GROUND IMPROVEMENT  
&  
STABILIZATION 
 
Editors: 
Nurly Gofar 
Khairul Anuar Kassim 

 
 
FACULTY OF CIVIL ENGINEERING 
UNIVERSITI TEKNOLOGI MALAYSIA 
2008 
 
 CONTENTS 

Contents

Preface

Chapter 1 Compaction Parameters’ Estimation for

Ground Improvement
Fauziah Kasim 1

Chapter 2 Modification & Stabilization of

Cohesive Soils with Lime
Khairul Anuar Kassim 27

Chapter 3 Ground Improvement by Preloading and

Vertical Drain
Nurly Gofar and Rosdi Mohamed 53

Chapter 4 Geotechnical Behaviour of Electronically

Treated Residual Soils
Kamaruddin Ahmad, Khairul Anuar Kassim,
Mohd Raihan Taha 71

Chapter 5 Ground Stabilization by Tree Induced

Suction
Nazri Ali 89
Chapter 6 Geosynthetic Reinforced Soil Structure
Nurly Gofar 106

Chapter 7 Reinforcement Mechanism of Rock Bolt

Mohd For Mohd Amin, Khoo Kai Siang and
Chai Hui Chon 135

Chapter 8 Soil Confinement System for Slope

Rehabilitation
Ramli Nazir 149

Index
 PREFACE 
 
 
 
 
 
 
 
 
 
The ground at a construction site is not always suitable for
supporting structures such as buildings, bridges, highways and
dams. In order to overcome this problem several methods have
been employed worldwide to improve engineering characteristics
of soils. The methods can be categorized as mechanical
stabilization, chemical stabilization, thermal and electrical
stabilization, or inclusion of materials such as geosynthetics into
the soil or inclusion of rock bolt into rock. Surface protection also
plays important role in preserving the soil characteristics against
climate. The role of vegetation and tree on stabilizing soil,
especially slope, has been studied. In the absence of tree,
confinement system can be used to stabilize the topsoil against
erosion.
This book contains eight chapters and each chapter presents the
research done by staff of Geotechnical Engineering Division of the
Faculty of Civil Engineering UTM over the past decade on the
topic of Ground Improvement and Stabilization.

Nurly Gofar
Khairul Anuar Kassim
Faculty of Civil Engineering
Universiti Teknologi Malaysia
2008
CHAPTER 1

COMPACTION PARAMETERS’
ESTIMATION FOR GROUND
IMPROVEMENT

Fauziah Kasim and Wan Rohaila Wan Ab Lah

Universiti Teknologi Malaysia

1.1 Introduction

Compaction of soil is pressing the soil particles tightly together by

expelling air from the void space. In most cases, dry soil can be
best compacted if certain amount of water is added in it. Water will
acts as a lubricant and allows soil particle to be packed together
better (Liu and Evett, 2004). However, if too much water is added,
it will result in a lesser density. Thus, for a given compaction
effort, there is a particular moisture content which gives the
greatest dry unit weight; this pair of values known as optimum
moisture content (OMC) and maximum dry density (MDD),
respectively. The compaction characteristics are first determined in
the laboratory by various compaction tests such as Standard
Proctor Test and Modified Proctor Test. The compaction of the soil
will give three important effects; an increase in the shear strength,
a decrease in future settlement of the soil and a decrease in the
permeability of the soil. The information on these effects is useful
for many types of earth construction.
Several studies have indicated that there is a relationship
between the compaction parameters and Atterberg Limit (Faizah,
2005). The objectives of this project were to determine the
relationships between optimum moisture content to liquid limit,
plastic limit and plasticity index and also to determine the

1
2

relationship of the maximum dry density to liquid limit, plastic

limit and plasticity index. These relationships can be used to
estimate the values of MDD and OMC based on Atterberg Limits,
without having to perform the laboratory compaction test.

1.2 Soil Compaction

Craig (1997) states that compaction is a process by which the soil

particles are artificially rearranged and packed together into a
closer state of contact by mechanical means, in order to decrease
the porosity (or voids ratio) of the soils and thus increases its dry
density.
Compaction is different from consolidation. Consolidation is a
gradual process of volume reduction under sustained loading while
compaction refers to rapid reduction mainly in air voids under a
loading of short duration. An example of compaction is the
reduction in voids produced in a layer of the subgrade by a rubber-
tired or steel-tired roller during construction. The compaction
process may be accomplished by rolling, tamping, or vibration.
Compaction of soil will give three important effects which include
the following:

i. An increase in the shear strength;

ii. A decrease in the settlement; and
iii. A decrease in the permeability of soil.

As a general of rule, the greater the compaction, the greater

these benefits will be. Compaction is actually rather cheap and it is
an effective way to improve the properties of soil.
The degree of compaction is traditionally assessed with
reference to its dry density which is the density that the soil would
have at the same void ratio but zero water content (Powrie, 2004).
The dry density is given by
 3

ρdry = ρ (1.1)
( 1 +ω )
where :
ρdry = dry density
ρ = soil density in the wet state
ω = water content

Figure 1.1 shows the generated compaction curves when the

results of the compaction test are plotted in the form of dry density,
ρ dry against water content , w. The compaction curves show that
as the water content increasing from a low value, the dry density
increases to maximum and then start to decrease. This is because
low water contents imply high suctions, resulting in a soil which is
stiff and not readily compactable. Powrie (2004) states that, at high
water contents, a limit to the degree of saturation, Sr, achievable by
compaction is reached. At Sr ≈ 0.9, any remaining air voids are
surrounded by water, and it is almost impossible for air removal by
compaction.

Figure 1.1: Proctor compaction data ρdry against water contents,

ω (Powrie, 2004).
4

1.3 Factors Affecting Compaction of Soil

There are several factors that affect the compaction of soil. Liu and
Evett (2004) state that the factors that affect compaction of soil are
moisture content, compaction effort and type of soil. It is also
mentioned that, addition of admixtures, and amount and type of
compaction will affect the compaction of soil.

1.3.1 Moisture Content

Moisture content, ω is the ratio of the weight of water , Ww to the

weight of solid, Ws as shown in the following equation:

ω = Ww (1. 2)
Ws

When water is added to the soil during compaction, it will act as a

softening agent for the soil particles. The soil particles slip over
each other and move into a densely packed position. Initially, the
dry unit weight after compaction, increases with increasing
moisture content. After a certain water content, any increase in
moisture content tends to reduce the dry unit weight. This
condition happens because the water takes up the spaces that
would have been occupied by solid particles. The moisture content
at which the maximum dry unit weight (MDD) is reached, is
referred to as the optimum moisture content (OMC).

1.3.2 Compaction Effort

Compaction effort can be measured in terms of the compaction

energy per unit volume. A function of number of blows per layer,
number of layers, weight of the hammer, height of the drop of the
hammer, and volume of the mold, compaction energy per unit is
600 kNm/m3 for the Standard Proctor test and 2700 kNm/m3 for
the Modified Proctor Test (Liu and Evett, 2004).
 5

Obviously, the greater the compaction energy per unit volume,

the greater the compaction will be. Blotz et al. (1998) reports that
a linear relationship exists between the maximum dry unit weight
and the base 10 logarithm of compaction energy based on the tests
conducted on a micaceous silty fine sand. The amount of
compaction energy greatly affects the maximum dry density and
optimum water content of a given soil as shown in Figure 1.2. The
maximum dry density increases, while the optimum water content
decreases with increasing compaction energy.

1.3.3 Types of Soils

Type of soil also affects the compaction of soil. The grain-size

distribution of soil, shape, and specific gravity of solids as well as
the type and amount of clay minerals present, affect maximum dry
unit weight and optimum moisture content for a given compaction
effort and compaction method. Mohd Ruslan (2005) stated that the
soil with high grade would have high density compared to the
lower grade (with respect to categories of subgrade materials).
Maximum dry unit weight may range from 9.42 kN/ m3 for
organic soils to about 22.78 kN/ m3 for well graded granular
material containing just enough fines to fill small voids and
optimum moisture contents may range from around 5% for
granular soil and 35% for elastic silts and clays (Liu and Evett ,
2004). It is also stated that well graded coarse-grained soils attain a
much higher density and lower optimum water content than fine
grained soils which require more water for lubrication because of
the greater specific surface.
Higher optimum moisture contents are linked with lower dry
unit weights. Uniformly graded sand, clays of high plasticity and
organic silts and clays typically respond poorly to compaction.
Moisture content versus density curves for various types of soils
are given in Figure 1.3. These curves were determined by the
Standard Proctor method (ASTM D 698). Both shapes and
6

positions of the curves changes as the texture of the soils varies

from coarse to fine grained soils.

Figure 1.2: Effect of compaction energy on compaction of a sandy clay

(Liu and Evett, 2004).
 7

Figure 1.3: Moisture-density relationships for various types of soil (Liu

and Evett, 2004).
8

1.4 Atterberg Limits

Atterberg limits define four states of consistency for cohesive soils

which are liquid, plastic, semi solid and solid. Consistency refers
to their degree of firmness. The dividing line between liquid and
plastic state is liquid limit; the dividing line between plastic and
semisolid state is the liquid limit and the dividing line between
semisolid and solid is the shrinkage limit as shown in Figure 1.4.

Liquid state

Liquid Limit ( LL)

Plastic state

Plastic Limit ( PL)

Semi solid state

Shrinkage Limit ( SL)

Solid state

Figure 1.4: Atterberg limits (Liu and Evett, 2004).

The moisture content affects the behaviour of a soil. When the

moisture content is too high, the soil will suspend in the water and
cause the soil in liquid state (Craig, 1997). As the soil is slowly
dried out a point is reached when the soil just begins to exhibit
small resistance. When this shear stress is removed, it is found that
the soil has experienced a permanent deformation: it is acting as a
plastic solid and not as a liquid.
 9

The moisture content in the soil when it stops acting as a liquid

and starts acting as a plastic solid is known as liquid limit (LL). As
further moisture is driven from the soil, resistance to large shearing
stressed becomes possible. The soil then exhibits no permanent
deformation and simply fractures with no plastic deformation for
example when it acts as brittle solid. The limit between plastic and
brittle failure is known as the plastic limit (PL). As further
moisture is driven from the soil, the volume of the soil will not
change. The limit when the volume of the soil not change is known
as shrinkage limit (SL).
The plasticity index ( PI) is the range of moisture content in
which a soil behaves a plastic; the finer the soil, the greater its
plasticity index (Craig, 1997).

PI = LL – PL (1.3)

The liquid limit and the plastic limit are very useful in identifying
and classifying the fine-grained cohesive soil. Sieve tests normally
provide quantitative data on the range of sizes of particles and the
amount of clay present, but not the type of clay. Even the clay
particles is very small to be examined visually, Atterberg limits
enable clay soils to be classified physically, and the probable type
of clay mineral to be assessed. Classification is usually
accomplished by means of the plasticity chart. The classification
for cohesive soils is shown in Figure 1.5.

1.5 Correlation Between Compaction Parameters and

Atterberg Limits

Linear regression is a procedure for finding an equation that gives

the average value of one variable as a linear function of the other
values of variables (Maxwell,1983). Regression analysis produces
an estimate of the relationship which exists on the average but does
not assess the strength of the relationship. The regression analysis
does not determine how well a linear pattern fits the data. While
correlation analysis is a procedure for assessing the strength of
10

such relationship, if they exist at all. The coefficient of

determination is used to measure the accuracy and strength of the
linearity between the variables in sample data which is denoted as
R2.

Figure 1.5: Classification system for fine grained soil, based on

plasticity index and liquid limit; modified from BS5930:1981 (Powrie,
2004),

Faizah (2005) states that there are linear relationships between

the compaction parameters; maximum dry density (MDD) and
optimum moisture content(OMC) versus the Atterberg limits;
liquid limit (LL), plastic limit (PL) and plasticity index (PI).The
summary for the results by Faizah (2005) is shown in Table 1.3.
Faizah (2005) concludes that the multiple-parameter relationships
between the optimum moisture content and the dry density with
liquid limit and plastic limit are more significant compared to the
single-parameter relationships. The correlation for the multiple-
parameter relationship is categorized as relatively good (R2 of
 11

0.517 and 0.588 for OMC and MDD, respectively). On the other
hand, the R2 values for the single-parameter relationships for OMC
and MDD were 0.313-0.493 and 0.347-0.559, respectively.

Table 1.1: The linear equations for the estimation of the compaction
parameters based on Atterberg limits (Faizah, 2005).
EQUATIONS
DATA COLLECTED LABORATORY
Y VERSUS Xs
MDD VS. LL MDD = 2.089 – 0.006 LL MDD = -0.006LL + 1.9744
R2 = 0.559

MDD VS PL MDD = 2.129 – 0.013PL MDD = -0.0069PL +

R2 = 0.535 1.9264
MDD VS. PI MDD = 1.935 – 0.008PI MDD = -0.0414PI + 2.262
R2 = 0.347
MDD = 2.132 - 0.004 LL MDD = 1.926 -0.007PL
MDD vs. LL+PL – 0.006 PL
R2 = 0.588
OMC VS. LL OMC = 0.203 LL + 5.383 OMC = 0.2816LL + 2.1873
R2 = 0.493
OMC VS. PL OMC = 0.399 PL + 4.157 OMC = 0.3286PL + 4.4686
R2 = 0.469
OMC VS. PI OMC = 0.263PI + 10.111 OMC = 1.9714PI - 11.5
R2 = 0.313
OMC = 4.065 + 0.125 LL OMC = 0.329PL +4.469
OMC vs. LL+PL + 0.180 PL
R2 = 0.517

Mohd Ruslan (2005) also investigated the relationship between

the compaction parameter with Atterberg limits and he found
similar linear relationship between the compaction parameters and
Atterberg limits, to findings of Faizah (2005). For the multiple-
parameter relationship, the following equations were derived by
Mohd Ruslan (2005):
12

OMC = 0.079 LL + 0.24 PL +3.577 with R2 = 0.458 (1.4)

MDD = -0.003 LL + 0.007 PL +2 .137 with R2 = 0.496 (1.5)

Mohd Ruslan (2005) stated that the non-uniformity and the

quality of the compiled data affected the coefficient of
determination, R2. The value of R2 (less than 0.5) showed that
there was a weak linear relationship between the compaction
parameter and Atterberg limits. Mohd Ruslan (2005) also
compared the results with Al Khafaji (1993) as shown in Figures
1.6 and 1.7.
Al Khafaji (1993) investigated the relationship between the
Atterberg Limits and the soil compaction parameters as measured
by the use of the standard proctor compaction test. The data was
collected from four locations in Iraq, i.e., Balad-Ruz, Kamaliya,
Suwaira, and Abu Ghraib and from the United States’s soil. By
applying the empirical technique that had been used by Maxwell
(1983), all the compaction parameters, i.e., maximum dry density
(MDD), optimum moisture content (OMC) were correlated to the
liquid limit, LL, and the plastic limit, LL. For Iraqi soils, the
following equations were derived:

MDD = 2.44 – 0.22 PL - 0.008 LL (1.6)

and
OMC = 0.4 LL + 0.63 PL – 3.13 (1.7)

For United States (US) soils, the following equations were derived
by Al Khafaji (1993);

MDD = 2.27 – 0.019 PL - 0.003 LL (1.8)

and
OMC = 0.14 LL + 0.54 PL (1.9)
 13

Ruslan
(2005)
Mohd

Figure 1.6: The comparison of OMC and LL correlation (for various

values of PL) between the findings of Mohd Ruslan (2005) and Al
Khafaji (1993).
Mohd Ruslan (2005)

Figure 1.7: The comparison of MDD and LL correlation (for various

values of PL) between the findings of Mohd Ruslan (2005) and Al
Khafaji (1993).
14

Figure 1.8: The distribution of Atterberg limits of Iraq’s soil data of Al

Khafaji (1993) with respect to Casagrande plasticity chart.

Figure 1.8 shows the distribution of the liquid limits and the
plasticity indexes of Iraq’s soil data used by Al Khafaji (1993) for
the correlation regression between the compaction parameters and
Atterberg limits. Al Khafaji (1993) also compared the compaction
parameters of Iraq soils to that of United States of America (US)
soils. The comparison of the values of MDD (ρd ) and OMC (wopt)
is shown in Table 1.2.
 15

Table 1.2: Comparison of MDD (ρd) and OMC (wopt) values for
between Iraqi soils and US soils (Al Khafaji 1993) for various liquid
limit (LL) and plastic limit (PL) values.

Al Khafaji (1993) concluded that the inconsistency of the

results represented the variability of the data for Iraqi soil. He
mentioned that if all Iraqi and US data were combined together, the
inconsistency of the results may be minimized. However, it
depends on the quality of the collected data and ranges of the
Atterberg limits..
Blotz et al. (1998) analyzed twenty two (22) clayey soil samples
in order to estimate MDD (γd max) and OMC (wopt ) for a given
liquid limit. Then, five clayey soils were used to validate the data.
The soil samples were compacted using two to four compaction
efforts including Standard Proctor (ASTM D 698), Modified
Proctor (ASTM D 1557), “Reduced” Proctor and “Supermodified”
Proctor. The results obtained were as the following equations:
16

⎛E⎞
γ dmax,E = γ dmax,k + (2.27LL− 0.94) log⎜⎜ ⎟⎟ (1.10)
⎝ Ek ⎠
and
⎛E⎞
w opt,E = w opt,k + (12.39−12.21LL) log⎜⎜ ⎟⎟ (1.11)
⎝ Ek ⎠
where E is compaction energy (unknown) (kJ/m3) and Ek is
compaction energy (known) (kJ/m3)

If only liquid limit is known, γ d max, E and w opt, E for compaction

effort can be predict by the following equation

γ d max, E = ( 2.27 log LL -0.94) log E – 0.16 LL + 17.02 (1.12)

and
w opt, E = (12.39 – 12.21 log LL) log E – 0.67 LL + 9.21 (1.13)

Figure 1.9 shows the relationship between γ d max, E , w opt, E and

LL with RP, SP, and MP referred as Reduced, Standard, and
Modified Proctor efforts, respectively. Blotz et al. (1998) states
that Equations 1.12 and 1.13, and Figure 1.9 are limited to soils
with similar characteristics, i.e., of liquid emit between 17 and 70.
The present research started with compilation of available data
and laboratory test results. The data for this research was obtained
from related private sectors, i.e., IKRAM Services Sdn. Bhd.
(South and East Branches). The collected data were also obtained
from previous students’ theses; Ibrahim hj Husaini (2002), Freddy
Teo (2002), Redzuan Md Yunus (2003), Khainoriyani Khalid
(2003), Mohd Ruslan Che Ghani (2005), Faizah Kamarudin
(2005), and Hafedz Zakaria ( 2007). The parameters considered for
the collected data were optimum moisture content, maximum dry
density, liquid limit, plastic limit and plasticity index. Then, the
regression equations were developed by using SPSS software.
 17

Figure 1.9: Maximum dry unit weight (γd max) and and optimum water
content (wopt ) versus liquid limit for Reduced, Standard and Modified
Proctor Compaction Efforts (Blotz et al., 1998).

The laboratory tests that had been performed were Standard

Proctor, liquid limit, and plastic limit tests. The laboratory tests are
in accordance to BS 1377:1990. The type of the soils that were
used for the tests were silt and kaolin. The silt sample was taken
from Kolej Tun Dr Ismail, Universiti Teknologi Malaysia, while
Kaolin sample was taken from Kahang, Johor. The parameters
obtained from the laboratory tests were liquid limit, plastic limit,
optimum moisture content, and maximum dry density. Then,
comparison of the compaction parameters between the laboratory
test results and the estimated values from the developed regression
equations was made.
18

1.6 Research Results and Discussion

1.6.1 Distribution of Atterberg Limits of Collected Data and

Laboratory Tests Results

Figure 1.10 shows the distribution of the liquid limit and plasticity
index values from the collected data and the laboratory test results used
in this study. The range of liquid limits for the collected data is from
23.5 % to 90 % while the range of plasticity indexes is from 2.9 %
to 60 %. The chart indicates that the data of the soil ranges from
low plasticity to very high plasticity soils. For laboratory tests’
results, the values of LL were 88.5 % for silt and 56.5 % for
kaolin. While the values of PL from laboratory tests were 53 % for
silt and 29.1 % for kaolin.

80
Low Medium High Very High Extremely
Plasticity Plasticity Plasticity Pasticity High
L I H V Plasticity
70
Collected Data
'B' LINE CE Silt (KTDI)
Kaolin (Kahang)
60
'A' LINE
CV
Plasticity Index (%)

50

CE

40
CH

30 MV

CI

20
CL MH

10 SF

MI
SC ML

0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
Liquid Limit (%)

Figure 1.10: The distribution of the liquid limit and plasticity index
values of the collected data and the laboratory test results for the study.
 19

1.6.2 Regression Correlation between Compaction Parameters and

Atterberg Limits

By using SPSS Version 11.5 software, several regression equations

had been developed to show the relationship between the
compaction parameters and Atterberg limits. The equations that
had been developed include both single-parameter and multiple-
parameter regression. The summary of the regression equations of
the collected data and their corresponding coefficients of
determination is tabulated in Table 1.5. The value of R2 for
multiple-parameter regression is higher than that of single-
parameter regression. This indicates that the relationship for the
multiple parameter regression is stronger compared to relationship
of single parameter regression.

Table 1.3: The regression correlation between the compaction

parameters and Atterberg limits.
Coefficient of
Y versus xs Correlation equation Determination
(r2)
MDD vs LL MDD = -0.006LL + 2.063 0.384
(Eq. 1.14)
MDD vs PL MDD = -0.011PL + 2.084 0.374
(Eq. 1.15)
MDD vs PI MDD = -0.007PI + 1.909 0.218
(Eq. 1.16)
MDD vs LL & PL MDD = -0.003LL- 0.005PL + 2.105 0.415
(Eq. 1.17)
OMC vs LL OMC = 0.175LL + 5.822 0.381
(Eq. 1.18)
OMC vs PL OMC = 0.343PL + 4.922 0.387
(Eq. 1.19)
OMC vs PI OMC = 0.206 PI + 10.706 0.215
(Eq. 1.20)
OMC vs LL & PL OMC = 0.092LL + 0.195PL + 4.330 0.421
(Eq. 1.21)
20

The results of the regression correlation analyses show that the

maximum dry density decreases with increasing liquid limit,
plastic limit and plasticity index values (Eqs 1.14 to 1.17 and
Figures 1.11 and 1.14). While the optimum moisture content
increases with increasing liquid limit, plastic limit and plasticity
index values (Eqs. 1.18 to 1.21 and Figures 1.12, 1.13 and 1.15).

2.400
Collected Data

Laboratory Results
2.200 (Standard Proctor)
Maximum Dry Density (Mg/m ³)

2.000

1.800
MDD = -0.006LL + 2.063
R² = 0.384

1.600

1.400

1.200

1.000
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
Liquid Limit (%)

Figure 1.11: Single-parameter regression for MDD with respect to liquid

limit.
 21

35.0
Collected Data

Laboratory Result
30.0 (Standard Proctor)
Optimum Moisture Content (%)

25.0
OMC = 0.175 LL + 5.822
R² = 0.381
20.0

15.0

10.0

5.0

0.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
Liquid Limit (%)

Figure 1.12: Single-parameter regression for OMC with respect to liquid

limit.

35.0
Collected Data

Laboratory Result
30.0 (Standard Proctor)

OMC = 0.206 PI + 10.706

Optimum Moisture Content (%)

R² = 0.215
25.0

20.0

15.0

10.0

5.0

0.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0
Plasticity Index (%)

Figure 1.13: Single-parameter regression for OMC with respect to

plasticity index.
22

2.2
Laboratory Results (Standard Proctor)

Calculated from Equation 4.4

2 Multiple-parameter Regression
Equation from Faizah (2005)
Maximum Dry Density (Mg/m³)

PL = 100

MDD = -0.003 LL- 0.005 PL + 2.108

PL = 80 R² = 0.415
(56.5, 1.79)
1.8

PL = 60
PL = 100

PL = 40 (56.5, 1.592) (88.5, 1.575) PL = 80

1.6

PL = 20 PL = 60

1.4 PL = 40
(88.5,1.403)

PL = 20

1.2 Equation 4.4 is the same as Eq. 17 of this article

Faizah (2005)

1
0 20 40 60 80 100 120
Liquid Limit (%)

Figure 1.14: The two-parameter regression for MDD with respect to

the liquid limit and the plastic limit in comparison to that of Faizah
(2005) and the results of the laboratory Standard Proctor compaction
tests. (Note: Equation 4.4 is the same as Eq. 1.17 in this article).

Figure 1.14 indicates that there is a negative linear relationship

between the maximum dry density and the liquid limit and the
plastic limit while there is a positive linear relationship between
the optimum moisture content and the liquid limit and the plastic
limit (Figure 1.15). The developed multiple regressions from
Faizah( 2005) were also plotted in the same graph ( Figures 1.14
and 1.15). Both Figures 1.14 and 1.15 show that the slopes of the
graphs of Faizah (2005) are steeper than that of the slope of the
authors’ graphs. It occurred due to the compiled data of higher
variations than that of Faizah (2005) data.
 23

1.6.3 Comparison between Laboratory Test Results and Developed

Regression Equations

The finding shows that the developed regression equations for the
relationship between the maximum dry density and the Atterberg
limits over predicted the laboratory results (Figures 1.11 and 1.14).
While the developed regression equations for the relationship
between the optimum moisture content and the Atterberg limit
under predicted the laboratory results (Figure 1.12, 1.13, and 1.15).

40
Laboratory Results (Standard Proctor)

Calculated from Equation 4.8 Faizah (2005)

35
(88.5, 33.2)
Multiple Regression Equation from PL = 100
Faizah(2005)
30
PL = 80
Optimum Moisture Content (%)

PL = 60
25
PL = 100 (88.5, 22.8)

PL =40
20 (56.5, 19.1)
PL = 80

PL = 20
PL = 60 (56.5, 15.2)
15

PL = 40
10 OMC = 0.092LL + 0.195 PL + 2.015
PL = 20 R² = 0.421

5
Equation  4.8  is  the  same  as  Eq.1.20  of  this 
0
0 20 40 60 80 100 120
Liquid Limit (%)

Figure 1.15: The two-parameter regression for OMC with respect to the
liquid limit and the plastic limit in comparison to that of Faizah (2005)
and the results of the laboratory Standard Proctor compaction tests.
(Note: Equation 4.8 is the same as Eq. 1.21 in this article).

1.6.3 Comparison between Laboratory Test Results and Developed

Regression Equations

The finding shows that the developed regression equations for the
relationship between the maximum dry density and the Atterberg
24

limits over predict the laboratory results (Figures 1.11 and 1.14).
While the developed regression equations for the relationship
between the optimum moisture content and the Atterberg limit
under predict the laboratory results (Figure 1.12, 1.13 and 1.15).
The results show that the two-parameter regression equations of
MDD is more accurate than that of OMC. The difference of MDD
values for silt and kaolin between the laboratory test results and the
calculated values from Eq. 1.17 are 10.9 % and 11.2 %,
respectively. These values are lower than the difference of OMC
values for silt and kaolin between the laboratory test results and the
calculated values from Eq. 1.21, i.e., 45.6 % and 25.6 %,
respectively. The authors believe that the higher difference
between values of OMC for laboratory results and calculated
values from equation occurred because of the collected data used
to develop the equation does not include the values of OMC that
higher than 30 %. Most of the OMC values for the collected data
are only between 10 % to 20 %, while several data is between 20
% and 30 %.

1.7 Conclusions

The conclusions that can be drawn from this study are as follows:

1. Maximum dry density decrease with increasing Atterberg

limits ( liquid limit (LL), plastic limit (PL) and plasticity
index (PI)).

2. Optimum moisture content increases with increasing

Atterberg limits ( liquid limit (LL), plastic limit (PL) and
plasticity index (PI)).

3. The two-parameter regression equation (variables of LL

and PL) gave higher value of R2 compared to that of the
single-parameter regression equation for the estimation of
 25

the soil compaction parameters, MDD (maximum dry

density) and OMC (optimum moisture content). This
indicates that the two-parameter regression correlation
equation is better than that of single parameter relationship
for the estimation of the compaction parameters.

References

Al-Khafaji, A.N. (1993). Estimation of Soil Compaction

Parameters by Means of Atterberg Limits, Quarterly Journal of
Engineering Geologist, 26:359-368.
Blotz, L.R., Benson C.H., and Boutwell G.P. (1998). Estimating
Optimum Water Content and Maximum Dry Unit Weight for
Compacted Clay, Journal of Geotechnical and
Geoenvironmental Engineering. :907-912.
Craig, R. F. (1997). Soil Mechanics, 6th ed. Chapman&Hall,
United Kingdom.
Faizah Kamarudin (2005). Estimation of Compaction Parameters
Based on Atterberg Limits, M.Eng Thesis, Universiti Teknologi
Malaysia, Skudai, Johor, Malaysia.
Freddy Teo (2002). Hubungan antara Kandungan Lembapan
Optimum dan Ketumpatan Kering Maksimum bagi Ujian
Pemadatan dengan Had Atterberg, B.Eng Thesis, Universiti
Teknologi Malaysia, Skudai, Johor, Malaysia.
Hafedz Zakaria (2007). Korelasi antara Ujian Mampatan dengan
Had Atterberg, B.Eng Thesis, Universiti Teknologi Malaysia,
Skudai, Johor, Malaysia.
Ibrahim hj. Husaini (2002). Kajian Trend Darjah Ketepuan pada
Ketumpatan kering Maksimum Bagi Ujian pemadatan tanah, B
Eng Thesis, Universiti Teknologi Malaysia, Skudai, Johor,
Malaysia.
Khainoriyani Khalid (2003). Korelasi antara Data Ujian
Pemadatan dan Had Atterberg, B.Eng Thesis, Universiti
Teknologi Malaysia, Skudai, Johor, Malaysia.
26

Liu, C. and Evett, J. B. (2004). Soils and Foundations, SI Edition,

Prentice Hall, Singapore.
Maxwell, E. A. (1983). Introduction to Statistical Thinking,
Englewood Cliffs, New Jersey, USA.
Mohd Ruslan Che Ghani (2005), Korelasi Regresi antara
Parameter Pemadatan Tanah dan Had Atterberg, B.Eng
Thesis, Fakulti Kejuruteraan Awam, Universiti Teknologi
Malaysia, Skudai, Johor, Malaysia.
Powrie, W. (2004), Soil Mechanics Concepts and Application,
Second Edition, Spon Press, London, United Kingdom.
Redzuan Md Yunus (2003), Korelasi antara Hasil Ujian
Pemadatan dan Had Atterberg, B.Eng Thesis, Universiti
Teknologi Malaysia,
CHAPTER 2

SOIL MODIFICATION AND

STABILISATION OF MALAYSIAN
COHESIVE SOILS WITH LIME

Khairul Anuar Kassim

Universiti Teknologi Malaysia

2.1 Introduction

Large urban centres in the western part of Peninsular Malaysia are

located in limestone regions. Limestone formations are widespread
in Pulau Langkawi, Kinta Valley, Gua Musang and Kuala Lumpur
area as illustrated by Neoh (1998) in Figure 2.1. Malaysia is rich
with limestone resources and by this local advantage, the using of
lime stabilization method in construction present a good prospect
for acceptance as a cost-effective soil stabilization On the other
hand, there are extensive deposits of very soft normally
consolidated cohesive clay in both East and West Malaysia. In
these areas, the depth of soft clay may exceed 40 m. The water
content of the soft clay is generally high; typically about 60 to
80%. The undrained shear strength is generally low from between
7 to 12kPa (Pitts, 1984). The undesirable properties include low
strength, high plasticity, poor workability, difficult compaction and
high swell potential. One of the techniques to enhance the usage
of this soil is to stabilize it insitu using active additive such as lime.

27
28

Figure 2.1 The Geology of Peninsular Malaysia. (after Neoh, 1998)

 29

Many of the researchers on lime stabilization only consider the

physical characteristic of the soil to establish the suitability of lime
stabilization. It is therefore important to investigate further on the
mineralogy aspect since the primary component of soil-lime
reaction involved the development of new products. Failing to
understand the mechanism of soil-lime reaction will leads into less
effectiveness in lime stabilization method.

2.2 Materials
Soil samples from various locations in Peninsular Malaysia were
collected for the study. The soil samples were chosen to represent
plasticity index ranging from 20 to 50 and clay fraction from 10 to
50 percent. Hydrated lime was used as an active additive. The
lime was kept in an air tight container to preserve its quality.

2.2.1 Soil
Soil samples from five locations in Peninsular of Malaysia were
tested. They are Kulai clay, Pelepas marine clay, Sg. Buluh clay,
Tapah Kaolin and Jerangau clay. The five locations are as shown
in Figure 2.1 and the properties of the soils are given in Table 2.1.

2.2.2 Water

Distilled water was used for mixing soil and lime in all
classification and strength tests.

2.2.3 Lime
Lime is an effective stabilizing agent for plastic soils and can be
used to improve their workability, reduce settlement and increase
strength. Two types of lime commonly used in stabilization are
hydrated lime [Ca(OH)2] and quicklime [CaO] (Ingles, 1970).
These limes will modify the soil through cation exchange and
stabilised it after the modification process has completed.
Hydrated lime from Lime Treat, Johor was used for the preparation
 30

of all test specimens. The chemical content of hydrated lime used

in this project is given in Table 2.2.

Table 2.1: Properties of cohesive soils

Pelepas Sg.
Kulai Tapah Jerangau
marine Buluh
Clay Kaolin clay
clay clay
1. Physical properties
Natural moisture
33 121 34 48 39
content (%)
Particle density 2.73 2.74 2.75 2.64 2.69
Liquid limit (%) 52 56 71 93 70
Plastic limit (%) 28 24 31 43 25
Plasticity index (%) 24 32 40 50 45
UCS (kPa) 38.1 24.5 35.5 24.4 150.1
2. Particle size distribution
Sand 43.2 34.0 0.5 6.0 0.4
Silt 45.4 27.7 68.2 57.6 55.4
Clay (CF) 11.4 38.3 31.3 36.4 44.2
Clay activity (Ac) 2.11 0.84 1.28 1.37 1.02
3. Soil classification
BSCS MH CH CV CE CV
USCS CH CH CH MH CH
ASCS A-7-6 A-7-6 A-7-5 A-7-5 A-7-6
4. Chemical properties
Organic content (%) 0.30 0.75 0.42 0.24 0.33
pH at 20% solids 4.3 8.4 3.5 4.5 4.9

Table 2.2: Chemical content of hydrated lime used

Chemical content Quantity
Calcium Hydroxide
[Ca(OH)2] 90 % min
Magnesium Oxide [MgO] 2 % max
Calcium Carbonate
[CaCO3] 6 % max
Arsenic [As] 10 p.p.m. max
Lead [Pb] 5 p.p.m. max
Source of result: from the manufacturer
 31

2.3 Laboratory Tests and Procedures

Laboratory tests carried out can be categorized into classification,
strength and mineralogical investigation. The classification tests
include the determination of optimum lime content and suitability
of lime and soil for stabilization. The strength test reflects the
effect of age of stabilised soils on the strength increase and the rate
of strength increase. The mechanism for the strength increase was
further investigated through mineralogical studies using XRD and
SEM.

2.3.1 Lime fixation capacity (LFC)

When lime is added to clay soil, it must first satisfy the
consumption of the soil for lime to modify, before it can proceed to
stabilization process. The minimum amount of lime to modify is
referred to lime fixation capacity (Hilt & Davidson, 1960). The
lime fixation point corresponds with the point where further
addition of lime does not bring about further changes in the plastic
limit.

2.3.2 Suitability of lime

This test determines the suitability of lime as stated in BS 1924:
Part 2 : 1990 Clause 5.4.6. The alkalinity of lime use must be in
the range of pH 12.35 to 12.40 tested at 25oC.

2.3.3 Initial consumption of lime (ICL)

The initial consumption of lime (ICL) test gives an indication of
the minimum amount of lime that is required to initially stabilized
the soil. Lime is added to the soil until the pH value reaches 12.40
at 25oC. The test procedure is explained in BS 1924 : Part 2 : 1990
: Clause 5.4.
32

2.3.4 Available lime content (ALC)

The available lime content (ALC) of a calcareous material is the
amount of calcium present in the hydrated lime or quicklime,
which is immediately available for reaction with the soil. It is
determined by using the sucrose extraction method as detailed in
BS 6463 : Part 2 : 1984.

2.3.5 Unconfined compression test

Unconfined compression test (UCT) specimens were prepared by
compacting the material into cylindrical moulds as specified in BS
1924 : Part 2 : 1990, Clause 4.14. Dry mixing method was used in
mixing lime and soil at various lime content before water at
optimum moisture content (OMC) is added. Stabilized soils
samples were cured from 7 to 56 days before being tested.

2.4 Results and Discusion

2.4.1 Lime fixation capacity (LFC)

The lime fixation point corresponds with the point where further
addition of lime does not bring about further changes in the plastic
limit. Very small quantities of lime normally between 1 to 3% by
dry mass are added, to bring about these changes in plasticity. The
changes in plastic limit of treated materials at different percentage
of lime are shown in Figure 2.2. It indicated that Kulai clay has
the lowest lime fixation capacity of 0.5%. The lime fixation
capacity for Pelepas marine clay, Tapah kaolin and Jerangau clay
is about 1.5% whereas for Sg. Buluh clay is 1%. The different in
lime fixation capacity is dependent on the clay content and its
capacity for cation exchange.
 33

2.4.2 Suitability of lime test

BS 1924: Part 2: 1990, Clause 5.4.6 suggested the average pH of

hydrated lime used should be within the acceptable limits of pH
value from 12.35 to 12.4 at the corrected temperature of 25oC. The
quality of the lime used in this study complies with the
requirement in BS 1924 as stated in Table 2.3.

5 5 .0

5 0 .0

4 5 .0
Plastic Limit (%)

4 0 .0

3 5 .0

3 0 .0 K u la i c la y
P e le p a s m a rin e c la y
S g . B u lu h c la y
2 5 .0
T a p a h K a o lin
J e ra n g a u c la y
2 0 .0
0 1 2 3 4
% o f H y r a te d L im e

Figure 2.2 Lime fixation capacity of various types of materials with

different additions of lime.

Table 2.3. Suitability of lime test result based on pH value.

pH test T oC pHT pH25

1 26.8 12.32 12.37
2 27.2 12.33 12.40
3 27.3 12.31 12.38
Average 12.38
34

2.4.3 Available lime content (ALC)

The results of ALC test are given in Table 2.4. The average ALC
in terms of equivalent CaO is 65.8%, which is above the minimum
requirement of 60%. The average Ca(OH)2 content is 86.8%,
which is also above the minimum requirement of 80% for hydrated
lime.

Table 2.4 Available lime content as CaO and Ca(OH)2 test results.

Volume of %
Test
hydrochloric acid % of CaO Ca(OH
No
(1M) used )2
1 23.7 66.5 87.8
2 21.8 61.1 80.8
3 24.1 67.6 89.3
4 24.3 68.1 90.0
5 24.9 69.8 92.3
6 23.4 65.6 86.7
7 24.2 67.9 89.7
8 22.5 63.1 83.4
9 23.2 65.0 85.9
10 22.1 62.0 81.9
Average 65.8 86.8

2.4.4 Initial consumption of lime (ICL)

Table 2.5 indicates the summary of ICL test results on various

types of soils at corrected temperature of 25o C. Based on the LFC
method and ICL test results, some general correlation can be
derived to predict the LFC and ICL values from the clay fraction
(CF) and the activity of clay (Ac) as presented in Figures 2.3 and
2.4.
Figure 2.3 presents a reasonable correlation between LFC and
CF with the R2 value of 0.91 as represented by the following
equation:
 35

LFC (%) = (CF/30) + 0.1 (2.1)

where CF is the percentage of clay (particle size < 2 μm)

Table 2.5 ICL test results on various types of soils at 20% solids at 25oC.

Soil Percentage of Lime Content

Description 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Kulai clay 4.3 11.7 12.3 12.4 12.4 12.4 12.4 12.4
Pelepas
marine clay 8.4 10.9 11.8 12.2 12.3 12.3 12.4 12.4
Sg. Buluh
clay 3.5 12.1 12.1 12.3 12.3 12.4 12.4 12.4
Tapah
Kaolin 4.5 10.6 11.4 12.3 12.4 12.4 12.4 12.4
Jerangau
clay 4.9 10.5 11.8 12.2 12.3 12.4 12.4 12.4

Figure 2.3 LFC versus clay fraction for various types of soils.
36

ICL has a good correlation with the clay activity as indicated in

Figure 2.5 with the R2 value of 0.94. A linear correlation between
the ICL values and the activity of clay as in equation (2) can be
used to predict the minimum amount of lime required for initial
stabilisation process.
ICL (%) = - 1.13Ac + 3.83 (2.2)

where Ac is the activity of clay (PI/CF)

Figure 2.4 Correlation of ICL with Ac for various types of soils.

2.4.5 Compaction characteristics

The addition of lime to clay increases their optimum moisture

content and reduces their maximum dry density for the same
compactive effort as shown in Figure 2.5.
The optimum moisture content tends to move towards higher
values, enabling soils in wetter than original condition to be
compacted satisfactorily (Bell, 1996).
 37

  1.6
0% Lime Addition

3% Lime Addition

6% Lime Addition
1.5
9% Lime Addition
3)

12% Lime Addition

Dry Density (Mg/m

1.4

1.3

1.2
0 10 20 30 40 50
Moisture Content (%)

Figure 2.5 Dry density versus moisture content relationship curves for
Sg. Buluh clay containing 0, 3, 6, 9 and 12% lime.

2.4.6 Determination of optimum lime content

In order to establish the mix design for lime stabilization, one has
to consider the two processes involved namely modification and
stabilization. Minimum lime content for modification was
established using lime fixation capacity (LFC). For stabilization
purposes, the minimum lime content to initiate the process can be
obtained using ICL test. More lime is required for full stabilization
and the lime required can be established from the strength test.
UCS tests were performed on lime stabilized soils at various age
and lime content.
38

Bell (1996) explained that the optimum lime content represents

the point of lime saturation and may move, according to clay
mineral content, towards higher increments of UCS with
increasing length of curing time. As can be seen from Figure 2.6,
the UCS of lime-treated soils develop rapidly with increasing
addition of lime until an optimum lime content is reached. Beyond
this, the strength continues to increase at the reduced rate or begins
to decline.
  P elepas m arine clay
200
Sg. Buluh clay
180
T apah Kaolin
Unconf ine d Compressiv e St rengt h ( kPa)

160 Jerangau cla y

140 Kulai clay
120

100

80

60

40

20

0
0 3 6 9 12
Lim e C on te n t ( %)

Figure 2.6 Unconfined compressive strength at 14 days with various

additions of lime.

Table 2.6 Mix design with the optimum lime content at optimum
moisture content by % of dry weight.
Hydrated Optimum Moisture
Soil Description Lime (%) Content (%)
Kulai clay 6.0 23.3
Pelepas marine clay 6.0 28.0
Sg. Buluh clay 6.0 28.0
Tapah Kaolin 6.0 39.0
Jerangau clay 3.0 18.2
 39

All the clays in the study exhibit a rapid initial increase in UCS
with the additions of lime after curing for 14 days except for Kulai
clay (Figure 2.7). Even though Kulai clay does not show
significant change in strength with increasing lime content, an
increase in strength of more than 100% was achieved with
increasing age (Figure 2.7). Jerangau clay and Tapah kaolin show
the most rapid increase in strength with age. Sungai Buluh clay
however shows the slowest rate in strength increase. The reasons
for slow and fast rate in strength increase will be discussed later in
the strength development. In general, the optimum lime content
for the clays were in the range of 3% to 6% (Table 2.6). It is
suggested here that excessive addition of lime more than the
optimum will produce a detrimental effect on the strength
development of stabilised soil due to the unreacted lime. The
unreacted lime only acts as a filler to the soil.

2.4.7 Unconfined compression test (UCT)

Mitchell and Hooper (1961) reported that the effect of lime on soil
strength is a function of time, temperature and relative humidity.
Figure 2.7 shows the results of UCT with different curing time
from 7 days to 56 days at room temperature of 27±2oC.
The strength of Kulai clay increases gradually at a more or less
constant rate, during the first 7 to 28 days of curing but after
28days, the increase is more rapid with more than 200% was
achieved. The Pelepas marine clay shows the highest early increase
in compressive strength of 3.6 times that of the untreated soil. Sg.
Buluh clay however shows a less significant increase in strength
with age even after 28 days of curing. It could be due to the
solubility of the clay minerals at that stage, which does not
encourage the reaction between lime and clay. More lime may be
required to maintain the alkaline environment so as to promote
pozzolanic reaction.
The strength of Tapah Kaolin increases rapidly with age at a
constant rate. With an optimum lime content of 6%, Tapah kaolin
40

has developed in strength to more than eleven (11) times higher

than that of the untreated soil at 56 days (Figure 2.7).
The untreated Jerangau clay, exhibits an average compressive
strength of about 150 kPa. With addition of 3% lime, the
stabilized Jerangau clay increases slowly at the first 7 to 14 days of
curing but the increase become more rapidly to more than 200% at
56 days.

Figure 2.7 Unconfined compressive strength at mix design with various

curing days in room temperature.
 41

2.4.8 Development of strength

A study on strength development was carried out using the X-ray

diffraction (XRD) and scanning electron micrographs (SEM).
New compounds formed due to pozzolanic reaction were identified
using the XRD and SEM techniques. While the effect of lime
content on the strength gain of the stabilized soil was established
through UCT at various curing ages.
 
2.4.9  Strength development and mineralogy  

The XRD results of clay materials are summarized in Table 2.7.

XRD investigation was conducted at 14 curing days to evaluate the
minerals transformation at optimum lime content. Both untreated
and lime treated soil minerals are listed in the descending order of
relative intensity in Table 2.7. In general, minerals composition
from the XRD results can be categorized as clay minerals, non-
clay minerals, additive and new reaction products.
Mineralogy plays an important role in establishing the character
of a soil. The clay minerals that are found in the soil samples are
kaolinite, chlorite and illite whereas the non-clay minerals are
quartz, gibbsite, goethite, orthoclase, aragonite and calcite.
Portlandite or lime was added to the soil to form soil-lime
mixtures known as calcium aluminate silicate hydrate (CASH) that
existed in the alkaline environment. Alkaline environment is
essential in ensuring the pozzolanic reaction to take places as the
silica and alumina became soluble. This partially crystallized,
CASH slowly converted into a well crystalline phase to form
calcium silicate hydrate (CSH) and calcium aluminium hydrate
(CAH). The crystallized compound of CSH and CAH hardened
with age to form a permanent compound binding the soil particles
thus increasing the shear strength of the stabilized soils.
42

Table 2.7. List of minerals in clay samples with intensity in descending

order obtained from XRD of untreated and 14 days lime treated soils.

M ineral
Soil
Untreated Lim e Com position
Description
Soil Treated Soil
Quartz SiO 2
Quartz Orthoclase KAlSi 3 O 8
Orthoclase K aolinite  Al 4 Si 4 O 10 (OH) 8
K aolinite  Portlandite ™ Ca(OH) 2
Kulai clay Gibbsite CASH * Ca Al 2 Si 3 O 10 3H 2 O
CASH Gibbsite Al 2 (OH) 6
Calcite Calcite ™ CaCO 3
CSH CSH * 3CaO 2SiO 2 3H 2 O
CAH * 3CaO Al 2 O 3 Ca(OH) 2 12H 2 O
Quartz SiO 2
Chlorite  (M g, Fe″) 10 Al 2 (Si, Al) 8 O 20 (OH, F) 16
Quartz Aragonite CaCO 3
Aragonite Goethite FeO OH
Pelepas Chlorite  Gibbsite Al 2 (OH) 6
marine clay K aolinite  K aolinite  Al 4 Si 4 O 10 (OH) 8
Gibbsite Portlandite ™ Ca(OH) 2
Goethite CAH * 3CaO Al 2 O 3 Ca(OH) 2 12H 2 O
Calcite ™ CaCO 3
CSH * 3CaO 2SiO 2 3H 2 O
Quartz SiO 2
K aolinite  Al 4 Si 4 O 10 (OH) 8
Quartz
Illite  K 2 Al 4 Si 6 Al 2 O 20 (OH) 4
K aolinite 
Sg. Buluh Portlandite ™ Ca(OH) 2
Gibbsite
clay Goethite FeO OH
Illite  Al 2 (OH) 6
Gibbsite
Goethite CaCO 3
Calcite ™
CSH * 3CaO 2SiO 2 3H 2 O
K aolinite  Al 4 Si 4 O 10 (OH) 8
CSH * 3CaO 2SiO 2 3H 2 O
K aolinite  Quartz SiO 2
Tapah
Gibbsite Gibbsite Al 2 (OH) 6
Kaolin
Quartz Calcite ™ CaCO 3
CASH * CaAl 2 Si 3 O 10 3H 2 O
Portlandite ™ Ca(OH) 2
Quartz SiO 2
Portlandite ™ Ca(OH) 2
Orthoclase KAlSi 3 O 8
Quartz
Goethite FeO OH
Orthoclase
K aolinite  Al 4 Si 4 O 10 (OH) 8
Jerangau K aolinite 
CASH * Ca Al 2 Si 3 O 10 3H 2 O
clay Gibbsite
Gibbsite Al 2 (OH) 6
Illite 
Illite  K 2 Al 4 Si 6 Al 2 O 20 ( OH) 4
Goethite CaCO 3
Calcite ™
CSH * 3CaO 2SiO 2 3H 2 O
CAH * 3CaO Al 2 O 3 Ca(OH) 2 12H 2 O

Note: : Clay Minerals; ™: Additive (Lime); *: Cementitious products; and without any mark:
Non-clay Minerals.
 43

The stabilized soils were cured for 14 days at 27±2°C with 3, 6,

9 and 12% of lime based on the dry mass of the soil. The
percentage of increase in strength as compared to the untreated
strength, Cst can be obtained from the following equation.

S st - S ut
Cst = ×100% (2.3)
S ut

where Sst is the compressive strength of stabilized soil and Sut is

the compressive strength of untreated soil.

From Figure 2.7, lime treated Kulai clay shows an initial slow
reactions for the first 28 days. Nevertheless, lime stabilization
process continues to proceed until 56 days of curing with a
significant gained in strength of more than 200%. The slow
increase in strength achieved during the lime stabilization phase
can be explained from the XRD results where CASH still remains
unconverted to CSH or CAH even after 14 days of curing.
However, after the period of 56 days, a dramatic increase in
strength was achieved due to the development of permanent well
crystallized CSH and CAH.
Pelepas marine clay responds more quickly to strength gain even
at the first 7 days of curing and the gain in strength continues to
develop in a progressing rate. The early gain in strength of 257%
is governed by the natural alkaline environment of the parent soil
with a pH of greater than 7. This will enhance the dispersion
process, i.e cation exchange, flocculation and agglomeration.
Lime stabilized sg. Buluh clay exhibit an increase in strength
from 70% to 120% compared to the untreated soil. The strength
increase with age was however not significant as compared to
other soil samples. Xrd test on sg. Buluh clay exhibit high content
of quartz. The high content of quartz in sg. Buluh clay provides a
higher strength of untreated soil compared to tapah kaolin and
pelepas marine clay. Quartz however does not assist in the
strength development but instead it acts as an obstacle from
44

achieving a good surface bonding between lime and the clay

particles. From the lime treated XRD results at 14 days, the new
compound of CSH were identified at a low relative intensity of
4.0% and 7.3%. On the other hands, the relative intensity of
kaolinite in the lime treated Sg. Buluh soil still remains unchanged.
This could be due to the nature of the soil which is quite acidic
thus causing the silicate and aluminate to be less soluble. This
could results in the lack of reaction between lime and clay.
Tapah Kaolin has a low XRD relative intensity quartz but high
in available kaolinite which is ready to react with lime. The
increase up to more than 10 times compared to the untreared soil
was achieved at 56 days of curing. The high XRD relative
intensity of CSH at 34.4% proves the success of the lime
stabilization in Tapah Kaolin.
Lime treated Jerangau clay was treated at mix design of 3%
lime instead of 6% based on the optimum strength gain (Figure
2.6). This is due to the higher available portlandite in the lime
treated Jerangau clay as compared to the other soils as shown in
Table 2.7. The above phenomenon shows that it is not necessary
the highest clay fraction of 44.2% in Jerangau clay will consume
the highest lime content for stabilisation. The lime stabilised
Jerangau clay produces new cementitious products such as CASH,
CSH and CAH as observed from XRD test. The lime treated
Jerangau clay presents an exponential trend of gain in strength over
the untreated soil.

2.4.10 Effect of Loss on Ignition (LOI) on Strength Development

Head (1984) described several ways in which water is held in

cohesive soils. Clay minerals which exist in plate-like particles of
less than 2 μm enable them to combine with or hold on to water by
several complex means. Water can be categorised into
gravitational water, capillary water, adsorbed water and chemically
combined water.
 45

Adsorbed and chemically combined water cannot be removed by

oven drying at 110oC and therefore considered to be a part of the
solid soil grain. Al-Khafaji & Andersland (1992) explained that
chemically combined water in the clay minerals will be released
only when the temperature is beyond 400oC.
Adsorbed water is held on the surface of the particle by
powerful forces of electrical attraction and virtually in a solid state.
Chemically combined water is in the form of water of hydration
within the crystal structure. Table 2.8 shows the determination of
the adsorbed and chemically combined water by oven dry at
105±5°C followed by 1000°C for at least one hour or until a
constant mass is attained.

Table 2.8. Determination of loss on ignition at 1000°C.

Pelepas Sg.
Kulai Tapah Jerangau
Untreated Soil Description marine Buluh
clay Kaolin clay
clay clay
Water Content Determination
Mass of porcelain dish
Mp 76.923 72.328 78.054 78.105 66.798
with aluminium foil (g)
Mass of porcelain dish
M pw 79.828 78.305 82.768 82.872 74.132
with aluminium foil + wet sample (g)
Mass of porcelain dish with
M pd 79.818 78.260 82.721 82.820 74.037
aluminium foil + oven dried sample (g)

Mass of oven-dried sample at 105 0C (g) Ms 2.895 5.932 4.667 4.715 7.239

Water content (%) w 0.35 0.76 1.01 1.10 1.31

Loss on ignition Determination

Mass of sample after burning at 1000 0C

M pb 79.678 77.739 82.415 82.165 73.537
for 1h + porcelain dish and foil (g)
Mass of burned sample (g) Mb 0.140 0.521 0.306 0.655 0.500

Mass loss due to ignition (%) XL 4.84 8.78 6.56 13.89 6.91

The amount of organic matter in all the cohesive soil samples

are small (< 1.0%), as previously presented in Table 2.1. Therefore
the values of LOI correspond mainly due to the crystallized water
46

in the clay minerals. LOI (%) was calculated by using the moisture
content determination method as shown in Table 2.8.

Figure 2.8 exhibit the influence of LOI on the Cst value at 28

days. This figure shows the trend of increasing in Cst with
increasing LOI. Inorganic soils with high LOI normally has thick
adsorb water layer which contributed to high cation exchange.
High cation exchange means high potential in pozzolanic reaction
between lime and clay.
600

500
Cst (%) at 28 days

400

300

200

100

0
0 5 10 15
LOI (%) of untreated soils

Figure 2.8 Loss on ignition (%) versus Cst at 28 days

2.4.11 Scanning electron micrographs observations

Scanning electron microscopy (SEM) technique has been used to

visualise the nature of clay particles present in the soil before and
after stabilization. The SEM of untreated Pelepas marine and
 47

Tapah Kaolin are shown in Figure 2.9 and Figure 2.10

respectively.

Figure 2.9 SEM of untreated Pelepas marine clay (20μm).

Figure 2.10 SEM of untreated Tapah Kaolin (5μm).

48

From the micrograph, it can be seen that in untreated Pelepas

marine clay there are lot of aggregated packets of clay particles.
The clay particles are assembled in a flocculated arrangement and
the porous nature of particulate arrangement can be seen. An open
type of fabric arrangement is expected in marine sediment samples
such as in Pelepas marine clay. The only clay particle in Tapah
Kaolin is kaolinite as confirmed by the XRD studies where there
are a lot of layers of clay particles.
Figure 2.11 shows the SEM of Pelepas marine clay treated with
6% lime after 28 days curing. The reaction of lime with the clay
minerals begins by replacement of the existing cations between the
silicate sheets with calcium ions. Clay lumps were being coated
and bounded by silicate gel, where the process of aggregation
starts to occur with a kind of spongy appearance. The gels
cemented the particles to form aggregated crumbs which make the
entire soil mass becoming porous. The cementation products
formed were identified by XRD known as CAH and CSH. Similar
formation of various cementation compounds such as CAH and
CSH due to soil-lime reactions in lime-treated soil systems has
been reported by Willoughby et al. (1968) and Rajasekaran &
Narasimha (1996).
Figure 2.12 shows the soil-lime mixtures micrographs illustrate
the new phase consists of an interlocking network of needle like
crystals. After curing for a year, bridges of interlocking network
are formed between adjacent soil particles. These interlocking
networks of needle like crystals have grown into the interstices to
form a continuous network.
 49

Cementitious

products

Figure 2.11 SEM of 6% lime treated Pelepas marine clay after 28 days
of curing (20μm).

Interlocking net
work of needle
like crystals

Figure 2.12 SEM of 6% lime treated Tapah Kaolin after one year of
curing (5μm).
50

2.5 Conclusions

The lower limit of lime for modification process could be

determined from LFC method whereas the minimum limit for
stabilization process to take place could be determined from icl
test. The lower limits of hydrated lime needed for modification and
initial stabilization process to take place is between 0.5 to 1.5%
and 1.5 to 3.0% by the dry mass of the soil respectively. Mix
design for the long-term stabilization however has to be based on
strength at 14 days curing. The optimum lime content for the long
term stabilization of the above Malaysian cohesive soils is in
between 3.0 to 6.0%.
Inspection on the quality of lime used is essential as it
determines the effectiveness of modification and stabilization
processes. Standard method of specifying the content of lime such
as available lime content test (ALC) should be conducted
regularly.
All soil materials give a similar trend with 6% lime addition to
achieve the maximum strength at 14 days except Jerangau clay,
which gives the optimum addition of lime at the initial amount of
3%. This phenomenon indicates that it is not necessary the highest
clay fraction will consume the highest lime content for
stabilization but depending on the particles surface area and the
alkaline environment for the pozzolanic reaction to take place.
The XRD results of treated soils cured for 14 days give an
indication of early reaction products such as CASH which leads to
an exponential growth in the Cst value for the first 56 days for
Kulai, Pelepas and Tapah clays. The success of the treatment
process is highly dependent on the available lime content, curing
time, soil type, soil pH and clay minerals. Clay minerals such as
kaolinite, chlorite and illite found in the untreated soils play an
important role in forming the character of the soils. Tapah Kaolin,
is the soil with the highest content of kaolinite produces new well-
crystallized compounds of CSH with a high relative intensity
according to XRD analysis. The increase in compressive strength
 51

for lime stabilized Kaolin is over 10 times higher than that of

untreated Kaolin. Pelepas marine clay responds more quickly to
the increase in compressive strength as the result of its alkaline
nature that enhances the dissolution process of silicates and
aluminates. Among the soil samples tested, lime treated Sg. Buluh
clay demonstrates a less significant improvement in the strength
over untreated soil due to its high content in quartz. High content
of quartz acts as an obstacles to particles surface bonding.
Soil with thick adsorb water layer i.e, high LOI, has a higher
capacity for cation exchange hence enhancing the process of
stabilization. This can be seen from the increasing trend of Cst with
increasing LOI.
From the scanning electron microscopy (SEM) technique on
stabilised Pelepas marine clay with 6% of lime, the changes in clay
particles due to flocculation was observed. The formation of
cementitious compound is confirmed in both stabilized Pelepas
marine clay and Tapah kaolin. Long term stabilization for one
year, has shown the development of new crystalline phases, which
consist of an interlocking network of needle like crystals. This
long-term improvement is defined as “stabilization” where the
formation of gels cemented the particles which increase in strength
with age. Evidence from the XRD analysis showed that the new
phase consists of a hydrate of calcium silicate or calcium
aluminates, which are the cementation products.

References

Bell, F.G. (1996). Lime Stabilisation of Clay Minerals and Soils.

Engineering Geology Publication 42: 223-237. London: The
Geological Society.
British Lime Association, (1990). Lime Stabilisation Manual,
Second Addition, January 1990. London: British Lime
Association.
52

British Standards Institution BS 1924 (1990). Methods of Test for

Stabilised Material for Civil Engineering Purposes, BSI
London, 1990.
British Standards Institution BS 6463 (1984). Methods of Chemical
Analysis, BSI London.
Hilt, G.H. & Davidson, D.T., (1960). Lime Fixation of Clayey
Soils, High. Res. Board, Bull. 262: 20-32. Washington, DC.
Ingles O.G. (1970). Mechanisms of Clay Stabilisation with
Inorganic Acids and Alkalis, J. Soil Research, 85: 81-96.
Australia.
Mitchell, J.K. and Hooper, D.R. (1961). Influence of Time
between Mixing and Compaction on Properties of Lime
Stabilised Expansive Clay. High. Res. Board Bull. 304: 14-31.
Washington, DC.
Neoh C.A. (1998). Design & Construction of Pile Foundations in
Limestone Formation, Journal of Institution of Engineers,
59(1), March: 30. Malaysia: Institution of Engineers.
Pitts, J. (1984). A Manual of Geology for Civil Engineers.
Singapore: Singapore National Printers.
Rajasekaran, G. and Narasimharao, S. (1996). Reaction Products
Formed in Lime-Stabilised Marine Clays. ASCE: Journal of
Geotechnical Engineering. Vol. 122, No. 5: 329-336.
Willoughby, D.R., Gross, K.A., Ingles, O.G., Silva, S.R. and
Spiers, V.M. (1968). The Identification of Reaction Products in
Alkali Stabilised Clays by Electron Microscopy, X-ray and
Electron Diffraction. Proc. of the 4th Conference of Australian
Board 4, Part 2, Paper No.430 :1386-1408.
CHAPTER 3
 

GROUND IMPROVEMENT BY
PRELOADING AND VERTICAL DRAIN
 
Nurly Gofar
Universiti Teknologi Malaysia

Rosdi Mohamed
Universiti Malaysia Pahang
 
 
3.1 Introduction

In Malaysia, soft soil deposits are widespread all over the country
and mostly found in the coastal area. In general, soft soil posses
low strength and high compressibility and thus having low bearing
capacity. The soil needs to be improved to avoid excessive
settlement and prevent stability failure that affecting the safety of
the infrastructure.
Preloading is one of the most effective and economical
methods to reduce settlement and improve the bearing capacity of
the soft soil. The application of preloading does help in applying
stress to the existing soil, thus increasing the pore water pressure.
Consolidation happens when the water in soil is squeezed out from
the soil matrix. With the elimination of water from the soil, the soil
strength is increased. However, this method does not have the
advantages of speeding up the process of pore water pressure
dissipation hence settlement might takes years to complete.
The preloading method is usually combined with vertical drain.
The use of prefabricated vertical drain in a soft soil layer can
shorten the treatment time of ground by increasing the rate of

53 
 
54 
 

consolidation. The consolidation time depends on the travelling

distance of the water to flow out from the soil matrix. The
installation of vertical drain shortens the drainage path by allowing
the water to flow horizontally towards the drain, and drain out to
the soil surface through the vertical drain.
Generally, the installation of vertical drain is carried out in an
extensive area. The behavior of vertical drain can be expressed as
axisymmetric unit cell. There are studies carried out by researchers
on vertical drain using linear model, e.g. embankment on soft
Bangkok clay (Bergado 1993). Besides, analytical methods are
also used by designers assuming that consolidation happens in
uniform soil column with linear compressibility characteristics
without any lateral movement.

3.2 Basic Concept

 
The benefits of using preloading and vertical drain as ground
improvement are shown in figure 3.2. Consolidation settlement
analysis is usually carried out based on 1-Dimensional
consolidation theory as proposed by Terzaghi. Since the soft soil is
usually in Normally Consolidated condition, the primary
consolidation settlement given as:
H σ' + Δσ'
S c = Cc log o (3.1)
1+ eo σ' o
where Cc/1 + e0 is the compression ratio (CR), σ’0, and Δσ are initial
effective stress and the stress increment respectively and H is the
thickness of the compressible layer.
There are three types of preloading i.e. simple preloading, surcharge
preloading and stage loading. Fore surcharge preloading, a temporary fill
(Δq) was applied to the permanent fill (Δp) so that the anticipated
settlement under Δp alone can be achieved in less time. The temporary
surcharge should be removed after the 90% of anticipated settlement was
achieved and no excess pore water pressure was left in the soil. In
practice, the time for surcharge removal was determined based on the
 55 
 

construction schedule and the height of the extra surcharge could be

calculated.

Figure 3.1 The beneficial effect of preloading and vertical drain (after
Haussmann, (1990)

The time to reach a certain percentage of the expected settlement

is:
2
H T
t= d v (3.2)
Cv
where Hd is the length of drainage, Tv is the time factor and Cv is
the coefficient of consolidation. Figure 3.2 shows the curve used
for time factor Tv.
56 
 

1.2
1.0
Time Factor, Tv

0.8
0.6
0.4
0.2
0.0
0 10 20 30 40 50 60 70 80 90 100
Percent consolidation (%)

Figure 3.2 Time factor for vertical consolidation

When vertical drain is used to speed up consolidation, the

overall consolidation process consists of horizontal and vertical
components. The combined effect is given by Carillo (1942) as:

U vh = 1 − (1 − U h )(1 − U v ) (3.3)

where Uh and Uv = average degrees of consolidation in the

horizontal and vertical directions, respectively. The equation for Uv
is given by Terzaghi and available in many literatures (e.g. Bardet,
1997; Whitlow, 1995). The modified general expression for
average degree of consolidation in horizontal direction is given by
Hansbo (1979, 1981) as:
⎛ − 8c t ⎞
U h = 1 − exp⎜⎜ 2 h ⎟⎟ (3.4)
⎝ De F ⎠
where De is the equivalent diameter of the soil cyclinder; t is the
time required to achieve Uh and F is the factor influencing the
consolidation which is a combination of F(n), Fs and Fr known as
the drain spacing factor, the smear factor and the well-resistance
factor, respectively:

F = F (n) + Fs + Fr (3.5)
 57 
 

⎛D ⎞
F (n) = ln⎜⎜ e ⎟⎟ − 0.75 (3.6)
⎝ dw ⎠
⎡⎛ k ⎞ ⎤ ⎛ d ⎞
Fs = ⎢⎜⎜ h ⎟⎟ − 1⎥ ln⎜⎜ s ⎟⎟ (3.7)
⎣⎢⎝ k s ⎠ ⎦⎥ ⎝ d w ⎠
⎛k ⎞
Fr = πz ( L − z )⎜⎜ h ⎟⎟ (3.8)
⎝ qw ⎠

In equations 3.6 to 3.8, dw is the equivalent drain diameter; kh is

the horizontal permeability of the undisturbed soil; ks is the
horizontal permeability of the smear zone; ds is the diameter of the
smear zone; L is the effective drain length; z is the distance below
the top surface of the compressible soil layer; and qw is the vertical
discharge capacity of the PVD at an hydraulic gradient of unity.
Yee (2000) suggested that the influence of Fr with discharge
capacity above 100m3/year could be neglected unless the drains
exceeds 20m long. Yee (2000) also recommended that smear factor
for the PVD design can be assumed to be zero for soils with
sensitivity less than 4. However the smear effect is also affected by
many other factors, such as the shape, the spacing of the drain, the
size of the mandrel for installation and the type of soil (Chu et al.,
2004). When the drains are installed too close to each other, the
smear effect may become too large to be ignored.
Determination of ch requires advanced laboratory tests such as
Rowe cell or other consolidometers. ch can be assumed as a
proportion of the cv, based on the following relationship (Bergado
et al. 1996):
⎛k ⎞
c h = ⎜⎜ h ⎟⎟cv                     (3.9)
⎝ kv ⎠
In practice, the general assumption to use ch/cv ranging between
1.5 and 2.
58 
 

3.3 Case Study

The pre-compression and vertical drain method were implemented

as ground improvement in the earthwork package of the
development of permanent campus for Universiti Malaysia
Pahang (UMP) in Kuala Pahang, Pekan, Pahang (Figure 3.3) to
improve the subsoil and to reduce the time required for soil
stabilization. Thus, relevant data such as the properties of soil both
before and after ground improvement work was collected from the
results of soil investigation performed prior to ground
improvement (2002) and after the ground improvement (2007).
Data on settlement monitoring of an observational embankment
(December 2003 to March 2004) was used for the achievement of
acceptance criteria.
The performance of this ground improvement method can be
evaluated by comparing the engineering properties of the subsoil
before and after stabilization. In addition, the achievement of
ground improvement criteria can be evaluated by observing the
time required to reach the predetermined criteria of the ground
improvement project through settlement monitoring. Settlement
analysis was performed for backfilling on original soil without and
with surcharge preloading and with surcharge preloading and
vertical drain following method suggested in Hausmann (1990).
Analysis of the performance of preloading and vertical drain was
done based on monitoring data and Asaoka procedure (Asaoka,
1978).
 59 
 

Figure 3.3: Site of UMP’s permanent campus

3.3.1 Soil Profile

A total of fourteen boreholes had been carried out at the proposed

site before the implementation of ground improvement project.
The subsoil information gathered from the boreholes carried out at
the proposed site showed that, in general, the soil profile comprises
of four layers.
Layer 1 is classified as very loose to medium dense silty sand.
This layer mainly consists of sand, silty sand and gravelly sand
with thickness of about 6 to 12 m. The Standard Penetration Test
results (SPT-N) was in the range of 5 to 15. Occasionally, there
were very loose sand having SPT-N below 3 near the ground
surface and medium to dense lenses of SPT-N value of 20 to 26.
From laboratory test result, the organic contents were in the range
of 3% to 5% within the depth of 6 m.
60 
 

Layer 2 consists of very soft to medium stiff clayey silt or silty

clay with some sand. The thickness varies from about 6 m to 14
m. The SPT-N values were generally less than 4. From the
laboratory test results, the moisture contents were in the range of
26 to 55%. The liquid limit (LL) and plastic index (PI) were about
29% to 66% and 8% to 31% respectively.
Layer 3 mainly consists of medium stiff to very stiff clayey slit
or silty clay with some sand. The SPT-N values generally were in
the range of 5 to 15 and increases with depth. The thickness varies
from about 12 to 20 m.
Layer 4 is hard or very dense soil layer. This layer generally
could be encountered at depth of about 35 to 42 m. The SPT-N
values were more than 50. In BH1, the depth was 30m whilst in
BH5, BH11 and BH13, the hard layer was not encountered within
borehole termination depth of 40m. In a few boreholes, namely
BH4, BH5, BH6 and BH14, dense lenses of sandy material of 2 to
6 m thick were encountered before the borehole was terminated
upon reaching SPT-N value greater than 50. The groundwater level
fluctuates from the ground surface to about 4 m below the existing
ground surface.
Based on site investigation, it can be summarized that loose
sandy soil was found below the existing ground surface followed
by a layer of soft clayey silt. Underlying the soft clayey silt layer
was medium stiff to stiff clayey silt inter-bedded with dense sandy
lenses before encountering hard soil layer. Figure 3.4 shows the
typical soil profile based on borehole log.
 61 
 

Figure 3.4: Typical soil profile based on borehole log

For the purpose of settlement analysis, the soil profile and soil
properties can be simplified as shown in Table 3.1.

Table 3.1 Simplified Soil profile

Soil layer Unit Thickness
Weight
Without With PVD
(kN/m3)
PVD
Sandy soil 18 8 6
Soft Silty soil 17 10 14
Sandy soil 18 3.5 8
Medium stiff soil 18 14.5 12
62 
 

3.3.2 Soil Properties

From the results of site investigation, the undrained shear strength

properties of the compressible layer (Su) as obtained by field vane
shear tests was about 6 to 30kPa. The compressible layers were
normally consolidated to slightly over-consolidated with OCR of
about 1.0 to 2.0. The compression index cc was about 0.2 to 0.6
while the compression ratio (CR) was between 0.1 and 0.3. The
average recompression index (cr) was 0.02 giving a recompression
ratio (RR) of 0.01. The coefficient of consolidation (cv) of the soft
soil layer as obtained from laboratory test was about 6 to 21m2/yr.

3.3.3 Preloading and Vertical Drain

The designed platform level approved by Jabatan Pengairan dan

Saliran (JPS) was at RL 5.5m while the existing ground level
varies from R.L. 3.7m to 5.7m. Therefore, backfilling of 1 to 2m
was required to reach the required level. The backfilling would
cause the soft soil layer to experience consolidation settlement and
the settlement would continue for years. Since time was of
concern in the project, surcharging was recommended. The
thickness of surcharge fill was about 1 to 2m.
There were areas where soft soil layer was very thick (up to
14m) that vertical drain was required to further speed up the
consolidation process. The prefabricated vertical drains were
installed with the average depth of 20m at 1.4m spacing in a square
pattern (Figure 3.5). The width and thickness of the PVD’s strip
used in this project were 100mm and 4mm respectively.
 63 
 

Figure 3.5: Installation of PVD at site

3.3.4 Acceptance Criteria

The construction of the buildings was scheduled to start at least

after 90% of the anticipated settlement has been achieved. The
post construction settlement should also be less than 50 mm.
To monitor the achievement of the settlement, instrumented
embankments were constructed at the site. The typical cross-
section of the observation embankments is shown in Figure 3.6.
The designated differential settlement across the embankment was
not to exceed 1% within 7 years.
64 
 

Settleme
nt gauge
Observation  Top 
embankment platform  
Water 
  Sand    level 

PVD   piezometer

Figure 3.6 Typical cross-section of the observation embankment

3.4 Results and Discussion

Settlement analysis was made for the anticipated settlement under

backfilling to the design platform and compensation fill. The
settlement under backfill and surcharge fill of 1 to 2 m was
evaluated to observe the effect of pre-compression. Effect of
vertical drain in speeding up the consolidation process was also
evaluated.
As mentioned above, the level of ground surface and the
thickness of compressible layer vary with the location of the
borehole.
The compressible layer is in normally consolidated condition
with CR of 0.21. A moderately cv value of 14m2/yr was used.
Calculation presented herein is the analysis for the worst case
where the platform fill was 2 m and the surcharge fill was also 2
m. Settlement analysis shows that for backfilling of about 2 m,
the estimated consolidation and immediate settlements due to
backfill only was 355 and 54 mm respectively. The time to reach
90% of consolidation (320 mm consolidation settlement) was 1.5
yrs. When surcharge of 2 m and compensation settlement height
was applied on top of the fill, the immediate and consolidation
settlement became 126 and 551 mm respectively giving a total
 65 
 

settlement of 677mm. Resting period of about 6 months was

required to reach the prescribed consolidation settlement of
320mm before the surcharge could be removed. Figure 3.7 shows
the typical settlement plot for the above case.

Figure 3.7 Settlement plot for embankment with and without surcharge
preloading

For platform fill of 1m, the total settlement was 261mm. When
surcharge pre-load of 1m and 2m high were used, the 90%
consolidation settlement was reached after 5.3 months and 3
months respectively. Table 3.2 shows the estimated time required
for various backfill and surcharge heights.
Vertical drain was installed in the swampy area where the
depth of compressible layer reaches 14m. For the worst case
analyzed above and with the installation of vertical drain, the
consolidation time was further reduced to one month. Figure 3.8
shows the achievement of settlement criteria when vertical drain
was used.
66 
 

Table 3.2 Surcharge Design for Platform without PVD

Platform Surcharge Total Consolidation Surcharge Time
Height Height Settlement with (month)
(m) Surcharge Load
(mm)
1.0 320 5.3
1.0 m Platform
2.0 427 3.0
2.0 m Platform
2.0 551 5.7

Figure 3.8 Settlement plot with and without vertical drain

The performance of ground improvement can be identified by

the achievement of the prescribed performance criteria and the
improvement in terms of engineering properties.
In order to verify the design analysis, an instrumented
embankment was constructed prior to the commencement of
earthwork. The achievement of the performance criteria was
evaluated by settlement measurement and the achievement of the
predicted settlement in terms of time. The settlement measurement
 67 
 

shows that in general, the embankment had reached the targeted

90% consolidation in less than 5 months after the construction. In
this case, the final settlement was predicted based on Asaoka
(1978) method. Figure 3.9 shows the typical data obtained from
settlement monitoring.
In 2006, soil investigation was carried out to collect detailed
subsoil information after the ground improvement. Results
gathered from laboratory tests on samples retrieved from ten
boreholes showed that the undrained shear strength (su) increased
to 67 kpa. The compression index (cc) based from the laboratory
test result was about 0.3 - 0.36 while the void ratio was about 1.04.
The coefficient of consolidation (cv) of the soft soil layer was
reduced to about 2.11m2/yr to 3.64m2/yr.

Table 3.3 summarizes the comparison of engineering properties of

the soil prior to and after the implementation of the ground
improvement project by pre-compression and vertical drain. The
comparison was made in terms of shear strength of the soil and its
compressibility characteristics such as cc, eo and cv.

Table 3.3 Comparison of soil properties prior to and after improvement

Prior to After improvement

Soil Properties
improvement
Su (6 – 30 kPa) 67 kPa
Average 10 kPa
cc 0.2 – 0.6 0.3 – 0.36
Average 0.5
e0 0.5 – 1.5 1.04
cv 6 - 21m2/yr 2.11 - 3.64m2/yr
68 
 

Figure 3.9 Typical data obtained by settlement monitoring

 69 
 

3.5 Conclusions

The case study proved that pre-compression has the beneficial

effect on increasing the bearing capacity and reducing the
compressibility of soil by forcing the compressible layer to
consolidate. Significant improvement was gained in terms of shear
strength and compressibility characteristics.
Based on the site investigation performed prior to construction,
ground treatment might not be required if sufficient time (12 to 18
months) were available for most of the settlement to be completed.
However as time was of concern, ground improvement by using
surcharging method was recommended to expedite the settlement.
It was achieved by placing a temporary surcharge on the top of
platform level prior to the construction of building works. The
surcharge time varies from 3 to 5.7 months for various conditions
encountered in the field. Vertical drain with surcharge was used
especially at swampy area to further expedite the settlement.
Calculation shows that the installation of vertical drain in the worst
soil profile further reduced the settlement to only one month.

References
Asaoka, A (1978). Observational Procedure of Settlement
Prediction. Soils and Foundation. 18 (4): 87-101.
Bardet, J.P. (1997). Experimental Soil Mechanics. New Jersey.
Prentice Hall.
Bergado D.T., Alfaro, M.C., and Balasubramaniam, A.S. (1993).
Improvement of soft Bangkok clay using vertical drains.
Geotextiles and Geomembranes , 12 (7): 615-664
Bumimetro Engineering Corporation Sdn Bhd. (2002). Factual
Report on SI Works for Cadangan Merekabentuk, Membina dan
Menyiapkan Kolej Universiti Kejuruteraan & Teknologi
Malaysia (KUKTEM) di Pekan, Pahang Darul Makmur
Carillo,N. (1942).Simple Two and Three Dimensional Cases in the
Theory of Consolidation. Journal of Mathematics s,21(1):1-5.
70 
 

Chu, J.,Bo, M.W., Chang, M.F. and Choa, V. (2002).

Consolidation and Permeability Properties of Singapore Marine
Clay. ASCE Journal of Geotechnical & Geoenvironmental
Engineering. 128(9):724-732.
Gofar, N and Kassim, KA. (2005) Introduction to Geotechnical
Engineering-Part I, Prentice Hall. Chapter 4
Goughnour, R. R. (1997). Lateral Drainage for Vertical Drain
Systems. Ground Improvement Ground Reinforcement Ground
Treatment Development 1987-1997 ASCE.
Green Effect Sdn. Bhd. (2002). Geotechnical Assessment Report
for the proposed design and build for Kolej Universiti
Kejuruteraan dan Teknologi Malaysia (Kuktem) Pekan,
Pahang. September
Green Effect Sdn Bhd. (2004). Preliminary Settlement Review on
Observation Embankments for the proposed design and build
for Kolej Universiti Kejuruteraan dan Teknologi Malaysia
(Kuktem) Pekan, Pahang. April
Hansbo, S. (1979). Consolidation of Clay by Band Shaped
Prefabricated Drains. Ground Engineering, 16-25.
Hansbo, S. (1981). Consolidation of Fine-Grained Soils by
Prefabricated Drains. Proc.10th Intl Conf on Soil Mechanics and
Foundation Engineering, 3:12-22, Balkema, Netherlands.
Hausmann, M. R. (1990). Engineering Principles of Ground
Modification, McGraw-Hill
Liu, C. and Evett, J. C. (2005). Soils and Foundations,
Pearson/Prentice Hall.
Progress Drilling Sdn Bhd. (2007). Soil Investigation Report for
Cadangan Pembangunan Kampus Tetap Fasa 1B Universiti
Malaysia Pahang di Mukim Kuala Pahang, Daerah Pekan,
Pahang Darul Makmur
Whitlow, R. (1995). Basic Soil Mechanics. United Kingdom.
Addison Weslet Longman Limited.
Yee, K. (2000). Consolidation of Soft Clay by Vertical Drainage
Method. Seminar on Ground Improvement - Soft Clay
(SOGISC). August 23-24. Dewan Jumaah UTM, Kuala Lumpur,
64-84.
 71

CHAPTER 4

GEOTECHNICAL BEHAVIOR OF
ELECTROKINETICALLY-TREATED
RESIDUAL SOIL

Kamarudin Ahmad and Khairul Anuar Kassim

Universiti Teknologi Malaysia

Mohd Raihan Taha

Universiti Kebangsaan Malaysia

4.1 Introduction

Several methods have been employed worldwide to improve

engineering characteristics of tropical granite residual soil. Among
them is the soil improvement techniques by diffusing chemical
substances into the soil with the assistance of a direct current (DC)
electrical potential was investigated. Chemicals were introduced at
the anodes or/and cathodes. The study was initiated since the
migration of chemical substances such as lime added into low
hydraulic conductivity soil is generally slow and restricted to small
distance from their sources (Glendinning and Rogers, 1996).
Enhancement of engineering properties is anticipated by
homogeneous precipitation of the species and appropriate ion
exchange mechanism. The selection of chemical substances is
based upon the types that have been utilised successfully on
temperate soils, particularly kaolinite.

71
72

Most residual soils in Malaysia are high in iron and aluminum

contents, which are very stable from further weathering (Taha et
al., 1997). A distinctive feature of the granitic residual soil is that it
is a mixture of sand, silt, and clay in varying proportions (Ting and
Ooi, 1976). One of the characteristics of such residual soils is the
existence of bonds between particles. de Carvallo and Simmons
(1997) attributed increased cohesional strength to the bridging or
bracing between quartz grains by clay minerals and/or
sesquioxides.
Barksdale and Blight (1997) found that all residual soils
behave as if overconsolidated but their compressibility is relatively
low at low stress. They also observed that once a threshold yield
stress has been exceeded, the compressibility increases. According
to Gidigasu (1976), Terzaghi’s theory of consolidation has been
found to be useful in predicting the settlement of structures on
residual soils. Investigations by Sridharan (1988) on
compressibility of tropical residual soil showed that shearing
resistance near contact points between soil particles controls the
compression behaviour.

4.2 Electrokinetic

Electrokinetic phenomena in porous medium are basically based

on the relative motion between a charged surface and the bulk
solution at its interface. The formation of an electric double layer
at the charged surface of clay particles is responsible for
electrokinetic phenomena of interest, namely electrophoresis,
electromigration and electroosmosis. In addition, oxidation and
reduction reactions took place at the electrodes as electrons are
transferred in and out of the system, resulting in ion diffusion, ion
exchange, development of osmotic and pH gradients, dessication
by heat generation at the electrodes, mineral decomposition,
precipitation of salts or secondary minerals, electrolysis,
hydrolysis, physical and chemical adsorption, and fabric changes
(Mitchell, 1993). At the same time, electrolysis of water at the
73

anode and the cathode produces oxygen and hydrogen,

respectively, which can be represented by the following equations;

2H 2 O − 4e − → 4 H + + O2 ↑ (anode) (4.1)

2 H 2 O + 2e − → 2OH − + H 2 ↑ (cathode) (4.2)

It is noteworthy that both H+ and OH- sweep across the soil

sample towards the cathode and the anode, respectively during the
course of electrokinetic processing. Since H+ travels approximately
two times faster than OH-, prolonged electrokinetic processing will
result in acidification of the soil.

3.2.1 Electroosmosis (EO) for soil improvement

The utilisation of EO to consolidate fine-grained soils has been

investigated by several researchers such as Mitchell and Wan
(1977) and Shang and Ho (1998). Comparison performed by and
Lo et al. (1991) before and after treatment revealed that the treated
clay was less compressible and the clay was virtually
“overconsolidated” by the EO process, indicating that the effect of
treatment is permanent.
According to Hausmann (1990), the stabilising effect of EO
can be enhanced by the appropriate choice of the electrode material
(e.g., aluminium instead of steel). In a pull out tests incorporating
electroosmosis, Hamir et al. (1997) demonstrated that the shear
strengths gained by incorporating the electroosmosis consolidation
vary from 70% to 150%, whereas the bond strength increment
resulted from the treatments range from 54% to 210%.

4.2.2 Electrochemical (EC) hardening for soil improvement

Introduction of additives to the anodes (Murayama and Mise,

1953; Harton et al., 1967; Gray, 1970 and and Hausmann, 1990)
74

besides the dissolved electrode materials may be utilised to ensure

the formation of irreversible cementing accretions, which
contribute to strength enhancement. There is a wide choice of
electrolytes that can be distributed into relatively impermeable
soils by electroosmotic flow. These chemical solutions can be fed
at the anodes or the cathodes depending upon the ions to be
transferred into the soil. The utilised chemical substances include
amongst others calcium chloride (Harton et al., 1967), aluminum
chloride (Gray, 1970), and aluminum sulphate and phosphoric acid
(Ozkan et al., 1999).
The initial stage of EC hardening is characterised by an
intensive development of molecular bonding resulted from ion
exchange, ion migration and electrolysis. Such chemical reactions
occur leading to the formation of irreversible compounds (Harton
et al., 1967; Gray, 1970; and Ozkan et al., 1999), which cause
cohesion in a clay soil. Subsequently, stronger and more stable
bonds arise between the particles, and the strength as a whole can
be attributed to a process of coagulation-crystallisation. In a study
utilising Al2(SO4)3 as the anolyte and H3PO4 as the catholyte,
Ozkan et al. (1999) observed that the treated soil gained up to
600% increase in strength. However, the strength increases were
not homogeneous throughout the sample.

4.3 Experimental Procedures

4.3.1 Sample preparation

Only soil passing 2.00 mm sieve was used in the electrokinetic
(EK) experiments. Soil samples with 90% of maximum dry density
and its corresponding optimum moisture content (33%) were
chosen as the basis of sample preparation for EK treatments. Every
mixture of soil and distilled water prepared for the EK treatment
was poured into the 10 cm diameter EK cell. Non-conductive
silicon grease was thinly applied onto the inner surface of the soil
compartment to avoid the formation of water film between the soil
75

and the inner wall of the cylinder. This precaution was important to
ensure that water would only flow through the soil during the
experiments. After the required quantity was obtained, the mixture
was pressed until a predetermined thickness of about 5 cm was
achieved so that a bulk density of 1695 kg/m3 was obtained. Then
each sample was subjected to saturation.
The basic physical and chemical properties of the soil are
tabulated in Table 4.1 and Table 4.2, respectively. This soil is
classified as silt of high plasticity with sand and gravel (MVG)
based on the British Soil Classification System (BSCS).

Table 4.1 Physical properties of soil

Specific gravity 2.71

Natural moisture content, % 36.5
Hydraulic conductivity, cm/s 3.60 x 10-6
Liquid limit, % 74.5
Plastic limit, % 49
Plasticity index, % 25.5
Shrinkage limit, % 18.31
Maximum dry density, kg/m3 1415
Optimum moisture content, % 33

Table 4.2 Chemical properties of soil

Organic 2.0%
Chloride 0.012%
Sulfate 5.85%
Carbonate 0.20%

4.3.2 Electrokinetic experiments

The electrokinetic experimental set-up is shown in Figure 4.1.

76

Titanium meshes were utilised as the electrodes due to its high

resistant to corrosion (Berkeley and Pathmanaban, 1990). For these
electrokinetic experiments, the only electrical parameter held
constant was the applied voltage, i.e. 30 V. No attempts were made
to maintain the voltage gradients and the current densities. Four
different electrokinetic systems were employed in this study,
which consisted of:

a) Distilled water as both the anolyte and catholyte (DW-DW).

b) 1.0 M aluminium chloride as the anolyte and distilled water as
the catholyte (Al-DW).
c) 1.0 M calcium chloride solution as the anolyte and distilled
water as the catholyte (Ca-DW).
d) Distilled water as the anolyte and 1.0 M phosphoric acid as the
catholyte (DW-PA).

Note that the left and right terms in each parenthesis represent
the anolyte and catholyte, respectively. The selection of the
chemical substances was based upon the types that have been
utilised successfully on temperate soils, particularly kaolinite.

4.3.3 Compressibility tests

After termination of electrokinetic processing, each treated sample

was extruded and divided into two parts namely the cathode side
and the anode side. Then each part of the treated sample was
subjected to one-dimensional consolidation test using 50 mm
diameter consolidation ring.
77

MARIOTTE BOTTLE BUBBLINGTUBE BUBBLINGTUBE MARIOTTE BOTTLE

GAS VENT

ELECTRODES

VALVE VALVE
INFLOW INFLOW
TUBE TUBE

SOILSAMPLE

GRADUATED
CYLINDER
GRADUATED
OUTFLOWTUBE OUTFLOWTUBE CYLINDER
MULTIMETER

DCPOWER
SUPPLY

Figure 4.1: Schematic diagram of the EK experimental set-up

4.3.4 Direct shear tests

After termination of electrokinetic processing, each treated sample

was extruded and divided into two parts namely the cathode side
and the anode sides. Then each part of the treated sample was
subjected to direct shear test using 60mm x 60mm shear box
apparatus. A normal stress of 40 kPa representing in-situ
overburden stress was applied to each specimen during shear. No
attempts were made to cure the treated samples with the
expectation that the treated samples should gain strength increase
during or immediately after treatment.
78

4.4 Results and Discussion

4.4.1 Compressibility

Figure 4.2 and Figure 4.3 show the normalised void ratios – log
pressures curves representing the treated samples near the cathodes
and anodes, respectively. Since disturbed samples were utilised for
the consolidation tests, yield stresses σy will be referred to in the
discussion instead of pre-consolidation pressures σc. In addition,
the consolidation curve of the compacted and saturated sample
(C&S) that was not electrokinetically treated would be utilised as a
reference to any changes in compression index (Cc) and swelling
index (Cs) resulting from the electrokinetic treatments. Table 4.3
summarises the values of indices and yield stresses obtained from
the consolidation tests.

4.4.1.1 DW-DW system

It is observed from Figure 4.2 that the normalised consolidation

curve of the DW-DW treated sample near the cathode resembles
that of the C&S sample. As a result, the Cc and Cs as well as the σy
of the DW-DW treated sample are similar to that of the C&S
sample.
As for the DW-DW treated sample near the anode, it is
observed from Figure 4.3 that the consolidation curve drops
significantly at the early stage of the consolidation process and
followed by a steady reduction in void ratios parallel to that of the
C&S sample. As a consequence, the σy is reduced to about 25 kPa.
Insignificant ionisation or desorption of cementing materials was
the most probable reason for the insignificant compression during
the final stage of the consolidation. Meanwhile, the resulting Cc
and Cs are quite similar to that of the C&S sample.
79

1.05
1.00
0.95
Normalised void ratio, e/eo

0.90 C&S
0.85 DW-PA
DW-DW
0.80 Ca-DW
0.75 Al-DW

0.70
0.65
0.60
1 10 100 1000 10000
Pressure, kPa

Figure 4.2: Normalised consolidation curves of electrokinetically treated

samples near cathodes

1.05

1.00
Normalised void ratio, e/eo

0.95

0.90 C&S
DW-PA
0.85
DW-DW
0.80
Ca-DW
0.75 Al-DW
0.70

0.65

0.60
1 10 100 1000 10000
Pressure, kPa

Figure 4.3: Normalised consolidation curves of electrokinetically treated

samples near anodes
80

Table 4.3: Values of Indices and Yield Stresses Obtained from

Consolidation Test

Treatment C&S DW-DW Ca-DW Al-DW DW-PA

system
Cc(anode) 0.25 0.24 0.23 0.25 0.24
Cc(cath) 0.25 0.25 0.24 0.23 0.14
Cc(anode) 0.03 0.03 0.03 0.04 0.02
Cc(cath) 0.03 0.03 0.01 0.04 0.01
σy(anode) 100 25 200 50 200
σy(cath) 100 100 100 100 400

Note: anode - anode side cath - cathode side

4.4.1.2 Ca-DW system

The treated sample near the cathode was slightly less compressible
than that of the C&S sample. The reduction in the Cc and Cs values
might be due to the formation of calcium hydroxide adjacent to the
cathode compartment as shown by the existence of a few thin
whitish hard spots. These hard spots might be highly resistant to
inter-particles sliding and deformation so that the treated sample
was slightly less compressible. In the meantime, no change was
observed in terms of σy value of the treated soil.
Similarly, the treated sample near the anode was less
compressible than that of the C&S sample. The reduction in Cc
could be due to changes in the pore water chemistry in addition to
reduction in the hydraulic conductivity during consolidation
process resulting in lower rates of pore pressure dissipation. On the
other hand, the Cs was not affected by the treatment. In addition,
the treated sample was apparently overconsolidated with σy of
about 200 kPa.
81

4.4.1.3 Al-DW system

The Cc of the treated soil near the cathode decreased, whilst near
the anode the value remained as 0.25. Meanwhile, the values of Cs
near the anode and cathode increased slightly. It was found that the
treated soil near the cathode was less compressible than the C&S at
higher stresses. As part of the explanation, the formation of
aluminium hydroxide gel (Murayama and Mise, 1953) might have
reduced the voids between soil particles (Shin et al., 2002) thus
reducing the sizes of flow paths for dissipation of excess pore
pressure during the consolidation process. It was also possible that
the increased viscosity due to changes in pore fluid chemistry was
responsible in reducing the pore pressure dissipation rate.
The σy values of the treated sample near the anode decreased to
about 50 kPa whilst near the cathode it remained approximately
100 kPa. The acidic environment that dissolved certain compounds
(Ozkan et al., 1999) might increase the lubrication effects between
soil particles (Ghazali, 1981), thus responsible for the reduction in
σy near the anode. Swelling of the treated sample near the anode
due to direct contact with the anolyte (Taha, 1996) as well as the
released of stress during sampling were additional reasons for the
reduction in σy value.

4.4.1.4 DW-PA system

The Cc and Cs values of the treated soil near the cathode decreased
by about 44% and 67%, respectively with respect to the C&S
sample. Near the anode, no significant change was observed in
term Cc, whilst the Cs was reduced by about 33%. Meanwhile, the
values of σy of the treated soils near the cathode and anode
increased to approximately 400 kPa and 200 kPa, respectively.
Such changes were associated to the formation of phosphate
compounds. These compounds encapsulated and bridged the soil
particles together to form large soil grains and exerted internal
effective stress under which the soil was consolidated due to the
confinement of the sample in the electrokinetic cells. According to
82

Lo et al. (1991), the samples were therefore virtually

“overconsolidated” by the electrokinetic processing, and this is a
reason why the effects of the treatment are permanent. It was also
suspected that the cementing compounds and thus the cemented
particles were highly resistant to sliding and deformation.

4.4.2 Strength

Figures 4.4 and 4.5 show the plots of shear stresses against axial
strains obtained from the direct shear tests on the samples within
the vicinity of the cathodes and anodes, respectively. Besides the
curves, Table 4.3 tabulates the ultimate shear stress values deduced
from these figures.

4.4.2.1 Compacted and saturated (C&S) sample without EK

treatment

It was found that the averaged maximum shear stress of the C&S
samples was about 31 kPa. The shear resistance achieved was due
to cohesion, surface friction as well as the packing and orientation
of the soil particles. The cohesion contributing to the shear
resistance was resulted from factors like cementation, which might
be oxides of iron or aluminium (Gidigasu, 1976; Blight, 1997 and
Fookes, 1997).
83

100

80
C&S
Shear stress, kPa

60 Ca-DW

DW-PA

40 Al-DW

DW-DW
20

0
0.000 0.050 0.100 0.150 0.200

Axial strain

Figure 4.4: Shear resistances of EK treated samples near the cathodes

100

80
C&S
Shear stress, kPa

60 Ca-DW
DW-PA
40 Al-DW
DW-DW
20

0
0.000 0.050 0.100 0.150 0.200
Axia l stra in

Figure 4.5: Shear resistances of EK treated samples near the anodes

84

Table 4.3: Ultimate Values of Shear Resistance After EK Processing

EK system Shear stress, % increase Shear % increase
kPa (Cathode) stress, kPa (Anode)
(Cathode) (Anode)
C&S (No EK) 31 0 31 0
DW-DW 30 -3 25 -19
Al-DW 12 -61 12 -61
Ca-DW 31 0 15 -52
DW-PA 83 168 40 29

4.4.2.2 Ca-DW system

The shear resistance at the cathode was about 31 kPa, similar to

that of the C&S sample. On the other hand, a reduction of about
52% was observed at the anode. Because of the acidic environment
within the soil near the anode (pH = 5.3) and slightly alkaline near
the cathode (pH = 7.53) resulted from the EK processing, no
cementitious materials contributing to strength increase were
formed. It is noteworthy that calcium ions will only react with clay
minerals to form cementitious materials under highly alkaline
environment, i.e. pH of 12.4 (Glendinning and Rogers, 1996). The
lower pH environment than 12.4 caused the calcium to remain
ionic. Thus the pore water containing calcium ions might act as
lubricant within the system (Ghazali, 1981).

4.4.2.3 Al-DW system

The strengths of the soil treated using this system were reduced
both at the cathode and the anode, respectively. Such observations
contradicted to the results obtained by other researchers such as
Murayama and Mise (1953), Harton et al. (1967) and Gray (1970).
Testing showed that the pH of the soil was reduced from 5.3 to
about 3.5 after the EK processing. Thus, the reduction of shear
resistance of the treated soil was associated with the acidic
85

environment as the migrated H+ from the anode compartment into

the soil altered the chemistry of the pore fluid, which led to
ionisation of certain minerals such as kaolinite besides desorption
of basic precipitates and surface ions from the soil surface (Ozkan
et al., 1999). Curing period in addition to introduction of
secondary chemical might be required to gain strength increases as
shown by Gray (1970). Murayama and Mise (1953) found that the
strengthening of clay soils treated electrokinetically by the
injection of Al3+ would only take place once the precipitated
Al(OH)3 was dehydrated to become bauxite (AlO(OH)) and
alumina (Al2O3).

4.4.2.4 DW-PA system

The maximum shear resistances of the soil near the cathode and
anode were increased to 83 kPa and 40 kPa, respectively after
treated electrokinetically. The formation of phosphate compounds
as coating and cementing agents might be the most probable reason
for the strength. Such compounds were formed as a result of
electromigration and advection of phosphate ions originated from
the cathode compartment across the soil sample during the course
of the EK experiment. Gidigasu (1976) anticipated that Al3+ was
released due to acidification of residual soil. Hence, the most
probable compound formed resulted from the injection of
phosphate ions was that of hydrated aluminium phosphate.
However, attempts made based on X-ray Diffraction (XRD)
method to identify the formation of new minerals or compounds on
the DW-PA treated samples immediately after completion of the
processing were unsuccessful. Hence, it is hypothesised that the
new and fresh cementing compounds were in a non-crystalline or
an amorphous form (Mohamed, 2003).
86

4.5 Conclusions

It was observed that the DW-PA system resulting in a significant

reduction in Cc of the treated soil near the cathode due to the
formation of phosphate compounds and subsequent internal
effective stress. The other systems did not significantly change the
compressibility of the treated soils both near the anode and the
cathode. The formation of phosphate compounds and calcium
hydroxides reduced the Cs values of the DW-PA and Ca-DW
treated soils near the cathode, respectively. The significant
reduction of Cc near the cathode and insignificant reduction of Cc
near the anode of the DW-PA treated soil showed the non-
homogeneity of chemical precipitation throughout the treated
sample. EK treatment also affected the yield stress values of the
treated soil.
In addition to compressibility, the utilisation of EK processing
combined with the introduction of chemicals led to increases and
decreases in strengths of the treated soil depending upon the ion
types and locations of the soils within the system during
treatments. The introduction of phosphoric acid at the cathode
increased the shear resistances near the cathode and the anode due
to the formation of phosphate compounds. Utilising Ca2+ as
anolyte resulted in decreases in strengths due to the acidic
environment within the soil mass. Curing period or dehydration
process might be required for aluminium compounds to crystallise
in the system that involved injection of Al3+ to contribute to
strength increase.

References

Barksdale, R.D. and Blight, G.E. (1997). Compressibility and

Settlement of Residual Soils. Mechanics of Residual Soil,
Blight, G.E (ed), Balkema, 95-154.
Bell, F.G. (1975). Methods of Treatment of Unstable Ground. Bell,
F.G. ed. Newnes-Butterworths. London, 26-36.
87

Berkeley, K.G.C. and Pathmanaban, S. (1990). Cathodic Protection

of Reinforcement Steel in Concrete. Butterwoths, London.
Blight, G.E. (1997). Origin and Formation of Residual Soils,”
Mechanics of Residual Soils. Blight, G.E. ed. Balkema,
Rotterdam, 1-15.
de Carvallo, J.B. and Simmons, J.V. (1997). Mineralogy and
Microstructure. Mechanics of Residual Soil, Blight, G.E. ed.,
Balkema, 31-40.
Fookes, P.G. (1997). Geological Society Professional Handbooks -
Tropical Residual Soils. A Geological Society Engineering
Group Working Party Revised Report, Fookes, P.G. ed. The
Geological Society London.
Ghazali, M.F. (1981). Soil Stabilization by Chemical Additives.
PhD Thesis, The University of Washington, USA.
Gidigasu, M.D. (1976). Laterite Soil Engineering: Pedogenesis and
Engineering Principles. Elsevier.
Glendinning, S. and Rogers, C.D.F. (1996). Deep Stabilisation
Using Lime. Lime Stabilisation, Rogers et al. eds. Thomas
Telford, London, 127 – 136.
Gray, D.H. (1970). Electrochemical Hardening of Clay Soils.
Geotechnique 20(1):81-93.
Hamir, R.B., Jones, C.J.F.P. and Clarke, B.G. (1997). Soil Nailing
Incorporating the Electro-osmostic Phenomenon. Proc. 4th
Regional Conf. In Geotechnical Engineering (GEOTROPIKA
’97), Universiti Teknologi Malaysia, 181-189.
Harton, J.H., Hamid, S., Abi-Chedid, E., and Chilingar, G.V.
(1967). Effects of Electrochemical Treatment on Selected
Physical Properties of a Clayey Silt. Engineering Geology. 2(3)
191-196
Hausmann, M.R. (1990). Engineering Principles of Ground
Modifications. McGraw-Hill Publishing Company, New York.
Lo, K.Y., Inculet, I.I., and Ho, K.S. (1991). Electroosmotic
Strengthening of Sensitive Clays. Canadian Geotechnical
Journal, 28(2):62-73.
Mitchell, J. K. (1993). Fundamentals of Soil Behavior. 2nd ed.,
John Wiley & Sons Inc., New York.
88

Mohamed, H. (2003). Professor of Geology, Univ. Kebangsaan

Malaysia, personal comm.
Murayama, S. and Mise, T. (1953). On the Electrochemical
Consolidation of Soil using Aluminium Electrodes. Proc. 3rd
International Conference on Soil Mechanics and Foundation
Engineering, Zurich, 156-159.
Ozkan, S., Gale, R.J., and Seals, R.K. (1999). Electrokinetic
Stabilization of Kaolinite by Injection of Al3+ and PO43- Ions.
Ground Improvement, 3:135-144.
Shang, J.Q. and Ho, K.S. (1998). Electro-Osmotic Consolidation
Behaviour of Two Ontario Clays, Geotechnical Engineering,
Journal of Southeast Asian Geotechnical Society, 29(2): 181-
194.
Shin, E.C., Omar, M.T., Tahmaz, A.A., Das, B.M. and Atalar, C.
(2002). Shear Strength and Hydraulic Conductivity of Oil-
Contaminated sand. Proc. of 4th Intl. Congress on
Environmental Geotechnics (4th ICEG), de Mello, L.G and
Almeida, M, eds. Brazil, Balkema Publishers, 9-13
Sridharan, A. (1988). General Report – Technical session II:
Engineering Properties of Tropical Soils. Supplementary
Documentation of the 2nd Intl. Conf. on Geomechanics in
Tropical Soils, Singapore, 42-56
Taha, M.R. (1996). Micellar Electrokinetic Remediation of TNT
from Soil. PhD Thesis, The Louisiana State Univ., USA
Taha, M.R., Acar, Y.B., Gale, R.J. and Zappi, M.E. (1994).
Surfactant Enhanced Electrokinetic Remediation of NAPLs in
Soils, Proc. of 1st Intl. Congress on Env. Geotechnics,
Edmonton, Canada, A.A. Balkema, Rotterdam, 373-377.
Taha, M.R., Sarac, D., Chik, Z, and Nayan, K.A. (1997).
Geotechnical and Geoenvironmental aspects of Residual Soils.
Proc. 4th Regional Conference in Geotechnical Engineering
(GEOTROPIKA 97), Univ. Teknologi Malaysia, 331-341.
Ting, W.H. and Ooi, T.A. (1976). Behaviour of a Malaysian
Residual Granite Soil as a Sand-Silt-Clay Composite Soil.
Geotechnical Engineering, 7: 67-79.
CHAPTER 5
GROUND STABILIZATION BY TREE-
INDUCED SUCTIONS
 

Nazri Ali
Universiti Teknologi Malaysia

5.1 Introduction

It is now recognized that variations in soil suction that may occur

in the presence of vegetation, and indeed those that can occur on
removal of vegetation, have an important role to play in the
performance of engineered and natural soil slopes. This is a
problem that is exacerbated by climate change and increasingly
intense rainfall events. Repair, maintenance and operation of
railway and road embankments are particular areas where this is an
important issue.
Previous experience in Malaysia indicates that removal of
vegetation can be the main cause for the failure of a slope
(IKRAM, 1994). In this tragedy, investigation of the substructure
and the surroundings revealed that clearing of trees on the adjacent
slope led to the water level in soil to rise, thus causing the
instability of slope.
Some researchers (Thorne, 1990; Simon et al, 2000; MacNeil,
2001) claim that vegetation is widely believed to increase the
stability of slope. Simon and Collison (2001) suggested that the
impact of vegetation on slope or bank stability can be divided into
mechanical and hydrologic effects. Mechanical effects come from
the root tensile strength and hydrologic effects increases slope
stability by extracting soil moisture for transpiration. In their work,

89
90 
 

it was found that hydrologic effects can be as important as

mechanical effects and in certain cases.
Ridley et al (2004) discussed the relationship between climate,
in the form of soil moisture deficit, the presence of trees and pore
water pressures in embankments. The study showed that most
embankment failures in London were due to an increase in soil
moisture content.
Greenwood (2006) considered the potential engineering
influences of vegetation and how it can be characterized on site
within a geotechnical framework for slope stability assessments.
Greenwood’s software (SLIP4EX - based on equilibrium of forces)
may be used for estimating the factor of safety (FOS) against slope
failure and can be readily adapted to include vegetation effects.
The program uses the method described by Greenwood et al (2004)
to include the influence of vegetation mass, effects on the
groundwater regime, enhanced cohesion due to fine roots, wind
forces and the anchoring effects of the larger roots. Changes in
ground water table due to vegetation were included however effect
of negative water pressure was excluded from the equation. For the
situations where the ground water table is deep and where tree root
activity is involved, tree-induced suctions may be significant. In
these cases, it may be appropriate to perform slope stability
analyses which include the shear strength contribution from the
negative pore pressure. A modified form of the Mohr-Coulomb
equation can be used to link shear strength to soil suction.
Therefore, this chapter explores the application of the water
uptake model described previously (Rees and Ali, 2006) to provide
a preliminary assessment of the significance of water content (and
therefore suction) changes on the stability of unsaturated soil
slopes. This work only considers hydrological effects (i.e. water
uptake) at this stage.
 
 91 
 

5.2 Theoretical and Numerical Background

5.2.1 Limiting Equilibrium Forces for Unsaturated Slope Stability

This study uses the theory of limiting equilibrium of forces and
moments to compute the FOS against failure. The limiting
equilibrium method of slices is widely used for its simplicity
particularly when compared to the finite element method (Fredlund
and Rahardjo, 1993; Renaud et al, 2003). The FOS is defined as
that factor by which the shear strength of the soil must be reduced
in order to bring the mass of soil into a state of limiting
equilibrium along a selected slip surface. Calculations for the
stability of a slope are performed by dividing the soil mass above
the circular slip surface into vertical slices. The limit equilibrium
formulation assumes that the factor of safety is to be equal for all
soils involved and for all slices.
The current work aims to explore the importance of suction
changes on shear strength. Therefore, in the example considered,
the water table is assumed to be below the zone of interest.
Tension cracks are, therefore, excluded from the current work. It is
also assumed that there are no interslice shear forces involved in
the equation for both horizontal and vertical forces. This
assumption has been made due to following reasons;
i) Vertical interslice forces can be assumed equal and opposite
(Bishop, 1955).
ii) The resultant of the interslice forces acting on a slice can be
assumed to act parallel to the base of the slice. By resolving
forces normal to the base of the slice, the interslice forces are
eliminated.

5.2.2 FOS for Unsaturated Slope

To calculate the FOS of an unsaturated soil slope and link this with
the effect of tree-root-water uptake, a force equation which
includes matric suction must be established first. The analysis is an
extension of conventional limit equilibrium analysis an equation is
formed using the basic principle of static equilibrium of forces and
92 
 

moments. The mobilized shear force at the base of a slice can be

written as (Lambe and Whitman, 1969)
τl
Sm =           (5.1) 
F
where τ is shear strength of unsaturated soil. Fredlund et al (1978)
provided the following expression for shear strength:

τ = c '+ (σ n − u a ) tan φ '+ (u a − u w ) tan φ b (5.2)

Where (u a − u w ) is the matric suction and φ b is an angle

indicating the rate of increase in shear strength relative to matric
suction. (σ n − u a ) is the net normal stress, c ′ is the effective
cohesion and φ ′ is angle of friction. Combining equations (5.1)
and (5.2), gives,
Sm =
[
l c'+(σ n − u a ) tan φ '+(u a − u w ) tan φ b ]
  (5.3)
F
From Figure 5.1, (taking point O as centre of the moments) the
summation of moments in the slope, yields:

∑ Wx − ∑ S m R = 0 (5.4)
Substituting equation (5.3) into (5.4) and substituting
(u a − u w ) = S (matric suction) and assume air pore pressure is
atmospheric, u a = 0 , equation (5.4) become

F=
∑ c' lR + NR tan φ '+ SRl tan φ b

      (5.5) 
∑Wx
Equation (5.5) has been used throughout the research for
calculating the FOS. Note that if the matric suction is zero (i.e. the
soil is saturated) equation (5.5) becomes the standard Fellenius’s
method (Fellenius, 1936).
 93 
 

5.2.3 The Water-Uptake Model

The two-dimensional water-uptake model provided by Rees and
Ali (2006) is used here. In Cartesian form, the sink term can be
written as:

⎛ z ⎞⎛ x⎞
α (ψ )⎜⎜1 − ⎟⎟⎜⎜1 − ⎟⎟
4T
S (ψ , x, z ) = (5.6)
z r xr ⎝ z r ⎠⎝ x r ⎠

where S (ψ , x, z ) is the sink term (cm3/cm3/s), T is the potential

transpiration rate (cm/s), xr is the maximum rooting horizontal
distance (cm), zr is the maximum rooting depth (cm), α (ψ ) is a
dimensionless water stress function (see Feddes et al, 1976), z is
the vertical coordinate (cm) and x is the horizontal coordinate
(cm).
Combining Richard’s equation (1931) and sink term in
equation (5.6), provides the following moisture transfer equation:

∂ψ ∂ ⎡ ∂ψ ⎤
C (ψ ) = ⎢ K (ψ ) +
∂t ∂x ⎣ ∂x ⎥⎦
(5.7)
∂ ⎡ ∂ψ ⎤ ∂K (ψ )
⎢ K (ψ ) + − S (ψ , x, z )
∂z ⎣ ∂z ⎥⎦ ∂z

A solution of equation (5.7) is obtained via a finite element spatial

discretisation procedure and a finite difference time-stepping
scheme. In particular, adopting a Galerkin weighted residual
approach yields:
94 
 

∂N s ∂N r
− ∫K ψ s ∂Ω e −
Ωe
∂x ∂x
∂N r ∂N s ∂K e
∫K ∂z ∂z
ψ s ∂Ω e − ∫ N r
∂z
∂Ω
Ωe Ωe (5.8)
+ ∫ N r λ∂Γ − ∫ N r S ( x, z )∂Ω − e

Γ Ωe

∂ψ s
∫N r N sC
∂t
∂Ω e = 0
Ωe

Using, Green’s formula and introducing boundary terms leads to

the final disctretised form:

ψs
Kψ s + C +J +S =0 (5.9)
∂t
Where

⎡ ∂N s ∂N r ⎤
⎢K ⋅ +
∂x ∂x ⎥ e
m
K = ∑ ∫⎢ ⎥∂Ω (5.10)
e =1 Ω e ⎢ ∂N s ∂N r ⎥
K ⋅
⎢⎣ ∂z ∂z ⎥⎦
m
C = ∑ ∫ [N r N s C ]∂Ω e (5.11)
e =1 Ω e

m
⎡ ∂K ⎤ e m
J = ∑ ∫ ⎢N x ⎥∂Ω − ∑ ∫ [N x λ ]∂Γ e (5.12)
e =1 Ω e ⎣ ∂x ⎦ e =1 Γ e

m
S = ∑ ∫ [N x S ( x, z )]∂Ω e (5.13)
e =1 Ω e
 95 
 

The time dependent nature of equation (9) is dealt with via a mid-
interval backward difference technique, yielding:

⎡ψ n +1 − ψ n ⎤
K n +1 2ψ n +1 + C n +1 2 ⎢ ⎥+
⎣ Δt ⎦ (5.14)
n +1 2 n +1 2
J +S =0
This finite element spatial discretisation procedure and a finite
difference time-stepping scheme has been coded in FORTRAN
and used throughout the simulation in this study. Further detail of
the water-uptake model can be found in Rees and Ali (2006).

5.3 Case Study

5.3.1 Slope Geometry and Slip Surface

For the purposes of this test case, a typical 1 in 2 slope has been
considered. In this study, location of the critical slip surface has
been determined using SLOPEW – employing 147 trial surfaces.
For stability analysis, the slope has been divided into eight slices,
numbered from 1 at the toe, to 8 at the crest of the slope. Figure
5.1 shows the geometry of the slope, the position of each slice and
the location of the critical slip surface.
5.3.2 Soil Properties

The soil chosen for consideration here follows on from the work of
Rees and Ali (2006). In particular, the behavior of Boulder Clay is
considered. The relevant shear strength properties of Boulder clay
are given in Table 5.1.
96 
 

o
 

 
8 350cm
7
  6
5

27°cm
4 8
  750cm 3

500cm
1 2
7
  1 6
250cm

2 5
3 4
 
Origin
1600cm
 

Figure 5.1 Test Slope Geometry

Table 5.1 Material Properties (Bishop et al, 1960)

Soil Type γ c′ φ′ φb
3
(kN/m ) (kPa) o
() (o )
Boulder Clay 22 9.6 27.3 21.7

The water retention curve and the hydraulic conductivity

relationship for this material are shown in Figure 5.2 and 5.3
respectively. The figure also shows measured data for three other
soils as references: typical Sand, Kimmeridge Clay and typical
Loam (Rees, 1990). A comparison of results would appear to
suggest that the assumed relationship for Boulder Clay is within
the range of previously published data for this soil type.
The water-uptake problem considered here represents drying of
the soil from a near-saturated initial state. Hysteresis is not
considered.
 97 
 

1.0E+05 Boulder Clay

Sand
Capillary Potential (negative value, cm)

Kimmeridge Clay
1.0E+04 Loam

1.0E+03

1.0E+02

1.0E+01

1.0E+00
0.05 0.20 0.35 0.50 0.65
Volumetric Water Content (%)

Figure 5.2 Water Retention Curve for Boulder Clay

Volumetric Water Content (%)

0.05 0.15 0.25 0.35 0.45 0.55 0.65

1.0E-05

1.0E-06
Hydraulic Conductivity (cm/s)

Boulder Clay
Kimmeridge Clay
1.0E-07
Silt

1.0E-08

1.0E-09

1.0E-10

1.0E-11

Figure 5.3 Hydraulic Conductivity for Boulder Clay

98 
 

Numerical representation
In the first instance the case study presented here considers the
effect of a mature lime tree located near the toe of the
embankment. Figure 5.4 shows a diagrammatic representation of
the tree, the extent of the root zone and the finite element mesh.

700 elements
Mature Tree
2231 nodes

5m

5m
2m

16m

Figure 5.4 Finite Element Mesh

The root zone is assumed to extend a depth of 2 m and a radial

distance of 5 m both left and right direction (Biddle, 1998). The
mesh consists of seven hundred, eight-node isoparametric elements
with 2231 nodes. The mesh was configured to offer some
refinement within the root zone area since this is the region where
the most significant moisture content variations were expected to
occur.
The simulated period covered a spring/summer soil-drying
phase of 9 months (270 days). The simulation employed a time-
step size of 21600 seconds, which was held constant for the entire
period considered. A transpiration rate of 5 mm/day is used forthis
tree (Biddle, 1998). In this case, water extraction is assumed to
 99 
 

take place at its potential rate. Therefore α (ψ ) in equation (6) has

been taken as unity.
A uniform initial value of capillary potential of – 20 cm is
assumed to apply throughout the domain. This indicates that this
soil is close to field capacity (Biddle, 1998). The drying phase was
represented via the application of the above transpiration rate
within the root-water uptake model. The surface boundary, the
lower boundary and the far-field vertical boundary were
unconstrained (natural) throughout the simulation.
The results of the drying phase water-uptake simulation were
first determined via the solution of equation (5.7). The suction
values this obtained were then fed into a slope stability analysis
based on equation (5.5).

5.4 Results and Discussion

Figure 5.5 summarizes the results of the water-uptake simulation.

The figure shows simulated contours in terms of capillary potential
after a 270 day drying period.
The numerical solution yields raw output in terms of capillary
potential. These values have been converted to matric suction for
each node at the centre of the base of each slice. The resulting
suctions have then been employed in equation (5) to calculate the
FOS of the slope.
Although the active root zone of the tree lies below the start of
the slope, Figure 5.5 reveals that suctions have been generated
within on the lower section of the slope itself. The water-uptake
model, only applies the sink (extraction) within the pre-defined
geometry of the root zone. However, moisture is free to migrate
towards this zone from the surrounding soil. Hence, a ‘drawdown’
of moisture from the slope can be expected.
100 
 

8
7
6
5
4
3
1 2

5m
12.311.1
9.9
Origin line

8.8
6.5
4.2
2.0

6m
7.5 m
10 m
16 m

Figure 5.5 Matric Suction (kPa) contours at 270 days

Figure 5.6 shows the changes of matric suction at nodes located

at the centre of the base of some selected slices (see Figure 5.3).
The most significant changes in matric suction occur near to the
centre of tree (i.e. slice no 1). This effect diminishes as the distance
from the centre of tree increases (slice no 2 > slice no 4 > slice no
8). For ease of interpretation, Figure 5.7 also shows the results in
terms of volumetric moisture content at these slices.
Figure 5.8 shows the corresponding changes in the FOS
computed at various times during the drying period. This figure
shows that FOS varies with time and increases with matric suction.
Further work has also been undertaken to examine the effect of
changing the position of the tree in relation to the slope.
Presentation of the detailed results is not possible within this
format of publication. However, a brief summary of the overall
result is presented for two cases; 1) tree located at mid-slope, and
2) tree located at crest of slope. Based on these variations, two new
water-uptake simulations have been performed. All other
characteristics of these simulations remain as described above.
 101 
 

12.0

Slice 1
10.0
Slice 2
Slice 4
Matric Suction (kPa)

8.0 Slice 8

6.0

4.0

2.0

0.0
0 50 100 150 200 250 300
Days

Figure 5.6 Matric Suction (kPa) at the base of selected slices (Refer to
Figure 5.3)

4.0E-01
Volumetric Moisture Content (%)

3.8E-01

3.6E-01

3.4E-01

3.2E-01 Slice 1
Slice 2
3.0E-01 Slice 4
Slice 8
2.8E-01

2.6E-01
0 50 100 150 200 250 300
Days

Figure 5.7 Volumetric moisture content (%) at the base of selected slices
(Refer to Figure 5.3)
102 
 

2.96

2.94
FOS

2.92
FOS

2.90

2.88

2.86

2.84
0 50 100 150 200 250 300
Days

Figure 5.8 Variation of FOS with time

Table 5.2 shows a comparison of the resulting FOS calculated at

the end of each simulation. The table also indicates the percentage
difference in the FOS as compared to the saturated base-line. The
coordinates are based on the origin at lower left corner of the
model (see Figure 5.1).
This comparison indicates that changes in matric suction
considered here can result in a variation in the FOS against slope
failure of between 5 % and 7 %. This result is indicative of one
particular possible benefit of root water uptake. It should be noted,
however, that this variation in FOS arises as a result of only one
specific aspect of the problem – i.e. suction induced variation in
shear strength.
These results are independent of related vegetation effects e.g.
weight of vegetation, windthrow, root strength etc. and must
therefore be treated only as a one component of the overall
problem. In addition, it is evident that vegetation may induce
much higher suctions than those considered here. Further work
 103 
 

therefore needed to validate the approach presented and to set-it in

a more general context.

Table 5.2 Comparison of FOS at various conditions

Conditions Description FOS Percentage

difference (%)
Fully Saturated Trial slope with no 2.74 0
tree water uptake

Tree near toe Position of tree; 2.95 7.7

x = 6 m, y = 2.5m
Tree at mid- Position of tree; 2.89 5.47
slope x = 10 m, y = 4 m
Tree near crest Position of tree; 2.88 5.11
x = 12.5 m, y = 5.0 m

References

Biddle, P. G. (1998). Tree Root Damage to Buildings.

Willowmead Publishing Ltd, Wantage.
Bishop, A.W. (1955). The use of the slip circle in the stability
analysis of earth slopes. Geotechnique. 5(1)7 – 17.
Bishop, A. W., Alphan I., Blight, G. E. and Donald, I. B. (1960).
Factors Controlling the Shear Strength of Partly Saturated
Cohesive Soils. ASCE, Colorado. 503 – 532.
Fellenius, W. (1936). Calculation of the Stability of Earth Dams.
Trans. 2nd Int. Cong. Large Dams, Washington, 445 – 459.
Fredlund, D.G., Morgenstern, N. R. and Widger, R. A. (1978). The
shear strength of unsaturated soils. Canadian Geotechnical
Journal, 15: 313 – 321.
Fredlund, D. G. and Rahardjo, H. (1993). Soil Mechanics for
Unsaturated Media. John Wiley & Sons:New York.
104 
 

GEO-SLOPE/W International Ltd. (2004). GEO–SLOPE ver 6.17

Software, Calgary Alberta. Canada.
Greenwood, J. R. (2006). SLIP4EX – A program for routine slope
stability analysis to include the effects of vegetation,
reinforcement and hydrological changes. Geotechnical and
Geological Engineering, 24:449–465.
Greenwood, J. R. BSc, Norris, J. E. and Wint, J. (2004). Assessing
the contribution of vegetation to slope stability. Geotechnical
Engineering, 157:199–207.
IKRAM. (1994). The Collapse of Block 1 and the Stability of
Blocks 2 and 3 Highland Towers Condominium. Report of the
Technical Committee, Hulu Klang, Malaysia.
Janbu, N., Bjerrum, L. and Kjaernsli, B. (1956). Soil mechanics
applied to some engineering problems. Norwegian
Geotechnical Institute Publication, 16.
Lambe, T. W. and Whitman, R. V., (1969). Soil Mechanics. Wiley,
New York, 363–365.
Morgenstern, N. R. and Price, V. E. (1965). The analysis of the
Stability of General Slip Surfaces. Geotechnique, 15:79 – 93.
Rees, S. W. (1990) Seasonal Ground Movement Effects on Buried
Services, PhD Thesis, University of Wales, Cardiff.
Rees, S. W. and Ali, N. (2006). Seasonal water uptake near trees:
A numerical and experimental study. Geomechanics and
Geoengineering: An International Journal, 1(2):129-138.
Renaud, J. P., Anderson, M. G., Wilkinson, P. L., Lloyd, D. M.
and Wood, D. M. (2003). The importance of visualisation of
results from slope stability. Geotechnical Engineering, 156
(1):27–33.
Richards, L. A. (1931). Capillary conduction of liquids in porous
media. Physics, 1:318 – 333.
Ridley, A., Ginnity, M. and Vaughan, P. (2004). Role of pore
water pressures in embankment stability. Geotechnical
Engineering. 157:193–198.
Simon, A. and Collison, A. J. (2004). Quantifying the mechanical
and hydrologic effects of Riparian vegetation on streambank
stability. Earth Surface Processes and Landforms. 27:527–546.
 105 
 

Simon, A., Curini, A., Darby, S. E. and Langendoen, E. J. (2000).

Bank and near-bank processes in an incised channel.
Geomorphology, 35:193–217.
Spencer, E. (1967). A Method of Analysis of the Stability of
Embankments Assuming Parallel Interslice Forces.
Geotechnique. 17:11 – 26.
Terzaghi, K. (1936). The Shear Resistence of Saturated Soils.
Proc. Conf. Soil Mech. Found. Eng. Cambridge. 54 – 56.
Thorne, C. R. (1990). Effects of vegetation on riverbank erosion
and stability. in Vegetation and Erosion: Processes and
Environments, John Wiley and Sons, 125–144.

 
CHAPTER 6

GEOSYNTHETICS REINFORCED
RETAINING STRUCTURES
Nurly Gofar
Universti Teknologi Malaysia

6.1 Introduction

Geosynthetics are being used in Malaysia for a wide range of

applicatios e.g. slope stabilization, construction of retaining
structures, bridge abutment walls and embankments. As
deformable material, geosynthetics have the effect of not only
increasing the strength and ductility of soil, but also creating a
more flexible structure. In the construction of geosynthetic
reinforced soil structures, sucessive layers of free draining soil are
compacted between sheets of reinforcement. This procedure results
in a stable composite structure that can extend to significant height.
Such structures can undergo fairly large deformation without
catastrophic collapse and often without their serviceability be
affected (Bell et al., 1983). From a mechanical standpoint,
reinforcing soil provide the benefit of stiffening earthwork
structures without increasing their mass.
Research showed that due to their drainage capability,
geosynthetics can be used as reinforcement for cohesionless as
well as cohesive soil. The drainage capability of geosynthetics in
reinforced soil structures helps reducing the pore-pressure built up
at the interface between soil and reinforcement, thus improving the
interface shear strength (Fourie and Fabian, 1987). Further
investigation (Fabian, 1990) confirmed that non-woven needle
punched geotextile used as reinforcement can effectively drain the
clay backfill, thus helps its consolidation and results in increase of
106
 107

undrained shear strength. Furthermore, because geotextiles are

highly permeable and have drainage capability, when they are used
as reinforcement of clayey soil, they can help accelerate the fill
consolidation and increase its shear strength (Gofar, 1995).
Soil–geosynthetics interface plays an important role in the
reinforced structures. Most laboratory studies dealt with stress
transfer mechanism at the interface between the soil and the
inclusions. Direct shear and pullout tests were performed for this
purpose. If the reinforcement is in the form of a sheet that
completely separates the soil above and below it, then the transfer
mechanism is strictly friction. Hence the interface resistance can be
readily determined by the direct shear test. However, if the
reinforcement contains a large number of transverse elements such
as geogrids, then the transfer mechanism would be the combination
of friction and passive resistance and should be determined by
pullout test. Pullout of reinforcing strip is a three dimensional
phenomenon in which the soil dilatancy plays a major role.
Field studies were focused mostly on the magnitude and
distribution of lateral stresses in the soil fill and the tensile stresses
developed in the reinforcement. The magnitude and distribution of
lateral force in the reinforcement are affected by factors such as the
construction procedure and the stiffness of the foundation soil.
Staged or combined incremental and immediate loading using strut
is considered effective to reduce the post construction lateral
deformation of the wall, especially on softer foundation, by
allowing partial active yielding to take place during construction.
The presence of water is not a problem, in reinforced soil structure
since the reinforcement has high permeability and good drainage
capability.
Since geosynthetics are made of polymeric material, creep is
potential problem for their long term performance as
reinforcement. However, research results suggest that problem
might be less critical than anticipated because for geotextile, the
soil confinement at the soil-fabric interface has the effect of
increasing the tensile modulus of the inclusions.
108

6.2 Mechanism of Reinforcement

6.2.1 Strength Increase

The basic mechanism of reinforced soil involves the generation of

frictional forces at the soil-reinforcement interface. These forces
induce an increase of confining pressure in the soil in the direction
parallel to the reinforcement, thus restricting the lateral strains of
the soil. This can be referred as an apparent anisiotropic cohesion
of the composite material (Figure 6.1). As strain increases, the
frictional resistance between soil and geosynthetics is mobilized,
thus the soil stress deviator does not increase as much as it would
without reinforcement and the apparent shear strength and axial
stiffness of the soil are increased. The magnitude of strain required
to develop the interface friction can be observed on the stress-
strain curve when the stiffness of the reinforced soil begin to differ
from the stiffness of unreinforced one.
n
Reinforced soil
gives apparent φreinforced
increase in φ
h
φunreinforced

Apparent
cohesion
Ka σv Ko σv σv

Δ σr = -T(n/h)

Figure 6.1: Basic mechanism of soil reinforcement and failure strength

envelope (After Schosser and de Buhan, 1990)
 109

The peak strength of the reinforced specimens was obtained for

ratios of the minor to major boundary stresses less than the
coefficient of earth pressure at active yielding state of unreinforced
specimens (Figure 6.1). These observations were explained
qualitatively using the concept of “enhanced confinement”. When
tensile forces are induced in the inclusion by the soil deformation
towards active yielding, the reinforcement tends to restrict this
lateral (or radial) deformation of the specimen and its yielding.
This results in apparent confining pressure at failure less than
given by the active state of earth pressure. These effects were more
pronounced when the number of reinforcement layers was
increased. Investigations by Fabian (1990) and Ling and Tatsuoka
(1993) have shown that at the same level of deformation, the
strength ratio between geotextile reinforced wall and the
unreinforced wall was about 1.8.
The response of geosynthetics reinforced structures was
modeled by finite element formulations in which the reinforced
soil system is represented either by a composite material
(composite formulation) or by modeling individual component
(discrete formulation). The discrete formulation requires more
refined meshes be used since each component and interface must
be presented, yet it provides an assessment of stress and strains of
each element in the structure as well as the localized deformation
near the edges of the reinforced soil mass. On the other hand, the
composite formulation allows only the assessment of boundary
deformation. Composite formulations are developed based on an
extension of continuum concepts to a macro level of observation
whereby the reinforced soil mass is treated as an equivalent
anisotropic, homogeneous material.

6.2.2 Soil-Reinforcement Interaction

The interface mechanism in reinforced soil is influenced by the

degree of irregularity that exist along the interface (geometric
pattern and surface properties of the inclusions) as well as grain
size and frictional characteristics of the fill material. It is also
110

affected by the initial state of compaction of the soil fill and

drainage capability of the fabric.
Depending on the type of geosynthetics inclusions, stresses are
transferred from the soil to the reinforcement by the friction and/or
passive resistance (Figure 6.2). Friction occurs when the stresses
are transferred through shearing along the interface. Passive
resistance exists when the stresses are transferred by bearing of the
transverse inclusions elements against the soil. The frictional
mechanism is dominant for linear and planar reinforcement such as
steel strips, geotextiles and the longitudinal bars of geogrids. This
mechanism is usually representated using an interface friction
angle and adhesion. Passive resistance is the dominant mechanism
for reinforcements containing transverse elements such as bar
mats, grids and wire meshes.
Direct shear tests are used to estimate the soil-reinforcement
interface friction angle and cohesion (Collios et al., 1980, and
Ingold, 1982). The frictional resistance is computed based on the
average stress required to produce sliding of the soil against
reinforcement under a given applied normal stress. Collios et al.
(1980) introduced a concept of contact efficiency for the interface
resistance developed between soil and reinforcement, defined as

Ec = (cα/c) 100 %
Eφ = (tanδ/tanφ) 100 % (6.1)

where Ec and Eφ are the efficiency with respect to cohesion and

friction respectively, cα is the interface adhesion of soil to
geosynthetics, c is the cohesion of the soil, δ is the interface
friction angle of soil to geosynthetics, and φ is the internal friction
angle of the soil. The procedure of the direct shear test for
geosynthtics reinforced soil is standardized in ASTM D5321-08.
 111

Normal pressure

Pullout
force

Frictional force Normal pressure

(a) Pure friction (Direct shear test)

Passive
resistance
Frictional
resistance
Pullout
force

(b) Friction and passive resistance (pullout test)

Frictional
resistance
Pullout
force Passive
resistance

(c) Friction and passive resistance (pullout test)

Figure 6.2: Stress transfer mechanism at the soil-reinforcement interface

(after Christopher et al., 1989)
112

The anchorage resistance of soil reinforcing inclusions can also

be modeled using pullout testing. The test is performed by pulling
the geosynthetic inclusion at a controlled rate of displacement until
a peak of the pullout resistance is reached. The pullout resistance is
obviously dependent on the normal force applied to the
surrounding soil that mobilizes shear forces on both sides of the
inclusions. Measurement of deformations during the pullout test
does not allow in general the determination of the tensile force
distribution along the inclusion unless the displacement along the
inclusion is also monitored. In this test, the, movement of the
geosynthetic is due to its stretching and progressive mobilization
of the interface strength.
It was observed (Kate et al., 1988) that under given normal
stresses, the failure shear stress and friction coefficient of a
sand-fabric interfaces obtained from pullout testing are less than
those obtained from modified direct shear tests. On the other hand,
the anchorage or pullout resistance of geogrids can far exceed their
interface direct shear strength (Koerner, et al., 1989). This is
because the pullout resistance of geogrids is the combination of the
resistance to shearing along the top and bottom surfaces of
longitudinal and transverse ribs and the passive resistance
mobilized against the front face transverse ribs, where the soil can
locally reach a passive yielding state. A detailed explanation of the
interface mechanism for geogrid reinforced soil was proposed
Jewell et al., (1984). The standard procedure for the pullout test is
in ASTM D6706-01 (2007).
Interaction between the soil and the reinforcement in the finite
element formulation is modeled using interface elements.
According to Rowe and Ho (1988), any modeling of the interface
behavior must consider the possible failure mechanisms as
observed in the direct shear or pullout tests. If the reinforcement is
in the form of a sheet that completely separates the soil above and
below reinforcement, the interface resistance can be readily
determined by direct shear test’s results (Rowe et al, 1985). In this
case, provision for slip at the interface is the same irrespective of
the mechanism of failure (i.e., direct shear or pullout). However if
 113

the reinforcement consists of geogrids (with large openings as

compared to the grain size of the soil) or reinforcing strips, then
special care is required to correctly model the failure mechanism.
For strip reinforcement, independent movement of the soil above
and below the plane of reinforcement can only occur during direct
shear mode of failure.

6.2.3 Mobilization of Lateral Earth Pressure

Fundamental to soil reinforcing in retaining structures is the role

played by reinforcement with regard to soil yielding and
mobilization of lateral earth pressures. Laboratory tests on the
effect of reinforcement on lateral earth pressure were conducted
using the triaxial apparatus (Holtz et al., 1982, and Wu, 1989) for a
variety of tensile reinforcements ranging from flexible and
extensible geosynthetics to stiffer metal inclusions. The specimen
was subjected to axial compression perpendicular to the
reinforcement layers. In addition, scale models were developed for
unreinforced and reinforced sand-backfilled structures (McGown
et al., 1988).
Field observations performed on geosynthetic reinforced
structures mostly dealt with the magnitude and distribution of
lateral earth pressures developed in the reinforced backfill and
stress transferred from soil to geosynthetics. The confinement
pressure resulted from lateral expansion of soil as a consequence of
the yielding of the structure. The movement of the wall generates
tension in geotextiles and the interaction between the soil and the
geotextiles induces a redistribution of compressive stress in the soil
mass and reduces the strain in the soil system.
The tensile stress in geosynthetics is obtained by measuring the
strain developed along reinforcement layers and backcalculating
them using the tensile modulus of the geosynthetics. Distribution
of tensile strain showed its maximum at some distance from the
front edge and decreases along the anchorage zone (e.g., Fannin &
Hermann, 1990, and Balzer et. al., 1990). The locus of maximum
114

tensile stress along a reinforcement layer is assumed to represent a

critical failure surface.
The most prominent factor influencing the distribution of strain
in polymeric reinforcement is the facing type (Bathurst et al,
1987). Their investigations suggested that wall with rigid facing
panels may be subjected to additional vertical loads due to the
settlement of the reinforced fill. This has been accounted for in the
numerical study performed by Gofar (1994). Observation the
influence of facing elements by Tatsuoka et al. (1986) and Gofar
(1994) showed that the use of continuous rigid facing reduces
deformation.
Construction procedures including fill placement and
compaction force as well as connections between the reinforcing
elements and the facing are important factors in geosynthetics
reinforced walls (Richardson & Behr, 1988, Gofar, 1994). For
extensible reinforcement, it is particularly important to utilize
incremental construction methods to allow for the lateral strain to
develop during construction, so that post-construction lateral
stresses are minimized. Chou, et al. (1993) showed that much
larger lateral wall displacements are developed under service load
when the wall is built on soft clay foundation.

6.3 Design Approach

Optimum design of a reinforced wall structure should satisfy two

conflicting requirements, i.e.: (1) geosynthetics have to develop
high tensile strain to mobilize its confinement effect, and (2) the
wall movement has to be limited to ensure satisfactory
serviceability of the structure. The interaction between the tensile
reinforcement and the soil fill depends on the strain developed
within the structure. The reinforced fill must undergo sufficient
deformation to mobilize its shear strength and the necessary
interface interaction with the reinforcement. On the other hand,
the serviceability criterion requires fairly low design tensile force
adopted in order to maintain the deformation level.
 115

The design analyses of reinforced soil structures are usually

based on modified limit equilibrium approaches accrued from the
analysis of conventional cantilever or gravity walls or unreinforced
slopes. The methods are presented in a number of textbook (e.g
Craig (2004), Gofar & Kassim (2005) and Koerner (2005)).
Reinforced soil walls design requires the determination of the
geometric and reinforcement characteristics to prevent external and
internal failure. There are two criteria of stability: (1) the
geometric and reinforcement requirements to prevent external
failure, and (2) reinforcement resistance and length necessary to
prevent internal or local failure.
The external stability analysis of the structure can be verified by
considering a rigid gravity structure. As with classical unreinforced
retaining structure, the external stability of reinforced soil
structures is verified with respect to sliding on the base,
overturning, bearing capacity failure of the foundation soil and
deep-seated slope instability (rotation slip-surface of slip along a
plane of weakness) (Figure 6.3). In such analysis, the reinforced
soil structure is considered as a rigid body. The stability analysis
for external failure can be referred to any textbook in Geotechnical
Engineering. Because of the flexibility of the reinforced soil
structure, the suggested factor of safety for external failure is lower
than those used for reinforced concrete cantilever or gravity walls,
i.e., 2.0 and 1.5 for overturning and sliding along the base
respectively.
For internal stability, the local equilibrium for each soil layer
around an element of reinforcement and the overall equilibrium of
the wedges of reinforced soil are considered. The local stability
analysis includes two possible modes of failure, i.e., breakage or
excessive elongation and pullout of the reinforcement. Each mode
of failure can be analyzed using the maximum tensile force
developed at the intersection with the critical slip surface inside the
reinforced fill. The length of reinforcement extending beyond this
line is the available anchorage length that resists pullout. Hence,
the internal stability analysis require a definition of an assumed
critical slip surface within the reinforced mass (Figure 6.4).
116

Figure 6.3: External failure mechanism in reinforced soil structures

(after Koerner, 2005)
 117

Tmax

To
Active
zone Resistance zone

Figure 6.4: Location of critical slip plane or locust of maximum tensile

force in reinforcement

In the current design procedure, different failure planes were

used with respect to different types of reinforcement because the
response of the reinforced structure depends on the extensibility of
the reinforcing inclusions. There is no general form of slip plane
which can be used for any types of reinforcement.
For inextensible reinforcement, a method was first introduced
by Steward (1977, revised in 1982) known as US Forest Service
Method. The method assumes that the wall is in ‘at rest’ condition
(Ko analysis) and will fail along Rankine failure plane.
For geotextile or extensible reinforcement, the failure
mechanism of a structure reinforced with extensible inclusions
resembles Coulomb's failure plane which starts at the toe and pas
through a line inclined at the angle of (45+φ/2) to the horizontal,
where φ is the angle of internal friction of the fill material. (Figure
6.5a). The method, known as Tie Back wedge, was first introduced
by Murray (1980). The movement of the wall is assumed to start at
the top of the wall generating an active state stress throughout the
reinforced wall.
Experiments performed at Laboratoire Central des ponts et
Chaussees (LCPC), France (Schlosser & Long, 1974) for
118

Reinforced Earth retaining structures on the basis of small scale

models and prototype full scale structures showed that the locus of
maximum tensile forces is essentially different from the classical
Coulomb’s failure plane. It is a curved surface that can be
approximated by a bi-linear failure plane (Figure 6.5b).
Back-calculated coefficients of lateral earth pressure varied from
Ko at the top of the walls to a value less than Ka in the lower
section of the walls. The observed shapes of the critical surfaces
suggest that the movement of the wall start at the toe by rotation
around the top. These findings led to the development of Coherent
Gravity Method for inextensible reinforcement (Schlosser, 1978).
Slight modification to this failure plane was proposed by Juran and
Schlosser (1978) and the method was followed in the textbook by
Craig (2005).
Reinforcing layers intersecting the potential failure surface are
assumed to increase the resisting force or moment based on their
tensile capacity and orientation. The tensile capacity of the
reinforcing layers is taken as the minimum of its allowable pullout
resistance behind the potential failure surface and its allowable
design strength.
The internal stability analysis includes the computation of
tensions developed in reinforcing layer. These forces should not
exceed the tensile resistance of the reinforcement and the pullout
anchorage capacity at the interface between the soil and the
reinforcement. The allowable tensile resistance in the
reinforcement (Tall) is

Tult (CRF)
Tall = (6.2)
F D FC FS

in which Tult is the ultimate strength of the reinforcement, CRF is

the creep reduction factor, and FD, FC, and FS are the reduction
factors that account for chemically and/or biological durability,
construction damage, and the uncertainty in the determination of
the reinforcement strength.
 119

a. Tie back wedge method

b. Coherent gravity method

Figure 6.5: Assumed failure planes for analysis of reinforced soil

structure
120

The allowable pullout force (Pa) at the soil reinforcement

interface should be less than the anchorage resistance:
PR
Pa ≤ r c (6.3)
FS po

where Pr is the available pullout resistance for particular type of

reinforcement, Rc is the coverage ratio, and FSpo is the prescribed
factor of safety against pullout.
The pullout resistance of the reinforcement (Pr) is mainly a
function of the type and the stiffness of the reinforcement and the
interaction mechanism at the interface (Christopher et al., 1989).
For linear reinforcement, the pullout resistance can be estimated
as:
Pr = 2 f* α σv Le (6.4)

where σv is the effective vertical stress, Le is the available

anchorage length, and a is the reinforcement effective unit.
The pullout resistance factor (f*) is the combination of friction
and passive bearing resistance at the soil-reinforcemeat interface,

f* = fq αβ + K μ∗αf (6.5)

where fq is the bearing capacity factor for embedment, αβ and αf

are structural geometric factors for passive resistance and friction
respectively, K is the ratio of actual normal stress to the vertical
stress, μ* is the apparent friction coefficient, and α is the scale
effect correction factor. The scale effect correction factor (α) is
defined as

τ av tan φm
α = = (6.6)
τp tan φ peak
where τav and τp are the average and ultimate interface lateral shear
stresses mobilized along the reinforcement, while φm and φpeak are
the average interface friction angle and peak interface friction
 121

angle mobilized along the reinforcement. The summary of the

pullout resistance for different type of reinforcing element for used
in practice is summarized in Figure 6.6.

Figure 6.6: Pullout resistance factor for different types of reinforcement

The lift thickness (Sv) for geosynthetic walls varies with the
strength of the inclusion and the maximum lateral earth pressure
developed in the backfill,
Tu Tu
Sv = = (6.7)
σ h FS (Kγb zFS )

where γb is the unit weight of the back-fill, z is the depth of the

reinforcement layer, FS is the prescribed factor of safety, and Tu is
the ultimate tensile strength. For Reinforced Earth walls, the lift
122

thickness is governed by the size of the facing panel while for

other types of reinforcement, the lift thickness can vary along the
wall height.

6.4 Analytical and Empirical Analysis of Factors Affecting the

Lateral Stress in Geosynthetic Reinforced Structure

Rational design procedures of reinforced soil structure are based

on the fundamental understanding of the interaction between soil
and geosynthetic reinforcement. As deformable material,
geosynthetics respond differently from metal strips (Reinforced
Earth) in their function as reinforcing elements. The deformability
of geosynthetics can vary over six orders of magnitude.
Theoretically the stiffness of reinforcing inclusions influences the
stress state both locally (at the interface) and globally (reinforced
mass a whole). Furthermore, construction procedure is known to
have an effect on the development of lateral earth pressure.
Therefore, a good approach for analyzing the reinforced soil
structures should consider the deformability of the geosynthetics,
interface mechanism and the construction procedure involved in
the construction of the structure. As a result, the design analysis
would allow for modulation relative to stiffness of the selected
geotextiles and the level of acceptable deformation.

6.4.1 Effect of reinforcement stiffness on lateral earth pressure

Christopher et al. (1989, 1993) proposed that the lateral earth

pressure in the reinforced soil to vary with the anticipated
movement of the wall. Therefore, the lateral earth pressure
coefficient varies as a function of the global stiffness of the wall
and the type of reinforcement as shown in Figure 6.7. The
reinforced soil mass is assumed to approach an active yielding
state at a depth of 6 m for all types of reinforcements except bar
mats and welded wire in which the earth pressure coefficient varies
from passive values to the 'at-rest' value at depth of 6 m. Wall
 123

reinforced with very flexible reinforcement such as geotextile and

woven meshes will reach an active yielding condition along its
height.

Figure 6.7: Relationship between K/Ka and the stiffness ratio for the
design of reinforced soil structures subjected to body force

6.4.2 Combined Effect of stiffness of reinforcement and interface

mechanism on lateral earth pressure

As discussed in the previous section of this chapter, the benefits

obtained from reinforcing soil originate in the generation of
frictional forces at the soil-reinforcement interface and the tensile
resistance of the reinforcement. The developed interaction
increases the soil confining pressure, thus restricting the lateral
124

strain in the soil and increasing its internal stability. Large lateral
strain in the soil and strain in the inclusions are necessary for
active yielding condition to develop.
With small strain, the soil remains close to the 'at-rest'
condition. For a soil element subjected to a uniform boundary
stress, this condition is illustrated by a Mohr circle shown in Figure
6.8a for unreinforced condition. In the reinforced soil wall,
movement is restrained by the confinement at the
soil-reinforcement interface. This effect can be represented by an
additional increment of lateral stress acting on the soil, ∆σr.
Therefore, the coefficient of lateral stress can be computed as

σ H 1 σ h + Δσ r
Kr = = (6.8)
σV σV
where σv and σh are the vertical and horizontal stresses for
unreinforced case, σH1 is the horizontal stress for reinforced soil,
and ,∆σr is the horizontal stress increment due to confinement.
This condition is illustrated in Figure 6.8b.

The confining pressure (∆σr) depends on the stiffness and the

density of the reinforcing element as well as the stress transfer
mechanism at the interface. The confining pressure for a perfectly
adherent interface is
n
Δσ r = −T (6.9)
h
where T is the maximum tensile force in the reinforcement, and
(n/h) is the density of the reinforcement (i.e., the number of
inclusion by unit height of the soil).
 125

a. Unreinforced soil

Unreinforced soil
σv τ Failure due to
active yielding
σh
φunreinforced

Ka σv Ko σv

b. Reinforced soil

Reinforced
τ φreinforced
σ

Ka σv Ko σv
Δ σr = -T(n/h)

Figure 6.8 Mohr circle for unreinforced and reinforced soil

126

The elastic solution for the relative improvement due to the

enhanced confinement is given by equation 6.10 a for plane stress
condition and equation 6.10b for plane strain condition (Bourdeau,
1991):

Δσ r 1/E
=
σ v (K o − K a ) 1 h
+
E nE r (6.10a)

Δσ r
=
( )
1 − μ 2 /E
(
σ v (K o − K a ) 1 − μ 2 )+
h
(6.10b)

E nE r

where E is the Young's modulus of the soil, Er, is the tensile

modulus of the reinforcement, and μ is the Poisson's ratio of the
soil. In terms of the ratio between, the coefficient of lateral stress
in the reinforced soil (Kr) to that in the unreinforced soil at yielding
(Ka), Equations 6.10 can be written as:

Kr E sin φ
= 1+ r (6.11a)
Ka Er + S v E

Kr
= 1+
(1− μ 2 ) sin φ Er
Ka (1− μ 2 ) Er + S v E (6.11b)

where φ is the soil friction angle, and Sv is the spacing between two
consecutive reinforcement layers (h/n).
Gofar (1994) performed finite element analysis for two different
types of uncompacted soil and six different types of reinforcement
(Table 6.1) to simplify the above equation.
 127

Table 6.1: Stiffness of the reinforcing element considered in the study

Type of reinforcing element Tensile modulus

(kN/m)
Non-woven Geotextile 25
Non woven Geotextile 118
Woven Geotextile 500
Geogrids 2000
Steel Strips 4000
Bar Mats 8000

The effect of reinforcement density on the enhanced confining

pressure was studied by varying the vertical distance between two
reinforcement layers (Sv) from 0.4 to 1.2 m. Note that the
confinement mechanism also depends on the stress transfer at the
interface between the soil and the reinforcement, therefore study
was made for different mechanism of interface mechanism. She
arrives at an empirical Equation 6.12 for normalized coefficient of
lateral earth pressure with the stiffness of reinforcement and
apparent coefficient of friction.

⎡ ⎛ ⎞ ⎤
0 .076
Kr Er
= 1.59⎢sin φ ⎜⎜ ⎟⎟ f ⎥ (6.12)
Ka ⎣ E
⎝ r + S v E ⎠ ⎦

                
6.4.3 Effect of Construction Procedure and Compaction Effort

When a reinforced soil structure is being constructed, the soil fill is

normally placed in successive lifts and compacted to optimize its
strength and compressibility properties. However, initial distortion
of the facing resulted by the fill compaction can generate
unexpected tensile form within the not be considered herein.
During compaction, the vertical stresses are temporarily
increased by an increment ∆σv’. The largest stress increments
occur at shallow depth immediately beneath the compaction
128

equipment. Broms (1971) assumes that at shallow depth, after

being compacted under Ko condition, a soil element will follow an
unloading path consecutive to the removing of the compaction
effort at the slope of 1/Ko yielding to a final horizontal stress (σhf)
(Figure 6.9a). At greater depth, where the change in stress beneath
the compaction equipment is small, the soil will not reach the same
unloading line. There, the final horizontal stress after compaction
is assumed to remain equal to the stress when the compaction
effort (σhc) was applied. The resulting pressure distribution is
shown in Figure 6.9a by curve 2 below a critical depth Zcr at which
the maximum pressure (σhrm) occurs and by line 3 above this level.
This approach was extended by Ingold (1979) to the case of
smooth walls that yields during compaction by substituting Ka for
Ko.
The result of finite element analysis (Gofar, 1994) shows that:
the lateral earth pressure coefficient in compacted fill is strongly
related to the compaction force and the applied pressure at that
point. This result agrees well with the available analytical
solution. The empirical relationship between the ratio of the lateral
earth pressure in compacted soil to the lateral earth pressure
coefficient at initial stage can be written as

Kc ⎛ Δσ hc ⎞
= 1 + 1.1⎜⎜ ⎟⎟ (6.13)
K ⎝ q + σ ovb ⎠

where Kc is the lateral stress coefficient after compaction, K is the

initial coefficient of lateral stress for the uncompacted case, σhc is
the compaction force which varies with depth (z), q is the applied
load, and σovb is the overburden pressure which also varies with
depth (z).
 129

Lateral earth pressure,

0 σ'hr

Critical A σ'hm = Ko σ’vm

depth 3
2 σ'h = 1/Ko σ’v
1

σ'h = Ko σ’v
depth below
surface of fill

b. Theoretical
Lateral earth pressure,
0 σ'hr

Critical
depth
σ'h = 1/Ko σ’v
Locus of point A

σ'h = Ko σ’v
depth below
surface of fill

a. Simplified distribution
Figure 6.9: Stress distribution in granular soil during compaction (after
Clayton et al. 1993)
130

Equation 6.13 shows that the effect of compaction is generally

limited to shallow depth, and the effect becomes less sensitive as
the surface applied load is increased. In the following
development, it will be assumed that the results represented in
Equation 6.13 obtained for unreinforced soil can also be extended
to the reinforced soil.

6.4.4 Summary

Table 6.2 summarize the coefficient of lateral earth pressure to be

used for the analysis of unreinforced and reinforced soil structures
based on the numerical study performed by Gofar (1994). The
coefficient of lateral earth pressure proposed herein could be used
in combination of the current design procedure when the effect of
the stiffness of reinforcement, interface mechanism and
compaction are to be considered.

Table 6.2: Summary of formulation lateral earth pressure coefficient for

soil structures

Case K Equation/comment
Unreinforced wall
Uncompacted Ka at yielding
Compacted ⎡ ⎛ Δσ hc ⎞⎤
Kac K ac = K a ⎢1 + 1.1⎜⎜ ⎟⎟⎥
⎣ ⎝ q + σ ovb ⎠⎦
Reinforced Wall
Uncompacted ⎡ ⎛ ⎞⎤
0 . 076
Er
Kr K r = 1.59 K a ⎢ f sin φ ⎜⎜ ⎟⎟⎥
⎣ E
⎝ r + S v E ⎠⎦
Compacted ⎡ ⎛ Δσ hc ⎞⎤
Krc K rc = K r ⎢1 + 1.1⎜⎜ ⎟⎟⎥
⎣ ⎝ q + σ ovb ⎠ ⎦

The use of the equations with the current design procedure was
validated with data obtained from full scale model test wall
 131

performed by Karchafi and Dysli (1993) and the comparison of the

predicted value and actual behavior could be referred to Gofar
(1994).

6.5 Conclusions

Most design methods for reinforced soil structure were developed

empirically and specifically for a certain type of reinforcement.
The fundamental differences in these design approaches are the
assumptions of the lateral earth pressure and the load transfer
mechanism at the soil-reinforcement interface, both are related to
the increase of confining pressure resulting from reinforcement.
The variation of lateral earth pressure coefficient for reinforced
wall depends on the deformability and the density of the
reinforcing element and the friction at the soil reinforcement
interface. In this case, the coefficient of lateral stress is larger than
Ka due to the enhancement of confining pressure. Note that the
enhancement of confining pressure at the soil-reinforcement
interface is the basic mechanism of soil reinforcement.
The variation of lateral earth pressure coefficient in
unreinforced and reinforced wall is also depends on fill
compaction. Construction procedures, including soil placement and
compaction, reduce the concentration of vertical pressure below
the loaded surface. However, initial distortion of the facing can
results from these compaction efforts. This distortion generates
higher forces transferred to the reinforcement.

References
Bonaparte, R., Holtz, R.D. and Giroud, J.P. (1987). Soil
Reinforcement Design Using Geotextiles and Geogrids. ASTM
Special Publication, 69-116.
Bourdeau, P.L. (1991). Membrane Action in a two-layer Soil
System Reinforced by Geotextile. Proc. of Geosynthetics’91
Conference, Atlanta. 1: 439-453.
132

Broms, B.B. (1971) “Lateral Pressure due to Compaction of

Cohesionless Soils”, Proc. 4th Conf. Soil Mech. Found. Eng.,
pp.373-384, Budapest.
Chou, N.N.S. and Wu, J.T.H. 1993. Effect of Foundations on the
Performance of Geosynthetics-Reinforced Soil Walls. Proc.,
Geosynthetics’93 Conf. Vancouver, Canada. pp. 189-201
Christopher, B.R. (1993). Deformation Respoase and Wall
Stiffness in Rx1ation to Reinforced Soil Wall Design. Ph. D.
Thesis, Purdue University.
Christopher, B.R. (1989) Reinforced Soil Structures, vol. 1, Design
and Construction Guidelines, Report no. FHWA-RD-89-043,
285 pp.
Clayton, C.R.I, Milititsky, J., and Woods, R.I. (1993). Earth
Pressure and Earth Retaining Structures. 2nd ed. Chapter 4.
Blackie Academic and Professional, London, pp. 63-80.
Collios, A.P., Delmas, J.P., Gourc, J.P. and Giraud, J.P. (1980).
Experiments on Soil Reinforcement with Geotextiles. Proc.
ASCE National. Conf. Use of Geotextiles for Soil
Improvements. Portland, Oregon, 53-73.
Craig, R.F. (2004). Craig’s Soil Mechanics. 7th ed. Spon Press. pp.
221-225
Duncan, J.M., Williams, G.W., Sehn, A.L., and Seed, R.B. (1991).
Estimation of Earth Pressures due to Compaction, ASCE
Journal. Geot. Eng. Div. 117(12):1833
Fabian, K. (1990) Time Dependent Behavior of
Geotextile-Reinforced Clay Walls, Proc. 4th Int. Conf.
Geotextiles, Geomembranes and Related Products, 1:33-38.
Farrag, K., Acar, Y.A., and Juran, I. (1993). Pullout Resistance of
Geogrid Reinforcements. Geotextiles and Geomembranes. 12:
133-139.
 Gofar, N and Kassim, KA. (2005) Introduction to Geotechnical
Engineering-Part I, Prentice Hall. pp. 93-99
Gofar, N. (1995). The Use of Geotextile as Reinforcement on
Embankment Constructed of Fine Grained Soil (in Bahasa
Indonesia). Majalah Sriwijaya, UNSRI. 31(2)
 133

Gofar, N. (1994). Modeling of Applied Boundary Load

Transmission in Geosynthetics Reinforced Structures, PhD.
Thesis, Purdue University, USA.
Ingold, T.S. (1979). The Effects of Compaction Upon Retaining
Walls. Geotechnique 29 (3):265-283.
Jewell, R.A., Milligan, G.W.E., Sarsby, R.W., and DuBois, D.D.
(1984). Interaction between Soil and Geogrids. Proc. Conf.
Polymer Grid Reinforcement, Thomas Telford, London, pp.
18-29.
Juran, I. and Schlosser. F. (1978) Theoretical Analysis of Failure in
Reinforced Earth Structures. Proc. Symposium on Earth
reinforcement, ASCE convention, Pittsburgh, New York. 528-
555
Kharchafi, M. and Dysli, M. (1993). Geotextiles Used as
Reinforcement - Full-Scale Tests. (In French), Research Report
No. 283, Swiss Federal Highway Administration, Soil
Mechanics Laboratory of the Swiss Federal Institute of
Technology, Lausanne, June 1993, 127 pp.
Kate, J. M. Ventakappa Rao, G. and Tyagi, S.K. (1988).
Evaluation of Soil Reinforcement Friction. Indian
Geotechnical Journal. 18(2).
Koerner, R.M. (2005), Designing with Geosynthetics, 5th ed.,
Prentice Hall, New Jersey.
Koerner, R.M., Wayne, M.H. and Carroll, R.G.jr. (1989).
Analytical Behavior of Geogrid Anchorage. Proc.
Geosynthetics '89 Conf., San Diego, CA. pp. 525-536.
Lee, K.L., Adams, B.D., and Vagneron, J.M. (1973). Reinforced
Earth Retaining Walls. J. Soil Mechanics Foundation Eng.
Div., ASCE, 99(SM10):745
Leschinsky, D., and Boedeker, R.H. 1989. Geosynthetics
Reinforced Soil Structures Journal Geotech Eng. Div., ASCE.
115 (10):1459-1479.
Ling, H.I., and Tatsuoka, F. 1993. Laboratory Evaluation of a
Non-woven Geotextile for Reinforcing on-site Soil. Proc.
Geosynthetics’93 Conf., Vancouver, Canada. pp. 533-546.
134

Mc.Gown, A., Andrawes, K.Z., and Murray, R.T. 1988. Controlled

Yielding of the Lateral Boundaries of Sod Retaining
Structures. Proc. Symp. Geosynthetics for Soil Reinforcement,
ASCE. Nashville, TN, pp. 193-210
Murray, R.T. 1980. Fabric Reinforced Earth Walls: Development
of Design Equations. Ground Engineering. 13(7)
Rowe, R.K,, and Ho, S.K. (1988). Application of Finite Element
Techniques to the Analysis of Reinforced Soil Walls.
Application of Polymeric Reinforcement in Soil Retaining
Structures, Kluwer Acad. Publishers, pp. 541-553.
Rowe, R.K., Ho, S.K., and Fisher, D. (1985). Determination of
Soil-Geotextile Interface Strength Properties. Proc. 2nd.
Canadian Symp. Geotextile, Montreal, Canada, pp. 24-34.
Schlosser, F. (1978). La Terre Armee, Historique Development,
Actuel and Futur, Proc. Symp. Soil Reinforcement and
Stabilizing Techniques, NSWSIT/NSW, 5-28.
Schlosser, F. and Elm, V. (1978). Friction in Reinforced Earth,
Proc. Symp. Earth Reinforcement, ASCE, Pittsburgh,
Schlosser, F. and Long, N.T. (1974). Recent Results in French
Research on Reinforced Earth. ASCE Journal of Const. Div.
100(C03): 223-237
Tatsuoka, F., Ando, H., Iwasaki, K, and Nakamura, K. 1986.
Performances of Clay Test Embankments Reinforced with a
Non-woven Geotextile. Proc. 3rd Int. Conf. Geotextiles,
Vienna, Austria. pp. 355-360.
Tensar Corp. 1986. Guidelines for the Design and Construction of
Reinforced Soil Retaining Wall using Tensar Geogrids, Netlon
Ltd.
CHAPTER 7

Reinforcement Mechanisms of Rock Bolt

 

Mohd For Mohd Amin, Khoo Kai Siang & Chai Hui
Chon
Universiti Teknologi Malaysia
 

7.0 Introduction

Failure in rock mass is implied as the incapability of a rock mass to

support its own weight. Installation of stabilising methods ensures
the inherent strength of the rock is maintained before excessive
failure takes place. The effectiveness of any stabilising method
usually depends on the type of instability in rock and the
stabilising mechanisms of the selected method. Various methods
are currently available for stabilising unstable rock and usually
more than one method of stabilisation are adopted to achieve the
required stability.
Methods for rock stabilisation are divided into two types
namely, support systems and reinforcement systems (Windsor and
Thompson, 1993). The former includes shotcrete and wire-mesh in
which, stabilising elements are installed on the rock surface. The
latter comprises reinforcing elements installed in the rock that
includes rock bolts and dowels. A rock bolt is a steel bar, which is
inserted into a hole drilled in the rock. Despite of its many
varieties, all rock bolts have in common the following elements: a
steel bar (shank), an anchoring device (resin or grout) at one end,
and a tensioning device (bearing plate and nut) at the other, as
shown in Figure 7.1 (Brady and Brown, 1985). The reinforcement
bar helps to mobilise the inherent strength of the rock mass by
modifying its internal strength and deformation characteristics.
However, the effectiveness of the bolt in mobilising the inherent
135
 
136
 

strength of the rock mass depends on several parameters. The

presence laboratory investigation is aimed at verifying the effect of
bolt inclination, type of bonding (full-bonded length and point-
anchored) and level of tension on the performance of a rock bolt.

REINFORCEMENT
ELEMENT: steel rebar

LOCKING BONDING AGENT:

NUT resin or grout

BEARING PLATE

Figure 7.1: Full-bonded rock bolt (after Brown and Brady, 1985)

7.2 Reinforcement Mechanisms of Rock Bolt

Improvement in the installation procedures and the associated

hard-wares (e.g. fast setting resin and corrosion protection) have
made rock bolt as one of the most popular permanent
reinforcement for rock engineering structures (Schubert and
Schubert, 1993). However, its effectiveness depends on several
factors as briefly discussed in the following sections.

7.2.1 Stabilising Effect of Rock Bolt

Rock bolt is most effective for stabilising planes of weakness in

rock, as it helps to mobilise the inherent strength of these unstable
planes. The reinforcement mechanisms of rock bolt include the
creation of radial compression, composition action and mobilising
friction resistance (Windsor and Thompson, 1993).
 137
 

Lets consider instability induces along a joint plane. The shear

strength and dilatation versus shear displacement for this joint can
be presented as Figure 7.2. As shown in this Figure, it is beneficial
to prevent slips between joint blocks before shear strength reaches
its residual stage. If joint reaches its residual strength (after certain
amount of shear displacement), the rock system then becomes less
capable of supporting itself and consequently, will require a
proportionally larger amount of reinforcement. This implies the
most important rule in rock bolting, i.e. the system should be
installed as soon as possible after excavation in order to prevent
the loss of an appreciable portion of the inherent strength of the
rock mass.
The importance of dilatation effect (denoted as δn in Figure
7.2) is related to the phenomenon of the build up of the peak shear
strength with shear displacement. Dilatation results from the
uneven nature of joint surface (Mohd Amin et al. 2001). In order
for the joint surface to ride over any asperities/undulations (along
its shearing path), work has to be done against the force acting
normally to the joint. The higher the force is, the more difficult it is
for shearing process to take place. This is due to the law governing
the friction acting between sliding joint surfaces:
τ = σn tan (φ + i) + c (7.1)
Where, τ is shear strength of joint; σn is normal stress acting on
joint; φ is angle of friction; i is angle of dilatation, and is c
cohesion of joint.
If dilatation is absolutely prevented, it becomes impossible for
shearing to occur along a joint plane and the shearing force must
be sufficiently high to produce crushing of the surface
irregularities. Thus, the prevention of dilatation, which if
permitted, removes all constraints on sliding. It can be seen in
Figure 7.2 that the maximum rate of dilatation is associated with
the build up to peak strength. This implies that it is better to
reinforce the joint before initial displacements and consequent
large dilatations have occurred. The shear behaviour of a joint
highlighted above can be used to appreciate the mechanisms of
reinforcement of a rock bolt as outlined below.
138
 

7.2.2 Bolt Inclination

The orientation of bolt to the joint plane has a pronounced

effect on shear resistance offered by rock bolt. The stabilising
action of a tensioned (active) bolt intersecting a joint can be
explained using Figure 7.3.

Shear strength, τ Peak shear strength

Residual strength

Normal
displacement, δn
Dilatation

Shear displacement, δh

Figure 7.2: Shear strength and normal displacement versus shear

displacement

tn T

tn
Tension in bolt, T Inclined bolt
th

Joint α

(a) (b)

Figure 7.3: Rock bolt intersecting a joint, (a) perpendicular and

(b) at an angle α, to the joint plane.
 139
 

Figure 7.3(a) shows a bolt installed at right angle to a joint. The

pre-tension in the bolt acts like a clamping force on the joint, thus
helps to inhibit joint dilatation. Resistance to shearing is
contributed by the shear strength of the joint and shear stiffness of
the bolt. The latter depends on bolt diameter and material type
(Bjurström, 1974). In addition, perpendicular bolt does not
experience a considerable amount of tensile stress when joint is
displaced, thus the additional shear resistance from the bolt is not
possible without significant bending and yielding of the bolt
(Spang and Egger, 1990). However, if the same joint is reinforced
with a similar bolt, but at an angle α to the joint as shown in Figure
3(b), significant improvement can be expected. Haas (1981) has
noted that it is most effective if the bolt is inclined in such a way
that it elongates as joint starts to shear. In this orientation, the pre-
tension in bolt and the additional tensile stress induces into the bolt
due to shearing of the joint act in two ways. The component
tangential to the joint surface (component th) will give a direct
contribution to the shear resistance. The component normal to the
joint surface (component tn) will increase the normal stress acting
on joint and help to counteract dilatation and thus, increasing the
frictional strength of the joint. Generally, the optimum angles of
bolt inclination (α) are between 300 and 600.

7.2.3 Tensioning of Bolt

An active bolt is when the bar is tensioned between the fixed end
and the plate (see Figure 7.1). This essential feature, if properly
installed, exerts a positive compression into the rock and maintains
the interlocking of rock blocks. Effective bonding between the bolt
and surrounding rock is essential hence, rock bolt is not suitable
for weak rock. The main advantage of tensioned bolt is that it
increases shear-stiffness of the bolted joint as soon as it is installed.
Fast-setting resin permits tensioning of the bolt a few minutes after
installation (Windsor and Thompson, 1993).
The amount of tension applied on the bolt (working load, Tw)
depends on the characteristic tensile strength (fpu) of the bolt
140
 

material. Typically, for high tensile steel, Tw is about 80% of fpu.

Creep or loss in bolt tension throughout its service life is common
hence, bolt is usually tensioned at about 110% Tw or re-tensioned
whenever required. It should be noted that when bolt is pre-
tensioned, some elastic deformation of the bar is utilised.
Additional deformation may be induced into the bolt when there is
movement in the reinforced joint. It is for this reason that Tw is
lower than the tensile strength of the bolt particularly, if the
reinforced joint is known to dilate.

7.2.4 Bolt Anchorage

Point anchored rock bolt (anchored using wedge and expansion

shell) is limited to hard rocks and for use as temporary
reinforcement (Brady and Brown, 1985). Local crushing of rock at
the anchorage point and slippage of anchorage are the main
weaknesses of point anchored bolt. When cement or resin fills the
annulus around the full length of the bar as means of anchorage,
the bolt is called full-bonded bolt. The cement or resin also acts as
protection against corrosion. The continuous anchorage ensures the
bond between bar and the surrounding rock remains effective
although the tensioning end becomes ineffective. Full-bonded bolt
requires only a small amount of movement in the rock mass to
mobilise its anchorage capacity. This allows for maximum
utilisation of the strength of reinforcing element immediately after
installation (Windsor and Thompson, 1993). In general, full-
bonded bolt displays a better resistance to shearing and more
superior in terms of anchorage ‘creep’ and consequently, it is
commonly use for permanent reinforcement. The anchored length
(L) of rock bolt to maintain stability is calculated based on the
following formula (Douglas and Arthur, 1983):

L = (FTf)/(πDτult)

where, F is factor of safety (2 < F < 3); Tf is ultimate load of rock

bolt; D is fixed anchor diameter and τult is ultimate bond stress.
 141
 

7.3 Laboratory Testing

 
The main objectives of this study are to verify the effect of bolt
inclination, type of bonding (full-bonded length and point-
anchored) and level of tension on the performance of a rock bolt.
Laboratory investigation was carried out using physical model of
bolted joint blocks. Joint block of dimensions
63mm×103mm×56mm was formed by gluing several layers of
4mm thick perspex. Two sets of blocks were prepared and each set
representing the upper and lower joint blocks. 5mm diameter holes
were drilled in these blocks at an angle 900, 700 and 600. Due to the
limited length of the block, hole with inclination angle of less than
600 was not possible. 5mm diameter bamboo rods, instead of steel
rod, were used as bolts. Shear tests on the bolted joints were
conducted using portable shear box apparatus model Roctest Phi-
10. Description of this equipment is given in Mohd Amin et al.
(2000).
The use of specific materials for the physical model is mainly
due to the following reasons:
(i) Perspex instead of rock - For ease of drilling of 5mm
diameter holes at various orientations in the blocks. Cutting
of cube rock sample was relatively difficult. Bamboo instead
of steel rod – Shear strength of steel may be too high to be
tested on the Phi-10 shear box. Steel rod may induce
excessive fracturing of the drill hole walls and bonding
material particularly at the vicinity of the joint plane.
(ii) This study is mainly to verify the behaviour and not the
ultimate strength of a bolted joint therefore, material
characteristics are not essential.
The various types of shear test conducted are listed in Table 7.1.
142
 

Table 7.1: List of laboratory shear test conducted on unbolted and bolted
joint

Test Bolt Type of Normal Test description

no. inclination, anchorage Stress,
α0 MPa
1 N/A N/A 0.5 Shear test on unbolted
joint
2 N/A N/A 1.0 Shear test on unbolted
joint
3 N/A N/A 2.0 Shear test on unbolted
joint
4 90 Full length 0.5 Joint with resin bonded
bolt
5 90 Full length 1.0 Joint with resin bonded
bolt
6 90 Full length 2.0 Joint with resin bonded
bolt
7 70 Un-bonded 0.5 Joint with point-
anchored bolt
8 70 Un-bonded 1.0 Joint with point-
anchored bolt
9 70 Un-bonded 2.0 Joint with point-
anchored bolt
10 70 Full length 0.5 Joint with resin bonded
bolt
11 70 Full length 1.0 Joint with resin bonded
bolt
12 70 Full length 2.0 Joint with resin bonded
bolt
13 60 Full length 0.5 Joint with resin bonded
bolt
14 60 Full length 1.0 Joint with resin bonded
bolt
15 60 Full length 2.0 Joint with resin bonded
bolt
 

The shearing rate employed was 1.0mm/min. and the maximum

shear displacement is twice the bolt diameter. Normal stresses
applied during shear were 0.5, 1.0 and 2.0 MPa. For the full-length
bonded bolt, epoxy resin was used to bond the full length of the
 143
 

bamboo rod into the drill hole. As for point-anchored bolt, the
5mm rod was just inserted into the drill hole without any bonding.
Since the rod and the hole were of similar diameter therefore, the
rod was in rigid position once inserted into the hole. To induce
some degree of pre-tension in the full-length bonded bolt, special
procedure had been adopted. The insertion of the bolt into the
blocks of perspex was undertaken while the blocks were under
compression machine (using Matest C22 Compression Machine) at
stress level similar to the normal stress employed during shear
tests. The compressive stress on the perspex blocks was released
after the bonding agent had fully set (about 20 minutes) and
assuming that this procedure was able to induce some level of
tension in the bolt. The effect of joint dilatation on bolt was not
verified. This is mainly due to the difficulty in preparing the
typical surface roughness on the perspex model. Detailed samples
preparation and testing procedures are discussed in Khoo Kai
Siang (2003) and Chai Hui Chon (2003).

7.4 Results and Discussion

Only selected plots are discussed in this section. Series of tests

conducted on the unbolted joint blocks are mainly to verify the
shear strength of the model joint surface. The strength represents
the ‘inherent strength’ of an unbolted joint and this is used as a
basis to substantiate any improvement in strength due to bolting.
Figure 7.4 shows the behaviour of same joint reinforced with
full-bonded bolt, at different inclination angle and at normal stress
of 0.5 MPa. An increase of about 50% in strength can be observed
after bolting. The highest and the lowest shear strength of the
reinforced joint are observed at bolt inclination of 600 and 900,
respectively. Similar improvement in joint strength can be
observed in shear test at higher normal stress (Figure 7.5), although
the increment in is less significant. This is probably due to the
strength increase induced by a higher normal stress over shadows
the effect of bolting.
 
144
 

1000

800
Shear strength (kPa)

600

bolted 60 deg
400 Unbolted
bolted 90 deg
bolted 70 deg
200

0
0 1 2 3 4 5 6 7

Horizontal displacement (mm)

Figure 7.4: Shear strength of unbolted and bolted joint (900, 700 & 600
bolt orientation) at normal stress 0.5 MPa.

1800
Bolted 60 deg

1500
Shear strength (kPa)

1200 Unbolted
Bolted 90 deg
Bolted 70 deg
900

600

300

0
0 1 2 3 4 5 6 7
Horizontal displacement (mm)

Figure 7.5: Shear strength of unbolted and bolted joint (900, 700
& 600 bolt orientation) at normal stress 2.0 MPa.
 145
 

Table 7.2 summarises the effect of bolt inclination on peak shear

strength and shear displacement to reach the peak strength. Joint
with bolt inclined at 600 exhibits the highest peak strength and the
lowest shear displacement to reach this strength compared to other
bolt orientations.

Table 7.2: Peak shear strength and horizontal displacement for full-
bonded bolt installed at different orientations.

Bolt Peak shear Horiz. displc.

orientation, α strength, Mpa at peak strength, mm
900 0.81 2.30
700 0.92 2.11
600 0.98 2.03

The effect of bolt anchorage on joint strength can be seen in

Figure 7.6, at normal stress 0.5 MPa. Joint reinforced with full-
length bonded bolt exhibits a higher strength compared to joint
reinforced with un-bonded bolt. Similar behaviour is observed in
tests at higher normal stresses. Table 7.3 highlights several
advantages of full-bonded bolt. In addition to a higher joint
strength, this bolt offers less shear displacement to mobilise the
peak strength hence, immediate utilisation of the reinforcing effect.
The joint also exhibits less dilatation, as indicated by the smaller
normal displacement at peak strength.

Table 7.3: Average peak strength, horizontal and vertical displacement

for different types of Anchorage (with bolt inclination 700 and normal
stress 0.5 MPa).

Bolt anchorage Peak shear Horiz. displc. at

strength, MPa peak strength, mm
Un-bonded bolt 0.76 2.56
Full-bonded length 0.92 2.10
146
 

Finally, Figure 7.7 shows the effect of bolt orientation and tension
level on joint strength for the three level of normal stress (i.e. 0.5,
1.0 and 2.0 MPa). As pointed out earlier, bolt tension induces a
clamping effect on the joint. Consequently, tests conducted on bolt
with pre-tension and at various levels of normal stress during
shearing, may be taken to characterise this effect, i.e. component tn
in Figure 3(b). For a given bolt orientation, Figure 7.7 indicates
that bolted joints tested at higher normal stress (i.e. higher bolt
tension) consistently exhibit a higher strength. At a given normal
stress, joints with bolt installed at a smaller angle of inclination (<
900) exhibit higher shear strength. If each of the trend-lines in
Figure 7 is extended to include smaller angle of bolt inclinations,
say inclination angle of 300, a higher joint strength can be
expected.

1000

800
Shear strength (kPa)

600

400 Point anchored

Full-length bonded
200

0
0 1 2 3 4 5 6

Horizontal displacement (mm)

 

Figure 7.6: Shear strength of joint with full-length bonded and point-
anchored bolt, at normal stress 0.5MPa.
 147
 
Peak shear strength (Mpa) 2.5

2.0
Normal stress
2.0MPa
1.5
Normal stress
1.0MPa
1.0

Normal stress
0.5 0.5MPa

0.0
20 30 40 50 60 70 80 90 100

Bolt orientation, (deg)

 

Figure 7.7: Peak shear strength versus bolt orientation at normal stress
0.5, 1.0 and 2.0 MPa.

7.5 Conclusions

Laboratory test results show that the reinforcement mechanisms of

rock bolt are affected by bolt inclination, anchorage type and level
of pre-tension on the bolt. Optimum reinforcement can be obtained
if bolt is inclined at an angle so that it elongates upon joint
displacement. Full-bonded bolt is superior to point-anchored bolt
in terms of mobilising the anchorage capacity and this allows for
immediate utilisation of the reinforcing element. Pre-tensioning of
an inclined bolt induces clamping effect on joint surface
consequently, helps to reduce joint dilatation and increases the
inherent shear strength of the joint.
 

References

Brady, B.H.G. and Brown, E.T. (1985) Rock Mechanics for

Underground Mining, London, George Allen and Unwin, 527p.
148
 

Bjurstrom, S. (1974) Shear strength of hard rock joints reinforced

by grouted untensioned bolts, Proc. of the 3rd. Conf., ISRM,
Denver, 1194-1199.
Chai Hui Chow (2003) Kesan tegangan ke atas mekanisme
pengukuhan bolt batuan - suatu kajian model di makmal. B Eng
Thesis, Universiti Teknologi Malaysia, Skudai, Johor.
Douglas, T.H. and Arthur, L.J. (1983) A guide to use of rock
reinforcement in underground excavations, CIRIA Report
No.101.
Haas, C.J. (1981) Analysis of rock bolting to prevent shear
movement in fractured ground. Mining Engineering, June, 698-
704.
Khoo Kai Siang (2003) Kesan orientasi bolt ke atas mekanisme
pengukuhan bolt batuan – suatu kajian model di makmal. B Eng
Thesis, Universiti Teknologi Malaysia, Skudai, Johor.
Mohd Amin, M. F., Teo, K.B. and Mustaffar, M. (2001) Joint
surface texture measurement using close-range
rd
photogrammetry. 3 Asian Symp. on Engng. Geology and the
Environment, 3 - 9 Sept., Yogyakarta, Indonesia, 294 – 301.
Mohd Amin, M. F., Mustaffar, M. and Teo, K.B. (2000)
Preliminary Design parameters based on laboratory shear test of
core samples. Annual Confc., Geological. Soc. of Malaysia, 8-9
Sept. Pulau Pinang.
Schubert, W. and Schubert, P., (1993) Tunnels in squeezing rock:
Failure phenomena and counteractions. Int. Symp. on
Assessment and Prevention of Failure Phenomena in Rock
Engineering, Rotterdam, 479-484.
Spang, K. and Egger, P. (1990) Action of fully-grouted bolts in
jointed rock and factors of influence. Rock Mechanics and Rock
Engineering, 23(3), 201-229.
Windsor, C.R. and Thompson, A.G., (1993) Rock reinforcement -
technology, testing design and evaluation. Comprehensive Rock
Engineering, Principles, Practice and Projects, Vol. IV,
Hudson et al. (eds.), 1st. edition, London, Pergamon Press,
483p.
Chapter 8

Soil Confinement System for Slope

Rehabilitation

Ramli Nazir &Lim Kee Ley

Universiti Teknologi Malaysia

8.1 Introduction
Erosion has always been one of the major sources of early damage
to both natural and man-made slopes. The erosion is caused by
different factors such as rainfall, runoff or wind. This can be
avoided only if protection measures are properly selected
according to its cause. Rainfall erosion occurs mainly on bare
slopes, while vegetation provides the natural and best protection.
Hence, the best way to protect a slope from this kind of erosion is
to establish a dense and uniform grass cover as soon as possible.
This condition cannot be achieved at early stage of construction
because vegetations root need time to grow and to reinforced soil
particles. Strengthening of soil particle at this stage is necessary to
protect the slope from erosion. This could be achieved by the use
of soil confinement system. The system divides large surface areas
into many small areas or cells. It is purposely engineered for
trapping a layer of topsoil, thus avoiding rainfall erosion by both
sheltering the surface against the raindrops impact and by reducing
the movement of the soil particles.
The topsoil is usually of poor geotechnical characteristics and
can easily slip down along slopes with an inclination greater than
30 degree. It is also prone to erosion due to heavy or sustained
rains occurring prior to grass growth. Slope with different length,

149 
 
150 
 

inclination, soil characteristics, can be properly protected against

erosion by the choice of the most suitable kind of soil
confinements (Rimoldi and Riccluti, 1994). In this application,
soil confinements provide a structural function because the
stability of the top layer relies only on the confinement strength;
hence all possible failure mechanisms shall be properly evaluated,
in order to define the minimum resistance required for the soil
confinement.
Soil confinement become more popular, as is evidenced by
increasing number of research work such as (Rimoldi and Riccluti,
1994; Anderson and Riemer, 1994; Michael and William Outcalt,
2002; Greg Northcutt, 2003; Miller, 2003) focusing on the topic of
bearing capacity improvement, erosion control and landslides.
However, certain aspect has been overlooked such as its
configurations, anchoring system, and correlation in terms of soil
strength with and without confinement system. In terms of slope
failure rehabilitation application, the system is still in its nascent
stage especially in our region. Besides providing slope stability, an
attractive feature of different type of soil confinement system is
preserving the natural look of the slope surface.
Soil confinement system is the first stage prevention to surface
erosion. The stability of the surface layer of topsoil provided by
the confinement allows the seeds to germinate quickly and the
vegetation to uniformly covers the slope. While growing, the roots
intertwine with the soil confinement system, thus creating a natural
synthetic network with a much higher resistance to the shear force
than roots alone.
In this study, the performance of the soil confinement system
for the selected site was compared with that of without the
confinement system. The performance of existing methods may be
improved by varying the anchoring, materials, compaction degree
and type of filling soil.
 151 
 

8.2 Small Scale Model test

The arrangement of smal scale model test was designed to make

optimum use of the space available in the laboratory. Considering
the above requirement, test package was adopted as shown in
Figure 8.2. The water storage tank was connected to spray nozzle
via a water pump using PVC pipes. The collection tank of size
1.20m in breadth, 0.75m in width, and 1.05m in depth use to
accommodate test specimen, measuring container was placed
underneath of it. A 20mm diameter hole located at lowest end of
this collection tank to allow wash water together with eroded soil
flow into measuring container. Three confinement configurations
(Figure 8.2) were used in small model test. In addition non-
confinement was also tested to provide the basis for comparison
and to evaluate the effectiveness of the confinement system.

 
Water supply

Water tank

Water pump

Spray nozzle (6 numbers)

Showering water

Various slope angle,(θ°) Test specimen

Water flow to measuring container

Measuring container
 
Figure 8.1: Schematic setting up of the model test
152 
 

0.45m 0.45m

1.0m
1.0m

h
b

0.45m

1.0m

Figure 8.2: (a) Typical circular confinements. d = diameter (b) Typical

triangular confinement. b = breadth, h = height, (c): Typical square
confinement. b = breadth, h = height.
 153 
 

8.3 Results and Discussion

The first section discuss with the test results on non-confinement
slope while the latter section deals with confined slope with three
configurations of confinement of sizes 50 mm, 100 mm, and 150
mm. The effect of slope angle, effect of rainfall intensity and effect
of rainfall duration are discussed in details for both tests.

8.3.1 Non-Confinement Slope Test Results

The non-confinement slope was designed to investigate the effect

of slope angle, rainfall intensity and rainfall duration on the slope
surface without protection. Thus it will form a basis for
comparison with various confinement slopes. Factors involving the
determination of soil loss mass were observed. Slope angle, θ,
rainfall intensity, I, rainfall duration, tr, are correlated with soil loss
mass with due respect to the slope surface.

8.3.1.1 Effect of Slope Angle

Summary of test results in term of soil loss mass on non

confinement slope of different angles is presented in 2(a) to Figure
2(d). The figures exhibit the influence of slope angle for rainfall
intensity ranging from 20mm/hour to 75mm/hour with rainfall
duration of 10, 30, 50 and 60 minutes respectively. Generally, the
soil loss mass increases with an increment of slope angle at various
rainfall intensities as expected. Similar trend was observed for
slope angle less than 60 degree for rainfall intensity approximate
equivalent to 40mm/hour. However, for slope angle less than 60
degrees with rainfall intensity more than 40mm/hour, an increasing
of soil loss mass is significant. This is due to the fact that higher
rainfall intensity will create more overland flow thus breaking-up
more soil particles. The trend was expected since steeper slope will
reduce frictional force in soil particles. An increase in slope angle
will also increase the gravity effect of soil particles. The breakup
soil particles at the top of the slope will move downwards and push
154 
 

the adjacent soil particles. This is an early process of surface

erosion. The retrogressive process will occurs via transportation
element such as water.

8.3.1.2 The Effects of Rainfall Intensity

Figure 3(a) to 3(d) shows the effect of rainfall intensity on soil loss
mass. Trends of maximum soil loss mass are expected as the
graph gradient shown increase with an increment of rainfall
intensity at constant slope angle. The trends are almost similar
except Figure 3(c) and 3(d) where the acute increase in gradient
was observed for slope angle of 60 degree. As shown in Figure
3(a), for slope angle 30 degree to 60 degree, the effects of rainfall
intensity is minor when the rainfall intensity is approximate
equivalent to 50mm/hour. However, there is evidence that as the
rainfall intensity increases to 60mm/hour and 75mm/hour, the
effect is more significant since further increase in soil loss mass is
clearly evident for all range of slope angle. It can be seen that the
rainfall intensity has significant effect on the soil loss mass. The
effect is getting more pronounced as the slope angle increases.

8.3.1.3 Effect of Rainfall Duration

Figure 4(a) to Figure 4(d) shows the relationship between soil loss
mass and rainfall duration ranging 10 minutes to 60 minutes and
rainfall intensity of 20 mm/hour to 75 mm/hour. Generally the
maximum soil loss mass increases with increasing rainfall duration
as expected. It can be seen in Figure 4(a) and Figure 4(b), that the
slope subject to rainfall, from first 10 minutes shows constant rate
of increasing in soil loss mass for all range of rainfall durations
with rainfall intensity 20mm/hour and 30mm/hour. While a sudden
increase in variation of soil loss mass especially from rainfall
duration approximate equivalent to 30 minutes with rainfall
intensity 40mm/hour to 75mm/hour. Similar trend were observed
for non-confinement slopes as shown in Figure 4(c) and Figure
 155 
 

4(d). Therefore, rainfall duration is significant when (i) rainfall

duration more than 30 minutes with rainfall intensity 20mm/hour
and 30mm/hour for all range slope angle; (ii) rainfall duration
more than 10 minutes with rainfall intensity more than 40mm/hour
for all range slope angle.

Soil Loss Mass Vs Slope Angle For Non-Confinement

(t r = 10 minutes)

100
Soil Loss Mass (g)

20mm/hour
80
30mm/hour
60 40mm/hour
50mm/hour
40
60mm/hour
20 75mm/hour

0
20 30 40 50 60
Slope Angle (Degree)

Figure 2(a): Variation of Soil loss mass between Non Confinements and various
slope angle for various rainfall intensity at rainfall duration 10 minutes.
Soil Loss Mass Vs Slope Angle For Non-Confinement
(t r = 30 minutes)

250
Soil Loss Mass (g)

20mm/hour
200
30mm/hour
150 40mm/hour
50mm/hour
100
60mm/hour
50 75mm/hour

0
20 30 40 50 60
Slope Angle (Degree)

Figure 2(b): Variation of Soil loss mass between Non Confinements and various
slope angle for various rainfall intensity at rainfall duration 30 minutes.
156 
 

Soil Loss Mass Vs Slope Angle For Non-Confinement

(t r = 50 minutes)

700
Soil Loss Mass (g)

600 20mm/hour
500 30mm/hour
400 40mm/hour
300 50mm/hour

200 60mm/hour

100 75mm/hour

0
20 30 40 50 60
Slope Angle (Degree)

Figure 2(c): Variation of Soil loss mass between Non Confinements and various
slope angle for various rainfall intensity at rainfall duration 50 minutes.

Soil Loss Mass Vs Slope Angle For Non-Confinement

(t r = 60 minutes)

1000
900
Soil Loss Mass (g)

20mm/hour
800
700 30mm/hour
600 40mm/hour
500 50mm/hour
400
300 60mm/hour
200 75mm/hour
100
0
20 30 40 50 60
Slope Angle (Degree)

Figure 2(d): Variation of Soil loss mass between Non Confinements and various
slope angle for various rainfall intensity at rainfall duration 60 minutes.
 157 
 

Soil Loss Mass Vs Rainfall Intensity For Non-Confinement

(t r = 10 minutes)
100
Soil Loss Mass (g)

80 slope angle 30 degree

60 slope angle 40 degree

40 slope angle 50 degree

slope angle 60 degree

20
0
10 20 30 40 50 60 70 80
Rainfall Intensity (mm/hour)

Figure 3(a): Variation of Soil loss mass between non confinements and
various rainfall intensity for various slope angle at rainfall duration 10
minutes.
Soil Loss Mass Vs Rainfall Intensity For Non-Confinement
(t r = 30 minutes)
250
Soil Loss Mass (g)

200 slope angle 30 degree

150 slope angle 40 degree

100 slope angle 50 degree

slope angle 60 degree

50

0
10 20 30 40 50 60 70 80
Rainfall Intensity (mm/hour)

Figure 3(b): Variation of Soil loss mass between non confinements and
various rainfall intensity for various slope angle at rainfall duration 30
minutes.
158 
 

Soil Loss Mass Vs Rainfall Intensity For Non-Confinement

(t r = 50 minutes)

700
600
Soil Loss Mass (g)

slope angle 30 degree

500
slope angle 40 degree
400
300 slope angle 50 degree

200 slope angle 60 degree

100
0
10 20 30 40 50 60 70 80
Rainfall Intensity (mm/hour)

Figure 3(c): Variation of Soil loss mass between non confinements and
various rainfall intensity for various slope angle at rainfall duration 50
minutes.

Soil Loss Mass Vs Rainfall Intensity Non-Confinement

(t r = 60 minutes)
1000
900
Soil Loss Mass (g)

800 slope angle 30 degree

700
600 slope angle 40 degree
500
400 slope angle 50 degree
300
200 slope angle 60 degree
100
0
10 20 30 40 50 60 70 80
Rainfall Intensity (mm/hour)

Figure 3(d): Variation of Soil loss mass between non confinements and
various rainfall intensity for various slope angle at rainfall duration 60
minutes.
 159 
 

Soil Loss Mass Vs Rainfall Duration

(Non-Confinement, Slope Angle 30°)

350
300
Soil Loss Mass (g)

20mm/hour
250 30mm/hour
200 40mm/hour
150 50mm/hour

100 60mm/hour

50 75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 4(a): Variation of Soil loss mass between Non Confinements and
various rainfall duration for various rainfall intensity at 30 degree slope
angle.
Soil Loss Mass Vs Rainfall Duration
(Non-Confinement, Slope Angle 40°)

500
Soil Loss Mass (g)

20mm/hour
400
30mm/hour
300 40mm/hour
50mm/hour
200
60mm/hour
100 75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 4(b): Variation of Soil loss mass between Non Confinements and various
rainfall duration for various rainfall intensity at 40 degree slope angle.
160 
 

Soil Loss Mass Vs Rainfall Duration

(Non-Confinements, Slope Angle 50°)

500
Soil Loss Mass (g)

20mm/hour
400
30mm/hour
300 40mm/hour
50mm/hour
200
60mm/hour
100 75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 4(b): Variation of Soil loss mass between Non Confinements and various
rainfall duration for various rainfall intensity at 50 degree slope angle.
Soil Loss Mass Vs Rainfall Duration
(Non-Confinement, Slope Angle 60°)

1200
Soil Loss Mass (g)

1000 20mm/hour
30mm/hour
800
40mm/hour
600 50mm/hour
400 60mm/hour
200 75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 4(d): Variation of Soil loss mass between Non Confinements and various
rainfall duration for various rainfall intensity at 60 degree slope angle.
 
 161 
 

8.3.2 Confinement Slope Tests Results

As described in Section 8.2, three confinement shapes of sizes

50mm, 100mm and 150mm has been tested for slope angle 30, 40,
50 and 60. The slopes are subjected to rainfall intensity of 20, 30,
40, 50, 60 and 75 mm/hr applied for 10, 20, 30, 40, 50 and 60
minutes. The results are presented in Figures 4(a), 4(b) and 4(c).

8.3.2.1 Effects of Slope Angle

Figure 5(a) to Figure 5(f) shows the variation of soil loss mass
against slope angle from 30 degree to 60 degree for three types and
sizes of confinements. The figures indicate the soil loss mass show
increases with an increment of slope, the soil loss mass decreases
as the size of confinement increases. The effect of slope angle is
not significant regards to 50mm confinement size for rainfall
intensity less than 75mm/hour and rainfall duration less than 60
minutes. Nevertheless, for 100mm and 150 mm confinement sizes,
the soil loss mass increases significantly for slope angle of 50
degree and 60 degree for all range of rainfall intensity. Apparently,
the maximum soil loss mass increases with an increment of
confinement size. Figure 5(a) to Figure 5(f) portrays the soil loss
variations at various slope angles. It is noticeable that in general,
triangular shape confinements possess least soil loss mass followed
by circular and square shape confinements. This is influenced by
surface plain area of confinements that the triangular shape has a
smaller plain surface area as compared to circular and square
shape.

8.3.2.2 Effects of Rainfall Intensity

Figure 6 (a) to Figure 6 (d) depicts the variations of soil loss mass
with rainfall intensity from 20mm/hour to 75mm/hour with respect
to confinement size and rainfall duration. The soil loss mass
increases gradually with rainfall intensity for 50mm size
162 
 

confinement slope. The Figures shows the effect of rainfall

intensity is minor for confinement size 100mm and 150mm size
when the rainfall intensity is approximate equivalent to
40mm/hour. Nevertheless, there is clear evidence that rainfall
intensity more than 40mm/hour is significant since maximum soil
loss mass continue to increase with an increment of rainfall
intensity until 75mm/hour.

8.3.2.3 Effects of Rainfall Durations

Figure 7(a) to Figure 7(f) presents the variation of soil loss mass
with rainfall duration. Two sizes of confinement width of 50mm
and 150mm portray the constant rate of increment with rainfall
intensity for all range rainfall durations. The variations of soil loss
mass is not significant for rainfall duration less than 30 minutes
and rainfall intensity ranging from 20mm/hour to 75mm/hour.
However, the variation of soil loss mass further increases for
rainfall duration 50 minutes and 60 minutes and rainfall intensity
from 40mm/hr to 75mm/hr. The Figures also shows that rainfall
duration is less significant for rainfall intensity less than
30mm/hour. The reduction of soil loss mass were found to be due
higher for 50mm size confinement than those of 150mm size
confinement slopes. This is due to the fact that larger surface area
will reduce surface stiffness of the soil. Disintegration of soil
particles will be more easier for those having large surface area in
comparison with the smaller one.
 163 
 

Soil Loss Mass Vs Slope Angle

For Three Types and Sizes Confinements
(t r = 10 minutes, I = 20 mm/hour)
6
50CF
5
Soil Loss Mass (g)

50T F
50SF
4
100CF
3 100T F
2 100SF
150CF
1
150T F
0 150SF

20 30 40 50 60
Slope Angle (Degree)

Figure 5(a): Variation of soil loss mass against slope angle for three types
and sizes confinements slope at rainfall duration 10 minutes and rainfall
intensity 20 mm/hour.
Soil Loss Mass Vs Slope Angle
For Three Types and Sizes Confinements
(t r = 10 minutes, I = 30 mm/hour)
8
50CF
7
Soil Loss Mass (g)

50T F
6 50SF
5 100CF
4 100T F
3 100SF
2 150CF
1 150T F
0 150SF

20 30 40 50 60
Slope Angle (Degree)

Figure 5(b): Variation of soil loss mass against slope angle for three types
and sizes confinements slope at rainfall duration 10 minutes and rainfall
intensity 30 mm/hour.
164 
 

Soil Loss Mass Vs Slope Angle

For Three Types and Sizes Confinements
(t r = 10 minutes, I = 40 mm/hour)
10 50CF
50T F
Soil Loss Mass (g)

8
50SF
6 100CF
100T F
4 100SF
150CF
2
150T F
0 150SF

20 30 40 50 60
Slope Angle (Degree)

Figure 5(c): Variation of soil loss mass against slope angle for three types
and sizes confinements slope at rainfall duration 10 minutes and rainfall
intensity 40 mm/hour.
Soil Loss Mass Vs Slope Angle
For Three Types and Sizes Confinements
(t r = 10 minutes, I = 50 mm/hour)
14
50CF
12
Soil Loss Mass (g)

50T F
10 50SF
8 100CF
100T F
6
100SF
4
150CF
2 150T F
0 150SF

20 30 40 50 60
Slope Angle (Degree)

Figure 5(d): Variation of soil loss mass against slope angle for three types
and sizes confinements slope at rainfall duration 10 minutes and rainfall
intensity 50 mm/hour.
 165 
 

Soil Loss Mass Vs Slope Angle

For Three Types and Sizes Confinements
(t r = 10 minutes, I = 60 minutes)
20
50CF
Soil Loss Mass (g)

50T F
15 50SF
100CF
10 100T F
100SF
5 150CF
150T F
0 150SF

20 30 40 50 60
Slope Angle (Degree)

Figure 5(e): Variation of soil loss mass against slope angle for three types and
sizes confinements slope at rainfall duration 10 minutes and rainfall intensity 60
mm/hour.
Soil Loss Mass Vs Slope Angle
For Three Types and Sizes Confinements
(t r = 10 minutes, I = 75 mm/hour)
30
50CF
25
Soil Loss Mass (g)

50T F

20 50SF
100CF
15 100T F
10 100SF
150CF
5
150T F
0 150SF
20 30 40 50 60
Slope Angle (Degree)

Figure 5(f): Variation of soil loss mass against slope angle for three types
and sizes confinements slope at rainfall duration 10 minutes and rainfall
intensity 75 mm/hour.
166 
 

Soil Loss Mass Vs Rainfall Intensity

For Three Types and Sizes Confinements
(t r = 10 minutes, Slope = 60 Degree)
30
50CF
25
Soil Loss Mass (g)

50T F
50SF
20
100CF
15 100T F

10 100SF
150CF
5 150T F
0 150SF

10 20 30 40 50 60 70 80
Rainfall Intensity (mm/hour)

Figure 6(a): Variation of soil loss mass against various rainfall intensity for
three types and sizes confinements slope at rainfall duration 10 minutes and
slope angle 60 degree.
Soil Loss Mass Vs Rainfall Intensity
For Three Types and Sizes Confinements
(t r = 30 minutes, Slope = 60 Degree)

70
Soil Loss Mass (g)

60 50CF

50 50T F

40 50SF

30 100CF

20 100T F

10 100SF

0 150CF
150T F
10 20 30 40 50 60 70 80
150SF
Rainfall Intensity (mm/hour)

Figure 6 (b): Variation of soil loss mass against various rainfall intensity for
three types and sizes confinements slope at rainfall duration 30 minutes and
slope angle 60 degree.
 167 
 

Soil Loss Mass Vs Rainfall Intensity

For Three Types and Sizes Confinements
(t r = 50 minutes, Slope = 60 Degree)
140
50CF
120
Soil Loss Mass (g)

50T F
100 50SF

80 100CF
100T F
60
100SF
40 150CF
20 150T F
0 150SF

10 20 30 40 50 60 70 80
Rainfall Intensity (mm/hour)

Figure 6 (c): Variation of soil loss mass against various rainfall intensity for
three types and sizes confinements slope at rainfall duration 50 minutes and
slope angle 60 degree.
Soil Loss Mass Vs Rainfall Intensity
For Three Types and Sizes Confinements
(t r = 60 minutes, Slope = 60 Degree)
140
50CF
120
Soil Loss Mass (g)

50T F
100 50SF
80 100CF
100T F
60
100SF
40
150CF
20 150T F
0 150SF

10 20 30 40 50 60 70 80
Rainfall Intensity (mm/hour)

Figure 6 (d): Variation of soil loss mass against various rainfall intensity for
three types and sizes confinements slope at rainfall duration 60 minutes and
slope angle 60 degree.
168 
 

Soil Loss Mass Vs Rainfall Duration

(Circular Confinement 50mm Size, Slope angle 60°)

50
Soil Loss Mass (g)

20mm/hour
40
30mm/hour
30 40mm/hour
50mm/hour
20
60mm/hour
10 75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 7(a): Variation of soil loss mass against rainfall duration for 50 mm
size circular confinements and slope angle 60 degree.
Soil Loss Mass Vs Rainfall Duration
(Circular Confinement 150mm Size, Slope angle 60°)

200
Soil Loss Mass (g)

20mm/hour
150 30mm/hour
40mm/hour
100 50mm/hour
60mm/hour
50
75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 7(b): Variation of soil loss mass against rainfall duration for 150 mm
size circular confinements and slope angle 60 degree.
 169 
 

Soil Loss Mass Vs Rainfall Duration

(Triangular Confinement 50mm Size, Slope angle 60°)

50
Soil Loss Mass (g)

20mm/hour
40
30mm/hour
30 40mm/hour
50mm/hour
20
60mm/hour
10 75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 7(c): Variation of soil loss mass against rainfall duration for 50 mm
size triangular confinements and slope angle 60 degree.

Soil Loss Mass Vs Rainfall Duration

(Triangular Confinement 150mm Size, Slope angle 60°)

120
Soil Loss Mass (g)

100 20mm/hour
30mm/hour
80
40mm/hour
60 50mm/hour
40 60mm/hour
20 75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 7(d): Variation of soil loss mass against rainfall duration for 150 mm
size square confinements and slope angle 60 degree.
170 
 

Soil Loss Mass Vs Rainfall Duration

(Square Confinement 50mm Size, Slope angle 60°)

60
Soil Loss Mass (g)

50 20mm/hour
30mm/hour
40
40mm/hour
30 50mm/hour
20 60mm/hour
10 75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 7(e): Variation of soil loss mass against rainfall duration for 50
mm size square confinements and slope angle 60 degree.

Soil Loss Mass Vs Rainfall Duration

(Square Confinement 150mm Size, Slope angle 60°)

200
Soil Loss Mass (g)

20mm/hour
150 30mm/hour
40mm/hour
100 50mm/hour
60mm/hour
50
75mm/hour

0
0 10 20 30 40 50 60
Rainfall Duration (minutes)

Figure 7(f): Variation of soil loss mass against rainfall duration for 150 mm
size square confinements and slope angle 60 degree.
 171 
 

8.3.3 Discussions 

Test on non-confinement slope gives the limiting soil loss mass

thus postulated as benchmarking for comparison. The preliminary
factors that could lead to slope failures was achieved by
conducting various laboratory model tests for non-confinement
slope on the clayey soil material obtained from the site.
Apparently, the behaviour of soil loss mass for non-confinement
and confinement slopes tests are of almost similar trend in both
characteristic. However some differences were observed in view
of soil loss mass generated form slope surface and time taken to
failure.
The results plotted show that the soil loss mass is greatly
affected by several factors such as, slope angle, rainfall intensity
and rainfall duration. Apart from that, soil loss mass induced from
confinements slope was affected by the confinement size,
configuration and plane surface area. Generally, soil loss mass
generated from non-confinement slope is more pronounced than
confinement slope. For non-confinement slope, rainfall intensity
equivalent to 30mm/hour and rainfall duration of 50 to 60 minute
gives significant attribution towards erosion. In contrast, for
rainfall intensity more than 40mm/hour, rainfall duration 10
minutes is significant. This finding is in agreement with Morgan
and Davison (1986) findings that erosion occur at low intensity but
prolong rainfall duration and intense rainfall with short duration. It
was observed that the effect of slope angle is significant for rainfall
intensity approximate equivalent to 40mm/hour or higher with
rainfall duration more than 10 minutes.
A sudden increase of soil loss mass was clearly seen at 60
degree slope angle, rainfall intensity of 75mm/hour with 50
minutes rainfall duration; 60mm/hour and 75mm/hour for rainfall
duration 60 minutes respectively. Significant effect of slope angle,
rainfall intensity and rainfall duration on erosion observed in this
study is in agreement with Anders Rapp, (1975); Tan et al., (1987)
findings that when wetting front advance into the soil, loss of
suction will occur and could be cause of slope failure.
172 
 

Slopes embedded with 50mm and 150mm soil confinement

tests show similar behaviour to that non-confinement slope tests.
Thus direct comparison between confinement slope and non-
confinement slope tests was based on this confinement size.
Graphs plotted show that triangular shape soil confinement posses
least soil loss mass than circular and square shape. This finding
explain that soil confinement system is effectively in reducing
slope erosion. The smaller the confinement size more effective
will be the confinement system. The tests results further confirm
the virtue of soil confinement system in increasing slope stability.

8.4 Conclusions

The main purpose of this research is to determine soil loss mass

from non-confinement slope and three types and sizes of
confinements slope influenced by rainfall intensity, rainfall
duration at various slope angles. Soil confinement system to
improve the slope stability was demonstrates by laboratory
experimental results for all three types of confinements were
proven to be useful in reducing soil loss mass significantly.
Envisage to scenario of sudden increase of soil loss mass on non-
confinement slope, there is no significant of soil loss for all
confinement slopes directly evident support improvising the slope
stability. Tests shows that time taken for confinement slope to
failure indeed required than non-confinement slope further confirm
soil confinement system virtually can improve the slope stability.
Soil confinement system confines and reinforces infill material,
greatly improving non-vegetated slopes resistance to erosive
forces. Confinement prevents down slope migration of individual
particles caused by gravity and hydraulic traction. The slope apply
with soil confinement system was found significantly reduced the
rate of erosion which lead towards slope stability rehabilitation. In
this regard, the slope can be permanently protected by soil
confinement system. Although the work does not dedicated to any
specific subgrade condition and confinement material, however the
 173 
 

finding was thought to be invaluable in venturing the prospect of

confining system as an early stage prevention of slope instability.

References

Anders R. (1975). Soil Erosion and Sedimentation in Tanzania and

Lesotho. The Royal Swedish Academy of Sciences. 4:154-163.
Anderson S. A. and Riemer M. F. (1994). Collapse of Saturated
Soil Due to Reduction in Confinement.” Journal of
Geotechnical Engineering. 121(2): 216.
Bradford and Huang (1993). Comparison of Interrill Soil Loss for
Laboratory and Field Procedures. Soil Technology 6:145-156.
Greg Northcutt (2003). Capture Top Performance from
Geosynthetics. Fleet Management Magazine. Grading and
Excavation Contractor. 3 - 7.
Hudson N. W. (1957). The Design of Field Experiments on Soil
Erosion. Journal Agriculture Engineering Research. 2: 56-65.
Magnus Karlsson and Magnus Rahmberg (1999). Assessment Of
Suspended Sediment Variability in Odzi River, Zimbabwe.
Uppsala University.
Michael B. and William Outcalt (2002). Final Report – Evaluation
of Slope Stabilization Methods. Colorado Department of
Transportation. USA. Report No. CDOT-DTD-R-2002-10, 3 -
7.
Miller S. (2003). Soil Retaining Structure. The Primer Magazine
For The California, Nevada & Arizona Residential Building
Industry. A Natural Approach to Grade Separations.
Morgan R.P.C and Davidson D. A. (1986). Soil Erosion and
Conservation. Longman Scientific & Technical.
Poesen J. and Torri D. (1988). The Effect of Cup on Splash
Detachment and Transport Measurements, II: Theoretical
Approach.” Catena Supplement 12: 127-137.
Ramli N. and Lim, K.Y. (2007). Shape Variabilty of Soil
Confinement System for Slope Protection Work, Proceeding
of the 4th. International Conference on Disaster Prevention and
174 
 

Rehabilitation, Diponegoro University, Semarang, Indonesia,

139 – 149.
Ramli N. and Lim, K.Y. (2008). Unit Gravity procedure in
Modelling of Soil Confinement System, International
Conference in Geotechnical Engineering
(GEOTROPIKA2008), Kuala Lumpur, Malaysia.
Rimoldi P. and Ricciuti A. (1994). Design Method for Three-
Dimensional Geocells on Slopes. 5th International Conference
on Geotextiles, Geomembrances and related Products.
Singapore.
So H. B., Sheridan G. J., Horn C. P. and Currey N. (1999).
Predicting Hill Slope Erosion on Disturbed Landscape From
Laboratory Scale Measurement. Proceeding of 10th
International Soil Conservation Organization Meeting. May 24
– 29. Perdue University.
Tan S. B., Lim T. L., Tan S. L. and Yang K. S. (1987). Landslide
Problems And Their Control In Singapore. Proceeding of The
9th Southeast Asian Geotechnical Conference. December 7 –
11 Bangkok 
 INDEX 

anchorage, 112–113, 115, 118, 120, 139, 141, 145, 147

atterberg limit, 1–2, 8–12, 14–15, 18–19, 23–24
compaction energy, 4–6, 16
compressibility, 53–54, 67, 69
electrokinetic, 72–79, 81–82, 85
electrolyte, 74
electromigration, 72, 85
electroosmosis 72–73
erosion, 149–150, 154, 171
friction, 107–110, 112, 117, 120, 123, 126–127, 131
geosynthetics, 106, 110, 113–114, 122
lateral earth pressure, 113, 118, 121–123, 127–128
limestone formation, 28
liquid limit, 1–2, 8–10, 12, 14–18, 20–24
maximum dry density,1–2, 5, 10,12, 17, 20, 22–25
optimum moisture content, 1–2,4–5,10,12, 16, 17, 20, 22–25
passive bearing, 120
plastic limit, 1–2, 9–10, 1, 15–17, 20, 22–23
plasticity index, 1–2, 9–10, 14, 16, 18, 20–21, 24
pore water pressure 89
pozzolanic reaction, 39, 41, 46, 50
preloading, 53-55, 58, 62, 65
pullout, 107, 112, 115, 118, 120-121
rainfall, 89
reinforced wall, 109, 114, 117, 130-131
reinforcement, 106-115, 117-118, 120-123, 126-127, 130-131,
135-137, 140, 147
residual soil, 71-72, 85
rock bolt, 134-139, 140
rock stabilization, 135
root tensile strength, 89
settlement criteria, 65
settlement monitoring, 58, 67-68
shear strength, 27, 41, 62, 67, 69
slope failure, 150, 171
soft soil, 53-54, 62, 67
soil confinement, 149-152, 172
soil-lime reaction, 31, 48
soil stabilization, 27
soil modification, 27
soil moisture, 89-91
tensile resistance, 118, 123
tree induced suction, 90
vertical drain, 53-56, 58, 62, 64-67, 69
water content, 2-5, 17