Sie sind auf Seite 1von 19

32 Geotechnical Journal Vol. 6 No.

1 2014

USE OF SHOCK MATS FOR MITIGATING DEGRADATION OF RAILROAD BALLAST


Geotechnical Journal Vol. 6 No. 11 2014

B. Indraratna , S. Nimbalkar2, S.K. Navaratnarajah3, C. Rujikiatkamjorn4 and T. Neville5

1 Professor of Civil Engineering and Research Director, Centre for Geomechanics and Railway Engineering,
USE OF SHOCK MATS FOR MITIGATING DEGRADATION OF RAILROAD BALLAST
Program Leader, ARC Centre of Excellence for Geotechnical Science and Engineering, University of Wollongong, Australia.
1 Geomechanics
3
4
2 Research Fellow,Centre
for
and
Railway Engineering,
University
Australia.
B. Indraratna
, S. Nimbalkar2, S.K.
Navaratnarajah
, C. Rujikiatkamjorn
andofT.Wollongong,
Neville5
Email: sanjayn@uow.edu.au
1
Professor ofCentre
Civil Engineering
and Research
for Geomechanics
andofRailway
Engineering,
3 PhD Candidate,
for Geomechanics
andDirector,
RailwayCentre
Engineering,
University
Wollongong,
Australia.
Program Leader, ARC Centre of Excellence for Geotechnical Science and Engineering, University of Wollongong, Australia.
4 Associate 2Professor,
School
of
Civil,
Mining
and
Environmental
Engineering,
University
of
Wollongong,
Research Fellow,Centre for Geomechanics and Railway Engineering, University of Wollongong, Australia. Australia.
5 Senior Geotechnical Engineer, AustralianEmail:
Rail Track
Cooperation Ltd, Broadmeadow, NSW 2292, Australia
sanjayn@uow.edu.au
3

PhD Candidate, Centre for Geomechanics and Railway Engineering, University of Wollongong, Australia.

Professor,
School of
Civil, Mining
and Environmental
Engineering,
of Wollongong,
Australia.
ABSTRACT: Associate
In Australia,
increasing
demand
for High
Speed Rail (HSR)
and University
heavier freight
transport
is a technical and
5
Senior
Engineer, Australian
Rail
Track
Cooperation
Ltd, Broadmeadow,
NSW 2292,
economic challenge
forGeotechnical
practicing engineers,
designers
and
researchers.
Because
of this increased
trainAustralia
speed and axle load,
high undue stresses are transferred to the ballast and underlying formation. Ballast degradation is a major factor affecting track
longevity
and stability. Use of energy absorbing shock mats to reduce noise and vibrations is an established practice. The shock
ABSTRACT:In Australia, increasing demand for High Speed Rail (HSR) and heavier freight transport is a technical and economic challenge
matforispracticing
sometimes
called designers
as Underand
Sleeper
Pad (USP)
and
Ballasttrain
Mat
(UBM)
depending
their
placement
position.
engineers,
researchers.
Because
of Under
this increased
speed
and axle
load, highupon
undue
stresses
are transferred
to
the ballast
and underlying
formation.
Ballast degradation
is a majorballast
factor degradation
affecting trackare
longevity
and
of energylaboratory
absorbing
However,
studies
to analyse
their effectiveness
in minimising
limited.
A stability.
series ofUse
large-scale
shock mats to reduce noise and vibrations is an established practice. The shock mat is sometimes called as Under Sleeper Pad (USP) and
tests
were conducted on ballast using a high-capacity drop-weight impact testing equipment to understand the performance of
Under Ballast Mat (UBM) depending upon their placement position. However, studies to analyse their effectiveness in minimising ballast
energy
absorbing
shock A
mats
in of
thelarge-scale
attenuation
of impact
and subsequent
mitigation
of ballast degradation.
A numerical
degradation
are limited.
series
laboratory
tests loads
were conducted
on ballast
using a high-capacity
drop-weight impact
testing
equipment
to understand
theon
performance
of energy
absorbing shock
mats in
attenuation
of breakage
impact loads
and subsequent
mitigation
of
model
was developed
based
the modified
stress-dilatancy
approach
to the
capture
particle
during
impact loading.
Model
ballast degradation.
A numerical
model wasresults.
developed
based
on presents
the modified
stress-dilatancy
approach
to capture particle
during
predictions
are compared
with laboratory
This
paper
state-of-the-art
review
of laboratory
studiesbreakage
and numerical
impact loading. Model predictions are compared with laboratory results. This paper presents state-of-the-art review of laboratory studies and
modelling
benefits ofbenefits
USPs of
and
UBMs
in the inpractice.
numericalillustrating
modelling illustrating
USPs
and UBMs
the practice.
4

Keywords: Ballast, Impact load, Shock mats, Degradation, Deformation


Keywords: Ballast, Impact load, Shock mats, Degradation, Deformation

1 INTRODUCTION
1. absorbing
INTRODUCTION
Energy
mats such as Under Sleeper Pad (USP) and Under
Ballast
Mats
(UBM) mats
are resilient
under
sleepers
and
Energy absorbing
such as pads
Underplaced
Sleeper
Pad the
(USP)
and Under
under
the
ballast,
respectively.
The
most
significant
applications
of
Ballast Mats (UBM) are resilient pads placed under the sleepers and
these
resilient
pads in
railways are:
reduce
the structure-borne
under
the ballast,
respectively.
The 1)
most
significant
applications of
vibration
and noise
to in
protect
nearby
reduce the
these resilient
pads
railways
are: structures
1) reduce and
the 2)
structure-borne
ballast
degradation
to improve
stability
maintainand
track
vibration
and noise
to protect
nearbyand
structures
2) geometry,
reduce the
thereby
the service
life ofstability
the rail track.
The resilient
ballastincreasing
degradation
to improve
and maintain
track
geometry,
increasing
the service
life of the
rail track.
The
material
used thereby
as the USP
and UBM
to improvethe
overall
vertical
resilient
as the USP
and UBM
to use
improvethe
overall
elasticity
of material
the trackused
substructure.
In recent
years,
of elastomeric
elasticity ofconcrete
the tracksleepers
substructure.
recent increasingly
years, use of
softvertical
pads underneath
have In
become
elastomeric
soft
pads
underneath
concrete
sleepers
have
become
popular and is the primary focus of track research (Marschnig
and
popular and is the primary focus of track research
Veitincreasingly
2011).
(Marschnig and Veit 2011).
The elastic
embedded
under theunder
sleeper
interface
The pad
elastic
pad embedded
theavoids
sleepera hard
avoids
a hard
with the
ballast,theallowing
ballast
bed padding
into the
withinterface
the ballast,
allowing
ballast the
to bed
intoto the
padding
material.
This
the contact
surface
area of
the
material.
This
increases
theincreases
contact surface
area of
the ballast
with
ballast
with other
as, USP
the contact
area
other
interfaces
suchinterfaces
as, USP such
(increases
the(increases
contact area
of ballast
ballast with
and UBM
the of
contact
withof sleeper),
and sleeper),
UBM (increases
the(increases
contact area
ballastarea
withof
ballast with
sub-ballast soil).
or formation
soil). Consequently,
this avoids
sub-ballast
or formation
Consequently,
this avoids excessive
excessive
forces
the and
interfaces
ballast particles,
contact
forcescontact
between
thebetween
interfaces
ballastand
particles,
leading
leading to increased stability, less settlement and reduced wear of
to increased stability, less settlement and reduced wear of the track
the track sub-structure. In the case of USP, it extends the bending
sub-structure.
In the case of USP, it extends the bending length of
length of the rails. Therefore, the axle load from the train is
the distributed
rails. Therefore,
axlenumber
load from
the train
is distributed
over
over a the
larger
of sleepers
compare
with sleepers
a larger
number
sleepers
compare withload
sleepers
without
USPs.
without
USPs.ofSince
the compression
distributed
over
large
Since
compression
load distributed
1), it
areathe
(Figure
1), it further
reduces the over
forcelarge
actingarea
on (Figure
sleeper-ballast
further
reduces
force acting
on sleeper-ballast
interface
inter-et
interface
andthe
inter-ballast
particle
forces(Bolmsvik
2005;and
Plek
ballast
particle
2005; Plek et al. 2007; Loy 2008;
al. 2007;
Loyforces(Bolmsvik
2008; Dahlberg 2010).
The2010).
wheel and rail irregularities such as wheel flat, rail
Dahlberg
corrugation, dipped rail, defective rail weld, insulation joints and rail
Theexpansion
wheel andgap
rail causes
irregularities
as wheel
highersuch
impact
load flat,
thanrail
thecorrugation,
cyclic load
dipped
rail,
defective
rail
weld,
insulation
joints
and
rail expansion
exerted by moving wheels (Nielsen and Johansson 2000;
Bruni et al.
gap2009;
causesNimbalkar
higher impact
thanChange
the cyclic
load exerted
by moving
et al.load
2012).
of stiffness
where
the track
wheels
(Nielsen
Johansson
2000;
Bruni approach,
et al. 2009;
Nimbalkar
passages
fromand
ballasted
track to
the bridge
track
transition
et al.
2012).such
Change
of stiffness
where
the track
passages
from
locations
as road
crossing and
change
of subgrade
condition
(weak track
subgrade
to bridge
bedrock)
is accelerating
track degradation
ballasted
to the
approach,
track transition
locations due
suchto
this high
impact
loading
(Liofand
Davis 2005;
Nimbalkar
al. 2012).
as road
crossing
and
change
subgrade
condition
(weaketsubgrade
Thereforeisthe
use of energy
in the
track
to bedrock)
accelerating
track absorbing
degradationresilient
due to pads
this high
impact
structure to attenuate the rail track degradation is becoming

loading (Li and Davis 2005; Nimbalkar et al. 2012). Therefore the use
of energy absorbing resilient pads in the track structure to attenuate
increasingly popular in rail road industries (Esveld 2001). This
the
rail track
degradation
increasingly
popular on
in rail
paper
presents
overviewisofbecoming
various methods
of analysis
useroad
of
industries
(Esveld
2001).
This paper
presents
overview
of various
shock mats
in track
structures
in recent
years.
Few preliminary
methods
analysis
of shock of
mats
in track
in recent
researchofstudies
on on
theuse
assessment
shock
matsstructures
using large
scale
years.
preliminary
research
studies onofthe
assessmentAustralia
of shock
impactFew
testing
equipment
at the University
Wollongong,
mats
using
large scale impact testing equipment at the University of
are also
presented.
Wollongong, Australia are also presented.

(a)

(b)

32

Figure 1. Distribution of Axle Load.


(a) Without Shock Mats; (b) With Shock Mats

rd
3Geotechnical
Proff 18-02-2015
Journal Vol. 6

No. 1 2014

Geotechnical Journal Vol. 6 No. 1 2014 33

2. LITERATURE
SURVEY
LITERATURE
SURVEY

comfort
of the
rail transport
the increase
soft padded
2011). The
comfort
of the also
rail increase
transportbyalso
by sleepers
the soft
in
the track
structure.
padded
sleepers
in the track structure.

2.1 History:
Development
of shock
mats
History:
Development
of shock
mats
The track improvements by using shock mats have been in use since
1980s and it is increasingly in use last 10 to 20 years specially in
Central Europe (Bolmsvik 2005; Schneider et al. 2011). Initially it
was used
usedtotoreduce
reduce
vibrations
transmitted
thetrack
rail totrack
to
vibrations
transmitted
fromfrom
the rail
nearby
nearby buildings,
thenbeen
it hasinbeen
wide
to reduce
the sleeperbuildings,
then it has
wideinuse
to use
reduce
the sleeper-ballast
ballast contact
Sincein2005,
in Austria,USPs
as a
contact
stresses.stresses.
Since 2005,
Austria,USPs
are used are
as aused
standard
standard
component
in
turnouts
to
improve
track
quality
and
reduce
component in turnouts to improve track quality and reduce rail
rail corrugation growth on small radius curves in category A tracks
corrugation growth on small radius curves in category A tracks (curve
(curve radius >250m and traffic load >30,000 tons/day)(Schneider et
radius >250m and traffic load >30,000 tons/day)(Schneider et al.
al. 2011). Recent studies by Loy (2008) and Marschnig and Veit
2011).
studies
Loyof
(2008)
Marschnig
Veit (2011)
(2011) Recent
confirms
thebyuse
USPandlessening
theandmaintenance
confirms
the
use
of
USP
lessening
the
maintenance
requirements
requirements and thereby dramatically reducing Life Cycle Cost
and
thereby
dramatically
reducing Life
Cost
(LCC)
found
thatofthetrack
use
(LCC)
of track
structure. Indraratna
et al.Cycle
(2012)
structure.
Indraratna
al.to(2012)
foundofthat
use layer
of shock
mats
of shock mats
reducedetup
50% strain
the the
ballast
subjected
reduced
to 50%
straintoofthethewheel
ballastraillayer
subjected toDetailed
impact
to impactupforces
owing
imperfections.
forces
owing
to the wheel
railon
imperfections.
Detailed
of
overview
of important
studies
the use of shock
mats overview
in rail track
improvement
is presented
in of
theshock
following
important
studies
on the use
mats sections.
in rail track improvement
is presented in the following sections.
2.2 Shock Mats for Vibration reduction
2.2 Shock Mats for Vibration reduction
In the beginning of 1980s, thin elastic pads as a USP material
In
theused
beginning
of 1980s,
thin elastic
pads to
as aminimize
USP material
was used
was
to cover
the wooden
sleeper
the vibration
to
cover the to
wooden
sleepernear
to minimize
the vibration
to
transmitted
the houses
the rail tracks.
Then transmitted
in 1990s, the
the
houses
near the
rail tracks.
Thenby
in 1990s,
the French
railway
French
railway
started
the testing
introducing
thin layer
of
started
the testing
introducing
of polyurethane
as a USP
polyurethane
as abyUSP
materialthin
to layer
minimize
the sleeper-ballast
contact stresses
(Bolmsvik
2005). A study
Auersch
(2006)
material
to minimize
the sleeper-ballast
contactbystresses
(Bolmsvik
suggest Athat
the ballast
mats (i.e.
UBM)
are anthat
efficient
measure
to
2005).
study
by Auersch
(2006)
suggest
the ballast
mats
reduce
the vibration
near themeasure
rail tracks.
In histhe
study,
numerical
(i.e.
UBM)
are an efficient
to reduce
vibration
near
method
track In
dynamics
using
three dimensional
an improved
the
rail of
tracks.
his study,
numerical
method ofand
track
dynamics
simplethree
two dimensional
weresimple
used totwo
analyse
ballast
using
dimensional FEM
and anmodels
improved
dimensional
track
with
and
without
ballast
mat.
Auersch
(2006)
reported
that
the
FEM models were used to analyse ballast track with and without
resonance frequency depend on the stiffness of the ballast mat and
ballast mat. Auersch (2006) reported that the resonance frequency
the insertion of an elastic mat under the ballast layer shifts the
depend on the stiffness of the ballast mat and the insertion of an
vehicletrack resonance frequency between 20 and 50 Hz, thereby
elastic
mat under
the ballast
layer shifts
the vehicletrack
resonance
considerably
improving
the reduction
of dynamic
forces. Loy
(2008)
frequency
between
20
and
50
Hz,
thereby
considerably
improving
the
reported that the use of USP significantly improve the ballast track
reduction
of dynamiccompared
forces. Loy
(2008)
reportedtrack
that without
the use ofUSPs,
USP
vibration behaviour
with
traditional
significantly
the ballast
vibration
especially theimprove
frequencies
above track
40 Hz.
Mediumbehaviour
frequencycompared
range of
with
track
USPs,
especially
the and
frequencies
50-150traditional
Hz vibration
tendwithout
to liquefy
the ballast
material
become
above
40 Hz.
Mediumthe
frequency
rangeisofa50-150
Hz vibration
unstable.
Therefore,
use of USP
beneficial
effect ontend
the
on
stability
ballasted
track. Aand
research
study
by Loy
(2012) the
to
liquefyofthe
ballast material
become
unstable.
Therefore,
mitigating
by USP
that appropriate
USPs
use
of USP vibration
is a beneficial
effectsuggest
on the stability
of ballasted
track.can
A
reduce the
vibration
alsoonimprove
thevibration
track bed
A
research
study
by Loyand
(2012)
mitigating
by geometry.
USP suggest
sandwich
type of
USPcan
consist
a soft
and and
acoustically
highlythat
appropriate
USPs
reduceofthe
vibration
also improve
the
effective
layerAembedded
theof
concrete
sleeperofona one
track
bedelastic
geometry.
sandwich to
type
USP consist
soft side
and
and a visco-plastic
material layer
the ballast
sidetorecommended
acoustically
highly-effective
elasticonlayer
embedded
the concrete
to cater for above two requirements.
sleeper on one side and a visco-plastic material layer on the ballast
side recommended to cater for above two requirements.
2.3 Reduction of Life Cycle Cost
2.3
Life Cycle
Cost sections data analysed by the
TheReduction
Austrian of
mainline
network

Standardized Annual Cost (%)

100
90
80
70
60
50
40
30
20
10
0

100%

Without Pad
With Pad (SLB 3007G)

68%
54%
41%
25%

21%

15%

+
Depreciation

12%

Operational Maintenance
hindrance cost
cost

Total LCC

Figure 2 Breakdown of Normalized Annual Cost


(data sourced from www.getzner.com)
2.4
Ballast
Degradation
2.4 Mitigating
Mitigating
Ballast
Degradation
Ballast
in the
bed which
also
Ballast is
is aa major
major load
load bearing
bearing layer
layer in
the track
track bed
which also
facilitate the water draining easily from top of the track bed to the
underlying formation
formation or
oradjacent
adjacentnative
nativeground.
ground.
speed
of
AsAs
thethe
speed
of the
the
the axle
increases,
the ballast
material
used
the
rail rail
and and
the axle
load load
increases,
the ballast
material
used in
theintrack
track
bed needs
considerable
maintenance
or aofway
of protect
the
bed needs
considerable
maintenance
or a way
protect
the ballast
from high
Limiting
the generated
stressstress
on ballast
is an
ballast
fromstresses.
high stresses.
Limiting
the generated
on ballast
is
economical
option
which
many
railway
agencies
and and
authorities
are
an
economical
option
which
many
railway
agencies
authorities
currently
more
interested
on.on.
This
can
are
currently
more
interested
This
canbebeachieved
achievedby
bythe
the use
use of
energy absorbing shock mats such as USP and UBM. As mentioned
previously, the
theuse
useof of
USP
in concrete
the concrete
sleepers
the
previously,
USP
in the
sleepers
reducesreduces
the ballast
ballast stresses
two mechanisms:
1) Increase
the contact
of
stresses
by twobymechanisms:
1) Increase
the contact
area area
of the
the
ballast
to
concrete
sleeper
interface,
and
2)
increase
the
number
ballast to concrete sleeper interface, and 2) increase the number of
of load
bearing
sleepers
per load
axle (Bolmsvik
load (Bolmsvik
2005; Dahlberg
load
bearing
sleepers
per axle
2005; Dahlberg
2010).
2010).
Each
of
two
mechanisms
reduces
the
maximum
loadby
carried
Each of two mechanisms reduces the maximum load carried
each
by each and
sleeper
and reduces
thereby reduces
the stresses.
ballast stresses.
Bolmsvik
sleeper
thereby
the ballast
Bolmsvik
(2005)
(2005) reported that USP increase the contact area of ballast to the
reported that USP increase the contact area of ballast to the sleeper by
sleeper by more than 36% for soft USP (stiffness 30 kN/mm) and by
more than 36% for soft USP (stiffness 30 kN/mm) and by more than
more than 18% for stiff USP (stiffness 70 kN/mm), which is
18% for stiff USP (stiffness 70 kN/mm), which is otherwise far lower
otherwise far lower than 12%. As of the study by Loy (2008), the
than 12%. As of the study by Loy (2008), the contact area between
contact area between the sleeper and the ballast increases 30-35%
the
thewhich
ballastisincreases
30-35%
with pads,
sleeperatpads
which
withsleeper
sleeperand
pads
5-8% without
sleeper
a bedding
is
5-8%
without
sleeper
pads,
at
a
bedding
modulus
C=0.2
N/mm
modulus C=0.2 N/mm and reducing the pressure on the ballast by
and
reducing
the pressure
the ballast
10-25%.
(2010)
found
that thebyhigher
stiffDahlberg
tracks transmit
10-25%.
Dahlberg
(2010) on
found
that thecontact
higher forces
stiff tracks
the wheelrail
contact
the wheelrail
to the transmit
ballast through
fewer number
of
forces
to the ballast
through
fewer number of
the sleepers.
the sleepers.
Therefore,
the ballast-sleeper
contact
stress isTherefore,
very high.
the
contactbystress
is very high.
can be
minimized
Thisballast-sleeper
can be minimized
introducing
USPsThis
which
distribute
the
by
introducing
USPsnumber
which distribute
the stresses
over more
number
stresses
over more
of sleepers
and thereby
decrease
the
of
sleepers
and thereby
decreasecontact
the ballast
The maximum
ballast
stresses.
The maximum
forcestresses.
57 kN without
USP is
contact
kN 32
without
USP22is kN
reduced
to 48pad
kN,(stiffness
32 kN and
22
reducedforce
to 4857kN,
kN and
for stiff
3000
kN
for stiffmedium
pad (stiffness
3000(stiffness
kN/mm),400
medium
stiff and
pad (stiffness
kN/mm),
stiff pad
kN/mm)
soft pad
(stiffness
50 and
kN/mm),
3).respectively
It was concluded
400
kN/mm)
soft padrespectively
(stiffness 50(Figure
kN/mm),
(Figure
fromIt the
by Dahlberg
(2010)
duebytoDahlberg
significant(2010)
reduction
3).
wasstudy
concluded
from the
study
due of
to
ballast stresses,
these
USPs stresses,
can be used
protect
ballast
significant
reduction
of ballast
thesetoUSPs
can the
be used
to
materialtheinballast
the track
bed inand
of hanging
protect
material
the the
trackdetrimental
bed and theeffects
detrimental
effects
sleepers
cansleepers
also be can
reduced
by reduced
these USPs.
of
hanging
also be
by these USPs.

Technical
University
Graz sections
shows that
installation
padded
The
Austrian
mainline of
network
datathe
analysed
by theof
Technical
sleepers
significantly
reduce
the
LCC
for
the
track
(Marschnig
and
University of Graz shows that the installation of padded sleepers
Veit 2011). This
achieved
three
main cost
(1)
significantly
reducecan
thebe
LCC
for the by
track
(Marschnig
andportions
Veit 2011).
prolonged service life by reducing depreciation, (2) higher track
This can be achieved by three main cost portions (1) prolonged
availability by reducing obstructions of operational cost and
service life by reducing depreciation, (2) higher track availability by
(3)reduced maintenance needs as shown in Figure 2. Therefore,
reducing obstructions of operational cost and (3)reduced maintenance
Marschnig and Veit (2011) concluded that the use of soft padded
needs
as shown
Figurestep
2. Therefore,
Marschnig
(2011)
track system
is ainmajor
towards cost
efficient and
and Veit
sustainable
concluded
that
the
use
of
soft
padded
track
system
is
a
major
step
ballasted track.
towards
Sincecost
theefficient
stiffnessand
of sustainable
the track is ballasted
reduced track.
by the installation of
USPs the
on stiffness
concrete ofsleepers
which
lesser by
thethecorrugation
Since
the track
is reduced
installationinofsmallUSPs
radii tight curves and reduce the higher maintenance cost required at
on concrete sleepers which lesser the corrugation in small-radii tight
the curves. Soft padded concrete sleepers reduce the ballast wear
curves
and reduce
the higher
maintenance
cost cycle
required
at the
and extend
the intervals
between
two tamping
by at
leastcurves.
2 and
Soft
padded
concrete
sleepers
ballast
wear andand
extend
thereby
increase
the service
lifereduce
of the the
ballast
(Marschnig
Veit
the intervals between two tamping cycle by at least 2 and thereby
increase the service life of the ballast (Marschnig and Veit 2011). The

2.5 Field
Field
Study
of Shock
Mats
2.5
Study
on on
UseUse
of Shock
Mats
An extensive
extensivefull-scale
full-scalefield
field
to investigate
the influence
of
An
testtest
to investigate
the influence
of under
under sleeper
pads(USPs)
trackand
quality
track was
dynamics
was
sleeper
pads(USPs)
on trackon
quality
track and
dynamics
conducted
Schweizerische
conducted
et the
al.Schweizerische
(2011) on the
by
Schneiderbyet Schneider
al. (2011) on
Bundesbahnen
test
Bundesbahnen
test site at This
Kiesen
Switzerland.
study
site
at Kiesen in Switzerland.
studyinconcluded
that theThis
placement
concluded
the placement
of USPs
in a ballasted
trackThe
changes
of
USPs in that
a ballasted
track changes
the track
performance.
track
the track performance. The track settlement increased with time
settlement increased with time when track was without USPs, and
when track was without USPs, and needed renewal of sleepers and
needed renewal of sleepers and re-tamping of ballast. The settlement
re-tamping of ballast. The settlement restarted over again when the
restarted over again when the track was loaded. But when USPs
track was loaded. But when USPs were used, the track settlement
were
used,tothedecrease
track settlement
appeared
decrease
with time.
appeared
with time.
The to
authors
reported
that The
the
authors
the varying
subgrade
condition
between
padded
varying reported
subgradethat
condition
between
padded
and unpadded
test
track
and unpadded test track sites made it difficult to draw any specific
33

Geotechnical Journal Vol. 6 No. 1 2014


Geotechnical Journal Vol. 6 No. 1 2014
Geotechnical
Journal
Vol. 6Vol.
No.61 No.
20141 2014
34 Geotechnical
Journal
sites made it difficult
to draw
any specificinrecommendation.
It was
recommendation.
It was
also mentioned
the study, the resilient
sites
made
it difficult
to study,
draw
any
specific
recommendation.
It was
also
mentioned
in
the
the
resilient
layer
reduced
sleeper
sites
made
it
difficult
to
draw
any
specific
recommendation.
It
was
layer
reduced sleeper
strains
but increased
rail and sleeper
also
mentioned
in increased
theflexural
study,rail
theand
resilient
layer
reduced
sleeper
flexural
strains
but
sleeper
accelerations
and
the
also
mentioned
in
the
study,
the
resilient
layer
reduced
sleeper
accelerations
and
the
contact
forces
between
the
USP
and
the
ballast
flexural forces
strainsbetween
but increased
railand
andthe
sleeper
accelerations
and the
contact
the USP
ballastaccelerations
bed were related
to
flexural
strains
but
increased
rail
and
sleeper
and
the
bed
wereforces
related
to the stiffness
thethe
USPs.
contact
between
the USPofand
ballast bed were related to
the
stiffness
of between
the USPs.the
contact
forces
USP
and
the
ballast
bed
were
related
to
the stiffness of the USPs.
the stiffness of the USPs.

Stiff USP
Stiff USP
Stiff
USP
No USP
No USP
No USP

track ordegradation
andrisesettlement
(Li energy
and Davis
2005).
The
track
vice versa, the
of high impact
accelerate
the track
track
degradation
and forces
settlement
(Li
and
Davis
2005).
The
magnitude
of
the
impact
is
very
high
within
the
very
short
(2
track
degradation
and
settlement
(Li
and
Davis
2005).
The
degradation
(Li isand
Davis
2005).
The
magnitude
of
magnitude
ofand
the settlement
impact
forces
very
high
within
the very
short
(2
10
msec)
impulse
duration
(Lee
et
al.
2005).
Therefore,
the
effects
magnitude
of
the
impact
forces
is
very
high
within
the
very
short
(2
the
impact
forces
is
very
high
within
the
very
short
(210
msec)
10
msec)
impulse
duration
(Lee
et
al.
2005).
Therefore,
the
effects
of impact
forces are
very significant
in2005).
the design
and utilization
of
10
msec)duration
impulse
duration
(Lee
et al.
Therefore,
the
effects
impulse
(Lee
et al.
2005).
Therefore,
the and
effects
of impact
of
impact
forces are
very
significant
in thetrack
design
utilization
of
concrete
sleepers
as very
parts significant
of the railway
structures
(Kumaran
of
impact
forces
are
in
the
design
and
utilization
of
forces
aresleepers
very significant
in the
the railway
design and
utilization
of(Kumaran
concrete
concrete
as
parts
of
track
structures
et al. 2002).
concrete
sleepers
of the
railway
track(Kumaran
structureset(Kumaran
sleepers
as
parts ofastheparts
railway
track
structures
al. 2002).
et al. 2002).
et
al.
2002).
3.2 Impact Forces
3.2 Impact
Forces
Impact
Forces
3.2
Impact
Forces
Usually,
the track
degradation is driven by the wheel/rail impact
Usually,
the track
is driven
by the
wheel/rail
impact
loads, referred
to asdegradation
static load and
peak loads.
Two
distinct types
of
Usually,
the track
degradation
is driven
by the
wheel/rail
impact
to
as static loadsharp
and
peak
loads.
Two
loads,
referred
distinct
types
of
peaks
(1)
an
instantaneous
peak;
and
(2)
a
much
longer
loads, referred
to
as
static
load
and
peak
loads.
Two
distinct
types
of
(1)anan
instantaneous
sharp
and
(2) alonger
muchduration
longer
peaks
(1)
instantaneous
sharp
peak;peak;
and (2)
a much
duration
peak of smaller
peaks
(1)gradual
an instantaneous
sharpmagnitude
peak; andwere
(2) observed
a much during
longer
duration
gradual
peak
ofetmagnitude
smaller
magnitude
were
observed
during
gradual
peak
ofJenkins
smaller
were observed
during
impact
termed
these
force
peaks
as P1
impact
loading.
al.
(1974)
duration
gradualJenkins
peak ofet smaller
magnitude
were force
observed
during
termed
these
peaks
as P1
impact
loading.
al.Ptermed
(1974)
loading.
Jenkins
et
al.
(1974)
these
force
peaks
as
P1
and
and
P
,
respectively.
These
and
P
are
respond
to
how
a
wheel
2
1
2
termed
these force
peaks
asP2,
P1
impact
loading.
Jenkins
et al.P(1974)
and
P
,
respectively.
These
and
P
are
respond
to
how
a
wheel
2
1
2
respectively.
These
P1
and
P2
are
respond
to
how
a
wheel
rolling
rolling
a short-pitch
These tonotations
were
and
P2, over
respectively.
Theseirregular
P1 and Pdefects.
how a wheel
2 are respond
rolling
over
a
short-pitch
irregular
defects.
These
notations
were
over
a short-pitch
irregular
defects.
These notations
were
adopted
adopted
by industry
and are
in common
use
today
to describe
rolling
over
a short-pitch
irregular
defects.
These
notations
were
adopted
byon
industry
and
are
in
common
use
today
to describe
by
industry
and
are inapplied
common
use
todaystructure
to use
describe
limitations
limitations
forces
to in
thecommon
track
(Indraratna
et on
al.
adopted
by
industry
and
are
today
to
describe
limitations
on
forces
applied
to
the
track
structure
(Indraratna
et
al.
forces
to2 the
track
structure
(Indraratna
et impact
al.
2011).
P1ettime
and
P
forces
observed
from
wheel/rail
force
2011). applied
P1 and
limitations
on
forces
applied
to
the
track
structure
(Indraratna
al.
and
P
forces
observed
from
wheel/rail
impact
force
time
2011).
P
1
2
P2
forces
observed
fromvehicle
wheel/rail
impact
forcerail
time
histories
when
histories
train
passes
awheel/rail
typical
joint
on
Chinese
and Pthe
observed
from
force
time
2011).
P1when
2 forces
histories
when
the
train
vehicle
passes
aare
typical
railimpact
joint
on
Chinese
mainline
tracksthe
at
various
train passes
speeds
injoint
Figure
5 tracks
(Zhai
the
train vehicle
passes
avehicle
typical
rail joint
onshown
Chinese
mainline
histories
when
train
a
typical
rail
on
Chinese
mainline
tracks at various train speeds are shown in Figure 5 (Zhai
andvarious
Cai 1997).
at
train speeds
are shown
in Figure
(Zhai in
andFigure
Cai 1997).
mainline
tracks
at various
train speeds
are 5shown
5 (Zhai
and The
Cai 1997).
P1 force is due to the inertia of the rail and sleepers resisting
and The
Cai 1997).
force
is
due
to
the
inertia
of
the
rail
and
sleepers
resisting
P
The
P1
force
is
due
to
the
inertia
of
the
rail
and
sleepers
resisting
the
the The
downward
wheel and
compression
of theresisting
contact
is due of
to the
inertia
of the
rail and sleepers
P11 forcemotion
the
downward
motion
of
the
wheel
and
compression
of
the
contact
downward
motion
of
the
wheel
and
compression
of
the
contact
zone
zonedownward
between motion
the wheel
andwheel
rail and
and compression
the force isofa the
very
high
the
of the
contact
zone
between
the
wheel
rail
and
the
a very
high
between
the
wheel
railand
thethan
force
isaaforce
very is
high
frequency
frequency
(>100
Hz)and
force
ofand
less
half
millisecond
in length.
zone
between
the
wheel
and
rail
and
the
force
is
a
very
high
frequency
(>100
Hz)
force
of
less
than
half
a
millisecond
in
length.
(>100
Hz) (>100
force
of
thanofhalf
athan
millisecond
insleepers,
length. Its
effects
Its effects
are
mostly
filtered
out
thehalf
rail aand
therefore,
frequency
Hz)less
force
lessby
millisecond
in length.
Its
effects
are
mostly
filtered
out
by
thesleepers,
rail and
sleepers,
therefore,
are
mostlyeffect
filtered
by the
railby
and
therefore,
its direct
its effects
direct
on out
ballast
or out
subgrade
settlement
is very minimum
Its
are
mostly
filtered
the
rail
and
sleepers,
therefore,
its
direct
effect
on
ballast
or subgrade
settlement
is P
very
minimum
effect
on ballast
or subgrade
is very
minimum
(Frederick
occurs
(Frederick
and Round
1985).
On the other
hand,
the
2 force
its
direct
effect
on
ballast
or settlement
subgrade
settlement
is very
minimum
force
occurs
(Frederick
and
Round
1985).
On
the
other
hand,
the
P
2
and
1985).
On the
other
hand,
thethan
P2hand,
force
at
a
lower
but
in
at aRound
lowerand
frequency
(30
On
90
Hz)
thethe
Poccurs
1 Pforce,
force
occurs
(Frederick
Round
1985).
the
other
2
but
in
at a lower(30to
frequency
(30 the
this
90
Hz)force
than
the
P1 force,
frequency

90
Hz)
than
P1
force,
but
in
comparison
to
static
still
classified
as
high
comparison
static
forces
P
2
force,
but
in
at
a
lower
frequency
(30

90
Hz)
than
the
P
1
force
still
classified
as
high
comparison
to
static
forces
this
P
2
forces
this force.
P2toforce
classified
high
force.
This
P2
frequency
Thisstill
P2 force
is due
the frequency
downward
movement
of
force
still classified
as high
comparison
static
this
Pas2 to
frequency
force.
This Pforces
is due
to
the downward
movement
of
2 force
the
vehicle
unsprung
mass
and
the
rail/sleeper
mass
and
causing
force
is dueforce.
to
theThis
downward
movement
of
the
vehicle
unsprung
mass
frequency
P
force
is
due
to
the
downward
movement
of
2
the
vehicle unsprung
mass
and
theunderneath
rail/sleeper the
masssleeper
and causing
compression
of
the
ballast
mass
which
and
the
rail/sleeper
mass
and
causing
compression
of
the
ballast
mass
the
vehicle unsprung
mass and
theunderneath
rail/sleeper the
masssleeper
and causing
compression
of the stresses,
ballast
mass
which
increases
thethecontact
and the
loads
on sleepers
and and
ballast.
underneath
which mass
increases
the contact
the
compression
ofsleeper
the stresses,
ballast
underneath
thestresses,
sleeper
which
increases
the
contact
and
the
loads
on
sleepers
and ballast.
forces
are
of
great
interest
to
the
track
designers.
Therefore,
the
P
2 andstresses,
loads
on
sleepers
ballast.
Therefore,
the
P2
forces
are
of
great
increases
the
contact
and
the
loads
on
sleepers
and
ballast.
Therefore, the P2 forces are of great interest to the track designers.
are of great interest to the track designers.
Therefore,
the track
P2 forces
interest to the
designers.

Soft USP
Soft USP
Soft USP
Medium USP
Medium USP
Medium USP

Figure 3Sleeper/ballast contact force


Figure
3Sleeper/ballast
contact
force
(data3Sleeper/ballast
sourced from Dahlberg
2010)
Figure
contact
force
(data sourced from Dahlberg
2010)
(data sourced from Dahlberg 2010)

3. DYNAMIC WHEEL-RAIL IMPACT FORCES


3.
DYNAMIC WHEEL-RAIL IMPACT FORCES
3.
DYNAMIC WHEEL-RAIL IMPACT FORCES
The wheel
and rail undergo significant
irregularities
during the life
3.
DYNAMIC
IMPACT
FORCES
The wheel and rail WHEEL-RAIL
undergo significant
irregularities
during the life
time
of the
the and
trackrail
structure.
irregularities
are discrete
discrete
inthe
nature
The
wheel
undergoThese
significant
irregularities
duringin
life
time
of
track
structure.
These
irregularities
are
nature
The
wheel
and
rail
undergo
significant
irregularities
during
the
life
time
of
the
track
structure.
These
irregularities
are
discrete
in
nature
and
usually
at
the
surface
of
the
rail
and
wheel.
The
higher
frequency
and of
usually
at the
surface
of irregularities
the rail and are
wheel.
Thein higher
time
the track
structure.
These
discrete
nature
and
usually
atbythe
surface
ofthese
the are
rail
and wheel.
The
higher
forces
created
these
irregularities
known
as dynamic
wheelfrequency
forces
created
byof
irregularities
are The
known
as
and
usually
at
the
surface
the
rail
and
wheel.
higher
frequency
forces
by
thesewhich
are
known
as
rail
impact
forces, created
which
higher
inirregularities
magnitude
than
dynamic
wheel-rail
impactare
forces,
are higher
in quasi-static
magnitude
frequency
forces
created
by
these
irregularities
are
known
as
dynamic
wheel-rail
impact
forces,
which
are
higher
in
magnitude
forces.
If
the
wheel
and
rail
surfaces
are
in
good
condition,
then
the
than quasi-static
forces.
If the
wheel
and rail
surfacesinare
in good
dynamic
wheel-rail
impact
forces,
which
are higher
magnitude
than quasi-static
forces.
If
the
wheel
and
rail
surfaces
are
in togood
wheel-rail
contact
force
would
be similar
towould
the
static
wheel
load
condition,
then
the
wheel-rail
contact
forcerail
be similar
the
than
quasi-static
forces.
If
the
wheel
and
surfaces
are
in
condition,
then
the
wheel-rail
contact
force
would
be
similar
togood
the
(Steffens
2005).
static wheelthen
loadthe
(Steffens
2005).
condition,
wheel-rail
contact
force
would
be
similar
to
the
static wheel load (Steffens 2005).
static
wheel
load
2005).
3.1
of impact
load
3.1 Sources
Sources
of(Steffens
impact
load
3.1 Sources of impact load
3.1
Sources of
impact
loadare caused by various sources such as
The wheel-rail
impact
forces
The wheel-rail
impact
forces
are
caused
by
various
sources
such as
wheel
flat,
wheel
shells,
worn
wheel
and by
rail,various
dippedsources
rails, turnouts,
The
wheel-rail
impact
forces
are
caused
such as
wheel
flat,insulated
wheel
shells,
worn
wheel
and
rail,
dipped
rails,
turnouts,
crossings,
joints,
expansion
gap
between
two
rail
segments,
expansion
gap
between
two
rail
segments,
wheel
flat,insulated
wheel shells,
worn
wheel and
dipped
rails,
turnouts,
crossings,
joints,
expansion
gap rail,
between
two
rail
segments,
rail
joint misalignment,
andtwo
railrail
corrugation
crossings,
joints,imperfect
expansionrail
gap weld
between
segments,
rail joint insulated
misalignment,
imperfect
rail
weld
and
rail
corrugation
of typical
sources of
(Indraratna
et al. 2011). Figure
4 shows
some and
rail
joint misalignment,
imperfect
rail weld
rail corrugation
(Indraratna
et al. 2011). Figure
4 shows
some of typical
sources of
irregularities.
(Indraratna
et
al.
2011).
Figure
4
shows
some
of
typical
sources
of
irregularities.
irregularities.

Worn wheel surface


Worn wheel surface
Worn wheel surface

Worn rail surface


Worn rail surface
Worn rail surface

Wheel flat
Wheel flat
Wheel flat

Dipped rail joint


Dipped rail joint
Dipped rail joint

Hollow rail weld


Hollow rail weld
Hollow rail weld

Humped rail weld


Humped rail weld
Humped rail weld

Figure 5 Wheel/Rail Impact Force

Figure
5 Wheel/Rail
Impact
(data
sourced
from ZhaiImpact
and CaiForce
1997)
Figure
5 Wheel/Rail
(data
sourced
from Zhai and CaiForce
1997)
(data sourced from Zhai and Cai 1997)

Since the P2 forces are of greater importance in the assessment of


are of greater
importance
in the assessment
of
Since
the P2 forces Jenkins
track the
degradation,
et al. importance
(1974) proposed
a theoretical
are of greater
in the assessment
of
Since
P2 forces Jenkins
track
degradation,
et
al.
(1974)
proposed
a
theoretical
equation
to calculateJenkins
P2 forces
dipped
The P2a force
in the
track
degradation,
et at
(1974)joints.
proposed
theoretical
equation
to
calculate
P2isforces
atal.dipped
joints.
Theunsprung
P2 force
in
the
equation
shown
belowP
dependent
on the
vehicle
mass,
equation
to
calculate
forces
at
dipped
joints.
The
P
force
in
the
2
2
equation
shown
below
is
dependent
on
the
vehicle
unsprung
mass,
track mass,
trackbelow
stiffness,
vehicle speed
and vehicle
joint dip angle. mass,
equation
shown
is dependent
on the
track mass,
track stiffness,
vehicle speed
and joint dipunsprung
angle.
track mass, track stiffness, vehicle speed and joint dip angle.

Mu
Ct
. Kt M u
P2
P0 2Vm
M u . 1
Ct

P
P 2V M M M . 1 4 K (C
M M ) . K M
P22
P00 2Vmm M uu uM tt . 1 4 Ktt ( Mt uu M tt ) . Ktt M uu (1)
M u M t 4 Kt ( M u M t )
(1)
(1)
where:
where:
P0 = Vehicle static single wheel load (kN)
where:
P
=
Vehicle
static
single
wheel
load (kN)
0
Mu
= Vehicle
Vehicle static
unsprung
mass
(kg)load
P
=
single
wheel
(kN)
0
Mu
=
Vehicle
unsprung
mass
(kg)
2 =
= Vehicle
Total joint
angle (rad)
Mu
unsprung
mass
(kg)
2 = Total joint angle (rad)
2 = Total joint angle (rad)

Rail corrugation
Rail joint misalignment
Rail corrugation
Rail joint misalignment
Rail corrugation
Rail joint misalignment
Figure 4 Wheel-Rail Irregularities causes impact forces
Figure 4 Wheel-Rail Irregularities causes impact forces
Figure 4 Wheel-Rail
impact forces
These abnormalities
on the Irregularities
wheel and railcauses
can generate
large impact
These
abnormalities
on
the
wheel
and
rail
can
generate
large
impact
forces abnormalities
between wheelonand
rail.
Theand
impact
loadgenerate
caused by
defects
on
These
the
wheel
rail
can
large
impact
forces
between
wheel androtates
rail. The
impact
load
caused
by defects
on
the
wheel
subsequently
with
each
wheel
rotation
and
roll
between wheel
wheeland
andrail.
rail.The
Theimpact
impact
load
caused
by
defects
forces
between
load
caused
by
defects
on
the
wheel
subsequently
rotates
with
each
wheel
rotation
and
roll
overthe
when
thesubsequently
defects arerotates
in
thewith
rail.
large
wheel
impactand
forces
on
wheel
rotates
withAeach
wheel
rotation
and
the
wheel
subsequently
wheel
rotation
roll
over
whenat
thethedefects
areand
in the
rail. each
Adue
large
wheel
impact
forces
generated
turnout
crossings
to wheel
traversing
of forces
wheel
in
the
rail.
over
when
the
defects
are
in
the
rail.
A
large
impact
generated
at the turnout and
crossings due toal.traversing
of wheel
over the rail
(Anastasopoulos
2009). Besides,
the
generated
at discontinuity
the turnout and
crossings dueet
toal.traversing
of wheel
over
the
rail
discontinuity
(Anastasopoulos
et
2009).
Besides,
the
rapidthe
change
of track stiffness
at the road crossing,
bridge
approach
over
rail
discontinuity
(Anastasopoulos
et
al.
2009).
Besides,
the
rapidtrack
change
of tracksuch
stiffness
at the slab
road track
crossing,
bridge
approach
and
transition
as
concrete
merging
to
ballasted
rapid
change
of tracksuch
stiffness
at the road
crossing,
bridge
and track
transition
as concrete
slab
track
merging
to approach
ballasted
track
or vice
versa,such
the rise
of highslab
impact
accelerate
the
and
transition
trackenergy
merging
to ballasted
tracktrack
or vice
versa, the as
riseconcrete
of high impact
energy
accelerate
the
track or vice versa, the rise of high impact energy accelerate the
34
34
34

rd
3Geotechnical
Proff 18-02-2015
Journal Vol. 6

No. 1 2014
Geotechnical Journal Vol. 6 No. 1 2014
Geotechnical Journal Vol. 6 No. 1 2014

Geotechnical Journal Vol. 6 No. 1 2014 35

V = Speed of vehicle (m/s)


V = Speed of=vehicle
(m/s)track stiffness (MN/m)
Equivalent
VKt =
of vehicle
(m/s)
2KSpeed
td
= Equivalent track stiffness (MN/m)
Kt 2Ktd
=
Equivalent track
track damping(kNs/m)
stiffness (MN/m)
Kt 32CKtdtd
=
Equivalent
Ct 3Ctd
= Equivalent track damping(kNs/m)
Ct 3C2td
=
Equivalent
track
damping(kNs/m)
Ct 3M
2
= Equivalent track mass (kg)
2 td
M t 3M
= Equivalent track mass (kg)
M t 3M2tdtd
= Equivalent track mass (kg)
M t 2 0.25
Ktd2 0.25 = Effective track length (m)
Ktd 0.25 = Effective track length (m)
4KEI
= Effective track length (m)

td
4 EI

Ballast
Stiffness per metre (MN/m/m)
Ktd =
4 EI

Ballast
Stiffness
Ktd =
=
Ballast
Dampingper
per metre
metre(MN/m/m)
(kNs/m/m)
C
=
Ballast
per
metre
(MN/m/m)
Ktd
td = Ballast Stiffness
Damping
perper
metre
(kNs/m/m)
C
td = Rail + Sleeper
mass
metre
(kg/m)
M
(kNs/m/m)
Ctdtd = Ballast Damping per metre
Mtd = Rail + Sleeper mass per metre (kg/m)
Mtd = Rail + Sleeper mass per metre (kg/m)
Internationally similar limits are placed for the safety of the track.
Internationally
similar
limits
placed forBoard
the safety
of the
track.
The British Rail
Safety
andare
(RSSB)
Railway
Internationally
similar
limits
areStandards
placed for the safety
of the
track.
The
British
Rail
Safety
and
Standards
Board
(RSSB)
Railway
GroupBritish
Standard
(GM/TT0088):Permissible
Forces
for
The
Rail Safety
and Standards BoardTrack
(RSSB)
Railway
Group
Standard
(GM/TT0088):Permissible
Track
Forces
for
Railway Standard
Vehicles (1993)
states that when a vehicle
(Class
55 Deltic
Group
(GM/TT0088):Permissible
Track
Forces
for
Railway Vehicles
(1993)
states that
when
a vehicle (Class
55 Deltic
locomotive)
negotiates
aa vertical
ramp
discontinuity
at
Railway
Vehicles
(1993)
states that
when
a vehicle (Class
55 Deltic
locomotive)
negotiates
vertical
ramp
discontinuity
at its
its maximum
maximum
locomotive)
negotiates
a
vertical
ramp
discontinuity
at
its
maximum
force
produced
design
operating
speed(160
(160
km/h)
the
total
P2 produced
locomotive)
negotiates
a vertical
ramp
discontinuity
at its maximum
design
operating
speed
km/h)
the
total
P2
force
should
produced
design
operating
speed
the totalAustralian
P2 force standards
should
not 322
exceed
322(160
kN km/h)
per
wheel.
produced
design
operating
speed
(160
km/h)
the total
P2 forcerecommend
not
exceed
kN
per
wheel.
Australian
standards
should
not Jenkinss
exceed formula
322 kNto per
wheel.
Australian
standards
forces
and
the
recommend
calculate
P 2 specify
should
exceed
322 kN P2per
wheel.
Australian
standards
Jenkinssnot
formula
to calculate
forces
and
the specify
guidelines
and
specify
the
recommend
Jenkinss
formula
tolimit
calculate
P 2 forces
guidelines
shown
in
Table
1
to
P
forces
as
a
function
of
track
2 function
and and
specify
the
recommend
Jenkinss
formula
to calculate
P 2 forces
shown
in
Table
1
to
limit
P2
forces
as
a
of
track
vehicle
guidelines
shown
in
Table
1
to
limit
P
forces
as
a
function
of
track
2
and
vehicleshown
characteristics
(Indraratna
al. 2011).
guidelines
in Tableet
1 al.
to limit
Pet
forces
as a function of track
2
characteristics
(Indraratna
2011).
and vehicle characteristics (Indraratna et al. 2011).
and vehicle characteristics (Indraratna et al. 2011).
Table 1 Limiting P2 forces (QR 2001; RIC 2002; ARA 2003 )
Table 1 Limiting P forces (QR 2001; RIC 2002; ARA 2003 )
Table 1 Limiting P22 forces (QR 2001; RIC 2002; ARA 2003 )
Track
Track
Class
Track
Class
Class
1
12
12
32
34
34
54
5
5

Maximum P2
Maximum
2
Force P
Maximum
Force P2
Locomotives
Force
Locomotives
(kN)
Locomotives
(kN)
295
(kN)
295
230
295
230
200
230
200
180
200
180
130
180
130
130

Maximum P2
Maximum
P2
Force Other
Maximum
P2
Force
Other
Rolling
Force
Other
Rolling
Stock(kN)
Rolling
Stock(kN)
230
Stock(kN)
230
230
230
230
230
230
230
180
230
180
130
180
130
130

Kt
Kt
(MN/m)
Kt
(MN/m)
(MN/m)
110
110
110
110
110
95.8
110
95.8
90.3
95.8
90.3
83.6
90.3
83.6
83.6

Ct
Ct
(kNs/m)
Ct
(kNs/m)
(kNs/m)
52.5
52.5
48
52.5
48
45.9
48
45.9
43.2
45.9
43.2
40
43.2
40
40

Drop
Drop
Hammer
Drop
Hammer
&
Hammer
Load&
&Cell
Load Cell
Load Cell
Impact
Impact
Load
Impact
Load
Quick
Load
Quick
Release
Quick
Release
System
Release
System
System

Mt
Mt
(kg)
Mt
(kg)
(kg)
135
135
117
135
117
106
117
106
95
106
95
85
95
85
85

Data
Data
Acquisition
Data
Acquisition
System
Acquisition
System
System

Figure 6 Impact load Testing Apparatus


Figure 6 Impact load Testing Apparatus
Figure 6 Impact load Testing
Apparatus
Impact

50 mm
50 mm
30
50 mm
mm
30 mm
30 mm

Load
Load
Cell
Load
Cell
Cell

Accelerometer
Accelerometer
Accelerometer

4.
LABORATORY
4. Laboratory
testing TESTING
4.
LABORATORY TESTING
4.
LABORATORY
TESTING
In
In this
this study
study the
the use
use of
of energy
energy absorbing
absorbing shock
shock mats
mats to
to mitigate
mitigate the
the
In
this
study
the
use
of
energy
absorbing
shock
mats
to
mitigate
the
ballast
degradation
under
impact
loading
was
assessed
aa series
of
In
this study
the useunder
of energy
absorbing
shock
mats toby
mitigate
the
ballast
degradation
impact
loading
was
assessed
by
series
of
ballast
degradation
under
impact
loadingstresses
was assessed
by a series
of
laboratory
testing.
The
typical
dynamic
in
the
range
of
400ballast
degradation
under
impact
loading
was
assessed
by
a
series
of
laboratory testing.
testing. The
typical
dynamic stresses
stresses in
in the
the range
range of
of 400400laboratory
The
typical
dynamic
600
by
wheel-flat
and dipped
dipped
rail (Jenkins
(Jenkins
et al.
al.of1974;
laboratory
testing.
typical dynamic
stresses
in the range
400600 kPa
kPa caused
caused
byThe
wheel-flat
and
rail
et
1974;
600
kPa
caused
by
wheel-flat
and dipped
rail
(Jenkinssimulated
et al. 1974;
Steffens
and
Murray
2005;
Indraratna
et
al.
2010)was
by
600
kPa
caused
by
wheel-flat
and
dipped
rail
(Jenkins
et
al. 1974;
Steffens
and
Murray
2005;
Indraratna
et
al.
2010)was
simulated
Steffenstheandlarge-scale
Murray 2005;
Indraratna
et al.facility
2010)was
simulated
by
using
impact
load
test
available
at
the
Steffens
and
Murray
2005;
Indraratna
et
al.
2010)was
simulated
by
by using
large-scale
impactload
loadtest
testfacility
facility available
available at
at the
the
using
the the
large-scale
impact
University
of
Wollongong,
Australia.
using
the of
large-scale
impact
load test facility available at the
University
Wollongong,
Australia.
Australia.
University of Wollongong, Australia.
4.1 Test
Test
apparatus,
Impact
Loading
Instrumentations
apparatus,
Impact
Loading
andand
Instrumentations
4.1 Test apparatus, Impact Loading and Instrumentations
4.1
Test
apparatus,
Impact
Loading
and
The impact loading test facility available Instrumentations
at the University of
The
impact (Figure
loading 6)test
availabledrop
at weight
the University
of
Wollongong
isisafacility
capacity
impact
test
(Figure 6)
ahigh
high
capacity atdrop
impact
The
impact loading
test
facility
available
the weight
University
of
Wollongong
(Figure
6)
is
a
high
capacity
drop
weight
impact
test
machine.
It (Figure
can
be 6)
hoisted
mechanically
totoweight
the
which
test machine.
It can
be
mechanically
the height
height
Wollongong
ishoisted
a high
capacity drop
impact
test
machine. It tocan
be
hoisted
mechanically
to theand
height
which
corresponds
the
required
impact
load load
magnitude
drop
machine.
It to
can
be
hoisted
mechanically
to
the height
which
corresponds
the
required
impact
magnitude
andheight
drop
corresponds
to
the
required
impact
load
magnitude
and
drop
height
through
guided
roller
onroller
vertical
to the
strong
concrete
corresponds
to the
required
impact
load fixed
magnitude
and
height
height through
guided
on column
vertical
column
fixed
todrop
the
strong
through
guided
roller
on
vertical
column
fixed
to
the
strong
concrete
floor.
The
efficiency
ofvertical
the hammer
velocity
dueconcrete
to
through
guided
roller
on
column
fixedvelocity
toisthe98%
strong
concrete
floor.
The
efficiency
the hammer
is 98%
duethe
to
floor.
The
efficiency
of the of
hammer
velocity
isRemennikov
98%
due to
the
friction
of
the
guiding
column
(Kaewunruen
and
2010).
floor.
The
efficiency
of
the
hammer
velocity
is
98%
due
to
the
the
friction
of
the
guiding
column
(Kaewunruen
and
Remennikov
friction
of the
columndrop
(Kaewunruen
Remennikov
2010).
is calculated
Therefore,
the guiding
actualthehammer
heightdrop
( and
friction
of the
guiding
column
(Kaewunruen
and
Remennikov
2010).
2010). Therefore,
actual drop
hammer
height
(h=V^2/2g)
is calculated
Therefore,
the
actual
hammer
height
(a factor
2is
multiplying
the
theoretical
drop
height
by
1.04
(i.e.,
1/0.98
calculated
Therefore,
the
actual hammer
drop height
( height by aisfactor
2).
calculated
multiplying
the
theoretical
drop
1.04
multiplying
thedrop
theoretical
drop
height
by aoffactor
1.04
(i.e.,it 1/0.98
2).
The
free
fall
hammer
is
a
weight
592
kg
and
can
be
multiplying
the
theoretical
drop
height
by
a
factor
1.04
(i.e.,
1/0.98
).
(i.e., free
1/0.982).
The free
fall drop
is a592
weight
of 592
kg and
The
fall drop
hammer
is a hammer
weight
of
kgtheand
it can
be
dropped
from
a
maximum
height
of
6
m
from
base
of
the
The
free
fall
drop
hammer
is
a
weight
of
592
kg
and
it
can
be
it can be from
dropped maximum
from a maximum
height
6 m from
the base
of
dropped
height
6 moffrom
base
of the
concrete
floor. aaThe
impact load
wasof
andthe
dropped
from
maximum
height
ofmeasured
6measured
m from
therecorded
base ofby
theaa
the concrete
floor.
The
impact
loadwas
was
and
recorded
by
concrete
floor.
The
impact
load
measured
and
recorded
by
dynamic
a capacity
1,200
kN mounted
at the bottom
concrete load
floor.cell
Theofimpact
loadofwas
measured
and recorded
by a
kN mounted
at the bottom
dynamic
load cell
a capacitytoofa1,200
of
the hammer
andof
data acquisition
Ballast
dynamic
load cell
ofconnected
a capacity of 1,200
kN mountedsystem.
at the bottom
of
the hammer
connected
to a data acquisition
system.
deformation
andand
transient
acceleration
of the impact
loadsBallast
were
of the hammer
and
connected
to a data acquisition
system.
Ballast
deformation
and
transient acceleration
ofofthe
impact loads
were
captured
by
a
piezoelectric
accelerometer
a
capacity
of 10,000g
deformation and transient acceleration of the impact loads
were
captured
bygravitational
piezoelectric
accelerometer
capacity
10,000g
aa piezoelectric
accelerometer
ofofa acapacity
10,000g
(g
(g
is theby
acceleration)
connected
at theof of
top
of the
captured
by a piezoelectric
accelerometer
of a capacity
of
10,000g
(g
is
the
gravitational
acceleration)
connected
at
the
top
of
the
is the gravitational
acceleration)
connected
at the top of the sample
sample
load
plate
shown
in
Figure
7.
(g is the gravitational acceleration) connected at the top of the
sample
load
plate in
shown
in 7.
Figure 7.
load plate
shown
Figure
sample
load
plateSpecifications
shown
in Figure 7.
4.2 Material
4.2 Material
Material
Specifications
Specifications
4.2 materials
Material
Specifications
The
used
in this study are the ballast, shock mats and the
The
materials
used
in thisspecifications
study are the ballast,
matsareand
the
weak
and
hard
base.
these shock
materials
given
The materials used inThe
this study are the of
ballast,
shock
mats and
the
weak
and
hard
base.
The
specifications
of
these
materials
are
given
weak
and
hard
base.
The
specifications
of
these
materials
are
given
in
following
sections.
weak
and hard
base. The specifications of these materials are given
in
following
sections.
in following
following sections.
sections.
in

Steel Plate
Steel Plate
ShockPlate
Mat
Steel
Shock Mat
Shock Mat

Accelerometer
Accelerometer
Accelerometer
To data logger
To data logger
To data logger

Ballast
Ballast
Ballast

300 mm
300 mm
300 mm

Rubber
Rubber
Membrane
Rubber
Membrane
30 mm
Membrane
30 mm

30 mm
mm
100
100 mm
100 mm
50 mm
50 mm
50 mm

Impact
Load
Impact
Load
Load

Shock Mat
Shock Mat
Shock
Mat
Base

Base
Base

Steel Plate
Steel Plate
Steel Plate
300 mm
300 mm
300 of
mmthe
Diagram

Rubber
Rubber
Membrane
Rubber
Membrane
Membrane

Figure 7 Schematic
Test Specimen
Figure 7 Schematic Diagram of the Test Specimen
Figure 7 Schematic Diagram of the Test Specimen
4.2.1 Ballast
4.2.1 Ballast
4.2.1railway
Ballastballast material commonly used in New South Wales
The
The
railway
ballastis material
commonly
used in
New South
Wales
(NSW),
Australia
Latite basalt,
a common
igneous
rock can
be
The
ballast material
commonly
used in
New South
Wales
4.2.1railway
Ballast
(NSW),
Australia
iscost
Latite
basalt,
a common
igneous
rock can
be
found
in
the
south
of
NSW
and
closer
to
Wollongong
City,
(NSW), Australia is Latite basalt, a common igneous rock can
be
found
in
the
south
cost
of
NSW
and
closer
to
Wollongong
City,
The railway
ballast
material
commonly
used to
involcanic
New South
Wales
Australia.
The
aggregates
from
crushed
basalt
are
found
in the
south
cost
ofmade
NSW
and closer
Wollongong
City,
Australia.
The
aggregates
made
from
crushed
volcanic
basalt
are
(NSW),
Australia
is Latite
a from
common
igneous
rock
canbasalt
becorners
found
dark,
fine
grained
and basalt,
very
with
sharp
angular
Australia.
The
aggregates
madedense
crushed
volcanic
are
dark,
fine
and very
dense
sharp
angular
corners
in the south
cost ofrailway
NSW
and
closer
towith
Wollongong
City, and
Australia.
suitable
for grained
fresh
ballast
material.
The
physical
index
dark,
fine
grained
and very
dense
with
sharp
angular
corners
suitable
foroffresh
railway
ballast
material.
Thebasalt
physical
and
index
properties
themade
freshfrom
ballast
were
evaluated
as per
2758.7
The aggregates
crushed
volcanic
areAS
dark,
fine
suitable
for fresh
railway
ballast
material.
The physical
and
index
properties
of
the fresh
ballast
were
evaluated
asa per
AS 2758.7
(1996)
and
discussed
by
Indraratna
et
al.
(1998)
in
previous
study.
grained
and
very
dense
with
sharp
angular
corners
suitable
for
fresh
properties of the fresh ballast were evaluated as per AS 2758.7
(1996)
and discussed
by this
Indraratna
et al.prepared
(1998) in
aaccordance
previous study.
The
ballast
material
for
was
in properties
with
railway
ballast
material.
Thestudy
physical
index
the
(1996)
and
discussed
by Indraratna
et al.and
(1998)
in
a previous of
study.
The
ballast
material
for this study
wasAS
prepared
in(1996).
accordance
current
practice
in
Australia
as
per
2758.7
The with
raw
The
for this as
study
prepared
in accordance
with
freshballast
ballastmaterial
were evaluated
per was
AS 2758.7
(1996)
and discussed
current
practice was
in Australia
as cleaned
per ASby2758.7
(1996).
The
raw
ballast
material
thoroughly
water
and
dried
before
current
practice
in (1998)
Australia
per ASstudy.
2758.7
The
raw
by Indraratna
et was
al.
in aasprevious
The(1996).
ballast
material
ballast
material
thoroughly
cleaned
by
water
and
dried
before
sieving.
The
particle
size
distribution
(PSD)
of
thecurrent
ballast
material
is
ballast
material
wasprepared
thoroughly
cleaned
by with
water
and dried
before
for
this
study
was
in
accordance
practice
in
sieving.
particle
sizebasic
distribution
of the from
ballast
is
shown
inThe
Figure
8.
The
martial(PSD)
parameters
thematerial
PSD was
are
sieving.
The
particle
size
distribution
(PSD)
of the
ballast
material
is
Australia
as
per
AS
2758.7
(1996).
The
raw
ballast
material
shown
in
Figure
8.
The
basic
martial
parameters
from
the
PSD
are
listed
ininTable
2. 8. The
shown
Figure
basic martial parameters from the PSD are
thoroughly
cleaned
listed
in Table
2. by water and dried before sieving. The particle
listed
in Table 2. (PSD) of the ballast material is shown in Figure 8.
size distribution
The basic martial parameters from the PSD are listed in Table 2.

35
35
35

Geotechnical
Journal
Vol. 6Vol.
No.61 No.
20141 2014
36
Geotechnical
Journal
Table 2Material Parameters of Ballast and Sand

100
90
80

Percentage Passing

70

Fresh ballast

60

Australian Standard
AS 2758.7 (1996)

50
40
30
20
10
0

10

Parameters

Fresh Ballast

Particle Shape
Type of Gradation
Max. particle size, mm (Dmax)
Min. particle size, mm (Dmin)
Effective size, mm (D10)
Uniformity Coefficient (Cu)
Coefficient of Gradation (Cc)

Angular
Uniformly graded
63.0
19.0
24.0
1.6
1.0

Sand
Subgrade
Poorly graded
4.75
0.075
0.24
2.3
1.0

4.3 Laboratory Test Setup


4.3 thickness
Laboratory
Setup
The
of theTest
ballast
layer in Australian rail track is 250-300
mm
(the lowerofthicknesses
at in
theAustralian
bridge deck).
Therefore
a 300
The thickness
the ballast are
layer
rail track
is 250-300
mm (the
thicklower
ballast
layer was are
selected
the specimen
height ina 300
this
thicknesses
at the as
bridge
deck). Therefore
mm thick
layer was
selected
as the to
specimen
in this
study.
300ballast
mm ballast
thickness
is found
be moreheight
realistically
study. 300 site
mmcondition
ballast thickness
found study
to be on
more
realistically
simulating
as per the is
previous
ballast
material
simulating on
site
condition
as per
the previous
study on
ballast
conducted
large
scale triaxial
or cubical
test apparatus
by Brown
material
conducted
on large scale
or cubical
test apparatus
et
al. (2007)
and Indraratna
et al.triaxial
(2007).The
inclusion
of shock
by Brown
al. (2007)
and Indraratna
al. (2007).The
of
mats
at theettop
and bottom
of ballast etlayer
brings the inclusion
total height
shock
mats foundation
at the top and
ofvalue.
ballastInlayer
the total
of
the track
morebottom
realistic
orderbrings
to simulate
the
heightdensity
of the(approximately
track foundation
realistic
value.
Intracks,
order the
to
field
1560more
kg/m3)
for heavy
haul
3
) for
heavy
simulatematerial
the field
(approximately
1560 by
kg/m
ballast
wasdensity
compacted
in several layers
using
a rubber
haul tracks,
the ballast
material
compacted
in several
layers
by
padded
hammer.
The low
lateralwas
confining
pressure
for the
ballast
using
a
rubber
padded
hammer.
The
low
lateral
confining
pressure
was simulated by placing a cylindrical rubber membrane around the
for the ballast was simulated by placing a cylindrical rubber
specimen. The rubber membrane (thickness of 7 mm) was capable of
membrane around the specimen. The rubber membrane (thickness of
prevent piercing or cutting the membrane by sharp corners of ballast
7 mm) was capable of prevent piercing or cutting the membrane by
particles.
sharp corners of ballast particles.
of base
condition
1) relatively
TheThe
two two
typestypes
of base
condition
used used
were,were,
1) relatively
weakweak
base
base represented
a 100
thick
sandlayer
layervibro-compacted
vibro-compacted to
to a
represented
by aby100
mmmm
thick
sand
base
density of
of 1620
1620 kg/m3and
kg/m3and placed
density
placed under
under the
the ballast
ballast bed,
bed, 2)
2) hard
hard base
condition represented
represented by
by aa rigid
plate of
of thickness
This
condition
rigid steel
steel plate
thickness 50
50 mm.
mm. This
hard
base
condition
is
represented
by
the
tracks
running
on
steel
hard base condition is represented by the tracks running on steel
bridge deck
deck or
or track
track foundation
located on
on hard
hard bed
bed rock.
rock. Three
Three
bridge
foundation located
layers
of
shock
mats
(total
thickness
of
30
mm)
were
used
at
the
top
layers of shock mats (total thickness of 30 mm) were used at the top
and
bottom
of
the
ballast
specimen
(Figure
7).
and bottom of the ballast specimen (Figure 7).

100

Sieve Size (mm)

Figure 8 Particle Size Distribution (PSD) of the ballast material


4.2.2 Sand Subgrade
In order to simulate a typical weak base condition, a thin layer of
sand subgrade cushion was used in the laboratory testing. The sand
parameters are listed in Table 2.
4.2.3 Shock Mats
There are many manufacturers of the USP and UBM around the
world and some of the manufacturers listed by (Bolmsvik 2005).
One of such manufacturers USP and UBM with its material
parameters are shown in Figure 9 (a). USPs are generally stiffer than
UBMs as they are placed adjacent to higher stress zones i.e. sleeperballast interface. The rubber shock mats used in this study was a 10
mm thick made of recycled rubber granulates of 1 to 3 mm particle
size, bound by polyurethane elastomer compound. A sample of
shock mat and its material parameters are shown in Figure 9 (b).

4.4 Test
Procedure
Test
Procedure

Under Sleeper Pad (USP)


Thickness including mounting
mesh = approx. 15 mm
Weight = 4.2 kg/m2
Bedding Modulus
Cstat= 0.22 N/mm3
Tear Strength of the connection
USP-Concrete Sleeper
Minimum = 0.4 N/mm2
Average = 0.5 N/mm2

Each test specimens were placed on the concrete floor under the
impact load hammer. The hammer was hoisted to the required drop
andreleased
releasedbyby
an electronic
release
system.
The
height and
an electronic
quickquick
release
system.
The ballast
ballast specimens
werewith
tested
with
and without
shock
mats at
placed
at
specimens
were tested
and
without
shock mats
placed
the top
the top
the of
bottom
of thelayer.
ballast
layer.
Theloading
impactwas
loading
was
and
the and
bottom
the ballast
The
impact
repeated
repeated
for 10
each sample.
It wasthat
found
thedue
strain
for
10 times
fortimes
each for
sample.
It was found
the that
strain
to
due to impact
is attenuating
after number
certain number
blows
impact
loadingloading
is attenuating
after certain
of blowsof(typical
8 or 9 blows).
Automatic
triggering
of impact
loading
8(typical
or 9 blows).
Automatic
triggering
of impact
loading
signalsignal
was
was
enabled
and
data
at
sampling
frequency
of
50,000
Hz was
enabled and data at sampling frequency of 50,000 Hz was collected
collected
byacquisition
the data acquisition
system.
Tothe
remove
noise
in the
by
the data
system. To
remove
noise the
in the
data,
the
data, the raw impact load-time history data were digitally filtered
raw impact load-time history data were digitally filtered using lowusing low-pass fourth order Butterworth filter with a cut-off
pass fourth order Butterworth filter with a cut-off frequency of 2,000
frequency of 2,000 Hz. Ballast deformation and transient
Hz. Ballast deformation and transient acceleration of the impact load
acceleration of the impact load data were collected by data
data
were collected
by the
datapiezoelectric
acquisition system
by the connected
piezoelectric
acquisition
system by
accelerometer
at
accelerometer
connected
at
the
top
of
the
sample
plate.
the top of the sample plate.

Under Ballast Mat (UBM)


Thickness = approx. 17 mm
Weight = 10.5 kg/m2
Specific Static Stiffness
Cstat= 0.15 N/mm3
Tensile Strength = 1.3 N/mm2

(a)

10 mm

(b)

4.5
load-time
history
4.5 Impact
Impact
load-time
history
The
impact
load
was
dropped
on the
and after
firstthe
impact
The impact load was dropped
on sample
the sample
and the
after
first
the
hammer
rebounded
on the sample
couple of
time of
then
thethen
impact
impact
the hammer
rebounded
on the sample
couple
time
the
load
attenuated
with timewith
as shown
distinct
impact
load attenuated
time in
asFigure
shown10.inTwo
Figure
10. types
Two
of
peakstypes
were observed
impact loading
named
as P1 and
distinct
of peaks during
were observed
duringand
impact
loading
and
P2
as per
et
al.
(1974).
The
peak
P1
related
to
the
and
P
as
per
Jenkins
et
al.
(1974).
The
peak
Pmultiple
named
as PJenkins
1
2
1 related
impacts
including
the first
impactthe
from
free fall
to the multiple
impacts
including
firstthe
impact
fromhammer
the freedrop
fall
and
the hit
from
hammer.
The hammer.
single peak
is related
hammer
drop
andrebounded
the hit from
rebounded
TheP2single
peak
to
mechanical
the ballast
to its
significant
is related
to the resistance
mechanicalofresistance
ofleading
the ballast
leading
to its
P2 the
compression
(Saxton et (Saxton
al. 1974).
The1974).
P2 peak
the
peak than
is lesser
significant compression
et al.
TheisP2lesser

Recycled Rubber Shock Mat


Tensile Strength = 600 kN/m2
Tensile Strain at Failure = 80%
Modulus at 10% compressive
strain = 3800 kN/m2

Figure 9 (a) Sample of USP and UBM; (b) Shock Mat used for
laboratory testings in this study
36

rd
Journal
3Geotechnical
Proff 18-02-2015
Geotechnical
Journal Vol.
Vol. 66

No.
No. 11 2014
2014
Geotechnical Journal Vol. 6 No. 1 2014

Geotechnical Journal Vol. 6 No. 1 2014 37


(2)
(2)
(2)
(3)
(3)
(3)
The variation
variation
shear
volumetric
with
the
of
variationofof
ofshear
shear
and
volumetric
strain
withnumber
the number
number
of
The
andand
volumetric
strainstrain
with the
of impact
impact
blows
are
shown
in
Figure
12a
and
12b,
respectively.
In
The
variation
of
shear
and
volumetric
strain
with
the
number
of
impact are
blows
are in
shown
in 12a
Figure
12b, respectively.
In
blows
shown
Figure
and12a
12b,and
respectively.
In general
general
both
the
shear
and
volumetric
strains
increased
in
the
initial
impact
blows
are
shown
in
Figure
12a
and
12b,
respectively.
In
general
both
the
shear
and
volumetric
strains
increased
in
the
initial
both the shear and volumetric strains increased in the initial impact
impact
loadings
and
eventually
become
constant
at
the
end
of
general both
the shear
volumetric
strains
increased
initial
impact
loadings
and and
eventually
become
constant
at in
thetheblows
end
of9
loadings
and eventually
become
constant
at the
end of impact
impact
blows
9 and
This
is
the
ballast
layer
displays
a
loadings
and10.
become
at the
end of
impact
and
10.eventually
This
is because
because
theconstant
ballast
layer
displays
and
10. blows
This
is9because
the
ballast
layer
displays
a strong
tendency
toa
strong
tendency
to
compact
under
repetitive
loading
due
to
impact
blows
9
and
10.
This
is
because
the
ballast
layer
displays
strong
tendency
to
compact
under
repetitive
loading
due
to
compact under repetitive loading due to rearrangement, reorientationa
rearrangement,
reorientation
andunder
breakage
of
corners
of
the
ballast
strong
tendency
to compact
repetitive
loading
to
rearrangement,
breakage
of (Lackenby
corners
of et
theal.due
ballast
and
breakage
of reorientation
corners
of 2007;
theand
ballast
particles
2007;
particles
(Lackenby
et
al.
et
al.
2010)
rearrangement,
reorientation
andIndraratna
breakage of
corners
ofand
thebecome
ballast
particles
(Lackenby
et
al.
2007;
Indraratna
et
al.
2010)
and
become
Indraratna
et al.
2010)
and become are
stable when the ballast
particles
stable
when
the
ballast
rearranged
and
particles
(Lackenby
et al. particles
2007; Indraratna
et al. 2010)
and become
stable
when
the
ballast
particles
are completely
completely
rearranged
and
are
completely
rearranged
and
densified.
densified.
stable when the ballast particles are completely rearranged and
densified.
densified.
24
24
22
24
22

Shear
Shear
Shear
Strain,
Strain,
Strain,
q (%)
qq(%)
(%)

16
18
16
14
16
14

Impact
Impact
Impact
Force,
Force,
Force,
P2 P
(kN)
P2 2(kN)
(kN)

12
14
12
10
12
10
8
10
8
6
8
6

0
2
0

0
00
0

45
40
40
40
35
35
35
30
30
30
25
25
25
20
20
20
15
15

050
0
00
0

11
1

22
2

Without
Without Shock
Shock mat
mat
Shock
at
and
Without
Shock
Shock mat
mat
at top
topmat
and bottom
bottom
Shock mat at top and bottom

33
44
55
66
77
3 Number
4
5
6
7
Number of
of Blows,
Blows, N
N

88
8

99
9

1
1
1

2
2
2

3
4
5
6
7
3
4
5
6
7
of
3 Number
4
5 Blows,
6 N
Number
of
Blows,
N7

8
8
8

9
9
9

10
10
10

Number of Blows, N

(a)
(a)
(a)

Hard
Hard base
base
Weak
base
Hard base
Weak
base
Weak base

15
10
10
1055

4
6
4
2
4
2

Figure
Histories
Time, t (sec)
Figure 10
10 Impact
Impact Load-Time
Load-Time
Histories (Hard
(Hard base)
base)
Figure 10 Impact Load-Time Histories (Hard base)

55
50
50
50
45
45

Without Shock mat

18
20
18

350
350
Hard
Hard base
base
350
300
Hard base
300
300
A Without Shock mat
250
A Without Shock mat
250
B Shock mat placed at top and bottom
A
B Shock
Without
Shock
matat top and bottom
250
mat
placed
200
B Shock mat placed at top and bottom
200
Multiple
200
Multiple Force
Force Peaks
Peaks (P
(P11))
150
Multiple
Force
Peaks
(P1)
150
A
A
150
A
100
Single
100
Single Force
Force Peak
Peak (P
(P22))
100
Single
Force
Peak
(P2)
B
50
A
B
B
50
A
B
B
50
A
B
00
0.00
0.02
0.04
0.06
0.08
0.10
0.12 0.14 0.16 0.18 0.20
00.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20
Time,
t
(sec)
0.00 0.02 0.04 0.06 0.08
0.10
0.12 0.14 0.16 0.18 0.20
Time, t (sec)

60
60
60
55
55

Hard
Hard base
base
Weak
base
Weak
base
Hard base
Weak base

With
With Shock
Shock mat
mat
Without
Shock
Without
Shock
mat
With Shock
matmat

20
22
20

Volumetric
Volumetric
Volumetric
Strain,
Strain,
Strain,
v (%)
v v(%)
(%)

Impact
Impact
Impact
Forces,
Forces,
Forces,
P1 P
&P1 1P&&
P
(kN)
P (kN)
(kN)
2 22

than
It
evident
Figure
10,
the
instantaneous
P1 peaks. P
evident
from
Figurefrom
10, the
shock
than the
the instantaneous
instantaneous
PIt11 ispeaks.
peaks.
It is
is
evident
from
Figure
10,mats
the
peak)
shock
mats
are
attenuating
the
impact
force
(reduces
the
P
2
than
the
instantaneous
P1 force
peaks.
It is evident
from
the
are
attenuating
impact
the P2
peak)Figure
and
shock
mats
arethe
attenuating
the (reduces
impact
force
(reduces
theextending
P 210,
peak)
and
extending
time
duration
impact
load.
shock
mats
arethe
the of
impact
force
the
duration
of
impact
load.
andtime
extending
theattenuating
time
duration
of
impact
load.(reduces the P 2 peak)
Railway
group
standards
recommends
and U.K.
extending
the time
of impact
load. considering
force
U.K.
Railway
groupduration
standards
recommends
considering P
P 22 force
U.K.
Railway
group
standards
recommends
considering
P2 Pforce
in
in
the
track
design
criteria
as
it
is
the
direct
influence
the
U.K.
Railway
group
standards
recommends
considering
in the track design criteria as it is the direct influence on
on
the
2 force
the
track
design
criteria
as
it
is
the
direct
influence
on
the
degradation
forces
variation
with
degradation
of
track
bed.
Therefore
the
P
2 forces
in the track ofdesign
as it is the
the Pdirect
influence
onwith
the
variation
degradation
track criteria
bed. Therefore
2
of
track bed.
variation
continuous
continuous
impact
loading
is
the
major
in
this
study
variation
degradation
ofTherefore
track
bed.theTherefore
theconcern
P 2 forces
continuous
impact
loading
is P2
the forces
major
concern
inwith
this
study with
with
impact
loading
isdegradation.
the
majoris concern
in concern
thisplotted
study
respect
to
each
blow
respect
to
ballast
The
P
continuous
impact
loading
the
in with
this study
with
forces
plotted
with
each
blow
respect
to
ballast
degradation.
Themajor
P 22 forces
is
Figure
11
showed
aa gradual
increase
with
the
increased
ballast
degradation.
The
P2 forces
plotted
with
each
blow
is shown
plotted
with
blow
respect
toin
The
P 2 forces
is shown
shown
inballast
Figuredegradation.
11
showed
gradual
increase
with
theeach
increased
number
As
ballast
get
rearranged
and
become
in
Figureof
showed
gradual
increase
the
increased
number
of
is
shown
inblows.
Figure
11athe
showed
a particle
gradualwith
increase
with the
increased
number
of11
blows.
As
the
ballast
particle
get
rearranged
and
become
aanumber
densely
packed
after
each
blow,
which
offer
aa higher
inertial
blows.
As
ballast
particle
get
rearranged
and become
densely
of the
blows.
As
the
ballast
particle
get rearranged
anda become
densely
packed
after
each
blow,
which
offer
higher
inertial
values.
When
ballast
resistance,
increased
2 force
packed
afterleads
eachto
which
a higher
leads
aresistance,
densely
packed
after
eachP
blow,
offer
aresistance,
higher
inertial
values.inertial
When
ballast particle
particle
leads
toblow,
increased
Poffer
2 force which
forces
become
rearrange
and
stabilise
completely,
the
changes
of
to
increased
P2
force
values.
particle
rearrange
and
force
values.
When
particle
resistance,
leads
to increased
P2When
forces
become
rearrange
and
stabilise
completely,
theballast
changes
of P
P 22 ballast
insignificant.
is
apparent
from
Figure
11,
the
changes
of
P
forces
become
rearrangecompletely,
and This
stabilise
completely,
the
changes
of
P
stabilise
changes
of
P2
forces
become
insignificant.
insignificant.
This
isthe
apparent
from
Figure
11,
the
changes
of
P22
2
th blow.from Figure 11, the changes of P
forces
minor
after
88th
insignificant.
This
is Figure
apparent
This
isvery
apparent
from
11, the changes of P2 forces very minor2
blow.
forces
very
minor
after
By
comparing
the
forces
very
minor after
8th blow.forces
By
comparing
the impact
impact
forces with
with and
and without
without shock
shock mats,
mats,
after
8th
blow.
the
results
shows
that
the
shock
mats
attenuated
the
impact
By
comparing
the
impact
forces
with
and
without
shock forces
mats,
the results shows that the shock mats attenuated the impact
forces
By
comparing
the that
impact
with
and
without
shock
mats,
the
for
base
conditions.
is
evident
from
the
results
shown
in
the
results
shows
theIt
shock
attenuated
impact
forces
for both
both
base
conditions.
Itforces
is also
alsomats
evident
from
thethe
results
shown
in
results
shows
that
the
shock
mats
attenuated
the
impact
forces
for
Figure
11
for
the
weak
base
without
shock
mat
and
hard
base
with
for both11base
It is without
also evident
shown
in
Figure
for conditions.
the weak base
shockfrom
mat the
andresults
hard base
with
shock
mat,
the
weak
base
acted
as
shock
absorbing
material.
both
conditions.
Itbase
isitself
also
evident
results
shown
in
Figurebase
11 for
weak
without
shock
matthe
and
hard base
with
shock
mat,
thethe
weak
base
itself
acted
as aafrom
shock
absorbing
material.
Therefore,
the
impact
forces
were
more
hard
base.
Figure
11 for
weak
base
without
shock
matfor
and
hard
base
with
shock
mat,
thethe
weak
base
itself
acted
asdistinct
a shock
absorbing
material.
Therefore,
the
impact
forces
were
more
distinct
for
hard
base.
shock
mat, the
the impact
weak base
itself
acted
asdistinct
a shockfor
absorbing
material.
Therefore,
forces
were
more
hard base.
Therefore, the impact forces were more distinct for hard base.

10
10
10
9
9
9
8
8
8
7
7
7
6
6
6
5
5
5
4
4
4
3
3
3
2
2
2
1
1
1
0
0
0
00
0

Hard
Hard base
base
Weak
base
Weak
base
Hard base
Weak base

With
With Shock
Shock mat
mat
Without
Shock
With Shock
matmat
Without
Shock
mat
Without Shock mat

1
1
1

2
2
2

3
4
5
6
7
3
4
5
6
7
of
Blows,
N
3 Number
4
5
6
Number of Blows, N7

Number of Blows, N

8
8
8

9
9
9

10
10
10

(b)
(b)
(b)Figure 12 Permanent strain response of ballast with and without
Figure 12 Permanent strain response of ballast with and without
shock
mat:(a)strain
shearresponse
strain; (b)
(b)ofvolumetric
volumetric
strain
Figure 12
Permanent
ballast withstrain
and without
shock
mat:(a)
shear
strain;
(data
sourced
from
Nimbalkar
et al.
al. 2012)
2012)
shock
mat:(a)
shear
strain;
(b) volumetric
strain
(data
sourced
from
Nimbalkar
et
(data sourced from Nimbalkar et al. 2012)
The inclusion of shock mats in the ballast bed reduced the shear and
The inclusion
inclusion
of shock
shock
mats in
in
the ballast
ballast
bed reduced
reduced
the shear
shear
volumetric
strain of
the ballast
layer.
The permanent
strains
were
The
mats
the
bed
the
and The
volumetric
strain
of hard
the
ballast
layer.
The
permanent
strains
more
pronounced
for
the
base
condition.
However
when
shock
inclusionstrain
of shock
mats
in the
ballastThe
bed
reduced
thestrains
shear
and
volumetric
of
the
ballast
layer.
permanent
werevolumetric
more
pronounced
forthe
theballast
hard base
base
condition.
However
when
and
strain
of
layer.
permanent
mats
are
placed
at
the top
and
bottom
of condition.
theThe
ballast
layer thestrains
shear
were
more
pronounced
for
the
hard
However
when
shock
mats
are
placed
at
the
top
and
bottom
of
the
ballast
layer
the
were
more
pronounced
for
the
hard
base
condition.
However
when
and
the
volumetric
strains
are
reduced
in
the
order
40%
to
50%.
shock mats are placed at the top and bottom of the ballast layerThe
the
shear and
and
the
volumetric
strains
are
reduced
inof
the
order
40%
to
shock
mats
arevolumetric
placed
the
toptoand
ofin
thethe
ballast
ballast
breakage
can beat related
thebottom
number
blows
aslayer
wellthe
as
shear
the
strains
are
reduced
order
40%
to
50%.
The
ballast
breakage
can
be
related
to
the
number
of
blows
as
shear The
and ballast
theimpulse
volumetric
are
reduced
innumber
theloading
order
40%
to
50%.
breakage
can bethe
related
to the
of blows
as
accumulated
(area strains
under
transient
impact
curve).
well
as
accumulated
impulse
(area
under
the
transient
impact
50%.
Thetoaccumulated
ballast
breakage
canofbe(area
related
thepaper,
number
of blows
as
well
as
under
the
transient
impact
In
order
abbreviate
inimpulse
view
scope
of to
this
Figures
11 and
loading
curve).
In order
order
to abbreviate
abbreviate
inunder
view of
of
scope
of this
thisimpact
paper,
well
accumulated
impulse
(areain
thescope
transient
loading
curve).
In
to
view
of
paper,
12
areasplotted
against
number
of blows.
Figures
11
and
12
are
plotted
against
number
of
blows.
loading
curve).
In
order
to
abbreviate
in
view
of
scope
of
this
paper,
Figures 11 and 12 are plotted against number of blows.
Figures 11 and 12 are plotted against number of blows.

10
10
10

variation
number
of
Figure
ofwith
Blows,
N
forceNumber
variation
with
number
of blows
blows
Figure 11P
11P22 force
(data
from
et
2012)
with number
blows
Figure
11Psourced
(data
sourced
from Nimbalkar
Nimbalkar
et al.
al. of
2012)
2 force variation
(data sourced from Nimbalkar et al. 2012)
4.6
4.6 Ballast
Ballast deformation
deformation and
and strain
strain response
response
4.6
Ballast
deformation
and
strain
response
Vertical
and
lateral deformation
deformation
data
were collected
collected after
after each
4.6
Ballast
deformation
and straindata
response
Vertical
and
lateral
were
each
and
volumetric
strain
(
for
impact
shear
strain
Vertical
and The
lateral
deformation
collected
and were
volumetric
strainafter
(vv)) each
for
impact blow.
blow.
The
shear
strain (
(qq))data
Vertical
and lateral
deformation
data
were
collected
after
each
impact
axisymmetric
loading
were
calculated
by
using
the
following
)
and
volumetric
strain
(
)
for
impact
blow.
The
shear
strain
(
axisymmetric loading were calculated
by using the following
q
v
blow.
Theby
shearloading
strain (qwere
) and volumetric
strain
(v) forthe
axisymmetric
equation
(1951).
axisymmetric
calculated
by using
following
equation
by Timoshenko
Timoshenko
and Goodier
Goodier
(1951).
loading
wereTimoshenko
calculatedand
byGoodier
using (1951).
the following equation by
equation by
Timoshenko and Goodier (1951).
37
37
37

Geotechnical Journal Vol. 6 No. 1 2014


38 Geotechnical Journal Vol. 6 No. 1 2014
Geotechnical Journal Vol. 6 No. 1 2014
4.7 Ballast Breakage under Impact Loading
4.7 Ballast Breakage under Impact Loading
BallastBallast
particleBreakage
breakage takes
under
repetitive impact loading.
4.7
underplace
Impact
Loading
Ballast particle breakage takes place under repetitive impact loading.
Initially, breakage of corners of the angular ballast at the interInitially,particle
breakage of corners
of theunder
angular
ballast at theloading.
interBallast
takesfollowed
place
particle contactsbreakage
takes place,
by repetitive
complete impact
fracture of the
particle contacts
place, followed
by complete
the
Initially,
breakagetakes
of corners
of the angular
ballastfracture
at the of
interparticles depends on the strength of the raw ballast and level of the
particles
depends takes
on theplace,
strength
of the by
rawcomplete
ballast and
level of the
particle contacts
followed
fracture
load increase.
increase. This
This affects the
the overall deformation
deformation characteristics
characteristics
load
particles
depends on affects
the strengthoverall
of the raw ballast and
level of the
and ultimate
ultimate strength
strength of
of the
the ballast
ballast layer(Selig
layer(Selig and
and Waters
Waters 1994;
1994;
and
load increase. This affects the overall deformation characteristics
Indraratna
et al.
al.
2011).This
of ballast
ballast particles
particles
contributes
Indraratna
et
2011).This
breakage
of
contributes
and
ultimate
strength
of thebreakage
ballast layer(Selig
and Waters
1994;
to increased
increased
vertical
and lateral
lateral
deformations
and
differential
track
to
and
deformations
differential
track
Indraratna
et vertical
al. 2011).This
breakage
of ballastand
particles
contributes
settlement.
quantify
the
particle
breakage
under
loading,
an
settlement.
To
quantify
the
particle
breakage
under
impact
loading,
to
increasedTo
vertical
and
lateral
deformations
andimpact
differential
track
evaluation
of
breakage
was performed.
After After
10
impact
blow,
an
evaluation
of
ballast
breakage
was
performed.
10loading,
impact
settlement.
Toballast
quantify
the
particle
breakage
under
impact
the evaluation
ballast
fromofthe
specimen
was recovered
and particle
blow,
the ballast
from
thebreakage
specimen
wasperformed.
recovered
and size
particle
size
an
ballast
was
After
10analysis
impact
was performed
tofrom
compare
the
degraded
withand
the particle
fresh
ballast
analysis
performed
compare
theballast
degraded
ballast
with
the
blow,
thewas
ballast
thetospecimen
was
recovered
size
initially
used
in the testing.
Tothequantify
the
breakage,
the
fresh
ballast
used
testing.
To ballast
quantify
thewith
ballast
analysis
was initially
performed
to in
compare
the degraded
ballast
following
equation
from
the
proposed
for
Ballast
Breakage
breakage,
theinitially
following
equation
from
theTomethod
proposed
for
fresh
ballast
used
inmethod
the testing.
quantify
the
ballast
Ballast
Breakage
Index (BBI)
by(2005)
Indraratna
et method
al. (2005)proposed
was used.for
Index (BBI)
Indraratna
et al.
wastheused.
breakage,
thebyfollowing
equation
from
Ballast Breakage Index (BBI) by Indraratna et al. (2005) was used.

5. NUMERICAL MODELLING
5.
NUMERICAL MODELLING
The dynamic
responseMODELLING
of this layered system attributed to transient
5.
NUMERICAL
The dynamic response of this layered system attributed to transient
impact load is analyzed by a 2-dimentional (2D) axisymmetric
impact
load is
analyzed this
by layered
a 2-dimentional
(2D) axisymmetric
The
dynamic
systemPLAXIS(PLAXIS
attributed
to transient
dynamic
finiteresponse
elementofanalysis
by using
2D:
dynamic
finiteiselement
analysis
using PLAXIS(PLAXIS
2D: Ver.
impact load
analyzed
by a by2-dimentional
(2D) axisymmetric
Ver. 8.6).The main features of this dynamic finite element analysis
8.6).The
main element
featuresanalysis
of thisbydynamic
finite element 2D:
analysis
dynamic finite
using PLAXIS(PLAXIS
Ver.
includes, introduction
introduction ofofmodified
modifiedstress-dilatancy
stress-dilatancy
relationship
includes,
relationship
to
8.6).The main
features of this dynamic
finite element
analysis
to capture
the ballast
particle
degradation
and incorporation
of
capture
theintroduction
ballast
particle
degradation
and incorporation
of material
includes,
of modified
stress-dilatancy
relationship
to
materialthe
damping
for track
various
track materials
tested.
Theofspecimen
damping
forballast
various
materials
tested.
The
specimen
of this
capture
particle
degradation
and incorporation
material
of thiswas
study
was
modeled
asmaterials
an elasto-plastic
model
of a composite
study
modeled
as track
an elasto-plastic
model of
a composite
layered
damping
for
various
tested.
The
specimen
of this
layered
system
including
ballast,mat,base
shockmodel
mat,base
and
steelAplate.
A
system
including
ballast,
shock
andofsteel
plate.
typical
study
was
modeled
as an elasto-plastic
a composite
layered
typical including
axisymmetric
specimen
model
simulated
finiteAelement
element
axisymmetric
specimen
modelmat,base
simulated
in inplate.
finite
system
ballast,
shock
and steel
typical
All 3element
layers
discretization
PLAXISmodel
2D is shown
in Figure
axisymmetric using
specimen
simulated
in 14.
finite
interaction
are
modeled using
elements
and14.
theAll
discretization
using15-node
PLAXIScubic
2D isstrain
shown
in Figure
3 layers
between
granular
and
shock
mats
are are
modelled
using
5-node
granular
media
andthe
the
shock
mats
modelled
using
5are
modeled
usingmedia
15-node
cubic
strain
elements
and
the
interaction
interface
elements.
The and
15-point
cubic mats
element
provides
a fourth
node
interface
elements.
Thethe15-point
cubicareelement
provides
between
granular
media
shock
modelled
using
5-a
fourth
order interpolation
displacements.
The
numerical
order interpolation
for displacements.
Thecubic
numerical
integration
bya
node
interface
elements.
Thefor
15-point
element
provides
integration
byscheme
the
Gaussian
scheme
involves
12 Gauss
points.
the Gaussian
involves
12
points.
fourth
order
interpolation
for Gauss
displacements.
The
numerical
The digitally
filtered (by
usinginvolves
a low-pass
Butterworth
integration
by the Gaussian
scheme
12 Gauss
points. filter)
The digitally
filtered
(by using
a low-pass
Butterworth filter)
transient
transient
impact
load-time
histories
the laboratory
The digitally
filtered (by
using aobtained
low-passfrom
Butterworth
filter)
impact load-time
histories
obtained
from
the
laboratory
are
testings
are
used load-time
for the dynamic
element
Lateral
transient
impact
historiesfinite
obtained
fromanalysis.
thetestings
laboratory
used
for
the
dynamic
finite
element
analysis.
Lateral
distributed
loads
distributed
to the finite
right boundary
to represent
the
testings
areloads
used are
for applied
the dynamic
element analysis.
Lateral
are appliedeffects
to theare
right
boundary
represent
the confining
effects
confining
of
thick
rubber
(Henkel
and Gilbert
distributed
loads
applied
to thetomembrane
right
boundary
to represent
the
of thick The
rubber
membrane
(Henkel
and
Gilbert are
1952).
The
following
1952).
following
boundary
adopted
for
the
confining
effects
of thick
rubber conditions
membrane
(Henkel
and
Gilbert
boundary
conditions
are
adopted
forconditions
theofnumerical
left
numerical
analysis.
The
left (axis
symmetry)
and The
bottom
1952).
The
following
boundary
are analysis.
adopted
for
the
(axis of symmetry)
and
boundaries
are vertical
restrained
inbottom
lateral
boundaries
are restrained
lateralof and
directions,
numerical
analysis.
Thebottom
leftin (axis
symmetry)
and
and vertical are
directions,
respectively.
The and
topare
and
right
respectively.
The restrained
top and
right
free
to boundaries
move.The
boundaries
in boundaries
lateral
vertical
directions,
are free
at
theof
left
corner
mesh
is
node
at to
themove.The
left
thebottom
meshare
is restrained
in both
respectively.
Thebottom
top node
andcorner
right
boundaries
freeoftothe
move.The
vertical
and
horizontal
directions
- standard
restrained
in left
both
verticalcorner
and horizontal
directions
(pinned
support
node
at the
bottom
of the(pinned
mesh
issupport
restrained
in
both
fixity).The
right
and bottom
boundaries
are considered
adsorbent
- standardand
fixity).The
rightdirections
and
bottom
boundaries
are -considered
vertical
horizontal
(pinned
support
standard
boundaries.
Two and
different
models
adopted
are (1)
fixity).The
right
bottom
boundaries
are been
considered
adsorbent boundaries.
Two soil
different
soilhave
models
have
beenadsorbent
adopted
classical
Mohr-Coulomb
modelmodel
for
boundaries.
Two Mohr-Coulomb
different elastic-perfectly-plastic
soil models
have been adopted
are the
(1)
are (1) classical
elastic-perfectly-plastic
for
basematerial
andand
(2)(2)
isotropic
Hardening
al.
classical
Mohr-Coulomb
elastic-perfectly-plastic
model foret
the basematerial
isotropic
HardeningSoil
Soilmodel(Schanz
model(Schanz
etthe
al.
1999)
for ballast.
ballast.
Theconstitutive
constitutive
model
parameters
adopted
basematerial
and The
(2)
isotropic
Hardening
Soil
model(Schanz
ethere
al.
1999) for
model
parameters
adopted
here
are
are
based
on
the
available
laboratory
test
results.
1999)
forthe
ballast.
Thelaboratory
constitutive
parameters adopted here
based
on
available
testmodel
results.
are based on the available laboratory test results.
Impact Load
Impact Load

(4)
(4)
The parameters defining the BBI are shown in Figure 13. The BBI
for both
hard anddefining
weak base
with and
without13.
shock
The
parameters
the condition
BBI are shown
in Figure
The mats
BBI
are both
summarized
Table
The values
in parentheses
in
for
hard andinweak
base3.condition
withshown
and without
shock mats
Table
3 are the percentage
of BBI
by in
theparentheses
use of shock
are summarized
in Table 3.reduction
The values
shown
in
mats
the bottomreduction
of the ballast
layer.
Tableat3the
aretop
theand
percentage
of BBI
by the use of shock
mats at the top and the bottom of the ballast layer.

Top Plate
Top Plate
Shock Mat
Shock Mat

30 mm
30 mm

Membrane Stress
Membrane Stress

Figure 13 Determination of Ballast Breakage Index (BBI) (after


Indraratna
et al.Breakage
2005) Index (BBI) (after
Figure 13 Determination
of Ballast
Indraratna et al. 2005)
Table 3Ballast Breakage after 10 impact blows
Table 3Ballast Breakage
after 10 impact
blows
Ballast Breakage
Index (BBI)
Base
Ballast
Breakage
Index
(BBI)
Without
Shock
With
Shock
Mats
(placed at
Condition
Base
Mats
top
and
bottom
of
ballast)at
Without
Shock
With
Shock
Mats
(placed
Condition
Mats
top and
bottom by
of ballast)
Hard
0.170
0.091
(reduced
46.5%)
Hard
0.170
0.091
(reduced
by
46.5%)
Weak
0.080
0.028 (reduced by 65.0%)
Weak

0.080

Ballast
Ballast

300 mm
300 mm

Absorbent boundaries
Absorbent boundaries

Shock Mat
Shock Mat

30 mm
30 mm

0.028 (reduced by 65.0%)

Base
Base

100 mm
100 mm

It isItevident
from
thethe
BBI
values
is evident
from
BBI
valuesreported
reportedininTable
Table33that
that the
the use
of shock
mats considerably
breakage
of ballast
It is evident
from the BBIreduced
values the
reported
in Table
3 that particle
the use
under
impact
The reduced
hard base
of shock
mats loadings.
considerably
thecondition
breakage induced
of ballastrelatively
particle
higher
breakage
than theThe
weaker
to the
under impact
loadings.
hard base
base condition.
condition induced
relatively
This is due
concentration
of than
non-uniform
stresses
at the
of
higher breakage
the weaker
base developing
condition. This
is corners
due to the
the
ballast
by
higher
hard
concentration
of non-uniform
stresses
developing
at the corners
of
the sharp
sharp angular
angular
ballast increasing
increasing
by the
the
higher resistive
resistive
hard base.
base.
When
shock
mats
placed
at
and
ballast
layer,
the
the
sharp
angular
thebottom
higherof
resistive
When
shock
matsballast
placedincreasing
at the
the top
topby
and
bottom
of
ballasthard
layer,base.
the
ballast
breakage
was
reduced
by and
about
46.5%
for
When
mats
placed
at the
top
bottom
of ballast
layer,base
the
ballast shock
breakage
was
reduced
by about
46.5%
for
hard
basehard
condition.
condition.
The
same
reduction
for
the
weak
base
condition
was
ballast
breakage
was
reduced
by
about
46.5%
for
hard
base
The same reduction for the weak base condition was relatively higher
relatively
and about
This
is due
the weak
base
itself
condition.
The same
reduction
for
the
weak
base
condition
was
and abouthigher
65%.
This
is
due65%.
to the
weak
baseto itself
act
as
a shock
act
as a shock
absorbing
layer.
relatively
higher
and about
65%. This is due to the weak base itself
absorbing
layer.
act as a shock absorbing layer.

300 mm
300 mm

Bottom Plate
Bottom Plate

CL
Figure
CL 14Finite Element Mesh for the typical test specimen
Figure 14Finite Element Mesh for the typical test specimen

38
38

rd
Journal
3Geotechnical
Proff 18-02-2015
Geotechnical
Journal Vol.
Vol. 66

5.1
5.1

No.
No. 11 2014
2014

Geotechnical Journal Vol. 6 No. 1 2014 39

Mohr-Coulomb Elasto-Plastic
Elasto-Plastic model
model
Mohr-Coulomb

Steel
E
Steel ::

E
Shock
Mat

:
E
Shock Mat

:E

The Mohr-Coulomb
Mohr-Coulomb (MC)
(MC) model
model is
is used
used to
to represent
represent the
the weak
weak base.
base.
The
The
The following
following key
key parameters
parameters and
and values
values were
were used
used to
to represent
represent aa
relatively
relatively weak
weak base
base (i.e.,
(i.e., poorly
poorly graded
graded sand).
sand).

The
The hardening
hardening soil
soil (HS)
(HS) model
model is
is used
used to
to simulate
simulate the
the strainstrainhardening
hardening behaviour
behaviour of
of ballast
ballast under
under impact
impact loading.The
loading.The mobilised
mobilised
friction
mmis
is defined
defined as
as follows:
follows:
friction angle
angle

q
q

q
q2
2
33

(5)
(5)

mmis
is given
given by(Nimbalkar
by(Nimbalkar et
et al.
al. 2012):
2012):
The mobilised
mobilised dilatancy
dilatancy angle
angle
The
sin
dBBI
sin
mm ))
dBBI ((11

sin

sin
pp
sin
mm
sin
cvcv
2 d 1 tan 22

30
30


cv
2 33 d 11 1 tan cv

sin
sin
mm

sin
dBBI
dBBI ((11
sin
mm ))

11 sin

pp
sin
mm sin
sin
cvcv
2 d 1 tan 22 cv
2 33 d 11 1 tan cv
(6)
(6)

Laboratory
Laboratory Data
Data
With
With Shock
Shock mat
mat
Without
Without Shock
Shock mat
mat

Vertical
VerticalStrain,
Strain,1 1(%)
(%)

25
25

The symbols
symbols are
are explained
explained in
in the
the notation
notation section
section of
of this
this paper.
paper.
The
Further
Further details
details of
of the
the HS
HS material
material parameters
parameters and
and breakage
breakage
parameters are
are given
given in
in Table
Table 4.
4.
parameters
Table 4Ballast
4Ballast Parameters
Parameters for
for HS
HS Model
Model Simulation
Simulation
Table
Hard Base
Base
Weak Base
Base
Hard
Weak
Material
Material
Sample
Sample
Sample
Parameters
Sample Sample Sample Sample
Sample
Parameters
11
22
33
44

FE
FE Predictions
Predictions
Hard
Hard base
base
Weak
Weak base
base

20
20
15
15
10
10
5
5
0
0

ref
ref
E
E50
50

11.04
11.04

13.12
13.12

12.43
12.43

15.10
15.10

ref
ref
Eoed
E
oed

11.04
11.04

13.12
13.12

12.43
12.43

15.10
15.10

Figure
Figure 15
15 Axial
Axial Strain:
Strain: Measured
Measured vs
vs FE
FE predicted
predicted values
values
(data
(data sourced
sourced from
from Nimbalkar
Nimbalkar et
et al.
al. 2012)
2012)

ref
ref
Eur
E
ur

10.20
10.20

12.09
12.09

12.53
12.53

14.80
14.80

pp

73.34
73.34

73.60
73.60

74.81
74.81

75.83
75.83

21.27
21.27

16.15
16.15

18.20
18.20

14.58
14.58

Pref
P
ref

19.70
19.70

12.67
12.67

10.65
10.65

6.06
6.06

0.81
0.81

0.68
0.68

0.73
0.73

0.47
0.47

882.44
882.44

728.54
728.54

664.45
664.45

674.72
674.72

714.78
714.78

495.41
495.41

485.05
485.05

317.12
317.12

6. CONCLUSION
6.
CONCLUSION
6.
CONCLUSION
The performance of ballasted track with shock mats has been
The
performance
of
track
mats
has
The performance
of ballasted
ballasted
track with
with
shock
matsmodels.
has been
been
described
through laboratory
experiments
and shock
numerical
The
described
through
laboratory
experiments
and
numerical
models.
described
through
laboratory
experiments
and
numerical
models.
impact load causes accelerated ballast breakage was confirmed
by
The
load
causes
ballast
breakage was
The impact
impact and
load numerical
causes accelerated
accelerated
ballast
was confirmed
confirmed
experiment
model data.
Twobreakage
base conditions
tested
by
experiment
and
model
data. Two
conditions
by this
experiment
and numerical
numerical
model
Two base
base
in
study confirm
that the hard
basedata.
conditions
such conditions
as bridge
tested
in
this
study
confirm
that
the
hard
base
conditions
such as
tested
in
this
study
confirm
that
the
hard
base
conditions
as
deck, rail track-road crossing and track on rock foundationsuch
cause
bridge
deck,
rail
track-road
crossing
and
track
on
rock
foundation
bridge
deck,
rail
track-road
crossing
and
track
on
rock
foundation
comparatively
higher higher
ballastballast
degradation
compare
to weak
base
cause
degradation
compare
to
cause comparatively
comparatively
ballast
degradation
compare
tois weak
weak
condition.
Initially,
thehigher
impact
induced
strainstrain
of the
ballast
very
base
condition.
Initially,
the
impact
induced
of
the
ballast
base condition. Initially, the impact induced strain of the ballast is
is
high
and itand
eventually
stabilizes and become
constant after certain
very
very high
high and it
it eventually
eventually stabilizes
stabilizes and
and become
become constant
constant after
after
number
of load application.
certain
certain number
number of
of load
load application.
application.
The
insertion
of
mats
at
of
TheThe
insertion
of shock
shock
mats mats
at the
theattop
top
and
bottom
of the
the ballast
ballast
insertion
of shock
theand
topbottom
and bottom
of the
reduces
the
induced
stresses
on
ballast
reducesreduces
the impact
impact
induced
stresses
on on
ballast
and
considerably
ballast
the impact
induced
stresses
ballastand
and considerably
reduces
As
the
hard
base
condition
produces
reduces the
the ballast
ballast
degradation.
As the
the hard
hard base
base condition
condition produces
produces
reduces
the
ballast degradation.
degradation. As
more
breakage,
the
benefits
of
shock
mats
are
greater
in
hard
base
more
breakage,
the
benefits
of
shock
mats
are
greater
in
hard
base
more breakage, the benefits of shock mats are greater in hard base
conditions
compared
to
weak
bases.
Weak
base
itself
act
as
aa shock
conditions
compared
to
weak
bases.
Weak
base
itself
act
as
shock
conditions compared to weak bases. Weak base itself act as a shock
absorbing
layer,
therefore
the
use of
shock
mats are
not
absorbing layer,
layer,therefore
thereforethe
the
of additional
additional
shock
aremore
not
absorbing
useuse
of additional
shock
matsmats
are not
more
pronounced
for
softer
foundations.
The
finite
element
model
more
pronounced
for
softer
foundations.
The
finite
element
model
pronounced
for softer of
foundations.
The
finiteresponses
element model
analysis
analysis
is
predicting
strain
measured
for
analysis
isofcapable
capable
of strain
predicting
strainmeasured
responses
measured
for
is
capable
predicting
responses
for
ballast
under
ballast
under
impact
loading
with
and
without
shock
mats.
It
is
ballast
under
impact
loading
with
and
without
shock
mats.
It
is
impact
loading
with
and
without
shock
mats.
It
is
evident
from
this
evident
evident from
from this
this study,
study, by
by placing
placing shock
shock mats,
mats, loads
loads on
on the
the ballast
ballast
study,
by placing
shockamats,
loads on the ballast
bed canthe
be reduced
bed
bed can
can be
be reduced
reduced by
by a more
more homogenous
homogenous mounting
mounting of
of the sleepers
sleepers
by
a more
homogenous
mounting
of
thebe
sleepers
and ballast
and track
and
ballast
and
track
stability
can
improved.
This
and ballast and track stability can be improved. This leads
leads to
to
stability
can be misalignment,
improved. This leads toinreduced
track misalignment,
reduced
reduced track
track misalignment, which
which in turn
turn leads
leads to
to aa reduced
reduced
which
leads to a reduced
number of maintenance operations.
number
maintenance
operations.
numberinof
ofturn
maintenance
operations.
P
Results
could
vary
for
different
PSD
and
impact
force
P
and
P22..
Results
could
varyvary
for different
PSDPSD
andand
impact
force
P11 and
Results
could
for different
impact
force
P1 and
Also in
in reality,
reality, material
material used
used for
for USP
USP and
and UBM
UBM can
can vary,
vary, usually
usually
Also
P2. Also in reality, material used for USP and UBM can vary, usually
stiffer
mats
preferred
sleeper.
No
study
has
yet
been
reported
stiffer mats
mats preferred
preferred under
under
sleeper. No
No study
study has
has yet
yet been
been reported
reported
stiffer
under sleeper.

(MPa)
(MPa)
(MPa)
(MPa)
(MPa)
(MPa)
(degrees)
(degrees)

(degrees)
(degrees)
2
(kN/m
(kN/m2))

dBBI
dBBI

d11pp
d

dE
dE

B
B

p
d
d11p

(kNm/m33)
(kNm/m )

5.3
5.3

3
77
77 kN
kN m
m3
3
12.04
12.04 kN
kN m
m3

Figure
15
shows
the
element
prediction
of
axial
using
data obtained
the laboratory
impact
testing.
Figurethe
15impact
showspulse
the finite
finite
elementinmodel
model
prediction
of the
the
axial
strain
using
the
impact
pulse
data
obtained
in
the
laboratory
impact
The
axial
strain
values
are
compared
with
laboratory
measured
data
strain using the impact pulse data obtained in the laboratory impact
testing.
The without
axial strain
strain
values are
are
compared
with
laboratory
for
with and
the placement
of shock
mats for
bothlaboratory
hard and
testing.
The
axial
values
compared
with
measured
data
without
placement
shock
for
weak
baseconditions.
from
Figurethe
the finiteof
analysis
measured
data for
for with
withAsand
and
without
the15,
placement
ofelement
shock mats
mats
for
both hard
hard
and weak
weak
baseconditions.
Asbehaviour
from Figure
Figure
15, the
the under
finite
able
to predict
the baseconditions.
strain hardeningAs
of ballast
both
and
from
15,
finite
element
able
predict
the
of
element analysis
analysis
able to
to
predict
the strain
strain hardening
hardening
behaviour
of
repeated
impact load.
The
FE simulation
is closely behaviour
captured the
ballast
under
repeated
impact
load.
The
FE
simulation
is
closely
ballast yielding
under repeated
impactwhich
load.influenced
The FE simulation
is vicious
closely
plastic
of the ballast
by amount of
captured of
thetheplastic
plastic
yielding
ofThe
theclose
ballast
which influenced
influenced
by
captured
the
the
ballast
which
by
damping
ballastyielding
material.of
comparison
of FE model
amount
damping
of
material.
The
close
amount of
ofandvicious
vicious
damping
of the
the
ballast
material.
Thethat
close
predicted
laboratory
measured
axialballast
strain values
reveal
the
comparison
of FE
FE
model
predicted
and
laboratory
measured
axial
comparison
of
model
predicted
laboratory
measured
axial
influence
of P1
forces
on the
responseand
of the
ballast is
negligible,
as
on
the
response
of
strain
values
reveal
that
the
influence
of
P
1 forces
forces
on
the
response
of
strain
values
reveal
that
the
influence
of
P
1
the digitally filtered P2 force load-time history was used as an input
force load-time
load-time
the ballast
ballast is
is negligible,
negligible, as
as the
the digitally
digitally filtered
filtered P
P 22 force
the
for the finite element analysis.
history was
was used
used as
as an
an input
input for
for the
the finite
finite element
element analysis.
analysis.
history

Hardening Soil
Soil Model
Model
Hardening

sin

sin
mm

0.15,

0.15,

0.48,

0.48,

5.4 Finite Element Model Predictions


5.4
Element
Model
Predictions
5.4 Finite
Finite
Element
Model
Predictions
Figure
15 shows
the finite
element
model prediction of the axial strain

o
E
,,
cc 0,

E 45
45 MPa
MPa
0.33,
0.33,
0,
24
24o and
and
0.
0.

5.2
5.2

210

,,
210 GPa
GPa

6.12
MPa

,
6.12 MPa

ff

ff

Linear
Linear Elastic
Elastic Model
Model and
and Interface
Interface Elements
Elements

Steel
Steel plates
plates located
located at
at the
the top
top and
and bottom
bottom of
of the
the test
test sample
sample are
are
considered
as
linear
elastic.
The
shock
mat
is
also
modelled
considered as linear elastic. The shock mat is also modelled as
as aa
linear
linear elastic
elastic material.
material. Zero-thickness
Zero-thickness interface
interface elements
elements are
are used
used to
to
model
model the
the frictional
frictional behaviour
behaviour between
between various
various layers
layers and
and are
are
simulated
simulated by
by 5-nodeline
5-nodeline elements.
elements. The
The following
following material
material
parameters
parameters were
were used
used for
for Steel
Steel and
and shock
shock mat.
mat.

39
39

0
0

2
2

4
6
4
6
Number
of
blows,
N
Number of blows, N

8
8

10
10

Geotechnical
Journal
Vol. 6Vol.
No.61 No.
20141 2014
40
Geotechnical
Journal
9. REFERENCES

on quantitative or qualitative analysis of ballast degradation by


on quantitative
qualitative
analysis
of ballast
degradation
placing
USP and or
UBM
under cycling
loading
condition.
Currently by
an
placing USP is
andundertaken
UBM under
cycling
loadingof condition.
Currently
investigation
at the
University
Wollongong
testing
an investigation
undertaken
at the
University
of in
Wollongong
facility
to evaluateisthe
effectiveness
of USP
and UBM
mitigating
testing
facility
to
evaluate
the
effectiveness
of
USP
and
UBM in
ballast degradation.
mitigating ballast degradation.
7. ACKNOWLEDGMENTS
7. authors
ACKNOWLEDGMENTS
The
wish to thank the Australian Research Council for its

Anastasopoulos, I., Alfi, S., Gazetas, G., Bruni, S. and Van Leuven,
9. (2009)
REFERENCES
A.
Numerical and Experimental Assessment of Advanced
Concepts
to Reduce
andGazetas,
Vibration
UrbanS.
Railway
Turnouts.
Anastasopoulos,
I., Noise
Alfi, S.,
G.,onBruni,
and Van
Leuven,
Journal
of
Transportation
Engineering,
135,
(5),
pp.
279-287.
A. (2009) "Numerical and Experimental Assessment of Advanced

Concepts
to Reduce
and for
Vibration
on Urban
Railway
ARA
(2003),
Code ofNoise
Practice
the Defined
Interstate
Rail
Turnouts."Volume
Journal4:
of Track,
Transportation
135, (5), pp. 279Network,
Civil andEngineering,
Electrical Infrastructure,
Part
287.
1:
Identification and Classification of Wheel defects, Australasian
Railway Association.
ARA (2003), Code of Practice for the Defined Interstate Rail
Auersch,
(2006)4:Dynamic
axleand
loads
on tracksInfrastructure,
with and without
Network, L.
Volume
Track, Civil
Electrical
Part
1: Identification
and Classification
of Wheel defects,vehicle-trackAustralasian
ballast
mats: numerical
results of three-dimensional
Railway
Association.
soil
models.
Proceedings of the Institution of Mechanical Engineers
Part F-Journal of Rail and Rapid Transit, 220, (2), pp. 169-183.
Auersch, L. (2006) "Dynamic axle loads on tracks with and without
Bolmsvik,
R. numerical
(2005) Influence
USP on the trackvehicle-trackresponsea
ballast mats:
results ofofthree-dimensional
literature
survey.
soil models."
Proceedings of the Institution of Mechanical Engineers
Part F-Journal
of RailJ.and
220, (2),
pp. 169-183.
Brown,
S. F., Kwan,
andRapid
Thom,Transit,
N. H. (2007)
Identifying
the key

financial
support.
support
by University
of Peradeniya,
The authors
wishThe
to thank
theprovided
Australian
Research Council
for its
Sri
Lanka support.
offering the
leave provided
to conductbydoctoral
work of
financial
Thestudy
support
University
Sinniah
K. Navaratnarajah
is gratefully
acknowledged.
The assistance
Peradeniya,
Sri Lanka offering
the study
leave to conduct
doctoral
work of Sinniah
K. technical
Navaratnarajah
is Alan
gratefully
The
provided
by senior
officers,
Grant,acknowledged.
Cameron Neilson
assistance
provided
bymuch
seniorappreciated.A
technical officers,
Alan
Grant,
and
Ian Bridge
is also
significant
portion
of
Cameron
Neilson
and reproduced
Ian Bridgewith
is also
appreciated.A
the
contents
have been
kind much
permission
from the
significant
portion of theand
contents
have been reproduced
with
kind
Journal
of Geotechnical
Geoenvironmental
Engineering
ASCE,
permission from
the Journal
of Geotechnical
and
Geoenvironmental
International
Journal
of Geomechanics,
ASCE,
ASTM
Geotechnical
Engineering
ASCE,
International
of Geomechanics,
ASCE,
Testing
Journal,
Geotechnique,
andJournal
Canadian
Geotechnical Journal.
ASTM Geotechnical Testing Journal, Geotechnique, and Canadian
Geotechnical Journal.
8. NOTATION
8.
NOTATION

parameters that influence geogrid reinforcement of railway ballast.


Bolmsvik, R.
"Influence 25,
of (6),
USPpp.
on326-335.
the track responsea
Geotextiles
and(2005)
Geomembranes,
literature survey."
Bruni, S., Anastasopoulos, I., Alfi, S., Van Leuven, A., Apostolou, M.
Brown,
S. F., G.
Kwan,
J. and
Thom, N. H. Vibrations
(2007) "Identifying
key
and
Gazetas,
(2009)
Train-Induced
on Urbanthe
Metro
parameters
that influence
geogrid
reinforcement
of railway 135,
ballast."
and
Tram Turnouts.
Journal
of Transportation
Engineering,
(7),
Geotextiles
pp.
397-405.and Geomembranes, 25, (6), pp. 326-335.

The symbols used in this paper are listed below:


A
B
Cc
Cu
c
D10
dEB
Dmax
Dmin
E
Eoed
Eur
E50

N
P1
P2
Pref
q

q
v
1
3

1
3

cv
m
p

= Shift in the PSD curve after the test


= Potential breakage or the area between the
arbitraryboundary of maximum breakage and
the final PSD
= Coefficient of curvature
= Coefficient of uniformity
= Cohesion (kPa)
= Effective particle size (mm)
= Incremental energy consumption by particle
breakage per unit volume (kNm/m3)
= Maximum particle size (mm)
= Minimum particle size (mm)
= Youngs modulus (kPa)
= Stress-dependent tangent stiffness modulus for
primary loading (kPa)
= Stress-dependent secant stiffness modulus for
unloading and reloading (kPa)
= Stress-dependent secant stiffness modulus for
primary loading (kPa)
= Constant of proportionality
= Number of blows
= High-frequency impact force (kN)
= Low-frequency impact force (kN)
= Reference pressure (kPa)
= Deviator stress (kPa)
= Unit weight (kN/m3)
= Shear strain
= Volumetric strain
= Average vertical strain (major principal strain)
in ballast layer
= Average lateral strain (minor principal strain) in
ballast layer
= Poissons ratio
= Major principal effective stress (kPa)
= Minor principal effective stress (kPa)
Friction angle (degree)
= Friction angle at critical state (degree)

Dahlberg,
T. (2010) Railway
Track
Variations

Bruni, S., Anastasopoulos,


I., Alfi, S.,
Van Stiffness
Leuven, A.,
Apostolou,
Consequences
and
Countermeasures.
International
Journal
of
Civil
M. and Gazetas, G. (2009) "Train-Induced Vibrations on Urban
Engineering,
8, (1),
pp. 1-12. Journal of Transportation Engineering,
Metro and Tram
Turnouts."
135,
(7),
pp.
397-405.
Esveld, C. (2001). Modern railway track. MRT-Production, The

Netherlands.
Dahlberg, T. (2010) "Railway Track Stiffness Variations
Consequences
andand
Countermeasures."
International
of Civil
Frederick,
C. O.
Round, D. J. (1985).
Vertical Journal
Track Loading.
Engineering,
8, (1), pp.
1-12.
Track
Technology:
Proceedings
of a Conference, University of
Nottingham, UK, Thomas Telford Ltd, pp. 135-149.
Esveld, C. (2001). Modern railway track. MRT-Production, The
GM/TT0088
Netherlands. Rail Safety and Standards Board (1993) Permissible
track forces for railway vehicles, Group Standard, Issue 1, Revision
A.
Rail Safety
andand
Standards
London.Vertical Track Loading.
Frederick,
C. O.
Round, Board,
D. J. (1985).
Track Technology:
Proceedings
of a The
Conference,
University
of
Henkel,
D. J. and Gilbert,
G. D. (1952)
Effect Measured
of the
Nottingham,
UK, Thomas
pp. 135-149.Strength of Clay
Rubber
Membrane
on theTelford
TriaxialLtd,
Compression

Samples. Gotechnique, 3, pp. 20-29.


GM/TT0088 Rail Safety and Standards Board (1993) Permissible
track forcesB.,
for Ionescu,
railway vehicles,
Group H.
Standard,
1, Behavior
Revision
Indraratna,
D. and Christie,
(1998) Issue
Shear
A. Railway
Rail Safety
and Standards
London.
of
Ballast
Based on Board,
Large-Scale
Triaxial Tests. Journal of
Geotechnical and Geoenvironmental Engineering, 124, (5), pp. 439Henkel, D. J. and Gilbert, G. D. (1952) "The Effect Measured of the
449.
Rubber Membrane on the Triaxial Compression Strength of Clay
Indraratna,
B., Lackenby,3, J.pp.and
Christie, D. (2005) Effect of
Samples." Gotechnique,
20-29.
confining pressure on the degradation of ballast under cyclic loading.
Gotechnique
325-328.
Indraratna, B.,55,
Ionescu,
D. and Christie, H. (1998) "Shear Behavior
of Railway Ballast Based on Large-Scale Triaxial Tests." Journal of
Indraratna, B., Nimbalkar, S., Christie, D., Rujikiatkamjorn, C. and
Geotechnical and Geoenvironmental Engineering, 124, (5), pp. 439Vinod, J. (2010) Field Assessment of the Performance of a Ballasted
449.
Rail Track with and without Geosynthetics. Journal of Geotechnical
and
Geoenvironmental
Engineering,
136, (7), pp.
Indraratna,
B., Lackenby,
J. and Christie,
D. 907-917.
(2005) "Effect of

confining pressure
on the S.
degradation
of ballast under
cyclic
Indraratna,
B., Nimbalkar,
and Rujikiatkamjorn,
C. (2012).
loading." Gotechnique
Performance
evaluation55,
of 325-328.
shock mats and synthetic grids in the
improvement of rail ballast pp. 47-62.
Indraratna, B., Nimbalkar, S., Christie, D., Rujikiatkamjorn, C. and
Indraratna,
Salim,
W. and
Rujikiatkamjorn,
(2011). Advanced
Vinod, J. B.,
(2010)
"Field
Assessment
of theC.Performance
of a
Rail
Geotechnology:
Ballasted
Track.
CRC Press/Balkema,
Ballasted
Rail Track with
and without
Geosynthetics."
Journal of
Rotterdam,
Netherlands.
Geotechnical
and Geoenvironmental Engineering, 136, (7), pp. 907917.
Indraratna, B., Shahin, M. A. and Salim, W. (2007) Stabilisation of
granular
media
formationS.soil
using
geosynthetics with
special
Indraratna,
B., and
Nimbalkar,
and
Rujikiatkamjorn,
C. (2012).
reference
to railway
engineering.
of the ICE
- Ground
Performance
evaluation
of shock Proceedings
mats and synthetic
grids
in the
Improvement
11,rail
27-43.
improvement of
ballast pp. 47-62.

= Mobilized friction angle (degree)


= Peak friction angle obtained from peak stress
ratio, 1 3 (degree)
f
= Dilatancy angle (degree)
= Mobilized dilatancy angle (degree)

40

3rd Proff 18-02-2015

Geotechnical Journal Vol. 6 No. 1 2014 41

Jenkins, H. M., Stephenson, J. E., Clayton, G. A., Moorland, J. W.


and Lyon, D. (1974) The effect of track and vehicle parameters on
wheel/rail vertical dynamic forces. Railway Engineering Journal, 3,
(1), pp. 2-16.

Plek, O., Svoboda, R. and Hruzkov, M. (2007).Assembly of


under sleeper pads in turnouts for the homogenization of vertical
rail deflectionsCentre for Integrated Design of Advanced Structures,
Czech Republic.

Kaewunruen, S. and Remennikov, A. (2010) Dynamic Crack


Propagations in Prestressed Concrete Sleepers in Railway Track
Systems Subjected to Severe Impact Loads. Journal of Structural
Engineering, 136, (6), pp. 749-754.

QR (2001), STD/0026/TEC Rollingstock Dynamic Performance,


Safety Management System, Version: 2, Queensland Rail.

Kumaran, G., Menon, D. and Nair, K. K. (2002) Evaluation of


dynamic load on railtrack sleepers based on vehicle-track modeling
and analysis. International Journal of Structural Stability and
Dynamics, 02, (03), pp. 355-374.
Lackenby, J., Indraratna, B., McDowell, G. and Christie, D. (2007)
Effect of confining pressure on ballast degradation and deformation
under cyclic triaxial loading. Gotechnique 57, 527-536.
Lee, M. L., Chiu, W. K. and Koss, L. L. (2005) A Numerical Study
into the Reconstruction of Impact Forces on Railway Track-like
Structures. Structural Health Monitoring, 4, (1), pp. 19-45.
Li, D. and Davis, D. (2005) Transition of Railroad Bridge
Approaches. Journal of Geotechnical and Geoenvironmental
Engineering, 131, (11), pp. 1392-1398.
Loy, H. (2008). Under Sleeper Pads: improving track quality while
reducing operational costs. European Railway Review, Issue 4.
Loy, H. (2012) Mitigating vibration using under-sleeper pads.
Railway Gazette International, 168, (4), pp. 40-42.
Marschnig, S. and Veit, P. (2011) Making a case for under-sleeper
pads. International Railway Journal, 51, (1), pp. 27-29.
Nielsen, J. C. O. and Johansson, A. (2000) Out-of-round railway
wheels - a literature survey. Proceedings of the Institution of
Mechanical Engineers Part F-Journal of Rail and Rapid Transit, 214,
(2), pp. 79-91.
Nimbalkar, S., Indraratna, B., Dash, S. and Christie, D. (2012)
Improved Performance of Railway Ballast under Impact Loads
Using Shock Mats. Journal of Geotechnical and Geoenvironmental
Engineering, 138, (3), pp. 281-294.

RIC (2002), RSU120 General Interface Requirements, Version: 2.0,


Rail Infrastructure Corporation (Rail Corp).
Saxton, H. J., Ireland, D. R. and Server, W. L. (1974). Analysis
and control of inertial effects during instrumented impact testing.
ASTM Spec. Tech. Publ., American Society of Testing and Materials,
Philadelphia, pp. 50-73.
Schanz, T., Vermeer, P. A. and Bonnier, P. G. (1999). The hardening soil
model: Formulation and verification. Beyond 2000 in Computational
Geotechnics Balkema, Rotterdam, A A Balkema Publishers, pp. 281296.
Schneider, P., Bolmsvik, R. and Nielsen, J. C. O. (2011) In situ
performance of a ballasted railway track with under sleeper pads.
Proceedings of the Institution of Mechanical Engineers, Part F:
Journal of Rail and Rapid Transit, 225, (3), pp. 299-309.
Selig, E. T. and Waters, J. M. (1994). Track Geotechnology and
Substructure Management. Thomas Telford, London.
Steffens, D. and Murray, M. (2005). Establishing meaningful results
from models of railway track dynamic behaviour. Proceedings of 8th
IHHA Conference, Rio de Janeiro, Brazil, pp. 41-50.
Steffens, D. M. (2005).Identification and development of a model
of railway track dynamic behaviour. Master of Engineering
Thesis,School of Built Environment,Queensland University of
Technology, Australia.
Timoshenko, S. and Goodier, J. N. (1951). Theory of Elasticity.
McGraw-Hill book Company.
Zhai, W. and Cai, Z. (1997) Dynamic interaction between a lumped
mass vehicle and a discretely supported continuous rail track.
Computers & Structures, 63, (5), pp. 987-997.

42 Geotechnical Journal Vol. 6 No. 1 2014

GROUND IMPROVEMENT TO MITIGATE EARTHQUAKE-INDUCED SOIL LIQUEFACTION


HAZARDS
D. Wijewickreme1
1Professor of Civil Engineering, Department of Civil Engineering
University of British Columbia, Vancouver, Canada E-mail: dharmaw@civil.ubc.ca
ABSTRACT: Ground improvement is commonly used as a means to enhance the geotechnical engineering performance of
structures under anticipated design loads. In essence, the objective is to increase the stiffness and/or strength properties of soil
through ground improvement, and in turn, to resist the system loads while keeping the deformations within acceptable limits
from a performance point of view. In addition to operational loads, permanent ground displacements and/or loss of bearing
capacity due to earthquake-induced liquefaction are key geotechnical hazards to structures founded on loose saturated granular
soils in seismically active areas. Typical methods of ground improvement against liquefaction-induced geotechnical hazards
include: dynamic compaction, vibro-replacement using stone columns, compaction piling, preloading, blast densification, and
compaction grouting. There is field evidence to suggest that ground improvement is effective in reducing seismic damage
to facilities. A number of published geotechnical case histories involving ground improvement for mitigating liquefaction
induced geotechnical hazards, originating from the Lower Mainland of British Columbia, Canada, which is located in a moderate
to high seismic risk region, are reviewed. In particular, current approaches for the prediction of earthquake-induced ground
deformations, considerations in governing the selection of suitable ground improvement methods, and applicability of such
methods to address typical engineering situations are highlighted using the case histories.
1 INTRODUCTION
Enhancing geotechnical stiffness and strength properties is critical to
controlling settlements and bearing capacity of structures founded or
supported on soils. A variety of techniques evolved in the past few
decades are commonly used for improving ground and, in turn, the
properties of soils (Mitchell, 1981; JGS, 1998).
Some of the ground improvement measures are directly aimed at
reducing the risk of soil liquefaction in seismically active areas. As
noted by Youd and Perkins (1978), regions underlain by relatively
young marine, deltaic, and alluvial soil deposits are considered to
be susceptible to liquefaction and large ground movements when
subjected to earthquake shaking. Liquefaction would trigger loss of
shear strength and reduction in deformation modulus (stiffness) in
soils. As such, earthquake-induced permanent ground displacements
and/or loss of bearing capacity are some key geotechnical hazards
to structures located at sites underlain by liquefiable soils (MCEER,
1999; ORourke and Hamada, 1992). The observed performance of
sites following major earthquake events [e.g., 1964 Niigata (Japan),
1995 Hyogoken Nanbu (Kobe, Japan), 1999 Kocaeli (Turkey), 2001
Nisqually (Washington State, USA)] indicates that the sites with
improved ground had generally less susceptibility to earthquakeinduced permanent ground deformations and resulting damage than
the sites that had not been densified (Mitchell et al., 1998; Hausler
and Sitar, 2001; Hausler and Koelling, 2004). Typical ground
improvement measures include dynamic deep compaction, vibroreplacement using stone columns, compaction piling, explosive
compaction, and compaction grouting.
In general, there are four ways to consider in improving the seismic
performance against an identified geotechnical hazard: (a) avoid
the hazard by relocation; (b) isolate the structure from the hazard;
(c) accommodate the hazard by strengthening the structure; and (d)
reduce the hazard using ground improvement. Typically, all of the
above options are considered in developing retrofit concepts. When
ground improvement is considered as the desired option, the selection
of the most suitable remedial option is governed by many factors
including: soil conditions, equipment/space restrictions, issues related
to the protection of existing structures during ground improvement,
operational constraints, environmental regulatory requirements, and
land availability.
In this paper, a number of published geotechnical case histories
involving ground improvement for mitigating liquefaction induced

geotechnical hazards, originating from the Lower Mainland of


British Columbia (BC), Canada, which is located in a moderate to
high seismic risk region, are reviewed. Considerations governing the
selection of suitable ground improvement methods and applicability
of such methods to address typical engineering situations are
highlighted using these case histories. The paper draws particularly
from publications of previous involvement by the author on ground
improvement against earthquake-induced soil liquefaction and
associated hazards (Wijewickreme et al. 2002; Wijewickreme et al.
2005; Wijewickreme and Atukorala, 2006).
2 BURIED GAS PIPELINE SITE
A case history on the seismic upgrading of a buried natural gas
pipeline gate station site in Vancouver, BC, is presented in this section.
The gate station is part of a major natural gas transmission system.
Prevention of loss of pipeline pressure integrity under earthquake
loading corresponding to an equivalent return period 2,000 years was
used as the performance criterion for the acceptability of pipeline
performance. Liquefaction-induced ground deformations were
identified as significant hazards to the pipelines entering the gate
station and associated facilities.
2.1 Site Description and Subsurface Soil Conditions
The gate station site is generally rectangular in plan (~100 m x 75
m) and located on the North Bank of the North Arm of the Fraser
River (see Fig. 1) in Vancouver, BC, Canada. As illustrated, two
transmission pipelines 20 in. and 24 in. diameter (i.e., NPS 20 and
NPS 24 pipes) enter the gate station below the riverbed from the south.
The site topography within the station compound and also in the eastwest direction is generally flat. Prior to ground improvement, the
river bank sloped down towards the south at slopes ranging 1H:1V to
3H:1V (horizontal: vertical) within the rip-rap area which extended
to about 6 m below crest level. The riverbed below this level sloped
southward at an average gradient of about 8% to the horizontal.
Fig. 2 presents the inferred soil stratigraphy at the site, developed
based on a geotechnical field investigation. A combination of the
methods of electric cone penetration testing (CPT), mud-rotary
drilling, and solid-stem auger drilling was used in this investigation.
The upper soils within the station consisted of about 2 to 3 m of loose
to compact sand to sandy silt fill material. The upper fill materials
in the northern part of the gate station were found to be underlain
by a layer of very soft to soft silt (Liquid limit, LL = 38%; Plasticity

was analysed, using both circular and non-circular failure surfac


to investigate the potential for a flow slide condition at the site. T
post-liquefaction shear strength parameters for potentia
liquefiable zones Geotechnical
were mainlyJournal
selected
based
a review
Vol.
6 No.on1 2014
43 of numb
of previously published results of laboratory post-cyclic monoton
simple
shear
test
Potential
slip zones
with significa
The
stability
of the
gatedata.
station under
post-liquefaction
condition
was
analysed, using
boththe
circular
and compound
non-circular failure
surfaces,
encroachment
into
station
(failure
zones extendi
to
investigate about
the potential
a flow
condition
the site.
The
landwards
30 mfor
from
theslide
river
bank) at
were
computed
to have
post-liquefaction
shear
strength
parameters
for
potentially
liquefiable
post-liquefaction factor of safety less than 1.0 even witho
zones were mainly selected based on a review of number of previously
application
ofthe
anygate
seismic
inertia
forces.
This suggested
a high ri
The
stability
of
stationpost-cyclic
under post-liquefaction
condition
published
results
of laboratory
monotonic simple
shear
of
a
flow
slide
as
a
result
of
earthquake
shaking
leading
to ve
was
analysed,
using
both
circular
and
non-circular
failure
surfaces,
test data. Potential slip zones with significant encroachment into
to
investigate
the
potential
for
a
flow
slide
condition
at
the
site.
The
large
deformations
for
the
southern
part
of
the
site.
the station compound (failure zones extending landwards about 30

asticity index, PI = 11%; Water content, w = 40%) extending to


pths in the order of 6 to 8 m below the ground surface. The silt
ne is underlain
by18-02-2015
a compact to dense sand stratum, which, in
3rd Proff
rn, was found to overlie a stratum of very dense sand and gravel at
depth of about 9 m below the ground surface. Within the southern
oreline ofindex,
the PI
gate
station,
the6 soils
underlying
theto depths
upperinfill
= 11%;
Water
content,
40%) extending
the
Geotechnical
Journal
Vol.
No.w1 =2014
order of 6consisted
to 8 m below
ground
Thesand
silt zone
is underlain
aterials primarily
ofthe
loose
to surface.
compact
extending
to
compact
to dense
stratum,
which, in
turn, was found
to
pths of upbytoa 12
m below
thesand
ground
surface.
Underlying
these
overlie
a
stratum
of
very
dense
sand
and
gravel
at
a
depth
of
about
ils, compact to dense sand with some gravel was encountered.
m below the ground surface. Within the southern shoreline of the
hese strata 9gate
are
underlain
by
dense
glacial
till-like
material
that
was
underlain
by athe
layer
of
very soft
tothe
soft
silt (Liquid
limit, LL
= 38%;
station,
soils
underlying
upper
fill materials
primarily
countered consisted
at a depth
of
about
14
m
below
the
ground
surface.
Plasticity
index,
PI
=
11%;
Water
content,
w
=
40%)
extending
to
of loose to compact sand extending to depths of up to 12 m
depths
in
the
order
of
6
to
8
m
below
the
ground
surface.
The
silt
PT testing below
within
the
river
adjacent
to
the
site
also
indicates
the
the ground surface. Underlying these soils, compact to dense
zone with
issoils,
underlain
by a
awas
compact
to denseof
sand
stratum,
esence of sandy
below
2-m
thickness
siltstrata
and clayey
silt,in
sand
some
gravel
encountered.
These
are which,
underlain
turn,
was glacial
found to
overliematerial
a stratum
ofwas
veryencountered
dense sand and
gravel
at
by
dense
till-like
that
at
a
depth
of
d extending
down to a depth of about 9 m below the riverbed.
a depth of about 9 m below the ground surface. Within the southern
about are
14 munderlain
below the ground
surface. CPT
testing soil
withinstratum.
the river
hese materials
a compact
dense
shoreline of
the gate by
station,
the soils to
underlying
the upper fill
adjacent to the site also indicates the presence of sandy soils, below
he groundwater
at the
site was
assessed
to sand
be located
materialslevel
primarily
consisted
of loose
to compact
extendingat
to
a 2-m thickness of silt and clayey silt, and extending down to a depth
depths
of 3upmtobelow
12 m below
the ground
surface.
Underlying
these
pths of about
1
to
the
ground
surface
at
the
site.
of about 9 m below the riverbed. These materials are underlain by a

post-liquefaction
shearwere
strength
parameters
for potentially
m
from the river bank)
computed
to have a post-liquefaction
liquefiable
zones
were
mainly
selected
based
on a review
of number
factor
of
safety
less
than
1.0
even
without
application
of any
seismica number
Ground
displacement
analyses
were post-cyclic
conducted
using
of
previously
published results
of laboratory
monotonic
inertia forces. This suggested a high risk of a flow slide as a result
methods
available
at the
time toslipassess
magnitude
simple
shear
test data.
Potential
zonesthe
with
significantand patter
of earthquake shaking leading to very large deformations for the
encroachment
intoground
the station
compound in
(failure
zonesnorth
extending
of
the
relative
movements
the
area
of the predict
southern part of the site.
landwards
about
30
m
from
the
river
bank)
were
computed
to
have
a
flow slide zone. In particular, the liquefaction-induced
free-fie
post-liquefaction
factor
of safety
than using
1.0 even
without
Ground
displacement
analyses
were less
conducted
a number
of
ground
displacements
were
calculated
using
the
computer
progra
application
of any at
seismic
inertia
forces.
suggested
high risk
methods
available
the time
to assess
the This
magnitude
and apatterns
of
developed
by
Houston
et
al.
(1987),
sliding
block
method
of
a
flow
slide
as
a
result
of
earthquake
shaking
leading
to
very
the relative ground movements in the area north of the predicted flow
large
deformations
for the
southern
part MLR
of the site.
Newmark
(1965),
thethe
empirical
method
developed
slide
zone.
In
particular,
liquefaction-induced
free-field
ground by Bartl
and Youd (1992),
and using
a mechanistic
displacements
were calculated
the computer finite
programelement
developedapproach
Ground
displacement
analyses
a number
of
by
Houston
et al.
(1987),
slidingwere
blockconducted
method byusing
Newmark
(1965),
Byrne
et
al.
(1992).
methods
available
at
the
time
to
assess
the
magnitude
and
patterns
the empirical MLR method developed by Bartlett and Youd (1992),
of the
relative ground
in the area
north et
ofal.
the(1992).
predicted
and
a mechanistic
finitemovements
element approach
by Byrne
The slide
predictions
all analysis
techniques indicated
flow
zone. Infrom
particular,
the liquefaction-induced
free-field that, for t
The
predictions
from corresponding
all
analysis
techniques
indicated
for the
ground
displacements
were
calculated
using
computer
program
seismic
loadings
to allthe
the
riskthat,
levels
considered
seismic
loadings
corresponding
to(1987),
all the risk
levelsblock
considered
the
developed
by
Houston
et
al.
sliding
method
by towards t
the study, large ground displacements (in excess of in
3 m)
study,
large(1965),
groundthe
displacements
(in excess
of developed
3 m) towards
river
Newmark
empirical MLR
method
bythe
Bartlett
river influence
would
influence
an area
extending
to from
about
m north fro
would
an area
to about
30 m
north
the30
crest
and Youd
(1992),
and extending
a mechanistic
finite
element
approach
by
thethecrest
theThe
river
bank.
The ground
displacements
Byrne
et al. of
(1992).
of
river
bank.
ground
displacements
for the
non-liquefiable for the no
liquefiable
siltythezones
within
half of to
the site we
silty
zones within
northern
half ofthe
the northern
site were computed
be
less
than about
Along
with 0.1
lateral
ground
movements,
The
predictions
allm.analysis
techniques
indicated
that,with
for the
computed
to from
be0.1
less
than
about
m.
Along
lateral grou
significant
vertical
ground movements
expected
toconsidered
seismic
loadings
corresponding
to all were
the risk
levels
in expected
movements,
significant
vertical
ground
movements
were
the study, large ground displacements (in excess of 3 m) towards the
river would influence an area extending to about 30 m north from
the crest of the river bank. The ground displacements for the nonliquefiable silty zones within the northern half of the site were
computed to be less than about 0.1 m. Along with lateral ground
movements, significant vertical ground movements were expected to

soils, compact to dense sand with some gravel was encountered.


compact
to dense
soil stratum.
The groundwater
level
at the that
site was
These strata
are underlain
by dense
glacial till-like
material
was
assessed
to
beatlocated
depths
about
1 to 3 m
thesurface.
ground
2 Geotechnical
Performance
under
Earthquake
Loading
encountered
a depthat of
aboutof 14
m below
thebelow
ground
surface
at thewithin
site. the river adjacent to the site also indicates the
CPT testing
he seismicpresence
response
of soils,
the below
site was
oneof
sandy
a under
2-m assessed
thickness
ofusing
siltLoading
and the
clayey
silt,
2.2 Geotechnical Performance
Earthquake
mensional and
wave
propagation
et al.,
extending
down to program
a depth of SHAKE
about 9 m(Schnabel
below the riverbed.
seismic
response
the by
site
was
assessed
using
the
oneThese
materials
are underlain
to dense
soilused
stratum.
72), and The
charts
developed
byof Seed
eta compact
al. (1985)
were
to
dimensional
wave
propagation
program
SHAKE
(Schnabel
et
al.,
The
groundwater
level
at
the
site
was
assessed
to
be
located
at
sess the liquefaction potential of the site soils. The results
1972),
charts1 developed
by the
Seedground
et al. (1985)
depths and
of about
to 3 m below
surfacewere
at theused
site.to assess
dicated that
the loose to compact sands in the southern portion of
the liquefaction potential of the site soils. The results indicated that
e site would
liquefy
under
of
seismic
loading
the
to compact
sands the
in thelevels
southern
portion
of the
site
would
2.2 loose
Geotechnical
Performance
under Earthquake
Loading
vestigated.liquefy
The under
loosethe
sandy
at theloading
site extending
a depth
levelssoils
of seismic
investigated.to The
loose
The seismic
response
of the to
sitea depth
was assessed
using
thefound
onesoilsfound
at the
site
extending
of about 12
m was
about 12sandy
m
was
to
be
potentially
liquefiable
(see
Fig.
2).
dimensional wave propagation program SHAKE (Schnabel et al.,
to bemagnitude
potentially liquefiable
(see Fig. 2). An10
earthquake
magnitude
n earthquake
of
M7
(representing
to
15
cycles
1972), and charts developed by Seed et al. (1985) were used of
to
M7 (representing
10 to 15 cycles
of loading) was used in the
ading) wasof
used
in liquefaction
the liquefaction
assessment
to
assess
the
potential
of the site corresponding
soils. The results
liquefaction assessment corresponding to the considered seismic
indicated
thathazard
the loose
to compact sands in the southern portion of
e considered
seismic
level.
hazard
the sitelevel.
would liquefy under the levels of seismic loading
investigated. The loose sandy soils at the site extending to a depth
of about 12 m was found to be potentially liquefiable (see Fig. 2).
An earthquake magnitude of M7 (representing 10 to 15 cycles of
loading) was used in the liquefaction assessment corresponding to
the considered seismic hazard level.

Fig. 1. Site plan showing existing structures, pipeline configurations, and geotechnical testhole locations.

Fig. 1. Site plan showing existing structures, pipeline configurations, and geotechnical testhole locations.
occur within the southern area of the site.

occur within the southern area of the site.

43

43

Geotechnical
Journal
Vol. 6Vol.
No.61 No.
20141 2014
44
Geotechnical
Journal
Geotechnical Journal Vol. 6 No. 1 2014

Fig. 2. Profile of soil stratigraphy, predicted zone of potential liquefaction, and alignment of 610 mm diameter pipeline.
Fig. 2. Profile of soil stratigraphy, predicted zone of potential liquefaction, and alignment of 610 mm diameter pipeline.
Maximum computed ground deformations derived from the
ground surface preventing the installation of some 20 stone
geotechnical
analyses were
with the
computed
columns.
Of these
locations,
columnsofwere
Maximum computed
groundcompared
deformations
derived
from pipe
the
ground surface
preventing
the some
installation
somesuccessfully
20 stone
Maximum
computed
ground
deformations
derived
from
the
some
columns
were
installed
at alternate
locations
by
structural
deformation
capacities.
Thewith
computed
large
ground
installed
at Of
alternate
locations
relocating
within
m of the
geotechnical
analyses were
compared
the
computed
pipe
columns.
these successfully
locations, by
some
columns
were1.5
successfully
geotechnical
analyses
were
compared
with
the computed
pipe
relocating
1.5
m
of theattempts
design
location.
In general,
displacements
and resulting
differential
at the
gate
design
In general,
were made
to relocate
structural deformation
capacities.
Thedisplacements
computed
large
ground
installedlocation.
atwithin
alternate
locations
by relocating
within
1.5 mattempts
ofstone
the
structural
deformation
capacities.
Thedisplacements
computed
large
ground
were
to relocate
stoneexcavating
columns
than
excavating
station
from
earthquake-induced
liquefaction
were found
to
well
columns
rather
than
the
obstruction.
displacements
and resulting
differential
at the
gate
designmade
location.
In locally
general,
attemptsrather
were
madelocally
to relocate
stone
displacements
and
resulting
differential
displacements
at
the
gate
the
obstruction.
exceed
the
estimated
capacity
of
the
pipelines;
this,
in
turn,
station from earthquake-induced liquefaction were found to well
columns rather than locally excavating the obstruction.
indicated
the risk capacity
of damage
to the
station
pipingintounder
station
earthquake-induced
liquefaction
were found
well
Field verification
selected
centroids
of the
exceed from
thethatestimated
of
the
pipelines;
this,
turn,
Field
verification testing
testingwas
wasperformed
performedat at
selected
centroids
of
earthquake
loading
would
above
acceptance
criteria.
exceed
the that
estimated
capacity
ofwell
the pipelines;
this,
in turn,
indicated
stone
column pattern
using
CPT testing
during the
progress
of
indicated
the risk
of bedamage
to thethestation
piping
under
Field
verification
testing
was
performed
at
selected
centroids
of the
the
stone
column
pattern
using
CPT
testing
during
the
progress
of
The
only
remedial
measures
deemed
practical
the gatecriteria.
station
that
the
risk
of damage
to thebe
station
under
earthquake
loading
densification.
The
results
of the
post-densification
testing
together
earthquake
loading
would
well piping
above
the for
acceptance
stone
column
pattern
using
CPT
testing
during
the
progress
of
densification. The results of the post-densification testing together
involved
thethe
ground
conditions.
would
beimproving
well above
acceptance
criteria. for
Thethe
only
with
review of the results
stone column
installation details indicated
that
The only
remedial
measures
deemed
practical
gateremedial
station
densification.
of the post-densification
together
with
review ofThe
the stone column
installation detailstesting
indicated
that
values installation
generally exceeded
pre-specified
the
cone
tip resistance
(Qt)column
involved improving
the ground
measures
deemed practical
for conditions.
the gate station involved improving
withcone
review
of the stone
details indicated
that
the
tip resistance
(Qt)
values
generally
exceeded
pre-specified
criteria
(ranged
between
100
and
125
bars
for clean
performance
Q
Theground
effectiveness
the
conditions.of ground improvement in reducing the
t
) values generally exceeded pre-specified
the
cone
tip
resistance
(Q
t
performance
criteria
(ranged
100 andout
125
bars for
clean
sand
zones). Qt
of the
initialbetween
CPTs, carried
within
about
14
liquefaction-induced
ground
displacements
at the in
site reducing
was assessed
performance
QSome
The effectiveness of
ground
improvement
the
t criteria (ranged between 100 and 125 bars for clean
The
effectiveness
ofstability
ground improvement
in reducing
the liquefactionsand zones). Some
of the initial CPTs, carried out within about 14
again
using
slope
and
finite
element
analysis.
The
results
sand zones). Some of the initial CPTs, carried out within about 14
liquefaction-induced ground displacements at the site was assessed
induced
ground
displacements
atoftheansite
was assessed
againbarrier,
using
indicated
that
the
introduction
in-ground
densified
again using
slope
stability
and finite
element
analysis.
The results
slope
stability
and
finite
element
analysis.
The
results
indicated
that
likely
in
the
order
of
15
to
20
m
wide,
would
reduce
the
expected
indicated that the introduction of an in-ground densified barrier,
the
introduction
of an
barrier,
likely
inthe
theexpected
order
of
large
earthquake-induced
ground
inreduce
the vicinity
of the
likely
in the order
ofin-ground
15 to
20 densified
m movements
wide, would
15
to
20
m
wide,
would
reduce
the
expected
large
earthquake-induced
gate
to a level below
the tolerable
ground
deformation
of the
largestation
earthquake-induced
ground
movements
in the
vicinity of
the
ground
movements
in below
thetovicinity
of the improvement,
gate
station
to the
a level
below
pipelines.
In
the
shoreline
gate station
to aaddition
level
theground
tolerable
ground
deformation
of the
the
tolerable
deformation
of slope
the improvement,
pipelines.
In addition
to the
slope
was configured
totoa the
gentler
to improve
the
pipelines.
Inground
addition
ground
the riverbank
shoreline
ground
improvement,
configured
a gentler
slope
slope stability.
was
configured the
to shoreline
a gentler slope
slopewas
to improve
thetoriverbank
slope to
improve the riverbank slope stability.
stability.
2.3 Ground Improvement Using Vibro-replacement
2.3 Ground Improvement Using Vibro-replacement
2.3 selection
Ground of
Improvement
Using ground
Vibro-replacement
The
the most suitable
improvement technique
The selection of the most suitable ground improvement technique was
was
governed
by the
several
factors
such
as soilimprovement
conditions, equipment
The
selection
of
most
suitable
ground
technique
governed by several factors such as soil conditions, equipment
space
space
restrictions,
pipeline
protection
issues,
environmental
was
governed
by
several
factors
such
as
soil
conditions,
equipment
restrictions, pipeline protection issues, environmental regulatory
regulatory
requirements,
land protection
availability issues,
etc. Based
on an
space restrictions,
pipeline
environmental
requirements,
land availability
etc. the
Based
on an
of these
evaluation
these
considerations,
method
ofevaluation
vibro-replacement
regulatory ofrequirements,
land availability
etc.
Based on an
considerations,
the method
vibro-replacement
was considered
to
was
considered
be theof most
suitable
of ground
evaluation
of thesetoconsiderations,
the
methodtechnique
of vibro-replacement
be
the
most
suitable
technique
of
ground
densification
for
use
at
the
densification
for use
gatemost
station
site. technique of ground
was considered
to atbethethe
suitable
gate station site.
densification for use at the gate station site.
Two hundred and seventy three (273) stone columns were installed
(using
the method
method
of
vibro-replacement)
in
triangular
pattern
at 33
Two hundred
and seventy
three (273) stone
were
installed
(using
the
of vibro-replacement)
in aacolumns
triangular
pattern
at
m
centre-to-centre
spacing
totocover
in in
Fig.
1atto31
(using
the method of
vibro-replacement)
in aarea
triangular
pattern
m
centre-to-centre
spacing
coverthe
theplan
plan
areashown
shown
Fig.
improve
thethe
overburden
AApoker
type
V-23
vibrator
with
m centre-to-centre
spacingsoils.
to
cover
the
plan
area
shown
in Fig.
1 toaa
to
improve
overburden
soils.
poker
type
V-23
vibrator
with
rated
energy
of165
165hphpwas
was
used
install
the
stone
columns
using
improve
the of
overburden
soils.
Ato
poker
type
V-23
vibrator
with
rated
energy
used
to
install
the
stone
columns
using
thea
the
feed
method
Fig.
3).
stone
columns
extended
to
the
ratedtop
energy
of 165
hp
was
to install
the
stone
columns
using
top
feed
method
(see(see
Fig.
3).used
All All
stone
columns
extended
to the
top
top
of
the
underlying
hard
stratum
to
depths
between
8
and
16
m
the
top
feed
method
(see
Fig.
3).
All
stone
columns
extended
to
the
of the underlying hard stratum to depths between 8 and 16 m below
below
ground
withdepths
an average
depth
of about
top existing
of the
the existing
underlying
hard surface,
stratum
between
8 and
16 m.
m
the
ground surface,
with antoaverage
depth
of about
14
14
m. theThe
average
amperage
outputan during
construction
of
below
ground
surface,
average
depth of about
The
averageexisting
amperage
output
duringwith
construction
of individual
stone
Fig. 3. Poker type V-23 vibrator (photo taken just before
individual
stoneaverage
columns was about
150 during
A, withconstruction
peak outputs
14 m. was
The
output
columns
about 150 amperage
A, with peak
outputs
ranging from 170 of
to
insertion
at the stone-column
location with
water
jets started).
ranging
from
170
to
260
A.
Boulders,
concrete
and
timber
Fig. 3. Poker
type V-23 vibrator
(photo
taken
just before
individual
stone concrete
columns and
wastimber
aboutobstructions
150 A, with
outputs
260
A. Boulders,
werepeak
encountered
obstructions
columnconcrete
installation
some
insertion
at
the
stone-column
location
with
water
jets
started).
ranging fromwere
170encountered
to 260 A. during
Boulders,
and attimber
days from the time of stone column installation, indicated that the
during
column
installation
at of
some
generally
depths
locations,
generally
at depths
somelocations,
3column
to 6 minstallation
below theatat
existing
obstructions
were encountered
during
some
wascolumn
not satisfied
in certain
zones that
of silty
specified
t requirement
days fromQthe
time of stone
installation,
indicated
the
of some 3 to 6 m below the existing ground surface preventing
locations, generally at depths of some 3 to 6 m below the existing
specified
Q
requirement
was
not
satisfied
in
certain
zones
of silty
the installation of some 20 stone columns. Of these locations, 44
t
44

Geotechnical Journal Vol. 6 No. 1 2014

3rd Proff 18-02-2015

Geotechnical Journal Vol. 6 No. 1 2014

Geotechnical Journal Vol. 6 No. 1 2014 45

Fig. 4. Results of post-densification electric cone penetration tests at the centroid of adjacent stone column triangular patterns (comparison
between results from tests conducted 13 and 34 days after installation).
Fig. 4. Results of post-densification electric cone penetration tests at the centroid of adjacent stone column triangular patterns (comparison
penetration resistance values in coarse granular soils. Therefore,
fine
sand results
(Note: from
Qt requirements
were13
corrected
the for
siltinstallation).
content);
between
tests conducted
and 34 days
after
BPT with a relatively larger diameter penetrating tool is often
however, repeat testing carried out in the same area after about five
adopted in the characterization of coarse-grained soils.
weeks from the installation of the stone columns indicated that the
penetration
resistance
values in coarse
soils. the
Therefore,
finesand
sand(Note:
(Note: Qt
Qt requirements
requirements were
were corrected
corrected the
fine
the for
for silt
silt content);
content);
The presence
of low permeability
layersgranular
of soil within
generally
Q
t values had increased significantly from the initial postBPT
with a relatively
larger
diameter
penetrating
tool and
is often
however,
repeat
testing
carried
out
in
the
same
area
after
about
five
however,
repeat
testing
carried
out
in
the
same
area
after
about
five
coarse-grained
overburden
soils
was
considered
a
concern,
this
densification values, and met the specified criteria for silty sands
The presence
of
low permeability
layers of soil within
the generally
adopted
in
the
characterization
of
coarse-grained
soils.
weeks
from
the
installation
of
the
stone
columns
indicated
that
the
weeks
from
(see
Fig.
4). the installation of the stone columns indicated that the Qt
coarse-grained overburden soils was considered a concern, and this
had increased
significantly
initial postQt values
values
had increased
significantly
from the from
initial the
post-densification
densification
values,
and
met
the
specified
criteria
for
silty
sands
The presence of low permeability layers of soil within the generally
and met theBRIDGE
specified criteria
for silty
sands (see Fig. 4).
3values,
HIGHWAY
CROSSING
SITE
(see Fig.
4).
coarse-grained overburden soils was considered a concern, and this
3 HIGHWAY BRIDGE CROSSING SITE
This
history describes
of the retrofit
adopted to
3 case
HIGHWAY
BRIDGEsome
CROSSING
SITE work
This
case
some
retrofit
work
to
enhance
thehistory
seismicdescribes
performance
ofofthethe
foundation
soils adopted
of a major
enhance
the
seismic
performance
of
the
foundation
soils
of
a
major
1.3 km long, 6-lane bridge on a highway in Vancouver, BC. The
Thiskmcase
history describes
some
of the retrofit
work adopted
to
1.3
long,
on
a highway
in Vancouver,
BC. The
primary
focus 6-lane
was to bridge
minimize
the
risk of bridge
collapse following
enhance
the
seismic
performance
of
the
foundation
soils
of
a
major
primary
focus
was to minimize
the risktoofa bridge
following
the
design
earthquake
corresponding
seismiccollapse
event having
an
1.3 km long, 6-lane bridge on a highway in Vancouver, BC. The
the design
earthquake
corresponding
to a (or
seismic
event
having
an
annual
probability
of exceedance
of 1/475
a return
period
of 475
primary focus was to minimize the risk of bridge collapse following
annual probability
of exceedance
of 1/475were
(or characterized
a return period
years).
The site-specific
ground motions
by of
a
the design earthquake corresponding to a seismic event having an
475 years).
The firm-ground
site-specific ground
were with
characterized
uniform
hazard
responsemotions
spectrum
a peak
annual probability of exceedance of 1/475 (or a return period of 475
horizontal
ground
acceleration
of 0.20
g and aspectrum
design earthquake
of
by
a uniform
hazard
firm-ground
response
with a by
peak
years).
The site-specific
ground motions
were characterized
a
magnitude
M7. Inacceleration
order to achieve
the
project
requirements,
a
horizontal
ground
of
0.20
g
and
a
design
earthquake
uniform hazard firm-ground response spectrum with a peak
displacement-based
design
was used,
and, in turn, grounda
of
magnitude
M7. acceleration
In orderapproach
to ofachieve
horizontal
ground
0.20 g the
andproject
a designrequirements,
earthquake of
improvement
was designed
to
limit ground
deformations
at locations
displacement-based
design
approach
was
used,
and, requirements,
in turn,
ground
magnitude M7. In order to achieve the project
a
that
are identified
as
critical.to limit ground deformations at locations
improvement
was
designed
displacement-based design approach was used, and, in turn, ground
that
are
identified
critical.
improvement
wasas
designed
to limit ground
3.1
Site
Description
and Subsurface
Soildeformations
Conditions at locations
that
are
identified
as
critical.
3.1 Site Description and Subsurface Soil Conditions
Fig. 5. Bridge site the alluvial fan can be seen on the left (i.e.,
The
north
side of the bridge
is underlain
byConditions
coarse granular soils
3.1 north
Siteside
Description
and is
Subsurface
Soil
The
ofsand
the bridge
underlain
coarse
north abutment).
(i.e.,
primarily
and gravel)
arisingbyfrom
angranular
alluvial soils
fan (i.e.,
of a
primarily
sand
and gravel)
arising
fan ofnature
a river of
stream
river
stream
deposit
(see Fig.
5).from
Dueantoalluvial
the coarse
site
aspect
investigated
the fan
method
ofseen
Sonic
where
Fig.
5.was
Bridge
site theusing
alluvial
can be
ondrilling,
the left (i.e.,
The north
side
of
the
is
underlain
by site
coarse
granular
soils
deposit
Fig.of
5).different
Duebridge
to the
coarse
nature
soils,
a number
of
soils,
a (see
number
techniques
wereof
utilized
to characterize
aspect
investigated
using
methodthat
of Sonic
drilling,
where
continuous
samples
of soils
weretheobtained
provided
information
north was
abutment).
(i.e.,
primarily
sand
and
gravel)
arising
from
an
alluvial
fan
of
a
different
techniques
utilized
to characterize
the subsurface
the
subsurface
soils.wereThese
included
conventional
drilling soils.
with
continuous
samples
of
soils
were
obtained
that
provided
information
on
soil
stratification.
This
was
considered
important
since
low
river stream
deposit
(see Fig.drilling
5). Duewith
to the
coarse
nature Becker
of site
aspect was investigated using the method of Sonic drilling, where
These
included
conventional
rotary
methods,
rotary
methods,
Becker Penetration
percussion
testing
with energy
on soil stratification.
consideredlead
important
sincepostlow
permeability
soil layersThis
canwas
potentially
to severe
soils, a number of different techniques were utilized to characterize
continuous samples of soils were obtained that provided information
Penetration percussion
testing
withwave
energyvelocity
measurements,
measurements,
downhole
shear
testing,downhole
seismic
permeability behaviour.
soil layers canBased
potentially
lead to severe
post-liquefaction
liquefaction
on ground
response
analyses, the
the subsurface soils. These included conventional drilling with
on soil stratification. This was considered important since low
shear wave velocity
testing, seismic
profiling, anddrilling
Sonic
refraction
profiling,
andrefractionSonic
foundation
were
as having
a high the
riskfoundation
of liquefaction
behaviour. soils
Based
on identified
ground response
analyses,
soils
rotary methods, Becker Penetration percussion testing with energy
permeability soil layers can potentially lead to severe post(http://www.prosoniccorp.com)
and sampling.
drilling (http://www.prosoniccorp.com)
and sampling.
under
the designasearthquake
ground
shaking.
The analytical
results
were
identified
having
a
high
risk
of
liquefaction
under
the
design
measurements, downhole shear wave velocity testing, seismic
liquefaction behaviour. Based on ground response analyses, the
indicated
soil liquefaction
extend results
to depths
varying
earthquakethat
ground
shaking.
Themay
indicated
thatfrom
soil
refraction
drilling
Becker
Penetrationprofiling,
Testing (BPT) and
consists of Sonic
driving a closed-toe
foundation
soils
were
identified
asanalytical
having a high
risk
of liquefaction
Becker
Penetration Testing (BPT)
consists
of driving a closed-toe
15
to
20
m
below
the
existing
ground
surface.
liquefaction
may
extend
to
depths
varying
from
15
to
20
m
below
the
(http://www.prosoniccorp.com)
and
sampling.
steel pipe that is 169 mm (6.65 in) in diameter using an ICE 180
under the design earthquake ground shaking. The analytical results
steel pipe that is 169 mm (6.65 in) in diameter using an ICE 180
existing
ground
surface.
diesel hammer that delivers 11 kN-m of rated maximum energy
indicated
thatofsoil
liquefaction
may extend
to depths varying from
3.2 Design
Ground
Densification
Configurations
diesel hammer that delivers 11 kN-m of rated maximum energy per
Becker
Testing
(BPT)
consists
driving
a closed-toe
per
blow.Penetration
The number
of blows
required
to of
drive
the steel
pipe are
15
20 m below
the existing
ground surface.
3.2toDesign
of Ground
Densification
Configurations
blow. The number of blows required to drive the steel pipe are
steel pipeover
that is 169 mmincrements
(6.65 in) of
in 0.3
diameter
using an
ICE 180
recorded
Rigorous
ground
responseDensification
analyses were
undertaken to optimize the
recorded
overpenetration
penetration increments
of m
0.3and
m these
and blow
thesecounts
blow
3.2
Design
of Ground
Configurations
dieselwith
hammer
that delivers
11 kN-m
oflater
rated
maximum
per
Rigorous
ground
response
analysestaking
were
undertaken
to optimize
along
hammer
energy levels
were
converted
to energy
equivalent
ground
densification
requirements
into consideration
the
counts
along
with hammer
energy
levels
were
later converted
to
blow. The
number Test
of blows
required
to following
drive the the
steel
pipe are
the
ground
densification
requirements
taking
into
consideration
Standard
Penetration
(SPT)
blow
counts
procedures
displacements that can be tolerated by the bridge foundations. the
In
equivalent Standard Penetration Test (SPT) blow counts following
Rigorous
ground
response
were
undertaken
to optimize theIn
recordedbyover
penetration
increments
0.3 m andSPT
these blow
displacements
that
canground
beanalyses
tolerated
by the
bridgeasfoundations.
outlined
Harder
and Seedby
(1986).
Theof
conventional
particular,
free-field
response
analyses
well as soilthe
procedures
outlined
Harder
and
Seed (1986).sampler,
The
ground
densification
requirements
taking
into
consideration
the
counts
along withsmaller
hammer energy[50
levels
later
converted
to
particular,interaction
free-field ground
response
analyses as to
well
as soil-structure
due
its relatively
mmwere
(2smaller
in.)
outer
diameter],
structure
analyses
were undertaken
assess
the bridge
conventional
SPT sampler,diameter
due its relatively
diameter
[50
displacements
that can
beundertaken
tolerated by
the
bridge
foundations.
In
equivalent
Standard
Penetration Test
(SPT)
blow counts
following
interaction
analyses
were
to
assess
the
bridge
foundation
could
give
rise
to
unrealistically
high
penetration
resistance
values
foundation
performance
under
the
475-year
ground
motions
using
mm (2 in.) outer diameter], could give rise to unrealistically high
particular,
free-field
ground
response
analyses
as
well
as
soilthe
procedures
outlined
by
Harder
and
Seed
(1986).
The
in coarse granular soils. Therefore, BPT with a relatively larger 45 performance under the 475-year ground motions using the computer
structure
interaction
were A
undertaken
to assess the
bridge
conventional SPT sampler, due its relatively smaller diameter [50
code FLAC
(Versionanalyses
3.4, 1998).
cyclic stress-strain
model
that
diameter penetrating tool is often adopted in the characterization of
foundation performance under the 475-year ground motions using
mm (2 in.) outer diameter], could give rise to unrealistically high
was capable of simulating sequential liquefaction in accord with
coarse-grained soils.
45 the commonly used liquefaction resistance chart (Seed et al. 1985;

the computer code FLAC (Version 3.4, 1998). A cyclic stress-strain


model that was capable of simulating sequential liquefaction in
46
Geotechnical
Journal Vol.
No. 1 2014resistance chart (Seed
accord
with the commonly
used6liquefaction
et al. 1985; Youd et al. 2001) was used in the analysis of ground
Youd
et al. 2001)
used were
in thecompared
analysis of
ground
deformations.
deformations.
Thewas
results
with
empirical
methods
of estimating
ground
deformations
and were
in good
The
results were
compared
with empirical
methods
of agreement.
estimating
For most
of the critical
liquefaction
predicted
ground
deformations
andfoundations,
were in good
agreement.was
Fornot
most
of the
below the
pile foundations,
and the
geotechnical
hazard
critical
foundations,
liquefaction
wasmain
not predicted
below
the was
pile
noted to ariseand
fromthe
liquefaction-induced
ground
foundations,
main geotechnicallateral
hazardspreading.
was notedAto
arise
improvement
program involving
the construction
in-ground
from
liquefaction-induced
lateral spreading.
A ground of
improvement
densified involving
zones (barriers)
at selectedofcritical
locations
along
the
program
the construction
in-ground
densified
zones
bridge
alignment
was
assessed
to
be
sufficient
to
reduce
the
lateral
(barriers) at selected critical locations along the bridge alignment
displacements
to be
acceptable
was
assessed to
sufficientlevels.
to reduce the lateral displacements to
acceptable
levels.
3.3 Ground Improvement Program
3.3 Ground Improvement Program
Ground improvement using vibro-replacement was selected as the
Ground
improvement
usingofvibro-replacement
selected
as the
most cost-effective
method
treatment in areas was
where
the available
most
cost-effective
method
of
treatment
in
areas
where
the
available
headroom was sufficient for the use of construction equipment. The
headroom
sufficient
for existing
the use settlement
of construction
equipment.
proximity was
of the
work to
sensitive
bridge
The
proximitydepth
of theofwork
to existing
settlement
sensitive
bridge
foundations,
treatment
required,
available
headroom
foundations,
depth
of treatment
available
beneath
beneath bridge
deck,
and cost required,
of treatment
were headroom
critical factors
in
makingdeck,
this and
decision.
ground
wasin required
bridge
cost ofSince
treatment
wereimprovement
critical factors
making
mostly
in over-water
stone columns
constructed
this
decision.
Since areas,
groundthe
improvement
was were
required
mostly
using
the bottom-feed,
wetcolumns
methodwere
of constructed
construction.
in
over-water
areas, the stone
using The
the
amperage build-up
depth
probe penetration
were monitored
bottom-feed,
wet and
method
of of
construction.
The amperage
build-up
during
stone
columnpenetration
construction.
contractor
elected
install
and
depth
of probe
wereThe
monitored
during
stonetocolumn
the stone columns
a square
grid topattern
a centre-to-centre
construction.
The using
contractor
elected
installatthe
stone columns
spacing
of 3 m.
total atofa 365
stone columns
wereofinstalled.
using
a square
grid A
pattern
centre-to-centre
spacing
3 m. A
Robust
equipment
that
can
sustain
amperage
readings
as highthat
as 300
total of 365 stone columns were installed. Robust equipment
can
A
for
longer
durations
in
the
order
of
10
to
30
seconds
were
sustain amperage readings as high as 300 A for longer durations
in
required
highly
coarse-grained
the
orderfor
of this
10 tosite
30with
seconds
were
required for soils.
this site with highly
coarse-grained soils.
The post-improvement testing was carried out using the Becker
The
post-improvement
Becker
5/8thusing
inch) the
casing
was
Hammer.
An HAV-180testing
rig withwas
169carried
mm (6out
Hammer.
rig with 169
mm (6 5/8th
inch)Bounce
casing
utilized forAnallHAV-180
post-improvement
verification
testing.
was
utilized
for allwas
post-improvement
verification
testing.
Bounce
chamber
pressure
measured with depth
of probe
penetration
in
order to obtain
necessary
parameters
the of
inference
of equivalent
chamber
pressure
was measured
withfor
depth
probe penetration
in
Typicalparameters
post-improvement
penetration
resistance
SPT (N
order
to1)obtain
necessary
for the inference
of equivalent
60 values.
measurements
obtained
at post-improvement
the centroid of the
vibro-columns
are
SPT
(N1)60 values.
Typical
penetration
resistance
shown in Fig. 6.obtained
The results
indicated
thatofthethe
stone
column spacing
measurements
at the
centroid
vibro-columns
are
and procedure
adequate
the target
SPT
shown
in Fig. 6.were
Themore
resultsthan
indicated
thattotheachieve
stone column
spacing
penetration
resistance
specified.
and procedure were more than adequate to achieve the target SPT
penetration
resistance specified.
4
FOUNDATIONS
AT AN INDUSTRIAL PLANT
4The
FOUNDATIONS
AT ANwork
INDUSTRIAL
ground improvement
undertakenPLANT
in the preparation of
foundations
for
a
kiln
line
and
two
large-diameter
storage
The ground improvement work undertaken in thematerial
preparation
of
silos (i.e., clinker
silo and
at a major cement
are
foundations
for a kiln
line cement
and twosilo)
large-diameter
materialplant
storage
presented
The and
peakcement
horizontal
acceleration
for
silos
(i.e., herein.
clinker silo
silo) firm-ground
at a major cement
plant are
the site was estimated to be 0.22g for the design seismic risk level
presented herein. The peak horizontal firm-ground acceleration for
corresponding to a 475-year return period. Foundation systems
the site was estimated to be 0.22g for the design seismic risk level
involved a combination of ground treatments including: piling,
corresponding to a 475-year return period. Foundation systems
replacement of weak soils, and densification of liquefiable soils.
involved a combination of ground treatments including: piling,
4.1 Site Description
andand
Subsurface
Soilof
Conditions
replacement
of weak soils,
densification
liquefiable soils.
TheSite
site Description
area is flat, and
and Subsurface
it is located Soil
on the
river bank of a major
4.1
Conditions
river, and the plan locations of the two silos in relation to the
The
site area
andare
it isidentified
located on in
theFig.
river7.bankThe
of ageneral
major river,
existing
plantis flat,
layout
soil
and
the planatlocations
the two silos
in relation
to thesoils
existing
conditions
the siteof comprise
soft/loose
deltaic
thatplant
are
layout
are identified
in Fig. 7.They
Thecan
general
soil conditions
at the
site
susceptible
to liquefaction.
be described
in terms
of four
comprise
soft/looseunits.
deltaicThe
soilsupper
that are
liquefaction.
main stratigraphic
unitsusceptible
comprisestogranular
river
They
can inbethe
described
terms
of four main
stratigraphic
sand fills
order of in
3m
in thickness.
Underlying
the fillsunits.
is a
The
upper
unit comprising
comprises granular
riverfine
sandsandy
fills in
thesilty
ordersands,
of 3
deltaic
deposit
interlayered
silts,
m
thickness.
fills isranging
a deltaic
deposit
comprising
andinclayey
silts Underlying
extending tothe
depths
from
4.0 m
to 6.5 m
interlayered
fine sandy
silts,
siltyon
sands,
and clayey silts
extending
below the ground
surface.
Based
the geotechnical
test data,
these
to depths ranging from 4.0 m to 6.5 m below the ground surface.
Based on the geotechnical test data, these deposits were considered 46
to be loose to compact and/or soft to firm. Underlying the deltaic
soils, river sand deposit with occasional loose silt/silty sand layering
was encountered. The measured cone penetration resistance indicated
that the sand deposit within about the upper 15-20 m is in a loose
to compact condition whereas the sands beneath this depth level,
generally, are inferred to be in a compact to dense state. The river

deposits were considered to be loose to compact and/or soft to firm.


Underlying the deltaic soils, river sand deposit with occasional loose
silt/silty sand layering was encountered. The measured cone
penetration resistance indicated that the sand deposit within about
sand
extends
to a depth
of aabout
35 m below
the ground
surface.
the upper
15-20
m is in
loose30totocompact
condition
whereas
the
sands
beneath
this depthmarine
level, generally,
are inferredinter-layered
to be in a
A
thick
compressible
deposit comprising
compact
to dense
extendslayers
to a depth
of about
sands,
silty
sands,state.
sandyThe
silts,river
andsand
occasional
of clayey
silt
30 to 35
m below
ground
surface.
A thick to
compressible
exists
below.
Thethe
marine
deposit
is inferred
extend to amarine
depth
deposit
comprising
sands,
sandy
silts,very
and
in
the order
of 250 inter-layered
m below the sands,
groundsilty
surface,
below
which
occasional
layers of
clayey silt
existstill
below.
Theare
marine
deposit
is
dense
Pleistocene
sediments
(glacial
or drift)
expected
to be
inferred to extend
a depth
in the
order
of inferred
250 m below
ground
encountered.
The to
ground
water
level
was
to be the
about
3m
surface,thebelow
which
veryondense
below
ground
surface
an Pleistocene sediments (glacial till
or drift) are expected to be encountered. The ground water level

Fig. 6. Typical post-improvement penetration resistance


profile compared with specification for required
densification.
was inferred
be about
3 m below the
surface
on an
average
basis, to
although
the groundwater
levelground
at the site
is expected
average
basis,thealthough
groundwater
expected
to
vary with
seasonalthe
precipitation
aslevel
well at
as the
the site
tidalisvariations
to the
varynearby
with the
seasonal precipitation as well as the tidal variations
of
river.
of the nearby river.
4.2 Clinker Silo
4.2 Clinker Silo
The clinker silo was built east of an existing silo foundation (i.e.,
called
B-silos)
shown
Fig. of
7. anThe
clinkersilo
silofoundation
has a diameter
The clinker
siloaswas
builtineast
existing
(i.e.,
of
45 B-silos)
m and as
theshown
foundation
to clinker
be designed
support an
called
in Fig. had
7. The
silo hastoa diameter
of
average
maximum
bearinghad
pressure
500 kPa
static
loading
45 m and
the foundation
to be of
designed
to under
support
an average
conditions.
Ground pressure
response of
analysis
indicated
a soil
zone
maximum bearing
500 kPa
underthat
static
loading
extending
depth ofresponse
about 18analysis
m belowindicated
the ground
the
conditions.to aGround
thatsurface
a soil atzone
clinker
silotolocation
potentially
under
the design
extending
a depthisof
about 18 liquefiable
m below the
ground
surfaceseismic
at the
loading.
Totallocation
vertical isground
settlements
up to 350
mmthe
duedesign
to the
clinker silo
potentially
liquefiable
under
dissipation
of excess
porevertical
pressures,
and lateral
groundup
displacements
seismic loading.
Total
ground
settlements
to 350 mm
duethe
to the
dissipation
excesspredicted
pore pressures,
and lateral
ground
in
order
of 0.5 mof were
under design
earthquake
displacements
in capacity
the orderrequirements
of 0.5 m were
design
loading.
Bearing
underpredicted
both staticunder
and seismic
earthquake
loading.
Bearing
capacity
requirements
under both
static
loading
could
be best
satisfied
if ground
improvement
measures
and seismic
loading
bewas
bestthen
satisfied
if ground
improvement
were
undertaken
and could
the silo
founded
within the
improved
measuresdeposits.
were undertaken
the silo was
then founded
within
the
alluvial
Variousand
foundation
options
including
timber,
improved and
alluvial
Various
foundation
options
including
concrete,
steel deposits.
piling as well
as shallow
foundation
options
with
timber, improvement,
concrete, andwere
steelconsidered.
piling as well
foundationa
ground
With as
the shallow
piling alternative,
optionsof
with
ground
improvement,
werepiles
considered.
Withto
theachieve
piling
pattern
closely
spaced
lower capacity
were needed
alternative,
a
pattern
of
closely
spaced
lower
capacity
piles
were
the required densification. If piles were to be used, it was also noted
needed
to
achieve
the
required
densification.
If
piles
were
to
that there might be a requirement for installation of vertical drains be
to
used, it was also noted that there might be a requirement for
relieve excess pore water pressures under earthquake loading.
installation of vertical drains to relieve excess pore water pressures
under earthquake loading.

3rd Proff 18-02-2015

GeotechnicalJournal
JournalVol.
Vol.6 6No.
No.1 12014
2014
Geotechnical

Geotechnical Journal Vol. 6 No. 1 2014 47

Fig.7.7.Plan
Planlayout
layoutofofstructures
structures
Fig.
Ground improvement using vibro-replacement was selected as
Ground
improvement
using
vibro-replacement
wasselected
selected
the
the
most
suitable option,
because
drainage iswas
also
improved
with
Ground
improvement
using
vibro-replacement
asasthe
mostsuitable
suitableoption,
option,
because
drainage
also
improved
with
this
this
technique.
It was
determined
thatisis
the
upper
deltaicwith
deposits,
most
because
drainage
also
improved
this
technique.could
wasbe
determined
thatthe
the
upper
deltaicdeposits,
deposits,
however,
not
effectivelythat
treated
using
thedeltaic
method
of
vibrotechnique.
ItItwas
determined
upper
however,could
could
not
effectively
treated
using
themethod
method
vibroreplacement.
As
such,
the
silty soils
within
the stress
influence
zone
of
however,
not
bebeeffectively
treated
using
the
ofofvibroreplacement.
As
such,
the
silty
soils
within
the
stress
influence
zone
replacement.
such, theabout
silty 2soils
withinthe
thefoundation
stress influence
zone
the structure As
(extending
m beyond
perimeter
to
thestructure
structure
(extending
about
beyond
the
foundation
ofaofdepth
the
about
2 2mmlevel)
beyond
the
foundation
of
about 5(extending
m
below the
foundation
were
sub-excavated
perimeter
depthof
ofabout
about
belowthe
the
foundation
level)
were
perimeter
totoa adepth
5 5mmbelow
foundation
level)
were
prior
to performing
ground
densification.
Because
of the
relatively
sub-excavated
prior
to
performing
ground
densification.
Because
sub-excavated
prior
to
performing
ground
densification.
Because
ofof
shallow groundwater level, the excavation work was performed with
therelatively
relativelyshallow
shallow
groundwater
level,the
theexcavation
excavation
work
was
the
groundwater
level,
work
was
dewatering.
The
excavation
was backfilled
using
locally
available
performedwith
withdewatering.
dewatering.
Theexcavation
excavationwas
was
backfilled
using
performed
The
backfilled
using
relatively
coarse
clean sand. Vibro-replacement
was
then undertaken
locallyavailable
availablerelatively
relativelycoarse
coarseclean
cleansand.
sand. Vibro-replacement
Vibro-replacement
locally
to treat a zone extending horizontally about 10 m beyond the perimeter
wasthen
thenundertaken
undertakentototreat
treata azone
zoneextending
extendinghorizontally
horizontally about1010
was
of the
clinker silo foundation
and also
extending vertically about
down to a
mbeyond
beyondthe
theperimeter
perimeterofofthe
theclinker
clinkersilo
silofoundation
foundation andalso
also
mdepth
of 18 m below the existing ground surface as shownand
in Figure
extendingvertically
verticallydown
downtotoa adepth
depthofof1818mmbelow
belowthe
theexisting
existing
extending
8.
The surface
vibro-replacement
would
densify
the upper portion
of
ground
shownininFigure
Figure8.not
8.The
The
vibro-replacement
would
ground
surface asasshown
vibro-replacement
would
the
sand
fill
due
to
lack
of
confinement
and
groundwater.
As
such,
notdensify
densifythe
theupper
upperportion
portionofofthe
thesand
sandfill
filldue
duetotolack
lackofof
not
about
1 m ofand
the groundwater.
sand fill was removed
upon
of vibro
confinement
such,about
about1completion
1mmofofthe
thesand
sand
fill
confinement
and groundwater.
AsAssuch,
fill

replacement work, and the excavation base was raised to the design
wasremoved
removedupon
uponcompletion
completionofofvibro
vibroreplacement
replacementwork,
work,and
andthe
the
was
slab
underside
elevation
usingto75the
minus
well-graded
crushed
rockfill
excavation
base
wasraised
raised
design
slabunderside
underside
elevation
excavation
base
was
to the design
slab
elevation
compacted
to 100%
of standard
Proctor
maximum
dry density.
using7575minus
minus
well-graded
crushed
rockfill
compacted
100%The
using
well-graded
crushed
rockfill
compacted
toto100%
ofof
1.5-m
thick
raft
for
the
clinker
silo
was
founded
on
the
compacted
standardProctor
Proctormaximum
maximumdry
drydensity.
density.The
The1.5-m
1.5-mthick
thickraft
raftfor
forthe
the
standard
rockfill
as
prepared
above (see
Fig.
8compacted
for a schematic
cross-section).
clinkersilo
silo
wasfounded
founded
thecompacted
rockfill
prepared
clinker
was
ononthe
rockfill
asasprepared
above
(seeFig.
Fig.
fora aschematic
schematiccross-section).
cross-section).
above
(see
8 8for
4.3
Cement
Silo
4.3 Cement
Cement
Silo
4.3
Silo
The
cement
silo
was constructed immediately south west of the new
clinker silo (see Fig. 7). The silo was constructed on a relatively
Thecement
cement
silo
was
constructed
immediately
south
westofofthe
new
The
silo
was
constructed
immediately
south
west
new
stiff
circular
raft
foundation
25 m
in diameter,
to withstand
athe
design
clinkersilo
silo
(seeFig.
Fig.
The
silowas
was
constructed
relatively
clinker
(see
7).7). The
silo
constructed
ononstatic
a arelatively
average
maximum
bearing
pressure
of 300
kPa under
loading
stiffcircular
circularraft
raftfoundation
foundation
25mthe
mininraft
diameter,
withstand
design
stiff
diameter,
totowithstand
a adesign
conditions.
The
underside25of
foundation
was located
at a
average
maximum
bearing
pressure
of
300
kPa
under
static
loading
average
maximum
bearing
pressure
of
300
kPa
under
static
loading
depth of about 3 m below the ground surface.
conditions. The
Theunderside
undersideofofthe
theraft
raftfoundation
foundationwas
waslocated
locatedatata a
conditions.
The
conditions
at
thethe
cement
silosurface.
are
similar to those at the clinker
depthsoil
about3 3mmbelow
below
theground
ground
surface.
depth
ofofabout
silo with the estimated potentially liquefiable soil zone
Thesoil
soilconditions
conditionsatatthe
thecement
cementsilo
siloare
aresimilar
similartotothose
thoseatatthe
the
The
clinkersilo
silowith
withthe
theestimated
estimatedpotentially
potentiallyliquefiable
liquefiablesoil
soilzone
zone
clinker

ZONE OF VIBRO-REPLACEMENT

47
Fig.8.8.Foundation
Foundationsystem:
system:clinker
clinkersilo
silo Section
SectionY-Y
Y-Y(see
(seeFig.
Fig.6-1
6-1
forsection
sectionlocation).
location).
47
Fig.
for

Geotechnical Journal Vol. 6 No. 1 2014

48 Geotechnical Journal Vol. 6 No. 1 2014


extending to a depth of about 20 m below the ground surface. Unlike
at the clinker silo, partial excavation of the weak upper silty soils
extending
depth
20 msubgrade
below the
surface.
and
supporttoofa the
raft of
on about
a densified
wasground
not considered
Unlike atfor
thetheclinker
silo,
excavation
theclose
weakproximity
upper silty
feasible
cement
silopartial
foundation
due toofthe
of
soils andexisting
supportfacilities
of theasraft
subgrade
wasuplift
not
adjacent
wellonas athedensified
need to resist
potential
considered loads
feasible
forthe
the
cement
siloreason,
foundation
to the
close
foundation
from
silo.
For this
it wasdue
decided
that
the
proximity
adjacent
facilities
as well
the need to
resist
cement
siloofraft
would existing
be supported
on piles.
Twoasalternative
options
potential
uplift foundation
fromthe
thecompact
silo. For
it
were
considered:
(a) install loads
piles into
to this
densereason,
alluvial
was decided that the cement silo raft would be supported on piles.
soils at depths below 21 m; (b) install relatively short piles 10 to 12
Two alternative options were considered: (a) install piles into the
m, with densification of the loose subsoils below the pile tip levels
compact to dense alluvial soils at depths below 21 m; (b) install
to a depth of 21 m prior to pile driving (see Zone A in Fig. 9). In
relatively short piles 10 to 12 m, with densification of the loose
both
options,
thethe
piles
would
be installed
at a relatively
subsoils
below
pile
tip levels
to a depth
of 21 m close
priorspacing
to pile
so
that the
soilsinbetween
pilesoptions,
would be
driving
(seelooser
Zone A
Fig. 9). the
In both
thecompacted,
piles wouldand
be
an
annular
of ground
treatment
the looser
foundation
installed
at zone
a relatively
close
spacing around
so that the
soils footprint
between
would
be undertaken
mitigate effects
of liquefaction
(similar
to the
the piles
would be to
compacted,
and an
annular zone
of ground
circumferential
ground
treatment atfootprint
the clinker
silo).be Option
(b) was
treatment around
the foundation
would
undertaken
to
identified
as the more
cost-effective(similar
and preferred
design.
mitigate effects
of liquefaction
to the
circumferential
ground
treatment at
thetheclinker
(b)for
was
as
It
was determined
that
criticalsilo).
lateralOption
loading
theidentified
cement silo
the more cost-effective and preferred design.
would occur under earthquake shaking conditions. The response of
the silo foundation under lateral loading was analysed accounting for
It wassoil-pile
determined
that group
the critical
lateraleffects
loadingtofor
the cement
silo
both
and pile
interaction
provide
necessary
would
occur
under
earthquake
shaking
conditions.
The
response
of
input for final structural design. It was specified in the construction
the silo foundation under lateral loading was analysed accounting
contract that the lower Zone A shown in Fig. 9 be densified with
for both soil-pile and pile group interaction effects to provide
suitable provisions to facilitate subsequent installation of expandednecessary input for final structural design. It was specified in the
base piles through the upper Zone B (i.e., the Zone A to be treated
construction contract that the lower Zone A shown in Fig. 9 be
without
densifying
the upper to
Zonefacilitate
B). A total
of 123
densifiedexcessively
with suitable
provisions
subsequent
expanded-base
piles with shaftpiles
diameter
of 508
andZone
with B
a design
installation of expanded-base
through
themm
upper
(i.e.,
compression
of 1070
kN/pile
were specified.
Thetheannular
the Zone A capacity
to be treated
without
excessively
densifying
upper
zone
soilAwithin
m expanded-base
outside the footprint
of the
Zone of
B).
total of10123
piles with
shaftfoundation
diameter
extending
a depth
21 m below
the existing
ground
surface
was
of 508 mmtoand
with of
a design
compression
capacity
of 1070
kN/pile
densified
using vibro-replacement
Zone10
C).
were specified.
The annular zone(see
of Fig.
soil 9,
within
m outside the
footprint of the foundation extending to a depth of 21 m below the
existing ground surface was densified using vibro-replacement (see
Fig. 9, Zone C).

After installing some 30 expanded-base piles, it was determined that


the construction progress was slower than anticipated; this appeared
After
30 expanded-base
piles,
it was determined
that
to
be ainstalling
result ofsome
not carefully
limiting the
densification
of the upper
the construction
was slower
thanlower
anticipated;
this
appeared
Zone
B during progress
the densification
of the
Zone A
prior
to the
to beinstallation
a result of not
carefully
limiting
the upper
pile
process.
Because
of the
thisdensification
difficulty, theofcontractor
Zone B the
during
the densification
the lower
Zone A prior
the
proposed
installation
of 508 mmofdiameter
(open-ended)
steeltopiles
pile installation
process.expanded-base
Because of this
difficulty,
the contractor
instead
of the specified
piles.
This alternative
steel
proposed
of the
508design
mm diameter
steel
pile
optionthe
wasinstallation
accepted (by
team and (open-ended)
the owner) subject
piles instead of the specified expanded-base piles. This alternative
to contractor achieving the required vertical capacity and lateral fixity
steel pile option was accepted (by the design team and the owner)
as per original design. As a result, the remaining 90 pile locations for
subject to contractor achieving the required vertical capacity and
the cement silo foundation was completed using steel pipe piles; pile
lateral fixity as per original design. As a result, the remaining 90
driving
analyser
testingsilo
wasfoundation
conductedwas
on completed
selected piles
to
pile locations
for(PDA)
the cement
using
confirm
thepiles;
axialpile
capacities
steel pipe
drivingachieved.
analyser (PDA) testing was conducted
on selected
piles to confirm the axial capacities achieved.
4.4
Kiln Line
4.4 new
Kilnkiln
Lineline was to be constructed immediately north of an
The
existing
kiln
located
immediately
northimmediately
of clinker silo.
Theofnew
The new kilnline
line
was to
be constructed
north
an
kiln
line
comprises
several
units:
raw
grinding
mill,
homogenizing
existing kiln line located immediately north of clinker silo. The new
silo,
pre-heater/coal
kilnunits:
piers,raw
andgrinding
cooler. The
prominent
kiln line
comprises mill,
several
mill,most
homogenizing
feature
of the kiln line mill,
is the pre-heater
has a height
the
silo, pre-heater/coal
kiln piers,tower
and that
cooler.
The inmost
order
of 115feature
m. of the kiln line is the pre-heater tower that has a
prominent
height in
the orderare
of 115
m. aligned from east to west extending
These
structures
linearly
over
distance ofareabout
150 aligned
m. Thefrom
kiln east
pier to
No.west
2, pre-heater,
Thesea structures
linearly
extending
and
silo 150
wasm.constructed
on No.
a pile-supported
raft
over homogenizing
a distance of about
The kiln pier
2, pre-heater, and
foundation
(~ 38silo
m xwas
70 m),
and the cooler
and the kiln
homogenizing
constructed
on a building
pile-supported
raft
foundation
(~ 38supported
m x 70 m),
the cooler
building
the kiln
pier
No. 1 was
on aand
similar
but separate
raftand
foundation
pier23No.
supported
on a similar
butrafts
separate
raft foundation
(~
(~
m 1x was
55m).
The underside
of the
was located
at a depth
23 2.7
m xm55m).
underside
of the rafts
was grinding
located atmill
a depth
of
of
below The
the ground
surface.
The raw
and its
2.7 m below
ground surface.
Thepile
raw foundations,
grinding millwith
and the
its
building
werethe
supported
on separate
building pile
were
supported
onlocations,
separate tied
piletogether
foundations,
with
the
building
caps,
at column
by grade
beams
building
pile caps, at column
tied
together
by grade
and
a suspended-grade
slab. locations,
The design
bearing
pressures
forbeams
each
and a suspended-grade
slab. The design
bearing
pressures for each
foundation
footprint is summarized
in Table
1.
foundation footprint is summarized in Table 1.

Approx. 21m

10 m

10 m

ZONE B

VIBROREPLACEMENT
ZONE C

VIBROREPLACEMENT
ZONE C
VIBRO-REPLACEMENT
ZONE A

48
Fig. 9. Foundation system - cement silo Section Z--Z (see Fig. 6-1 for section location).

Geotechnical
Journal
Vol.- 6Kiln
No.Line
1 2014
Table
1. Foundation
details

3rd Proff 18-02-2015


Kiln Line Structure

Gross Bearing Pressure (kPa)

Cooler
80
Table
1.
Foundation
details
Kiln Pier # 1
280 - Kiln Line
Kiln
Line
Kiln
Pier
# 2Structure

Gross Bearing Pressure (kPa)


280

Cooler
Pre-heater/Coal
Mill

80
115

Kiln Silo
Pier # 1
Homo

280
240

Kiln
Pier # 2
Raw
Mill

65280

Pre-heater/Coal Mill

115

upper 10 to 12 m, (i.e. approximately above the pile base levels)


have a high risk of liquefaction and it would lead to free-field
Geotechnical
Journal
6 No. 1 2014
49
lateral ground movements
in the order
of 0.5 Vol.
m. Detailed
soil-pile
interaction analyses indicated that expanded-base piles would be
subjected
to unacceptably
largeground
bending
moments under
these
bending moments
under these
displacements.
Reduction
upper 10 to 12 m, (i.e.
approximately
above the
pile
base
levels)
ground
Reduction
of the
potential
forwas
lateral
ground
of the displacements.
potential for lateral
ground
displacements
considered
have
a
high
risk
of
liquefaction
and
it
would
lead
to
free-field
displacements
was considered
essential,
and installation
ground
essential, and installation
of ground
improvement
barriers of
around
the
lateral ground
movements
in the
the order
of 0.5
m.
improvement
barriers
around
footprint
of most
the Detailed
kiln
linesoil-pile
was
footprint
of
the
kiln
line
was
judged
to
be
the
effective
form
interaction
analyseseffective
indicated
thatofexpanded-base
piles would
judged
to be the
form
achieving
this objective.
Asbe
ofsubjected
achieving
thismost
objective. Aslarge
a result,
groundmoments
improvement
using
to
unacceptably
bending
under
these
avibro-replacement
result, ground improvement
using
vibro-replacement
(installation
(installation
of stone
columns)
undertaken
ground
displacements.
Reduction
of the
potential was
for outside
lateral
ground
of
stone
columns)
was
undertaken
immediately
the
immediately
outside
the foundation
footprints
extending ofover
a
displacements
was considered
essential,
and installation
foundation
footprints
over
a horizontal
distance
of ground
6 the
m
horizontal
distance
ofextending
6m
and to
a depth
of about
12 m
below
improvement
barriers
around
the
footprint
of
the
kiln
line
was
and
to asurface
depth of
about
12 m below the ground surface (see Figure
ground
(see
Figure
to be the
most
effective form of achieving this objective. As
10judged
for a schematic
cross-section).
a result, ground improvement using vibro-replacement (installation
of stone columns) was undertaken immediately outside the
foundation footprints extending over a horizontal distance of 6 m
and to a depth of about 12 m below the ground surface (see Figure
10 for a schematic cross-section).

Homo
Silo
240
It was
identified
that the performance
of the kiln line is sensitive to
post-construction
settlements,
Raw Mill
65 especially differential settlements
between adjacent foundations and across the larger foundations.
The allowable differential settlements are quite limited because of
the
requirement
for
tolerances onof tilt.
In addition
to the
It Itwas
thatstrict
the
the
kiln
line isis sensitive
sensitive
wasidentified
identified
the performance
performance
of strata,
the kiln
line
expected
settlementsthat
within
the upper
soil
settlements
due to to
topost-construction
post-construction
settlements,
especially
differential
settlements
especially
differential
settlements
the compressibility ofsettlements,
the deep marine
deposit (located
below
~30 m
between
adjacent foundations
and
the
between
andacross
acrossThe
thelarger
largerfoundations.
foundations.
depth)
wereadjacent
also a foundations
key consideration.
anticipated
lateral
The
differential
settlements
are quite limited
limited because
becauseof
Theallowable
allowable
differential
settlements
loadings
from wind
and earthquake
shaking quite
on the tall pre-heater
ofthe
therequirement
requirement for strict
strict tolerances on tilt.
tilt. InIn addition
additiontotothethe
tower
were the otherfor
critical concerns.
expectedsettlements
settlementswithin
withinthe
theupper
uppersoil
soilstrata,
strata,settlements
settlementsdue
duetoto
expected
In
consideration
of of
the
sensitivities
todeposit
differential
settlements
and
the
compressibility
ofthe
the
deepmarine
marine
deposit(located
(located
below~30
~30
the
compressibility
deep
below
mm
the
lateral
loadings,
was
that
it The
wouldanticipated
be prudent
to
depth)
were
key concluded
consideration.
lateral
depth)
were
alsoalso
a keyitaconsideration.
The
anticipated
lateral
loadings
support
all the
kiln
lineand
structures
ononpiled
foundations.
Expandedloadings
from
wind
earthquake
shaking
on the tall
pre-heater
from
wind
and
earthquake
shaking
the
tall
pre-heater
tower
were
base
cast-in-place
concrete
piles
or closed-end steel pipe piles were
tower
were
the concerns.
other
critical
concerns.
the
other
critical
considered most suitable for the support of the kiln line structures.
Inconsideration
consideration
ofexisting
the
sensitivities
differential
settlements
and
sensitivities
to differential
and the
ItInwas
noted that of
thethe
kiln line
attothe
plantsettlements
is also
supported
the
lateral
loadings,
it
was
concluded
that
it
would
be
prudent
lateral
loadings, it was
concluded
be prudent
to supportto
on
expanded-base
piles;
becausethatofit would
this valuable
site-specific
Fig.10. Foundation system kiln line
support
all
the
kiln
line
structures
on
piled
foundations.
ExpandedFig. 10. Foundation system - kiln line
all
the
kiln
line
structures
on
piled
foundations.
Expanded-base
castexperience combined with the assessed advantage in terms
of
base cast-in-place
concrete
orsteel
closed-end
steel
pipe
piles were 5 CONCLUSIONS
in-place
concrete
piles
or closed-end
pipe piles
were
considered
potential
cost savings,
the
usepiles
of expanded-base
piles
were
selected
considered
most
suitable
for
the
support
of
the
kiln
line
structures.
mostthe
suitable
for the
support
of the
kiln linethat
structures.
It was
noted
over
steel pile
option.
It was
estimated
an allowable
vertical
It was
notedload
that
the
existing
kilntoline
at the
also
supported
Liquefaction is one of the main causes of geotechnical hazards to
geotechnical
capacity
of up
kNplant
per ispile
could
be
that
the existing
kiln
line
at the
plant
is1070
also
supported
on
expandedon
expanded-base
piles;
because
of
this
valuable
site-specific
facilities
located
in seismically
Fig.10.
Foundation
systemactive
kilnregions.
line The options to improve
allowed
on because
a 508-mm
shaft
diameter
expanded-base
pile combined
installed
base piles;
of this
valuable
site-specific
experience
10. Foundation system - kiln line
5the Fig.
CONCLUSIONS
experience
combined
with
the
assessed
advantage
in
terms
of
seismic
performance
of
a
given
structure
against liquefactionwith
its base
at aadvantage
depth of inabout
12 m cost
below
the existing
with the
assessed
terms10
of to
potential
savings,
the use
potential
cost savings, the use of expanded-base piles were selected induced geotechnical hazard are: (a) relocate to avoid the hazard; (b)
ground
surface.
of expanded-base piles were selected over the steel pile option. It was Liquefaction is one of the main causes of geotechnical hazards to
over the steel pile option. It was estimated that an allowable vertical isolate from the hazard; (c) tolerate the hazard (by strengthening); and
estimated
that an
allowable
vertical
of upbe facilities located in seismically active regions. The options to
Based
on these
design
considerations,
the1070
kiln line
foundations
were
geotechnical
load
capacity
of upgeotechnical
to
kNload
percapacity
pile could
(d) eliminate
hazardperformance
(using ground
With against
respect
improve
the the
seismic
of improvement).
a given structure
to
1070
kN
per
pile
could
be
allowed
on
a
508-mm
shaft
diameter
supported
using
over 600
expanded-base
piles (shaftpile
diameter
allowed on
a 508-mm
shaft
diameter expanded-base
installed to the elimination of the hazard, ground improvement is emerging
liquefaction-induced
geotechnical
hazard
are:
(a)
relocate
to
avoid
5
CONCLUSIONS
expanded-base
pile
with
itswith
base
at a12
depth
of about
to 12
varying
from
406
to 559
mm)
horizontal
with its
base
at mm
ainstalled
depth
of
about
10 centre-to-centre
to
m below
the10existing
as one
of the
measures
the
the
hazard;
(b)widely
isolate adopted
from themitigative
hazard; (c)
tolerate totheminimize
hazard (by
m
below
the
existing
ground
surface.
pile
spacing
varying from 2.0 m to 2.6 m. Pile spacing was chosen
ground
surface.
Liquefaction
isand
oneCommonly
of the
main
causes
geotechnical
hazards
risk
of liquefaction.
used
ground
improvement
strengthening);
(d)
eliminate
the of
hazard
(using measures
ground to
to
avoid significant
reduction
in pile group
efficiency,
and to
facilities
located
in respect
seismically
active
regions.
The
options
Based
these design
design
considerations,
thekiln
kiln
foundations
include
dynamic
compaction,
vibro-replacement
usingofstone
With
to the
elimination
the columns,
hazard, to
Basedon
on
considerations,
the
lineline
foundations
were improvement).
minimize
thethese
riskusing
of damage
to adjacent
piles
during
installation.
improve
the
seismic
performance
of
a
given
structure
against
were
supported
over
600
expanded-base
piles
(shaft
diameter
compaction
piling,
blast
densification,
and
compaction
grouting.
supported using over 600 expanded-base piles (shaft diameter ground improvement is emerging as one of the widely adopted
liquefaction-induced
to avoid
from
mm
horizontal
mitigative
measures geotechnical
to minimizehazard
the are:
risk (a)ofrelocate
liquefaction.
Avarying
pile-supported
system
themm)
kilnwith
line
essentially
eliminated
the
varying
from406
406
mmtofor
to559
559
mm)
withcentre-to-centre
centre-to-centre
horizontal
Using
a number
of case
histories
from (c)
thetolerate
GreatertheVancouver
the hazard;
(b) ground
isolate
from
the hazard;
hazard (by
pile
varying
from
2.0
m to
2.6
m. m.
Pile
spacing
was was
chosen
to Commonly
used
improvement
measures
include dynamic
settlement
concerns
arising
from
the
layers.
However,
itchosen
was
pilespacing
spacing
varying
from
2.0
m
toupper
2.6
Pile
spacing
Region
of British Columbia,
located in
a moderate
strengthening);
and (d) Canada,
eliminatewhich
the ishazard
(using
ground
avoid
significant
in pile
group
efficiency,
and to
minimize
recognized
that reduction
settlements,
below
the
pile tips,
within
the to compaction, vibro-replacement using stone columns, compaction
to avoid
significant
reduction
in pile
group
efficiency,
and
improvement).
With
respect
to
the
elimination
of
the
hazard,
to
high
seismic
risk
region,
several
engineering
considerations
in
the
risk of damage
toofadjacent
duringdeposits
installation.
piling, blast densification, and compaction grouting.
occasional
looser
of thepiles
would
still be a
minimize
the riskzones
damage
toalluvial
adjacent
piles
during
installation.
ground toimprovement
is emerging
as
one ofliquefaction-induced
the widely adopted
relation
ground
improvement
to
mitigate
concern. Since these looser zones within the alluvial deposits would
A pile-supported
system forforthethekiln
essentially eliminated
the Using
a number
of casetohistories
fromthethe risk
Greater
mitigative
measures
minimize
of Vancouver
liquefaction.
pile-supported
kilnline
eliminated
notA be
far below system
the piled zone,
it line
wasessentially
considered
that anythe geotechnical hazards are illustrated. The case histories correspond
Region
of British
Canada, out
whichusing
is located
in a moderate
Commonly
usedColumbia,
ground carried
improvement
measures
include
dynamic
settlement
concerns
arising
from
the
upper
layers.
However,
it
was
settlement
concerns
arising
from
the
upper
layers.
However,
it
was
to
ground
improvement
vibro-replacement,
settlements arising within these zones could potentially translate into
to compaction,
high seismicvibro-replacement
risk region, several
engineering
considerations
in
using
stone columns,
compaction
recognized
settlements,
below
pile
within
recognized
thatthat
settlements,
the line.
pilethe
tips,
withintips,
the occasional
differential
settlements
alongbelow
the kiln
In consideration
of this,the preloading, piling, and combinations of one or more of the above
relation
ground
improvement
to and/or
mitigate
liquefaction-induced
piling, to
blast
densification,
and compaction
grouting.
occasional
looser
zones
of
the
alluvial
deposits
would
still
be
a
looser
zones
of
the
alluvial
deposits
would
still
be
a
concern.
Since
methods.
The
protected
foundations
structures
belong
to
key
preloading of the site, up to heights in the order of 6 m, was
geotechnical
hazards
are illustrated.
The systems.
case histories
concern.
zones
within
thewould
alluvial
would
these
looserSince
zones
within
the alluvial
deposits
notdeposits
beatfarleast
below
industrial
highways,
and
pipeline
Somecorrespond
ofVancouver
the key
recommended,
andthese
this looser
work
was undertaken
to
remove
a
Using
aplants,
number
of casecarried
histories
from
thevibro-replacement,
Greater
to
ground
improvement
out
using
not
be
far
below
the
piled
zone,
it
was
considered
that
any
the piled
it was considereddifferential
that any settlements
observations/findings
are
as
follows:
portion
ofzone,
the post-construction
settlementsarising
due towithin
local
Region of British Columbia, Canada, which is located in a moderate
settlements
arising
within these
zones could
potentially settlements
translate into preloading, piling, and combinations of one or more of the above
these
zones
could
potentially
translate
into differential
variations
in soil
conditions.
to high The
seismic
risk
region,
several
considerations
1.
In
seismic
mitigation
works,
theengineering
design
philosophy
methods.
protected
foundations
and/or
structures
belongrevolves
to key in
differential
settlements
along
the
kiln
line.
In
consideration
of
this,
along the kiln line. In consideration of this, preloading of the site,
relation
to
ground
improvement
to
mitigate
liquefaction-induced
around displacement-based
plants, highways, design.
and pipeline systems. Some of the key
Itupwas
acceptedof
slabs
and
pileand
caps,
well
preloading
thebelow-ground
site, of
up6 to
inwalls,
the order
of this
6asm,
was industrial
to heights
inthat
the
order
m, heights
was recommended,
work
geotechnical hazards are illustrated. The case histories correspond
as
piles
would provide
thework
passive
resistance
against
recommended,
this
was
undertaken
remove at
least a observations/findings are as follows:
was
undertaken
toand
remove
atrequired
least
a portion
ofsoil
theto
post-construction
2. to ground
The improvement
ground improvement
commonly
carried outconfigurations
using vibro-replacement,
lateral
loads
primarily
arising
due differential
tovariations
wind andin
earthquake
loading.
portion
ofsettlements
the
post-construction
settlements
due
to local
differential
due to local
soil conditions.
1.
Inin seismic
mitigation
works,
the design
used
practice
include:
(i)
in-ground
densified
aligned
preloading,
piling,
and combinations
of
one philosophy
or barrier(s)
more of revolves
the
above
It variations
was noted
liquefaction-induced horizontal ground
in soil that
conditions.
around displacement-based
design.
methods.
The
protected
foundations
and/or
structures
belong to key
perpendicular
to
the
direction
of
ground
movement;
(ii)
densification
It was acceptedare
that
slabs
and walls,
caps,
well
displacements
thebelow-ground
key controlling
parameter
inpile
terms
of as
lateral
The ground
improvement
configurations
commonly
used
plants,
highways,
and
pipeline
systems.
Somefootprints
of
thein
key
It was
accepted that
slabsSeismic
and walls,
pile caps,
as well 2.
ofindustrial
wide-area
footprints
beneath
and/or
around
foundation
load
performance
of below-ground
thethefoundation.
ground
response
as piles
would provide
required passive
soil resistance
against
practicebearing
include:capacity
(i) as
in-ground
densified
barrier(s)
aligned
observations/findings
are
follows:
as pilesindicated
would provide
theloose
required
passive
soil
resistance
against
to
improve
failures
and
the
impacts
from
lateral
analyses
that
the
Fraser
River
sands,
within
the
lateral loads primarily arising due to wind and earthquake loading. It
perpendicular to the direction of ground movement; (ii)
lateral loads primarily arising due to wind and earthquake loading. spreading.
was noted that liquefaction-induced horizontal ground displacements
1. In seismic mitigation works, the design philosophy revolves
It was noted that liquefaction-induced horizontal ground
49
around
design.
are
the key controlling
in termsparameter
of lateral load
performance
The displacement-based
selection of the most
suitable ground improvement
displacements
are theparameter
key controlling
in terms
of lateral 3.
2. The
ground improvement
configurations
used in
ofload
the performance
foundation. of
Seismic
ground
response
analyses
indicated
option
is
governed
by
many
factors
including:commonly
soil conditions,
the foundation. Seismic ground response
practice
include:
(i)
in-ground
densified
barrier(s)
aligned
that
the
loose
Fraser
River
sands,
within
the
upper
10
to
12
m,
analyses indicated that the loose Fraser River sands, within the equipment/space restrictions, issues related to the protection
perpendicular
the direction
ground movement;
(ii)
(i.e. approximately above the pile base levels) have a high risk of
of existing
structuresto during
ground of
improvement,
operational
liquefaction and it would lead to free-field lateral ground movements 49constraints, environmental regulatory requirements, and land
in the order of 0.5 m. Detailed soil-pile interaction analyses indicated
availability.
that expanded-base piles would be subjected to unacceptably large

50 Geotechnical Journal Vol. 6 No. 1 2014


4. The method of vibro-replacement using stone columns can be
effectively used to densify coarse-grained soils within about 25 m
below existing ground level. The method has also become attractive
because of the potential availability of drainage through stone
columns for the dissipation of excess pore water pressures in addition
to the densification effect.
5. Heavily-loaded structures can be founded on compressible deposits
by careful selection of site treatments such as preloading, piling, and
vibro-replacement. In addition to reducing the compressibility of
soils, such treatments can be used to effectively reduce the liquefaction
susceptibility of upper loose soils under seismic loading. .
6 ACKNOWLEDGEMENTS
The author is grateful to the BC Ministry of Transportation, Terasen
Gas Utility Ltd., Lafarge Canada Inc., and Golder Associates Ltd. for
granting permission to publish the technical information associated
with the case histories. The authors involvement in the projects
presented herein took place during his tenure at Golder Associates
Ltd. between 1990 and 2000. The important contributions to the
engineering projects cited herein made by the authors former
colleagues Mr. Trevor Fitzell, P.Eng. and Dr. Upul Atukorala, P.Eng.
of Golder Associates Ltd., as project team members, are respectfully
acknowledged.
7 REFERENCES
Bartlett, S.F. & Youd, T.L. (1992). Empirical analysis of horizontal
ground displacement generated by liquefaction-induced lateral
spreads. Technical Report NCEER-92-0021, National Center for
Earthquake Engineering Research, State University of New York,
Buffalo, N.Y.
Byrne, P.M., Jitno, H. & Salgado, F. (1992). Earthquake induced
displacements of soil-structures systems. Proceedings 10th World
Conference, Earthquake Engineering, 19-24 July. Madrid, Spain,
A.A. Balkema, Rotterdam, pp. 1407-1411.
FLAC. (1998). Fast Lagrangian Anlysis of Continua. Itasca
Consulting, Ithaca, Minnesota, USA.
Harder, L.S. & Seed, H.B. (1986). Determination of penetration
resistance for coarse-grained soils using the Becker hammer drill.
Earthquake Engineering Research Center Report No. UCB/EERC86/06, University of California, Berkeley, California, U.S.A.
Hausler, E.A. & Koelling, M. (2004). Performance of improved
ground during the 2001 Nisqually Earthquake. Proceedings of the
5th International Conference on Case Histories in Geotechnical
Engineering, Paper No 3.27.
Hausler, E.A. & Sitar, N. (2001). Performance of soil improvement
techniques in earthquakes. Proceedings of the 4th International
Conference on Recent Advances in Geotechnical Earthquake
Engineering and Soil Dynamics, Paper No. 10.15.
Houston, S. L., Houston, W. N. & Padilla J. M. (1987).
Microcomputer-aided evaluation of earthquake induced permanent
slope displacements. Microcomputers in Civil Engineering 2,
Elsevier Science Publishing Co. pp. 207-222.
Japanese Geotechnical Society (eds.). (1998). Remedial measures
against soil liquefaction: From investigation and design to
implementation, A.A. Balkema, Rotterdam.
MCEER. (1999). Proceedings of the Seventh U.S.-Japan Workshop
on Earthquake Resistant Design of Lifeline Facilities and
Countermeasures for Soil Liquefaction, San Francisco, California,
Technical Report MCEER-99-0019, Multidisciplinary Center for
Earthquake Engineering Research, State University of New York,
Buffalo, N.Y.
Mitchell, J.K. (1981). Soil improvement - State-of-the art report.

Proceedings of the 10th International Conference on Soil Mechanics


and Foundation Engineering, Stockholm, Vol. 4, pp. 509-565.
Mitchell, J.K., Cooke, H.G. & Schaeffer, J.A. (1998). Design
considerations in ground improvement for seismic risk mitigation.
Geotechnical Special Publication No. 75, Geotechnical Earthquake
Engineering and Soil Dynamics III, Specialty Conference, American
Society of Civil Engineers, pp. 580-613.
Newmark, N.M. (1965). Effects of earthquake on dams and
embankments - Rankine Lecture. Gotechnique, Vol. 15, No. 2, pp.
139-160.
ORourke, T.D. & Hamada, M.(eds.). (1992). Case studies of
liquefaction and lifelines performance during past earthquakes:
United States case studies. Technical Report NCEER-92-0002, Vol.
2, National Centre for Earthquake Engineering Research, State
University of New York, Buffalo, N.Y.
Schnabel, P.B., Lysmer, J. & Seed, H.B. (1972). SHAKE: A computer
program for earthquake response analysis of horizontally layered
sites. University of California, Berkeley, California, Earthquake
Engineering Research Center, Report No. EERC 72-12.
Seed, H.B., Tokimatsu, K., Harder, L.F. & Chung, R.M. (1985). The
influence of SPT procedures in soil liquefaction resistance evaluations.
J. Geotech. Eng., ASCE, Vol. 111, No. 12, pp. 1425-1445.
Wijewickreme, D., Fitzell, T.P. & Smith, D. (2002). Foundation
Systems for the Modernization of Lafarge Canada Richmond
Cement Plant - Geotechnical Engineering Considerations, Proc. 55th
Canadian Geotechnical Conference, Niagara Falls, Ontario, October
2002, Paper No. G131.
Wijewickreme, D., Honegger, D.G., Mitchell, A., & Fitzell, T.P.
(2005). Seismic Vulnerability Assessment and Retrofit of a Major
Natural Gas Transmission System: A Case History, Earthquake
Spectra, Journal of the Earthquake Engineering Research Institute
(EERI), Vol. 21, No. 2: 539-567.
Wijewickreme, D. & Atukorala, U. (2006). Ground Improvement
For Mitigating Liquefaction-Induced Geotechnical Hazards, Eds. B.
Indraratna & C. Jian, Elsevier Publishers (UK) (43 pages).
Youd, T.L. & Perkins, D.M. (1978). Mapping of liquefaction severity
index. J. Geotech. Eng., ASCE, Vol. 113, No. 11, pp. 1374-1392.
Youd, T.L., Idriss, I. M., Andrus, R. D., Arango, I., Castro, G.,
Christian, J.T., Dobry, R., Finn, W.D.L., Harder Jr., L.F., Hynes,
M.E., Ishihara, K., Koester, J.P., Liao, S.S.C, Marcuson III, W.F.,
Martin, G.R., Mitchell, J.K., Moriwaki, Y., Power, M.S., Robertson,
P.K., Seed, R.B., & Stokoe II, K.H. (2001). Liquefaction Resistance
of Soils: Summary Report from the 1996 NCEER and 1998 NCEER/
NSF Workshops on Evaluation of Liquefaction Resistance of Soils,
J. Geotech. Geoenvir. Eng., ASCE, Vol. 127, No. 10, pp. 817-833.