2010 crc showcase - performance - ballast design r3.106

25
Integrated BallastFormation-Track Design and Analysis including the Implications of Ballast Fouling and High Impact Loads Buddhima Indraratna Professor of Civil Engineering Director, Centre for Geomechanics & Railway Engineering Faculty of Engineering, University of Wollongong Other Researchers: Dr Sanjay Nimbalkar; Dr Cholachat Rujikiatkamjorn, Nayoma Tennakoon (PhD student) Industry Partners: David Christie and Sandy Pfeiffer (RailCorp); Mike Martin and Damien Foun (QR), Tim Neville (ARTC) Cooperative Research Centre (CRC) for Rail Innovation Showcase Event Thursday 30 September 2010 CRC Projects R3.106 - Ballast Design

Post on 14-Sep-2014

1.702 views

Category:

Documents


8 download

DESCRIPTION

 

TRANSCRIPT

Page 1: 2010 CRC Showcase - Performance - Ballast Design R3.106

Integrated Ballast–Formation-Track Design

and Analysis including the Implications of

Ballast Fouling and High Impact Loads

Buddhima IndraratnaProfessor of Civil Engineering

Director, Centre for Geomechanics & Railway Engineering

Faculty of Engineering, University of Wollongong

Other Researchers: Dr Sanjay Nimbalkar; Dr Cholachat

Rujikiatkamjorn, Nayoma Tennakoon (PhD student)

Industry Partners: David Christie and Sandy Pfeiffer (RailCorp); Mike

Martin and Damien Foun (QR), Tim Neville (ARTC)

Cooperative Research Centre (CRC) for Rail Innovation

Showcase Event

Thursday 30 September 2010

CRC Projects R3.106 - Ballast Design

Page 2: 2010 CRC Showcase - Performance - Ballast Design R3.106

Problems in Rail Track Substructure

Differential Settlement

Degradation

Clay Pumping

Void Clogging

Poor DrainageCoal Fouling

Page 3: 2010 CRC Showcase - Performance - Ballast Design R3.106

VCI =

(1+ef)

eb

x

Gs.b

Gs.f

xMf

Mb

x 100

eb = Void ratio of clean ballast

ef = Void ratio of fouling material

Gs-b = Specific gravity of ballast material

Gs-f = Specific gravity of fouling material

Mb = Dry mass of clean ballast

Mf = Dry mass of fouling material

Void Contaminant Index (VCI)

Ballast Fouling

100

b f

f b f

k kk

VCIk (k k )

Page 4: 2010 CRC Showcase - Performance - Ballast Design R3.106

Permeability Test Measurements and Predictions

1.E-06

1.E-05

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

0 20 40 60 80 100

Void Contaminant Index,VCI /(%)

Hy

dra

uli

c C

on

du

cti

vit

y /

(m/s

) Clay fouled ballast-Theoretical

Coal fouled ballast-Theoretical

Clay fouled ballast-Experimental

Coal fouled ballast-Experimental

Hydrulic conductivity of

coal fines

Hydraulic conductivity of

clayey fine sand

100

b f

f b f

k kk

VCIk (k k )

Page 5: 2010 CRC Showcase - Performance - Ballast Design R3.106

Seepage model with SEEP-W

0.3m

4m

45o

Degree of Fouling

VCI (%)

Hydraulic conductivity

k (m/s) – Lab data

0% 0.3

25% 0.02

50% 0.00012

100% 2.3 x 10-8

Clay fouled ballast

Zero pore water

pressure

Total Head =0.5m

Free Drainage Q/Qc>50

Good drainage 5<Q/Qc<50

Acceptable drainage 1<Q/Qc<5

Poor Drainage 0.25<Q/Qc<1

Very Poor 0.0005<Q/Qc<0.25

Impervious Q/Qc<0.0005

Equivalent Maximum Flow rate ,Qc = 0.4 litres/sec.

(based on an extreme precipitation event of

300mm/hour)

Drainage capacity of the track, Q

Drainage Criteria – PhD work of Ms. Nayoma Tennakoon

Page 6: 2010 CRC Showcase - Performance - Ballast Design R3.106

Shoulder ballast maintenance requirement

Shoulder ballast with

0% VCI

Poor Drainage

(k2,k3,k4)

Min. VCI = (50,50,50)

Shoulder ballast with

25% VCI

Poor Drainage

(k2,k3,k4)

Min. VCI =

(50,50,50)

Shoulder ballast with 50% VCI

L=0.2m

Poor Drainage in

all cases

L=0.1m

Poor Drainage

(k2,k3,k4)

Min. VCI =

(25,25,25)

Shoulder ballast with 100% VCI

Impervious in all cases

k2

k3

k4

L

Track Drainage Assessment

Bottom Ballast layer

Middle ballast layer

L

Top ballast layer

Page 7: 2010 CRC Showcase - Performance - Ballast Design R3.106

Height of the ballast sample = 300 mm

Diameter of the ballast sample= 300 mm

Performance of ballast upon impact – use of shock mats

Weight of drop hammer = 5.81 kN (0.6 t)

Maximum Height = 6 m

Maximum drop velocity = 10 m/s

Dynamic load cell capacity = 1200 kN

Drop Hammer - Impact Testing equipment

Low confining pressures in track are

similar to rubber membrane encasement

Page 8: 2010 CRC Showcase - Performance - Ballast Design R3.106

Impact Response

In the application of continuous blows on the same specimen, multiple

instantaneous P1 peaks are followed by a longer duration P2 peak.

It is force P2 that causes predominate ballast damage.

With greater breakage and subsequent compression, P2 peak

becomes more distinct with increasing number of blows.

1st Blow 9th Blow

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14

0

40

80

120

160

200

240

280

320

360

Impact force excitation during 1st Blow

Fast Fourier Transform:

Low Pass Filter (cut-off frequency 50000 Hz)

Imp

act

forc

e (k

N)

time (sec)

Separation between the impactor and sample

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18

0

40

80

120

160

200

240

280

320

360

Imp

act

forc

e (k

N)

time (sec)

Impact force excitation during 9th Blow

Fast Fourier Transform:

Low Pass Filter (cut-off frequency 50000 Hz)

Multiple P1 type peaks

P2 type peak

Page 9: 2010 CRC Showcase - Performance - Ballast Design R3.106

Assessment of ballast breakage during impact

Subgrade

type

Position of shock

mat

Ballast Breakage

Index (BBI)

Without shock mat

Stiff - 0.170

Soft - 0.080

With Shock mat

Stiff Above ballast 0.145

Stiff Below ballast 0.129

Stiff Above & below

ballast

0.091

Soft Above ballast 0.045

Soft Below ballast 0.056

Soft Above & below

ballast

0.028

Sieve Size (mm)

0

1

Fra

ctio

n P

ass

ing

Initial PSD

Final PSD

Arb

itrar

y bo

unda

ry o

f max

imum

bre

akag

e

2.36

d95iA

B BAABBI

2.36 = smallest sieve size

0 63

PSD = particle size distribution

d95i = d95 of largest

sieve size

Shift in PSD caused by degradation

dmax

Indraratna et al., 2005

Page 10: 2010 CRC Showcase - Performance - Ballast Design R3.106

Prismoidal Triaxial Rig to

Simulate a Track Section(Specimen: 800600600 mm)

Use of Geosynthetics – Process Simulation Testing

0

5

10

15

20

25

0 100000 200000 300000 400000 500000 600000

Number of load cycles, N

Sett

lem

en

t, S

(m

m)

Fresh ballast (wet)

Recycled ballast (wet)

Recycled ballast with geotextile (wet)

Recycled ballast with geogrid (wet)

Recycled ballast with geocomposite (wet)

Stabilisation

Rapid increase

in settlement

-6

-4

-2

0

2

4

0 10 20 30 40 50 60 70

Grain size (mm)

DW

k (

%)

Fresh ballast (wet)

Recycled ballast (wet)

Recycled ballast with geotextile (wet)

Recycled ballast with geogrid (wet)

Recycled ballast with geocomposite (wet)

Effect of geosyntheticsHighest breakage

Effect of Geosynthetics on Ballast Degradation

Settlement of ballast with and without geosynthetics

Page 11: 2010 CRC Showcase - Performance - Ballast Design R3.106

Details of instrumented track

Section of ballasted track bed with geocomposite layer

From Theory to Practice: Use of Geosynthetics in Bulli Track

Page 12: 2010 CRC Showcase - Performance - Ballast Design R3.106

Preparation of Fully Instrumented Trial Track in Bulli

Geocomposite layer

(geogrid+geotextile)

before ballast

placement8 October 2006

Ballast placement

over the geocomposite

Geotextile

Bonded Geogrid

Recycled Ballast

from Chullora Quarry, Sydney

Fresh Ballast

Bombo Quarry, Wollongong

Page 13: 2010 CRC Showcase - Performance - Ballast Design R3.106

Field Instrumentation in Bulli

Settlement pegs

installed at ballast-

capping interface

Displacement

transducers installed at

sleeper-ballast interface

Page 14: 2010 CRC Showcase - Performance - Ballast Design R3.106

Deformation of Ballast

The recycled ballast performs well, if a well-graded PSD is adopted (Cu = 1.8) and

stabilised with geogrids.

A well-graded recycled ballast (Cu>2) can provide a higher placement density, hence

a reduced settlement compared to a Uniform ballast (Cu<1.5) .

Mean settlement (Sv)avg and

average vertical strain (1)avg

Average lateral displacement (Sh)avg

and average lateral strain (3)avg

(Indraratna et al, ASCE, JGGE, 2010)

0 2 4 6 8 10 12 14

-14

-12

-10

-8

-6

-4

-2

-0

0 1x105

2x105

3x105

4x105

5x105

6x105

7x105

-0.56

-0.48

-0.40

-0.32

-0.24

-0.16

-0.08

-0.00

Fresh Ballast (uniform graded)

Recycled Ballast (well graded)

Fresh Ballast with Geocomposite

Recycled Ballast with Geocomposite

Number of load cycles, N

Av

era

ge l

ate

ral

dis

pla

cem

en

t o

f b

all

ast

, (S

h) av

g (

mm

)

Av

era

ge l

ate

ral

stra

in o

f b

all

ast

, 3

) avg (

%)

time, t (months)

0 2 4 6 8 10 12 14 16 18

18

15

12

9

6

3

0

0 1x105

2x105

3x105

4x105

5x105

6x105

7x105

8x105

9x105

6.00

5.00

4.00

3.00

2.00

1.00

0.00

Fresh Ballast (uniform graded)

Recycled Ballast (well graded)

Fresh Ballast with Geocomposite

Recycled Ballast with Geocomposite

Number of load cycles, N

Mean

sett

lem

en

t o

f b

all

ast

, (S

v) av

g (

mm

)

Av

era

ge v

ert

ical

stra

in o

f b

all

ast

, (

1) av

g (

%)

time, t (months)

Page 15: 2010 CRC Showcase - Performance - Ballast Design R3.106

Soft Subgrade: Embankment fill Stiff Subgrade: Hard rock cutting

Use of Shock Mats & Geogrids in Practice: Singleton (NSW) – R3.117

Types of Geosynthetic

Biaxial Geogrid - TerraGrid TG3030 (Polyfabrics)

Biaxial Geogrid - Tensar Geogrid SSLA30

(Geofabrics Australasia)

Biaxial Geogrid - EnkaGrid MAX 30 (Maccaferri)

Geocomposite - Combigrid 40/40 Geogrid +

Geotextile (Global Synthetics)

Shock mat (10 mm thick)

Page 16: 2010 CRC Showcase - Performance - Ballast Design R3.106

Instrumented Track for Performance Monitoring - Singleton

Geogrid layer placed

above the capping

Mudies Creek Bridge

pressure cells installation

Settlement pegs

placement in the track

Pressure cells below

the sleeper

Page 17: 2010 CRC Showcase - Performance - Ballast Design R3.106

Because of symmetry, adequate to consider half of the

track

Axle load of 25 tonnes and dynamic impact factor of

1.43 (@ speed of 80 km/h on standard gauge)

PLAXIS - Finite Element Analysis

Ultimate redistribution of vertical stress

Ultimate settlement with depth

If breakage is captured with associated plastic flow, then the settlement prediction will be more

accurate.

450

300

150

0

0 50 100 150 200 250 300

Elasto-plastic Model

Field Data

Vertical stress under rail, v (kPa)

Dep

th b

elo

w b

ase

of

sleep

er,

z (

mm

)

Ballast layer

Sub-ballast layer

450

300

150

0

0 5 10 15 20 25 30 35

Settlement under rail, Sv (mm)

Dep

th b

elo

w b

ase

of

sleep

er,

z (

mm

)

Elasto-plastic Model

Field Data

Ballast layer

Sub-ballast layer

FEM predictions are underestimated because

breakage is not captured well

Page 18: 2010 CRC Showcase - Performance - Ballast Design R3.106

New Design Procedures – UoW method(Systematic Method of Analysis of Rail Track – SMART)

Criterion 1: Critical Shear Strength (ballast or subgrade)

Conventional Li and Selig approach UoW Ballast Parameters

UOW

Page 19: 2010 CRC Showcase - Performance - Ballast Design R3.106

Criterion 2: Critical track deformation

(Plastic vertical strains for (a) ballast = 8%; (b) subgrade = 2%)

Conventional Li and Selig approach UoW ballast parameters

UOW

Page 20: 2010 CRC Showcase - Performance - Ballast Design R3.106

Single Subgrade Layer Multiple Subgrade LayersFormulation of

SMART Approach (to be completed in 2012

under R3.117 project)

Page 21: 2010 CRC Showcase - Performance - Ballast Design R3.106

• The track drainage is assessed using a new parameter, ‘Void

Contaminant Index’ - VCI that takes into account the specific gravity of

different fouling materials.

• Recycled ballast stabilised with Geosynthetics can perform as well as

fresh ballast

• Shock mats improve the performance of ballast by reducing the

breakage caused by impact loads. Effectiveness depends on the

subgrade stiffness.

• Field trials conducted in Bulli and Singleton (NSW) demonstrate the

advantages of Field Performance Monitoring, apart from calibrating

FEM-based design technique.

• UOW research outcomes are continually captured in a MATLAB based

design approach: SMART (Systematic Method of Analysis of Rail

Tracks).

Conclusions

Page 22: 2010 CRC Showcase - Performance - Ballast Design R3.106

Australian Research Council (2 Discovery Projects and 3 Linkage

Projects since 1993).

Cooperative Research Centre for Railway Engineering and Technologies

(Rail CRC) (Project 6/139) from 2000-2007

Cooperative Research Centre (CRC for Rail Innovation)

ARC Centre of Excellence for Geotechnics (funded in 2010).

Industry Partners:

RailCorp, QR, and ARTC.

David Christie (RaiCorp, Sydney)

Tim Neville (ARTC, Newcastle)

Michael Martin, Damien Foun (QR, Brisbane)

Sandy Pfeiffer (RaiCorp, Sydney)

UOW Researchers: Dr Joanne Lackenby, Dr Wadud Salim, Dr. Sanjay

Nimbalkar, Ms. Nayoma Tennakoon, Dr Cholachat Rujikiatkamjorn

UOW Technical Staff: Alan Grant, Cameron Neilson, Ian Bridge

Acknowledgement

Page 23: 2010 CRC Showcase - Performance - Ballast Design R3.106
Page 24: 2010 CRC Showcase - Performance - Ballast Design R3.106

Test materials – Specifications

Shock mat (10 mm thick)

made of recycled rubber

(polyurethane)

Damping Ratio = 0.08

Fine sand as subgrade

Material Particle

Shape

dmax

(mm)

dmin

(mm)

d50

(mm)

Cu

Fresh

Ballast

Highly

angular

63.0 19.0 35.0 1.6

Subgrade

(sand)

- 4.75 0.075 0.48 2.3

0

10

20

30

40

50

60

70

80

90

100

0.010 0.100 1.000 10.000 100.000

% p

assi

ng

Particle Size (mm)

Australian StandardAS 2758.7 (1996)

Sand

Fresh Ballast

UoW new gradationIndraratna and Salim (2005)

Page 25: 2010 CRC Showcase - Performance - Ballast Design R3.106

Multiple Impact Loading with Shock Mats

The P2 force shows a significant increase with the extent of cumulative impact energy.

For stiff subgrade, shock mat is more effective in reducing P2 when located at the

bottom of ballast than at the top.

A soft subgrade itself serves as an energy absorber, hence the benefits of the shock mat are

generally marginal. However, if the shock mat is placed at the top of ballast to attenuate the

impulse waves, then P2 is reduced (less breakage).

Very Stiff Subgrade – steel plate Natural Softer Subgrade - sand

0.0 0.6 1.2 1.7 2.3 2.9 3.5 4.1 4.6 5.2 5.8

20

24

28

32

36

40

44

48

520 1 2 3 4 5 6 7 8 9 10

No Shock mat

Shock mat at top

Shock mat at bottom

Shock mat at top and bottom

Number of blows, N

Max

imu

m I

mp

act

Fo

rce,

P2 (

kN

)

Cumulative Impact Energy, E (kNm)

0.0 0.6 1.2 1.7 2.3 2.9 3.5 4.1 4.6 5.2 5.8

20

24

28

32

36

40

44

48

520 1 2 3 4 5 6 7 8 9 10

No Shock mat

Shock mat at top

Shock mat at bottom

Shock mat at top and bottom

Max

imu

m I

mp

act

Fo

rce,

P2 (

kN

)

Number of blows, N

Cumulative Impact Energy, E (kNm)