Ground Improvement in Transport

Geotechnics - from Theory to Practice

Prof. Buddhima Indraratna Professor of Civil Engineering & Research Director

Centre for Geomechanics and Railway Engineering

Program Leader, ARC Centre of Excellence in Geotechnical Science & Engineering

University of Wollongong, NSW 2522 Australia

14th IACMAG 2014, Kyoto, Japan

A/Prof. Cholachat Rujikiatkamjorn Associate Professor, School of Civil, Mining & Environmental Engineering

University of Wollongong, NSW 2522 Australia

Dr. Sanjay Nimbalkar Research Fellow, Centre for Geomechanics and Railway engineering

University of Wollongong, NSW 2522 Australia

Dr. Ana Heitor Lecturer, School of Civil, Mining & Environmental Engineering

University of Wollongong, NSW 2522 Australia

Figures from “Road and rail freight: competitors or complements?” Bureau of

Infrastructure, Transport and Regional Economics, Australian Government Canberra.

Demand for freight and passenger transport has increased in the past decade.

Increased traffic tonnage necessitates use of ground improvement techniques in

Rail, Road and Port Infrastructure.

2

Part A

Part B

PART A:

Railways – Granular media stabilisation

• Effect of confining pressure on track design

• DEM - Particle degradation modelling

• Cyclic loading and FEM modelling

• Track Contamination and design implications

3

Particle degradation Track buckling Subgrade clay pumping

Prismoidal Triaxial Rig to Simulate a

Track Section

(Specimen: 800x600x600 mm)

Cylindrical Triaxial Equipment

(Specimen: 300 mm dia.x600 mm high)

Dynamic Process Simulation Test Facilities, Designed and Built at UoW

4

Effect of High Impact Loads and Track Degradation

Subgrade

type Location of shock mat

Ballast Breakage

Index (BBI)

Without shock mat

Stiff - 0.170

Soft - 0.080

With Shock mat

Stiff Above ballast 0.145

Stiff Below ballast 0.129

Stiff Above & below ballast 0.091

Soft Above ballast 0.055

Soft Below ballast 0.056

Soft Above & below ballast 0.028

Shock Mat

Nimbalkar, Indraratna, Dash & Christie (2012). JGGE, ASCE, Vol. 138(3), 281-294

5

Effect of Confining Pressure on Strain Behaviour of Ballast

(Indraratna, Lackenby and Christie (2005), Geotechnique, Vol. 55(4), 325-328)

0 5 10 15 20 25 30 35

Axial Strain (%)

6

5

4

3

2

1

0

-1

-2

-3

-4

Vo

lum

etr

ic S

tra

in (

%)

3 kPa

10 kPa

20 kPa

30 kPa

45 kPa

60 kPa

90 kPa

120 kPa

180 kPa

240 kPa

3

qmax = 500 kPa

compression

dilation

Cyclic Loading

0 5 10 15 20 25

Axial Strain (%)

4

2

0

-2

-4

-6

-8

-10

-12

-14

-16

Vo

lum

etr

ic S

tra

in (

%)

1 kPa

8 kPa

15 kPa

30 kPa

60 kPa

90 kPa

120 kPa

240 kPa

3

dilation

compression

Monotonic Loading

Peak friction angle, p, of fresh ballast

Dilatancy (+)

(-) Compression

Particle breakage

f (excludes particle breakage and dilatancy)

0 100 200 300 400

Effective confining pressure (kPa)

36

40

44

48

52

56

60

Fri

ctio

n a

ngl

e,

(de

gree

)

0 50 100 150 200 250

Confining Pressure (kPa)

Fri

ctio

n A

ng

le

()

f (excludes particle breakage and dilatancy)

fb (includes breakage but excludes dilatancy)

Compression

Dilation

p

44

6

Increasing Confining Pressure using: Intermittent

Lateral Restraints or Embedded Winged Sleepers

Rail

Lateral restraints

Intermittent lateral

restraints

Sleepers

Winged sleepers

Lackenby, Indraratna, McDowell and Christie (2007) Geotechnique, ICE, UK. Vol. 57(6), 527-536 7

Effect of Confining Pressure on Particle Degradation

(Cyclic Loading)

Sieve Size (mm)

0

1

Fra

ctio

n P

ass

ing

Initial PSD

Final PSD

Arb

itrar

y bo

unda

ry o

f max

imum

bre

akag

e

2.36

d95iA

B BAABBI

2.36 = smallest sieve size

0 63

PSD = particle size distribution

d95i = d95 of largest

sieve size

Shift in PSD caused by degradation

dmax

Ballast Breakage Index (BBI)

0 50 100 150 200 250

Effective Confining Pressure (kPa)

0

0.02

0.04

0.06

Ba

llast B

rea

kag

e In

de

x,

BB

I

qmax = 500 kPa

qmax = 230 kPa

(I) (II) (III)

Optimum Contact

0 50 100 150 200 250

Effective Confining Pressure (kPa)

0

0.02

0.04

0.06

Ba

llast B

rea

kag

e In

de

x,

BB

I

qmax = 500 kPa

qmax = 230 kPa

(I) (II) (III)

0 50 100 150 200 250

Effective Confining Pressure (kPa)

0

0.02

0.04

0.06

Ba

llast B

rea

kag

e In

de

x,

BB

I

qmax = 500 kPa

qmax = 230 kPa

(I) (II) (III)

Optimum Contact

Unsta

ble

Dila

tion

Zone

Optim

um

Degra

dation Z

one

Compressive Stable

Degradation Zone

Indraratna, Lackenby and Christie (2005)

Geotechnique, Vol. 55(4), 325-328

8

Constitutive Modelling Incorporating Ballast Breakage

– Energy Approach

245tan1

3

1

3

2

sin1/

245tan1

3

1

3

2

12

45tan1

2

1

1

2

1

2

1

fv

fB

fv

fv

d

dp

ddE

d

d

d

d

p

q

f = basic friction angle

p = Effective mean stress

q = Deviator stress

dEB = increment of energy consumption due to particle breakage

Indraratna and Salim (2002)

Geotechnical Engineering,

ICE Proceedings, UK.

Conventional theory

Stress-Strain behaviour Volume Change Behaviour

**912

1

*23912

2

)(

)(

Mp

BM

p

peM

dMMp

p

p

p

do

i

ics

io

csps

MM

M

p

B

MM

M

d

dps

pv

*239

*

*239

9

Model Parameters need to

be determined by large-

scale testing

Salim & Indraratna (2004), Canadian Geotechnical Journal, Vol. 41(4), 657-671

0.0 5.0 10.0 15.0 20.0 25.0

Distrortional strain, s (%)

0

400

800

1200

1600

2000

Dis

tort

ion

al str

ess,

q

(kP

a)

3 = 300 kPa

100 kPa

200 kPa

50 kPa

Test data for crushed basalt(Indraratna and Salim 2001)

Model prediction

Deviatoric Strain, εs (%) 0.0 5.0 10.0 15.0 20.0 25.0

Distrortional strain, s (%)

8.0

6.0

4.0

2.0

0.0

-2.0

-4.0

-6.0

-8.0

Vo

lum

etr

ic s

tra

in, v

(%

) 3 = 50 kPa

100 kPa

200 kPa

300 kPa

Dilation

Contraction

Test data for crushed basalt(Indraratna and Salim 2001)

Model prediction

Deviatoric Strain, εs (%)

Constitutive Modelling of Particle Breakage

10

Particle Breakage near the top plate

Model particle shapes and sizes

Hossain, Indraratna, Darve, & Thakur (2007).

Geomech & Geoengg, Vol. 2(3), 175-181

Distinct Element Modelling: Train Velocity vs Breakage

11

Constitutive model: Critical State capturing particle breakage Indraratna, B., Sun, Q. D. & Nimbalkar, S. (2014). Can. Geotech. J. in press.

0 1c cM M exp BBI

c ref a exp b BBI ln p'

Mc0 is critical state stress ratio for BBI = 0

12

DEM Modelling Geogrid-reinforced Ballast under Shearing Loads

Comparison of shear stress and displacements for

DEM simulation of reinforced ballast DEM particle shapes and sizes

Ngo, Indraratna, and Rujikiatkamjorn (2014). Computers & Geotechnics, Vol. 55, 224-231

DEM - geogrid

13

DEM Model for Geogrid-reinforced Ballast under Direct Shearing

14

Role of Ballast Fouling on Track Performance

Infiltration of coal

VCI = (1+ef)

eb

× Gs.b

Gs.f

× Mf

Mb

× 100

Void Contaminant Index (VCI) proposed by UOW eb = Void ratio of clean ballast

ef = Void ratio of fouling material

Gs-b = Specific gravity of clean ballast

Gs-f = Specific gravity of fouling material

Mb = Dry mass of clean ballast

Mf = Dry mass of fouling material

Slurried Clay infiltration

15

Bounding Surface Model for Fouled Ballast

(Indraratna et al., 2014; Geotechnique, in press)

Bounding Surface

0lnR

'p'pln( 'p M p'pαqF

N1

coa

p

v

e

vv dεdεdε

ηMdε

dεfp

q

p

v

Non-Associated Flow rule

Mf=f(VCI)

Hardening modulus h=hb + hf

σF

m'p

κλ

e1

p'pln'pN

lnR

p'plnp p'pMα

hp

co

cc

1

coa

b

N

cf p

c maxv

1 p'h

p εf VCI

16

Loading Surface

f=0ij

p'c

M

q

n

n

Bounding Surface

F=0

p'P’c/R

Impeded Track Drainage due to Ballast Fouling

Large-scale permeability test apparatus

Variation of hydraulic conductivity vs. Void Contaminant Index

kb = Hydraulic conductivity of clean ballast

kf = Hydraulic conductivity of fouling material

Hydraulic Conductivity (k) of fouled ballast

100

b f

f b f

k kk

VCIk (k k )

Tennakoon, Indraratna, Cholachat, Nimbalkar and Neville

(2012) ASTM Geotechnical Testing Journal, Vol. 35(4), 1-12

0 20 40 60 80 10010

-7

10-6

10-5

10-4

10-3

10-2

10-1

100

Sydenham Site

VCI=22%

Coal-fouled ballast: Experimental

Coal-fouled ballast: Theoretical

Sand-fouled ballast: Experimental

Sand-fouled ballast: Theoretical

Hyd

rau

lic C

on

du

cti

vit

y, k (

m/s

)

Void Contaminant Index, VCI (%)

Bellambi Site

VCI=33% Rockhampton Site

VCI=72%

hydraulic conductivity

of clayey fine sand

hydraulic conductivity

of coal fines

17

Cyclic loading (sinusoidal form) of Track

18

Initial static loading to reach the minimum cyclic deviator stress.

Frequency conditioning phase (f = 1 Hz, N = 10) to prevent any loss of actuator

contact with the specimen.

Cyclic loading phase ( f = 5, 10, 20 and 40 Hz, N = 500 000).

Effect of frequency on the axial strain of ballast

Sun, Q. D., Indraratna, B. & Nimbalkar, S. (2014). Géotechnique, doi: 10.1680/geot./14-T-015.

Range I: Plastic shakedown (5Hz ≤ f ≤ 20 Hz)

Range II: Plastic shakedown followed by Ratcheting (30 Hz ≤ f ≤ 50 Hz)

Range III: Plastic collapse (f = 60 Hz) 19

Field Trial on Instrumented Track in Bulli and Singleton

Geocomposite layer (geogrid+geotextile)

before ballast placement

Ballast placement over the

geocomposite

Geogrid layer placed

above the capping

Settlement pegs

placement in the track 20

Field Instrumentation - Bulli

Settlement pegs

installed at ballast-

capping interface

Displacement

transducers installed at

sleeper-ballast interface

Field Deformation Response

The recycled ballast performed well

because, it was broadly graded compared

to the relatively uniform fresh ballast.

Indraratna et al. (2010). JGGE, ASCE, Vol. 136(7), 907-917

Indraratna et al. (2014). ICE Proc. Ground Improvement, Vol. 167(1), 24-34

0 2 4 6 8 10 12 14 16 18

18

15

12

9

6

3

0

0 1x105

2x105

3x105

4x105

5x105

6x105

7x105

8x105

9x105

6.00

5.00

4.00

3.00

2.00

1.00

0.00

Fresh Ballast (uniformly graded)

Recycled Ballast (broadly graded)

Fresh Ballast with Geocomposite

Recycled Ballast with Geocomposite

Number of load cycles, N

Aver

age

ver

tica

l def

orm

atio

n o

f bal

last

, (S

v) av

g (

mm

)

Aver

age

ver

tica

l st

rain

of

bal

last

, (

1) av

g (

%)

time, t (months)0.0 5.0x10

41.0x10

51.5x10

52.0x10

52.5x10

524

18

12

6

00 20 40 60 80 100

8

6

4

2

0

Vert

ical

Str

ain

of

Ball

ast

, v

(%

)

Time, t (days)

Vert

ical

Defo

rmati

on

of

Ball

ast

, S

v (

mm

)Number of Load Cycles, N

Type of Section Aperture

Size (mm)

Geogrid 1 44 44

Geogrid 2 65 65

Geogrid 3 40 40

Geocomposite 31 31

Optimum aperture size of geogrids is

about 1.15D50 of ballast.

Bulli Track Singleton Track

22

Plane Strain FEM Analysis of Track Substructure

(Confining Pressure @ 50 kPa by geocells)

0 5 10 15 20 25

0

100

200

300

400

500

600

700

800

900

1000

Effective confining pressure 3

' = 50 kPa

Devia

tor

Str

ess

, q =

' -

' (kP

a)

Axial Strain, a (%)

E50

1 asymptote

23

Track transverse section deformation

Track longitudinal section deformation

PART B:

Road Embankments and Port Reclamation – Soft Soil

Stabilisation

• Concepts of vacuum consolidation

• Factor affecting vacuum consolidation

• 2D and 3D FEM modelling of vacuum

consolidation

26

Membrane-type Vacuum Application (Courtesy Austress-Menard)

Drain Installation

Horizontal drain

installation Peripheral bentonite

trench

Connection between horizontal

drainage and vacuum pump Membrane installation

Time

-150

-75

0

75

150

Str

ess/

Pre

ssu

re (

kP

a)

Time

-150

-75

0

75

150

Str

ess/

Pre

ssu

re (

kP

a)

p (preloading pressure)

p (preloading pressure)

p0 (Vacuum

pressure)

Time

-150

-75

0

75

150

Excess p

ore

pre

ssure

(kP

a)

Time

-150

-75

0

75

150

Exce

ss p

ore

pre

ssu

re (

kP

a)

Time

-150

-75

0

75

150

Ve

rtic

al e

ffe

ctive

str

ess (

kP

a)

Time

-150

-75

0

75

150

Ve

rtic

al e

ffe

ctive

str

ess (

kP

a)

Maximum excess pore pressure

Maximum excess pore pressure

Principle of Vacuum Consolidation

Consolidation: (a) conventional surcharge loading;

(b) idealised vacuum preloading (Indraratna et al. 2005).

2 2

2 2

1( )h v

u u u uc c

r r r z t

Governing Equation

VP directly adjusts the initial

pore pressure boundary

conditions.

Vertical consolidation term

can be ignored if Z is

very large.

28

Pore Pressure generation and retarded dissipation within the Smear Zone

Mandrel Driving INCREASES effective vertical stress, hence, the lateral permeability

decreases within the smear zone (Sathananthan, Indraratna, & Rujikiatkamjorn (2008),

ASCE J. of Geomechanics, Vol. 8(6), 355-365).

Excess Pore Pressure is

rapidly created during

mandrel intrusion

Excess PWP dissipates

very gradually after

mandrel withdrawal in

spite of the drain.

29

Analytical and Numerical Simulation Multi-drain Analysis and Plane Strain Conversion

Field condition: Axisymmetric

Reduce the

convergence time

and require less

computer memory

Must give the same

consolidation response

Maintain geometric

equivalence

2D plane strain FEM

30

Conversion of an Axisymmetric Unit Cell into Plane Strain

Indraratna et al., 2000 & 2005

Conversion must give

the SAME time-

settlement curve

w

h

wp

hp

q

kls

hkh

k

s

n

q

k

B

l

hpk

hpk

hk

hpk

3

275.0lnln

3

4

2

2

31

0.1 1 10 100 1000Time (days)

0

0.2

0.4

0.6

0.8

1

Ave

rag

e e

xce

ss p

ore

pre

ssu

re r

atio

ch=0.32m2/year

de=0.45m

Axisymmetric

Plane strain

2 graphs coincide

0

vach

o

vac

o u

uT8exp

u

u1

u

u

Normalized average excess pore pressure in axisymmetric

condition with vacuum (Indraratna et al., 2005), CGJ

= pore pressure at time t (average values)

= time factor

= undisturbed horizontal permeability

= smear zone permeability

0u = initial pore pressure

u

hT

w

h

h

h

q

kls

k

k

s

n

3

275.0)ln(

'ln

2

hk

h'k

vacu = average applied vacuum pressure kh ks

-p0

-k1p0

z

l

ds/2

de/2

Smear zone

Undisturbed zone

Vacuum pressure

distribution

C L

32

Sea wall and

development area

33

WD5A WD5BWD1

VC2

WD4

WD2 WD3

VC1

155m

35m

70m 41m 84.5m 84.5m

70m

70m

50m

169m

210m

MS15-1

MS17-1 MS18-1

MS16-1MS22-1 WD5B

VC1-2

VC2-1

Surface settlement plates

Piezometers

VWP2-WD1

VWP1-WD2

VWP4-WD4

MS20-VWP5

MS19-VWP5

MS28-VC1

MS27-WD3

Inclinometers

0 40 80

Liquid and plastic limit (%)

-30

-20

-10

0

10

Ele

va

tio

n (

m)

PL

LL

Water Content

0.4 0.6 0.8 1

Cc

2 4 6 8

Cv, Ch (m2/yr)

Ch

Cv

20 40 60 80

Su (kPa)

Dredged mud

Upper Holocene sand

Holocene Clay

Pleistocene

Plan view Soil Properties

PL w LL

Tough Luck! No Vertical Drains here

Applications at Port of Brisbane

3D modelling at corner of embankment or at marine boundary

Effect of vacuum application (negative movements) may extend more than 10 m

from the edge of the embankment

A

A

34

Time-Settlement, Pore Pressure & Lateral Yielding Response

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

11/0

5/0

7

08/0

6/0

7

06/0

7/0

7

03/0

8/0

7

31/0

8/0

7

28/0

9/0

7

26/1

0/0

7

23/1

1/0

7

21/1

2/0

7

18/0

1/0

8

15/0

2/0

8

14/0

3/0

8

11/0

4/0

8

09/0

5/0

8

06/0

6/0

8

04/0

7/0

8

01/0

8/0

8

29/0

8/0

8

26/0

9/0

8

24/1

0/0

8

21/1

1/0

8

Sett

lem

ent

(m)

Prediction

Field

UOW Predicted DOC (%)

0

10

20

30

40

50

60

70

80

90

100

11/05/07 19/08/07 27/11/07 06/03/08 14/06/08 22/09/08

De

gre

e o

f C

on

so

lid

ati

on

(%)

-80

-60

-40

-20

0

20

40

60

11/05/07 19/08/07 27/11/07 06/03/08 14/06/08 22/09/08

Excess p

ore

pre

ssure

(kP

a)

Dredged MudUpper Holocene SandUpper Holocene ClayLower Holocene ClayMeasurement

(a) Settlement and (b) excess pore pressure

for a typical vacuum site

35

0 1 2Lateral displacement/Total change in effective stress

(mm/kPa)

-30

-28

-26

-24

-22

-20

-18

-16

-14

-12

-10

-8

-6

-4

-2

0

2

4

6

8

10

De

pth

(m

)Section/Plate No.

VC1/MS28

WD3/MS27

PlatformPlatform

Dredged mud Dredged mud

HS Holocene sand

UHCUpper Holocene Clay

LHC

Lower Holocene Clay

(b) Reduction of lateral displacement for a

typical vacuum site

Part A + Part B: PVD Applications to Rail Embankment at Sandgate and FEM Analysis

Class A Prediction (Indraratna et al. 2010; ASCE, JGGE, Vol. 136(5), 686-696)

Soft Alluvial Clay

Stiffer Silty Clay

0 10 20 30 40 50Lateral displacement (m)

-20

-16

-12

-8

-4

0

De

pth

(m

)

No PVD

PVDs @ 1.5m spacing

Field

Reduction in

lateral displacement

0 100 200 300Time (days)

0.25

0.2

0.15

0.1

0.05

0

Se

ttle

me

nt

(m) Field Data

Prediction-Class A

36

• Geogrids increase confining pressure and reduce particle dilation

and breakage in rail tracks.

• Vacuum preloading effectively controls excess pore pressure and

lateral displacement of soft soil. VP is often the ideal choice, where

lateral movement at a marine boundary needs to be curbed.

• PVDs effectively mitigate the build-up of excess PWP under high

cyclic loading (e.g. applications to railways and airport runways).

• DEM models are most useful to analyse track degradation,

compared to current empirical assessments.

• Fully-instrumented Field trials are imperative to study complex

issues of track behaviour and for performance verification.

Conclusions

37

Australian Research Council (ARC) for substantial funding

Past and Present research students, Research Associates and

Technical Staff

Industry Organisations: RailCorp (NSW), ARTC, QLD Rail,

ARUP, Coffey Geotechnics, Douglas Partners. Roads & Traffic

Authority, Queensland Department of Main Roads, Port of

Brisbane Corporation

Acknowledgment

Thank You!

38