Faculty of Engineering

Faculty of Engineering - Papers

University of Wollongong Year

Predictions and Performances of

Prefabricated Vertical Drain Stabilised
Soft Clay Foundations

B. Indraratna C. Rujikiatkamjorn

University of Wollongong, [email protected]

University of Wollongong, [email protected]

This conference paper was originally published as Indraratna, B and Rujikiatkamjorn,

C, Predictions and Performances of Prefabricated Vertical Drain Stabilised Soft Clay Foun-
dations, Proceedings of the Symposium on Rigid Inclusions in Difficult Soft Soil Conditions
International Society for Soil Mechanics and Geotechnical Engineering (ISSMGE TC36), 2006.

This paper is posted at Research Online.

http://ro.uow.edu.au/engpapers/195
 Predictions and performances of prefabricated vertical drain sta-
bilised soft clay foundations
Buddhima Indraratna
Professor, Division of Civil Engineering, University of Wollongong, City of Wollongong, NSW
2522, Australia.

Cholachat Rujikiatkamjorn
Associate Research Fellow, Division of Civil Engineering, University of Wollongong, City of Wol-
longong, NSW 2522, Australia.

ABSTRACT: In this paper, the analytical solution for radial consolidation of soft soils is proposed considering the impacts of the
variation of volume compressibility and permeability. The Cavity Expansion Theory is employed to predict the smear zone caused by
the installation of mandrel driven vertical drains in soft clay. The smear zone prediction is then compared to the data obtained from
the large-scale radial consolidation tests. Furthermore, the advantages and limitations of vacuum application through vertical drains in
the absence of high surcharge embankments are discussed using the proposed solutions. The applied vacuum pressure generates nega-
tive pore water pressure, resulting in an increase in the effective stress, which leads to accelerated consolidation. Analytical and Nu-
merical analysis incorporating the equivalent plane strain solution are conducted to predict the excess pore pressures, lateral and verti-
cal displacements. The equivalent plane strain solution can be used as a predictive tool with acceptable accuracy due to the significant
progress that has been made in the past few years through rigorous mathematical modelling and numerical analysis developed by the
primary author and co-workers (Indraratna et al., 1992 2005).
Several case histories are discussed and analysed, including the site of the 2nd Bangkok International Airport. The predictions are
compared with the available field data, confirming that the equivalent plane strain model can be used confidently to predict the per-
formance with acceptable accuracy. Difficulties in assuring good performance can also be analysed and interpreted through mathe-
matical modelling, thereby enabling due caution in the design and construction stages. The research findings verify that the role of
smear, drain unsaturation, and vacuum distribution can significantly affect soil consolidation, hence, these aspects need to be mod-
elled appropriately in any numerical analysis to obtain reliable predictions.

1 INTRODUCTION C
L
Rapid urban development has compelled engineers to con- Inclinometer
struct earth structures such as major highways and rail- Benchmark and Surface settlement plate
Dummy piezometer Sand Blanket
ways over soft clay deposits which usually show exces-
sive settlement and low bearing capacity characteristics
(Johnson, 1970; Indraratna et al., 1992). In the coastal re-
gions of many countries, soft clays are widespread in the
vicinity of capital cities. When such areas are selected for
infrastructure development, ground improvement tech-
niques for the underlying soft soils is essential. The appli-
cation of prefabricated vertical drains (PVDs) with pre-
loading has become common practice, and it is one of the Piezometer Sub-surface settlement plate
most effective soft clay improvement methods. Usually, a
surcharge load equal to or greater than the expected foun- Figure 1: Vertical drain system with preloading (Indraratna et al.
dation loading is applied over the soil surface to accelerate 2005d).
consolidation by rapid pore pressure dissipation through In order to increase vertical drain system performance,
the vertical drains (Hansbo, 1981; Jamiolkowski, 1983; the vacuum preloading method was initially introduced in
Indraratna et la., 1994; Indraratna and Redana., 2000). Sweden by Kjellman (1952) for cardboard wick drains. It
Figure 1 illustrates a typical plan of vertical drain sys- has been used extensively to accelerate the consolidation
tem installation and essential instruments required to process for improving soft ground, such as Tianjin port,
monitor the performance of soil foundation beneath an China (Yan and Chu, 2003). When a higher surcharge
earthfill embankment. Before installing the vertical drains, load is required to meet the expected settlement and the
general site preparations including the removal of vegeta- cost of surcharge becomes costly, a combined vacuum and
tion and surficial soil, establishing site grading and plac- fill surcharge can be employed (Cognon et al., 1994; In-
ing a compact sand blanket are nominally required. The draratna et al. 2005a). In areas of very soft clay, where a
sand blanket is employed to expel water away from the high surcharge embankment cannot be raised, the vacuum
drains and to provide a sound-working mat for the vertical application remains the ideal choice. Recently, the PVD
drain rigs. system has also been employed to distribute vacuum
pressure to deep subsoil layer, thereby increasing the1
consolidation rate of reclaimed land from the sea (e.g. 2 DEVELOPMENT OF THE THEORY OF
Qian et al., 1992; Chu et al., 2000; Tang and Shang, CONSOLIDATION WITH VERTICAL DRAINAGE
2000). AND VACUUM PRELOADING
To investigate the behaviour of vertical drains, the unit
cell theory representing a single circular drain surrounded During the consolidation process, both volume compressi-
by a soil annulus in an axisymmetric condition has usually bility and permeability vary according to the stress level,
been used (e.g. Barron, 1948; Yoshikuni and Nakanodo, which imparts a direct influence on the shape of e-log'
1974; Hansbo, 1981). Under embankment loading, verti- and e-logkh relationships (Lekha et al. 2003). Moreover,
cal drains not only accelerate the consolidation rate, but the nature of subsoil stress history (preconsolidation pres-
also influence the deformation pattern of the subsoil. It is sure) gives different consolidation responses (Seah and
only at the embankment centreline where lateral dis- Juirnarongrit 2003). Therefore, in order to obtain the accu-
placements are zero that the unit cell analysis provides ac- rate prediction of the behaviour of a vertical drain system,
ceptable accuracy. Elsewhere, especially towards the em- Indraratna et al. (2005b) have incorporated the relation-
bankment toe, single drain analysis cannot provide an ships of e-log and e-logkh with radial consolidation.
accurate prediction due to lateral yield and heave com-
pared to plane strain multi-drain analysis (Indraratna, et
al., 1997). 2.1 Key assumptions
Modelling of prefabricated vertical drains in plane The following assumptions are based on the equal vertical
strain finite element analysis has become popular (Hird et strain solution given by Hansbo (1981) and the variation
al., 1992, Indraratna et al., 2004). The basic principle is to of soil permeability and soil volume compressibility. The
match the average degree of radial consolidation in axi- following is a summary taken from Indraratna et al.
symmetric condition with the proposed equivalent plane (2005b).
strain analysis. This equivalence enhances computational (1) Soil is homogeneous and fully saturated, whereby
efficiency by reducing the convergence time and required Darcys law is adopted. At the outer boundary of the unit
computer memory, while still providing the correct time- cell, flow is not permitted to occur. For relatively long
settlement response. Various researchers have described vertical drains, only radial flow is allowed (i.e. no vertical
the advantage of such plane strain solutions for field stud- flow).
ies where a large number of drains are used, for which a (2) Soil strain is uniform at the upper boundary of the
true 3D analysis may become cumbersome or impractical unit cell and the small strain theory is valid.
(e.g. Hird et al., 1992, Chai et al., 1995, Indraratna et al., (3) The relationship between the average void ratio
2005f). and the logarithm of average effective stress in the
In this paper, the smear zone prediction based on The normally consolidated range (Fig. 2) can be expressed by:
Cavity Expansion Theory is discussed based on the large
scale laboratory results. The effects of the compressibility e = e0 C c log( ' / ' i ) . If the current vertical effective
indices, the variation of soil permeability and the magni- stress () is less than pc, then for the overconsolidated
tude of preloading are examined through the consolidation range, the recompression index (Cr) is used rather than Cc.
process. The equivalent 2-D plane strain solution is de-
scribed which includes the effects of smear zone caused e
by mandrel driven vertical drains. The equivalent (trans- Slope Cr
e0
formed) permeability coefficients are incorporated in fi-
nite element codes, employing the modified Cam-clay
theory. Numerical analysis is conducted to predict the ex-
cess pore pressures, lateral and vertical displacements. Slope Cc
Two case histories are discussed and analysed, including
the site of the New Bangkok International Airport (Thai-
land) and the predictions are compared with the available ef
field data. The research findings verify that the impact of
smear and vacuum pressure can significantly affect soil
consolidation, hence, in order to obtain reliable consolida-
tion predictions, these aspects need to be simulated appro- 'i p'c 'i+p log v
priately in the selected numerical approach.
Figure 2: Soil compression curve (after Indraratna et al., 2005b)

(4) For radial drainage, the horizontal permeability of

soil decreases with the average void ratio (Fig. 3). The
relationship between these two parameters is given by
Tavenas et al. (1983): e = e0 + C k log(k h / k hi ) The per-
meability index (Ck) is generally considered to be inde-
2
pendent of stress history (pc) as explained by Nagaraj et The void ratio (e)-effective stress () and the void ra-
al. (1994). tio-permeability relations for normally consolidated clays
(5) The vacuum pressure distribution along the drain can be express as (Tavenas et al., 1983):
boundary is considered to vary linearly from - p 0 at top of
e = e0 C c log( ' ' i ) (2)
the drain to - k1 p0 at the bottom of the drain, where k1 is
e = e0 + Ck log(kh khi ) (3)
a ratio between vacuum pressure at the bottom and the top
of the drain (Fig. 4).
where, Cc is the compression index and Ck is the perme-
e ability index ( C k 0.5eo ). If the current vertical effective
e0 stress () is smaller than preconsolidation pressure (pc),
the recompression index (Cr) is used instead of Cc for the
overconsolidated range.
Differentiating Eq. (2) with respect to the effective
stress () gives:
Slope Ck
ef mvi mv = 1 + p / 'i Ru p / 'i (4)

Combining Eqs. (3) and (4) yields:

kh khi log kh
(
kh khi = 1 + p / 'i Ru p / 'i )
Cc / C k
(5)
Figure 3: Semi-log permeability-void ratio (after Indraratna et
al., 2005b) Substituting Eqs. (4) and (5) into (1) gives:

CL Smear zone p (1 + k1 )
-p0 Ru + 0 Th = 8 P(Ru + p0 (1 + k1 ) 2p )
p 2
z (6)
Vacuum (
where, P = 1 + p / 'i Ru p / 'i )
1 C c / C k

pressure It can be seen that Eq. (6) is a non-linear partial differ-

distribution ks kh l ential equation for radial consolidation under immediate
loading, incorporating the e-log and e-logkh relations.
The nonlinear differential Eq. (6) with variable Ru does
not have a general solution and, P varies between 1 and
-k1p0
Undisturbed zone
(
the value of 1 + p 'i + p0 (1 + k1 ) 2 'i )
1C c / C k
. Hence, an
ds/2 average value for P is adopted:
de/2

(
P = Pav = 0.51 + 1 + p 'i + p0 (1 + k1 ) 2 'i

)
1 Cc / C k
(7)

Figure 4: An axisymmetric unit cell with vacuum pressure distri-
bution (modified after Indraratna et al., 2005c) Integrating Eq. (6) after incorporating Eq. (7) subjected
to the boundary condition that Ru = 1.0 at Th=0, gives the
following algebraic expression:
2.2 Solution for an axisymmetric unit cell
According to Indraratna et al. (2005b), the dissipation rate p (1 + k1 ) 8Th * p0 (1 + k1 )
Ru = 1 + 0 exp (8)
of average excess pore pressure ratio ( Ru = u t / p) at any p 2 p 2
time factor (Th) can be expressed as: where, Th * = PavTh (9)

p (1 + k1 ) When the value of Cc/Ck approaches unity and p0 be-

Ru + 0
p 2 = 8 mvi k h R + p0 (1 + k1 ) comes zero, the authors solution converges to the conven-
(1)
mv k hi tional solution proposed by Hansbo (1981):
u
Th p 2
Ru = exp( 8Th ) (10a)

in which,
3
 n kh (10b) 1C / C
= ln + ln s 0.75 c h, pc t p' c r k
s ks where, Th = , c h , pc = c hi
'

d e2 i
where, = a group of parameters representing the Since the relationship between effective stress and
geometry of the vertical drain system and smear effect, n strain is non-linear, the average degree of consolidation
= de/dw, s = ds/dw, de = equivalent diameter of cylinder of can be described either based on excess pore pressure
soil around drain, ds = diameter of smear zone and dw = (stress) (Up) or based on strain (Us). Up indicates the rate
diameter of drain well. In Eq. (10b), kh = average of dissipation of excess pore pressure whereas Up shows
horizontal permeability in the undistrubed zone (m/s), and the rate of development of the surface settlement. Nor-
ks = average horizontal permeability in the smear zone mally, U p U s except when the effective stress and
(m/s). Th is the dimensionless time factor for strain is a linear relationship, which is in accordance with
consolidation due to radial drainage. Terzaghis one-dimensional theory. Therefore, the average
Substituting Th * in Eq. (8), the expression for excess degree of consolidation based on excess pore pressure can
pore pressure ratio for normally consolidated clay be- be obtained as follows:
comes:
U p = 1 Ru (12)
1Cc / Ck
p Th The average degree of consolidation based on settle-
Ru = exp 4 1 + 1 + (11a)
'i ment (strain) is defined by:


Us = (13)
For overconsolidated soil, '< p'c (e.g. topmost layer
close to surface)
The associated settlements () are then evaluated by the
1Cr / Ck following equations:
p Th
Ru = exp 4 1 + 1 + (11b)
'i '
=
HC r
log , ' i ' p' c (14a)

1 + e0 'i
When the effective stress equals the preconsolidation
H p'c '
pressure ( '< p'c ), the corresponding time factor Th,pc can = C r log + C c log
p' ,

then be determined by: 1 + e0 'i c (14b)
p' c ' ' i + p
p
Th , pc = ln HC c '
1Cr / Ck ' i + p p' c = log for normally consolidated clay

p 1 + e0 'i
41 + 1 +
' (14c)
i

(11c) It is noted that can be obtained by
substituting ' = 'i + p into the above equations.
When ' > p' c along the slope of the normally con-
In the above equations, = settlement at a given
solidation curve (Cc), the expression for excess pore pres-
sure ratio when Th>Th,pc is given by: time, c = total primary consolidation settlement, Cc =
compression index, Cr = recompression index and H =
1 + compressible soil thiskness.
Depending on the location of the initial and final
1 +
1 Cc / C k
effective stresses with respect to the normally
4
' i + p p ' c consolidated and overconsolidated domains, the following
Ru = exp ' i + p p ' c
is a summary of the relavant computational steps.
p
p' c If both the initial and final effective stresses are in the

(
Th Th, pc )
normally consolidated range, Eqs (11a) and (12) are em-
ployed to calculate Up, whereas Eqs. (13) and (14c) are
used to compute Us.
(11d) If both the initial and final effective stresses are in the
overconsolidated range, Eqs. (11b) and (12) are employed

4
to calculate Up, and Eqs. (13) and (14a) are used to deter- r

mine Us.
If the initial effective stress falls on the overconsoli- z
Drain
kwp
l kw
dated domain and the final effective stress is on the nor- kh
ks
Smear zone
khp
k'hp

mally consolidated domain, then Eqs. (11b)-(11d) and

bw
(12) are employed to calculate Up, Eqs. (13) (14a) and rw
rs bs
R
(14b) are employed to calculate Us. l B

3 EQUIVALENT PLANE STRAIN CONSOLIDATION

MODEL

Most finite element analyses on embankments are carried

out based on the 2D plane strain assumption. However,
the consolidation around vertical drains is mainly axi-
symmetric. Therefore, to employ a realistic 2-D finite
element analysis for vertical drains, the equivalence be- (a) Axisymmetric (b) Plane strain
tween the plane strain analysis and axisymmetric analysis
needs to be established. Indraratna and Redana (1997, Figure 5: Conversion of an axisymmetric unit cell into plane
1998, 2000) and Indraratna et al. (2005a) converted the strain condition (after Indraratna and Redana, 2000)
vertical drain system shown in Fig. 5 into an equivalent
parallel drain well by adjusting the coefficient of perme- At a given effective stress level and at each time step,
ability of the soil, and by assuming the plane strain cell (a the average degree of consolidation for both axisymmetric
width of 2B). The half width of the drain bw and half width ( U p ) and equivalent plane strain ( U p, pl ) conditions are
of the smear zone bs may be kept the same as their axi-
made equal, hence:
symmetric radii rw and rs, respectively, which suggests
bw = rw and bs = rs . U p =U p , pl (18)
Indraratna et al. (2005a) represented the average degree
of consolidation in plane strain condition as follows: Combining Eqs. 15 and 17 with Eq. 10 of original
Hansbo theory (Hansbo, 1981), the time factor ratio can
p 0 p (1 + k1 ) 8T
0 p (1 + k1 ) (15)
u p be represented by the following equation:
= 1 + exp hp
2 p 2
u0 u o u0
Thp k hp R 2 P
and = = (19)
k hp Th kh B 2
p = +( ) (16)

k hp By assuming the magnitudes of R and B to be the same,
Indraratna and Redana (2000) presented a relationship be-
where, u 0 = initial excess pore pressure, u = pore , as follows:
tween k hp and k hp
pressure at time t (average values) and Thp = time factor in The influence of smear effect can be modelled by the
ratio of the smear zone permeability to the undisturbed
plane strain., are the undisturbed horizontal
k hp and k hp permeability, as follows:
and the corresponding smear zone equivalent permeabili-
ties, respectively. The geometric parameters and , are
k hp
= (20)
given by: k hp k hp n k h
ln + ln (s ) 0.75
k h s k h
2 2b s bs b2
= 1 + s2 (17a)
3 B B 3B If smear and well resistance effects are ignored in the

above expression, then the simplified ratio of plane strain
=
1
B2
(bs b w )2 +
bs
3B 3
(3b 2
w b s2 ) (17b) to axisymmetric permeability is readily obtained, as also
proposed earlier by Hird et al. (1992):

k hp 0.67
= (21)
kh [ln(n ) 0.75]

5
 For vacuum preloading, the equivalent vacuum pres-
sure in plane strain and axisymmetric are the same.

ln1
a 2 a02 ( ) = 2(1 + )


r2 3 3 (1 2 )
(24)

2 3 f ( , , OCR )
4 EVALUATION OF SMEAR ZONE SURROUNDING
MANDREL DRIVEN VERTICAL DRAINS USING and
CAVITY EXPANSION ANALYSIS
f ( , , OCR ) =
1 ( + ) 1 OCR 1
ln
( )
Indraratna et al. (2005g) and Sathananthan (2005) have 2 ( ) 1 + OCR 1 ( ) (25)
made an attempt to estimate the extent of smear zone,
caused by mandrel installation using the Cylindrical Cav- tan 1 + tan 1 OCR 1

( )
ity Expansion theory incorporating the modified Cam-clay
model (MCC). This technique is commonly used to ana-
lyse pile driving, tunnelling and soil testing. When a man- where, a = radius of the cavity, a0 = initial radius of the
drel is driven into soil, it will initially displace a volume cavity, = Poissons ratio, = slope of the overconsolida-
of soil equal to the volume of the mandrel. A heave of tion line, = specific volume and = 1 ( is the
soil can occur at the soil surface, up to about ten times the slope of the normal consolidation line). Finally, the corre-
radius of the mandrel. At a greater depth, the soil is dis- sponding mean effective stress, in terms of deviatoric
placed predominantly outwards in the radial direction. stress, total stress and excess pore pressure, can be ex-
Therefore, the expansion of a cylindrical cavity with a fi- pressed by the following expressions:
nal radius equal to that of the mandrel is appropriate to
predict the extent of smear zone.
'
OCR
The following is the summary from the doctoral re- p = p0' 2
(26)
search of Sathananthan (2005). After the initial yielding at 1 + ( )
the cavity wall, a zone of soil extending from the cavity q = p' (27)
wall to a radial distance (rp) will become plastic as the rp
cavity pressure continues to increase (Figure 6). For a soil q 2 q
obeying the MCC model, the yielding criterion is given
p = rp
3
+
3 r dr
r
(28)
by:
Employing Equations (26)-(28), the excess pore pres-
pc' sure due to mandrel driving (u) can be determined by:
= 1 (22)
p'
u = ( p p0 ) p ' p0' ( ) (29)
where, = stress ratio q/p (q is the deviatoric stress
(1-3)/2 and p is the effective mean stress (1+23)/3), where, p 0 = initial total mean stress. The extent of the
M = slope of critical state line projected to q-p plane and smear zone can be suggested as the region in which the
pc = effective preconsolidation pressure.
excess pore pressure is higher than the initial overburden
The stress ratio at the elastic-plastic boundary can be
found as follows: pressure ( v' 0 ) (Fig. 6). This is because, in this region, the
soil properties, such as permeability and soil anisotropy,
q qp are disturbed severely at radial distance where u = v' 0 .
p = ' = ' = OCR 1 (23)
p r = rp p0

where, p = stress ratio at the elastic-plastic boundary, r

Excess pore pressure

u = v' 0
= distance from central axis of the drain, qp = (1-3)/2 at
the elastic-plastic boundary and OCR is the isotropic over
Vertical drain

consolidation ratio, defined by pc' 0 p0' ( pc' 0 is the initial Smear

zone
preconsolidation pressure and p0' is the initial effective
mean stress). Stress ratio at any point can be determined
rp
as follows: Distance (r )

Figure 6: Smear zone prediction by the Cavity Expansion Theory

6
5 LABORATORY TESTING USING LARGE-SCALE
CONSOLIDOMETER

At the University of Wollongong, the performance of pre-

fabricated vertical drains has been extensively studied by
employing large-scale laboratory consolidation tests (Fig.
7). Initially, this apparatus was developed by Indraratna
and Redana (1995,1997) to study the effect of smear due
to vertical drain installation including sand drains and pre-
fabricated vertical drains (PVDs). The apparatus was later
modified to accommodate the vacuum application ( Indra-
ratna et al. 2004). The internal radius and the overall
height of the assembled cell are 225 mm and 1000 mm,
respectively. The loading system was applied by an air
jack compressor system via a piston. The settlement was
measured by a Linear Variable Differential Transducer
(LVDT) placed at the top of this piston. An array of strain
gauge type pore pressure transducers complete with wir-
ing to supply recommended 10 V DC supply was installed
to measure the excess pore water pressures at various
points.

5.1 Smear zone determination due to vertical drain

installation
Figure 7: Schematic of large-scale consolidation apparatus (after
Indraratna and Redana (1997) found that even though a Indraratna and Redana, 1995)
larger width of the drain may cause a greater smear zone,
for PVDs, the measurements and predictions indicate 2.00

slightly increasing settlements due to the increased surface

Horizontal/Vertical permeability ratio

area, facilitating efficient pore water pressure dissipation.

1.50
The smear zone extent can be determined either by per-
meability variation or water content variation along the
radial distance (Indraratna and Redana, 1997; Sathanan- 1.00
than and Indraratna, 2006). Figure 8 shows the variation Mean Consolidation Pressure:
drain
of the ratio of the horizontal to vertical permeabilities Smear zone 6.5 kPa
0.50 16.5 kPa
(kh/kv) at different consolidation pressures along the radial 32.5 kPa
64.5 kPa
distance, obtained from large-scale laboratory consolida- 129.5 kPa
260 kPa
tion. The variation of the water content with depth and ra- 0.00
0 50 100 150 200
dial distance is shown in Figure 9 for an applied pressure Radial distance, R (mm)
of 200 kPa. As expected, the water content decreases to-
wards the drain, and also the water content is greater to-
wards the bottom of cell at all radial locations. Figure 8: Ratio of kh/kv along the radial distance from the central
drain (after Indraratna and Redana, 1995)
Based on these curves, the extent of smear zone can be
estimated to be around 2.5 times the equivalent mandrel
radius. This agrees well with the estimated extent of smear 5.2 Effects of vacuum preloading
zone based on both the kh/kv ratio (Fig. 8). It can be seen Indraratna et al. (2004) studied the effects of application
that the average value of kh/kv starts to decrease considera- of vacuum pressure through prefabricated vertical band
bly from 1.65 (outside smear zone) to 1.1 (inside smear drains (PVDs). The vacuum pressure was set to 100 kPa
zone). It implies that the permeability ratio between undis- and applied to the PVD and soil surface through the cen-
turbed and smear zone (kh/kh) is approximately 1.5 and tral hole of the rigid piston. The subsequent surcharge
the extent of smear zone (rs) is 4-5 times the radius of the load was applied instantaneously in two stages, 40 kPa
vertical drain (rw). It should be noted that kh/kh ratio in and 100 kPa with duration of 14 days between each stage.
the field can vary from 1.5 to 10, depending on the type of During the test, the vacuum pressure was released in two
drain and soils as well as installation procedures (Saye, stages for short periods to investigate the effects of vac-
2003). uum unloading and reloading. The suction measured close

7
to the PVD and the surface settlement measured from 6 PERFORMANCE OF TEST EMBANKMENT
LVDT are shown in Figure 10. STABILISED WITH VERTICAL DRAINS
INCORPORATING VACUUM PRESSURE ON SOFT
70
CLAY, BANGKOK, THAILAND
wmax = ws = 69%

68
Water content, w (%)

6.1 General
Drain

66 Location of the sample

smaple The Second Bangkok International Airport or Suvarnab-
from bottom (mm)
0 (bottom) humi Airport is situated about 30km from the city of
200 Bangkok. In the past, the site was occupied by rice fields
64 400 for for agricultural purposes. The area is often flooded
600 during the rainy season and the soil generally has very
800
Smear zone high moisture content. Therefore, soft marine clays, often
62
0 1 2 3 4 5
present considerable construction problems, which require
r/rm ground improvement techniques to prevent excessive set-
Figure 9: Variation of water content with depth and radial dis- tlement and lateral movement.
tance for an applied pressure of 200 kPa (after Sathananthan and The subsoil profile at the site consists of 1m thick
Indraratna, 2006) weathered crust (highly overconsolidated clay) overlying
a very soft to medium clay, which extends about 10m be-
20 low the ground surface. Underneath the medium clay
0 (a) layer, a light-brown stiff clay layer is found at a depth of
Suction (kPa)

-20 10-21m (AIT, 1995). The ground-water level fluctuates

-40 between 0.5 and 1.5m below the surface. The water con-
-60 tent of the very soft clay layer varies from 80 to 100%,
-80 whereas in the lower parts of the stratification (10-14m) it
-100
0 changes from 50 to 80%. The plastic limits and liquid lim-
0 vacuum removal 10 Time (Days) 20 30 its of the soil in each layer are similar and found to be in
the range of 80 to 100% and 20 to 40%, respectively.
vacuum reloading
surcharge load doubled
(b)
-20
Settlement (mm)

6.2 Embankment details and vacuum pressure

Stage 1
measurement
-40 Surcharge load vacuum removal
40kPa vacuum reloading Several trial embankments were built at this site, two of
which, (TV1 and TV2) were built with PVDs and vacuum
-60 application (Asian Institute of Technology, 1995). Total
Stage 2
base area of each embankment was 40 40 m2. Figures 11
Surcharge load and 12 present the vertical cross sections and field instru-
100kPa
-80 ments positions for embankments TV1 and TV2, respec-
tively. For TV1 (Fig. 11), 15m long PVDs with a hyper-
Figure 10: a) Suction in the drain (240mm from bot- net drainage system were installed, and for TV2 (Fig. 12),
tom) and b) surface settlement surface settlement associ- 12m long PVDs with perforated and corrugated pipes
ated with simulated vacuum loading and removal (Indra- wrapped together in non-woven geotextile were used. The
ratna et al. 2004) drainage blanket (working platform) was constructed with
sand 0.3 m and 0.8 m for embankments TV1 and TV2, re-
This also indicates that the suction head decreases with spectively with an air and water tight Linear Low Density
the drain depth, as the maximum suction of 100 kPa could Polyethylene (LLDPE) geomembrane liner placed on top
not be maintained along the entire drain length. The set- of the drainage system. This liner was sealed by placing
tlement associated with combined vacuum and surcharge its edges at the bottom of the trench perimeter and covered
load is shown in Figure 10b, and it clearly reflects the ef- with a 300mm layer of bentonite and then submerged with
fect of vacuum removal and re-application by the corre- water. The array of instrumentation of the embankments
sponding change of gradient of the settlement plot (Indra- includes piezometers, surface settlement plates, multipoint
ratna et al. 2004). extensometers, inclinometers, observation wells and
benchmarks.
Table 1 shows the vertical drain properties for em-
bankment TV1 and TV2. The PVDs were installed in a
triangular pattern at a spacing of 1m. The type of drain in-

8
stalled in both embankments is Mebra drains (100 mm x 3 0
mm). In these embankments, the drains were installed us-

D epth (m)
ing a mandrel, which was gradually pushed into the soil
by a static weight without any vibration. This method was 5
employed to reduce the extent of smear zone as much as
possible. 10
.
10 m 10 m 5m 15 m
Geomembrane (LLDPE) Geonet with Geotextile
15 4 4.4 4.8 5.2 5.6 6
2.5 m.
LBM Vacuum Pump S3 S2 S1 0.3 m
0m rs/rw
Bentonite
-3 m
Legend
Surface settlement plate -6 m
Figure 13: Variation of smear zone with depth by the Cav-
Stand-pipe piezometer
Extensometer
ity Expansion theory (where, rs: radius of smear zone, and
-9 m
Electrical piezometer rw: radius of drain)
Inclinometer
-12 m

-15 m Table. 2 Soil parameters used in the analysis

Depth Cr Cc kh e0 Pc
Figure 11: Cross section of embankment TV1 and location
m 10-9 kN/ kPa
of monitoring system (Indraratna et al., 2005d)
m/s m3
15 m 5m 10 m 10 m 0.0-2.0 0.06 0.37 30.1 1.8 16 58
2.0-8.5 1.6 12.7 2.8 15 45
Perforated Pipe Geomembrane (LLDPE)
8.5-10.5 1.7 6.02 2.4 15 70
2.5 m
0.8 m
S1 S2 S3 Vacuum Pump S4 LBM 10.5-13 0.95 2.56 1.8 16 80
0.0 m 13-15 0.88 0.60 1.2 18 90
-3 m Bentonite
Legend 3
Embankment height (m)

-6 m Surface settlement plate

Stand-pipe piezometer
TV1
-9 m
Extensometer TV2
-12 m Electrical piezometer 2
Inclinometer
-15 m
PVD, S=1.00 m.
1
= 18 kN/m3
Figure 12: Cross section of embankment TV2 and location
0
of monitoring system (Indraratna et al., 2005d)
0 50 100 150
Time (days)
Table 1. Vertical drain parameters Figure 14: Multi-stage loading for embankments TV1 and
TV2 (Indraratna et al., 2005d)
Spacing, S 1.0 m (triangular)
Ratio of kh/kh 10 In each embankment, a vacuum pressure up to 70kPa
Diameter of drain, dw 50 mm could be achieved using the available vacuum equipment.
Discharge capacity, qw 50 m3/year (per drain) This pressure is equivalent to a fill height of 4m. After 45
Length of vertical drain 15 m for TV1 and 12 m days of vacuum application, the surcharge load was
for TV2 applied in 4 stages upto 2.5m high (the unit weight of
surcharge fill is 18 kN/m3) as illustrated in Fig. 14. The
The extent of the smear zone with depth was predicted settlement, excess pore water pressure, and lateral
using the cavity expansion theory as explained in Section movement were observed for 5 months. During the appli-
4, incorporating the Modified Cam clay parameters given cation of vacuum pressure, it was found that the suction
in Table 2. The predicted normalized pore water pressure head transmitted to the soil could not be constantly main-
(i.e., u/v0) variation with radial distance for each soil tained as shown in Fig. 15 (Sangmala, 1997). This fluctua-
layers is shown in Figure 13. tion has been was attributed to air leaks through the sur-
face membrane or the loss of suction head beneath the
certain depth for long PVD. Intersection of natural macro-
pores with drains at various depths also lead to suction
head drops, at various times.

9
 0 follows the normally consolidated line as indicated by the
Vacuum pressure (kPa)
values in Table 2. The time-dependent vacuum pressure is
-20 assumed to vary linearly to zero along the drain length
(k1= 0).
-40 Figures 16 and 17 compare predicted surface centreline
settlement with the measured data for Embankment TV1
TV1
-60
and TV2, respectively. As expected, the predicted results
TV2
based on the authors solutions agree well with the meas-
0 40 80 120 160 ured results, whereas the prediction based on the constant
Time (days) kh and smear zone overestimates settlement after 80 days,
because, the actual soil permeability decreases signifi-
cantly at higher stress levels. It is verified that the com-
Figure 15: Time dependent vacuum pressure (Indraratna bined vacuum application and the PVD system accelerates
et al., 2005d) consolidation, while the vacuum pressure acts as an addi-
tional surcharge load. As shown in Figs. 16 and 17, no
leakage condition gives more settlements, whereas the
6.3 Single drain analysis at the embankment centreline prediction without any vacuum application gives less set-
using proposed analytical model tlement. It indicates that performance of this system de-
In the field, at the embankment centreline (exploiting pends entirely on preventing airleaks and the distribution
symmetry), the condition of 3-D axisymmetric of vacuum pressure along the length of the drain.
consolidation without lateral strain assumed in the
proposed analytical model can be justified. The soil pa- 0

rameters, the in-situ effective stress and the soil perme-

ability for soft Bangkok clay subsoils are shown in Table
2. The relevant soil properties were obtained from CKoU 0.4
triaxial tests (AIT, 1995). As suggested by Tavenas et al.
Settlement (m)

(1983), the slope of e-logkh (Ck) can be determined by:

0.8
C k = 0.5e0 (30)
Field measurement
In the analysis, each subsoil layer was divided into 1.2 Proposed solution (with leakage)
smaller sublayers to derive a more accurate effective Constant kh and mv

stress distribution with depth. The value of soil com- No vacuum pressure
Proposed solution (no leakage)
pressibility index (Ck) is associated with the actual stress 1.6
state within a given region of the foundation, where the 0 40 80 120 160
Time (days)
working stress range must be considered in relation to the
preconsolidation pressure of soil at that particular depth
(Indraratna et al., 1994). The values of kh/ks for this case Figure 16: Surface settlement predictions at the centerline
study were assumed to be 2, whereas ds/dw is detemined of Embankment TV1 (after Rujikiatkamjorn, 2005)
by the cavity expansion theory (Fig. 13). 0
The embankment loading was simulated by assuming
an instantaneous loading at the upper boundary. Settle-
ment predictions were carried out at the embankment cen- 0.4
terline using the authors analytical model (e.g. Eqs. 8-
14). At the beginning of the subsequent stage, the initial
Settlement (m)

in-situ effective stress and initial coefficient of horizontal 0.8

consolidation ( c hi ) were calculated based on the final de-
gree of consolidation of the previous loading stage. An Field measurement
EXCEL spreadsheet formulation can be conveniently em- 1.2 Proposed solution (with leakage)
Constant kh and mv
ployed in the calculation. The value of soil compressibility
No vacuum pressure
(Cc or Cr) in association with the correct working effective Proposed solution (no leakage)
stress plays a very important role for predicting settle- 1.6
0 40 80 120 160
ment. For Stage 1 loading, where the effective preconsoli- Time (days)
dation pressure (pc) is not exceeded, the value of recom-
pression index (Cr) may be used. In particular, the surface
crust is heavily overconsolidated (upto about 2 m depth). Figure 17: Surface settlement predictions at the centerline
Once pc is exceeded, the value of compression index (Cc) of Embankment TV2 (after Rujikiatkamjorn, 2005)

10
6.4 Multi-drain analysis using FEM incorporating Model A: Conventional analysis (i.e., no vacuum ap-
proposed equivalent plain strain model plication),
Model B: Vacuum pressure varies according to field
In order to investigate the performance of a whole em-
measurement and decreases linearly to zero at the bottom
bankment, the consolidation behaviour was analysed using
of the drain (k1= 0),
the finite element software ABAQUS (Indraratna et al.,
Model C: No vacuum loss (i.e. -60kPa vacuum pres-
2005d). The equivalent plane strain model (Eqs.. 20-21) as
sure was kept constant after 40 days); vacuum pressure
well as the modified Cam-clay theory (Roscoe and Bur-
diminishes to zero along the drain length (k1= 0),
land, 1968) were incorporated in the analysis. According
Model D: Constant time-dependent vacuum pressure
to Indraratna and Redana (1997), the ratios of kh/ks and
throughout the soil layer (k1=1).
ds/dw determined in the laboratory are approximately 1.5-
2.0 and 3-4, respectively, but in practice these ratios can All the above models included the smear effect but ne-
vary from 1.5 to 6 depending on the type of drain, mandrel glected well resistance as previous studies indicated that
size and shape and installation procedures used (Indra- well resistance was not significant for drain lengths
ratna and Redana, 2000; Saye, 2003). The constant values shorter than about 20m (Indraratna & Redana, 2000). For
of kh/ks and ds/dw (Table 6.3) for this case study were as- embankment TV2 with 12m drain length, pore pressure at
sumed to be 2 and 6, respectively (Indraratna et al., 2004). the drain interface was fixed up to 12m depths only.
For the plane strain FEM simulation, the equivalent per- Figures 19 and 20 compare surface settlement between
meability inside and outside the smear zone was deter- prediction and measurement (centreline) for embankment
mined using Eqs. (20) and (21). The discharge capacity TV1 and TV2, respectively. Model B predictions agree
(qw) is assumed high enough and can be neglected (Indra- with the field data. Comparing all the different vacuum
ratna and Redana, 2000). pressure conditions, Models A and D give the lowest and
highest settlement, respectively. A vacuum application
The finite element mesh contained 8-node bi-quadratic
combined with a PVD system can accelerate the
displacement and bilinear pore pressure elements (Fig.
consolidation process significantly. With vacuum
18). Only the right hand side of the embankment was
application, most of the primary consolidation is achieved
modeled due to symmetry, as shown in Fig. 18. For the
around 120 days, whereas conventional surchage (same
PVD zone and smear zone, a finer mesh was implemented
equivalent pressure) requires more time to complete
so that each unit cell represented a single drain with the
primary consolidation (after 150 days). It is also apparent
smear zone established on either side. The finer mesh was
that a greater settlement can be obtained, if any loss of
imperative to prevent an unfavorable aspect ratio of the
vacuum pressure can be minimised (Model C).
elements (Indraratna and Redana, 2000). The different
drain lengths for the two embankments were modelled by 0
fixing the pore pressure boundary along the appropriate
depths. The incremental surcharge loading was simulated
-0.4
at the upper boundary.
Settlement (m)

drain smear zone 8-node biquadratic displacement, -0.8

bilinear pore pressure Field measurement
Model A
Displacement node -1.2 Model B
Pore pressure node Model C
Model D
-1.6
0 40 80 120 160
Time (Days)

15m Figure 19: Surface settlement time curves of embankment

TV1 (modified after Indraratna et al., 2005f)

Figures 21 and 22 illustrate the predicted and measured

excess pore pressures for embankments TV1 and TV2
20m 40m respectively. The field observations are closest to Model
B, implying that the writers assumption of linearly
Figure 18: Finite element mesh for plane strain analysis decreasing time-dependent vacuum pressure along the
(modified after Indraratna et al. 2005f) drain length is justified. Excess pore pressure generated
from the vacuum application is less than the conventional
The following 4 distinct models were numerically ex- case, which enables the rate of construction of an
amined under the 2D multi-drain analysis (Indraratna et al. embankment to be higher than conventional construction.
2005f):
11
 0 lateral displacement (m) lateral displacement (m)
Inward Outward Inward Outward
-0.3 -0.2 -0.1 0 0.1 0.2 -0.3 -0.2 -0.1 0 0.1 0.2
0 0
-0.4
Weathered crust layer
2 2
Settlement (m)

-0.8
4 Soft clay layer 4

Field measurement 6 6

Depth (m)

Depth (m)
-1.2
Model A
Model B 8 8

-1.6 Model C
Model D 10 10

0 40 80 120 160 12 12
Time (Days)
14 14
Figure 20: Surface settlement of embankment TV2 (modified
after Indraratna et al., 2005f)
(a) (b)
Model A
20 Model B
Model C
Excess pore water pressure (kPa)

0 Model D
Adjusted vacuum
Field measurement
-20

-40 Field measurement Figure 23: Calculated and measured lateral displacements
Model A distribution with depth (a) embankment TV1 (b)
-60
Model B embankment TV2 (modified after Indraratna et al., 2005f)
Model C
Model D
-80
The predicted and measured lateral displacements (at
0 40 80 120 160 the end of embankment construction) are shown in Fig.
Time (Days)
23. The observed lateral displacements do not agree well
with all vacuum pressure models. In the middle of the
Figure 21: Variation of excess pore water pressure 3m deep be-
low the surface and 0.5m away from centreline for Embankment very soft clay layer (4-5m deep), the predictions from
TV1 (modified after Indraratna et al., 2005f) Models B and C are closest to the field measurements.
Nearer to the surface, the field observations do not agree
20 with the inward lateral movements predicted by Models
B and C. The discrepancy between the finite element
Excess pore water pressure (kPa)

0 models and the measured results is more evident in the

topmost weathered crust (0-2 m).
-20

-40 7 SUMMARY AND CONCLUSIONS

Field measurement
Model A
-60 Model B
The use of prefabricated vertical drains, their properties
Model C and associated merits and demerits have been discussed.
Model D
-80 In this paper, the extent of smear zone can be predicted by
0 40 80 120 160 the cavity expansion theory and validated by large scale
Time (Days)
laboratory results based on the variation of water content
Figure 22: Variation of excess pore water pressure at 3m below
and soil permeability. It is found that soil unsaturation at
the surface and 0.5m away from centreline for Embankment the vertical drain boundary due to mandrel driving could
TV2 (modified after Indraratna et al., 2005f) delay the excess pore pressure dissipation in the early
stage of consolidation process. The behaviour of soft clay
under the influence of PVD and vacuum application was
described on the basis of numerous case histories where
both field measurements and predictions were available.
An analytical model incorporating vacuum preloading
has been described for both axisymmetric and equivalent
plane strain conditions. A finite element code (ABAQUS)
was employed to analyse the unit cell and compare the
12
numerical results with the analytical approach. A conver- REFERENCES
sion procedure based on the transformation of permeabil-
ity and vacuum pressure was introduced to establish the AIT. 1995. The Full Scale Field Test of Prefabricated Vertical Drains for
relationship between the axisymmetric (3D) and equiva- The Second Bangkok International Airport (SBIA). AIT,Bangkok,
Thailand.
lent plane strain (2D) conditions. The plane strain solution Barron, R.A., 1948. Consolidation of fine-grained soils by drain wells.
was applied for case history analysis, proving its validity Transactions ASCE, Vol 113, paper 2346: 718-724.
to predict real behaviour. Field behaviour as well as model Chai, J. C., Miura, N., Sakajo, S., and Bergado, D. T. 1995. Behavior of
predictions indicate that the efficiency of vertical drains vertical drain improved subsoil under embankment loading. J. Soil
and Foundations, Japanese Geotechnical Society, 35(4), 49-61.
depends on the magnitude and distribution of vacuum Chu, J., Yan, S. W., and Yang, H. 2000. Soil improvement by the
pressure. vacuum preloading method for an oil storage station. Geotechnique,
An accurate prediction of lateral displacement depends 50(6), 625-632.
on the careful assessment of soil properties including the Cognon, J. M., Juran, I and Thevanayagam, S., 1994. Vacuum
consolidation technology- principles and field experience,
overconsolidated surface crust. This compacted layer is Proceedings of conference on vertical and horizontal deformations of
relatively stiff, and therefore it resists inward movement foundations and embankments deformations, College station, Texas.
of the soil after vacuum application. Clearly, the modified Hansbo, S. 1981. Consolidation of fine-grained soils by prefabricated
Cam-clay model is inappropriate for modeling the behav- drains. In Proceedings of 10th International Conference on Soil
Mechanics and Foundation Engineering, Stockholm, Balkema,
ior of a weathered and compacted crust. It may be better Rotterdam, 3, pp. 677-682.
modeled as an elastic layer rather than a soft elasto- Hird, C.C., Pyrah, I.C., and Russell, D. 1992. Finite element modelling
plastic medium. An analysis of the case histories showed of vertical drains beneath embankments on soft ground.
that the vacuum application via PVD substantially de- Geotechnique, 42(3): 499-511.
Indraratna, B., Balasubramaniam, A. S., and Balachandran, S. 1992.
creases lateral displacement, thereby reducing potential Performance of test embankment constructed to failure on soft
shear failure during rapid embankment construction. marine clay. J. Geotech. Eng., ASCE, 118, 12-33.
There is no doubt that a system of vacuum-assisted Indraratna, B., Balasubramaniam, A. S., and Ratnayake, P. 1994.
consolidation via PVD is a useful and practical approach Performance of embankment stabilized with vertical drains on soft
clay. J. Geotech. Eng., ASCE, 120(2), 257-273.
for accelerating radial consolidation. Such a system re- Indraratna, B., Balasubramaniam, A. S., and Sivaneswaran, N. 1997.
duces the need for a high surcharge load, as long as air Analysis of settlement and lateral deformation of soft clay foundation
leaks can be eliminated in the field. Accurate modeling of beneath two full-scale embankments. Int. J. for Numerical and
vacuum preloading requires laboratory and field studies to Analytical Methods in Geomechanics, 21, 599-618.
Indraratna, B., Bamunawita, C., and Khabbaz, H. 2004. Numerical
investigate the exact nature of vacuum pressure distribu- modeling of vacuum preloading and field applications. Canadian
tion within a given soil formation and PVD system. In ad- Geotechechnical Journal, 41, 1098-1110.
dition, a resilient system is required to prevent air leaks Indraratna, B., and Redana, I. W. 1995. large-scale, radial drainage
that can reduce the desirable negative pressure (suction), consolidometer with central drain facility. Australian Geomechanics,
29(103-105).
with time. Indraratna, B., and Redana, I. W. 1997. Plane strain modeling of smear
effects associated with vertical drains. J. Geotech. Eng., ASCE,
123(5), 474-478.
8 ACKNOWLEDGEMENT Indraratna, B., and Redana, I. W. 1998. Laboratory determination of
smear zone due to vertical drain installation. J. Geotech. Eng., ASCE,
125(1), 96-99.
The authors wish to thank the CRC for Railway Engineer- Indraratna, B., and Redana, I. W. 2000. Numerical modeling of vertical
ing and Technologies (Australia) for its continuous sup- drains with smear and well resistance installed in soft clay. Canadian
Geotechnical Journal, 37, 132-145.
port and the Thailand Airport Authority for providing the Indraratna, B., Rujikiatkamjorn C., and Sathananthan, I. 2005a.
trial embankments data. The Cavity Expansion Theory Analytical and numerical solutions for a single vertical drain
application to PVDs was the result of a PhS study by Dr. including the effects of vacuum preloading. Canadian Geotechinical
I. Sathananathan. In addition, assistance of Prof. A.S. Journal, 42: 994-1014.
Indraratna, B., Rujikiatkamjorn C., and Sathananthan, I. 2005b. Radial
Balasubramaniam (formerly at Asian Institute of Technol- consolidation of clay using compressibility indices and varying
ogy, Bangkok, Thailand) is gratefully appreciated. A horizontal permeability. Canadian Geotechinical Journal, 42: 1330-
number of past research students of Prof. Indraratna, 1341.
namely, Sathananathan, Redana, Bamunawita, Bala- Indraratna, B., Rujikiatkamjorn C., and Sathananthan, I., 2005c.
Analytical modeling and field assessment of embankment stabilized
chandran, and Ratnayake, have also contributed to the with vertical drains and vacuum preloading. The Proceedings of the
contents of this keynote paper through their research 16th International Conference on Soil Mechanics and Geotechnical
work. Engineering, 12-16 September 2005, Osaka, Japan, Edited by the 16th
Much of the contents reported in this paper are de- ICSMGE committee, Millpress, Rotterdam, the Natherlands, 1049-
1052.
scribed in greater details in a number of Canadian Geo- Indraratna, B., Rujikiatkamjorn C., Balasubramaniam, A. S. and
technical and ASCE Journal of primary author as listed in Wijeyakulasuriya, V. 2005d. Predictions and observations of soft
the references. clay foundations stabilized with geosynthetic drains and vacuum
surcharge. Ground Improvement Case Histories Book (Volume 3),
Edited by Indraratna, B. and Chu, J., Elsevier, London, 199-230.
Indraratna, B., Sathananthan, I., Bamunawita, C. and A.S.
Balasubramaniam. 2005e. Theoretical and Numerical Perspectives
and Field Observations for the Design and Performance Evaluation of
13
 Embankments Constructed on Soft Marine Clay. Ground
ImprovementCase Histories Book (Volume 3), Edited by Indraratna,
B. and Chu, J., Elsevier, London, Chapter 2:61-106.
Indraratna, B., Sathananthan, I., Rujikiatkamjorn C. and
Balasubramaniam, A. S. 2005f. Analytical and numerical modelling
of soft soil stabilized by PVD incorporating vacuum preloading.
International Journal of Geomechanics, Vol. 5 No. 2, 114-124.
Indraratna, B., Rujikiatkamjorn C., Sathananthan, I., Shahin, M.A., and
Khabbaz, H. 2005g. Analytical and numerical solutions for soft clay
consolidation using geosynthetic vertical drains with special
reference to embankments. The Fifth International Geotechnical
Engineering Conference, Cairo, 55-86.
Jamiolkowski, M., Lancellotta, R., and Wolski, W. 1983.
Precompression and speeding up consolidation. In Proceedings of 8th
European Conference on Soil Mechanics and Foundoundtions,
Helsinki, Finland, 3, 1201-1206.
Johnson, S.J. 1970. Precompression for improving foundation soils. J.
Soil. Mech. Found. Div., ASCE, 96(1):111-114.
Kjellman, W. 1952. Consolidation of clayey soils by atmospheric
pressure. Proceedings of a conference on soil stabilization,
Massachusetts Institute of Technology, Boston, 258-263.
Lekha, K. R., Krishnaswamy, N. R., and Basak, P. (2003). Consolidation
of clays for variable permeability and compressibility. Journal of
Geotechnical and Geoenvironmental Engineering, ASCE, 129(11),
1001-1009.
Nagaraj, T. S., Pandiam, N. S., and Narasimha Raju, P. S. R. 1994.
Stress-state-permeability relations for overconsolidated clays.
Geotechnique, 44(2), 349-352.
Qian, J. H., Zhao, W. B., Cheung, Y. K., and Lee, P. K. K. 1992. The
theory and practice of vacuum preloading. Computers and
Geotechnics, 13, 103-118.
Roscoe, K.H., and Burland, J.B. 1968. On the generalized stress strain
behavior of wet clay. Engineering plasticity, Cambridge Univ. Press;
Cambridge, U.K., 535-609.
Rujikiatkamjorn. 2005. Analytical and numerical modelling of soft clay
foundation improvement via prefabricated vertical drains and
vacuum preloading. PhD Thesis, Universtiy of Wollongong, 263p.
Sangmala, S., 1997. Efficiency of drainage systems of vacuum
preloading with surcharge on PVD improved soft Bangkok clay, ME
Thesis, Asian Institute of Technology, Bangkok, Thailand.
Saye, S. R. 2003. Assessment of soil disturbance by the installation of
displacement sand drains and prefabricated vertical drains.
Geotechnical Special Publication No. 119, ASCE, 325-362.
Sathananthan, I. 2005 Modelling of Vertical Drains with Smear Installed
in Soft Clay. PhD Thesis, Universtiy of Wollongong, 264p.
Sathananthan, I. and Indraratna, B. 2006. Laboratory Evaluation of
Smear Zone and Correlation between Permeability and Moisture
Content. Submitted to Journal of Geotechnical and Geoenvironmental
Engineering, ASCE (in press).
Seah, T. H., and Juirnarongit, T. 2003. Constant rate of strain
consolidation with radial drainage. Geotechnical Testing Journal,
26(4), 1-12.
Tang, M., and Shang, J. Q. 2000. Vacuum preloading consolidation of
Yaoqiang Airport runway. Geotechnique, 50(6), 613-623.
Tavenas, F., Jean, P., and Leroueil, S. 1983. The permeability of natural
soft clays, Part 2: Permeability characteristics. Canadian
Geotechechnical Journal, 20, 645-660.
Yan, S. W., and Chu, J. 2003. Soil improvement for a road using a
vacuum preloading method. Ground Improvement, 7(4), 165-172.
Yoshikuni, H., and Nakanodo, H. 1974. Consolidation of Fine-Grained
Soils by Drain Wells with Finite Permeability. Japan Soc. Soil Mech.
and Found. Eng, 14(2), 35-46.

14