AGS VIC 2023 Symposium Proceedings WEB

PROCEEDINGS
2023 AUSTRALIAN GEOMECHANICS SOCIETY

VICTORIAN SYMPOSIUM
Novel Solutions in
Geotechnical Engineering
Wednesday 25 October 2023, 8:30am – 6:00pm
State Library Victoria, 328 Swanston Street Melbourne
VICTORIA CHAPTER
PREFACE
The Victorian Chapter of the Australian Geomechanics Society (AGS) is pleased to
announce the 2023 AGS Victorian Symposium titled “Novel Solutions in Geotechnical
Engineering” to be held on 25 October 2023.
The symposium forms part of the continuing program of events organised by the Victoria
Chapter of the AGS. The event is designed to showcase recent novel solutions and
innovations to address geotechnical challenges in practice. This symposium will be held as
a face-to-face event in Melbourne.
For the past decade, Australia and particularly Victoria has seen an unprecedented boom
in infrastructure investment involving mega-projects, transport and rail developments.
These significant infrastructure developments have generated new geotechnical
challenges due to complex geological condition, spatial variability of the ground and urban
construction. Adding to these challenges are the need for higher efficiency, cost saving and
consideration for sustainable developments. All these challenges and requirements have
spurred novel and innovative solutions across all aspects of geotechnical engineering.
Fuelling the geotechnical creativity and innovations is the advancement in technology
and knowledge, which allows novel solutions and tools to be developed to tackle
these geotechnical challenges using a variety of approaches (e.g. advanced numerical
modelling, physical testing, data-driven solutions, field instrumentation and monitoring).
This symposium seeks to bring together the local geotechnical community to share
and discuss recent geotechnical advances, with a particular emphasis on tackling the
geotechnical challenges in the Melbourne region and regional areas of Victoria. To
promote engagement with young engineers, the Symposium also hosts an inaugural
Young Geoprofessionals Poster Exhibition. In addition, a Heritage Time Capsule (HTC)
session has been included to collect feedback from the geotechnical community
regarding Geo-Education (current state, needs and challenges) in an Australian context.
This valuable feedback will contribute towards the development of our next generation
geotechnical workforce.
All right reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form without the permission of the Australian Geomechanics Society.
© 2023 Australian Geomechanics Society.
2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING i

SYMPOSIUM ORGANISING
COMMITTEE*
Adam Lander (Editor)
Clare Bridgeman
Marvin Bedary
Shiaohuey Chow (Chair)
Yunlong Tang (Editor)
Yuqi Tan
*a sub-committee of the AGS Victoria committee
TECHNICAL REVIEWERS
All papers have been peer reviewed in accordance
with the AGS Symposia review procedure.
SPONSORS
PLATINUM
Chadwick Geotechnics
GOLD
Insitu Geotech Services Pty Ltd
Global Synthetics Pty Ltd
SILVER
Black Insitu Testing
Insitutek
Rocscience
Cover image: Don Valley Brick Works Park, courtesy of Zhenyu Liu
ii 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

CONTENTS
Preface i
Contents iii
About the Keynote Speakers iv
Schedule v
Young Geoprofessional Poster Exhibition vi
Sponsors vii
SESSION 1 1 SESSION 3 41
NOVEL GEOTECHNICAL ANALYSIS ROCK BEHAVIOUR IN MINING
APPLICATIONS
Numerical Analysis of the Melbourne 3
Metro State Library Station Cavern Novel Solutions in Utilising Underground Coal 43
B. Coombes, D. Sainsbury and R. Storry Seam Gas Reservoirs for Gas Storage/Recovery
M. S. A. Perera
Interpretation of Shape Array Inclinometer 9
Data to Assess Trigger Action Levels Vein Effect on Mechanical Properties and 49
in Tailings Dams – A Case Study Fracture Behaviour of Porphyry Rock
M. Naeini and B. G. Nejad H. J. Wang and Q. B. Zhang
Back Analysis, Monitoring and Remediation 15

of Instability in a Soil Nail Wall in
Fissured Basaltic Clay
A. E. M. Dupleix, R. A. Day and D. N. Hagues
SESSION 2 21
CHALLENGING GEOTECHNICAL
CONDITIONS
Ground Improvement Works for Cutter Head 23

Intervention and Cross Passages for Klang
Valley Mass Rapid Transit (KVMRT)
L. H. Ooi and Y. S. Teo
Earthquake Site Subsoil Classification in 28

Melbourne’s South-west – A Crack in the Code?
A. McIntyre, C. Bridgeman and C. Crowe
Gas Permeability of Cement-Bentonite 35

Cut-off Wall Materials with Powder
and Granular Bentonite
S. E. Al-Baiaty and A. Bouazza
2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING iii

ABOUT THE KEYNOTE
SPEAKERS
DR SAMINTHA PERERA DR LEAN-HOCK OOI
Senior Lecturer in Geotechnical Geotechnical Director
Engineering Gamuda Berhad
University of Melbourne Dr. Ooi is a professional civil
Dr Samintha Perera is a senior engineer with more than 30 years of
lecturer at the University of experience in the field of geotechnical
Melbourne. Her research focuses engineering design and construction
on rock-fluid interactions in deep implementation. He has been
underground reservoirs during various field applications, managing geotechnical designs and construction of major
including CO2/H2 storage and natural gas extraction. Her infrastructural projects in South East Asia and Middle
research findings led her to win ROCHA Medal 2014 from East. These projects include runways, railways, dams and
ISRM, the Philip Law Medal 2016 from the Royal Society highways. Apart from that, he also has more than 10 years
of Victoria, and the Dorothy Hill Medal 2022 from the of experience in Metro underground works, which includes
Australian Academy of Science. She has recently been deep excavation, cut and cover structures and tunnelling.
featured among the “top 2% scientists in the world in the The tunnelling methods employed in the projects that he
Energy field” in the Stanford University Review in 2020 and was involved in included, TBM tunnelling, drill and blast
Engineers Australia “Create” Magazine in 2022. tunnelling and box jacking. Recently he was relocated to
Australia and is fortunate to be involved in both metro and
highway related projects and a firsthand experience on
road header tunnelling.
BEN COOMBES
Principal Engineer Some of the major projects that he was involved included
NOMA Consulting the Klang Valley Mass Rail Transit (KVMRT) Lines 1
and 2, Electrified Double Track Project (Ipoh-Padang
Ben Coombes is a Principal Engineer Besar), Stormwater Management and Road Tunnel
with over 20 years of experience in (SMART) Project, Pergau HEP Dam in Malaysia; Hamad
both the mining and civil tunnelling International Airport, and Dukhan Highway in Qatar; Sitra
industries. He has worked in Causeway Bridges in Bahrain; and Nam Theun 1 HEP
Australia, Papua New Guinea, India Dam in Laos.
and Kazakhstan. As part of his career, he has enjoyed
using his numerical analysis skills in conjunction with his Dr. Ooi graduated with a PhD in Geotechnical Engineering
site based experience. Ben was involved with the Detailed from the University of Sydney in 1990. He is a Member
Design and Construction Phase Supervision of the mined of Board of Engineers Malaysia (BEM), Member of the
tunnels (caverns, cross passages and back stubs) on Institution of Engineers, Malaysia (MIEM), Member of
Melbourne Metro. Engineers Australia and CPEng.
iv 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING
Go to table of contents
SCHEDULE
8:30 am Welcome and opening remarks SESSION 3
ROCK BEHAVIOUR IN MINING
SESSION 1 APPLICATIONS
NOVEL GEOTECHNICAL ANALYSIS Session Chair: Yunlong Tang (Tonkin+Taylor)
Session Chair: Shiaohuey Chow (University of Melbourne) 1:35 pm KEYNOTE ADDRESS
8:40 am KEYNOTE ADDRESS Novel Solutions in Utilising Underground
Numerical Analysis of the Melbourne Metro Coal Seam Gas Reservoirs for Gas Storage/
State Library Station Cavern Recovery
Ben Coombes (NOMA Consulting) Samintha Perera (University of Melbourne)
9:20 am Presentation 2:15 pm Presentation

Interpretation of Shape Array Inclinometer Vein Effect on Mechanical Properties and
Data to Assess Trigger Action Levels in Tailings Fracture Behaviour of Porphyry Rock
Dams – A Case Study Haojun Wang (Monash University)
Mahdi Naeini (ATC Williams) 2:30 pm Presentation
9:35 am Presentation Some Geotechnical Properties of the
Back Analysis, Monitoring and Remediation Bendigo Goldfield
of Instability in a Soil Nail Wall in Fissured Andrew Campbell (Mott MacDonald)
Basaltic Clay 2:45 pm Gold Sponsor Presentation
Dominic Hagues (Mott MacDonald)
Insitu Geotech Services
9:50 Platinum Sponsor Presentation 3:00 pm Q&A
Chadwick Geotechnics 3:15 pm Afternoon Tea & Young Geoprofessionals’
10:05 Q&A Poster Exhibition
10:15 Morning Tea & Young Geoprofessionals’ 4:00 pm HERITAGE TIME CAPSULE
Poster Exhibition & DISCUSSION SESSION
Negin Yousefpour (University of Melbourne)
SESSION 2
CHALLENGING GEOTECHNICAL CONDITIONS 4:30 pm Closing remarks and thanks
Session Chair: Mike Shackleton (McConnell Dowell) 4:40 pm Canapés and drinks
11:00 am KEYNOTE ADDRESS
Ground Improvement Works for Cutter Head
Intervention and Cross Passages for Klang
Valley Mass Rapid Transit (KVMRT)
Lean-Hock Ooi (Gamuda Berhad)
11:40 am Presentation
Earthquake Site Subsoil Classification in
Melbourne’s South-west – A Crack in the Code?
Andrew McIntyre (Douglas Partners)
11:55pm Presentation
Gas Permeability of Cement-Bentonite Cut-
off Wall Materials with Powder and Granular
Bentonite
Sarah Al-Baiaty (Tonkin + Taylor)
12:10 pm Gold Sponsor Presentation
Global Synthetics
12:25 pm Q & A
12:35 pm Lunch & Young Geoprofessionals’ Poster
Exhibition
2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING v
YOUNG GEOPROFESSIONAL
POSTER EXHIBITION
A Speargun Penetrometer Modelling rainfall-induced Forecasting Offshore

for Soil Characterisation slope failure using SPH with Turbidity Currents Using
Junlin Rong (RMIT) an advanced unsaturated Deep Learning Models
soil constitutive model Farid Fazel Mojtahedi
Dinh Nguyen (Monash University) (University of Melbourne)
Microbial Stabilization Preliminary design guidelines for Internal Erosion:

of Recycled Sand in geothermal pavements under From Micromechanical Insights
Geotechnical Fills different climatic conditions to Macro Constitutive Models
Sina Fouladi (Swinburne University) Xiaoying (Annie) Gu Tien V. Nguyen (Monash University)
(University of Melbourne)
Ground improvement using Biochar as a sustainable Geopolymer Stabilized

demolition wastes as geotechnical fill material Washed Recycled Sand in
semi-rigid inclusion columns Yueji Bai (Swinburne University) Pavement Constructions
Youli Lin (Swinburne University) Yunxin (Sophie) Xue
(Swinburne University)
vi 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING
SPONSORS
PLATINUM SPONSOR
GOLD SPONSORS TECH SERVI

EO
CE
G
INSITU
S
re d u
cin
ge
ain
ty
g
ote
c h n i c a l u n c e rt
SILVER SPONSORS
2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING vii
OVER 25 YEARS
OF ENGINEERING
TESTING EXCELLENCE
Proud sponsors of the

AGS Victoria Symposium 2023
chadwickgeotechnics.com.au
Australia’s growing investment in major new
infrastructure projects across a range of sectors
is driving a surge in demand for high-quality
geotechnical data.
Chadwick Geotechnics operates NATA accredited

testing laboratories in Dandenong South (VIC),
Western Melbourne (VIC), and Sunshine Coast
(QLD). Our laboratories are equipped with state-
of-the-art equipment and staffed by highly
qualified and experienced technicians.
Contact our team to learn how we can help you

to obtain the data you need to make informed
decisions and ensure the success of your project.
Call: 03 8796 7900

Email: info@chadwickgeotechnics.com.au
Innovative
Geosynthetic
Solutions for
Geotechnical
Problems
Global Synthetics is a 100% Australian-owned company,

proud to offer a complete range of high-quality geosynthethic
products backed by over 200 years of combined staff
experience in the industry.
We have supplied products to some of the largest recent
infrastructure works in Australia. Global Synthetics provides
major benefits to any geotechnical engineering project with
the right products and our technical expertise.
Stay on top of the latest Global Synthetics information

and visit our website at globalsynthetics.com.au
IGS can do anything
(almost)
TECH SERV
EO I
CE
G
INSITU
COMPREHENSIVE RANGE OF
redu
IN SITU TESTING, SAMPLING

cin
AND GEOTECHNICAL SERVICES ge nt

y
g
ote ai
c h nical u n c ert
mick@insitu.com.au mark@insitu.com.au
0407 467 025 0437 824 776 www.insitu.com.au
BRISBANE | MELBOURNE | SYDNEY | HOBART | PERTH | TOWNSVILLE | PNG
Applied Landslide
Risk Assessment
Course
The Australian Geomechanics Society is
running the Applied Landslide Risk Assessment
Course in Victoria in February 2024.
The seminar will teach participants the

fundamentals of landslide risk assessment
working through real world practical examples.
To learn more, visit:

https://australiangeomechanics.org/alra/
SESSION 1
NOVEL
GEOTECHNICAL
ANALYSIS
Numerical Analysis of the Melbourne Metro State Library Station Cavern
B. Coombes1, D. Sainsbury2 and R. Storry3
1
NOMA Consulting, Melbourne; email: ben.coombes@noma-consulting.com
2
Geotechnica, Melbourne; email: david.sainsbury@geotechnica.com
3
Bouyugues Construction, Melbourne, email: roger.storry@bouygues-construction.com
ABSTRACT
The Melbourne Metro Tunnel Project is delivering twin nine-kilometre rail tunnels in Melbourne, Australia. In addition
to the tunnels, five new underground stations are being constructed. Two of the new stations – State Library and
Town Hall – are complex cavern and adit excavations located in Melbourne’s City Centre. The State Library Station,
located predominantly underneath Swanston Street and a busy tram route, is surrounded by a mixture of modern,
educational and heritage developments requiring the excavation sequence and primary support to be designed to
ensure minimal surface impacts. To simulate the anisotropic rock mass response to the excavation of the State
Library Station, FLAC3D numerical analysis was undertaken. The analysis adopted the ubiquitous joint constitutive
model approach and was used to assess the performance of the primary lining design and to determine the impacts
the predicted ground displacements may have on the surrounding structures. Calibration and validation of the
ubiquitous joint modelling parameters for the Melbourne Formation have been completed against the measured
and observed responses throughout excavation. This was also completed against the structural elements, which
are often overlooked with respect to calibration and validation.
Keywords: anisotropic, numerical analysis, calibration, tunnelling
1 INTRODUCTION 2 GEOTECHNICAL CONDITIONS
The Melbourne Metro Tunnel Project is a The Melbourne Formation (also known as the
metropolitan rail infrastructure project located in Melbourne Mudstone) is an anisotropic rock mass,
Melbourne, Australia. The project comprises the both in terms of its intact properties and in terms of
construction of twin nine-kilometre tunnels with five its structural geological features.
underground stations. The two stations under
Swanston Street, State Library and Town Hall, were The predominant rock type in the Melbourne Formation
constructed as mined trinocular form caverns with the is siltstone, which typically com-prises 70% to 80% of
central ‘station’ tunnel excavated prior to the outer the formation. Bedding in the siltstone is typically well-
‘platform’ tunnels using sequential excavation developed and varying in the range 10 mm to 300 mm
methods (Figure 1). (Storry et al. 2023), provides a thick, but more typically in the range 30 mm to 60 mm
broad description of the project and construction thick. The remainder of the formation comprises
techniques used. sandstone and rare (<2%) beds of con-glomerate and
claystone. The bedding defects are highly persistent,
limited only by fold hinge axes, major fault structures
and igneous dykes (Figure 2), which resulted in
anisotropic strength and deformation behaviour.
Figure 1. Arrangement of State Library Station
To simulate the anisotropic rock mass response to

the excavation of the State Library Station, a three-
dimensional numerical analysis was undertaken with
FLAC3D. The analysis adopted the ubiquitous joint
constitutive model approach and was used to assess Figure 2. Geological Plan View of Major Structures
the performance of the primary lining design and to
determine the impacts the predicted ground dis- When un-weathered, the siltstone rock is a moderately
placements may have on the surrounding strong rock with uniaxial compressive strengths in the
infrastructure. order of 20 to 50 MPa. The weathering of the Melbourne
Formation ranges from extremely weathered to fresh
and is represented in the oxidized environment by a
2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING 3
Numerical Analysis of the Melbourne Metro State Library Station Cavern B. Coombes, D. Sainsbury and R. Storry
colour change from light yellow and grey through yellow design packages prior to construction. The
browns to dark grey (Paul et al. 2017). At the State advantage of using this technique was the improved
Library Station, the weathering profile of the encountered understanding of the rock mass response with the
Melbourne Formation was complex because of the inclusion of structural elements, allowing for
geological structure (intrusive igneous dykes and folds) improved accuracy with respect to predicted
and varied significantly over short distances. structural element forces, allowing greater efficiency
with the design of these elements. This numerical
3 NUMERICAL ANALYSIS analysis approach also had the benefit of a more
realistic rock mass response for inclusion in the
3.1 Background ground movement and building impact assessments
given the dense, urbanised environment that the
Geotechnical numerical modelling techniques can be State Library Station was constructed.
broadly grouped into continuum, discontinuum, and
equivalent continuum methods. The most common 3.2 Constitutive Model
equivalent continuum method is the ubiquitous (or
subiquitous) joint constitutive model which has been The Ubiquitous Joint model corresponds to a Mohr-
implemented within the finite difference codes FLAC, Coulomb material that exhibits a well-defined strength
FLAC3D, UDEC and 3DEC (Itasca, 2018). Due to the anisotropy due to embedded planes of weakness. As
inherent nature of the ubiquitous joint formulation, it shown in Figure 3(a), the planes of weakness can be
has been found that direct use of discontinuum assigned different orientations for each zone in the
properties in a ubiquitous joint model without careful model. The criterion for failure on the plane of
calibration of the material response can provide weakness consists of a composite Mohr-Coulomb
misleading model results by not capturing the rock envelope with a tension cut-off. The propagation of
mass response accurately (Sainsbury et al. 2017; damage within a Ubiquitous-Joint model can be
Perras and Diederichs, 2009). observed through the progressive degradation of
matrix cohesion and ubiquitous joint-failure plots at
Various discontinuum modelling techniques are various stages of loading in a simulated unconfined
available that explicitly simulate joints and compressive strength test, illustrated in Figure 3(b).
discontinuities within an anisotropic rock mass. Due
to the computational intensity of these numerical The Ubiquitous-Joint model formulation assumes
techniques, it is not practical to explicitly simulate the infinitesimal spacing and no length scale to their
joint fabric of an entire rock mass for routine analyses implementation. As such, a ubiquitous-joint material
of large-scale excavations. To overcome this, the cannot account for the bending stiffness of the
continuum based ubiquitous joint constitutive model individual layers of rock. As demonstrated by (Leitner
is commonly used to represent the strength and et al. 2006; Sainsbury and Sainsbury 2017), the
deformation behaviour of anisotropic rock masses selection of matrix and joint properties based on direct
(Clark 2006; Leitner et al. 2006). input of the measured block and joint strength will
result in a simulated material response that does not
This approach to numerical analysis was completed represent the true rock mass strength or deformational
for all adits and cavern primary ground support profile and provide misleading model results.
Figure 3. (a) Ubiquitous joint model: matrix and joint properties, (b) stress-strain response of simulated ubiquitous
joint rock mass (Sainsbury et al. 2008)
4 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING
To provide meaningful modelling results, careful During the tender and detailed design phases of the
calibration of the matrix and ubiquitous joint Metro Tunnel project, detailed UJRM parameter
parameters to the emergent behaviour from calibration was conducted based upon laboratory
discontinuum modelling techniques and in situ testing and field measurement data (including
monitoring and observation is required. A detailed bedding spacing and persistence measurements at
Ubiquitous Joint Rock Mass (UJRM) calibration notable exposures of Melbourne Formation).
procedure to account for rock mass anisotropy in
open pit rock slopes, block cave mines, deep mine Figure 4 illustrates the response of a large-scale,
access development and civil tunnelling has been discontinuum (3DEC), synthetic rock mass specimen
developed continuously since 2006 (Clark, 2006; on Melbourne Formation with horizontal bedding
Sainsbury and Sainsbury 2017; Sainsbury et al. (β=0). Figure 5 presents the calibrated continuum
2016; Johnson et al. 2016). This included back- (FLAC3D) strength response to the discontinuum
analysis of historical underground excavations in the (3DEC) model response at bedding orientations from
Melbourne Formation, notably the Melbourne horizontal to vertical. The emergent strength
Underground Rail Loop and City Link Domain Tunnel anisotropy in both discontinuum and continuum
were conducted and reported by Sainsbury and modelling approaches can be observed.
Amon (2017).
Figure 4. 3DEC model of simulated Melbourne Formation, β = 0, σ3 = 0 MPa
Figure 5. Calibrated UJRM anisotropic strength response compared to the discontinuum anisotropic response for
Melbourne Formation
4 MODEL GEOMETRY interface. The primary steel columns within the

central tunnel were simulated as a single beam
Multiple three-dimensional FLAC3D models were element connected to the shotcrete liner and
constructed to cover the entire State Library Station permanent base slab.
Cavern and the associated access shaft
infrastructure. Figure 6 shows the model geometry 5 GEOTECHNCIAL PARAMETERS
constructed to simulate the northern end of the
station. The primary support simulated, which The State Library Station Geotechnical Interpretive
consists of fibre reinforced shotcrete with local Report (GIR) was the basis for the fundamental
reinforcement, canopy tubes, spiles, rock bolts and geotechnical material properties (Wilson, 2018).
steel columns, is presented in Figure 7. Sainsbury and Amon (2017) previously reported the
calibrated UJRM input properties that were used in
the design development, while further refinement of
the UJRM properties have been reported by
Coombes et al. (2023). The calibrated UJRM input
parameters used throughout these analyses are
shown in Table 1.
An example of the disturbed nature of the rock mass

is evident in Figure 8, where Figure 8(a) is a face
photograph taken during geological mapping and
Figure 8(b) is the numerical analysis representation
of the same location.
Figure 6. FLAC3D Model Geometry
Figure 7. FLAC3D Model Structural Elements
Within the global models, the fibre reinforced

shotcrete is simulated with two-dimensional, linear
elastic shell elements installed in-cycle after each Figure 8. (a) Photograph of excavated face, (b)
1.5m excavation advance, while the rock bolts have FLAC3D central tunnel (top heading) model
been modelled with one-dimensional cable elements, geometry (disks showing ubiquitous joint orientation)
that simulate the axial behaviour of the rock bolt itself
and the shearing resistance along grout / rock
Table 1: Melbourne Formation modelling parameters (calibrated UJRM input properties)

Matrix Properties Joint Properties
Unit Weight Modulus Poisson’s
Domain
[kN/m3] [GPa] Ratio
c [kPa] φ [deg] σt [kPa] c [kPa] φ [deg]
MF3 23.5 1.6 0.26 126 42.0 5 0 20
MF2 24.5 4.2 0.25 336 44.8 30 0 32
MF1 26.0 9.2 0.24 964 47.7 150 0 32
6 CALIBRATION ASSESSMENT (Figure 9(b)) above Franklin Adit was predicted to be

22mm. At this excavation stage, it also becomes
Ground displacement calibration in the form of apparent the impact dyke orientation has with a skew
vertical settlement at the ground surface along the in the predicted settlement orientation matching the
centreline was completed for the central tunnel and dyke orientation. The maximum predicted settlement
the cavern excavation stages. Additionally, from the excavation increased from 20mm to 34mm.
calibration based on displacement of the primary
steel columns within the central tunnel cavern is The overall settlement profile with increasing
discussed in detail by Coombes et al. (2022). chainage demonstrated a settlement trough
associated with Franklin Adit before settlement
6.1 Instrumentation and Monitoring decreased. With the first intersection of the dykes,
the settlement progressively increased with
Instrumentation and monitoring of the excavation and settlement greatest between CH9280 and CH9305.
primary support elements was monitored with an Figure 10 shows the measured settlement along the
extensive array of in-tunnel convergence prisms, and centreline of the central tunnel alignment compared
surface monitoring points (SMPs) which were to the settlement predictions with both the best-
surveyed via manual survey techniques and ATS. estimate and lower-bound K0 simulations.
The State Library Station surface was monitored with
CH9200
CH9335
nearly 300 SMPs plus a suite of utility monitoring CL CL
0.0
points and building monitoring points.
-5.0 Predicted
6.2 Surface Settlement

(Design K0)
-10.0
Predicted
SETTLEMENT (mm) (Low K0)
The predicted surface settlements after the -15.0
Measured
excavation of the central tunnel and the eastbound -20.0
and westbound platform tunnels, to form the
-25.0
trinocular cavern, are shown in Figure 9 for the lower-
A’BECKETT ADIT
bound K0 simulations. -30.0 9220 FRANKLIN ADIT
-35.0
(a)
-40.0
FRANKLIN EAST
9190
9200
9210
9230
9240
9250
9260
9270
9280
9290
9300
9310
9320
9330
9340
9350
9360
SHAFT
CHAINAGE
CENTRAL TUNNEL
Figure 10. Calibration of predicted versus measured

CH9200
CH9335
settlement for trinocular cavern (ground

displacement calibration)
CH9305
DYKE(S)
6.3 Structural Element Settlement
PROFILE
EXTENT
Settlement
(mm) FRANKLIN ADIT Figure 11 shows the measured settlement of the
0.0 western steel columns compared to the settlement
-5.0
-10.0 predictions with both the best-estimate and lower-
-15.0
-20.0
bound K0 simulations.
-25.0
FRANKLIN EAST
-30.0 EASTBOUND PLATFORM CL

SHAFT
-35.0 TUNNEL
CL
0
-40.0
-2
Predicted
(Design K0)
-4
Predicted
CH9200
CH9335
SETTLEMENT (mm)
(Low K0)
-6
Measured
CH9305
-8
DYKE(S)
PROFILE -10
EXTENT
A’BECKETT ADIT
FRANKLIN ADIT
-12
WESTBOUND PLATFORM
TUNNEL
(b) -14
-16
Figure 9. Estimated settlement for the (a) central
9220
9230
9240
9250
9260
9270
9280
9290
9300
9310
9320
9330
9340
9350
9360
tunnel and (b) trinocular cavern CHAINAGE
Predicted settlement from the central tunnel (Figure Figure 11. Calibration of predicted versus measured
9(a)) above Franklin Adit was predicted to be 15mm settlement for western steel columns (structural
with a maximum predicted settlement at CH9295 of element calibration)
20mm. With continued excavation of the platform
tunnels, the predicted settlement from the excavation
The additional primary steel column settlement REFERENCES

measured may also be partially attributed to
compressibility of the waterproofing membrane and Coombes, B., Storry, R. and Sainsbury, D. (2022)
the displacement from engagement of the primary Melbourne Metro Tunnel Project Numerical Analysis
steel columns onto grout beds and the permanent of Anisotropic Rock Mass for State Library Station.
base slabs. This settlement contribution is minor. Australian Geomechanics Journal. Vol 57, No. 4,
December 2022.
Clark, I. (2006). Simulation of rock mass strength using
While not presented here, there was also very good ubiquitous-joints. In Hart, R. & Varona, P. (Eds.) Proc.
correlation with the strain gauges installed on the 4th Int’l. FLAC Symposium on Numerical Modeling in
steel columns compared to the predicted column load Geomechanics – 2006, Ma-drid, Spain. Paper No.
from 3DEC and FLAC3D simulations. In these 08–07. Minneapolis: Itasca.
simulations, the steel columns were explicitly Johnson, T. Pere, V., Dixon, R., de Graaf, P., Wines, D.
modelled as the six rectangular hollow sections that Hebert (2016) Geotechnical optimisation of Southern
formed each of the main steel columns. The details Ridge Cutback 3 at Tom Price mining operations, in
Proceedings of the First Asia Pacific Slope Stability in
of the steel column configuration and installation
Mining Conference, Australian Centre for
within the trinocular cavern are presented in Geomechanics, Perth.
Coombes et al. (2022). Leitner, R, Potsch, M. & Schubert, W. (2006). Aspects on
the Numerical Modeling of Rock Mass Anisotropy in
7 CONCLUSION Tunneling, Felsbau 24, No.2.
Paul, D. R., Barret, S. V. L., Jones, T. O. and Bennett, A. L.
Discontinuum analyses will always provide the most (2017). Development of Geotechnical Units and
Geotechnical Design Parameters for the Melbourne
rigorous assessment of anisotropic rock mass
Formation. Challenging Underground Space: Bigger,
strength and deformation behaviour. However, when Better, More: Proceedings of the 16th Australasian
large-scale analysis of anisotropic rock masses Tunnelling Conference, 30 Oct. – 1 Nov., Australasian
dictates the need for a continuum-based model, Tunnelling Society.
careful calibration of the material response to a series Perras, M. and Diederichs, M. (2009) Tunnelling in
of discontinuum numerical experiments provides horizontally laminated ground. ROCKENG09:
significant insight, understanding and robust Proceedings of the 3rd CANUS Rock Mechanics
modelling results. The critical step, where possible, Symposium, Toronto, May 2009.
Sainsbury, D.P., Sainsbury, B.L. and Sweeney (2016)
is careful calibration of the geotechnical parameters
Three-Dimensional Analysis of Complex Anisotropic
against a measured response. The calibration Slope Instability at MMG’s Century Mine.
process of the numerical analysis of the State Library Transactions of the Institutions of Mining and
Station was considered successful when compared Metallurgy, Section A, Mining Technology. Vol. 125.
against both rock mass response and structural No. 4. pp. 212-225.
element response. Sainsbury, D.P and Amon, A. (2017) Advanced Analysis of
Tunnel Excavation within the Anisotropic Melbourne
The complex geological conditions encountered at Formation. Challenging Underground Space: Bigger,
Better, More: Proceedings of the 16th Australasian
State Library Station between Franklin and A’Beckett
Tunnelling Conference, 30 Oct. – 1 Nov., Australasian
Adits highlights the need for good geotechnical Tunnelling Society.
understanding within the numerical analyses. Sainsbury, D.P., and Sainsbury, B.L. (2017) Practical Use
Achieving a calibration would be almost impossible of the Ubiquitous-Joint Constitutive Model for the
without this understanding. All data can be useful, Simulation of Anisotropic Rock Masses, Rock
even passing conversations with superintendents Mechanics and Rock Engineering. DOI
can reveal good geotechnical knowledge to improve 10.1007/s00603-017-1177-3.
the geotechnical understanding. Storry, R., Tricot, V., Barito, A. and Auvergne, S. (2023)
Approaches to Large-Span Cavern Construction in an
Urban Environment. WTC 2023. Athens Greece.
The UJRM numerical analysis approach adopted Wilson, R., 2018, CBD North Geotechnical Interpretative
provided the project with a robust design that gave all Report (TAS-COF-CBN-ZWD-REP-CGT-TNC-
stakeholders confidence in the station cavern design. C0008-00), Coffey, Abbotsford.
As anisotropic rock masses represent a significant
proportion of the worlds near surface rock masses,
this is an advance in the design of large scale,
complex underground excavations. While this project
was focussed on the Melbourne Formation, the
UJRM approach can be adopted for any anisotropic
rock mass.
8 ACKNOWLEDGEMENTS
The authors wish to acknowledge Rail Projects

Victoria and Cross Yarra Partnership for permission to
publish this paper and to our many project colleagues
who worked professionally to help deliver this project.
Cross Yarra Partnership brings together three of
Australia's leading construction partners, Lendlease
Engineering, John Holland, and Bouygues
Construction Australia to deliver the Metro Tunnel.
Interpretation of Shape Array Inclinometer Data to Assess Trigger Action
Levels in Tailings Dams – A Case Study
M. Naeini1 and B. G. Nejad2
1
Senior Geotechnical Engineer, ATC Williams, Melbourne Victoria 3195. Email: MahdiN@atcwilliams.com.au
2
Chief Technical Officer, ATC Williams, Melbourne Victoria 3195. Email: BehroozG@atcwilliams.com.au
ABSTRACT
This article investigates the displacement data from Shape Array Inclinometers in a tailings dam constructed on a
bed of clay and silty sand overlying the bedrock. During embankment construction, two inclinometers indicated
rapid movements at the interface of sand and clay layers, which triggered further action with the operation being
temporarily ceased. Consequently, a numerical simulation was conducted to properly simulate the development of
stresses within the foundation materials during construction and operation, and to better understand the extent of
foundation/embankment deformation and its pattern. Staged construction of the embankment was simulated by
employing FLAC software (Itasca group), adopting suitable constitutive models and associated input parameters
for each material. The model parameters were calibrated based on a series of field and laboratory tests to replicate
the observed movements in the inclinometers. It was found that the inclinometer's movements were localised and
caused by the adopted staged construction technique and not indicating a global sliding mechanism through the
foundation clay. It is concluded that despite the significant benefits of monitoring displacements by inclinometers,
adopting trigger action levels purely on this basis might be misleading.
Keywords: inclinometer data, trigger level assessment, numerical simulation, FLAC, tailings dam
1 INTRODUCTION performance of structure during various stages of

construction and operation. In a study by S. T.
Tailings dams are engineered structures constructed Kachanov et al. (2018), shape arrays were used to
to contain mining waste materials or tailings. These monitor and calibrate the deformation of a tailings
structures are often constructed in the form of an dam foundation. The study showed that the
embankment and are commonly used in the mining calibration process helped to improve the accuracy
industry to store and manage tailings. The stability of of the data obtained from the shape arrays and
the tailings dam structure is essential to ensure the provided valuable information on the stability of the
safety of the surrounding environment and structure. Another study by Li et al. (2018) used a
communities. One of the key factors that contributes combination of finite difference modelling and shape
to the stability of the tailings dam is its foundation. In array data to assess the stability of a tailings dam.
recent years, shape arrays have emerged as a useful The study showed that the combination of these two
tool for monitoring embankment and foundation provided a comprehensive assessment of the dam’s
deformations. stability and helped to identify potential instability
issues.
A shape array is a series of sensors placed vertically
in a borehole to measure the deformation of the The behaviour of the soil adjacent to a shape array
surrounding soil. The sensors are arranged in a can limit its ability to provide information on the
particular shape, such as a circle or a square, and overall performance of the structure. This is
are used to detect any changes in the shape of the demonstrated in a case study in which a tailings dam
soil. They are also useful for monitoring the stability is constructed on high plasticity clay and silty sand
of dams and their foundations. materials overlying bedrock. During construction, two
inclinometers showed rapid movements at the
Several studies in the literature have investigated the interface of sand and clay layers, leading to a
use of shape arrays in tailings dams. In a study by temporary halt in operations. The methodology and
Zavodni et al. (2016), a shape array was used to results of numerical modelling (using FLAC software)
monitor the deformation of the foundation of a tailings to better understand the recorded displacements (by
dam. The study showed that the shape array was shape arrays) are discussed in this paper.
able to detect the deformation of the soil and
provided valuable information on the stability of the 2 NUMERICAL MODELLING
dam’s foundation. Similar studies were performed by
Hagan et al. (2017), Williams et al. (2015), and 2.1 Methodology
Abuel-Naga and Abdel-Gawad (2018) on other
tailings dams. Overall, these studies demonstrate the The simplified embankment profile, used in the
usefulness of shape arrays in assessing the numerical modelling is shown in Figure 1. This cross-
foundation deformation of tailings dams. section was chosen given the complexity of the
foundation stratigraphy, groundwater conditions,
The data from shape arrays can be calibrated against availability of inclinometers, as well as the slope
the numerical modelling results to evaluate the stability consideration (noted as the critical section).
Interpretation of Shape Array Inclinometer Data to Assess Trigger M. Naeini and B. G. Nejad
Action Levels in Tailings Dams – A Case Study
The foundation beneath the embankment consists of possibility of static liquefaction with the increase in
fresh and weathered bedrock, lower clay, sand the stresses induced by embankment construction
(upstream side and downstream side), and upper and tailings deposition. The MCC constitutive model
clay. The complexity of the foundation layering has was implemented for the clayey foundation material
resulted in several scenarios being analysed with to adequately simulate the deformation
changing material characteristics and constitutive characteristics of this layer throughout the staged
models throughout the development of embankment construction loading.
construction and tailings deposition stages.
For all analyses undertaken, an embankment model
Two-dimensional plane strain, stress-deformation was generated, material properties were assigned
analyses simulating the development of the Tailings and then the model was run so stress states within
Dam to final configuration were undertaken using the embankment and foundation materials could be
FLAC (ITASCA group, 2022) and its built-in and user- progressively established. Sensitivity analyses were
defined constitutive models. Staged construction undertaken to optimise the mesh size within the
simulations were undertaken, from the starter model. Square zone (mesh) sizes of 0.5 m and 1 m
embankment construction and tailings deposition to were assessed, and it was determined that a mesh
the final design height. Considering the design of all size of 1 m provided reasonably close results to the
raises as well as the available piezometer data during mesh size of 0.5 meters while maintaining
construction, three piezometric lines were defined at reasonable runtimes for the analyses. This zone size
the end of the construction of each stage, to simulate was adopted to also allow conformance to wave
the pore pressure development within the tailings, propagation criteria required for subsequent dynamic
sandy and clayey foundation zones and deformation analyses, which are not the subject of
embankment materials. this paper.
A range of constitutive models from Mohr-Coulomb Five static models are presented in this study based
(M-C), NorSand, and Modified Cam-Clay (MCC) was on the results of the sensitivity analyses for different
employed to simulate the behaviour of embankment constitutive models and the adopted material
zones and foundation materials. The M-C elastic- parameters. These models have been assessed in
plastic model has been adopted for embankment terms of total deformation, stress distributions, and
materials and was initially used for both sandy and shear strain development for the Embankment
clayey foundation zones (i.e., the starting point from Raise 7 (existing) and Raise 8 (future) design
which to improve). Several sensitivity analyses were arrangements. Most importantly, the Raise 7 models
completed to calibrate the results with the overall were calibrated against inclinometer monitoring data
deformations of the available inclinometer records. (Figure 1).
The constitutive model for the sandy foundation
materials was replaced with NorSand to assess the
Figure 1. Schematic geometry of embankment and foundation
Table 1 summarises the key differences between the The results were assessed in terms of the following:
model runs for different scenarios. Run 1 (preliminary • Overall deformations for the embankment and
analysis) was carried out to evaluate the functionality foundation materials;
of staged construction simulation, verify the definition • Comparison of predicted foundation
of piezometric lines, and assess the suitability of deformations with the inclinometer records with
material parameters. In Run 1, the M-C elastic-plastic the focus being on foundation deformations; and
constitutive model was used for all materials, except • Developed shear strains at the end of each
for bedrock layers with adopted elastic model. construction stage.
The Run 1 results were considered the starting point, compared the results against monitoring data from
including the embankment Raise 7 geometry inclinometers S7 and SN17 (see Figure 1). The only
(existing) with subsequent model runs including difference between Runs 4 and 5 was the constitutive
changes in constitutive models, and adopted model adopted for the clay layers.
parameters for Embankment Raises 7 and 8.
2.2 Material parameters
Table 1: Staged construction runs
Foundation Table 2 presents the material properties adopted for
Constitutive Model the numerical modelling. Most parameters for
Upper Interface different zones were determined from extensive field
Run No. Sand
and Layer and laboratory investigations, including field CPTu
and
Lower with shear wave velocity measurement, triaxial,
Gravel
Clay monotonic and cyclic simple shear, oedometer and
1 Mohr- Mohr- CRS, direct shear, and ring shear testing. However,
No
(preliminary) Coulomb Coulomb if data were not available, reasonable assumptions
Mohr were made based on other data published in the
2 NorSand No literature.
Coulomb
Modified
3 Cam- NorSand No The large-strain shear moduli adopted in the staged
Clay construction simulations were factored from the small
Mohr strain shear modulus at approximately one percent
4 NorSand Yes shear strain. The maximum shear modulus, Gmax, is
Coulomb
Modified a measure of the stiffness of the material at very
5 Cam- NorSand Yes small strains. The Gmax values were formulated
Clay (based on the test data where applicable) using the
empirical relationship developed by Seed and Idriss,
Runs 2 to 5 were conducted to assess the foundation as described in Kramer (1996).
sand susceptibility to static liquefaction, predict upper
bound deformations of the foundation clay layers, The variations in Gmax with 𝜎𝜎𝜎𝜎′𝑚𝑚𝑚𝑚 (mean effective stress
and evaluate the performance of different constitutive or p′) for foundation sand and gravel materials were
models. Run 2 examined the functionality of the determined using the available field Seismic Cone
NorSand model, while Run 3 used the MCC model to Penetration Testing (SCPT) data. Similarly, the Gmax
predict overall deformation in the foundation clay values for the foundation clay layers were determined
layers. Runs 4 and 5 analysed the effects of from the shear wave velocities obtained during field
foundation deformation at the interface layer investigation as well as bender element tests on
between the sand and upper clay layers and undisturbed triaxial samples.
Table 2: Material parameters for MC and elastic Constitutive models

Material Constitutive 𝝆𝝆𝝆𝝆𝒅𝒅𝒅𝒅𝒅𝒅𝒅𝒅𝒅𝒅𝒅𝒅 Gmax c′
μ G/Gmax φ′
Type Model (kg/m3) (kPa)
𝑲𝑲𝑲𝑲𝟐𝟐𝟐𝟐𝒎𝒎𝒎𝒎𝒎𝒎𝒎𝒎𝒎𝒎𝒎𝒎 m
Uncompact
2,000 0.3 80 0.5 1 38
Fill
Konno et al. (1994)
Zone 4B 2,000 0.3 80 0.5 1 38
Zone 3B/3C 2,000 0.3 150 0.5 1 38
Drain 2,100 0.3 120 0.5 1 38
Shake-PI=0 (Vucetic
Filter 1,900 0.3 100 0.5 1 40
and Dorby 1991)
Tailings 1,800 0.35 30 0.5 2 26
MC Shake-PI=15 (Vucetic
Starter Emb. 1,800 0.3 150 0.5 1 38
and Dorby 1991)
Shake-PI=15 (Vucetic
Upper Clay 1,800 0.35 45 0.64
and Dorby 1991) Su(min) = 70
Shake-PI=30 (Vucetic Su/σ′vo = 0.27
Lower Clay 1,800 0.35 45 0.64
and Dorby 1991)
Sand & Shake-PI=0 (Vucetic
1,800 0.35 125 0.5 0.1 30
Gravel and Dorby 1991)
Weathered
1,900 0.4 160 0.5 - -
Bedrock
Shake-Rock
Slightly Elastic
(Schnable 1973)
Weathered 2,400 0.4 180 0.5 - -
Bedrock
MC: Mohr-Coulomb, μ: Poisson’s ratio, Gmax: small strain shear modulus, K2max: small strain shear modulus coefficient, m:
small strain shear modulus exponent, G/Gmax: Shear modulus reduction factor, c′: cohesion, φ′: friction angle, PI: Plasticity
Index, Su/σ′vo: Undrained shear strength ratio, Su(min): Minimum undrained shear strength.
The stiffness of a material reduces at larger strains, During construction, S7 inclinometer consistently
and this is described by the Modulus Reduction presented larger horizontal movements at the
function. To estimate the large strain shear modulus contact of the sand and upper clay layers. In model
(G), the Gmax values were converted using different Runs 4 and 5, an interface zone with a friction angle
modulus reduction curves for each material type. of 15° was adopted at this location across the model
This procedure was used only for elastic and M-C to assess if a global slippage at the contact can be
constitutive models, while other models such as justified. In these model runs, the stiffness and
NorSand and MCC can effectively simulate the strength properties of the foundation soils were also
non-linear stress-strain response of soil. adjusted to predict similar results to the Raise 7
displacement records from S7.
Table 3 presents the MCC model parameters used in
Runs 3 and 5. The parameters were selected based 3 RESULTS AND DISCUSSIONS
on the laboratory test results and sensitivity analyses
with further calibration undertaken using the field This section provides a summary of the results of the
displacement records. static stress-strain analyses (staged construction) for
the identified critical cross-section, with a particular
Table 3: Modified cam-clay constitutive model focus on comparing the results with the available
parameters deformation data for inclinometers S7 and SN17.
Run No.
Parameters
3 5 Figure 2 presents the foundation horizontal
λ 0.03 displacements at the end of Embankment Raise 7
κ 0.01 (existing conditions) for different model runs together
M with those recorded at S7 inclinometer (i.e., actual).
1.03 1.33 The first three runs were aimed at replicating the
overall foundation deformation profile recorded by S7
νλ 0.840
inclinometer. As can be seen from Figure 2, all three
Pre-consolidation runs provide a conservative (upper bound)
400 600
stress, p´0 (kPa) deformation profile, particularly in the sand layer.
p′0: preconsolidation pressure,
λ: slope of critical state line (CSL) or isotropic normal Runs 4 and 5 were to evaluate the effect of the
consolidation line (INCL) in Specific Volume (ν) vs
interface layer on the predicted deformation profile.
Mean Effective Stress (p´) plane,
κ: slope of swelling (unload/reload) line in ν-p´ plane, The stiffness of the sand and upper clay layers was
M: frictional constant, slope of the CSL in q-p´ plane also adjusted to closely simulate the overall
νλ: reference specific volume on the INCL deformation response in the first three runs. The
corresponding to p′ = 1 kPa results show that Runs 4 and 5 can predict similar
displacements to Runs 1-3 at S7 inclinometer.
Table 4 presents the NorSand model parameters
adopted for the foundation sand layer in Runs 2 to 5. Based on the preliminary analyses, Run 2 was adopted
The NorSand model parameters were calibrated and progressed as the seed model for prediction of
based on multiple triaxial tests on loose/dense embankment and foundation deformations due to
samples under both drained and undrained construction of subsequent embankment raises. Run 2
conditions. model predicted similar responses to Runs 1 and 3,
and can effectively assess the potential for static
Table 4: NorSand constitutive model parameters liquefaction of the foundation sand.
Run No.
Parameter
2 3 4 5
Gref 27,500
Elastic m 0.5
µ 0.2
Γ 0.53
Critical state λ 0.024
M 1.32
N 0.3
Plasticity
H 50 50 200 200
parameters
χtc 4.5
State
ψ -0.2
parameter
Gref: reference shear modulus,
m: exponent of mean effective stress ratio,
µ: Poisson’s ratio,
Γ: Reference void ratio on the CSL corresponding
to p´ = 1 kPa, Figure 2. Comparison of results with S7 inclinometer
N: Volumetric coupling coefficient, slope of stress data
dependency plot
H: Plastic hardening modulus, Figure 3 compares the recorded horizontal
χtc: Constant relating dilatancy to the state displacements of the foundation materials in Raises 5
parameter.
to 7 with those predicted by Run 2. As can be seen,
the magnitudes of the predicted deformations at S7, Figure 5 illustrates the predicted shear strain
although relatively small, are in reasonable agreement increment contours for the final geometry (Stage 10)
with the recorded displacements at this location. of embankment construction. The vertical scale has
been exaggerated with the embankment boundary
The predicted deformations for Raise 8 are also schematically shown.
depicted in Figure 3, which are slightly higher than
those for Raise 7, as expected. During the The localised shear strain increments are set to
construction of Raise 8, the S7 inclinometer was develop in two main locations of the foundation, one
damaged, and as a result, the predicted deformation under the downstream side of the starter
for Raise 8 could not be verified against the actual embankment and the other near the S7 inclinometer
data at this location. location. It is evident that the developed shear strains
are localised and not continuous across the entire
foundation (i.e. a potential failure plane), as also
confirmed by the recorded displacements at SN17.
The possibility of a static liquefaction within the

foundation sandy zones was examined considering
variations in the state parameter (ψ) due to staged
construction (i.e. increase in confining stresses). The
ψ is defined as the difference between the current
void ratio and the void ratio at the critical state, at the
same mean effective stress. A negative ψ means the
state of the material is dilative and hence not
susceptible to static liquefaction. Figure 6 plots the ψ
contours at the end of construction for the upstream
and downstream sandy zones (Refer to Figure 1).
As shown in Figure 6, while the ψ has increased with

the construction of embankment raises, the minimum
recorded state parameter of -0.14 indicates that the
foundation sand will remain dilative and hence will
Figure 3. Comparison of S7 inclinometer data with
not be susceptible to static liquefaction under the
Run 2 at different stages of construction
stresses applied by tailings and embankment raises
at the final height.
In order to investigate whether the observed
foundation displacements in S7 represent local or
global movements, the data from the SN17 4 CONCLUSION
inclinometer was reviewed. This inclinometer was
installed before the Raise 8 construction. This article investigated the foundation response of a
tailings dam as constructed in stages and compared
As shown in Figure 4, the runs with the interface the predicted values with field measurements. A
elements across the entire model yielded significantly summary of the process and outcomes of this study
higher horizontal deformations compared to those is provided below:
recorded at SN17 or those predicted by Run 2. This
indicates that the recorded interface displacements at • The MCC model parameters using the upper range
S7 are more likely localised. of the laboratory test results can yield similar
deformations to those from S7 inclinometer data for
the foundation clay.
• The inclusion of an interface layer does not provide
a deformation pattern that could be representative
of the overall response of the foundation to the
staged construction. This is demonstrated by
comparing the predicted deformations with the
SN17 inclinometer records.
• The developed shear strains through the
construction process indicate only localised
shearing of the foundation clay at limited locations
due to stress concentrations induced by the
adopted staged construction methodology.
• NorSand constitutive model for the foundation
sandy materials together with the MC model for the
clay layers provides a proper yet conservative
deformation response for the foundation. The
model indicates that the foundation sandy
materials remain dilative (i.e. negative state
parameter) after construction of the embankment
Figure 4. Comparison of results with SN17 to full height and final tailings deposition, and hence
inclinometer data for mid-raise 8 construction
are not susceptible to the static liquefaction during adopting trigger action levels purely based on a
construction and operation stages. limited number of inclinometer monitoring data
• The movements recorded by the inclinometers are might be misleading. Employing numerical
localised and caused by the adopted staged simulation methods has the potential to evaluate
construction technique and not indicating a global the degree of movements and its influence on the
sliding mechanism through the foundation clay. overall performance of the structure.
• It is concluded that despite the significant benefits Shear strain increment
of displacement monitoring by inclinometers,
Figure 5. Predicted shear strain increment contours at the end of construction (not in scale)
Figure 6. Predicted state parameter contours at the end of construction
REFERENCES Konno, T., Hatanaka, M., Ishihara, K., Ibe, Y. and Izuka, S.
(1994), “Gravelly Soil Properties Evaluation by Large
Abuel-Naga, M. A., & Abdel-Gawad, F. M. (2018). Application Scale In-situ Cyclic Shear Tests, Ground Failures Under
of shape arrays in the evaluation of the deformation Seismic Conditions” Geotechnical Special Publication,
behavior of the foundation of a tailings dam. Journal of No. 44, ASCE Reston, Va., 177-200.
Rock Mechanics and Geotechnical Engineering, 10(3), Li, H., Chen, Y., Zhang, J., & Li, G. (2018). Numerical analysis
496-504. and field monitoring of the stability of a tailings dam
Hagan, P. C., Walton, G., Galvin, J., & Gu, Z. (2017). The use foundation. Arabian Journal of Geosciences, 11(10), 231.
of shape arrays in the monitoring of tailings dams. In Schnabel, P.B. (1973), “Effects of Local Geology and Distance
Proceedings of the 20th International Seminar on Paste from the Source on Earthquake Ground Motions”, PhD.
and Thickened Tailings (pp. 237-246). Thesis, University of California, Berkeley, California.
ITASCA Consulting Group, “FLAC Version 8.1 Manual”, Vucetic, M. and Dobry, R. (1991), “Effect of Soil Plasticity on
Minnesota, USA. Cyclic Response”, Journal of Geotechnical Engineering,
Jefferies, M. and K. Been, “Soil Liquefaction - A Critical State Vol. 117, No. 1. ASCE, pp. 89-107.
Approach (Second Edition)”, Taylor & Francis Group, Williams, J. R., Harrison, J. P., & Young, R. P. (2015). Seismic
2016: p. 712. deformation monitoring of a tailings dam foundation
Kachanov, S. T., Zhu, Y., & Yin, J. (2018). Application of shape using a shape array. Soil Dynamics and Earthquake
array calibration in the monitoring of tailings dam Engineering, 72, 90-101.
foundations. In Proceedings of the 21st International Zavodni, B., Maloney, S., Eitzenberger, A., & Hagan, P. C.
Seminar on Paste and Thickened Tailings (pp. 33-43). (2016). Monitoring of a tailings dam foundation using a
shape array. In Proceedings of the 19th International
Seminar on Paste and Thickened Tailings (pp. 101-108).
Back Analysis, Monitoring and Remediation of Instability in a Soil Nail Wall in
Fissured Basaltic Clay
A. E. M. Dupleix1, BEng GradIEAust, R. A. Day2, MEngSc FIEAust CPEng NER
and D. N. Hagues3, MSc DIC CGeol
1
Geotechnical Engineer, Mott MacDonald Pty Ltd, Melbourne austen.dupleix@mottmac.com
2
Associate Principal, Arup Australia Pty Ltd, Melbourne email: rob.day@arup.com
3
Principal Geotechnical Engineer, Mott MacDonald Pty Ltd, Melbourne email dominic.hagues@mottmac.com
ABSTRACT
Soil nail walls are increasingly used as one of the few feasible construction methodologies for constrained urban
environments. The top-down construction methodology used for soil nail walls is inherently unstable during
construction and there are documented instances of instability occurring during construction staging. Temporary
stability during excavation stages relies heavily on short-term undrained strength of clay, the adhesion of shotcrete
to the face, and the ability of soil nails to act in shear, none of which are relied on in the completed soil nail wall for
long term stability. Fissuring in basaltic clay further exacerbates this, as it can lead to over-break in the excavation
face and rapid onset of drained strength behaviour. For the design and construction of the soil nail wall at Ferguson
Street Level Crossing Removal, laboratory testing had identified the presence of low drained strength in the basaltic
clay indicative of fissuring. Confined space for construction and a limited occupation window for construction also
put tight constraints on construction staging. These factors were addressed with design measures to reduce reliance
on nail shear and shotcrete adhesion during construction, construction methodology, close monitoring and
supervision, and a toolbox of treatments developed to ensure construction could take place quickly before onset of
drained conditions. However, an unprecedented industry-wide two-week shut during the Covid epidemic, followed
by several weeks of reduced shift work occurred mid-way through construction of the soil nail wall, leading to
significant delays in the construction program. Significant overbreak and wall movements were subsequently
observed, and additional propping and remedial measures were implemented to prevent further movement and
allow completion. Back analysis was carried out on the movements, which was able to demonstrate the behaviour
was consistent with the development of drained conditions and that there was no long-term detrimental effect to the
completed wall.
Keywords: soil nail fissured clay instability
1 INTRODUCTION of approximately 15m long were excavated and soil

nailed while leaving a 1H:1V berm approximately
The Ferguson Street Level Crossing Removal is one 15m long in adjacent sections as recommended by
of several rail level crossing removal projects carried Gniel et al. (2019). However, due to time constraints
out in the western suburbs of Melbourne in ground and access issues (single point of access for plant
conditions comprising two to five metres of fissured and spoil removal), a compromise was made
basaltic clay overlying relatively competent basalt between the design and delivery teams, allowing as
rock. Removal of the level crossing required lowering much wall to be opened as could be completed within
of the existing twin track alignment and North a 48-hour period.
Williamstown Station by up to 9m to pass under a
new Ferguson Street Bridge. The alignment was 2 LABORATORY TESTING
tightly constrained by existing property boundaries
and building development, particularly in the area to There is a tendency for many projects in recent years to
the north of Ferguson Street (refer Figure 1), have relatively little site-specific laboratory test data,
meaning that train services would need to be relying instead on published typical strength values.
suspended for a planned 10 weeks under a full However, there had already been previous soil nail
occupation to facilitate the construction of the 650m projects in basaltic clays where unstable conditions had
long rail trench. developed during construction. Based on this history,
drained strength parameters of the residual clay soils
A combination of solider pile walls and soil nail walls had been identified as a key design sensitivity, and a
were selected as the retaining solution. Solider piles total of six multi-stage consolidated undrained triaxial
were selected in areas with adjacent buildings or tests with pore water pressure measurement were
proposed development. As a time and cost saving carried out for this project to establish effective strength
exercise, soil nail walls up to 6m high with up to five parameters. The plasticity index of the residual basaltic
rows of soil nails were utilised in areas where there clays is very high (average plasticity index of 60%) and
was sufficient space and there was no planned future low effective strength results were expected.
development or adjacent industrial infrastructure. Nevertheless, these tests, both individually and
(refer Figure 2). combined, showed a surprisingly low effective friction
angle. The line of best fit through all data, as shown in
Initially, as a part of the design the adoption of a hit Figure 3, gave a drained cohesion (c’) of 10kPa and
and miss approach was proposed where hit sections drained angle of friction (φ’) of 18o. Some of the test
Back Analysis, Monitoring and Remediation of Instability in a Soil Nail Wall in Fissured Basaltic Clay A. E. M. Dupleix, R. A. Day and D. N. Hagues
stages showed quite high excess pore pressures during was well below peak friction angle documented for
shearing, but removing the individual tests with high soils of this plasticity and was more representative of
excess pore pressures did not change the line of best fit critical state friction angle. This suggested that the clay
significantly. had undergone significant reworking through fissuring
and shrink/swell movements. It was recognised that
multi-stage testing may still have resulted in some
increase in apparent cohesion. As small increases in
cohesion can have a big impact on the stability of soil
nail walls, a slightly lower cohesion of c’=6kPa was
adopted with a corresponding slight increase in friction
to 22o adopted to maintain a line of fit with almost the
same correlation coefficient as the best fit line.
Figure 3. Combined effective stress Mohr plot
While the long-term stability of soil nail walls is

typically governed by drained effective strength
parameters, construction of soil nail walls relies
heavily on short-term undrained strength parameters
to allow excavation and installation of soil nails and
shotcrete before drained strength conditions can
develop. Direct measures of undrained cohesion
from hand vane and laboratory unconsolidated
undrained tests, as well as correlated results from
pocket penetrometer and SPT are shown on Figure
4. Based on these results a design undrained
Figure 1. Layout of northern retaining walls strength (cu) profile was adopted as:
cu (kPa) = 50 + 12*depth ………………………(1)
TOP OF ROCK
Figure 2. Typical highest soil nail wall section
Another known issue with multi-stage tests is that Figure 4. Undrained Strength Results
progressive softening of the sample can occur
between the stages of the test with the third stage in It is important to note that the triaxial testing
particular giving a low strength closer to residual consistently gave lower undrained strengths than
strength. Discounting the third stage of all tests to inferred from either shear vane or pocket
investigate this possibility also had negligible impact penetrometer tests. At the time this was attributed to
on the line of best fit. On this basis, it was considered sample disturbance, but with the benefit of hindsight
that the results were accurate. The low friction angle it seems likely that small scale fissuring in the clay
contributed to this result as triaxial testing can shear shotcrete lip and facing shotcrete. Stage 2 was the
the sample along fissures, while the other tests excavation and strip drain placement for the subsequent
require shearing through the intact clay. drop, Stage 3 is installation of nails and shotcrete. Steps
2 and 3 were repeated until wall the wall was completed.
3 SOIL NAIL DESIGN AND DETAILS
Previous projects in basaltic clays had identified that

undermining of shotcrete and slumping of the
shotcrete face had been a key issue leading to
instability. Localised face instability and breakout due
to a combination of stress relief, clay fissuring and
“blowout” occurring when compressed air from soil
nail drilling was blocked by drilling spoil had also
been identified as an issue. To address these issues
as well as the low drained strength results, the
following were incorporated into the design:
• Relatively long soil nails were required to give the

required factor of safety due to the low drained
strength parameters.
• Long term, vertical support was provided to the
shotcrete by extending the base below finished
level as a form of footing, reducing reliance on
shotcrete adhesion.
• Restrictions were placed on traffic loads above the Figure 5. Reinforced shotcrete lip to prevent face
soil nail wall during construction, and works were slumping
limited to areas that could be treated within 48hrs.
• Trenching and backfilling with clay ahead of bulk
excavation was included to remove granular fill
materials in the top part of soil nail wall which could
rapidly behave in a drained manner.
• A 700mm wide reinforced shotcrete lip was
adopted to support the shotcrete vertically in the
temporary excavation stages, reducing reliance on
shotcrete adhesion (see Figure 5).
• Steep dipping of nails was avoided wherever
possible to reduce the vertical component of load
on the shotcrete face. (However, existing services/
structures forced adoption of a 20° dip in first nail
row).
• Shotcrete reinforcement was designed to
accommodate up to 300mm of overbreak in the
base case, with a drawing note requiring additional
support designed on a case-by-case basis for
larger overbreak.
• Widespread instrumentation and monitoring was
adopted with trigger levels for each construction Figure 6. Provisional staging if movements observed
stage set in advance along with an action plan in
the event of red trigger levels. Construction of the northern soil nail wall began on the
• Geotechnical site presence was included 16th September 2021 with replacement of granular
throughout construction to make observations on material on top of the wall and excavation of Stage 1,
face stability and movements and direct provisional leaving the batter 200-300mm fat of final cut face.
measures. Works continued on-schedule with nails being
• Provisional details were provided on the drawings installed and grouted over the subsequent 3 shifts.
for hit and miss construction and/or installation of During these 3 shifts, desiccation of the face and
soil nails ahead of excavation in the event unstable opening of fissures was observed as early as 10 hours
face conditions or excessive soil movements were after completion of Stage 1 excavation (Figure 8).
observed (see Figure 6).
• Provisional weep drains were included in the On the 18th of September in response to the increasing
drawings if seepage was observed. case numbers of the Covid-19 pandemic, shifts on site
were reduced from 12 hours to 6 hours, trimming of the
4 CONSTRUCTION AND MONITORING wall continued and was completed during the dayshift of
20th September. Preparation of the shotcreting platform
As per typical soil nail wall construction, as shown in was commenced on that nightshift. At the end of
Figure 7, Stage 1 consisted of the initial bench of 1.2m nightshift, Victoria went into a 2-week mandatory
deep, installation of first row of nails, installation of construction shutdown.
Figure 7. Construction Staging
There was concern that the un-shotcreted face would

soften and stress relieve if left uncovered for two
weeks, so emergency works were approved to
complete the shotcreting of Stage 1 during the
shutdown. Availability of work crews and shotcrete
hampered these works, which took three days to
complete.
Following the reopening of construction sites across

the state on 4th October, wall monitoring points were
installed and Stage 2 excavation began on the
dayshift of 9th October with bulk excavation being
completed on half of the wall that nightshift.
Opening of fissures and slumping was observed that

Figure 8. development of drying fissures in
nightshift. Works progressed with large blowouts up to
excavation face over 10 hours
0.5-0.8m deep caused by air pressure from soil nailing
being forced along fissures in the clay (Figure 9).
Further slumping on fissures was noted over the
following two days. These undermined support to the
completed section of shotcrete facing where this
occurred. At this point, the half of the wall initially
excavated on the 9th showed movements reaching the
red trigger limit set within design (6mm). To prevent
instability of the wall, instruction was given to construct
a 1V:2H wide earth batter in front of the wall.
Figure 9. extensive blow-out along fissures in clay

caused by air pressure from soil nailing
Figure 10. Wall movement with time and excavation
depth
At the point of shutdown, approximately 160m length
of wall had been excavated to Stage 1 with soil nails
Construction continued on the remaining section of
installed but shotcrete not yet applied except for a
exposed wall until red trigger limits were exceeded
25-40mm thick flash coat applied to a 50m section of
on the nightshift of the 12th. Instruction was given to
face where surface instability had been observed.
construct a batter in front of the remaining sections of
wall with these works completed by the end of night
shift of the 13th and works were suspended until a
cause of the movements and a remedial works plan 5.3 Impacts on Construction and
could be determined. At the point of intervention and Completed Design
backfilling, maximum movements of 20mm and a rate
of increasing movement of 1mm/hr were observed Having identified the failure mechanism, the key
(Figure 10). Over the following week, despite questions to be addressed were then:
backfilling, movements continued to increase before
stabilising at a maximum of approximately 50mm • Had the stability of the wall been compromised
lateral movement and 30mm of vertical settlement. by the ground movements?
• How would the remainder of construction be
5 BACK ANALYSIS achieved without causing collapse?
In order to understand the causes of the ground

movements and enable design of remedial measures
to allow completion of the walls, a series of back
analyses and sensitivity analyses were undertaken,
as described below.
5.1 Undrained “Blow-out” Modelling
Initial Plaxis back analysis models of the blow-outs at

stage 2 excavation level using undrained parameters
and the shotcrete ledge gave a factor of safety just
above 1 and large movements consistent with Figure 12. Movement vectors as near fully drained
observations were predicted (see Figure 11). conditions develop
However, blow-outs only occurred locally and would
not explain the movements observed over As the original design parameters had been
considerable stretches of wall. demonstrated to be consistent with the observed
behaviour, drained analysis of the completed soil nail
wall with the shotcrete supported vertically by the base
of the excavation gave a factor of safety unchanged
from the initial analysis.
There were initial concerns that the ground

movements could have caused yield and softening of
the soil nails, reducing their capacity. However, as can
be seen in Figure 12, the ground movements are
roughly perpendicular to the nails. Detailed queries of
the modelled stresses and strains in the soil nails
showed small soil nail strains with the relative axial
movement between the soil and the nail generally less
than 4mm (max 6mm), and the peak shaft resistance
was not mobilised anywhere along any of the nails.
Figure 11. Undrained displacements from blow-out
Additional soil nails tests to failure were carried out to
model
establish the resistance vs axial extension
relationship. These nails demonstrated that peak pull
5.2 Partially Drained Modelling out resistance was reached at an overall nail axial
strain of 0.2% (a total movement of 16mm for an 8m
By re-running the design models with drained friction long nail). From this testing, it was concluded that soil
angle and with drained cohesion gradually reduced nail capacity had not been impacted by the ground
until failure occurred, it was possible to replicate the movements.
observed deformations (see Figure 12) and
determine that a factor of safety of 1 occurred when It was apparent from stability analysis of the
the drained cohesion was reduced to 9kPa (only completed soil nail wall that if the vertical weight of
3kPa above the design value of 6kPa). This analysis the shotcrete could be supported through the
suggested that the originally adopted design remainder of construction, the construction stage
parameters were still reasonable and the most likely stability would be improved significantly. In practice,
reason for the observed movements was the this could be achieved by hit-and-miss construction
development of nearly fully drained conditions. It is by supporting the shotcrete at intervals and relying
the authors’ opinion that the development of near fully on the shotcrete to span as a deep beam between
drained conditions was due to a combination of the sections of shotcrete supported on the base of the
delays in construction and the extensive presence of excavation. In terms of the design model, this was
fissures which allowed more rapid pore pressure simulated by installing a compression only prop
dissipation. It was therefore apparent that further between the base of the shotcrete and the base of
excavation of Stage 4 and beyond would have an the excavation. These models showed an increase in
unacceptably high risk of wall collapse unless changes factor of safety for construction stages with drained
to the construction methodology were made. parameters from 1.0 to greater than 1.5.
6 REMEDIAL ACTIONS
Following back analysis and the workshopping of

remediation measures, removal of the emplaced
berm and continuation of Stage 3 was begun on 17th
October, 32 days after construction on Stage 1
began.
In conjunction with the delivery team, a hit and miss

construction methodology was implemented. Section
widths were limited to 13m hit sections and 11m miss Figure 14. localised propping to support movement
sections. This spacing was selected to allow joint in Stage 1 shotcrete
streamlined placement of reinforcement. A
completed hit section is shown in Figure 13, which
shows a block out for a test nail, allowing shotcrete
to be placed prior to production testing. Hit and miss
construction was able to be stopped once at least
one row of soil nails/rock bolts were installed within
rock, allowing valuable schedule gains. Monitoring of
wall movements was increased to twice per shift, with
daily updates fed back to the delivery team.
Fulltime geotechnical supervision continued, allowing

rapid identification of any issues to be discussed with
the delivery team and on-site development of solutions
to allow construction to continue rapidly. An example of Figure 15. trial alternative soil nail installation prior to
this is shown in Figure 14, where the shotcrete on one trimming to reduce blowout
side of an isolation joint was observed to drop 25mm
after excavation of the supporting berm. Within 3 hours, 7 CONCLUSION
the area was cleaned out and local prop utilizing Kelly
blocks were installed. Monitoring demonstrated that this Good quality site-specific laboratory testing is
remediation stopped wall movements after the initial essential for identifying potential stability issues with
drop. soil nail walls. The construction stability of soil nail
walls is heavily reliant on short-term undrained
Ongoing monitoring of the wall (Figure 10) showed behaviour of the soils, and delays in construction can
that enacting the hit and miss strategy, completing result in significant instability. Design details and
sections as soon as practical and localised propping construction methodology need to be employed to
slowed the rate of movement to within expected minimise the time that exposed faces are left open
levels. Movements began to stabilise after and reduce the reliance on shotcrete adhesion for
completion of the final level of shotcrete. Lessons temporary stability. Detailed instrumentation and
learnt from the northern wall were applied to the monitoring with clearly defined trigger levels and
southern soil nail wall. This reduced wall movements action plans need to be developed to provide early
to one third of those experienced on the northern warnings of potential failures and allow actions to be
wall. taken to prevent development of failures. For this
project, unavoidable delays were demonstrated to
have resulted in nearly fully drained conditions
developing during the excavation stages, ultimately
resulting in wall movements of greater than 100mm.
Despite this, it was possible to avoid wall collapse
and demonstrate that long term stability was not
compromised by adopting hit-and-miss construction
and detailed back-analysis.
8 ACKNOWLEDGEMENTS
Figure 13. Completed "hit" shotcrete panel supported The authors would like to acknowledge the member
on base of excavation organisations of the Western Program Alliance for
permission to present this paper, specifically our
owner and construction partners: LXRP, Metro
Trains and McConnell Dowell.
REFERENCES
Gniel, J., Lenthall, C., Paul, D. (2019). “Recent experiences
with soil nailing in deep cuts as part of level crossing
removal works in Melbourne.” Australian
Geomechanics Journal, vol. 54 (1), 41-51.
SESSION 2
CHALLENGING
GEOTECHNICAL
CONDITIONS
Photo courtesy of Narges Khajavi (ATC Williams)
Ground Improvement Works for Cutter Head Intervention and Cross Passages
for Klang Valley Mass Rapid Transit (KVMRT)
L. H. Ooi1, CPEng, PEng, MIEM and Y. S. Teo2, CPEng, PEng, MIEM.
1
Gamuda Engineering Australia, Victoria, Level 12 West, 303 Collins Street, Melbourne, VIC 3000, Australia;
email: ooi.leanhock@gamuda.com.au
2
Gamuda Engineering Australia, Victoria, Level 12 West, 303 Collins Street, Melbourne, VIC 3000, Australia;
email: teo.yeesoon@gamuda.com.au
ABSTRACT
Klang Valley Mass Rapid Transit (KVMRT) consists of 3 main lines, namely Sungai Buloh Kajang (SBK Line 1),
Sungai Buloh Serdang Putrajaya (SSP Line 2) and Circle Line to date. SBK Line 1 and SSP Line 2 have been
completed with Circle line at planning stage. A total of 23km tunnel length from both Line 1 and Line 2 were
constructed. The general geology of the underground section is in Kenny Hill formation and Kuala Lumpur (KL)
Limestone formation. This paper discusses the ground improvement works carried out for cross passages and
scheduled or planned intervention for cutter head inspection and maintenance. The performance of the treated
blocks for SBK line 1 was reviewed and the mixed successes will be presented. A case history of a cross passage
with highly complex karst in SSP Line 2 involving extensive investigation and ground treatment works is presented
as well.
Keywords: cutterhead intervention, ground improvement, cross passage, Kenny Hill, karstic limestone
1 INTRODUCTION
Klang Valley Mass Rapid Transit (KVMRT) consists

of 3 main lines namely Sungai Buloh Kajang (SBK
Line 1), Sungai Buloh Serdang Putrajaya (SSP
Line 2) and Circle Line. SBK Line 1 and SSP Line 2
have been completed with Circle line at the planning
stage. SBK Line 1 is approximately 51km in length,
comprising 7 underground stations, 9.5km twin
tunnel and 24 elevated stations with 41.5km elevated
tracks. Whereas SSP Line 2 is approximately 57.7km
in length, comprising 9 underground stations, 13.5km
twin tunnel and 27 elevated station with 44.2km
elevated tracks.
The general geology of the underground sections is

in Kenny Hill formation (approximately 55%) and KL
Limestone formation (approximately 40%).
The permeability of Kenny Hill generally ranges from

10-6 to 10-7 m/s while KL Limestone itself is generally
impervious except when faults, join swamps, solution
channels and cavities are encountered. In poor
ground and complex geological conditions, ground
improvement to facilitate cutter head intervention and
maintenance is required. Ground improvement works
will mitigate the inherent geological and Figure 1. SBK Line 1 and SSP Line 2 Alignment Map
hydrogeological risks when lowering the face support
pressure for maintenance of the cutter face or to 2.1 Kenny Hill Formation
excavate the cross passages in free air.
Kenny Hill formation mainly comprises sandstone,
2 GROUND CONDITIONS quartzite and phyllite. This formation is highly folded,
faulted and fractured. In the Kuala Lumpur basin, this
The alignments of KVMRT SBK Line 1 and SSP formation has been subjected to regional
Line 2 overlaid on Kuala Lumpur Geological map are metamorphism from interbedding of sedimentary
shown in Figure 1 below. The underground section of rocks of sandstone and shale to metamorphic rocks
SBK line traverses through relatively well define of quartzite and phyllite. In general, the tunnel is in
Kenny Hill and Karstic KL Limestone formation. The extremely weathered to highly weathered Kenny Hill
SSP line traverses through a more complex geology formation (see Figure 2).
overlaid with thick alluvium. A brief summary of the
geology is presented in sections below.
Ground Improvement Works for Cutter Head Intervention and Cross L. H. Ooi and Y. S. Teo
Passages for Klang Valley Mass Rapid Transit (KVMRT)
Figure 4. Example of exposed KL Limestone

excavated at one of KVMRT station box
3 SELECTION OF INTERVENTION
LOCATIONS AND GROUND
IMPROVEMENT WORKS
The mining of the KVMRT metro tunnels was carried

Figure 2. Extremely weathered to highly weathered out using TBMs. TBM tunnelling delays are usually
Kenny Hill Formation related to breakdown, damage and unforeseen ground
conditions causing face collapse, heaving and
2.2 KL Limestone Formation excessive cutter wear leading to low advance rates.
The KL Limestone Formation along the alignment is In all tunnelling works, including SBK and SSP line,
composed of fine to coarse grained, white to grey, to achieve the best productivity and trouble-free
predominantly recrystallised limestone, with local mining with the TBM, maintenance of the TBM cutter
developments of dolomitic limestone. The highly head is a major consideration.
irregular bedrock is overlaid by alluvium. The
limestone/dolomite rock has steep-sided valleys and The selection of the planned intervention location
an intricate three-dimensional network solutions considered the land use, impact to the community
channels and voids. These features are consistent and stakeholders and feasibility/accessibility of the
with Extreme tropical Karst classification according to location if ground treatment from surface is required.
Waltham (2003). This Karstic feature creates various The desired intervention block should facilitate free
problems and considerations for design and air intervention, if possible. The spacing of the
construction of cross passages and maintenance or intervention blocks also considers the geology and its
intervention blocks. impact on the degree/rate of wear and tear due to
mixed face condition and the hardness/abrasiveness
Figure 3 provides an indicative nature and formation of the mined material.
of KL limestone for a better understanding.
There were several unplanned interventions for SBK
Line 1 due to:
• Low productivity due to wear in quartzite

• Removal of left in place ground anchor tendons.
• Crusher/sizer unable to deal with blocky rocks
when traversing through a highly faulted/jointed
zone.
This resulted in significant downtime of the TBM

operations and recovery works of the TBM is highly
risky. Due to the prolonged stoppages and variation
or lack of confinement, these may lead to aggravate
settlement and impact to buildings and utilities.
Figure 3. Possible formation of KL Limestone
(Waltham 2003) It the light of the above, it is a good practice to plan
for intervention blocks along the tunnel alignment to
Figure 4 shows examples of the karsts features mitigate any untoward incidences.
exposed in the excavation of a shaft in KL Limestone.
The excavation revealed the complexities of KL 3.1 Intervention Locations in Kenny
limestone that include clean solution channels, Hill Formation
infilled caves, deep weathered valleys and the
challenge is to identify and employ an effective Based on the mechanical and hydraulic properties of
treatment to achieve the desired objectives. Kenny Hill formation obtained from soil investigations,
no ground treatment works was expected however, the intervention blocks in mixed face condition is
scheduled inspection and maintenance around 340m presented in Figure 7. The sectional view represents
intervals was adopted as part of the Standard the approximate location of the TBM cutter head
Operating Procedure (SOP) for SBK Line 1. intervention. The longitudinal section reflects the
condition variation of the mixed face condition over
3.2 Intervention Locations in KL the shield section. The different shade demarcated
Limestone the grout sections, and the number is the grout take
in litres.
In the extreme karst conditions, it is anticipated that
the highly variable rock head, steep-sided valleys
and solution channels have a significant impact on
the wear and tear of the disc cutters. Intervention
blocks coinciding with cross passage locations are
treated with appropriate ground improvement
methods at approximately 240m intervals and are
often tweaked or adjusted to locations with least
impact to utilities, public and traffic.
4 GROUND IMPROVEMENT METHODS
The intervention blocks were carried out where the

TBM traverses through limestone formation, as the
associated risks to tunnelling and mining of the cross Figure 7. Example of Mixed Face Condition for
passages are the highest. Figures 5 and 6 show the Intervention Block IB03D (N)
planned extent of the jet grout blocks at intervention
or cross passage locations. These schematics also In mixed face condition, ground improvement with jet
show that jet grouting was carried out in soil and grouting was carried out in the alluvium. The treated
fissure grouting in rock and the treated block is 5m block was carried out 5m around the extrados. The
from the extrados of the tunnel or cross passage. jet grout block was carried out based on the column
diameter, spacing and jetting parameters derived
from trial jetting. For quality control, a 2MPa strength
was specified for the jet grout column. The primary
emphasis of the jet grout block was water tightness
and cap grouting. Fissure grouting was carried out in
the balance of the intervention block in rock. The
object of the grouting to seal open fissure and
solution channels that may be connected to the
alluvium layer beyond the jet grouted block.
In cases where full face rock was encountered, only

fissure grouting was carried out as the ground
improvement for the treated block. Fissure grouting
was carried in a grid pattern and the edge of the
treated block was 5m from the extrados of the tunnel.
The grouting was based on the conventional fissure
grouting approach from a light to thicker mix when it
Figure 5. Plan View of the intervention block exceeded certain a threshold volume. Grouting was
carried out with a packer in bottom-up approach. The
first stage was 3m to 5m and subsequent stages
were carried out in 5m stages. For grouting points
that exceeded 10m3, secondary or additional grout
points were added until refusal criteria.
Figure 8 summarised some of the ground

improvement works that were carried out at the
cross-passage with intervention locations for SBK
Line 1. It can be seen from Figure 8 that majority of
the treatment works were carried out in full face rock
with the balance in mixed face conditions.
Figure 6. Section View of the invention block
An example of the treatment block and the

associated ground improvement carried out in one of
Figure 8. Intervention Block Summary Data for SBK

Line 1
Figure 9. Treated points locations for IB03
5 MONITORING AND TESTING
Typically, ground settlement and monitoring wells

were installed to measure the ground settlement and
ground water response during intervention and cross
passage construction.
As part of the quality control, coring would be carried

out in selected jet grout column to verify the uniaxial
compressive strength of the soilcrete in the alluvium.
In rock Lugeon test was performed on randomly
selected or preferred location to demonstrate the
water tightness of the treated block.
In cases where the intervention block was located

near to sensitive/heritage structure or sites with Figure 10. Summary of grout take in graphical form
cultural and historical concerns in extreme karsts,
additional investigation such as geophysical surveys In Figure 10, the diameter gives a visual impression of
using microgravity and electrical resistivity would be the total grout volume injected. The primary grout
carried out. Any major anomalies detected will be points with prefix P with high grout have secondary
further investigated using intrusive drilling like grout points prefix S introduced in between. When the
additional boreholes to evaluate the risks and grout intake of secondary, tertiary points was low, the
treatment requirements. grouting regime would be terminated. Although all the
treatment was carried out to the specified
6 PERFORMANCE OF TREATED requirements, some of the cutter head intervention
BLOCKS was still carried out under compressed air to control
groundwater inflow. This could be due to solution
Figure 8 shows that all the treated blocks in mixed features that evaded grouting or that the interface
face conditions require some level of compressed air between the jet grouting and rock is leaky. Overall, the
intervention. In full face rock, fissure grouting seems ground treatment provided a safe working condition for
to work well with more than 65% of the intervention intervention.
carried out in free air. At locations where intervention
was undertaken in compressed air, higher volume of 7 INTERVENTION AND CROSS PASSAGE
grout intake was observed during fissure grouting. IN KARSTS
This observation could be related to interconnected
fissures and cavities in the limestone. The following presents the challenges of ground
improvements at an extreme karstic location in SSP
A case history of the treatment for IB03 is used to Line 2. The preliminary investigation indicated that
demonstrate the process and outcome observed. the limestone bedrock was relatively shallow. Hence,
The blue eclipses in Figure 9, shows the sections the routine ground treatment started with a standard
presented in Figure 7. Location of the treated points regime of fissure grouting in accordance with the
and grout consumption is presented in Figure 9 and standard operating procedure and acceptance
Figure 10 respectively. criteria. When the North bound TBM arrived at this
location for scheduled maintenance, the face was
found to be “unstable” with soil and water inflow. The
maintenance stoppage was abandoned and the TBM 8 CONCLUSION

mining continued to the next intervention location.
The emergency response team was put on alert during Based on the actual interventions done for KVMRT
tunnelling and managed the traffic away from the SBK Line 1, significant observations which could be
tunnel corridor. After mining not too far from this concluded are:
location, a sink hole occurred. This led to a scrutiny of
the information of the grout points. The data was a) Mixed face ground condition has a higher
analysed, and the depths of the grouting points and occurrence of compression air intervention. A
the contours generated were plotted in Figures 11 low-risk approach was taken as mixed face
and 12. ground condition are usually closer to the
surface and the alluvial soil is usually
The top of rock generated from these data revealed unconsolidated.
several localised deep depressions in the bedrock b) Full face rock condition with successful free air
with soil penetration into the tunnel horizon and even intervention is generally related to better quality
below the tunnel invert near the cross adjacent to the bedrock as shown by the lower grout intake.
south bound tunnel intersection. In response to the c) Full face rock ground condition with
above observation, more probing was carried out to unsuccessful free air intervention shows higher
map the extent of the depressions and in view of the grout intake. This could be due to the joints and
risk posed to the mining of the cross passage, it was fissures and possible solution features that are
decided that jet grouting was to be carried out to cap not fully grouted.
the localised depression near the SB tunnel as
shown in Figure 13. This exercise proved to be highly 9 ACKNOWLEDGEMENTS
beneficial as the cross passage was successfully
completed without further incidence. The Authors would like to thank MMC-Gamuda JV
Sdn Bhd for permitting this paper to be published.
Special appreciation to those who were involved in
this project for their contributions and efforts to
complete this project.
REFERENCES
A.C. Waltham and P.G. Fookes (2003). “Engineering
classification of karst ground conditions”
Figure 11. Contour of the Rock Level
Figure 12. Sectional View Generated from the

Contour
Figure 13. Mapped 3D view of the rock level
Earthquake Site Subsoil Classification in Melbourne’s South-west
– A Crack in the Code?
A. McIntrye1, C. Bridgeman2 and C. Crowe3
1,2,3
Douglas Partners Pty Ltd, 231 Normanby Road, South Melbourne 3205, PO Box 5051
ABSTRACT
Design for earthquake actions in Australia is guided by the Australian Standard (AS 1170.4-2007) where the
design earthquake structural actions are a function of the size and nature of the structure being assessed, the
behaviour of the local ground profile during a seismic event and the probability of an earthquake occurring. To
assist designers, AS 1170.4-2007 outlines a simplified approach to model the behaviour of the ground profile,
where the ground profile is assigned to one of five subsoil classes. There are geographical areas west and south-
west of Melbourne where the local ground profile does not always align with one of the five ground profile (or
subsoil) classifications defined in AS 1170.4-2007. Strict adherence to the subsoil definitions may lead to either
a conservative or an optimistic assignment of subsoil class, both of which can have an impact on the design
earthquake loading and resulting seismic-force-resisting system. This paper discusses the nature of the ground
profile in the west and south-west of Melbourne and how a site subsoil class could be applied in this region,
including discussion on site natural period.
1 INTRODUCTION
For most sites in the broader Melbourne area,

assigning a site subsoil class in accordance with
AS 1170.4-2007 is a relatively straightforward
process. It can be undertaken with an understanding
of the expected geology and can be confirmed with
standard geotechnical investigation techniques. The
areas to the west and south-west of Melbourne,
including Footscray, Altona and Werribee, are
situated on the Werribee Plains Lava Field. The Port Phillip
ground profile in Werribee Plains Lava Field does not Bay
always align into one of five subsoil classes, which
assume that the ground profile typically increases in
strength / stiffness with depth.
Figure 1. Surface Geology map (GSV) showing
If a site subsoil class is not appropriately assigned, it approximate location of the Werribee Plains Lava
may result in an insufficient design during a seismic Field
event. Most of the development to date in this
geographical region has been low to medium rise The surface geology, as shown on the Melbourne
buildings, with an Earthquake Design Category I or II. geology map (GSV 1974) is dominated by the
As such, the impact of assigning an inappropriate Tertiary / Quaternary age Newer Volcanic Group,
subsoil class for these lower risk structures is although this is locally overlain with windblown
reduced. However, as the population of Melbourne sediments and alluvial / estuarine / lacustrine
and Victoria continues to expand, taller and more sediments near watercourses. The Newer Volcanic
significant structures (Earthquake Design Category II Group is not on an unusual geology for Melbourne.
and III) are likely to be developed in the west and The geology of the Werribee Plains Lava Field differs
south-west. in that it also forms the western extent of the Port
Phillip Bay Sunkland. There are different models
available for the geological setting (refer to the notes
2 THE WERRIBEE PLAINS LAVA FIELD
of the geology map, Geology of Melbourne District
(GSV 1967), and papers (Rosengren 1994) and
The geology of the west and south-west of
(Boyce 2012) as well as other sources).
Melbourne is referred to as the Werribee Plains Lava
Field. It extends inland from Port Phillip Bay to
Broadly speaking, the Port Phillip Bay Sunkland was
Bacchus Marsh, and down to Geelong, as shown in
created during the Tertiary period (circa 65 million
Figure 1.
years ago or Mya) when Australia was drifting away
from Antarctica. During the drift, subsidence of
Silurian age bedrock / basement rock (siltstone and
sandstone) occurred along faults that were orientated
in an approximate 100km block shape. The block
was bound by Rowsley Fault (in the west) and
Selwyn Fault (in the south-east), and a large
Earthquake Site Subsoil Classification in Melbourne’s South-west – A Crack in the Code? A. McIntyre, C. Bridgeman and C. Crowe
northwest trending fault (Melbourne Warp) to mark 3 SUBSOIL DEFINITIONS AS PER AS

the north. Depth to the Silurian age bedrock is now in 1170.4-2007
excess of 200m below ground surface level in
locations (discussed later). In the southern extent of AS 1170.4-2007 defines five subsoil classes and
the Werribee Plains, there is Devonian age granite states that a site subsoil class shall be assigned to
instead of the Silurian age bedrock. the subsoil class it most closely resembles. The site
subsoil classes are:
In addition to the plunging Silurian age bedrock, the
sea levels fluctuated significantly during the Tertiary • Class Ae – Strong rock
period. This resulted in the deeper parts of the • Class Be – Rock
sunkland experiencing extended periods of time in a • Class Ce – Shallow soil
submarine depositional environment and also in a
• Class De – Deep or soft soil
subaerial/ terrestrial depositional environment, e.g.,
• Class Ee – Very soft soil
Werribee Formation, Gellibrand Marl, Sandringham
Sandstone. Some volcanic activity occurred in this
For the purposes of this paper, Class Ee (very soft
period and is referred to as the Older Volcanics
soil site) is excluded from further discussion as it is
Group. A deep sedimentary basin of multi-layered
not deemed relevant to the geology. The remaining
deposits thus formed over the bedrock.
subsoil definitions are presented in Table 1 (see next
page) and can be summarised as:
Towards the end of the Tertiary period, extensive
volcanic activity arising from points of eruption
located north and northwest Melbourne flowed south • The ‘rock’ subsoil classes (Ae and Be) are
and created vast lava plains. The Werribee Plains assigned primarily based on rock strength and/or
Lava Field differs to the northern lava fields as the shear wave velocity; and
extensive early lava flows entirely eradicated the pre- • The ‘soil’ subsoil classes (Ce ad De) are based
existing Tertiary terrain. In addition, subsidence on natural site period, soil strength and soil layer
along the Rowsley Fault was ongoing during these thicknesses.
early lava flows, resulting in some downfaulting and
warping of the lava/basalt. Pauses in volcanic activity A high-level overview of the subsoil class definitions
created zones of localized weathering and erosion, would suggest it is possible for any of these four
creating an uneven distribution of basalt rock. subsoils to exist in the Werribee Plains Lava Field.
The basalt rock can often have uniaxial compressive
Later lava flows in the Quaternary period filled in strength (UCS) of 50MPa or more and be exposed at
valleys and depressions. The last lava flows to occur surface resulting in Class Ae. The basalt rock can
in the region were late Pleistocene. The basalt rock often have surficial layer of soil that is less than 3m
that then cooled at the surface was then differentially thick, resulting in a Class Be. The surficial soil profile
weathered and eroded over time. This resulted in a can be deeper than 3m, resulting in a Class Ce. There
highly irregular rock surface with a mantle of residual can be deep soils underneath the basalt that exceed
soils, extremely weathered materials with common the maximum depth limits for a Class Ce site and can
basalt boulders or floaters. return a natural site period well in excess of 0.6s. In
this scenario, a Class De applies. This is further
Rivers that had their flow path altered during the lava discussed in Section 6.
flows resumed course or cut new paths, and ponded
in lakes and estuaries, depositing recent sediments 4 WHAT IS THE IMPLICATION OF A
of clay, silt, and sand. The strong cold winds that SUBSOIL CLASS?
Victoria experienced during the last ice age maxima
deposited silts and fine sands drawn from floor of the Subsoil class is a fundamental component in the
dry (at the time) Port Phillip Bay and other coastal estimation of horizontal forces arising in an
areas. earthquake event as per Earthquake Design
Categories II and III defined in AS 1170.4-2007. For
Thus, the ground profile can be ‘simplified’ as from Earthquake Design Category II, the minimum
surface down comprising an intermittent and variable horizontal static force (Fi) applied to each storey of
thickness of soil overlying a highly irregular basalt the structure is given by the following equation from
rock surface that can outcrop. The rock has variable the Standard:
thickness and strength formed from multiple volcanic
flows, overlying variable Tertiary age layered soils
down to the top of the Silurian and Devonian age The factor Ks is obtained from Table 5.4 of the AS
bedrock, which can be quite deep. 1170.4-2007 and is a function of the site subsoil class
and storey under consideration. Ks ranges from 0.6
to 3.9 for a Class Be site. In comparison, a Class De
site which has a Ks factor of between 1.2 and 6.1,
approximately twice that of a Class Be site.
Table 1: Summary of Subsoil Class Definitions as per AS 1170.4-2007

Subsoil Description Definition
Class Ae Strong • UCS >50MPa or an average shear wave velocity over the top 30m of >1500m/s;
rock AND
• Not underlain by materials having a compressive strength less than 18MPa or
having an average shear wave velocity <600m/s
Class Be Rock • UCS between 1 and 50MPa or an average shear-wave velocity, over the top 30m
of >760m/s; AND
• Not underlain by materials having a compressive strength <0.8MPa or an
average shear wave velocity <300m/s.
• A surface layer of no more than 3m depth of highly weathered or completely
weathered rock or soil (a material with UCS <1MPa).
Class Ce Shallow • Not Class Ae or Class Be (i.e., not rock site) or Class De; AND
soil • The low-amplitude natural site period is less than or equal to 0.6s; or the depths
of soil do not exceed those listed in Table 4.1. The low-amplitude natural site
period may be estimated from:
(i) four times the shear-wave travel time from the surface to rock;
(ii) Nakamura ratios;
(iii) recorded earthquake motions; or
(iv) evaluated in accordance with Clause 4.1.3 (i.e. weighted summation of site
periods)
Where more than one method is used, the value determined from the most preferred
method given in Clause 4.1.2 shall be adopted. (i.e. preferentially use site period).
Class De Deep or • Not Class Ae, Class Be or Class Ee; AND
soft soil • Underlain by less than 10m of soil with an undrained shear-strength less than
12.5kPa or soil with Standard penetration test (SPT) N-values less than 6; AND
• either
(i) the low-amplitude natural site period is greater than 0.6s; or
(ii) the depths of soil exceed those listed in Table 4.1, where the low-amplitude
natural site period is estimated in accordance with Clause 4.2.3 (i.e., use the
methods described for Class Ce above).
For Earthquake Design Category III, the horizontal 5 WHAT EARTHQUAKE?

equivalent static design force (also Fi) applied to each
storey is proportional to the spectral shape factor, The Werribee Plains Lava Field is not known to be a
Ch(T). The spectral shape factor is a function of tectonically active region and building damage from
subsoil class and natural period of the structure (T1). an earthquake is rarely reported. The earthquake
An extract from AS 1170.4-2007 showing the hazard map in the AS 1170.4-2007 shows that the
normalized response of spectra for subsoil class is Werribee Plains Lava Field has a variable hazard
shown in Figure 2. factor (z) of between 0.08 and 0.10 for a 1-in-500-
year annual probability of exceedance. The hazard
factor increases moving southwest away from
Melbourne (0.08) towards Geelong (0.10). The
application of the hazard map zoning may not
completely reflect the seismic hazard of the area, as
the map does not necessarily account for the local
conditions such as faults or epicentres. Small shallow
earthquakes of small magnitude are known to cause
minor damage in an epicentral area.
The National Earthquake Alerts Centre at

Geoscience Australia reports that at least 38
earthquakes have been recorded in the Werribee
Figure 2. AS 1170.4-2007 Figure 6.4 Normalized Plains Lava Field from 1976 and up until as recently
Response Spectra for Site Subsoil Class as August 2023. The epicentres of the earthquakes
in the region are typically located within 15km of the
For a given structural period, soil sites (Class Ce, De Rowsley Fault, although 4 earthquake epicentres
and Ee) yield a higher spectral shape factor have been recorded within 8km of Werribee town
compared to rock sites (Class Ae and Be). Re-visiting centre. The magnitude of the measured earthquakes
the comparison between Class Be and De sites, the typically ranged from 0.7 to 3.5. These magnitude
ratio of Ch(T) between Be and De again is earthquakes would rarely cause damage to buildings
approximately 2. Thus, the size (and cost) of the unless the building was close to the epicentre and in
seismic-force-resisting system required in the poor condition. One earthquake near Mount Anakie
nominated building is influenced by the subsoil class. in 1977 had a measured magnitude of 4.7. Some
building damage could potentially occur at this
magnitude for inadequately designed structures or basalt rock in the Werribee Plains Lava Field is
those in a deteriorated condition. variable. A review of 55 published boreholes put the
basalt thickness at between 1m to 125m. Figure 3
Even though the first seismograph in Melbourne was shows the depth to the base of the basalt.
installed in 1888, more comprehensive recording of
seismic activity only commenced in 1976. Analysis of
seismic data in the Melbourne area suggests that
earthquakes are considered shallow (less than
17km) and are due to horizontal compressive stress
from south-east to northwest with movement along >100m
on reverse faults (Gibson et al. 1981). Given the
limited time scale of seismic recordings in Victoria in
comparison to a notional return period of moderate or >50m >20m
large earthquakes, it is possible that larger
earthquakes may occur in the future. Noting that the
general area would still be considered a region of
lower seismicity.
6 DIFFICULTIES ASSIGNING SUBSOIL Figure 3. Contour map of base depth Newer

Volcanics Basalt from 55 published boreholes
CLASS
(geo.vic)
6.1 Lack of a Definition for Rock The ‘rock’ subsoil classes have a secondary
requirement where the ground profile cannot be
Bedrock or rock is not defined in AS 1170.4-2007 and underlain with materials that have shear wave
as such the intent behind the subsoil class definitions velocities of less than 600m/s or 300m/s, or 18MPa
is unclear in the context of the Werribee Plains Lava or 0.8MPa UCS, for Class Ae and Be, respectively.
Field. The hierarchy for subsoil definitions stated in The Tertiary age soils underlying the basalt rock are
AS 1170.4-2007 is based on “surface geology and old (5 to 60 Mya) and are over-consolidated. The in-
estimates of depth to underlying rock”. This would situ confined strengths of these ancient sediments
suggest that for the Werribee Plains Lava Field, the are not well understood and can be difficult to assess.
nearest underlying rock is the basalt. A subsoil class In boreholes, these soils can range from stiff to hard
of Ae or Be could be assigned where the ground for cohesive soils and medium dense to very dense
profile also meets the other requirements of the for cohesionless soils. In some instances, the
classification. sediments could easily be assessed as very low or
low strength rock. Limited testing of shear wave
A possible definition of bedrock for an earthquake velocity has been undertaken on the Tertiary age
action standard could be the geological layer where soils. Roberts et al. (2004) has undertaken
the fault that is undergoing deformation or movement measurement of shear wave velocity of ground profile
is located and ergo, the source of the earthquake. For at one site in Altona. The results for the shear wave
the Werribee Plains Lava Field, the basalt rock is velocities are repeated in Table 2 for the inferred
relatively young and is unlikely to be tectonically materials encountered at that location, which are
stressed. Geologically speaking, the basalt is just shown in Figure 3. The results of the testing suggest
another layer of material in an otherwise deep that shear wave velocities are higher than a typical
sedimentary basin. The underlying Silurian age soil. However, the shear wave velocities and UCS
siltstone and sandstone are locally referred to as the requirements for Class Ae are unlikely to be met if the
‘bedrock’. The Silurian age bedrock is well known for basalt layer is thin and Tertiary age soils are
being folded and faulted, including by the previously encountered.
discussed Rowsley Fault. The Silurian age bedrock
is also known to be under high horizontal
compressional stress, as such, it could be the source
of an earthquake or at least be overlying the rock that
is deforming. The Devonian age granite is also
considered locally to be bedrock. If both Silurian and
Devonian age geologies are the ‘rock’ intended by
AS 1170.4-2007, then full ground profile through the
Tertiary sedimentary basin would need to be
considered in assigning a subsoil class.
6.2 Upper 30m and Shear Wave

Velocity of Materials
A Class Ae or Be subsoil can be assigned if the top

30m of the ground profile has an average shear wave
velocity of >1500m/s or > 760m/s, respectively.
Basalt rock is often high to very high strength with
zones of weaker rock. The average shear wave
velocity can easily be > 1500m/s. The thickness of
Table 2: Summary of Measured Shear Wave Velocity for Inferred Geological Units
Geological Unit Approximate Material Measured Shear
Age of Unit Description Wave Velocity
(Mya) (m/s)
Newer Volcanic Group (Qvn) 1-5 Basalt 1800
Newport Formation (Tmn) 8-25 Silt, sand, clay 500
(currently referred to as Gellibrand Marl)
Altona Coal Seam (Tea) 23-60 Brown coal, clay 1300
(a subunit of the Werribee Formation)
Werribee Formation (Tew) 23-60 Sand, silt, clay 500
Note: The table does not include all known Tertiary age deposits e.g., Brighton Group / Sandringham Sandstone.
Figure 4. Shear Wave Velocity from Roberts et al at Altona with extract from GSV 1974
Class Be could be assigned based on assumption the Table 3: Estimate of the Permissible Depth to
Tertiary soils have a shear wave velocity >300m/s Bedrock to Achieve a Certain Site Period
and a UCS >0.8MPa (or approximate undrained Average Maximum depth to top of
shear strength of 400kPa); and there is sufficient Shear Wave bedrock
basalt thickness to achieve the average shear wave
Velocity of Site Period of Site Period of
velocity of 760m/s in the top 30m.
Ground 0.2s 0.6s
6.3 Can Class Be apply if the natural Profile
site period is >0.6s? 1500m/s 75m 225m
760m/s 38m 114m
AS 1170.4-2007 is silent on maximum depth limits 300m/s 15m 45m
and natural site periods for the rock classes. Authors, Profile as per
including Poulos (1993), note that structural damage Roberts et al. 45m 103m
in buildings can be more prevalent when the natural (2004) Table 2
period of the building coincides with the natural period
of the ground. Designers may assume that by
assigning a Class Be a reasonably low (or fast) natural
site period would apply e.g., earthquake standards >50m
elsewhere in the world refer to “rock sites” as having >100m
natural site periods in the range of 0.1 to 0.2s.
>200m
AS 1170.4-2007 provides an upper limit on Class Ce
as a natural site period of 0.6s. If this site period of
0.6s is exceeded, then a Class De applies. A method >250m
for estimating site period is provided in AS 1170.4-
2007 as four times the shear wave velocity travel time
from surface to the underlying rock. Using this
methodology, Table 3 presents a summary of
estimated permissible depth to bedrock to achieve Figure 5. Depth to Top of Silurian and Devonian
certain site periods, e.g., where the ground profile Bedrock based on 13 published boreholes (geo.vic)
has an average shear wave velocity say of 760m/s,
then the estimated permissible depth to top of 6.4 The significance of a nominal 3m
bedrock is about 114m to remain within Class Ce surficial layer
definition of having a site period less than 0.6s. As
shown in Figure 5, the depth to Silurian and Devonian In AS 1170.4-2007, the difference between assigning
age bedrock is deep and typically exceeds the depths a subsoil class of Be or De could potentially rely on
in the table. the presence of a 3m thickness of soil strength
materials overlying the basalt rock, i.e., residual
clays, weathered materials or any alluvial or
windblown silts that were discussed previously. A 3m on a case-by-case basis. If earthquake load was a
thick layer of a surficial soil would not be expected to critical design case, then the building could be at risk
change the behaviour of the ground in an of damage from being under-designed, particularly if
earthquake. However, the impact from a design the building was located near an earthquake
perspective can be significant when considering epicentre.
force calculations presented in Section 4 of this
paper, where the resulting force estimates could Where Class Ce has been assigned and depending
double. on the size of the building and the Earthquake Design
Category, it is possible that earthquake loads were
With reference to the specific ground profile under-estimated by 10-20%. As a Class Ce seismic-
encountered by Roberts et al. (2004) in Altona force-resisting system would have been constructed,
(Table 2), the method of assigning a subsoil class this additional load may not induce as much damage
would be as follows: the basalt has a shear wave as the previous scenario where Class De site
velocity of approximately 1,800m/s over the top 30m, assigned as a Class Be.
so subclause (a) for both Class Ae and Be are
satisfied. For sub-clause (b), the strength of the 8 SUGGESTIONS MOVING FORWARD
Tertiary soils is not provided, but the shear wave
velocity profile indicates values between 500m/s and AS 1170.4-2007 could provide a definition of
1300m/s, so the clause (b) for Class Be is satisfied rock/bedrock so that an appropriate ground profile
but not Class Ae. Therefore, provided there is no can be determined. AS 1170.4-2007 could also
more than 3m of soil or highly/completely weathered provide a consistent criterion across all of the subsoil
material above the basalt (thickness not reported by classes, such as site period. A proposal for a site
Roberts et al. (2004)), then a Class Be applies. If the subsoil classification scheme has been prepared by
thickness of soil is exceeded, then the site period for the University of Melbourne for regions of lower
this profile did not encounter Silurian bedrock within seismicity (Tsang et al. 2015).
150m and is well above the 0.6s i.e., Class De
applies. The spatial extent of the issue can be defined by the
depth to the basement rock. At the north-eastern
6.5 Calculating Natural Site Period extent of the Werribee Plain Lava Field near
Footscray, where the development density and
It is acknowledged in AS 1170.4-2007 that the building height increases, the basement rock also
proposed method of estimating natural site period rises up and the calculated site period may be less
(four times shear wave velocity) is simplified. A deep than 0.6s. Where the site period is less than 0.6s,
sedimentary basin can either attenuate or amplify there is less uncertainty over the site subsoil class. It
seismic energy depending on the wave amplitude of may be advantageous to conduct investigations to
the nominated earthquake and the strength of the locate the top of the Silurian age bedrock as either
soils. A thick cap of basalt rock creates a shear wave part of the master planning for these areas or as part
velocity inversion that could potentially confine or of an AGS initiative. Drilling deep geotechnical
dampen the effects of seismic wave energy. Both of boreholes for an individual development on a case-
which could be considered in estimating the site by-case basis is unlikely to occur or be cost effective.
period, as suggested by Fabozzi et al. (2021).
However, it is not easy to predict seismic wave The properties of the underlying Tertiary age
energy in near-surface sediments without materials are not well understood. This is partly due
measurement. Unfortunately, buildings are not to most geotechnical investigations not penetrating
typically installed with seismographs to measure the basalt and the geological investigations that do
performance during an earthquake. penetrate the basalt are primarily interested in coal
and mineral deposits, so do not provide detailed
A separate issue with estimating natural site period strength data. As a result, properties of the Tertiary
using the method suggested in AS 1170.4-2007 is materials may need to be derived from non-intrusive
the basalt thickness is highly variable in the Werribee methods, such as seismic profiling (Roberts et
Plains Lava Field. The depth to sound rock can vary al. 2004).
significantly over short distances and the unit is
composed of multiple volcanic layers that each have To assign a subsoil of Class Be in critical areas within
experienced variable erosion and weathering. The Werribee Plains Lava Field, a borehole drilled to at
underlying base of the basalt is also variable, as it least 30m should be considered a minimum
reflects the relic topography of the Tertiary basin. requirement.
7 POSSIBLE IMPLICATIONS The impacts of earthquakes on buildings located on

deep sedimentary basins with shear wave velocity
For sites that have been assigned Class Be but may inversions are not well understood. Seismographs
have a natural site period in excess of 0.6s, then could be installed either on new buildings or
depending on the size of the building and the retrospectively installed on existing buildings to
Earthquake Design Category, it is possible that provide data to aid future design processes. Local
earthquake forces were under-estimated by 20-50% seismologists could review the data to provide
as discussed in Section 4. The earthquake load case guidance to designers.
may or may not be the critical load case from a design
perspective and impacts would need to be assessed
9 CONCLUSIONS
The guidance provided in AS 1170.4-2007 to assign

subsoil classes may require additional comment to
address natural site periods for the rock classes to
make it clearer to designers. Further investigations
and studies into methods of appropriately estimating
site period for Melbourne’s west and south-west are
required.
REFERENCES
AS 1170.4. (2007). Structural Design Actions, Part 4:
Earthquake Actions in Australia. Reconfirmed 2018.
Incorporating Amendments 1 & 2: Standards
Australia.
Boyce, Julie. (2012). Victoria erupts! The Newer Volcanics
Province of south-eastern Australia: a new database
of eruption points and an updated map.
Engineering geology of Melbourne : proceedings of the
Seminar on Engineering Geology of Melbourne,
Melbourne, Victoria, Australia, 16 September 1992 /
edited by W.A. Peck ... [et al.]
Eruption points of the newer volcanics province of Victoria :
an inventory and evaluation of scientific significance /
Neville Rosengren. Book. Bib ID, 2412822.
Fabozzi, S. Catalano, G. Falcone, G. Naso, A. Pagliaroli, E.
Peronace, A. Porchia, G. Romagnoli, M. Moscatelli,
Stochastic approach to study the site response in
presence of shear wave velocity inversion:
Application to seismic microzonation studies in Italy,
Engineering Geology, Volume 280,2021,
Geoscience Australia website https://www.ga.gov.au/
accessed September 2023
Gibson, G., Wesson, V., & Cuthbertson, R. (1981).
Seismicity of Victoria to 1980. Journal of the
Geological Society of Australia, 28(3–4), 341–356.
https://doi.org/10.1080/00167618108729173
GSV (1967) Geology of the Melbourne District, Victoria:
Explanatory Notes on the Stratigraphy, Structure and
Economic Geology to Accompany the Geological
Map of Melbourne and Suburbs (1959, Scale 1:31,
680) Mines Department, 1967
GSV (1974). Melbourne Sheet SJ55-1. Melbourne:
Geological Survey of Victoria and accompanying
notes
Geo.vic database website accessed September 2023
Leonard, John. (2022). GEOLOGY OF VICTORIA.
Poulos, H.G. (1993). Effect of earthquakes on settlements
and axial pile response in clays. Aust. Civil Eng.
Transactions, IEAust, Vol. CE35(1): 43-48.
Roberts, J., Asten, M., Tsang, H., Venkatesan, S., & Lam,
N. (2004). Shear Wave Velocity Profiling in Melbourne
Silurian Mudstone Using the SPAC Method. AEES
Conference. Mount Gambier, SA.
Tsang, H.H., Wilson, J.L., Lam. N.T.K., 2015.
Recommended Site Classification Scheme and
Design Spectrum Model for Regions of Lower
Seismicity. In: Proceedings of the Tenth Pacific
Conference on Earthquake Engineering, Sydney,
Australia, November 6-8, 2015.
Gas Permeability of Cement-Bentonite Cut-off Wall Materials with Powder and
Granular Bentonite
S. E. Al-Baiaty1, A. Bouazza2
1
Ph.D. Candidate, Department of Civil Engineering, 23 College Walk, Monash University, Vic. 3800, Australia;
PH +61 3 9905 0538; email: sarah.al-baiaty@monash.edu; Geotechnical Engineer, Tonkin + Taylor Pty Ltd,
Kings Technology Park, Level 3, 99 Coventry Street, Southbank, Vic 3006: SAl-Baiaty@tonkintaylor.com.au
2
Professor, Department of Civil Engineering, 23 College Walk, Monash University, Vic. 3800, Australia;
PH +61 3 9905 4956; email: malek.bouazza@monash.edu
ABSTRACT
A gas flow unified measurement system (UMS-G) was used to examine the gas permeability of cement-bentonite
(CB) mixtures prepared with powder and granular bentonite over various saturation states. The influence of
bentonite particle size was evident at volumetric water content > 50%; the CB mixture with granular bentonite
experienced less gas migration. This is due to the interfacial transition zone (ITZ) that could be present in the
microstructure of the CB mix with granular bentonite. Also, the granular bentonite contains a higher content of quartz
than the powder bentonite, which can provide a further pozzolanic reaction and denser microstructure to the CB
mixture. A reverse effect for the ITZ was found at 50% volumetric water content (VWC), resulting in severe fissure
cracks within the CB matrix compared to the CB mixture with powder bentonite that achieved a stable mixture
without cracks. At volumetric water content <50%, no significant difference was observed in gas migration results
between the CB mixes with powder and granular bentonite. The CB mixture with granular bentonite was disregarded
from the test at 40% VWC compared with 30% VWC for the CB mixture with powder bentonite due to the formation
of apparent cracks. Such findings can enhance the confidence in assessing the effect of bentonite particle size on
the gas permeability properties of the CB cut-off wall materials.
1 INTRODUCTION gradient of total pressure and diffusion in response to

a gas concentration gradient (Vangpaisal and
Cement-bentonite (CB) cut-off walls are widely used Bouazza 2004, Rouf et al. 2016). In a land-fill context,
to encapsulate contaminated sites (Xanthakos 1979, gas migration’s primary driving force is the pressure
D’Appolonia 1980, Jefferis 1981). The CB materials differential due to the water table/leachate level
provide a hydraulic barrier against pollution migration changes and the fluctuation in atmospheric pressure
by controlling contaminants/gas movements (Jefferis (Vangpaisal and Bouazza 2004). This paper explores
1981, Ryan and Day 1986, Jefferis 1993, Jefferis the advective gas flow through typical CB material.
1997, Tedd and Jefferis 2000, Evans et al. 2002,
Ryan and Day 2002). The slurry mixes generally The paper explores the influence of bentonite
contain bentonite, cement, blast furnace slag or fly type/mineralogical properties and their particle size
ash and other fine materials such as silt and sand on CB gas permeability and identifies the relevant
(Ryan and Day 2002, Daniel and Koerner 2007, parameters that might influence their gas
Daniel 2012, Jefferis 2013). The gas sealing integrity permeability. The study was conducted on two CB
of the section of the cement-bentonite (CB) cut-off mixtures prepared using powder and granular
wall above the groundwater table can become bentonites and investigated the effect of cracking
problematic due to its unsaturated state, as the caused by drying on gas permeability.
potential presence of a dry area can result in
microcracks formation. Consequently, this can 2 MATERIALS AND METHODS
degrade the serviceability of the CB wall as a barrier
against gas migration. The microcracks potentially 2.1 Materials
develop a gas-filled system of vapour or gases that
easily migrate through the air void. Two types of bentonite, Volclay powder sodium
bentonite and VSP granular bentonite and a special
Previous studies (Tedd and Jefferis 2000, Ratnam purpose cement commercially available in Australia,
2002, Opdyke and Evans 2005, Evans and Opdyke known as marine cement, were used in this study.
2006, Joshi et al. 2008, Joshi et al. 2010, Jefferis Bentonite mineralogy was obtained using the
2012, Soga and Joshi 2015) have identified the quantitative X-ray diffraction (XRD) analysis of the
presence of potential drier sections in CB walls; their bulk and <2μm fraction (Table 1). The Volclay
work has only considered the strength, stress-strain bentonite contained the highest percentage of
behaviour and permeability of CB materials under montmorillonite, which was within the typical range of
saturated conditions. The formation of micro-cracks montmorillonite content of sodium bentonite (>70%)
in the part of the cut-off wall close to the ground used in hydraulic barrier systems (Gates et al. 2009),
surface can exacerbate gas or vapour migration (ICE compared with the VSP bentonite, which contained
1999, Joshi et al. 2008, Joshi 2010, Cermak and Law 66% of montmorillonite. VSP bentonite had a higher
2011, Cermak et al. 2012, Soga and Joshi 2015). The percentage of cristobalite (i.e. quartz family rich with
mechanism of gas movement in a porous system is silica) than the Volclay bentonite. The percentage of
mainly governed by advective flow in response to a particle size at <2μm fraction is higher in VSP
Gas Permeability of Cement-Bentonite Cut-off Wall Materials with Powder and Granular Bentonite S. E. Al-Baiaty and A. Bouazza
bentonite than the Volclay bentonite, indicating the Jefferis 2013, Carreto 2014). After 48 hours of
presence of a high percentage of granular bentonite hydration, the bentonite suspension's rheological
in VSP. properties were measured to identify the minimum
viscosity achieved during hydration, as Jefferis
The cement used in the current investigation (2013) recommended. Two methods were used: the
complies with AS3972, LH/SR type, special purposes indirect method (Marshal funnel) and the direct
cement, and it is composed of cement clinker and method (Viscometer). Table 2 summarises the
ground granulated blast furnace slag (GGBSF) (40 viscosity results. The indirect method was found
cement/60 GGBSF) and gypsum. It is typically used unreliable for both mixtures (i.e. foam from bentonite
in applications where an aggressive sulphate-rich suspensions blocked the flow of the suspensions
environment and other aggressive salt and solution from the funnel). The apparent viscosity from the
are expected. direct method was used as the indicative parameter
for the suspension rheological properties.
Table 1: Mineralogical composition of the bentonites
Volclay VSP In stage two, cement was added to the bentonite
Chemical
(P) (G) suspension and mixed for a time not less than 5
Bulk Ca saturated minutes to form a homogenous CB mixture. The
Quartz % 12 11 fluidity and workability properties of the CB slurry
Cristobalite % 1 10 were measured using a viscometer (Table 2). It can
Montmorillonite % 78 66 be seen from Table 2 that the apparent viscosity was
Calcite % <1 - significantly increased for both CB-P and CB-G due
Albite/ Anorthite % 7 8 to the increase of solid content in the mixture after
Kaolin % 1 <1 adding the cement. Also, the viscosity results at both
Anatase % - <1 stages are higher in the CB mixture with powder
bentonite (CB-P) than the CB mixture with granular
Zeolite % 1 5
bentonite (CB-G) due to the highest montmorillonite
Particle Size >2µm 24.4 39.2
content in powder bentonite.
<2µm fraction
Quartz % 3 2 The set properties of the CB mixture were
Cristobalite % <1 9 characterised by measuring the mixture’s slurry
Montmorillonite % 95 86 density and filter loss. The density measurement was
Calcite % - - conducted using a known cylindrical volume, and the
Albite/ Anorthite % 1 <1 filter loss test was by using a cylinder graduated in 20
Kaolin % 1 1 ml up to 100ml as recommended by ICE (1999).
Anatase % <1 <1 Typically, the density of CB materials should be
Zeolite % - 1 around 1220kg/m3, and the filter loss not be more
Particle Size <2µm 75.1 55.7 than 2% within 24 hours. The CB mixes set
CEC (cmol/Kg) by Ba and properties were within the requirement (ICE 1999).
93 77
XRF
Mineralogical test was done by Mineralogical Services The CB mixtures were poured inside PVC tubes
(CSIRO), Land and Water Division, South Australia closed from both ends to prevent moisture losses.
The PVC tubes were kept inside a fog room
2.2 Experimental methods (RH=100%, T=20oC) for curing. The mechanical
properties of the CB materials were measured after
2.2.1 CB mixture preparation curing at 28 days. The hydraulic conductivity was
measured at 90 curing days as recommended by ICE
The CB slurry mixture was prepared to a standard (1999). ICE (1999) specifies that unconfined strength
specification according to the Institution of Civil at 28 days should be more than 100kPa, and
Engineers, UK guidelines (ICE 1999), aided by other hydraulic conductivity is recommended to be 10-8 to
studies (Daniel and Koerner 2007, Jefferis 2012, 10-9 m/s at 90 days. In the current study, the strength
Jefferis 2013, Lam and Jefferis 2018). The mixture for both CB mixtures was within the recommended
included 7% bentonite and 20% cement of the total value; the strength of the CB-P mixture was 792kPa
weight of water. The preparation was done in two and 1515 kPa for the CB-G mixture. The permeability
stages; the first stage aimed at forming a bentonite to water of CB-P was 4.1 x 10-10 m/s and 3.06 x 10-9
suspension (BS) by mixing the bentonite with water m/s for CB-G; both values were within the specified
for not less than 5 minutes. Once formed, the range recommended by ICE (1999).
suspension was kept inside closed containers for
about 24 hours for hydration purpose. It should be 2.3 Gas permeability measurements
noted that granular bentonite (VSP) was remixed
after 24 hours of hydration and rehydrated for The gas permeability test was conducted using the
another 24 hours, as the accumulated hydrated Monash unified measurement system (MUMS-G)
bentonite particles were observed due to the developed by Rouf et al. (2016). The set-up
bentonite particle size. Sufficient hydration time measures sequentially gas diffusion and
allows the bentonite particles to be dispersed, permeability.
supporting the cement grains later during the
cementitious system hydration (Jefferis 2012,
Table 2: Rheological properties of bentonite suspensions and cement-bentonite slurries.

Mixture with powder bentonite Mixture with granular bentonite
Viscosity Tests (CB-P) (CB-G)
Bentonite Cement- Bentonite Cement-
Suspension Bentonite slurry Suspension Bentonite slurry
cP, 600 (rpm) 55 116 15 60
cP, 300 (rpm) 39 89 9 49
Gel strength, (N/m2) 7.6 52.8 4.3 24.5
Apparent viscosity, cP
27.5 58 7.5 30
600/2
The set-up comprises three major parts made of The results were plotted against volumetric water
acrylic: a top cylinder of 100mm inner diameter and content (VWC). One can observe that there is a
70mm height, a middle cylinder to accommodate the marked difference between CB-P and CB-G at VWC
sample, and a bottom cylinder of 100mm inner range between 75% up to 50%; where the gas
diameter and 95mm height. The bottom chamber permeability of the CB-P mixture is higher than the
included two valves and the oxygen and pressure CB-G mixture; beyond 50% VWC, there is no marked
transducer sensors. difference between CB-P and CB-G. This is
attributed to factors influencing the gas permeability
One of the valves in the bottom chamber was used of CB mixtures such as porosity, bentonite type and
to pressurise the gas into the bottom chamber. In the mixture sensitivity to the external environment; for
contrast, the second valve was used to expel it example, drying can cause cracks that affect the
during the gas diffusion test (i.e. oxygen diffusion). permeability process. As a cement-based material,
The top cylinder contained a movable rod and six the porous system interconnectivity can affect its gas
ports. Only one port can be opened during the gas transport properties (Jennings 2008, Jennings et al.
permeability test, and an interchangeable gas flow 2015, Wu et al. 2017). The pores system in cement-
meter was attached to this port to measure the gas based materials is referred to as capillary pores (i.e.
flow through the sample. Different ranges of flow between 8nm to 10μm),
meters were connected to this port to measure the
gas flow through the sample, depending on the CB
1E-6
degree of saturation.
CB-P
CB-G
Nitrogen gas (N2) was used for the gas permeability
1E-7
measurement because it is inert and has low water
Gas Permeability, m/sec
solubility (μ=1.76x10-5 N s m-2, ρ=1.165 kg m-3 at room

temperature). A differential pressure range of 1E-8
1-10kPa was used to measure the gas permeability,
as the pressure is unlikely to exceed 10kPa in
containment settings (McBean et al. 1995). A 1E-9
pressure transducer measured the inlet pressure
once the gas was injected into the bottom chamber.
1E-10
A gas flow meter was attached to an outflow port to
monitor the flow of gas that migrated through the CB
sample. The gas permeability was calculated using 1E-11
20 30 40 50 60 70 80 90
the modified Darcy law; more details about the setup
VWC, %
and testing procedure can be found in Bouazza and
Vangpasial (2003) and Rouf et al. (2016). Figure 1. Variation of gas permeability versus VWC
of CB mix
The CB sample preparation involved several steps;
first, the CB sample was extruded from the PVC tube small pores are known as gel pores (i.e. <10 nm), and
once it reached 28 curing days. The sample was then the space between the hydration products sheets
sized and placed inside a humidity-controlled (i.e.<2 nm) (Jennings 2008, Jennings et al. 2015).
desiccator (RH=60%) to dry the CB sample. It should The physio-chemical activities between bentonite
be noted herein that the humidity-controlled and cement during the hydration process are
desiccator drying method allows the CB sample to be believed to affect the CB pores' structure formation
dried gently to avoid any cracks; the duration was significantly. Dissolution/precipitation and
around four months and started when the CB sample decalcification (i.e. calcium leaching) are the
completed the 28 days of curing. Once the sample standard chemical activities processes that induce
achieved the required moisture state, it was moved changes to the porous system microstructure (i.e.
to another desiccator for one day for moisture physical changes) and influence the materials
homogenisation. transport properties (John et al. 2019). More details
on the chemical reaction between bentonite and
3 RESULTS AND DISCUSSIONS cement can be found in Carreto (2014)
The gas permeability (kg) for both CB-P and CB-G The granular bentonite and its mineralogical
mixtures exposed to drying is presented in Figure 1. properties in the CB-G mixture can provide high
resistance against gas transport compared to the accumulate high ettringite, which could cause
CB-P mixture. The bentonite of the CB-G mixture expansion and is considered the main reason for
contains a high percentage of cristobalite (i.e. quartz the cracking in ITZ. In the case of the CB materials,
family), which can facilitate the pozzolanic reaction ITZ was observed under the secondary electron
and develop a denser microstructure. In an microscopic (SEM) (Figure 3). This is evidence that
environment rich with pozzolanic, ettringite forms bentonite is not entirely dissolved during the
quickly; this can provide more filling and denser interaction with cement, which contradicts what was
internal structure to the porous system and observed by other researchers (Evan 2021, Jefferis
decelerate calcium leaching through a denser 2012). Granular bentonite can provide a better
structure network, developing a chemically and barrier against the gas flow at high saturation levels
physically stable barrier. In contrast, the bentonite of because it has high porosity between (10 and 50
CB-P contains a high percentage of montmorillonite um) that can facilitate moisture ingress through the
that, in turn, can increase the calcium percentage in formation of ITZ. Still, it can cause fissure cracks at
CB pore solution, causing high carbonation. a moderate saturation level.
Carbonation can cause dimensional changes to the
porosity (i.e. disordered structure) of the porous At 40% VWC, the CB-G mix was disregarded from
system through high dissolution/precipitation to the the test due to the apparent cracks (i.e. around 2 mm,
cementitious system hydration products and see Figure 4). In contrast, the CB-P mix experienced
influence the porous system gas transport properties fissure cracks at 40% VWC and was disregarded
(Jain and Neithalath 2009). Carbonation can also from the test at 30% VWC.
induce finer cracks in the porous system and
increase the porous system’s gas migration.
At VWC varying from 75% to 60%, the gas

permeability of the CB-P mixture was about one
order of magnitude higher than the CB-G mixture.
This is due to the effect of granular bentonite on the
CB-G mixture’s internal structure matrix and gas
permeability properties. This study infers that the
granular bentonite required more time to dissolve
than the powder bentonite; this, in turn, can develop
an interfacial transition zone (ITZ) within the CB-G
porous system. The ITZ is the weakest phase in a
cementitious system, located between cement grain
and aggregate (Care and Derkx 2011, Sicat et al.
2014). The ITZ is treated as a separate phase from
the bulk matrix; it contains different microstructure
(i.e. morphology and density) and other properties Figure 2. Fissure cracks appear in the CB-G mixture
from the matrix (Sicat et al. 2014). Previous studies at 60% volumetric water content (VWC)
(Care and Derkx 2011, Sicat et al. 2014) found this
phase is weak and retains high moisture due to its
large thickness, which can facilitate the pozzolanic
reaction and provide less gas migration. The
thickness of the ITZ is described as between 10 and
50 µm; the high porosity is considered the key to the
weakness and the high moisture present within this
phase (Care and Derkx 2011, Sicat et al. 2014); the
porosity of ITZ depends on water/binder ratio,
curing days and the content of granular materials
within the mixture.
At a VWC of less than 60%, the difference in gas

permeability of both CB mixes becomes less than
one order of magnitude. The CB mixes start offering
the same resistance against the gas flow due to the
fissure cracks appearance in the CB-G mixture
(Figure 2). The CB-G mixture experienced a severe
fissure crack (Figure 2) at VWC of around 60%
compared to the CB-P mixture, which achieved a
stable mix without cracks. The appearance of
cracks in the CB-G mix is due to the interfacial Figure 3. Interfacial Transition Zone (ITZ) in CB-G
transition zone (ITZ) effect. Golewski (2018) and mixture.
Sun et al. (2018) indicated that ITZ could
Carreto, J. (2014). Self-hardening slurry walls. A contribution

for design, quality control and performance monitoring
(Doctoral dissertation, Ph. D. thesis, Dept. of Civil
Engineering and Architecture from IST, Lisbon Univ.,
Lisbon, Portugal).
Cermak, J., J. Evans and G. J. Tamaro (2012). Evaluation of
Soil-Cement-Bentonite Wall Performance - Effects of
Backfill Shrinkage. Grouting and Deep Mixing 502-511.
Construction Industry Research, Information Association, &
Building Research Establishment. (1999). Specification
for the construction of slurry trench cut-off walls as
barriers to pollution migration. Thomas Telford.
D’Appolonia, D. J. (1980). “Soil-bentonite slurry trench cut-
offs.” Journal of Geotechnical and Geoenvironmental
Engineering 106(ASCE 15372).
Daniel, D. and R. Koerner (2007). “Vertical cut-off walls.” A.
Press, Waste containment facilities: Guidance for
construction quality assurance and construction quality
Figure 4. Apparent cracks were identified in the CB-G control of liner and cover system: 281-311.
mixture at 40% volumetric water content (VWC) Daniel, D. E. (2012). Geotechnical practice for waste
disposal, Springer Science & Business Media.
Evans, J. C. and S. M. Opdyke (2006). Strength,
4 CONCLUSION Permeability, and Compatibility of Slag-Cement-
Bentonite Slurry Wall Mixtures for Constructing Vertical
The gas permeability of CB materials was measured Barriers. 5th ICEG Environmental Geotechnics:
using a gas unified measurement system (UMS-G). Opportunities, Challenges and Responsibilities for
Gas permeability was conducted on two CB mixtures Environmental Geotechnics: 118-125.
prepared with two types of bentonite, powder and Evans, J. C., Larrahondo, J. M. & Yeboah, N. N. N. (2021)
granular bentonite. Fate of bentonite in slag–cement–bentonite slurry
trench cut-off walls for polluted sites. Environmental
Geotechnics, 40(XXXX), 1-13.
It was found that at VWC higher than 50%, the effect Care, S. and Derkx, F., 2011. Determination of relevant
of granular bentonite was more pronounced on CB parameters influencing gas permeability of
gas permeability. The CB-G mixture with granular mortars. Construction and Building Materials, 25(3),
bentonite experienced reduced gas permeability than pp.1248-1256.
the CB-P mixture with powder bentonite. The Evan J. C., A. R. Dawson and S. Opdyke (2002). Slurry walls
granular bentonite can develop an ITZ within the CB for groundwater control: a comparison of UK and US
material microstructure, increasing water ingress and practice. Geotechnical conference, ASCE/PENNDOT
facilitating a pozzolanic reaction. Also, granular Central PA, USA, Citeseer.
Gates, W.P., Bouazza, A. and Churchman, G.J., 2009.
bentonite contains more silica than powder Bentonite clay keeps pollutants at bay. Elements, 5(2),
bentonite, which can provide a denser pp.105-110.
microstructure. Golewski, G.L., 2018. Evaluation of morphology and size of
cracks of the Interfacial Transition Zone (ITZ) in
At a moderate saturation level of 60%, the CB-G concrete containing fly ash (FA). Journal of hazardous
mixture experienced fissure cracks due to the ITZ's materials, 357, pp.298-304.
susceptibility to moisture loss. However, the CB-G Jefferis, S. (1981). Bentonite-cement slurries for hydraulic
mixture still experienced slightly less gas migration cut-offs. Proceedings, Tenth International conference
on soil mechanics and foundation engineering,
than the CB-P mixture. At VWC of less than 50%, no Stockholm, Sweden.
difference was observed between the gas Jefferis, S. (2012). Cement-Bentonite Slurry Systems.
permeability of both CB mixtures. The CB-G was Grouting and Deep Mixing New Orleans, Louisiana,
disregarded from the test at 40% VWC due to United States: 1-24.
apparent cracks. The CB-P mixture achieved a stable Jefferis, S. (2013). Grouts and slurries. Construction
mix without cracks, and it was disregarded from the materials reference book. D. Doran, & Cather, B.
test at 30% VWC due to the evident cracks. Routledge, Taylor and Francis: 165-194.
Jefferis, S. A. (1993). In-ground barriers. Contaminated Land-
Problems and Solutions. D. M. Hobson, & Cairney, T.
5 ACKNOWLEDGEMENTS London and New York, Champan & Hall: 111-140.
Jefferis, S. A. (1997). The origins of the slurry trench cut-off
Our appreciation is extended to Bentonite’s and a review of cement-bentonite cut-off walls in the
supplying companies and Boral Pty Ltd for supplying UK. United States: 52-61.
the cement. The first author would like to thank Joshi, K. D. (2010). Long-term engineering performance and
Tonkin+Taylor Ltd Pty for their support in finishing In-Situ assessment of cement-bentonite cut-off walls,
this paper. University of Cambridge.
Joshi, K., C. Kechavarzi, K. Sutherland, M. Y. A. Ng, K. Soga
and P. Tedd (2010). “Laboratory and In Situ Tests for
REFERENCES Long-Term Hydraulic Conductivity of a Cement-
Bentonite Cut-off Wall.” Journal of Geotechnical and
Bouazza, A. & Vangpaisal, T. (2003) An apparatus to Geoenvironmental Engineering 136(4): 562-572.
measure gas permeability of geosynthetic clay liners. Joshi, K., K. Soga, M. Y. A. Ng and C. Kechavarzi (2008).
Geotextiles and Geomembranes, 21(2), 85-101. “Durability Study of Eleven Years Old Cement-
Cermak, J. and M. Law (2011). Cut-off performance of a Bentonite Cut-Off Wall Material.” Proc., GeoCongress
soil-cementbentonite wall. Proceedings of 14th Pan- 2008: The Challenge of Sustainability in the
American Conference on Soil Mechanics and Geoenvironment: 620-627.
Geotechnical Engineering, Toronto, Canada.
Jennings, H. M. (2008) Refinements to colloid model of Wu, Z., Wong, H. S. & Buenfeld, N. R. (2017) Transport
CSH in cement: CM-II. Cement and Concrete properties of concrete after drying-wetting regimes to
Research, 38(3), 275-289. elucidate the effects of moisture content, hysteresis
Jennings, H. M., Kumar, A. & Sant, G. (2015) Quantitative and microcracking. Cement and Concrete Research,
discrimination of the nano-pore-structure of cement 98, 136-154.
paste during drying: New insights from water sorption Xanthakos, P. P. (1979). Slurry walls, McGraw-Hill Book
isotherms. Cement and Concrete Research, 76, 27- Company, New York.
36. Zhang, Y., & Zhang, M. (2014). Transport properties in
John, E., Epping, J. D. & Stephan, D. (2019) The influence unsaturated cement-based materials–A
of the chemical and physical properties of CSH seeds review. Construction and Building Materials, 72, 367-
on their potential to accelerate cement hydration. 379.
Construction and Building Materials, 228, 116723.
Jain, J. and N. Neithalath (2009). “Analysis of calcium
leaching behavior of plain and modified cement
pastes in pure water.” Cement and Concrete
Composites 31(3): 176-185.
McBean, E. A., Rovers, F. A. & Farquahar, G. J. (1995)
Solid waste landfill; engineering and design. Prentice
Hall.
Opdyke, S. M. and J. C. Evans (2005). “Slag-Cement-
Bentonite Slurry Walls.” Journal of Geotechnical and
Geoenvironmental Engineering 131(6): 673-681.
Ratnam, S. (2002). Development of a novel self-boring
permeability measurement technique, University of
Cambridge.
Rouf, M. A., A. Bouazza, R. M. Singh, W. P. Gates and R.
K. Rowe (2016). “Gas flow unified measurement
system for sequential measurement of gas diffusion
and gas permeability of partially hydrated
geosynthetic clay liners.” Canadian Geotechnical
Journal 53(6): 1000-1012.
Ryan, C. R. and S. R. Day (2002). Soil-cement-bentonite
slurry walls. Deep Foundations Congress 2002 -
International Perspective on Theory, Design,
Construction, and Performance_, Orlando, FL.
Sicat, E., Gong, F., Ueda, T. and Zhang, D., 2014.
Experimental investigation of the deformational
behavior of the interfacial transition zone (ITZ) in
concrete during freezing and thawing
cycles. Construction and Building Materials, 65,
pp.122-131.
Soga, K. and K. Joshi (2015). “Cement bentonite cut-off
walls for polluted sites.” Multilevel Modeling of Secure
Systems in QoP-ML: 149.
Sun, D., Wu, K., Shi, H., Zhang, L. and Zhang, L., 2018.
Effect of interfacial transition zone on the transport of
sulfate ions in concrete. Construction and Building
Materials, 192, pp.28-37.
Rouf, M. A., Bouazza, A., Singh, R. M., Gates, W. P. &
Rowe, R. K. (2016a) Gas flow unified measurement
system for sequential measurement of gas diffusion
and gas permeability of partially hydrated
geosynthetic clay liners. Canadian Geotechnical
Journal, 53(6), 1000-1012.
Tedd, P. and S. A. Jefferis (2000). The use of slurry trench
cut-off walls to repair embankment dams in the UK.
Proceedings of the biennial conference of the BDS
held at the University of Bath on 14-17 June 2000:
425-438.
Vangpaisal, T. and A. Bouazza (2004). “Gas permeability of
partially hydrated geosynthetic clay liners.” Journal of
Geotechnical and Geoenvironmental Engineering
130(1): 93-102.
Wu, Z., Wong, H. S., & Buenfeld, N. R. (2015). Influence of
drying-induced microcracking and related size effects
on mass transport properties of concrete. Cement
and Concrete Research, 68, 35-48.
SESSION 3
ROCK BEHAVIOUR IN
MINING APPLICATIONS
Novel Solutions in Utilising Underground Coal Seam Gas Reservoirs
for Gas Storage/Recovery
M. S. A. Perera1
Department of Infrastructure Engineering, Building 175, Grattan Street, Parkville, The University of Melbourne, Victoria 3010
1
Australia; PH +61 3 9035 8649; email: samintha.perera@unimelb.edu.au
ABSTRACT
Today, the world is moving to a green energy future by targeting net zero carbon emissions, which wouldn't be
successful without the support of underground reservoirs. Geological reservoirs, such as saline aquifers and coal
seam gas reservoirs, play an essential role in decarbonisation by providing ideal candidates to store CO2. On the
other hand, during the transition period of moving to renewable energy from fossil fuels, natural gas extracted
from underground can play a major role as a low-emission fuel. However, when underground reservoirs,
particularly low-stiffness geological reservoirs such as coal seams, are used for gas storage and extractions,
numerous issues may arise due to extreme pressures, temperatures, and geological and environmental
complexities. This paper focuses on potential novel solutions to overcome such challenges. Our research focuses
on minimising reservoir stimulation-related issues, such as effective proppants usage with minimising embedment
and poor distribution, focusing on damage mechanisms, fluid dynamics (e.g., flow complications caused by
inertia) and the geological reservoirs' characteristics. Also, according to our research, some CO2 storage issues in
coal seam gas reservoirs can be minimised by reducing the swelling effect by controlling injection scenarios and
introducing a stream of inert gas (e.g. nitrogen) after each CO2 cycle. This paper thus provides an excellent
summary of the issues and opportunities to overcome those issues during underground coal seam gas reservoirs
utilisation during the gas storage and recovery.
Keywords: underground utilisation, gas storage, gas recovery, challenges, novel solutions
1 INTRODUCTION 2 NATURAL GAS EXTRACTION FROM

UNDERGROUND
The world has abundant suitable geological
sinks/sources for natural gas (e.g., coal seam, 2.1 Issues arise during gas extraction
shale, and sandstone reservoirs). Importantly,
natural gas extracted from deep underground Coal Seam Gas (CSG) extraction has been a
reservoir rocks (e.g., coal seam gas (CSG), shale significant focus in Australia due to its economic
gas, tight gas) is the second largest energy source benefits and availability. Importantly, CSG can be
currently used for global power generation (24% extracted from relatively shallow seams, and CSG
contribution), which is, thus, play a critical role in the reservoirs contain a greater amount of gas in place
worldwide energy mix (IEA 2023). Also, with its high in a given mass than other natural gas reservoirs,
energy-carrying capacity and comparatively lower such as shale. According to URS (Ltd 2014), the
carbon emission (e.g., 50% less than coal), natural available economic demonstrated resources (EDR)
gas can significantly contribute to achieving the of coal seam gas in Australia estimated in 2007
green energy future expected by the world today. (35,905PJ) is equivalent to around ten times the
However, there is currently a huge mismatch total yearly energy used in Australia in that year.
between supply and demand. However, much of the
gas is trapped in the pore space of essentially The hydrostatic pressure given by the pore water in
impermeable rocks, and fracture stimulation or the coal seams and the overlying saturated rock
hydraulic fracturing is therefore required to achieve hold the gas in place, and therefore the reduction of
commercially viable production rates. However, pore pressure releases or desorbs the CSG (CH4),
successfully applying such stimulation techniques enabling its extraction for use (Ltd 2014). This pore
has many complications to overcome. pressure reduction can be achieved by pumping out
the pore fluid (mostly water) through a well
On the other hand, deep underground is rich with traditionally used for CSG extraction. However, this
geological sinks for energy/waste storage (e.g., traditional CSG production technique harvests less
Hydrogen/CO2). Deep coal seam gas reservoirs than 50% gas in place due to the associated
have been utilised for the CO2 storage process increased effective stress issue, which eventually
worldwide. However, these reservoirs have unique causes the reservoir permeability to be significantly
issues that substantially hinder their storage depleted (Day et al. 2000).
capacity. For example, the injection of CO2 into coal
seam gas reservoirs causes the coal matrix to swell, This indicates the importance of applying
resulting in significant pore shrinkage. Suck appropriate reservoir permeability enhancement
modification in the coal structure causes a techniques, such as hydraulic fracturing of the coal
substantial reduction in its CO2 storage capacity. seam gas reservoir. During fracturing, a pressurised
fluid, typically water, is injected to create a new
Go to table of contents 1
Novel Solutions in Utilising Underground Coal Seam Gas Reservoirs for Gas Storage/Recovery M. S. A. Perera
fracture network in the targeted reservoir to According to our research, it is not worth using
enhance permeability. However, this fracture proppants at shallow depths as it reduces the
network is rapidly closed due to the extreme conductivity through proppants clogging in fractures.
pressure at deep underground reservoirs. To For example, proppants should only be used after
overcome this issue, tiny particles known as around 300m depth for bituminous coal. Likewise,
proppants are mixed with the fracturing fluid. These conducting a preliminary investigation on the
proppants can keep open the fracture after effectiveness of proppants for any given reservoir is
releasing the fracking fluid upon completion of the worthwhile to do. Further investigation showed that
fracking process, offering permeable pathways for proppants with irregular shapes, like sand, are
the CSG to flow to the wellbore. effective in shallow depths, and regular, spherical-
shaped ceramics proppants are effective at deeper
2.1.1 Improper proppants usage issues depths (Figure 3).
Coal is a very soft material with a very small elastic

modulus compared to other reservoir rocks. This
causes proppants to embed into the surface,
shrinking the fracture opening and reducing
conductivity (Ahamed 2021). On the other hand,
proppants occupy the fracture space by obstructing
the flow paths. Thus, the optimum usage of
proppants in coal seams has always been
challenging.
Figure 1. Fracture conductivity reduction is caused

by proppants a) embedment and b) clogging
Figure 4. Damage mechanisms of sand proppants

under different confining pressures (Ahamed 2021)
Figure 2. The impact of confining pressure or depth

on the effectiveness of proppants (Ahamed 2021)
Figure 5. Damage mechanisms of ceramic

Figure 3. Improvement of effectiveness by using a proppants under different confining pressures
suitable proppant type (Ahamed 2021) (Ahamed 2021)
As can be seen from Figure 4, under low As shown in Figure 7a, slowdown or minimised flow
confinement, sand proppants can rotate and re- through the branching fractures minimises the
arrange to provide spacing for fluid flow and under proppants entering them, resulting in the distribution
high confinement, in addition to proppants of proppants only along the major fractures, which
embedment, sharp contact points cause stress explains the reason for poor proppants distribution.
concentration and breaking the sand particle,
generating fine and blocking the flow paths. Due to The natural gas industry spends millions of dollars
its high strength and regular shape, this does not on proppants and injection schemes with
happen in ceramic proppants (Figure 5), so it works hydrodynamics that might not be favourable. For
better under high confinements. example, fracture stimulation of a typical
unconventional reservoir well may consume over 4
2.1.2 Inertia effect on fracture flow behaviour million pounds of proppants, costing over 1 million
dollars (Earthworks 2023). Thus, understanding the
The fluid inertia caused non-Darcy flow or fluid flow real fluid dynamic in a fracture network is essential.
transition from regular linear flow to nonlinear flow Opening this black box will provide the crucial
behaviour, significantly impacting the flow behaviour scientific foundations to rigorously engineer effective
through fractured geologically reservoirs (naturally fracture stimulation and proppant usage to direct
fractured or hydraulically fractured). The inertia proppants into secondary/tertiary fractures, reducing
causing flow irregularities (Figure 6) creates the wasteful disposition of proppants.
complicated flow behaviours in underground
reservoirs, making it hard to describe the fluid flow Further, formation damage and pore plugging in
through rock fractures through simple theories. reservoirs due to fines migration are commonly
associated with oil/gas wells. Several studies have
shown the contribution of fluid inertia effect on fines
migration mechanism with high flow rates employed
in reservoirs (Zeinijahromi et al. 2012, Bai et al.
2015, Miri et al. 2021).
This aspect has been thoroughly investigated

experimentally by developing a Hele show cell
experimental set-up (Figure 8) and numerically by
creating a large-scale fracture network in ANSYS
and CSMP++ software platforms (Figs.9-10).
Figure 6. Various complex flow patterns develop in

fractures at high fluid velocities
During the hydraulic fracturing process, fluid injects

at very high flow rates. Therefore, highly
complicated flow behaviours involving irregularities
such as eddies, vortices, and resulting backflows
should be expected. Importantly, inertia causes
resistance to movement and moves significantly at
bends, thus obstructing the fluid flow through
branching fractures in a fracture network. This
impacts the fracturing process, avoiding fluid, and Figure 8. The Helle-show cell experimental set-up
hence, the fracture network distribution and poor used to visualise the inertia effect on fluid flow
proppants distribution. through the fracture network and that impact on the
particle transport process
Importantly, our modelling study on the ANSYS
platform shows that such flow irregularities caused
by fluid inertia can happen in very low velocities in
the presence of sudden flow path variations, such
as bend and fracture intersects (Figure 7).
Figure 9. Simulated large-scale fracture network

Figure 7. a) The uneven proppants distribution and with the used boundary condition in ANSYS
velocity distribution in a 2D model of a branching software (Hansika 2023)
fracture when; inlet velocity is, b) 0.01ms-1, c) 1ms-1
Moreover, it was identified that flow partitioning

between sets and velocity spectra varies with the
flow rate, as flow patterns in fracture networks are
highly dependent on the inertial effect. The overall
outcome of this research provided a comprehensive
knowledge of fluid flow dynamics in real fracture
networks by considering the inertial impact. These
findings will benefit the petroleum industry,
geothermal energy extraction, waste disposal and
groundwater management to enhance the well-
performance, maintaining commercially viable rates.
The study recommends having a complete bed
model to study the flow patterns of the selected
reservoir to understand such issues so that some
controlling measures can be taken with controlling
fluid velocities and proppants properties (size,
shape, concentration, etc.).
3 GAS STORAGE IN UNDERGROUND

COAL SEAM GAS RESERVOIRS
Figure 10. Inertia causing instabilities at fracture
intersections (10-3 m3m-2s-1 inlet flux) (Hansika 2023) 3.1 Issues arise during gas storage
The effect of intersection angle and aperture on flow Among multiple ways of mitigating atmospheric CO2
in simple fracture intersections was analysed for a content, CO2 sequestration has gained growing
broader range of injection rates. The study found attention in various carbon capture and storage
that nonlinearity originates from the inertia effect (CCS) options in multiple geological formations. In
enhanced by the intersection angle, which becomes this context, CO2 sequestration in coal reservoirs is
critical with the increasing flow rate (Figure 11). identified as an effective method because the
process is associated with enhanced coalbed
methane (ECBM) extraction – which further offsets
the cost of CO2 storage in coal seams. Even though
a century-long exploration of physics has shaped
the fundamental scientific underpinnings of CO2
sequestration, some specific aspects remain
unexplored and thus require further investigation
and improvement. For instance, the injection of CO2
into coal affects its structural properties and,
consequently, the efficiency of the sequestration
process, as well as the long-term integrity of the
reservoir. CO2 adsorption-induced swelling causes
permeability reductions, hindering CO2 injectivity,
and the CO2 interaction-induced strength alterations
may cause the gas to leak out into the atmosphere,
rendering the storage efforts unproductive and risky.
A deep understanding of the structural alterations in
coal hydro-mechanical properties is thus essential
to optimise the process while minimising reservoir
damage.
Figure 11. Geometries considered (a) Y intersection
3.1.1 Physical structure and gas flow in coal
(45o intersection angle), and (b) T intersection (90o
intersection angle) (Hansika 2023)
Coal has properly developed cleat systems made of
two types of cleats; Butt and face cleats orthogonal
Notably, the outcomes of the flow simulation done
to each other. Further, coal has a dual porosity
for a natural fracture network mapped in the outcrop
system (micropores <2nm and macropores >50nm).
show that the inertial effect in real fracture outcrops
becomes critical at very low flow compared to small
3.1.2 Issues
fracture networks with a single intersection, causing
flow regime transition.
Once CO2 is injected, it travels through the cleats as
an advective flux, adsorbing into the cleats' wall and
Though there is a cubic relationship between flow
diffusing into the matrix. During this adsorption, a
rate and pressure drop over the entire domain at
strain develops, causing coal matrix swelling.
low flow rates, it turns into a quadratic relationship
According to our UCS/tri-axial strength and tri-axial
with the increment of flow rate with distinct
permeability tests and modelling done on CO2-
coefficients for laminar and turbulent regimes
saturated coal under various reservoir
because of the inertial effect.
environments, this can cause a significant strength
and permeability reduction in coal, where the impact

is much greater when the injected CO2 is at its
super-critical condition due to the associated greater
coal matrix swelling caused by the greater degree of
polymerisation of organic matter and greater energy
exchange (Figure 12).
Figure 13. Effect of CO2 adsorption pressure and

phase on tri-axial strength of bituminous coal for
different biot values (Sampath 2020)
Regarding the fluid injectability or flow ability of CO2

through the seam, our findings showed a significant
permeability reduction upon CO2 saturation,
especially under super-critical condition. Figure 14
shows our observation on fracture shrinkage upon
CO2 adsorption-induced swelling.
a)
Figure 12. The physical appearance of failed, CO2

saturated, bituminous coal samples after UCS tests
(CO2 saturation under various pressures under
33 °C (Perera 2011)
CO2 converts to its super-critical state when the

temperature exceeds 31.8oC and pressure exceeds
7.38MPa, typically over 800m depths. According to b)
this finding, shallow seams seem more effective;
however, they are unsuitable for CO2 storage due to
potential back migration issues and storage
capacities. Interestingly, we found that the impact of
CO2 adsorption on coal mechanical properties
reduces after the critical depth, showing the safety
of the deeper seams for CO2 storage. On the other
hand, deeper seams have well matured high-ranked
coal with well-developed cleat systems and are thus
more vulnerable to CO2 interaction-caused
changes.
So, it is required to conduct a simulation comparing

these two controversial effects to decide about the
suitability of the seam for the storage process if the
risk of leakage through the seam's mechanical
failure is considered. Such a study has been
conducted considering modifying cohesion and
friction angle with the CO2 saturation process Figure 14. Temporal variation of fracture aperture
(Figure 13). with- and without-swelling conditions (Sampath
2020)
All these studies showed the importance of

recovering the swelling in the CO2 sequestration
process. According to our research, this can be
done by injecting a stream of inert gas as nitrogen
into the coal seam after each CO2 injection cycle.
This has been discovered by conducting a series of the CO2 storage process in deep coal seam gas
tri-axial permeability tests. After the first cycle of reservoirs causes a coal matrix swelling,
CO2 injection, N2 gas was injected into coal samples significantly reducing the seam's storage capacity
for 24h, and permeability tests for CO2 were then and mechanical stability, where the damage rate
repeated to investigate changes in CO2 flow ability depends on the coal maturity and reservoir location.
after the N2 injection. As shown in Figure 15, CO2 According to our studies, the impact is much more
permeability showed an increment in permeability. significant in the super-critical CO2., a highly
reactive physical state, than in gas CO2. Thus,
Around 20% of permeability recovery could be seen selecting a suitable type of reservoir located at
at 20MPa confining pressure. Thus, it is clear that desired depth plays an important role here. Also, the
there is a significant impact coming from N2 flooding swelling area is found to be partially recovered by
after the first cycle of CO2 storage on CO2 injectivity injecting a stream of N2 after the first injection cycle
into the seam and can be used in commercial CO2 of CO2.
storage sites.
5 ACKNOWLEDGEMENTS
This study acknowledges the financial contribution

of the Australian Research Council Discovery
Projects Scheme Grants (DP220100851,
DE130100124).
REFERENCES
Ahamed, M.A.A. (2020). " Efficient ways of proppant usage
in hydraulic fracturing of coal seam gas reservoirs."
PhD Thesis, The University of Melbourne, Australia.
Bai, T., Z. Chen, S. M. Aminossadati, Z. Pan, et., (2015).
"Characterisation of coal fines generation: A micro-
scale investigation." Journal of Natural Gas Science
and Engineering, 27, 862-875.
IEA, Gas-Fuels and Technologies,
http://www.iea.org/fuels-and-technologies/gas (Jul.
10, 2023)
Figure 15. Effect of N2 flood on CO2 permeability at Day, S., L. Connell, D. Etheridge, et al. (2000). Fugitive
20 MPa confining pressure (Perera et al. 2011) greenhouse gas emissions from coal seam gas
production in Australia, Energy Technology, CSIRO,
4 CONCLUSION Australia.
Earthworks. "Hydraulic fracturing 101."
https://earthworks.org/issues/hydraulic-fracturing-
The study focuses on novel solutions discovered to
101/ (July. 10, 2023)
optimise the usage of underground geological Hansika H. (2023). Influence of inertial and centrifugal
reservoirs for gas storage and recovery. forces on the flow rate and patterns of flow in
fracture networks, MPhil Thesis, The University of
In the gas storage and recovery processes, Melbourne, Australia.
reservoir stimulation techniques, such as hydraulic Ltd, U.A.P. (2014). Background review: Hydraulic
fracturing, play a significant role. The effectiveness fracturing ('fraccing') techniques, including reporting
of proppants used in the fracking process depends requirements and governance arrangements,
Department of the Environment of Australian
on the type and location of the reservoir rock and Government.
the fluid dynamic there. Thus, the optimum usage of Miri, R., M. Haftani and A. Nouri (2021). "A review of fines
proppants in underground should clearly understand migration around Steam Assisted Gravity Drainage
these aspects for maximum storage capacity and wellbores." Journal of Petroleum Science and
gas recovery. In particular, low-stiffness rocks like Engineering, 205,108868.
coal causes proppants to embed while shrinking the Perera, M. S. A. (2011). "Investigation of the effect of CO2
flow paths and tend to settle at junctions and avoid sequestration on hydro-mechanical properties of
branching fractures due to the inertia causing flow coal." PhD Thesis, Monash University, Australia.
Sampath K. H. S. M. (2020). " A comprehensive analysis
complexities.
of the structural alterations occur in CO₂- interacted
coal: An insight into CO₂ sequestration in coal." PhD
According to our studies, identifying the desired Thesis, The University of Melbourne, Australia.
depths and the type of proppants through proper Zeinijahromi, A., A. Vaz and P. Bedrikovetsky (2012).
preliminary investigation can significantly enhance "Well impairment by fines migration in gas fields".
the effectiveness. On the other hand, identifying Journal of Petroleum Science and Engineering, 88–
potential inertia causing flow instability and 89, 125–135.
controlling the gas injection rates accordingly helps
to improve the proppants distribution while avoiding
unnecessary proppants clogging at fracture
intersections.
The gas storage process in deep underground

reservoirs faces a lot of challenges. For example,
Vein Effect on Mechanical Properties and Fracture Behaviour of Porphyry Rock
H. J. Wang1 and Q. B. Zhang2
1
Department of Civil Engineering, Monash University, VIC 3800, Australia; email: haojun.wang@monash.edu
Department of Civil Engineering, Monash University, VIC 3800, Australia; email: qianbing.zhang@monash.edu
2
ABSTRACT
Veins are commonly found in porphyry copper deposits and can greatly impact rock mass stability during block caving
processes. Understanding the mechanical properties and fracture behaviours of veined rocks under various
compression tests can help reduce potential hazards in rock engineering. In this study, laboratory tests were conducted
on veined porphyry from mining operations in Australia under uniaxial compression and triaxial compression with
different confinements. The results show that the orientation of calcite veins is the main inherent factor in reducing rock
strength, with the least favourable angle being between 15° and 45° relative to the normal loading direction. Among
these values, the weakest specimen was found at 30° in uniaxial compression tests. The real-time cracking process
of compression tests was monitored by an acoustic emission (AE) system, and surface cracking was captured by high-
speed cameras for uniaxial tests. AE events were concentrated at the final failure stage for veined samples, while
noticeable signals were detected during the early stages for intact samples. Furthermore, early-stage event energy
increased as confining pressure increased. X-ray computed microtomography (CT) was also used on tested samples
to obtain and reconstruct fracture patterns after peak load. The results show that veins dominate failure planes at low
confinement, while brittle fractures in the rock matrix increase at high confinement.
Keywords: veined porphyry, fracture pattern, high-speed camera, computed microtomography
1 INTRODUCTION falling within the range of 2.0 to 2.5. Prior to testing, the
cylindrical surfaces of the samples were meticulously
Veins form through the deposition of minerals carried prepared to ensure they were flat and smooth. In this
by hydrothermal fluids circulating through paper, we focused on investigating the influence of two
mechanical discontinuities within rock masses. This key geometric factors: vein number and vein
natural formation process underscores the orientation shown in Figure 1. Vein number was
complexity of vein geometry, mineralogy, mechanical defined as the presence of continuous veins that were
properties (Day et al. 2019, Shang 2022). Veins play visible to the naked eye. Vein orientation was
a fundamental role in modulating the mechanical, measured as the angle between the veins and the
geophysical, and geochemical behaviour of rocks. loading direction (Turichshev and Hadjigeorgiou 2015).
Understanding the influence of veins on the
deformation and failure of veined rock is of
paramount importance for various applications in the
field of underground engineering. One such industry ①
is copper mining, where veined porphyry has proven
θ
to be a common and valuable deposit (Robbiano et
al. 2022, 2023). ②
Despite the natural distribution of veins being complex,

this study finds some specimens with simple patterns ③
characterising geometric distribution by both their 60mm
number and orientation within the host rock (Clark and
Day 2021, Turichshev and Hadjigeorgiou 2017).
Investigating veined rock mechanical behaviour, (a) Prepared (b) Sample (c) Orientation
particularly in the context of uniaxial compressive porphyry with three respect to
strength and triaxial compressive strength tests, would specimen veins loading
be helpful for both open-pit and block-caving operations. Figure 1. Veined porphyry with geometric factors
2 EXPERIMENTS PROCEDURES 2.2 Test setup and procedure

2.1 Specimen preparation Uniaxial compressive strength (UCS) tests were
conducted using the Instron 600 DX apparatus that is
Porphyry specimens were selected for this study due specifically designed for high-capacity compression,
to their distinctive vein appearance. Core samples from testing. It boasts an impressive ±600kN loading capacity,
Australia had a diameter of 60mm. These samples features dual test spaces within a single footprint, and is
adhered to the ASTM (American Society for Testing constructed using high-quality materials. To ensure
and Materials) Standards (ASTM 2010) and ISRM quasi-static strain rates, a loading rate of 0.1 mm/min
(International Society for Rock Mechanics and Rock was employed.
Engineering) Suggested Methods (Bieniawski and
Bernede 1979), with a length-to-diameter ratio (L/D)
Vein Effect on Mechanical Properties and Fracture Behaviour of Porphyry Rock H. J. Wang and Q. B. Zhang
For capturing the fracture process during the uniaxial

compression tests, high-speed CMOS cameras
(Phantom V2511) were utilised. These cameras were
mounted on tripods and captured grey-scale images
of the specimen surface from various angles. Image
acquisition was carried out at an impressive rate of
100,000 frames per second (fps) (Liu et al. 2019),
resulting in a series of deformed images (384 × 288
pixels) of the specimen.
In addition to surface visual data, acoustic emission

(AE) signals were collected using Elsys Data
Acquisition Instrument TraNET EPC in conjunction
with NANO 30 sensors. These sensors were chosen
for their excellent frequency response within the (a) UCS loading apparatus + high-speed cameras + AE
range of 125 - 750kHz. Each rock sample was data acquisition system
equipped with six AE sensors, all without
preamplifiers. These sensors were affixed using
Olympus couplant, and the sampling rate for data
collection was set at 5MHz. The six sensors were
divided into three sets, with two sensors in each set.
Two sets of sensors were symmetrically positioned
on both sides (left and right) of the specimen, with a
distance of 30mm from the top and bottom ends. The
third set was attached 40mm away from the ends on
the back of the specimen, allowing the front face to
remain available for high-speed camera recording. (b) Nano 30 (c) Nano 30 (d) X-ray CT
Furthermore, two pairs of strain gauges were sensors on sensors on facility at
adhered to two sides of the specimen, precisely in the UCS samples Hoek cell Australian
middle of each set of AE sensors. These strain synchrotron
gauges were used to measure axial and lateral strain
during the tests. Before initiating the compression Figure 2. Experimental facilities utilised in this study
tests, P-wave velocities of each specimen were
recorded. These velocity measurements were crucial During the triaxial tests, the photographs were taken
for calculating AE event locations, a process that will to document the failure mode of each sample.
be carried out using InSite software. Additionally, the failed samples had the opportunity
to be scanned using X-ray Computed Tomography
Triaxial compressive strength (TCS) tests were (CT) to extract 3D visualisations of the final fracture
conducted by using an Avery-Denison servo-hydraulic patterns (Liu et al. 2019, Li et al. 2021, Wang et al.
universal testing machine with a 2000kN capacity. In 2022). Such scanning was employed at Australian
this setup, a Hoek cell was used to apply lateral Synchrotron - Imaging and Medical Beamline (IMBL).
confining pressure to the samples. However, direct
observation of the fracture process was not possible 3 RESULTS AND ANALYSIS
due to the design of the Hoek cell, which prevented
high-speed capture of the fracture process. To 3.1 Veins effect on uniaxial test
address this limitation, eight sensors were attached
outside the pressure chamber using couplant. The axial compressive strength of porphyry exhibits
a wide range, spanning from 57 to 145 MPa. These
Since AE sensors were not directly attached to the UCS results underscore the significant influence of
porphyry samples, it was anticipated that many AE failure mode on the observed failure stress levels,
signals would be attenuated, potentially introducing with values displaying considerable variability due to
errors in location tracking. Therefore, AE waveforms the differing textures resulting from sample
were primarily used to capture the history of crack discontinuities. Notably, the rock's behaviour is
propagation in the time domain. Different strongly correlated with its failure mode. For UCS
preamplifiers were employed for this purpose, with values exceeding 120 MPa, the rock tends to
low or no amplification used for capturing global manifest longitudinal tensile cracks approximately
signals and high amplification used for detecting parallel to the loading direction. Conversely, shear
local, small signals that might exceed the measuring failure is typically observed in veined samples.
range of the sensors. The sampling rate for AE data
collection was set at 2MHz, and preamplifiers with Categorizing the test samples based on material and
10dB and 40dB amplification were used. This failure mode, they are divided into two distinct groups:
allowed AE signals to be amplified by either 10 times intact porphyry and veined porphyry. When
or 100 times, respectively, facilitating the detection of considering failure mode, specimens are further
micro-crack propagation. The monitoring plan categorized as either tensile-governed or shear-
involved two channels at ×1 amplification, two governed. It is noteworthy that shear failure tends to
channels at ×10 amplification, and four channels at reduce the strength by half compared to tensile failure.
×100 amplification. This distinction is evident in the failure processes
captured by high-speed cameras, with typical cases spheres represent AE event locations. Larger sphere
shown in Figures 3 and 4. For intact case, surface radii correspond to higher energy magnitudes.
spalling and rock pillar splitting up can be observed
around peak load. For veined case, failure mode is In alignment with high-speed images, the behaviour
sliding shear along veins then splitting follows. of the intact sample unfolds with the initial
observation of spalling near the top, followed by
160 2500 crack propagation towards the bottom. Internal
(ii)(iii) cracks are also detected, contributing to the eventual
(i) (iv) splitting of the sample. On the other hand, in the
2000
veined sample, initial observations reveal bottom
120 ③
cracks, succeeded by shear failure on the right side,
along with subsequent cracking along the plane of
1500
the vein.
σ1 (MPa)
AE hit
80
160 2500
1000
① ④ ④
40 2000
500 120
② (ii)
(iii) 1500
(iv)
σ1 (MPa)
0 0
0.00 0.05 0.10 0.15 0.20 0.25
AE hit
80
ε (%)
1000
40
①② 500
③
(i) (ii) (iii) (iv)
0 0
0.00 0.05 0.10 0.15 0.20 0.25
ε (%)
(i) (ii) (iii) (iv)

① ② ③ ④
Figure 3. Intact porphyry – stress strain curve, AE

hits, high-speed camera images & AE locations
In both intact and veined samples, AE hits are

consistently focused at the peak load, leaving no
room for doubt regarding their correlation with the
mechanical behaviour. However, the veined sample ① ② ③ ④
exhibits a cleaner signal during the breakage
process, offering clearer and more distinct Figure 4. Veined porphyry – stress strain curve, AE
signatures. Notably, strong pre-signals are detected hits, high-speed camera images & AE locations
before the stress drop and impending failure,
providing valuable insight into the failure It is crucial to acknowledge that the presence of veins
mechanisms. plays a pivotal role in shaping the overall properties
and behaviour of the rock. Moreover, our statistical
Traditionally, AE location calculations assume analysis highlights the significance of two key
homogeneity in material properties, particularly in geometric factors: vein number and vein orientation,
terms of P-wave velocity across all directions. The as shown in Figure 5. Vein number exhibits a positive
intact sample serves as a validating reference for this correlation with the veined volume ratio. This
algorithm, as it confirms a close match between the relationship is important, as cemented veins typically
calculated AE locations and the observed crack possess lower strength than the surrounding rock
patterns. However, in the case of the veined sample, matrix, in this case, the porphyry. Consequently, as
the presence of veins introduces increased material the vein number increases, the overall strength of the
heterogeneity. This deviation from homogeneity sample decreases. Regarding vein orientation, we
affects the accuracy of the location calculation have chosen to designate the thickest vein as the
method, albeit remaining within a reasonable range. dominant vein. When we categorize the samples into
This still allows for the demonstration of crack two groups based on vein thickness, distinguishing
positions. In the provided visual representation, black between those greater or less than 1mm, a distinct
triangles signify the sensor locations, while red trend emerges. The relationship between vein
orientation and strength displays a U-shaped predominantly concentrate around the point of final
distribution. Strength tends to decrease within the failure, other information is hard to collect. However,
0-30° range, reach its minimum value at around 30°, in the veined sample, the presence of veins
and then increase from 30 to 60°. Therefore, the substantially enhances AE response during the early
orientation of veins is a critical factor with significant stages of testing, suggesting a progressive fracturing
effects on the mechanical behaviour of the rock. process taking place.
160 350
Intact
240
210
300
120 250
180
Before After CT
150
σ1 (MPa)
σ3-σ1(MPa)
200
AE hits
80 120
150
90
100
40 60
50
30
0 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0
0 1 2 3 4 ε(%)
Veins number
350 240
Veined
160
210
300
180
250
120
150
σ3-σ1(MPa)
200
AE hits
σ1 (MPa)
120
80 150
90
100 Before After CT

60
40
50
30
Thickness > 1mm
Thickness < 1mm 0 0
0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
0 30 60 ε(%)
Thickest vein orientation (°)
Figure 6. Typical TCS results - stress strain curves,

Figure 5. UCS test summary for veined porphyry AE hits, fracture patterns, CT reconstruction images
3.2 Veins effect on triaxial test Triaxial strengths are graphically represented in
Figure 7. An intriguing observation is that intact
Figure 6 presents the typical triaxial test results samples normally follow Mohr-Coulomb failure
conducted under a confinement of 10MPa, and it criterion, giving a linear positive correlation between
reveals a significant observation: the intact sample σ3 and σ1, whereas the veined samples present
demonstrates a strength and peak strain approximately discrete strength distribution. The difference more
double that of the veined sample, yet the elastic evident is under low confinement range (<10MPa)
modulus is only slightly weaken. This performance indicating the importance of vein study during
mirrors the trends observed in the UCS tests. In the excavation activities.
case of the intact sample, typical brittle failure manifests
along an oblique direction. The computed tomography
(CT) extraction reveals a fracture surface that is notably
thin, indicating a relatively smooth failure surface.
Conversely, in the veined sample, the presence of
weaker surfaces influences the direction of shear
damage, resulting in a CT-reconstructed fracture that
appears wide and rough.
Concerning AE data, it is evident that the Hoek cell

setup diminishes the sensitivity of AE detection, with
the hit count in UCS and TCS differing by an order of
magnitude. For the intact sample, AE signals
While their presence can complicate analysis, it also

offers valuable insights into the behaviour of rocks
under different loading conditions. Future research
efforts should continue to explore the effects of veins
and pre-existing faults to enhance our understanding
of rock properties in practical applications like
underground infrastructure, block caving.
5 ACKNOWLEDGEMENTS
The authors would also like to acknowledge the

support from Newcrest Mining Limited and Australian
Synchrotron (Project No.: 18984). A special thanks
goes to Dr. Gisela Viegas, Dr. Anto Maksimenko and
Dr. Matthew Cameron for their kind support.
REFERENCES
Figure 7. TCS results of intact and veined porphyry
ASTM, A. (2010). Standard test method for compressive
Upon comparing the stress results derived from the strength and elastic moduli of intact rock core
loading test with the observed failure patterns, it specimens under varying states of stress and
becomes evident that vein distribution wields temperatures. In.
significant influence over triaxial strength. This Bieniawski, Z. T., & Bernede, M. J. (1979). Suggested
methods for determining the uniaxial compressive
observation aligns harmoniously with the findings
strength and deformability of rock materials: Part 1.
from UCS tests, where it was consistently observed Suggested method for determining deformability of
that the majority of shear failures occurred along the rock materials in uniaxial compression. International
veins. These observations underscore the dominant Journal of Rock Mechanics and Mining Sciences &
role of structural features, such as veins, in dictating Geomechanics Abstracts, 16(2), 138-140.
the mechanical behaviour of the rock in triaxial Clark, M. D., & Day, J. J. (2021). Mineralogical and sample
testing conditions. selection implications for geomechanical properties of
intact heterogeneous and veined rocks from the
Legacy skarn deposit. Engineering Geology, 285,
4 CONCLUSION 106067.
Day, J. J., Diederichs, M. S., & Hutchinson, D. J. (2019).
In conclusion, our research has highlighted the critical Composite Geological Strength Index Approach with
role of veins in the mechanical behaviour of rocks, Application to Hydrothermal Vein Networks and Other
particularly in the context of uniaxial and triaxial tests. Intrablock Structures in Complex Rockmasses.
We observed that veins contribute to a wide range of Geotechnical and Geological Engineering, 37(6),
results in terms of compressive strength, and their 5285-5314.
Li, J., Zhao, J., Gong, S. Y., Wang, H. C., Ju, M. H., Du, K.,
presence significantly influences failure modes. In
& Zhang, Q. B. (2021). Mechanical anisotropy of coal
uniaxial compression tests, we noted distinct patterns under coupled biaxial static and dynamic loads.
of failure, with shear failures occurring primarily in International Journal of Rock Mechanics and Mining
veined samples and tensile cracks developing in intact Sciences, 143, 104807.
samples with higher UCS values. Liu, K., Zhang, Q., Wu, G., Li, J., & Zhao, J. (2019). Dynamic
mechanical and fracture behaviour of sandstone
Acoustic Emission monitoring revealed that veined under multiaxial loads using a triaxial Hopkinson bar.
samples exhibited cleaner AE signals during Rock Mechanics and Rock Engineering, 52, 2175-
2195.
breakage, and strong pre-signals provided insight into
Robbiano, F., Liu, K., Zhang, Q.-B., & Orellana, L. F. (2022).
impending failure. While traditional AE location Dynamic uniaxial compression testing of veined rocks
calculations assume material homogeneity, our study under high strain rates. International Journal of Rock
demonstrated that deviations caused by veins still Mechanics and Mining Sciences, 153, 105085.
allow for reasonably accurate crack position Robbiano, F., Liu, K., Zhang, Q.-B., & Orellana, L. F. (2023).
determination, although with slightly reduced Dynamic mechanical properties of veined rocks under
precision. biaxial confinement. International Journal of Rock
Mechanics and Mining Sciences, 170, 105538.
Shang, J. (2022). Stress path constraints on veined rock
Triaxial tests further underscored the impact of veins,
deformation. Rock Mechanics Bulletin, 1(1), 100001.
with the Hoek cell setup reducing AE sensitivity but Turichshev, A., & Hadjigeorgiou, J. (2015). Experimental
veins increasing AE response, signifying gradual and numerical investigations into the strength of intact
fracturing. The relationship between triaxial strength veined rock. Rock Mechanics and Rock Engineering,
and confining pressure was not always 48, 1897-1912.
straightforward, emphasizing the dominance of Turichshev, A., & Hadjigeorgiou, J. (2017). Quantifying the
structural factors. This aligns with UCS test results, effects of vein mineralogy, thickness, and orientation
where most shear failures were observed along veins. on the strength of intact veined rock. Engineering
Geology, 226, 199-207.
Wang, H. C., Zhao, J., Li, J., Braithwaite, C. H., & Zhang,
In summary, veins are a crucial factor in understanding Q. B. (2022). Progressive fracturing of concrete under
and predicting the mechanical behaviour of rock biaxial confinement and repetitive dynamic loadings:
materials. They introduce variability in strength, From damage to catastrophic failure. International
influence failure modes, and affect AE monitoring. Journal of Impact Engineering, 165, 104232.
NOTES

2023 AUSTRALIAN GEOMECHANICS SOCIETY
VICTORIAN SYMPOSIUM
Novel Solutions in
Geotechnical Engineering
VICTORIA CHAPTER

AGS VIC 2023 Symposium Proceedings WEB

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AGS VIC 2023 Symposium Proceedings WEB

Uploaded by

Copyright:

Available Formats

PROCEEDINGS

2023 AUSTRALIAN GEOMECHANICS SOCIETY

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING i

ii 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

Back Analysis, Monitoring and Remediation 15

Ground Improvement Works for Cutter Head 23

Earthquake Site Subsoil Classification in 28

Gas Permeability of Cement-Bentonite 35

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING iii

iv 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

9:20 am Presentation 2:15 pm Presentation

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING v

A Speargun Penetrometer Modelling rainfall-induced Forecasting Offshore

Microbial Stabilization Preliminary design guidelines for Internal Erosion:

Ground improvement using Biochar as a sustainable Geopolymer Stabilized

vi 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

GOLD SPONSORS TECH SERVI

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING vii

Proud sponsors of the

Chadwick Geotechnics operates NATA accredited

Contact our team to learn how we can help you

Call: 03 8796 7900

Global Synthetics is a 100% Australian-owned company,

Stay on top of the latest Global Synthetics information

IN SITU TESTING, SAMPLING

AND GEOTECHNICAL SERVICES ge nt

The seminar will teach participants the

To learn more, visit:

Keywords: anisotropic, numerical analysis, calibration, tunnelling

1 INTRODUCTION 2 GEOTECHNICAL CONDITIONS

Figure 1. Arrangement of State Library Station

To simulate the anisotropic rock mass response to

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING 3

4 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

Figure 4. 3DEC model of simulated Melbourne Formation, β = 0, σ3 = 0 MPa

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING 5

4 MODEL GEOMETRY interface. The primary steel columns within the

An example of the disturbed nature of the rock mass

Figure 6. FLAC3D Model Geometry

Figure 7. FLAC3D Model Structural Elements

Within the global models, the fibre reinforced

Table 1: Melbourne Formation modelling parameters (calibrated UJRM input properties)

6 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

6 CALIBRATION ASSESSMENT (Figure 9(b)) above Franklin Adit was predicted to be

The State Library Station surface was monitored with

6.2 Surface Settlement

Figure 10. Calibration of predicted versus measured

settlement for trinocular cavern (ground

-30.0 EASTBOUND PLATFORM CL

tunnel and (b) trinocular cavern CHAINAGE

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING 7

The additional primary steel column settlement REFERENCES

The authors wish to acknowledge Rail Projects

8 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

1 INTRODUCTION performance of structure during various stages of

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING 9

Figure 1. Schematic geometry of embankment and foundation

10 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

Table 2: Material parameters for MC and elastic Constitutive models

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING 11

12 2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING

The possibility of a static liquefaction within the

As shown in Figure 6, while the ψ has increased with

2023 AGS VICTORIAN SYMPOSIUM NOVEL SOLUTIONS IN GEOTECHNICAL ENGINEERING 13