR is

inversely proportional to the saturated water content

( sat ) i.e.

30.4tan;1

tconsisCBRSoakedsatd

sat

Based on the reasons discussed in Section (2.6.1.3) and

the data analysed of Figures 4.36 and 4.37, it can be

suggested that a linear relationship may exist between

the soaked CBR and the calculated saturated dry density

of the samples having the same saturated moisture

content. thus, the soaked CBR can be assumed to have a

direct proportional relationship with the saturated dry

density (satd ) i.e.

31.4tan; tconsisCBRSoaked satdsat

As reported by Yoder (1975) and shown in figure 2.25. It

can be seen that increasing in the amount of swelling of

expansive soils during the soaking period reduces the

soaked CBR value. Whereas in granular soils swelling has

no influence on the soaked CBR values during the soaking

period. Hence, for different types of soil, it can be

assumed that the soaked CBR is inversely proportional to

the amount of Swelling (S)(as a fraction of one) as given

by the term (1+S) and expressed thus:

32.41

1

SCBRSoaked

Hence from the above relationships and using

satd instead of

satd to make the term dimensionless results in:

33.4)1(

1

SCBRSoaked

sat

dsat

101

From the above equation it can be suggested that the soaked

CBR is directly proportional to the soaking state factor

(Fs). This relationship can be expressed thus:

34.4sFCBRSoaked

4.5.3 Soaked CBR and factor Fs relationship

Based on the direct linear proportional relationship exists

between the soaked CBR and the soaking state factor as

indicated by equation 4.34. Therefore the soaked CBR versus

the soaking state factor (Fs) relationship can be expressed

in the linear equation as:

35.4cFmCBRSoaked s

Where m and c are the constants of the linear equation.

The present study results and the data reported by El-Tayeb

(1991) and Yoder (1975) were analysed. The plots of the

soaked CBR versus soaking state factor (Fs) for the data

analysed are shown in Figures 4.36 to 4.39. The straight

line of this relationship drawn in each figure and the

magnitudes of the correlation coefficient (R) given in Table

4.15(a) have clearly demonstrate the validity of the direct

linear relationship between soaked CBR and the soaking state

factor.

Each straight line in these plots can be expressed as:

36.4)( 0FFMCBRSoaked s

Where:

0F is the value of Fs at zero soaked CBR value

M is the gradient of the straight line.

For each of the data analysed in plots of Figures 4.36 to

4.39 the magnitudes of the line intercept (Fo), the line

102

gradient (M) and the correlation coefficient (R) of the

linear relationship between the soaked CBR and the factor

(Fs) are given in Table 4.15(a) and collectively illustrated

in Figure 4.40. From this figure it can be noticed that

differences in soils expansiveness result in different Fo

and M values. The soils with relatively low liquid limits

and plasticity indices such as soils (3) and (2) seem to

have higher values of M and Fo than soils with relatively

high liquid limits and plasticity indices such as soils (1)

and (4). Therefore it can be concluded that expansive soils

with high degree of expansiveness have low M and Fo values

and vice versa. Figure 4.41 shows the relationship of M and

Fo with plasticity index. The regression equations of the

curve and line drawn in figure are expressed as follows:

37.486.1255238.81443.02

0 PIPIF

38.4)(10*425.811 PIM

By substituting the above two expressions of Fo and M in the

general equation 4.36 and rearranged to express soaked CBR

thus:

39.486.1255238.8)(1443.0)(10*4225.811

PIPIFPICBRsoaked s

Where:

sF is the soaking state factor

PI is the plasticity index (%)

The analysis results obtained when using equation (4.39) are

given in Table 4.15(b). It is clear that the ratio of

calculated soaked CBR using equation (4.39) and measured

soaked CBR within the range 0.75 1.20, which strongly

verify the validity of the relationship expressed by

equation (4.39). This result allows the definition and

measurement of soaked CBR as a function of soaking state

factor and plasticity index.

103

4.6 ANALYSIS OF SHEAR STRESS RESULTS

A linear relationship between the shear stress and the

initial state factor exists. The present study results and

data reported by O’Connor (1994) were analysed and used to verify the validity of this relationship.

The shear stress () versus initial state factor (Fi) plots for the analysed data are given in Figures 4.42 to 4.47. The

straight line shown in each of the plots is the linear

regression of the shear stress and the factor and can be

expressed thus:

40.4)( 0FFM i

The results of all the data analysed is presented in Table

4.16(a) and collectively drawn in Figure 4.48. The results

and the magnitude of the correlation coefficient (R)

indicate that for all the data analysed a direct linear

relationship exists between the shear stress and the initial

state factor. As can be seen in Table 4.16(a) and Figure

4.48 that the magnitudes of M and Fo are different for

different types of soils and cell pressures. The differences

in soil types (i.e. different composition or expansiveness)

seem to result in different Fo and M values. For soils

tested at a constant cell pressure, the higher the clay

content, liquid limit and plastic limit, the higher the

value of M and the lower the value of Fo (such as soil (6))

and vice versa. The results of the studied samples as shown

in Figure 4.48 and Table 4.16(a) suggest that, as the cell

pressure increases from 50 to 200 kPa, the value of M

increases from 30.54 to 45.68 and the value of Fo decreases

from 8.06 to 6.64. The dependency of Fo and M values on

plasticity index and cell pressure is illustrated in Figures

4.49 and 4.50. For the Fo relations, the regression

equations of the best fit lines are expressed by:

41.415.352.10 PILogFLog

42.414.114.0 30 LogFLog

104

By adding equations 4.41 & 4.42 and rearranged to express Fo

as follows:

43.4)(93.670314.0

3

52.1

0

PIF

And similarly when considering the M value relationships

with plasticity index and cell pressure, the equations of

the best fit lines are given by:

44.4033.004.1 PILogMLog

45.4995.029.0 3 LogMLog

By adding the above two equations and rearranged to present

M thus:

46.4)(94.454.029.0

3

04.1 PIM

When substituting the expressions of Fo and M (as given by

equations 4.45 and 4.46) in equation 4.40 gives the shear

stress as a function of initial state factor, cell pressure

and plasticity index. This result can be expressed in the

following equation:

47.4)(93.6)(703)(94.4)(54.014.0

3

52.129.0

3

04.1 PIFPI i

Where:

is the shear stress (kPa)

iF is the initial state factor

3 is the cell pressure (kPa)

PI is the plasticity index (%).

The calculated values of Fo and M using the developed

equations 4.45 & 4.46 and the ratio of the calculated and

measured shear stress values (2/1) for each of the data analysed are given in Table 4.16(b). As can be seen in table

that the values of the ratio (2/1) demonstrate the validity

105

of the developed equation 4.47. This result suggests the

definition and measurement of shear stress as a function of

initial state factor, cell pressure and plasticity index.

106

CHAPTER FIVE

DESIGN CHART

5.1 INTRODUCTION

Presented in this chapter the development and applications

of new design chart for expansive soils as road subgrade. As

the aim of this research was to investigate the possibility

of characterising the swelling behaviour and strength of

expansive clay soil at different initial states, a new chart

that contains isolines of both swelling and strength will be

developed. Detailed descriptions of characterising swelling

and strength of a given expansive soil by producing isolines

charts of swelling, CBR and shear stress will be given in

this chapter. The developed chart can be of great use in

pavement design, field compaction control and pavement

evaluation of roads constructed on expansive clay soils. The

chart may be useful also in prediction of swelling and

strength properties of a given soil at certain initial state

parameters.

5.2 CHARACTERISATION OF SWELLING AND STRENGTH

The study of swelling behaviour and strength of fine grained

soils requires the development of isolines of swell percent

and swelling pressure to characterise the swelling behaviour

of a given expansive soil at different initial states (water

content and dry density). On the other hand, the study of

the strength variation with soil initial state parameters

will be characterised as well by using isolines of CBR and

shear stress. The linear relationship exists between swell

percent, swelling pressure, CBR or shear stress and initial

state factor Fi (or soaking state factor Fs)(as described in

chapter 4) can be of great use in the characterisation of

the swelling and strength variation with water content and

dry density of a given expansive soil.

5.2.1 Iso-swelling lines

The isolines for the swelling and initial state behaviour of

a given soil can be produced by two different methods; these

107

methods can be classified into a direct method and indirect

method. The isolines of swelling can be produced directly if

a limited number of swell percent tests (three to five) at a

certain surcharge pressure, and swelling pressure tests can

be performed, then the values of Fo and M can be determined.

By using these values and equations 4.9 & 4.17, and by

selecting various values of Fi (Fi1, Fi2, Fi3, …………… , Fin), the lines of swell percent and swelling pressure can be

produced. Then a plot of swelling relationship with initial

state variations can simply be drawn.

The indirect method of producing iso-swelling lines is

simply by substituting the measured basic properties of a

given soil such as plasticity index, clay content and

surcharge pressure, in equations 4.16 & 4.24, and by

selecting various values of Fi, the isolines of swell

percent and swelling pressure can be drawn.

To investigate the validity of this iso-swelling lines

concept, the results obtained in this project and shown in

Figures 4.2 to 4.4 and Figure 4.22 were analysed and plotted

in Figures 5.1 and 5.2. In these plots iso-swelling lines

were obtained for selected values of Fi. For each value of

Fi the corresponding swell percent at different surcharge

pressures (S1, S2, S3) and the swelling pressure (SP) were

calculated using equations 4.16 & 4.24. The calculated

values of swell percent and swelling pressure are indicated

on the lines. The data points shown in Figures 5.1 and 5.2

are the measured initial state (water content and dry

density) of the samples. The measured swell percent and

swelling pressure values of these samples are given

alongside. The comparison of the measured and calculated

swell percent and swelling pressure values in these figures

clearly demonstrates the validity of the iso-swelling lines

concept.

5.2.2 Iso-CBR lines

The isolines for the CBR (soaked and unsoaked) variation

with the soil state factors (initial state factor Fi and

soaking state factor Fs) for a given soil can be produced.

As previously mentioned these isolines can be produced by

108

two different methods either the direct method or indirect

method. The indirect method, which is the simple way of

producing isolines, is more preferred to be used in analysis

of CBR data. By substituting the plasticity index of a given

soil in equations 4.28 & 4.39 and by selecting values of Fi

and calculating the corresponding Fs using equation 5.1,

lines of unsoaked and soaked CBR can be drawn on a graph of

water content as x-axis and dry density as y-axis.

1.5

100*

01.01i

s

sF

S

eGF

Where:

sF is the soaking state factor

iF is the initial state factor

S is the swell percent during soaking (can be

calculated using equation 4.16)(%)

sG is the specific gravity

is the initial water content(%)

e is the initial void ratio.

To prove the validity of this concept, the experimental

results of this research as shown in Figures 4.29 and 4.36

were analysed and then plotted in Figure 5.3. The isolines

drawn in this plot were obtained for selected values of Fi

and calculated Fs. For each value of Fi the corresponding

unsoaked CBR was calculated using equation 4.28. Similarly

for each value of Fs the corresponding soaked CBR was

calculated using equation 4.39. The data points shown in

Figure 5.3 are the measured initial state parameters of the

analysed samples. The measured values of soaked and unsoaked

CBR of the samples are given alongside the points, whereas

the calculated values are indicated on the lines. The

comparison of the measured and calculated CBR values in this

plot demonstrates clearly the validity of the developed iso-

CBR lines concept.

Comparing the developed iso-CBR lines concept with those

suggested by TRL team (1981) in Road Note No. 31 and shown

109

in Figure 2.35. It appears that the two concepts are

conflict with one another. A close examination, however, of

the TRL team concept reveals that the parts of the 40%, 60%

and 70% curves which are upwards are erroneous, because they

suggest that increasing the dry density for the same

moisture content has no influence on the unsoaked CBR, this

is not correct as it contradicts the results obtained in

this project (Section 4.3.2) and those reported in the

theory (Section 2.6.1).

5.2.3 Iso-shear stress lines

The developed linear relationship of shear stress and

initial state factor Fi can be used to investigate the iso-

shear stress lines variation with initial state parameters

of a given soil. These isolines of shear stress can be

produced directly if a limited number of triaxial tests at a

certain cell pressure can be performed and then the values

of Fo and M can be determined. By substituting these values

in equation 4.40, and by selecting different values of Fi,

the shear stress lines can be drawn. Another way of

producing iso-shear stress lines of a given soil simply can

be done by applying the plasticity index and cell pressure

values in equation 4.47, and by selecting various values of

Fi, lines of shear stress can be obtained. Then a chart to

show the relationship of shear stress characterisation lines

with water content and dry density variations can be drawn.

The validity of iso-shear stress lines concept can be

verified by analysing the results of present study shown in

Figures 4.42 to 4.44. The analysis result is illustrated in

Figure 5.4. In this plot iso-shear stress lines were drawn

as described above. The data points shown in the plot are

the measured initial water contents and dry densities of the

analysed samples. When comparing the calculated shear stress

values that indicated on the lines and the measured values

located alongside the data points, their values seem to be

equal. This result verifies the validity of the iso-shear

lines produced.

110

5.3 DEVELOPMENT OF DESIGN CHART

The main objective of this project is to investigate the

swelling behaviour and strength characteristics at various

initial states in order to develop new design chart. Based

on the linear relationship between swell percent and

swelling pressure as well as CBR and shear stress versus the

soil states factors (Fi and Fs). This chart can be developed

as outlined hereunder.

An important feature of the type of characterisation lines

approach developed in the previous section is the

possibility that the data regarding soil strength (i.e. CBR

and shear stress) and swelling behaviour (i.e. swell percent

and swelling pressure) of expansive soils can be

superimposed in one characterisation line. Thus

characterisation lines (isolines) of CBR, shear stress,

swelling pressure and swell percent at different surcharge

pressures (P1, P2, P3, etc.) - Initial state (water content

and dry density) behaviour of a given expansive soil could

be produced. By substituting the basic properties of the

soil, surcharge pressure and cell pressure in equations

(4.16), (4.24), (4.28), (4.39) and (4.47) and selecting

different values of Fi and calculating the corresponding Fs

values using equation 5.1, characterisation lines indicate

swell percent, swelling pressure, CBR (as soaked or

unsoaked) and shear stress values can be produced. Then a

chart of characterisation lines indicating swelling, CBR and

shear stress relationship with soil initial water content

and dry density variations can be plotted.

The experimental results of this project were analysed and

used to demonstrate the validity of the developed design

chart. The result of analysis was plotted in Figure 5.5, in

which lines indicated swelling and strength parameters were

obtained for selected values of Fi. For each value of Fi the

corresponding swell percent, swelling pressure, unsoaked CBR

and shear stress were calculated using equations 4.16, 4.24,

4.28 and 4.47. The calculated values of swelling and

strength parameters are indicated on the lines. The data

points shown in Figure 5.5 are the measured initial water

contents and dry densities of the samples. The measured

values of swell percent, swelling pressure, CBR and shear

111

stress are shown alongside the points. The comparison of the

measured and calculated values in this figure clearly

demonstrates the validity of the design chart produced.

5.4 PRACTICAL APPLICATIONS OF THE DEVELOPED CHART

The design chart developed in the previous section can be of

great use when applied for roads constructed on expansive

soils. The chart can be used in the following aspects:-

Pavement design

Field compaction control

Pavement evaluation.

These three main practical applications of the developed

design chart will be described in some length hereunder.

5.4.1 Pavement design

Pavement thickness is almost universally determined by

subgrade CBR strength. The fundamental difference between

the common practical codes of design lies in the

determination of the subgrade CBR, where the subgrade is

defined as the upper 15 cm layer immediately beneath the

subbase. These design methods consider the moisture content

as the main factor of subgrade strength. In the Kenya Road

Design Manual it is stated that subgrade of expansive soil

should be compacted at the Equilibrium Moisture Content

(EMC) so as to control the moisture content of this soil.

But the EMC is not the only factor that controls strength

and swelling of expansive soils. The soil state factor that

relates moisture content and dry density, is the main

control of the strength and swelling.

The TRL Road Note No. 31 suggests that the strength of the

subgrade to be taken as that of the subgrade soil at a

moisture content equal to the wettest moisture condition

likely to occur in the subgrade after the road is open to

traffic. It recommends three cases of estimating the

subgrade moisture content as described below.

112

The first case where water table exists to the ground

surface, the ultimate subgrade moisture content can be

determined by measuring the moisture content in the subgrade

below existing pavement. The second case when the water

table is not near the ground surface and where the average

annual rainfall is greater than 250 mm a year, the ultimate

moisture content for design purposes can be taken as the

optimum moisture content. The third case in regions where

the climate is arid throughout most of the year (annual

rainfall 250 mm or less), the ultimate moisture content of

the subgrade will be virtually the same as that of the

uncovered soil at the same depth. This is the moisture

content that should be used for design purposes.

Having estimated the ultimate subgrade moisture content it

is then possible to determine the design subgrade CBR at

different densities and thus indicates the value of

achieving the specified density in the subgrade at which the

soil can be compacted without subjected to objectionable

volume change. This value of CBR is then used to determine

the required subgrade thickness from the design charts.

The TRL Road Note No. 31 uses a chart of iso-CBR curves

Figure 2.35 in pavement design. This chart can be used to

get a suitable moisture content and density of the soil that

give a certain design strength but not indicates swelling

behaviour so it can only be used for nonexpansive soils.

Also the iso-CBR curves are not correct as discussed

previously in Section 5.2.2.

In stead of all above defects of the iso-CBR curves

suggested by the TRL (1981), however it seems to be the best

of the conventional methods that used in pavement design.As

discussed in Section (2.8) that all the technical design

methods used for expansive soils only consider the swelling

behaviour of the soil but not consider the strength although

it is the main factor of pavement design. On the other hand,

the conventional methods of flexible pavement design

consider the subgrade strength using the CBR or shear stress

alone and not swelling.

113

For all the above reasons, it seems that there is a great

need for a new approach of design to take care of both the

swelling behaviour and strength characteristics of expansive

soils as road subgrade. The method developed in this

research will be described in the following design steps:-

1) The road to be designed is subdivided into certain

sections according to the variations in the subgrade

soils properties.

2) For each section samples of the subgrade soil should be

taken to a soil laboratory in order to measure the basic

engineering soil properties mainly index tests, clay

contents and specific gravity. A surcharge pressure

equivalent to an anticipated overburden pressure (i.e.

the pressure exerted by the pavement layers above the

subgrade) can be estimated.

3) By using the information obtained in step (2), then

design chart of characterisation lines of swelling and

strength for the subgrade expansive soil can be produced

as described previously in Section (5.3).

4) As specified by the Highway Authorities, the

permissible swell percent of the subgrade soil is in the

range 1% to 2% depending on the class of the road.

5) From the design chart as produced in step (3) the

design strength (using soaked or unsoaked CBR and shear

stress) of the subgrade can be determined as the strength

at water content and dry density (soil initial state)

around which the swell percent is 50% less than the

permissible swell percent.

6) As the thickness of pavement on expansive soils depends

on the subgrade strength and swelling measured at

suitable moisture content and dry density and on the

traffic loading expressed in terms of equivalent standard

axles expected under the road after construction. Thus it

is recommended to continue the pavement design stages

using the suitable conventional pavement design methods

such as the AASHTO and CBR design method. Adopt the

114

procedure described in Section 2.7.1 or 2.7.2 and the

design chart developed by the TRL or AASHTO shown in

Figures 2.34(b) and 2.36 to determine the pavement

thickness and proceed to do the other design steps.

5.4.2 Field compaction control

This method is used for the material quality control of

compacted embankment and/or subgrade. The conventional

pavement compaction control specifications are generally

given in terms of strength determined in the laboratory and

the field control density expressed in terms of relative

compaction. The field compaction control method presently

used in pavement engineering practice is briefly outlines

below:-

1. The compaction characteristics of the soil are studied

according to the standard specifications, compaction

curve is plotted and the maximum dry density and optimum

moisture content are determined.

2. After field compaction of embankment and/or subgrade,

the field density and the moisture content of the

material are measured.

3. The values of the densities and moisture contents in the

above two steps (1&2) are compared. For the densities,

the field density (measured in (2)) should not be less

than 90% sometimes 95% of the maximum dry density

measured in (1). No standard method is used for the limit

of acceptability for moisture content.

This method as described above has some defects and is not

acceptable especially in expansive soils embankment and/or

subgrade for the following reasons: -

The method set a lower limit for the dry density but

not set an upper limit. As previously discussed in

Section (2.5.4) the higher the dry density the higher the

swell percent, in other words a soil compacted to very

high dry density is very sustainable to swell leading to

damages than another compacted to a lower density.

115

The field compaction specification in terms of maximum

dry density alone may be misleading. The moisture content

at which the desired relative compaction to be achieved

must also be specified. As previously discussed in

Sections (2.5.4) and (2.6.1.2) that both CBR and swelling

are much affected by the moisture content.

It appears that a new method of field compaction control

must be developed for expansive soils embankment and/or

subgrade. This method is briefly described below: -

(1) Soil samples are taken from the field to perform some

basic tests such as compaction tests, index tests and

clay contents. From tests results the soil compaction

characteristics (optimum moisture content and max dry

density) can be determined and the design chart is

formed as described in Section (5.3).

(2) After field compaction of embankment and/or subgrade,

the field density and moisture content of the material

are measured.

(3) A point to indicate the measured field density and

moisture content is drawn in the produced chart.

(4) The point indicated in the chart as described by the

above step, if it is located within the range of design

CBR and permissible swelling then the compaction is to

accepted otherwise will be rejected.

5.4.3 Pavement Evaluation

The evaluation of embankment and/or subgrade material is to

check the performance and the strength of the road. The

methods normally used are the Benklman Beam and the Dynamic

Cone Penetrometer (DCP) test.

116

The DCP method of pavement evaluation suffers from the

following limitations: -

The CBR obtained from the correlation or chart is for

the field as compacted only and this may not give a

realistic CBR value.

This DCP test need to be done for many sections of a

highway and this is very expensive and not economical

especially in undeveloped countries.

The charts used to get the layer thickness for a

certain design CBR do not indicate the swelling so it is

unsuitable to be used for expansive soils.

The Benklman method of evaluation is very expensive to be

used in undeveloped countries and does not give the exact

strength components.

For all these reasons, it seems that a new method needed to

be developed as outlined below: -

1) The road is subdivided into different sections

depending on expansive soil properties. By taking soil

samples from each section, the basic soil properties can

be measured and the chart formed in the way described in

Section (5.3).

2) The field density and moisture content of the subgrade

soil along each section are measured. From these data the

field dry density can be calculated.

3) The field measured data of dry density and moisture

content obtained in the above step (2) are used to

determine the swelling (i.e. swell percent and swelling

pressure) and the strength (i.e. soaked or unsoaked CBR)

from the chart.

4) By comparing the determined values of swelling and CBR

with the design values of swelling and CBR values of this

road, then sections with excessive swelling and lower CBR

117

will be excluded out (i.e. sections or subsections in

areas of weak subgrades or subgrade sustainable to

swelling will be located).

118

CHAPTER SIX

CONCLUSIONS

6.1 EXPERIMENTAL WORK

The study of the swelling behaviour and strength of

compacted expansive clay soil has performed on standard

equipment to provide data at different initial state and

testing conditions. All the experiments were performed in a

way to represent the boundary conditions of a road subgrade

of swelling soil. Samples of expansive soil prepared at a

wide range of water content and were compacted to different

dry densities.

In the conventional oedometer cell both the swell percent

and swelling pressure were measured. In order to simulate

the field conditions of the subgrade soil, the swell percent

of compacted soil was measured at three different surcharge

pressures. The swelling pressure was measured as the

pressure required to consolidate a pre-swelled sample until

it retains back to its initial volume. The experimental

results obtained indicated that both the swell percent and

swelling pressure are much influenced by water content and

dry density.

The CBR tests were performed on expansive soil to measure

the soil strength at different testing conditions. The tests

performed on compacted samples for measuring both the soaked

and unsoaked CBR. Some of the samples were tested directly

as remoulded to measure unsoaked CBR while others were

subjected to soaking process prior to the penetration test

for measuring the soaked CBR.

The triaxial tests provided values of the shear stress for

an element of soil tested at a constant cell pressure. The

tests were performed on compacted samples at three different

cell pressures. The results showed that the shear stress

values depend on the sample moisture content and dry density

as well as the cell pressure used in testing.

119

6.2 DEPENDENCE OF THE SWELLING AND STRENGTH ON THE INITIAL

STATE AND SOIL BASIC PROPERTIES

The experimental work has been carried out to study the

swelling behaviour and strength characteristics of expansive

clay soil. Several tests to measure the swell percent,

swelling pressure, CBR and shear stress were performed on

samples having different initial water contents and dry

densities.

A study of the swelling behaviour of expansive soil samples

compacted over a wide range of water contents was conducted.

It was found that for the expansive soil studied, initial

water content, dry density and void ratio greatly influenced

the swell percent and swelling pressure. On the other hand

the strength of expansive soil as indicated by the CBR and

shear stress measurements was studied. The samples were

prepared over a full range of water contents and compacted

to different dry densities. The results of the experiments

confirmed that; initial water content, dry density and void

ratio greatly influenced the unsoaked CBR and shear stress.

It was found that for the soaked (saturated) soil studied;

saturated moisture content, saturated dry density and amount

of swelling during the soaking period greatly influenced the

soaked CBR.

Initial water content, dry density and void ratio were

combined in a way reflecting the influence of each of them

on swell percent, swelling pressure, unsoaked CBR and shear

stress. This combination was termed the initial state Factor

(Fi). On the other hand the combination of the saturated

water content, saturated dry density and amount of swelling

was termed the soaking state Factor (Fs).

Analysis of the experimental results and data reported by

previous researchers demonstrate very clearly that a direct

linear relationship exists between swell percent, swelling

pressure, unsoaked CBR or shear stress and the initial state

Factor (Fi). Similarly the relationship between soaked CBR

and the soaking state Factor (Fs) is linear as well. The

coefficients of this linear relationship (i.e. constant and

slope) were found to depend on plasticity index and clay

contents (i.e. basic properties) of soil, applied surcharge

120

pressure and cell pressure. The influencing factors are

related to each of swell percent, swelling pressure, CBR and

shear stress by simple empirical mathematical expressions.

The results showed that the swelling behaviour is a function

of the initial state factor as well as being dependent on

the plasticity index, clay content and applied surcharge

pressure used in swell percent measurements. The results

suggested that the following equations link swell percent

and swelling pressure to initial state factor and basic soil

properties:

16.4)()()(0474.0)()(01.056.058.043.032.133.132.0

CPIPFCPIPSi

24.4)(719)(5707)(0042.0)(0004.054.11.221.281.2 CPIFCPISP

i

The CBR results proved that the measured soaked and unsoaked

CBR values are dependent on the initial state factor and the

plasticity index of the soil. The results suggested the

following equations to relate unsoaked and soaked CBR to

initial state factor and plasticity index of soil:

28.43022.2)(035.0)(1615282.1

PIPIFPICBRunsoakedi

39.486.1255238.8)(1443.0)(10*4225.811

PIPIFPICBRsoakeds

The dependence of shear stress on the initial state factor,

plasticity index and cell pressure was also investigated and

expressed by the equation:

47.4)(93.6)(703)(94.4)(54.014.0

3

52.129.0

3

04.1 PIFPIi

It is suggested that with equations 4.16, 4.24, 4.28, 4.39

and 4.47 the swelling and strength of the soil can be

predicted for any type of soil at any water content and dry

density.

121

6.3 PRACTICAL IMPLICATIIONS OF THE RESEARCH

The use of expansive clay soils as road subgrade is

beneficial. The costs of removing this type of soil and

importing expensive granular fill are greatly reduced if

naturally occurring expansive soil is used. The aim of this

research was to improve the current specifications used in

pavement design, evaluation and field compaction control of

roads on expansive soils. The method suggested takes care of

both the swelling behaviour and strength characteristics of

road subgrade of expansive soils.

The results of this research suggested that the developed

linear relationship between swelling as well as strength and

initial state factor Fi (as described in chapter 4) can be

used in the characterisation of the swelling and/or strength

variation with initial state parameters of a given expansive

soil. The analysis results verified the possibility of

superimpose all the data regarding soil swelling and

strength in one design chart. This chart can be utilised in

design and evaluation of pavement as well as in compaction

control in the field.

6.4 LIMITATIONS AND FUTURE WORK

The experimental work has been focused on the soil initial

state changes in saturated and unsaturated state and has

provided useful information regarding the swelling behaviour

and strength of compacted expansive clay soils. However, as

discussed in Chapter two, the testing method and measuring

devices used influence the swelling pressure and shear

stress of soil. A useful extension of this work would be to

examine the effects that testing methods and equipment used

have on measuring the swelling pressure and shear stress.

For the measurements of swelling, CBR and shear stress it is

important that the relationships shown in equations 4.16,

4.24, 4.28, 4.39 and 4.47 be tested further by performing a

wider series of swelling and strength testing measurements

on a variety of soils. These tests should be undertaken at

different water contents, dry densities and different

testing conditions.

122

It is recommended for the future researches to apply the

developed design chart in the field and any other road

engineering aspects.

6.5 SUMMARY

It has been demonstrated that the swelling behaviour of

expansive soil compacted over a wide range of water

content different dry densities is depended on: the soil

initial state (water content, dry density and void

ratio), soil expansiveness (plasticity index and clay

content) and applied surcharge pressure.

A study of the strength by the CBR and shear stress

measurements of compacted expansive soil at different

water contents and dry densities was conducted. It was

found that the soil initial state, plasticity index and

testing conditions greatly influenced the unsoaked CBR

and shear stress. The results of the soaked (saturated)

soil samples suggested that the soil soaking state

(saturated moisture content, saturated dry density and

amount of swelling during the soaking period) and soil

plasticity index greatly influenced the soaked CBR

value.

Initial water content, dry density and void ratio were

combined in a way reflecting the influence of each of

them on swell percent, swelling pressure, unsoaked CBR

and shear stress. This combination was termed the

initial state Factor (Fi). On the other hand the

combination of the saturated water content, saturated

dry density and amount of swelling during the soaking

period was termed the soaking state Factor (Fs).

It has been proved that a direct linear relationship

between swell percent, swelling pressure, unsoaked CBR

or shear stress and the initial state Factor (Fi).

Similarly the relationship between soaked CBR and the

soaking state Factor (Fs) was verified to be linear

relationship as well. Based on this relationship and

simple mathematical expressions developed for the

123

swelling and strength, design chart of swelling and

strength variations with initial water content and dry

density were developed. This chart can be applied in

design and evaluation of pavement as well as in

compaction control in the field.

The work demonstrates that it is possible to improve the

current specifications and to include expansive clay

soils as road subgrade provided that the placement water

content and dry density are carefully controlled.

109

REFERENCE

1. American Association of State Highway Officials

(AASHO) (1972): AASHO interim guide for the design of

flexible pavement structures. Washington.

2. Atkinson, J.H., Evans, J.S. and Scott, C.R. (1985):

Developments in micro computer controlled stress path

testing equipment for the measurement of soil parameter.

Ground Engineering, Vol. 18, No. 1, pp. 15-22.

3. Blight, G.E. and De Wet, J.A. (1965): The

acceleration of heave by flooding. A symposium in print.

Moisture Equilibria Moisture Changes in Soils Beneath

Covered Areas. Butterworths, Australia. pp. 89-92.

4. Brackley, I.J.A. (1973): Swell pressure and free

swell in a compacted clay. Proc. 3rd International

Conference on Expansive Soils, Haifa.

5. Brackley, I.J.A (1975): A model of unsaturated clay

structure and its application to swell behaviour. 6th

Regional Conference for Africa on Soil Mechanics and

Foundation Engineering, Durban, South Africa.

6. British Standards Institution (1975): Methods of

testing soils for civil engineering purposes. BS 1377,

London.

7. Casagrande, A. (1948): Classification and

identification of soils. Proceedings of the American

Society of Civil Engineers (A.S.C.E.), Vol. 113, pp.

901-991.

8. Chen, F.H. (1975): Foundation on Expansive Soils.

Developments in Geotechnical Eng. Elsevier Publg. Co.

9. Ellis, C.I. (1980): Soil compaction at low moisture

contents, field trials in Sudan. Proc. 7th Regional Conf.

for Africa on Soil Mech. and Found. Eng., Accra, Ghana.

10. El-Tayeb, M.M. (1991): Characterisation of expansive

soils as road subgrade. M.Sc. thesis, Building and Road

Research Institute, University of Khartoum.

11. Glanville, W.H. (1951): Soil Mechanics for Road

Engineers. Her Majesty’s Stationery Office, London. 12. Grim, R.E. (1962): Applied clay mineralogy. London,

McGraw-Hill.

13. Herzong A. and Brock R. (1964): Some factors

influencing the strength of soil-lime-flashing mixtures.

Proc. 2nd Conf. of Australian Road Research Board,

Melboune, Vol. 2.

110

14. Holtz, W.G. and Gibbs, H.J. (1956): Engineering

properties of expansive clays. Trans. American Society

of Civil Engineers (A.S.C.E.), Vol. 121.

15. Holtz, W.G. (1959): Expansive soils, properties and

problems. Quart. Colorado School of Mines, Vol.54, No.4.

16. Ingles, O.G. and Metcalf, J.B. (1972): Soil

Stabilization, Butterworths, Sydney.

17. Kassif, G., Komornik, A., Wiseman, G. and Zeitlen,

J.G. (1965): Studies and design criteria for structures

on expansive clays. Proc. 1st Res. Conf. on Exp. Soils,

Texas.

18. Kassiff, G., Livneh, M., and Wiseman, G. (1969):

Pavements on Expansive Clays. Jerusalem Academic Press,

Jerusalem, Israel.

19. Lewis, W.A. (1959): Investigation of the performance

of pneumatic-tyred rollers in the compaction of soil.

Road Research Laboratory. Technical Paper No. 45. HMSO.

20. Ministry of Transportation and Communication (MOTC)

(1981): Botswana Road Design Manual.

21. Ministry of Works and Construction (MOWC) (1982):

Kenya Road Design Manual.

22. Mitchell, J.K. (1976): Fundamentals of soil

behaviour. New York. J. Wiley and sons.

23. Murphy, C.P. (1966): Problems of road design and

construction on the clay soils of the Fitzroy Basin.

Australian Road Research Board Proc. Vol. 3, No. 2.

24. Mustafa, A.E. (1982): Cement stabilization of black

cotton soil. M.Sc. Thesis, Building and Road Research

Institute, University of Khartoum.

25. National Institute for Transport and Road Research

(NITRR) (1978): Construction of road embankments. Draft

Technical Recommendations for Highways (TRH9), Pretoria,

CSIR.

26. Nicholas J. Garber and Lester A. Hoel (1988): Traffic

and highway engineering.

27. O’Connor, K. (1994): Swelling behaviour of unsaturated fine grained soils. Ph.D. thesis, City

Univ., London.

28. O’Flaherty, C. A. (1988): Highway engineering. Volume 2, third edition.

29. Porter, O. J. (1942): Foundations for Flexible

Pavements, Highway Research Board Proceedings of the

Twenty second Annual Meeting.

30. Rengmark, F. and Eriksson, R. (1953): Apparatus for

111

investigation of swelling, compression and elastic

properties of soils. Proc. 3rd Int. Conf. of Soil

Mechanics and Foundation Engineering Zurich.

31. Seed, H.B. and Chan, C.K. (1959): Structure and

strength characteristics of compacted clays. Proc.

American Society of Civil Engineers (A.S.C.E.), No. SM5,

pp. 87-127.

32. Seed,H.B., Woodward,R.J. and Lundgren,R.(1962):

Prediction of swelling potential for compacted clays.

Proc. American Society of Civil Engineer. Journal of

Soil Mechanics and Foundation Engineering. Vol. 88.

33. Shell International Petroleum Company Limited (1978):

Shell pavement design manual. London.

34. Skempton, A.W. (1953): The colloidal Activity of

clay. Proc. 3rd Int. Conf. on Soil Mechanics, Vol. 1, pp.

33-62.

35. Skempton, A.W. and Northey, R.D.(1953): The

sensitivity of clays. Geotechnique, Vol. 3, pp. 30-53.

36. Snethen, D.R. (1984): Evaluation of expedient methods

for identification and classification of potentially

expansive soils. Proc. 5th Int. Conf. on Expansive Soils,

Adelaide, South Australia.

37. Sridharan,A., Sreepada, R. and Sivapullaiah,

P.V.(1986): Swelling pressure of clays. Geotechnical

Testing Journal, Vol. 9, No. 1, pp. 24-33.

38. Terzaghi, K. (1936): The shearing resistance of

saturated soil and the angle between the planes of

shear. Proc.1st Int. Conf. Soil Mechanics and Foundation

Eng., Vol. 1, pp. 54-56.

39. Transport and Road Laboratory (TRL)(1981): A guide to

the structural design of bitumen surface roads in

tropical and subtropical countries. Road Note 31,

London.

40. U.S.A. Army Corps of Engineers (1945): The California

Bearing Ratio Test as applied to the Design of Flexible

Pavements for Airports. U.S. Army, Engineer Waterways

Experiment station.

41. Worth, C.P. and Wood, D.M. (1978): The correlation of

index properties with some basic engineering properties

of soils. Canadian Geotechnical Journal, Vol. 15, No. 2,

pp. 137-145.

42. Yevnin A. and Zaslavasky D. (1970): Some factors

affecting compacted clay swelling. Canada Geotechnical

Journal, Vol 7.

112

43. Yoder, E.J. and Witezak, M.W. (1975): Principal of

Pavement Design. 2nd edition. John Wiley and Sons, New

York.

44. Zein, A.K.M. (1985): Swelling characteristics and

microfabric of compacted black cotton soils. Ph.D.

thesis, University of Strathclyde, Glasgow.

45. 中华人民 和国行业标准，公路土工试验规程 JTJ051-

93 ，人民交通出版社，1993年。

46. 王学武等著， 丹膨胀土工程地质特性及基础设计处理，中国石油天然

气华东探矿设计研究院，CNPC，胶洲市，1997年。

47. 李生林等著，中国膨胀土工程地质研究，江 科学技术出版社，1992年

。

48. 肖荣久等著，陕南膨胀土及其灾害地质研究，陕西科学技术出版社，199

4年。

49. [美] J.K.

Mitchell著高国瑞等译 岩土工程土性分析原理 南京工学院出版社，1987年

。

50. 张咸恭，工程地质学， 州大学地质理系，地质出版社，1979年。

51. 李文武，特殊地区公路膨胀土地区，人民交通出版社1992年。

52. 同济大学土力学与基础工程研究室等编,土质学及土力学,人民交通出版

社，1980年。

53. 朱照宏，许志鸿编著，柔性路面设计理论和方法，同济大学出版社，198

5年。

54. 林绣贤，柔性路面结构设计方法，人民交通出版社，1988年。

Colloid Content

(%)

Plasticity Index

(%)

Shrinkage Limit

(%)

Probable Volume

Change

(%)

Degree of

Expansion

(expansiveness)

>28

20 – 13

13 – 23

<15

>35

25 – 40

15 – 30

<18

<10

7 – 10

10 – 15

>15

>30

20 - 30

10 - 30

<10

Very high

High

Medium

Low

Table 2.1 Classification of expansive soils using index tests (after Holtz and Gibbs, 1956).

Method Definition Remark

A

[ASTM D4546]

Pressure required to bring soil back to

the original volume after the soil is

allowed to swell completely without

surcharge (except for a small seating

pressure)

May lead to larger swell pressures

because the method incorporates

hysteresis that tends to overcome

specimen disturbance

B

Pressure applied to the soil so that

neither swell or compression takes

place on inundation; a specimen may be

confined and pressure inferred from

deflection of the confining vessel

A null test in which measured swell

pressures are influenced by apparatus

stiffness; apparatus of higher stiff-

ness leads to less expansion on

swelling; large swell pressures can be

relieved with small specimen expansion;

therefore stiffer can provide improved

control over one-dimensional changes

and can lead to improved measurements

of swell pressure

C

[ASTM D4546]

Pressure necessary to permit no change

in volume upon inundation when

initially under applied pressure equal

to the overburden pressure; various

loads are applied to the soil after

inundation to maintain no volume change

Must be corrected for specimen

disturbance; one dimensional consolido-

meter swell tests are influenced by

lateral skin friction especially in

tests conducted on stiff clays or

shales

D

Pressure required for preventing volume

expansion in soil in contact with

water; various loads are applied to the

soil after inundation to maintain no

volume change

Requires correction of swell pressure

similar to method C above but a

standard correction procedure is not

available

Table 2.2 Definitions of swelling pressure (after Johnson, 1989).

115

Test

Item

Liquid

Limit

(%)

Plastic

Limit

(%)

Specific

Gravity

(g/cm3)

Clay

Content

(%)

Test

Value

58.5

58.0

60.1

59.0

25.9

25.4

26.8

26.3

2.71

2.70

2.73

-

29

33

27

-

Average

Value

58.9

26.1

2.72

29.7

Table 4.1 Index tests, specific gravity and clay

content results of soil tested.

Standard compaction

modified compaction

Water

Content (%)

Dry density

(Mg/m3)

Water

Content (%)

Dry density

(Mg/m3)

12.62

13.61

17.21

18.60

22.81

24.74

27.39

1.363

1.421

1.703

1.720

1.640

1.576

1.449

12.69

14.13

16.72

19.49

21.26

23.25

25.74

1.818

1.868

1.883

1.884

1.858

1.802

1.768

Table 4.2 Compaction tests of soil tested.

116

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Swell

percent

(%)

S1

S2

S3

S4 S5 S6

S7

S8

S9

S10

S11

S12

S13

11.94

11.99

14.97

15.29

15.46

16.76

16.76

17.38

19.76

19.76

23.03

23.56

28.73

1.565

1.732

1.583

1.789

1.574

1.572

1.445

1.734

1.686

1.610

1.609

1.448

1.492

0.738

0.570

0.718

0.520

0.728

0.730

0.882

0.569

0.613

0.689

0.690

0.878

0.823

13.40

21.60

10.10

19.90

11.50

9.39

7.84

15.80

8.70

8.35

7.05

5.95

5.30

Table 4.3 Soil initial state and oedometer test data of

swell percent at 7 kPa surcharge pressure.

117

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Swell

percent

(%)

S14

S15

S16

S17 S18 S19

S20

S21

S22

S23

S24

S25

S26

11.83

13.33

13.54

14.73

16.23

17.38

19.04

19.04

21.37

22.58

26.26

26.41

28.91

1.500

1.590

1.521

1.673

1.817

1.561

1.472

1.648

1.721

1.598

1.580

1.489

1.460

0.813

0.711

0.788

0.626

0.497

0.742

0.848

0.650

0.580

0.702

0.722

0.827

0.863

8.28

7.90

6.40

10.71

13.25

6.60

3.77

5.90

6.43

4.19

3.35

3.20

2.75

Table 4.4 Soil initial state and oedometer test data of

swell percent at 25 kPa surcharge pressure.

118

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Swell

percent

(%)

S27

S28

S29

S30

S31

S32

S33

S34

S35

S36

S37

S38

S39

11.83

13.39

13.55

14.73

15.06

16.23

17.38

18.98

21.37

22.58

26.41

28.73

29.22

1.696

1.487

1.683

1.801

1.599

1.747

1.741

1.633

1.686

1.674

1.578

1.490

1.442

0.604

0.829

0.616

0.510

0.701

0.557

0.562

0.666

0.613

0.625

0.724

0.826

0.886

11.58

5.25

7.85

11.55

5.70

8.64

7.22

3.86

4.30

3.43

2.70

2.20

1.80

Table 4.5 Soil initial state and oedometer test data of

swell percent at 50 kPa surcharge pressure.

119

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Swelling

pressure

(kPa)

SP1

SP2

SP3

SP4 SP5 SP6

SP7

SP8

SP9

SP10

SP11

SP12

SP13

11.94

11.99

14.97

15.29

15.46

16.76

16.76

17.38

19.76

19.76

23.03

23.56

28.73

1.565

1.732

1.583

1.789

1.574

1.572

1.445

1.734

1.686

1.610

1.609

1.448

1.492

0.738

0.570

0.718

0.520

0.728

0.730

0.882

0.569

0.613

0.689

0.690

0.878

0.823

220

300

150

260

115

125

28

200

120

95

112

50

38

Table 4.6 Soil initial state and oedometer test data of

swelling pressure.

120

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Unsoaked

CBR

(%)

UC1

UC2

UC3

UC4 UC5 UC6

UC7

UC8

UC9

UC10

UC11

UC12

7.11

9.78

12.69

14.01

16.72

19.49

21.26

23.58

25.74

28.29

30.99

35.34

1.450

1.519

1.745

1.580

1.816

1.884

1.763

1.669

1.575

1.549

1.495

1.344

0.876

0.791

0.559

0.722

0.498

0.444

0.543

0.630

0.727

0.756

0.819

1.024

50.0

42.0

59.4

32.1

44.6

29.3

25.0

20.5

3.8

1.9

2.2

0.5

Table 4.7 Soil initial state and unsoaked CBR test data.

121

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Soaked

CBR

(%)

SC1

SC2

SC3

SC4 SC5 SC6

SC7

SC8

SC9

SC10

SC11

SC12

5.88

9.48

10.71

14.18

15.08

17.68

17.93

19.78

24.30

25.02

27.08

29.72

1.701

1.695

1.814

1.893

1.862

1.766

1.835

1.804

1.699

1.648

1.592

1.510

3.50

2.10

2.28

2.00

1.95

1.25

1.15

1.11

0.58

0.50

0.45

0.40

0.63

0.61

0.94

1.11

1.03

0.81

0.88

0.90

0.68

0.60

0.52

0.33

Table 4.8 Soil initial state and soaked CBR test data.

122

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Shear

stress

(kPa)

Sh1

Sh2

Sh3

Sh4 Sh5 Sh6

Sh7

Sh8

14.80

15.28

15.52

16.65

17.23

19.77

20.95

22.64

1.586

1.765

1.578

1.594

1.725

1.598

1.571

1.556

0.715

0.541

0.724

0.706

0.577

0.702

0.731

0.748

220

425

196

142

248

131

68

29

Table 4.9 Soil initial state and triaxial test data of

shear stress at cell pressure of 50 kPa.

123

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Shear

stress

(kPa)

Sh9

Sh10

Sh11

Sh12 Sh13 Sh14

Sh15

Sh16

Sh17

Sh18

14.90

15.25

15.40

16.56

17.35

19.70

21.20

23.20

25.42

26.21

1.582

1.784

1.568

1.588

1.732

1.608

1.585

1.571

1.520

1.510

0.719

0.525

0.735

0.713

0.570

0.692

0.716

0.731

0.789

0.801

282

605

274

203

381

143

130

70

32

25

Table 4.10 Soil initial state and triaxial test data of

shear stress at cell pressure of 100 kPa.

124

Test

No.

Initial

water

content

(%)

Initial

dry

density

(Mg/m3)

Initial

void

ratio

Shear

stress

(kPa)

Sh19

Sh20

Sh21

Sh22 Sh23 Sh24

Sh25

Sh26

Sh27

Sh28

14.32

14.80

15.04

16.54

16.67

17.29

19.81

20.55

25.29

25.95

1.668

1.675

1.754

1.687

1.725

1.690

1.741

1.732

1.680

1.655

0.619

0.612

0.539

0.600

0.565

0.598

0.551

0.559

0.607

0.631

560

540

700

453

554

462

392

351

232

155

Table 4.11 Soil initial state and triaxial test data of

shear stress at cell pressure of 200 kPa.

No.

Reference

Type of soil

LL

(%)

PL

(%)

C

(%)

P (kPa)

M

F0

R

1

2

3

4

5

6

7

8

9

10

11

12

13

14

Present study

Present study

Present study

Chen (1975)

Zein (1985)

Yevnin and Zaslavsky(1970)

Yevnin and Zaslavsky(1970)

Yevnin and Zaslavsky(1970)

Yevnin and Zaslavsky(1970)

Kassif et al (1965)

Kassif et al (1965)

Holtz and Bara (1965)

Holtz and Bara (1965)

Barckley (1973)

Expansive soil(China)

Expansive soil(China)

Expansive soil(China)

Expansive clay soil(USA)

Black cotton soil(Sudan)

Expansive clay(Israel)

Expansive clay(Israel)

Expansive clay(Israel)

Expansive clay(Israel)

Clay soil(Israel)

Clay soil(Israel)

Clay soil(USA-California)

Clay soil(USA-California)

Expansive Clay(S. Africa)

59

59

59

44

77

64

64

64

64

85

40

57

45

89

26

26

26

30

43

32

32

32

32

45

18

23

26

35

30

30

30

20

50

61

61

61

61

60

30

38

30

62

7

25

50

48

7

2.5

10

40

160

1

1

7

7

1

0.90

0.63

0.53

0.17

1.15

2.47

1.54

1.23

0.98

3.73

1.56

1.29

0.52

4.50

1.72

2.76

3.73

3.43

0.35

2.53

2.26

3.68

7.11

1.33

0.70

1.63

2.09

0.72

0.972

0.974

0.976

0.925

0.975

0.934

0.968

0.978

0.961

0.901

0.800

0.944

0.932

0.896

Table 4.12(a) Summary of soil properties of the data analysed and the swell percent analysis

results.

No.

Reference

LL

(%)

PL

(%)

PI

(%)

C

(%)

P (kPa)

M2

F02

S2/S1

R

1

2

3

4

5

6

7

8

9

10

11

12

13

14

Present study

Present study

Present study

Chen (1975)

Zein (1985)

Yevnin and Zaslavsky (1970)

Yevnin and Zaslavsky (1970)

Yevnin and Zaslavsky (1970)

Yevnin and Zaslavsky (1970)

Kassif et al (1965)

Kassif et al (1965)

Holtz and Bara (1965)

Holtz and Bara (1965)

Barckley (1973)

59

59

59

44

77

64

64

64

64

85

40

57

45

89

26

26

26

30

43

32

32

32

32

45

18

23

26

35

33

33

33

14

34

32

32

32

32

40

22

34

19

54

30

30

30

20

50

61

61

61

61

60

30

38

30

62

7

25

50

48

7

2.5

10

40

160

1

1

7

7

1

1.04

0.69

0.55

0.25

1.52

2.44

1.57

1.01

0.65

3.58

1.50

1.24

0.75

4.34

1.57

2.71

3.65

2.50

1.82

1.23

2.23

4.04

7.34

0.87

0.60

1.69

1.34

0.96

1.16

1.09

1.03

1.41

1.08

1.14

1.01

0.77

0.70

1.34

0.87

0.93

1.40

0.87

0.970

0.972

0.975

0.908

0.960

0.863

0.966

0.978

0.961

0.800

0.810

0.939

0.910

0.887

Note: M2, Fo2: are the calculated values.

S1, S2 : are measured and calculated swell percent values respectively.

Table 4.12(b) Summary of soil properties of the data analysed and the swell percent analysis

results.

No.

Reference

Type of soil

LL

(%)

PL

(%)

C

(%)

M

F0

R

1

2

3

4

Present study

O’Connor (1994) O’Connor (1994) O’Connor (1994)

Expansive soil (China)

Brickearth clay soil (UK)

Wadhurst clay soil (UK)

London clay soil (UK)

59

38

57

78

26

18

24

28

30

13

40

45

15.65

2.80

29.63

34.14

6.27

22.22

7.43

3.28

0.971

0.934

0.995

0.993

Table 4.13(a) Summary of soil properties of the data analysed and the swelling pressure analysis

results.

No.

Reference

LL

(%)

PL

(%)

PI

(%)

C

(%)

M2

F02

SP2/SP1

R

1

2

3

4

Present study

O’Connor (1994) O’Connor (1994) O’Connor (1994)

59

38

57

78

26

18

24

28

33

20

33

50

30

13

40

45

15.12

3.03

21.98

42.69

7.51

24.42

6.15

3.59

0.85

0.94

0.84

1.16

0.968

0.934

0.973

0.990

Note: SP1, SP2 : are measured and calculated swelling pressure values respectively.

Table 4.13(b) Summary of soil properties of the data analysed and the swelling pressure analysis

results.

No.

Reference

Type of soil

LL

(%)

PL

(%)

C

(%)

M

F0

R

1

2

3

4

5

Present study

El-Tayeb (1991)

El-Tayeb (1991)

Murphy (1966)

Glanville (1951)

Expansive soil(China)

Khartoum black cotton soil(Sudan)

Gezira black cotton soil(Sudan)

Black earth(Australia)

Expansive clay soil(USA)

59

70

65

57

67

26

28

40

35

28

30

61

55

--

--

2.66

1.83

4.55

6.06

2.12

5.13

0.67

2.79

2.43

3.25

0.957

0.978

0.973

0.987

0.982

Table 4.14(a) Summary of soil properties of the data analysed and the unsoaked CBR analysis

results.

No.

Reference

LL

(%)

PL

(%)

PI

(%)

C

(%)

M2

F02

CBR2/CBR1

R

1

2

3

4

5

Present study

El-Tayeb (1991)

El-Tayeb (1991)

Murphy (1966)

Glanville (1951)

59

70

65

57

67

26

28

40

35

28

33

42

25

22

39

30

61

55

--

--

2.78

1.79

4.61

5.82

2.05

5.15

1.50

3.63

1.90

3.35

1.01

0.91

0.90

1.30

0.94

0.911

0.978

0.969

0.897

0.982

Note: CBR1, CBR2 : are measured and calculated unsoaked CBR values respectively.

Table 4.14(b) Summary of soil properties of the data analysed and the unsoaked CBR analysis results.

No.

Reference

Type of soil

LL

(%)

PL

(%)

C

(%)

M

F0

R

1

2

3

4

Present study

El-Tayeb (1991)

Yoder (1975)

Yoder (1975)

Expansive soil (China)

Gezira black cotton soil(Sudan)

Silty clay soil (USA)

Silty clay soil (USA)

59

54

44

50

25

30

20

24

30

60

--

--

0.11

1.29

1.68

1.10

1.75

4.34

4.51

1.81

0.963

0.963

0.900

0.812

Table 4.15 (a) Summary of soil properties of the data analysed and the soaked CBR analysis results.

No.

Reference

LL

(%)

PL

(%)

PI

(%)

C

(%)

M2

F02

CBR2/CBR1

R

1

2

3

4

Present study

El-Tayeb (1991)

Yoder (1975)

Yoder (1975)

59

54

44

50

25

30

20

24

33

24

24

26

30

60

--

--

0.12

1.64

1.64

0.85

1.72

4.41

4.41

1.79

1.10

1.16

0.96

0.75

0.960

0.963

0.875

0.800

Note: CBR1, CBR2 : are measured and calculated soaked CBR values respectively.

Table 4.15(b) Summary of soil properties of the data analysed and the soaked CBR analysis results.

No.

Reference

Type of soil

LL

(%)

PL

(%)

C

(%)

3 (kPa)

M

F0

R

1

2

3

4

5

6

Present study

Present study

Present study

O’Connor (1994) O’Connor (1994) O’Connor (1994)

Expansive soil (China)

Expansive soil (China)

Expansive soil (China)

Brickearth clay soil (UK)

Wadhurst clay soil (UK)

London clay soil (UK)

59

59

59

38

57

78

26

26

26

18

24

28

30

30

30

13

40

45

50

100

200

100

100

100

30.54

38.24

45.68

25.72

38.60

67.45

8.06

7.20

6.64

13.66

7.17

3.33

0.985

0.992

0.990

0.943

0.988

0.970

Table 4.16(a) Summary of soil properties of the data analysed and the shear stress analysis

results.

No.

Reference

LL

(%)

PL

(%)

PI

(%)

C

(%)

3 (kPa)

M2

F02

2/1

R

1

2

3

4

5

6

Present study

Present study

Present study

O’Connor (1994) O’Connor (1994) O’Connor (1994)

59

59

59

38

57

78

26

26

26

18

24

28

33

33

33

20

33

50

30

30

30

13

40

45

50

100

200

100

100

100

35.86

39.28

43.46

30.96

39.28

50.35

7.47

7.09

6.76

11.04

7.09

5.48

1.03

1.25

0.94

1.26

1.03

0.65

0.991

0.978

0.989

0.981

0.988

0.878

Note: 1, 2 : are measured and calculated shear stress values respectively.

Table 4.16(b) Summary of soil properties of the data analysed and the shear stress analysis results.

Standard Compaction

1.300

1.400

1.500

1.600

1.700

1.800

10.00 15.00 20.00 25.00 30.00

Moisture content (%)

Dry

den

sit

y (

Mg

/m3)

Modified Compaction

1.700

1.750

1.800

1.850

1.900

10.00 15.00 20.00 25.00 30.00

Moisture content (%)

Dry

den

sit

y (

Mg

/m3)

Figure 4.1 The compaction curves of the expansive soil

samples tested.

(%)

d

(Mg/m3)

e

iF

S

(%)

11.94

11.99

14.97

15.29

15.46

16.76

16.76

17.38

19.76

19.76

23.03

23.56

28.73

1.565

1.732

1.583

1.789

1.574

1.572

1.445

1.734

1.686

1.610

1.609

1.448

1.492

0.738

0.570

0.718

0.520

0.728

0.730

0.882

0.569

0.613

0.689

0.690

0.878

0.823

17.8

25.3

14.7

22.5

14.0

12.8

9.8

17.5

13.9

11.8

10.1

7.0

6.3

13.40

21.60

10.10

19.90

11.50

9.39

7.84

15.80

8.70

8.35

7.05

5.95

5.30

The swell percent and initial state data

analysed of the studied compacted specimens

tested at 7 kPa surcharge pressure.

(%)

d

(Mg/m3)

e

iF

S

(%)

11.83

13.33

13.54

14.73

16.23

17.38

19.04

19.04

21.37

22.58

26.26

26.41

28.91

1.500

1.590

1.521

1.673

1.817

1.561

1.472

1.648

1.721

1.598

1.580

1.489

1.460

0.813

0.711

0.788

0.626

0.497

0.742

0.848

0.650

0.580

0.702

0.722

0.827

0.863

15.6

16.8

14.3

18.1

22.5

12.1

9.1

13.3

13.9

10.1

8.3

6.8

5.9

8.28

7.90

6.40

10.71

13.25

6.60

3.77

5.90

6.43

4.19

3.35

3.20

2.75

The swell percent and initial state data

analysed of the studied compacted specimens

tested at 25 kPa surcharge pressure.

(%)

d

(Mg/m3)

e

iF

S

(%)

11.83

13.39

13.55

14.73

15.06

16.23

17.38

18.98

21.37

22.58

26.41

28.73

29.22

1.696

1.487

1.683

1.801

1.599

1.747

1.741

1.633

1.686

1.674

1.578

1.490

1.442

0.604

0.829

0.616

0.510

0.701

0.557

0.562

0.666

0.613

0.625

0.724

0.826

0.886

23.7

13.4

20.2

24.0

15.1

19.3

17.8

12.9

12.9

11.9

8.3

6.3

5.6

11.58

5.25

7.85

11.55

5.70

8.64

7.22

3.86

4.30

3.43

2.70

2.20

1.80

The swell percent and initial state data

analysed of the studied compacted specimens

tested at 50 kPa surcharge pressure.

(%)

d

(Lb/in3)

e

iF

S

(%)

5.84

9.95

10.77

12.48

12.92

14.84

17.97

18.59

19.37

106.97

105.93

106.27

105.60

106.47

106.37

105.46

105.73

106.35

0.546

0.561

0.556

0.566

0.553

0.555

0.568

0.564

0.555

53.8

30.4

28.4

24.0

23.9

20.7

16.6

16.2

15.9

7.7

5.6

5.0

4.3

3.5

3.3

2.2

1.4

0.8

The swell percent and initial state data

analysed of dynamic compacted clay samples (as

reported by Chen, 1975).

(%)

d

(Mg/m3)

e

iF

S

(%)

15.60

21.00

21.20

24.30

24.60

28.00

31.30

34.50

41.30

1.640

1.380

1.640

1.630

1.340

1.340

1.380

1.370

1.260

0.707

1.029

0.707

0.718

1.090

1.090

1.029

1.044

1.222

14.9

6.4

10.9

9.3

5.0

4.4

4.3

3.8

2.5

16.9

8.4

11.4

9.7

7.0

5.6

4.0

3.0

1.3

The swell percent and initial state data

analysed of black cotton soil samples (after

Zein, 1985).

(%)

d

(Mg/m3)

f

(Mg/m3)

e

iF

S

(%)

14.30

16.91

20.80

20.48

20.69

20.79

23.93

23.76

24.67

29.01

28.36

32.95

1.549

1.465

1.358

1.460

1.568

1.640

1.370

1.458

1.458

1.371

1.445

1.376

1.183

1.233

1.206

1.275

1.338

1.387

1.239

1.317

1.402

1.293

1.373

1.325

0.795

0.898

1.047

0.904

0.773

0.695

1.029

0.907

0.907

1.028

0.924

1.020

13.6

9.7

6.2

7.9

9.8

11.3

5.6

6.8

6.5

4.6

5.5

4.1

29.8

18.8

12.6

14.5

17.2

18.2

10.6

10.7

4.0

5.8

5.2

3.8

The swell percent and initial state data analysed of

static compacted clay specimens tested at 2.5 kPa

surcharge pressure (source Yevnin and Zaslavsky,

1970).

(%)

d

(Mg/m3)

f

(Mg/m3)

e

iF

S

(%)

14.30

14.45

20.60

20.68

21.00

20.72

23.92

24.09

24.63

28.13

29.33

33.14

1.545

1.453

1.359

1.461

1.567

1.641

1.371

1.458

1.558

1.471

1.362

1.370

1.305

1.289

1.258

1.329

1.402

1.454

1.291

1.364

1.449

1.420

1.316

1.342

0.799

0.913

1.046

0.903

0.774

0.694

1.028

0.907

0.784

0.890

1.041

1.029

13.5

11.0

6.3

7.8

9.6

11.4

5.6

6.7

8.1

5.9

4.5

4.0

18.4

12.7

8.0

9.9

11.8

12.9

6.2

6.9

7.5

3.6

3.5

2.3

The swell percent and initial state data analysed of

static compacted clay specimens tested at 10 kPa

surcharge pressure (source Yevnin and Zaslavsky,

1970).

(%)

d

(Mg/m3)

f

(Mg/m3)

e

iF

S

(%)

13.64

14.78

16.97

20.70

20.72

20.65

20.47

23.41

23.94

24.47

29.23

28.27

1.553

1.520

1.457

1.362

1.453

1.558

1.645

1.369

1.452

1.550

1.352

1.457

1.373

1.360

1.347

1.313

1.379

1.457

1.517

1.332

1.399

1.485

1.338

1.438

0.790

0.829

0.908

1.041

0.913

0.784

0.690

1.031

0.915

0.794

1.056

0.908

14.4

12.4

9.5

6.3

7.7

9.6

11.6

5.7

6.6

8.0

4.4

5.7

13.1

11.8

8.2

3.7

5.4

6.9

8.4

2.8

3.8

4.4

1.1

1.3

The swell percent and initial state data analysed of

static compacted clay specimens tested at 40 kPa

surcharge pressure (source Yevnin and Zaslavsky,

1970).

(%)

d

(Mg/m3)

f

(Mg/m3)

e

iF

S

(%)

14.71

16.95

20.60

20.80

20.67

20.50

23.65

24.11

24.62

28.94

28.36

32.90

1.531

1.456

1.357

1.444

1.555

1.638

1.369

1.451

1.554

1.352

1.452

1.372

1.442

1.434

1.396

1.437

1.516

1.572

1.393

1.455

1.538

1.382

1.469

1.400

0.816

0.909

1.049

0.925

0.788

0.697

1.031

0.916

0.789

1.056

0.915

1.026

12.8

9.4

6.3

7.5

9.5

11.5

5.6

6.6

8.0

4.4

5.6

4.1

6.2

1.5

-2.8

0.5

2.6

4.2

-1.7

-0.3

1.0

-2.2

-1.2

-2.0

The swell percent and initial state data analysed of

static compacted clay specimens tested at 160 kPa

surcharge pressure (source Yevnin and Zaslavsky,

1970).

The data of swell percent and dry density of

expansive clay samples compacted at different

water contents (after Kassif et al, 1965).

The data of swell percent and moisture

content of expansive clay samples compacted

at different dry densities (after Kassif et

al, 1965).

The swell percent, dry density and moisture

content data of remoulded expansive clay

specimens (from Holtz and Bara, 1965).

The swell percent, dry density and moisture

content data of expansive clay specimens

(from Holtz and Bara, 1965).

The data of swell percent and initial

void ratio of expansive clay samples

compacted at different water contents

(after Brackley, 1973).

(%)

d

(Mg/m3)

e

iF

SP

(kPa)

11.94

11.99

14.97

15.29

15.46

16.76

16.76

17.38

19.76

19.76

23.03

23.56

28.73

1.565

1.732

1.583

1.789

1.574

1.572

1.445

1.734

1.686

1.610

1.609

1.448

1.492

0.738

0.570

0.718

0.520

0.728

0.730

0.882

0.569

0.613

0.689

0.690

0.878

0.823

17.8

25.3

14.7

22.5

14.0

12.8

9.8

17.5

13.9

11.8

10.1

7.0

6.3

200

310

130

250

95

105

10

180

100

75

92

30

20

The swelling pressure and initial state data

analysed of the studied compacted specimens.

(%)

d

(Mg/m3)

e

iF

SP

(kPa)

8.52

9.79

9.86

10.14

12.10

12.25

12.82

13.80

14.16

15.26

18.59

19.27

2.017

1.964

1.963

1.917

1.924

1.916

1.978

1.875

1.868

1.887

1.778

1.729

0.349

0.385

0.386

0.419

0.414

0.420

0.375

0.451

0.456

0.441

0.530

0.573

67.9

52.1

51.6

45.1

38.4

37.3

41.1

30.1

28.9

28.0

18.1

15.7

136.7

111.2

84.5

63.7

17.6

23.6

40.3

18.5

16.2

8.3

3.3

2.4

The swelling pressure and initial state data

analysed of compacted Brickearth Clay samples

(after O’Connor, 1994).

(%)

d

(Mg/m3)

e

iF

SP

(kPa)

15.79

20.38

20.38

20.38

20.38

20.38

20.64

22.73

22.86

22.87

23.16

23.28

23.86

23.95

23.95

1.881

1.588

1.549

1.495

1.483

1.397

1.730

1.650

1.734

1.727

1.640

1.655

1.627

1.642

1.621

0.435

0.700

0.743

0.806

0.821

0.933

0.561

0.636

0.557

0.563

0.646

0.631

0.659

0.644

0.666

27.4

11.1

10.2

9.1

8.9

7.3

14.9

11.4

13.6

13.4

11.0

11.3

10.3

10.6

10.2

583.3

113.5

77.2

50.9

41.0

15.6

228.7

121.5

215.0

191.3

100.4

107.8

68.1

88.9

52.5

The swelling pressure and initial state data

analysed of compacted Wadhurst Clay samples

(after O’Connor, 1994).

(%)

d

(Mg/m3)

e

iF

SP

(kPa)

24.03

24.17

24.43

25.19

30.00

30.11

31.05

31.62

31.96

34.17

34.40

1.653

1.600

1.625

1.607

1.447

1.457

1.434

1.452

1.398

1.384

1.362

0.682

0.738

0.711

0.730

0.921

0.908

0.939

0.915

0.989

1.009

1.041

10.1

9.0

9.4

8.7

5.2

5.3

4.9

5.0

4.4

4.0

3.8

226.0

195.2

208.0

200.2

69.5

65.0

58.0

37.5

34.6

35.0

28.0

The swelling pressure and initial state data

analysed of compacted London Clay samples

(after O’Connor, 1994).

(%)

d

(Mg/m3)

e

iF

Unsoaked

CBR

(%)

7.11

9.78

12.69

14.01

16.72

19.49

21.26

23.58

25.74

28.29

30.99

35.34

1.450

1.519

1.745

1.580

1.816

1.884

1.763

1.669

1.575

1.549

1.495

1.344

0.876

0.791

0.559

0.722

0.498

0.444

0.543

0.630

0.727

0.756

0.819

1.024

23.3

19.6

24.6

15.6

21.8

21.8

15.3

11.2

8.4

7.2

5.9

3.7

50.0

42.0

59.4

32.1

44.6

29.3

25.0

20.5

3.8

1.9

2.2

0.5

The unsoaked CBR and initial state data

analysed of the studied compacted specimens.

(%)

d

(Mg/m3)

e

iF

Unsoaked

CBR

(%)

8.70

11.30

11.60

13.60

15.20

15.90

16.20

17.70

19.40

20.50

21.50

21.50

22.00

26.00

30.20

34.50

1.650

1.718

1.361

1.414

1.339

1.355

1.347

1.316

1.742

1.711

1.696

1.322

1.380

1.539

1.394

1.333

0.673

0.607

1.028

0.952

1.061

1.037

1.049

1.097

0.584

0.613

0.627

1.088

1.000

0.793

0.980

1.071

28.2

25.1

11.4

10.9

8.3

8.2

7.9

6.8

15.4

13.6

12.6

5.7

6.3

7.5

4.7

3.6

44.2

47.9

21.1

19.8

11.4

12.1

15.3

11.7

27.9

29.1

23.6

8.3

8.3

10.9

5.9

4.7

The unsoaked CBR and initial state data

analysed of compacted Khartoum black cotton

soil samples (as reported by El-Tayeb, 1991).

(%)

d

(Mg/m3)

e

iF

Unsoaked

CBR

(%)

13.69

20.56

20.10

20.00

23.24

25.43

29.45

29.96

1.654

1.635

1.291

1.293

1.356

1.321

1.499

1.357

0.657

0.676

1.122

1.119

1.021

1.074

0.828

1.019

18.4

11.8

5.7

5.8

5.7

4.8

6.1

4.4

66.5

50.3

17.8

9.5

6.8

8.3

16.9

8.2

The unsoaked CBR and initial state data

analysed of compacted Gezira black cotton soil

samples (as reported by El-Tayeb, 1991).

The unsoaked CBR, dry density and

moisture content data of compacted black

earth samples (after Murphy, 1966).

The unsoaked CBR and compaction curves

of expansive clay samples (after

Glanville, 1951).