University of Wollongong

Research Online
Faculty of Engineering and Information Sciences -
Faculty of Engineering and Information Sciences
Papers: Part A

2016

Analytical solution and numerical simulation of

vacuum consolidation by vertical drains beneath
circular embankments
Buddhima Indraratna
University of Wollongong, [email protected]

Mojtaba E. Kan
University of Wollongong, [email protected]

David Potts
Imperial College London

Cholachat Rujikiatkamjorn
University of Wollongong, [email protected]

Scott W. Sloan
University of Newcastle

Publication Details
Indraratna, B., Kan, M. E., Potts, D., Rujikiatkamjorn, C. & Sloan, S. W. (2016). Analytical solution and numerical simulation of
vacuum consolidation by vertical drains beneath circular embankments. Computers and Geotechnics, 80 83-96.

Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library:
[email protected]
Analytical solution and numerical simulation of vacuum consolidation by
vertical drains beneath circular embankments
Abstract
This paper presents an analytical solution and numerical simulation of vacuum consolidation beneath a
circular loading area (e.g. circular oil tanks or silos). The discrete system of vertical drains is substituted by
continuous concentric rings of equivalent drain walls. The effectiveness of the vacuum as distributed along the
drain length and the well resistance of the drains are considered. A rigorous solution of radial drainage towards
cylindrical drain walls is presented and compared to numerical FEM predictions. The model is then
successfully adopted to analyse the vacuum consolidation of a circular embankment in the Ballina field testing
facility in Australia.

Disciplines
Engineering | Science and Technology Studies

Publication Details
Indraratna, B., Kan, M. E., Potts, D., Rujikiatkamjorn, C. & Sloan, S. W. (2016). Analytical solution and
numerical simulation of vacuum consolidation by vertical drains beneath circular embankments. Computers
and Geotechnics, 80 83-96.

This journal article is available at Research Online: http://ro.uow.edu.au/eispapers/5688

 Analytical Solution and Numerical Simulation of Vacuum
Consolidation by Vertical Drains beneath Circular Embankments
Buddhima Indraratna
BSc (Hons., Lond.), MSc (Lond.), DIC, PhD (Alberta), FTSE, FIEAust., FASCE, FGS
Distinguished Professor of Civil Engineering, Faculty of Engineering,
Director, Centre for Geomechanics and Railway Engineering; University of Wollongong,
Wollongong City, NSW 2522, Australia

Mojtaba E. Kan
BSc (Hons), MSc, PhD (UNSW)
Research Associate, Centre for Geomechanics and Railway Engineering, School of Civil
Engineering, Faculty of Engineering,
University of Wollongong, Wollongong City, NSW 2522, Australia

David Potts
BSc (Hons, Lond.), PhD (Cambridge), FRAE, FICE
GCG Professor of Geotechnical Engineering, Faculty of Engineering,
Head of Geotechnics, Department of Civil and Environmental Engineering,
Imperial College, London, UK

Cholachat Rujikiatkamjorn
BEng (Hons), MEng (AIT), PhD
Associate Professor, Centre for Geomechanics and Railway Engineering, School of Civil
Engineering, Faculty of Engineering,
University of Wollongong, Wollongong City, NSW 2522, Australia

Scott W. Sloan
BE (Hons I, Monash), MEng, MPhil, PhD (Cambridge), FRS
Laureate Professor of Civil Engineering and Director of Australian Research Council Centre
of Excellence for Geotechnical Science and Engineering,
The University of Newcastle, Callaghan, NSW 2308, Australia

Prepared for submission to: Computers and Geotechnics

No. of Words (without abstract and references): 5788, No of Figures: 16, No of Tables: 3

Page 1 of 50
 Analytical Solution and Numerical Simulation of Vacuum
Consolidation by Vertical Drains beneath Circular Embankments
Buddhima Indraratna, Mojtaba E. Kan, David Potts,
Cholachat Rujikiatkamjorn, and Scott W. Sloan

Abstract

This paper presents an analytical solution and numerical simulation of vacuum consolidation

beneath a circular loading area (e.g. circular oil tanks or silos). The discrete system of vertical

drains is substituted by continuous concentric rings of equivalent drain walls. The

effectiveness of the vacuum as distributed along the drain length and the well resistance of

the drains are considered. A rigorous solution of radial drainage towards cylindrical drain

walls is presented and compared to numerical FEM predictions. The model is then

successfully adopted to analyse the vacuum consolidation of a circular embankment in the

Ballina field testing facility in Australia.

Keywords: Analytical solution, Vertical drains, Consolidation, Vacuum preloading, Circular

Embankments

Page 2 of 50
1 Introduction

Very soft clays are widespread in many coastal regions of Australia and other parts of the

world. These soft clay deposits normally show unfavourable soil properties such as low

bearing capacity and high compressibility. For such soils, appropriate ground improvement

techniques are usually employed to minimize the post-construction settlements and lateral

displacements which may threaten the stability of infrastructure built on them (Indraratna and

Chu, 2005).

Application of prefabricated vertical drains (PVDs) is one of the most widely used ground

improvement techniques for improving the mechanical properties of soft clay deposits. The

PVDs expedite the progress of soil consolidation by shortening the length of the drainage

path and enhancing radial drainage. PVDs are artificially created vertical drainage boundaries

which accelerate lateral (radial) flow from the surrounding soil, thereby increasing the rate of

consolidation significantly. The consolidation of the soil results in higher shear strength and

helps to reduce the post-construction settlement of the super-structure (Holtz et al., 1991).

However, preloading with PVDs can be relatively slow, as a staged construction is generally

required to prevent instability. In such cases, the application of vacuum pressure would

significantly reduce the consolidation time (e.g. Indraratna et al., 2005, Dam et al., 2006).

The behaviour of vertical drains was first solved analytically by Barron (1948) and Richart

(1957) based on a unit cell concept. The unit cell represents a single drain surrounded by a

soil annulus under axisymmetric conditions (three-dimensional, 3D). Hird et al. (1992)

extended the unit cell concept to the plane strain condition (two-dimensional, 2D). The unit

cell concept is accurate when applied at the embankment centreline, where the lateral

displacements are negligible. In practice, the subsoil is usually not uniform, and the process

of consolidation is not always a one-dimensional problem (Indraratna et al., 1992). Numerical

modelling of multidrain systems in plane strain was further improved by Indraratna and
Page 3 of 50
Redana (2000), who introduced a mathematical technique to convert axisymmetric properties

to equivalent 2D plane strain condition and also considered the smear effect caused by

mandrel intrusion.

The vacuum preloading (VP) method was initially introduced in Sweden by Kjellman (1952)

for cardboard wick drains. The 2D analysis is also applicable for VP in conjunction with

vertical drains (e.g. Gabr and Szabo, 1997). Mohamedelhassan and Shang (2002) proposed an

analytical solution for radial consolidation with vacuum application. Indraratna et al. (2005)

extended the unit cell radial consolidation theory for vacuum application, based on the

equivalent plane strain condition and considering the potential vacuum loss along the drain

length. It can be noted that most of the previous studies have been devoted to modelling the

multidrain systems corresponding to an embankment strip loading in 2D plane strain. So far,

no study has been conducted to model vacuum consolidation via PVDs beneath a circular

loaded area, where the system conforms to an axisymmetric problem. Ground improvement,

however, may be needed for most heavy circular structures, such as oil or water storage

tanks, silos, and heavily loaded roundabouts in commercial areas. Indraratna et al. (2008) first

introduced the concept of concentric rings of equivalent drain walls for simulating circular

embankments with PVDs, but this work was limited to conventional preloading with vertical

drains and no vacuum pressure.

The main objective of this paper is to introduce a comprehensive analytical solution for

vacuum preloading in conjunction with vertical drains beneath a circular foundation. The

theory of ring walls is extended to accommodate vacuum preloading, considering the possible

loss of vacuum with depth and the effect of well resistance on consolidation. The finite

element model (FEM) in PLAXIS (Brinkgreve et al., 2015), incorporating the author’s

solution, is then validated for a single ring situation. The FEM code PLAXIS is then used to

analyse the performance of a full-scale test embankment to be constructed near the Pacific

Page 4 of 50
Highway in Ballina, NSW, Australia. The effect of vacuum pressure termination on the

consolidation response and post-construction settlements are also analysed and a novel

approach to determine the optimum time for terminating the vacuum application is presented

and examined for the design of the Ballina circular embankment.

2 Mathematical Formulation

Vertical drains are generally installed either in equilateral triangular or square patterns. It is

noted that the square pattern can be more easily controlled in the field, although the triangular

pattern may give more uniform settlement (Rixner et al., 1986). The vacuum consolidation of

soil around a single vertical drain can be readily analysed as a unit cell (Indraratna et al.,

2005). However, to analyse a multidrain system under an axisymmetric condition, the

equivalent soil parameters that give the same time-settlement response in the field must be

determined. Indraratna et al. (2008) proposed a transformation scheme for axisymmetric

problems in which each drain is assumed as a part of the concentric cylindrical drain wall

with an increasing perimeter with the radial distance from the centreline, as shown in Figure

1. In this section, the ring wall theory is extended to incorporate the concept of vacuum

consolidation for the analysis of circular loading, such as silos or oil and water storage tanks.

2.1 Assumptions

The main assumptions made in developing the analytical solutions are summarized below:

 Equal strain assumption and small strain theory are valid. The flow in the soil mass is

assumed to be laminar and Darcy’s law is adopted.

 Only vertical strains are allowed, i.e. at the centreline of a relatively large loading

area, volume change is due to settlement only, and lateral displacements are

negligible.

Page 5 of 50
  The soil is fully saturated and homogeneous, and the permeability of the soil is

assumed to be constant during consolidation.

 Well resistance is taken into account. It is assumed that well resistance is constant

during consolidation.

 Each set of vertical drains located at the same radial distance from the line of

axisymmetry is modelled as a continuous cylindrical drain wall of radius ri  i.S ,

where S is the spacing of the drains and i is the number of that set, as shown in

Figure 1(c)

 It is assumed that the cylindrical drain wall has a negligible thickness.

 Each cylinder is assumed to be impermeable with respect to the outer and inner

boundaries, and has an internal horizontal (radial) flow.

 Smear effects around the drain walls are not incorporated directly in the equations. It

is assumed that the smear effect due to drain installation can be taken into account in

by calculating the reduced lateral permeability of the soil (Indraratna and Redana,

1997).

 The vacuum pressure distribution along the drain boundary is considered to vary

linearly from  p0 at the top of the drain to k1 p0 at the bottom of the drain. The

vacuum pressure is assumed to be evenly distributed in the horizontal direction

between the adjacent drains.

2.2 Analytical Solution

Considering the inner hollow cylindrical soil wall, the flow rate in the radial direction from

the inner impermeable boundary to the hollow cylindrical drain wall is expressed by Darcy’s

law, i.e. in zone (1) of Figure 2(a):

Page 6 of 50
 Q k h u
 A (1)
t  w r

where Q is the flow in the soil mass, u is the excess pore pressure due to preloading, and A

is the cross-sectional area of the flow at distance r and can be expressed as (2 rdz ) where

dz is the height of an arbitrary thin layer of the soil as shown in Figure 2(b).  w is the unit

weight of water, t is time and kh is the horizontal permeability coefficient of soil.

The rate of volume change in the vertical direction of the soil mass can be expressed by:

V  v  
2
S
   ri    r 2  dz (2)
t t  2 

where V is the volume of the soil mass and  v is the volumetric strain.

The rate of radial flow is assumed to be equal to the rate of volume change of the soil mass in

the vertical direction, therefore, by rearranging equations (1) and (2), the gradient of excess

pore pressure can be derived as:

u1  w  v 1  
2
S
  ri    r 
2
(3)
r 2kh t r  2 

where the index 1 refers to zone 1 in Figure 2(a).

Similarly, the gradient of excess pore pressure in Zone (2) of Figure 2(a) is determined by:

u2  w  v 1  
2
S
  ri    r 
2
(4)
r 2kh t r  2 

Page 7 of 50
Considering a horizontal cross-sectional slice with thickness dz in Figure 2(b), the change of

flow in the z direction of the drain from the entrance to the exit of the slice, dQz , is then

expressed by:

qw  2 u
dQz  dzdt (5)
 w z 2

where the flow term qw represents well resistance.

The total change in flow from the inner impermeable boundary to the exit face of the slice

(dQ1 ) is given by:

2 ri kh u1
dQ1  dzdt (6)
 w r

Similarly, the total change in flow from the outer impermeable boundary to the exit face of

the slice (dQ2 ) is equal to:

2 ri kh u2
dQ2  dzdt (7)
 w r

If the water is assumed to be incompressible, then the following equation should be satisfied:

dQz  dQ1  dQ2  0 (8)

where dQz is the total change in flow in the vertical direction. At the drain boundary, it is

assumed that a sudden drop in pore water pressure does not occur. Therefore, substituting

equations (5), (6) and (7) in equation (8) yields:

 u1   u2  qw   2u 
       2  0 (9)
 r r ri  r r ri 2 ri kh  z r ri

Page 8 of 50
Equation (9) governs the dissipation of pore pressure in the horizontal direction, and can be

solved and simplified to yield the average excess pore water pressure, u , in a given cylinder.

Details of the derivations are given in the Appendix, where it is shown that u can be

expressed as:

 w  v 1 po
u d e 2  ring  (1  k1 ) (10)
k h t 8 2

where

2   i 
ring   2i 3  1.5i  2(i  0.5)4 ln  
i   i  0.5 
(11)
 i  0.5  4kh L i 
2
+2(i  0.5)4 ln    
 i  3qw 

and S   de , with   0.887 and 0.952 for drains installed in a square pattern and an

equilateral triangular pattern, respectively. Here de is the equivalent diameter of soil cylinder

around a single drain.

If the well resistance is ignored, i.e. when qw is large enough, ring becomes:

2   i   i  0.5  
ring   2i 3  1.5i  2(i  0.5)4 ln   +2(i  0.5) ln 
4
 (12)
i   i  0.5   i 

The value of ring in equation (12) converges to 2 2 / 3 for all values of i  4 , as shown in

Figure 3, and therefore equation (11) can be rewritten as:

2 2  2kh L2 
ring  1   (13)
3  qw 

Note that for small number of ring walls ( i  4 ), equation (12) is still required to determine

the value of ring rather than the limiting value of 2 2 / 3 .

Page 9 of 50
Combining equation (10) with the well-known compressibility relationship

 v / t = – mv u / t gives:

w u 1 2 p
u  mv d e  ring  o (1  k1 ) (14)
kh t 8 2

where m v is the coefficient of volume compressibility for one-dimensional compression.

Rearranging equation (14) and then integrating by applying the boundary condition u  1 at

t  0 gives:

u  (1  k1 ) p0   8Tring  (1  k1 ) po
 1   exp     (15)
1  21   ring  21

ch ,ring t
where Tring  and ch ,ring is the coefficient of consolidation in radial direction for the
d e2

ring. The average degree of horizontal consolidation can be found from (Indraratna et al.,

2005):

1  u 1 1  u
Uh   (16)
1  u  1  1  (1  k1 ) po 2

where u 1 can be calculated by equation (15) when t   and can be taken as

 (1  k1 ) po 21 .

2.3 Equivalent Parameters for Multidrain Analysis under 2D Axisymmetric

Condition

Equivalent parameters can be determined either by geometric transformation or permeability

transformation (or both) to minimize the disparity between the two methods, plain strain and

axisymmetric (Indraratna and Redana, 2000). For circular loading, the proposed procedure

Page 10 of 50
for conversion can be based on the equivalent average excess pore pressure by maintaining

the geometric equivalence. In this procedure, the size and relative distance of the drain wall

from the centre of the embankment under 2D conditions are assumed to be the same as those

for the 3D condition.

For a vertical drain with vacuum preloading under axisymmetric conditions, the average

excess pore pressure in a unit cell can be given by (Indraratna et al., 2005):

 w  v 1 po
u d e 2  ax  (1  k1 ) (17)
kh t 8 2

where

 n k 3 2kh L2 
ax  ln    h ln  s     (18)
  s  ks 4 qw 

and n  de dw , s  ds dw , and ds and d w are the diameter of the smear zone and vertical

drain, respectively. ks is the coefficient of horizontal permeability in the smear zone.

At a given stress level and at each time step, the average excess pore pressure for both the

unit cell and a unit of revolving prism of soil are made the same by equating equation (17)

with equation (10). The equivalent permeability for the multidrain system under

axisymmetric conditions can then be expressed by:

2 2  2kh L2 
1  
kh , ring 3  qw 
 (19)
kh   n  kh 3 2kh L2 
ln
    ln   
s  
  s  ks 4 qw 

where the equivalent vacuum pressure for both the two and three dimensional axisymmetric

conditions are assumed to be the same.

Page 11 of 50
3 Model Verification Using FEM

The finite-element program PLAXIS (Brinkgreve et al., 2015), based on Biot’s consolidation

theory, was used to validate the proposed analytical solution. Indraratna et al. (2005) have

reported the results of analytical and numerical modelling of unit cells of soft estuarine

Sydney clay with and without vacuum for the corresponding material parameters. Using the

same soil in the current study, a single ring drain of soft clay is simulated under axisymmetric

conditions. Two cases are considered, Case A: a 10 m deep ring with constant vacuum

pressure with depth (i.e. k1  1.0 ), and Case B: a 50 m deep ring with decreasing vacuum

pressure to be halved at 50 m depth (i.e. k1  0.5 ). In both cases, the width of the ring is

1.2 m and the distance between the drain and axis of symmetry is assumed to be 6 m, i.e.

ri  6 m or i  6 /1.2  5 . The drain is 50 mm in diameter and surrounded by a smear zone of

170 mm in diameter. A total of 1840 and 8666 axisymmetric triangular elements (six-node

quadratic displacement and linear pore pressure) were used in the finite element discretization

to simulate the 10 m and 50 m deep rings, respectively (see Figure 4(a)). Each PVD is

simulated using a linear drain element which acts as an internal free drainage boundary. The

top, bottom, and outer boundaries were set as impermeable, and therefore drainage was

limited to the horizontal direction towards the PVD. The relevant boundary conditions are

shown on Figure 4(b). After establishing the in-situ stresses in the model with K0=1, first the

installation of PVD was simulated by activation of the drain element in the model and then a

surcharge load of 1  50 kPa was applied on top of the ring. To capture the equal-strain

condition, rigid elements were used at the top of the soil surface where only vertical

displacement was only allowed to prevent any rotation. Application of surcharge was

followed by a vacuum pressure of p0  50 kPa applied along the drain length. For the

Page 12 of 50
current FEM simulations conducted using PLAXIS, the fully implicit time-marching scheme

with a default error tolerance of 1% was adopted to ensure adequate and swift convergence.

The soil behaviour is assumed to be linear elastic with mv =0.001 m 2 /kN , and the case of

zero lateral displacement was imposed (Poisson’s ratio = 0) for the unit cell. The values of ch

= 0.32 m2/year and de =1.354 m (equivalent to S=1.2 m for square drain pattern) were used.

The horizontal undisturbed soil permeability ( kh ) was taken as 10−10 m/ s, and the ratio of the

undisturbed permeability to the smear zone permeability kh / ks was assumed to be 3.0

(Indraratna and Redana, 2000).

Figure 5 shows a comparison between the analytical and numerical models for Cases A and

B. The normalized dissipation rate of excess pore water pressure with elapsed time is

presented in Figure 5. The analytical solution presented is based on the calculations using

equation (15), employing relevant material parameters for the clay. The ratio of pore water

pressure dissipation in the numerical simulations is calculated in two different ways: first,

using the average values of pore pressure at key points of the model (dashed line); second,

using the ratio of the current settlement (St ) to the total settlement (S f ) , shown with a

continuous line:

St  u u S  u 
 1   1 t 1   (20)
S f  1  u 1 Sf  1 

where:

u p0 (1  k1 )
 (21)
1 2 1

and u represents the ultimate average pore water pressure. It can be seen in Figure 5 that the

results of the numerical simulations are very close to the analytical solutions. The simulated
Page 13 of 50
curve for excess pore pressure based on the settlement is closer to the analytical results for

t  350 and t  500 days, for cases A and B respectively. As the elapsed time exceeds these

thresholds, the curves based on average excess pore pressure show better agreement with

those predicted by the analytical solution. Small deviations are noted for the range 15-300

days with a maximum error of less than 3% in the degree of consolidation. The above

matching procedure therefore confirms the reliability of the proposed procedure after

transforming the true field condition to the equivalent axisymmetric drain rings.

4 Application to a circular embankment currently being constructed

along Pacific Highway in Ballina

Design, construction and monitoring of the performance of a vacuum preloading circular

embankment at the National geotechnical Field Testing Facility (BFTF) in Ballina NSW are

now considered. The conversion method developed in this paper was used to perform

numerical simulations of this VP circular embankment in the preliminary design stage. The

VP circular embankment is being constructed and has a total height of 3.5 m and a crest

diameter of 20 m, with side slopes of 1.5H:1V. The plan and typical cross section of the

embankment are shown in Figure 6. The loading response of the VP circular embankment is

simulated by a 2D equivalent axisymmetric model in PLAXIS (Brinkgreve et al., 2015) using

the theory in the current study.

4.1 Material parameters

The material parameters are selected based on reported properties for the soil layers in the

literature, mainly using the parameters reported in Indraratna et al. (2012) for the Ballina

Bypass project which was located approximately 1 km away from the intended location for

construction of the VP circular embankment. Results of a series of more recent site

Page 14 of 50
investigations at the Ballina Site, as reported in Li et al. (2014), Kelly et al., 2014 (2014), and

Pineda et al. (2014, 2016), are also taken into account in current study.

Based on the published data regarding soil layers in the Ballina area, the required material

parameters in the present study are listed in Table 1. The material parameters for the crust,

sand/clay transition and fine sand layers are based on a Mohr-Coulomb model, while for the

soft clay layer the ‘Soft Soil’ model of PLAXIS is selected for simulations. The permeability

of the soil layers in the horizontal direction ( kh ), and the equivalent ring drain ( k h , ring ), are

listed in Table 1 and will be discussed further in the following sections. The embankment fill

is assumed as Mohr-Coulomb material with the properties given in Table 2.

4.2 Mesh discretization and model set up

The proposed VP circular embankment at the Ballina Site was simulated in PLAXIS, using a

2D axisymmetric model and the associated conversion procedure (Eq. 19). The vertical drains

are installed in a square pattern. The model is extended laterally for 50 m in order to

minimize the effects of the lateral boundaries, as shown in Figure 7. The smear effect of the

PVD installation is taken into account indirectly by reducing the horizontal permeability ( kh )

of the soil in the surrounding area. The corresponding mesh, boundary conditions and

selected control points of the VP circular embankment are shown on Figure 7. A total number

of 4535 triangular 6-node elements are used for the discretization. The bottom and side

boundaries were set as impermeable to simulate the axisymmetric condition. The water table

is assumed to be at the ground surface and the initial stress state in the ground is simulated

using the K0 values listed in Table 1. PVDs are simulated using drain elements with the

vacuum feature. The vacuum feature in PLAXIS is based on reduction of the pore-water

pressure head to simulate the applied vacuum.

Page 15 of 50
The fully implicit time marching scheme of PLAXIS was used in the simulations (Brinkgreve

et al., 2015). Preliminary analyses show that large deformations may occur in soft soil layers,

therefore the influence of the mesh geometry change on the equilibrium conditions is

considered in the FEM simulations capturing the finite strain approach. The finite strain

computation procedure in PLAXIS is based on a Lagrangian formulation (Bathe, 1982)

which updates the finite element mesh and the stiffness matrix at the beginning of each load

step as the iteration proceeds.

4.3 Smear characteristics and conversion of the permeability

The extent of the smear zone created by drain installation can be obtained by assessing the

variation of compressibility, permeability and water content of the soil samples obtained

along the radius from the PVDs (Indraratna and Redana, 1998, Indraratna et al., 2014). In the

present work, the size of the smear zone is calculated based on the available information in

the literature. The diameter of the drains (dw) is 34 mm, and the size of the rectangular

mandrel is 125×60 mm with an equivalent diameter (dm) of 92.5 mm. Indraratna et al. (2012)

have used a smear size ratio of s  ds / dw  3 for the past Ballina Bypass trial embankment.

Parsa-Pajouh et al. (2014) showed that the ratio of s  4 gives the best match between the

numerical model and field observations of the Ballina Bypass trial embankment. In

multi-layer soils in the field, the smear effects might be due to combination of both mandrel

action and the down-dragged soils from the upper layers (Casagrande and Poulos, 1969).

Indraratna et al. (2014) performed field sampling and laboratory investigations to characterise

the smear zone during installation of band shaped vertical drains at the Ballina Site. Based on

the variations in the permeability, they found that the smear zone could be at least 6 times

greater than the equivalent dimension of the mandrel, while the horizontal permeability

Page 16 of 50
within the smear zone varied from 0.2 to 0.9 of the lateral permeability of the undisturbed

soil.

Based on past background, the diameter of the smear zone was taken as 200 mm in the

current study, which resulted in a smear size ratio of s  5.9 . The horizontal permeability in

the smear zone was assumed one third of the horizontal permeability in the undisturbed zone.

To calculate the horizontal permeability of the equivalent drain rings in the 2D axisymmetric

model, the conversion procedure described earlier in this paper has been used. Considering

the short length of the drains, the effect of well resistance can be ignored so that equation (19)

simplifies to:

2 2
k h , ring 3
 (22)
 n  kh
ln    ln  s   0.75
kh
 s  ks

where n  de / dw  (1.2 1.128) 0.034  39.8 , s  ds / dw  5.9 , and   1/1.128  0.887 .

Assuming kh / ks  3 , the permeability of the ring drain can be calculated. The conversion

equation (Eq. 22) is used to calculate the corresponding values of the permeability of the

equivalent rings in Table 1. This simplification significantly reduces the number of elements

for the domain discretization, and accelerates the speed of convergence extensively.

4.4 Loading history

The proposed time history of fill construction and application of vacuum was considered, and

this comprised of constructing a 0.6 m thick working platform (sand) in 5 days, installation of

PVDs and monitoring instruments in 20 days, placing a 0.6 m thick sand blanket and the final

layer of random fill with total height of 2.3 m in 25 days, and then the application of a 70 kPa

vacuum. The idealized time history of embankment construction and vacuum loading adopted

Page 17 of 50
in the FEM simulations is shown in Figure 8(a). Based on the unit weights of the fill

materials given in Table 2, the total preloading due to both fill and the vacuum will be around

130 kPa, i.e. equivalent to a 7.5 m high embankment. Construction of such a high

embankment on this soft soil would require staged construction and could be very time-

consuming as shown in Figure 8(b). Based on the material parameters adopted in Table 1, the

mobilized shear strength will range from 1.5 to 3 kPa at the top to the bottom of the soft clay

and rapid construction of a high embankment on this ground condition will cause slide

failure. The time history for construction of the 7.5 m high embankment in Figure 8(b) is

therefore based on a required elapsed time between each fill placement stage to gain enough

strength in the underlying soil strata. By applying a vacuum head however, the duration of

pre-consolidation could be reduced. The effect of vacuum application on the required time to

achieve a certain level of consolidation will be further discussed in the following sections.

4.5 Results and discussions

Two possible field conditions are considered herein: application of vacuum after construction

of a 3.5 m high embankment following the loading history shown in Figure 8(a), and the

construction of a 7.5 m high embankment (no vacuum) to provide the same equivalent total

pressure following the loading history shown in Figure 8(b) – introduced as ‘Vacuum+Fill’

and ‘Fill Only’, respectively. In order to avoid any boundary effects and to keep the geometry

similar for both conditions, the 7.5 m high embankment is simulated by applying a

corresponding uniform load on the crest of the 3.5 m high embankment, as shown in Figure

7. The staged construction of 7.5 m high embankment is simulated in such a way that, at each

stage, the application of the new load increment (surcharge) will take place after 60%

dissipation of the excess pore pressure at the middle of the soft clay layer. This is to ensure

that the soft clay has gained adequate shear strength to bear the associated load without

failure or significant increase in settlement.

Page 18 of 50
The settlement patterns of the ground surface after one year of consolidation with and without

vacuum are shown in Figure 9. It can be seen that the total maximum settlement after one

year of vacuum application is 1.99 m, i.e. approximately twice that of the maximum

deformation without vacuum during the same period of time under the surcharge loading of

the 3.5 m high embankment.

The time history of the ground settlement at point A, located at the centreline of the

embankment, is shown in Figure 10(a). It can be seen that for ‘Vacuum+Fill’ condition, the

total settlement is around 1.8 m after 280 days, while for the ‘Fill Only’ condition this

settlement takes almost one year longer. From the ‘Vacuum+Fill’ curve it is obvious that the

rate of settlement decreases after almost three months of vacuum application. The same

observation can be seen from the time histories of excess pore pressure at the middle of the

soft clay (point C) as shown in Figure 10(b). As expected, the final excess pore pressure at

point C for the ‘Fill Only’ condition approaches zero, and for the ‘Vacuum+Fill’ condition it

eventually reaches -70 kPa.

The contours of the horizontal displacements after one-year of vacuum application are shown

in Figure 11. This shows that the horizontal displacements at the toe of the embankment are

inward due to the effect of isotropic vacuum compression. This is further highlighted in

Figure 12, where the profiles of the horizontal displacements at the toe of the embankment at

various vacuum periods (up to 12 months) are compared with the corresponding profiles for

the ‘Fill Only’ condition. Figure 12 shows that if a 7.5 m high embankment is constructed

without any vacuum pressure, then the entire horizontal displacement is outward with a

maximum displacement close to 0.21 m. Application of the vacuum decreases the outward

displacements and causes inward displacements in the soft clay (e.g. Chai et al., 2006, Chai

and Carter, 2011). The maximum horizontal displacement after one year is 126 mm which

occurs at a depth of 3.2 m below the surface. Only 3 months of vacuum application will be

Page 19 of 50
sufficient to neutralize the outward displacement along much of the soil profile, and this will

control the maximum outward displacement to around 41 mm. The time-dependent horizontal

displacements for five selected key points located beneath the embankment toe are shown in

Figure 13. Point D, located at the interface of surface crust and upper soft clay, shows

outward horizontal displacement during the first 3 months of vacuum application and inward

displacement afterwards. All other four points mostly show inward horizontal displacements

during the period of vacuum application. Point E (located at a depth of 4 m) always has the

maximum inward displacements among the other points.

4.6 Optimum time for the termination of vacuum

The advantage of applying vacuum pressure diminishes with time as the primary

consolidation progresses. As can be seen in Figure 10, the vacuum pressure accelerates the

consolidation settlement significantly at the start, but subsequently shows marginal returns

after about 120 days.

The cost savings for switching off of the vacuum pump early can be significant. This was

demonstrated through a field trial at the Port of Brisbane reclamation site (Indraratna et al.,

2011). Kianfar et al. (2015) performed a series of Rowe cell tests to study the suitable time

for terminating the vacuum, and it was found that the amount of undissipated pore water

pressure and its current rate of dissipation give a good indication of the best time to switch

off the vacuum pumps. Kianfar et al. (2015) concluded that if the excess pore pressure is

negative before vacuum removal, then the excess pore water pressures attain a relatively

steady condition with no further settlement thereafter due to an existing fill. In the present

study, first the unit cell analysis mimicking the laboratory testing by Kianfar et al. (2015) was

conducted. It is confirmed that the numerical predictions can simulate the removal of vacuum

application with acceptable accuracy compared with the laboratory observations.

Page 20 of 50
To study the effect of vacuum removal on the performance of vacuum consolidation in

prototype scale, three cases were considered:

 Case A: Vacuum removal after 1 month

 Case B: Vacuum removal after 3 months

 Case C: Vacuum removal after 12 months

In each case, the vacuum was switched off after a certain time and the remaining

consolidation was assumed to take place under the effect of surcharge pressure only. Results

of the FEM simulations for these three cases are shown in Figure 14.

Figure 14(a) shows the time history of excess pore pressure at the middle of the soft clay

(Point C) while the settlement curve at Point A is given in Figure 14 (b). In Case A, the

vacuum removal takes place when the excess pore water pressure at Point C is still positive,

i.e. the vacuum pressure does not have adequate time to lower the total pore pressure below

the hydrostatic pore pressure at the same depth. So after vacuum removal, the pore water

pressure first decreases swiftly due to the redistribution of pressure and then approaches the

zero condition after almost 30 days. Meanwhile an extra settlement of 210 mm takes place. In

Case B, the excess pore water pressure upon vacuum removal is negative. The redistribution

of stress after vacuum termination causes a much smaller change in the excess pore water

pressure and just 30 mm of extra settlement occurs during a period of 650 days. In Case C,

the vacuum is applied for 12 months, and the extra settlement after vacuum termination is

negligible while the excess pore water pressure after vacuum termination reach zero within

150 days. In summary, this exercise clearly shows that for the Ballina circular embankment,

the vacuum pumps can be stopped after 3 months and the subsequent limited settlement can

occur under surcharge pressure.

Page 21 of 50
The effectiveness of vacuum application is further investigated considering the

pre-consolidation pressure increments. Figure 15 shows the increments of pre-consolidation

pressure ( pc ) with respect to the initial conditions, along a profile passing through the

centerline of the embankment. As expected, as the duration of vacuum application increases,

a higher pre-consolidation pressure is observed. However, the rate of increase in the pre-

consolidation pressure in the soft clay layer is not constant. While an average increment of

around 25 kPa can be observed during the first 3 months (8.3 kPa per month), the increment

during the next 9 months is similar, at around 25 kPa (2.8 kPa per month), and about 15 kPa

during the next 12 months (1.3 kPa per month). The diminishing rate of increment in the

pre-consolidation pressure implies that the effectiveness of the vacuum is reduced

significantly after almost 3 months.

It should be noted that in terms of analytical solutions for vacuum termination, there are only

limited documented solutions in the literature, attributed to inspiring works of Pariseau

(1999), Belotserkovets and Prevost (2011) and Selvadurai and Suvorov (2012, 2014) , which

may be considered to extend the Biot type models to reflect the role of unloading due to

vacuum termination.

4.7 Effect of smear characteristics on vacuum termination

The size of smear zone for the circular VP embankment was assumed to be 200 mm in

diameter around each drain, and the smeared horizontal permeability was taken as one-third

of the undisturbed soil. A sensitivity analysis is performed to investigate the range of

expected outcomes for a variety of smear characteristics.

Parsa-Pajouh et al. (2014) showed that, based on past studies, the extent of the smear zone

may vary from 1.6 to 7 times the equivalent drain size and the range of the undisturbed to

smeared permeability ratios ( kh / ks ) might be 1.3–10. Very high values of kh / ks are mainly

Page 22 of 50
reported for band shaped drains (e.g. Bergado et al., 1993) and for circular drains lower

values are reported in the literature (e.g. Hansbo, 1997, Indraratna and Redana, 1998).

In the current study, a sensitivity analysis is performed for ratios of d s / d w in the range of 2

to 6, and kh / ks varied from 1.5 to 5. In each analysis, the smear zone characteristics are

altered accordingly to observe the performance of vacuum pressure application. Results of

this sensitivity analysis in terms of the time history of excess pore water pressure at the

middle of the soft clay layer are shown in Figure 16. The required time to achieve the zero

excess pore water pressure condition is indicated by the bullet points in Figure 16, and the

corresponding values are summarized in Table 3. It is clearly evident that as the permeability

of the smear zone increases or the size of the smear zone decreases, the desirable time for

terminating the vacuum pressure decreases. For instance, the range of required time to switch

off the vacuum pressure varies from a period of 38 days for kh / ks  1.5 and d s / d w  2 to

121 days for kh / ks  5 and d s / d w  6 . This broad range was deliberately selected to examine

the worst possible and the optimum scenarios when adopting the smear zone parameters in

preliminary design. However, this aspect requires further insightful investigation that has not

been covered within the scope of this study, but will be elaborated in the future, when the

relevant field data become available for the vacuum stabilized circular embankment.

5 Conclusions

Construction of circular loaded structures such as oil tanks, silos, heavily built-on

roundabouts, and circular embankments on soft soils requires effective consolidation to

achieve higher shear strengths to minimize the post construction settlements. This paper

presented a new technique to model vacuum consolidation via vertical drains beneath a

circular loaded area, where the pattern of vertical drains could be transformed to a series of

Page 23 of 50
equivalent concentric (cylindrical) drain walls, and an analytical model for radial drainage

was formulated. The proposed analytical formulation was validated using FEM simulations in

the finite element code PLAXIS.

Application of the proposed model in practice was further examined by FEM simulation of a

multidrain full-scale circular embankment at the Ballina Site, NSW, Australia. The

settlements, excess pore-water pressures, and lateral movements in the embankment were

analysed by employing the proposed conversion procedure. The termination time of the

vacuum was also studied using the FEM simulations to provide insights into the optimum

duration for application of the vacuum. It was concluded that the criteria of achieving zero or

small negative excess pore water pressure prior to vacuum terminations works well for the

Ballina circular embankment. Using these criteria, it was shown that the vacuum termination

can occur in a 3 months period and the corresponding post-vacuum settlements are negligible

under the existing fill. The vacuum termination was further studied by tracking the rate of

increase in the pre-consolidation pressure along the centreline profile of embankment and it

was shown that after the first 3 months, this significantly reduces and therefore the

effectiveness of vacuum application would be insignificant. Sensitivity analyses were also

performed to show the effect of smear characteristics on the optimum time of the vacuum

switch off. Application of the suggested method for vacuum termination in practice requires

selection of an appropriate representative point within the clay layer, and this selection

requires further investigation.

The analytical solutions, validated by FEM modelling, provide a sound basis for the design of

heavy circular foundations on soft clay. Moreover, the validation of the Authors' theoretical

formulations for the proposed circular embankment near the Pacific Highway in Ballina with

field measurements, is expected to begin from September 2016, and a Class A prediction will

be submitted by the Authors as a follow up Technical Note in early 2017.

Page 24 of 50
6 Acknowledgements

This research has been funded under the Australian Research Council (ARC) Linkage scheme

as a collaboration among the Universities of Wollongong and Newcastle in Australia and

Imperial College of Science, Technology and Medicine, University of London, UK. The

authors acknowledge the support of the industry partners Menard-Bachy, Coffey

Geotechnics, Douglas Partners, the National Jute Board of India, and Soilwicks Pty Ltd.

7 Appendix: Derivation of the analytical solution

By substituting equations (3) and (4) in equation (9), we have:

 w  v 1  
2 2
S  S qw   2u 
 i
r    r 2
 ir    r 2
i    2  0 (A1)
2kh t ri  2
i
 2  2 ri h  z  r  ri
k

Therefore:

  2u   w S 2  v
 2   (A2)
 z r ri 2qw t

Integrating equation (A2) in the z direction and subject to the boundary conditions (i) at

z  0 , ur  ri   p0 (applied vacuum pressure), and (ii) at z  L ,  u z r  r  p0 (1  k1 ) / L , the

i

excess pore water pressure at r  ri can be determined by:

 w S 2  v  z
ur  ri  
4 qw t
z 2

 2 zL  p0 1  (1  k1 ) 
 L
(A3)

Integrating equations (3) and (4) in the radial direction with the boundary conditions given in

equation (A3), u1 and u2 can be expressed as:

Page 25 of 50
  w  v  
2
S   r  1 2 2 kh S 2 2
u1   ri   ln    r  ri 
2kh t  2   ri  2 2 qw

z  2 zL    


 z
 p0 1  (1  k1 )  (A4)
 L

where ri  S 2  r  ri , and

 w  v  
2
S   r  1 2 2 kh S 2 2
u2   ri   ln    r  ri 
2kh t  2   ri  2 2qw

z  2 zL    


 z
 p0 1  (1  k1 )  (A5)
 L

for ri  r  ri  S 2 .

The average excess pore pressure u is determined from:

L ri L ri  S / 2
u  L[( ri  S ) 2  ( ri  S ) 2 ]    2 u1rdrdz    2 u2 rdrdz (A6)
2 2 0 ri  S / 2 0 ri

Integrating equation (A6) after substituting equations (A4) and (A5), the average excess pore

water pressure can be determined by the following:

 w  v S 2   i 
u  2i 3  1.5i  2(i  0.5)4 ln  
kh t 8i   i  0.5 
(A7)
 i  0.5  4kh L i  po
2
+2(i  0.5) ln 
4
   (1  k1 )
 i  3qw  2

Page 26 of 50
8 References

Barron, R. A. 1948. Consolidation of fine-grained soils by drain wells. Transactions of

American Society of Civil Engineers, Paper No. 2346, 113, 718–742.
Bathe, K. 1982. Finite element analysis in engineering analysis. Prentice-Hall.
Belotserkovets, A. & Prevost, J. H. 2011. Thermoporoelastic response of a fluid-
saturated porous sphere: An analytical solution. International Journal of
Engineering Science, 49, 1415-1423.
Bergado, D., Mukherjee, K., Alfaro, M. & Balasubramaniam, A. 1993. Prediction of
vertical-band-drain performance by the finite-element method. Geotextiles and
Geomembranes, 12, 567-586.
Brinkgreve, R., Kumarswamy, S. & Swolfs, W. (eds.) 2015. PLAXIS 2D 2015, User’s
Manual, Netherlands: Plaxis bv.
Casagrande, L. & Poulos, S. 1969. On the effectiveness of sand drains. Canadian
Geotechnical Journal, 6, 287-326.
Chai, J.-C., Carter, J. & Hayashi, S. 2006. Vacuum consolidation and its combination
with embankment loading. Canadian Geotechnical Journal, 43, 985-996.
Chai, J. & Carter, J. P. 2011. Deformation analysis in soft ground improvement, Springer
Science & Business Media.
Dam, T. K. L., Sandanbata, I. & Kimura, M. 2006. Vacuum Consolidation Method–
Worldwide Practice and the Latest Improvement in Japan.
Gabr, M. & Szabo, D. J. 1997. Prefabricated vertical drains zone of influence under
vacuum in clayey soil. In: EVANS, J. C. (ed.) In Situ Remediation of the
Geoenvironment. Minneapolis, Minnesota: American Society of Civil Engineers,
Reston, VA (United States).
Hansbo, S. 1997. Aspects of vertical drain design: Darcian or non-Darcian flow.
Geotechnique, 47, 983-992.
Hird, C., Pyrah, I. & Russel, D. 1992. Finite element modelling of vertical drains
beneath embankments on soft ground. Geotechnique, 42, 499-511.
Holtz, R. D., Jamiolkowski, M., Lancellotta, R. & Pedroni, R. 1991. Prefabricated
vertical drains: design and performance, London, Construction Industry Research
& Information Assoc.
Indraratna, B., Aljorany, A. & Rujikiatkamjorn, C. 2008. Analytical and numerical
modeling of consolidation by vertical drain beneath a circular embankment.
International Journal of Geomechanics, 8, 199-206.
Indraratna, B., Balasubramaniam, A. & Balachandran, S. 1992. Performance of test
embankment constructed to failure on soft marine clay. Journal of Geotechnical
Engineering, ASCE, 118, 12-33.
Indraratna, B. & Chu, J. 2005. Ground Improvement—Case Histories, London, UK,
Elsevier.

Page 27 of 50
Indraratna, B., Perera, D., Rujikiatkamjorn, C. & Kelly, R. 2014. Soil disturbance
analysis due to vertical drain installation. Proceedings of the Institution of Civil
Engineers - Geotechnical Engineering, 168, 236-246.
Indraratna, B. & Redana, I. 1997. Plane-strain modeling of smear effects associated with
vertical drains. Journal of Geotechnical and Geoenvironmental Engineering, 123,
474-478.
Indraratna, B. & Redana, I. 1998. Laboratory determination of smear zone due to vertical
drain installation. Journal of geotechnical and geoenvironmental engineering,
124, 180-184.
Indraratna, B. & Redana, I. 2000. Numerical modeling of vertical drains with smear and
well resistance installed in soft clay. Canadian Geotechnical Journal, 37, 132-
145.
Indraratna, B., Rujikiatkamjorn, C., Ameratunga, J. & Boyle, P. 2011. Performance and
prediction of vacuum combined surcharge consolidation at Port of Brisbane.
Journal of Geotechnical and Geoenvironmental Engineering, 137, 1009-1018.
Indraratna, B., Rujikiatkamjorn, C., Kelly, R. & Buys, H. 2012. Soft soil foundation
improved by vacuum and surcharge loading. Proceedings of the ICE-Ground
Improvement, 165, 87-96.
Indraratna, B., Rujikiatkamjorn, C. & Sathananthan, I. 2005. Analytical and numerical
solutions for a single vertical drain including the effects of vacuum preloading.
Canadian Geotechnical Journal, 42, 994-1014.
Kelly, R., O’loughlin, C., Bates, L., Gourvenec, S., Colreavy, C., White, D., Gaone, F.,
Doherty, J. & Randolph, M. 2014. In situ testing at the National Soft Soil Field
Testing Facility, Ballina. New South Wales. Australian Geomechanics, 49, 15-26.
Kianfar, K., Indraratna, B., Rujikiatkamjorn, C. & Leroueil, S. 2015. Radial
consolidation response upon the application and removal of vacuum and fill
loading. Canadian Geotechnical Journal, 52, 2156-2162.
Kjellman, W. 1952. Consolidation of clayey soils by atmospheric pressure. Proceedings
of a Conference on Soil Stabilization. Boston: MIT Press.
Li, J. H., Huang, J., Cassidy, M. J. & Kelly, R. 2014. Spatial variability of the soil at the
Ballina National Field Test Facility. Australian Geomechanics, 49, 41-48.
Mohamedelhassan, E. & Shang, J. 2002. Vacuum and surcharge combined one-
dimensional consolidation of clay soils. Canadian Geotechnical Journal, 39,
1126-1138.
Pariseau, W. 1999. Poroelastic‐plastic consolidation—analytical solution. International
journal for numerical and analytical methods in geomechanics, 23, 577-594.
Parsa-Pajouh, A., Fatahi, B., Vincent, P. & Khabbaz, H. 2014. Trial Embankment
Analysis to Predict Smear Zone Characteristics Induced by Prefabricated Vertical
Drain Installation. Geotechnical and Geological Engineering, 32, 1187-1210.
Pineda, J., Suwal, L. & Kelly, R. 2014. Sampling and laboratory testing of Ballina clay.
Australian Geomechanics, 49, 27-40.
Pineda, J., Suwal, L., Kelly, R., Bates, L. & Sloan, S. 2016. Characterisation of Ballina
clay. Géotechnique, 1-22.

Page 28 of 50
Richart, F. E., Jr. 1957. A review of the theories for sand drains. Journal of the Soil
Mechanics and Foundations Division, ASCE, 83, 1-38.
Rixner, J., Kraemer, S. & Smith, A. D. 1986. Prefabricated vertical drains: Engineering
guidelines, FWHA/RD-86/168, Washington DC, Federal Highway Administration.
Selvadurai, A. & Suvorov, A. Boundary heating of poro-elastic and poro-elasto-plastic
spheres. Proceedings of the Royal Society of London A: Mathematical, Physical
and Engineering Sciences, 2012. The Royal Society, 2779-2806.
Selvadurai, A. & Suvorov, A. Thermo-poromechanics of a fluid-filled cavity in a fluid-
saturated geomaterial. Proceedings of the Royal Society of London A:
Mathematical, Physical and Engineering Sciences, 2014. The Royal Society.

Page 29 of 50
 List of Tables

TABLE 1: SELECTED STIFFNESS AND STRENGTH MATERIAL PARAMETERS FOR BALLINA SITE ........................................................ 31 

TABLE 2: MATERIAL PARAMETERS SELECTED FOR SIMULATION OF FILL EMBANKMENT .............................................................. 32 

TABLE 3: EFFECT OF SMEAR ZONE ON EXPECTED TERMINATION OF VACUUM .......................................................................... 33 

Page 30 of 50
 Table 1: Selected stiffness and strength material parameters for Ballina site

Depth  c kh k h , ring
Soil type E (MPa)  Cc Cs OCR 3 eo Ko   
(m) (kN/m ) (kPa)
(×10-9 m/s) (×10-9 m/s)
Crust (Sandy
0-1.5 4 0.35 N/A N/A 6.5 16.0 N/A 1.27 5 30 0 2 0.2
clayey silt)
Soft
Holocene 1.5-12 N/A N/A 1.4 0.157 1.7 14.5 2.9 0.9 10 25 N/A 1 0.1
Clay
Transition
12-15 4 0.3 N/A N/A N/A 16.0 N/A 0.85 5 30 0 3 N/A
Clay to Sand
Fine Sand 15-20 6 0.25 N/A N/A N/A 18.0 N/A 0.6 0 35 5 5 N/A
Medium Stiff
Pleistocene >20 Assumed unaffected, treated as bedrock
Clay

E: Elastic modulus,  : Poison’s ratio, C c : Compression index, in e  log  plane, C s : Swelling index in e  log   plane, OCR: Over-consolidation ratio,
 : Saturated unit weight, eo: Initial void ratio, Ko: Coefficient of lateral pressure at-rest, c : Drained cohesion,   : Drained friction angle,  : Dilation angle,
kh : Coefficients of permeability in horizontal direction, and kh , ring : Equivalent coefficients of permeability for converted drain wall

Page 31 of 50
 Table 2: Material parameters selected for simulation of fill embankment

Fills layer E (MPa)   (kN/m3) c (kPa)  

1 (Working platform) 6 0.3 16 0 30
2 (Sand Blanket) 7 0.3 17 0 32
3 (Final fill) 8 0.3 18 5 35

Page 32 of 50
Table 3: Effect of smear zone on expected termination of vacuum

kh/ks
1.5 3 5
ds/dw Time to achieve zero excess pore pressure (days)
2 38 51 66
4 42 66 101
6 44 76 121

Page 33 of 50
 List of Figures

FIGURE 1: CONVERSION FOR MULTIDRAIN SYSTEM UNDER CIRCULAR LOADING ADOPTED FOR ANALYTICAL SOLUTIONS IN (A) SQUARE 

PATTERN, (B) TRIANGULAR PATTERN, AND (C) ELEVATION VIEW OF A SINGLE RING ......................................................... 35 

FIGURE 2: (A) A SINGLE HOLLOW CYLINDER OF SOIL‐DRAIN, (B) VERTICAL CROSS SECTION OF A DRAIN WALL AND FLOW CONDITION IN 

VERTICAL DIRECTION, (C) DISTRIBUTION OF VACUUM PRESSURE IN ANALYTICAL MODEL ................................................... 36 

FIGURE 3: VARIATION OF  ring /  2 WITH  i  ............................................................................................................... 37 

FIGURE 4: FINITE ELEMENT DISCRETIZATION FOR AN AXISYMMETRIC DRAIN RING: (A) NODES AND INTEGRATION POINTS FOR A SINGLE 

6‐NODE ELEMENT; (B) MESH DISCRETIZATION, BOUNDARY CONDITIONS, AND VACUUM PRESSURE DISTRIBUTION FOR A SINGLE 

DRAIN RING (MESH SHOWN FOR L=10 M) ............................................................................................................. 38 

FIGURE 5: RESULTS OF SIMULATIONS OF SINGLE RING OF SYDNEY CLAY IN (A) CASE A: L=10 M, K1=1.0, AND (B) CASE B: L=50 M, 

K1=0.5, IN COMPARISON WITH ANALYTICAL SOLUTIONS ........................................................................................... 39 

FIGURE 6: (A) PLAN, AND (B) TYPICAL CROSS SECTION OF CIRCULAR VP EMBANKMENT IN BFTF ................................................ 40 

FIGURE 7: MESH, BOUNDARY CONDITIONS AND CONTROL POINTS OF THE VP CIRCULAR EMBANKMENT ...................................... 41 

FIGURE 8: ASSUMED TIME HISTORY OF EMBANKMENT CONSTRUCTION AND APPLICATION OF VACUUM:  (A) VACUUM AND SURCHARGE 

OF 3.5 M HIGH FILL, AND (B) NO VACUUM AND SURCHARGE OF 7.5 M HIGH FILL .......................................................... 42 

FIGURE 9: SETTLEMENT PATTERN OF THE GROUND SURFACE BENEATH VP CIRCULAR EMBANKMENT WITH AND WITHOUT VACUUM .. 43 

FIGURE 10: COMPARISON BETWEEN NON‐VACUUM AND VACUUM CONSOLIDATION OF CIRCULAR EMBANKMENT:  (A) SETTLEMENT, 

AND (B) EXCESS PORE PRESSURE TIME HISTORIES ..................................................................................................... 44 

FIGURE 11: CONTOURS OF HORIZONTAL DISPLACEMENTS, AFTER ONE YEAR VP CONSOLIDATION ............................................... 45 

FIGURE 12: PROFILES OF HORIZONTAL DISPLACEMENTS AT THE TOE OF CIRCULAR EMBANKMENT (X=14.4 M) .............................. 46 

FIGURE 13: TIME HISTORIES OF HORIZONTAL DISPLACEMENT IN SELECTED KEY POINTS IN CASE OF CONTINUES VACUUM 

CONSOLIDATION FOR ONE YEAR ........................................................................................................................... 47 

FIGURE 14: EFFECT OF VACUUM TERMINATION ON CONSOLIDATION PERFORMANCE: (A) EXCESS PORE PRESSURE, AND (B) SETTLEMENT 

TIME HISTORIES ................................................................................................................................................ 48 

FIGURE 15: PROFILE OF PRE‐CONSOLIDATION PRESSURE ALONG THE EMBANKMENT CENTRELINE ............................................... 49 

FIGURE 16: TIME HISTORY OF EXCESS PORE PRESSURE AT THE MIDDLE OF THE SOFT CLAY LAYER FOR A VARIETY OF SMEAR 

CHARACTERISTICS: (A) DS/DW=2, (B) DS/DW=4, AND (C) DS/DW=6 .............................................................................. 50 

Page 34 of 50
 Figure 1: Conversion for multidrain system under circular loading adopted for analytical
solutions in (a) square pattern, (b) triangular pattern, and (c) elevation view of a single ring

Page 35 of 50
 Outer impermeable
boundary
S
ri -po
i
z
ri -S/2 Q+dQ z
i
dz dQ1 dQ2 L L
O 1 2
Inner impermeable
boundary Q

ri +
S/2
ri +S/2 -k1po
Drain wall
Drain wall
(b) (c)
(Thin wall with infinitesimal thickness)

(a)

Figure 2: (a) A single hollow cylinder of soil-drain, (b) vertical cross section of a drain wall and
flow condition in vertical direction, (c) distribution of vacuum pressure in analytical model

Page 36 of 50
 0.695

0.690

0.685
ring/2

0.680

0.675

0.670

0.665
0 20 40 60 80 100
i

Figure 3: Variation of ring /  with i

2

Page 37 of 50
  po

 k1 po

Figure 4: Finite element discretization for an axisymmetric drain ring: (a) nodes and integration
points for a single 6-node element; (b) mesh discretization, boundary conditions, and vacuum
pressure distribution for a single drain ring (mesh shown for L=10 m)

Page 38 of 50
 1.0
(a) (b)

0.8

0.6

L
L
0.4
Average excess pore pressure ratio  (ū/1)

0.2

0.0
Case A Case B
L  10m L  50m
‐0.2 k1  1.0 k1  0.5

‐0.4

‐0.6
Analytical Solution‐Curent Study

‐0.8 Simulation, Average PWP

Simulation, Settlement curve

‐1.0
0.1 1 10 100 1000 100.1
10000 1 10 100 1000 10000
Time (days) Time (days)

Figure 5: Results of simulations of single ring of Sydney clay in (a) Case A: L=10 m, k1=1.0, and
(b) Case B: L=50 m, k1=0.5, in comparison with analytical solutions

Page 39 of 50
Figure 6: (a) Plan, and (b) typical cross section of circular VP embankment in BFTF

Page 40 of 50
  0 Vacuum+Fill
p0  
70kPa Fill Only

Figure 7: Mesh, boundary conditions and control points of the VP circular embankment

Page 41 of 50
 4.0 ‐80

3.0 Fill embankment ‐60

Vacuum
2.0 ‐40

Vacuum pressure (kPa)
Fill height (m)
1.0 ‐20

0.0 0
Working platform 
‐1.0 Vacuum 20
PVD  Sand blanket
‐2.0 installation and final fill 40

‐3.0 60

‐4.0 80
0 10 20 30 40 50 60 70
Time (days)
(a) 

8.0

7.0

6.0
First Stage
Fill height (m)

5.0
Third Stage
4.0

3.0
Second Stage
2.0
Equivalent Fill Only
1.0 Treatment

0.0
0 50 100 150 200 250 300
Time (days)
(b) 

Figure 8: Assumed time history of embankment construction and application of vacuum:

(a) vacuum and surcharge of 3.5 m high fill, and (b) no vacuum and surcharge of 7.5 m high fill

Page 42 of 50
 0.0

‐0.5
Settlement (m)

‐1.0

‐1.5

‐2.0 After construction
One year after construction, No Vacuum
One year after construction, Vacuum
‐2.5
0 5 10 15 20 25
Distance from the centreline (m)
Figure 9: Settlement pattern of the ground surface beneath VP circular embankment with and
without vacuum

Page 43 of 50
 0.0
0
‐0.2
2
Settlement at point A (m) ‐0.4
4
‐0.6
6
‐0.8
8
‐1.0
0
‐1.2
2 t = 345 days
‐1.4
4
‐1.6
6
‐1.8
8
0 120 240 360 480 600
Time (days)

(a)
80
0
Fill Only
Excess pore pressure at point C (kPa)

60
0 Vacuum + Filll

40
0

20
0

‐20
0

‐40
0

‐60
0

‐80
0
0 120 240 360 480 600
Time (days)

(b)
F
Figure 10: Comparison
C between non
n-vacuum and
a vacuum consolidatioon of circula
ar
eembankmen
nt:
(a) settlemeent, and (b) excess pore pressure tim
me historiess

Page 44 of 50
0
Figure 11: Contours of horizontal displacements, after one year VP consolidation

Page 45 of 50
 Horizontal displacement (m)
‐0.2 ‐0.1 0 0.1 0.2 0.3
0

8
Depth (m)

10

12

14

16
Fill Only (Stable final condition)
Vacuum‐1 month
18
Vacuum‐3 months
Vacuum‐12 months
20

Figure 12: Profiles of horizontal displacements at the toe of circular embankment (x=14.4 m)

Page 46 of 50
 0.10

0.05 D
Horizontal displacement (m)

0.00
H
D
‐0.05
E
G
F
‐0.10
G
F H
‐0.15
E
‐0.20
0 120 240 360 480 600 720 840
Time (days)

Figure 13: Time histories of horizontal displacement in selected key points in case of continues
vacuum consolidation for one year

Page 47 of 50
 100
80

Excess pore pressure at point C (kPa)
60
40
20
0
‐20
‐40
‐60
0.0
‐80
During construction
Vacuum applied continuesly
Vacuum terminated after 1 month
‐0.5
Settlement at point A (m)

Vacuum terminated after 3 months
Vacuum terminated after 12 months

‐1.0

‐1.5

‐2.0

‐2.5
0 120 240 360 480 600 720 840
Time (days)

Figure 14: Effect of vacuum termination on consolidation performance:

(a) excess pore pressure, and (b) settlement time histories

Page 48 of 50
 Pre‐consolidation pressure increment, Δp'c (kPa)
0 50 100 150
0
1
2
3
4 After construction
Depth (m)

5 After 3 month VP
After 12 month VP
6
After 24 month VP
7
8
9
10
11

Figure 15: Profile of pre-consolidation pressure along the embankment centreline

Page 49 of 50
 60

40

Excess pore pressure  (kPa)
20

‐20
kh/ks=1.5
‐40
kh/ks=3
‐60 kh/ks=5

60
‐80

40
Excess pore pressure  (kPa)

20

‐20

‐40

‐60

60
‐80

40
Excess pore pressure  (kPa)

20

‐20

‐40

‐60

‐80
1 10 100 1000
Vacuum time (days)

Figure 16: Time history of excess pore pressure at the middle of the soft clay layer for a variety
of smear characteristics: (a) ds/dw=2, (b) ds/dw=4, and (c) ds/dw=6

Page 50 of 50