Professional Documents
Culture Documents
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data
and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The
authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright hold-
ers if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know
so we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any elec-
tronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information
storage or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or con-
tact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that
provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system
of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation
without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
Printed and bound in Great Britain by Antony Rowe (A CPI Group Company),
Chippenham, Wiltshire
All rights reserved. No part of this publication or the information contained herein may be
reproduced, stored in a retrieval system, or transmitted in any form or by any means,
electronic, mechanical, by photocopying, recording or otherwise, without written prior
permission from the publisher.
Although all care is taken to ensure integrity and the quality of this publication and the
information herein, no responsibility is assumed by the publishers nor the author for any
damage to the property or persons as a result of operation or use of this publication
and/or the information contained herein.
Table of Contents
Preface XXIII
Organisation XXV
Müller lecture
Keynote lectures
Mine-by experiment performed in the Callovo-Oxfordian claystone at the Meuse Haute Marne underground
research laboratory (France) 157
G. Armand, A. Noiret, M. Cruchaudet & N. Conil
Burgers rock creep around axisymmetric tunnels 159
S.D. Athanasopoulos, P.P. Nomikos & A.I. Sofianos
Multi-parameter responses of soft rocks during deformation and fracturing and their implications in
geomechanics and geoengineering 161
Ö. Aydan, Y. Ohta & H. Tano
Fractal approach to determine rock mass strength and deformation 162
M.N. Bagde, A.K. Raina, P.B. Choudhury & A. Sinha
A theoretical analysis of core extrusion and squeezing behaviour in tunnelling 163
L. Cantieni, G. Anagnostou & R. Hug
Applications of strain based damage criterion in geotechnical engineering 165
Y. Chang
Solving some problems of geomechanics on the base of defining relations of post-limit deformation of rocks 167
A.I. Chanyshev, I.M. Abdulin & O.E. Belousova
The structural geology contribution to rock mechanics modelling and rock engineering design 169
J.W. Cosgrove & J.A. Hudson
Rock mass instability caused by incipient block rotation 171
A.V. Dyskin & E. Pasternak
Attempt to advanced observational construction considering predictive analysis of long-term deformation 173
T. Fukuda, T. Takahashi, H. Yamada & Y. Jiang
Vector sum method: A new method for anti-sliding stability analysis 174
X.R. Ge
Design of a railway tunnel parallel to an existing tunnel 176
R.K. Goel, A. Swarup, R.K. Dayal & V. Bhardwaj
Shear failure mechanism in oil wells due to reservoir compaction “A case study in Gachsaran formation, Iran” 177
M. Gorjian, M. Moosavi, H. Memarian & H. Ajamzadeh
Fractal analysis on the fracture development of sandstone using AE measurement 179
Z.Y. Yang, T.H. Huang, Y.X. Zhao & M.C. Tsai
Static and dynamic response analyses of rock mass considering joint distribution and its applicability 181
N. Iwata, T. Sasaki, & R. Yoshinaka
Numerical study on zonal disintegration of rock mass around deep underground openings 182
P. Jia, C.A. Tang & Y.B. Zhang
The secondary lining design of subsea tunnel due to water inflow 184
K.-S. Joo, Y.-H. Suh, K.-H. Park, D.-H. Huh, E.-J. Hong, S.-J. Moon & S.-C. Lee
An assessment of rock pillar considering excavation damaged rock zone 186
W.-B. Kim, T. Ha & H. Yang
Response of granite under strain controlled loading and effect of support system on behaviour of
large underground cavern 187
Rakesh Kumar, K.G. Sharma & A. Varadarajan
Determination of backfill requirements in multi-reef mining using numerical modelling 188
K.B. Le Bron & R.A. Johnson
Comparisons between two empirical yield criteria for rock masses 190
A.J. Li & R.S. Merifield
Buckling failure of highly-stressed hard rocks surrounding deep underground openings and confining effects of
backfilling material 191
D.Y. Li, X.B. Li, Z.X. Liu, C.D. Ma & L.N.Y. Wong
Bemposta II powerhouse shaft: Geotechnical characterization, design and construction 193
C. Lima, E. Resende, C. Esteves & J. Neves
VI
Numerical modelling
Coupled flow-deformation simulation for mine scale analysis of cave initiation and propagation 223
D.A. Beck & G. Putzar
A numerical analysis of a four-way coal mine intersection with primary and secondary supports 224
Y.P. Chugh & B. Abbasi
3-D modeling and parameters research of irregular shape particles for DEM based on computerized tomography 225
X. Du, Y.W. Zeng, R. Gao & J. Yan
Three-dimensional thermomechanical modelling of backfilling of old mining rooms in the Morsleben repository 227
S. Fahland, R. Eickemeier & St. Heusermann
Development of a non-contact survey method of tunnel excavation face for DEM modelling 229
A.M. Ferrero, G.F. Forlani, R. Migliazza & R. Roncella
Numerical simulation of crack propagation in rock by clumped particle model 231
T. Funatsu & N. Shimizu
Stability modelling of isolated rock blocks at the surface of underground excavations taking into
account in-situ stresses 232
R. Ghazal, F. Hadj-Hassen & M. Tijani
VII
VIII
IX
XI
A new method for visualizing stresses in rock support measures by using a light emitting sensor 403
S. Akutagawa, Y. Minami, H. Yamachi, T. Haba, T. Hiroshima & T. Nozawa
Microseismic monitoring strategy as a key component of post-mining risk management:
Review and feedback experience over the past decade 404
P. Bigarré, M. Bennani, I. Contrucci, E. Klein, H. Baroudi, R. Hadadou, I. Vuidart & J.P. Josien
Factors dominating deterioration of clastic sedimentary rocks 406
P.-C. Chan, S.-J. Lyu, T.-T. Wang, F.-S. Jeng & T.-H. Ueng
Image analysis system for areal measurement to preserve historic masonry wall, Japan 408
H. Chikahisa, R. Ohta, T. Kawahara, N. Kitayama & Y. Okamoto
Property evaluation of beachrock based on field investigation in Okinawa Island, Japan 409
T. Danjo & S. Kawasaki
An engineering view of the seismic prediction during tunneling 410
T. Dickmann & F. Pittard
Application of acoustic emission measurement on assessment of loosened zone around underground cavern 411
L. Fan, H.M. Zhou, Y.H. Zhang & Z.W. Zhong
Investigation of 3D terrestrial laser scanning techniques for potential application to rock mechanics 412
Q. Feng, G. Wang & K. Röshoff
Interpretation of rock mass deformability measurements using large flat jack tests 413
B. Figueiredo, F. Bernardo, L. Lamas & J. Muralha
Abandoned roadways aged up to 50 years observed in Kushiro Coal Mine, Japan 415
Y. Fujii, T. Sugawara, J. Kodama, Y. Ishijima, T. Kiyama, M. Takada, Y. Ichihara, S. Kumakura,
T. Narita, M. Sawada & E. Nakata
The role of instrumentation and monitoring system on the stability of a slope 416
M. Gharouni Nik
Determination of mechanical properties of a large deformation zone 417
R. Glamheden, I. Olofsson & A. Fredriksson
Spalling in a field experiment of heated boreholes in hard rock 419
R. Glamheden, B. Fälth & R. Christiansson
About the determination of the joint set intensity 421
N.F. Grossmann
Determination of in-situ stress orientation by thermally induced spalling 423
E. Hakami & R. Christiansson
Site investigation in volcanic rock mass for building projects 425
L.E. Hernández, M.J. Pomares, J.A. Rodríguez-Losada, A. Eff-Darwich & C. Olalla
XII
XIII
Damping analysis with depth and influence of geology on the transmission of vibrations generated by blasts 471
M.I. Alvarez-Fernandez, C. Gonzalez-Nicieza & A.E. Alvarez-Vigil
Comparison of blast-induced ground vibration predictors in Seyitomer coal mine 472
E. Arpaz, Ö. Uysal, Y. Tola, K. Görgülü & M. Çavuş
A new testing station for rock impact on full-scale structures 473
M. Bost, S. Loock, L. Lenti, F. Rocher-Lacoste, J. Gineys & L. Dubois
Mechanical behaviour of rock discontinuities under seismic conditions 474
L. Canelli, A.M. Ferrero, M. Migliazza & A. Segalini
Dynamic responses of tunnels subjected to Rayleigh wave 476
C.-H. Chen, T.-T. Wang & T.-H. Huang
Dynamic tensile-splitting tests of rocks 478
S.H. Cho, J.R. Ahn, M.S. Kang, Y. Nakamura, Y. Ogata & K. Kaneko
Characterization of dynamic rock fracture parameters using Notched Semi-circular Bend (NSCB)
method and Cracked Chevron Notched Brazilian Disc (CCNBD) method 480
F. Dai & K. Xia
Numerical simulation of train-induced vibrations in rock masses 482
A. Eitzenberger, P. Zhang & E. Nordlund
Effect of micro-defects and macro-joints on stress wave propagation in rock mass 484
L.F. Fan, F. Ren & G.W. Ma
Progress in study of rockburst experiments and control countermeasures 486
M.C. He, X.N. Jia, W.L. Gong, G.J. Liu & J.T. Zhang
Evaluation of dynamic behaviors of bedrock foundation based on FEM and DEM simulations 488
Y. Jiang, L. Yang, Y. Tanabashi, B. Li & S. Li
Application of discrete element method for validity evaluation of rock dynamic fracture
toughness measured by semi-circular bend technique 489
T. Kazerani & J. Zhao
The mechanism behind seismic damage to railway mountain tunnels and assessment of their aseismic performance 490
Y. Kojima, K. Yashiro, T. Funahashi, H. Miyabayashi, T. Iura & T. Asakura
Destress rock blasting as a rockburst control technique 491
P. Konicek, P. Konecny & J. Ptacek
Dynamic test of a high energy-absorbing rock bolt 492
C.C. Li
A 1D equivalent medium method for wave propagation across parallel jointed rock mass 494
J.C. Li, H.B. Li, G.W. Ma & J. Zhao
Effects of decoupled charge blasting on rock fragmentation efficiency 496
W. Liang, H.Y. Liu, X. Yang & D.J. Williams
Modelling earthquake-induced rock failures using NMM and DDA 497
Y.J. Ning, X.M. An, Z.Y. Zhao & G.W. Ma
XIV
Experimental and numerical study on the behavior of supercritical CO2 injected to low permeable rocks 521
A. Arsyad, Y. Mitani, H. Ikemi, K. Kuze & S. Oura
A fully coupled thermoporoelastic model for drilling in HPHT formations 522
A. Diek, L. White, J.-C. Roegiers & D. Blankenship
Hydraulic stimulation of geothermal wells: Modeling of the hydro-mechanical behavior of a stimulated
fractured rock mass 524
S. Gentier, X. Rachez, M. Peter-Borie & A. Blaisonneau
Investigation of thermal effect on stability of deep wellbores in low and high permeable formations 526
Z. Hou & X. Luo
Modelling and optimization of multiple fracturing along horizontal wellbores in tight gas reservoirs 528
Z. Hou & L. Zhou
Dependence of fracture geometry and spatial variation in pressure on hydraulic conductivity in rock fractures 530
K. Kishida, T. Hosoda, A. Sawada, H. Sato, S. Nakashima & H. Yasuhara
Analytical solutions for poroelasticity within a finite two-dimensional domain 531
P.C. Li & D.T. Lu
Mechanical effect of the presence of gas on faults modeled as a sandwiched Cam-Clay material 532
V. Maury, J.-M. Piau & D. Fitzenz
Impact of fluids and formation anisotropy on acoustic, deformation and failure characteristics of
reservoir shales and pure clay minerals 534
A.N. Tutuncu & A.I. Mese
Change of joint aperture during formation of ice-ring around underground LNG storage cavern 535
E.-S. Park, S.-K. Chung, J.-H. Synn, W.-C. Jeong, T.-K. Kim & S.-H. Bae
A parametric study to investigate the effect of injection and production well operation on the CO2
sequestration process in deep unmineable coal seams 537
M.S.A. Perera, P.G. Ranjith & S.K. Choi
Development of the confined pressure type of multi-purpose rock flow system and its application
to the evaluation of permeability of various geomaterials 538
A. Sato, H. Yoshitake, T. Shiote & K.Sasa
Near-field application of aperture back calibrated distinct fracture network 539
M. Sharifzadeh & M. Javadi
XV
XVI
XVII
XVIII
XIX
XX
Interdisciplinary
XXI
XXII
Preface
This proceedings volume contains papers of ISRM 2011 Beijing – the 12th International Congress on Rock Mechanics of the
International Society for Rock Mechanics (ISRM), held at the China National Convention Centre in Beijing, China, from 17–21
October 2011. It contains the full papers of the Müller lecture by Dr Nick Barton, the Rocha Medal lecture by Dr Dohyun Park,
13 keynote papers and the extended abstracts of 384 technical papers. A CD-ROM contains all the full papers.
The ISRM will be 50 years old in 2012. This congress had an added significance because the ISRM Board decided to kick
off the 50th Anniversary celebrations at this congress. For the congress proceedings, we have three special keynote lectures
marking this major milestone in the ISRM. Two keynote lectures by Prof Ted Brown and Prof John Hudson deliberated on the
progress in rock mechanics of the past 50 years, and how rock mechanics will develop in the next 50 years. The third keynote
lecture by Prof. Qihu Qian provided an overview on developments of rock engineering technology in China.
The theme of the ISRM 2011 Congress is Harmonising Rock Engineering and the Environment, emphasing the critical
role of rock mechanics in meeting rock engineering requirements for development and the need to protect and preserve our
environment. The Congress received a total of 384 contributed papers from National Groups, other regions, and individual
members of the ISRM. Abstracts and full papers were reviewed by the National Groups. Papers submitted by individual members
were received and reviewed by the Scientific and Publications Committee.
The ISRM 2011 Congress was jointly organised by the Chinese Society for Rock Mechanics and Engineering and the Society
for Rock Mechanics & Engineering Geology Singapore, with support from the Korean Society for Rock Mechanics.
On behalf of Organising Committee, the editors express their appreciation for the contribution of all supporters, sponsors,
ISRM National Groups, authors, and staff of secretariat from both China and Singapore, which has made this congress a success.
Finally, we would like to thank Cheng Hoon Khoo and Zaiyi Guo for their excellent work in managing the paper submission.
The Editors
XXIII
Organisation
Congress Chairs:
Qihu Qian, Jian Zhao
Congress Co-Chairs:
Xia-ting Feng, Chung-In Lee, Yingxin Zhou
Organising Committee
Chairman: Qihu Qian
Co-chairman: Xia-ting Feng, Yingxin Zhou
Secretariat Members
Ting Feng Wei Hu Chun-Guang Li Wen-Lin Lu
Zhong Ren Yan-Jun Shang Xiao-Jie Yang Shuang-Liang Zheng
Yan Zhou
XXV
XXVI
N.R.Barton
Nick Barton and Associates, Oslo, Norway
ABSTRACT: The behaviour of the jointed-and-faulted-anisotropic-water-bearing media that we call rock masses, was an
abiding pre-occupation of Leopold Müller. The author has been similarly pre-occupied. So starting with modest developments
from tension-fractured physical models, and progressing via simple UDEC models to the real jointed and three-dimensional
world in due course, a few of the lessons learned and subsequently applied in rock engineering practice will be described.
The failure of jointed rock masses involves mobilization of the intact rock strength at smaller strain than that for the jointing.
This was also discussed by Müller in 1966. Continuum modelling with conventional strength criteria is questioned. A simple
solution is proposed. Excavation disturbed zones caused by joint shearing are compared with log-spiral failures. The reality of
jointing as affecting the deformation of a large cavern, as affecting NMT pre-injection design in rock masses, and as affecting
the compaction and subsidence behaviour of a reservoir are each illustrated. Discontinuous behaviour provides rich experiences
for those who value reality, even when reality has to be simplified by some empiricism.
Subject: General
Keywords: Rock joints, rock masses, physical modelling, empiricism, site characterization, tunnelling, rock failure
Table 2. Illustration of parameters CC (MPa) and FC◦ for a declining sequence of rock mass qualities, with simultaneously reducing σc
(MPa). Estimates of VP (km/s) and Em (GPa) are from Figure 9, whose derivation was described in Barton, 2002.
RQD Jn Jr Ja Jw SRF Q σc QC FC ◦ CC Vp Em
Figure 10. Over-break of 5 m and near break-through into the sapro- Figure 11. Systematic pre-injection and NMT permanent support
lite, following lack of pre-injection, and due to a cost-driving shallow (one final S(fr) layer is needed), gives an average 20 m per week for
metro choice by owners that took many years to complete. the completed tunnel, due to Q-improved rock mass properties.
10
rock joints. His studies, and some of the author’s multi-block Figure 19. NGI and Itasca UDEC-BB ‘uniaxial-strain’ M-H mod-
studies, are summarised in Bandis et al., 1981 and 1983. Fig- elling of a vertical 1.5 × 1.5 m 2D-element of jointed Ekofisk chalk,
ure 16 shows a graphic summary of the joint closure (N) and which had a porosity of 40%. Input data was obtained from JRC and
shear components (S) of deformation, which were utilised in JCS characterization of numerous joint samples that operator Philips
Barton, 1986 to explain the measured concave, linear and recovered as part of the extensive exploration process. A vertical total
convex load-deformation trends of three different styles of stress of 62 MPa was applied, and internal pore pressure was reduced
rock mass. The dominance or combination of the N and S from 48 to 28 MPa, causing shrinkage of the porous matrix, and con-
components are shown in the lower half of Figure 16, and sequent space for down-dip joint shearing, despite one-dimensional
confirmatory numerical modelling with UDEC-BB is shown strain. Barton et al., 1986.
in Figure 17.
from this giant off-shore field projected to last for 80 years,
there was naturally concern when a supply-boat captain appar-
7 MODELLING A COMPACTING SUB-SEA ently reported that a reduced number of wave-baffle holes (to
RESERVOIR protect the temporary oil storage tank) appeared to be above
sea levels in 1985, than at the beginning of production. These
The Ekofisk reservoir in the North Sea was discovered dur- holes are spaced at 2 m c/c. (See effect of 13 years time interval
ing the very last attempts to find oil in this ‘unpromising’ in Figure 18).
petroleum province, at the end of the 1960’s. With reservoir The down-dip shearing that can occur despite one-
development from about 1971, and large-scale production dimensional strain, is a fundamental necessity for the
11
12
REFERENCES
Addis, M.A., Barton, N., Bandis, S.C. & Henry, J.P. 1990. Laboratory
studies on the stability of vertical and deviated boreholes. 65thAnn.
Tech. Conf. & Exhibition of the Society of Petroleum Engineers,
New Orleans.
Bandis, S., Lumsden, A. & Barton, N. 1981. Experimental studies
of scale effects on the shear behaviour of rock joints. Int. J. Rock
Figure 25. The tilt test result can be extrapolated from 0.001 to Mech. & Min. Sci. 18, 1–21.
10 MPa, and can be performed on samples of 10 cm to 1 m in size. Bandis, S., Lumsden, A.C. & Barton, N. 1983. Fundamentals of rock
The same method has also been used on 5 m long as-built rockfills. joint deformation. Int. J. Rock Mech. & Min. Sci. 20: 6: 249–268.
Barton, N. 1971. A model study of the behaviour of steep excavated
rock slopes. Ph.D. Thesis, Univ. of London.
Barton, N., Lien, R. & Lunde, J. 1974. Engineering classification of
and JCS0 = 50 MPa, some aperture can be maintained even rock masses for the design of tunnel support. Rock Mechanics, 6,
without the benefits of shearing. 4: 189–236.
Barton, N. 1976. The shear strength of rock and rock joints. Int. Jour.
Rock Mech. Min. Sci. 13, 9: 255–279.
Barton, N. & Choubey, V. 1977. The shear strength of rock joints in
9 SHEAR STRENGTH AT EXTREME STRESS LEVELS theory and practice. Rock Mechanics, Vienna: Springer; 1/2:1–54.
Barton, N. & Hansteen, H. 1979. Very large span openings at shal-
In 1976 the writer proposed a ‘critical state’ concept for the low depth: Deformation magnitudes from jointed models and
shear strength of intact rock at high stress, which involved F.E. analysis, 4th Rapid Excavation and Tunnelling Conference,
both the expectation and the actual horizontal orientation of RETC; Atlanta Georgia, Vol. 2: 1131–1353. A.C. Maevis and
W.A. Hustrulid (Eds), American Inst. of Min. & Metall, &
the Mohr strength envelope. This condition was shown or sug- Petroleum Engineers, Inc. New York, New York, 1979.
gested by tectonophysicists’ triaxial tests of various rocks at Barton, N. & Bandis, S. 1982. Effects of block size on the shear
extreme confinement. Recently, this concept has been applied behaviour of jointed rock. Keynote, 23rd US Symposium on Rock
to better define the curvature of intact rock strength envelopes. Mechanics, Berkeley, California.
A few tests at low confining pressures provide all the data Barton, N., Bandis, S. & Bakhtar, K. 1985. Strength, deformation
needed for extrapolation to high levels of confinement. The and conductivity coupling of rock joints. Int. J. Rock Mech. &
elegant Singh et al., 2011 criterion heralds a new era in rock Min. Sci.. 22, 3, 121–140.
mechanics understanding. Barton, N. 1986. Deformation phenomena in jointed rock. 8th Laurits
Since a blend of theory and empiricism has been promised Bjerrum Memorial Lecture, Oslo. Publ. in Geotechnique, Vol. 36,
in the title, with possible application to problem solving in 2: 147–167.
Barton, N., Hårvik, L., Christiansson, M., Bandis, S., Makurat, A.,
rock engineering, the final figure to be presented will be of Chryssanthakis, P. & Vik, G. 1986. Rock mechanics modelling
the tilt testing concept. This is at the other end of an extreme of the Ekofisk reservoir subsidence. Proc. 27th US Rock Mech.
stress range.Tilt tests on joints of 10 cm and up to 1 m length, at Symp., Univ. of Alabama.
normal-stress-at-failure of 0.001 to 0.01 MPa, have provided Barton, N. & Grimstad, E. 1994. The Q-system following twenty
JRC0 and JRCn values for defining strength envelopes three years of application in NMT support selection. 43rd Geomechanic
to four orders of magnitude higher. Colloquy, Salzburg. Felsbau, 6/94. pp. 428–436.
13
14
Dohyun Park
Underground Space Department, Korea Institute of Geoscience and Mineral Resources, Daejeon, South Korea
ABSTRACT: This paper presents two different kinds of blasting techniques for reducing blast-induced vibration in tunnelling,
which are applicable according to the location of nearby structures: line-drilling; and bottom-air-deck. Numerical investiga-
tions through a non-linear hydrocode, which can numerically model chemical explosives, were carried out in order to obtain
quantitative information on the effectiveness of both methods, and then the numerical results were verified through field
experiments. The numerical and experimental results agreed well with each other, and indicated that both methods effectively
reduced blast-induced vibration according to the location of structures. From the present numerical and experimental study,
it was revealed that when both methods are applied to tunnelling sites, the vibration-reduction effects, that were predicted
through the numerical study, can be expected, and thus based on the numerical results, design conditions of both methods for
the reduction of blast-induced vibration in tunnelling were proposed by introducing the concept of factor of safety.
1 INTRODUCTION
17
18
19
20
Figure 10. Advances in the cases with and without the bottom-air-
deck. (a) length of the bottom-air-deck = 0 mm, and (b) length of the
bottom-air-deck = 300 mm.
Figure 8. Relationship between the vibration-reduction effect and where ARF (%) = amplitude reduction factor; Vno-red and
bottom-air-deck ratio. Vred = PVS values (mm/s) of the two cases (without and
21
22
23
24
Vibration-reduction effect
D and S = diameter and spacing of drill holes, respectively; DLD-BH = distance between line-drilling and blasthole
∗
The value was determined by applying a factor of safety of 1.2 to the numerically predicted effect.
25
Design conditions
Vibration reduction effect
Drilling length & Bottom-air-
Ground condition blast initiation deck ratio Predicted Designed
blasthole, the blast initiation that employs one hole per delay conditions of the ground and blasting are needed to extend the
was proposed. engineering application.
26
E.T. Brown
Golder Associates Pty Ltd, Brisbane, Australia
ABSTRACT: The 12th Congress of the International Society for Rock Mechanics (ISRM) is being held in the fiftieth year of
the life of the Society which was established formally on 24 May 1962 in Salzburg, Austria. This paper discusses the emergence
of rock mechanics as a distinctive engineering and scientific discipline, the state and achievements of the discipline at the time
the ISRM was founded in 1962, the events leading up to the formation of the Society, the development and achievements of
the Society in the 50 years since 1962, and the progress made in the discipline of rock mechanics and rock engineering since
that time.
Subject: Fifty years of the ISRM and rock mechanics
Keywords: General
This 12th Congress of the International Society for Rock To begin, it is worth recalling what the world was like in 1962.
Mechanics (ISRM) falls in the fiftieth year of the life of the In world politics, the Cold War was in full swing. The Pre-
Society which was established formally on 24 May 1962 in mier of the then U.S.S.R, Nikita Kruschev, and the young
Salzburg, Austria (Fairhurst 2010). This Congress therefore President of the U.S.A., John F. Kennedy, were exercised over
provides an important opportunity for the Society to recall its what became known as the Cuban missile crisis. Elsewhere,
foundation, reflect on its achievements and on the progress Harold Macmillan was the Prime Minster of the U.K., Charles
made in our discipline in the 50 years since its formation, and de Gaulle was still the President of France, Jawaharlal Nehru
to look forward to the next 50 years. was the Prime Minister of India and Chairman Mao Zedong
This paper attempts the rather large and complex task of was the leader of China. It has to be said that, at that time, the
reviewing and synthesising the emergence of rock mechanics People’s Republic of China was not at all well known to those
as a distinctive engineering and scientific discipline, the state living in what is usually referred to as “the West”.
and achievements of the discipline at the time the ISRM was In culture – popular and otherwise – the American writer,
founded in 1962, the events leading up to the formation of the John Steinbeck, won the Nobel Prize for Literature. There was
Society, the development and achievements of the Society in an unhappy link to one of the life-long interests of the ISRM’s
the 50 years since 1962, and finally, the progress made in the Founder and First President, Professor Leopold Müller, when
discipline since that time. Although too young to have been the world-famous Austrian violinist, Fritz Kreisler, died on
involved in the work of the Society in its earliest years before 29 January. In popular music, The Beatles released their first
the First Congress held in Lisbon in 1966, the writer has fol- recording and The Rolling Stones made their public debut.
lowed the developments discussed here with keen interest for In the writer’s favourite form of music – jazz – the pianist
about 48 of these 50 years. He is honoured and delighted to Herbie Hancock’s tune Watermelon Man topped not only the
have been invited by the Society’s current President, Professor jazz charts but the popular music charts as well for several
John Hudson, to prepare and present this paper. In a compan- weeks in 1962.
ion paper and lecture, Professor Hudson will look forward to In sport, Brazil won the World Cup in football; the NewYork
the next 50 years of the Society and the discipline (Hudson Yankees won the World Series in baseball; theAmerican golfer,
2011). Arnold Palmer, won his second consecutive British Open
Throughout this paper, the term rock mechanics will be used and his third U.S. Masters titles; and an Australian, Rod
in the sense given in the ISRM Statutes: “The field of rock Laver, won the Grand Slam in tennis.
mechanics is taken to include all studies relative to the phys- In science and technology, the first observation of laser
ical and mechanical behaviour of rocks and rock masses and activity was reported; the first practical light emitting diode
the applications of this knowledge for the better understand- was produced; the world’s first active telecommunications
ing of geological processes and in the fields of engineering”. satellite, Telstar 1, was launched on 10 July; the felt-tipped
Thus, the term rock mechanics will be generally taken to pen came into being; and, although digital computing was in
include rock engineering, although on occasion, the two may its infancy and the personal computer was still some years
be expressed independently as in the now commonly-used off, in a sign of things to come, the first computer game was
term, rock mechanics and rock engineering. developed for a PDP-1 computer.
29
30
31
32
33
Figure 4. Flowchart of rock mechanics modelling and rock engineering design approaches (Feng & Hudson 2004).
34
35
36
37
2010, Hawley et al. 2009, Schubert 2006) represents a great tunnels Sakurai (1997, 1999). Hoek (2001) discusses a simi-
advance on the comparable processes available 50 years ago, lar approach based on the normalised radial displacement at
although the purposes and principles of good monitoring the tunnel boundary.
programs remain essentially unchanged (e.g. Franklin 1977,
Kovari & Amstad 1993). There can be little doubt that modern
7.7 Rock engineering applications
monitoring systems have contributed significantly to the rock
engineering achievements to be outlined in Section 7.7 below. Over the last 50 years, the pre-existing rock mechanics knowl-
One of the common objectives of monitoring formalised edge outlined in Section 4, and the further development of that
by the seventh ISRM President, Dr John Franklin of Canada knowledge outlined thus far in Section 7, have contributed
(Franklin 1977), is to check the validity of the assumptions, to some remarkable rock engineering achievements in the
conceptual models and values of rock mass properties used general area of earth resources engineering (Fairhurst 2011),
in design calculations. Originally, and often still today, these including civil, mining, petroleum, energy resources and envi-
checks were made, and values of rock mass properties and the ronmental engineering. Space limitations preclude reference
in situ stresses estimated, using a process of trial and error or to all but a very few of these achievements.
curve fitting using forward analyses based on the theory of In civil engineering, the scale of rock engineering projects
elasticity. However, in the late 1970s and early 1980s, it was has steadily increased in terms of project dimensions, costs
realised that it was possible to improve this process by devel- and environmental significance (Feng & Hudson 2004). We
oping and applying more formal back analysis techniques. might think, for example, of achievements such as the 62 m
These techniques used essentially two approaches, the inverse span cavern for the Olympic Ice Hockey stadium at Gjøvik,
and the identification or calibration methods (Gioda 1980). Norway (Barton et al. 1994); the helical underground car
In the earliest use of the inverse method in rock mechanics park for the Sydney Opera House, Australia, possibly the
known to the author, Kirsten (1976) used measured displace- widest shallow-cover rock cavern in the world (Pells 2002,
ments in a slope and displacements calculated using the FEM Pells et al. 1991); the Three Gorges Dam and the associ-
to back calculate values of the elastic moduli. ated ship locks and hydro-electric developments in China
The ninth ISRM President, Professor Shunsuke Sakurai of (Feng & Hudson 2004, Liu et al. 2003a,b); and the long
Japan, was heavily involved in the development of back anal- European alpine tunnels such as the Lötschberg and Gotthard
ysis methods early in their history (Sakurai & Takeuchi 1983) base tunnels excavated under difficult conditions including
and subsequently (Sakurai & Akutagawa 1995, Sakurai et al. squeezing and faulted ground and brittle rock under high
2009). Although formal back analysis methods were applied stress at great depth (Hagedorn et al. 2007, Rojat et al.
originally to 2D linear elastic problems, the theory was soon 2008).
extended to the three dimensional and non-linear elasto-plastic In mining engineering, greater numbers of larger, deeper
problems encountered in rock engineering. A wide range of and more productive open pit mines with optimised slope
innovative techniques for solving the sets of non-linear equa- angles have been designed and operated effectively using a
tions involved have since been developed (Feng et al. 2004, range of advances in rock mechanics and rock slope engi-
Gioda & Sakurai 1987, Sakurai 1993). neering dating back to the seminal work of Hoek & Bray
Professor Sakurai has also emphasised the importance of (1974) and earlier. Read & Stacey (2009) provide a detailed
monitoring the performance of excavations in rock for pur- account of the state of this particular art. In underground
poses other than the use of the measurement data in back metalliferous mining over the last 50 years, rock mechanics
analyses. For example, in order to assess the stability of tun- has contributed significantly to the mining of highly stressed,
nels from measured displacements, he proposed the concept brittle, tabular orebodies in the deep-level gold mines of
of critical direct and shear strains, and on the basis of the South Africa (Cook et al. 1966, Ryder & Jager 2002), and
critical strain, proposed hazard warning levels for the mea- to the development and implementation firstly of cut-and-
sured displacements as well as for the measured strains in fill and then of long hole open stoping and bench stoping
38
39
40
41
42
43
44
45
The next 50 years of the ISRM and anticipated future progress in rock mechanics
J.A. Hudson
Department of Earth Sciences and Engineering, Imperial College of Science, Technology and Medicine,
London, UK
President International Society for Rock Mechanics (2007–2011)
ABSTRACT: This paper on the 50-year anticipated future of the ISRM and rock mechanics forms part of the ISRM 50-year
anniversary celebrations and complements the preceding paper by E T Brown in these proceedings on the previous 50 years of the
ISRM. The current status of rock mechanics and the unsolved problems are summarised—so that the results of extrapolating our
current capabilities can be considered. The problems are noted under the headings of geology, rock stress, intact rock, fractures,
water flow, modelling and design. The anticipated future developments are then highlighted under the headings of information
access, site investigation, subject integration, international co-operation, ‘intelligent’ computer programs, increased excavation
speed and larger excavations, environmental aspects, the future of the ISRM itself, and the implications of future computing
power. It is predicted that the greatest changes to rock mechanics and the ISRM itself will result from the continuing growth
in computing power.
Keywords: Rock failure, rock joints, lab testing, field measurements, numerical modelling, neural networks
1 INTRODUCTION
47
Despite the major progress that has been made in rock mechan-
ics and rock engineering over the last 50 years as described
by Brown (2011), there are still outstanding problems. In fact,
Leopold Müller’s motivation for founding the ISRM in 1962
was encapsulated in his May 1962 comment, “We do not know
the rock mass strength, and that is why we need an Interna-
tional Society”, yet we still have problems estimating the rock
mass strength!
In this Section, some of the major unsolved problems in
rock mechanics will be outlined. These are presented under
the subject headings of
Figure 4. More irregular fracturing in the Coal Measures strata,
• Geology South Wales, UK (∼1 m wide sample).
• Rock stress
• Intact rock
• Fractures
• Water flow
• Modelling
• Design
At the end of each of these secondary headings, italicised text
is included concerning a key point related to that particular
subject.
3.1 Geology
The roles of geology, particularly structural geology, and rock
mechanics for engineering are highlighted in Figure 2.
In both cases, structural geology and rock mechanics/
engineering, we are interested in deciphering the mechanisms
that led to the current circumstances and then in predicting Figure 5. Even more irregular fracturing in the Coal Measures
what will happen in the future. In fact, the quintessence of strata, South Wales, UK (∼1 m wide sample).
rock engineering is predicting the future. We wish to predict
what will happen in the future so that rock engineering designs
are appropriate; if such prediction is not possible, there can stretched without expert geological advice in obtaining the
be no coherent design. For example, we might wish to predict correct data for modelling and identifying the hazards that may
what will occur if a tunnel is driven in a certain direction and accompany certain geological formations. For example, many
depth as opposed to other directions and depths. rock mechanics modelling exercises require knowledge of the
In this context and because the engineered structure will rock fractures, but this may be difficult to obtain. The three
be founded on or contained within a rock mass, it is essential photographs in Figures 3–5 illustrate three different types of
to have an adequate understanding of the geological circum- fracturing—all observed in close proximity within the same
stances relating to the host rock, yet the engineer will be sorely Lower Middle Coal Measures strata in South Wales, UK.
48
49
50
51
In this Section, the future for rock mechanics and the ISRM
is predicted.
52
4.6 Larger, deeper and longer excavations Thus, when planning a new rock engineering project, we
should ask the question, “Given the existing system (whether
As the demand for materials and underground space continues a virgin rock mass or a rock mass already containing devel-
to increase, it is anticipated that there will be larger, deeper oped space), is it acceptable to locate a new facility in the
and longer excavations. Currently, one of the largest caverns rock mass, i.e., within the context of sustainable development
open to the public has a span of 61 m: the Gjøvik ice hockey as defined above? The whole subject of sustainable develop-
stadium in Norway. But how large can a cavern be? The Kirsch ment in rock engineering has not yet been properly tackled in
solution for the stress concentrations around a circular hole the above context and the subject urgently needs attention.
includes neither the absolute size of the hole nor the elastic
properties of the rock. In a rock mass free of fractures, such 4.9 The future of the ISRM
as can be found in the Canadian Shield, it should be possible
to construct caverns with spans considerably larger than 61 m. For the ISRM to have any purpose, it must provide a bene-
How far can we go? A 100 m span? A 1000 m span? No doubt fit to the members. Accordingly, the ISRM must continually
the record will soon be broken by some entrepreneurial group develop to provide the required benefits in our changing world,
of engineers. especially in relation to ‘immediate’ access to information.
The current Board of the ISRM, operating from 2007–2011,
has been engaged in a modernisation process to ensure that
4.7 Increased rate of mechanised excavation
the necessary benefits are provided to members. This has
It is difficult to imagine how rock support can be improved. included a survey of ISRM Members to establish what they
The rock blocks can be held together with rockbolts so that want, increasing the downloadable material from the ISRM
a fractured rock mass behaves literally more like a mono- website, the availability of papers from all the previous ISRM
lithic continuum. A resisting force can be induced through the symposia in pdf form via the OnePetro website, emailing a
use of introduced supports. Shotcrete can be used to provide digital newsletter to members on a regular basis, increasing
additional resistance and to combat deterioration of the rock the appreciation of the ISRM members’ achievements, ratio-
surface induced by weathering. nalising the style of the ISRM meetings, and organising lecture
Conversely, overall energetic principles indicate that mech- tours and field trips. There are many further possibilities for
anised excavation can be greatly increased. In Figure 15, the the future, including electronic conferences.
two ways that energy is input during current methods of tun- The advantages of interacting with other researchers and
nel construction are illustrated. When blasting, large amounts engineers has already been mentioned in previous sections
of energy are input for a few seconds at widely spaced inter- on coupled models and international co-operation. The same
vals, e.g., eight hours. This is represented in Figure 15 by the applies to increasing the co-operation with other Societies.
narrow vertical white bands. However, when a tunnel boring Additionally, cross-promotion of websites and events is being
machine (TBM) is cutting, a lower level of energy is contin- developed, together with interaction between similar com-
uously input, represented by the lower horizontal white band. missions and interest groups. However, some new method of
It is extraordinary that we have only these two excavation interaction needs to be developed. As is evident from the fact
methods characterised by the two extreme conditions. that the vast majority of papers published on rock mechanics
Since there is generally no problem in providing energy and rock engineering are by more than one author, interac-
to the tunnel construction, it is likely that new methods of tion is to everyone’s advantage. Perhaps, the current method
53
54
55
Q.H. Qian
Chinese Society for Rock Mechanics and Engineering
ABSTRACT: In terms of rock engineering and technology in hydropower construction, the slope stability and monitoring
techniques for high slopes of Three Gorges Project, the stability and support technology for high slopes of hydropower projects
in deep river valleys, the stabilization techniques for underground cavern group with large span and high side walls are
introduced in this paper. As for rock engineering and technology in highway and railway construction, the Qinghai-Tibet
Railway – new construction techniques in permafrost, the support techniques for large squeezing deformation in Wuqiaoling
Tunnel, the construction techniques for tunnels in alpine and high-altitude region, the geological prediction techniques for
tunnels in karst region, the microseismic monitoring and early warning techniques for rockbursts in deep and long tunnels are
presented. For rock engineering and technology in mining engineering, the innovative techniques for roadway support in mines,
the simultaneous extraction technique of pillarless coal and gas in coal seams with low permeability, the safe and efficient
deep open mining technology, advances in monitoring, early warning and treatment of mine dynamic disasters are presented.
In addition, the new anchorage techniques and precision blasting technique in rock engineering are introduced.
Keywords: Rock engineering; Hydropower engineering; Tunnel engineering; Mining engineering; Anchorage technique;
Blasting technique
1 INTRODUCTION
57
58
The reinforcement and treatment techniques for high slopes The shear-resistance tunnels were constructed on eleva-
mainly included: tions 1834 m, 1860 m and 1885 m. The cross section was
9 m×10 m. The fault and shear zones were removed. They
1) Water cut-off and drainage system: Cut-off ditches around
were then back filled by concrete after the surrounding rock
the slope, surface drainage ditches, underground drainage
masses were grouted.
holes and drainage adits were constructed so as to avoid
5) Grouting: The unloaded fractures and unloaded rock
water seepage into the excavated slope, which may induce
masses were grouted by utilizing the available anchor
slope failure.
holes. Grouting was carried out in the slope by using the
2) Reinforcement by rock bolts and anchor bars: In order to
hole for anchor bars in the slope berm and the pre-grouting
ensure rock mass stability and construction safety, system-
holes between berms.
atic rock bolts and anchor bars were employed to reinforce
the loose rock masses in the shallow excavated slope. A monitoring system for the slope was established. The
3) Layout of anchor cables: For the unloaded sandy slate monitoring program included: (1) horizontal and vertical
above elevation 1820 m and the rock mass at the outcrop displacements by geodetic survey; (2) monitoring of the dis-
of fault f42-9 near elevation 1800 m, anchor cables with placement in deep boreholes; (3) monitoring of displacement
prestress of 2000 kN were adopted (prestress of 3000 kN in deep rocks by the graphite rod convergence method and
may be applied when necessary). The spacing of anchor the distance measurement pier method; (4) monitoring of the
cables was 4 m×4 m and the length varied between 40– reinforcement effect, including stress in rock bolt and load
80 m. The unloaded marble rock masses were reinforced in anchor cable; (5) inspection monitoring; (6) monitoring of
by randomly distributed anchor cables with prestress of opening and closing of cracks, shear and dislocation of faults;
2000 kN–1000 kN. The shallow rock masses near the slope (7) monitoring of seepage pressure. More details are listed in
open line were reinforced by anchor cables with prestress Table 1.
of 2000 kN and spacing of 5 m×5 m. The layout of monitoring instruments is shown in Figure 3.
4) Replacement by shear-resistance tunnel: Fault f42-9 was In order to effectively analyze and manage the vast amount
the potential slip surface for overall slope failure. Shear- of monitoring data, a high slope stability monitoring system
resistance tunnels were constructed for reinforcement. for hydropower engineering was developed. With this system,
59
Dip slope Toppling deformation at Surface displacement, 10 observation points along the external border of the top
structure prone the back slope of the shallow rock excavation line, 3 monitoring sections in the excavated
to toppling cableway platform above deformation, load in slope (A-C), 15 external observation points, 13 multi-point
elevation 2000 m anchor cable (rock bolt) extensometers, 13 rock bolt dynamometers, 74 anchor line
dynamometers based on 5% of the number of anchor cables
Wedge-shaped Rock mass with Surface displacement, 8 monitoring sections below elevation 1960 m, 55
structure with tensile fracture (large relative displacement external observation points, 20 multi-point extensometers
double sliding block between elevation of structural planes passing through fault f42-9 and lamprophyre dyke (X) and
surface 1800–2020 m) 54 rock bolt dynamometers
Block-shaped Small local blocks Surface displacement, 8 monitoring section below elevation 1960 m, 41
slope structure outside the rock mass displacement in shallow multi-point extensometers between elevation 1885–1960 m,
with tensile fracture rock mass, load in anchor 82 cable line dynamometers, 27 rock bolt dynamometers,
and shallow rock masses cable (rock bolt) 3 monitoring points for groundwater, 4 boreholes for
inclinometers. Monitoring points for surface displacement
of other blocks
Deep fractures Five-layer monitoring Horizontal and vertical Exploratory pits PD44, PD42 at elevation 1930 m, drainage
pits displacements in the tunnel at elevation 1915 m and curtain grouting pits at
directions perpendicular elevation 1785 m, 1829 m and 1885 m as monitoring pits,
to the river 1–2 measurement lines for horizontal displacement
perpendicular to the river, 9 convergence graphite rods,
10 measurement piers, 3 sliding micrometers
Shear-resistance Three-layer shear- Deformation in the 25 monitoring sections in 8 caverns, 24 three-point
replacement resistance tunnel for surrounding rock mass, extensometers, 8 single-point rock bolt dynamometers,
tunnel fault f42-9 , two-layer deformation in faults and 16 three-point rock bolt dynamometers, 28 dislocation
shear-resistance tunnel dykes, stress and strain in meters, 18 joint meters, 16 rebar strain meters, 12 osmometers,
for fault f5 , three-layer steel-reinforced concrete 10 sets of 5-D strain gauges, 10 non-stress meters
replacement tunnel for
lamprophyre dyke (X)
Resisting force Four-layer monitoring Horizontal and vertical 12 monitoring sections in certain grouting pits at elevations
body in the pits displacements 1829 m, 1785 m, 1730 m and 1670 m, 3–4 measurement lines
left bank for horizontal displacement perpendicular to the river,
27 multi-point extensometers, 27 rock bolt dynamometers,
36 rebar strain meters, 4 joint meters, 12 convergence sections
management and integration of monitoring data, construction walls exists not only during construction, but also during the
information, geological information, hydrological and meteo- long-term operation period.
rological information can be implemented. The system has the The main caverns for a large-scale underground hydropower
functions of data mining and analysis, prediction and graphic station in the powerhouse, the transformer chamber, the surge
output of analysis results. The system can realize comprehen- chamber or tailrace gate chamber, as shown in Figure 4.
sive and dynamic management of distributed network for field Different excavation sequences for underground cavern
monitoring data. group would result in different stress paths. Research on
With the above slope reinforcement measures, systematic the optimum construction sequence is one of the key issues
monitoring indicated that the slope deformation gradually in the study on the stability and support measures for
converged within a few years, which proves that the reinforce- large-scale underground cavern groups. Currently, for the
ment measures were effective. Successful implementation of study on excavation sequence of underground cavern groups,
the slope construction and monitoring indicates that the slope the analysis method has been developed from planar anal-
engineering in China has achieved a new leap and the level of ysis to quantitative analysis reflecting three-dimensional
engineering geological investigation, stability analysis, rein- nonlinear characteristics of large-scale underground cavern
forcement techniques, monitoring system for excavation and groups. In combination with the optimization of construc-
construction has been advanced. tion organization, the mechanical analysis and construction
system engineering can be integrated closely for quantitative
2.3 Stabilization techniques for underground cavern group assessment.
with large span and high side walls By summarizing the construction conditions of the under-
ground cavern groups which have been completed or are under
In recent years, a great number of underground caverns construction, the principle of excavation sequence is proposed
(groups) have been constructed for energy resources, trans- as “arch first and then side walls, from top to bottom, excava-
portation, mining and defense industries. Taking the under- tion in layers, support installation in layers, multiple processes
ground hydropower station as an example, the span, height in the horizontal direction, multiple layers in the vertical
and scale have been in the front ranks of the world. direction, cross operation of multiple work faces”. Figure 5
The stability problem of the surrounding rock masses of shows a sketch of stratified excavation of the powerhouse and
underground caverns (groups) with large span and high side transformer chamber.
60
Figure 4. Sketch of cavern group for a large-scale underground Figure 5. Sketch of stratified excavation of powerhouse and trans-
powerhouse. former chamber (unit: cm).
61
62
63
64
65
66
67
68
Xia-Ting Feng
State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics,
Chinese Academy of Sciences, Wuhan, China
ABSTRACT: This paper reviews new rock mechanics developments in China in the last ten years. A series of devices for
laboratory tests, in situ tests and monitoring have been developed. Also, strength criteria based on energy and for hard rocks,
progressive and anisotropic damage models, coupled THM models, new creep models, the vector sum method, reinforcement
theory, and a multi-scale and non-linear thermodynamic theory were established. There have been major developments in
studying zonal disintegration, rock dynamics, fractals in rock mechanics, rock fracturing process simulation methods, and
intelligent rock mechanics methodology. The 3-D limit equilibrium methods, strength reduction method, FEM, BEM, NMM,
Meshless method, and DDA were updated. There have been considerable efforts on mechanism understanding through lab
and real time field monitoring, prediction and the mitigation of rockbursts. The need for rock engineering projects has largely
promoted the development of rock mechanics; and rock mechanics has provided a practical tool to support complicated rock
engineering problems. Further development of rock mechanics in China is anticipated and more applications of rock mechanics
to rock engineering are expected.
Keywords: Lab tests, field measurements, back analysis, neural network, numerical modelling, dynamics
1 INTRODUCTION This paper just provides a brief review of the new devel-
opments in rock mechanics in China over the latest ten years
Rock Mechanics is a discipline having strong practical appli- and covers lab and field tests, mechanical models and crite-
cations. The construction of large-scale rock engineering ria, analysis methods for rock samples, slopes, tunnels, and
projects in China, including high/steep slopes, large cavern cavern groups, rock dynamics and zonal disintegration of a
groups, mining stopes, tunnels, oil wells, and foundations, deep rock mass, and applications of rock mechanics to rock
has promoted the development of rock mechanics in China. engineering. In the years ahead, further development of rock
Oil and gas underground storage, nuclear waste disposal, mechanics in China is anticipated and more applications of
CO2 sequestration, and geothermal development have pro- rock mechanics in rock engineering are expected.
moted the development of coupling THMC processes in rocks.
Sun & Wang (2000) have summarised the development and
the state-of-the-art of rock mechanics in China before 2000. 2 LABORATORY & FIELD TESTING TECHNIQUES
The ‘century achievement’ in rock mechanics in China has
been summarised in the book published by Hohai University 2.1 Laboratory testing techniques
Press (Wang, 2004).
In memory of Tan Tjong Kie, who initiated the Interna- In order to enhance understanding of the rock fracturing
tional Society for Rock Mechanics (ISRM) China National process under complicated environmental conditions, test-
Group in 1979, was the first President of the Chinese Soci- ing systems were developed using a medical or industrial
ety for Rock Mechanics and Engineering (CSRME) and was CT scanner with a loading system and a microscope with
the ISRM Vice-President for Asia during 1983–1987, the Tan CCD cameras. Ge (2004) developed firstly a CT real-time
Tjong Kie Lecture series was initiated in 2008. Three lectures testing technique for triaxial testing and study of the meso-
have been given by: damage evolution law. Through CT scanning, clear CT images
can be obtained which include the micro-hole (microcrack)
• Sun Jun (2007), ISRM Vice-President at Large 1995–1999
compressed ⇒ the microcrack growth ⇒ the microcrack
and President of CSRME 1994–1998; bifurcation ⇒ the microcrack development ⇒ crack frac-
• Wang Sijing (2009), President of CSRME 1999–2003; and
ture ⇒ rock sample destroyed ⇒ unloading. This technique
• Qian Qihu (2009), ISRM Vice-President at Large 2003–
was further developed by Xie Heping’s group using an indus-
2007 and President of CSRME 2003–present. trial CT scanner. Chen Houqun led a group in developing a
The 4th Tan Tjong Kie Lecture will be given by Ge Xiuren portable real-time dynamic loading apparatus with the sup-
this year on the measurement of in situ stress and its estima- port of the medical CT (Tian et al., 2010). The apparatus is
tion. Following the development of Chinese Journals, such as available for performing tests such as dynamic tension, com-
the Chinese Journal of Rock Mechanics, the Chinese Jour- pression test with impact, triangular and sine wave with the
nal of Underground Space and Engineering, Rock and Soil greatest load output being 100 kN, and frequency of 5 Hz.
Mechanics, a new English language journal, the Journal of Under dynamic loading, the whole process of crack propaga-
Rock Mechanics and Geotechnical Engineering, was launched tion and coalescence, CT images are obtained. It is laying the
in 2009. foundation for dynamic CT testing of brittle materials.
71
72
73
74
75
76
77
78
79
A.J. Gonzalez-Garcia
School of Engineering, National University of Colombia, Bogota D.C
ABSTRACT: In this paper two issues about rock strength are reviewed by the Author: a common one (a) rock strength and
brittleness and a supposedly uncommon one (b) intact rock, its definition and methods to obtain its value. For the first issue,
a proposal is presented for a new simple definition for brittle-ductile transition behaviour. For the second issue an alternative
laboratory method to obtain intact rock strength is proposed, instead of the recommended triaxial tests; also a refined definition
of intact rock is recommended, which leads to briefly examine the topic of rock sample disturbance.
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 SHEAR STRENGTH AND ITS COMPONENTS mainly from density and angularity of particles in granular
soils and from roughness in rock discontinuities. It is usually
A characteristic of ceramic materials, such as earthen materials represented by ν = tan i.
(soils and rocks) is that they are very weak in tension and Solid friction, or surface resistance, is a very complex phe-
strong in compression, as compared, for instance, with metals, nomenon. It is totally ductile and plastic, and it is not possible
which can be more or less equally strong under tension and to eliminate it, unless contact between surfaces is lost. It is the
compression, all of these being very well known facts. last component of strength in all solid materials and in earthen
Since solids only can fail under tension or shear stresses, materials it can be considered a property of minerals and it is
other types of failure being really an indirect result of these represented by µm = tan φm
stresses (compression and torsion failures are really shear Usually, the components of dilatancy v and friction µm
failures, flexure failure is a combination of tension failure are combined into the so-called effective friction µ =
and shear failure, etc), shear strength is the preferred strength tan(φm + i) = tan φ = dτ/dσ , and cohesion is represented by
parameter for geotechnical materials. the cohesion intercept c (τ for σ = 0). However, in curved
Shear strength can be postulated as composed of three envelopes, such as those of coarse granular materials and rock
main parts: cohesion c , dilatancy ν and friction µ , with the discontinuities (in which σt = 0) or rock materials and masses
apostrophe to denote effective values (Figure 1) (in which σt > 0), c tends to lose its meaning, that comes from
Cohesion, or internal strength, is given by crystal molecu- the Mohr-Coulomb linear equation (Eq. 1).
lar bonds, electrochemical surface forces or cements, and it
really comes from tensile strength σt . In preconsolidated clays,
residual soils and rock materials it is the main component of 2 STRENGTH CRITERIA
strength: it is usually rigid, somewhat elastic but very brittle,
so it is quickly lost with deformation. Prior to its maximum In this paper the following well-known criteria will be used:
value it provides elasticity to the earthen material. Mohr-Coulomb as a general one, Hoek-Brown (Hoek et al,
Dilatancy, or resistance to volume change, is brittle, not 2002) for rock materials and masses and Barton-Bandis
very rigid, elasto-plastic and can be positive or negative. It is (Barton&Bandis, 1982, 1990) for rock discontinuities.
one of the main strength components of granular materials and 2.1 Mohr-Coulomb Strength Criterion
rock discontinuities. It varies with normal stress and comes
where
where
81
where
82
Figure 4. Mohr-Coulomb diagram for quartzitic sandstone rock masses and discontinuities–close-up of Figure 3.
Examining these quartzitic sandstone envelopes: curves 2 3.2 Rock strength and brittleness
(intact rock), 3 (average core sample) and 4 (good rock mass),
Considering brittleness and using the common Mogi crite-
all show a clear behaviour, brittle condition and descending
rion for brittle-ductile transition σ1 /σ3 = 3.4 (Mogi, 1966 in
strength.
Mogi, 2007), equivalent to τ/σ = tan(33.06◦ ), and examining
However, curves 5 (poor rock mass), 6 (discontinuity-
Figure 4, the following facts emerge:
lab), 7 (discontinuity-field) and 8 (discontinuity-residual
friction) seem intermingled and without a clear distinctive a) Quartzitic sandstone poor rock mass strength becomes
behaviour, so a closer look is taken at these last four curves ductile for normal stress σ > 40.7 MPa (upper circle)
(Figure 4). beyond which H-B strength criterion is no longer valid.
83
84
85
86
87
Figure 10. Hoek-Brown rock strength envelopes for example – Figure 12. Hoek-Brown rock strength envelopes for example – poor
good rock mass (GSI = 70) – distorted σ3 axis. rock mass (GSI = 30) – close-up with distorted σ3 axis.
88
89
J.V. Lemos
LNEC – Laboratório Nacional de Engenharia Civil, Lisbon, Portugal
ABSTRACT: The safety assessment of concrete dam foundations entails the examination of the potential failure mechanisms,
typically defined by natural rock discontinuities or the concrete-rock interface. Numerical models which represent the rock
mass as a discontinuous medium, in particular discrete element models, are particularly adequate for the analysis of these
failure scenarios, given their ability to represent the geologic structure of the rock mass, as well as the concrete structure.
The application of discrete element deformable block models to the safety assessment of arch and gravity dam foundations is
discussed with reference to specific examples. The issues involved in model generation, such as the representation of the rock
discontinuities or the application of joint water pressures, are examined, as well as the procedures for safety factor evaluation.
91
or gap elements (e.g. Alonso et al. 1994). The development 3 ARCH DAM FOUNDATION ANALYSIS
of the distinct element method by Cundall in 1971 was aimed
at a direct representation of a discontinuum. Presently, the 3.1 The case of Baixo Sabor dam
designation of “discrete elements” (DE) covers a wide family
of numerical methods (distinct elements, discrete finite ele- The main topics involved in the application of DE models to
ments, DDA, etc.), all sharing the concept of representing a the analysis of failure mechanisms in arch dam foundations
discontinuous medium as an assembly of blocks or particles. will be discussed resorting to a few examples, in particular to
These idealizations are applicable at various scales of analy- Baixo Sabor dam, presently under construction. The Hydro-
sis. At one end, we have the micromechanical models based on electric Project of Baixo Sabor, owned by EDP, is located in
many rigid particles or blocks employed, for example, in frac- the north-east of Portugal in the lower branch of the Sabor
ture analysis at lab test scale. At the other end, the deformable river, a tributary of the right bank of the Douro river. It is
block models of large civil and mining engineering works. composed of two dams, a 123 m high arch dam upstream, and
Although the increase of computer power is expanding the a 45m high gravity dam downstream. Both powerhouses will
range of the micro-models (Cundall 2001), it is the latter type have reversible units to enable pumping from the Douro river
of DE model that is most suitable for dam foundation analysis to the reservoir created by the upstream dam with a capacity of
(Fig. 1). 1.1 × 106 m3 . The design was performed by EDP (Matos et al.
The analysis of collapse mechanisms in dam foundations 2007). The arch dam has a crest length of 505 m and a total
involves the representation of the discontinuities where sliding concrete volume of 670000 m3. The 3DEC model for foun-
may take place. In this field of application, the block structure dation safety assessment, shown in Figure 1, is described in
thus defined is better replicated in a numerical model by means more detail by Lemos & Antunes (2011). The full model com-
of deformable blocks. In this way, a more realistic simulation prises about 2300 deformable blocks with 26100 grid-points
of the distribution of structural loads is obtained, influenced by (the internal block mesh is not shown in the figure).
the foundation properties and their spatial variation, even with
a fairly coarse block system. In the code 3DEC (Itasca 2006),
3.2 Model generation
deformable blocks are obtained by internal discretization into
a finite element mesh of tetrahedra. For dam foundation stud- Various techniques and software are now available for the gen-
ies, these rock blocks are typically assumed elastic, with all eration of numerical models, with most analysis codes capable
the nonlinear behavior concentrated on the joints. For arch of using a combination of built-in tools and data input from
dams, the correct bending behavior is more easily achieved external geometric modelers. For dam foundation analysis
with higher order elements, thus 3DEC allows 20-node bricks with the code 3DEC, the model generation typically starts by
to be used for the concrete vault. In this type of model, the reading a FE mesh of the dam. The mesh of the Baixo Sabor
vertical contraction joints and the concrete-rock interface are dam model is plotted in Figure 2, with the contraction joints
also discontinuities which may be assigned general constitu- separating the structure into the cantilever blocks. This mesh
tive models. The Mohr-Coulomb model is the most widely consists of higher order brick elements (in the plot the faces
used, but many other rock joint models exist. are triangulated). More than one element across the thickness
The variety of numerical techniques presently available is normally recommended to provide a finer discretization of
often brings the question of what are their real differences, the foundation surface. Block interaction mechanics in most
e.g. how does a DE deformable block model differs from a FE DE codes is based on sets of point contacts, rather than inter-
model with joint elements. If both models share the constitu- face elements, which allows the contact between blocks with
tive assumptions regarding block material and joint behavior, unmatched meshes, as happens in the present case along the
then their response should be not be dissimilar. It is mainly foundation surface (e.g. Lemos 2008).
the numerical approach that sets them apart. FE models rep- The global model geometry may be created, in the simplest
resent block interaction by means of joint elements, while DE case, by extending the model upstream and downstream from
models typically use point contacts. FE packages favor matrix the dam-rock interface, assuming a cylindrical valley shape,
and implicit solvers, while DE codes obtain static solutions as shown at the top of Figure 3. Alternatively, the actual sur-
by dynamic relaxation. The same explicit algorithm, but with face topography may be input, as a new layer of blocks placed
real values of damping, is employed in time domain dynamic above, as shown at the bottom of the same figure. The model
analysis. Most standard FE analyses assume geometric linear- in the top figure may be used for the analysis of sliding mecha-
ity, while DE codes are designed to extend the solution into nisms on the dam-rock interface, in the simplified scenario of
92
93
94
95
96
97
98
G.W. Ma
School of Civil and Resource Engineering, University of Western Australia, Crawley WA, Australia
Keywords: Numerical modeling; Rock mass; Stability analysis; Rock failure; Rock joints.
1 STATE OF THE ART OF NUMERICAL METHODS IN nonlinearity, thus is more suitable for solving problems in
DISCONTINUOUS DEFORMATION ANALYSIS homogeneous and linearly elastic bodies.
The finite element method (FEM) has been the most popu-
Rock mass, as a typical natural geological material, are com- lar numerical method in rock mechanics and rock engineering
plex with different extent discontinuities including bedding due to its flexibility in handling material heterogeneity and
planes, faults, shear zones, joints and fractures. To explic- anisotropy, complex boundary conditions and dynamic prob-
itly characterize such mechanical discontinuities, various lems, and efficiency in dealing with complex constitutive
numerical methods have been developed. These methods are models and fractures. Representation of fractures in the FEM
generally classified into three categorizes, i.e., continuum has been motivated since the late 1960s. Various joint element
based methods (e.g., the FDM, the BEM, the FEM, the mesh- or interface element models such as ‘Goodman joint element’
less methods, etc), discontinuum based methods (e.g., the (Goodman et al., 1968) have been developed. Despite these
distinct element method (DEM), the discontinuous deforma- efforts, explicitly modeling fractures and their growth remains
tion analysis (DDA)), and combined continuum-discontinuum limited in the FEM. In the FEM, the finite element mesh must
methods. be constructed in such a way that element edges/faces coin-
The finite difference method (FDM) is the earliest mem- cide with the crack surfaces and nodes must be placed on each
ber in the family of numerical methods. The implementation side of the crack to allow material separation. The construction
of the FDM is simple in both two-dimensional and three- of such meshes becomes more difficult, especially in three-
dimensional cases. However, the conventional FDM using dimensional cases. When discontinuity evolution needs to be
regular grid system suffers from inflexibility in dealing with modeled, remeshing is inevitable, which makes the simulation
material heterogeneity, fractures, and complex boundary con- tedious and time-consuming. Moreover, during remeshing
ditions, which restricts its application in rock mechanics. process, variables such as displacement, stress and strain need
In order to overcome such shortcomings, the finite volume to map to a new set of nodes and quadrature points, which may
method (FVM) was developed, which allows for the use of result in additional inaccuracies.
unstructured meshes, and is flexible in non-linear material In order to overcome such inconveniences in meshing and
models and boundary condition enforcement. Explicit repre- remeshing processes, a variety of modifications to the conven-
sentation of fractures is possible in the FVM by incorporating tional FEM have been made within the framework of partition
special ‘fracture elements’ as in the FEM, such as reported in of unity (PU). One typical example is the extended finite ele-
Granet et al. (2001). ment method (XFEM). It was first realized in Belytschko and
The boundary element method (BEM) requires only dis- Black (1999) by enriching the nodes of the finite elements near
cretization at the boundary of the solution domain, thus the crack tips and along the crack surfaces with the asymp-
reduces the problem dimensions by one and results in fast totic crack tip functions. Later, Moes et al. (1999) adopted
computing speed and easy mesh generation. One typical BEM the generalized Heaviside function H(x) (H(x) = 1 when x
is the displacement discontinuity method (DDM), which has above crack surface, and −1 when x below the crack surface)
been widely applied to simulate rock fracture propagation instead of asymptotic crack tip functions to introduce a dis-
problems by incorporating single fracture elements (Wen and continuous field across the crack surface away from the crack
Wang, 1991). However, in general, the BEM is not effi- tips. The finite element mesh is allowed to be constructed
cient as the FEM in dealing with material heterogeneity and totally independent of the cracks and remeshing is completely
99
100
101
102
103
104
Figure 9. Discontinuous rock mass. Figure 11. Manifold treatments of a Rock slope.
105
106
REFERENCES
107
108
ABSTRACT: A new approach to the prediction of rock mass behaviour, called Synthetic Rock Mass (SRM) modeling, has
been developed in response to the demands of the mass mining industry, where successful caving and fragmentation of the rock
mass demands a greater understanding of the joint fabric and its impact on rock mass strength brittleness and fragmentation.
The method involves embedding a Discrete Fracture Network (DFN) into a Bonded Particle Model (BPM) of rock. Much of
the work in producing a realistic SRM sample centers on accurate reproduction of the joint fabric present in the real rock mass.
This paper focuses on some of the challenges associated with this process in the context of mass mining applications. It also
touches on some of the challenges associated with representation of intact rock (and embedded joints) and concludes with a
short review of SRM validation efforts.
Subject: Modelling and numerical methods
1 INTRODUCTION
109
gravity flow models. A number of 3D SRM samples normally Figure 5. Anisotropy in rock mass modulus derived from lab tests
are produced for each geomechanical domain at representative on 12-m diameter SRM samples with the joint orientations shown in
scales (normally around 5–20 m, depending on the joint spac- Figure 4.
ing) and inputs are derived from standard core logging, lab
testing and underground mapping. The flow chart in Figure 3 model. These are well-suited to situations in which there is
outlines the general methodology developed within the MMT often insufficient data to justify the use of more complex (e.g.,
research project to obtain predictions of cave growth and mate- terminating) joint models and have (as a result) been employed
rial flow at the mine scale from a combination of empirically in most of the SRM studies conducted to date. The goal is to
derived and SRM-derived rock mass properties. This process produce joint networks that honor the measurements of open-
also includes validation of both the SRM-scale and cave-scale joint orientation, density (spacing) and size distribution (trace
models through comparison with measurements of microseis- length) that are obtained from drilling and mapping. The fol-
micity, time domain reflectometry (TDR) breakages (or open lowing sections discuss a few of the challenges associated with
hole blockages) and drawpoint fragmentation made on site. measuring each of these three properties, as well as their poten-
tial impacts on SRM and rock mass behavior. Hadjigeorgiou
3 DFN SIMULATION (2011) outlines some further challenges associated with the
data gathering and DFN generation process for mining appli-
Much of the work in producing SRM sample centers on accu- cations (e.g., the identification of open versus sealed joints).
rate reproduction of the joint fabric present in the real rock
mass. Discrete Fracture Network (DFN) simulation provides
3.1 Joint orientation
a convenient and practical means to build joint networks for
this purpose. There are many different approaches to DFN When there is a wide range in joint orientations (e.g., Fig. 4),
simulation and they vary considerably in their complexity and the process of joint-set selection is bound to exclude a large
underlying assumptions (Staub et al., 2002, Hadjigeorgiou, number of randomly orientated joints. As a result, the orien-
2011). Research is being conducted within MMT to inves- tations are generally bootstrapped, i.e., sampled directly from
tigate how sensitive the strength and deformation behavior the entire range of mapped orientations. Results of SRM test-
of SRM samples is to the underlying DFN model, and how ing suggest that marked anisotropies in rock mass strength
important it may be to consider more state-of-the-art models and modulus still emerge from weak concentrations of joint
such as the Universal Fracture Model (Davy et al., 2010). This orientations (Fig. 5).
paper focuses on the challenges associated with the devel- Joint orientations generally are the easiest of the three
opment of DFNs employing a relatively simple random disk main inputs to obtain from mapping. The main challenge is
110
111
112
113
114
T. Ramamurthy
Department Angron Geotech Pvt. Ltd., Naraina Industrial Area, Phase II, New Delhi, India
A. Ghazvinian
Rock Mechanics Division, Faculty of Engineering, Tarbiat Modares University, Tehran, Iran
ABSTRACT: Strength and modulus of rock mass suggested from RMR, Q and GSI have been examined with modulus ratio,
Mrj , for their reliability. The design parameters adopted in some case studies based on these rock mass classifications are
presented. The modulus ratios of rock masses are found to be much higher than those of intact rock, even with back analyses.
Based on joint factor, Jf , compressive strength, modulus, cohesion and friction angle were estimated and applied in the analysis
of a few cases. The predictions of deformations agreed well with the field measurements. Estimation of normal (Kn ) and shear
(Ks ) stiffness of rock mass under various conditions of joint system is covered in some detail. When dilatancy component is
separated in direct shear tests, the scale effect appears to be insignificant. Further, when to adopt equivalent continuum model
for rock mass is suggested.
Keywords: numerical modeling, rock joints, rock mass, rock mass classification, rock properties and site characterisation.
1 INTRODUCTION
115
Bieniawski (1973) suggested shear strength parameters, cj For estimating the deformation modulus, Hoek (1994) rec-
(cohesion) & φj (friction angle), subscript j refers to rock mass, ommends the use of Eq. (2) as per Serafim & Pereira using
for five levels of rock mass classes. With these values of cj & RMR as per Bieniawski (1973) and not the GSI system. The
φj the uniaxial compressive strengths (σcj ) of the mass are values of GSI, sj & Ej given in Table 5 of Hoek (1994) have
calculated as per Mohr-Coulomb criterion, Eq. (1). been adopted in calculating the values of modulus ratio, Mrj .
The values of Mrj are surprisingly high, ranging from 1500 to
1875 with an average value of 1642. These values of Mrj do not
decrease, with the decrease of GSI values, (refer Ramamurthy
The values of Ej are estimated from Eq. (2) as per Serafim & 2004).
Pereira (1983),
7 PARAMETERS USED IN ANALYSIS AND DESIGN
116
(vi) For Masjed – e – Soloiman, HEPP, Iran (Stabel & Samani is given by Eq. (9) and the corresponding modulus by Eq. (10).
2003), 2D – elasto-plastic hybrid FEM & BEM applied Equation (9) as per Ramamurthy (2001),
using the properties of the layer mass.
Conglomerate: Mri = 789, Mrj = 1136, Mrj /Mri = 1.44
Sandstone: Mri = 373, Mrj = 1022, Mrj /Mri = 2.74
Siltstone: Mri = 333, Mrj = 2372, Mrj /Mri = 7.12
Claystone: Mri = 348, Mrj = 3896, Mrj /Mri = 11.2.
(vii) Intake tunnel, Karuna III, HEPP, Iran: the analysis of the wherein Jf is a joint factor defined by Eq. (11)
tunnel was carried out using 2D, UDEC and 3D elastic
programmes (Tabanrad 2003). Measured deformations
agreed with the estimated values from block analy-
sis. Back analysis was carried out using direct method where, Jn = joint frequency, i.e., number of joints/meter,
UDEC software and σcj and Ej were estimated. For Marly which take care of RQD and joint sets and joint spacing;
limestone/marl: Mri = 221, Mrj = 2885, Mrj /Mri = 13.05 n = inclination parameter depends on the inclination of sliding
(viii) A stochastic analysis was carried out to estimate σcj and plane with respect to the major principal stress and anisotropy;
Ej of three grades of Ankara andesites by calculating the joint or set which is closer to (45 – φj /2)◦ with the major
the influence of correlations between relevant distribu- principal stress will be the most critical one to experience
tions on the simulated RMR values (Sari et al. 2010). sliding; r = a parameter for joint strength; it takes care of
The model was also used in Monte Carlo simulation to the influence of closed or filled up joint, thickness of gouge,
estimate possible ranges of the Hoek – Brown strength roughness, extent of weathering of joint walls and cementa-
parameters. tion along the joint. This factor could be assessed in terms
From minimum strength & modulus values of an equivalent value of friction angle along the joint as
Grade A: Mri = 520, Mrj = 7000; Mrj /Mri = 13.46 tan φj = τj /σnj obtained from shear tests, in which τj = shear
Grade B: Mri = 470, Mrj = 2113; Mrj/ Mri = 5.89. strength along the joint under a normal stress, σnj . The values
Grade C: Mri = 359, Mrj = 1311; Mrj/ Mri = 3.65. of n and r are given in Tables 1, 2 and 3.
From maximum strength & modulus Now from Eqs. (9) & (10), the modulus ratio of the jointed
Grade A: Mri = 425, Mrj = 1568; Mrj/ Mri = 3.69. mass with respect to that of the intact rock is given by Eq. (12)
Grade B: Mri = 431, Mrj = 1378; Mrj/ Mri = 3.20.
Grade C: Mri = 326, Mrj = 963; Mrj/ Mri = 2.90.
(ix) Power house Cavern, Rogun project, Kajikistan:
To predict deformations of roof and slide walls using 3D
FEM with M-C criterion (Bronshteyn et al. 2007) adopted For values of σcj & Mrj for different values of Jf varying
reduced ci , from 0 to 500 for σci = 100 MPa & Mri = 500 of intact rock,
φi and Ei as indicated here. the Mrj values rapidly decrease to 87 with the increase of Jf .
For sandstone: c = 3.0 to 1.1 MPa (F.S. = 2.7) The relationship between Ej & σcj (i.e., Mrj ) can not be taken
φ◦ = 50 to 38 (F.S. to tan φ: 1.53) as constant when the rock mass is experiencing continuous
E = 9000 to 5000 MPa (F.S. = 1.4) fracturing and undergoing change in its quality.
Mri = 546, Mrj = 1109; Mrj /Mri = 2.03 Even though the original classification due to Deere &
For aleurolites: c = 2 to 0.5 MPa (F.S. = 4.0) Miller was suggested only for intact rocks, it was modified to
φ◦ = 45 to 32.5 (F.S. to tan φ: 1.56) classify rock masses as well (Ramamurthy 2004). The main
E = 5500 to 2670 (F.S. = 2.06) advantage of such a classification is that it not only takes into
Mri = 569, Mrj = 1467; Mrj /Mri = 2.5. account two important engineering properties of the rock mass
(x) Waitaki dam block No. 10, New Zealand: A 3D FEM but also gives an assessment of the failure strain (εf ) which
analysis was carried out for the Waitaki dam (Richards & the rock is likely to exhibit in the uniaxial compression, when
Read 2007). The tiltmeter deformations under the block the stress-strain response is near linear. That is,
No. 10 was matched to obtain in situ modulus Ej . The
ratio Ej /Ei was 0.15 for GSI = 20 of class II greywake
and found to be very high even for disturbance factor,
D = 0 as per Hoek & Diederichs (2006). Further, the ratio of the failure strain of the intact rock to
that of the jointed rock is given by,
117
Table 3. Suggested values of r for values of σci (Ramamurthy 1994). Figure 3. Influence of weathering on modulus ratio of rocks.
118
where, σ1 and σ3 are major and miner principal stresses,
respectively, σcj is the uniaxial compressive strength of jointed
rock obtained from Eq. (9), αj and Bj are strength parameters
of the jointed rock. The values of αj & Bj are obtained from
Eqs. (18 & 19),
119
120
121
122
123
124
125
W. Schubert
Graz University of Technology, Austria
ABSTRACT: Accidents, cost- and time overruns at tunnel projects are still rather common. It appears that design and con-
struction procedures still do not comply with the requirements of an up to date engineering approach. Too often dominant
behaviors during the design process are ignored or the models simplified in a way that critical mechanisms cannot be detected.
To improve the situation, design and construction processes need to be adopted, which focus on the inherent risks of tunneling.
The paper presents such an approach, which has been used successfully over the last decade.
1 INTRODUCTION
Hazard is commonly understood as an unfavorable event. Risk In the course of this paper we are only addressing geotechnical
then is defined as the product of the consequences of such and environmental hazards, while hazards during operation
an event and the probability of occurrence. In engineering are not dealt with.
it is understood that a risk assessment process has to start Hazards during tunnel construction may be various modes
with the definition of the system followed by hazard iden- of failure, large displacements, excessive water inflows and
tification, evaluation of the probability of occurrence and a other effects, like intolerable emissions by blasting vibrations
consequence analysis. The next steps include the risk evalu- or dust. Like in other risk assessment processes, also in tunnel
ation and the check whether the risk is acceptable or not. In design those hazards and their consequences should be known
case of an unacceptably high risk, mitigation measures have before deciding on hazard or risk mitigation measures in the
to be implemented and the risk re-evaluated. This is done until form of construction method and supports.
an acceptable risk level is obtained (Figure 1).
It is quite clear that without hazard identification and
4 RISK ORIENTED DESIGN APPROACH
evaluation of the consequences, as well as the determi-
nation of probability of an event, all further steps are
4.1 Basic idea
meaningless.
This principle now can be relatively easily used in tunnel The generation of a cavity in the ground always triggers a reac-
design. tion from the ground. The “quantity” and the sort of reaction
127
128
Min. Max.
Joint set 1
(vertical)
Joint roughness (small scale) Js [−] 0.75 2.00
Joint roughness (large scale) Jw [−] 1.00 3.00
Friction angle ϕ [◦ ] 15 30
Normal joint spacing [cm] 20 100
Joint alteration factor Ja [−] 2.00 4.00
Joint set 2
Joint roughness (small scale) Js [−] 0.75 2.00
(horiz.)
Joint roughness (large scale) Jw [−] 1.00 3.00
Friction angle ϕ [◦ ] 25 35
Normal spacing [cm] 20 100
UCS [MPa] 15 50
Intact
rock
Hoek-Brown Constant mi [−] 10 20
Unit weight γ [MN/m3 ] 0.028
Eq. friction angle ϕ [◦ ]
Granite
40 60
Gravel Hom.
Eq. cohesion c [MPa] 0.10 0.25
Unit weight γ [MN/m3 ] 0.028
Friction angle ϕ [◦ ] 30 40
Cohesion c [MPa] 0.00 0.05
Unit weight γ [MN/m3 ] 0.018
129
130
131
132
133
134
All examples shown above are from the daily practice and
not academic. Precondition for successful monitoring and
using the results to the benefit of the project and risk reduc-
Figure 20. Utilization of the lining of an urban railway tunnel with tion is, besides quality monitoring, the awareness of owner
multiple drifts (Moritz et al, 2008). and contractor of the benefits.
6 CONCLUSIONS
135
136
T.R. Stacey
School of Mining Engineering, University of the Witwatersrand, Johannesburg, South Africa
ABSTRACT: Rockbursts are violent events that manifest in the ejection of rock from excavation walls, often at high velocity.
Although attempts have been made to link the seismic source magnitude via peak particle velocities to predicted ejection
velocities, back analyses of ejection velocities from rockburst damage indicate inexplicably large magnifications of the peak
particle velocities obtained from seismic measurements. This indicates a lack of understanding of the mechanisms of rock
mass behaviour in seismic conditions. There is an equal lack of understanding of the mechanisms of action and interaction of
support elements in a rockbursting environment. The design of rock support for these conditions is therefore problematic, since
neither the demand imposed on the support, nor the capacity of support systems under dynamic loading, can be defined. A
conventional engineering design approach is therefore not possible owing to the design indeterminacy. Since testing of support
has shown that both rockbolts and containment support systems such as wire mesh plus wire rope lacing can absorb large
amounts of energy, an alternative philosophy is to specify the support conservatively. A very important requirement is that the
containment support, and the connecting elements between this support and the retainment support, should not fail – they must
ensure that the dynamic loading is passed fully into the rockbolts. The conservative specification approach may be perceived to
be an expensive solution. However, if all costs of consequences are taken into account, containing rockburst damage is almost
certain to create improved safety and increased value for the operation.
1 INTRODUCTION crush; shear rupture; and fault slip. Of these five mechanisms,
the first three are events in which the source and damage loca-
Research into the seismicity in mines, and consequently rock- tions are coincident. The remaining two are events in which
burst research, began in South Africa a century ago (SAIMM, the source location and resulting damage location can be sepa-
2010), and possibly even earlier in other countries. Despite this rated by significant distances, sometimes hundreds of metres.
lengthy period of research, solutions to the problem of rock- The magnitudes of these events can be very large, of earth-
bursts appear to be elusive. Mining is progressing to greater quake scale, whereas the magnitudes of the former group are
depths, in higher stress environments, and rockbursting is per- usually much smaller. When seismic source location and rock-
haps an increasing problem world-wide, rather than a problem burst damage location are not coincident, even if one knew
whose solution is understood and known. Although advances exactly the source location, the prediction of the possible dam-
have been made, rockbursts remain a hazard. In their book age location is an additional problem. The philosophy of the
dealing with safety in the mining industry, Simpson et al design of rock support for this severe rockburst problem is the
(2009) quote from Simpson (1998), “Most of the required focus of this paper.
science for most mining hazards is now well understood (with
one or two notable exceptions, for example, rock bursts in
deep hard rock mines) ….” 2 MANIFESTATIONS OF ROCKBURSTS
As has been done for earthquakes, locations of greater seis-
mic hazard can be identified, but predictions of exactly when Rockbursts in mines and civil engineering excavations are
and where the seismic event will occur, and what its magni- very violent events that commonly result in considerable dam-
tude will be, are not known. One of the recent large magnitude age to excavations, well illustrated by Ortlepp (1997). Rock
earthquakes in Chile in March 2010 was actually predicted in is usually ejected, and when this is the case, the ejected rock
terms of both location and magnitude (Ruegg et al, 2009), but, is observed to be fragmented into relatively small blocks and
although there was an expectation of such an event, the tim- slabs, as illustrated in Figure 1.
ing was unknown. The result was unfortunate loss of life due In rockburst events, rock reinforcement such as rockbolts,
mainly to an associated tsunami, as well as widespread dam- and surface support elements such as wire mesh and shotcrete,
age to structures. The rockburst problem, similarly, remains often fail. In such events, conventional rockbolts and cables
one in which the “critical” seismic event location, magni- commonly exhibit brittle failures. Another common observa-
tude and time are unknown. In mines, much of the significant tion is that when the surface support fails, the ejection of rock
seismic activity is often associated in time with blasting. Per- often leaves the reinforcement elements exposed, protruding
sonnel can therefore be removed from the mine at such times. out of the rock as shown in Figures 1 and 2.
However, “rogue” seismic events do occur at other times and Ejection in a rockburst event is not dictated by the action
these events represent a hazard to life. Ortlepp and Stacey of gravity, as is the case with rock falls, but can be in any
(1994a) suggested a classification of seismic event sources direction, and floor-heave and sidewall ejections are common,
that included the following: strain bursting; buckling; face examples being shown in Figures 3 and 4. The latter figure
137
138
139
140
141
142
143
144
145
J. Zhao
Ecole Polytechnique Fédérale de Lausanne (EPFL), Laboratory for Rock Mechanics (LMR), Lausanne, Switzerland
Y.X. Zhou
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
K.W. Xia
Department of Civil Engineering, University of Toronto, Canada
1 INTRODUCTION mature, e.g., LS-DYNA, ABQUS, FLAC and RFPA are com-
mercial codes capable to model dynamic fracturing. However,
Rock dynamics deals with the responses of rock (materials the continuum based methods are not the most suitable meth-
and masses) under dynamic stress fields, where an increased ods for dealing with complete detachment and large-scale
rate of loading (or impulsive loading) induces a change in the fracture opening problems. Because of the models are con-
mechanical behaviour of the rock materials and rock masses tinuum bases, they are not seen as an ideal tool dealing
(Zhou & Zhao 2011). with complex discontinuity, typically represented in naturally
Dynamic loads include explosion, impact, and seismic, that jointed rock masses and rocks in post-failure fracturing state.
are typically in the form of time histories of particle accelera- Discontinuum based methods treat modelling material as
tion, velocity, or displacement. Distribution of dynamic stress an assembled model of blocks, particles or bars, e.g., the Dis-
field is in the form of propagating stress waves. Wave prop- tinct Element Method (DEM) (Cundall 1971), Discontinuous
agation in rock masses is governed by the wave transmission Deformation Analysis (DDA) (Shi 1988) and Distinct Lattice
and transformation across the rock discontinuities in the rock Spring Model (DLSM) (Zhao 2010). In these methods, the
masses, which forms a major topic of rock dynamics. detachment of rock mass and fracturing process of rock mate-
Response of rock materials and rock masses under dynamic rial is represented by the breakage of inter-block contacts or
stress are at different scales, including rock material fracturing inter-particle bonds. Specifically, discontinuum based meth-
and failure, sliding along the rock joints and rock block move- ods can reproduce realistic rock failure process at the post
ments. Rock material fracturing, for example, is a dynamic failure stage. Available commercial computer codes which
micro-scale process leading to macro-scale deformation and based on DEM are UDEC/3DEC and PFC (ITASC 2010) and
failure. Rock fracturing is a dynamic process often associated DDA (Shi 1988). There also exist some research codes, for
with sudden energy release forming dynamic stress waves. example, DLSM (Zhao 2010).
Rock dynamics has applications in civil, mining, energy In order to optimize the advantages of both continuum and
and environmental engineering encountering dynamic loads discontinuum methods, coupled methods are developed in
and behaviours, e.g., rock excavation and fragmentation by recent years. For example, the Numerical Manifold Method
blasting, tunnelling and slope stability and support under (NMM) (Shi 1991) was developed to integrate DDA and
earthquakes, protection of rock falls, rock burst in deep mines, FEM, the FEM/DEM method (Munjiza 2004) is to couple
fracturing of hot rock in geothermal fields, hazard and risk FEM with DEM and the Particle based Manifold Method
control due to explosion and blast. (PMM) (Zhao 2009) was proposed to combine DLSM and
This keynote addresses advancements in some of the topics NMM. The coupled method is capable of capturing both the
of rock dynamics and applications, specifically on numerical pre-failure and the post-failure behaviour of rock materials.
modelling methods, laboratory testing techniques and tunnel However, its implementation is difficult.There only exist some
stability under explosive loading. research codes, e.g., NMM (Shi 1991), Y2D (Munjiza 2004),
LDEC (Morris et al. 2006), m-DLSM (Zhao 2010), and no
2 MULTISCALE CONTINUUM-DISCONTINUUM commercial codes are readily available yet. Table 1 provides
NUMERICAL MODELLING a summary on these numerical methods and corresponding
computer codes.
Numerical methods used in rock mechanics are classified into The coupling of DEM and FEM emerged in the late 1980s
continuum based, discontinuum based and coupled contin- and various coupled models have been developed. The most
uum/discontinuum methods (Jing 2003). direct coupling methodology is to model the discontinuous
The continuum based methods are based on continuum zone with a DEM model and the continuous zone with a FEM
assumptions. They include the Finite Element Method (FEM), model, e.g. the model used by Pan & Reed (1991) and Chen &
the Finite Difference Method (FDM), and the Smoothed Zhao (1998).Yet, this simple FEM/DEM coupling cannot fully
Particle Hydrodynamics (SPH).The methods utilize the macro take advantage of the strength of each method and is limited
mechanical parameters to model the stress-strain state. Com- on handling complex rock failure problems. Recently devel-
puter codes for continuum based methods are relatively oped coupled FEM/DEM methods aims at modelling failing,
147
148
Figure 3. Multiscale model for the blasting wave propagation 3.3 Dynamic Brazilian disk (BD) method
interact with a tunnel. The BD specimen in the SHPB system is shown schematically
in the insert of Figure 5, where the sample disc is sandwiched
strain rates from 10 to 103 s−1 , the split Hopkinson pressure between the incident bar and the transmitted bar. This method
bar (SHPB) is an ideal choice. SHPB was invented in 1949 by was first used to measure the dynamic tensile strength of con-
Kolsky for testing metallic materials (Kolsky 1949). Shortly cretes (Ross et al. 1989, Ross et al. 1995), and was used for
after that, SHPB was attempted by researchers to test brittle studying marbles (Wang et al. 2006) and argillites (Cai et al.
materials such as concretes, ceramics and rocks. We will dis- 2007). However, it is only recently found that the dynamic
cuss in this section about some recent applications of SHPB tensile strength can be determined using the quasi-static equa-
to rocks on the quantification of rock dynamic tensile strength tion only if the dynamic force balance has been achieve in the
and fracture toughness. dynamic test (Dai & Xia 2010). This method was then used to
generate the data shown in Figure 5.
149
≤102 Hydraulic machines Dynamic load applied by Uniaxial compression (e.g., Green & Perkins 1969,
movement of a piston Zhao et al. 1999a) ; dynamic triaxial compression
hydraulically driven by (e.g., Li et al. 1999)
gas or oil Direct tension (e.g., Yan & Lin 2006, Asprone et al.
2009); dynamic Brazilian indirect tension (e.g.,
Zhao & Li 2000)
Punch shear test (e.g., Zhao et al. 1998)
Shear of rock joints (e.g., Barbero et al. 1996,
Kana et al. 1996)
100 –103 Drop-weight machines Gravitational potential Flexural loading (e.g., Banthia et al. 1989)
energy Impact and fragmentations (e.g., Whittles et al. 2006)
101 –103 Hopkinson bar One dimensional stress Uniaxial compression (e.g., Li et al. 2000, Li et al.
wave propagation theory 2005, Cai et al. 2007, Zhou et al. 2010) ; Triaxial
compression (e.g., Christensen et al. 1972, Li et al.
2008, Frewet et al. 2010)
Direct tension (e.g., Cadoni 2010, Huang et al. 2010a);
Brazilian indirect tension (e.g., Wang et al. 2006,
Cai et al. 2007, Dai & Xia 2010);
Flattened Brazilian disk (FBD) tension (e.g., Wang
et al. 2009);
Semi-circular bend (SCB) test (e.g., Dai et al. 2008);
One-point impact test (e.g., Belenky & Rittel, in press);
Spalling test (e.g., Erzar & Forquin 2010)
>103 Gas gun High-pressure gas driven Equations of state (e.g., Shang et al. 2000)
projectile
150
151
152
153
154
Keywords: mine-by experiment, in situ experiment, claystone, hydro mechanical behavior, Excavated damaged zone
157
Figure 2. GCS drift: (a) Road header, (b) section, (c) view of GCS
80% of the convergence has been reached during the first 100
drift.
days. The convergence rates decrease as a function of time.
Other results in terms of deformation, loading of the support
and in term of pore pressure are presented and discussed.
As a conclusion, a new drift, parallel to the horizontal in
situ major stress, has been excavated at the main level of
the Meuse Haute Marne URL in order to perform a mine-
by experiment. Among the 200 sensors installed previously
to the digging of the GCS drift, more than 95% working
during the shaft sinking, thus providing a important data
base to understand the coupled hydro mechanical behaviour
of the Callovo-Oxfordian claystone. Measured parameters
(mainly displacements and pore pressures) are consistent.
Further analyses of the hydro mechanical behaviour have to be
conducted, accounting the result on the ongoing EDZ charac-
terization in the GCS drift. The long term displacements and
pore pressure evolution will continue to be recorded in order
Figure 3. Concept of the GCS experiment. to study the long term behaviour.
158
Squeezing in tunnelling may provoke various severe problems installation of a shotcrete lining during excavation. These
during construction, and therefore it is of high importance for are installed in stages, following the support installation
tunnel design. The squeezing phenomenon is closely asso- procedure of an actual tunnel case.
ciated with creep, which is essentially a time – dependent Figure 3 displays the normalized radial displacement pro-
behaviour of the rock mass. In this study, the time – depen- files (normalized LDPs) along the tunnel axis for various
dent rock deformation in tunneling is examined numerically, times up to 1 year time period, for an unsupported tunnel. It
using the Burgers viscoelastic rheological model. Burgers’ is observed that both the elastic and the viscoelastic normal-
body is a four – constant viscoelastic model, formed by com- ized deformation profiles are almost identical. Furthermore,
bining in series Kelvin’s and Maxwell’s creep bodies. The a good correspondence of these profiles with an empirical
four viscoelastic parameters are the Maxwell’s and Kelvin’s LDP is observed.
shear moduli G M and G K and viscous coefficients nM and nK
respectively. This idealized model is considered to be a first
approximation of the actual behaviour of rocks, as it is able to
represent both primary and secondary creep.
The problem of the Burgers rock creep around an advanc-
ing axisymmetric tunnel is investigated numerically with a
three dimensional finite differences code. The model prepared
is shown in Figure 1. Appropriate boundary conditions are
applied in order to take into consideration the axisymmetric
properties of the problem (Fig. 2).
Two model series are prepared and examined. In the first
series (A–series model), the excavation is performed incre-
mentally without considering the rock mass creep. Then the
rock mass creep is activated and the model response is exam-
ined for a time period of 1 year. In the second model series
(B–series model), the rock mass creep during excavation is Figure 2. Boundary conditions of the three dimensional axisym-
taken into account. Both supported and unsupported mod- metric models (xy–plane and zx–plane).
els are examined. Shell elements are used to simulate the
159
160
Multi-parameter responses of soft rocks during deformation and fracturing and their
implications in geomechanics and geoengineering
Ö. Aydan
Department of Marine Civil Engineering, Tokai University, Shizuoka, Japan
Y. Ohta
Japan Nuclear Energy Safety Organization, Tokyo, Japan
H. Tano
Department of Civil Engineering, Nihon University, Koriyama, Japan
Keywords: rock failure, lab testing, dynamics, monitoring, rock stress, rock bursts
161
Rock tests constitute one of the bases for theoretical and The rock mass strength were determined averaging three
numerical analysis and design, and are important tools for most popular methods recommended by Sheorey (1997),
studying rock mechanics and engineering. Rock mass strength Truemann (1999) and Laubscher (1990). The following rela-
and modulus have got several applications in rock excavation tion has been obtained between the pre and post-blast Frac-
engineering. Presently it is difficult to determine accurately tal Dimensions, specific charge (q) and average rock mass
the relevant parameters used in computation, since, rock mass strength.
is characterized by its discontinuities in the form of joints, frac-
tures, and other micro-fissures. Many empirical approaches
are presently available for this purpose. The data collection
procedure for this purpose requires considerable field study
and laboratory work. Also it was felt many a time that the end where,
result is person specific. Hence there is a possibility of loos-
ing preciseness and reliability of the conclusions drawn on the
basis of the data.
The term “fractal” is perhaps best described as a general
concept which is relevant in describing the geometry of irreg-
ular objects or patterns. The fractal concept was originally Similarly, the average Deformation modulus (Ed ) was esti-
introduced by Mandelbrot (1967). A simple definition of a mated from three popular relations given by Serafim and
fractal is a shape made of parts similar to the whole in some Pereira (1983), Hoek & Brown (1997) and Mitri (1994). The
way. The fractal concept has been applied to many aspects of following relation is obtained between the pre and post-blast
nature. Fractal concept has been applied to rock mechanics Fractal Dimensions, specific charge (q) and average Ed .
since its initiation in the early 1980s. A number of researchers
have been devoted to the fractal characterization of fragmen-
tation. The fragmentation method can be well understood by
in-situ and blasted block size distribution using digital image where,
analysis technique. The Fractal Dimensions can be well related
with the distribution.
In the present study, fractal approach to estimate various
rock and rock mass properties from the Fractal Dimensions
of the in-situ rock mass and blasted rock fragments will be
presented.
The study includes the field investigations to collect the rock
mass and rock samples, blasting information and photographs
of the in-situ and blasted fragments, evaluation of the rock and The above relation is in line with that proposed by Mitri
rock mass properties, estimation of Fractal Dimensions of the (1994).
in-situ blocks (pre-blast) and blasted fragments (post-blast) In the present study all the observation were made in the
and relating those with the rock and rock mass properties. A field using the available resources and without much varia-
total of 22 sites were covered in the field investigations. The tion in the on-going blast design parameters. Studies in the
various formations covered are sandstone, limestone, granite, models may be explored with single hole blasting varying one
manganese, basalt, sillimanite, flourite, iron ore (hematite), or more design parameters for further corroboration of the
laterite and bauxite. The pre and post-blast data were collected derived conclusions.
in the referred sites.
162
R. Hug
Ernst Basler + Partner AG, Zurich; formerly ETH Zurich, Switzerland
1 INTRODUCTION
2 OBSERVATIONS
Figure 1. Axial strain εy as a function of the distance d between the
The ground may respond more or less quickly to tunnel exca- face and the measuring point for advance rates of 3.2 and 0.8 cm/h
vation, depending on its rheological properties. Slow ground (after Myer et al. 1981).
response may reduce the extrusion of the core significantly,
thus making it difficult to predict squeezing intensity. Myer
et al. (1981) investigated the effect of the advance rate on the is modelled as linearly elastic (according to Hooke’s law)
axial strain ahead of the face through the use of physical mod- and perfectly plastic (according to the Mohr-Coulomb yield
els. Figure 1 shows the measured effect of the advance rate on criterion, with a non-associated flow rule).
extrusion. It shows clearly that the faster the advance rate, the As a reference point, the case of an unsupported tunnel
smaller will be the extrusion of one and the same material. excavated through ground with time-independent behaviour
will be discussed first. Figure 2 shows the tangential strain
εt,c as a function of the axial strain at the centre of the face
3 THEORETICAL ASPECTS εy (0) for different values of the normalized uniaxial compres-
sive strength fc /p0 and the friction angle φ. The conditions
The present section analyses the response of the core ahead of that lead to high axial strains at the face lead also to larger
the face numerically in order to gain a better understanding of convergences of the tunnel. As the relationship is unique, pre-
the observed behaviour, and to investigate whether there is a diction is theoretically possible. The dashed lines in Figure 2
correlation between extrusions and convergences. Emphasis show that most values are in the range of εt,c /εy (0) = 1 to 2.
will be placed on the effect of the time-dependency of ground Squeezing ground often exhibits a pronouncedly time-
behaviour. dependent response to tunnelling. The deformations in a cavity
The numerical analyses will focus on an axisymmetric may continue for several weeks or even months after exca-
model of a deep, cylindrical tunnel crossing a homogeneous vation. As the time scales for core extrusion (a short-term
and isotropic ground which is subject to uniform and hydro- phenomenon) are different to those for convergence (a long-
static initial stress. The mechanical behaviour of the ground term phenomenon), it is interesting to investigate the extent to
163
4 CONCLUSIONS
164
Yanting Chang
Geton Teknik, Stockholm, Sweden
Keywords: Rock failure, rock bursts, stability analysis, monitoring, filed measurements
165
Figure 4. Different domains of failure behavior defined in the Figure 6. Illustration of stress and strain cycles associated to
ε1 − εv space. seismological events.
166
1 INTRODUCTION
2 STARTING POSITIONS
167
6 CONCLUSIONS
ACKNOWLEDGEMENT
are presented. It is shown that conditions of a continuity
of pressure and deformations on these borders should be The study is financially supported by the Russian Founda-
satisfied. tion for Basic Research, Grant No. 09-05-00327a, and the
Integration Projects Nos. 61, 69 and 74 of Siberian Branch,
Russian Academy of Sciences.
168
Keywords: Structural geology, numerical modelling, fractures, scale, rock engineering design
1 INTRODUCTION
169
REFERENCES
Feng, X.T. & Hudson, J.A. 2011. Rock engineering design. London:
Figure 3. Flowchart of rock mechanics modelling and rock engi- CRC Press, Taylor & Francis.
neering design approaches (Feng & Hudson, 2011). Hudson, J.A. & Cosgrove, J.W. 1997. Integrated structural geology
and engineering rock mechanics approach to site characterisation.
Int. J. Rock Mech. Min. Sci., 34, 3–4, paper 136.
Hudson, J.A., Cosgrove, J.W. & Johansson, E. 2008. Estimating
the mechanical properties of the brittle deformation zones at
Olkiluoto. Posiva Work Report 2008-67. Downloadable from
www.posiva.fi
Jing, L. 2003. A review of techniques, advances and outstand-
ing issues in numerical modelling for rock mechanics and rock
engineering Int. J. Rock Mech. Min. Sci. 40, 3, pp. 283–353.
Latham, J.-P., Guo, L., Wang, X. & Xiang, J. 2011. Modelling the
evolution of fractures using a combined FEM-DEM numerical
model. Proc. ISRM Congress, Beijing. London:Taylor and Francis.
Price, N.J. & Cosgrove, J.W. 1990. Analysis of Geological Structures.
Cambridge University Press, UK, 502p.
Turner, S.A, Liu, J.G. & Cosgrove, J. W., 2011. Structural evolution
of the Piqiang Fault Zone, NW Tarim Basin, China, Journal of
Asian Earth Sciences 40, 1, pp. 394–402.
170
A.V. Dyskin
School of Civil and Resource Engineering, University of Western Australia, Australia
E. Pasternak
School of Mechanical and Chemical Engineering, University of Western Australia, Australia
171
172
T. Fukuda
Geoscience Research Laboratory, Co., Ltd., Tokyo, Japan
T. Takahashi
East Nippon Expressway Company, Hokkaido, Japan
H. Yamada
Konoike Construction Co., Ltd., Osaka, Japan
Y. Jiang
Nagasaki University, Nagasaki, Japan
Keywords: back analysis, tunnelling, rock properties, rock mass, rock support
1 INTRODUCTION
173
X.R. Ge
Institute of Rock and Soil Mechanics, Chinese Academy of Sciences, Wuhan, China
State Key Laboratory of Geomechanics and Geotechnical Engineering, Wuhan, China
Institute of Geotechnical Engineering, Shanghai Jiao Tong University, Shanghai, China
174
Figure 5. The stress state at point A on potential slip surface. where R is the projection of resultant anti-sliding force on
potential sliding direction d̂; T is the projection of resultant
force is a vector sum of anti-sliding normal force and anti- sliding force on potential sliding direction d.
sliding shear force acting on the Si of the potential sliding The VSM is used to analyze the anti-sliding stability of a
surface by bed rock. dam foundation, which is one of the typical dam foundations of
Shown as Figure 2, let σ s , σ τ , σ n stand for stress vector, a hydraulic power station in China. The main natural faults and
shear stress and normal stress, respectively at the point A on the potential sliding interfaces of the representative section are
the sliding interface. n is the unit normal vector of tangent shown in Figure 3. These sliding paths are ABCD , ABCD,
plane at point A (positive pointing to outside of the sliding EFD and EFGI, and the corresponding results by 2D vector
mass). d is the unit vector of the potential sliding direction. sum method are listed in figure 3. For this example 3D VSM
S is the slip surface. Then, result is also given.
6 CONCLUSIONS
The safety factor of VSM is derived from the real stress state
rather than the virtual stress state. The safety factor is defined
as the ratio of the projection of resultant anti-sliding and slid-
where σ is the stress tensor at the point A on the potential slip ing force vector on the potential sliding direction, which has
surface. clear physical meaning.
175
Keywords: Site characterization; Rock mass classification; Numerical modeling; Tunnelling; Rock support
Mumbai Rail Vikas Corporation Limited has planned to con- basaltic lava flows. Basalts are predominantly the amyg-
struct 5th and 6th railway lines from Thane to Diva. On the daloidal basalts. In the area near Tunnel 1, the rocks are
route, the rail line passes through various geological features grayish to brownish fine grained amygdaloidal basalt rock.
like creeks, lineament, hills and flood creek plains. Because The empirical and the numerical approach using FLAC3D were
of the varied topography, various structures like bridges, foot used for the analysis and design of the tunnel.
over bridges and tunnels are to be constructed. As per the In the empirical approach the Q-system has been used to
Geotechnical investigation report, two tunnels are planned on classify the rock mass for the design purposes. The Q value of
the route. amygdaloidal basalt is about 25 indicating a ‘Good Rock’ for
The proposed new tunnel having length 165 m, width 12 m tunnelling. Different Q values were obtained for portal areas
and height 8.5 m is to be driven parallel to the existing tunnel. and intersection of trolley refuse. With factor of safety of 1.5,
The tunnel is to accommodate double track. There is also a the Q-value is back calculated and accordingly the supports
provision of trolley refuse in the tunnel. The existing Tunnel have been designed. Using the Q values, the shotcrete and
1 is around 110 m long having a gentle curve. The tunnel is rock bolts support is suggested in the tunnel. The slope above
aligned to N120◦ at Thane end and N290◦ at Diva end. The portal is suggested to be protected by erecting the fence or
tunnel is around 100 years old. The roof of the tunnel upto protection wall.
springing level throughout its length is brick masonry lined. At The numerical analysis using the FLAC3D code was also
the portal upto about 20 m inside the tunnel the walls are also carried out to study the effect of adjacent tunnel on the induced
brick lined. At places inside the tunnel also the walls are brick stresses. Accordingly it was found out that a minimum 16 m
lined. On the uphill side walls in the tunnel, small seepage pillar width between the two tunnel (existing and the proposed)
of water is also observed. The existing tunnel is horse-shoe is required. The results of analysis also show that the induced
shaped. stresses are within reasonable limits.
Geologically the region falls in Deccan volcanic province Details about the tunnel design and numerical analysis are
that is made up of a number of practically horizontal sub-aerial given in the full length paper.
176
A large number of casing collapses are reported in one of the For this reason, the shallow cores had been heated by oven
largest carbonate oil fields in south of Iran. Most of the casing and kept in constant temperature for at least 24 hrs. Then the
collapses have been occurred in Gachsaran overburden forma- laboratory tests were carried out promptly. From these data,
tion. Compaction of reservoir is one of the prime suspects of relations between E and UCS with temperature were generated
this phenomenon. It is very difficult to take core samples bel- both for marl and anhydrite which are summarized in Table 1.
low 200 m depth in Gachsaran formation due to high solubility Additional triaxial compressive tests were carried out on
and weakness of the rocks. Moreover there are not adequate the anhydrite and marl cores at various temperatures to obtain
logs in this formation. So obtaining the geomechanical param- Coulomb failure criterion. The coulomb failure criteria in
eters of rocks in depth associated with casing collapses (about anhydrite and marl at various temperatures were determined
3000 m) is virtually impossible. As a result, the main reason as equation 1 and 2 respectively:
of casing collapse in Gachsaran formation has been not clear
yet; however, reservoir compaction has been named as eminent
cause by some researchers.
To quantify the compaction, the lithology across the fail-
ure depth (3200 m for well M-42, as the first failed well) was
reviewed and an interface between anhydrite (as a stiff rock)
and marl (as a soft rock) was detected. One dimensional reser-
voir compaction was calculated to be as high as 0.15–0.3 m in
50 years due to 9 MPa reservoir depletion.
Anhydrite and marl cores were taken from depths of 12 m
to 195 m of Gachsaran formation at the Gotwand-E-Olya dam
Moreover, the relation between failure angle (β) and tempera-
site (Fig. 1). However the thermal condition at shallow level
ture (T◦ C) in Gachsaran anhydrite and marl were obtained as
is quite different from high depths (collapse depth), which
expression 3 and 4 respectively: The triaxial compressive test
requires testing at elevated temperatures on the samples.
results were consistent with those from uniaxial compressive
test in anhydrite and marl.
177
178
Z.Y. Yang
Department of Civil Engineering, Tamkang University, Taipei
T.H. Huang
Department of Civil Engineering, National Taiwan University, Taipei
Y.X. Zhao
Department of Civil Engineering, China University of Mining & Technology, Beijing, China
M.C. Tsai
Department of Civil Engineering, National Taiwan University, Taipei
1 INTRODUCTION
179
REFERENCES
Chang, S.H. & Lee, C.I. 2004. Estimation of cracking and dam-
Figure 3. Development of AE hypocenters before peak strength age mechanisms in rock under triaxial compression by moment
during uniaxial compression. tensor analysis acoustic emission, International Journal of Rock
Mechanics and Mining Science 41: 1069–1086.
Coughlin, J. & Kranz, R. 1991. New approaches to studying
rock burst-associated seismicity in mines. In Proceedings of the
32nd U.S. Symposium on Rock Mechanics as a Multidisciplinary
Science, Oklahoma, July 10–12, 1991. Rotterdam: Balkeman.
Turcotte, D.L. 1997. Fractals and chaos in geology and geophysics,
Cambridge University Press: 106–108.
180
N. Iwata
Chuden Engineering Consultants Co. Ltd., Hiroshima, Japan
T. Sasaki
Rock Engineering Institute, Suncoh Consultants Co. Ltd., Tokyo, Japan
R. Yoshinaka
Saitama University, Saitama, Japan
Keywords: numerical modelling, rock joints, case studies, rock slopes and foundations, dynamics
The authors are also applied MYM to large scale slope col-
lapse due to earthquake and compared the collapse mode.
The analytical results indicate that the collapse was occurred
beginning at the early time, and shear failures extended to
the middle of slope along bedding joints and tension failures
were occurred along orthogonal joints, it caused toppling at Figure 2. Comparison of response spectrum of horizontal acceler-
the surface. These results agree with eyewitness reports. ation at 60 m in depth.
181
P. Jia
School of Resources and Civil Engineering, Northeastern University, Shenyang, China
With the increasing of the axial load, cracks are first initiated
around the periphery of the opening and propagated gradually
until the first order of fracture ring is formed (Fig. 5b). Then
the second and third order are generated sequentially. These
fracture rings are not closed rings, but discontinuous fracture
rings, i.e., in the fracture zones, there includes intact zones;
while in intact zones, there includes fracture zones too.
The fracture rings are not continuous both in radial and
axial direction, which is different from our former cognition
that fracture rings are continuous rings. This discontinuity is Figure 6. Configuration of Zonal disintegration in different cross
sections.
mainly due to the heterogeneity of rock mass.
Numerical results indicate that the elastic modulus as well
as compressive strength have a very important impact on ZD
phenomenon. In layered materials, the fracture ring is mainly
initiated and propagated from weak layers. There are only few
integrated fracture rings in the model and the number of the
fracture rings is very limited compared with that in the intact
model. As for the rock mass with lower strength and higher
deformation ability, the radial strain is relatively high and the
radial tensile stress thus decreased. Due to the intensive het-
erogeneity of layered materials, the distance between fracture
rings is also longer than that in intact model. Figure 7. Longitudinal section of numerical specimen after loaded.
182
183
Seok-Cheon Lee
Department of Tunnel, Taeso Engineering Co.
Keywords: numerical modeling, rock stress, rock support, stability analysis, subsea tunnel, secondary lining
184
185
Won-Beom Kim
Department of Energy and Resources Engineering, Chonnam National University, Gwangju, Korea
Taewook Ha
Hakyong Engineering Corporation, Seoul, Korea
Hyung-sik Yang
Department of Energy and Resources Engineering, Chonnam National University, Gwangju, Korea
As the tunnel construction in the urban areas gets vital- In this study, considering the excavation damaged rock
ized, the planning and execution of construction reflecting zone, the behaviors of the rock pillars have been assessed
a variety of factors such as linear conditions, restrictions on by the 3-dimensional numerical analysis. In order to assess
land procurement, the protection of some obstacles, etc. are the damaging area of rock pillars based on excavation meth-
required. For example, the distance between two horizontal ods, the study adopted a seismic exploration method to the
tunnels (pillar width) increases the construction example of a blasting section as well as the mechanical excavation section
very near parallel tunnel. The excavation damaged rock zone with the breaker. After the size and degree of the damage
which is inevitably occurred when excavating the underground were re-evaluated as the decrease in the strength parame-
structures is the major cause of the stability loss during the ters of rock pillars was applied to the results of numerical
construction of the rock structures. In case that the parallel analysis, it appeared that the decrease did not influence the
tunnels of which distance between the two horizontal tunnels safety factor of mechanical excavation while it was certainly
is very near, this affects the stability of the tunnels depend- causing a decline in the safety factor a general blast. Added
ing on the size of the excavation damaged rock zone occurred to this, the damaging area of the rock resulting from an
to the rock pillars if the rock pillars themselves between the excavation method proved to be influential in improving the
tunnels act as the supporting points. range of the minimized distance between the two horizontal
tunnels.
In particular, the result showed that the safety factor reduced
to maximum 28% in case that the excavation damaged rock
zone depending on the excavation methods was considered
than the result of numerical analysis without considering the
excavation damaged rock zone.
In this study, as the size of the excavation damaged rock
zone occurred by blasting appeared to be bigger than that by
the mechanical excavation, the minimized distance between
the two horizontal tunnels to which the mechanical excavation
and blasting could be applied in constructing the very near
parallel tunnels was assessed by the numerical analysis.
Figure 1. Schematic diagram of SHT wall transmission method.
Safety factor
Pillar width Ground depth EDZ
Excavation method RMR (m) (m) (m) Without EDZ Considered EDZ
186
Response of granite under strain controlled loading and effect of support system on
behaviour of large underground cavern
Rakesh Kumar
GMR Consulting Services Pvt. Ltd., Delhi, India
The hard rocks like granite, basalt, quartzite etc. show the The rate of softening is less under low confining pressures and
sudden failure due to loading and unloading. The phenom- high under high confining pressures.
ena of rock bursting during underground excavation for such The paper also presents the application of softening behav-
hard rocks under high rock cover is quite common. The rocks ior of granite to the analysis of large underground caverns
in general show strain softening behavior under loading and (Power house and transformer yard) under plane strain con-
unloading. The strain softening behavior of rocks is captured ditions using FLAC. The power house cavern is 22 m wide
using strain controlled tests in the laboratory. The rate of and 44 m high and transformer hall cavern is 16 m wide and
strain softening (residual strain/peak strain) is generally more 22 m high. The analyses of large underground caverns have
in softer rocks than that of harder rocks. been carried out using conventional Mohr-Coulomb model
The paper presents the post peak response of granite and Mohr-Coulomb model with strain softening effect using
under strain controlled loading for confining pressures of FLAC. The support system has been installed in various stages
0 to 30 MPa in the laboratory. The strength and deformation in the form of shotcrete and rock bolts and effect of softening
parameters are determined from laboratory test results. The on cavern behavior under each stage is studied. The effect of
peak and residual strength parameters are determined using support system on the softening behavior is also studied in
Mohr Coulomb criteria. The variation of cohesion and friction detail.
angle with plastic strain is also determined. The rock shows It is seen from the analyses that there is significant effect
the vertical splitting under confining pressures of 0 and of the softening on caverns behavior and support systems also
10 MPa and shear failure under other high confining pressures. play a major role in softening behavior.
187
Keywords: stability analysis, rock stress, mine design, mining, numerical modelling, multi-reef mining
1 INTRODUCTION
1.1 Background
Primary Ground Control Districts have been established for
the UG2 based on the middling between reefs (UG2 and
Merensky), and rock mass conditions or quality. The mid-
dling between reefs varies between ∼14 m (UG2 underneath
Pothole Merensky) and ∼38 m (UG2 underneath Normal
Merensky). In between the Pothole Merensky and the Normal
Merensky, the transition zone is present where the middling
is between 14 m and 38 m.
The higher stress concentration on UG2 faces when min-
ing underneath Merensky pillars normally results in Poor
(RMR1976 ) ground conditions that have in the past led to large
scale falls of ground in the UG2 stopes. In many instances,
this resulted in the abandonment of the mining panels and
consequently reduced the UG2 extraction ratio between 10 Figure 1. Schematic showing stress driven failure mechanism when
Level (∼500 m below surface) and 20 Level (∼1000 m below mining UGII underneath a Merensky pillar.
surface). Experience has indicated that ground conditions
become worse in panels undermining Merensky pillars from
16 Level (∼800 m below surface). Experience gained with mining of the UG2 under Merensky
reef pillars over the last 10 years at Union Mine, indicates
that this mining to be most difficult for pillar widths varying
1.2 Postulated failure mechanism between 15 m and 65 m. The influence of different Merensky
pillar widths on the stability of the UG2 stope panels does
The large falls of ground that occurred may be referred to as not form part of this analysis. Merensky pillar widths of 35 m
back-breaks. A number of theories have been put forward by (Unpublished consulting report, 2002) and 50 m have been
numerous investigators as to the main cause that lead to the modelled and these results are presented below.
occurrence of back-breaks on the UGII mining underneath the
Merensky at low middlings (14–38 m). The authors agree that
at Union Mine, although it may be a combination of rock mass 2 REVIEW OF PREVIOUS WORK
stress and structure, the failure is primarily driven through
the changes in the stress field. Figure 1 shows a schematic 2.1 Review of numerical modelling
of the postulated failure mechanism as a result of stress.
Previous numerical modelling results have been reviewed and
are presented below to highlight the influence of backfill as
1.3 Objectives well as the stress levels on a UGII stope below a 35 m wide
Merensky pillar. A UDEC model was setup to simulate a 35 m
The objective of this analysis is to:
wide Merensky pillar under which the UGII was extracted
(i) identify the type of failure of the UG2 stope hanging wall in 20 m face advance increments (approximately 1–2 months
through numerical modelling, mining). The stope in the UDEC model was at a depth of
(ii) quantify the level of potential failure of the UG2 stope 1200 m (approximately 23 Level) below surface.
hanging wall through numerical modelling, Ultra-deep mining conditions may be experienced in the
(iii) determine whether the current in-panel support strategy UGII panel when the UGII stope face is mining close to the
would be able to cater for the predicted failure, centre of a 35 m Merensky pillar at a middling of 16 m at a
(iv) determine if and where backfill support would be suit- depth of approximately 1200 m below surface.
able for the conditions experienced through numerical The influence of leaving a 5 m wide pillar underneath the
modelling. Merensky pillar with only elongates as stope support was also
188
189
1 INTRODUCTION
190
D.Y. Li
School of Resources and Safety Engineering, Central South University, Changsha, China
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
L.N.Y. Wong
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
Buckling is a type of failure that has been observed around shown in Figure 2. It was evident that the smaller the slen-
underground openings in highly stressed as well as in jointed derness ratio, the larger the required stress for buckling. The
rock masses. Slabs can be formed due to the presence of effects of the slenderness ratio and eccentricity on the q/σθ
joint planes parallel to the excavation surfaces in the case of ratio are demonstrated in the parametric analysis presented in
highly jointed rock masses, or under a high compressive stress Figure 3. It could be seen that for a slenderness ratio of 10, a
that acts parallel to the excavation boundaries in intact rock confining pressure of no more than 5% of the axial stress was
masses. Buckling stability analysis is an important concept adequate to stop any deflection on the column. The effects of
for slim column under compression in material mechanics. the eccentricity of maximum tangential stress and the radial
The buckling failure of highly-stressed hard rocks surround- stress gradient were considered in the analysis. The maximum
ing deep underground openings is analyzed in this paper. The lateral displacement and the maximum moment on the rock
stress redistribution of surrounding rock masses around the slabs were calculated by using vector superposition method. It
underground opening can be shown in Figure 1. The tangen- showed that the slenderness, the eccentricity, the elastic mod-
tial stress (σ θ ) will be increasing and the radial stress (σ r ) ulus of the slabs and the loading stress played important roles
will be decreasing for the surrounding rock masses near the on the buckling failure of rock slabs. The confining pressure
excavation surfaces. Under the approximately uniaxial com- provided by filling material could prove adequate to control
pression, the rock masses may split and form a number of buckling failure under certain loading conditions. The way to
surface-parallel slabs. obtain the basic parameters in buckling analysis was also dis-
The buckling analysis was carried out by considering the cussed in the paper. These parameters can be basically divided
axial loading condition, the eccentric loadings and the stress into three categories: mechanical parameters (P, σθ and σr ),
gradient in the rock slabs. It indicated that the buckling stress material parameters (E and q) and geometric parameters
is a function of the modulus of elasticity and the slenderness (l, l0 , e, t and T ).
ratio. The effect of these parameters on the buckling stress is
191
Table 1. In-situ stress measurement results at the testing locations in Maluping Mine.
σ1 σ2 σ3
Testing Stress Dip Dip Stress Dip Dip Stress Dip Dip
location /MPa direction/◦ angle/◦ /MPa direction/◦ angle/◦ /MPa direction/◦ angle/◦
M700-1# 34.49 207.16 16.95 8.53 111.30 18.50 0.61 336.81 64.47
M750-2# 8.45 143.47 17.05 7.46 28.48 54.03 −2.53 243.92 30.60
M750-3# 8.94 123.54 3.79 4.11 31.58 27.32 −1.80 220.80 62.38
M600-4# 17.02 44.08 28.46 2.83 313.67 0.64 −2.64 222.54 61.53
According to the buckling theory, a case study was pro- into the filling materials to partially replace the cement.
vided at the Maluping Mine in the Guizhou Kailin Group, Several groups of compressive strength of the phosphorus
China. Buckling and slabbing failure has occurred in the gypsum tests with different ratios of consolidated filling were
level +700 m at the Maluping Mine. Figure 4 shows the obtained. It was found that the confining pressure with 1.5–
buckling failure of the surrounding rock masses even after 2.1 MPa provided by filling material is sufficient to stop
shortcreting at the Level +700 m in the subsidiary company of buckling and slabbing failure in the surrounding rock masses.
Maluping Mine. By measuring the in-situ stresses, it was It can realize the stability of the highly-stressed underground
found that the maximum principal stress is as high as 35 MPa stopes.
at the Level +700 m, shown in Table 1. Cutting-and-filling
method is widely used nowadays in underground mines. How-
ever, the reinforcement and support mechanism of the filling ACKNOWLEDGEMENTS
material has not been fully understood by the researchers yet.
The back filling material is mainly composed by the phos- The paper was financially supported by the National
phorus gypsum, 425# ordinary Portland cement and fly ash. Basic Research Program of China (973 Program No.
To reduce the cost and to improve the flow properties of 2010CB732004) and the Natural Science Foundation of China
the paste, an appropriate combination of fly ash was added (grant No. 50934006 and 51074177).
192
193
Initiation and evolution of failure zones in the vicinity of heterogeneities in rock masses:
Seismology and deep oil well stability problems
1 INTRODUCTION
Earthquakes in the Earth crust mostly take place nearby faults The surface GPS-monitoring allows quantitative assess-
so we associate the hypocenral area of an incipient seismic ment of increments in horizontal strain tensor components
event with an section S of a fault F that differs in properties. Dkm (xn , yn ) (k = 1, 2; n = 1, …, N ) on a discrete point set. Is it
Even a slight variation in natural tectonic stresses due to, for possible to find parameters of an anomalous area with such
instance, movement of lithospheric plates change geomechan- data at hand? Evidently, it is practically impossible to deter-
ical fields, and it is possible to record this disturbance on the mine shape, spatial position and properties of S based on the
Earth surface with the help of space geodesy methods. The “passive” information Dkm , as peculiarities of the additional
estimates of displacement variations in the Altai earthquake strain field smoothed over already at the distance of few diam-
epicenter (27.09.2003, M = 7.5) showed that the strain range eters of S. For this reason, we use the equivalent point source
accuracy (10−8 ) of contemporary GPS techniques is sufficient concept, which is applied to inversion of seismic data, and
to identify the preparation of the moderate and strong crustal formulate another boundary-value problem: find a source that
earthquakes. generates such a strain field in the weightless domain P that
We analyzed a typical situation (Fig. 1): the Earth’s crust is close to Dkm on the daylight surface. Introduce an objective
part occupied by a parallelepiped P (x, y, z-coordinates; z = 0 – function:
the daylight surface) contains an inclined discontinuity F.
Geomedium deformation in P\F is described by the incre-
mental elastoplastic model with the Mohr-Coulomb condition
(Nikolaevsky 1996), whereas state equations for F were set in
pursuance of (Barton 1986). where Ekm = Gi ekmi ; ekmi is the strain field generated in P by
As the anomalous area S (Fig. 1) we take a rectangle with a unit force applied at the point (px , py , pz ) and oriented in the
the center (sx , sy , sz ) so that stiffnesses in this rectangle are line of i(i = x, y, z). The required parameters of the equivalent
less than corresponding one of the remaining part of F. The source are the arguments such that is minimal: components
deformation and strength properties were typical of the rocks of the force Gi and the location pi .
that compose the modeling object, i.e. the hypocenter area of Minimization of is implemented with the original com-
the Altai earthquake. bination of methods of steepest descend and least squares.
194
Deep hydrocarbon exploration and extraction boring usually Nikolaevsky, V.N. 1996. Geomechanics and Fluid Dynamics,
faces such negative issues as deviation of a well bore from Moscow: Nedra (in Russian).
a pre-set trajectory, failure and permeability variations in the Barton, N.R. 1986. Deformation phenomena in jointed rock.
well bore zone. Based on the same constitutive relations, we Geotechnique, 36(2): 147–167.
195
1 INTRODUCTION
196
5 CONCLUSION
197
1 INTRODUCTION The rock mass strength differs from that of the intact rock
because of the discontinuities such as joints, fissures, faults,
Castrovirreyna Compania Minera SA operates the San Genaro etc. present in the former, which is why the failure crite-
mine in Central Peru, is planning the expansion of its exis- ria for the rock mass were defined using the classification
tent tailings deposit, which will require the construction of index proposed by Hoek & Brown (1988), updated by Hoek,
an underground mine plug to prevent the flow of water and Carranza-Torres and Corkum (2002).
tailings into the mine. The plug should be in operation dur- This approach takes into account the strength of the intact
ing the mine life and after mine closure. The aim of the study rock and the constants mb, s and a, which were estimated
carried out was to assess the geomechanical conditions of the Based on the structure and condition of rock mass discontinu-
area where the plug is to be located and design it to ensure its ities, being finally reflected by the geological strength index
stability in the long term. GSI. By using the generalized form of the failure criterion pro-
posed by Hoek and Brown, the value of the elastic modulus for
2 GEOMECHANICAL MODELS the rock mass (Em) at the site proposed for the construction
of the plug was estimated at Em = 5.296 MPa, a value which
San Genaro ramp was excavated in the Caudalosa volcanic corresponds to a competent andesite.
formation of the Miocene, which is the one with the largest During the field study, some areas showing water flows and
extension in the area. This formation consists of volcanic filtrations were identified, these being the result of seepage of
rocks composed of lava flows and breccias of dark gray and rain water through the fractured rock mass located above the
green andesitic flows, with inter-bedded lenticular pyroclastic ramp. It was anticipated that along major fault zones, due to
and tuffaceous sandstones. Some additional sub-units such as the increase in permeability of the rock mass, the flow of water
porphyritic andesites, andesitic tuffs, basalt and pyroclastic and transport of tailings material would preferably follow the
tuffs have also been identified in the surrounding areas. direction of the structures within the rock mass and would
The main structures found in the Castrovirreyna mine have eventually fill the fractures, which may lead to a decrease the
a general EW trend, with another system showing a NW-SE permeability of the rock mass.
and to a lesser extent a NE-SW trend. The structures in the During the field research stage, a geomechanical cell and
study area of San Genaro ramp have the following characteris- detail lines mapping was carried out from the mine entrance
tics: major fault zones and some shear zones (between 0 + 000 (0 + 000) to chainage 0 + 180.
and 0 + 088 m), generally sub-vertical and following a general The study area was divided into four structural domains
trend EW, having been affected by hydrothermal alteration, the showing similar behaviour, lithological, structural and alter-
shear zones being up to 20 m thick. The minor faults located ation characteristics.
between 0 + 088 and 0 + 180 m have thicknesses varying Table 1 summarizes the main geomechanical character-
from 50 to 100 mm, being filled with gypsum, calcite and istics of the sections that were assessed during the field
fractured rock with some quartz crystals. study.
198
199
Y.J. Shang
Key Lab of Engineering Geomechanics, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
Y.Y. Shi
Sichuan Ertan International Engineering Consulting Co. Ltd, Chengdu, China
1 INTRODUCTION Statistic results of over 200 nos dykes shown that the main
strike of basic veins is nearly SN, dip westward, with dip angle
In China the magmatic activity is frequently occurred espe- of 30◦ (Figure 2).
cially in the Yanshan Movement, and various igneous rocks
are often appeared in most of China (Sun & Peng, 1985).
China sits at west rim of Pacific Plate, the granite is widely 3 DIFFERENTIAL WEATHERING OF SHALLOW
spread and various magma activities were so intensive that ALTERNATIVE ROCKS
they form different kinds of lower hydrothermal metal deposit.
In a long time, the mining experts paid more attentions on In shallow depth or even in ground surface, the alternative
the ore deposit formation with alteration and alterative rocks. rocks were further weathered to gray green, red brown clays,
Besides, the alteration makes smectite to form from many which are very similar to red clays in open fissures or joints.
intermediate to basic igneous rocks. As a result, more types Figure 3 is red clays weathered from basic dykes and filled
of expansive rocks are appeared with wide distribution and in gently dipping joints. In shallow depth or the surface, the
distinctive features, which incurred many serious engineering original alternative rocks dominated by smectite are weath-
geological problems in engineering construction and mining ered to illite/smectite mix layer and kaolinite clay minerals.
(ISRM, 1994). They constitute the weak rocks to form weak intercalated lay-
ers, which control the natural slope stability. For instance, in
field survey, it was observed that the alternative rock as weak
2 DISTRIBUTION OF INTERMEDIATE-BASIC interclated layer, directly controlled the stability of rock mass
DYKES AND ALTERNATIVE MINERALS above it, or even covering as smooth slope surface. In natural
IN JINZHOU GRANITE status, the slope angle is less than 40◦ , and height is less than
30 m.
During field investigation and drilling works at the Jinzhou
granites at Tianqiao, it was observed that lots of acid and
intermediate-basic dykes and sills are intruded into granite 4 ALTERNATIVE ROCK STABILITY IN WATER AND
(Liao, 1989). From survey and thin section identification of 25 ENGINEERING MEASUREMENTS
nos dyke samples, it was found that the dykes include diabase,
aplite, granophyre, diorite porphyrite, andesite pophyrite, and The rock engineering practice shows that for swelling rocks,
lamproite, etc. The distribution of dykes and sills is not well especially those under groundwater table, the swelling is
uniform, which is several centimeters to more than 10 meters
thick. Among them, the dark green basic dykes (diabase) are
well emerged, in the form of sills in a steady thickness and an
altitudes appearing in granites (Figure 1).
Figure 1. Diabase vein (sill) intruding into completely decomposed Figure 2. Rose map and equal density chart of intermediate-basic
granites in shallow depth and keeping well. veins.
200
difficult to occur without previous drying or the water lost 4) Strengthen discharging groundwater measures to prevent
no more than boundary values. Thus drying and water loss groundwater from accumulated in caverns.
are necessary conditions for alternative rocks to expand with
large deformation. Understanding of this phenomenon is sig-
5 CONCLUSIONS
nificant to adjust shotcrete time and closing measures in due
course (Qu, 1984).
There exist various kinds of dykes in Jinzhou granite of NE
When encounter with smectitic swelling rocks, the follow-
China, especially smectitic intermediate-basic dykes, which
ing methods are useful for prevention of geohazards in rock
constitute one challenge for construction of the rock caverns
caverns.
there. The altitude of basic veins is steady and behaves cer-
1) Advance geological prediction, make clear the distribu- tain magnitude sills with a strike of SN, dipping SWW and
tion, magnitude and features of expansive alternative rocks, angle 30◦ . In ground surface these alterative rocks become
adopt measures to keep away from their affects; weak intercalated layer to control rock slope stability. When
2) In sections with alternative rocks, micro-step excava- the buried depth is over 90 m the cored smectitic alternative
tion, weak explosion, more times excavation, decreasing rocks behave moderate expansion, which threaten the rock
disturbance, setting up shotcrete and anchor-bolts in time; cavern stability. Quick closing and isolating inflow of ground-
3) Drying wash before shotcrete, if not, clean out of alterna- water in time are main engineering measures for prevention
tive rocks with obvious swelling; of associated geohazards.
201
Keywords: case studies, numerical modelling, rock mass classification, rock support, stability analysis, tunnelling
1 INTRODUCTION
202
203
A study on mechanical behaviors of concrete lining and rock caused by shaft sinking
at the Horonobe underground research laboratory
204
C. Vibert
Coyne et Bellier-TE., Gennevilliers, France
P. Vaskou
Géostock, Rueil-Malmaison, France
Keywords: numerical modelling, rock mass classification, rock failure, rock support, stability analysis, tunnelling
1 EXTENDED ABSTRACT of the rock mass. The GSI method is now widely used, but
the complexity of the phenomena occurring in the deforma-
The rapid spread of the use of the geomechanical rock mass tion of a rock mass is such as, obviously, a single rate can
classifications at an international scale is a worldwide and not pretend represent the whole range of possible behaviours
irreversible trend driven by its applications in the profession. depending on the mineralogy, geology, tectonic features, etc.
In particular, rock mechanical engineers value these tools in Moreover, the behaviour is also variable with the type of the
assessing characteristics and failure criteria of rock masses. Works.
The implications on design are therefore obvious. Having witnessed tendencies to a blind application of the
Using case studies, the authors present an update on the ben- recommended methods, especially related to GSI, the authors
efits of the rock mass classifications actually used at an inter- have pointed out some of the most frequent misuses, which
national scale, namely the Q-system, RMR and GSI, although most of the time actually discarded the limitations or posterior
other classifications, such as the Chinese and Japanese one, notices from the originators of the rock mass classifications
exist. With regard to the limitations of these classifications, themselves. They recall that the method is to be used with
much has already been published; therefore attention of the judgement, and that, despite the comfortable assurance the
Reader is focused on the specific aspects where it is felt that strict application of formulas provides, it might have to be
strong recommendations are to be done for a correct, conve- adapted to fit the requirements of the Project, and that critical
nient and efficient utilisation, from on-site estimation of rock acumen shall apply in any case.
mass ratings to the use made of results for design purposes. Different examples from the experience of the two authors
First, a review is made of the positive aspects of the rock are listed to illustrate the limitations in the application of the
mass classifications, and principally: rock mass classifications.
• to impose a minimum description of the characteristics of
The consequences of the simplification of rock mass char-
acteristics to a single rate is a first one, since similar ratings
the rock masses, which have to be recorded,
• to allow, in similar contexts, direct comparison between
may be obtained for fractured hard rock on one side, and quasi-
homogeneous soft rock on the other side, while deformation
different rock masses, at various scales, between different
pattern at tunnel excavation is likely very different.
sites, or within a same project, and/or to monitor the varia-
The scale effect, when assessing ratings from a small pilot
tion in the geotechnical characteristics of the rock with the
gallery, and comparing it to ratings finally obtained once the
progress of the Works.
cavern is excavated, which demonstrates that the Q-index is
A review of the most frequent difficulties encountered when actually not an intrinsic characteristics of the massif, but more
assessing the rock mass classes is made, with insistence on the or less depends upon the relative size of the Work with regard
absolute need of a minimum of geological and geotechnical to the rock mass features.
skills for the one performing the assessment. Some important Improper use of serial correlations between Q-index, RMR
aspects may be neglected if the responsible staff in charge of and GSI is pointed out, too.
the Project does not have an experience in rock mechanics, The presence and frequency of specific zones – karsts,
such as occurrence of alterable or swelling material. Ability faults – requiring special treatment like grouting ahead of
to 3D visualisation is also required for a proper assessment. the face, can not be assessed from the only rock mass
Assessment of the different parameters used for ratings, classification.
such as RQD, Jn and SRF, also require caution and engineering The effect of saturation on the strength of the rock matrix
judgement. is to be checked in any case, and especially in hydropower
But the most sensitive aspect – the most interesting, too – projects, where the rock foundation is to be saturated. Some
is the possibility given to the Engineer to extrapolate the rocks exhibit a very significant drop in strength in presence
geotechnical characteristics of the rock matrix and disconti- of water, which cannot be assessed from the tests performed
nuities until obtaining a range of geotechnical characteristics on dry samples.
205
206
1 INTRODUCTION used more and more widely in pug design now, but the mechan-
ics effect of plug is still not studied systematically. Based on
The hydro diversion tunnel should block with a permanent the consideration mentioned above and combined with the
concrete plug before impoundment of the reservoir while diversion tunnel project, the mechanical characters of the plug
finished diversion. Although the amount of the project of are analyzed, and the optimization method of plug design has
hydraulic tunnel plug generally small, but as one of the main been studied in the theoretical meaning and practical level.
control conditions of hydropower engineering, its construc- (Wu, et al. 2006, Xu, et al. 2007).
tion are urgent. It has an extremely important significance for
the latter part of the project whether the blocking job to finish
promptly. At present, the Hydraulic Tunnel Design Specifica- 2 ANALYSIS OF THE MECHANICAL EFFECTS OF
tion (SL279–2002) gives only general principles for design HYDRAULIC TUNNEL PLUG
of the plug, the actual project has been formulated in their
design principles and selecting parameters, also does not have 2.1 Project profile
uniform standard. It is very important for the safe operation
of hub to block diversion tunnel, but design length of many Jin-Jia-Ba hydropower project diversion tunnel cross section
plugs tend to over-indulge as result of the irrational design and the concrete plug section are shown in Figure 1 and
principles and any other reasons (Zhao, 2004, Chen, 2007, Figure 2.
Su, 2008).
At present, the major method of sealing body design for
construction diversion tunnel are as follows: 2.2 Surrounding rock and plug three-dimensional finite
element numerical simulation
1) The length of the plug is determined according to the
multiples of the sealing diameter. In order to reduce the effect of boundary element analysis
2) The length of the plug is determined according to the model for calculating results, take cross-sectional area as:
experiential formula, the major formulas are as follows: 109 m × 111.5 m, which consider 50 m around the diversion
L = (3∼5)H /100 or L = mHD. tunnel cross-sectional area. The length of the model was
3) The length of the plug is determined by all kinds of for- 120 m, which take 30 m upstream and downstream of the plug.
mulas, such as cylindrical punching shearing formula and Finite element model shown as Figure 3.
against shearing formula. But still many factors of the The simulation of the support effect of bolts is considered
plug’s stability are not considered sufficiently. according to equivalent principle. The cohesive strength of
4) The interaction between plug and surrounding rock can be anchorage body after applying bolts is as follows (Li, S.C.
simulated with 3-D finite element method. et al. 2002):
Although, the development of finite element theory and
computing software is great, the 3-D finite element method is
Figure 1. Dimension of diversion tunnel cross-section (m). Figure 2. The dimension and shape of the plug section (m).
207
208
P.L.P. Wasantha, P.G. Ranjith, D.R. Viete, A. Haque, J. Kodikara & A. Bouazza
Department of Civil Engineering, Monash University, Melbourne, Victoria, Australia
Keywords: Lab testing, Mine design, Rock joints, Stability analysis, Rock failure
Civil and mining engineering projects that deal with rock loading direction) has a significant influence on failure mech-
are confronted with difficulties that arise due to the inherent anisms, with failure occurring by brittle ‘splitting’ for bedding
heterogeneous behavior of jointed rock masses. Prediction of angles between 0◦ and 30◦ , ‘shearing along bedding planes’
the mechanical behavior of jointed rock masses can be dif- for bedding angles between 45◦ –90◦ and by a combination
ficult due to uncertainties in the independent contribution of of these mechanisms for bedding angles between 30◦ –45◦
intact rock properties and discontinuity properties to the rock (Table 1). Strain concentrations on the surface of the samples
mass behavior. Currently, models such as the Hoek-Brown cri- (as shown by the ARAMIS camera results of Table 3) display
terion and the Mohr-Coulomb criterion are popularly used to distinct differences in the location and orientation of zones of
determine the strength of jointed rocks, but these treatments incipient failure for the various failure modes associated with
do not provide any indication of the expected failure mecha- different bedding orientations. Distinctions between the three
nisms for a jointed rock mass. A thorough understanding of failure modes can also be observed in the AE results of Table 1,
the influence of discontinuity properties on failure mecha- which show a sudden increase in AE counts for ‘splitting’
nisms remains elusive. This study addresses this shortcoming failure, a two-stage increase in AE counts for ‘mixed’ fail-
by focusing on the role that discontinuity geometry plays in ure (representing initial splitting and later failure by ‘shearing
governing failure mechanisms in natural rock. Uniaxial Com- along bedding planes’or vice versa) and intermittent increases
pressive Strength (UCS) testing was carried out on natural in AE counts (representing numerous discrete episodes of fail-
sandstone samples with various bedding orientations (Fig. 1). ure that contributed to the overall failure) for the ‘shearing
To better identify the operation of the various mechanisms along bedding planes’ failure mechanism.
available for failure, an optical deformation and strain measur-
ing (ARAMIS camera) and Acoustic Emission (AE) system
were employed for all the UCS tests (Fig. 2).
The results of the UCS tests revealed that bedding orien-
tation (as measured from the direction perpendicular to the
209
Failed specimens
ARAMIS camera
AE system
210
J.-H. Wu
Graduate Institute of Geophysics, National Central University, Taoyuan
S.-S. Lin
Department of Harbor & River Engineering, National Taiwan Ocean University, Keelung
Subject: Rock material and rock mass property testing (laboratory and in situ)
211
212
J.Y. Xu
Engineering College, Air Force Engineering University, Xi’an, China;
School of Mechanics and Civil Architecture, Northwest Polytechnic University, Xi’an, China
Subject: Rock material and rock mass property testing (laboratory and in situ)
ABSTRACT: The research of constitutive model of rock- Through research of this paper, four aspects of conclusions
like materials is one of the most important aspects in the are demonstrated as follows:
rock mechanics study at all times. These models are essen-
1. By using the theory of new rock micro-unit strength and
tial for studying blasting mechanism, stress wave propagation,
Drucker-Prager failure criterion, a statistical damage con-
destruction effect analysis of underground structure, etc. At
stitutive model which satisfies weibull distribution was
present, constitutive model of rock-like materials in static
developed. The model can reflect confining pressure effect
uniaxial stress state has been investigated widely, and can
and variation law of rock strength with confining pres-
be proved by experiment accurately. However, as for compli-
sure. It can also simulate the deformation after pear point
cated stress state, especially static-dynamic coupling loading,
and residual strength of rock perfectly. Moreover, it reflects
the investigation of constitutive model is much more com-
that the damage of rock failure proceeding not only changes
plicated, and still at exploring stage. The development of
with rock micro-unit strength, but also depends on the stress
constitutive model and impact mechanical properties tests
state of rock.
of rock are closely linked. At present, Split Hopkinson
2. Through discussing the physical meaning of parameters
Pressure Bar and light air gun instrument were mainly used
in the constitutive damage model, it can be believed that
to study the uniaxial dynamic characteristic of rock. But there
F0 can judge the strength and m can judge the brittleness
were rare reports on the dynamic mechanical properties of
of rock, and damage correction coefficient q can reflect
rock-like materials under confining pressure. In fact, rock
the residual strength of rock. The relationship between
is under the condition of high static stress or ground stress
parameter F0 , q and strain-rate is approximately linear,
before bearing dynamic stress in engineering practice and
and parameter m obeys the quadratic multinomial function
nature. So it is meaningful to investigate constitutive model
change law with strain rate.
and dynamic mechanical properties of rock under loading with
3. Experimental results show that stress peak and integral
active confining pressure.
level of stress-strain curves of amphibolites under con-
Firstly, on the basis of damage mechanics and statistical
fining pressures were higher than uniaxial dynamic load.
theory, through combining visco-elastic model and statistic
Under the same confining pressure, the dynamic compres-
damage model, a new statistical damage constitutive model
sive strength of amphibolites are strain-rate dependent.
of rock based on weibull distribution under dynamic loading
The strengthening effect becomes more prominent with the
is developed. Then, in order to simulate real stress state of
increase of confining pressure at the same range of strain
rock in underground project, the traditional SHPB apparatus
rate.
was improved. The impact test of amphibolites with differ-
4. Stress-strain curves calculated from proposed model has a
ent confining pressure and strain rate was carried out by using
good agreement with that from experimental data directly,
improved 100 mm diameter SHPB. Thus parameters of consti-
which indicates the constitutive model is reliable. What’s
tutive model were obtained, and the influences of parameters to
more, the new model with less parameter is simple and easy
constitutive model were analyzed. In order to improve the pre-
for application, and it can offer reference for both further
cision, the statistical damage constitutive model of rock was
research of rock mechanical properties and engineering
modified reasonably, so as to set up dynamic damage constitu-
application of underground engineering.
tive model of rock which corresponds to reality better. Finally,
compared with SHPB experimental curve, the rationality of
constitutive model has been tested and verified.
213
Global stability analytical method of cavern group in salt rock based on energy criterion
Keywords: numerical modeling; oil reservoir; rock caverns; rock failure; stability analysis
214
215
Serious shaft lining ruptures have often occurred in the east- The UCIC is constructed in the bottom aquifer, can interrupt
ern part of China, such as Datun, Xuzhou, Huaibei, Yanzhou, the connection with the seepage path, and eliminate the effect
Yingxia, Hebi, Dongrong, etc, since 1987. The geological of bottom aquifer drainage to shaft lining. Furthermore, the
conditions of all these spots are the almost same. All these UCIC has the strength to reinforce the aquifer around the shaft
shaft linings pass through deep topsoil of Quaternary strata lining, so that the effect of subsidence can be decreased and
for which the composition of the bottom aquifer is complex; the additional force can be exterminated from this point.
and before shaft lining rupture, the water head of the bot- Finite element software was used to simulate the rein-
tom aquifer has fallen dramatically and large scale subsidence forcement effect of the UCIC method. Phase2 7.0 is a 2D
occurred. During the water head falling, the soil layer applies elasto-plastic finite element analysis program for underground
additional force to the outer shaft lining. As the additional or surface excavations in rock or soil. An axisymmetric model
force is increasing with the effect of other joint forces, the was built for the simulation and the UCIC is set near the
condition of shaft lining goes into a plastic state gradually shaft lining. There are four stages in this model for simu-
and then can be ruptured. As the maximum additional stress lation. Stage 1: Initial Model; Stage 2: Formation of a shaft
occurs in the bottom aquifer and near the bedrock surface, (diameter 8 m); Stage 3: Formation of UCIC (thickness 1 m);
most of the ruptures have happened in this zone. Stage 4: Increasing pressure around the shaft. The location of
Based on an analysis of other methods and the mechanism of the UCIC is mostly built in the aquifer and in this analysis the
the shaft lining rupture, a new treatment method was proposed range and width of the UCIC are expanded for simulation.
for shaft lining reinforcement with underground continuous The widths of the UCIC were in three widths as 1 m, 2 m
impervious curtain. From the results of a series of numeri- and 3 m. From the results of various conditions, the most
cal analyses, the UCIC built in bottom aquifer can avoid the effective width of the UCIC is 1 m. When the UCIC was only
stress concentration in shaft lining. The current researches are built in aquifer or built in aquifer and 10 m above were more
mostly focused on the causes, mechanism and solutions for effective than built in other ranges.
these kinds of geotechnical issues in the engineering projects The changes of the safety factors show that the effect of
and they still have been the hot topics in recent 20 years in the UCIC depends on the UCIC materials.
the shaft construction. In this research, the various range and In this study, the failure criterion is Mohr-Coulomb fail-
width of UCIC built in aquifer were analyzed, and the effect ure criterion. The safety factor of the shaft lining was the
of the divers range was discussed, then the optimum condition evaluation criterion for the effect of the UCIC.
was determined. When the UCIC were built around the shaft lining, the safety
In this analysis, we choose Baodian coal mine as a typical factor in the boundary between upper layer and aquifer didn’t
example for the mines which located in the eastern China. change with the various conditions. The lowest safety factor
This mine belongs to Yanzhou Mining Group in Shandong was happened in the boundary between the aquifer and the bed
province, and is a concealed coalfield which the shafts pass rock. When the range and width were both larger, the safety
through the Quaternary stratum and Jurassic stratum. factor of the shaft lining had obvious reducing.
Underground continuous impervious curtain (UCIC) is an The most effective width of the UCIC was 1 m near the shaft
underground wall that has the anti-seepage capacity. This lining. When the UCIC was only built in aquifer or built in
method for shaft lining rupture treatment is based on the aquifer and 10 m above were more effective than built in other
mechanism of shaft lining rupture and borrows ideas from ranges. Material’s property played an important role in shaft
the existing treatment method such as strata grouting. The lining treatment using the UCIC.
UCIC is constructed outside of the shaft lining in a certain
area, adopting the vertical cutting mixing technique, mixing
cement grout with soil or concrete to be a wall.
216
X.P. Zhang
Key Laboratory of Engineering Geomechanics, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
L.N.Y. Wong
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
S.J. Wang
Key Laboratory of Engineering Geomechanics, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
Anisotropic and foliated rocks such as slates, schists and to provide data for predicting the behavior of rock masses
phyllites, the behavior of which is dominated by closely spaced after the excavation of the underground powerhouse and main
planes of weakness, cleavages or planes of schistosity, present transformer hall of the DanBa hydroelectrical power project,
particular difficulties in the determination of rock strength an exploration adit, which was 298 m long and 3 m diameter,
parameters. In choosing the strength values for anisotropic was first excavated for survey. Convergences after the exca-
rocks, one has to decide whether the highest, the average or the vation were measured, based on which a back-analysis was
lowest values obtained from testing should be used. Accord- performed to determine the strength parameters.
ing to our best knowledge, a widely adopted guidance on the To ensure the uniqueness of the back-analysis solution
choice of anisotropic rock mass parameters is unavailable. and to minimize the number of iterations of back-analysis
This paper presents a displacement based back-analysis required, the parameters for back-analysis were chosen based
method (flow chart as shown in figure 1) for the determina- on the following principles: (1) Select parameters that are of
tion of the elastic modulus (E), cohesive strength (C), friction greater influence on the deformation of underground exca-
angle (ϕ) and in-situ stress of a schistose rock mass. In order vations. (2) Select parameters that are very difficult to be
obtained accurately by other methods. (3) The number of
unknown parameters should be as small as possible. There
are a number of parameters affecting the rock mass defor-
mation which include elastic modulus (E), cohesive strength
(C), friction angle (ϕ), Poisson’s ratio (ν), tensile strength
(σt ), dilation angle (ψ) and in-situ stress. According to the
above-stated principles, E, C, ϕ and in-situ stress were cho-
sen for back-analysis in the present study. A preliminary
assessment revealed that the rock mass deformation was rel-
atively less sensitive to the variation of the Poisson’s ratio,
tensile strength and dilation angle. Constant single values of
ν = 0.25, σt = 2.7 Mpa, ψ = 10◦ were thus assigned based on
preliminary laboratory test results and recommendations in
the literature. In a 2D calculation model, there are three in situ
stress components σxx , σyy and σxy . In the present study, the
rock mass is assumed to be isotropic. Including E, C and ϕ,
a total of six parameters thus took part in the back-analysis.
The second step is to define a range for each of the six cho-
sen parameters to be used for back-analysis. The range of
each parameter was decided with reference to the preliminary
laboratory test results, which was then slightly broadened to
different extents: E from 1 GPa to 20.5 GPa, C from 1 Mpa
to 8 Mpa, ϕ from 24◦ to 49◦ , σxx from 8 MPa to 12 MPa,
σyy from 8 MPa to 12 MPa, and σxy from 0 MPa to 2.4 MPa.
Figure 1. Flow chart of back-analysis calculation. The designed range of each parameter is then respectively
217
218
Different analytical models and methods to study wave propagation across rock joints
Rock Masses are different from other engineering materials reflections are taken into account. Based on one-dimensional
by the existence of a great number of joints. They results in wave equation, relations between particle velocity and stress
the discontinuous characteristics of rock masses, and greatly along right- and left-running characteristics can be built.
affect the properties of rock masses. When a wave propa- Response at certain point and time can be derived from those
gates in jointed rock masses, it is both slowed and attenuated, on previous times.
mainly due to the existence of joints. The effects of rock With the RM, relations among different layers with respect
joints on wave propagation are of fundamental interest in geo- to potential amplitude or stresses and displacements can be
physics exploration, rock dynamics, earthquake seismology, established with a standard algorithm. The standard algorithm
and hydrocarbon production. is a process to take into account the effects of all the joints on
This paper conducted an overview of different models to seismic response at certain point in the form of a product of a
describe the physical properties of joints in wave propagation lot of functions or matrices.
and methods to take into account multiple wave reflections With the VWSM, virtual wave source exists at the joint
among joints. The joint models include the layered medium position and represents the mechanical properties of the joint.
model (LMM), the displacement discontinuity model (DDM), It produced reflected transmitted waves each time an incident
the wave scattering model (WSM) and the equivalent medium wave arrives at the joint. The overall reflected wave and trans-
model (EMM). Bases on these models, the effects of joints on mitted wave are the results of all the reflected waves and all
wave propagation can be studied. However, when multiple par- the transmitted waves arriving at different times, which are
allel joints exist, wave propagation is more complicated due created by VWSs.
to multiple wave reflections among joints. There are mainly The DDM can be regarded as an approximation of the
three methods to take into account multiple reflections among LMM or WSM. Parameters in DDM can be obtained from
joints, i.e., the method of characteristics (MC), the recursive those in the LMM or WSM. Similarly, parameters in the LMM
method (RM), and the virtual wave source method (VWSM). can be deduced from those in the WSM. Changes of equivalent
Comparison of these models and methods is also performed moduli of the EMM due to the presence of joint are function
to detect the relationship and difference among them. of parameters used in boundary equations of the LMM, DDM
With the LMM, both the stresses and displacements across or WSM.
the interface are continuous. There are two kinds of treat- The LMM can be used to study wave propagation across
ment of joints within the LMM. One is that the joint is treated filled joints or perfectly bonded joints. The DDM is applica-
as a perfectly bonded interface, and both the displacements ble to non-perfectly bonded joint with thickness much smaller
and stresses across the interface are continuous. The other is than wavelength. With the WSM, the effects of cracks in joints
that the joint is treated as a layer of the filled weak medium on wave propagation can be taken into account. The EMM
sandwiched between two fully-bonded interfaces. has advantages over other models in engineering estimate
The basic assumption of the DDM is that, as a wave prop- the overall influence of joints on wave transmission and
agates through a joint, the stresses are continuous, while the disadvantages of lose of joint discreteness.
displacements are not. The MC works only in time domain, while the RM and
The WSM treats the joint as a plane boundary with a dis- VWSM function in both time and frequency domains. The
tribution of small cracks and voids. The wave reflection and RM and VWSM can be adopted to study two dimensional
transmission across a joint is the results of wave scattering wave propagation across joints, while it is difficult for the MC.
through all cracks. Due to the discretization of time, the MC is semi-analytical.
The EMM treats problems from the viewpoint of entirety. The VWSM is also semi-analytical because of the prescript
From the EMM, the rock material and the contained joints of times of wave propagation among joints. The RM is purely
together are approximated by an equivalent continuous, analytical method without any approximations.
homogeneous and isotropic medium. The effect of joints is All models and methods have their own advantages, dis-
lumped into effective moduli of the equivalent medium. advantages and application areas. Depending on the problem
The MC is a mathematical tool to study wave propagation to be solved, specific model or method can be chosen and
across different layers or discontinuities, where multiple wave adopted.
219
D.A. Beck
Beck Engineering Pty Ltd, Sydney, Australia
G. Putzar
Beck Engineering Pty Ltd, Berlin, Germany
Keywords: numerical modelling, monitoring, mine design, back analysis, stability analysis, rock mass classification
In most stress-deformation simulations of block or sub computes a new solution for stress, deformation, damage and
level cave induced rock mass damage and subsidence, the fault movement. As a consequence of the draw, an unstable
cave growth is approximated using the outputs of empiri- zone in the cave back – measured using nodal velocity – may
cal or stochastic tools that assume propagation without due thus develop at each coupling step and slough into the cave,
regard for the stress, strain, strength and structural conditions at which point these elements become available to be drawn.
required for cave propagation. The cave growth is usually The process is repeated following the draw schedule, and is
assumed to occur following simple rules related to draw, and able to simulate most cave propagation phenomena including
will propagate in the stress-deformation model whether cave stalling and chimneying.
back instability is indicated to occur or not. Example results are shown from an example mine where a
To overcome the problem and better estimate the likely comparison of the model results and field observations over
performance of caves, a tool that properly accounted for the 12 months showed a good correspondence between forecast
physical coupling of the cave material to the un-caved rock and actual induced seismicity, extraction level damage and
mass and the draw schedule, driven by the known physics of cave growth. A similar result has been achieved at several
both parts of the problem was needed. other mines.
After analysing a number of alternatives, a coupled Dis- It is concluded that the coupled DFE-LGCA simulation
continuum Finite Element (DFE) – Lattice Grain Cellular procedure enables rapid simulation of cave propagation, flow
Automata (LGCA) scheme was developed. In this scheme, and induced deformation driven by the cave draw schedule.
the LGCA part computes the velocities within the cave and The method can be calibrated directly using observations of
changes in airgap geometry due to an increment of draw. cave back location, grade and recovery, seismicity, tunnel
Particles are removed through drawpoints, element by element damage, tomography or ground movement. The close match
following the draw schedule. A relation between average par- between results of coupled flow deformation analysis and field
ticle flow velocity and modulus for cells within the cave is measurements suggests the technique is useful for forecasting
then used to imposed cave state changes node by node in the of cave induced deformation, and is especially useful for sim-
DFE model. These state (modulus) and geometry changes in ulating cave propagation and assessing risk related to the draw
the cave mass in the DFE model disturb the quasi-equilibrium schedule.
between the cave and the rock mass and the DFE model then
223
Keywords: mine design, numerical analysis, bolts and anchors, rock mass, stability analysis
Partial extraction room-and-pillar mining systems provide the value of the ratio of the horizontal stress to vertical stress
about 60% of the underground coal mined in the USA. This (M), width of opening to height of opening (Wo/Ho), and
mining system develops 3-way and 4-way intersections. About whether the immediate roof layers are bonded or un-bonded,
70% of these falls occur at intersections. Similarly, over 80% 2) The ratio of the Wo/Ho is most critical for intersection
of these falls occur at intersections in Illinois. Thus, there is a stability, 3) The peak value of the compressive stress due to
significant need to improve stability of intersections. The goal curvature is located about 0.2 m from the edge of the open-
here was to analyze stress distribution and instability around ing toward the center of the intersection, 4) The effect of
a 4-way intersection in a typical geologic environment in the de-bonding the layers is to increase the SCFs values and the
Interior Coal Basin, USA. An additional goal was to evaluate effect of increasing the M1 (M-value in the E-W direction)
primary and secondary support layouts for improving their is also to increase the HSCF and SSCF values, 5) The effect
stability. The results of numerical analyses of stress distribu- of de-bonding the layers is much more significant than the
tions and associated instabilities with and without primary and M1 value and explains why more failures are associated with
secondary supports installed are presented. thinly laminated strata in the immediate roof, 6) Increasing
Three-dimensional finite element analyses were performed the M1 values increases the peak VSCF only slightly but the
with estimated Hoek-Brown failure criterion parameters for HSCF and SSCF values are significantly increased, 7) Fail-
different lithologies associated with the coal seam. Vertical ures are likely to initiate around the edges of intersection
stress of 2.5 MPa (100 m mining depth) was applied at the top pillars, 8) Stability of an intersection would be significantly
of the model. Different lateral stress ratios were applied in improved if additional primary and secondary supports were
the E-W (M1 ) and N-S directions (M2 ), where M1 and M2 provided around the corners of an intersection to minimize
represent the ratios of lateral stress to vertical stress in the two failure initiation, 9) Increasing mining depth results in initi-
orientations. The immediate roof strata above the 2.0 m thick ation of shear and tension failure at intersection corners, 10)
coal seam are Black Shale (2.0 m), Gray Shale (0.6 m), weak Weak floor strata results in slight increase in the yielded zones
limestone (0.6 m), weak shale (0.9 m), competent limestone in the roof, and pillar ribs and significant yielding around
(1.2 m) and 6.7 m thick shale. The strata below the coal seam the floor corner, 11) Primary supports are not adequate to
consist of Gray Shale (1.0 m), weak limestone (1.0 m) and ensure stability of intersections, and 12) Fully-grouted bolts
10 m thick shale below the weak limestone. Models 1 to 6 with pre-tension, located around the corners of an intersec-
used above lithologies. In model 7, a 0.6 m weak claystone tion demonstrate the best potential to improve stability of an
layer below the coal seam was included. For unbonded layers intersection. A 5-bolt pattern around intersection corners, with
analysis, the cohesion between different layers was assumed 2.4 m long vertical, #7 rebar, pre-tensioned bolts appears best
to be zero. as secondary support. A coal company has agreed to demon-
The results indicate that 1) The peak value for each SCF is strate the effectiveness of proposed secondary supports and
located around the edge of the opening and its precise location planning is underway at this time.
shifts toward the opening or toward the pillar depending upon
224
3-D modeling and parameters research of irregular shape particles for DEM
based on computerized tomography
225
4.1 Conclusions
where: djm is the distance between B-S j and m; Rj , Rm are In this paper, the 3-D coordinates of irregular shape parti-
the radii of B-S; min and max indicate the smaller one and the cle had been collected by dealing with CT images. A model
bigger one in this two B-S. generate program had been compiled based on the spatial
Finally, the split-radius Rd is introduced. The function of coordinates. The program built particle model by searching
Rd is dividing the B-S into two groups. The B-S which radii the minimum distance between the inner points and sur-
larger than Rd are given a relatively high value of K, and the face points. And a quantitatively evaluate method of model
B-S which radii smaller than Rd are given a relatively low value accuracy has be proposed.
of K. By this approach, the number of B-S will be reduced By the error analysis, some meaningful rules are found:
extra and the accuracy of characterization will be decreased (1) the function of Rlim error-reduction goes to strengthen
insignificantly. progressively, while the function of K goes to weaken (2) Rlim -
decrease should be the first choice for improving accuracy
and if the accuracy is not advanced obviously, the K-decrease
3.2 Model generation and parameters research
method should be executed. (3) The parameter Rd can miti-
Base on the software platform of Visual C Studio, the mod- gate the conflict of error and B-S number partly. Finally, the
elling programme is developed. The general programme is balance of B-S number and model errors had been searching
divided into four modules corresponding to the four steps by adjusting three governing parameters.
of modelling ideas. Moreover, particle models simulated in
different accuracy grade can be obtained by adjusting the
three governing parameters. 4.2 Discussions
The purpose of introducing parameters is reducing the num- Comparing with the method of inscribed sphere, the boundary
ber of B-S in models and the meantime guaranteeing the error in our study is circumscribed sphere error, and the two
accuracy of simulation.As the B-S number is positively related methods are similar. But in the sight of calculation speed, the
with the accuracy of model, reducing the number of B-S radius searching is more effective and faster than the inscribed
will cause the accuracy of model shape loss, adding the B-S sphere method.
number will cause computer resource waste. Therefore, three Three governing parameters are interrelated, and it makes
governing parameters are introduced for the aim of finding a the optimization between B-S number and accuracy becomes
suitable balance between B-S number and accuracy of model. more complex, which needs more deep study.
If the accuracy has been determined, the parameters should As the method introduced in this paper is overlap-able
maximize the reduction of B-S number. Then, the number of algorithm, the model’s density distribution is probably dif-
B-S and accuracy of model can be accepted simultaneously. ferent from the real particles. The enhancement of similar-
Comparing the B-S number and the errors of the models ity between numerical model and archetype needs further
which change with the values of parameters, it is found that research.
with the increasing of K and Rlim , the B-S number is reduced,
and this reducing has a high pertinence and sensibility with
K. The B-S number reduced with the increasing of K, how- AUTHORS
ever, the errors of model increased correspondingly, and vice
versa. The third parameter Rd is trying to mitigate the intense Du, X. School of Civil Engineering, Wuhan University, Wuhan,
effect caused by parameter K, moreover, it can mitigate the Hubei, 430072, China. duxin0218@126.com.
increasing of model errors which are caused by the reducing Zeng, Y. School of Civil Engineering, Wuhan University, Wuhan,
of B-S number. Hubei, 430072, China. zengyw@whu.edu.cn.
226
227
REFERENCES
228
229
ACKNOWLEDGMENTS
Authors would like to thanks Italferr spa for the constant sup-
port and data and in particular: the Project Manager Ing. Giulio
Marcheselli; the Project Manager Assistant Ing. Giuseppe
230
T. Funatsu
Institute for Geo-Resources and Environment, National Institute of Advanced Industrial Science and Technology, Japan
N. Shimizu
Department of Civil and Environmental Engineering, Yamaguchi University, Japan
In order to evaluate the application of particle flow code with 2. The fracture toughness obtained by the numerical modeling
clump model to crack problem on rock, we conduct the sim- is lower than that obtained by the experiments. However,
ulation of fracture toughness testing using SCB specimen, the difference is within 10%.
and conventional uniaxial compressive testing and Brazilian 3. The deformation behavior during fracture toughness testing
testing. The simulation result of uniaxial compressive test- is different from that of numerical modeling. This differ-
ing is shown in Figure 1. In addition, the Brazilian testing ence may be caused by the difference of particle of element
shows that the mechanical properties obtained by the simula- and mineral particle of natural rock. The further study is
tions are close to those obtained by the experiments. And then needed to solve the difference.
SCB specimen for evaluating fracture toughness is modeled
by using the input parameters collected above. The SCB model
is shown in Figure 2. And load versus displacement curve is
recorded shown in Figure 3. The clump model can reproduce
the value of fracture toughness as well as fracture process zone,
followed by the brittle failure. In conclusion, the particle flow
code with clump model is able to model the rock with strength,
elastic property, and crack propagation property. The main
results obtained from this work are summarized as follows:
1. The distinct element method can model Kimachi sandstone
by giving the appropriate micro-parameters. Especially,
matching the properties of compressive strength and ten-
sile strength at the same time is successfully conducted by
adopting clumped particle model.
Figure 2. Model of SCB specimen.
Figure 1. Comparison of stress-strain curves obtained by simulation Figure 3. Load versus displacement curve obtained by numerical
to experiment. simulation.
231
Keywords: numerical modelling, rock joints, stability analysis, rock stress, rock failure
232
Back analysis of tunnel response using the Differential Evolution Genetic Algorithm
(DEGA)
Marte Gutierrez
Division of Engineering, Colorado School of Mines, Golden, CO, USA
Sotirios Vardakos
Parsons Brinckerhoff Geotechnical and Tunneling, New York, NY, USA
Caichu Xia
Department of Geotechnical Engineering, Tongji University, Shanghai, PRC
233
234
The assessment of the hydro-jacking risk of the pressurized water tunnel by theoretical and
numerical method
H.G. Lee
Doosan Engineering & Construction, Seoul, Korea
H.S. Yang
Dept. of Energy & Resources Engineering, Jeonnam Nat’l Univ. Gwangju, Korea
Keywords: case studies, fluid flow, hydraulic fractures, risks and hazards, stability analysis, pressurized water tunnel
Different from the common tunnel such as road/railroad tun- occurrence of hydro-jacking in the condition of before and
nel, a pressurized water tunnel is known to be exposed to water after constructing emergency spillway tunnel, as well. Safety
pressure applying on an inside of tunnel. The internal water factor by Norwegian theory for the diversion tunnel at present
pressure acts on the concrete liner of tunnel, and it causes state was evaluated as 2.4 which were evaluated over allowable
damage in the liner as well as leakage into the surrounding safety factor of 1.3. The minimum safety factor evaluated by
rock mass. Apart from damages in the liner, the surround- numerical analysis is 1.47 which calculated from proportion
ing rock mass also can be damaged from crack initiation and between pore pressure with minimum principal stress at the
propagation by water leakage. Therefore, a careful analysis on current state. An also, the safety factors calculated by theo-
the hydraulic damage should be carried out when the stabil- retical and numerical method at the state after constructing
ity of internal water tunnel is needed to be concerned against
“hydro-jacking” which status that pre-existing joint is opened
and grown by leakage is predicted when stability analysis of
the water tunnel is carried.
In this study, hydro-jacking related analysis has been per-
formed on the stability of old drainage tunnel at dam site with
theoretical and numerical methods. It was predicted from the
stability analysis that hydro-jacking can be occurred at the
existing pressurized tunnel when new drainage tunnel is con-
structed nearby. On close examination it was found that the
safety factor of the old drainage tunnel was higher than a
permissible range.
The Norwegian criterion among theoretical method was
adopted for evaluating hydro-jacking hazard of existing
drainage tunnel. The numerical method was carried to predict
Figure 1. Schematic diagram of Norwegian Criterion. Figure 2. Present and planned state of construction site.
235
236
Dynamic analysis of rock tunnel failure by using 3-D Numerical Manifold Method
L. He & X. Huang
Department of Structures and Mechanics, Nanyang Technological University, Singapore
G.W. Ma
Department of Civil and Resource Engineering, The University of Western Australia, Australia
Keywords: Rock dynamics; Numerical modeling; Rock tunnel; Rock failure; Key block; Stability analysis; NMM
A number of methods are available to analyze tunnels, from NMM simplifies the modeling process by avoiding the tedious
the earliest closed-form analytical solutions to the most meshing operation. The numerical description on the irregu-
recent numerical modeling methods. Normally, we have two lar joint network in rock engineering is easily achieved. In
branches of approaches to rigorously analyze tunneling prob- this paper, weak form of governing equation, NMM interpo-
lem (Barla & Barla 2000): Equivalent continuum approach lations, and discrete equations are introduced for the dynamic
and Discontinuum approach. The continuum assumption has analysis. Rock falling from the side and roof of a horseshoe-
obvious disadvantages since the rock mass failure is largely shaped tunnel is studied using the proposed 3-D NMM, setting
dominated by the joint structure. Strong tectonic discontinu- as in Figure 1, and it is an important task to find and describe
ities cut the surrounding rock blocks of a tunnel surface into the most critical key block around an excavation in such
discrete state.An equivalent continuum cannot reflect the main engineering problem.
aspects of such blocky system. It might be improper to ana- The calculation consists of two stages. At first, the fric-
lyze rock tunnel with only emphasizing the deformation or tion angle is set to 59 degree. The designed system reaches
strain based on the continuum assumption. Block deformation a steady state with proper systematic damping. Secondly,
could not substitute the displacement in the block system com- through reducing the friction gradually, it could reach the case
pletely. The enforcement scheme design, based on continuous until one of key blocks starts to detach out from surrounding.
assumption, might not be safe especially in 3-D situation. Such Gradually reduce the value of friction angle, there are two
averaging design of bolt and lining might not be sufficient for
locally loose blocks.
In general sense, the NMM can be thought as a bridge
connecting two branches of numerical method: continuum
method (e.g. FEM) and discontinuum method (e.g. DDA). The
achievements on “The description of irregular joint network”
in rock mass and “merging FEM and DDA” as a combined
method, are through the concept of “Dual-Cover system” in
this method. Different with traditional finite element method,
it divides the solution domain into two domains: mathematical
and physical domain. More specifically, it builds interpo-
lation function in the mathematical domain,meanwhile the
functional integration is done in the physical domain. The sep-
aration also frees the shape of cover-based element. Therefore, Figure 2. Blocks with diverse shapes in tunnel.
237
potential falling blocks, indexed with Blk #11 (Critical fric- moving downward. From the curvature change along the
tion angle for the roof key block is 11.1070◦ ) and Blk #12 curve, we can also identify that block #11 will be completely
(Critical frictionangle for the sidewall key block is 17.6450◦ ) detached from surrounding at time 0.776 s. It later falls to
(see Fig. 3). the bottom of the tunnel at 1.172 s. Similarly, side block #12
Fig. 4 gives rock tunnel failure process with a frictional is detached at 0.709 s, and falls to the bottom at 1.605 s.
angle of 8◦ , which means that both two key blocks in this This paper strongly provides evidence for 3-D NMM anal-
tunnel model will be detached from the surrounding blocks ysis. The proposed program can simulate the whole process of
and fall down to the bottom. failure, and it can also get more detailed information, despite
The maximum collapsing volume is 0.148 m3 under the the critical friction angle (including collapse time, collapse
setting condition, which will not collapse as a whole. Enforce- mode, volume velocity, etc.). These provide the fundament
ment scheme design should focus on how to prevent small for forward and backward analysis to engineering scenario.
block from falling. Generally speaking, lining and rock bolt It also demonstrates the potential ability to analyze the interac-
should be enough for stabilization. Fig. 5 plots the displace- tive contact interaction among complex blocks in the platform
ment history and indicated that the position of them is mainly of 3-D NMM.
238
1 EXTENDED ABSTRACT
239
REFERENCES
240
Keywords: Numerical modeling, nuclear repository, rock stress, site characterization, stability analysis
1 INTRODUCTION
2 THERMAL EVOLUTION
2.1 Forsmark host rock thermal properties Figure 1. Contours of temperature increase in the plane of the
repository at four instances of time. Positions of vertical scan-lines
The thermal property model accounts for the distribution as are indicated in the upper right.
well as for the spatial variability of the heat transport prop-
erties. The effective equivalent global thermal conductivity
corresponding to the conditions in the hottest positions at the distribution of thermal properties are not important on
the time of the peak is called the dimensioning conductivity the scale of the large scale model mesh resolution, mean-
(Hökmark et al. 2010). Some distance away from the canisters ing that mean values of the thermal properties are valid
the mean conductivity is an adequate approximation. approximations. In high resolution near-field models, rock
wall temperatures are monitored at canister mid-height. For
the dimensioning case, the peak buffer temperature reaches
2.2 Calculations the 100◦ C design limit, less a 5◦ C margin specified in the
All numerical analyses of the thermo-mechanical evolution dimensioning rules to account for uncertainties, some 10 years
related to the safety assessment are performed using the dis- after deposition. The corresponding maximum rock wall peak
tinct element code 3DEC (Itasca 2007). The 3DEC thermal temperature amounts to about 66◦ C.
logic is based on analytical solutions and can be used to
analyze temperature fields generated by thousands of time-
dependent heat sources on all scales in time and space. 3 MECHANICAL EVOLUTION
Variations in heat transport properties cannot, however, be
taken into account. Here, the mean thermal conductivity is The thermally-induced stresses are calculated on three model
used for large-scale analyses, whereas small-scale models are scales: The large, the medium and the small scale. All analyses
analyzed for two different values: the mean conductivity and are performed using the 3DEC code.
the dimensioning conductivity. The large scale models include the entire repository. The
Figure 1 shows the large-scale thermal evolution at 460 m horizontal cross section area is about 60 km2 and the model
depth calculated using the 3DEC thermal logic. The details of height about 3 km.
241
5 DISCUSSION
242
W. Jiang
Key Laboratory of Geological Hazards on Three Gorges Reservoir Area, China Three Gorges University,
Ministry of Education, Yi Chang, Hubei, China
H. Zheng
State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics,
Chinese Academy of Sciences, Wuhan, Hubei, China
243
4 SOLUTION PROCEDURE
with the set of its zeros being the two nonnegative semi axes of
the planar system. That means that ψ(a, b) = 0 if and only if the To this point, we can combine system-M with system-C and
following complementarity condition on two real numbers, a obtain a system of nonlinear equations, called system-MC,
and b, is satisfied
244
Y.D. Jiang
China University of Mining & Technology, Beijing, China
State Key Laboratory of Coal Resources and Safe Mining (CUMTB), Beijing, China
W.J. Wang
China University of Mining & Technology, Beijing, China
Y.X. Zhao
China University of Mining & Technology, Beijing, China
State Key Laboratory of Coal Resources and Safe Mining (CUMTB), Beijing, China
L. Zhang
Bluocean Technology Inspection Co., Ltd, Tianjin, China
J.L. Han
China University of Mining & Technology, Beijing, China
Subject: Rock material and rock mass property testing (laboratory and in situ)
ABSTRACT: There are many researches on studying char- corresponding to different substances. Hounsfield units were
acteristics of coal and rock by using X-ray Computed Tomo- converted into by using experience formula and density values
graphy (CT) technology and finite element method (FEM). were assigned to the corresponding volume mesh. Then elastic
However, there are only a few papers to combine the two modulus and poisson coefficient were defined for every sub-
methods together to study this kind of issues. Combining stance. Finally volume mesh which was assigned the material
X-ray CT and FEM together is a new analysis method. We would be output to ABAQUS files, as inp file.
can get the physics model of coal and rock based on real The commercial FEA suite ABAQUS was used to simulate
material with three-dimensional reconstruction of CT images. the process of uniaxial compression of coal sample. Input the
Finite numerical modeling of materials can directly predict its inp file which was converted from MIMICS to create one
behavior under different loading conditions based on simula- model with meshing situation of 270460 tetrahedron units.
tion rather than indirectly prediction by measured parameters. The loading path was 0-5MP-10MP-15MP-20MP.
Therefore, in this paper we present an example to combine From the simulation results which were shown in figure 1
these two methods to study the X-ray CT technology-based and 2, we can observe the development of maximum princi-
finite element simulation of coal samples under uniaxial pal stress in the vertical section of model during the loading
compression. process and get the following points:
In order to study the effect of compressive loads on coal
1) Stress was firstly produced at some points of inner model;
in finite analyze software based on real materials, we used
2) Secondly, with increasing loading force, high stress points
X-ray microCT scanner to scan the original sample. Then, CT
gradually formed an area;
images of coal sample was import and threshold.The coal sam-
3) Thirdly, high stress area would continue to expand to both
ple was divided into two types of material based on different
ends of the model, which mainly expanded outward along
gray value range, namely coal matrix and mineral impurities.
the interface between two kinds of materials;
The 3D model was established according to the scan data and
4) Finally, stress increased more quickly in high stress area,
threshold classification. After 3d reconstruction of coal sam-
distributed along with 45◦ direction.
ple, in order to facilitate the follow-up finite element analysis,
the meshing on the surface and its optimization and remeshing The variation of maximum principal stress in the middle
was conducted in MIMICS FEA module, and then generated cross-section of model was similar as it in vertical section
the volume mesh. After loading volume mesh data, gray value in the early step, namely stress was firstly produced at some
of hounsfield units was calculated for each mesh unit based points of inner model, but with increasing loading force, high
on scan images data in FEA module. Then hounsfield units stress points gradually formed an area from the outside of
in volume mesh were divided into equal areas, and each area model to the interior, as shown in Figure 2.
245
Figure 2. Distribution of the maximum principal stress in the middle cross-section of model during different analytical steps (unit: Pa).
The micro FEA simulation agrees well with the experiments of the effect of coal under uniaxial compression load. This
and theory, which present the point that the interface between method can be more specific responses of physical properties
two materials will more likely get stress concentration result of small specimen. It has certain significance to study the
in finally fracture. The methods described in this paper can be basic mechanical properties of rock and failure mechanism,
applied to a variety of coal samples in an attempt to predict but it is limited to apply this method for large area (such as
their destruction areas, and to gain a deeper understanding coal seam) to study its stress and strain.
246
T. Kitaoka
Graduate School of Science and Engineering, Kansai University, Osaka, Japan
1 GENERAL INSTRUCTIONS the fluctuation of water level in the shallow monitoring wells.
The amount of the rainfall greatly influences the fluctuation
In the management and a usage of a groundwater resource, of water level in the shallow monitoring wells, and the water
we need to work on an accurate grasp of the current situation level tends to rise in summer when the amount of the rainfall
and to make a future vision in the region where the ground- is large, and to decrease in winter when the amount of the
water exploitation is active. In the past research, a common rainfall is a little. We used monitoring data when we do the
aquifer was examined in one municipality. This simulation analysis. Fig. 3 shows the analysis result. As a result, it can be
was undertaken where the scope was expanded, the subject of seen from this that the fluctuation of water level group wells
which was a pumping well for water service – water levels cause accuracy by using our model made in the present study.
for one pump were large. To use the groundwater resource in If the pumping well is newly established, using this model
this region effectively, we examine the influence on periph- based on the result of measurement of the water level for 20
eral groundwater properties especially where a large amount years or more, we want to forecast its future influence on the
of pumping from the wells for waterworks is being practiced. water level. We also want to use it as a model that can propose
an appropriate groundwater control for the long-lasting use of
ground water in this region, and for the prevention of troubles
2 CONCLUSIONS concerning the wells, subsidence, and uplifts.
Fig. 1 shows the 3D mesh model. In the examination, we con-
struct the 3D mesh model in detail, analyze the seepage flow
by the finite element method, and compare the change of the
groundwater level that the group of pumping wells caused with
the actual measurement values. Fig. 2 shows the relation of
247
Abutment stability assessment of the Hongrin arch dam using 3D distinct element method
A. Koliji
Stucky Ltd, Switzerland
Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland
T. Bussard
Norbert SA, engineering geology and hydrogeology, Switzerland
A. Wohnlich
Stucky Ltd, Switzerland
J. Zhao
Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland
Keywords: rock joints, stability analysis, rock support, numerical modeling, monitoring
The Hongrin north dam is a double curvature concrete arch role of rock discontinuity in the failure mechanisms. The rock
dam located in western Swiss Prealps, which attains 125 m mass is defined as deformable distinct blocks which inter-
high. The right bank abutment of the dam mainly consists of act along frictional discrete discontinuities representing the
intensively jointed Neocomian limestone and exhibits zones of rock joint sets. The water pressure is introduced as fluid pres-
potential instability. At the time of construction (1965–1969), sure boundary condition along the discontinuities, and the
this slope was reinforced with rock anchors. Subsequent rock reinforcement is modeled as structural elements working
hydrogeological study and groundwater monitoring revealed across the discontinuities. The dam reaction forces, derived
the presence of water pressure due to a slight seepage flow from a separate finite element analysis, are evaluated for their
through the rock joints in the dam foundation. This lat- possible effects on the stability. The model examines the sen-
ter evidence raised an additional concern about the stability sitivity of the abutment stability to the presence of joint water
of the abutment. In a dedicated study, the stability of the pressure and evaluates the improving effects of rock reinforce-
right abutment the Hongrin north dam abutment has been ment. The results of the analysis allow achieving an enhanced
assessed using continuum-discontinuum numerical analysis. understanding of potential failure mechanisms and helps in
3DEC (3-Dimensional Distinct Element Code) has been used proposing further suitable measures to improve the stability
to model the complicated slope geometry and to explore the of the abutment.
248
K. Takahashi
Suimon-Gijyutu Consultant, Co. Ltd., Chiba, Japan
For the construction of underground facilities and dams, the and advection-dispersion equations using 2D finite element
control and sealing groundwater is one of the key issues. method (FEM) assuming the grout is Newtonian fluid. In the
Grouting (grout injection) is one of the commonly used meth- numerical model, the time dependent physical properties of
ods to seal groundwater during the constructions. Grout is the grout (increase of viscosity) and the reduction of porosity in
mixture of water and cement with a certain mixing rate and is the porous media were considered to express the reduction
injected into the rock fractures from the injection boreholes of hydraulic conductivity during grout injection (improving
drilled into rock masses in order to improve hydraulic con- hydraulic conductivity field). In the numerical simulations,
ductivity fields of target rock masses. Recently, underground two different grout injection methods mentioned above were
facilities have been built under various geological conditions simulated using the in-situ grout injection data including the
with limited budget, and for grouting system, more effec- change of the grout injection pressure and grout density. The
tive and economical injection method has been required. For simulation results were compared with the ones obtained from
grout injection, so far the injection method which changes the in-situ grout injection tests. The simulation results clearly
the water/cement mixing ratio (W/C) stepwise from high to showed that the change of the grout viscosity as well as the
low has been commonly used in order to inject cement into decrease of porosity in time played important roles for the
the fractured rock masses effectively. Unlike this conventional grout injection mechanism. The simulation results also agreed
method, in the newly developed method, the W/C is changed well with the ones obtained from the in-situ grout injection
from high to low continuously. This method is more effective tests. Figure 1 shows the comparison of hydraulic conductivity
injection method and starts being applied to the construction of field after 18, 30 and 180 minutes between the two injection
various underground facilities. In this study, to investigate the methods. The newly developed injection method was more
grout injection mechanism and processes, the numerical sim- effective and was able to inject more cement in shorter time
ulations were carried out by solving unsteady seepage flow comparing with the conventional injection method.
Figure 1. The distribution of hydraulic conductivity after 18, 30 and 180 minutes, a) conventional injection method and b) newly developed
injection method.
249
250
251
L. Li
SINTEF Petroleum Research, Trondheim, Norway
1 INTRODUCTION
Figure 2. Triaxial compression test results: E50 and ν50 vs. confin-
ing stress.
252
253
J.-S. Lin
U.S. Department of Energy, National Energy Technology Laboratory, Pittsburgh PA, USA
Department of Civil and Environmental Engineering, University of Pittsburgh, Pittsburgh PA, USA
Subject: Rock material and rock mass property testing (laboratory and in situ)
Rock cutting is at the core of all construction on rocks or into successfully. In the discrete element analysis, assemblies of
rocks. A cutting action, by its nature, is defined by causing circular disk particles, in 2-D, and spherical particles, in 3-D,
failure in a rock so that part of the rock can be removed. For a were created following the grain size distribution of a par-
practical application the numerical solution is often the only ticular rock. The assemblies were then packed under a preset
credible approach. From modeling perspectives, rock cutting level of compressive stress and bonded together afterwards at
constitutes a sequence of difficult problems: A contact prob- the points of contact between particles. The micro-parameters
lem first arises as a cutter advances and interacts with a target were calibrated with laboratory mechanical tests.
rock. This is followed by the problem of determining when We have found both the discrete element and the finite
and if the rock would fail. If the rock does fail, a modeler element methods were capable of providing cutting force
is subsequently faced with the problem of how to initiate the estimates and capturing the important characteristics of the
fragmentation process. Moreover, this sequence of problems fragmentation process. One of the major remaining issues is:
repeats itself each step of the way. how to select a velocity in the modeling that is a proper refec-
In this study, we employed both the finite element method tion of the actual cutter velocity? This question can only be
and the discrete element method for modeling rock cutting. resolved through calibration provided sufficient test data are
Our immediate goals were to develop a framework within the available. Our study also showed that particle crushing could
two methods so that a laboratory rock scratch test could be have important impacts both on the cutting forces and fail-
well modeled. The reasons for the focus were twofold: first, a ure modes. Again, the answers require the aid of a carefully
scratch test possessed all essential characteristics of a general planned experiment program. For drilling applications, the
rock cutting problem; second, there were test data available logical next steps of rock cutting research include incorpo-
for validation. We used LS-DYNA for the finite element anal- rating high pressure effects and addressing rock-fluid-cutter
ysis, PFC2D and PFC3D for the discrete element analysis. interaction problems. These types of research will have appli-
In the finite element analysis, the Continuous Surface Cap cations over a wide range of problems beyond drilling. A
Model, a visco-elastic-plastic damage model, was the mate- research on the modeling of complete drill bit mechanics is
rial model we used. After substantial tuning, this material also desirable as the experiences should be readily applicable
model together with element erosion that removed elements to other areas such as modeling the mechanics of a complete
with excessive damages, modeled rock cutting fragmentation tunnel boring machine.
254
X.D. Ma
Shandong University, Shanghai, China
R.J. Rodriguez
ETSI Caminos, C. y P. Technical University of Madrid, Spain
255
256
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 INTRODUCTION
257
The structure in a rock mass can occur at several scales: Figure 3. Effect of fracture intensity and fracture orientation on
(1) centimetre scale where the micro-defects are healed with confined rock mass strength.
infilling that is usually weaker than the intact rock and (2)
metre-scale where the fractures generally lack infilling and are 4 CONCLUSIONS
classed as open. In the synthetic rock mass, the rock fractures
are described using the geometrical descriptions associated Laboratory results were used in conjunction with physical
with the Discrete Fracture Network (DFN). The fracture inten- and numerical model results to examine (1) the strength of
sity description in a DFN is given by: P10 = No of fractures the intact rock and (2) the strength of a fractured rock mass.
per unit length, #/m, P21 = Trace length of fractures per unit The findings suggest that scale effects in intact rock are lim-
area, m/m2 , P32 = Area of fractures per unit volume, m2 /m3 . ited to 80% of the standard laboratory uniaxial compressive
The rock mass at the Äspö Hard Rock Laboratory (HRL), strength. The physical and numerical modelling results sug-
Sweden, is typical of the granitic rock masses found in the gest that the rock mass strength decreases as the fracture
Scandinavian Shield. It contains several joint sets that vary in intensity, expressed as fracture area per unit volume, increases.
both trace length and fracture frequency. A discrete fracture However, the lower limit to this decrease is controlled by the
network model was developed for the Äspö HRL and the three fracture orientation and persistence, regardless of the fracture
fracture sets with their fracture intensity values (expressed as intensity. Hence fracture intensity alone may not be an ade-
P32 ) are provided in Figure 2. quate parameter when estimating rock mass strength using
Park et al. (2004) used the Particle Flow Code to examine discrete fracture networks.
the effect of fractures on rock mass strength. In their mod-
elling the fracture sets were incrementally added to assess the
effect of each fracture set on rock mass strength. The results REFERENCES
from their modelling are provided in Figure 3 as a function
of fracture intensity, and show that as the fracture intensity Blanks, R. and McNamara, C. 1935. Mass concrete tests in large
increases the rock mass strength decreases. If the rock mass cylinders. ACI Journal Proceedings, 31(1):280–303.
was intact the expected strength using the results from Figure Brown, E. T. and Trollope, D. H. 1970. Strength of a model of jointed
rock. ASCE Journal Soil Mechanics and Foundations Division,
1 would be 80% of the peak strength while the a lower bound
96(2):685–704.
strength would be that provided by an optimally oriented sin- Einstein, H. H., Nelson, R. A., Bruhn, R. W., and Hirschfeld, R. C.
gle through-going fracture as shown in Figure 3. These limits 1969. Model studies of jointed-rock behaviour. In Proc. 11th U.S.
appear to adequately bound the rock mass strength determined Symposium on Rock Mechanics, Berkeley, pp. 83–103.
by Park et al. (2004). Hoek, E. and Brown, E. T. 1980. Underground Excavations in Rock.
Brown and Trollope (1970) and Einstein et al. (1969) con- The Institution of Mining and Metallurgy, London.
ducted a number of physical model tests to establish the effect Jackson, R. and Lau, J. S. O. 1990. The effect of specimen size on
of fractures on rock mass strength. Their tests were carried the mechanical properties of Lac du Bonnet grey granite. In Proc.
out using a plaster model with continuous fracture planes. The 1st. Int. Workshop on Scale Effects in Rock Masses, Loen, Norway,
results from their model studies are also shown in Figure 3. pp. 165–174.
Park, E-S., Martin, C. D., and Christiansson, R. 2004. Simula-
The fracture intensity in Figure 3 is expressed as fracture area
tion of the mechanical behavior of discontinuous rock masses
per unit volume. For the model tests the volume of each test is using a bonded-particle model. In CD-ROM Proc. 6th NARMS04,
considered the unit volume. The results in Figure 3 show that Houston, pp. 1–8. NARMS 04–480.
regardless of the fracture intensity, the persistence of a single Pierce, M., Mas Ivars, D., Cundall, P. A., and Potyondy, D. O. 2007. A
fracture can reduce the strength of the rock mass to its lower synthetic rock mass model for jointed rock. Proc. 1st Canada-US
bound strength. Rock Mechanics Symposium, Vancouver, vol. 1, pp. 341–349.
258
Keywords: case studies, lab testing, mining, numerical modeling, rock properties, rock stress
Today many mines are built in higher depths to prospect for the tensile strength and the abrasivity (e.g. Cerchar abrasivity
needed resources and obtaining them. Not only mines but also index, LCPC abrasivity coefficient, petrographic composition
many construction and tunneling projects like the Gotthard with thin section analysis) were determined. In addition to the
base tunnel or the Brenner base tunnel advances into these laboratory testing the performance and the wear data are col-
deep laying rock zones where typical high stress phenomena lected directly in the mines and will be generated in additional
like rock bursts can occur. Higher stress can also lead to prob- drilling tests. For the evaluation and the interpretation all gath-
lems like high costs due to high wear or less performance of the ered data will be combined in a MySQL database and will be
rock cutting tools. Due to the complexity of the measurement analyzed statistically using Matlab (The MathWorks).
methods the stress rates were seldom directly quantified. In Three 2D cross sections of a regular stope geometry were
most cases the stress is only determined by numerical models. modeled and calculated with Phase2 (Rocscience) to define
This paper and the associated research project deals with the the occurring stress and its distribution. For a detailed analysis
influence of these high stress rates on wear and performance further complex 3-D models (Flac 3D, Itasca) are calculated.
of rotary hammer drills in deep-seated mines in South Africa The previous collected field data will be compared to the cal-
up to a depth of 3900 m. The encountered rocks are high abra- culated stress data from the numerical models to benchmark
sive quartzites with high uniaxial compressive strength. These the influence of stress on wear and performance of the used
rocks were mined to reach the decimeter thick gold bearing tools. The aim of this research work is to take the advantage
layers – the so called reefs. of these effects to increase the efficiency and to optimize the
The most common geomechanic properties including the mining progress.
rock strength like uniaxial compressive strength or brazilian
259
Numerical modeling for the evaluation of grout penetration in fractured rock masses
Y. Mito
Department of Urban Management, Graduate School of Engineering, Kyoto University
M. Ashraf
School of Civil Engineering, Universiti Sains Malaysia
T. Yamamoto
Kajima Corporation
Grouting aims to reduce the permeability of rock mass below model. The goal achievement probability under the specific
the required value and has been widely used in rock engineer- target permeability is computed with the help of geostatisti-
ing field for a long time. However, the injection procedure cal interpolation method. The applicability of the method is
needs several empirical decision makings and it is rather verified through the several field applications.
difficult to reliably confirm the achievement of improve-
ment, because the grout penetration process is not sufficiently
clarified.
This study proposes a stochastic prediction method of spa-
tial distribution of rock hydraulic properties after grouting in
fractured rock mass by using the grout penetration process
260
Numerical method for determining contact areas of a rock joint under shear load
Keywords: rock joints, roughness, contact area, back analysis, lab testing
261
262
J.J. Peura
VTT Technical Research Centre of Finland, Finland
263
the phase of “a new rock mass properties” calculation it was During the excavation work the rock quality in pilot tunnel
noticed that this method can not be used these rock param- was checked continuously on daily basis by design engineer
eters because the new calculated rock mass properties were and careful observations showed that the rock quality was
only few hundredth parts smaller than the original properties. better than predicted and no fore poling was required. Exten-
Fore pole umbrella-method can be used only if the strength someter measurements showed displacements from 0.2 mm
and the deformation modulus of the rock are very diminutive up to 4 mm depending on the location of the extensometer.
in relation to quantity of the support.
Usage of the soft material in the excavation: The proposed
method is based on softening behaviour of the material within REFERENCES
the excavation. In this method the soft material describes
deformations before the support system is installed using Curran J. H., Hammah R. E & Yaboub T. E. 2003. A two-dimensional
minor Young’s modulus in the excavation area than outside Approach for Designing Tunnel Support in Weak Rock. Proceed-
excavation. In this case calculated results were not reliable ings of the 56th Canadian Geotechnical Conference. October
because in the all other calculations displacements were 2003. Winnipeg, Manitoba, Canada. 6 p.
much smaller. The failure was caused by unsupported tunnel Hoek E. 2003. Numerical Modelling for Shallow Tunnels in Weak
modelling; the multi-element model collapsed and further cal- Rock. www.rocience.com (2007–2008). 12 p.
culations were done by using only 2 phase’s method to achieve Peura J. J. 2008. Likimääräismenetelmät kalliomekaanisessa ele-
the balance. Despite the failure the method itself seems to be menttilaskennassa (Approximate Methods in Numerical Rock
Mechanics Modelling). Master thesis. Helsinki University of
promising. Technology. Espoo, Finland. 114 p.
When the fore poles are formulated in plane in Kuula, H., Remes, H., Somervuori P, Syrjänen P., Hakala M.
2-dimensional space, the results of the displacement are the 2006. Mestarintunnelin rakennettavuusselvitys (Rock mechanical
same than 3-dimensional model. The results are presented in feasibility study). WSP Gridpoint Finland.
Table 1 and 2. Horizontal displacements were insignificant.
264
A. Giraud
Laboratoire LaEGO, ENSG INPL, Vandoeuvre-lès-Nancy, France
D. Hoxha
Institut PRISME, Université d’Orléans, Orléans, France
The long term behavior of underground excavations is a social Creep phenomenon has already been studied for COx-
and economic stake, in particular in the context of nuclear argillite (Gasc-Barbier et al. 2004 for instance), but the
waste storage in deep geological formation. Several experi- association of both creep and hydro-mechanical effects are
mental galleries have been dug in the underground research quite rare in the literacy (Jia et al. 2008). The specific elas-
laboratory (URL) of Bure in the Meuse/Haute-Marne region toviscoplastic model used in this study is the L&K model
in France, where studies are leaded in order to understand the (Kleine 2007): it offers a coupling between instantaneous and
global behavior of the constitutive rock which is the Callovo- delayed behavior, and an influent variation of dilation which
Oxfordian argillite (Cox argillite) (see Jia et al. 2008; Martin governs the volume strains of the material during a solicita-
et al. 2010 for instance). tion. Rock strain is strongly coupled with fluid flow thanks
The purpose of this paper is to establish a modeling with to Biot’s equations (Coussy 1995). Even if taking unsaturated
Code_Aster of one of the Bure laboratory galleries taking into conditions would be possible afterwards, modeling is limited
account non local approach, creep effect and hydromechanical for this study to saturated ones.
coupling in the framework of the mechanics of porous media, Furthermore, another main novelty of this work concerns
and then to compare numerical results with available experi- the coupling between such complex rheological behavior and
mental data. An example of result can be seen in Figure 1: we non local approaches. Among available methods of regular-
compare water pressure changes in a point of the rock mass ization, the second gradient of dilation (Fernandes et al. 2008)
with and without activating the viscosity in the model. is well fitted to dilatant geomaterials. Its aim is to avoid mesh
dependency and numerical localized solutions, and allow the
simulations to be achieved.
REFERENCES
265
266
267
Stochastic model generation for discontinuous rock mass media and numerical analysis
using the NSCD method (case study: Saint-Béat gallery in the southwest of France)
Ali Rafiee
Zanjan University, Iran
Muriel Gasc-Barbier
CETE du Sud Ouest, LRPC de Toulouse, France
Marc Vinches
Alès School of Mines, France
1 EXTENDED ABSTRACT
268
M. Rinne
Aalto University School of Engineering, Espoo, Finland
B. Shen
CSIRO Earth Science and Resource Engineering, Brisbane, Australia
O. Stephansson
Helmholtz Center Potsdam, GFZ German Research Centre for Geosciences, Potsdam, Germany
Keywords: numerical modelling, hydraulic fractures, fluid flow, rock failure, geothermal
269
5 APPLICATION EXAMPLES
270
Keywords: Meshless Methods, Element Free Galerkin Method (EFGM), Rock Fracture Mechanics, Three point bending
(3PB) load test
1 INTRODUCTION
271
3 CONCLUSIONS
Figure 4. Variation of crack initiation angle θ0 with the crack angle β
under uniaxial tension (Experimental results, BEM and EFGM). Meshless EFG method has shown a great promise in rock
mechanics and rock engineering applications due to their flex-
ibility in solving those problems with discontinuities within
Method (BEM) with the same domain geometry and bound- fracture mechanics and stress analysis. The evaluated frac-
ary conditions. The angle of fracture initiation in different ture growth trajectory using EFG developed code in uniaxial
crack orientation angle in uniaxial tensional test is calculated tensional test and also the three point bending (3BP) test
and is depicted in Figure 4. Regarding to this figure, the cal- demonstrated the ability of the code for solving fracture prop-
culated values of initiation angles are closer to experimental agation problems in geomechanics. As a future research, this
results than the numerical boundary element method results. code can be extended to the anisotropic materials which con-
For example, in the case of β = 30◦ , BEM has about 5% error tain numerous discontinuities like faults and joints within a
and EFG code has about 1% error relative to the experimental rock mass to make it more applicable in rock engineering
data. practical problems.
272
G.H. Shi
DDA Company, Belmont, California, USA
273
274
The analysis of random stress-strain fields and the rock struc- of crack simulation. The integral equations and solutions
ture reliability are important in the rock engineering design. are established by the partial derivation of the original inte-
Since the joint systems of rock mass are always with uncer- gral equations and solutions. And the fundamental solutions
tainties, it is very difficult to solve the random stress-strain were derived and implemented to a computer program. For
field of a rock structure in a jointed rock mass. Stochas- calibration, an example problem was analyzed by the devel-
tic finite element methods have been applied to solve the oped model, and the results were compared with those from
problems of random stress-strain fields. However, boundary Monte Carlo analysis. The results show that developed model
element methods provide alternatives with the advantages can accurately analyze fracture reliability problems. And the
including efficiency and simplicity. In this study, a stochastic developed model is much more efficient than the Monte Carlo
boundary element was developed for the analysis of fracture analysis approach. The developed stochastic boundary ele-
reliability. Displacement discontinuity method, an indirect ment method can be applied to the reliability analysis of rock
boundary element method, was adopted due to its suitability structures.
275
B. Shen
CSIRO Earth Science and Resource Engineering, Queensland, Australia
M. Rinne
Fracom Ltd, Kirkkonummi, Finland
1 INTRODUCTION under thorough research for many years, but there are still
some uncertainties, related to the in situ stress and to the rock
1.1 Fracture propagation code (FRACOD) spalling strength. (Siren et al. 2011a).
To answer these questions, an in situ experiment called
Most numerical methods for rock engineering are based on
Posiva’s Olkiluoto Spalling Experiment (POSE) was started
methods that cannot genuinely describe the fracture initiation
(Aalto et al. 2009). The objective of POSE is to establish the
and propagation. Fracture propagation code (FRACOD) is a
in situ spalling strength of the rock in Olkiluoto and also to
two-dimensional Displacement Discontinuity Method (DDM)
establish the state of in situ stress at the −345 metre depth
computer code that was designed to simulate fracture initiation
level.
and propagation.
The code development was started in early nineties by
Shen & Stephansson (1993). The FRACOD code is based on
the principles of the Boundary Element Method (BEM). Since 2 DEVEPLOMENT OF ANISOTROPY OF
the BEM doesn’t cope with problems with a fracture – that ROCK MEDIUM
is, two surfaces coinciding with each other – Crouch (1976)
developed a method called the Displacement Discontinuity In FRACOD, the fracture initiation occurs when the combi-
Method (DDM). nation of two principal stresses reaches a critical value. More
FRACOD uses a modified G-criterion, which is also known closely, the tensile and shear stresses and strengths are used to
as the F-criterion, introduced by Shen & Stephansson (1993). determine the initiation of a new fracture. The fracture prop-
The problem with the original G-criterion is that, in some agation is, however, determined by using fracture toughness
cases, it predicts shear failure when the failure, in fact, is parameters.
tensile. However, the F-criterion can predict the propagation To take anisotropy of the rock into account, the parame-
of a crack system in both mode I and mode II, which is crucial ters (σt , c, φ, KIC , KIIC ) have an elliptical variation from θ to
for such material as rock in which shear and tensile failures θ + 90◦ . For each point, calculations are done in all anisotropy
are common. directions, and the fracture will initiate in the direction of
The latest development, described in later chapters, intro- maximum tension or shearing, and will propagate in the
duced in the code allows the possibility to simulate anisotropy direction where the maximum F value is reached.
of rock medium using strength anisotropy instead of just
explicit joints and bedding planes. Anisotropy related to new
fracture initiation is described by direction dependent Mohr- 3 CASE EXAMPLE OF POSE
Coulomb and direct tensile strength criteria.The fracture prop-
agation function is converted to anisotropic by formulating The input parameters for the case example are determined
F-criterion to be direction dependent. by using existing test results for pegmatitic rock (PGR),
which is assumed to be isotropic, and for migmatitic gneiss
(MIGN.GN), which is assumed to be anisotropic. Both
1.2 Posiva’s Olkiluoto Spalling Experiment (POSE) types are assumed to be homogeneous and linearly elastic.
For migmatitic gneiss, individual rock strength and fracture
Currently in Olkiluoto, the construction of the underground strength parameters are determined for two perpendicular
rock characterisation facility for the final disposal of spent foliation directions. The input parameters, as listed in Table 1,
nuclear fuel named ONKALO is on the way. The site has been are used.
276
277
G.F. Zhao
School of Civil and Environmental Engineering, The University of New South Wales, Sydney, Australia
In this paper, methods and technology of contact description Contact detection is an important issue which is the prereq-
are shortly reviewed. A new spherical contact model using uisite to perform the contact response. Actually detection of
numerical manifold method is presented to describe the col- contact between arbitrary shapes is very difficult and it is still
lision between two balls. This algorithm is based on penalty the puzzle in modern applied mathematics. In modern con-
theory and open-close iteration. The test results show that it is tact detection algorithm, contacts are detected in two ways: a
capable to simulate both elastic and inelastic collision. With post-detection (after the collision occurs) or a pre-detection
the given maximum interpenetration rate and coefficient of (before the collision occurs).
restitution, the penalty number can be chosen wisely. The benefits of the pre-detection algorithm are increased
Analytical contact theory dates back to 1881 by Heinrich fidelity and stability. But it needs to integrate the physical
Hertz who tried to solve the problem involving contact simulation algorithm with the detection algorithm and this is
between two elastic bodies with curved surfaces. In the very difficult. The main benefits of the post-detection method
next one hundred years, the framework of contact mechan- are as follows. In this case, the contact detection algorithm
ics was established step by step, for example, the well-known need not be aware of the myriad physical variables; a sim-
DMT model and JKR model which are used to represent ple list of physical bodies is fed to the algorithm, and the
adhesive elastic contact. For numerical implementation, uni- program returns a list of intersecting bodies. This means
lateral constraints are added to restrict the motion of bodies the contact detection algorithm does not need to understand
but this process is challenging especially including stress friction, elastic contact and even motion and deformations. In
analysis. addition, the post algorithm is in effect one dimension simpler
The second model is fully based on physical laws (e.g. than the priori algorithms. The pre algorithm must deal with
Newton’s and Coulomb’s) and is mainly used in computa- the time variable, which is absent from the post algorithm. On
tional graphics. To achieve high efficiency, this model usually the other hand, the post algorithm causes problems in the fixed
deals with rigid bodies and gives less accuracy. There are time step, where intersections, which may not be physically
mainly impulse-based method and constrain-based method. correct, need to be corrected.
The former models all forces between bodies through a series Open-close iteration algorithm was proposed by Shi (1991).
of impulses and it cannot completely handle a prolonged static One of its advantages is to overcome the disadvantage of the
contact. The latter works by applying a series of constrains on post-detection algorithm. This algorithm adopts an iteration to
rigid bodies but its simulation is still plausible. check the fixing positions of bodies. And this guarantees that
The so-called parameterized model is widely used in energy all the contact responses really exist during the whole time
principle based numerical methods (e.g. FEM and Manifold step.
Method). There are mainly two ways to parameterize the Using the penalty method and open-close iteration algo-
contact: Lagrange multiplier method and penalty method. rithm, a spherical contact model is presented. This model
The former introduces Lagrange multiplier λ into the energy adopts the numerical manifold method (NMM) to establish
system to describe the reaction force. The new variables the system equation.
are treated as unknowns which are solved together with For a 3D problem, 12 unknown variables are used to
displacements. The latter minimizes the potential energy with describe the motion and deformation of a body. They are trans-
an additional penalty term. Although penalty method does lation (u0 , v0 , w0 ), rotation (θx , θy , θz ), normal strain (εx , εy , εz )
not require additional independent variables, the parameter, and tangential strain (γxy , γyz , γzx ). So the displacement of an
penalty number α, needs to be defined by users and it influ- arbitrary point in a body is:
ences the simulation a lot. These contact models above have
different applicability, accuracy and efficiency.
The choice of contact models is a compositive considera-
tion depending on computational scheme, quantity of poten- The dynamic system equation without damping is:
tial contacts, simulation time, material properties, required
accuracy, etc.
278
279
Takumi Onuma
JGI, Inc., Tokyo, Japan
Tsukasa Kumagai
JGC Corporation, Yokohama, Japan
Jonny Rutqvist
Lawrence Berkeley National Laboratory, Berkeley, USA
Keywords: numerical modelling, fluid flow, case studies, monitoring, field measurements
CCS (carbon dioxide capture and storage) is a promising TOUGH-FLAC (Rutqvist et al, 2002), that links the mul-
approach for reducing the greenhouse gas content in the tiphase fluid flow simulator TOUGH2 with a commercial
atmosphere, through capturing carbon dioxide (CO2 ) from geomechanical code FLAC3D . As numerical grid of the three
large emission sources and injecting it into reservoirs (such dimensional region of 10 km × 10 km wide and 4 km deep
as deep saline aquifers). Large-scale storage projects will around the horizontal injection well KB-501 was used. The
likely involve very long-term storage of huge amounts of geological model and properties basically follow Rutqvist
CO2 , potentially exceeding hundreds of millions of tonnes et al. (2010). A sensitivity study was performed for investigat-
(Mt). Effective site monitoring is crucial for demonstrating ing the effect of model parameters on surface deformation pat-
the effectiveness and safety of a site, calibrating predictive terns, and for investigating how coupled fluid-geomechanical
models, as well as providing warning and identify measures models can be calibrated in the light of surface-deformation
of any potential surface leakages (Chadwick, 2010). temporal and spatial changes detected in detail by DInSAR.
Differential interferometry SAR (DInSAR) is a promising As the results, the coupled fluid flow and geomechanical
remote sensing technique for detecting surface deformation on model successfully simulated the temporal and spatial surface
the order of millimeters. The technique has been applied to the deformation, and illustrated and demonstrated that:
CO2 injection site at the In Salah CO2 storage project, Algeria,
1) As expected, information on the CO2 flow is indirectly
and successfully demonstrated its usefulness and effectiveness
transferred to the surface displacement through pressure
for measuring, in remarkable detail, spatial surface defor-
propagations. Even if the CO2 flow were highly localized
mations due to pressurization of the reservoir during CO2
along with reservoir heterogeneities such as faults, surface
injection (Onuma et al., 2010). The detected surface deforma-
deformation would not be sharp but smoothened through
tion patterns are anisotropic, likely due to fluid-geomechanical
the overburden.
effects of fractures and faults within the site.
2) Temporal and spatial changes in surface deformation data
In the In Salah project, since August 2004, CO2 removed
acquired by DInSAR potentially provide a unique opportu-
from natural gas production has been injected from three injec-
nity for inferring reservoir spatial permeabilities, including
tion wells into the storage reservoir formation, at a depth of
anomaly features such as hydraulic compartments (low-
about 1.8 km and at a rate of nearly 1 million tonnes per/year.
permeability barriers), which are important economic
The surface deformation detected by a DInSAR (Figure 1)
performance factors for reservoirs storing CO2 .
shows that the uplifts occur near the injection well with a
surface uplift rate up to 5 mm/year.
For simulating surface deformation due to CO2 injec-
tion, we use a coupled fluid-geomechanical simulator,
280
J. Zhao
Laboratory of Rock Mechanics, Ecole Polytechnique Federale de Lausanne (EPFL), Lausanne, Switzerland
281
5 CONCLUISION
REFERENCES
282
H.J. Zhao, F.S. Ma, Y.M. Zhang, J. Guo & A.H. Wei
Key Laboratory of Engineering Geomechanics, Institute of Geology and Geophysics, China Academy of Sciences, Beijing, P.R. China
ABSTRACT: The geo-stress field of rock mass is a complex of which boundary conditions are established, the state of the
system affected mutually by various factors (such as topo- back-calculated initial stress field should correspond to that
graphy, lithology, and geologic structures). Currently, there of the practical stress field. In fact, no matter which kind of
are many kinds of in situ stress measurement methods, but the boundary conditions is used, the state of initial stress should
precise description of the geo-stress field remains difficult. be the same as that of the actual stress state, horizontal or
The initial geo-stress field is the basic factor and important nearly horizontal. However, previous studies have shown that
basis for the study of rock mechanics problems. To inverse the simulation results always vary considerably under different
initial stress field of rock mass reasonably is an important issue boundary conditions. If fixed displacement boundary condi-
in rock mass mechanics.The degree of approximation between tion was applied to both sides, the simulated horizontal stress
the inversed initial geo-stress field and the actual stress field contours are horizontal and smooth. This stress field is consis-
is one of basic conditions for assessing the validity of the tent with the expected distribution of the stress field. However,
numerical simulation. Especially excavating in high tectonic if stress boundary condition was applied to both sides, the dis-
stress areas, the correctness of the initial geo-stress obtained tribution contours of horizontal stresses are irregular. Aside
by numerical simulation directly influences the final results. from the horizontal distribution trend near the lateral borders,
The boundary conditions of the continuous mechanic large fluctuations can still be observed in the inner model.
models are always divided into 3 types: (1) stress boundary These simulated stress fields highly differ from the theoret-
condition, which means surface force or concentrated force ical stress field. Therefore, boundary conditions are optimal
of the body surface is given, and the force can be variable or for initial geo-stress field inversion.
not; (2) displacement boundary condition, which means dis- To study the special effects of high tectonic stress on exca-
placement component of the boundary points is known, it can vating works, a typical mining-induced ground movement
be a constant, equivalent to zero or equivalent to zero in some caused by underground mining in Jinchuan No. 2 mine is
directions. When the displacement in some direction is zero, carried out. Based on the results of years GPS monitoring
it is what is called constraint boundary condition; (3) mixed data on the ground surface, inverse analysis of mining-induced
boundary condition, which means the displacement is known rock mass movement characteristics under high tectonic stress
in some directions, and the boundary force is given in the other was studied, and a comparative analysis of the surface move-
directions. ment, deformation characteristics were given out. The results
During numerical simulation, the reasonable initial geo- indicated that the back calculated method can achieve reason-
stress field is usually acquired by regression fitting method able and reliable initial geo-stress field. Because of different
using limited on terrain, geomorphology and observed geo- mechanisms of rock mass movement in high tectonic stress
static stress as base. For the underground excavation in field, the mining-induced ground movement has distinctive
geologic body with high horizontal tectonic stress, to simulate characteristics of ground movement and deformation in the
the actual geo-stress state rationally is the key of numerical following aspects.
simulation.
To discuss the inversion of initial geo-stress field, suppose a) The volume of subsidence trough is much smaller than
there is a homogeneous geologic body with high-level tectonic the volume of the underground goaf. The slow subsidence
stress, in which the direction of the maximum principal stress ratio will not result in a sudden collapse.
is horizontal. Then, the distribution of the initial tectonic stress b) The scope of ground movement, deformation and damage
contours should be horizontal under the influence of tectonic expand distinctively outward of the subsidence center. The
stress field. ground damage types vary with different places. Internal
In numerical simulation, if the stress boundary condition region of the basin, it shows shear dislocation scarp dam-
is set, additional stresses are applied to lateral boundaries to age, which corresponds to the mining-induced subsidence
simulate the corresponding tectonic stresses. During numer- and deformation caused by its self-weight stress. However,
ical simulation, additional stresses are constant, whereas the in the outside region of the basin, it is mainly tensile fail-
boundaries are moveable during excavation. If the displace- ure, which corresponds to the subsidence and tension effect
ment boundary condition is set, tectonic stresses can be under the tectonic stress field.
assigned as internal stresses in the calculated area. Regardless
283
284
V.M. Zhigalkin, V.N. Semyonov, O.M. Usoltseva, P.A. Tsoi, A.I. Chanyshev & I.L. Abdulin
Siberian Branch of Russian Academy of Sciences, Institute of Mining, Novosibirsk, Russia
285
286
Multiple scale numerical modeling of growth process of multiple cracks in rock mass
Rock mass is neither continuous material nor totally discrete a process of crack propagation for quasi brittle material under
one, it’s a kind of defect material which contains many cracks. tensile stress by Wells. A process of crack propagation under
The nonlinear deformation behavior of rock mass is induced tensile stress is obtained by Du, and the feasibility of this
by the unstable propagation of cracks which existed in the rock method is verified too. The relationship among extended
mass under external loads. Therefore, it is key important for finite element method, meshless method and traditional finite
designing rock engineering to study the propagation process element method are summarized by Jin.
of cracks existed in rock mass under compressive stress. It is well known that there are many joints, cracks and
They are crucial problems for complex system that the phe- faults in rock mass. The dimensions of these discontinuous
nomenon of cross-scale and the coupling of multiple scale. are varied from meters to millimeters. In the engineering prac-
Moreover, It is pointed that they mirror the intrinsic behavior tices, the failure of surrounding rock mass mostly related to
of material. These crucial problems have rich contents. For the interaction of discontinuous interface with varied dimen-
rock mass, it contains many discontinuous interfaces (such as sions. However, it is lack of effective numerical simulation
joints, cracks and faults), and these discontinuous interface method for crack-weakened rock mass which can take the
varied in dimension. In addition, the deformation behavior multiple scale mechanical behavior or rock mass into account.
of rock mass is proved time-dependent and temperature- In order to solve this troublesome, some breakthroughs should
dependent too. So, the behavior of rock mass is multiple be made in the fundamental theory field of numerical simu-
scale problem. In this paper, only the dimension aspect of lating method for crack-weakened rock mass. On the basis
cross-scale problem is discussed. For example, the interaction of extended finite element method, a multiple scale model for
problem between microcracks and marco faults are usually simulating the damage evolution in crack weakened rock mass
observed in rock engineering. However, no effective modeling is proposed. The coordination between the internal bound-
method for solving this problem has been proposed till now. ary (such as cracks and holes) and mesh is not needed in
It is nearly impossible for traditional numerical method which this multiple scale model. By combining the displacement
needs mesh for calculation to modeling multiscale cracks projecting method and displacement loading method, the con-
existed in rock mass. nection between different sizes of mesh behave smoothly. This
There are some numerical methods for modeling crack multiple scale model indeed surmounts difficulties in mesh-
growth which remeshing is not needed, such as mani- ing and it save the storage space. The contact constraint on
fold method, meshless method and extended finite ele- crack surface is embedded within the total stiffness matrix by
ment method. Among these, extended finite element method the penalty method, in which the load vectors are modified
(XFEM), which can be used to solve discontinuous mechan- too. The path of crack growth and stress field are determined
ical problems, is proposed by Belytschko et al. in the year of through iterative computations. Because additional discontin-
1999. It is proved that the extended finite element method is uous functions and enriched tip elements are added in the
effective in modeling discontinuous problem, such as growth displacement field, the geometry of cracks is independent of
of interface, propagation of crack and complex fluid mate- the finite element mesh. As a result, remeshing is not neces-
rial. There are three main advantages for the extended element sary to model crack propagation by using this multiple scale
method. Firstly, the details of substructure are neglected in this model. Because the frictional contact of crack surface is taken
method, and the mesh for calculating are generated according into account, the present method is fit for modeling the growth
to the geometry shape of structure. Secondly, in order to track of multiple cracks in geomaterials under compressive loads.
the growth of cracks, other methods such as level set method Finally, the method is employed to simulate multiple crack-
are adopted. Then, the location of crack is accurately deter- growth when frictional contact exists on the crack surfaces,
mined. Finally, in order to model the propagation of cracks, and the results show good agreement with the experimental
only the shape function of element is modified. Therefore, the ones.
matrix of total stiffness, which is sparse and zonal, has the The paper is organized as follows. The model problem for
same advantages as traditional finite element method. an elastic body with microcracks with varied dimensions is
At present, the method of extended finite element is adopted formulated in Sect. 2. In Sect. 3 the total stiffness matrix is
in solving the propagation problem of single crack under obtained. The asymptotic method for calculating stress inten-
axial tensile stress. For example, a object-oriented program sity factor in the crack tip is introduced in Sect. 4, and the
for extended finite element method is compiled by Stephane fracture criterion is also adopted in this section. Numerical
in C++ language. By using extended finite element method, results are presented in Sect. 5.
287
F.L. Pellet
INSA – University of Lyon, Villeurbanne, France
The French national radioactive waste management Agency possibility that shearing might develop at the interface due to
(ANDRA) studies the feasibility of geological repository the difference in the thermal expansion coefficients of steel
for radioactive wastes. Since 2000 it has been engaged and rock. It should also be noted that during placement of the
in the development of the underground laboratory of steel lining in the micro tunnel, there was no perfect contact
Meuse Haute Marne to study the properties of a claystone, between the lining and the claystone. Therefore normal stress
called “Callovian-Oxfordian argillite”. At the Underground on the steel lining will vary over time. However, despite the low
Research Laboratory, the claystone layer lies between 420 permeability of the rock, water will appear at the wall of the
and 550 m below ground and is the target layer for the repos- micro tunnel, which means that the interface could become
itory of high-activity and long-life (HA) radioactive wastes. more or less saturated. Therefore, it is necessary to under-
Along the main gallery, smaller horizontal blind micro-tunnels stand the mechanical behavior of the steel–claystone interface.
were excavated. These subsidiary micro tunnels are approxi- This research was an initial attempt to study the properties of
mately 40 m long and 700 mm in diameter lined with an about friction and dilatancy in this type of interface at room temper-
25 mm thick steel lining along the entire length, and designed ature. Several tests on claystone-steel interfaces in both dry
to accommodate waste canister. (meaning at the water content of core sample which may be
In this study, direct shear tests were carried out using a slightly desaturated) and wet (meaning after water injection at
prototype machine that was developed in the 3S-R labora- the interface) were performed for this study. Based on the test
tory, Grenoble, and named BCR3D. This machine operates results, the following conclusions can be drawn:
with five electro-mechanical servo controlled jack and was
1. The maximum normal stress for studying the behavior of
designed to measure the hydro mechanical properties of nat-
the claystone is 12 MPa. Above this stress level, the prop-
ural or artificial joints/fractures as well as the interfaces
erties of claystone degrade due to the presence of crushed
between different materials with 3-D stresses (two orthogo-
material and removed particles.
nal shear directions and normal compression). This machine
2. The presence of water alters some of the claystone proper-
can also be used to create a hydraulic conductivity interface
ties and this inevitably affects the mechanical behavior of
by injecting a fluid at the center of the specimen, creating a
the interface. We have highlighted the fact that the shear
radial flow at the interface. Symmetrical movement of the two
resistance of the “wet” interface is significantly less than
parts of the shear box during testing ensures the quality of
for the “dry” interface. We obtained the same results for
tests. The relative normal and shear displacements (tangential
claystone-claystone interfaces, although these test results
[w] and normal [u]), between the two parts of the shear box
are not presented here.
as well as the stresses (shear τ and normal σn ) are measured
3. For the wet and dry interface tests, the values of shear
by various gauges. It is also possible to measure the flow (q)
stresses in the loading direction are more reliable than the
and injection pressure (p) of fluid or gas during the test.
parameters obtained from the reverse shearing direction.
The concept for high level, long life radioactive wastes is
4. The general behavior of the claystone – steel interface
to place the waste canisters in horizontal steel casing of micro
remains virtually iso-volumetric during shearing, except
tunnels excavated in Callovo Oxfordian claystone. It is impor-
during reverse shearing that causes damage at the edges of
tant to be aware of the Thermo-Hydro-Mechanical behavior
the specimens (about 20% of the surface for “dry” and over
of the steel-claystone interface in order to understand the
30% for “wet” specimens).
long-term modifications that can be expected around these
cells. In particular, wastes are exothermic, and this raises the
291
Subject: Rock material and rock mass property testing (laboratory and in situ)
292
Keywords: transversely isotropic rock, elastic constant, apparent elastic moduli, uniaxial strength, anisotropy ratio
This paper reports the experimental investigation of deforma- planes (E/E ) were determined to be 1.3, 2.1 and 3.4 for
tion and strength anisotropy of Asan gneiss, Boryeong shale Asan gneiss, Boryeong shale and Yeoncheon schist, respec-
and Yeoncheon schist in Korea which show clear evidence tively. The anisotropy ratios of maximum to minimum elastic
of transverse isotropy. For each rock type, three sets of seven modulus (Emax /Emin ) were 1.7, 2.7 and 4.0 for Asan gneiss,
cylindrical specimen cored with angles of 0, 15, 30, 45, 60, 75, Boryeong shale and Yeoncheon schist, respectively, which
and 90 degrees with respect to transversely isotropic planes were higher than E/E due to the fact that the minimum elas-
were obtained by using laboratory directional coring system tic modulus exist when the transverse isotropic plane were
established for this study. Five independent elastic constants inclined with respect to loading direction. The applicability of
of transversely isotropic rocks were determined using various transversely isotropic model to the chosen rocks was quantita-
combinations of strain measurements obtained from speci- tively investigated by comparing the apparent elastic moduli
mens having different coring directions. The anisotropy ratios with the theoretical apparent elastic moduli predicted by tenso-
of elastic modulus parallel to and perpendicular to isotropic rial transformation of compliance matrix. The mean prediction
error (MPE) defined as the average relative difference between
measured and predicted elastic moduli were lower than 17 %,
which indicates that transversely isotropic model is a reason-
able assumption for the rock types chosen in this study. The
mean prediction error becomes lower as more strain measure-
ments were employed and this demonstrates that more strain
measurement can increase the accuracy of the determined
elastic constants even if five independent strain measurement
is theoretically sufficient. The anisotropy ratios of maximum
to minimum uniaxial compressive strength were 2.6, 2.6 and
18.6 for Asan gneiss, Boryeong shale and Yeoncheon schist,
respectively. The anisotropy ratios of maximum to minimum
tensile strength determined by Brazilian Tensile test were 3.2,
2.2 and 7.1 for Asan gneiss, Boryeong shale and Yeoncheon
schist, respectively. The strength anisotropy were greater than
deformation anisotropy for Asan gneiss and Yeoncheon schist
and the opposite were observed for Boryeong shale. The extent
of anisotropy observed in this study is rather significant and
this study reconfirms the importance of anisotropy in rock
Figure 1. Photograph of the tilt coring system. engineering.
Table 1. The five elastic constants of Asan gneiss, Boryeong shale, and Yeoncheon schist in Korea.
293
294
N. DeDontney
Harvard University, Cambridge, MA, USA
Subject: Rock material and rock mass property testing (laboratory and in situ)
All geomechanical models from simple analytical calcula- from production-induced fluid pressure decline and strongly
tions to complex fully-coupled finite-element simulations impacts reservoir performance.
require some knowledge of mechanical properties in order to In this study we introduce a new predictive algorithm
describe the deformation behavior of the stressed rock units. for evaluating the mechanical strength of siliciclastic rocks.
For petroleum-related applications two ubiquitous experi- The relationship is sufficiently generic that it can be applied
mental techniques include triaxial compression testing for throughout sandstone-to-shale sequences and requires only
rock strength measurement and compaction testing under basic lithology and porosity wireline logs as data input. It
uniaxial strain boundary conditions for rock compressibil- is simple and ideally suited for rock strength prediction in
ity measurement. Knowledge of compressive rock strength data-sparse regions where, for example, no acoustic logs or
is essential for modeling near-wellbore deformation while associated core measurements are available for calibration
rock compressibility controls pore volume strains resulting purposes (Fig. 1).
Single element ABAQUS modeling has been used to simu-
late rock compressibility evolution with uniaxial compaction
and the onset of pore collapse, benchmarked against an exten-
sive database of core measurements spanning unconsolidated
sands to tight gas reservoir rocks (Fig. 2). Empirical corre-
lations relating Cam-Clay elastic-plastic material parameters
and other petrophysical properties potentially derivable from
geophysical wireline logs are discussed.
Figure 1. Mohr-Coulomb cohesion and friction angle derived from Figure 2. Modeling impact of yield surface eccentricity on forma-
wireline logs. tion compressibility magnitude following pore collapse.
295
Subject: Rock material and rock mass property testing (laboratory and in situ)
A true triaxial testing device has been developed in the nineties circular section. The range of mean pressures corresponding
in the mining department of the Faculty of Engineering, to a particular shape varies according to the competence of
Mons-Belgium. Since then data have been gathered from sev- the rock. For Sorcy limestone, we also went further by trying
eral rocks, among which various carbonated rocks exhibiting to build the octahedral envelopes on the cap surface, indicat-
different mechanical behaviours. In this work, we compare ing a triangular shape with apexes in the direction of triaxial
the behaviour of three limestones: a compact and strong extension states.
one (Soignies limestone, UCS = 100–120 MPa), a compact These observations can be associated with the evolution of
one with medium strength (Moca limestone, UCS = 60 MPa) the rocks behaviour. In triaxial conditions, perfectly plastic
and a soft and porous one (Sorcy limestone, UCS = 40 MPa). flow was observed for P between 10 and 20 MPa for Sorcy
Figure 1 compares the failure envelopes for those three limestone, between 80 and 90 MPa for Moca limestone and
rocks in the (P, Q) plane. In addition to the meridian sections, is expected beyond 240 MPa for Soignies limestone. For the
some octahedral envelopes are also sketched on this figure. In first two cases at least, strain hardening therefore occurs before
each case, we observe an evolution of octahedral sections from the deviatoric stress attains its maximum value on the meridian
an almost triangular shape to a hexagonal one, and then a quasi section.
296
L.M. Ferreira Gomes, C.R.V. Batalha, P. Gabriel de Almeida, L.J. Andrade Pais & E. Mendes
Department of Civil Engineering and Architecture, University of Beira Interior, Covilhã, Portugal
Subject: Rock material and rock mass property testing (laboratory and in situ)
The investigation zone of Foz-Côa schists, known as the quar- The rocks in study corresponds to slate schist unweathered
ries of Poio is located in Portugal, near the city of Foz-Côa, (W1), coloured from dark grey to black, sometimes striped,
in Guarda’s district (Europe). The rocks exploited in quarries as shown in Figure 2.
integrate the geological unit named Formação de Desejosa, When broken by cleavage S1 frequently presents a black
belonging to the Douro Group of Super-Grupo Dúrico Beirão, color, being very soft to the touch. When broken by cleavage
corresponding to ante-Ordovician rocks. S2 is dark grey and a little rough to the touch. When changed,
The quarry area is located in a syncline, diving ESE, with and in particular on the faces corresponding to fractures, it
layering (So) varying from sub-horizontal, South, to N54◦ E presents chestnuts colorations.
60◦ SE, in the northernmost area. The results of tests of physical indices are presented in
The main schistosity, S1 , is parallel to S0 . The secondary table 1, being noted the low values of porosity and water
schistosity, S2 , is generically directed to N43◦W sub-vertical. content when samples are saturated.
The joints, sometimes parallel, sometimes perpendicular to The slate schist of Foz Côa, in the 20th century were known
S2 , makes the start-up and the transformation of rock easier. with much regional importance, being exploited for the pro-
In Figure 1 are presented typical images of the quarries, where duction of props for alignment and verticality of vineyards.
it can be observed: the foot of the quarry semi-horizontal and
parallel to S0 (and to S1 ); the schistosity of the second phase,
S2 , semi-vertical; and yet the joints perpendicular to S1 and S2 .
γt γd WS WN na
Parameter (kN/m3 ) (kN/m3 ) (%) (%) (%)
297
298
There have been extensive efforts to develop empirical criteria so when the anisotropy is of a mechanical type (induced
by a number of rock mechanics researchers as the theoretical anisotropy).
ones are not adequately capable to demonstrate the response Studies on Ramamurthy criterion reviled that it is possible
of intact or jointed rock masses at a wide range of confining to modify the expression with the help of the term defined as
pressures. Amongst such expressions proposed Ramamurthy’s Anisotropic Effect (Ae ) given as:
criteria is easy to use and yields sufficiently accurate response
for intact and jointed rocks. However this research indi-
cates that this criterion could be further refined to estimate
even more precise compressive strength of anisotropic rocks.
An intensive laboratory investigation was undertaken to test where, σia is the compressive strength of the intact samples
jointed specimen at various discontinuity angle. Fabricated under ‘a’ confining pressure, σja is the compressive strength
artificial samples which include two perpendicular joints were of the anisotropic samples.
prepared by Plaster of Paris. These specimens were used for The anisotropic effect parameter Ae could only vary
simulation of mechanical anisotropy subjected to uniaxial and between 0 to tending 1. Hence a parameter called strength
triaxial tests (Fig. 1). reduction factor was derived as:
It was noticed that Ramamurthy’s criterion can be well
applicable for the anisotropic rocks only when the uniaxial
compressive strength is noticeable. On the other hand deter-
mination of the uniaxial compressive strength of the jointed where, Cj : strength reduction factor, σ1i : compressive strength
rocks in the laboratory or in-situ requires expenses and time. of isotropic samples under the confining pressure of σ3 , σ1j :
Furthermore, preparation of the samples that could describe compressive strength of anisotropic samples under confining
the various anisotropic conditions is a difficult task, more pressure of σ3 .
299
Now, considering factor Cj and the uniaxial compressive It is evident from Figure 2 that the proposed modification
strength of the intact rock, a new relationship could be derived generates a good conformity with the results obtained from
as below: the tests. Even the conformity is more pronounced when the
discontinuity angle is 30◦ or 45◦ . It can be concluded that for
the specimens which were tested for this research work the
proposed modification provides adequate precision.
300
Effects of fracture size and normal stress on the shear behavior of a fracture
as estimated by mortar replica
A.A. Giwelli
Technology & Research Center, JOGMEC, Chiba, Japan
Graduate School of Environmental Studies, Tohoku University, Sendai, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
In this study, the effects of fracture size and normal stress aperture, which can be a measure for the matedness of the frac-
on the shear behavior of a fracture were investigated in the ture surfaces, increases with shear displacement and fracture
direct shear tests by using mortar fracture replicas of three size. Thus, the matedness of the fracture surfaces decreases
different sizes of 200 × 200 × 80 mm (M200), 150 × 150 × with shear displacement and fracture size. Furthermore, the
80 mm (M150) and 100 × 100 × 80 mm (M100), which were SD of the initial aperture also increases with normal stress.
produced from a tensile fracture created in granite. Monotonic This suggests that an increase in normal stress causes more
shear loading was conducted under constant normal stresses damage on the asperities, which decreases the matedness of
of about 0.3 and 1 MPa with the maximum shear displace- the fracture.
ment of about 20 mm. To observe the evolution of surface Figure 2 shows the effect of normal stress on the non-
damage in the shearing process, the fracture surface heights linearity of the closure curve for M100, M150 and M200 at
were measured before and after the experiment, and the mean, shear displacement of 20 mm. Since the matedness decreases
the standard deviation (SD) and the power spectral densi- with normal stress, the non-linearity in the closure curve
ties (PSDs) of both the surface height and the initial aperture increases with normal stress. Figure 3 shows a log-log plot of
were determined at designated shear displacements of 2 and the PSD of the upper surface height (US) at normal stresses of
20 mm. In addition to the shear behavior, the closure curve 0.3 and 1 MPa and shear displacements (δs ) of 0 (initial) and
was measured by applying only normal stress at the desig- 20 mm for M150. This figure indicates that smaller asperities
nated shear displacements to clarify the effect of shearing on with shorter wavelengths were more damaged by shearing and
the non-linearity in the closure curve. that larger asperities with longer wavelengths are damaged as
The SDs of the initial aperture were plotted against shear the normal stress increases. Thus, the non-linearity in the clo-
displacement for different fracture sizes and normal stresses, sure curve during shear increases not only with fracture size,
as shown in Figure 1. It can be seen that the SD of the initial
Figure 1. Relation between shear displacement and the SD of the Figure 2. Effects of normal stress on the closure curve at shear
initial aperture as a function of normal stress and fracture size. displacement of 20 mm.
301
302
Study of the short and long-term brittle behavior of a porous rock with continuous
monitoring of (P- and S-) wave velocities under different hydrous conditions
D. Grgic
LAEGO, Nancy University (INPL-ENSG), Vandoeuvre-lès-Nancy, France
D. Hoxha
Laboratoire PRISME (EA4229), Polytech’ Orléans, Orléans, France
J. Eslami
LAEGO, Nancy University (INPL-ENSG), Vandoeuvre-lès-Nancy, France
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 EXTENDED ABSTRACT
303
304
The addition of water to rock is known to change its mate- A similar ameliorating effect trend was found with the use of
rial properties. These changes can be both beneficial as well a CuSO4 solution but it somewhat less falling midway between
as detrimental in terms of rock strength, stability of rock the results observed for water and for the KCl solution.
excavations and the process of mechanical rock excavation. It is postulated that the underlying mechanism is associated
This paper presents the results of a study to exam what effect with an exchange of cations in solution with those in the clay
certain chemical solutions have in ameliorating the changes minerals which was confirmed by analysis of the concentra-
in material properties brought about after clay bearing rock is tions of elements in solution before and after treatment of the
exposed to water. rock samples. Furthermore, the preferential cation exchange
Two chemical solutions, potassium chloride (KCl) and cop- with potassium and copper ions interferes with absorption of
per sulphate (CuSO4 ), were chosen on the basis that they water molecules and hence reduces swelling of the clay matrix.
would result in a chemical exchange with elements in the clay As well as altering the strength of rock, immersion in a
minerals. chemical solution also changed the mechanical rock cutting
The study confirmed earlier work whereby clay bearing parameters as shown in Figure 2. The cutting and normal
rock when exposed to water weakened the fabric of the mate- forces and specific energy reduced by 59%, 48% and 68%
rial. After immersing rock in distilled water for as little as respectively after immersion in water. The effect of KCl
30 minutes, rock strength as indicated by the average UCS though was found to differ depending on the period of immer-
reduced by nearly 50%. A similar weakening of the rock fab- sion with the longer immersion time of 72 hours resulting
ric was observed using the Slake Durability test and Schmidt in the cutting parameters retreating back towards the levels
Hammer. observed with untreated rock.
Immersion of rock samples in a solution of KCl was found Overall the results of this study indicate there is significant
to counteract the degradation effects of distilled water. As potential to alter the degradation in properties of clay bearing
shown in Figure 1, the UCS was of a similar level being only rocks brought about by exposure to water and hence improve
8% less than the untreated rock. the stability of excavations in rock and rock cutting conditions
with the use of certain chemical solutions.
305
Creep test of tuff for over thirteen years and its know-how
Subject: Rock material and rock mass property testing (laboratory and in situ)
Many researchers have studied creep of rock to understand its Occasionally the air pressure was adjusted and the pneumatic
time dependent behavior. However, important unsolved ques- lines were cleaned. During the second test, the air compressor
tions of long-term creep have still remained. In this study, a and the air conditioner malfunctioned and then were replaced
pneumatic creep testing machine was developed for long-term with new ones. The test had no other troubles and has been
continuous use. In the autumn of 1994, a uniaxial compression conducted for over thirteen years.
creep test was started, and up to now two sets of creep data Figure 1 shows the creep strain and stress from May 23,
have been obtained in two creep tests for over a year. One of 1997 to December 20, 2010 in the second test. From 1997 to
the tests is now under conducted and its period is over thirteen 1999, the final times when the displacement took some prede-
years. In this paper, the know-how and details of long-term termined values were automatically recorded in the computer
creep tests for rock were described. memory. After that the displacement and the air pressure were
In the testing machine, air was supplied from an air com- manually recorded on Monday and Friday every week. Except
pressor to a pressure-reducing valve and then delivered to an for small and large fluctuation probably caused by changes of
air cylinder with constant pressure. Displacement of a speci- temperature and creep stress, the creep strain increased under
men was measured with two cantilever-type sensors. Sample such a low stress along the thick solid line which has the slope
rock was Tage tuff yielded in Tochigi prefecture, Japan. A of 7 × 10−13 /s. From 2008 to 2010 the creep strain decreased
creep test was conducted under water-saturated condition, and though the creep stress was almost constant. The air cylin-
creep stress was 2.8 MPa, which is approximately 30% of uni- der probably had some kind of troubles and was dealt with in
axial compressive strength. The first creep test was started on October 2010.
November 22, 1994. Unnatural movement in the measured dis- Comparison of creep strain and strain-rate indicated that the
placement values was observed in August 1996 and therefore trends in the long-term creep tests with two specimens were
the test was halted. With this experience, countermeasures consistent with those for primary creep in short-term creep
against some troubles in measurement apparatus were pre- tests conducted in previous studies. These findings suggested
pared not to halt the test. The second creep test was started that deformation mechanism of long-term creep was similar to
on May 23 in 1997. During the test, the specimen, its dis- that of short-term one, and that common theory and constitu-
placement and the air pressure have been checked everyday. tive equation can be applied to both long- and short-term creep.
Figure 1. Creep strain and stress in the creep test for over thirteen years with Tage tuff under water saturated condition. The slope of the
thick solid line is 7 × 10−13 /s. Creep stress was adjusted to the predetermined value several times.
306
M.C. He
School of Mechanics and Civil Engineering, China University of Mining and Technology (Beijing), Beijing, China
State Key Laboratory for GeoMechanics and Deep Underground Engineering, Beijing, China
W. Nie
School of Mechanics and Civil Engineering, China University of Mining and Technology (Beijing), Beijing, China
State Key Laboratory for GeoMechanics and Deep Underground Engineering, Beijing, China
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
Z.Y. Zhao
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
C. Cheng
School of Mechanics and Civil Engineering, China University of Mining and Technology (Beijing), Beijing, China
State Key Laboratory for GeoMechanics and Deep Underground Engineering, Beijing, China
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 INTRODUCTION K-feldspar (beige) and 5.0% clay minerals. The most abun-
dant clay minerals include 56% illite and 44% chlorite. The
Rock displays splitting fractures under uniaxial compression grains are visible to naked eyes, displaying a crystalline
and shear fractures under moderate confining pressure in lab. interlocking texture. The average density, uniaxial compres-
Rock exhibits plasticity under high confining pressures. Once sive strength and elastic modulus of granite are 2.57 g/cm3 ,
unloading occurs, the rock suffers a change from plasticity 131MPa and 21GPa, respectively. Both the UC test and the
to brittleness. Rockburst is one kind of sudden and violent CC test had a loading rate of 0.002 mm/s. All the tests on
changing progress from high confinement to no confinement granite were conducted on prismatic specimens of dimensions
involving large energy release in rock mass. Generally, the 150 mm (height) × 60 mm (breadth)×30 mm (thickness) by a
rock mass is under a static stress distribution before excava- true triaxial compression facility (He et al., 2007).
tion, and excavation results in the tractions on the excavation The one-dimensional variogram function, 2γ(x, h), can be
surface being reduced to zero within a short time period. expressed in the discretized form as:
Either the combined dynamic and static stresses exceeding
the strength of the rock mass, or the loss of rock structural
stability will lead to the occurrence of rockburst.
The application of laser profilometers in the rock surface
measurement offers a great tool to analyze the fracture prop- where x is the horizontal distance along a roughness profile,
erties. In this paper, the variogram method is used with a z(x) is the height of the roughness profile from the datum, M
suggested refining procedure to analyze the fracture surface of is the total number of pairs of toughness heights of the profile
granite specimens after rock burst, and compare the specimens that are spaced at a lag distance h. A program was developed to
failed under uniaxial compression with those under polyaxial apply the variogram fractal technique to analyze the profiles.
tests. The fractal dimension Drld and another parameter Kv The data, measured by laser profilometers, were available in a
are calculated and fitted by ellipses for quantifying the stress grid pattern in the form of z = f (x, y), where x-y is a referred
field. The distribution of these fractal parameters are used to coordinate system on the horizontal plane and the z coordinate
compare the rock deformable behavior under different stress at each point is the height of the fragment surface from the
boundary conditions. x-y plane in the normal direction.
Granite specimens, taken from Laizhou in Shandong province, The variogram method was applied to the two-dimensional
China, were used for analysis. This granite specimen consists roughness profiles that were calculated for the fragments sur-
of 27.0% quartz (gray), 31.0% plagioclase (white), 37.0% faces as a measurement in all directions (Kulatilake et al.,
307
4 CONCLUSIONS
308
V. Valle
Institut PPRIME CNRS UPR3346, Université de Poitiers, Futuroscope Chasseneuil, France
J. Cabrera
DEI-SARG Department, Institute for Radiological Protection and Nuclear Safety, Fontenay-aux-Roses, France
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 EXTENDED ABSTRACT In order to apply the desaturation path, our samples were
placed in an oven (with glass doors). The desiccation pro-
Fractures associated with the process of desaturation of cess was achieved by heating the samples to temperatures (T)
argillaceous medium were observed on faces and fronts of between 40◦ C and 200◦ C. The loss of water (w) was calculated
the galleries of the experimental platform of Tournemire. at the end of each thermal stage.
This hydric fracturing process was evidenced in situ by sub- The duration of each thermal stage (t) varied between from
horizontal cracks spaced out by about twenty centimeters, on one to a few hours. Prior to this desaturation path, samples
all vertical walls in contact with ambient air. were carefully (air) cut in two cores of about seven centimeters.
The aim of this experimental study (supported by GNR The first core was used to calculate (w) whereas strain fields
TRASSE, Transfert des Radionucléides dans les Sols, les (ε1 , ε2 ) were measured by DIC method through the glass doors
Sous-sols et vers les Ecosystèmes and IRSN, French Institute on the second core.
for Radioprotection and Nuclear Safety) was to investigate The results of this experimental investigation using optical
in laboratory this fracturing process by using a non invasive method (DIC) are the following:
and non destructive optical method (Digital Image Correla-
tion, DIC). This method consisted in (a) recording with IDS i) The cracks (aperture ≥80 µm) induced by heating began
µEye SE (CMOS 1/2 sensor) camera (1280 × 1024 pixels2 ), to be observable at 80◦ C. As observed on the gallery walls,
digital images of a sample undergoing a mechanical transfor- these cracks were parallel to the bedding planes and spaced
mation (heating) and (b) using an image correlation algorithm out by twenty-twenty five millimeters.
in order to obtain full-field displacements on the core surface. ii) These cracks appeared also at the interface between
The image correlation algorithm usually uses a random the clay matrix and the macroscopic non-clayey hetero-
speckle located at the sample surface. In our case, small geneities (e.g., calcareous nodules).
particles (<100 µm) deposed on a small thickness of glue iii) Before the observation of such cracks at a macroscale,
and covering the surface sample were chosen as the random strain localizations were measured close to the locations
speckle. The measuring resolution depends on the marking of the future cracks.
(speckle) and varies generally from 0.05 to 0.1 pixel. In this iv) When water was fully removed from samples near 100–
study, the measuring resolution was about 0.08 pixel. 150◦ C according to the duration of the thermal stage, the
Two sets of two samples, in an undisturbed state were used cracks were no more observable and seemed to be closed.
for this experimental study. These two sets were taken from This original sealing process is likely to be associated with
the boreholes GSM135 and GSM180, which were drilled in the thermal expansion of the solid phase.
2009 perpendicular to the wall of the North gallery of the v) The strain localizations were organized following two net-
Tournemire URL. The borehole GSM180 was subvertical and works: the main network was associated with the first
perpendicular to the bedding and GSM135 was drilled with cracks (parallel to the bedding planes) and a secondary net-
an angle of 135◦ from vertical. At the beginning of this study, work connected to the main network occurred in a second
the samples were in a near-saturated state. The samples were step, with a complex spatial organization.
200 mm high and had a 78 mm diameter.
309
L.E. Hernández
Regional Ministry of Works, Government of the Canary Islands, Spain
Subject: Rock material and rock mass property testing (laboratory and in situ)
310
311
K. Yamamoto
Technology and Research Center, Japan Oil, Gas and Metals National Corporation (JOGMEC), Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
312
Direct shear and tensile test on cemented healed joints from El Teniente mine, Chile
ABSTRACT: Direct shear and direct tensile tests have been conducted on specimens taken from drill cores of porphyry ore
from the El Teniente mine containing infilled veins and bands of e.g. copper sulphide. The direct shear tests were conducted
at different normal stress levels in order to determine the failure shear strengths. The direct tensile tests were conducted using
stiff load frame in order to be able to obtain the complete stress-deformation. The tensile failure stress and fracture energy was
computed in the cases when the failure was localized to the vein.
Subject: Rock material and rock mass property testing (laboratory and in situ)
313
314
Y.Z. Jiang
Key Laboratory of Geotechnical Mechanics and Engineering of the Ministry of Water Resources,
Yangtze River Scientific Research Institute, Wuhan, China
R.H. Wang
Key Laboratory of Geological Hazards on Three Gorges Reservoir Area, China Three Gorges University,
Ministry of Education, Yichang, China
Subject: Rock material and rock mass property testing (laboratory and in situ)
For investigating the rheological properties of rock mass in samples became more and more obvious under the condition
the loaded state, there’re a lot of researches both abroad and at of unloading confining pressure sequentially and maintaining
home, which makes great progress, however, there are rarely stress level and confining pressure constant, and sandstone
experiments and theoretical studies on rheological mechani- rock samples, when in the final level of confining pressure
cal properties under the condition of unloading. Engineering levels, showed the three typical rheological stages over time,
practice shows that the instability and destruction of rock which are primary attenuating rheological stage, the steady
excavation does not occur immediately after the completion uniform rheological stage and accelerating rheological stage,
of the excavation, and the deformation due to excavation and finally it came into rheological destruction of non-linear
unloading has the time effect, and unloading is a time-related accelerating.
concept, including the instantaneous unloading, as well as Based on the results of triaxial compression rheological
long-term unloading. It is quite necessary to carry out rheolog- experiments, the parameters of Burgers rheological model
ical test of rock mass under unloading for the study of the rock were identified with the optimized algorithms, the non-
mass long-term stability, which has an important engineering stationary law between rheological parameters and confining
significance. pressure, stress level were studied. From the regression anal-
The approach of step loading on the same rock was adopted ysis results, it can be seen that the precision of the results
to carry out indoor rheological test. Load hydrostatic pressure of the exponential function is slightly lower than the return
to 30 MPa at the rate of 0.05 MPa/s firstly; when deformation fitting power function fitting precision. However, the correla-
was stable, maintain confining pressure at 30 MPa, continue tion coefficients of two functions have reached more than 0.95.
applying axial force to the set value at the rate of 0.5 KN/s, and The trend function could explain the regularity of feature vari-
maintain it constant for hours; when rock rheological deforma- able parameters to some extent. But for specific rheological
tion under the level of former class became stable, maintain parameters, the follow factors should be taken into consid-
axial force σ1 constant, and unload confining pressure step eration: the physical meaning of the rheological parameters;
by step at the rate of 5 MPa each level. Pay attention to that the corresponding combination of physical and mechanical
when each level of confining pressure unloaded, the emer- mechanisms. The rheological parameter η1 of Burgers rheo-
gence of rock steady-state creep are required to be observed, logical model reflects the size of the flow rate when sandstone
and then unloaded next level of confining pressure, until the is in a steady flow stage. When the confining pressure less or
destruction of rock flow occurred. The rate for unloading equal to zero, the rheological parameter η1 will corresponding
confining pressure were all 0.05 MPa/s. Control temperature decrease, but will not reduce or close to zero, because when
and humidity condition maintained constant during the exper- the rheological parameters η1 tends to zero, the corresponding
iment. By using of rock servo-controlling rheology testing uniform steady state flow stage of the flow rate will tend to
machine, triaxial compression rheological experiments were infinity, so ,as the confining pressure tends to zero, the rhe-
carried out, the variation laws of rock deformation with time ological parameters η1 tends to a constant greater than zero.
went was analyzed under different confining pressure, the The exponential function can be used to describe the relation-
deformation properties of different rocks in the rock triaxial ship between the rheological parameters η1 of sandstone and
rheological process were studied, and the development trends confining pressure.
and characteristics of rock deformation in stage of acceler- The phenomenon that for rock nonlinear rheological will
ated rheology were discussed. At the same time, the rupture happen in long term has gradually be recognised and
forms and mechanisms were analyzed. It can be seen from concerned; the development of rock rheology damage and
tests that, the rheological phenomena of sandstone samples unloading capacity are related to action time, represented
took place significantly over time in the long-term external in non-linear rule. When the exterior loads a rock suffered
loads, the rheological deformation characteristics of sandstone is greater than the long-term strength, the rock will fall
315
316
1 INTRODUCTION
317
M. Kuruppu
Western Australian School of Mines, Faculty of Science and Engineering, Curtin University,
Kalgoorlie, Western Australia, Australia
Subject: Rock material and rock mass property testing (laboratory and in situ)
In order to investigate the influence of anisotropic charac- On the other hand, in 3-2 type, the crack propagates along grain
teristics of rock and water vapor pressure of surrounding boundaries smoothly.
environment on fracture toughness, a series of Semi-Circular The results obtained in this paper are summarized as
Bend (SCB) tests under various water vapor pressures were follows:
performed.
1) The granite has anisotropic property. The fracture tough-
SCB specimen is a semi-circular with an artificial notch and
ness shows the directional dependence;
subjected to three-point bend loading. The setup of specimen
2) The fracture toughness depends on water vapor pressure
for SCB test is shown in Figure 1. The load is applied through
and decreases with increasing water vapor pressure; and
one upper and two lower rollers. The setup of specimen is
3) Based on the X-ray CT images of the fractured specimen,
placed inside a vacuum chamber used to control surrounding
the fracture toughness is influenced by the type of crack
environment of specimen. The load is recorded using a load
propagation.
cell equipped on a loading plate.
The rocks used in the test were African granite and Korean
granite. The measurement of elastic wave velocity was per-
formed to make clear orientation of inherent micro cracks.
Considering the direction of the artificial notch, two types of
specimen were prepared, 1-2 type and 3-2 type as shown in
Figure 2.
The relation between fracture toughness and water vapor
pressure is shown in Figure 3. The values of fracture tough-
ness of 1-2 type are larger than that of 3-2 type for both
granites. The fracture toughness varies almost linearly against
water vapor pressure on the logarithmic graph and increases
with decreasing water vapor pressure. The equation of the
approximated lines is:
Figure 3. Relation between fracture toughness and water vapor
pressure.
where β is a constant and m is a slope of the line.
Furthermore, crack propagation was discussed based on the
images of fracture visualized within rock specimen after SCB
test by means of X-ray CT scanner. The gray-level CT images
near the artificial notch tip of the fractured specimen are shown
in Figure 4. It is found that the crack propagates near the notch
tip along boundaries around grains or across grains in 1-2 type.
Figure 1. Setup of specimen. Figure 2. Two types of specimens. Figure 4. CT images near notch tip of fractured specimen.
318
M. Kato
Faculty of Engineering, Hokkaido University, Sapporo, Hokkaido, Japan
K. Kaneko
Faculty of Engineering, Hokkaido University, Sapporo, Hokkaido, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
The relation between rock mass deformation and groundwater kept constant during the test, second, the condition of triaxial
behavior is of importance for several fields in earth science compression, which the axial stress is suddenly subjected to
and engineering including geophysics, structural geology, both ends while confining pressure and pore pressure at both
rock physics, engineering geology, and geomechanics. The ends are kept constant, and third, the pore pressure is suddenly
knowledge of this relation is needed particularly for repos- changed at both ends while confining pressure is kept constant.
itories design for radioactive waste disposal. In terms of a During these tests, axial and circumferential strains are mea-
geological barrier of the repository system, sedimentary for- sured by the extensometers and discharge of pore water from
mations are predominant to guarantee that the radioactivity a specimen by the displacement of piston of syringe pump
cannot escape for several centuries. (Figs. 1, 2).
In this study, poroelastic and hydraulic parameters of sed- Hydraulic constants obtained from triaxial tests agreed
imentary rocks were estimated simultaneously from triaxial well with those from transient pulse permeability tests. The
compression tests with instantaneous change either of axial experimental results for discharges from the specimen dur-
stress, confining pressure, or pore pressure. The triaxial tests ing the tests permit theoretical interpretations. Therefore,
were conducted under the drained condition but under the besides the behavior of pore water, the transient response
undrained condition because of the difficulty of strict real- of rock deformation corresponding to pore water move-
ization. Stress conditions adopted in the tests are grouped ment can be theoretically treated. Accordingly hydraulic
into three types: first, the condition of hydrostatic compres- constants can be estimated from the strain data of triax-
sion, which all principle stresses are changed equally and ial tests with instantaneous change either of stress or pore
suddenly while pore pressure at both ends of a specimen is pressure.
Figure 1. Variations of strains and specific discharge with time from Figure 2. Variations of strains and specific discharge with time from
triaxial test with sudden change of axial stress. triaxial test with sharp change of pore pressure.
319
M. Kohno
Cold Regions, Environmental and Energy Engineering Course, Graduate School of Engineering,
National University Corporation Kitami Institute of Technology, Kitami, Japan
H. Maeda
Department of Civil and Environmental Engineering, Faculty of Engineering,
National University Corporation Kitami Institute of Technology, Kitami, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: weak rock, rock properties, rock failure, rock mass, lab testing, field measurements
320
It has been well established experimentally that dilatancy The Brenna sandstone was selected intentionally since it is
may appear at relatively early stages of the deformation pro- a rock that exhibits a considerably high compressive strength
cess under compressive loading conditions. In particular, in and, at the same time, behaves in a highly dilatant manner.
medium- and coarse-grained clastic sedimentary rocks, the The threshold of absolute dilatancy of the Brenna sandstone
onset of dilatancy may not exceed several percent of the peak amounted to, on average, 50% of the ultimate strength and the
differential stress under low confining pressure conditions, mean value of the dilatant volumetric strain at strength failure
and an absolute increase in volume often begins at stress lev- reached almost 1.4%.
els as low as 40 percent of the ultimate strength. Unfortunately, For modeling purposes, the stress-strain curves as obtained
these effects are often neglected in the mathematical model- from laboratory compression tests, were simplified to a piece-
ing of rocks and rock masses. Even those researchers who wise linear form (Fig. 1). The proposed idealized model has
acknowledge the pre-failure dilatancy in rocks tend to dimin- the following features: (i) it is a bilinear model in the pre-
ish its importance and when building a model, they allow failure domain which allows for the compaction and dilation
for the increase in volume only in the post-peak domain,
that is they stick to the very unrealistic assumption that it is
solely compactant behavior that rocks manifest in the pre-peak
domain (see Brown et al. 1983,Vermeer & de Borst 1984,Yuan
& Harrison 2004, Alejano & Alonso 2005, Zhao & Cai 2010).
In order to lay the foundation for a future challenge of this
common practice of excessive simplification of the volumetric
deformation mode of rocks in the pre-peak domain, a study
has been undertaken by the present authors with the aim of
creating a simple yet realistic model capable of capturing both
of the phenomena that are usually observed in the pre-failure
domain: compaction at the early stage of deformation and the
dilation that follows compaction and precedes strength failure.
In order to reveal the main features of the volumetric defor-
mation mode of a rock in the pre-peak domain and, first of all
to determine the values of the dilatancy angle in this domain,
samples of a medium-grained sandstone from Brenna in the
˛ Mountains in southern Poland were tested under
Beskid Ślaski Figure 1. Simplified, piecewise linear characteristics representing
uniaxial compression conditions. the behavior of rock in the pre-failure domain.
Figure 2. Slopes of the volumetric strain vs axial strain characteristic (a) and the lateral strain vs axial strain characteristic (b) in the
contractant domain (0-TD) and in the pre-peak dilatant domain (TD-F).
321
REFERENCES
or from the slope of the lateral strain vs axial strain character-
istic in the same domain (Fig. 2b): Alejano, L.R. & Alonso, E. 2005. Considerations of the dilatancy
angle in rocks and rock masses. International Journal of Rock
Mechanics & Mining Sciences 42:481–507.
Brown, E.T., Bray, J.W., Ladanyi, B. & Hoek, E. 1983. Ground
Similarly, the dilatancy angle can be determined either from response curves for rock tunnels. Journal of Geotechnical Engi-
the slope of the volumetric strain vs axial strain characteristic neering 109(1):15–39.
Vermeer, P.A. & de Borst, R. 1984. Non-associated plasticity for
in the pre-peak dilatant domain (TD-F range in Figure 2a): soils, concrete and rock. Heron 29(3):3–64.
Yuan, S.-C. & Harrison, J.P. 2004. An empirical dilatancy index for
the dilatant deformation of rock. International Journal of Rock
Mechanics & Mining Sciences 41:679–686.
Zhao, X.G. & Cai, M. 2010. A mobilized dilation angle model
or from the slope of the lateral strain vs axial strain character- for rocks. International Journal of Rock Mechanics & Mining
istic in the same domain (Fig. 2b): Sciences 47:368–384.
322
S. Kwon
Department of Energy Resources, Inha University
Keywords: rock properties, lab testing, nuclear repository, underground research tunnel
1 INSTRUCTIONS
323
4 COMPUTER SIMULATION Chaudhary, D.R. & Bhandari, R.C. 1968. Heat transfer through a
three-phase porous medium. Brit. J. Appl. Phys 1:815–817.
Using the rock properties at KURT, the thermal and mechan- Cho, W.J. & Kwon, S. 2010. Estimatino of the thermal properties for
ical conditions before and after the construction of the tun- partially saturated granite, Engineering Geology 115:132–138.
nel could be predicted. In the computer simulation using Kwon, S., Choi, W.J. & Choi, J.W. 2010. Initial Thermal Conditions
FLAC3D, the topography above the tunnel, rock property around an Underground Research Tunnel at Shallow Depth, Int. J.
variation, geothermal gradient, seasonal temperature varia- of Rock Mechanics and Mining Sciences 48:86–94.
tion, and sequential tunnel excavation were considered. Fig. 2
shows the model mesh.
324
Shear behavior model of rock joints applicable to CNL and CNS conditions
Keywords: lab testing, rock failure, rock joints, rock properties, rock stress
Mechanical properties of rock mass and its stability are highly The peak friction coefficient (µp ) was formulated asequation
affected by the mechanical and geometrical characteristics of (1), and the peak shear displacement (up ) was estimated as
discontinuities. One of the most important characteristics of equation (2). The post-peak behavior of the friction coefficient
discontinuities is shear strength and shear behavior observed was modeled using the peak friction coefficient and peak shear
in direct shear test. The direct shear test is divided into constant displacement. The overall modeling process of the behavior
normal load (CNL) test and constant normal stiffness (CNS) of friction coefficient is as follows. The peak friction coef-
test according to loading condition. The CNL condition is ficient is calculated by equation (1), and then the peak shear
suitable for rock slope while the CNS condition is suitable displacement is obtained by equation (2). The post-peak curve
for underground rock mass where the shear behavior of dis- intersecting the peak friction coefficient and the peak shear
continuities depends on the stiffness of rock mass. Reliable displacement at the initial normal stress is increased contin-
prediction of shear behavior of rock joints for both conditions uously in equation (1) and (2). The normal displacement in
has been an important issue in rock engineering field. the pre-peak range was assumed to be zero based on the test
In this study, an empirical model of shear behavior of rock results. The post-peak behavior was modeled as equation (3).
joints was proposed from direct shear tests using a servo-
controlled shear testing machine. Cement mortar replicas of
rock joints were tested with various normal stiffnesses (Kn ),
initial normal stresses (σ0 ), joint roughness coefficients (JRC)
and joint wall compressive strengths (JCS). Joint roughness
was measured using a 3-D laser roughness profiling machine.
The shear test results of 120 specimens were analyzed to inves-
tigate the effects of loading condition and material properties The suggested models can be applied to predicting the
on friction coefficient and normal displacement of the joints. shear behavior of rock joints including pre-peak and post-
Empirical models of the friction coefficient and normal dis- peak shear stress levels regardless of the loading condition.
placement were suggested based on the test results. In the To verify the suggested model, additional shear tests of rock
empirical models, dimensionless terms were adopted to avoid joints were carried out and the performance of the models was
the scale effect and therefore to enhance the applicability of compared with those of other model. Based on the additional
the suggested models. In a pre-peak range, the friction coef- test results, the suggested model of peak friction coefficient
ficient showed a linear relationship with shear displacement. showed lower error of prediction than Barton’s empirical shear
Figure 1. The behavior of friction coefficient under various initial normal stresses.
325
Figure 3. Verification of suggested shear model at applied test result of previous studies.
strength model (Barton, 1973) and modified Barton’s model size were selected from the previous studies for the compar-
suggested by Jang (Jang et al., 2010). Table 1 showed the ison to avoid the scale effect. The previous studies were only
comparison result. It was also verified by comparison with for CNL condition. In overall stress levels, the prediction by
previous studies of Bandis (Bandis, 1980), Homand (Homand suggested models showed good agreement with the shear test
et al., 2001), and Son (Son, 2005). Figure 2 shows comparison results. The suggested models can be effectively applied to
with additional tests while Figure 3 shows comparison with predicting peak friction coefficient and overall stress levels
previous studies. Only the test results of specimens similar in for both CNL and CNS conditions.
326
Evaluation of shear velocity dependency of rock fractures by using repeated shear tests
B. Li & Y. Jiang
Department of Civil Engineering, Nagasaki University, Nagasaki, Japan
G. Wang
College of Civil Engineering and Architecture, Shandong University of Science & Technology, Qingdao, China
Subject: Rock material and rock mass property testing (laboratory and in situ)
Understanding the deformational behaviors of rock masses friction of a fracture surface has a positive relation with the
subjected to seismic loads is an important issue to ensure shear velocity. The dilation of fracture decreases with sheared
the stability and safety of some important structures such as loops since the asperities on fracture surface are crushed more
dam and nuclear power plant, which requires adequate evalu- and more with the increase of shear loops.
ation of the mechanical properties of discontinuities existing No obvious relation between peak shear stress and shear
in rock masses. The mechanical properties of rock fractures velocity can be observed. Since rock is one kind of brittle
are usually estimated through laboratory test on rock samples material, the failure of major asperities at peak shear stress
containing single rock fracture. Direct shear test in labora- may depend more on the material properties, rather than the
tory is considered as an effective method to estimate the shear velocity that the failure is induced. Therefore, the behavior of
behavior of rock fracture, which is commonly conducted at fracture at residual stage will be more representative to the
relatively low shear velocity (e.g. 0.5 mm/min). When earth- characteristics of fracture subjected to seismic loads. At here,
quake happens, the shear velocity on rock fractures can vary in we take the shear stress and normal displacement at shear
a large range depending on the seismic strength. In this study, displacement of 10 mm as residual shear stress and residual
repeated shear tests on artificial rock fracture samples were normal displacement, respectively, to estimate the mechanical
conducted with shear velocities ranging from 0.5 mm/min properties at residual stage.
to 50 mm/min to estimate the effects of surface roughness, The residual shear stresses at different shear velocities
shear velocity and shear history on the shear behavior of rock exhibit clear linear relations with the normal stress. The larger
fractures. the shear velocity, the steeper the straight lines connecting
A servo-controlled direct shear apparatus for rock fractures residual shear stresses become, demonstrating that in the resid-
using virtual instrument (VI) software was used in this study.A ual stage, the fracture has larger friction at higher shear veloc-
rock fracture, labeled as J10, was created by splitting a granite ity. At the first shear loop with shear velocity of 0.5 mm/min,
block, and was used as prototype to produce artificial replicas the major asperities have not been totally crushed, thus cohe-
of rock fractures. Its surface is very rough with a few major sion can be observed by extending the straight line to the
asperities on various locations, and the JRC range is 18∼20. y-axis. In the following loops, the cohesion become very small
Totally 15 tests were conducted under constant normal load approaching 0 due to the damage of the asperities. The resid-
boundary conditions with normal stresses of 1, 2 & 4 MPa. ual normal displacements decrease proportionally with the
Repeated shearing was conducted in these tests with increas- increase of normal stress and shear velocity. For the cases with
ing shear rates for each loop of 0.5, 5, 10, 20 & 50 mm/min. the initial shear velocities (velocity at the first loop) of 0.5 &
The samples were sheared with displacement up to 20 mm, 5 mm/min, the friction angles increase significantly with the
then the shear direction was reversed and the samples were shear velocity before 20 mm/min, after that, the friction angle
sheared back to 0 mm to complete a loop. trends to be constant. For the cases with initial shear veloc-
In the repeated shear test results, the peak shear stress can ities larger than 5 mm/min, the friction angles increase with
only be observed at the first shear loop. After the peak, the the increase of shear velocity. These results indicate that the
shear stress undergoes a slow decrease until around 5 mm friction angle of a fracture without shear history depends on
of shear displacement, then, the shear stress keeps almost the behavior of major asperities, which don’t have clear rela-
constant value in the residual stage. The shear stress in the tion with shear velocity. The friction angle at more shear loops
backward shearing is much smaller than that of the forward always has larger value than that with fewer loops, in the con-
shearing for all the loops, since the backward shearing is dition of the same shear velocity. During a shear, the asperities
merely a restoration process on the shearing tracks of the for- are being crushed, which generate plenty of gouge materials.
ward shearing. The shear stresses at residual stages (e.g. at The gouge will fill the void spaces between the two walls of
10 mm of shear displacement) increase with the shear veloc- the fracture, which increases the contact areas between the two
ity for both forward and backward shearing, indicating that the walls, thus gives rise to the friction angle.
327
Q. Li
Department of Civil Engineering, The University of Hong Kong, Hongkong
The Research Center for Material Failure Modeling, Dalian University, Dalian
D.G. Wang
The Research Center for Material Failure Modeling, Dalian University, Dalian
Two numerical tests have been carried out to validate the new
2 CONSTITUTIVE MODEL developed constitutive model in the paper. One is the extension
of cracks in the experiment of pre-existing crack board under
In general, rocks are treated as a quasi-brittle material which uniaxial tension. It can be shown that the development path
means there is little plastic effect especially at a macro scale. of existing crack is first along the initial direction, and then
But as the different definition of plasticity in metals, the vertical to the horizontal tension stress (Shown in Figs. 1–2).
irreversible deformation is equal to the plastic deformation The other one is the uniaxial compression rock samples. It
328
is not difficult to find both the parallel and conjugate shear proposed, in which the multiple yielding criterion based on
band in the figures of failure (Shown in Fig. 3). With the stress space and ductile failure criterion based on strain space
increase of compression stress, the relatively weaker elements are adopted. What is more, the program for numerical model
first get yielded, and then fail which lead to release of their and calculations is self-developed which is open and can be
own stresses, resulting in a stresses redistribution of local area extended, therefore makes it easy to develop the corresponding
(Shown in Fig. 4). constitutive model with different material in further research.
Through the comparison with the relevant physical exper- According to the simulation of two tests including the ten-
iment and numerical method, it is shown that the present sion test of pre-existing crack board and the compression test
model is rational and accurate. of rock samples, it is shown that the present model is ratio-
nal and accurate, which can simulate the extension mode of
cracks and the development of localized shear band.
4 CONCLUSIONS
329
D.M. Lin
Key lab of engineering geomechanics, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
Graduate University, The Chinese Academy of Sciences, Beijing, China
Y.J. Shang
Key lab of engineering geomechanics, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
W. Zhang
Key lab of engineering geomechanics, Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
Graduate University, The Chinese Academy of Sciences, Beijing, China
ABSTRACT: GSI is the methodology that developed by E. Hoek and other geologists throughout the world. It is particularly
suitable for weathered and heterogeneous rock. However, there are not quantitative parameters for GSI, which leads to sub-
jectivity when getting value. Based on the study of rock mass in the region where is abundant of gneiss during geotechnical
investigation, the author propose a new parameter RCL to help the quantitative GSI; combine with geological effect of alteration
which is very normal in metamorphic rock, the author try to utilize the photos of the intact core in the field to substitute the
sketch of rock mass structure in the chart of GSI’s value, Ultimately, set up a chart of GSI’s value on the basis of gneiss. In
terms of the contrast of modulus of elasticity between calculation by the author and the value given by the unit of investigation,
the author finds that the result fit well, which means it is scientific of this chart.
Subject: Rock material and rock mass property testing (laboratory and in situ)
Drilling core is very important in the study of parameters In Fig. 1 the GSI chart given by Hoek is not highly applicable
for GSI; though RQD is one of most popular description in drilling project, especially the sketch of rock mass struc-
of drilling core, for drilling core can expose faults, joints ture, with the core box the author substitute the sketch with
in underground engineering, and Rock Core Length can typical photographs, in that way, it can look much more intu-
reflect the joint spacing, joint direction, joint sets and so itive. In the description of the drilling core, we combine with
on; from Fig. 2, you can find the obvious difference of the RQD, drilling core rate to have a good understanding of GSI
two underground geological condition in weathering, RQD, in drilling core.
joint roughness and filling, so the author refer the scholars Such as Fig. 3, you can see the differences obviously
around the world on rock mass structure, and find the rela- compare with Fig. 1.
tionship between core drilling and rock mass structure, at last The intact rock block length RCL that suggested by the
try to introduce RCL as one important parameter to help the author is an important parameter in rock mass structure’s study
quantitative value of GSI. at present which can reflect abundantly underground informa-
Table 1 is one the description and process of getting joint tion. Quantificative research of GSI combine with RCL can
condition value, we use this parameters to divide surface con- avoid effect of subjective factors to some extent. Many schol-
dition into several grand to help the quantitative value of GSI. ars research the GSI in underground engineering care about
At last we set up a chart of GSI’s value on the basis of gneiss’ the volume of rock, but is the same that consider the rock core
category such Fig. 3. length in drilling engineering. So this parameters is useful in
the research of GSI.
The structure surface is the most likely weathered surface,
so compile with joint roughness, filling can be practical in
Figure 2. The core core of different structure for gneiss (a: block structure b: sheared and fractured structure).
330
Keywords: rock mechanics, water-rock interaction, true/pseudo tri-axial tests, limit yield stress, anisotropic Biot coefficient,
action mechanism
332
REFERENCES
333
S.D. McKinnon
The Robert M. Buchan Department of Mining, Queen’s University, Kingston, Ontario, Canada
Keywords: rock mass classification, rock joints, rock properties, site characterization, field measurements
334
Figure 2. The 2D surface profile averaged from binned point cloud data. The maximum asperity amplitude is calculated based on
perpendicular distance to a straight line through the two highest local maxima. The 2D surface profile is from site #7, along the strike.
4 CONCLUSIONS
335
Understanding failure mode in uniaxial and triaxial compression for a hard brittle rock
V.B. Maji
Department of Civil Engineering, IIT Madras, Chennai, India
1 INTRODUCTION
336
5 SUMMARY
There is no single failure criterion which can properly jus- Basu, A., Celestino, T.B., and Bortolucci, A.A. (2009), Evaluation
tify the brittle rock fracture process and failure. Often Mohr’s of rock mechanical behaviours under uniaxial compression for
different weathering grades, Rock. Mech. Rock. Engg, Vol. 42,
theory is used in predicting the failure of brittle rock and can No. 1, 73–89.
be opted for 2D stress but is unable to explain the tensile Klien, E., Baud, P., Reuschle, T. and Wong, T.F. (2001), Mechanical
brittle fracture in rock. Griffith theory better suits for brittle behavior and failure mode of Bensteim Sandstone under triaxial
rock but unable to follow the in-homogeneity in tock material. compression, Phys. Chem Earth (A), Vol. 26, 21–25.
It explains the two dimensional relationship between shear Mogi, K. (1971) Fracture and flow of rocks under high triaxial
and normal stress at point of failure and the mechanism of compression. J. Geophys. Res. 76, 1255–1269.
failure is based on the formation, propagation and joining of Mogi, K. (1974) On the pressure dependence of strength of rocks
microscopic cracks leading to macroscopic failure. Griffith and the Coulomb fracture criterion. Tectonophysics, 21, 273–285.
cracks edges concentrate huge stresses and eventually failure Peng, S. and Johnson, A.M. (1970) Crack growth and faulting in
occurs. The von-Mises criterion may be modified and used cylindrical specimens of chelmsfold granite, Int. J. Rock Mech.
Min. Sci. Vol. 9, 37–86.
as satisfactory criterion for fracture of brittle material like Szwedzicki, T. (2006), A Hypothesis on Modes of Failure of Rock
rock. Because of the variable distortion strain energy (τoct ) Samples Tested in Uniaxial Compression. Rock. Mech. Rock.
during high pressure dependent brittle fracture, applicability Engg. 40 (1), 97–104.
of the von-Mises criterion for brittle failure may be ques- Wawersik, R.R. and Fairhurst, C. (1969) A study of brittle rock frac-
tionable as the criterion assumes, yielding occurs when the ture in laboratory compression experiments, Int. J. Rock Mech.
distortional strain energy (τoct ) reaches a constant value. Mogi Min. Sci. Abstr. vol. 7, 227–229.
337
Keywords: coal, gas flow and gas bursts, lab testing, rock mass properties
1 INTRODUCTION
338
Keywords: lab testing, numerical modelling, physical modelling, rock failure, rock properties
In this study, creating realistic specimen models of monomin- In the initial stage (stage I), tensile failure occurs at random
eral polycrystalline rock on a computer, we performed a 3D at grain boundaries with a small angle of the normal direc-
finite element analysis of uniaxial tensile fracture under the tion relative to the loading axis and the connectivity between
assumption that fracturing occurs only at the grain bound- them is poor. Unloading always occurs before the peak at
aries. The grain boundaries were modeled by quadratic grain boundaries that do not compose the final failure plane.
joint elements and the constitutive law was derived by During the completion of the final failure plane after the peak
applying the associated flow rule in the plastic theory and (stage III), the conversion of failure mode from tension to shear
a tension-shear-softening curve to an extended Coulomb occurs at grain boundaries with a large angle of the normal
criterion. direction relative to the loading axis to form a single failure
Figure 1 shows an example of the change in the number of plane.
failure mode with the normalized axial strain, including both The normal stresses on the grain boundaries that have a
elastic and unloading modes (Model 2). Open circles with small angle of the normal direction relative to the loading
Arabic numerals in the normalized σz − εz curve indicate (1) axis are always tensile until the final stage (7 & 27 in Figure
onset of tensile fracture, (2) onset of unloading, (3) the peak 3). On the contrary, the failure mode of the grain boundaries
strength, (4) onset of shear failure, (5) the point where the that have a large angle of the normal direction relative to the
number of tensile mode is maximum, and (6) the point where loading axis (21 & 39) was changed from tension to shear.
a failure plane is completely formed, after which no fracturing For such grain boundaries, the normal stress was in compres-
newly occurs. Figure 2 shows the failed grain boundaries at sion in the elastic stage of loading, and changed to tension
the stages indicated by open squares with Roman numerals to reach the failure condition in tension. However, the shear
in the normalized σz − εz curve in Figure 1. The dotted area, stress did not decrease significantly after the failure, since
the cross-hatched area, the broken line and the thick solid line the axial displacement given at the top end of the specimen
indicate partial failure in tension, shear failure, unloading and
the grain boundaries that penetrate to the side surfaces of the
specimen, respectively.
Figure 1. The change in the number of failure mode with the nor-
malized axial strain (εz Eeff /T0 ) ( Model 2, T0 = tensile strength of Figure 2. Intergranular cracks at stages I, II, III and IV in Figure 1
grain boundaries, Eeff = effective Young’s modulus). (Model 2).
339
340
H.J. Mendieta
Geotechnical Laboratory, Universidad Mayor de San Simón, Cochabamba, Bolivia
Keywords: lab testing, unconfined compressive strength, point load test, index parameters
1 UNCONFINED COMPRESSIVE STRENGTH TEST The failure is caused by cracks parallel to the loading axis.
Point load resistance is defined by the Point Load Index, Is ,
The concept of Uniaxial Compressive Strength refers to the given by Equation 1, where P is the load al failure and D the
maximum stress in the direction of a single axis that rock can distance between the tips.
withstand before breaking. The standard direct method for
determining compressive strength of intact rock is specified
by ISRM (1979) “Suggested Methods for Determining the
Uniaxial Compressive Strength and Deformability of Rock
Materials”. The expression that relates the Point Load Index, Is(50) , with
In this test, a right circular cylinder specimen is com- Unconfined Compressive Strength, σ ci , obtained according to
pressed between two plates of specific material and shape study carried by Bieniawski (1975) and Broch and Franklin
characteristics until failure is reached. (1972) is as follows (Equation 4) where K is the correlation
The Unconfined Compressive Strength, σci , is obtained coefficient.
dividing the force that caused failure, P, by the cross sec-
tion area of the specimen tested, A, (Equation 1). At least 5
specimens have to be tested in order to obtain representative
results.
3 INDEX PARAMETERS OF INTACT ROCK
341
4.3 Results
DMR 01 15.51
5 RESULTS ANALYSIS
DMR 02 15.18
DMR 03 14.90 Statistical analysis was performed in two stages. In the first
DMR 04 11.59 stage bivariate analysis was performed in which facing each of
DMR 05 18.15 the independent variables with dependent (point load index)
DMR 06 24.47 separately. To For this purpose the results were analyzed in
DMR 07 20.86 pair series.
DMR 08 24.58 In most research problems where it is applied regression
DMR 09 21.91 analysis requires more than one independent variable in the
regression model. In the second stage a multivariate analy-
sis that investigated the influence two or more independent
variables was performed
The characterization included strength tests like Uncon-
Both the bivariate and multivariate analysis used least
fined Compressive Test, UCS, and Point Load Test, PLT. The
squares method for the determination of the trend lines.
Point Load Test included the determination of the anisotropy
From the multivariate analysis we obtained the following
index, Ia , relating the maximum Point load Index and the min-
equation where the unconfined compressive strength depends
imum Point load index obtained perpendicular and parallel to
upon the shore hardness value, density of intact rock and
stratification respectively.
anisotropy index as obtained from the point load test.
4.1 Test samples
The samples were obtained from the study area of a project
located in the capital of Bolivia, Sucre. Complementary, two In the same way, the relation between the correlation coeffi-
more samples were characterized, one from El Abra Tunnel cient K and index parameters was investigated. The result of
and the other from a highway, both located in Cochabamba, this study is the following.
Bolivia. The samples are conformed by clastic sedimentary
rocks (sandstones).
342
L. Meng
China University of Mining & Technology, Beijing, China
State key lab of Coal Resources and Safe Mining, Beijing, China
M.J. Liu
Henan Polytechnic University, Jiaozuo, China
Y.G. Wang
Henan Polytechnic University, Jiaozuo, China
Keywords: dynamics, lab testing, rock failure, rock mass, rock properties
Coal or rock dynamic catastrophe is one kind of natural dis- Based on the previous research achievements and the test
aster induced by mining activities and phenomenon that coal principle of LCR meter, we design and make an experimental
or rock exerted by external stress results in dynamic effects system of electrical parameters of coal/rock mass under load
and disaster consequences in a very short time. And coal or as shown in the Figure 1, which can be able to conduct low
rock dynamic catastrophe has become worse than ever with frequency measuring for electrical parameters of coal mass in
deep coal seams being mined at higher production rate in a the process of coal mass fracturing. Electrical parameters of
more complex underground deep mining environment, so it is coal mainly include resistivity, dielectric constant and mag-
very important for disaster prevention and reducing the losses netic conductivity. However in view of less magnetic medium
caused by dynamic disaster to the minimum to accurately contained in coal bearing formation, magnetic conductivity
forecast coal or rock dynamic catastrophe. of which produces no changes and basically maintains at 1,
Geophysical methods forecasting coal or rock dynamic so electrical resistivity and dielectric constant are selected
catastrophe can overcome some traditional static prediction as main electrical parameters. In order to make experimental
methods’ shortcomings, such as not to exert great influence results more representative, these experimental samples that
over normal production, to authentically achieve real-time this paper uses include two structural types of coal samples,
dynamic monitoring and so on. Given the above mentioned such as the tectonic coal and the un-deformed coal respec-
analysis, it is necessary for cognizing coal or rock dynamic tively taken from three different mining communities, each
catastrophe to make an intensive study of geophysical meth- type of which includes 3 kinds of metamorphic grade coals
ods. For the reason that LFEW, Acoustic-Emission, EMR and and tectonic coal of which are made by using special mold
geophysical methods’ theoretical principles are all based on because this kind of coal is soft broken and can not be sam-
the electrical parameters change laws of coal or rock mass pled as block specimen. In addition, all the experimental coal
in the process of coal or rock fracturing, so in one sense, samples are made into standard specimen (5 × 10 cm) by
study on the electrical parameters change laws of coal or rock using special mold to relieve the additional friction caused by
mass under load is the same as study on the internal rela- ends of coal samples.
tion between the stress condition or energy accumulation and The experimental results show that in the process of instabil-
release in the coal or rock mass and the anomalous change of ity of coal, there is a clear anti-symmetry between the change
electrical parameters of coal or rock mass in the process of law of stress-strain curve and that of resistivity with its min-
dynamic disaster breeding, which means that research on the imum value corresponding to the cracking stress of coals,
electrical parameters of coal mass under load has great the- and the resistivity of the low metamorphic grade coal and
oretical significance and application prospect for monitoring high metamorphic grade gently decrease without appearance
and forecasting coal or rock dynamic catastrophe. Study on of abrupt change no matter what the structural type of coal
the electrical parameters measurement in the lab has attracted samples belong to, however the resistivity of the medium meta-
the attention of scholars, partly because of the development morphic grade coals abruptly decrease when external load
need of the discipline and partly because it can provide micro- reaches approximately 75% and 85% of compressive strength
cosmic mechanism explanation for change laws of electrical value for the tectonic coal and the un-deformed coal, respec-
parameters measured on spots to some extent. tively, the part of which is shown in Figure 2. In addition, it
343
Figure 2. Resistivity (a) and dielectric constant (b) change curve of coal samples with outburst prone taken from Hebi mining community
in the process of uniaxial compression.
is also observed that there is a maximum value correspond- the fracture revolution laws of coal under load and providing
ing to the cracking stress of coals for dielectric constant in the information of forecasting for the instability and fail-
all dielectric constant strain curves and the dielectric con- ure of coal and the coal/rock dynamic catastrophe. However,
stant strain curves of the tectonic coal are better fitting to the further investigation is necessary in order to verify the validity
stress strain curves, which is independent on the metamorphic of the above analysis to the electrical parameters of coal/rock
grade of coal. Therefore, study on the change laws of electri- contained gas under load.
cal parameters of coal under load is promising in revealing
344
T. Yamaguchi
Toho University, Funabashi, Chiba, Japan
S. Okubo
The University of Tokyo, Bunkyo, Tokyo, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
Methane hydrate (MH) is anticipated to be a promising energy Assuming that a specimen has been fractured at the time when
resource in the near future. Time-dependent characteristics the parameter D expressed by the following equation reaches
of MH-bearing rock are essential in constructing the con- a certain value Df :
stitutive equation to predict the long-term behaviors of a
reservoir. In this study, we conducted creep tests for artificial
MH-bearing rock under triaxial compression and discussed
the time-dependent properties. Figure 1 shows the strain rate
(dεcr /dt) plotted against the elapsed time t for MH-bearing where σ* = normalized deviator stress by σCε . Based on this
∗ relationship from the depen-
hypothesis, we obtain the tcrf –σcr
rock specimens. Larger (dεcr /dt) is observed under higher σ cr .
The slope of each log(dεcr /dt)-log(t) relationship in primary dence of σ Cσ on Cσ as shown in Figure 2. The tcrf in the
creep region varies with σ cr in the range of −1 to −0.4. The creep tests was found to be approximately estimated from the
time-dependent behavior of a MH-bearing rock specimen was Cσ -dependence of σ Cσ in constant-stress-rate tests.
found to be more apparent than that of a sand specimen. Some
specimens experience the increasing (dεcr /dt) and finally
ACKNOWLEDGEMENTS
fracture. Figure 2 shows the creep life tcrf plotted against
the creep stress level σcr∗ normalized by the strength σ
Cε in This work was financially supported by the Research Consor-
a constant-strain-rate test. The tcrf tends to decrease with
∗ . We separately conducted constant-stress-rate tium for Methane Hydrate Resources in Japan in the Japan’s
increasing σcr
Methane Hydrate R&D Program by the Ministry of Economy,
tests for MH-bearing rock specimens. The strength σ Cσ in
Trade and Industry.
constant-stress-rate tests increased with the stress rate Cσ .
Figure 1. Strain rate dεcr /dt versus elapsed time t for MH-bearing
∗
rock specimens. Figure 2. Creep life tcr versus creep stress level σcr .
345
Keywords: numerical modeling, constitutive model, damage mechanics, damage yield function, strain softening
346
Compressive Tensile
fracture fracture
Bulk Shear Compressive Tensile energy for energy for
modulus modulus Density strength strength unit volume unit volume
(GPa) (GPa) (kg/m3 ) (MPa) (MPa) (KPa) (KPa)
347
K. Morimoto
National Institute of Advanced Industrial Science and Technology, Tsukuba, Ibaraki, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
3 CONCLUDING REMARKS
Figure 1. Relations between stress intensity factor and crack Figure 2. Change of stress intensity factor when crack velocity is
velocity. a: KA, b: BS, c: SS. 10−5 m/s with increasing relative humidity.
348
Subject: Rock material and rock mass property testing (laboratory and in situ)
Rock fills are used for multi-purposes in coastal engineer- fills is expected to be related to the dilatancy measured during
ing. They form the foundation base and the core of seawater shearing. It is proposed to calibrate the R and S parameters
breakers. They serve as run ways for cranes that lift-up with dilatancy measurements made during direct shear box
large armourstones to erect seawater breakers. They are testing. It is also recommended to compare the BK model
also dumped on beaches to function as dynamic sea water inspired from the mechanics of rock discontinuities to soil
defence and mitigate coastal erosion. As testing rock fills models. It would be interesting to examine whether parameters
at the field scale requires metric to decametric set-ups and recognized to be of importance for sands, such as the relative
is therefore challenging, engineers favour simple models density, the state paramter and the gradation are indirectly
that capture reasonably well the physics of the behaviour accounted for in the BK model.
of rock fill and give a first estimation of the strength of The basic friction angle is a fundamental parameter of intact
a rock fill. Barton & Kjaernsli (1981) and Barton (2008) rock that contributes largely to the strength of rock fills. It is
took advantage of the analogy between the behaviour of dis- a key parameters of the Barton and Kjaernsli model for rock
continuities and rock fills, and proposed an elegant model fills. Experimental difficulties encountered when measuring
inspired from the Barton and Choubey model for joints the basic friction of a very strong meta-sandstone with the
(Barton & Choubey, 1977), that allows to predict the strength help of a Golder shear box are described. Values measured on
of rock fills for a wide range of stresses, included the very low diamond saw cut surfaces are very low and reflect the quality
stresses. of cutting rather than the natural rock texture. Values mea-
The Barton & Kjaernsli model, named here the BK model, sured on sand blasted surfaces are low as sand blasting caused
requires the determination of few parameters, the basic friction micro-damage to the rock surface. Values measured on rough
angle of the rock material (φb ) and the equivalent particle tensile cracks after correction for dilatancy are high and higher
roughness (R) and crushing strength (S). R and S play for than 40◦ which seems to be a upper bound for the basic fric-
rock fills the role of JRC and, respectively JCS for joints. They tion angle of silicate rocks, based on results published in the
contribute to the rock fill structural strength that is caused by literature. Such high values are explained by the very tight
particle interlocking and decreases when crushing of asperities interlocking and very high strength of the small scale surface
takes place. As for rock joints, the structural strength of rock asperities.
349
Nowadays, acoustic emission testing based on the Kaiser AE signal parameter. It is also possible to use averaging, but
Effect is increasingly used for estimating in-situ stress in lab- we need to have a sufficient number of samples. Since the
oratories. The Kaiser effect can be observed in rocks and other number of initially estimated values in the proposed system
material during their cyclic uniaxial loading by AE moni- is limited, using the averaging method cannot be justified.
toring and corresponds to a stress memory. As soon as the Weighted averaging is another method that can be used. In
load achieves its previously reached peak value, a noticeable this method, each estimator output is assigned a weight in a
increase in acoustic emission activity takes place. Different range of (0, 1] which indicates its importance. The important
methods have been proposed for determining the KE point point in using this method is determining the weight of each
by different researchers. There are cases where KE cannot be output. This study uses expected pre-stress range to assign
observed clearly by plotting data and using the different meth- weights to each output. A trapezoid function has been used
ods. Consequently, the KE results can be doubtful because for weighting outputs. The background knowledge about how
determining the KE value mostly is difficult. In this paper, sure we are that the real Kaiser point is around PG appears
AE testing and monitoring of AE signals have been accom- in the trapezoid figure. Sigari et al. (2008) have proposed a
plished under uniaxial compressive testing on the cylindrical method for combining estimator outputs which works based
samples of sandstone. After each preloading stage, AE test- on clustering data. In their method which was proposed for
ing was carried out without delay by applying a loading stress more accurate estimation of the signal period time, first out-
more than the previous loading stress level. Pattern recogni- puts from different estimators are clustered, then the cluster
tion is the study of how machines can observe the environment, with the highest frequency is chosen as the winner as in the
learn to distinguish patterns of interest from their background, majority voting method. After choosing the winning cluster,
make sound and reasonable decisions about the categories of in order to determine the final output accurately, weighted
the patterns. It is the science of the methods of data analy- averaging is used with the data in this cluster. This method
sis for the purpose of separating, clustering and classifying makes it possible to use the majority voting method for com-
them based on a number of features. One of the most impor- bining estimators whose outputs are in the continuous space
tant issues in the science of pattern recognition which has because by clustering the data the continuous space changes
recently attracted considerable attention in different sciences into a discrete one. Therefore, the majority voting method can
is combining classifiers. be used to choose the winning cluster. The proposed method
One of the simplest methods for combining results obtained uses a method similar to that of Sigari et al. In our method,
from several classifiers is Majority Voting which can be used in order to cluster estimator outputs, first data histogram is
in three different ways: a) Unanimity, b) Simple Majority Vote, computed in 20 ranges with equal distances from the range
and c) Plurality Vote. For combining estimators, averaging [0,2 PG]. Thus, by computing the histogram the data are clus-
can be used. The averaging method can also be used in two tered in a simple manner. The method proposed by Sigari
ways: a) Simple Averaging and b) Weighted Averaging. We et al. uses the Isodata method for clustering; using the Iso-
can assign weights to each estimator output and thus change data or other clustering methods in the current application
their effect on the final output. This method is called weighted can lead to producing clusters with different sizes in the range
averaging. Simple averaging is a type of weighted averaging in [0,2 PG]. Producing clusters with unequal sizes can lead to
which all the estimators have the same weights. There are two inaccurate estimation of the KE point. Therefore, we need to
ways to assign weights to estimators. The first is the manual use the histogram method and cluster the estimator outputs
weighting method in which each estimator is weighted based into the clusters. Now, each cluster is assigned a weight and
on an expert individual’s background knowledge. The second the weighting system is the same as in weighting data which
method is the automatic method in which a criterion is set was mentioned earlier. In other words, using a trapezoid func-
for estimators’ level of effectiveness. Then, each estimator is tion the clusters are weighted in such a way that those closer to
weighted based on this criterion. The most important stage in the center are assigned a greater weight. After weighting the
the proposed method is the combination of estimator outputs clusters, the cluster whose frequency multiplied by it weight
from different parameters. gives the highest value is chosen as the main cluster. If the
In the proposed method, outlier data could occur as a result center of this cluster is chosen as the final estimation, the error
of an estimator’s error in estimating the KE point from a certain of estimation could be great. In order to determine the final
350
351
Acoustic emission has been extensively utilized over the past capabilities. For all AE signals received the AE features were
decades to quantify microstuctural damage of rock both in determined automatically by the recording system.
laboratory and field studies. In this study the acoustic emission In Figure 2 typical stress-strain diagrams obtained from the
signature of Dionyssos marble under uniaxial compression is uniaxial compression tests of the prismatic samples is shown.
examined, and the stages of damage are identified combining For this test, a preload of 20 kN, corresponding to a stress of
acoustic emission and deformation measurements. Further, 8 MPa, was applied. In these diagrams the axial stress differ-
the evolution of some AE parameters with the applied load is ence σa -σa,preload is normalized with the maximum axial stress
examined whereas special attention is given to their usefulness difference σ max -σ a,preload obtained during the test. The axial
as precursors of the forthcoming rock failure. and lateral strain differences (ε-εpreload ) are also normalized
The Dionyssos marble is a calcitic medium crystalline with respect to their values at the maximum stress (εmax -
marble, with mechanical properties varying with the angle εpreload ). Also, the normalized volumetric strain is shown in
of the loading direction to the main plane of anisotropy. the diagram.
The marble specimens prepared for the uniaxial compres- In Figure 3 the normalized cumulative AE (excluding
sion tests are either prismatic or cubic with a cross section of any the events recorded during the preload stage) is shown
50 × 50 mm2 and 100 × 100 mm2 respectively, and 100 mm with respect to the normalized axial stress σa /σmax . In the
height. The specimens were loaded under constant loading same diagram the normalized axial strain-axial stress curve is
rate until macroscopic failure. The applied load, the axial and shown.
lateral strains were continuously monitored simultaneously From the diagrams of Figures 2 and 3 the elastic and crack
with the AE monitoring. development stages may be identified. The crack initiation
To capture the AE signals piezoelectric transducers were stress is estimated by the onset of the acoustic emission activ-
used. The transducers were attached on the surface of the ity at a normalized stress level of about 0.3. The crack damage
specimens using silicon vacuum grease as a coupling agent threshold is estimated at a normalized axial stress difference
(Fig. 1). Before each test, coupling of the sensors was exam- of about 97%. This stress level corresponds to the volumet-
ined using the PLB (pencil lead break) technique. Wideband ric strain curve reversal and to a drastic increase of AE rate
AE sensors were used. The AE signals were preamplified as shown in Figure 3. The main increase of AE activity is
by external amplifiers, with 40 dB preamplification. The AE observed near the macroscopic failure point.
triggering threshold was set to 50 dB. In each case full wave- Energy released by damage during loading may be deter-
form data were recorded for each AE signal with a multichan- mined either by rigorous or simplified approaches. A sim-
nel data acquisition system with real time signal processing plified measure of the released energy is the ASL (Average
Figure 1. Photograph of a marble prism tested in uniaxial compres- Figure 2. Typical normalized stress-strain diagrams obtained from
sion and the AE transducers mounting. the uniaxial compression tests of the prismatic marble specimens.
352
Signal Level) which is a measure of the continuously varying Figure 6. Normalized b-values with respect to the normalized load
and “averaged” amplitude of the AE signal. In Figure 4 the calculated for the prismatic marble specimens.
normalized ASL values (divided by the maximum ASL value)
of the AE signals for the prismatic marble specimen #1 with
respect to the normalized axial stress are shown. An increas- after 95% of the maximum load. This may be an indicator of
ing (almost linear) trend of ASL values with the normalized the damage state of the specimen. Also, this point corresponds
load is observed up to a normalized load of approximately to accelerated AE activity observed in the experiments and
0.95. An abrupt increase of the slope of a hypothetic bi-linear to the abrupt increase of the ASL values.
envelope is observed at 97% of the normalized axial stress. In conclusion, for the Dionyssos marble specimens tested
Similar response may be observed for the other prismatic spec- under uniaxial compression the onset of the AE activity is
imens tested. For the cubic specimens tested a precursory trend observed at a stress level lower than 30% of the maximum
increase is only identified very close to macroscopic failure. load. The crack damage threshold may be correlated with the
In Figure 6 the b-values calculated for the prismatic marble ASL values of the recorded AE signals. An abrupt increase
specimens #1 to #3 are shown with respect to percent loading. of the ASL-loading curves is observed at this stress level.
An increasing trend of the b-values up to approximately 90% A significant percentage of the AE activity and an appar-
of the maximum load is shown. Thereafter, b-values decrease. ent precursory drop in b-value, obtained from an amplitude
In all cases, an apparent precursory drop in b-value is observed distribution analysis, are also observed at this load.
353
Keywords: lab testing, neural network, problem rocks and rock properties
1 INTRODUCTION Table 1. Descriptive statistics of the rock sample data for 126 sets
(including training and test sets).
Mechanical properties of rocks such as uniaxial compressive
strength (UCS) and modulus of elasticity (E) of intact rocks Parameter Mean Min Max Std. dev
are of vital importance in mining and civil activities. Rock Vp (m/s) 5.34 2.2 6.82 1.03
engineers extensively use the UCS and E in designing tun- SHR 36.12 18.22 58.9 7.69
nels, foundations, slopes and other rock engineering projects. T (MPa) 8.7 3.2 20.15 3.28
In addition, analytical and numerical solutions require both IS(50) (MPa) 7.57 1.96 17.61 3.46
parameters. Measuring these parameters has been standard- Dry (gr/cm3 ) 2.51 2.05 2.77 0.16
ized by both the ISRM (1981) and ASTM (1984) which based Sat (gr/cm3 ) 2.56 2.22 2.79 0.12
on these standards, UCS test requires specimens prepared n (%) 5.03 0.15 27.3 6.21
accurately but it is often extremely difficult and time con- UCS (MPa) 61.72 3.36 133.54 25.62
suming to obtain such samples from weak, highly fractured E (GPa) 6.45 1.99 13.24 2.42
and thinly bedded rocks. Therefore, some predictive models
utilizing simple index test such as schmidt hammer, point load
index, sound velocity, physical properties, and other tests were
developed by many researchers (Hassani et al. 1980,Alvarez & desired output of models and other parameters (engineering
Babuska 1999, Meulenkamp & Alvarez 1999, Kahraman index tests) are input parameters.
2001, Lashkaripour 2002, Sonmez et al. 2004, Gokceoglu &
Zorlu 2004, Fener et al. 2005, Karakus & Tutmez 2006,
Tercan & Ozcelik 2006) owing to the fact that these index 3 NETWORK ARCHITECTURE
tests require less or no sample preparation when compared
with the uniaxial compressive strength test. In addition, they The prediction capability of the optimizedANN used for deter-
can be used easily in the field. mining the mechanical properties of rocks is assessed using
This paper presents the application of a neural network for different data pairs than those that were not used for training
the simultaneous prediction of the UCS and modulus of elas- the ANN.
ticity from index tests. To investigate the suitability of this After many trials, the optimal ANN is determined to be a
approach, the results of the network were compared with pre- model with a 7-25-15-2 ANN architecture, that there are two
dictions obtained by conventional statistical relations. For this hidden layers with 25 and 15 nodes in each layer for predicting
purpose different rock types have been tested based on ISRM the mechanical properties of rocks. The criteria used to deter-
suggested methods. These tests consist of point load, sound mine the optimal ANN model is the lowest root mean square
velocity, schmidt hammer, porosity, tensile strength, uniax- errors (RMSE) of the predicted values.
ial compressive tests, modulus of elasticity, dry and saturated Figures 1, 2 show the estimated values versus measured val-
density. ues for UCS and E from optimized ANN model. They indicate
that designed ANN model has sufficient ability to predict the
desired variables.
2 EXPERIMENTAL INVESTIGATIONS
4 COMPARISON WITH CONVENTIONAL METHODS
Rock blocks have been collected and 126 sample sets for rock
mechanics tests were obtained. Each data set consists of UCS, Several regression models were constructed in order to assess
E, point load index (IS(50) ), porosity (n), sound velocity (Vp), the ability of ANN models. Linear, power, logarithmic, and
schmidt hammer (SHR), tensile strength (T), dry and satu- exponential functions were employed so as to build a uni-
rated density. The procedures suggested by ISRM [2] for rock variate regression models. Strongest simple correlations were
mechanics tests were considered. Table 1 lists the descriptive obtained between UCS and point load index (R2 = 0.54) and
statistics of various parameters. In this paper UCS and E are between E and dry density (R2 = 0.46). Afterwards several
354
Figure 2. Predicted versus measured E from ANN. Figure 4. Predicted and measured value of E for MR.
multiple regression (MR) models were performed in the sec- values to predict the UCS from ANN models on test data sets
ond stage of the regression analyses (Figs 3, 4). Another were 9.64 and 7.96 respectively. These indices were 1.26 and
method namely canonical correlation has been developed in 0.95 for modulus of elasticity. As seen from Figures 1, 2, the
the next step. Whereas multiple regression predicts a sin- ANN model shows a good relation with high correlation of
gle dependent variable from a set of multiple independent coefficient (r = 0.89 for UCS and r = 0.87 for E). Consider-
variables, canonical correlation simultaneously predicts mul- ing performance indices and mentioned figures the prediction
tiple dependent variables from multiple independent variables. capacities of ANN to estimate UCS and E of intact rocks are
Some researchers utilized this method to study the relation- higher.
ship between mechanical properties of rock and engineering In addition, considering all indices, improvement can bee
index tests (Tercan & Ozcelik 2006, Noorani & Ghazvinian seen in ANN model rather than univariate models and also
2008). multiple regression. According to the results, artificial neural
Several indices were calculated to control the capacity of network is a robust means to estimate mechanical properties
the developed equations in this study. The RMSE and MAE of intact rocks simultaneously.
355
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 INTRODUCTION
2 METHODOLOGY
Figure 1. Relation between peak radius and maximum volumetric
shrinkage.
2.1 Drying experiment
Laboratory experiments of drying were carried out with the
cylindrical core samples. Four strain rosettes are attached on 3 RELATION BETWEEN MAXIMUM VOLUMETRIC
side surface of a cylindrical specimen in a symmetrical manner SHRINKAGE AND PEAK RADIUS OF PORE SIZE
against the cylindrical axis in order to measure successive DISTRIBUTIONS
deformation with the change in direction of principal strains.
Strain variations and weight loss of a sample were con- If the mechanism of drying-induced deformation is related
tinuously measured with logging of temperature and relative to the pore structure of a rock, pore size distribution must
humidity of environment up to that no weight loss is recorded. have adequate information. Then, as the first approximation,
The maximum shrinkage values in volumetric strain for four the relation between peak radius and maximum volumetric
rock types are different for each rock type. In case that vol- shrinkage is plotted on log-log scale as shown in Figure 1.
umetric strain does not reach to a steady state even after 15 All data are plotted on the right-upper half space of this
days passing like Opalinus clay and Shirahama sandstone, the graph and the smaller peak values seem to be lined up.
final point is taken as the maximum shrinkage value. The maximum shrinkage can be roughly estimated using this
relation.
2.2 Mercury intrusion porosimetry
Mercury intrusion porosimetry was used to evaluate the pore 4 SUMMARY
structures of four rock types. The absolute pressure applied
to inject the mercury into the dried sample is used to evaluate In this paper, we introduce drying-induced deformation expe-
the pore size from Young’s equation. riments performed in our laboratory. In addition, the first
Opalinus clay and pumiceous tuff have unimodal pore size attempt has been made to estimate the maximum shrinkage
distributions with different peak values, while Shirahama values by using MIP results.
sandstone and Tage tuff have bimodal ones. The ranges of Change in mechanical properties during drying is also
each peak radius for four rock types are summarized with the reported in this paper. The desiccation-driven hardening of
median values. Tage tuff is clearly shown.
356
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: lab testing, problematic rocks, rock properties, weak rock, clay mineral
Fault is a geological crack in the earth’s crust resulting from the located from fault planes, contain more clay minerals. Slak-
displacement of one side with respect to the other by being ing durability test and reflectance spectroscopy wasapplied to
forced, and crushed rocks around a fault plane show weak confirm assumption (Fig. 1).
physical properties. Among materials located between fault We collected rock samples from faults located inside of ◦◦
planes, especially, clay minerals, derived from weathering of mine, Samcheok-Si, Kangwon-Do, South Korea. Three kinds
bedrock or existence inside rocks, the clay has been become of samples were collected from fault gauge located between
one of the biggest problems during underground construction fault planes, rocks of fault planes, and fresh rocks located
by combining with water. For this reason, investigation of clay 1 m∼2 m apart from the fault planes (Fig. 2). Among many
minerals is essential to improve stability of underground con- faults inside the mine, 3 faults were selected and sampled.
struction because clay minerals exist inside of rocks lead to WG1, 2, and 3 were given to each fault, and then subnumber
not only physical expansion but also shotcrete malfunction 1, 2, and 3 were named to classify the samples according to
and injection efficiency reduction by bonding with water. distance from fault planes.
Slake durability test (SDT) can estimate rock’s durability SDT suggested by international society for rock mechanics
against slaking phenomenon whichis observed easily for clay- (ISRM) was used, and total 4 cycles were repeated to measure
ish rocks due to repeat drying and wetting. In this study, we slake durability index (SDI). Collected rock samples were cut
used X-ray diffraction analysis (XRD) to classify clay miner- to 10 spheroidal lumps of 40∼60 g for each specimen, total
als in rock samples collected and assumed that the closer rocks weight of 450∼550 g. Ten lumps were placed in meshed drum
Figure 2. Left, location of the OO mine in South Korea. Right, rock sampling area around fault in the mine.
357
ACKNOWLEDGEMENT
Reflectance spectroscopy is a non-destructive and in-situ
analysis method which can be applied to geomaterials, and This work was supported by the National Research Foundation
has been successfully applied to detect the clay minerals of Korea (NRF) Grant (NRF-2010-616-C00044).
358
Keywords: lab testing, rock failure, rock joints, rock properties, rock stress
359
Young-Ho Park
Department of Energy and Resources Engineering, Chosun University, Gwangju, Korea
Yuzo Obara
Graduate School of Science and Technology, Kumamoto University, Kumamoto, Japan
Seong-Seung Kang
Department of Energy and Resources Engineering, Chosun University, Gwangju, Korea
Keywords: uniaxial compressive strength, weak rock, needle penetration resistance, curing period
The various mechanical properties of rock materials, which indication ring, the UCS-NPR correlation chart, the remov-
are commonly used in rock engineering, are required for ana- able cap, and the penetration needle. Needle penetrometer can
lyzing and evaluating the stability of geo-structures such as be adapted for indirectly estimating the uniaxial compressive
tunnels, dams, and slopes. In particular, the uniaxial compres- strength of weak rocks from the needle penetration resistance
sive strength (UCS) has been used as one of the important (NPR) without sampling. The needle penetration resistance
mechanical properties in rock materials. The uniaxial com- can be calculated by an equation (1) as the ratio of load (F)
pressive strength of rock materials can be generally obtained and depth of penetration (D). An equation (2) was suggested
by the convensional uniaxial compression test. This conven- by a manufacturer from the correlation between the uniaxial
tional testing method can be obtained the value of exact compressive strength and needle penetration resistance, based
uniaxial compressive strength, whereas a lot of time and effort on data from artificial and natural rock specimens.
to prepare specimen for its tests are needed. In addition, a test The purpose of this study is to estimate the uniaxal compres-
specimen of the appropriate size demanded for the uniaxial sive strength for the weak rocks from the needle penetration
compression test is sometimes difficult to manufacture. Due resistance measured by needle penetrometer. For this purpose,
to these problems, some testing methods such as the point load we firstly made the artificial cemented-based specimens and
test, Brazilian test, Schmidt hammer test, and block punch curing period of each specimen is six steps with the ratio of
index test have been developed and suggested for estimating maxed materials of three cases. Then, the needle penetrometer
the uniaxial compressive strength. However, these methods are test and uniaxial compression test were performed to obtain
also difficult for preparing the appropriate testing specimens, the needle penetration resistance and the uniaxial compres-
especially in the case of weak rocks. That is, during coring sive strength. Secondly, the relationship between the measured
process for weak rocks, core specimens for the experiment needle penetration resistance and the estimated and measured
can be splitted into smaller sizes. uniaxial compressive strengths was compared. Finally, the
For solving above difficulties, a new testing device has been uniaxial compressive strength was estimated. In relationship
developed by a manufacturer of Maruto Corporation in Japan between NPR and UCS, the UCS gradually increased until
as non-destructive and portable testing methods for laboratory the NPR was less than 10 N/mm, then drastically increased
and field. This device is called needle penetrometer applied from more than 10 N/mm. The NPR was widely distributed
for weak rocks, and consists of eight parts, namely presser for at the curing period of 3-day to 14-day, but was narrowly dis-
measurement of the penetration and protection of the penetra- tributed at the curing period of 21-day to 28-day. In addition,
tion needle, the chuck used for fixing the penetration needle, the NPR of the curing period between 21-day and 28-day was
the penetration scale of 0∼10 mm, the load scale, the load nearly constant, but the UCS during those periods showed a
360
Curing NPR UCSe UCSm NPR UCSe UCSm NPR UCSe UCSm
Period (N/mm) (kN/m2 ) (kN/m2 ) (N/mm) (kN/m2 ) (kN/m2 ) (N/mm) (kN/m2 ) (kN/m2 )
3-day 1.86 0.77 0.49 2.87 1.18 1.02 4.94 2.02 1.59
7-day 4.97 2.03 1.66 6.03 2.46 2.67 11.03 4.47 3.50
14-day 7.63 3.10 3.48 8.70 3.54 3.87 14.00 5.66 5.22
21-day 9.25 3.76 4.06 11.30 4.58 4.68 16.71 6.74 6.91
28-day – – – 11.57 4.69 6.86 17.26 6.96 8.92
361
Gerhard Pittino
Chair of Subsurface Engineering, Department Mineral Resources and Petroleum Engineering, University of Leoben, Austria
Yuewei Bian
Department of Geotechnical Engineering Tongji University, Shanghai, P.R.C
362
2.2 Executed slab tests The comparison of the plate stiffnesses as expected leads to
very different magnitudes, expressed in force per vertical dis-
First of all, a reference test was carried out accordingly to the placement. The analytical solution is for a simply supported
standard ÖNORM EN 14488-5. Therefore a series of three plate and thus exaggerates the stiffness. The analysis yields
slabs was concreted with the addition of synthetic fibers as approx. 600 kN/mm. If the bedding material is modeled very
reinforcement. Because of the influence of the bedding mate- stiffly in the numerical analysis, we get obtain stiffness of
rial on the test result, which also should be investigated, the approx. 450 kN/mm. The stiffness found from the slab test
slabs were milled at the top side and underside and tested is for different bedding materials between 50 kN/mm and
without bedding material. 100 kN/mm.
In addition plain concrete slabs were tested for compari- An annular support and a square or circular slab was pro-
son purposes. Some slab tests are executed very fast with a vided in the draft of the ÖNORM EN 14488-5, issue 2003.
displacement rate of 10 mm per minute. The Swiss Standard SIA 162/6 (SN 562162-6, February 1999)
Numerous numerical analyses with FLAC3D show the influ- still specifies the slab test with round specimens and support
ence of the support condition on the failure pattern. The in addition to the test with square specimens and support.
analysis results are validated by laboratory tests using the ASTM C 1550-03a suggests a round slab supported on three
same support conditions as shown in Fig. 4. It appears that symmetrically arranged pivots. Therefore the geometric mod-
the supporting conditions have a significant influence on the els shown in Fig. 4 are examined.
results. Some results of the slab tests and numerical analyses are
described in the extended version of this contribution.
3 ANALYTICAL AND NUMERICAL ANALYSES
363
S.A.L. Read
GNS Science, Lower Hutt, New Zealand
L. Richards
Rock Engineering Consultant, Canterbury, New Zealand
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 INTRODUCTION the stress space would be σ1 < σ3 < σ1 /3.4 on the basis of
Mogi’s original definition of the brittle/ductile transition.
The original Hoek-Brown failure criterion (Hoek & Brown Subsequently Hoek & Brown (1997) recommended that the
1980) was defined in terms of major and minor total stresses range should be 0 < σ 3 < 0.5σ ci .
as follows: Sheorey (1997) advocated the inclusion of tensile strength
data along with unconfined and triaxial compression data, as
better regression is obtained with data point(s) closer to the
σ 3 axis of the principal stress plots.
The calculation method to analyse the testing data has
where the relationship between the principal stresses (σ 1 and an effect on the mi values, with the RocData and RocLab
σ 3 ) at failure is defined by two constants (m and s) and the programs from RocScience (2010) using the non-linear
uniaxial compressive strength of intact rock (σ ci ). Levenberg-Marquardt method of least-squares curve fitting as
In the case of intact rock material where s = 1.0, Hoek & the default, with options for the simplex reflection and linear
Brown (1980) noted that m ≈ σ ci / |σ t | where σ t is the uniaxial regression methods.
tensile strength of the intact rock.
The generalized failure criterion (Hoek & Brown, 1997) in 2.1 Guideline values for mi
effective stress terms included three constants (mb , s and a) in
As part of the original failure criterion, Hoek & Brown (1980)
the equation as:
suggested that m for intact rock (mi ) reflected rock type,
with values increasing ≈7 for carbonate rocks to ≈25 for
coarse-grained rocks. Some 15 years later, Hoek put forward
guideline mi values, based on a more detailed lithological
classification, with subsequent revisions including ranges of
In the case of intact rock material (i.e. s = 1) the generalized values.
equation becomes:
2.2 Values for mi calculated from test result databases
Valuable databases of laboratory test data have been provided
by Sheorey (1997) and Douglas (2002). In general, the cal-
culated database values show little correlation with guideline
The full paper first summarizes procedures used to calculate values, with calculated values having a much greater range
intact material constant mi and then compares values obtained than those in the guidelines.
from a large database of testing data with guideline tables. Correlations of mi with other rock parameters, (such as
After considering further testing results, suggestions are made unconfined compressive strength) are mostly very poor.
for a simple method of estimating mi by means of a combina-
tion of unconfined compression and indirect Brazilian tensile 2.3 Estimates of mi values using tensile testing
tests.
The only rock material parameter with which mi has any rea-
sonable correlation is the ratio of the unconfined compressive
2 CALCULATION OF mi VALUES strength σ c to the tensile strength σ t . Sheorey (1997) defined
this ratio as R = σ c /|σ t .|.
The Hoek-Brown failure criterion is intended for use in the The laboratory tensile data in the Douglas database com-
brittle range of rock behaviour, and this controls the range prised results from both pure uniaxial and indirect (Brazilian)
of minor principal stress (σ3 ) values over which triaxial test- testing. Using these data (59 of the 521 sets in the database),
ing to determine mi is carried out. Initially Hoek considered the intact rock parameter mi , shows a fairly good relationship
364
4 CONCLUSIONS
REFERENCES
365
M. Romana
Technical University of Valencia, Spain
Subject: Rock material and rock mass property testing (laboratory and in situ)
366
367
Effect of heat on the mechanical properties of selected rock types – a laboratory study
C. Saiang
Luleå University of Technology, Sweden
K. Miskovsky
Luleå University of Technology, Sweden
Envix Nord AB, Sweden
Subject: Rock material and rock mass property testing (laboratory and in-situ testing)
Keywords: effect of heat, increased temperatures, rock mechanical properties, mineralogical changes, micro-cracks, micro-
cracks distribution
A laboratory study was conducted to study the effect of of the rocks at the different heat treatment temperature levels,
heat on the mechanical properties of diabase, granite and determine the micro-crack intensity within the major minerals
quartzitic schist at temperatures of 400◦ C, 750◦ C and 1100◦ C. and in the rock specimen as a whole which assisted in explain-
Unheated samples were also studied. The reasoning behind ing the reasons for the mechanical behaviour observed from
this study was to understand the effect of elevated temper- the mechanical test results. Scanning Electron Microscopy
atures on the rock mass, such as in the event of a fire in combined with Energy Dispersive Spectrometry (SEM/EDS
a rock tunnel. Other authors who have undertaken similar method) was employed to identify mineral phases with grain
studies include for example, Smith & Pells, (2007), who stud- size below the resolution limit of the polarizing microscope.
ied how heat exposed sandstone is likely to respond in a The uniaxial compression test showed that the strength of the
tunnel fire by means of laboratory and field tests, Hajpal rock specimens increased by 6% for granite to 29% for diabase
(2002), who presented results of the changes in the physi- at 400◦ C when compared to the UCS values of the unheated
cal and mechanical properties of sandstone when heat treated specimens. From 750◦ C to 1100◦ C the decay in the strength
at temperatures up to 900◦ C and Zhang et al. (2010) who was very rapid. From the microscopic analyses it was con-
conducted experimental study of thermal spalling of rock cluded that the increase in the strength of the rock specimens
blocks exposed to fire. Samples from the aforementioned rock at 400◦ C is attributed to the initial reaction of the rock forming
types were heat treated at controlled low heating rates, 0.8◦ C minerals, hence the rock specimens were less brittle but more
to 1.5◦ C per minute in an electric furnace at temperatures plastic. The rapid drop in the strength from 750◦ C to 1100◦ C
shown above. After heat treatment at the desired tempera- is attributed to the mineralogical changes, micro-cracking and
ture levels the rock samples were allowed to cool slowly to dehydration due to the loss of crystal bound water. At 1100◦ C
room temperature and then subjected to uniaxial compression the rocks were highly friable and crumbled very easily when
and Brazilian tests. The uniaxial compression was performed tested mechanically. The effect of mineralogical changes was
using a servo controlled hydraulic testing machine at a dis- obvious in diabase where the physical appearance of the sam-
placement rate of 5 µm/s. For the Brazilian tests the disk ples mimicked that of natural iron, which is believed to be
shaped specimens were loaded using a 0.5 MN capacity servo due to the alteration of pyroxene. The result was an increase
controlled Instron machine at the same rate as for the com- in strength by 29% at 400◦ C compared to the unheated spec-
pression tests. Thin sections were extracted from the heat imens. Even at 750◦ C the strength was slightly higher than
treated samples for microscopic analyses. The microscopic unheated specimens. In summary; the mechanical behaviour
analysis conducted employed polarizing microscopy com- of the rock specimens depended on the temperature level and
bined with petrographic digital image analysis technique to the mineralogical and physical changes that occurred at that
identify micro-cracks and determine the mineral compositions temperature.
368
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: weak rock, rock properties, lab testing, field measurements
369
Keywords: modulus of rupture, flexural strength, Building stones, Granite, Calcite, Travertine
Estimation of tensile strength of stones is very important in between correlation of linear and nonlinear equations of
industry, because its measurement using direct tensile strength Calcite stone is approximately 5%. All stones of Sannandaj –
test is very difficult. 3 point flexural loading and four point Sirjan zone are classified in three groups consisting of gran-
flexural loading are two methods for measurement of flexural ite, calcite and travertine. The relationship between flexural
strength. The flexural strength is the maximum tensile stress strength and modulus of rupture in each group is considered.
on the bottom of the rock corresponding to peak load. It is cal- Relationship between flexural strength and modulus of rup-
culated from simple beam theory assuming elastic conditions ture are direct but difference between correlation of linear
throughout. ASTM C880-06 covers the procedure for deter- and nonlinear equations on travertine and Granite groups are
mining the flexural strength of stone by using a simple beam smaller than Calcite groups. All samples of Sahand – Bezman
using quarter-point loading. The flexural strength is found to zone, are classified in three groups consisting of granite, cal-
be greater than direct tensile strength. ASTM C99-87 is used cite and travertine, then relationship between flexural strength
for measurement modulus of rupture of building and deco- and modulus of rupture upon granite stones was considered
rative stones. To prevent various names of stones are used but it was not considered upon calcite and travertine stones
in different classifications, the Dunhan System for notation because samples were not enough. Relationship between both
of sedimentary stones the volcanic stones, it is considered the of the strength properties are direct and difference between
main and subsidiary materials, and also the tissue of the stones. correlation of linear and nonlinear equations of Granite is very
Based on obtained results from these tests all decorative little. Therefore, the linear equation which is simpler can be
stones in this project can be divided such as that mentioned. preferred. Taken samples from the quarries have been cut in
Based on geological map of Iran, it is divided in different cutting workshop and sent to the laboratories for tests. ASTM
nine zones such as Alborz, Central Iran, Zagros, Sannandaj – standards are considered during sampling. Based on standard
Sirjan, Mokran, Sahand – Bezman, Loot block, Eastern Iran, codes, physical and mechanical tests should be done in each
Kopedagh.The Building and decorative stones are one of those sample. All tests are carried on bases ASTM standards in lab-
important natural resources of Iran. According to recent statis- oratory of construction research center of Iran. The purpose
tics, Iran has been listed among the five important producers of this paper is considering relationship between the modulus
of building and decorative stones in the world. Approximately of rupture with the flexural strength on three types of building
all of decorative and building stones located in the following stone as Granite, Calcite and Travertine groups in Iran. Based
five zones such as Alborz, Central Iran, Zagros, Sannandaj – on results relationship between both of properties is direct
Sirjan, Sahand – Bezman. All available data of Alborz zone and difference between correlation of linear and nonlinear
is only four, so it is not enough to set equation upon this. All equations of stones is very little it is possible that is assumed
stones of Central Iran zone are classified, in two groups, gran- relationship between modulus of rupture and flexural strength
ite and calcite. Relationship between both strength on Granite of Granite, travertine and Calcite stones is linear and direct.
and Calcite groups is linear but difference between correlation Ultimately, relationship between modulus of rupture and flex-
of linear and nonlinear equations of Granite group is smaller ural strength for all samples are considered without assuming
than Calcite group. All of stones in Zagros zone are classified type of stones that it is direct and linear. The reason is usage
in Calcite stones group. In this case relationship between flex- of linear equation, because applicability of linear equations is
ural strength and modulus of rupture is direct and difference simpler than nonlinear equations.
370
A.K. Shrivastava
Department of Civil Engineering, Delhi Technological University, Delhi, India
Subject: Rock material and rock mass property testing (laboratory and in situ)
The shear behavior of rock joints greatly influence by the pres- affecting the shear behaviour of rock joint. The influence
ence of asperity and the normal stiffness of the surrounding of these parameters on shear strength of jointed rock has
rock mass. Correct evaluation of the shear strength of rock been studied by different researchers such as Patton (1966),
joints plays an important role in the design of deep under- Ladanyi & Archambault (1977), Barton (1973 and 1976),
ground openings, stability analysis of rock slopes, socketed Barton & Choubey (1977), Indrartna & Haque (1998, 1999),
piles in rock and anchored rock slopes. To study influence of Yang & Chiang (2000) and Weildeniya (2005). To study the
stiffness of the surrounding rock mass on the shear behav- influence of normal stiffness of the surrounding rock mass on
ior of non planar rock joints, test were performed on an the shear behavior of the rock joint, tests have been performed
automated servo controlled large scale direct shear testing on non planar joints of asperity angle of 15◦ -15◦ of 5 mm
machine on rock, which has been developed and fabricated thickness under CNS conditions at different normal stiffness.
at IIT Delhi. This equipment is capable of performing the A detailed account of shear deformation behaviour of non pla-
test to study the effect of joint roughness, scale, stiffness nar rock joint under different normal stiffness is discussed in
of the surrounding rock, shear rate, condition of the joint this paper.
i.e. unfilled joint/infilled joint, infill type, and infill thickess,
371
L.O. Suarez-Burgoa
Department of Civil and Environmental Engineering, University of Brasilia, Brasilia, Brazil
Faculty of Mines, National University of Colombia, Medellin, Colombia
Subject: Rock material and rock mass property testing (laboratory and in situ)
In literature one can find plenty of formulas that relate the presented in a graphic which shows up to 100% of estimation
point load index (Is(50) ) with the uniaxial compressive strength error. It is finally recommended avoid the use of the point
of rock material (σci ). In this article it is shown that the uniax- load index as a unique tool to assess the uniaxial compressive
ial compressive strength of rock material obtained with these strength, even though the equipment and the method to obtain
kinds of empirical formulas may be in most of the cases unre- it give to the engineer an economic and in-hand information
liable. From 37 formulas proposed by different authors, values of σci . In contrast, it is incentivized to use directly the uniaxial
of Is(50) ranging between 1 to 20 were used as input values in compressive tests of rock material, because it can give many
order to obtain σci . The result of the response values of σci was other important and reliable values apart of the strength.
372
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 EXTENDED ABSTRACT
373
B.G. Tarasov
The University of Western Australia, Perth, Western Australia
Subject: Rock material and rock mass property testing (laboratory and in situ)
374
T. Tsutsumi
Department of Urban, Environmental Design & Engineering, Kagoshima National College of Technology,
Kirishima, Japan
Subject: Rock material and rock mass property (laboratory and in situ)
The diametrical compression test is a simple and relatively two principal elastic axes) and the direction of the principal
inexpensive test for measuring the tensile strength of brittle elastic axis. Therefore, the orthotropy ratio and the direction
material. This test is performed by placing a disk between two of the principal elastic axis are needed to estimate the tensile
(rigid) plates and applying a diametrical compressive load. strength.
The test induces a biaxial stress state in which the stress at the In this study, the relation between the angle between the
center of the disk is compressive in the x-direction (σ x ), and maximum principal strain and the loading axis, and the angle
tensile in the y-direction (σ y ). Theoretically, for an isotropic between the principal elastic axis and the loading axis, is calcu-
material, the tensile stress reaches a maximum at a constant lated using the theoretical solution, and the results are shown
magnitude of P/(πR), where P represents the applied load as a graphical representation for the diametrical compression
and R represents the radius of the disk (Sokolnikoff, 1956) test as shown in Figure 1. The relation between the absolute
(Timoshenko, 1970). Because tensile strength is smaller than value of the ratio of the normal strain in the loading direction
compressive strength for many brittle materials, the mate- to that in the orthogonal direction at the center of the disk,
rial undergoes tensile failure first. However, for orthotropic and the angle of the principal elastic axis on the loading axis
materials, Lemmon et al. (Lemmon, 1996) and the authors is also shown in graphical form as shown in Figure 2. Fur-
(Tsutsumi, 2000) showed that the distribution of tensile stress thermore, by using the strain values obtained using Rosette
on loading radius is not same as that for isotropic materials. gages in the diametrical compression test and these graphi-
According to these researchers, the maximum tensile stress cal representations, a method of determining the orthotropy
which occurs on the loading radius depends on the orthotropy ratio and the direction of the principal elastic axis was
ratio (the ratio of eachYoung’s modulus in the directions of the shown.
Figure 1. Relation between strain ratios and direction of maximum Figure 2. Relation between the direction of principal elastic axis E1
principal strain. and the direction of maximum principal strain.
375
Benoît Valley, Bo-Hyun Kim, Fidelis T. Suorineni, Navid Bahrani & Rob P. Bewick
MIRARCO/Geomechanics Research Center, Sudbury, Canada
Peter K. Kaiser
CEMI – Center for Excellence in Mining Innovation, Sudbury, Canada
1 INTRODUCTION
376
Figure 2. Sigmoid failure envelope for rock masses with hypo- using smaller W /H ratios and still insuring the carrying capac-
thetical reduced degradation under high confinement overlain on ity of the pillar, and 2) pillars that would be assumed to be in
laboratory strength data of quartzite with variable intensity of
the yielding state based on the Hoek-Brown/GSI failure enve-
micro-defects and confinement.
lope could have a core that actually is in pre-peak conditions,
rough joints will be strongly interlocked and failure through i.e. still accumulating stress and potentially entering into a
intact rock (shearing of asperities and coalescence through burst prone state.
rock bridge failure) will be controlling the strength of the
rock mass. Based on this, we hypothesize that the degradation 4 CONCLUSIONS
factors back-analyzed from inner shell behavior underestimate
the rock mass strength in the outer shell. This is also supported The change of failure mode of rock and rock masses with
by the analysis of the laboratory test data presented in Fig. 2 increasing confinement has been recognized for a long time,
of a quartzite with variable intensity of micro-defects, that but assessing its impact on the strength of intact rock and
is considered as an analog for rock mass. These results sug- particularly of rock masses is not embedded in commonly
gest that there is significant strength increase under confined used failure criteria. In order to remedy this situation, we pro-
conditions compared to the Hoek-Brown/GSI relationships. pose to differentiate the approach for strength determination
For these reasons, we recommended to use, for design, a dif- when solving problems affected by low confinement, what
ferentiated approach to obtain the failure envelope and to we call inner shell problems, and problems controlled by high
assess potential failure modes in the inner or the outer shell. confinement or outer shell problems (Fig. 3).
Such differentiated approaches are discussed in the following
sections.
ACKNOWLEDGMENTS
3.1 Inner shell problems: tensile failure leading to spalling
This research is supported by CEMI’s industrial sponsors, the
In the inner shell, under low confinement, the failure mode Government of Ontario through its Ministry of Research and
will be dominated by tensile cracking leading to spalling and Innovation and by NSERC (Natural Sciences and Engineering
related geometric dilation processes. A number of field studies Research Council of Canada).
have shown that for brittle rocks, spalling initiates at tangential
stress levels significantly below the UCS (down to 0.4UCS).
The reason for such a difference in strength between labora- REFERENCES
tory and in situ measurements still remains uncertain as in
both cases the same mechanisms, unstable tensile crack prop- Kaiser, P. K., Kim, B. H., 2008. Rock mechanics advances for under-
agation, leads to final failure. Various explanations have been ground construction in civil engineering and mining. In: Korean
proposed, but none of them have been fully verified. Symposium of Rock Mechanics. Seoul.
Nonetheless, practical approaches to assess the degree to Kaiser, P. K., Kim, B.-H., Bewick, R. P., Valley, B., 2010. Rock mass
which a rock mass is affected by brittle processes (Suorineni strength at depth and implications for pillar design. In: Deep Min-
et al., 2009) as well as the practical implications for modeling ing 2010 – 5th international seminar on deep and high stress
(Martin et al. 1999) have been proposed. mining, Van Sint Jan, M and Potvin, Y. (eds.), Santiago, Chile.
Australian Center for Geomechanics: 463–476.
Martin, C. D., Kaiser, P. K., McCreath, D., 1999. Hoek-Brown
3.2 Outer shell problems: crack damage leading to shear parameters for predicting the depth of brittle failure around
failure tunnels. Can. Geotech. J. 36: 136–151.
Suorineni, F. T., Chinnasane, D., Kaiser, P. K., December 2009.
The degree to which strength degradation should be modi- A Procedure for Determining Rock-Type Specific Hoek-Brown
fied for outer shell problems is the focus of current research. Brittle Parameter s. Rock Mech. Rock Eng. 42 (6): 849-881.
For now, we tentatively propose to use a degradation constant Valley, B., Suorineni, F. T., Kaiser, P. K., 2010. Numerical analyses of
Cs = 50 for the confined conditions (see Fig. 3). The conse- the effect of heterogeneities on rock failure process. In: 44th U.S.
quences are most relevant for pillar design (Kaiser et al., 2010) Rock Mechanics Symposium and 5th U.S.-Canada Symposium.
and are two-fold: 1) pillar design can possibly be optimized, Salt Lake City. ARMA 10–648.
377
G. Vasconcelos
ISISE, Department of Civil Engineering, University of Minho, Portugal
C.A.S. Alves
CIGOVR, Department of Earth Sciences, University of Minho, Portugal
T. Miranda
C-TAC, Department of Civil Engineering, University of Minho, Portugal
1 INTRODUCTION
378
379
T. Miranda
C-TAC, Department of Civil Engineering, University of Minho, Portugal
Masonry is the oldest building material that survived until Granite Petrologic description Loading directions
today, being used all over the world and being present in the
BA Fine to medium-grained Parallel to the rift plane
most impressive historical structures as an evidence of spirit of porphyritic biotite granite
enterprise of ancient cultures. Portugal is rich of monumen- GA, Fine to medium-grained, Parallel to the rift plane
tal buildings and there is a considerable amount of existing GA* with porphyritic trend, two
residential masonry buildings in rural areas. Conservation, mica granite
rehabilitation and strengthening of the built heritage and pro- RM Medium-grained biotite Parallel to the rift plane
tection of human lives are clear demands of modern societies. granite
This requires identification of deficiencies of existing struc- MC Coarse-grained porphyritic Parallel to the rift plane
tures and appropriate intervention techniques. Being masonry biotite granite
a composite material, composed by units and mortar, its behav- AF Fine to medium-grained two Parallel and perpendicular
ior necessarily depends on the mechanical properties of the mica granite to the foliation plane
MDB, Fine to medium-grained two Parallel and perpendicular
components and on the masonry bond (arrangement of the MDB* mica granite to the foliation plane
stones). PTA, Fine to medium-grained two Parallel and perpendicular
Thus, the present article aims at giving an overview of the PTA* mica granite to the foliation/rift plane
tensile and compressive behavior of different types of gran- PLA, Medium to coarse-grained Parallel and perpendicular
ites that are characteristic of ancient masonry buildings in PLA* porphyritic biotite granite to the rift plane
the northern region of Portugal. Besides the presentation of
the details of experimental testing, discussion of the complete
by its random texture, are likely to have isotropic properties,
tensile and compressive behavior based on the complete stress-
a more careful analysis is needed when visual evidence of
displacement diagrams and the main engineering properties
heterogeneities and preferential orientations of the minerals
are presented. The discussion of the variation of the elastic
are detected. The latter assumptions, which were identified
and fracture properties among the different types of granites
on some granitic lithotypes under study, led to the necessity
as a function of the weathering and internal texture (planar
of conducting mechanical tests according to different load-
anisotropy) and physical properties was made based on the
ing directions. Three orthogonal planes can be identified with
experimental results. Finally, the relation between tensile and
rock splitting planes (quarry planes) defined as planes of pre-
compressive behavior as well as between compressive and
ferred rupture. The rift plane is the plane corresponding to the
tensile main mechanical properties is presented based on a set
easiest splitting in the quarry being easily recognized by the
of statistical correlations.
quarryman since it is further associated to the plane of easiest
finishing. A brief description of the adopted different litho-
2 DESCRIPTION OF MATERIALS types is indicated in Table 1. The more weathered types of the
same granite facies are distinguished with an asterisk (*).
The granites adopted in the present work were mostly col-
lected from the Northern region of Portugal. The selection
of the granitic types was based on the mineralogical com- 3 RELATION BETWEEN TENSILE AND
position and grain size, aiming at providing a comprehensive COMPRESSIVE PROPERTIES
sample of the Portuguese granites. In addition to these criteria,
the presence of preferential orientation planes and weathering From the experimental research, it was possible to derive
condition were also taken into account. f some granitic rocks, the compressive and tensile behavior of the distinct types of
380
381
Meng-Chia Weng
Department of Civil and Environmental Engineering, National University of Kaohsiung, Kaohsiung
Hung-Hui Lee
Chung Cheng Institute of Technology, National Defense University, Taoyuan
Shang-Shu Zhang
Graduated Institution of Engineering Technology-Doctorial, National Taipei University of Technology, Taipei
Fu-Shu Jeng
Department of Civil Engineering, National Taiwan University, Taipei
382
4 CONCLUSION
3 RESULTS AND DISCUSSION Erosy, A., Waller, M.D. (1995): Textural characterization of rock.
Engineering Geology 39: 123–136.
Hsieh, Y.M., Li, H.H., Huang, T.H. Jeng, F.S. (2008): Interpreta-
3.1 Strength
tions on how the macroscopic mechanical behavior of sandstone
The results indicate that the cohesion is significantly influ- affected by microscopic properties—revealed by bonded-particle
enced by both GAR and porosity. The greater GAR and model”, Engineering Geology, 99: 1–10.
porosity values will lead to the less cohesion. On the other Itasca Consulting Group Inc. (2002): PFC2D (Particle Flow Code
hand, the lower porosity induces the higher friction angle, and in 2 Dimensions), Version 3.0. Minneapolis, MN: ICG.
Jeng, F. S., Weng, M. C., Lin, M. L. and Huang,T. H. (2004): Influence
the influence of GAR on the friction angle is insignificant. of petrographic parameters on geotechnical properties of Tertiary
sandstones from Taiwan. Engineering Geology, 73: 71–91.
3.2 Characteristics of elastic deformation Lin, M. L., Jeng, F.S., Tsai, L.S., Huang, T.H. (2005): Wetting
weakening of Tertiary sandstones – microscopic mechanism.
For the elastic deformation, it can be found that the modulus Environmental Geology, 48: 265–275.
increases with decreasing porosity and increasing GAR. The Potyondy, D. O. and Cundall, P. A. (2004): A bonded-particle model
numerical results are similar with the tendency of Young’s for rock. Int. J. Rock Mech. Min. Sci., 41: 1329–64.
modulus with different GAR and porosity. Weng, M.C., Jeng, F.S., Huang, T.H., Lin, M. L. (2005): Character-
izing the deformation behavior of Tertiary sandstones. Int. J. Rock
3.3 Characteristics of plastic deformation Mech. Min. Sci., 42: 388–401.
383
Various crack types have been reported to initiate from a propagation processes for rock specimens containing a single
pre-existing flaw in rocks under compression. In addition to pre-existing open flaw. The variations of pressure, Mises stress
the most commonly observed tensile wing cracks, anti-wing and yield stress with time in response to external loading in
cracks propagating in an opposite direction to that of the ten- various pre-selected elements (labeled in figure 3) located on
sile wing cracks are also observed in the field (figure 1) and in the crack paths were monitored. The numerical study revealed
the laboratory testing (Wong and Einstein, 2006, 2009) (figure that for the tensile wing crack and the anti-wing crack, the
2). Although a substantial portion of the anti-wing crack path initiated crack segment adjacent to the flaw tip was found to
was observed to be associated with tensile cracking, signs of be “tensile” and “shear” in nature respectively. The propa-
shearing such as the production of shearing dust were iden- gation of both crack types, which later curved towards the
tified adjacent to the flaw tip. The objective of this paper is vertical loading direction, was tensile in nature as shown in
to numerically study the cracking processes involved in the the numerical study.
anti-wing crack formation and compare it with those in the A detailed stress analysis for the anti-wing crack initia-
tensile wing crack formation. tion and propagation in a rock specimen containing an open
In the present study, numerical analysis was performed by straight flaw under uniaxial (vertical) compression was illus-
a dynamic analysis software by incorporating the static damp- trated in Figures 3 and 4. Three elements (#7, #33 and #50)
ing technique to achieve a quasi-static loading condition. The located along an anti-wing crack propagating path and two ele-
numerical tool was then used to model the crack initiation and ments (#13 and #30) along a tensile wing crack were selected
for mechanics analysis. In elements #7 and #33, yield stress,
Mises stress and pressure all increased with time in response
to the external loading before the occurrence of cracking. The
rate of increase of the Mises stress was higher than those of the
yield stress and pressure (Figure 4a). When the Mises stress
reached the yield stress, yielding commenced. It was then fol-
lowed by an increase of plastic strain until failure (cracking)
occurred in the element. Due to the prevalence of the posi-
tive pressure in the element, the cracking can be classified as
“shear cracking”. When the anti-wing crack propagated, the
pressure in element #50 gradually increased, while staying
Figure 1. Development of tailcracks (tensile wing cracks) in the
extensional quadrants and anticracks in the compressional quadrants.
The right-lateral sense of shearing is indicated by the central arrow
pair (Kattenhorn & Marshall, 2006).
384
positive. When the propagating crack tip was very close to the Wong, L.N.Y. & Einstein, HH. 2006. Fracturing behavior of pris-
element, the pressure dropped and subsequently became neg- matic specimens containing single flaws. Golden Rocks 2006, The
ative (tensile pressure) as shown in Figure 4b. The cracking 41st U.S. Symposium on Rock Mechanics (USRMS): “50 Years of
event in element #50, which was far away from the flaw, was Rock Mechanics – Landmarks and Future Challenges”, Golden,
Colorado, June 17–21, 2006.
thus “tensile” in origin. Similar detailed mechanics analysis
Wong, L.N.Y. & Einstein, HH. 2009. Systematic evaluation of
was performed along the tensile wing crack for elements #13 cracking behavior in specimens containing single flaws under uni-
and #30. The modeling results revealed that the cracking pro- axial compression. International Journal of Rock Mechanics and
cesses in these two elements, respectively adjacent to and far Mining Sciences, 46(2): 239–249.
away from the flaw tip, was “tensile” in nature.
REFERENCES
385
G. Wu
Department of Civil Engineering, Shanghai Jiao Tong University, Shanghai, P.R. China
J. Chen
Department of Civil Engineering, Shanghai Jiao Tong University, Shanghai, P.R. China
386
387
W. Wu, J.B. Zhu, Q.B. Zhang, J.F. Mathier & J.P. Dudt
Laboratoire de mécanique des roches (LMR), École Polytechnique Fédérale de Lausanne (EPFL),
Lausanne, Switzerland
Subject: Rock material and rock mass property testing (laboratory and in situ)
A split Hopkinson pressure bar (SHPB) is a dynamic experi- top of each ball contacts the polished bottom of the specimen
mental technique that applied to determine mechanical prop- to minimize the friction. The plate is cut into two identical
erties of materials at high strain rates (102 −104 s−1 ). The quadrilaterals, and the interface is used to simulate a rock
application of the SHPB technique has been an active research joint. An artificial rock joint is determined in 20 degree with
topic to estimate dynamic properties of rock material, and respect to the front end of the specimen. A granite plate, with
gradually becoming a standard dynamic testing technique in the same cross-sectional area as the specimen, is designed
rock mechanics. However, the technique has been subjected as the striker plate and immediately launched by a couple
of less attention on rock joints. One possibility is the configu- of identical springs, storing equal elastic potential energy, to
ration of long bars especially suitable for the one-dimensional impact the front end of the specimen and generate a seismic
longitudinal wave propagation. In fact, joint patterns often wave. The specimen is pre-compressed by an axial compres-
display higher variability rather than discontinuities formed sion system to make the plate subjected to coupled static and
from purely tensile stress. For a general case, non-welded rock dynamic loads. Two groups of LVDTs are arranged perpendic-
joints are commonly subjected to an obliquely incident wave, ular and parallel to the joint to measure its normal deformation
resulting in not only the reflected and transmitted longitudinal and slip displacement, respectively. The incident, reflected
waves but also the frictional slip and the shear wave generation and transmitted waves are recorded by two groups of strain
and propagation. gauge rosettes attached along the two sides of the joint, and
The study discusses modified Hopkinson bars that conceive the normal and shear stress distributions along the joint can be
the design concept, and introduces a two-dimensional rock obtained.
physical model to simulate a seismic wave across an inclined The experimental technique has some other applications,
rock joint and allow the frictional slip and P- and S-waves e.g. the mechanism of stick-slip friction under seismic trig-
propagation simultaneously. In the two-dimensional config- gering, a seismic wave across an infill rock joint, etc. For the
uration, a Dark Impala granite plate is employed and laid new experimental technique, there are a few critical issues to
over an aluminum table to avoid the self-weight effect due be studied, for instance, the pulse shaping technique that suit-
to the large-scale and easily apply the axial pre-stress load- able for the two dimensional case, the precise control of the
ing. The table surface consists of half-buried balls, and the striker plate to generate a plane wave, etc.
388
G.S. Yang
College of Architecture and Civil Engineering, Xi an University of Science and Technology, Xi an, Shaanxi, China
H.M. Zhang
Department of Mechanics, Xi an University of Science and Technology, Xi an, Shaanxi, China
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: lab testing; rock failure; rock properties; stability analysis; weak rock
389
390
Fractures (faults, joints, cracks) which extend from the ground from 0◦ to 135◦ (45◦ increment) and the bridge length 2b
surface to a certain depth of rock mass are the common geo- (Fig. 1b) was changed from 16.5 mm (half of pre-existing sur-
logical structures and this kind of fractures is defined as face crack length) to 26 mm (about one pre-existing surface
surface fractures. Under the compression stress, crack will ini- crack length).The captured images from the CCD camera were
tiate, propagate and coalesce with the neighboring fractures. used to analyze the displacement field of the pre-existing sur-
However, unlike the 2-D fracture the sliding of the surface face cracks area using Digital Speckle Correlation Method
fracture is restricted by the rock under the surface fracture. (DSCM). The DSCM technique was used for the first time
Thus, the crack growth process and coalescence mechanism to classify the crack types and coalescence patterns. Accord-
of the surface fracture are different with the 2-D case. Up to ing to experimental results, six different crack types (Fig. 1)
now, the coalescence mechanisms are still not fully investi- were identified basing on their growth trajectories and initi-
gated. In order to understand it, granite specimens with two ation mechanisms. Three types are tensile cracks, other three
pre-existing parallel 3-D surface cracks (here and the fol- types are mixed mode (tensile and shear) cracks. Meanwhile,
lowing of this paper, “fracture” will be replaced by “crack”) five crack coalescence patterns (Fig. 2) were identified in
were tested under uniaxial compression. The pre-existing sur- accordance with the initiation mechanisms.
face crack length and angle were fixed at 33 mm and 45◦
respectively. But the bridge angle β (Fig. 1b) was varied
Figure 1. Six types of cracks observed in this study. T = tensile Figure 2. Crack coalescence patterns observed in this study.
crack, S = shear crack. T = tensile crack, S = shear crack.
391
Keywords: numerical modeling, rock joints, back analysis, bolts and anchors
392
These values are close to the lower expected range because of the behaviour of rock joints under small normal stress and
the way that the specimens were prepared. at different levels of compression due to bolting action. It is
The Barton-Bandis joint model has been found to yield therefore selected to be used for any further simulations of
very satisfactory results. It provided good approximation of laboratory-tested bolted rock beams.
393
X. Yun
Cameco Corporation, Saskatoon, Saskatchewan, Canada
H.S. Mitri
McGill University, Montreal, Quebec, Canada
X. Yang
Luoyang Institute of Science and Technology, Luoyang, Henan Province, China
Keywords: Lab testing, Rock properties, Rock stress, Rock failure, Rock bursts
1 INTRODUCTION
394
5 CONCLUSIONS
where Ai = 0.998, Bi = 1.873 and Ci = −1.533. Laboratory test results indicate that the UCS of cubic granite
As can be seen in Figure 2, the biaxial strength of intact samples is larger than that of cylindrical samples by as much
granite can be as much as 1.57 times the uniaxial strength as 22% for 75 mm cubes. The reduced strength of cylindrical
σ c when the confining pressure is 61% of σ c . It should be samples is mainly caused by their larger aspect ratio of 2.5,
noted however that the number of biaxial tests conducted in which in turn lessens the influence of sample end friction with
this study is too limited to warrant the use of Equation (1) as the loading platen. Biaxial tests on cubic granite samples of
an empirical failure criterion for intact granite. More tests are 75, 100 and 125 mm show that size effect is insignificant for
required to validate the coefficients Ai , Bi and Ci . the range of sample sizes investigated. This may be attributed
Combined with the Geological Strength Index (GSI ), a new to the reduced pore and flaw volume under the high confin-
biaxial failure criterion for the corresponding rockmass can ing pressure. The test results indicate that using MoS2 powder
be generally proposed as, is more effective in reducing end frictional effect than brush
platens. The results suggest that while the loading path has less
influence on biaxial strength, confinement increases biaxial
strength to its maximum at σ 2 ≈ 100 MPa. The observed fail-
ure mechanism of biaxially loaded samples is characterized
where Aj , Bj and Cj are constants for rockmass. by a varying degree of spalling initiated at the free faces in the
a) Constant Aj : σ 1 -σ2 plane when the stress level reaches approximately σ c /2
The constant is unified UCS, Aj = Ai . in both directions. The intensity and volume of spalling plates
b) Constant Bj : is found to increase with confining pressure. A formula for an
Bj is the slope of the failure envelope for rockmass. empirical failure criterion for biaxially loaded intact granite
and rockmass is derived based on test result, and combining
with GSI.
395
K. Zhang
Institute of Geotechnical Engineering, School of Civil Engineering, Dalian University of Technology, Dalian, China
Department of Civil and Environmental Engineering, The University of Tokushima, Tokushima, Japan
Q. Yang
Institute of Geotechnical Engineering, School of Civil Engineering, Dalian University of Technology, Dalian, China
The State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian, China
J.C. Jiang
Department of Civil and Environmental Engineering, The University of Tokushima, Tokushima, Japan
S.B. Tian
Institute of Geotechnical Engineering, School of Civil Engineering, Dalian University of Technology, Dalian, China
396
3 NUMERICAL STUDIES
Figure 1. (a) The geometry of a specimen (left); (b) The model of
a flaw under compression (right). The Extend Finite Element Method (XFEM) originally pro-
posed in 1999, which is different from the Conventional Finite
Element Method (CFEM), is very powerful for discontinu-
ous problems in fracture mechanics. The major improvement
between XFEM and CFEM is that the mesh in XFEM is
independent of the internal geometry and physical interfaces,
so that the meshing and re-meshing difficulties after crack
propagation in discontinuous problems can be overcome.
The XFEM technique is used in this paper to simulate
the propagation of cracks in the specimens, and the obtained
results are compared with the laboratory test ones. The numer-
ical results seem to be close to the testing ones, which show
that above criterions and methods can provide a satisfactory
simulation for crack initiation and propagation.
ACKNOWLEDGEMENTS
397
1 INTRODUCTION
398
4 CONCLUSIONS
ACKNOWLEDGEMENTS
Figure 4. Calculation results of fractal dimension. The research is financially supported by the Major State
Basic Research Development Program Fund (Grant Nos.
Table 1. Estimation results of fractal dimension based on two- and 2010CB226804), State Key Laboratory for Geomechanics
three-dimensional box-counting method.
and Deep Underground Engineering Open Research Fund
Average D of (No. SKLGDUEK1021) and Fundamental Research Funds
Stress (MPa) Slices (ADS) ADS + 1 Volumetric D for the Central Universities.
399
Keywords: bolts and anchors, field measurements, monitoring, rock slopes and foundations, rock support
403
J.P. Josien
Formely Geoderis, France, presently Rovconsult, Luxemburg
EXTENDED ABSTRACT. In France, abandoned anthro- management strategy focusing on public safety and based on
pogenic underground excavations underlie tens of thousands permanent monitoring was introduced.
of squared kilometres, disseminated through most regions Risks related to large-scale subsidence and ground collapse,
in various geological settings, at depths ranging from a few involving usually large surface areas, conducted to the choice
meters down to hundreds of meters. A significant part of of the passive microseismic monitoring method, enabling
those ancient mines have been exploited by room and pil- down-hole monitoring of extended mining voids based on a
lar techniques, with extended, multi-level, varied geometries limited number of stations. Its proven capabilities to detect
and interlaced voids, creating uncountable specific situations. distant early warnings related to brittle failures of mine pillars
Most of the mines are ancient and were closed with no or very and their overlying rock strata is documented through many
poor considerations about long term stability, at a time when other underground industrial activities worldwide. This tech-
knowledge in rock mechanics was not as advanced as nowa- nique was extensively used in France in the deep active mines
days. But time adversity is ineluctable with ageing man-made till 2002. Specific field experiments were undertaken to assess
geostructures and key issues posed by abandoned room and its suitability to the different geological settings concerned.
pillar mines has focused an important societal attention with In the salt ore basin, one of these experiment consisted,
a series of large-scale ground collapses and subsidence phe- along with Solvay mining company means and courtesy, in the
nomena that occurred in the old iron basin of Lorraine from unprecedented field monitoring experiment of the controlled
1996 to 1999, triggering emergency evacuations and reopen- collapse of a large solution mined cavern underlying 190
ing brutally socio-economic scars let by the mine industry meters deep. This resulted in the recording of a considerable
decline after-effects. Numerous sinkholes were also reported scientific databank of measurements, including thousands
during that same period of time unveiling eventually the whole of microseismic events and numerous strain, displacement
mining industry inheritance and its potential toll for the next and pressure time series. Relevant observations and results
decades to come. provided a crucial feedback experience (Fig. 1). The major
These accidents were rapidly followed by new regulatory observed phases of the phenomena appear to fit satisfacto-
reforms introduced in 1999, through the so-called “Post- rily the scenario and alarm procedure established prior to the
mining Act”. This included the reinforcing of the last existing event.
mining companies’ obligations during and after mine clo- Between 1998 and 2008, 35 local microseismic monitor-
sure, including long term monitoring whenever needed, land ing systems were deployed. Complementary field experiments
planning constraints and last but not least the setting up were carried out to calibrate the monitoring. Procedures and
of a specific public organization to undertake the expertise seamless numerical routines applied to the data processing
and the works related to these pervasive risks. Then, a con- were enhanced, while the Cloud monitoring technology was
siderable hazard assessment work was undertaken, showing introduced to improve collaborative work.
progressively numerous densely inhabited areas with rated A decade passed, 35 monitoring systems have been
risks involving public safety issues. However, considering deployed over subsidence and ground collapse prone areas
the infrequency of catastrophic ground collapses and giv- and are being run on a routine basis. Since nor catastrophic
ing careful consideration on social, technical and economical ground collapse or subsidence phenomena occurred to date,
aspects, mitigation investment strategies based on hazard or neither false alarm has been triggered along with erroneous
vulnerability reduction, as mine voids filling or expropria- evacuation, this strategy proves to be the most relevant by
tion, appeared in most cases clearly inappropriate. Then, a risk far. The cost benefit assessment related to this deployment
404
405
Fu-Shu Jeng
Department of Civil Engineering, National Taiwan University, Taipei
Tzuu-Hsing Ueng
Department of Materials and Mineral Resources Engineering, National Taipei University of Technology, Taipei
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 GENERAL INSTRUCTIONS test results for the sampled rocks in dry state are referred to as
fresh ones, which are distinguished from the deteriorated sam-
Weathered condition is one of important geological investiga- ples, perhaps after some artificial weathering process, tested
tion items for rock engineering. Conventionally, the qualitative in saturated state. In this manuscript UCS dry and Edry repre-
classification defining six weathered degrees suggested by sent the uniaxial compressive strength and Young’s modulus
ISRM is wide employed for geotechnical site evaluation. And for the rock samples, and UCS det and Edet for deteriorated
the slaking and durability test is usually adopted to evaluate samples, respectively. Two indices are defined to measure the
the rock resistance for weathering. The popular used investi- deterioration of rock samples. The RUCS , the ratio of UCS det
gation approaches focus on the existed in-situ situation and the to UCS dry , measures the strength reduction, and the RE , the
mechanical resistance to short term impact on the rock sam- ratio of Edet to Edry , measures the deformation reduction.
ples. For a rock formation and thus a rock engineering projects,
the long-term weathering potential are mostly disregarded,
perhaps due to previously limited problem caused by long-
term weathering of a rock after completion of engineering 3 STATISTICAL ANALYSES AND DISCUSSION
structure, or lacking of adequate criterion for evaluation.
Focusing on clastic sedimentary rocks, the factors affecting 3.1 Multivariate statistical analyses results
the deterioration of rock samples are first investigated in this This study conducts the multivariate statistical analysis to
study. Many factors influence mechanical characteristics of a identify major factors affecting the deterioration of clastic
rock, such as the diagenesis degree, cement type and degree sedimentary rocks. In the statistical analysis, the strength
between mineral grains, water content, and stress condition, reduction and deformation reduction indices are treated as
etc. The physical indices and petrographic compositions and dependent variables, which may be influenced by the physical
geometries of a rock are related with its uniaxial compressive indices, petrographic composition and geometry factors, and
strength. Do these factors also dominate the strength deterio- fresh strength and deformation modulus, i.e. the independent
ration and deformational softening of a rock? Which factors variables. To obtain the most important factors without double
are important ones? These issues will be distinguished in this accounted variables, correlation analyses between these vari-
study. ables are performed. Accordingly, seven variables among the
initial 11 ones are remained for multivariate statistical analy-
sis, which are n (porosity), GAR (grain area ratio), Quartz
2 ROCK SAMPLES DESCRIPTION (quartz content), Gs (specific gravity), dmean (mean grain
size), FF (form factor) and Feldspar (feldspar content).
A total of 33 clastic sedimentary rocks are sampled from Statistic forward and backward selections are utilized to
Western Foothill Range in northwest Taiwan and form Coastal distinguish the major or the minor factors between the seven
Range in east Taiwan. The rock samples, mostly lithic ones influencing the deterioration of clastic sedimentary
greywacke or quartzwack, were deposited in marine, marine- rocks. Table 1 lists the major factors obtained from of multi-
terrestrial, littoral sedimentary environments in the geological ple regression analyses. For UCS dry , the seven variables are
ages from Oligocene to Pliocene. ranked by r 2 in order as Feldspar, Quartz, n, FF, Gs, dmean and
The rock samples are tested not only in dry and fresh state GAR. A regression function with an r 2 of 0.79 can be deter-
but also in saturated or weathered state to obtain the mechan- mined. The number of major factors is further selected for
ical parameters for fresh rocks and deteriorated rocks. The multiply regression analysis by if the r 2 for the regression can
rocks may be minor weathered while sampled. However, the be significantly increased or not while the additional factor is
406
Ranking +1 +2 +3 +4 +5 +6 +7
considered. Usually four factors are used at most for sake of Figure 2 shows the variation of RE with n and GAR. The
simplicity. variation is scatter. Again the factors considered cannot inter-
The factors influencing the deformation reduction, as well pret the deformation reduction of clastic sedimentary rocks.
as the strength and deformation in both fresh and deterio- Further study is necessary for this issue.
rated state, are also analyzed. Note the r 2 s for regressions
of deformation characteristics and associated reduction are
always low. The r 2 for the most important factors (Quartz) 4 CONCLUSIONS
only is 0.20, and is limited increased by taking into additional
factors; implying that the deformation reduction cannot be By means of multivariate statistical analysis, the factors dom-
interpreted by the factors analyzed. inating the mechanical characteristics of clastic sedimentary
rocks in Taiwan are found. Seven factors including n (poros-
ity), GAR (grain area ratio), Quartz (quartz content), Gs
3.2 Discussion
(specific gravity), dmean (mean grain size), FF (form factor)
Based on the multivariate statistical analysis results, this study and Feldspar (feldspar content) can effectively describe the
further investigates the variation of strength deterioration with UCS dry through a statistical regression function with an r 2 of
first two major factors. Figure 1 shows the variation of RUCS 0.79. And two factors, i.e. the n and GAR can interpret the
with n and GAR. The variation can be divided into two classes RUCS (strength deterioration) of the study rock samples. The
by the dotted lines. For class I the porosity n < 14–18%, RUCS variation of RUCS with n and GAR can be divided into two
rapidly decreases as n increases, no matter the value of GAR. classes with the boundary of n ranging 14–18%. The mul-
And RUCS also decreases as GAR decreases, no matter the tivariate statistical analysis results can interpret the strength
value of n. Both n and GAR play important roles for the and associated deterioration well. Nevertheless, the statistical
strength deterioration of the clastic sedimentary rocks. For results are not good enough for deformational characteristic
class II the porosity n > 14–18%, RUCS decreases slightly as and related reduction. It is obviously that the factors domi-
n increases. Especially as GAR < 50%, the increasing of n nating the deformation reduction of clastic sedimentary rocks
seems do not change the RUCS . are different from those of strength deterioration.
407
Image analysis system for areal measurement to preserve historic masonry wall, Japan
H. Chikahisa
Innovation Center, University of Yamaguchi, Yamaguchi, Japan
408
T. Danjo
Graduate School of Engineering, Hokkaido University, Sapporo, Hokkaido, Japan
S. Kawasaki
Faculty of Engineering, Hokkaido University, Sapporo, Hokkaido, Japan
Beachrock is a coastal deposit that is cemented by cal- this figure, the relation between qu , and formative age, t, is
cium carbonate and/or silica cement in the tidal zone of expressed with the following equation:
sandy beaches in the tropical and the subtropical zones
(Fig. 1). Considering the application of man-made beachrock
to preserve submerged-looking islands, we performed a liter-
ature survey and a field investigation to understand essential From the above data, we obtained sufficient information to
beachrock characteristics. build man-made beachrock.
The methods are as follows. The literature survey was per-
formed from the viewpoints of regional distribution, formative
age, and mineral content based on 113 papers about beachrock
worldwide. The location sites were A (Atobihamabaru), M
(Maeda) and G (Gima) on Okinawa Island, Japan. In-situ tests
included P-wave velocity measurements, and Schmidt ham-
mer and Equotip hardness tests. The collected rocks from
the sites were cored and shaped. In the laboratory, we per-
formed unconfined compression tests, density measurements,
P- and S-wave velocity measurements, and elemental analyses
using EDX.
The results of the literature survey are as follows. Over
90% of beachrocks globally exists between about 40 degrees
north and southern circle. It is found that beachrock formed
from several decades to around 7350 years ago. Furthermore,
beachrock around the world differs in composition. Moreover,
the results of the field investigation are as follows. Beachrocks Figure 1. Beachrock in Atobihamabaru, Okinawa.
from various sites are mechanically and physically heteroge-
neous. For the field tests, the P-wave velocities range from
1.58 to 6.02 km/s. The Schmidt hammer rebound hardness and
the Equotip hardness range from 13 to 45, and from 127 to
456, respectively. As for the laboratory tests, wet and dry den-
sities range from 1.96 to 2.50, and from 1.57 to 2.39 g/cm3 ,
respectively. The water content differs in rocks from different
sites. For example, in sites A, M, and G the water content is
9.3–12.5%, 25.2–26.0%, and 3.4–4.2%, respectively. Uncon-
fined compression strength, qu , values are from 5.63 to
57.80 MPa. P- and S-wave velocities are from 3.00 to 5.14
and from 2.05 to 4.16 km/s, respectively. The compositions
of the beachrock are mainly Ca and C. The concentra-
tions of CaO and C were 50.23–59.65 and 35.60–46.49%,
respectively.
In wet conditions, the relation between qu and formative
age that were researched by Omoto (2005) is shown in Fig-
ure 2. It is shown that qu increases with formative age. From Figure 2. Relation between qu and formative age.
409
T. Dickmann
Amberg Technologies AG, Regensdorf, Switzerland
F. Pittard
Shimizu-Nishimatsu-UEMB-IJM (SNUI) Joint Venture, Kuala Lumpur, Malaysia
A major problem for the driving of a tunnel is the knowledge of noise ratio and the least restrictive conditions for recording
the geological environment and its geotechnical parameters. and processing. The body waves travel as compression or
Significant geological boundaries, which especially cut the shear waves through the ground and are being reflected at
tunnel axis very suddenly can cause serious problems and interfaces with different mechanical properties like density or
risks during tunnel advancement. elasticity. Thus, by separation of the different wave types using
Common exploration methods like exploratory drillings three-component-sensors it is possible to deduct information
or geophysical predictions from the surface can lead to about the mechanical properties of the ground such as elastic
insufficient results due to frequent changes of geological modules.
conditions, high overburden, dense infrastructures above the To perform reliable seismic measurements and thus miti-
tunnel track etc. Besides exploratory drilling from the tunnel gating risks for tunnel construction in hard rock conditions,
face, non-destructive geophysical methods can detect litho- the Tunnel Seismic Prediction (TSP) proved its efficiency in
logic heterogeneities within sufficient distances ahead of the hard rock tunnel projects worldwide. It provides an important
tunnel face. These methods need to be robust and reliable and impact on logistic optimization that the contractor himself can
shall not disrupt the tunneling process while yielding results manage and implement in the tunneling workflow.
quickly and at moderate costs. Fuerthermore it is discussed, what are remaining problems
Seismic reflection imaging is the most effective prediction raised by the engineers on-site, when dealing with the seis-
method because of its large prediction range, high resolu- mic prediction during tunneling. Various important aspects
tion and ease of application on a tunnel construction site. It on data evaluation and interpretation are responded and will
is based on evaluating elastic body waves, which are being be explaind by real data case histories of tunnelling sites.
excited by detonation charges providing the best signal to
410
Subject: Rock material and rock mass property testing (laboratory and in situ)
411
Keywords: field measurement, site characterization, rock mass, 3D terrestrial laser scanning, monitoring
Site characterization of a rock exposure in rock mechanics is 5) More development of hardware suitable to different sit-
one of the important steps, which is required to collect the uations, such as scanning with explosion-proof, mobile
input data for further rock mechanics analysis, rock engineer- scanning by loading a scanner on the car/train etc., remote-
ing design and numerical modeling. However, in the current controlled by a robot deployment system, cave-scanning
practice, these data is still obtained by hand including by using system with the remote-control.
compass and inclinometer for fracture mapping, measurement 6) Software development focusing more on standard data for-
with a ruler, and documentation by recording information on mat and compatible to other software, and simple and
a notebook and photographing with a camera, so the quality user-friendly to the users. Especially much more devel-
and quantity of the data are sometimes unable to meet the opment of software to create the interface to exchange
requirement in rock engineering projects. scanning data with other software.
3D laser scanning becomes now more popular to be used for 7) In order to make the whole procedure simple, the new
capturing 3D digital data to documentation and measurement. development is focusing now on the key-door solution,
This technique shows the great potential to rock mechanics which develops a system including both scanning hardware
application. In order to investigate the practical possibility, and software. A typical example is a tunnel scanning sys-
ISRM has set up a national task together with Swedish national tem, TMS, developed by Amberg AG, and specially applied
group since 2007, and aims to evaluate the state-of-the-art for to tunneling projects to control the geometry of a tunnel.
the development of laser scanning techniques and its potential
In the case study, laser scanning has been tested to different
application for rock mechanics practice. Based upon the inves-
applications related to rock mechanics concerns. With the high
tigation of ISRM-Swedish national task during 2007–2011,
resolution and accuracy scanning data, position and dimension
this paper presents the current state of laser scanning tech-
control are tested to check out the tunnel dimension, monitor
niques on both hardware and software, and summarizes the
deformation of a rock mass. Fracture mapping with two differ-
project examples related to rock mechanics, and then discusses
ent methods, both semi-automatic and full-automatic methods
the needs for further development.
are tested, which is potential to improve the input data for fur-
In current market, there are different scanning systems
ther analysis and numerical modeling. Intensity laser image
developed for different applications, but for rock mechanics,
has much information to be useful for site mapping also. Water
terrestrial laser scanner (TLS) is mostly used, and grouped
leakage is detected in intensity laser image, and possible to be
as pulse-based and phase-based scanners. By reviewing of
enhanced by image processing to identify different amount
the TLS development, the current development shows the
of water leakage. Similarly, laser image has also been tested
following features and development trend:
to identify different rock types, which is useful if no color
1) Hardware have been improved in their functions, such as scanning data available.
scanning range extended from several hundred meters to 3D laser scanning can not only be taken as a new 3D survey-
6000 m for the pulse-based scanner, and up to 187 m for ing method, but also a powerful tool for documentation and
phase-based scanner. visualization. With the infra-red laser, a high resolution 3D
2) Scanning speed has much improved also, from few thou- image can be obtained from the scanning data, which is par-
sands of points per second at the beginning, and up to ticularly important for documentation in a dark situation like
now with 122 000 points/second for the pulse-based scan- in the tunnel. These 3D information both 3D model and 3D
ner, and up to 976 000 points/second for the phase-based image have the potential to improve both quality and quantity
scanner. of site mapping data in rock mechanics.
3) The scanner become more stable and robust with working However, there are still some limits with current techniques,
temperature −10–+45, smaller and light with a size of such as colour scanning still limited to a good illumination,
24 × 20 × 10 cm with 5 kg, and also easy to operate by difficult to process the large amount of scanning data with
in-built computer, internal battery and blue-tooth control. high resolution, especially no so much development for the
4) Scanning with color become more alternative by real color application software to rock mechanics. These problems can
scanner limited for scanning small objects, and possible to be solved with the further development.
scan a large object by scanning with in-built color camera
and extra photographing.
412
Interpretation of rock mass deformability measurements using large flat jack tests
413
REFERENCES
414
Y. Ichihara
Kushiro Coal Mine, Kushiro, Japan
S. Kumakura
Docon, Sapporo, Japan
T. Narita
AIST (Advanced Industrial Science and Technology), Tsukuba, Japan
Keywords: case studies, coal mines, field measurements, lab testing, monitoring, nuclear repository
Excavation Damaged Zones (EDZs) appear around deep rock and 1.5 m from the deformability distribution along a bore-
repositories for geological disposal. Permeability in EDZs hole in the sidewall. In the triaxial compression test on roof
would be much higher than in intact rock masses, and the sandstone from the site, permeability increased with speci-
sealability of a rock cavern is supposed to be significantly men failure. It gradually decreased with time in the residual
compromised by high permeability. Mechanical properties of state and reached almost the same level as that for the intact
medium-hard clastic rocks in EDZs are unknown. To clarify rock specimen. Permeability decreased with the increase in
the sealability of abandoned and closed rock caverns in clastic confining pressure and then increased with its decrease. Frac-
rocks, old roadways were observed at 95 sites during the min- ture closures and poorly connected tortuous fractures were
ing of longwall panels in Kushiro Coal Mine. Rayleigh wave seen in the micrograph of the thin section. The sealabil-
velocities and permeabilities in abandoned and closed road- ity which is recovered to approx. 1/40 the virgin rock mass
ways as old as 50 years, which can be regarded as very severe would be much better than rock cavern in crystalline rocks
EDZs with zero initial sealability, and those in EDZs and in which much less recovery in sealability is expected when
EdZs (Excavation disturbed Zones) around recently excavated collapsed.
roadways in clastic rocks as deep as 300 m were measured
and compared. Most roadways were completely closed by
deflection of roof (Fig. 1) and/or floor heave due to plastic
deformation of clay and siliceous rock. There observed no
significant fractures in the closed parts. On the other hand,
a few roadways were left open by the wooden support. No
back filling was reached to the open roadways. Rayleigh
wave velocity in the closed roadway EDZ and EDZ around
the present roadway are in the range of 0.3–1.2 km/s and
1.1–1.8 km/s, respectively. Permeability in the closed road-
way EdZ, closed roadway EDZ, EdZ and EDZ around the
present roadway were 5.1 to 6.0 × 10−14 , 2 to 10 × 10−13 ,
1.4 × 10−15 and 5.3 × 10−15 m2 , respectively. It is estimated
that the rock mass within 4 m from the sidewall was an EDZ Figure 1. Old roadway observed in Kushiro Coal Mine. Roadway
from permeability distribution along a borehole in roof sand- which is denoted by the broken line was closed by the deflected and
stone. The EDZ boundary was estimated to be between 1 m sagged roof. Height of old roadways is approx. 2.5 m.
415
M. Gharouni Nik
School of Railway Engineering, Iran University of Science & Technology (IUST), Iran
Monitoring of slope movements have been considered by points of view, as well as generically of the types of equipment
researchers, since the investigation of effective factors on the and instrumentation that are available. Legal issues relevant to
start, continuation and stopping of natural and artificial slides slope monitoring have also been dealt with. From the recorded
movement as a result of various internal factors such as slope results of monitoring and detected effects of instrumentation
proportional to inherent soil conditions and the type of its in controlling the slope, one can easily show that conventional
components and also stimulant external factors such as dis- geotechnical instrumentation and survey will continue to have
ruption of forces equilibrium due to the forces of loading and their place in slope monitoring.
unloading, which practically may not be modeled easily in The inclinometer data indicated movement along distinct
laboratory conditions. shear planes whenever there was a built up in water pressure.
Principally monitoring systems are set of instruments which However, slope movement generally ceased upon completion
are capable of recording the variable parameters existing in the of blasting and excavation operations, although some incli-
nature in terms of time, and provide it in order to be used in nometers have shown several cycles of minor movement and
probable events analysis. Some of these variables are natural subsequent stabilization attributed to drained water. As the
and some others are the results of the changes made in nature instrumentation showed some changes in displacements and
due to natural changes taking place. pore pressure, specially after the construction of the slope,
The main magnitudes measured in monitoring are displace- countermeasures consisting of drainage system, rock anchors,
ments, stress variation, water pressure, load exerted on rock meshes and shotcrete against the slope surface were employed
mass and support systems. Measurement of displacements and to arrest slope movements and coming to a standstill. There-
deformations of rock mass, and support systems are the most fore, the use of an instrumentation program for this part of the
effective tools for investigation and inspection of a geotechni- project allowed the owner to make a secure condition for both
cal structure. These tools may have different types according personnel and equipment while there was no need to evacuate
to their application. the area.
In this research, the factors affecting the instability of In summary, instrumentation on its own to some extent can
slopes, the solutions for stabilization of slopes, behavioral predict the failure of the slopes. However, together instru-
variation of sliding, the necessity of behavior measurement mentation and finite element analysis can be used to assess
of slopes movements and investigation of the various tools if the work has the conditions required for a failure to occur.
related to the slopes are briefly introduced and their perfor- Once this has been established the most effective method of
mance and also the method of measurement and the general detecting the failure as it occurs are combination of inclinome-
points related to interpretation of the monitoring data have ters, crack meters and piezometers that have been correctly
been presented. installed at the appropriate locations, but the programme of
Site characterization for a surface mine in design stages monitoring is likely to be long-term. In the short term signifi-
suggested the presence of some instability. Excavation of cant savings could be made by slope monitoring as opposed to
the lower parts of the mine would aggravate the instability. site investigation with immediate remediation in mind. Moni-
Excavation has exposed rock strata that exhibit considerable toring and analysis, done in the correct way, should relatively
horizontal and vertical variation in strength and deforma- quickly identify those slopes that are most at risk of a failure
bility but with a general tendency for vertical repetition of occurring and should afford the opportunity to stagger spend
behavioral characteristics. Right-of-way restrictions limited on remediation, whilst providing the peace of mind that the
rock cut slope angles to those steeper than traditionally used problems are not being ignored.
for these slide prone and highly erodible geologic strata. Cut
slopes were designed to fit within the proposed right-of-way
by developing countermeasures to minimize weak rock degra-
dation, installing sub-horizontal drains to lower groundwater
levels and developing an instrumentation program consisting
of inclinometers, crack meters, extensometers and piezome-
ters to monitor slope movement, tension crack opening and
water levels. Contingency plans to arrest global movement
were to be employed if slope movements occurred beyond
tolerable values. The slope was monitored during and post
construction. If displacement of the material of the slope or
the pore water pressure exceeded preset tolerances, the system
was capable to alert personnel to evacuate the area.
In this paper an overview has been given of slope mon- Figure 1. Water pressure built up readings by the piezometers before
itoring requirements, from both the geotechnical and survey and after rainfalls.
416
Rune Glamheden
Golder Associates AB
Isabelle Olofsson
Svensk Kärnbränslehantering AB
Anders Fredriksson
Sweco AB
Keywords: Nuclear repository, Numerical modelling, Site characterization, Rock Properties, Weak rock
1 INTRODUCTION boundary of the area considered for the repository and was
chosen as a reference object to verify the application of the
For the purpose of studying the possibilities of a repository for modelling strategy used for evaluation of rock mass proper-
spent fuel at approximately 500 m depth in saturated granitic ties for large deformation zones, as monitoring values were
rock, the Swedish Nuclear and Fuel Management Company available from tunnels penetrating the zone.
(SKB) has carried out a series of site investigations between
2002 and 2008.
Data collected from these site investigations have been 3 DESCRIPTION OF THE DEFORMATION ZONE
interpreted and analysed to achieve a site description, which
is built up of models from several disciplines of importance The Singö deformation zone is the most pronounced structures
(SKB, 2008). One of the models is related to rock mechan- in the current region. The deformation zone is a regional fault
ics, which aims to characterize the mechanical properties as characterized as a 200 ± 50 m thick steeply dipping zone that
a basis for predicting the short- and long-term stability of the exist over a length of at least 30 km (Glamheden et al. 2007a).
rock mass. The Singö deformation zone is intersected by four tunnels,
two discharge tunnels from the Forsmark nuclear power sta-
tions and two parallel access tunnels to the SFR waste reposi-
2 MODELLING STRATEGY FOR ROCK MECHANICS tory for low and medium radioactive waste (SFR operational
and construction tunnels).
During the site investigations a modelling strategy was applied The appearance of the Singö deformation zone differs
for determining the rock mass properties for each dominant somewhat between the tunnels, but a core zone with transi-
rock type. This article deals with the theoretical approach of tion zones have been observed in all four tunnels. The width
the rock mechanics modelling.The modelling strategy is based of the core zone is 15–35 m. It is characterized by a 2–12 m
on Discrete Element Method for the modelling of behaviour wide zone of crushed rock, showing a high degree of alter-
of rock masses. The rock mass is modelled as a discontinu- ation and disintegration. On one or both sides of the crushed
ous geometry in which contacts between blocks are fractures rock, several clay filled fractures are found, with a thickness
(Olofsson and Fredriksson, 2005), and the discontinuities in of centimetre up to a metre.
the rock are provided by a Discrete Fracture Network model The transition zones are defined as zones having between
(DFN). The numerical code used was 3DEC. Besides fracture two and ten fractures per metre (Glamheden, 2007a).
geometry, the mechanical properties of fractures and intact
rock are required for the modelling.
In order to catch the natural variability of rock properties
4 MODELLING OF EQUIVALENT PROPERTIES
as well as the variability related to sampling procedure many
samples of intact rock and fractures were collected and tested
4.1 Mechanical properties estimated from numerical
in laboratory. More details are provided in (Glamheden et al.
modelling
2007b). In the same way 20 realizations of a DFN were sim-
ulated for each rock type and stress level to account for the Input data required for the modelling of the zone are the DFN
variability in the fracture sets. for each sector of the zone as well as the intact rock and frac-
When looking forwards to the planning of the construction ture mechanical properties. For the host rock mean values of
of the deep repository, determining mechanical properties of mechanical properties of intact rock and fractures as evaluated
major deformation zones are an important issue. The Singö in the site descriptive model of Forsmark were used (SKB,
regional deformation zone in Forsmark defines the northern 2008). In transition and core zones no tests were available
417
for intact rock and fractures, and their properties were there- The results presented in the paper demonstrate that the
fore estimated from empirical experiences and values given methodology used for simulating equivalent mechanical prop-
in the literature. Due to the high fracture density in the core of erties is an applicable and adequate method, even in case of
the zone only 5 DFN realizations for each sector were simu- large deformation zones.
lated. For the host rock 20 DFN realizations (based on fracture Typical equivalent rock mechanical parameters of the Singö
statistics of the host rock) were used as a comparison. Addi- deformation zone evaluated for use in a regional stress model
tional details regarding the fracture network generation and of the Forsmark area are: 200 MPa/m in normal stiffness, 10–
the modelling are presented in (Glamheden et al, 2007a). 15 MPa/m in shear stiffness, 0.4 MPa in cohesion and 31.5 deg
The five DFN realizations for each of the sub-sectors were in friction angle.
analyzed at three different stress levels: 5 MPa (equivalent
to the maximum principal stress σ1 ), 2.5 and 0.5 MPa. The ACKNOWLEDGMENT
mechanical models were loaded with a constant velocity in
the vertical direction while the horizontal stresses were con- The authors wish to acknowledge that this paper is based
stant during the loading test. Evaluated mean values of the on work funded by the Swedish Nuclear Fuel and Waste
mechanical properties of the deformations in the SFR tunnel Management Co (SKB).
passage are presented in Table 1.
The rock mass properties that best fitted the recorded deforma-
tions in a back-calculation were used for numerical modeling
418
Rune Glamheden
Golder Associates AB, Sweden
Billy Fälth
Clay Technology AB, Sweden
Rolf Christiansson
Svensk Kärnbränslehantering AB, Sweden
1 INTRODUCTION
419
420
N.F. Grossmann
Foundations and Underground Works Division, Concrete Dams Department, Laboratório Nacional de Engenharia
Civil (LNEC), Lisbon, Portugal
1 DEFINITION function f(ω, ε), where ω and ε are, respectively, the longitude
and the colatitude of the attitude (ω, ε), in a spherical coor-
For a homogeneous rock mass, the intensity I corresponds to dinate system whose revolution axis is normal to the mean
the ratio between the sum of the individual areas of the N attitude of this joint set.
joints of the considered set which occur in a given volume V The total intersection becomes
of the rock mass, and that volume
2 TOTAL INTERSECTION
421
If, by chance, the distribution of the attitudes of the joints the scanline, β the angle between the scanline and any normal
of the considered set is isotropic, equation (10) becomes to the attitude (σ, δ), and (cos β) the mean of (cos β).
As one has always to use the absolute value of (cos β)
5 SCANLINE
In order to give an idea of the values which are obtained with
If the observation surface is reduced to a scanline, the intensity
equation (15), Fig. 2 presents the variation of the arc cosine
becomes
of (cos β) with σ, for different values of ε1 .
The value of ε1 for each one of the curves, is, obviously, its
abscissa for the standard deviation 0.
422
E. Hakami
Geosigma AB, Stockholm
R. Christiansson
Swedish Nuclear Fuel and Waste Management AB, Stockholm
Keywords: rock stress, field measurements, nuclear repository, rock failure, site characterisation
423
424
C. Olalla
E.T.S.I.C.C.P., Universidad Politécnica de Madrid, Spain
425
426
In this study, the geostatistical method is tested for its goodness quality at un-sampled areas by means of geostatistical analy-
in prediction of rock mass conditions along alignment of two sis. For Lam ta Khong tunnels, anisotropy variograms RMR
tunnel projects in Thailand excavated in differential geologic of of Phu Kradung rock formation, tunnel hosting rock for-
settings. The first case was tunnels of Lam Ta Khong Pumped mation, were used in the prediction of rock mass quality of
Storage project excavated in a simple geological setting. The the two tunnels selected for study for the project because the
second case was tunnels of Tha Dan dam project which was tunnels traversed only the rock formation. The kriging estima-
situated in a more complex geological setting. In the first case, tion showed similar RMRs to field observation data, except
sample data were from vertical drillholes made in horizontally for shallow zones within 500 m from the portal as shown in
bedded sedimentary rock formations whereas for the second Fig. 1.
case they were from vertical drillholes orientated more or For Tha Dan Dam Tunnels, analysis of drillhole data showed
less sub-parallel to predominant discontinuity set of volcanic that a good variogram for kriging could not be determined
rocks. Parallel faults, shears and fractured zones in hard rocks in horizontal direction, indicating poor spatial correlation of
are characteristics of the volcanic rocks at the site. Rock mass data. This was likely due to effect of the predominant set of
quality along the tunnel alignments is interpreted based on steeply dipping weak discontinuities traversing tunnel align-
borehole data and surface mapping data. It is then compared ment at high angle. Thus the omni-directional variogram was
with the actual ones as mapped at tunnel face during construc- used in the kriging. The comparison between the actual and
tion, so that the goodness the geostatistical predictions can be the predicted RMR along two tunnel of the project as shown
assessed. Fig. 2, indicates that RMRs from field observation consis-
Core logs of drillholes made prior to construction were used tently showed abrupt fluctuations with distance along the
in determination of the rock mass quality in term of RMRs tunnel length with could not be captured by the kriging esti-
along depths of the drillholes so that statistical properties mation. For this geologic setting, the geostatistical method has
and variograms were determined for prediction of rock mass limitation.
Figure 1. Actual versus predicted RMR of Lam Ta Khong Tunnels. Figure 2. Actual versus predicted RMR of Tha Dan Dam Tunnels.
427
Keywords: Rock stress; Back analysis; Numerical methods; Field measurement; Case study; Simulation
here σ0 and S stand for the in situ measured and the tectonic
stress state, respectively. Also, α is a 3-by-3 matrix of coeffi-
cients and k is referred to as a 3-by-3 matrix of coefficients for
a linear distribution of horizontal stresses (Cornet FH, et al, To be able to proceed along the algorithm, the second pro-
1984 & G. Li, et al, 2009). vision must be investigated, too. If β ≤ 0.1, the process is
terminated and we will have the values of in situ stresses in
3 CONTRIBUTION every arbitrary point; otherwise, according to the second pro-
vision, another set of measured points must be chosen from
In this section, the proposed algorithm is described step- set N to restart the algorithm.
by-step and its applicability is shown by highlighting the
most important stages. Figure (1) illustrates the suggested 4 ALGORITHM APPLICABILITY
flowchart.
To show the applicability of this algorithm, it is applied for
3.1 Description of the first provision
Seymare Dam (SD) to determine its stress field using the pro-
To increase the ability of this algorithm in estimation of in situ posed method and based on the local stresses measured by HF.
stress, it is necessary to compare the measured data with those Fig. 2 shows the location of SD in Iran.
428
Young’s
Density Poisson’s modulus
Description t/m3 ratio GPa
σh σH σxy
γ = 0.08, β = 0.098
429
M.A. Kanji
University São Paulo, Brazil
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 INTRODUCTION that the maximum principal stress was also measured addi-
tionally to the minimum principal stress. The paper describes
This paper aims at the presentation of a new in situ test to the testing equipment, procedures and its interpretation.
determine directly the minimum principal stress (σ3 ) and the
maximum principal stress (σ1 ) by means of cheap and simple
hydraulic jacking tests performed in boreholes in fractured 2 EQUIPMENT REQUIRED AND TEST SECTION
rock masses.
Hydro jacking tests are usually performed when studying The required equipment consists of a hydraulic pump with
pressure tunnel sites under possible insufficient rock cover. a minimum flow capacity of 100 l/min under a pressure of
In the event of hydrofracturing, severe damage can occur to 20 to 25 MPa, double packers separated by about 1 to 1.5 m,
the rock mass, as shown by an actual drastic example of rock pressure and flow gages with automatic recording. A drill rig
damaged by hydrofracturing is shown in Figure 1. or a winch is also required do lower and raise the packers. If
The tests are run in boreholes on stretches of fractured rock the rock is too permeable the pump may not be able to reach
separated by double packers (usually separated by 1 to 1.5 m), the required pressure. In this case, it is recommended to select
until opening and closure of fractures are observed, indicating a new borehole stretch.
the value of the minimum stress acting in the rock mass.
Conventional hydrofracturing tests (Haimson, 1978 and
ASTM, 2008) require a section of intact rock for the test, 3 TESTING PROCEDURE AND INTERPRETATION
and the HTPF (Cornett, 1986) needs the determination of the
spatial orientation of at least six known pre-existing fractures After setting the packers in position within the borehole, the
to be testes individually. On the contrary, the test herein con- pressure is raised by steps of 2 atm up to 5 atm at the most,
cerned does not require a borehole section of intact rock but a waiting between 1 and 2 min in each pressure stage, as the
jointed rock mass and do not need inspection of the borehole joints opening pressure is to be determined as close as possible.
walls to detect the strike and dip of the joint. This is called the “slow” cycle. In these hydro jacking tests it
In carrying on hydro jacking tests in several hydro power is not necessary to spend more time in each stage, as in the
projects in the Andean region, the tests were conducted well Lugeon tests, as the response of the ground to the pressure is
beyond the usual pressure ranges for hydraulic jacking tests, very quick.
since a higher capacity pump was available. Several behav- When equaling the minimum confining pressure of the
iors during the tests were noticed, allowing the interpretation ground, plus eventual joint strength that might exist, the joints
will open.This is noticed by the change in relationship between
pressure and flow, as exemplified by the graph of Figure 2,
where the points of opening and closure of fractures are
marked respectively as “O” and “C”.
The pressure is continued to be increased in similar steps
to the maximum value or capacity of the pump, and then
decreased under the same pressure steps. The closure pres-
sure of a joint is a direct measure of the minimum principal
stress, or σ3 . In increasing the pressure, it is possible to notice
more than one changes of pressure to flow ratios, as shown in
Figure 3, indicating that other joints, with different orientation
opened up on their turn.
Following a “slow” cycle, a “fast” loading and unloading
cycle, was carried on, increasing and decreasing continuously
the pressure, within about 5 min. The results were surprisingly
very close to the “slow” cycle. Figure 3 show the curves of
both “slow” and “fast” cycles, with very good reproducibility.
Figure 1. Rock fracture opened by hydrofracturing in a pressure Several other particular behaviors may occur during the test,
tunnel with only shotcrete as support. consisting either as an oscillation (or jerking) of the flow or of
430
4 CONCLUSIONS
431
ABSTRACT: Parallel deformability tests, three rigid Plate Loading (PLT) and one extra Large Flat Jack (LFJ) were performed
side by side in slightly jointed marlstone in order to compare their results. The pressure-displacement data obtained from LFJ
and PLT tests are presented and processed using different formulas to determine overall (by the least-squares method), instant
and interval moduli. Different moduli interpreted from PLTs are compared to moduli from LFJ test data. The case shows typical
problems encountered in PLT data interpretation.
Three plate load tests and one large flat jack test were carried
2 TEST LOCATION
out in total. Two PLTs were set-up in a horizontal direction and
one in a vertical direction. The LFJ test was set-up as a vertical
The test area was located on the right bank of Karun River near
single slot test in the heading wall of test adit that enables a
the inlet of the Karun 3 dam diversion tunnel in sedimentary
good comparison with the horizontaly oriented PLTs.
rocks belonging to Pabdeh formation that comprises moder-
First PLT was set up in a horizontal direction perpendic-
ately strong, slightly jointed (joints spaced from 20 to 60 cm)
ularly to the adit axis and located 4 m far from the adit end
marlstone of low durability with RMR values ranging from
(heading wall). Second PLT was carried out immediately after
40 to 45. The Uniaxial Compressive Strength (UCS) of tested
the first. It was set up in a vertical direction and located 18 m
marlstone was between 15 and 25 MPa depending on water
far from the adit end. Seven months later third PLT was set up
content. The lower UCS limit corresponds to wet (saturated)
in a horizontal direction side by side to the first PLT, 5 m far
and the higher to dry rock.
from the adit end. It was carried out in order to compare its
The deformability tests were performed at the end of an
results with the first PLT.
exploratory adit that was excavated by drill and blast tech-
As large differences between the first and third PLT were
nique. The blasting was carerfully performed in order to
noticed, it was decided to carry out a large flat jack test. A year
decrease damage of the rocks.
later then the first PLT, the LFJ test was set-up as a vertical
single slot test in the heading wall of the test adit.
3 TEST PREPARATIONS AND EQUIPMENT Tests are evaluated on the basis of the measured pressure-
displacement curves. Different deformability moduli were
3.1 Plate loading test calculated from PLT and LFJ test data. A distinction was made
Plate loading (jacking) tests were carried out by means of a between moduli of loading (Eml ), unloading (Emul ) and peak
pair of circular rigid plates as a double load plate test. The to peak moduli (Empp ).
432
433
K. Kemppainen
Posiva Oy, Olkiluoto, Finland
M. Hakala
KMS Hakala Oy, Nokia, Finland
E. Johansson
Saanio & Riekkola Oy, Helsinki, Finland
H. Kuula
WSP Finland Oy, Helsinki, Finland
J.A. Hudson
Imperial College and Rock Engineering Consultants, London, UK
Subject: Rock material and rock mass property testing (in situ)
1 INTRODUCTION
434
435
436
Hyunwoo Kim
School of Civil, Urban and Geosystem Engineering, Seoul National University, Seoul, South Korea
Seokwon Jeon
Department of Energy Resources Engineering, Seoul National University, Seoul, South Korea
Monitoring of the ground condition ahead of subsea tunnel displacement and settlement at tunnel crown, when the tunnel
face is indispensable for safety and efficiency of excavation. face approaches a zone of different stiffness. The preceding
Since a sea floor is covered with water, geotechnical investiga- displacements are continuously measured by horizontal incli-
tion is usually carried out in the limited extent prior to the exca- nometers installed at tunnel crown in front of tunnel face, and
vation, resulting in insufficient geotechnical data available for their monitoring results are able to indicate the change of rock
tunneling. Hence, geotechnical investigation should be made mass condition. Lastly, it was considered that the monitoring
continuously during excavation with monitoring items speci- of water inflow at tunnel crown and bottom could inform us a
fied especially in subsea tunnels. Probe drilling has been one change of ground under the assumption that the permeability
of the most widely used methods to determine the ground heterogeneity of rock mass would change the amount of water
condition during subsea tunnel construction. The quality and inflow entering working space immediately after excavation.
applicability of its results are however largely dependent on the In the present study, we observed the trend of vector orienta-
number of drilling holes. The continuously stable excavation tion and preceding displacement at tunnel sidewall as well as
could make a contractor be over-optimistic and decrease the at crown.
number of probe drilling for monitoring, so that the detection A three-dimensional finite element analysis program was
of unexpected hazard may be too lately. Probe drilling some- used to simulate the process of subsea tunnel excavation. With
times requires an additional surveying plan to understand the the results of numerical simulation as shown in Figure 2, three
overall structure of surrounding rock mass. monitoring items were compared in terms of the capability of
The purpose of the present study is to compare monitor- categorizing type of hazardous ground and the early-warning
ing items to find out which of them can provide reliable and time. We assessed their capability to identify the shape of
supplementary information on the ground and can reduce hazardous ground ahead of face and examined which item
uncertainty when being adopted with probe drilling. From deviated from normal trend at the earliest time by the influence
the case studies on construction of subsea tunnels, we first of hazardous ground.
extracted the main geotechnical factors affecting tunnel sta- Figure 3 presents the typical characteristics of monitoring
bility such as type and property of hazardous ground, condition items in four different ground types. It was assessed that the
of surrounding rock mass, in-situ stress and rock cover. Six vector orientation could find out the type of hazardous ground
representative hazardous ground types were then obtained in all cases. The preceding displacement was assessed to detect
by the combination of the factors. Depression and weakness ‘Weak−45’ type clearly from increase of trends both at crown
zone type of ground were selected among them as the major and at sidewall. While it could notice the presence of ‘Depres-
hazardous ground type, which are shown in Figure 1. sion’ or ‘Weak+45’ type, no distinction was made between
Displacement vector orientation, preceding displacement two ground types. The existence of ‘Weak+00’ type could
and water inflow were chosen as monitoring items. The mon- be expected from the decrease of preceding displacement at
itoring with vector orientation can detect the weaker zone crown and the slight change at sidewall. In the case of water
ahead of tunnel face by tracing the ratio between longitudinal inflow, ‘Weak+00’ and ‘Weak−45’ type could be detected
437
438
Determination of principal stress orientations from wireline logging caliper data and
borehole images in deep drilling
W. Lin
Kochi Institute for Core Sample Research, Japan Agency for Marine-Earth Science and Technology (JAMSTEC)
439
Keywords: field measurements, risks and hazards, rock failure, rock mass, stability analysis.
This paper examines a new method of evaluating the stabil- the effectiveness of LDV for evaluating the stability of rock
ity of a rock slope using a remotely positioned laser Doppler slopes and cleared that the scaling effects of block size have
vibrometer (LDV). We conducted an experiment using phys- to be considered for the evaluating stability of rock slope.
ical models and performed a numerical analysis to evaluate This study determined that LDV can make accurate block
the new method. vibration measurements, and that the dominant frequency and
Figure 1 shows one of the physical model, which included amplitude of vibrations in a block are related to the block
concrete blocks on an artificial soil slope with two block sizes stability. Furthermore, we found that the dominant frequency
and three slopes. Figure 2 shows one of the result of the exper- of vibrations in the block was inversely proportional to the
iments. The LDV measurements agreed with conventional size of the block.
seismometer measurements.
The experiment results have cleared that the dominant fre-
quency of the blocks varied with the stability, and dominant
frequency and the amplitude varied with the block size (Fig-
ure 3). The numerical model was used to examine a concrete
block adhered to a concrete base with different contact areas.
The dominant frequency of the blocks determined using the
numerical model agreed with those obtained from the physical
experiments (Figure 4 shows the results by numerical model).
Furthermore, we analyzed different sized blocks to examine
the scaling effects. The dominant frequency of the blocks was
inversely related to the block size. These results demonstrated
Figure 2. Comparison of the LD and seismometer results. Figure 4. Example of the block vibration analysis.
440
F. Moosavi
School of Metallurgy & Materials Engineering, The University of Tehran, Tehran, Iran
Keywords: Rock stress, Numerical modeling, Field measurements, Spalling, Under Coring Technique
Spalling is a mode of failure which occurs in brittle materials. as the average of the values, is more reliable. The measured
In this paper, the effective factors on spalling phenomenon at maximum principle stress is about 8.8 MPa. The amount of
Nayriz Marble Mine (Iran) are investigated (Fig. 1). induced stress at mine wall cannot provide essential conditions
Earlier experiments have proven that this type of failure can for this type of failure, so some other factors that are mentioned
occur when the stress is as high as one-third of the material’s above, might play noticeable roles.
uniaxial compressive strength. Anisotropy and fluctuations of The variation of temperature during a day was measured,
temperature also contributes to weakening of the rock as well and numerical models were carried out to analyze the effect of
as inducing thermal stresses which helps spalling to occur at it on redistribution of stress state at mine wall. This effect can
lower stress levels. In this study, the main emphasize is put on be as high as 15 MPa due to 20 degrees Celsius temperature
the stress measurement to investigate if only this parameter change (Fig. 2).
can explain spalling in the mine or there are other factors The results showed that the concentrated stresses at the wall
contributing to this phenomenon. Samples from the mine were due to in situ stress filed can not by itself provide necessary
taken and their mechanical parameters were determined in conditions for spalling failure to occur, and other factors (such
the laboratory, its results can be found in table 1. The in situ as temperature variations, anisotropy, and/or time dependent
stress state was determined at site by using a new modified behavior etc) should also play a role to cause spalling.
Under Coring Method which is suitable for brittle materials.
This modified approach uses 12 pins around borehole, and
instead of one sets of diagonal for stress determination, 20
states are available, so the obtained result, which is calculated
441
Keywords: tunnelling, stability analysis, numerical modeling, soft ground, monitoring, support
1 INTRODUCTION
3 PROJECT OVERVIEW
5 MONITORING AND EVALUATION
Increasing the capacity of the main railway axes crossing
Vienna is an essential concern both at the national and the A comprehensive monitoring programme for the observa-
European level. The Lainzer Tunnel project – a 12.8 km tion of surface settlements and of displacements in the
new strategic railway link is necessary due to the increasing tunnel was executed. Tunnel displacements were systemati-
demand for passenger and freight transport and will improve cally obtained in regular intervals by absolute 3D monitoring.
442
A grid of surface targets were arranged in cross and longitu- Two case studies illustrate the successful application of
dinal sections and on buildings. the observational method within the geotechnical safety
The displacements of the targets in the area and along the management.
railway tracks were automatically obtained using the Online- In the first considerable surface settlements in soft ground
3D-geomonitoring system recorded continuously every 15 were measured in a section where the overburden decreases
minutes from stationary monitoring stations. Displacements to a minimum of approx. 6 m and before the excavation works
of the tunnel face, correlating with pre-face surface settle- succeeded in breaking through the concrete pile wall of the
ments, are monitored by an innovative, fully automatic face adjacent lot LT 44.
monitoring system called OBM. The second deals with unfavourable system behaviour
Advanced tools and methods for monitoring data evalua- during bench and invert excavation observed in flysh rocks.
tion assist in the continuous interpretation of the behaviour In both cases the spatial deformation analysis and the
of ground and support interaction, timely detection of devi- comparison between predicted and monitored displacements
ations and check against the warning and alarm criteria. The formed the basis for the selection and management of the
programs allow the evaluation and visualizing of displace- mitigation measures to successfully cope with the hazardous
ment versus time, distance-displacement, deflection curve situation.
diagrams, trends of displacements and ratios of different
displacement components. Furthermore contour plots of REFERENCES
surface settlements can be evaluated. A main feature was the
determination of the stress intensity in the shotcrete lining Moritz, B.; Schubert, W.: Application of the Observational Method in
by calculating the level of loading based on the measured Geotechnical Safety Management, Geomechanik und Tunnelbau
displacements (Figure 3). (2009a) No. 3, Verlag Ernst & Sohn.
Overall these methods allow a much better assessment of the Vavrovsky, G. M.; Ayaydin, N.; Schubert, P.: Geotechnisches Sicher-
geomechanical process during tunnelling, than it was possible heitsmanagement im oberflächennahen Tunnelbau, Felsbau 19
with traditional methods. (2001) Nr. 5, pp. 133–139.
443
T. Yamamoto
Department of Environmental and Urban Engineering, Kansai University, Osaka, Japan
Kinki Regional Development Bureau, Ministry of Land, Infrastructure, Transport and Tourism, Osaka, Japan
T. Tsuji
Department of Environment and Resource System Engineering, Kyoto University, Kyoto, Japan
Keywords: monitoring, field measurements, rock slopes and foundations, neural network, rock properties
1 INTRODUCTION
444
S. Kubota
Civil Engineering Division of Yamaguchi Prefectural Government, Yamaguchi, Japan
T. Nakano
Tokiwa Chika Kogyo Co., Ltd., Ube, Japan
445
A. Noorzad
Faculty of Water and Environmental Engineering, Power and Water University of Technology, Tehran, Iran
Subject: Rock material and rock mass property testing [laboratory and in-situ], information system, artificial intelligence and
other advanced techniques
Keywords: field measurements, lab testing, rock mass, case studies, neural network, Asmari formation
Numerous large dams of Iran have been constructed in Asmari Table 1. Equations between CYC1 to CYC5 and E.
formation. Knowing the geomechanical characteristics of this
formation can help significantly to reduce costs and increase Equation R2 F Sig.
accuracy of parameters used in future studies. Deformation
CYC1 = E 0.886 .82 179.70 .000
modulus represents the rock mechanical behavior before fail-
CYC2 = E 1.001 .89 363.33 .000
ure and is used as the most important data for understanding
CYC3 = E 1.029 .90 433.44 .000
deformation behavior of rock mass. Two large concrete dams, CYC4 = E 1.022 .91 470.89 .000
namely Karun IV and Khersan III located in Asmari forma- CYC5 = E 0.934 .97 1.04E3 .000
tion in the south of Iran, have been examined as case studies.
The relation between elasticity and deformation modulus in
two mentioned dams has been investigated by artificial neural
network and statistical method. Dam and hydro power plant of methods for training, validation and neural network testing,
Karun IV is located in Bakhtiari region, one of the branches of are utilized in order to control the results achieved by MAT-
Zagros Mountain. Rock mass surrounded the dam site is com- LAB software. If the obtained results are in agreement with
posed of units ofAsmari formation including different types of the measured values, then the equation between the predicted
limestone, porous limestone, marly limestone, limy marl and values of CYC5 and measured values of E will be presented.
marl. Khersan III dam foundation and abutments have been The characteristics of the developed program are:
situated in Asmari formation. The formation is composed of
• Number of transfer function: Tansig, Purelin
limestone with a medium thickness, marly limestone, marl and
• Learning steps: 6862
small amounts of dolomite. Because of variety in lithology and
• Gradient descent: 0.005
stratigraphy characteristics, both formations are divided into
• Number of input layer: 1
two parts of lower Asmari and upper Asmari.
• Number of hidden layer: 1
In the present research, the data sets of elasticity and defor-
• Number of neurons in the hidden layer: 2
mation modulus have been utilized. It is common by many
• Number of neurons in output layer: 1
researchers to apply RQD, RMR and GSI to present the dif-
• Type of network: Back-Propagation Neural Network
ferences in the behavior of intact rock and rock mass in their
relations to consider the effect of discontinuities. While in this (BPNN)
• Learning algorithm: Traingdm
study only the relations between the 1st to 5th loading cycles
in plate loading test with elasticity modulus have been con- The results indicate that the predicted values of statisti-
sidered. The results indicate that the correlation coefficient cal and neural network methods have the highest accuracy
(R2) and the F values from F tests are improved by increas- by nonlinear relations and have a very good agreement with
ing loading cycles as given in Table 1 with power model. In the in-situ measurements. The accuracy of the equations
other words, increasing the loading cycles leads to decrease is expressed by R2 and F values, which are defined by
the effect of discontinuities in relations between elasticity and the correlation relation between independent and dependent
deformation modulus. Therefore, it is obvious that there is no parameters, respectively. The most desirable equation between
need to utilize RQD, RMR and GSI in order to develop equa- the measured and predicted values of CYC5 is a power equa-
tions. Thus in this paper, the relation between elasticity and tion. The R2 and F values resulted from neural networks
Asmari deformation modulus obtained from the 5th cycle of are more desirable than the obtained values from statistical
plate load test has been employed. method. To conclude the above discussion, using neural net-
In the statistical method, the relation between CYC5 and E work power relation equation is recommended. Therefore, the
is determined considering with and without intercept in the 5th equation is more appropriate as there is a real prediction
equation, in which CYC5 is a dependent parameter while E between E and the CYC5 value resulting more validity in com-
is an independent parameter. In neural network model, E and parison to other equations. On the other hand, the difference
CYC5 values are used as input and output data, respectively. between E and CYC5, which is caused by discontinuities of
Initially, all 26 data sets, which are employed in statistical the rock mass, decreases by increasing E.
446
The Cross-sectional Borehole Deformation Method (CBDM) Based on the estimation of stress changes, it was made
is a method that two dimensional state of stress change within clear that the stress components increased in former period
rock mass in a plane perpendicular to a borehole axis can be of excavation and decreased in latter period gradually and
measured. This method was applied to estimate stress change continuously, then that the CBDM is available for measuring
of in mediate rock mass of a rock cavern in Figure 1. This is stress change.
the plan view of measurement site in Kamioka Mine. A cavern
was excavated at a depth of 900 m within gneiss. The Young’s
modulus and Poisson’s ratio are 30 GPa and 0.2 respectively.
The dimension of the cavern is 15 m by 21 m and 15 m in
height. The borehole with a length of 5 m for measurement of
stress change was drilled from the gallery to the cavern before
the start of its excavation. The width of the rock between the
gallery and the cavern is about 7 m. The measuring point is
located at a depth of 4 m.
The measurements were performed at nine stages before,
during and after excavation. The results of stress change are
shown in Figure 2. The vertical stress change σY is zero
until stage II, but increases at stage III, then reaches the
maximum value at stage V. After that, the stress decreases
gradually with the progress of excavation. The tendency
of the horizontal stress change σX is almost the same
that of σY . This means that the rock near the measur-
ing point was damaged and became the loosened zone.
However, as the change of all stress components is con-
tinuously, it is considered that that damage did not happen Figure 1. Location of borehole for measurement and cavern in the
suddenly. plan view of measurement site in Kamioka Mine.
Figure 2. Changes of all components of stressin the X -Y coordinate system: X -axis is defined in the horizontal direction as shown in Figure
1 and Y -axis is defined vertically.
447
Rock investigations for the construction of railway tunnel through Slovenian Karst
Keywords: site characterization, rock properties, rock mass classification, Karst formations, field measurements, monitoring
448
449
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: field measurements, rock properties, site characterization, tunnelling, weak rock
450
451
A.W. Ren, X.G. Wang, Z.X. Jia, Y.J. Wang & Q.W. Duan
Department of Geotechnical Engineering, China Institute of Water Resources and Hydropower Research
The Suoduoxi accumulation body is an aggregation of accu- the accumulation are got. Then, on the basis of geomorphic
mulation. It can be seen everywhere along the upper reaches features, the accumulation is divided into four domains. Next,
of Jinsha river. It is a big problem. So many studies have its stability is evaluated through detail rock structure analysis.
been done to resolve the problem. However, they were all And the geological origin is analyzed by integrating all the geo-
focused on single accumulation features. The entirety charac- logical information. Finally, four results are achieved. (1) Most
ter of an aggregate cann’t be achieved by this way. Therefore, of the accumulation is composed of big rock blocks and gravel-
the integrated geological analysis method is applied to study bearing soil, dominated by big rock blocks. So there is a rock
the ancient accumulation in this paper. The study is based on a block arching in the accumulation. But domain 2 has obvi-
project of Suwalong Hydropower Station in the upper reaches ous different structure from other domains. (2) Domain 2 is
of the Jinsha River. The accumulation body is located at an composed of gravel-bearing soil and rock blocks, dominated
area strongly affected by geological structure and climate. It by gravel-bearing soil. Furthermore, there is a grading feature
lies on the right-bank of the Jinsha River and is only about from up to down. This is a typical character of debris-flow. So,
one kilometer far from Suwalong Hydropower Station. It is a domain 2 came from anti-shore. (3) The stability of the accu-
very huge geological body, almost three million tons weight. mulation is largely good. But collapse will probably happen
Its geological and origin is not known up to now, and so its at domain 2. So some control measures should be adopted to
stability will be a big problem for the Suwalong project. In domain 2 before the construction of the hydropower station.
this paper the problem is analyzed by two ways: engineering (4) The product period of domain 1 is earlier than that of the
geological exploration and rock structure theory. There are other domains. But river blocking did not happen during the
two geological thoughts through the paper, not only grasping production period of domain 1. The production of domain 1
the whole, but also doing analogy analysis in detail. First of made just the river diverting. The true river blocking happened
all, by on-site exploration, the whole geological characters of after production of domain 2, domain 3 and domain 4.
452
The Downward Compact Conical-ended Borehole Overcoring An in situ test was carried out in a gallery of the Kamaishi
(DCCBO) technique is a method for the in situ measurement mine about 270 m below the surface. In situ stress measure-
of stress that is based on the Compact Conical-ended Borehole ments were carried out at depths of 10.1 m (No. 1), 18.0 m
Overcoring (CCBO) technique. This measurement technique (No. 2), 18.5 m (No. 3), 19.0 m (No. 4) and 19.6 m (No. 5)
can be applied to a water-filled vertical borehole for the in situ from the gallery floor. Through the use of in situ tests, we ver-
measurement of stress. ified that this improved measuring apparatus for a HQ-size
Sakaguchi et al. pointed out that the time-dependence of the borehole works well in a water-filled vertical borehole.
mechanical properties of the adhesive used to bond the strain The measured vertical stress σ v varied in the range 2.9 MPa
cell to the bottom of a borehole filled with water decreases the to 7.9 MPa. However, σ v at No. 5 was 7.9 MPa, and this was
sensitivity of strain measurement. This decrease in the sensi- approximately equal to the overburden pressure 7.3 MPa. With
tivity of strain measurement directly reduces the accuracy of regard to the magnitude of the principal stress, the minimum
stress measurement. To avoid such a decrease in the accuracy principal stress σ 3 at No. 4 shows the influence of some tensile
of stress measurement, they stated that we must consider the stress, but the other results were consistent with each other.
elapsed time after gluing the strain cell and estimated strain The maximum principal stress σ 1 tended to have a north-
coefficients by using a simulation model that incorporated a south orientation. It has been reported that the maximum
layer of adhesive and a strain cell consisting of epoxy resin and principal stress around the Kamaishi mine shows north-
brass. Moreover, for practical use, downsizing of the DCCBO south compression, which was consistent with the present
technique, such as downsizing of the borehole diameter and results.
the measurement equipment, is necessary. We verified the practicality of the measurement method
In this study, first, we downsized the DCCBO technique by comparing our results to those obtained at nearby sites
and performed in situ stress measurement at Kamaishi mine. by JNC (Japan Atomic Energy Agency). JNC conducted in
Second, we estimated in situ stress by using strain coefficients situ stress measurement by the CCBO, Hydraulic Fracturing
that were estimated according to the procedure of Sakaguchi method and Core-based methods (AE and DRA). When we
et al. Finally, we compared the results of this study with compared the magnitude of the principal stress estimated by
those obtained at a nearby test site by other methods, and we this study with those obtained by the CCBO, the intermediate
demonstrated the practicality of the proposed measurement principal stress σ 2 and the minimum principal stress σ 3 were
method. very similar. On the other hand, the magnitude of the maxi-
The main modification points for downsizing are as fol- mum stress estimated by the proposed method was less than
lows:(1) The outer pipe of the measuring apparatus was half of that by the CCBO, and about 70% of those by core-
downsized to HQ-size. (2) The dip/azimuth device consisting based methods. With regard to the difference in comparison
of a small digital video camera and a compass was changed to the CCBO, one possible explanation was that the mechan-
to a tilt/azimuth sensor. This tilt/azimuth sensor had a built-in ical properties of the adhesive in the actual test environment
data-logger for strain measurement. could not be accurately reproduced in the numerical simula-
We selected an epoxy resin-based adhesive (Bond E250: tion. With regard to the core-based methods, this difference
Konishi Co., Ltd.) to use. We performed a laboratory experi- might be due to the difference in the measurement princi-
ment to evaluate the measurement sensitivity under the same ple for the DCCBO technique, which is based on elasticity,
conditions as at the in situ test site. The sensitivity of mea- and that for the core-based methods, which are based on the
surement clearly fell. Therefore, we estimated the strain strain of inelasticity. We concluded that in situ stresses mea-
coefficients by using a simulation model that incorporated sured by the proposed measurement method were sufficiently
both the adhesive and the strain cell. reliable.
453
F. Sandrone
SBB, Lausanne, Switzerland
R. Wissler
Spacetec Datengewinnung GmbH, Germany
Regular inspections of tunnels are necessary for documenting niches, handrails, etc.) and to draw them on a separate layer.
their actual conditions, clarifying the evolution rate of detected This operation allows estimating the serviceability level of
problems and planning repairs. According to the traditional the tunnel installations and equipments. Moreover, when the
inspection procedure for tunnels belonging to the Swiss rail- tunnel is not newly constructed, it is possible to compare suc-
ways, (SBB 2007a, b) all the identified problems should be cessive analyses results as well as successive scan data, which
mapped in a specific inspection drawing once each 6 years. allows assessing the evolution rate of observed damages.
This means that according to the tunnel size, length and con- Due to the application of these new techniques, a new
ditions a visual inspection may often require a quite long inspection methodology (SBB 2011) has been developed for
time to be completed. Today, in spite of an increasing traf- improving and optimising the duration of the tunnel inspec-
fic and demand for safe operations the time for performing tion. The use of scanner systems leads to a considerable
inspections and maintenance of the infrastructure is constantly reduction of the time spent in the tunnel for a visual inspec-
reduced. Therefore, it is necessary to develop new systems for tion. As a matter of facts, according to this new process, only
tunnel survey and maintenance. Those systems should be able to the hammering and verification (i.e. measurements directly
to obtain more accurate results in a shorter time. done in the field by the inspection team, e.g. crack opening
Though the possibilities of using laser scanning techniques width, depth of empty joints) phases need to be realised in the
for detecting problems affecting the tunnel lining was already tunnel while the damages identification, mapping and tunnel
discussed by Haack et al. (1995), it is only today, with the conditions analysis phases can be done in the office. Fur-
improvements in resolution and quality of the measurements, thermore, working in the office with a scanned image in the
that laser scan can be considered one of the quickest and most drawing background strongly reduces the problems affecting
accurate systems for survey and monitoring conditions of the inspection quality observed in the past and mainly due to the
infrastructure (Yoon et al. 2009). Swiss railways currently use unfavourable working conditions in tunnel at night.
laser scan data for clearance profile analyses. Additionally
to the coordinates (i.e. tunnel profile), the used scanner sys-
tem (Spacetec Datengewinnung GmbH) provides in one single REFERENCES
recording process also the temperature and the grey value of
each point. Thus, together with the tunnel profile, the scan- Haack, A., Schreyer, J., and Jackel, G. 1995. State-of-the-Art of
ner produces a visual and a thermal image of the tunnel (i.e. Nondestructive Testing Methods for Determining the State of a
Tunnel Lining. Tunnelling and Underground Space Technology,
a record of the temperature distribution on the tunnel lin- 10(4): 413–431.
ing intrados surface). These images can be analysed with a Inagaki, T., and Okamoto, Y. 1997. Diagnosis of the leakage point on
specific tool which allows damages mapping and statistical a structure surface using infrared thermography in near ambient
analyses of recorded findings. Temperature patterns may give conditions. NDT & E International, 30(3): 135–142.
additional information if compared to visual images. In par- SBB 2007a. Conservation des tunnels (I R 20020): 37 pp.
ticular since cavities or water leaking in the structure appear SBB 2007b. Inspection principale des tunnels. Cahier des charges
in a thermal scan as temperature anomalies, thermal data can (D I-PS-IB 04/07): 16 pp.
be very helpful in identifying the main causes of observed SBB 2011. Unterhaltsrichtlinie Tunnels. Arbeitsanweisung: 39 pp.
damages affecting the shotcrete lining (Inagaki & Okamoto Yoon, J.-S., Sagong, M., Lee, J.S., and Lee, K.-S. 2009. Feature extrac-
1997). These analyses provide a clear image of the structural tion of a concrete tunnel liner from 3D laser scanning data. NDT &
E International, 42(2): 97–105.
conditions of the tunnel.
By analysing the visual images it is as well possible to iden-
tify all the installations characterising the tunnel (i.e. catenary,
454
Hydraulic fracturing is a common technique used in the oil creation of a strong bond between the mortar and the sand.
and gas industry for stimulation of tight formations in order to The samples were preserved in water bath for a period of 28
increase productivity. When a hydraulic fracture approaches days in order for the cement to reach to its final strength. The
an interface, which could be a second formation or a small blocks kept in room temperature for one month and were given
pocket of shale accumulation or sand lens, it may get arrested adequate time to loss their water gradually. The oil coated
or cross the interface. This depends on the state of in-situ interface in both blocks was separated for a fraction of mil-
stresses, rock mass properties and the interface characteristics. limeter in order to generate an open interface. A 6 mm hole
From practical point of view, sometime the induced fracture is was drilled in the centre of each block, i.e. in sandstone for
expected to be contained within the reservoir section: the cap sample I and mortar for sample II. A fracturing fluid with
rock integrity is the main objective in this case. The interaction a high viscosity of 100 Pa.s was injected at a constant flow
of an induced hydraulic fracture and an interface with different rate of 50 cc/hour using computer-controlled high pressure
rock types in each side of the fracture is the problem to be syringe pumps. The magnitude of stresses used for testing
studied in this case. However, the fracture may need to intersect sample I were σv = 2500 psi, σH = 2000 psi and σh = 500 psi.
and cross a number of sand packs in order to propagate to the Hydraulic fracture directly crossed the interface at both sam-
required length. An example of this is in tight gas formations ples except for one wing of fracturing test II that crossed the
where gas bearing sandstones are lenses distributed randomly interbed by offsetting. High normal stress on the interface
between shale formations: crossing different layers is the main (σH = 2000 psi) would be the main reason for crossing as it
objective in this case. provide enough fractional force on interface surface.
Simulation of hydraulic fracturing in an unconventional The numerical simulation carried out to investigate the
reservoir which is likely to have lenticular sedimentation, effect of different parameters on the interaction mechanism
directional anisotropy and heterogeneity would be more com- using PFC2D. PFC2D is a district element method (DEM)
plicated than conventional formations. For simplicity, such based numerical code developed by Itasca group and it uses the
formations may be considered as a combination of number of assembly of number of circular elements (2D disks) bonded
individually uniform and homogeneous sections being sepa- together to generate the model. The macro properties of the
rated with one or more interfaces. These interfaces may be samples (UCS, friction coefficient etc) used for this study were
any kind of discontinuity like fractures, interbed, or a bound- estimated from the model micro properties (bond strength,
ary across which the rock type changes. An ideal fracture is friction) which are the input to PFC models. This was done by
expected to propagate longitudinally far enough to satisfy the performing several simulations including bi-axial, direct and
production needs and cross any interfaces that may exist along Brazilian tensile tests from which the rock properties were
its propagation path. At the same time, the hydraulic frac- determined through the plot of Mohr Circles corresponding
ture should be contained within the reservoir section, i.e. does to different stress levels. Similarly, micro and macro hydrauli-
not cross the top and bottom interfaces. This indicates the cal properties were estimated using permeability simulation
multi-functionality expected from an initiated fracture. tests. A hydraulic fracture was initiated in the centre of the
This paper presents the result of numerical simulations sample and model response was monitored.
using PFC2D corresponding to two lab experiments that were Modeling the natural discontinuity may be the most dif-
performed using a true triaxial stress cell (TTSC). The inter- ficult part of this study. This is due to the limited infor-
action mechanism was simulated when a fracture approaches mation being available for the two open (oil wetted) and
an interface across which the formation properties are differ- close cement-sandstone interfaces. Modeling such scenarios
ent. Changing the location of formations with respect to each requires detailed interface characteristics including normal
other and also the interface properties the simulations result bond strength, shear and normal stiffness, friction and dila-
in different model responses. Sample I includes a sandstone tion angles, and cohesion of the discontinuity plane. The
block occupies a volume of 15 × 13.5 × 8 cm in the middle cement-sandstone bond across the closed interface was so tight
of a 15cm mortar cube. In sample II, the mortar fills in the and therefore assigning the average cement and sandstone
volume between two blocks of sandstone of 15 × 13.5 × 4 cm mechanical properties appears to be appropriate. Available
size located on both sides of the mould. The interfaces are fracture properties in the literature were used for modeling
parallel and perpendicular to the sample base. In both sam- the oil wetted (not fluid saturated) interface. This kind of
ples, before pouring the mortar in the mould, one side of the fracture is commonly referred to as gas filled fracture. The
sandstone blocks was coated with oil in order to prevent the experimental set up for estimating the normal stiffness is more
455
456
GPS displacement monitoring system for the precise measuring of rock movements
N. Shimizu
Department of Civil and Environmental Engineering, Yamaguchi University, Japan
T. Masunari
Furuno Electric Co., Ltd., Japan
T. Iwasaki
Kokusai Kogyo Co., Ltd., Japan
457
T. Shidahara
NEWJEC Inc., Formerly at Nuclear Waste Management Organization of Japan (NUMO), Japan
T. Hatano
J-POWER EPDC, Formerly at Nuclear Waste Management Organization of Japan (NUMO), Japan
S. Arai
Geosphere Science Lab., Formerly at Nuclear Waste Management Organization of Japan (NUMO), Japan
T. Miwa
Tohoku Electric Power Co., Formerly at Nuclear Waste Management Organization of Japan (NUMO), Japan
For the selection of the Detailed Investigation Areas for HLW at Quarternary volcanic zone, fault crush zone and hydrother-
disposal in Japan, predicting the tunnel constructability is one mally altered zone of Green Tuff area, mudstone and fault
of the requirements together with assessing long-term safety. crush zone of sedimentary rock of Neogene and later.
This paper dealt with the geological factors relating to difficult Information useful for predicting rockburst has been gath-
tunneling such as squeezing and rockburst. Also it dealt with ered from previous reports. In the Preliminary Investigation
the prediction of rockburst and squeezing. Another paper in stage, geological survey, geophysical survey and borehole
this Congress (Sawada et al. 2011) deals with the engineering survey from the surface are the source of information. There-
characteristics of rock mass through rock mass classification fore rock type, P-wave velocity from seismic exploration
based on rock core. and in-situ rock stress from hydrofracturing have been con-
This paper about difficult tunneling has been based upon sidered. Majority of rockburst events occurred at granitic
analysis of more than 500 tunneling reports about 280 tunnel rock, excluding coal mine where different kind of rockburst
construction. The causes of difficult tunneling are related to occurred at pillars. And P-wave velocity was around 5 km/s at
(i) underground water, (ii) mechanical properties of the rock, the rock of rockburst events. Horizontal maximum and min-
or (iii) others such as gas. The geological factors for exces- imum stresses σH and σh measured by hydrofracturing have
sive water inflow are porous volcanic product of Quarternary, been tested as a criterion for rockburst. It has been inferred
fault crush zone and hydrothermally altered zone of Green that σH − σh > 30 MPa is one of a good criterion for rockburst
Tuff area, and degenerated mixed rock in accretionary com- occurrence. When rock stress data is not available, continu-
plex. The geological factors for squeezing are solfataric clay ous occurrence of borehole breakout is also a good indicator
of rockburst.
This paper also dealt with the prediction of squeezing of
rock in tunneling. For this study, 187 Japanese tunnel projects
which encountered squeezing of rock were collected. And the
relationship between the degree of squeezing (δ/D) and geo-
logical structure, results of seismic exploration and various
rock properties from laboratory tests have been analyzed. As
a result, the followings have been found. i) Many squeezing
incidents are found near anticline, syncline and thrust fault.
Within the data collected, the hanging wall had more incidents
of squeezing. ii) Many squeezing incidents occur in the region
where Vp is less than 2.5 km/s. iii) Cation Exchange Capacity
has not been found to have correlation with δ/D. The parame-
ters which show correlation with δ/D include the ratio of UCS
to overburden pressure (Gn), unit weight (γ) and sand con-
tent (S). (See Figure). Using these 3 parameters, a criterion
Figure 1. Distribution of 3 ranks of δ/D in the 3D-space. Stereo for predicting the squeezing has been established. These find-
graph by parallel method (left graph for left eye and right for right) ings can be used during the following each stage of surface
S (%), γ(kN/m3 ), Gn (-), δ/D = convergence divided by diameter of geological survey, surface geophysical survey and borehole
a tunnel (degree of squeezing). survey.
458
Rajbal Singh
Central Soil and Material Research Station, New Delhi, India
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: field measurement; hydraulic fractures; rock caverns; rock mass; rock stress; tunneling
The Central Soil and Materials Research Station (CSMRS), jack tests were 2.67 and 2.96 MPa, respectively with a stress
New Delhi has conducted a number of stress measurements ratio of 1.11. The overburden stress with a height of 70 m over
for many hydroelectric and highway projects being con- the testing site was 1.89 MPa which was less than the mea-
structed/completed in the Himalayan region of India, Nepal sured vertical stress. The average vertical stress as measured
and Bhutan. The methods of measurements were flat jack, by hydraulic fracturing test was 3.795 MPa which was almost
over coring and hydraulic fracturing. There is lot of scatter in two times the overburden stress. It is, therefore, clear from the
the data of these measurements. The horizontal stresses were in-situ stress tests conducted at the site that the in-situ stresses
higher than the vertical stresses. The stresses estimated based in Himalayas can not be estimated based on the overburden.
on overburden were not matching with in-situ measurements. The measured vertical stress was more than the stress due to
Results of all the tests at different project sites have been overburden and horizontal stress was higher than the vertical
analysed and will be presented in this paper. The result of the stress. It is also concluded that there was very large variation
flat jack tests should be used with cautions, as the tests are con- in the results of in-situ stress by flat jack test as compared to
ducted on the surface of the drift, which is highly influenced hydraulic fracturing.
by the disturbances caused during excavation of the testing The case study of another project in the same vicinity of the
drifts. However, flat jack tests are not being conducted these Himalayan region will be presented along with interpretation
days. of data and a comparison with estimated stresses based on
This paper deals with the comparison of flat jack tests with overburden. A comparison of in-situ stresses based on over-
the tests conducted by hydraulic fracturing tests. The aver- burden and measured by flat jack and hydraulic fracturing will
age vertical and horizontal stresses as determined from flat be discussed in the paper.
459
Ki-Seog Kim
Heesong Geotek Co., LTD, Seoul, Korea
In in-situ stress measurement by hydraulic fracturing method, and pre-existing fracture. The typical pressure-time record
the values of the fracture reopening pressure (Pr ) and the obtained by hydraulic fracturing and the typical fracture pat-
fracture shut-in pressure (Ps ) are the governing parameters terns are represented in Figure 1. The fracture patterns are
in stress analysis. These pressure parameters are commonly classified according to the open-closed condition of the frac-
determined from the pressure-time (P-T) record obtained by ture plane at the unpressurized state and the flow condition
the conventional hydraulic fracturing (HF) and/or sometimes through the fracture plane at the pressurizing state. The three
the pressure-flowrate (P-Q) record obtained by the hydraulic fracture patterns are selected from the observation of the pres-
testing on pre-existing fracture (HTPF). However, it is not so sure drop tendency in the shut-in cycle of P-T record and the
easy to take the exact pressure levels in the P-T record corre- investigation of the fracture features by a borehole televiewer.
sponding to the actual reopening and shut-in behavior of the The closed fracture is defined as Frac-C, in which the frac-
fracture. According to whether the fracture plane is open or ture is completely closed at the unpressurized condition and
closed, the pressure level, which is apparently regarded as the shows the closed-loop flow within the fracture plane during
point of the reopening or shut-in in P-T record, may indicate the reopening to shut-in cycle. The slightly open fracture is
a different meaning in the actual pressure-fracture behavior. defined as Frac-S, in which the fracture is open at just tip at the
In this study, the validity of the determination of the pressure unpressurized condition and shows a closed-loop flow within
parameters (Pr and Ps ) in P-T record and the selection of the the fracture plane during the reopening to shut-in cycle. The
governing equation for stress analysis are discussed with the wide open fracture is defined as Frac-W, in which the fracture
fracture pattern and the stress estimation results. is open to some extent of the fracture plane at the unpres-
The field hydraulic fracturing is carried out in depths of surized condition and shows an open-loop flow through the
100–200 m of a vertical borehole in a mine region where rock fracture plane during the reopening to shut-in cycle.
mass is more or less fractured. The test sections have vari- In hydraulic fracturing using a vertical borehole, the hori-
ous fracture patterns, including new hydro-induced fracture zontal stresses are calculated in general by Equations 1 and 2
Figure 1. Governing parameters of the stress analysis and the fracture pattern in hydraulic fracturing.
460
461
M. Tsesarsky
Department of Structural Engineering and Department of Geological and Environmental Sciences,
Ben Gurion University of the Negev, Beer-Sheva, Israel
G.H. Wust-Bloch
Department of Geophysical and Planetary Sciences, Tel Aviv University, Tel Aviv, Israel
Keywords: field measurements, monitoring, rock caverns, stability analysis, weak rock
462
R. Ulusay
Department of Geological Engineering, Hacettepe University, Ankara, Turkey
Ö. Aydan
Department of Marine Civil Engineering, Tokai University, Shizuoka, Japan
M. Geniş
Department of Mining Engineering, Zonguldak Karaelmas University, Zonguldak, Turkey
H. Tano
Nihon University, Koriyama, Japan
Keywords: monitoring, numerical modelling, stability analysis, lab testing, rock caverns
There are numerous antique and modern rock structures in acoustic emission (AE) devices, crack monitoring devices and
the Cappadocia Region of Turkey, which are excavated by temperature, humidity and pressure sensors were installed
mankind. In addition to historical and archaeological charac- onto the outside walls of the hall to observe the behavior of
teristics of the rock structures in the region, their long-term surrounding rock. This research project still continues and the
behavior is also an important data source in terms of rock paper describes the preliminary findings.
engineering. Even today, modern underground openings of The Cappadocia Region is mainly covered by volcanic units
different sizes and shapes in the region are being used for multi of Neogene-Quaternary period. The surrounding tuff of the
purposes. In several locations of Cappadocia, some stability Congress Center is the fall-out product (pumice flow) of the
problems of different scales were observed. One of the modern Kavak member. It is white in color and has a thickness of
underground openings in the region is the Avanos Congress about 7–8 m in the vicinity of the Congress Center. Except 2 m
Center. Avanos municipality planned to utilize a hill remnant thick deposits of thinly layered marly rocks and top soil, the
from an abandoned quarry as an underground congress center surrounding rock has no discontinuities and is very massive
in 1980s. The rough excavation of the congress center was and the bedding is widely spaced. In fully saturated condition,
done. However, the construction has not been completed due a decrease of about 40% occurs in strength of the surrounding
to financial problems and also the change of governors and tuff when compared to that in dry condition. Although the
mayors. The excavation has been completed by more than 10 RMR rating of the rock mass ranges between 70 and 75, in
years. The surrounding rock, in which the Congress Center terms of strength and deformability of the intact rock it is a
has been excavated, is a very soft Cappadocian tuff and sub- weak rock. Comparison of the wetting-drying and freezing-
jected to degradation of material properties, resulting from thawing test results indicates that freezing-thawing cycles are
cycles of wetting-drying and freezing and thawing processes. more effective on decrease in weight loss and UCS of the
Furthermore, it is known from previous experimental works tuff studied. Based on the 4-cycle slake durability tests, Id
that the tuffs of Cappadocia exhibit strong creep behavior. values from the 1st to 4th cycles are between 81.5–98, 69.3–
The observations done by the authors inside and outside of 94, 56.7–88.3 and 40.7–77.1%, respectively. Particularly after
this uncompleted congress center revealed that there are some the 2nd cycle a considerable disintegration of the samples
new fracture formation and propagations at several locations was observed. Freezing-thawing shows its effect as the flaking
in the underground halls as well as at the perimeter of the hill. and/or slabbing of rock surface as observed by the authors.
These cracks are easily recognizable during drying process of The acoustic emissions induced by fractures coincide with the
rock soon after a rainfall. observation periods
This study aims to clarify the short- and long-term stability The experimental results indicated that the surrounding
of the Avanos Congress Center in terms of rock engineering. rock was quite vulnerable to cyclic freezing and thawing and
For the purpose, experimental studies concerning short-term wetting-drying processes.The strength of rock is reduced dras-
properties as well as the effects of freezing-thawing, cyclic tically under saturated conditions and the process of freezing
wetting-drying were carried out, and the stability of the open- and thawing accelerates further degradation of rock under such
ing was assessed by analytical and numerical techniques by a condition. The simple short-term stability analyses by ana-
considering the data from experimental studies. Two complete lytical methods and finite element analyses clearly showed
463
464
Evaluation study of nozzle design selection for hydraulic fracturing test on concrete,
limestone and andesite with triaxial loading at the laboratory
As it has been well known that hydraulic fracturing test is a Table 1. Physical properties of the samples.
method whereby in-situ stress at a borehole can be predicted
and this method employs the principles of hydraulic. When Sample ρnat ρdry ρsat W S n
the fluid pressure is greater than the tensile strength of the Type (gr/cm3 ) (gr/cm3 ) (gr/cm3 ) % % % e
borehole wall fractures will then be initiated. Concrete 1.72 1.54 1.95 11.60 44.65 40.62 0.69
In this research, the effect from different types of fluid Limestone 2.37 2.37 2.38 0.28 50.00 1.32 0.01
burst to the result of hydraulic fracturing was studied. The Andesite 2.20 2.13 2.27 3.62 53.89 14.13 0.17
hydraulic fracturing test in this research used different con-
figuration of nozzles. In order to create various fluid burst
directions, three different nozzles were designed and used and
each of these enables vertical spray, horizontal spray and a Table 2. Mechanical properties of the samples.
combination of vertical and horizontal sprays. The hydraulic
fracturing tests were conducted in three types of cubical sam- Sample σt σc E c φ
ples such as concrete, limestone and andesite with dimension Type (MPa) (MPa) (MPa) ν (MPa) (◦ )
of 25 × 25 × 25 cm, and triaxial loading was applied for each
Concrete 1.08 10.93 4.07 0.13 3.71 33.29
sample. Limestone 6.06 34.57 9.85 0.25 9.17 55.14
The physical and mechanical properties of the samples can Andesite 6.54 62.36 10.73 0.19 12.38 56.62
be seen at Table 1 and Table 2.
The summary of the testing results can be seen in Table 3.
There were nine tests in total, consisting of each type of nozzle
used in each type of sample once. Table 3. Summary of breakdown pressure.
The tests revealed that the most efficient type of nozzle that
could initiate fracture in all of those samples was the nozzle Breakdown pressure (MPa)
that bursts a combination of vertical and horizontal spray. The
tensile strength properties of the sample also have influence Sample Vertical Horizontal Combination
to the breakdown pressure, that the higher tensile strength will Type nozzle nozzle nozzle
have the higher breakdown pressure. The direction and shape
Concrete 6.01 4.94 0.76
of fracture induced by hydraulic fracturing test in this research
Limestone 18.37 12.17 11.86
is parallel to the maximum horizontal stress and perpendicular Andesite 22.22 18.95 18.80
to the minimum horizontal stress.
465
Robert J. Whiteley
Coffey Geotechnics, Sydney, Australia
Keywords: case studies, coal mines, rock stress, risks and hazards
466
467
A.E. Alvarez-Vigil
Department of Mathematics, University of Oviedo, Asturias, Spain
At the time to evaluate the environmental impact created by is a strong damping with depth. Therefore, in order to study
blasts at quarries, in general is considered the noise effect the damage on underground structures, suitable controls were
and the vibrations, which can cause damage to neighboring carried out by means of systems set out that allow register data
buildings. of the actual structure liable to be affected.
In recent decades the execution of blast vibration control Another important effect on the transmission of vibrations
measures has become general. The reason is the increase pre- is caused by the structure of the rock mass which induces
occupation shown by the administrations over environmental anisotropy to the transmission laws. The presence of an impor-
and social impact caused by civil and mining works. tant fault crossing the quarry defines the difference between
In some Quarries the proximity the road slopes or under- two areas, so it has been measured the absorption of the blast
ground structures like tunnels, water or gas conductions, even wave trough the fault. Consequently, the blasts made at each
another underground exploitations, it turns in a problem for the side of the fault have different levels of risk respect the build-
possibles complaints to these elements. Blasts complaints to ings situated in the area, and we traduce this to directional
the rock mass are: the decrease in the resistance of the joints transmission laws.
that can instabilize wedges or blocks, permebility changes, Deep vibrations due to blasts have been measured using two
without forgetting the dinamic effect of a shock wave, that three-axial geophones for boreholes. Surface vibrations have
introduces an instantaneous force that it takes part directly in been measured using three-axial stations type ETNA of high
the balance of the forces that define the wedges and blocks dynamic range.
stability. Initially, an analysis of the vibrations generated by blasting
This paper describes a set of control of vibrations produced was performed. To do so, there was a set of measurements
by the blasts in a limestone quarry in Spain. Also it is analyzed in surface, recording 17 blasts. In order to measure deep
if they possibly can affect an underground mine sited near the vibrations two geophones were placed into two boreholes
quarry. This mine is a experimental exploitation of limestone near the underground mine and at its level. An ETNA sta-
with three pillars. tion was sited too in the underground mine and another two in
By means of a comparative analysis of the measured data different places of the quarry. These equipments registered
of vibration in surface and with depth it is verified if there 31 blasts.
471
E. Arpaz
Vocational school of Kocaeli, Kocaeli University, Kocaeli, Turkey
Ö. Uysal
Mining Engineering Department, Dumlupinar University, Kütahya, Turkey
Y. Tola
Vocational school of Kocaeli, Kocaeli University, Kocaeli, Turkey
K. Görgülü
Mining Engineering Department, Cumhuriyet University, Sivas, Turkey
M. Çavuş
Seyitomer Coal Enterprise, Kutahya, Turkey
Keywords: blasting, blast induced vibrations, vibration predictors, peak particle velocity
In sectors such as mining, quarrying, excavating tunnels and Totally, 93 particle velocity data were taken during the blast-
pipeline trenches, blasting is indispensible. Thus, these sec- ing from the three panels (Panel-1 33 recordings, panel-2 46
tors are always prone to suffer from adverse conditions caused recordings, panel-3 14 recordings).
by ground vibrations, air shock, and flying rocks. Therefore, The data obtained from peak particle velocity recordings
when planning the blasting, besides crashing performance, were exposed to evaluation on the basis of panels the whole
size distribution, and cost and other economic issues, the recordings and peak velocity prediction methods that are
likelihood of the problems mentioned above must also be found in the literature (based on 11 different methods).
reckoned. It is of utmost importance to predict and measure From the evaluation data, it can be seen that different meth-
blast-related problems in the way to reduce environmental ods have different regression values. These variations raise
concerns that may arise. doubts concerning the reliability of the methods.
The issue is being studied extensively. Nevertheless, an As a result of the evaluation, the following were found:
agreement does not seem to have been reached on the pre-
1. All methods produced different reliability rates based on
diction and prevention of blast-induced vibrations. The lack
different panels.
of agreement can be explained by the variety of conditions of
2. Scaled distance method, which is widely accepted and
blast environments.
favored in the literature, did not produce clear rates in the
This study is intended to the results of studies that have
recordings.
been carried out at TKİ Seyitomer Coal Enterprise in order
3. No method surpassed by a high regression rate in all of the
to measure the performances of equations that can be used in
panels.
predicting blast-induced vibrations
4. The conclusion that can be drawn from the results is that
The experiments were performed in the dragline panel of
no single method provides the highest reliability rate in
Seyitomer Coal Enterprise, Kutahya, Turkey. The panel height
predicting blast-induced vibrations.
was 24 m. The geological formation over the coal in the panel
5. It can also be seen that the best method must be selected
was marl. ANFO was used for rock fragmentation. The blast
after investigating the effectiveness of all predictors based
holes were 24 m long, with 228 mm diameter. The distance
on their regression value.
between the blasting holes was 10 m. ANFO, power gel dyna-
mite and nonel capsule were used for the explosion. The delay
between each hole was 42 ms and in-hole delay, 25 ms.
472
Keywords: back analysis, dynamics, numerical and physical modeling, risks and hazards
473
1 INTRODUCTION
Tests on smooth discontinuities were performed by applying This can be considered as an ultimate cycling angle that does
900 displacement cycles. Waveform of displacement was sinu- not coincide with the ultimate monotonic friction angle since
soidal in time with each cycle characterized by an amplitude it is conditioned by the cycle amplitude. The cycle amplitude
of 2 mm, a frequency of 0.5 Hz. and a duration of 30 minutes. involves a limited discontinuity portion and consequently it
Displacement for each cycle was 8 mm, and data acquisition determines a localized damage of the asperities that it is dif-
frequency was 5 Hz. Different confining stresses were applied. ferent by the damage caused by a larger displacement induced
Normal stress was constant throughout the test. Results have by a monotonic load.
shown that the friction angle determined by cyclic tests is Based on Homand’s formulation for the assessment of the
higher than the value obtained with the monotonic tests (40◦ damage index Dw (2000) under cyclic shear tests, a new index
and 33◦ , respectively). This difference is due to a higher veloc- has been defined as follows:
ity of load application if compared to standard monotonic
shear tests.
Following the same procedure adopted for smooth dis-
continuities tests were carried out on rough samples, during
those tests the following parameters were varied as: maxi- where σ0 is the initial compressive stress and σι is the mean
mum displacement (cycle amplitude u0 ) ranged between 1 and shear stress at the i-th cycle. C0 is the uniaxial compres-
4 mm; cycle frequency (0.013–3.9 Hz); initial normal stress sive strength of the intact material, βd is the progressive
(1–1.5 MPa) and total number of cycles (112–450, so as to degradation function defined by Belem et al. (2007) as
obtain a final total shear displacement equal to 1800 mm). follows:
Generally, tests showed a normal rapid decrease in stress in
the first cycles, reaching an asymptotic value for a total dis-
placement of 900 mm (50% of the total number of the applied
cycles). Due to the geometric anisotropy of the discontinuity
along the shear direction, the shear stress showed an asym-
metric behaviour as in figure 1 (negative values correspond to where us-tot is the total shear/relative displacement, a0 is the
upwards movement of the shear apparatus, whereas positive roughness amplitude, and ka , DRr and θs are the joint geomet-
values correspond to downwards movement). rical parameters. This formulation is valid for a cycle number
The data obtained were analysed by considering the peak N > 1.
stress values, both for upwards and downwards shearing. Let us highlight that this index is strictly related to the cycle
At the end of each cyclic shear test, the friction angle was amplitude, values of ID close to 1 indicate the reach of ultimate
computed. Its average value was 45.2◦ , higher than the one dynamic friction angle, while other values indicate a partial
obtained from cyclic shear tests along a smooth discontinuity. degradation of the asperities.
474
475
Cheng-Hsun Chen
CECI Engineering Consultants, Inc., Taipei
Tai-Tien Wang
Institute of Mineral Resources Engineering, National Taipei University of Technology, Taipei
Tsan-Hwei Huang
Department of Civil Engineering, National Taiwan University, Taipei
1 INTRODUCTION
By means of the 2D finite element method this study inves- Figure 1. Idealized scheme of numerical model.
tigates the stress increment of a tunnel lining subjected to
seismic wave caused by an earthquake. Figure 1 shows model
configuration and associated setups for boundaries and seis- tunnel vault (σNθ , σVθ and σM θ ) is recorded during simulation.
mic wave input. The incident wave is applied by a harmonic The maximum and minimum values can be determined after
sinusoidal displacement right along the left boundary. The the time history of seismically induced stress has been com-
upper boundary of the model is a free surface, while the left and pleted. These maximum magnitudes of seismically induced
right boundaries are set to be “absorbent” boundaries which axial-, shear- and flexural stress are noted as (σNθ )max , (σVθ )max
are allowed to move with minimal reflecting wave from the and (σM θ )
max , respectively. Accordingly, the maximum mag-
two sides. The lower boundary is also an absorbent one. To nitudes of seismically induced stress, for instance, the axial
enhance the precision of numerical simulation, various model stress, in the liningcan be obtained by comparing (σNθ )max all
scopes for distinct tunnel depths have been tested to minimize over the profile of the tunnel. Then the maximum seismi-
the interference from two lateral boundaries. cally induced axial stress on the whole tunnel lining can be
In semi-infinite domain in case the incident angle of an inci- determined and written as σN . So as the maximum seismi-
dent S-wave is greater than the critical angle and the distance cally induced shear- and flexural stress on the whole tunnel
between the incident position and the tunnel is large enough, lining, σV and σM , respectively. These maximum seismically
the incident S-wave will reflect in a form of R-wave in addition induced stresses are further normalized by the maximum stress
to S- and P-waves. This study generates a harmonic S-wave amplitude of the incident wave σ0 , for discussion.
with the frequencies of 1 and 3 Hz as an incident wave in the Figure 2 shows the variation of maximum seismically
left boundary of the model, R-wave is then generated within induced axial- and flexural stress in the lining for various
some region in the model. The simulation results fit in with the tunnel depths caused by an incident wave with a frequency
analytical solution and the simulation model is then verified. of 3 Hz. The maximum seismically induce stresses on the
lining increases as the tunnel depth increases significantly
3 SIMULATION RESULTS AND DISCUSSION and peaks at H /λ = 0.06. While H /λ > 0.06 the seismically
induce stresses vary minor. Figure 3 shows the variation of
The seismically induced axial-, shear- and flexural stress in the maximum seismically induced axial- and flexural stress in the
point on the lining with a θ angle clockwise measured from lining for various tunnel depths caused by an incident wave
476
477
Y. Nakamura
Kumamoto National College of Technology, Japan
Y. Ogata
National Institute of Advanced Industrial Science and Technology (AIST), Japan
K. Kaneko
Graduate School of Engineering, Hokkaido University, Japan
Dynamic tensile properties of rocks are a factor of consid- dynamic events and the motion analysis of the digital images
erable importance in assessing the stability of underground was available to programmable operations. Kimachi sandstone
structure under dynamic loadings such as blasting, rocks was prepared as 50 mm in diameter and 26 mm thickness.
burst, earthquake etc. It is also essential to understand The density, elastic wave velocity and elastic modulus of the
that fractures initiate and propagate under different loading sandstone were 2 g/cm3 , 2710 m/s and 6.5 GPa respectively.
conditions. Tage tuff was also prepared in the same geometry. The den-
This study used the underwater explosive-driven loading sity, elastic wave velocity and elastic modulus of the tuff were
apparatus and split Hopkinson pressure bar (SHPB) system 1.76 g/cm3 , 2380 m/s and 4.5 GPa respectively. Note that the
to measure dynamic tensile strengths of rocks. The underwa- selected rocks belong to medium and soft rock respectively
ter explosive-driven loading apparatus consist of steel frame because of the capacity of the dynamic loading apparatus. The
and an electric detonator, and impact load cells and high- loading velocities were estimated from the position of upper
speed video technique were adopted to evaluate the fracture load plates with time. The average ratio of Sd /Ss (dynamic
behavior of rocks. SHPB system involving a pulse shaping strength to static strength of static tensile strength) was 1.79
technique was used to impact the same geometry rock sam- for sandstone and 1.84 for tuff.
ples under dynamic stress equilibrium and constant strain rate. Microfocus X-ray computerized tomography (CT) system
The microfocus X-ray Computerized Tomography was used to has been used to observe microstructure of the rock test sam-
visualize dynamic damages inside the impacted samples non- ples. The sandstone samples were scanned covering entire
destructively. The dynamic fracture processes of rocks under diameter of 50 mm and height of 25 mm. Figure 4 shows
indirect tensile loading were analyzed using the dynamic a schematic diagram of the micro CT system. X-rays from
fracture process analysis (DFPA) code. microfocus X-ray generator are partially attenuated by a spec-
It is also essential to understand how fractures initiate and imen that is made to rotate in equal steps in a full circle
propagate under different loading conditions in order to deter- about a single axis close to its center. The system can achieve
mine rock breakage and fragmentation in rock blasting and 2048 × 2048 pixel reconstructions and obtain data needed for
percussive excavation. The dynamic behavior of brittle mate- 3-D image in one scanning session. The slice thickness was
rials, including rocks and concrete, show complex fracture set to be 0.076 mm for 370 slices. The cone beam can be
patterns and fracture processes when they are subjected to used to scan the entire diametric sample. The specimen table
high strain rate loading. Fractures may occur from pre-existing was placed at 400 mm away from the X-ray tube. The setting
flaws or micro-cracks under applied stress condition. The condition provided a magnification of 5.21. Figure 2 shows
present study used underwater explosion-driven loading by the 3-dimensinal volume-rendering image from 370 X-ray atten-
detonation of an electric detonator in a water filed chamber, uation slice scan images obtained by cone-beam geometry of
to fracture along the diametric tuff and sandstone specimens. X-ray tube. The open arrow shows the direction of load. The
Figure 1 shows the schematic diagram of the dynamic appa- contrast in the images shows differences in the density of min-
ratus. High-speed videography system was used to observe eral; that is, the white portions indicate the places where high
the fracture processes in the rock samples. The frame rate of density minerals exist while the black indicates low density
the video can was variable from 30 to 40,500 frames per sec- such as air. The dynamic tensile-splitting behavior of the sand-
ond. The system was capable of recording 49,152 frames of stone samples was observed by high-speed video camera and
478
479
1 INTRODUCTION
2 EXPERIMENTAL SETUP
where Pmax is the peak load, S is the distance between the
A 25 mm diameter SHPB system is used in this study (Fig- two loading pins, a is the crack length, B is the thickness and
ure 1). Assuming one-dimensional stress wave propagation, R is the radius of the specimen. The key factor Y in the two
the forces on the incident end (P1 ) and transmitted end (P2 ) equations can be calibrated using finite element analysis.
of the sample can be calculated (Kolsky 1953). The fracture energy and fracture toughness can be quanti-
The LGG system is developed to monitor the CSOD. By fied using the energy conservation principle. The total energy
recording this voltage output of the photodiode, the gap open- absorbed by the specimen (W) can be estimated from the
ing distance (i.e., CSOD) ant thus the opening velocity can energy difference during the stress wave propagation. Part
be obtained with pre-calibration. Laurentian granite (LG), is of this energy creates new crack surfaces, the total fracture
selected in this study to demonstrate the two methods. The energy (WG ); the remaining is the residue kinetic energy (K)
geometry of the NSCB and CCNBD specimen is shown in of cracked samples. With the aid of LGG, the total rotational
Figure 1 as well. kinetic energy for NSCB and CCNBD tests can be measured.
Thus, the average propagation fracture energy is Gc = WG /Ac ,
where Ac is the area of the crack surfaces created. The average
3 MEASUREMENT PRINCIPLES dynamic propagation fracture toughness is:
The mode-I initiation fracture toughness can be calculated
using equation 1 for NSCB specimen and equation 2 for
CCNBD specimen following ASTM standard E399-06e2.
480
481
1 INTRODUCTION depth and the unit weight of the overlying rock, and the ratio
between horizontal and vertical stress was set to two.
A train moving along an underground railway will generate The bulk modulus was set to 30 GPa, the shear modulus to
vibrations in the rail and the underlying track structure. These 18 GPa, and the density to 2700 kg/m3 . The friction angle of
vibrations will radiate into the surrounding ground as elas- the discontinuity was set to 30◦ .
tic waves and will, in densely populated areas, reach nearby The dynamic load used in the analyses was one period of
buildings and its residents. a sinusoidal wave with a duration of 0.01 sec and a particle
The propagation of train-induced vibrations is complex due velocity of 100 mm/s at a frequency of 100 Hz.
to the scale and the many uncertainties involved. In the past, The UDEC-model was 100 m high and 100 m wide. The
empirical methods were used to analyse problems related to tunnel was shaped as a square with a height and width of
train-induced vibrations, while today the analyses are per- 10 m. Due to the tunnel shape, symmetry was used in order to
formed with the aid of numerical models in combination with reduce the model size. The damping ratio was set to 2%.
field measurements. During the analyses history points on the ground surface
A common assumption in numerical analyses is to repre- were used to monitor the maximum and minimum (most neg-
sent the ground by a homogenous and isotropic material. For ative) vibration levels in y-direction (vertical). The largest
situations where the ground consists of clay and soil assuming absolute value was determined as the vibration level at each
homogenous and isotropic material may be correct, but for a point.
discontinuous rock mass it is not. The vibration levels on the ground surface near a tunnel was
In a rock mass the presence of discontinuities will have a analysed for the following parameters:
significant impact on the transmission of waves, since a wave
• overburden – four different overburdens were studied; 10,
propagating across a discontinuity will be partly reflected and
25, 40, and 55 m, respectively. No discontinuities were
partly refracted. How much that is refracted and reflected,
included in the model during these analyses.
respectively, depends mainly on the normal and shear stiff-
• position of a single horizontal discontinuity near the
ness of the discontinuity and wave frequency, but also on
tunnel – three different positions were studied: 5 m above
the angle of incidence and the number of discontinuities. The
the tunnel roof, 5 m below the tunnel floor, and 5 above
propagation of waves in discontinuous rock masses has mainly
the tunnel floor (i.e. middle of the wall). The normal and
been studied with regard to blasting and explosions, i.e. high-
shear joint stiffness was 1 and 10 GPa/m and the overburden
frequency waves. Hence, there is lack of knowledge regarding
was 55 m.
the influence of discontinuities on low-frequency wave prop-
• stiffness of discontinuity – three values of the normal and
agation in discontinuous rock masses (i.e. for the frequency
shear stiffness were studied; 1, 10 and 100 GPa/m. The joint
range associated with train-induced vibrations).
was positioned in the middle of the wall and the overburden
This paper addresses how (i) overburden, (ii) position of
was 55 m.
a horizontal discontinuity in relation to the tunnel, and (iii)
the normal and shear stiffness of a discontinuity influence the
wave propagation in a rock mass near a tunnel. The model,
methodology, and input data are described as well as the results 3 MODELLING RESULTS AND DISCUSSION
and the conclusions of our findings.
Figure 1 shows the vertical vibrations levels at different posi-
tions on the ground surface for different the four overburdens.
2 MODEL SETUP AND ANALYSES The highest vibration levels occur above the tunnel (0 m). The
vibration levels decreases with increasing horizontal distance
Wave propagation analysis was conducted using the Univer- from the tunnel until they reach a minimum and then increases
sal Distinct Element Code (UDEC). Input data used in the to levels similar to those observed above the tunnel. It is evi-
analyses were (i) rock stresses, (ii) rock mass properties, (iii) dent that the overburden has a great impact on the vertical
mechanical properties of discontinuities, and (iv) dynamic vibration levels on the ground surface and that increasing
load. overburden will reduce the vibrations.
The virgin rock stresses were based on Swedish conditions, The distance from the centre of the tunnel to the point
where the vertical stress was assumed to be the product of the where the minimum occurs on the ground surface increases
482
4 CONCLUSIONS
483
G.W. Ma
School of Civil and Resource Engineering,
The University of Western Australia, Crawley, WA, Australia
Keywords: Dynamics Rock dynamics; Stress wave propagation; Rock mass; viscoelastic behavior
A rock mass is a material with various discontinuities different the dynamic complex viscoelastic modulus (Figure 2) of the
from other traditional structural materials in civil engineering. sedimentary rock was also derived. Measured stress waves
Ideally, a rock mass is composed of a system of rock blocks show that micro-defected rock is viscoelastic.
separated by discontinuities forming a material in which all The effect of micro-defect on the viscoelastic response
elements behave in mutual dependence as a unit. There- of rock is further investigated using numerical manifold
fore, the dynamic properties of rock mass are characterized method (NMM) (Figure 3). Complex viscoelastic modulus of
by mechanical and geometrical properties of discontinuities.
Investigating the mechanical effect of rock blocks and discon-
tinuities on the stress wave propagation is important in solv-
ing problems of seismology, mining, tunneling, underground
constructions, etc.
An investigation of the effects of discontinuities on stress
wave propagation through rock mass is carried out. The dis-
continuities in the rock mass are divided into two groups. The
primary discontinuity set is the one with relatively large or
the same order spacing to the wavelength, known as “macro-
joint”, while the secondary discontinuity set, characterized by
high density and relatively small spacing to the wavelength,
known as “micro-defect”.
The rock with micro-defects is modeled as an equivalent
viscoelastic medium. The effect of micro-defects on the wave
propagation is evaluated by an impact test on defected sedi-
mentary rock using a pendulum (Figure 1). The sedimentary
rock bar was cored from an underground cavern construction
site. The strain pulses were measured at the middle of the bar Figure 2. Complex viscoelastic modulus of sedimentary rock.
(Figure 1). Wave propagation coefficients (attenuation coef-
ficient and wave number) and phase velocity were obtained
from the frequency spectra of the recorded strain pulses and
Figure 1. Scheme for impact tests. Figure 3. NMM model for simulation of micro-defected rock.
484
485
J.T. Zhang
School of Mechanics and Civil Engineering, China University of Mining and Technology, Beijing, China
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: rock bursts; lab testing; bolts and anchors; rock support
1 INTRODUCTION
486
include the loading/unloading device, high speed data acquisi- (LDCR), can accommodate the large deformations of the rock
tion system and AE detecting system. The loading/unloading masses and absorb the impact by the sudden release of the
device of the DURATM is composed of the main stand; During rockburst energy while output a constant resistance subjected
the test, one surface of the specimen can be unloaded immedi- to the external loading. Therefore, LDCR can prevent a large-
ately from the true triaxial compression condition, simulating scale deformed rock mass from breaking and failing, and suit
the stress condition for rock mass at the free excavation bound- for preventing and mitigating the engineering disasters in deep
ary in underground excavations. (2) The distribution law of the ground such as coal bumps and rockbursts.
dominant frequency bandwidth (DFB) for different rock spec-
imens under varied stress paths during rockburst simulation
tests was obtained based on the time-frequency analysis and ACKNOWLEDGEMENTS
the discrete Fourier transformation (DFT) from the AE wave-
form data. The rock specimens with different lithology under This work was supported by the National Basic Research
rockburst tests have their unique frequency features of the AE Program of China (2006CB202200), the Major Program of
signals, including the DFB distribution in the LF, LMF, MHF, National Natural Science Foundation of China (50490270),
and HF bandwidths. (3) Laboratory experiments and field and Innovative Team Development Project of the Ministry of
engineering applications demonstrate that the novel anchor, Education of China (IRT0656).
i.e. the Large Deformation Cable with Constant Resistance
487
S. Li
Research Center of Geotechnical and Structural Engineering, Shandong University, Jinan, China
Keywords: dynamics, numerical modelling, rock joints, stability analysis, rock mass, rock slopes and foundations
488
Application of discrete element method for validity evaluation of rock dynamic fracture
toughness measured by semi-circular bend technique
1 INTRODUCTION the specimen surface friction influences the SCB test mea-
surements. The results of the RD model, labeled as RD (φ: 5),
This study investigates the influence of fracture constitutive suggest the necessity of introducing rate-dependency into the
law, loading rate and frictional boundary condition on the contact law, as the RD model can reproduce the test data.
rock dynamic fracture toughness, KICd , measured by the semi- Figure 3 presents the variation of predicated stress intensity
circular bend (SCB) method. Using UDEC, rock material is factor, KId , versus time and the observed fracture patterns.
modeled as an arbitrarily produced assemblage of elastic trian- Like the laboratory observations, fracturing begins from the
gular particles interacting at their edges. The size of particles is crack tip, then propagates straight, and eventually causes
taken as the average grain size of rock. The interface between the specimen to break into two pieces. However, the frac-
the particles is numerically represented as a contact surface ture starts propagating 55 µs after √ the loading bar touches
formed between two particle edges. the specimen, when√ K Id = 1.40 MPa m. In fact, the peak of
KId (=2.40 MPa m), which is experimentally reported as the
dynamic fracture toughness, does not correspond with the
2 CONTACT STRESS-DISPLACEMENT LAWS instant of fracture initiation. This is a point of concern as
it means that the dynamic fracture toughness of material is
The contact models developed are presented in Figure 1, overestimated by the test.
where σ = stress acting on contact surface, δ = contact dis-
placement, kt = contact initial stiffness, tc = contact strength,
δct = contact critical tensile displacement beyond which cohe-
sive softening happens, and δut = contact ultimate displace-
ment at which contact entirely loses its strength. Two models
are adopted: in the rate-independent (RI) model, δut is held
fixed, whilst in the rate-dependent (RD) model, it is assumed to
change with the contact opening rate through being multiplied
by a factor of α.
3 CALCULATION RESULTS
Figure 1. Contact behavior in RI and RD models. Figure 3. Stress intensity factor versus time.
489
T. Asakura
Department of Civil and Earth Resources Engineering, Kyoto University, Kyoto, Japan
490
Rockburst in the Upper Silesian Coal Basin (USCB) has rep- Destress rock blasting is designed according to natural and
resented a very dangerous phenomenon for a long period mining conditions. Boreholes with a diameter of 75–105 mm
of time (e.g. Staube 1972, Petuchov & Zamarski 1990, are usually drilled from the gateways, with a borehole incli-
Takla & Ptacek 1990, Holecko et al. 1999, Konecny 2005, nation of up to 30◦ and length of 30–120 m. The spacing of
Dubinski & Konopko 2000, Drewiecki & Kabiesz 2008). The parallel boreholes is usually from 5 to 12 m. Explosives are
rockburst problem relates primarily to the existence of compe- pneumatically charged into boreholes (in cartridges). The total
tent rocks in rock mass and the irregular stress fields induced amount of explosive charges ranges from 400 to 4781 kg. The
by long-term excavation. Destress rock blasting in overlying predominant charge percentage of the length of boreholes
competent rocks is a very important rockburst control tech- varies from 45 to 85% (increasing with an increase in hole
nique especially in the most complicated stress conditions. length).
The presented system of destress rock blasting as a modifica-
tion of the South African method (e.g. Comeau et al. 1999) is
a unique European system used in the Czech part of the USCB 5 MAIN GOALS AND THEIR EVALUATION
in difficult stress conditions in rock mass.
Properly designed and realized destress rock blasting (location
and spacing of boreholes, diameter of boreholes, length of
2 NATURAL AND MINING CONDITIONS charge, number of fired boreholes, total explosives charge,
etc.) reduces the strength and deformation properties of rocks
The USCB is a carboniferous multiseam sedimentary deposit. and releases a high stress concentration in the area surrounding
The sequence of competent sandstone beds and other much the blasting.
softer rocks and the network of faults form a very complicated Stress release is evaluated according to calculated seismic
set of natural conditions which results in an irregular natural effect (SE). The SE of destress rock blasting is the ratio of
stress state. seismic energy released in the rock mass when blasting to
Coal mining in the USCB started in the eighteenth century. the considered energy of the particular detonated charge. The
The irregular stress fields induced by long-term excavation are evaluation of stress release from the rock mass due to SE was
very complicated and in many cases they are very dangerous proposed by Knotek (1985) and verified by Konicek (2009).
for mining (increasing the rockburst risk).
6 CONCLUSIONS
3 CONCEPTUAL MODEL AND HISTORY OF
More than 30 years of experience with destress rock blasting
DESTRESS BLASTING
in hardcoal longwall mining in the Czech part of the USCB
have shown that it is a very useful method of rockburst control
A rockburst model has been formed since the second half of
in difficult stress conditions.
the twentieth century. The predominant types of rockbursts in
More than 2000 cases of destress rock blasting have been
the Czech part of USCB are rockbursts with a focal area in the
realized since the 1980s. Many of the cases which have been
main roof which originates from the high stress concentration
carried out show high SEs which document the required stress
in competent rocks. Destress rock blasting seems to be the
release in the rock mass. In the future, problems such as the
best measure against these. The competence of rock layers is
distribution of stress fields in the vicinity of the boreholes after
influenced by the application of this control technique.
blasting, the range of the resulting rock deformation, and more
The system of destress rock blasting has been developed
precise determination of the stress concentration area must
since the 1980s. The technique of destress rock blasting has
be solved. Nevertheless, destress rock blasting contributes to
improved (drilling, length of boreholes, diameter of boreholes,
safe longwall mining of hardcoal in conditions of substantial
weight of blasted charges, design, evaluation of effectiveness,
rockburst risk.
etc.). Since it began there have been more than 2000 destress
rock blasts and more than 3.5 million tons of explosives have
been used.
491
C.C. Li
Department of Geology and Mineral Resources Engineering,
The Norwegian University of Science and Technology (NTNU), Norway
Keywords: bolts and anchors, lab testing, rock support, rock bursts
492
1.5 m long stretch section was approximately 220 mm. The The test results of all the specimens have shown that the load
impact load for the 22 mm specimens was at a level of about transferred to the plate/thread of the D-bolt is only a small
280 kN. Figure 2 shows the test results of a Ø22 mm × 1.5 m portion of the impact load, indicating the anchor provides a
specimen. It was tested with an input energy of 56 kJ (i.e. a good protection to the thread/plate of the bolt.
mass of 2900 kg falling from a height of 1.97 m). The input In conclusion, the D-bolt fully utilizes the load and defor-
energy was just beyond the critical input energy (55 MPa) so mation capacities of the bolt steel material so that it is both
that it failed. Another specimen that was tested with an input strong and deformable. The high strength of the bolt provides
energy of 50 kJ did not fail. a high resistance to rock dilation so that the rock deformation
For conventional rock bolts using thread/plate, it is often is profoundly restrained. The high deformability of the bolt
observed in the field that the thread/plate fails when the bolt makes it possible that the bolt would not prematurely fail when
is heavily loaded. The thread is the weakest portion of the bolt subjected to large seismic bursts. A D-bolt can absorb up to
since the effective area of the cross section of the threaded 50–100 kJ of deformation energy, depending on the length of
portion is smaller than the solid bar. For the D-bolt, one of the the bolt. The energy absorption of the D-bolt is repeatable and
anchors is positioned close to the thread. It is seen in Figure 2 predictable since it is purely dependent upon the mechanical
that the plate load is very low compared to the impact load. properties of the bolt material, i.e. the steel.
493
G.W. Ma
School of Civil and Resource Engineering, University of Western Australia, Crawley, Australia
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
J. Zhao
Ecole Polytechnique Federale de Lausanne (EPFL), Laboratory for Rock Mechanics (LMR), Lausanne, Switzerland
Keywords: dynamic equivalent method, rock joints, rock mass, wave propagation, viscoelastic model
Rock mass usually consists of multiple, parallel planar joints, to be a completely elastic medium, which is effective only if
known as joint sets, which govern the mechanical behavior the frequency-dependence and the discreteness of joints, or
of the rock mass. The dynamic behavior and wave prop- multiple reflections among the joints are negligible.
agation across jointed rock mass are of great interest to The one-dimensional equivalent medium method proposed
geophysics, mining and underground constructions. It is also in the present paper is an extension of the study by Li et al.
significant to assess the stability and damage of rock struc- (J.C. Li, G.W. Ma and J. Zhao. An Equivalent Viscoelas-
tures under dynamic loads. Because of the discontinuity by tic Model for Rock Mass with Parallel Joints. Journal of
the joints, the dynamic response of jointed rock mass is a Geophysical Research, 115(B03), 2010). By analyzing the
complicated process. It is of significance to develop an effi- normal and shear equivalent property of a parallel jointed
cient and explicit method to analyze longitudinal- or shear- rock mass, the improved method is mainly to investigate
(P- or S-) wave normally across a rock mass with parallel P- and S-wave propagation across parallel jointed rock mass.
joints. For P- or S-wave, the method includes two portions: a vis-
Currently there are two common methods to analyze the coelastic medium model which is shown in Figure 1 and a
wave propagation in jointed rock masses; namely, the dis- concept of “virtual wave source (VWS)”. The viscoelastic
placement discontinuity method (DDM) and the effective medium model is to describe the normal and shear property
moduli method. Generally, the displacement discontinuity of the representative volume element (RVE) for a rock mass,
method (DDM) treats joints, particularly the dominant sets as shown in Figure 2(a). The VWS is proposed to take into
as discrete entities. The effective moduli methods predict the account the wave reflections between the joints. Figure 2(b)
aggregate effects of many joints or systems of joints within
a representatively elemental volume (REV), which creates a
continuum analysis of the problems. However, The conven-
tional effective moduli methods ignore the discreteness of
joints in rock masses and simplify the discontinuous rock mass
Figure 1. Equivalent mechanical model of an auxiliary spring in Figure 2. Scheme of jointed rock mass and equivalent medium
series with Voiget model. (k = p for incident P-waves and k = s for incident S-waves).
494
shows the one-dimensional equivalent medium for RVE in of transmitted waves obtained from displacement discontinu-
Figure 2(a). From the viewpoint of all, the normal or shear ity method (DDM) and equivalent medium method (EMM)
equivalent property of a rock mass is properly expressed from with different joint number when the joint spacing is one
the proposed method. Meanwhile, the derived wave propaga- tenth of the wave length. The results from the equivalent vis-
tion equation is explicit, which simplifies the analytical study coelastic medium method are very similar to those determined
of wave propagation. from the displacement discontinuity method. By analysis, it
The 1D equivalent medium method can be effectively is found that the P- or S-wave propagation normally across
applied to analyze P- or S-wave propagation normally an equivalent medium is related to the VWS number and the
across discontinuous rock mass with a set of parallel joints. ratio of the VWS spacing to an incident wave length. For a
Parameters in the viscoelastic medium model are derived given VWS number, if the wave length of the incident wave
analytically based on P- or S-wave propagation across a RVE. is sufficiently long compared to the joint spacing, the effect
The proposed method is then verified by applying identical of the VWS on wave propagation across rock mass is promi-
incident waves to the discontinuous rock mass and the cor- nent. Finally, frequency dependence and joint spacing effect
responding equivalent medium at one side to compare the on the wave propagation across parallel jointed rock mass are
output waves at the other side. Figure 3 shows the comparison discussed.
495
Wei-min Liang
School of Civil Engineering, Henan Polytechnic University, Jiaozuo, China
H.Y. Liu
School of Engineering, The University of Tasmania, Australia
Xiao-lin Yang
Luoyang Institute of Science and Technology, Luoyang, China
D.J. Williams
School of Civil Engineering, The University of Queensland, Australia
1 EXTENDED ABSTRACT that the gap between the explosive and the borehole wall in
decoupled charge blasting decreases the stress wave loading
Rock blasting has been the popular method for rock break- rate on the wall and extends the action time of stress wave and
age in mining and civil engineering. In this paper, laboratory gas, which correspondingly decreases the size of the crushed
experiments are first carried out in cement mortar samples zone formed, as shown in Figure 3, results in better fragmen-
with various air-decoupled charges to investigate effect of tation efficiency and improves explosive energy utilisation
decoupled charge blasting on rock fragmentation efficiency. ratio. Throughout this study, it is concluded that the rock frag-
The geometry of the specimen used in decoupled charge mentation efficiency can be greatly improved by adopting the
blasting is depicted in Figure 1. decoupled charge blasting with the optimal decoupled coeffi-
Two types of charge structures are adopted: 1) air-decoupled cients, which can be determined using the formulas proposed
charge with decoupled coefficients of 1.33, 1.60, 2.00 and in this study.
2.67, and 2) coupled charge. The resultant fragments from the
experiments were screened and statistically analysed using the
so-called Gates-Gandin-Schumann fragment model to obtain
fragment size distribution and then calculate the mean size
of fragments K50 and the boulder fragment ratio Bfr. The
obtained K50 and Bfr for the different tests with various decou-
pled coefficients k are plot in Figure 2. As it can be seen from
Figure 2, the fragmentation efficiency in decoupled charge
blasting can be improved if the decoupled coefficient is in
an optimal decoupling coefficient range (ODCR). After that,
analytical formulas are derived to predict ODCR in decouple
charge blasting from the lab to the field.
Numerical simulations are finally conducted to investigate
the mechanisms in decoupled charge blasting and it is found
Figure 2. K50 versus k and Bfr versus k for the various tests.
496
G.W. Ma
School of Civil and Resource Engineering, University of Western Australia, Crawley, Australia
497
498
Xia-Ting Feng
State Key Laboratory for Geomechanichs and Geotechnical Engineering, Wuhai, China
499
500
R. Resende
Concrete Dams Department, National Laboratory for Civil Engineering, Lisbon, Portugal,
currently at ISCTE-Lisbon University Institute, Portugal
R. Calçada
Faculty of Engineering of the Porto University, Porto, Portugal
501
502
Drilling and blasting are two necessary stages in mining engi- and β = charge constant and transmitting medium constant
neering and construction works.Although drilling and blasting respectively.
prove useful, have a few detrimental impact on structures Although these formulae are accepted in literature to predict
located at the neighboring township and residential spots. PPV from available data, in most cases they result in unaccept-
To name a few, these problems are fly rock, air blast, back able values compared to posterior measured ones i.e. in nearly
break and ground vibration. Amongst all problems originated all cases predict erroneous values. Also these formulae actu-
in blasting, because of stringent statutory regulations blasting ally are obtained from setting a correlation between the charge
induced ground vibration can lead into early mine closure. per delay value, the distance value and PPV value through sta-
When an explosive charge is detonated in a solid medium tistical analysis. So human intervention on statistical analysis
like rock, a family of waves is generated. These waves have results in a predictor which its general formula is arbitrarily
different particle motions and travel at different wave veloc- chosen and inevitably will predict different PPVs compared
ities through the medium (Sanchidrian, J. A. et al., 2007). to measured ones. However, nowadays computer based soft-
As the seismic waves travel through the rock, they excite wares prove viable and accurate computations on which PPV
particles to move. The motion of the ground particles (vibra- values can be obtained with maximum accuracy in prediction
tion) occurs in three directions which are vertical, radial and results.
transverse. Using a seismograph the particle velocity in three New studies on soft computing techniques offer good
directions are captured and maximum velocity of particles in approaches such as artificial neural networks (ANN), fuzzy
either direction is referred to peak particle velocity (PPV). regression and evolutionary programming. Some researchers
Several researchers have suggested that peak particle velocity proposed ANN to predict ground vibration components (PPV
(PPV) manifests the most appropriate and accurate indicator and frequency) to an acceptable level (Khandelwal & Singh
of possible damage to environment and structures. 2007).
During the evaluation and optimization process of a blast- In the present study the possibility of creating a model
ing operation, it is important to predict vibration levels (via through soft computing techniques and especially evolution-
PPV) at a specific point according to a given blasting design. ary programming is examined to find a predictor and its
The magnitude of PPV at a determined point varies with the constants to overcome aforementioned problems.
explosive charge amount that is detonated per delay (not total In GP, individuals are represented as trees. Elements of the
charge) and the distance of a specific point from the charge trees are functions and terminals. Terminals are the variables
center (Sanchidrian, J. A. et al., 2007). Certain empirical for- and the functions are operations applied to these variables
mulae were proposed by the different researchers to predict forming the model together. For example, Fig. 1 shows the
PPV using the charge per delay and the distance to shot center representation of the simple expression, x − y + x/5.
quantities (Duvall & Foghelson 1962, Langefors & Kihlstrom In applying genetic programming to a problem, there
1963, Ambreseys & Hendron 1968, Indian Standard Institute are five major preparatory steps. These five steps involve
1973). determining:
Table 1 lists four frequently used formulae called herein
1. predefined set of terminals,
empirical predictor equations.
2. predefined set of primitive functions,
where Q = charge weight (kg); R = distance between the
3. a fitness measure (called a “fitness function” to specify
charge center and the PPV measurement point (m); PPV =
what needs to be done)
peak particle velocity of ground vibration (mm/s); and K
503
504
T.N. Singh
Indian Institute of Technology, Mumbai
Drill-and-blast cycle is still the most economic and convenient of uniform size and well below the handling limit of mining
method of obtaining ore / mining / tunneling. Much advances machineries (Figure 3 and Figure 4).
related to the blasting technology has been studied which is In another case study, enhancement in the explosive energy
termed as ‘Blasting Engineering’. It is, however, an efficient utilization through application of SPARSH is clearly demon-
application of the technology makes a blasting as ‘Engineering strated. The mine management for a rock phosphate mine
Blasting’ which shall result into minimal deleterious impacts undercharged their blastholes due to proximity of habitats and
of blasting and maximum energy utilization from the blasting.
Further, little studies have done to understand impact of proper
stemming for obtaining ‘Engineering Blasting’.
We have invented a stemming plug named SPARSH
(A Stemming Plug Augmenting Resistance to Stemming in
Holes) which has many unique features and phenomenally
improves the blasting. This paper discusses initial results
of SPARSH and it is demonstrated that the application of
SPARSH results into NO FLY ROCK BLASTING (Figure 1
and Figure 2). The explosive energy utilization increases
by 30% and more with the use of SPARSH which results
into fewer requirements of boreholes drilling and lesser con-
sumption of explosive thus ground vibrations are lower in
magnitude than the conventional blasting. Another major
advantage of the application of the SPARSH is minimal
requirement of secondary blasting as the muck produced is
Figure 1. NO FLY ROCK BLAST with SPARSH in a limestone Figure 3. Fragmentation before SPARSH application in a
mine. manganese mine.
505
506
Rock support subjected to dynamic loading: Field testing of ground support using
simulated rockburst
L. Malmgren
Mining Technology R&D, LKAB Kiruna Mine, Kiruna, Sweden
Increasing the mining depth at LKAB Kiirunavaara mine support systems and the interaction with the rock mass can
located in the northern part of Sweden is leading to higher be investigated. A number of studies using explosives to gen-
stress magnitudes, resulting in increased seismic activity and erate the dynamic load have been carried out by for example
more rockburst damage. The effectiveness of various ground Milev et al. (2001), Espley et al (2004), and Heal & Potvin
support systems under dynamic loading conditions has there- (2007). A total number of twelve tests are planned to be car-
fore become of prime interest to LKAB for successful and ried out. Tests 1 and 2, called Zero tests-1 and 2, were planned
safe mining at deep levels. Therefore series of rockburst sim- in order to design the charge density, the instrumentation and
ulations will be conducted using explosives to generate the the influence of blast gases. This paper covers the set up and
dynamic load, on a number of support systems. In large-scale results from the first trial test called Zero test-1. In this test, the
dynamic tests of ground support the performance of ground explosive used to generate the dynamic load was of military
type, NSP711 with a high VoD compare to other commer-
cial explosives. The blasthole was approximately 15 m long
with the diameter of 115 mm. A nominal burden of 3 m was
selected between the blast hole and the test panel (see Fig-
ure 1). Two different charge diameters (76 mm and 45 mm)
were used in the blasthole to simulate different magnitudes
of the seismic event. The LKAB Standard support system for
seismically active mining areas was installed in the test drift.
The field monitoring included ground motion measurement
with a number of accelerometers, fracture investigation with
observation holes, ground and support motion imaging with
high speed camera, as well as the deformation measurements
(see Figure 2). Results from accelerometers and gas pressure
gauges indicate that the wall was loaded by a shockwave first
and then a gas pressure, which proves that the methodology
507
508
Subject: Rock material and rock mass property testing (laboratory and in situ)
Uniaxial and triaxial compression tests have been performed by the Office of Atomic Energy for Peace. This study is part
to assess the influence of loading rate on the compressive of a research effort on determination of the safe maximum
strength and deformability of the Maha Sarakham salt. The withdrawal rate of compressed air stored in salt caverns. The
tested specimens were prepared from 60 mm salt cores drilled salt specimens are prepared to have a nominal dimension of
from depths ranging between 140 m and 250 m by Pimai Salt 5.4 × 5.4 × 5.4 cm3 and compressed to failure using a polyax-
Co. in the northeast of Thailand. The salt cores belong to the ial load frame. The lateral confining pressures are maintained
Middle Salt member of the Maha Sarakham formation. This constant at 0, 3, 7, 12, 20 and 28 MPa while the axial stresses
salt member hosts several solution-mined caverns in the basin. are increased at constant rates of 0.001, 0.01, 0.1, 1.0 and
The Maha Sarakham formation is also being considered as a 10 MPa/s until failure occurs.
host rock for compressed-air energy storage caverns by the Table 1 shows the test results. The measured stress-strain
Thai Department of Energy, and for chemical waste disposal curves indicate that the salt behaves as a nonlinear material
Table 1. Salt strengths and elastic properties under various loading rates and confining pressures.
509
510
Ö. Uysal
Mining Engineering Department, Dumlupinar University, Kutahya, Turkey
E. Arpaz
Vocational school of Kocaeli, Kocaeli University, Kocaeli, Turkey
M. Cavus
Seyitomer Coal Enterprise, Kutahya, Turkey
S. Beyhan
Mining Engineering Department, Dumlupinar University, Kutahya, Turkey
Y. Tola
Vocational school of Kocaeli, Kocaeli University, Kocaeli, Turkey
S. Yuvka
Mining Engineering Department, Dumlupinar University, Kutahya, Turkey
In the case of blast induced vibrations it is very common that 10 holes were drilled at 2.5 meters distances and 10 meters of
mine managers come across with complaints from local res- length. The distance between the barrier holes and the blast
idents. This may lead to decrease or total cessation in the holes is 360 meters. This was intended to provide a discontinu-
production. Therefore, there are a lot of studies that focus on ity that comprises of holes, which was expected to eliminate or
eliminating the effects of blast induced vibrations. The first reduce the vibrations. The holes were 225 mm in diameter and
method that comes to mind for the elimination of blast induced 10 meters in length. For the recordings two different devices
vibrations is to reduce the amount of explosive. However, were used (Instantel mini mate plus, Instantel Blasmate III).
decrease in the amount of the explosive leads to a decrease Within the scope of the study, 395 vibrations were recorded
in the pre-planned production. Studies apart from this focus from 150 instances of blastings. Recorders were placed in front
on pre-split, barrier holes, and trench methods. While some of and 25, 50, and 75 meters at the back of the discontinuity in
of these researchers claim that blast induced vibrations are panel 1 and panel 2. For panel 3, the recorders were placed in
largely eliminated by the help of these methods, some others front of the discontinuity line and 5, 10, 15, and 25 meters at
claim that they have no positive influence on the issue. It can the back of it. Thus, the changes in the front line and backline
be said that there does not exist a common agreement on the of discontinuity were recorded according to distance.
effects of these methods on the elimination of blast induced Findings were evaluated based on the location of recorders
vibrations. as to the discontinuity lines. The values in panel 1 seem to
The primary purpose of this study is to eliminate the adverse have been unevenly distributed due to the fact that they have
effects of blast-induced vibrations at Seyitomer Coal Enter- not been subjected to any categorization. When categorized,
prise by forming artificial discontinuities. In order to do this, it can be seen that the values recorded in the front of the
an artificial discontinuity was formed in the area between discontinuity have higher particle velocity compared to the
blasting area and measurement point in the dragline panel. The ones taken from the back of the discontinuity.
panel height was 24 m. The rock type existing over the coal in In the light of the evaluation of the values in panel 2, it can
the panel was marl. ANFO was used for rock fragmentation. be seen that the data gathers around a narrow line when they
The blast holes were 24 m long, with 228 mm diameter. The are not grouped. After the grouping, it can be seen that the
distance between the blasting holes was 10 m. ANFO, power values taken from the front of the discontinuity are higher
gel dynamite and detonator were used for the blasting. The than the other values and the values taken from behind the
delay between each hole was 42 ms and in-hole delay, 25 ms. discontinuity have slower particle velocity.
The area has been separated into three sections comprising of In the light of the evaluation of the data from panel 3, the
panel 1, panel 2, and panel 3. At the end points of each panel data is distributed unevenly when they are not grouped. If they
511
512
J.J. Liao
Department of Civil Engineering, National Chiao-Tung, Hsin-Chu
1 INTRODUCTION hence, Cij (i, j = 1–6) can be presented in terms of the nine
independent elastic constants as:
Wave propagation in anisotropic materials is of great inter-
est in geophysics and other branches of applied sciences
such as rock mechanics/rock engineering, earthquake engi-
neering, and petroleum engineering, etc. It is known in the
present day that for most rocks, such as foliated metamor-
phic rocks, stratified sedimentary rocks, and regularly jointed
rocks, their properties might exhibit both anisotropic and
inhomogeneous characteristics. The mechanical responses of
anisotropic media with spatial gradients in composition are
called anisotropic functionally graded materials (FGMs), and
their characteristics are widely employed in various fields of
applied mechanics. Hence, in this study, the solutions of body-
wave velocities for continuously inhomogeneous orthotropic
rocks with Young’s moduli (Ex , Ey , Ez ), shear modulus (Gxy ,
Gyz , Gxz ), and the medium density (ρ) varying exponentially
with depth are derived.The governing equations for the present
solutions can be combined with the generalized Hooke’s
law, the strain-displacement relationships, and the equilibrium
equations. Then, supposing the direction of wave propagation
is on the x-z plane, the solutions of three quasi-wave veloci-
ties can be yielded. The proposed solutions are identical with
those of Tsvankin (1997) when the inhomogeneity parameter Based on the inhomogeneity parameter, α, the following
(α) is equal to 0. situations exist:
• α > 0 indicates a hardened surface; all Young’s and shear
moduli decrease as the depth increases.
2 THE PHASE VELOCITY SOLUTIONS
• α = 0 is the conventional homogeneous condition.
• α < 0 indicates a soft surface; all Young’s and shear moduli
The expression of the stress-strain relationship for a continu-
increase as the depth increases.
ously inhomogeneous orthotropic medium is:
Then, the strain-displacement relationship for a small strain
in a Cartesian co-ordinate system, and the differential forms of
equations of motion can be presented in terms of displacement
vector components, ux , uy , uz as:
513
Figure 1. Definitions of phase angle and group angle. Eventually, the three body waves, VSH , VSV , and VP , are
derived for an inhomogeneous orthotropic medium as:
where:
514
ABSTRACT: In mining, blasting, earthquakes, rock bursts by mechanical drilling, cutting and grinding. For impact flat-
and defence engineering involving penetration of missiles, tened Brazilian disc specimens, two parallel flat ends which
rock failure pattern, strength and breakage are often affected central angle is 25◦ are introduced to the disc for load bearing.
by loading rate. Therefore, it is of great importance to inves- This loading manner is superior to the original line loading of
tigate the effect of the loading rate on rock failure parameter the Brazilian test in avoiding local cracking, breakage or yield-
and fragmentation. It is a well-known fact that rock is an brit- ing around the loading point caused by stress concentration.
tle porous medium compositing of mineral grain, microcrack A SHPB apparatus which is 74 mm diameter was used to test
and pore space, this is mainly due to difficulty in theoreti- dynamic mechanical properties of granite. Brass pulse shapers
cal analysis and experimental study on rock at high loading with 12 mm diameter and 2 mm thickness were designed to
rates. So it is active demand on systematically dynamic testing improve incident wave forms, so that granite specimen has
of granite. The split Hopkinson pressure bar technique orig- enough time to obtain stress equilibrium, at same time.
inally were developed by Kolsky (1949), as one of method Besides, brass pulse shapers can smooth wave forms and
for testing dynamic mechanic character, the Split Hopkinson eliminate wave form oscillations. For test of impacting flat-
Pressure Bar (SHPB) technique which based on assumption of tened Brazilian disc, incident bar and transmission bar close
one-dimensional stress wave propagation and the assumption contact with specimen on two parallel flat ends, the specimen
of stress uniformity along the specimen thickness, has been is subjected to diametral impact. Strain gauges are mounted at
widely used in studies on the dynamic mechanical behav- the center of two sides of the disc respectively, and the average
ior of rock materials at high strain rates. Shan (2000), Liu of the two measured values is taken for analysis.
(2000), Li (2005) and Zhai (2006) have reported the result Before impact testing, the quasi-static compressive tests of
of granite under dynamic compression. The tensile strength granite specimens are performed with MTS810 test machine.
of rock material is typically an order of magnitude smaller Five kinds strain rate are selected in impact compression, and
than the compressive strength, tensile failure often occurs in four specimens at least are used in each kind strain rate impact
rock masses. However, a direct tension test of rock is very compression. The experiment results showed that the dynamic
difficult to perform, so that an indirect tension, such as the strength raised quickly and specimens damage significantly
Brazilian test, also called the split tensile test, has been a pop- contest with static compression. The stress increases linearly
ular choice. This splitting test method has been widely used initially and then decreases after the stress achieves the peak
in static indirect tension for rocks, concrete and many other value. The strain rate effect on the peak stress and strain can
materials. WANG and WU (2004) proposed introducing two be seen clearly. That is to say, granite is strain rate depen-
parallel flat ends to the disc specimen for loading, in this way a dent materials, the dynamic increase factor (DIFC) is directly
BD becomes the Flattened Brazilian Disc (FBD). The FBD is depend on the strain rate, and increase approximately lin-
favorable for reducing stress concentration effect, thus ensur- early with the increase of average strain rate. The compression
ing crack initiation from the center of the disc instead of from experiments show that the failure of the granite specimen was
the loading point. typical axial cleaving failure mode under the direction of uni-
However, as we know, few systematically dynamic testing axial compressive loading at different strain rates. Besides,
results about granite at relatively high strain rates in compres- it was concluded not only the strength of granite increased,
sion and tension were ever reported, despite it is very important but also the fragment size decreased and fragment numbers
for design of underground structure. In this study, the impact increased with the increasing strain rate, the specimens will
compressive and impact flattened Brazilian behaviors of gran- be crushed under higher strain rate.
ite at different strain rates are performed by split Hopkinson The impact splitting tension experiment results showed that
pressure (SHPB) techniques, the dynamic strength, failure the dynamic tension strength, peak strain will raise quickly
strain, failure pattern and other parameter at strain rates of with strain rate increment, the ratio of tension strength of
2.3 s∼108.4 s are obtained from impact test. dynamic and static is about 1.82∼2.82, the tension strength
The reddish-brown granite taken from northwest of China increase approximately linearly with the increase of average
were tested; its Poisson’s ratio is 0.24, Young’s modulus is strain rate. It shows that granite specimen will split into two
50 GPa, and density is 2650 kg/m3 . The cylindrical samples almost complete halves along the loading diameter after test
with 60 mm in diameter and 25 mm in thickness are obtained when specimen was subjected to a relatively low strain rate.
515
X.G. Wang
China University of Mining and Technology, Beijing, China
Jizhong Energy Group Company Limited, China
1 INTRODUCTION Considering the factors of frozen and due to the relevant exper-
iment, the empirical formula concerning the calculation of the
The vertical shaft in Xuandong is about 832.0 m in depth and Protodikonov’s hardness coefficient is made as follows:
6.0 m in diameter and its frozen foundation rock is 36.5 m
high.
The method of staggered freezing is adopted in order to
freeze to the rock of shaft. In the method of staggered freezing,
the long tube is 85 m and the short tube is 70 m, and the single Here, f is frozen Protodikonov’s hardness coefficient and f1
hole and the double hole are arranged respectively. There are is Protodikonov’s hardness coefficient before freezing. T1
26 frozen holes altogether, placed around the circle diameter is absolute value of the frozen stone temperature and T0 is
of 9.8 m. absolute value of the freezing point temperature for that stone.
The modification of frozen Protodikonov’s hardness coef-
ficient is as fallows: andesites: f = 6, orthophyre: f = 8,
2 GEOLOGY AND HYDROGEOLOGY dolerite: f = 7.
The blasting parameters are designed basing on the radius
In the 41.5 m depth of the main shaft, there are the andesites of 3800 mm of the net section. When the compound cut is
with growing cranny in length of 27.2 m, covered with the executed, depth of the burn cut is L1 = 3.2 m, with the diam-
watermark of iron rust in the joints. It also includes the amphi- eter of arranging for burn cut D1 = 1.4 m. The number of
bole and calcite which fills the cranny. Down to the 68.7 m holes arranged is N1 = 6. With the distances between holes
of the shaft, a yellow orthophyre of 4.8 m appears in thick is E1 = 733 mm. Every hole packs 6 volumes of water-gel
layers structuring in cryptocrystalline. The phenocryst is com- explosives. Depth of the second cut hole is L2 = 3.2 m. With
posed of the orthoclase, filled the cranny with calcite. In the the diameter of arranging is D2 = 2.20 m. The line of least
depth of 73.5 m, there is diabase composed by the pyroxene, resistance is W1 = 400 mm. The number of holes arranged is
anorthose and chlorite. The thick-layered celadon diabase is N2 = 9, and the distance between holes is E2 = 768 mm and
4.5 m in length. Water inflow of the bedrock before freezing is 5 volumes of water-gel explosive are packed every hole.
207 m3 /h, while no inflow water is there after freezing except Depth of the first, second and third satellite hole is
the “dead water” in the shaft. After frozen, the shaft walls are L3 –L5 = 3.0 m.With arranging diameters are D3 = 3.60 m,
also in good shape. D4 = 5.20 m, D5 = 6.60 m respectively. The line of least resis-
According to the “Geological Survey Report on Shaft tance are W2 = 700 mm, W3 = 800 mm, W4 = 700 mm. The
Review of Xuandong Shaft”, uniaxial compressive strengths number of arranged holes are N3 = 12, N4 = 15, N5 = 24.The
of the andesites, orthophyre and dolerite are 52.1 MPa, distances between holes are E3 = 942 mm, E4 = 1089 mm,
70.1 MPa and 65.9 MPa respective, while the rock E5 = 864 mm respectively. The first and second satellite holes
Protodikonov’s hardness coefficients are 5.2, 7.1, and 6.6 are packed 4 volumes of water-gel explosives and 7 volumes
respectively. of ammonium nitrate explosives for the third satellite holes.
Depth of the periphery holes is L6 = 3.0 m, with arranging
diameter of D6 = 7.40 m. The tilt angle is 81◦ 10 . The line
of least resistance is W5 = 400 mm. The number of arranged
3 CONSTRUCTION METHOD IN FROZEN ROCK
holes is N6 = 45.The distance is E3 = 942 mm and 6 volumes
of water-gel for every hole. Since the distance between the
3.1 Analysis of blast-ability
periphery holes is larger than the line of least resistance, so that
It shows some increasing in rock compressive strength and the empty holes arrayed in the proper places. In the whole sec-
Protodikonov’s hardness coefficient when the rocks are frozen. tion periphery, the number of holes is N = 111 and 156.87 kg
516
4 CONCLUSIONS
517
Over the past several hundred years, CO2 emissions in the A new developed flow pump permeability test was set up
atmosphere has increased steadily and risen over the pre- in a laboratory system where pressure and temperature can
industrialized level due to intensified human activity produc- be controlled. The entire laboratory system was isolated from
ing greenhouse gas emissions such as burning fossil fuels for external temperature change. All the equipments were placed
electricity generations, industries, transportations and domes- in a greenhouse chamber controlled with air conditioning.
tic uses. The increase of atmospheric CO2 emission has In case of internal temperature effect induced by laboratory
become a major contributing factor of a gradual raise of measurement heat, the pressure vessels and pipes includ-
the earth’s temperature, popularly known as global warming. ing syringe pumps were double insulated and connected to
Recently, the methods of reducing CO2 emission have been hemathermal circulation tanks and controller, so theirs tem-
developed including carbon capture and geological storage perature changes can be minimized as low as possible. The
(CCS). So far, CCS in depleted oil and gas reservoirs is the rock specimen used in this study was isotropic Ainoura sand-
most readily applicable since the technique is similar to the stone obtained from Nagasaki Prefecture Japan. The porosity
enhanced oil recovery (EOR) commonly applied in petroleum and specific gravity of the sandstone is 12.6% and 2.88 respec-
industries (IPCC, 2005). Even though oil and gas reservoirs tively. At the initial condition of 35◦ C temperature, 20 MPa
remain worldwidely available, these reservoirs are inequitably confining pressure and 10 MPa hydraulic pressure, supercrit-
distributed around the world. Therefore, deep saline aquifers ical CO2 was injected to the rock specimen saturated with
are recently investigated since their estimated storage capac- water. The upstream and downstream pressures were mea-
ities is far largest compared to the other formations. sured. To investigate the effect of CO2 state on the behavior of
A number of researchers have conducted the study of CO2 CO2 injected to the specimen, we also conducted the injection
storage in deep saline aquifers. In a core scale, Perrin and of CO2 in liquid state in the same initial condition, except the
Benson (2009) undertook the injection of CO2 and brine temperature of 20◦ C.
to sandstone obtained from the CO2CRC-Otway Project in The experimental results of supercritical CO2 injection
Australia. By employing a CT-scan, they observed the migra- show that the hydraulic pressure increases gradually. In com-
tion of CO2 along the sandstone pores and estimated CO2 parison to the liquid CO2 injection, the magnitude of the
saturation that changed during the injection. They found that hydraulic pressures generated by the supercritical CO2 injec-
heterogeneity of sandstone porosity has significant impact tion is lower than that generated by the liquid CO2 injection.
on the efficiency of CO2 migration. Shi, et al., (2010) per- However, the supercritical CO2 injection can yield a higher
formed CO2 injection to Tako Sandstone saturated with brine. differential pressure across the specimen than that by liquid
They also utilized a CT-scan to map the saturation profiles CO2 injection.
during the injection. Their findings show the heterogeneity Based on the pattern of differential pressure obtained from
effect of porosity is insignificant for the case of high flow rate the experiment, it can be suggested that there are four stages
employed in injecting CO2 . However, their studies do not pro- of conditions occurred during the injection of supercritical
vide adequate information about the relation of the increase CO2 into the specimen. At the pre-injection condition, the
of hydraulic pressure and the rise of CO2 saturation driven by water occupies most of the specimen pores and no CO2 exists
very low hydraulic gradient as representation of laminar flow (initial stage). After the injection started, CO2 begin to flow
in deep underground. and penetrates the bottom of the specimen (Stage 1). As the
In this paper, we present a new developed flow pump perme- injection continuing, more fractions of CO2 penetrate and
ability test to investigate the behavior of CO2 in low permeable occupy the specimen pores while more saturated water dis-
rocks. The method was originally introduced by Olsen et al., placed out (Stage 2). In this period, supercritical CO2 has
(1985), which has been able to determine permeability char- been able to penetrate the top of the specimen indicated by a
acteristics of low permeable. However, several developments drop of hydraulic pressure measured in the downstream gauge.
are needed in order to enhance the method’s ability of creating Finally, most of CO2 occupies the specimen pores and just a
similar physical conditions of geological site for storing CO2 , few amounts of water trapped (Stage 3).This stage is the period
expected to located in the depth of 800–1200 meters, where of the differential pressure declines and stabilizes to a steady
CO2 performs as supercritical fluid. Moreover, the develop- state. It seems to take a long time to accomplish due to very
ment of the method is conducted so as to work in very low low flow rate of the injection employed in the experiment.
hydraulic gradient of CO2 flow, due to groundwater flow in
deep underground is generally laminar.
521
D. Blankenship
Sandia National Laboratories, Albuquerque, NM, USA
Keywords: Geothermal, oil reservoir, physical modeling, numerical modeling, fluid flow
Coupled processes in porous media impact drilling and bore- evolution of the potential energy is obtained by combining the
hole stability. The coupling of the matrix deformation, and localized internal energy and entropy balance equations which
fluid and heat diffusions results in a time dependent response satisfy the first and second law of thermodynamics, respec-
in the formation. During the drilling of an oil or gas well, tively. The rock constitutive equations describe the response
temperature gradients between the drilling mud and the for- of the rock to mechanical and thermal loading. They express
mation can significantly change the pore pressure and stress the temporal evolutions of the total stresses, the variation of
distributions around the borehole. A general theory of ther- the fluid content, and the rock entropy in terms of the above
moporoelasticity is developed that fully couples the three pro- five independent variables. The transport relations that couple
cesses: mechanical, hydraulic, and thermal in porous media the influx and efflux of fluid and heat to their driving forces
saturated by a fluid. The theory satisfies the first and second are derived from the definition of the internal entropy produc-
laws of thermodynamics and is based on concepts of irre- tion rate and its associated generalized forces. The transport
versible thermodynamics, a novel rock constitutive relation, fluxes of fluid and heat are established through Onsager’s phe-
and Onsager’s transport phenomenology. Fully coupled field nomenology and concepts of irreversible thermodynamics.
equations are developed in terms of five independent vari- The temperature is locally equilibrated between the fluid and
ables, namely the three solid displacements, the pore pressure, the solid.A finite element model is developed for the fully cou-
and the absolute temperature. These equations are obtained pled processes consisting of: thermoporoelastic deformation,
by implementing the rock constitutive and transport equations hydraulic conduction, thermal osmosis, heat conduction, pres-
into the momentum, fluid mass, and energy balance equations. sure thermal effect, and the interconvertibility of mechanical
The rock constitutive equations are derived from a state func- and thermal energy. The model is used to analyze the problem
tion that describes the manner in which the potential energy of of a wellbore subjected to a hydrostatic in situ stress field.
the porous media changes with respect to time. The temporal Distributions of total stresses, pore pressure, and temperature
Figure 1. Induced pore pressure distribution around the borehole Figure 2. pressure distribution around the borehole for the cooling
for the heating case due to thermal loading at times of 0.05 (blue), case due to thermal loading at times of 0.05 (blue), 0.1 (green), 0.5
0.1 (green), 0.5 (red), and 1 (cyan) day. (red), and 1 (cyan) day.
522
523
Keywords: geothermal, numerical modeling, hydraulic fractures, fluid flow, case studies
1 INTRODUCTION
2 THE NUMERICAL MODEL The shapes of the curves are slightly different. Three of
the six fault zones in which most of the flowrate is injected
To simulate the behavior of the fractured rock mass in the (3 to 5 l/s), are roughly parallel to the maximum horizontal
vicinity of the wells during their hydraulic stimulations, a 3D stress. But, among the three others, one is oblique (at around
Distinct Element Method code was used. 30◦ ) and the two others are sub-perpendicular to the maximum
The numerical model is a 400 m × 400 m × 1000 m paral- horizontal stress. The three fault zones which react first to the
lelepiped volume centered on the stimulated open hole and stimulation, between 6 and 10 MPa, are the fault zones sub-
in which a network of nine fault zones has been defined from parallel to the maximum horizontal stress. For overpressures
various well data. The fault zones are roughly sub-vertical and greater than 10 MPa, the permeability increase is higher in two
four of them are sub parallel to the maximum horizontal stress. other fault zones whose one is sub perpendicular to the max-
The distribution of the initial fluid pressure in the frac- imum horizontal stress. The maximum shear displacement is
tured network is assumed to behave like a hydrostatic field. around 2 cm. The analysis of the evolution of the shear and
Zero displacements are imposed at the North (x = z = 0), West normal displacements shows that the stimulation cannot be
(x = z = 0) and bottom faces (y = 0) and stresses are applied considered in each fault zone independently of the other. The
on the South, East and top faces. combination of shearing in some fault zones can lead to irre-
All fault zones have the same mechanical constitutive law. versible opening of another fault zone which intersects them
The normal mechanical behavior is elastic linear, while the and not necessary sub-parallel to the maximum horizontal
fault zone is in compression, and tensile strength is null. The stress.
tangential mechanical behavior is elasto-plastic. It follows
a Mohr Coulomb failure criterion with associated dilation 3.2 GPK4 Well-shutdown modeling
effects. The effects of dilation appear as soon as the maxi-
mum shear strength is reached. The normal displacement and After the application of several stages of increasing pressure
the hydraulic apertures are directly related on the basis of an in the well, the pressure is shut down by returning to a zero
initial hydraulic aperture. flowrate at the wellhead.
At the end of the stimulation test, only a specific part of
the fault zone network shows a significant irreversible perme-
3 RESULTS
ability increase (Fig. 2), but the locations of these areas do
not systematically permit a connectivity increase. Above the
3.1 GPK4 Hydraulic stimulation test
potential local permeability increase which depends on the
A plot of the overpressure in the open hole versus the flowrate fault zone orientation regarding to the maximum horizontal
calculated in each fault zone is shown in Figure 1. stress, the whole network must be considered.
524
Figure 3. Calculated pressure at wellhead vs. flowrate in some fault Figure 4. Pressure vs. total flowrate at wellhead during stimu-
zones during stimulation test and post-stimulation injection. lation test and post-stimulation injection test (model and in-situ
experimental curves of the site).
525
Z. Hou
Institute of Petroleum Engineering, Clausthal University of Technology, Clausthal-Zellerfeld, Germany
Xuan Luo
Energy Research Centre of Lower Saxony, Goslar, Germany
1 INTRODUCTION
526
527
Z. Hou
Institute of Petroleum Engineering, Clausthal University of Technology, Clausthal-Zellerfeld, Germany
L. Zhou
Energy Research Centre of Lower Saxony, Goslar, Germany
1 INTRODUCTION
528
3 CONCLUSION
529
S. Nakashima
Department of Civil and Environmental Engineering, Yamaguchi University, Ube, Japan
H. Yasuhara
Department of Civil and Environmental Engineering, Ehime University, Matsuyama, Japan
530
P.C. Li
College of Mechanical Engineering, Shanghai University of Engineering Science, Shanghai, China
D.T. Lu
Department of Modern Mechanics, University of Science and Technology of China, Hefei, China
531
V. Maury
IFP School Rueil-Malmaison and Université de Montpellier, France (formerly with Total)
J.-M. Piau
IFSTTAR, formerly Laboratoire Central des Ponts et Chaussées, France
D. Fitzenz
Universidade de Evora, Centro de Geofisica, Portugal
Keywords: fluid flow, gas flow, physical modeling, risks and hazards, rock failure, stability analysis
EXTENDED ABSTRACT while the fault core is contractant during interseismic periods.
This dilatancy in undrained regime is prone to pore pressure
The role of fluids on earthquake initiation is now fully rec- fall-off and degassing of the fluid of these surrounding zones
ognized, but the effect of gas is still disputed. In particular, and of the much less thick fault core. In addition, the fault
contrary to what was hoped for in the seventies from Chi- core can be isolated or in relationship with an external source
nese observations, we now know that the observation of gas of over- or under-pressurized fluid, also a possible cause of
leakage around faults before a large earthquake is far from variation of the pore fluid pressure and its gas content. These
systematic and that using gas leakage as earthquake precur- mechanisms make the presence of gas a frequent issue in active
sor would lead to missed earthquakes as well as false alarms. fault settings, as confirmed by numerous observations.
Modeling efforts addressing explicitly both the mechanical Based on known properties of usual rocks and porome-
role of compressible fluids in faulting, but also the role of chanic consideration (Part II), we focus on the fact that the
fluid-assisted interseismic processes, and the changes in the presence of a tiny percentage of gas makes the Kempton’s
hydraulic structure of the fault through time (e.g., Fitzenz and coefficient decrease to zero. Thus porous rocks exhibit quasi
Miller, 2003) are still scarce. Here we want to go further in this drained geomechanical characteristics when saturated with a
direction and investigate in details the effects of the presence fluid including a tiny gas content (few per thousands), even if
of gas in the pore fluid, both on the stress paths and on the they are in undrained conditions (no exchange of fluid with
stability of faults. outside of the rock volume).
The paper builds upon the recent detailed descriptive work During interseismic periods (Part III), assuming the fault
on fault zone structure and deformation mechanisms to inves- core undrained, saturated with liquid and submitted to tec-
tigate several aspects and consequences of gas appearance tonic compressive (and shear) stress components, the pore
and/or disappearance on the geomechanical behaviour of pressure increases limiting the effective normal stress induced,
active faults. Some conclusions are drawn from an analyti- and hence the shear stress.
cal principle model aiming at describing and investigating the For faults in elastic behaviour, the presence of gas induces
behaviour of a fault within two bedrocks submitted to far-field greater normal effective stresses without inducing pore pres-
relative displacements. sure increase leading to greater possible shear stress and
The paper reviews some features of the fault zone struc- storage of elastic energy in surrounding and virgin zones.
ture and deformation mechanisms in Part I. The effects of If the fault is already active (in plastic contractant
gas on fluid compressibilities and geomechanical properties behaviour), the presence of gas in the fault core leads to a
of porous rocks are given Part II. The role of fluid and more more advanced compaction, with decrease of porosity and
particularly gas appearance and/or disappearance on the stress improvement of geomechanical properties. In addition the
paths acting on faults during interseismic period is presented permeability decreases. That again allows the active fault to
Part III. To evaluate the occurrence of instability and the effect sustain larger shear tectonic stress.
of gas, we propose an analytical model assuming a Cam-Clay If a communication between the fault core panel and an
type behaviour for the fault core, briefly described and dis- external over-pressurized fluid source is possible and fast
cussed in Part IV. Conclusions and perspectives are presented compared to the tectonic loading rate, the stress path is still dif-
in Part V. The considerations of this paper would also apply to ferent. The effective normal stress on the fault now decreases,
situations such as landslides, mine roofs, borehole stability. leading to conditions much more prone to rupture triggering
Seismological field observations (Part I) show that in some but at smaller shear stress values. Small overpressures are then
cases, zones surrounding the fault core may exhibit dilatancy sufficient to reach unstable conditions (Maury 1989, 1993).
532
533
A.N. Tutuncu
Colorado School of Mines, Golden, CO, US
A.I. Mese
Geomechanics Engineering & Research, Houston, TX, US
534
Seong-Ho Bae
Geogeny Consultants, Seoul, Korea
In general, either above-ground tanks or in-ground tanks been The phenomena involved in phase change during formation
used worldwide to store LNG. These types of tanks have a of an ice-ring can be divided into the followings: cryo-suction
number of problems such as the need for a large site for storage and swelling. Cryo-suction refers to suction pressure occurred
tanks, ground settlement, high operating costs, an excessive during phase change in the coexistence zone of ice and
boil-off ratio due to radiant heat from the sun, and so on. Many water. The zone where water and ice coexist is from 0◦ C
attempts have therefore been made in the past to store LNG isotherms up to the penetration zone of groundwater inside
underground in unlined containment, though without success. of the 0◦ C isotherms. In previous studies of Jeong et al.
The new system for storing LNG in rock caverns had been (2007), the temperature of rock mass frozen completely by
developed and verified through the design, construction and the groundwater infiltration was identified as groundwater
operation of a LNG pilot plant in Daejeon, Korea. Among freezing temperature. According to Lee et al. (2004), it was
the key technologies of the LNG storage system, drainage of known that the groundwater freezing temperature is in the
groundwater and formation of an ice-ring are very important range of −3◦ C to −2.5◦ C. However, it can be varied depending
ones. These are strongly related to the complex mechanism on the recovery rate of groundwater and average aperture of
between thermal and hydraulic characteristics of rock mass joints.
and groundwater. Another phenomenon that happens during freezing of water
Chung et al. (2005) performed a numerical calculation is swelling effect, which increases 9% in volume after phase
for the thermo-hydraulic modeling of the ice ring using the change from water to ice. The phase change to ice in a very
FLOW3D code. The coupled analysis reveals that an ice-ring tightly closed state can generate enormous swelling pressure
can be easily formed by controlling the groundwater drainage due to volume expansion. However, it is expected that the ice
system; this is based on the assumption that the average dis- in rock joints swell into a cavern at an atmospheric pressure
tance of the 0◦ C isotherm reaches 3 or 4 m from the cavern rather than opens joints during the phase change.
wall in the rock mass with the hydraulic conductivity of 10−7 The effect of phase change of groundwater on rock mass
to 10−6 m/s. Through the previous researches, the core of ice- near LNG storage caverns can be summed up the two fol-
ring design is summarized as follows: propagation distance of lowings. First, the residual water near the cavern is frozen,
0◦ C isotherms after the injection of LNG; groundwater pen- and frost heave pressure is generated after injection of LNG.
etration distance within 0◦ C isotherm; and change of joint Second, when an ice-ring is formed due to the recovery
aperture during/after ice-ring formation. of groundwater after the stop of drainage, cryo-suction is
It was confirmed that the propagation distance of 0◦ C occurred in rock joints.
isotherms after the injection of LNG affects the operation In this paper, thermo-hydraulic-mechanical coupled anal-
period of the drainage system, and can be predicted from the ysis with UDEC code is performed to find out the effect
thermal properties of rock mass, and initial temperatures of of cryo-suction on joint apertures. The following two mod-
LNG and surrounding rock mass. Also the following facts els were used for the T-H-M analyses: simple joint and pilot
were verified that in the sense that the stability of contain- plant models. In order to make cryo-suction effect clearer, it
ment system and rock mass around caverns can be affected by is assumed that phase change occurs in a temperature range
frost heave pressure, the penetration distance of groundwater of −20 to 0◦ C. In addition, the permeability of joints decrease
during/after ice-ring formation is an essential checklist for the linearly to zero from 0◦ C to −2.5◦ C.
design of the drainage system, and is mainly affected by the From the numerical results of a simple joint model, it was
recovery rate of groundwater and aperture of rock joints after verified that the phase change occurred and resulted in cryo-
the drainage work. suction in a temperature range of −20 to 0◦ C. Furthermore,
535
the aperture of the joint at the center of the model was reduced As temperature dropped, the joint aperture was gradually
by about 0.4 mm. increased by shrinkage of rock mass as like the results of mon-
During the operation of the LNG pilot plant, joint meters itoring. However, the aperture was reduced in an instant after
were installed and measured the change of aperture of joints at the temperature falls below 0◦ C. The aperture was decreased
the cavern. As a result, the suction phenomenon was observed to 0.01∼0.035 mm in a temperature range of −2.5 to 0◦ C. It
by the joint meter at the top of the cavern around 0◦ C. As can be thought that this result comes from the suction in the
shown in Figure 1, joint apertures were opened in a linear joint.
manner during cooling, although their magnitudes were small, The change of aperture by numerical analysis was smaller
and they were closed as temperatures were raised during the than that by direct measurement from LNG pilot plant. This
thawing stage. It means that groundwater flowed into a joint difference is caused by the following reasons: There are
after the stop of drainage, and the phase change of groundwater different measurement points between LNG pilot plant and
caused the cryo-suction in joints. It is important for the joint numerical analysis. And operating LNG pilot plant at the
opening to show the elastic hysteresis even at the temperature time, the groundwater penetrated deep into the cavern during
cycling, as this implies that the rock around the cavern is elastic ice-ring formation.
even at the low temperature of −30◦ C. In summary, the effect of cryo-suction on aperture of rock
Numerical analysis for LNG pilot plant was performed by joints was verified by the analysis results of the single joint and
the following four stages: First, Excavation of a cavern, sec- pilot plant model. The generalized Clapeyon equation defined
ond, start of drainage for drawdown of groundwater, third, in porous media was used to quantify the phenomena of cryo-
Build-up of containment system and start of LNG supply, suction during ice-ring formation. Since then, it was applied
and finally, Stop of drainage and formation of ice-ring. The to discrete element modeling code such as UDEC and simu-
change of joint aperture was measured at the top, bottom, lated to find out the change of joint aperture during ice-ring
and wall of storage caverns shown in Figure 2. It can be seen formation.
that the aperture of the joint was increased slightly and then Through these results, it was revealed that the cryo-suction
decreased again after the stop of drainage (June 10th). The occurs as groundwater is recovered and changed to ice for
result is caused by the following sequential processes: the forming an ice-ring, and can affect the joint aperture instantly.
increase of pore pressure due to penetration of groundwater Furthermore, the joint apertures around the cavern were
and cryo-suction in the joints. However, it can be expected decreased up to 0.01∼0.035 mm, and the change of the joint
that the increment of pore pressure and cryo-suction occur at aperture had little effect on the stability of the cavern.
the same time in the LNG underground storage cavern.
536
A parametric study to investigate the effect of injection and production well operation on
the CO2 sequestration process in deep unmineable coal seams
S.K. Choi
CSIRO Earth Science and Resource Engineering, Clayton, Australia
Injection of CO2 into a coal seam can cause changes to both the The model results show that when injection is introduced
chemical and physical structures of the coal. It has been shown from a single injection well at the bottom mid-point of the coal
that CO2 adsorption can cause swelling of the coal (Ates and seam, there is a higher rate of CO2 spread at the beginning of
Barron, 1988) and this may lead to reductions in permeabil- injection and the rate decreases with time. This is due to the
ity which will cause difficulties in injection of the CO2 into fact that the CO2 injection rate greatly depends on the pres-
the coal. This swelling phenomenon introduces complexity to sure gradient at the injection well and the gradient decreases
the design of injection systems for CO2 sequestration in deep with time as the CO2 is sequestered.The CO2 injection rate can
coal seams. Drilling an injection well into a deep coal seam be significantly increased by increasing the injection pressure.
(at depths on the order of 1000 m) is an expensive and time This may be related to expansion of coal cleats with increasing
consuming process. Therefore a good understanding of the injection pressure and/or due to increased permeability with
influence of design parameters on the effectiveness of CO2 the formation of new fractures. When more than one injection
sequestration in deep coal seams prior to drilling is impera- well are introduced into the coal seam, CO2 injection capacity
tive. Based on existing knowledge and current understanding is influenced by the pressure distribution around the injection
of the major processes, numerical modeling, which takes into wells. For the modeling scenario used in this work, the max-
account the interaction of the various processes, can be used imum injection capacity was observed for the two-well case
to conduct parametric studies in order to determine the oper- and a reduction in CO2 storage capacity was observed when
ational parameters to achieve optimal response from the coal the number of injection wells was increased. The CO2 storage
seam in the sequestration process. In this study the COMET3 capacity can be greatly increased by incorporation of a pro-
numerical simulator (Sawyer et al., 1990) has been used to duction well. Where a production well is included the optimal
develop a model to study the flow behavior inside a coal seam distance between the two wells will be influenced by a trade-
following CO2 injection. off between the effect that fluid migration toward a proximal
The main objective of this study was to conduct a paramet- production well from the injection well has on the storage
ric study using a 3D numerical model to optimize the CO2 capacity of the coal seam in the vicinity of the injection well,
sequestration process in a 540 m × 500 m × 20 m coal seam, and the time required to achieve pore pressure reduction in
located at 1000 m below the ground surface. The effect of the vicinity of the injection well when the production well
three model parameters on the CO2 storage capacity of the is farther away. From the model results, it is clear that just
coal seam was considered. These were injection pressure, the increasing the injection pressure or the number of injection
number and location of the injection wells and the location of wells will not necessarily increase the CO2 storage capacity
the production well. Firstly, one injection well with 0.1 m inter- of a deep coal seam and that the location of injection and
nal diameter was introduced to the bottom of the coal seam production wells can heavily influence CO2 storage capacity.
from the midpoint and CO2 was injected at 10 MPa pressure A numerical model can be used to assess the effects of oper-
for ten years. Then the injecting pressure was changed from ational parameters such as injection pressure and the number
10 MPa to 20 MPa and corresponding possible CO2 storage of wells and their location on the performance of a deep coal
capacities for ten years of injection were determined. Next, seam CO2 sequestration system.
the number of injection wells was increased to four and cor-
responding storage capacities were observed. Following this,
an injection pressure of 20 MPa was used for the two injec- REFERENCES
tion wells scenario and the distance between the injection
wells was changed to investigate the response in CO2 stor- Ates, Y., and Barron, K. 1988. Effect of gas sorption on the strength
age capacity. Finally, a production well was added and the of coal. Mining science & technology, 6(3): 291–300.
Sawyer, W. K., Paul, G. W., and Schraufnagel, R. A. 1990. Develop-
corresponding storage capacity was observed for a condition ment and application of a 3D coalbed simulator. Paper CIM/SPE,
whereby the water was pumped from the production well at Petroleum Society CIM, Calgary: 90–119.
a rate of 15 m3 /day. The distance between the injection and
production wells was changed and the corresponding storage
capacity variation was observed.
537
Development of the confined pressure type of multi-purpose rock flow system and its
application to the evaluation of permeability of various geomaterials
1 INTRODUCTION
The permeability measurements of bentonite during the with time, please refer the full paper. The X-ray CT image dur-
swelling process are conducted. The bentonite used in this ing swelling process is also shown in Figure 2. These images
study consists of large and small particles. Using the devel- are taken 10 days after commencement of swelling and taken
oped flow system, the intrinsic permeability k are evaluated. in the installed condition in the developed pressure vessels
The value of k is converged around the order of 10−19 m2 during permeation tests. Some original shape of bentonite
after 3 days. This value is almost same as the low perme- particles is observed, it is found that the bentonite has been
ability Granite and bentonite shows enough performance as a almost homogenized at the region around inlet surface, and it
water barrier. As for the relation between the evaluated values is thought that this region works as water barrier.
538
Keywords: fluid flow, hydraulic fractures, back analysis, numerical modelling, field measurements
539
540
H. Soroush
Well Engineering Centre of Excellence (WECOE), Weatherford Oil Tool M.E., Dubai, UAE
V. Rasouli
Department of Petroleum Engineering, Curtin University of Technology, Perth, Australia
541
M. Takahashi
National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan
C. Ahn
Saitama University, Urawa, Japan
M. Kato
Hokkaido University, Sapporo, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: case studies, fluid flow, lab testing, oil reservoir, rock mass, rock properties
In this paper, we use the 3DMA method to quantify the confining-pressure region, permeability of Berea sandstone
flow-relevant geometrical properties of the pore structure in decreased with increasing effective confining pressure. The
Berea sandstone, We measure distributions of pore size, throat rate of change of the permeability weakens gradually with
size, channel length, coordination number, and correlations increasing effective pressure. In horizontal direction, X and Y
between pore and throat geometry of Berea sandstone under axis, the tortuosity under the stressed condition were greater
zero and pressurized conditions for vertical and horizontal than those under zero pressure. In vertical direction, Z axis,
to the bedding plane. In addition, we present the quantita- tortuosity distributions under the stressed condition were com-
tive characterization of tortuosity distribution for intact Berea plex, but its variation distributed within narrow region. These
sandstone and Berea sandstone pressurized to 25 MPa. We phenomena are attributed to the decrease of the number of
discuss relation between permeability reduction by pressuriza- connecting paths or the increase of tortuosity for individual
tion and pore connecting or tortuosity information for different directions.
3 mutually perpendicular directions. In the low effective
Table 1. Connecting information and tortuosity distribution for X, Y and Z directions under zero and 25 MPa pressures.
0 MPa 25 MPa
X start face 226 end face 324 start face 279 end face 371
35816 connecting paths 51072 connecting paths
tort – min 1.50954 med 1.79514 max 2.34023 tort – min 1.54513 med 1.79524 max 2.57653
Y start face 337 end face 358 start face 278 end face 358
70992 connecting paths 51865 connecting paths
tort – min 1.4422 med 1.79058 max 2.46075 tort – min 1.5003 med 1.81598 max 2.59283
Z start face 422 end face 373 start face 357 end face 330
87680 connecting paths 72988 connecting paths
tort – min 1.49172 med 1.74476 max 2.40027 tort – min 1.48993 med 1.77543 max 2.4921
542
Algorithm of coupled seepage and normal stress processes in fractured rock mass by
composite element method
L.L. Xue
School of Civil Engineering and Transportation, South China University of Technology, Guangzhou, China
S.H. Chen
State Key Laboratory of Water Resources and Hydropower Engineering Science, Wuhan University, Wuhan, China
I. Shahrour
Laboratoire de Mechanique de Lille (UMR 8107), Université des Sciences et Technologies de Lille, France
ABSTRACT: There is an important and complicated cou- equivalent continuum) approach (Long, etc.,1982; Hsieh &
pling between seepage and stress-strain fields in fractured Neuman, 1985; Oda, 1985, 1986), which takes into account the
rock mass: the stress-strain field is influenced by the seepage; deformation and permeability characteristics of fractures but
however, the stress-strain field affects the permeability char- neglects their exact positions, therefore, leading to some calcu-
acteristics and in turn the seepage field as well. In the seepage lation errors. Another one is the explicit (or discrete) approach
and stress-strain coupling system, fractures play a fundamen- (Schwartz, etc. 1983; Dershowitz & Einstein, 1987; Cacas,
tal role because the seepage and deformation of fractured rock etc. 1990; Andersson & Dverstorp, 1987), which considers the
mass are mainly dominated by them: stresses acting on the geological and mechanical properties of each fracture deter-
fracture surface lead to the change of fracture aperture, which ministically. Usually there are a large multitude of fractures
causes the change of conductivity matrix of the fractured rock with small aperture. Therefore, the explicit approach often
mass and the seepage field. According to this coupling mecha- makes the pre-process work time costly and tedious, even
nism, seepage and stress-strain coupling analysis of fractured makes the generated mesh with bad topological configuration,
rock mass can be achieved by applying the iterative algorithm. which brings some calculation errors or makes calculation
It is also widely recognized that in the geotechnical engineer- nonconvergent.
ing, shear influence on fracture is less important than tension Composite element method (CEM) has been proposed by
or compression influence (Zhang, 2005; Raven & Gale, 1985; Chen (2003). One composite element is defined to cover sub-
Tsang & Witherspoon, 1981). domains with different shapes and characteristics, and these
Seepage and normal stress coupling in fractured rock mass sub-domains are named as sub-elements. Unknown variables
has been studied from various aspects including experimental within each sub-element (e.g. φ, {u}) can be interpolated
technology, mechanic parameters, numerical models, and so from the mapped nodal variables defined on the composite
on. Snow (1968) built the relationship between permeability elements (e.g. {φ}, {δ}), and the shape function is the same
coefficient and fracture aperture for horizontal fracture. Louis as that of conventional FEM defined over the whole com-
(1974) obtained a famous semi-empirical formula between posite element. The governing equation for the solution to the
permeability coefficient and normal stress through the packer mapped nodal variables can be established by applying Virtual
test in borehole with different depth in the homogeneous frac- Work Principle or Variational Principle. Based on the solved
tured rock mass. Gale (1982) proposed experiential formula variables on the mapped nodes on the composite elements,
between permeability coefficient and normal stress through the unknown variables within the relevant sub-element can be
the laboratory experiments on three kinds of rock mass. Bar- interpolated through the shape function. The proposed CEM
ton (1985) proposed joint roughness coefficient (JRC), and can discrete the fractured rock region regarded as one homo-
founded one empirical formula for seepage and stress-strain geneous medium, then the fractures can be embedded within
coupling. Malama and Kulatilake (2003) proposed a general the elements but be simulated explicitly. In this way the com-
exponential model by defining half-closure stress σ1/2 , in putation mesh generation based on the CEM is more simple
which fracture deformation is a function of the maximum frac- and convenient, especially for the complicated fractured rock
ture closure and σ1/2 . Tsang and Witherspoon(1981), Chen problems.
(1989) proposed physical models to simulate the coupling Since flow rate is proportional to the third power of frac-
behavior between seepage and normal stress-strain. ture aperture, and most deformation of fractured rock masses
The numerical methods used in simulation of rock frac- is generated by fracture deformation, therefore, fracture defor-
tures contain two categories. The first one is the implicit (or mation governs the flow capability of fractured rock masses.
543
544
S. Nakashima
Department of Civil and Environmental Engineering, Yamaguchi University, Ube, Japan
K. Kishida
Department of Urban Management, Kyoto University, Kyoto, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
EXTENDED ABSTRACT
545
Physical clogging model using stick rate of calcite on CO2 mineral trapping
Y. Kuroda
International Petroleum Exploration Co., Japan
A. Ueda
Graduate School Division of Science and Engineering, University of Toyama, Japan
Keywords: fluid flow, field measurements, lab testing, rock properties, geothermal, numerical modeling
Flow experiments through porous and fractured media were upward. According to results, an order of magnitude differ-
performed to examine the clogging phenomena by carbonate ence of stick rate between the column test and the in-situ
mineralization at high temperature. Flow rates of the column test is distributed to 2 orders at the similar temperature. This
test decreased gradually due to calcite precipitation. The clog- discrepancy might create different transportation that deposit
ging model with a stick rate of calcite explains the observed particles, once migrated and settled in pores, remain under
permeability changes on the in-situ test. On the other hand, constant head conditions of the column test, whereas under
a numerical simulation suggests an alternative how to eval- constant flow rate conditions of the in-situ test they might
uate the deposition rate of mineral due to CO2 -water-rock undergo re-entrainment. If we accelerate the stick rate of
interaction. This carbonate mineralization on geologic seques- calcite precipitation during injecting carbonated water, the
tration will decrease the possibility of CO2 leakage risk at permeability reduction of reservoirs will occur.
hydrothermal conditions and prevent buoyancy from moving
Figure 1. (a) Schematic experimental setup for column tests (b) schematic experimental setup for in-situ tests.
Figure 2. Column test: (a) Fitting flow rate curve. In-situ test: (b) head and flow rate curve near the bore hole (c) fitting permeability curve.
546
As a part of the overall IRSN R&D program that provides the Five performance tests using wireless intra-core instru-
bases for scientific expertise on disposal safety, the SEALEX mentation allow exploring conditions that may impact the
project is focused on sealing systems efficiency. Thanks to long-term performance of a clay-based seal, changing a single
limited-size in situ experiments on long-term performance parameter at a time with respect to a base case:
(i.e. under isothermal and water-saturated conditions) of
• The first test is a base case test itself, with a core build of
clay-based seals, SEALEX is dedicated to:
monolithic precompacted disks (70/30 MX80/sand mix).
• Test the long-term hydraulic performance of sealing sys- • Three tests designed to quantify, with respect to this base
tems (in normal conditions, i.e. non altered), for different case, the impact of the technological choices retained
core compositions (pure MX80, sand/MX80 mix) and for the engineered barrier. Modifications concern the
conditionings (pre-compacted blocks or in situ compacted); intra-core geometry (jointed disks), the core composition
• Quantify the impact of intra core geometry (construction (MX80/sand ratio) and the core conditioning (precom-
joints in the case of pre-compacted blocks) on the hydraulic pacted vs. in situ compacted):
properties of sealing systems;
– The intra-core geometry (jointed vs. monolithic disks,
• Quantify the effect of altered conditions (an incomplete
both precompacted);
saturation of the swelling clay or an incidental decrease
– The core composition and conditioning (core made of
of the swelling pressure caused by a failure of the concrete
pure MX80 pellets/powder compacted in situ vs. 70/30
confining plugs) on its performance, which tests the concept
mix precompacted monolithic disks);
robustness with respect to the hydraulic characteristics of
– The core composition (low MX80/sand ratio vs. reference
the system.
one 70/30);
This paper presents the main specifications of this test pro-
• Finally, a test to quantify the influence of altered conditions
gram, simulation results of the full hydration phase and the
(saturation defect and confining plug failure) with respect
first experimental results of the initial hydration phase.
to this base case.
All the experiments are emplaced in the Tournemire
URL (Toarcian argillites), via horizontal boreholes (diameter Pre-design simulations were carried out to optimize the
60 cm) excavated from recent drifts (2008). Each experiment dimension of the clay core seals with respect to the hydra-
represents a generic seal mock-up (i.e. either at a relevant tion phase duration and to the tests representativeness. Main
scale, or with relevant characteristics), except the presence of results are presented.
an artificial resaturation system. In parallel, a laboratory characterization program was ini-
Regarding the intra-core instrumentation, it was chosen to tiated to optimize the clay core properties with respect to the
limit as much as possible the disturbance induced by instru- specified target swelling pressure and permeability values.
mentation on the core performance, thus to avoid the potential This laboratory program also aims to characterize the long-
occurrence of flow pathways along signal wire paths. Accord- term intrinsic behavior of the clay core that will be mobilized
ingly, two types of tests have been designed: reference tests especially in the confinement loss test.
and performance tests. Finally, the first results of the initial hydration stages are
The reference tests aims at quantifying the hydro- briefly presented.
mechanical fields (swelling pressure, pore pressure, relative The SEALEX in situ experiments is an original testing pro-
humidity) within the core and their time-evolution as the core gram that should give valuable results for long-term safety
saturation proceeds, using classical wired sensors. evaluation. Preliminary results are encouraging.
549
F. Deleruyelle
IRSN, DSU, SSIAD, BERIS, Fontenay-aux-Roses, France
Keywords: nuclear disposal facility, numerical modelling, fluid flow, mine design
The handling drifts of an underground nuclear waste disposal Firstly, under simple geometric hypotheses, a complete
facility will be backfilled at the end of their operating time. simplified life cycle of the cavity is assumed (Fig. 2): (1) exca-
The lining of these tunnels will thus no longer be maintained. vation; (2) waiting for the liner; (3) lining support installation
Inevitably, they will lose their mechanical integrity with time and exploitation; (4) post-closure stage after backfilling.
and in a creeping rock, this may induce an inward, converg- Initial and boundary conditions are also presented.
ing movement of the surrounding rock mass. After filling up Secondly, fundamental equations describing coupling phe-
all the voids left behind by the construction and backfilling nomena are shown. From the fluid mass conservation equa-
operations, the rock mass would compress the backfill until a tions, the hydraulic diffusion equation is derived, linking pore
new equilibrium state is reached. In saturated porous media, pressure to volumetric strain. Poroelasticity equations are then
this could induce a water outflow, which could accelerate the recalled, giving notably incremental stress and porosity versus
dissemination of radioactive elements. This important issue incremental strain and pore pressure. Viscoplastic damage-
needs to be investigated for the purpose of long term safety able behaviour is finally described. From these equations are
assessment. derived the characteristic times which allow us to analyse
The present study aims mainly at estimating the hydraulic hydromechanical phenomena from a qualitative point of view.
disturbances inferred by this load redistribution between the Moreover, within the framework of saturated poroelasticity,
liner and the backfill around a typical gallery of an under- a quasi-analytical solution of this problem is developed using
ground nuclear waste disposal facility. The evolution kinetics Laplace transform. Solution in space and time domains can
of different hydromechanical variables (water pressure, dis- be obtained via numerical computations using the Stehfest’s
placements…) depends on a number of factors: rate of lining algorithm. It will be used as a benchmark test in subsequent
deterioration, hydraulic diffusivities of the ground and the stages. It shows for example the dissipation phenomenon dur-
backfill, initial backfill compaction as well as the general ing phases (2) and (3) (Fig. 3) or the outward displacement of
behaviour of the rock mass. The latter may exhibit a com- the host medium after backfilling (Fig. 4).
plex hydromechanical behaviour involving creep, hardening, Finally, a numerical solution is highlighted using finite ele-
and even damage. ments method via Cast3M software. It allows considering
To analyse this difficult problem, a progressive step-by- more realistic but complex behaviour of the host rock, sup-
step approach has been adopted from analytical modellings to posed to be a porous medium, such as creep (with or without
numerical simulations. hardening), coupled or not with damage effects. Comparisons
between numerical results and previously obtained quasi-
analytical solutions or published results show consistency and
allow to gain confidence in this numerical modelling.
Considering influence of liner damage or backfill com-
paction on the post-closure hydromechanical behaviour of the
cavity, principal results are presented.
550
551
He Manchao
Stake Key Laboratory for GeoMechanics and Deep Underground Engineering, Beijing, China
Zhang Na
Stake Key Laboratory for GeoMechanics and Deep Underground Engineering, Beijing, China
There are several ways of mitigating greenhouse gases (GHG) safely injected and stored at well characterized and properly
emissions to the atmosphere, however, the storage of large managed sites. Injecting carbon dioxide in deeply geological
quantities of carbon in geological formations is presented formations can store it underground for long periods of time.
today as one of the most effective and visible results. Carbon At depths below about 800–1000 m, CO2 has a liquid-like
dioxide (CO2 ) capture and storage (CCS) is a process consist- density that permits the potential for an efficient use of the
ing of the separation of CO2 from industrial and energy-related underground reservoirs in porous of sedimentary rocks.
sources. Fig. 1 brings together, in schematic form, the main Geological storage requires constructing facilities to cap-
sources and some of the possible storage sites. ture in large emission sources of CO2 such as power for
Storage of CO2 in deep geological formations uses many electricity production or cement, steel, ethanol, among others.
of the technologies developed by oil and gas industry and has It is also possible to act on fuel, that is, instead of proceeding
been proved to be economically feasible under specific condi- to the capture of CO2 after combustion of coal or natural gas
tions in oil and gas fields and saline formations (IPCC, 2005). can be drawn primarily engaged in industrial units, the car-
CO2 can also be stored in carboniferous formations, either in bon present in them. The captured CO2 is then transported by
unminable coal seams or in abandoned coal mines. CO2 can be pipelines or in ships, for underground storage sites. Most of
the mechanisms related to this technology are not new, since
they are already employed by the oil industry by contract-
ing for management and distribution of natural gas, for some
industries in the food sector, among others.
Currently CO2 capture is a costly and energy consuming.
The costs obviously depend on the dimension of the industrial
unit and of the type of fuel used. There are four basic systems
for capturing CO2 from use of fossil fuels and/or biomass
(IPCC, 2005): capture from industrial process streams; post-
combustion capture; oxy-fuel combustion capture; and pre-
combustion capture.
The environmental impacts from geological storage of CO2
can be integrated into two types of categories, i.e. local envi-
ronmental effects and global effects from the release of stored
CO2 to the atmosphere. Global effects may be viewed as the
uncertainty in the effectiveness of CO2 storage. Local hazards
arise from causes like direct effects of elevated gas-phase CO2
concentrations in the shallow surface or near surface, effects of
dissolved CO2 on groundwater chemistry and effects caused
Figure 1. Capturing and storage of CO2 (NETL, 2007). by displacement of fluids by the injected CO2 .
552
553
H. Kunz
Federal Institute for Geosciences and Natural Resources (BGR), Hanover, Germany
H.-G. Zhao
Beijing Research Institute of Uranium and Geology (BRIUG), Beijing, P.R. China
J. Wang
Beijing Research Institute of Uranium and Geology (BRIUG), Beijing, P.R. China
V. Bräuer
Federal Institute for Geosciences and Natural Resources (BGR), Hanover, Germany
554
2D 3D 3D 3D
Distance T-Model T-Model TM-Model THM-Model
REFERENCES
555
Y.M. Liu, J. Wang, X.G. Zhao, D. Ke, J.L. Xie, S.F. Cao, L.K. Ma, W.J. Jiang & L. Chen
CNNC key laboratory on geological disposal of high level radioactive waste, Beijing Research Institute of Uranium Geology, Beijing, China
1 INTRODUCTION
556
557
1 INTRODUCTION AND BACKGROUND fashion. The glacially induced pore pressure is influenced by
the coupling in two ways: 1) parts of the glacially induced
The Swedish Nuclear Fuel and Waste Management Co. (SKB) mechanical load may be transferred to the pore pressure and
are planning to construct a repository for permanent storage of 2) variations in effective stress may alter the hydraulic con-
high-level nuclear waste in the Forsmark area at approximately ditions in the rock. An assessment, in one dimension, of the
450 m depth in crystalline rock. impact of the mechanical load during SKB’s reference glacial
During its life-time, the repository host rock will be sub- period has shown that with a coupled analysis, the pressure is
jected to a number of different loads, which, under certain cir- increased more quickly during the advance of the ice sheet and
cumstances, will be accompanied by an increase in pore pres- reduced more efficiently during the retreat compared with an
sure. If the pore pressure in a fracture exceeds the sum of the uncoupled analysis. In the upper kilometre of rock the influ-
fracture’s normal stress and tensile strength, the result would ence is small. However, at very large depths, the pore pressure
be a significant and difficult to quantify increase in the frac- is dominated by the mechanical load. Typical variations in
ture’s transmissivity – a phenomenon known as hydraulic jack- transmissivities during the glacial phase are estimated to be in
ing. Therefore, if hydraulic jacking was likely to occur in the the range 0.4–3 of present-day values (Hökmark et al. 2010).
near vicinity of the repository, this would have to be included These should be compared with the large natural variability
in the risk assessment required by the regulating authorities. among fractures at Forsmark (Follin et al. 2007). However,
Hydraulic jacking is unlikely to be initiated at other times without explicit modelling, it is difficult to estimate the impact
than during a glacial period. After excavation, there will be on the effective large-scale permeability. Based on these anal-
atmospheric pressure within the excavated openings leading yses, hydro-mechanical couplings are not further considered
to a reduction in the pore pressure of near-field fractures. The in the assessment of hydraulic jacking.
hydraulic conductivity (Follin et al. 2007) at all depths is suf-
ficiently high to prevent large pore pressures from building up 3 MEANS TO ADDRESS HYDRAULIC JACKING
during the heated phase (Selvadurai & Nguyen 1996). How-
ever, during combined glacial and permafrost conditions, the The pore pressure is estimated by use of a combination of
low permeability of the frozen rock will prevent an upwards analytical solutions and numerical models in two dimensions
flow of water and increase the pore pressure. At the end of based on a continuum representation of the rock mass. The
the glaciation, a large proportion of the pore pressure may be ice sheet is represented by a generic steady state ice profile
retained by the rock. When the ice sheet covers the site, the (Paterson 1994) with a maximum thickness of 3 km. Frontal
glacially induced stresses are sufficiently high to balance the advance and retreat rates are set according to typical values in
increase in pore pressure. Several studies of the potential for SKB’s reference glaciation scenario (SKB 2010). The hydro-
hydraulic jacking during a glacial cycle have been conducted static water pressure at the ice/bed interface is set to 90% of
over the years (e.g. Chan et al. 2005, Vidstrand et al. 2008) the ice thickness. For combined glacial and permafrost condi-
but there is no commonly agreed model of the maximum jack- tions, seasonal variations in the pressure beneath the ice sheet
ing depth. The need to include hydro-mechanical couplings in are also considered. These are only noticeable within a dis-
the assessment of hydraulic jacking beneath a warm-based ice tance of a few kilometres from the meltzone. Further away,
sheet has been debated. Although it is theoretically possible the pore pressure is governed by the average pressure over a
for hydraulic jacking to be initiated at large depths, the Fors- seasonal cycle.
mark site does not appear to have the necessary geological The jacking criterion will first be fulfilled in fractures ori-
or hydrological conditions (Hökmark et al. 2006). Observa- ented perpendicular to the least principal stress, which, at
tions of sediment-filled sub-horizontal fractures, which may the Forsmark site, has been shown to be the vertical stress
be indications of hydraulic jacking events in the past, have (Hökmark et al. 2010). Therefore, it is sufficient to consider
only been made to depths of a few tens of metres. This paper the potential for hydraulic jacking of horizontal fractures. It
will illustrate the impact of hydro-mechanical couplings and is assumed that the tensile strength is zero such that hydraulic
establish bounding estimates of the maximum jacking depth jacking occurs when the pore pressure in a fracture exceeds the
during SKB’s reference glacial cycle. normal stress. The glacially induced vertical stress at all depths
is set equal to the mechanical load at the ground surface.
2 IMPACT OF HYDRO-MECHANICAL COUPLINGS
4 RESULTS
In order to determine the hydro-mechanical (H-M) evolution
of the rock mass, it is generally required that equations of During combined glacial and permafrost conditions as the ice
fluid flow and stress or deformations are solved in a coupled front is advancing, the hydraulic conditions of the rock mass
558
559
T. Nishimura
Department of Civil Engineering, Tottori University, Japan
Keywords: numerical modelling, risks and hazards, rock slopes and foundations
560
S. You
School of Civil and Environmental Engineering, University of Science & Technology Beijing, Beijing, China
G.F. Zhao
School of Civil and Environmental Engineering, The University of New South Wales (UNSW), Sydney, Australia
V. Labiouse
Laboratory of Rock Mechanics, Ecole Polytechnique Federale de Lausanne (EPFL), Lausanne, Switzerland
Within the 6th EURATOM Framework Program of the Euro- excavated in these two mediums. However, mechanism of why
pean Community, the TIMODAZ project (Thermal Impact the two apparent different failure patterns are formed is still
on the Damaged Zone Around a Radioactive Waste Disposal not clear, it also cannot be directly obtained from existing labo-
in Clay Host Rocks) aims at studying the fracturing and ratory tests or analytical solutions. This paper is to understand
self-sealing processes that develop in the Excavation Dam- the mechanism of the failure process of two different clay for-
aged Zone (EDZ) around disposal galleries for heat emit-ting mations during excavation. In that case, the micro mechanics
radioactive waste and at assessing the impact of the ther- by means of numerical modeling based on the DLSM with a
mal phase on their evolution. In this framework, simulation two phase model is performed. The numerical modeling are
tests are performed at the LMR-EPFL on hollow cylinders try to reproduce different failure patterns observed in exper-
with mechanical loadings fairly similar to the reality that will iments through choosing different micro failure laws of the
occur around galleries for heat emitting radioactive waste. Two micro structure. From the modeling results, it shows that the
different potential geological formations for deep radioactive micro tensile failure between particles is the main reason of
waste repositories are considered, i.e. Boom Clay (a kind of anisotropic failure for Opalinus Clay, whereas the micro shear
plastic clay) and Opalinus clay (an indurated clay) are promis- failure is the main failure mechanism for Boom Clay. Finally,
ing options for deposit nuclear wastes in Switzerland and the role of mechanical anisotropy on the failure process in the
Belgium respectively. The results from laboratory tests show near field of the underground galleries in those clay formations
that different failure patterns will be formed for the galleries are analyzed and discussed.
561
J.J. Zhang
North China Institute of Water Conservancy and Hydroelectric Power, Zhengzhou, Henan, China
B.J. Fu
Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
Z.K. Li
Tshinghua University, Beijing, China
S.W. Song
Chengdu Investigation and Design Institute, China Hydropower Consulting Group Co., China
Y.J. Shang
Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, China
562
For verifying the suggested method mentioned above, 5 case At present, different methods for rockburst criteria and clas-
histories relating to underground works of Jinping I and II sification put forward by different scholars are available.
Hydropower Stations, Cangling Highway tunnel, diversion With reference to different methods available, combining
tunnels of Tianshengqiao No.2 and Taipingyi Hydropower sta- engineering experiences in China, this paper presents a five-
tions in China are provided. The results obtained seem to be factor comprehensive method for criterion and classification
satisfied. of strain-mode rockbursts as well as some case histories for
Especially, in Jinping I Hydropower Station, the average further discussion.
value of Rc /Rt is 11 (<15), so there was no rockburst, but a
series of failure happened in the study area (Chengdu Investi- ACKNOWLEDGEMENT
gation, Design and Research Institute, 2009; Chinese Society
for Rock Mechanics and Engineering, 2009a,b). Sincere thanks are due to Prof. QIAN Qihu, Fellow of Chinese
Academy of Engineering for his kind direction and revision.
The five-factor rockburst criterion and classification can be Chinese Hydroelectric Consulting Group, Chengdu Investigation,
explained through diagram as shown in Fig. 1. Design and Research Institute. 2009. Report on Deformation
It is indicated that the heavy rockburst region is located at and Failure Mechanism as well as Reinforcement Measures of
the left upper part with inclined parallel lines when all the five Surrounding Rock Masses in Underground Caverns Complex of
Jinping No.1 Hydroelectric Power Station on Yalongjiang River in
factors belong to heavy rockburst class simultaneously. Sichuan Province. Sept. 2009.
The area with vertical parallel lines at the middle upper Chinese Society for Rock Mechanics and Engineering, Chengdu
part shows that only 3 among the 5 factors belong to heavy Investigation, Design and Research Institute, Chinese Hydro-
rockburst class. electric Consulting Group. 2009a. Consultation Report for the
Stability and Support Technology of Surrounding Rock Masses of
Underground Complex for Jinping I Hydropower Station.
5 DISCUSSIONS ON SOME PROBLEMS Chinese Society for Rock Mechanics and Engineering. 2009b. The
International Consultation Report for the Key Technology of Safe
and Rapid Construction for Jinping II Hydropower Station.
Some key problems concerning hardness, brittleness of rock
Zhang Jingjian & Fu Bingjun. 2008. Rockburst and its Criteria and
material, intact index of rock mass and so on are discussed. Control. Chinese Journal of Rock Mechanics and Engineering,
The term RC /Rt is commonly used to evaluate the brittleness Vol. 27 No. 10, pp 2034–2042.
of rock in the world. Obviously, if RC /Rt = 1 the rock should be
equivalent to metal, having no brittleness, but large ductility.
Judging from this, to adopt the brittle index RC /Rt for
evaluating the brittleness of rock is acceptable.
563
L. Zhou
Energy Research Centre of Lower Saxony, Goslar, Germany
M.Z. Hou
Institute of Petroleum Engineering, Clausthal University of Technology, Clausthal-Zellerfeld, Germany
Keywords: case studies, dynamics, geothermal, hydraulic fractures, numerical modeling, risks and hazards
1 INTRODUCTION tensile strength left after the initial tensile failure. So the sec-
ond tensile failure behind the fracture front contributes hardly
In the deep heat mining (DHM) project Basel, a seismic to energy releasing. On the contrary certain residual shear
event with a local magnitude of 3.4 was measured after about strength always exists in the rock after the initial shear failure.
12,000 m3 water injection, which caused building damages in There is probably a big shear failed volume (e.g. 1,000,000 m3 )
the city of Basel, the stop of the project for ever and suspicion behind the fracture front during the injection, which could be
against deep geothermy. For this reason it is very impor- the mechanism for the large amount of energy release in deep
tant to investigate the mechanism and to simulate seismic geothermy.
events induced by reservoir stimulation of an EGS (enhanced
geothermal system). Through the numerical modeling, the
process of energy release and the influencing factors on the 3 COMPUTATIONAL PRINCIPLE IN FLAC3D TO
energy release can be better understood, so that the procedure CALCULATE KINETIC ENERGY AND LOCAL
of the stimulation can be optimized and the geo-risks can be MAGNITUDE
minimized.
The kinetic energy of an object is the energy which it possesses
due to its motion. It is defined as the work needed to accelerate
2 MECHANISM OF INDUCED SEISMIC EVENTS BY a body of a given mass from rest to its current velocity. If the
RESERVOIR STIMULATION velocity of the object is known, then the kinetic energy is also
known. To calculate the kinetic energy the velocity must be
The induced seismic event by reservoir stimulation in principle firstly computed, which is the native function in FLAC3D, if
belongs to a thermal-dynamic process, in which the energy the dynamic option is turned on. The theoretical background
conservation rule must be obeyed. In other words the energy of dynamic calculation in FLAC3D bases on linear elasticity.
cannot be lost but transform from one type to another. The The key is to solve the equations of motion. During the calcu-
transformation can be described by eq. (1). lation the velocity of all grid points will be computed at every
calculation step. So the kinetic energy can be calculated with
the known velocity and mass. By use of an empirical equation
the kinetic energy can be converted to local magnitude.
564
5 CONCLUSION
oil boreholes (Hou & Zhou 2009). The reasons are the differ- REFERENCES
ent in-situ stress state, shear and tensile strength, as well as
Häring, M.O., Schanz, U., Ladner, F. & Dyer, B. C. 2008. Character-
softening behavior of rocks.
isation of the Basel 1 enhanced geothermal system. Geothermics
During the injection the kinetic energy in the model at each 37 (5): 469–495.
time step is continuously registered and evaluated as local Hou, Z. & Zhou, L. 2009. Vorausberechnung der gespeicherten und
magnitude, which is shown in Figure 3. The biggest calculated freizusetzenden Deformationsenergie bei Frac-Behandlungen in
value is 2.95, which is smaller than the biggest measured value der Tiefengeothermie. Presentation at the Geothermal Congress
3.4. However the tendency of both is comparable. Furthermore 2009, Bochum, 17–19 November 2009.
565
M. Akiyama
Advanced Technology Development Project Group, Geoscience Research Laboratory, Co., Ltd., Yamato, Japan
S. Kawasaki
Faculty of Engineering, Hokkaido University, Sapporo, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
Figure 1. Flow chart of the study. Figure 4. SEM images of test piece segments.
569
R. Migliazza
Dipartimento di Ingegneria Civile, dell’Ambiente, del Territorio e Architectura, Universitá de gli Studi de Parma, Italy
Keywords: mine design, site characterization, case studies, mining, numerical modeling, rock caverns
1 INTRODUCTION
In the recent times, the high cost of land and the stringent
environmental legislation, together with the good prices of
some minerals and rocks, such as marble and roofing slate,
has forced miners to foster the underground mining (Figure 1).
Even if the mining costs increased in comparison to quarrying,
by means of a cheap method like room & pillar, mining costs
can be kept at low levels (10–20 a/ton) in such a way that the
mining of these minerals is still profitable.
We have been involved in the design and stability study of
some of these mines, and in the preset article we seek to present
some of the basic methodology for the design of these mines.
This includes the selection of empirical simple methods for
initial design and the application of numerical techniques to Figure 2. Design and detail of numerical models of a quartz and a
simulate the proposed models (Figure 2). magnesite room & pillar exploitation in Spain.
The combination of empirical, analytical and numerical
techniques result quite useful for these design processes, since
when one single method is applied, some topics maybe over- 2 BASIC DESIGN CONSIDERATIONS
looked, so the combination of methods provide more confident
designs. To provide a good design the field and laboratory investiga-
tions are of relevance.A basic knowledge of the geology is first
needed. Since the underground mines we are talking about are
excavated near open pit mines of the same mineral, we have
available excellent outcrops to measure a good number of dis-
continuities following standards. These data are interpreted to
obtain joint sets and their features.
In our case, we usually have core-drilling in the areas to be
mined, since it is important to locate the walls of the seam to be
mined and to control the mineral grade or quality. Rock cores
are used to perform laboratory characterization tests. Labo-
ratory and field information is gathered to obtain the basic
figures of empirical approaches including Q, RMR and GSI.
Lab tests are interpreted to obtain peak strength envelopes,
deformability pre-failure data and, if necessary, post failure
parameters, following published characterization techniques.
In the room & pillar mining method, the extent of the rooms
is directly related to the mining costs. So in our designs, rooms
tend to be as large as possible. In this way, our initial room
Figure 1. Pictures of some room & pillar exploitations of designs, including the support needs, have been based on the
non-expensive minerals and ornamental rocks in Spain. application of the stability graph method (Potvin, 1988). This
570
The initial designs are based as illustrated on empirical We currently have available a number of sufficiently devel-
approaches and traditional rules of thumb. Once this initial oped rock mechanic tools to design room & pillar mines
design is proposed, and it has been proved that the mine is in shallow mineral or ornamental rock deposits. These tech-
economic, this design is checked out in detail by analysing the niques can foresee varied instability mechanisms, which may
different geomechanical issues which could produce problems occur in this type of exploitations. A proper rock mechan-
during the exploitation stage. ics approach should start with careful field data recovery and
painstaking laboratory testing. Empirical methods are conve-
3.1 Structurally control instabilities nient to propose a basic design, but analytical and numerical
techniques are needed to fine-tune these designs. Moreover,
With the field information and according to block theory or different types of control and monitoring are needed, to check
standard underground wedge analysis the factors of safety of the viability of the design implementation. Some of these
the wedges formed around the rooms are estimated. The initial mines have largely shown to be economically feasible.
support may be increased in particular zones of the rooms if
needed. It is important to correctly quantify the actual persis-
tence of discontinuities for this is a highly relevant issue in ACKNOWLEDGEMENTS
wedge failure control. The projection of joint traces derived
from field data in the roofs and walls of the room, with the The authors acknowledge the Spanish Ministry of Science &
help of codes able to represent joints can be helpful to quantify Technology, Spain, for financial support of the project under
continuity. reference number BIA2009-09673, which has been of help in
some of the research topics presented in the paper.
3.2 Stress-strain behavior
Stress-strain instability studies can be performed with the help REFERENCES
of the various continuous numerical methods available in the
market. In the mines we are designing, due to its low depth, Alejano, L.R., Taboada, J., Bastante, F.G., Rodríguez, P. 2008. Multi-
plastic zones do not usually extend far from the room surface, approach back-analysis of a roof collapse in a mining room
however this stress strain analyses are convenient, for they excavated in stratified rock. Int. J. of Rock Mech. Min. Sci. &
can also put forward the occurrence of tensile zones and reveal Geomech. Abstr. 45: 899–913.
relaxation problems. The role of faults can also be enlightened Hedley, D.G.F. & Grant, F., 1972. Stope and pillar design for the Elliot
by performing numerical models. Lake uranium mines. Can. Inst. Min. Metall. Bull 65:37–44.
Potvin, Y. 1988. Empirical open stope design in Canada. Ph.D. The-
sis. Dept. of Mining and Mineral Processing, University of British
3.3 Study of stability of rooms in stratified media Columbia, 1988.
For the design of rooms in stratified rock masses, particu-
lar attention is given to the immediate roof, which includes
571
Double shearing of rebar steel and cable bolts for effective strata reinforcement
Hossein Jalalifar
Department of Mining and Petroleum Engineering, Shahid Bahonar University of Kerman, Iran
An experimental study was undertaken to examine the shear- These bolts were tested in 20, 40, 50 and 100 MPa con-
ing behaviour of both steel rebar and hollow strand cable bolt crete blocks, subjected to different pretension loads of no axial
across joints and shear planes. Double shearing testing using tension (0 kN), 5, 10, 20 50 and 80 kN respectively.
double shear box as shown below, were made on rebar steel A particular aspect of this study was focused on the influ-
and cable bolts in concrete blocks. Five steel rebar bolt types ence of rebar (bolt) profile configuration on the load transfer
were used for the study. Three bolts were 22 mm core diam- capacity between bolt, resin and rock, under shear. It was found
eter bolt types with different profiles, the forth was 11 mm that; the level of maximum shear load and displacement were
threaded bars and the fifth was a 12 mm construction bolt. different because of different pretension loads; the strength
of the concrete had significant influence on the shear load
behaviour and shear deformation of the concrete in the vicin-
ity of the sheared joint planes; the shearing load across the
joint planes were less in 20 MPa concrete in comparison with
higher strength concrete of 40 and 100 MPa concrete. The
strength of the concrete and the level of bolt pretension were
found to influence the elastic limit and stiffness of the bolt.
The testing of the 28 mm diameter cable bolt was carried
out in a 50 MP concrete with the cables being tested under
two initial axial loading of 50 and 100 kN respectively. The
cable consisted of a hollow mild steel core surrounded with
nine 7 mm diameter wire strands.
Two tests were carried out using two 1.5 m long cable sec-
tions. The first test was limited to a 50 mm travel of the testing
machine loading plates and produced a shear load of 900 kN
(92 t) at the maximum 50 mm displacement. The axial load
generated on the cable bolt was 238 kN (24.3 t). In the second
test the machine travel was increased to 75 mm, and first strand
failure of the cable, due to shear loading, occurred at 1 354 kN
(138 t) vertical load at a vertical displacement of 59 mm. The
cable axial load was in the order of 385 kN (39.3 t). Analysis
of the failure mode and loads achieved showed that the cable
strands bent and the concrete crushed along the shear plane,
the shear loading across the concrete and grouted cable then
Figure 1. Double shear testing of bolts and cable bolts in concrete. reached the tensile strength of the steel wires.
572
The aim of this paper is to present the numerical evaluation of There, an internal pressure pi = 10 MPa is applied at the tun-
the load distribution along fully grouted rock bolts. In order nel boundary, stresses are initialized in the finite differences
to achieve that, three series of models are prepared using the zones and the model is allowed to reach an initial equilibrium
explicit finite difference codes FLAC2D and FLAC3D (Itasca, state. After the initial reproduction of the stress conditions,
2009). four stages follow which simulate the excavation and the rock
Firstly, plane strain and three dimensional models are pre- reinforcement process. The tunnel excavation is simulated by
pared, which intend to simulate a rockbolt installed in the rock reducing the internal pressure at the tunnel boundary. At the
walls of a circular tunnel with radius R = 5.0 m. Secondly, two second stage, the internal pressure is reduced to pi = 3 MPa
dimensional axisymmetric models are also prepared, which and the model is once again let to reach an equilibrium state.
simulate a rockbolt installed in the rock walls of a spherical At this point the rockbolt is installed along with the grout and
cavity of the same radius. the face plate. In the third and forth stages the internal pres-
Four different rock masses are considered, with the proper- sure is reduced to pi = 2 MPa and pi = 1 MPa respectively. In
ties shown in Table 1. The intact rock properties correspond the last stage no internal pressure is applied.
to a rock of medium to high strength. The rock mass is In the plane strain model series, the rock bolt is simulated
assumed to be elastic-perfectly plastic following the “general- using the “cable” model built in the FLAC2D code. The influ-
ized” Hoek-Brown strength criterion (Hoek et al. 2002), which encing area of the bolt in the rock is modeled as a sector of
is implemented in the FLAC code. For the present simulations circular annulus.
associated flow conditions are approximated. The axial load distribution along the rockbolt is shown in
A rockbolt pattern of 1 m2 is assumed. The length of the Figure 1. These diagrams indicate that the anchoring part of
rockbolt is 3.0 m and its diameter is 25.0 mm. The bolt is con- the rockbolt is reduced with a movement of the neutral point
sidered to be an elastic material with Young’s modulus value towards the rockmass with decreasing rock mass quality. Also,
equal to 200 GPa. The grout has an annulus of 11.5 mm. A the distribution of the load along each bolt is monotonically
Young’s modulus of 10 GPa is assumed and the material is con- decreasing, as expected, either side of the neutral point. Out-
sidered to be either elastic-perfectly plastic, with cohesion and side the plastic zone, it is clear that the bolts practically cease
friction angle 7.0 MPa and 20◦ respectively, or elastic-brittle to carry load.
plastic with the residual cohesion being reduced to 10% of the In the three-dimensional model series, the rock bolt instal-
peak. lation is simulated using the FLAC3D code. The influencing
The in-situ stress field is considered hydrostatic with field area of the bolt in the rock is modeled as a prism and due
stress equal to 10 MPa.This value is applied at the outer bound- to the symmetry conditions only one quarter of the prism is
ary of all models. In all three model series, the simulation is simulated. The z-axis of the model coincides with the axis of
performed in five stages. In the first stage, the model is allowed the rockbolt which is simulated by 1 zone along its radius, 10
to reach an initial equilibrium state before any excavation. zones along its quarter circumference and 60 zones along its
length.
Table 1. Rock mass properties used in the numerical models. Three zones along the grout annulus are used to simulate the
grout. The first two zones near the bolt are assumed elastic and
Rock mass properties the zone close to the grout-rock interface elastoplastic with a
strain softening material model. Peak strength parameters are
σ ci (MPa) 50 the same as those used in the plane strain models, while the
mi 10 residual cohesion is reduced to 10% of the peak.
Ei (GPa) 55 In Figure 2 the results from the three-dimensional models
GSI 90 70 50 30 show that when the quality of the rockmass is low, the axial
mb 6.997 3.425 1.677 0.821 load along the bolt is higher and that in very good quality
s 0.3292 0.0357 0.0039 0.0004
a 0.500 0.501 0.506 0.522
rockmasses the rockbolts are practically inactive.
Erm (GPa) 52.8 40.3 8.55 4.48 Finally, two-dimensional axisymmetric models, such as
*σ cm /p0 2.84 1.41 0.87 0.55 shown in Figure 3, are prepared with the FLAC2D code.
The influencing area of the bolt in the rock is modeled as
*σ cm is calculated with the method of Hoek et al. (2002). a truncated cone.
573
Figure 2. Axial load along the rock bolt for various values of GSI. Figure 4. Axial load along the rock bolt for various values of GSI.
The bolt, the grout and the rock mass are modeled using In Figure 4 the axial load along the grouted rockbolt for the
finite differences zones. The y-axis of the model coincides first model configuration is shown for the four cases of GSI
with the axis of the rockbolt which is simulated by one zone values. It is observed that the neutral point moves towards to
along its radius and 120 zones along its length. rock mass for lower rock mass qualities.
Two different model configurations are used. In the first To conclude, the maximum load in the rockbolt is evaluated
configuration, interface elements are used to simulate the bolt- to be significantly affected by the quality of the rockmass,
grout and the grout rock interface. These are assigned very being higher for a poorer rockmass. For very good quality
high values of normal and shear stiffness equal to 1000 GPa/m. rockmasses the rockbolts are practically inactive.
The grout is modeled as an elastic material with aYoung’s mod- Further, the bolts outside the plastic zone practically cease
ulus value Eg = 10 GPa. The bolt-grout interface is assigned a to undertake load. This illustrates their rather ground rein-
Coulomb slip criterion with 7 MPa cohesion and 20◦ friction forcement action, than any suspension from deeper more
angle. The grout-rock interface is given the same properties competent strata.
as the rockmass (using equivalent Mohr-Coulomb strength Finally, the neutral point of any rockbolt is found to move
parameters). In the second model configuration, interface ele- deeper with decreasing the quality of the rockmass. Moreover,
ments are not used. The two zones of the grout near the bolt this point does not go much deeper than half a metre for most
are assigned an elastic material model. The third zone near the of the cases examined. Such findings are in accordance with
rock is assigned the properties of an elastoplastic brittle mate- in situ measurements.
rial, with peak strength parameters cp = 7.0 MPa, ϕp = 20◦
and residual strength parameters cr = 0.7 MPa, ϕr = 20◦ .
574
M. Cai
School of Engineering, Laurentian University, Sudbury, Ontario, Canada
P.K. Kaiser
CEMI, Centre for Excellence in Mining Innovation, Sudbury, Ontario, Canada
Keywords: rock support; bolts and anchors; rock bursts; dynamics; tunneling; mine design
1 INTRODUCTION
575
576
Current developments in theoretical calculations of stress surface profile geometry indicate a significant variation in
distribution within the resin encapsulating roof bolts in under- peak bolt performance. Both profile height and spacing have
ground mines are presented. Monitoring of steel bolt perfor- been found to affect the load transfer mechanism of bolts
mance indicate that, amongst many parameters, bolt profile (Aziz et al., 2003, 2008). In particular, and as shown in
plays an important role in load transfer between the bolt and Figure 1, wider spacing of 37.5 mm bolt was found to be the
the surrounding strata. To optimise the bolt profile geometry most effective profile in bolt resin encapsulation interaction
a better understanding of the failure mechanisms, that occur when compared with similar height profile bolts of different
within the resin and bolt interaction, must be achieved. The spacing of 12.5 mm, and 50 mm bolts respectively.
short encapsulation pull out tests of roof bolts with various The analytical calculations supported by the numerical
modelling are used here to predict a probable mode of resin
failure while the laboratory and the in situ short encapsula-
tion pull out tests enable direct comparison to the theoretical
investigations.
Particular emphasis was directed to the study of:
• The stress distribution in infinite elastic media,
• Modelling of fully grouted bolt profiles,
• Stress distribution in the resin, and
• Mohr-coulomb failure study along the plane of weakness.
The important outcome of the study was to show that there
is another way to examine resin failure for different pro-
file configurations that can be compared with the laboratory
tests and numerical modelling. The ultimate aim and objective
of the theoretical predictions is to evaluate and complement
the experimental studies, enabling research to the best profile
geometries and to reach the optimum load transfer between
the bolt and the surrounding strata.
Figure 1. Pull out tests of bolts with various profile configuration Figure 2. A schematic drawing of a single spacing between two bolt
and spacing. profiles showing parameters used in calculations.
577
Layout assessment of underground tunnel for the stability of gangway in non-coal mine
Sung O. Choi
Department of Energy & Resources Engineering, Kangwon National University, South Korea
A lot of non-coal mines in Korea are planning to change first stage of its development. Furthermore, the gangway in
their mining method from the open-pit system to the under- non-coal mine have become bigger and more irregular, so the
ground system, not only to reduce expenses for environmental stability of gangway have become poor with the stress concen-
services but also to resolve an economic problem by low tration on the corner of safety pillar and the shear displacement
productivity. In underground mine operation, however, the along the rock blocks in safety pillar.
essential mechanical issues lie in keeping the stability of gang- In this study, the parametric analysis has been performed
way, and therefore an optimal design of gangway is most for the optimal design of safety pillar in underground non-coal
important to secure a mine stability as well as its productivity. mine. For this purpose, the theoretical solution by Obert and
For the optimal design of gangway space, the rock mass Duvall (1967) has been evaluated numerically, and the numeri-
classification method for mine has been adopted in this study. cal analysis with changing the width and height of safety pillar
It was modified from the conventional rock mass classifica- has been done to consider the correlation of gangway stability
tion method, considering the stand-up time and the equivalent with the space of safety pillar. Through the numerical anal-
support ratio in mine gangway. Because unlike most trans- ysis, the optimal space of safety pillar has been decided to
portation tunnels, mine gangway should reflect its direction minimize the stress concentration phenomena at the corner of
change, duration time, number of open space, dimension gangway.
change, and so on. The rock mass classification for mine can Figure 1 shows the numerical model used in this study. To
be suggested by RMR and Q , which are the simplified forms compare results with the theoretical solution by Obert and
of RMR and Q, respectively and namely elements related to Duvall (1967), the UDEC-2D program has been adopted for
rock mass only are considered neglecting the direction of this numerical model. The dimension of numerical model is
discontinuities and gangway. From the field investigation for 10 m × 10 m in case of single opening, and 20 m × 10 m in
two limestone quarries, the minimum width of unsupported multiple openings. In every numerical case, the diameter of
span was obtained by Eq. (1). opening is 5 m and the roller boundary condition has been
applied to the left and right boundary. Also the stress bound-
ary condition has been applied to the top and bottom boundary
with applying the 2 MPa compressive stress on the top bound-
where W is the minimum width of unsupported span in meter, ary. The numerical model was assumed to be elastic, and two
and RMR is basic RMR. joint sets were reflected in direction of 45◦ and 135◦ counter-
Also, the correlation of RMR with the stand-up time was clockwise from the horizontal line. The mechanical properties
investigated because the total stand-up time without any for rock mass as well as joints are summarized in Table 1.
supports can act as a negative factor on the stability of
gangway.
578
Properties Value
579
This paper assesses anchorage performance of fully grouted not effective. These coupled with well-documented studies
bolts in different lithologies in two coal mines in Illinois on moisture sensitivity of coal seams associated shales led
using the Short Encapsulation Pull Test (SEPT). These tests to the current study whose goal is to assess the long-term
were done at two different times (0 and 120 days) after bolt performance of different bolts used as primary supports.
installation. Additional testing is planned after much longer A roof fall data base for the period January 1, 2005 through
periods (180 and 270 days). A borehole was also drilled May 15, 2010 was also developed for the two cooperating
in each test area about 2.6 m above the coal seam prior to mines. At Mine 1 most falls occurred in 4-way intersections.
bolts installation and examined with a camera to document Most falls had stand-up time of less than 60 days or greater
changes in de-laminations. Load cells monitored changes in than 120 days. Most falls with stand-up time of over 180 days
few selected bolts. In addition to the Grip Factor (GF), anchor- were supported by 1.8 m long fully-grouted, passive #6 rebar,
age stiffness data is also developed that provides interesting Grade 40 bolts. Lower number of falls for the same time
insights. Anchorage stiffness is defined as the slope of the period supported by 1.2 m long fully-grouted passive bolts
load-displacement curve at 50% of the bolt anchorage capac- was most likely due to the presence of substantial limestone
ity or “Anchorage Stiffness around 50% of the failure load or in the immediate roof. Mine 2 experienced most falls in 3- and
AS (50)”. GF values for all lithologies are similar to those 4-way intersections in isolated areas of transitional lithologies
reported in literature. They typically increased after time. and they occurred between 0 and 90 days or after 180 days.
Several findings here corroborate the observations in the field Very few falls occurred where limestone was present within
on roof falls. 1.8–2.5 m above the coal seam.
The principal author analyzed the US Mine Safety and The results of SEPT show that 1) # 6 rebar installation in
HealthAdministration (MSHA) reportable roof-falls data base 35 mm boreholes should be avoided; 2) Longer bolts seem
of 642 falls for Illinois mines for the period 2004 to 2008. to result in lower GF values, 3) AS values vary much more
The data accounted for about 100% of the underground coal than GF values; 4) Authors think that performance testing
production. Although normalization of data was not possible, of grouted bolts should be based both on GF and AS val-
the relevant observations and conclusions made were: 1) Fall ues because AS values represent the rock mass reinforcement
height is typically larger than bolt length by 0.3–0.6 m for potential while GF values represent the load carrying poten-
all types of bolts used, 2) About 35% of the falls had stand- tial, 5) Mine 2 primary roof support should be reassessed since
up time less than 90-days, and 40% less than 120-days, 3) the AS values for larger size bolts with longer grouted length
Fully-grouted resin bolts may be more effective than other are similar to those for smaller size bolt and smaller grouted
bolt types, 4) Longer fully-grouted bolts (beyond 1.8 m length) length at Mine 1. Additional data will be developed over the
should be avoided, and 5) Point-anchored bolts were generally next six months.
580
M. du Plessis
Lonmin Platinum, South Africa
D.F. Malan
Department of Mining Engineering, University of Pretoria, South Africa
581
582
L.J. Gardner
Mining Projects Department, Impala Platinum Limited, Rustenburg, South Africa
Tunneling in South African platinum mines has traditionally while reducing their exposure to unsupported rock by using
been conducted using hand-held drilling and blasting meth- areal coverage with the tendon support. While presenting a
ods, with tunnels being supported using tendons installed minimum of technical detail, this paper attempts to explain
on at regular intervals. Rocks can and do fall out between how Impala has continually optimized its tunnel support sys-
the individual tendons, often with fatal results. To reduce tem to reduce rock-related risk in a challenging environment.
levels of rock-related risk in line with a “Zero Harm” pol- Although this story is by no means unique, it provides an exam-
icy, Impala Platinum has adopted a two-pronged strategy for ple of how such changes should be ongoing, and the resulting
mining project tunnels. This includes limiting the number of improvements that can be brought about in both safety and
personnel in tunnels by the increased use of mechanization, productivity.
583
Keywords: coal mines, case studies, field measurements, mine design, monitoring, numerical modelling
In Indian coal mines, blasting gallery method of extraction lower seam extraction leading the upper seam extraction by
has emerged as a very successful method of extraction for 30–60 m. With this technology the simultaneous extraction of
thick seams above 7 m. The paper deals with an innovative the two seams with upper seam using blasting gallery method
technology where two seams were extracted simultaneously and the lower seam by bord and pillar method of depillaring
with the upper seam by blasting gallery method.The two seams with stowing were tried at one panel.
exist with a parting of 9 m. The upper seam is 10 m thick and While extracting by this method the panels were instru-
the lower one is 3.5 m thick. The upper seam was extracted mented with different rock mechanic instrumentation like
with caving using blasting gallery method.The lower seam was multipoint borehole extensometer, load cell, convergence sta-
developed by bord and pillar method and it was experiencing tions etc. The instruments were monitored regularly and data
severe ground control problem due to the blasting and caving were analysed. The critical convergence value for predicting
effect of the upper seam. An attempt was made to extract the the roof fall of the blasting gallery panel could be determined
lower seam with stowing and extraction of upper seam by experimentally. All predictions related to the stability of the
blasting gallery method over the stowed pack of lower seam. upper seam and lower seam extraction could be made well in
But due to sub-surface subsidence cracks were generated at advance based on the instrument data.
the floor of the upper seam and it became extremely difficult The method has emerged as a very successful method for
to extract the upper seam using blasting gallery method. As a the mine and all nearby mines have accepted this method as
result lot of coal were lost. the sole method for high percentage of extraction in similar
Then an innovative technology of simultaneous extraction condition.
of both the seams were thought of. A thorough numerical The paper deals in detail the method used, the detailed anal-
modeling analysis using FLAC 3D software were attempted. ysis of data obtained from numerical modeling studies and the
It has been observed that the extraction is possible with the analysis of data obtained from rock mechanic instrumentation.
584
Keywords: back analysis, bolts and anchors, field measurements, lab testing, numerical modelling, rock support
The study deals with the interpretation of in situ pull-out tests argillite/grout and grout/bolt interfaces. In this first attempt,
on instrumented grouted rock bolts, which took place at the the argillite was modelled with a simplifying Mohr-Coulomb
ANDRA underground research laboratory of Meuse/Haute- constitutive law. The elastic behaviour of the interfaces is char-
Marne (France). It aims at characterizing the behaviour of a acterized by a normal and a shear stiffnesses, while the plastic
single rock bolt in argillite, taking into account both the differ- behaviour, of Coulomb-type, is characterized by a cohesion,
ent materials (bolt, argillite, grout) and the different interfaces a friction angle and a dilatancy angle. Both the grout and
(argillite/grout and grout/bolt). 6 tests were carried out in situ, the bolt were considered as elastic materials, on the basis of
on two types of bolts: 3 tests on HA-25 rock bolts, and 3 the lab tests on grout. The whole sequence of excavation of
tests on CT-Bolt rock bolts. For each kind of rock bolt, one the gallery, then drilling, installation and grouting of the bolt,
short term and two long term pull-out tests were performed. and of the pull-out test is taken into account in the numerical
The paper focuses on the interpretation of the long term tests. modelling.
During these tests, the end of the bolt is pulled with different Even if the orders of magnitude of experimentally mea-
successive load stages, each load being maintained constant sured and calculated displacements and strains are similar, the
during various times. Two different loading paths were con- numerical modelling did not allow to carry out a precise quan-
sidered in this work. During the tests, the displacement of the titative interpretation of the in situ tests so far, that is to say that
head of the rock bolt is classically measured. In addition, in the parameters set that best fits the experimental curves was
the tests considered in this work, the bolts were instrumented not identified until now. Nevertheless, these numerical mod-
with axial strain gauges, fixed to the rock bolts every 15 cm, ellings allow a qualitative interpretation of the tests, enabling
so that the axial strains are recorded during the whole test (as the understanding of a part of the experimental observations.
well as before the test) along the grout/bolt interface. In addi- The elastic properties of the grout/bolt interface seem to be
tion to the in situ pull-out tests, laboratory tests were carried of fundamental importance. The quantitative analysis should
out in order to characterize the mechanical behaviour of the then focus on these parameters. The quality of the grout seal-
grout. ing seems to be heterogeneous, with a rather good quality in
The analytical models developed by several authors (Cai the far end of the drilling (from where the grout is injected)
et al., 2004; Farmer, 1975; Li and Stillborg, 1999; Ren et al., and a rather poor quality at the head of the drilling, maybe due
2010) were used to try to interpret the in situ tests but did not to a more intensive fracturation due to the development of the
allow to explain the experimental results, due to several simpli- EDZ at the wall of the gallery.
fication hypothesis. So, numerical modelling was considered From these results, the future developments should focus on
to interpret the tests, with the finite differences code FLAC3D the numerical modelling, and particularly on the calibration of
(Itasca). We made use of three-dimensional modelling in order the parameters of the grout/bolt interface. Moreover, labora-
to take into account all the possible geometrical configura- tory tests should be performed to improve the characterisation
tions (respective orientation of the transverse isotropy planes of the two interfaces, which could lead to the development of
and the principal stresses, orientations of the bolts and the a more suitable constitutive law. In parallel, a more appropri-
strain gauges). The numerical modelling takes into account ate constitutive law for the argillite will be considered, on the
suitable constitutive laws for the bolt (steel), the grout (charac- basis of the knowledge acquired on this material from other
terized through the lab tests) and the argillite, as well as for the numerous studies on it, among others to account for the EDZ.
585
A. Russo
CODELCO (National Copper Corporation), Santiago, Chile
C. Carranza-Torres
Department of Civil Engineering, University of Minnesota, Duluth Campus, MN, US
586
Keywords: rock mass; rock support; bolts and anchors; case studies; monitoring
587
588
Keywords: coal mines, problem rocks, bolts and anchors, case studies, stability analysis
1 INTRODUCTION theory has been proposed. The main functions of bolting are
recognized as controlling the separation and slippage of sur-
Most coal mines in China operate by underground mining, rounding rock in bolted zone, crack opening and development
the geology and production conditions are complicated in associated with dilatant displacement, limiting the discontin-
many coal mines. The total length of roadways driven per uous displacement, and keeping the integrity and continuity
year amounts to 10000 km in Chinese coal mines, and 80% of of surrounding rock, making it in the compressive state, and
those are driven in coal seams. reducing the strength loss of rock mass. The pretension on
With increasing mining depth and extending mining area, bolts and its spreading to surrounding rock play a key role, and
the proportion of roadways with complicated conditions surface protecting parts are very important for the spreading of
becomes higher and higher, which poses a tough challenge bolt pretension. There is a critical reinforcement stiffness for
to ground control, and makes greater and strict demands pretension bolting system. For complicated roadways, the high
on reinforcement techniques. Therefore, reliable reinforce- pretension and intensive bolting system should be applied to
ment techniques for roadways with complicated conditions control the deformation and damage of rock mass once and for
have a special significance for the safety and high efficient all, avoiding the secondary reinforcement and maintenance.
production of coal mines.
4 REINFORCEMENT TYPES FOR COMPLICATED
2 TYPES OF COMPLICATED ROADWAYS ROADWAYS
Based on the natural and physical properties of surrounding There are many reinforcement forms for complicated road-
rock, the stress states in it, and the production conditions, ways. In recent years, the high pretension and high strength
complicated roadways in Chinese coal mines can be divided reinforcement system with bolts and cables as basic parts and
into 8 types: sometime complemented by grouting, has been the first choice
for complicated roadways.
1. Roadways with soft and swelling rock;
2. Roadways with extremely fractured rock mass;
3. Roadways with high stresses in deep mines; 4.1 High pretension and high strength bolting techniques
4. Roadways with large section; A series of high strength threaded steel rods used for bolts
5. Roadways affected by violent mining activities; were developed (see Tab. 1) to meet the needs of the compli-
6. Roadways with coal roof or all around the coal seam; cated roadways. The structures and shapes of the rods were
7. Roadways driven along next gob; optimized to improve the anchorage ability of bolts; and spe-
8. Roadways retained along gob-side or in the gob area. cial steel materials were developed to reach the level of high
or extra high strength. In addition, a series of resin capsules,
W-shaped and M-shaped straps were also developed to form
3 SUPPORT AND REINFORCEMENT THEORIES the high strength resin bolting system.
FOR COMPLICATED ROADWAYS
4.2 High pretension and high strength cable bolts
With the fast development of high strength bolting tech-
niques, the rock bolting mechanism for complicated roadways High extension and intensive cables were developed specially
has been studied deeply and extensively. The nature of rock for complicated roadways. The cables are made from 1 × 19
bolting has been recognized as improving the mechanical per- strands of steel wire, which increase the broken strength
formance of surrounding rock and the stress state in it, and and extension rate of cables substantially. To match with the
controlling the displacement and damage of surrounding rock. strength of rock bolts, the diameter of cables is divided into
It has been realized that it is very important to bolt surrounding three types: φ18 mm, φ20 mm and φ22 mm respectively. The
rock immediately after excavation, exert enough pretension broken load of φ22 mm cables amounts up to 550 kN, and the
on bolts and increase the stiffness of bolted body. Based extension rate is about 7%, obviously higher than that of 1 × 7
on these new insights, high pretension and intensive bolting strands. The cable pretension can amount up to 300 kN.
589
590
The predominant amount of European coal deposits exploited of lifetime of a mine working. The practical experience in
underground is extracted by the longwall method with con- the application of the combined supports leads to opinion that
trolled caving. Experience shows that roadways, which ensure even in in the conditions of disturbed rock mass roof bolts play
all transport and ventilation in coalfaces, restrict both out- a positive role, which can be expressed in the methodology of
put and safety, as well as economy in coal production. The determining the dimensions for standing supports.
problem is urgent especially in deep mining. The contribution The suggested and simple geo-mechanical model permit to
describes procedures for designing the supports and stabiliza- determine approximate dimensions of the combined support
tion measures using combined support in the already planned in a long mine working, so as the standing support is designed
roadways on the base of system developed in the Ostrava with regard to the favourable residual co-effect of roof
Karvina Coalfield (OKR) in Czech Republic. bolting.
The combined support in roadways combines roof bolting From the calculations carried out, it is apparent that the
and standing (conventional) supports. The calculation of a so called bonus of bolts with the application of the com-
load bearing capacity of roadway supports contains period of bined support – i.e. the reduction of the normal load of the
drivage and period of minig in the vicinity of roadways. The standing arch support, is 15–45% of the original load in OKR
load bearing capacity of supports during drivage is based on mine conditions. The most frequent values for load reduction,
the size of the natural arch, which comes into existence in the corresponding to the width of 5–6 m of a mine working, vary
vicinity of a roadway. between 20–30%.
The supports in roadways are dimensioned to the load The effect of roof bolting, i.e. the reduced requirements for
applied during the roadway’s life. In the first period of drivage bearing capacity for the standing support, rises with a reducing
of a roadway, the minimum bearing capacity of the supports in width of a mine working.
must correspond to the load of the loosened rock in its vicin- The greater efficiency of roof bolting is shown in condi-
ity or, as the case may be, to a portion of this load, which the tions of firm overburden, which is usually associated with the
support must transfer with respect to a certain coherence of manner of disturbance of rock mass – the firmer rock mass
the loosened rock. The input data for the calculation method is is disturbed in larger clusters, which can be better stabilized
deduced from external measurements of convergence and dis- with roof bolting.
placement in roadways. The loading of supports during mining Monitoring, together with the behaviour of the combined
comes out from a stress wave in rock mass in the forefront of support, now more frequently installed in OKR mines, primar-
the coalface and from the caving area of the mined out panel. ily in connection with the dual application of roadways, will
The great advantage of the application of the combined enable further verification and obtaining more precise results
supports is the smaller convergence in the important period from the theoretical solution presented.
591
Keywords: coal mines, risks and hazards, weak rock, problem rocks
Acid Mine Drainage (AMD) generation from sulfide min- The XRD analysis conclude that the samples consist of clay
eral containing rock constitutes a significant environmental mineral such illite and kaolinite as well as secondary mineral
problem in coal, and poly-metallic sulphide mine due to the such as ferrihydrite, jarosite (KFe3 (SO4 )2 (OH)6 ) and gypsum
generation of sulfate and iron salts and possibly mobilizing (CaSO4 · 2H2 O), while Pyrite (FeS2 ) and siderite (FeCO3 ) are
trace metals present in the rock. AMD generation rate were not detected. XRF analysis result shows that the sample has
depend on many factor, both of inherent stoichiometry of acid high sulfur content as much as 3.6% and low alkali content
and base mineral and external factor such as climate which which is concordant with the result of static geochemical test
make the AMD characteristic would be site specific. (paste pH, paste EC, Acid Neutralizing Capacity test, and also
Water plays an important role in acid mine drainage, other- Net Acid Generating test).
wise as reactant as well as contaminant carrier (Bussière et al. Five columns with diameter 50 mm and 150 mm in height
2004, Hollings et al. 2001). On non-durable slaking rock mate- were used in the experiment. Amount of 150 gram sample with
rial such as coal measure rock, the existence of water in pore certain grain size distribution (d10 = 0.2 mm, d50 = 0.85 mm)
space will control the oxygen diffusion, slaking, and further- was used. Leachate from the columns following flushing were
more the reaction rate, both of sulfide oxidation and alkali dis- collected and sampled which was then measured for pH, EC
solution, of mineral contained in rock (Gosselin et al. 2007). and Eh as well as major sulfate and metal content such as
At the surface layer of uncovered waste rock dump, rocks calcium, magnesium, sodium, potassium, iron, aluminum and
will interact with climate components involving rainfall, solar manganese.
light, temperature and humidity ambient that will control the As result of various water pouring interval, each column
weathering and water balance in the surface material. With has different water content fluctuation behavior. In general,
the assumption that the average temperature and humidity are maximum water content of each cycle in each column was
relatively same, the interval of rainfall will be the biggest decreased by time due to self compaction of rock sample in
influence that will determine the wetting and drying cycles regards with re-packing of the loose grain which can observed
of material, and furthermore will affect to the weathering rate by the increase of the bulk density.
and water balance of material. In relation with pouring interval, certain interval of water
This study is intended to investigate AMD release behavior pouring has resulting in lower pH of leachate water while
of sulfide mineral containing rock with different water content the EC has the opposite order. In addition, the reduction –
condition affected by different water pouring interval by using oxidation potential (Eh) were increase along the experiment
a column leach test in the laboratories. cycle for all sample meant the increase of the oxidation-
In this study, we used sample which taken from a coal reduction activity. Similar to the pH, sulfate and metal content
mine, PT Berau Coal, which is located in Berau Regency, (Al, Mn and Fe) were higher in the column with certain (one
East Kalimantan Province, Indonesia. Sample was identified days interval) pouring interval.
as partially weathered conditions since it was taken from the This experiment result suggests that acid generation is
surface of existing uncovered waste dump with more than one controlled by several factors that influence one another, both
year old. physically and geochemically.
592
L. Alejano Monge
University of Vigo, Spain
The Aguas Teñidas mine is located in the municipality of exceeds 600kPa. The remaining properties obtained in the
Almonaster La Real, Huelva. laboratory are, on average, as follows:
The mining method uses large sublevel stopes and fan blast- • Elastic module, E = 2–30 MPa
ing. The stopes are 20 m wide, 30 m high and between 20 m • Poisson’s ratio, µ = 0.2–0.4
and 60 m long. The method used to exploit the area consists • Density, 2400–2860 kg/m3
in extracting the primary stopes with a height of 30 m, and • Permeability, 2.38–4.83−6 cm/s
then extract the secondary stopes, of the same size, located
between the primary ones, after these have been filled with A FLAC 3D software programme was used for the numeri-
paste. Paste is a mix of mill tailings and cement that can be cal modelling. The model was constructed using paralepipedic
considered as non-segregating slurry which means that it has blocks of the same length and width, both variable, and a con-
negligible excess water when stationary and remains essen- stant height of 2.5 m. The total number of blocks is 36864.
tially as a homogeneous single phase product (Henderson, A. Two primary stopes have been successively excavated and
et al., 2005). filled into the model, each of which has been exploited in
First of all the ore of the first primary stope is fun blasted three sections. There is a secondary stope, located between
and extracted, afterwards the stope is filled with paste; when the primary ones, which has also been exploited in three sec-
the stope is too long so that the stability of its back can not tions. The height of the primary and secondary stopes is 30 m.
be guaranteed, it is convenient to extract and fill it in several The width varies between 12 m and 28 m and the length varies
sections of shorter length; every section must be filled before between 36 m and 84 m.
the next one is blasted. The other primary stope is extracted in The model has been firmly secured to its base, while the
the same way. In the primary stopes the walls are in ore and sides have only been fastened in the direction perpendicular to
in the secondary ones are in paste. the face. Before emptying and filling the stopes in the model,
Paste fill is a mix of mine tailings, with between 78 and the prior stresses were introduced and were assumed to be
85% solids by mass, and 3 to 7% cement. It uses some water hydrostatic, i.e. equal in vertical and horizontal directions.
for cement hydration and retains the rest within its matrix, so Extraction and filling of the stopes was carried out in 18
that it will not create any bleed water. phases as defined below:
The stability analysis of paste fills has been undertaken Phase 1: Extraction of first primary section.
by several authors. Mitchel et al. (1982, 1988) developed a Phase 2: Filling of first primary section.
three dimensional limit equilibrium solution for the stability Phase 3: Extraction of second primary section.
of exposed vertical face. Winch (1999) proposed an analytical Phase 4: Filling of second primary section.
solution to the total vertical stress within a three dimensional Phase 5: Extraction of third primary section.
back fill mass. Bloss (1992, 2001) used TVIS modelling pack- Phase 6: Filling of third primary section.
age to study the underground stability of cemented hydraulic Phases 7 to 12: Extraction and filling of primary sections, 4,
fills at Mount Isa mines. Numerical modelling with FLAC3D 5 and 6.
was used by Rankine et al. (2001, 2003, 2004) to examine the Phases 13 to 18: Extraction and filling of secondary sections,
stability of back filed stopes. 1 to 3.
Trials have been carried out using paste samples made with
The properties allocated to the paste are as follows:
flotation tailings and quantities of cement, usually varying
between 3% and 7%, with a slump between 7 and 10 . • Cohesion, C = 300 kPa (different cohesions have also been
Uniaxial compressive strength of the paste, after 28 days, applied)
593
594
Keywords: stability analysis, mine design, mining, numerical modeling, surface protection, undermining
1 INTRODUCTION Pipes. For this study, the categories for buildings will be con-
sidered. Stacey and Bakker (1992) further suggested that of
1.1 Background the available parameters such as vertical subsidence, horizon-
tal strain, tilt, radius of curvature and differential settlement,
Union Mine is located in the Limpopo Province of South
only two (horizontal strain and tilt) should be adopted for
Africa on the western limb of the Bushveld Complex. Two
subsidence guidelines. The reasons that Stacey and Bakker
reefs are currently being extracted, namely the Merensky reef
(1992) suggests for the choice of these two parameters is
and the UGII reef. The current mining depth extends from
that they are both readily measured, both can easily be cal-
surface to approximately 1.5 km below surface. The plan is to
culated from mining simulation stress analysis programs and
extend mining to a depth of approximately 2 km below surface.
that they have been commonly used. The horizontal strain and
A portion of the UG2 reef has been left un-mined at the
tilt criteria limits suggested by Stacey and Bakker (1992) were
Declines Section, Union Mine, Anglo American Platinum,
used for this analysis. A conservative approach was adopted
possibly because of the overlying surface structures includ-
for this study by using the Building Category 1 superficial
ing buildings (closest being an engineering workshop) and
damage limits.
railway tracks. Union Mine is planning to mine the UG2 reef
in the vicinity of the existing surface infrastructure.
2.2 Pillar design criteria
1.2 Objectives The Factor of Safety (FOS) approach for pillar design is well
The primary objectives of the study are to ensure that the stabil- established for South African hard and soft rock mines and is
ity of the underground excavations and surface infrastructure not repeated in this report. For normal operating conditions
is maintained and in so doing ensure the safety and health of (i.e. no protection of surface infrastructure required) it is typ-
workers and to satisfy the requirements of the aforementioned ical for mines to use a FOS = 1.6 for pillar design. When
legislation by: designing pillars in workings where subsidence control is a
consideration, higher FOS’s are advocated since:
(i) Determining the optimum extraction depth below the
surface structures with due cognisance taken of: – The life of the surface infrastructure typically far exceeds
that of the underground work space
– damage to any of the surface infrastructure (workshop – The pillars are required as far as possible to be permanent
and railway tracks), and unchanging with the passage of time
– the ability of the systematic pillar system to maintain – The decay of pillar strength with time must be considered
its integrity for the length of period that subsidence – The influence of substandard mining controls such as
control is required, drilling and blasting on FOS’s must be considered
– permissible mining spans to ensure appropriate pillar
loading and tolerable surface settlement. A FOS of 2.5, suggested by Bakker (1992) for ‘buildings or
places where people congregate’ was used for this analysis.
(ii) Recommending procedures to be adopted prior to extrac-
tion which will include but not be limited to:
3 UNDERGROUND GEOTECHNICAL ASSESSMENT
– instrumentation and monitoring,
– pillar and panel design. The rock mass rating system (RMR1976 ) introduced by
Bieniawski (Bieniawski, 1976) and Norwegian Geotechnical
Institute’s Q-system (Barton et al., 1973) was applied to clas-
2 CRITERIA FOR ASSESSMENT
sify the rock mass conditions. No major geological structures
were mapped in this area. The RMR1976 rating for the rock
2.1 Structural damage criteria
mass conditions observed in the ASD was approximately 90,
Stacey and Bakker (1992) described deformation limits for which could be described as Very Good. The Q-rating for the
four main groups of surface structures applicable in the South rock mass conditions observed in the ASD was approximately
African mining context as Buildings, Roads, Railway lines and 180, which could be described as Extremely Good. The ASD
595
Where h is the pillar height and weff the effective pillar width. REFERENCES
Assuming a UGII strength value of 75 MPa, a stoping width
of 1.5 m and a pillar width of 5 m, the pillar strength equate to Anon. 2000. UDEC User Manual. ITASCA Consulting Group, Inc.
∼41 MPa. The FOS, in turn is simply the ratio between pillar 1st Edition, Minnesota, USA.
strength and average pillar stress (APS). Bakker D. 1992. The undermining of surface structures and construc-
tion/erection over undermined ground. Symposium on construc-
tion over mined areas. Pretoria. South Africa. 1992.
5 SUMMARY Barton, N., Lien, R. & Lunde, J. 1974. Engineering classification of
rock masses for the design of tunnel support. Rock Mechanics. 6:
The analysis indicates that a FOS design limit of 2.5 for pil- 4: 189–236.
lars is achieved for mining depths down to a depth of 105 m
below surface, for a 5 m × 10 m pillar system and mining spans
of 30 m.
596
D.F. Malan
Department of Mining Engineering, University of Pretoria, South Africa
1 INTRODUCTION
Historic examples of monitoring in mines are useful as it illus- Figure 1. An area of a collapse in a platinum mine. Note the failed
trates that the early miners realised the value of measurements elongate support.
and that they were able to implement monitoring programmes
without the sophisticated electronic systems available today.
The earliest recorded example of in-stope monitoring in the 4 CASE STUDY ILLUSTRATING THE VALUE OF
South African mines was the use of “sag” meters at Crown MONITORING
Mines in 1927. The meters were originally manufactured from
timber. These were installed in all remnant areas. Systematic A case study is included in the paper to illustrate the value of
measurements of convergence (“sag”) were conducted at the monitoring programmes. A Platinum Mine in the Marikana
mine and the data was recorded on special forms. As described area of the Western Busheld Complex recently experienced
in the paper, other measurements of convergence were also a number of large panel collapses. The workings are rela-
conducted by various workers. tively shallow and current mining focuses on the UG2 reef
horizon. A number of monitoring sites were established at the
mine and important observations regarding the large collapses
were made. It should be noted that these observations are not
3 THE NEED FOR MONITORING applicable to small falls.
Large collapses are often preceded by time-dependent
Ryder & Jager (2002) summarises the need for rock engineer- deterioration of the hanging wall and an increase in the rate
ing monitoring as follows: of convergence. Broken elongate support (Figure 1) has tra-
“In geotechnical engineering work, there is inevitably a dis- ditionally served as early warning of the collapses, but these
crepancy between prediction and performance. …The instru- elongates only fail after approximately 30–50 mm of conver-
mentation and monitoring of real underground behaviour is, gence. The falls do not occur instantaneously, but appear to
therefore, an essential component of the rock engineering be preceded by a lengthy period of deterioration.
discipline.”
4.1 Panel convergence measurements
Even after many years of research, there are still a large num-
ber of unknowns in rock engineering. Some of the issues are Elongate deformation is not a very useful means of providing
described in the paper and these are pillar strength, under- early warning of the instabilities as the rate of convergence
ground de-rating of elongate strength and determination of preceding collapses at the mine was very small. Too much
stable spans. movement needs to occur before the support gave a clear
597
Ryder, J.A. & Jager, A.J. 2002. A textbook on Rock Mechanics for
5 THE DEVELOPMENT OF CONVERGENCE tabular hard rock mines. SIMRAC, Johannesburg.
TELLTALES
598
1 INTRODUCTION
2 THE EFFECT OF WEAK PARTINGS ON PILLAR Figure 1. Scaling facilitated by a weak contact at the pillar/hanging
STRENGHT wall contact.
599
Hedley, D.G.F. & Grant, F., 1972. Stope-and-pillar design for Elliot
Lake Uranium Mines. Bull. Can. Inst. Min. Metal., 65, pp 37–44.
Malan, D.F. 2010. Keynote Lecture – Pillar design in hard rock
mines: Can we do this with confidence? In: Hagan, P and Saydam,
S., (eds.) Proc. Second Ground Control in Mining Conference,
Sydney, November 2010, pp 15–30.
Martin, C.D. & Maybee, W.G., 2000.The strength of hard rock pillars,
Int. J. Rock Mech Min. Sci, vol. 37, pp 1239–1246.
Napier, J.A.L. & Malan, D.F., 2007. The computational analysis of
shallow depth tabular mining problems, J. S. Afr. Inst Min Metall.,
vol. 107, Nov 2007, pp 725–742.
Ozbay, M.U., Ryder, J.A. & Jager, A.J., 1995. The design of pillar
Figure 3. ESS values on a plane at constant 400 m depth as a systems as practiced in shallow hard-rock tabular mines in South
function of dip with a friction angle of 20◦ . Africa, J.S. Afr. Inst. Min. Metall., Jan/Feb 1995, pp 7–18.
600
Current development in roof bolting automation requires fast of integrity where the polymer was not damaged. The type of
and effective skin support of underground coal mine roadways. failure indicated a substantial toughness and strength of the
A strong and tough fibre reinforced polymeric thin spray-on TSL material.
liner (TSL) is emerging as a logical substitute for the old steel
mesh support system. This paper investigates the effectiveness
of a reinforced polymer skin liner in coal mines subject to high
horizontal stress. Investigations using numerical modelling
and laboratory experiments suggest that a polymer based TSL
can outperform steel mesh and significantly improve stability
of the strata skin between the bolts. Strength tests performed
on 5 mm thick polymer sheets indicate load bearing capacities
comparable to the heavy duty steel mesh currently used as a
skin support underground. Puncture tests and polymer adhe-
sion to the substrata indicate that the polymer TSL can provide
additional benefits to the skin support.
A puncture test was performed using a series of steel discs to
predict whether the standard steel bolt plate system that is cur-
rently used underground does not place excessive loads onto
the polymer TSL. Discs of various diameters were placed onto
a 5 mm thick fiber reinforced polymer sheet and loaded until
the polymer surface was compromised. The summarised test
results in Figure 1 show that a 5 mm thick reinforced polymer
sheet can withstand substantial loads, e.g. the typical bearing
capacity of the polymer when loaded using a 120 mm disc was
up to 3300 kN before permanent damage occurred. The plates,
which are commonly used in the underground bolting system,
yielded the polymer sheet at approximately 3300 kN a load Figure 1. Summary of the load bearing capacity tests using round
comparable to the 120 mm diameter steel disc load shown in discs and steel plates.
Figure 1.
In order to determine the ultimate strength of the reinforced
polymer a series of tests were performed to load the poly-
mer TSL sheets to failure as shown in Figure 2. During the
first test the ultimate load was not reached due to the com-
pressibility of the airbag that was used for loading, absorbing
the 100 mm loading stroke of the Avery loading machine. A
load of 68 kN at a deflection of 35.3 mm was reached without
any visible damage to the polymer sheet. On unloading, the
sheet deflection recovered 23.9 mm with 11.4 mm of perma-
nent deformation, indicating that the sheet was mostly within
the elastic region of deformation during the loading stage.
Another similar test reached 100 kN without polymer failure.
Since the polymer sheet did not fail, the loading method was
modified and two sheets were reloaded using a 150 mm steel
spherical seat underlined with a dense rubber mat (Figure 2) to
minimise the stress concentrations at the plate edge. The sam-
ple was loaded at 2 mm/minute until the polymer failed at a
load of 45 kN and an average deflection of 52 mm in both tests.
As expected the load at failure was lower due to the smaller
loading area that was applied to the sheet. The location of
failure was directly under the point of loading. The polymer
sheet failed in a non-brittle manner with a tear and no loss Figure 2. Loading of glass reinforced polymer sheet to failure.
601
Rock slope reinforcement in open cut mining using grouted rock pins
Keywords: bolts and anchors, mining, problem rocks, mine design, rock failure
1 INTRODUCTION masses two symmetrical plastic hinges are formed; one above
and one below the slip surface. These hinges act to focus axial
Reinforcement of rock slopes is a common practice in both force in the bolt until tensile failure occurs in the bolt at some
civil and mining applications. The reinforcement methods point between the hinges. The failure is primarily axial with
employed are generally based on the fundamental mechan- little shear force evident. The other mode occurs in stiffer less
ics although the design goals may differ. These differences deformable rock and also softer weaker rock. In this mode,
are reflected largely by the required longevity of the design. axial and shear force act at the joint intersection. The bolt
Civil applications often require design lives of 50 to 100 years fails through shear. Research conducted on the effects of soil
whereas mining applications can be anywhere from a few nail inclinations on a reinforced soil mass suggests that certain
weeks to a decade or more. The difference in the designed inclination angles can be detrimental to the normal stress on
lifespan can greatly affect the chosen reinforcement scheme. the slip plane. This has direct relevance to rock masses as the
Current reinforcing schemes utilized in the mining industry mechanics involved are similar in application. The method of
heavily rely on empirical data and experience gathered by the reinforcement pin installation currently used at the Cadia Hill
engineers in charge of the design and as such do not always rep- Gold Mine results in pin inclinations with greater dip than the
resent the best economical and efficient solution. Research has intersecting discontinuity plane. With reference to the typi-
been conducted on a number of different factors that influence cal shear pin installations most encountered at the Cadia Hill
the performance and efficiency of rock slope reinforcement Gold Mine and the angles at which the reinforcing elements
using grouted rock anchoring system. Traditional empirical meet the discontinuity, the research has indicated that some of
approaches for rock reinforcement design are examined in these angles may lead to a reduction in strength of the shear
conjunction with case studies. plane due to the mobilization of compression forces within
the element. The compression forces act to reduce the normal
stress on the plane and it is through this action that the overall
2 DESIGN PRACTICE AND ANALYSIS resisting force of the system is reduced. It is generally agreed
that the reinforcement pin increases the strength of the system
There are two types of common methods for slope rein- through two primary mechanisms. These mechanisms are the
forcement: one involves the use of active grouted anchors dowel effect and the friction effect. This increase in strength
installed into the face and pre-tensioned and the other utilizes is dependent on rock deformability and surrounding material.
un-tensioned steel bolts or shear pins. Existing knowledge on The friction effect on the other hand is the contribution of the
the consequences of rock bolt inclination in relation to the slip pin to the normal force on the discontinuity plane.
surface is rather limited. The mechanism of the grout – rock –
steel interaction on the slip surface is prohibitively complex
for straight forward closed-form analytical solutions. Current 3 DISCUSSION
research indicates that the complexity of stress interaction
problem can only be analyzed by computer programs such The assessment of the shear pinning techniques used at the
as Finite Element or Discrete Element Modeling techniques. Cadia Hill Gold Mine involved the analysis of the different
A number of authors have had a various degree of success with factors and mechanisms that interact to provide the required
modeling the problem utilizing the Finite Element model in resistance to an unstable geological structure.
three dimensions. The models used different failure criterions The results produced indicated that compression induced
for steel, grout and surrounding material. within the reinforcing element during the early stages of mobi-
Computer models are also able to identify that the bond lization will lead to a reduction in the normal stress acting on
between the grout and the steel at the intersection of the slip the shear plane. This reduction in normal stress will weaken
surface broke at low values of load. The model was able to the resistive force generated from friction between the two
accurately predict the size of the gap between the grout and the shear surfaces. A design methodology for pullout failure of
steel with the laboratory testing confirming this effect without the pinning installations was developed. The developed rela-
exception. It is observed that the maximum strength at failure tions used to redesign an existing setup at the Cadia Hill Gold
is generally achieved in deformable and weaker rock masses Mine could potentially lead to approximate 25% saving in
due to the larger plastic deformations occurring, with greater required grouted steel. Further investigation will be required
steel ductility also adding to the effect. There are two possible to determine the need for such relation or if it in fact holds
modes in which the steel reinforcing element can fail once true for the types of rock encountered.
the mass has mobilized. In weaker and more deformable rock
602
Geotechnical design of the sublevel stoping method for the Poder and
Trabajo Veins in the San Genaro Mine – Peru
603
604
Numerical study on disturbance of longwall mining to stress state in the rock mass
605
Changshou Sun
Newmont Mining Corp, Elko, NV, USA
Keywords: Geotechnical data collection, Ground support system, Rock bolts, Shotcrete
The Leeville underground mine has been in operation since The value of ESR is related to the intended use of the exca-
2006 and currently produces 3176 tons (3,500 short tons) of vation and to the degree of security which is demanded of
ore per day with an average gold grade of 10 g/ton (0.32 opt). the support system installed to maintain the stability of the
The mine has a large horizontal extent of 762 m (2,500 feet) excavation. (Barton et al., 1974). Then, a support system is
by 152 m (500 feet) and a vertical extent of 6.1 to 61 m (20 to determined by a chart, the equivalent dimension De versus
200 feet). Based on the characteristics of the ore bodies, the the value of Q. Q values in Leeville are 0.1 to 4. The average
deposits are divided into several mining zones. The geotech- Q value is 0.4. The spans of drifts are between 4.5 to 12 m (15
nical properties of mining zones are also different from one to 40 ft). So, for short term excavations, De is 1.52 to 4.06 m,
zone to another. The typical ground condition at Leeville is and long term excavations, De are 2.86 to 7.62. All numbers
poor to very poor. So, an appropriate ground support system is are plugged into Grimstad and Barton’s chart. The support
key for the successful production at the Leeville underground system can be determined. The spacing of bolts is 1 to 1.4 m
mine. (3 to 4.6 ft) for both long term and short term structures. The
Based on the available geotechnical data, the empirical and shotcrete thickness is 2 inches for short term structures and
numerical methods are used to construct the ground support 50 to 100 mm (2 to 4 inches) for the long term structures.
system at the Leeville underground mine. The main ground Several types of rock bolts were used at the Leeville mine,
support system includes in the following. including Split Sets, Swellex, Resin bolt, Cable bolt, and MAI,
Geotechnical data collections, core logging, site investiga- a self-drive grout bolt. Based on the geotechnical investiga-
tion and extra geotechnical drills. Based on the characteristics tion in Leeville, the rock condition in Leeville is poor to very
of the ore bodies, the deposits are divided into several mining poor. Although the intact rock is hard, it is extremely fractured.
zones. The geotechnical properties of mining zones are also The average RQD of rock in Leeville is only 21. The ground
different from one zone to another. To classify and evaluate condition is weak. To select an appropriate support system,
the rock mass condition, a complete geotechnical data must the concept of the interaction of ground reaction curve and
be collected. support load-displacement characteristics is applied. As rock
mass in Leeville is very weak, a yield support system should
• Data collected by core logging. Most geotechnical data is
be selected. The yield support system can allow the ground to
collected by this method from the exploration drills. Param-
deform more and decrease the support pressure requirement.
eters of RQD (Rock Quality Designation), FF (Fracture
To select the appropriate rock bolt, the anchorage character-
Frequency), Compressive strength, and RMR (Rock Mass
istics of rock bolts are evaluated. Stillborg (1994) carried out
Rate) are estimated
a number of tests on rock bolts and dowels installed across
• Data collected by site investigation. More detail geotechni-
a simulated ’joint’, using two blocks of high strength rein-
cal data and geo-structure data are collected by this method.
forced concrete. Based on Stillborg’s test results, the Swellex
Parameters of RQD, FF, Compressive strength, RMR, Q
and Split Sets are good for the yield support system. However,
(Rock Tunneling Quality Index) and geo-structure data are
the anchorage of the Split Sets is too low (1.5 to 4.5 ton/m)
estimated in the underground drifts and stopes.
and also it is hard to be installed for the extremely fractured
• Data collected by extra drill holes for the critical infras-
rock. So, the Swellex bolt is selected in Leeville. The Swellex
tructure facilities. To know more detail geotechnical infor-
bolt has a high anchorage (12 to 18 ton/m), easy to be installed
mation in some areas in which the critical infrastructure
in the fractured rock, and installation quality control is easy
located, this method is used.
to be guaranteed by operators
Leeville ground support systems are established based on Based on the Hoek and Brawn, Stillborg I, modified
the Tunnelling Quality Index (Q) for the determination of rock Barton, and Lang methods, a new equation to determine the
mass characteristics and tunnel support requirements, (Barton bolt length in Leeville is proposed. L is the length of the
et al., 1974). In relating the value of the index Q to the stabil- bolt and B is the span of the drift:
ity and support requirements of underground excavations, an
additional parameter is defined called the Equivalent Dimen-
sion, De , of the excavation. This dimension is obtained by
dividing the span, B, diameter or wall height of the excava- Bolt lengths in current in use at Leeville are based on
tion by a quantity called the Excavation Support Ratio, ESR. the new method. The new method is verified by a numeri-
Hence: cal method. In the numerical approach, the minimum stress
method is used to determine the length of the rock bolt. The
analysis also was proven by a ground control monitoring
system in Leeville underground.
606
607
Surface settlement due to deep mining with the Numerical Manifold Method
This paper examines the expected settlement at the ground Table 1. Maximum HD before surface heave, and settlement
surface due to circular cavities in layered rock. We begin with profile width in four different cavity configurations.
a verification of the Numerical Manifold Method (NMM) for
the case of an infinite plate with a circular cavity (Kirsch solu- Cavity depth [m] 40 60 80 60
Cavity diameter [m] 20 20 20 40
tion). The relative errors obtained for the vertical displacement Maximum HD [m] 40 80 120 80
change from 7.65% to 0.64%, depending on the position of the B [m] 80 120 140 120
measurement point. The Kirsch solution assumes constant far
field stresses and ignores the influence of gravity. We there-
fore verify the NMM for the case of a layered plate subjected
to gravity but with no cavity under plane stress conditions,
using a self-developed analytical solution for this problem.
A very good agreement is obtained, with relative errors never
exceeding 0.32%.
This set of verification experiments allows us to proceed
safely with exploration of boundary effects on the com-
puted surface settlement. The amount of surface settlement
is controlled by two opposite factors: 1) stress concentra-
tions due to cavity which increases the amount of settlement,
2) elastic rebound due to stress relaxation of the underlying
strata as a response to removal of material in the formation
process of the cavity, which decreases the amount of settle-
ment. Considering item 2 above, the influence of the distance
between the center of the cavity and the lower boundary of
the model (HD ) on the surface settlement is explored here Figure 1. Profiles of surface settlements above a 20 m diameter
in four different configurations (Table 1). In each config- circular cavity embedded in an elastic medium at depth of 60 m
uration, the distance HD is increased until surface heave is (E = 10 GPa, ν = 0.2, γ = 18 kN/m3 ) for various HD distances. In
obtained. this case surface heave is obtained when HD is greater than 80 m.
The results indicate that the distance HD has a significant
influence on surface settlements above the cavity. The total cavity depth as well as the settlement profile width (B)
surface settlement decreases significantly with increasing HD (Table 1). The diameter of the cavity does not seem to influ-
due to the elastic rebound effect (Fig. 1). The maximum HD ence the maximum HD or the settlement profile width B, but
before surface heave is obtained increases with increasing only the total amount of surface settlement V (Table 1).
608
Tashi Tshering
University of Oslo, Oslo, Norway
Rajinder Bhasin & Roger Olsson
Norwegian Geotechnical Institute, Oslo, Norway
Valerie Maupin
University of Oslo, Oslo, Norway
Keywords: Tunnel, seismicity, rock support, Q-system, Maximum Axial Force on Lining (MAFL)
609
4 CONCLUSIONS
For elastic good quality rocks, the MAFL, for both static and
seismic loading, is independent of depth (or in-situ stress) and
Figure 1. MAFL for both static and seismic loading (h = 0.5) for tunnel dimension. On the other hand, for EPP models (weaker
the four rock classes used for this study. The MAFL for elastic and rocks), the MAFL under static and seismic loading appears
EPP models overlap for good quality rocks with higher Q. to depend on tunnel dimension, depth of the tunnel and rock
mass quality Q. Consequently, the seismic force on the lining
For this study, a fixed pseudo-static seismic loading with is also dependent on depth and the surrounding rock mass
horizontal coefficient = 0.5 and vertical coefficient v = 0 is quality Q.
used for all the experiments. As the tunnel diameter increases from 5 m to 20 m, the
absolute magnitude of the MAFL increases with increasing
3 RESULTS tunnel dimension, but the net increase in MAFL due to seis-
mic loading (seismic force) does not vary significantly with
3.1 Effect of rock mass quality tunnel dimension.
Figure 1 shows the variation in the axial force on the shotcrete Although the absolute magnitude of MAFL for both static
liner of a tunnel in the four rock classes at 60 m depth. It can and seismic loading increases with depth, the seismic force
be seen that the MAFL, for both elastic and EPP models, is on the lining decreases with depth even though the seismic
dependent on the rock mass quality, where the MAFL for both coefficient is fixed for different depths, thereby simulating
static and seismic loading increases as Q decreases. the same level of shaking at different depths.
As the rock mass quality decreases from Q = 40 to 1, the For a 20-m diameter tunnel at depths from 30–200 m, the
absolute MAFL for both seismic and static loading increases seismic force on the lining decreases from about 80% of the
by about 3–4 times. static MAFL at 30 m depth to 2% of static MAFL at 200 m
For a particular depth, it is observed that the MAFL, at depth for Q = 1 and Q = 40; not only the percentage, but the
least for static loading, for elastic and elastic-perfectly-plastic net increase also has a decreasing trend with depth. This indi-
models overlap for good quality rocks (Q ≥ 4 for 60 m depth). cates that although the level of shaking is same for different
Conversely, for rocks with lower Q, the MAFL values deviate depths, there is still some reduction in effect of earthquakes
for the two models. This is because, within the elastic limit, at depth compared with shallower tunnels. It is generally
the rock mass behaves like elastic materials irrespective of the accepted that the major reason for the reduction of under-
choice of the post-failure characteristics. ground damage at depth is due to lower level of shaking with
For EPP materials, the seismic force on lining increases with respect to motion at the surface (Sharma and Judd, 1991).
decreasing rock mass quality. For the four rock classes, there However, from this study it can be inferred that there could be
is an about 12% increase in MAFL due to seismic loading at least another, probably stress-controlled, phenomenon that
for elastic models, and the % increase due to seismic loading contributes to the reduction of impact of ground shaking at
for EPP materials is about 20–26%. depth, in addition to the commonly accepted “lower level of
shaking at depth”.
3.2 Effect of tunnel dimension
For elastic-perfectly-plastic models, the MAFL increases with REFERENCES
tunnel diameter for Q = 1, whereas on the same scale, the
increase in MAFL with tunnel diameter for Q = 40 is insigni- Barton, N., R. Lien and J. Lunde. 1974. Engineering classification
ficant. Similarly, the increase in MAFL for elastic models is of rockmasses for the design of tunnel support, Rock Mechanics,
lower than for elastic-perfectly-plastic models. This shows that Vol. 6, No. 4, pp. 189–236.
for good rocks, behaving elastically, the MAFL does not vary Barton, N. 2002. Some new Q-value correlations to assist in site
significantly with tunnel dimension, and whereas the MAFL characterization and tunnel design. Int. J. Rock Mech. & Min. Sci.
increases with tunnel dimension for poor quality rocks, such Vol. 39/2:185–216.
as for Q = 1. On the other hand, the seismic force on the lin- Dowding, C. & Rozen, A. 1978. Damage to rock tunnel from earth-
quake shaking, Journal of Geotechnial Engineering ASCE, vol.
ing seems independent of the tunnel dimension. The increase
104: 175–191.
in MAFL due to seismic loading is ∼0.1 MN for Q = 1 and Singh, B. & Goel, R.K. (First Edition) 2006.Tunnelling in weak rocks.
∼0.02 MN for Q = 40. Oxford, UK: Elsevier.
3.3 Effect of depth Sharma, S. & Judd, W. R. 1991. Underground opening damage from
earthquakes, Engineering Geology, vol. 30: 263–276.
The magnitude of the seismic force on lining is insignifi- Waltham, T. 2009. (Third Edition). Foundations of Engineering
cant for elastic models when compared with EPP models. Geology. Taylor & Francis.
In addition, the seismic force for elastic models appears to
610
Investigation into the strength of iron ore pillars in the Lorraine basin
Keywords: risks and hazards, stability analysis, back analysis, case studies, rock failure
611
Figure 1. Contour plot of the areas of overlap between the distribu- and
tions of safety factors of failed and stable pillar populations.
6 CONCLUSION
612
Zhi-guo Wang
Institute of Rock Mechanics and Fractals, China University of Mining and Technology (Beijing), Beijing, China
College of Mining Engineering, Hebei United University, Tangshan, China
He-ping Xie
Institute of Rock Mechanics and Fractals, China University of Mining and Technology (Beijing), Beijing, China
Sichuan University, Chengdu, China
Hong-wei Zhou
Institute of Rock Mechanics and Fractals, China University of Mining and Technology (Beijing), Beijing, China
Keywords: mining induced crack network, deep mining, distinct element method, fractal dimension, evolutional law
With the increasing of coal mining depth, a better understand- spatial distribution of mining-induced crack network in over-
ing of evolution of mining-induced crack network plays an burden rock strata under the condition of deep mining at
important role in deep safe mining. The spatial distribution different coal seam thickness, according to the factors of engi-
of mining-induced crack network in overburden rock strata neering geology and mining technology on the coal seam
affect the stability of overburden rock strata to some extent No. 11-2 in Panyi Coal Mine, based on results of physical
under the condition of deep mining. So it is very important simulation experiment, taking the mined rockmass in over-
to research the spatial distribution of mining-induced crack burden rock strata as the crack rockmass of continued broken
network and its evolutional behavior for understanding the and fracturing. The rockmass material constitutive model is
evolution characters of overburden rock strata structure, the defined as Mohr-Coulomb plastic model, the coal seam is
subsidence law and stability of mined rockmass. The crack defined as strain softening model, and the joints is defined
network of rockmass can be described by the fractal geometry as joint area contact model with Coulomb slip failure in the
method, and the studies shows that the spatial distribution of numerical model. The original stress in the numerical model is
mining-induced crack network has the fractal characters, the defined by the actual measured stress in the mine. For model-
evolution law of the fractal crack network may be researched ing the mining-induced crack network better, the setting of the
with numerical simulation method for understand the break- horizontal joint set is according to the borderline in the hori-
age and movement law in the overburden rock strata. This zontal rock strata, the virtual joints is added in thick rock strata,
paper researches the statistical self-similarity and the evolu- and the space between every two vertical joints is defined as
tional law of mining-induced crack network by adopting the 9m base upon the pace of roof periodic caving in the phys-
fractal geometry theory in combination with the experiment ical simulation and field measurement. The software UDEC
results of the numerical simulation in the coal seam No. 11-2 is employed to simulate the mining-induced crack network
of Panyi Coal Mine. by considering the main fault Fe7 and it’s branches, and sim-
The study object of the paper is based on the working face ulated a series of models with different coal seam thickness
No. 2171(1) of the coal seam No. 11-2 in Panyi Coal Mine which equal to 1.8 m, 4 m, 6 m, 8 m for comparatively study the
of Huainan coal district. The coal seam is a near horizontal coal seam thickness effects to mining-induced crack network
coal bed, with an average thickness of 1.8 m. The dimension distribution.
of working face is 1780 m long in strike, 206 m in dip. The Box-counting dimension is one of the most used fractal
advancing velocity of the working face is 3.6 m/d by longwall dimension, so it is used to calculate the fractal dimension of
coal mining method with integrated mechanization coal cut- mining-induced crack network. The software Fractal Dimen-
ting system. The coal field is located in the plain, the terrain sion adopted may calculate the fractal dimension of all kinds of
is flat and the average mining depth is 742 m. The geologic two-dimension digital image quickly. Firstly the open mining-
structure in the area is middle complicated, belongs to a mon- induced crack network can be output from the numerical
oclinal structure, and is crossed at an angle by the fault Fe7. simulation experiment, then the images with different min-
The immediate roof of the coal seam No. 11-2 is mudstone, ing width is processed by the software Photoshop to extract
sandy mudstone and coal seam No. 11-3, the main roof is a mining-induced crack network, and finally the processed
medium to fine grained sandstone, immediate bottom is mud- images is inputted in the software Fractal Dimension with
stone, sandy mudstone, mudstone, carbon mudstone and coal box-counting dimension to calculate the fractal dimension of
seam No. 11-1, and the main bottom is fine grained sandstone. mining-induced crack network.
The DEM (discrete element method) was applied to estab- By studying of mining-induced crack network characters
lish discontinued medium models for the simulation on the in different thickness, different mining width with fractal
613
614
Hyung-sik Yang
Department of Energy and Resources Engineering, Chonnam National University, Gwangju
Myong-Hwan Jang
Korea Resources Corporation (KORES)
Keywords: mining, numerical modeling, rock joints, rock support, stability analysis
With the development of industry and increased awareness of Paddies and ordinary fields are located on the upper part of
environmental issue, eco-friendly operation of mine is gain- the facility, and the surface water may flow into the pit along
ing importance. The process of rock crushing generates noise, the joints from the upper part to the lower part during the rainy
vibrations and enormous dust, which pollutes the surround- season. Therefore, water must be cut off to properly guide the
ing environment. Mitigation of these environmental nuisances water in the pit.
generated by the crusher is essential. In this aspect, the present The stereographic-projection analysis results showed that
study focuses on geo-mechanic based safety investigation for pillars 1, 2, and 4 were exposed to planar failure, but that
operating crusher in an underground mine. pillar 3 hardly was. Risk of wedge failure was found in pillars
Geo-mechanical data of rocks (Q, RMR value) were inves- 1-3 and was low for pillar 4.
tigated and fed to appropriate FLAC3D analysis for evaluating According to the RMR analysis results for the pillars and
safety of the given underground excavation for installation of excavation site, the engineering properties of the rock was
crusher. Safety evaluation was undertaken in 3 steps and the “ordinary” or “satisfactory,” with an average RMR of 58 and
salient outcome is given as follows. a Q-system value of 4.4.
The stability test results based on the numerical analysis
i) Safety evaluation of existing underground excavation
showed that after the application of the recommended sup-
without removing any pillars to provide the benchmark-
port amount, pillars 1 and 3 had high safety factors above 3,
ing results; This step revealed that the underground
but pillars 2 and 4 still had low safety factors (0.97 and 1.0,
excavation was safe.
respectively). The numerical analysis for the extended pillars
ii) Safety evaluation of an underground excavation assuming
2 and 4 increased the safety factors to 1.83 and 1.92, respec-
some part of the pillars to be removed, this step showed
tively. This indicates that the pillar size must be increased
the generation of stress concentration in the span between
to improve the safety factor. In addition, the pillar must be
the pillars and in the pillar sidewalls.
supported according to the rock class.
iii) Safety evaluation of an underground excavation was done
assuming some pillar removed. However, after installation
of suitable supports the situation was safe.
615
N. Zhang
Institute of Hydroelectric and Geotechnical Engineering, North China Electric Power University, Beijing, China
S.C. Li
The Research Center of Geotechnical Engineering, Shandong University, Jinan, Shandong, China
A.Z. Lu
Institute of Hydroelectric and Geotechnical Engineering, North China Electric Power University, Beijing, China
M.T. Li
The Research Center of Geotechnical Engineering, Shandong University, Jinan, Shandong, China
Department of Civil Engineering, Shandong Jiaotong University, Jinan, Shandong, China
L. Yang
The Research Center of Geotechnical Engineering, Shandong University, Jinan, Shandong, China
Keywords: bolts and anchors; lab testing; rock failure; rock support
1 INTRODUCTION
616
617
F. Amberg
Lombardi Engineering Ltd., Minusio, Switzerland
For a given underground excavation geometry, the ground in particular if a rigid lining is placed. The most evident cases
reaction curve represents the displacement of a selected point are swelling and creep, but other possible conditions might be
on the excavation boundary as a function of the support pres- contact grouting with high pressure, excavation in a low per-
sure. The classical method refers to the simple case of axial meable and soft ground or placement of a rigid lining near the
symmetry in an infinite elasto-plastic medium, but in fact a tunnel face in case of squeezing ground. In fact, also the loos-
ground reaction curve can be defined for any type of geom- ening up of rock beyond the roof may lead to an equilibrium
etry and loading condition. It is thus possible to include any above the ground reaction curve.
desired effect, e.g. any shape of the cross section, the tunnel Recent developments have shown that even the usual con-
face, the presence of a fault beside the tunnel, a shallow tun- struction procedure of a tunnel tends to lead to an equilibrium
nel, the gravity, the schistosity, the stratification of the rock or lying above the ground reaction curve. This behavior can be
the anisotropy of the initial stress state. The great advantage of pointed out by simulating the tunnel advance in 3D-analyses.
the ground reaction curve is that the result is not limited to a Initially, this statement might sound peculiar, since it puts
single equilibrium, but it allows immediately to verify how the the classical theories aimed to define the equilibrium in tun-
equilibrium of the excavation can change for different support neling somehow into question. However, in every tunnel the
conditions. This is an essential aspect to analyze and to opti- radial pressure at excavation radius is in fact never only always
mize the design of a tunnel in case of expected convergences. decreasing. From the initial stress ahead of the face the radial
It allows to dimension the tunnel lining with respect to the pressure decreases to the minimum value in the excavation
construction procedure and to assess the safety condition of zone just behind the face. Subsequently, an increase of radial
the excavation. stress occurs while support measures are put in place. Any
The ground reaction curve has also some restrictions: in additional support measure, if placed at a distance from the
fact it must be considered that it does not represent all pos- face where convergences still occurs, produces a certain radial
sible equilibrium points, but only the lower boundary. The pressure increase causing a re-compression of the surrounding
ground reaction curve is determined with a progressive reduc- ground. In case of squeezing ground this phenomenon might
tion of the support pressure and thus the stress state within lead to a very sensitive rising of the radial pressure acting on
the failure zone around the tunnel is at failure limit. In case the lining
of an increasing support pressure after lining placement, a After a brief review of the convergence-confinement
re-compression of the ground in radial direction can be method, the paper presents various possible applications of
observed. The final equilibrium will be above the ground the ground reaction curve showing its usefulness for the anal-
reaction curve, i.e. the support pressure is greater than the ysis of underground excavation problems. Finally, the concept
expected one for a certain displacement. The mentioned of the ground reaction curve as boundary for a region of the
re-compression of the ground might arise for several reasons, possible equilibriums is also presented and discussed.
621
B. Brouard
Brouard Consulting, Paris
Keywords: numerical modelling, rock mass classification, rock failure, rock support, stability analysis, tunnelling
1 EXTENDED ABSTRACT
In the long term, salt behaves as a viscous fluid and salt cav-
erns gradually shrink. Creep closure rates in shallow caverns
are slow. They must be assessed through shut-in pressure tests,
which consist of closing the cavern and measuring the pres-
sure evolution at the wellhead as a function of time; or through
brine outflow tests, which consist of opening the cavern and
measuring the flow of fluid (brine or hydrocarbon) expelled
from the wellhead. In this paper, we describe a brine outflow
test and a shut-in pressure test performed in the 250-m deep
SG13-14 cavern of the Gellenoncourt brine-field operated by
Compagnie des Salins du Midi et Salines de l’Est (CSME) in
Lorraine, France. The SG13 and SG14 7”-wells were operated
as brine-production caverns from 1976 to 1980. After some
time, the two caverns coalesced, and, in 1980, SG13-SG14 vol-
ume was V = 240,000 m3 . At SG13-14 depth, creep closure
rate can be expected to be ε̇cr ≈ −10−5 /yr. Brine thermal- Figure 1. Cumulated expelled mass as a function of time.
expansion coefficient is αb = 4.4 × 10−4 /◦ C. A brine tem-
perature decrease rate of Ṫc = −0.02◦ C/yr would generate a
relative brine volume decrease rate of αb Ṫc ≈ −10−5 /yr: tem- The cavern was shut-in from May 25, 2009 to November19,
perature evolution must be carefully assessed. By December 2009. During a shut-in test, cavern convergence rate is pro-
2008, a temperature gauge was lowered into the SG13 well. portional to pressure build-up rate, ε̇cr = −βṖc . As build-up
The cavern temperature remained perfectly constant during rate due to cavern creep closure is Ṗwh = 47.1 Pa/day, cav-
the period December 2008–June 2010. In 2000 the cavern ern closure rate is ε̇cr = − 0.93 × 10−5 /yr (Cavern complete
had been shut-in after a sonar survey. Eight years later, before closure is reached after more than 100,000 years). Wellhead
the test began, wellhead pressure had built up to approxi- pressure experience small fluctuations due to Earth tides and
mately 0.08 MPa. On July-3, 2008, the cavern was opened atmospheric pressure changes, which are transmitted to the
and wellhead pressure dropped to zero. Cavern compress- cavern through the rock mass. Creep closure rate is slower
ibility was measured to be βV = 130 m3 /MPa. The outflow during the shut-in test. In fact transient effects triggered by the
test began on July 23, 2008 and was completed by May 25, initial Pc = −0.08 MPa pressure drop are important at the
2009, Figure 1. As temperature rate is exceedingly slow, the beginning of the outflow test. They mainly include transient
average brine outflow rate or q = 9.5 liters/day is governed creep and salt crystallization. The steady-state creep closure
by cavern-creep closure and the relative creep closure rate is rate, as observed during the shut-in test, is slightly slower than
ε̇cr = −q/V = −1.45 × 10−5 yr−1 . Large fluctuations can be 10−5 /year or 2 m3 /year. This value proves that even in the long
observed: periodically, the brine flow rate is several hundreds term (several centuries) subsidence and possible brine leaks
of liters per day. Conversely, for most of the time, the flow from the cavern should have negligible impact from the point
rate is nil and the air/brine interface drops down into the well. of view of environmental protection.
In fact, the opened cavern behaves as an extremely sensitive
barometer.
622
Convergences and core extrusions during construction of the Gotthard Base Tunnel
R. Hug
Ernst Basler + Partner AG, Zurich; formerly ETH Zurich, Switzerland
1 INTRODUCTION 2 GEOLOGY
Squeezing intensity in tunnelling often varies over short The tunnel crosses the northern TZM formation and the
distances, even where there is no apparent change in the Clavaniev zone for about 1150 m at a depth of 800 m. Both the
excavation method or lithology. This variability makes the TZM formation and the Clavaniev zone are characterized by
prediction of the conditions ahead of the face very diffi- alternating layers of intact and more or less kakiritic gneisses,
cult, even after some experience has been gained with a slates, and phyllites. The term “kakirite” denotes a broken or
specific geological formation during excavation. However, intensively sheared rock, which has lost a large part of its
reliable predictions of the ground conditions ahead of the face original strength.
are essential in order to enable adjustments to be made dur-
ing construction to the temporary support or the excavation
diameter. 3 CONSTRUCTION METHOD
An assessment of the behaviour of the core ahead of
the face provides some indications concerning the mechan- The tunnel was excavated full-face. Squeezing was tackled
ical characteristics of the ground and seems to hold out through a yielding support system consisting of two rings of
promise with regard to ground response predictions (Cantieni sliding steel sets (TH 44/70) lying one upon the other and
2011). connected by friction loops (Figure 3, 2f and 2g).
The present case history investigates the convergence and
extrusion data monitored during construction of the western 4 MONITORING AND DATA ANALYSIS
tunnel of the new Gotthard Base Tunnel, which crosses the
northern intermediate Tavetsch-formation (the so-called TZM The core extrusion was monitored by reverse-head-
formation) and the adjacent Clavaniev zone (Figure 1). The extensometers (Thut et al. 2006). The position of the measur-
aim of this study is to find out whether there is a correlation ing head and the six measuring points of each extensometer
between the extrusion of the core and the convergences of with respect to the tunnel alignment are shown in Figure 2i.
the tunnel and, accordingly, if it would have been possible to The convergences of the opening were monitored optically
predict the convergences solely on the basis of the monitored (Figure 2h).
extrusions. In order to assess the behaviour of the ground along the
tunnel it is necessary to compare deformations that occur
under similar conditions. Therefore, the analysis considers,
as a measure of the extrusion, the strain that develops due to
the advance of the face from a distance of 6 m to a distance of
2 m in respect of each ground interval. The radial displacement
of the tunnel crown is used as a measure of convergence. The
analysis takes into account only the monitored displacements
occurring due to a face advance of 25 m.
Figure 2a shows the axial strain and radial displacements of
the crown along tunnel reaches 1 and 2. The convergences in
tunnel reach 2 exhibit a weak correlation with the extrusions.
No correlation can be observed between the convergences and
the extrusions in tunnel reach 1.
5 CONCLUSIONS
Figure 1. a) Gotthard Base Tunnel: schematic representation of the
longitudinal geological section which includes the squeezing TZM The case history of the Gotthard Base Tunnel shows that
formation (after Kovári 2009). there is a weak correlation between the axial extrusions and
623
ACKNOWLEDGMENTS
REFERENCES
624
Li-Hsien Chen
Department of Civil Engineering, National Taipei University of Technology, Taipei
Zhe-Xun Yang
Department of Civil Engineering, National Taipei University of Technology, Taipei
To speed up construction of mass transit subway and the Meanwhile, in-situ experimental results from shield tunnel-
popularity rate of sewage, Taiwan’s underground excava- ing and pipe jacking construction were used to examine this
tion works, especially for mechanical cutting cases, show an model and it shows a nice agreement between both. From
ascendant tendency. This study presents a generalized solution this analytical approach, a proposed “oval-shaped cutting
for underground geological-mechanical interaction. By using ellipsoid”, including its center (O), area (A), and long/short
dimensional analysis, This model generalizes geological char- axis (ax /by , or ay /bx ), can be used not only to estimate the
acteristics grouped into three categories: (1) brittle (rock-like), functionality and efficiency of cutting machine adopted for
(2) the ductile (soil-like), and (3) brittle-ductile (gravel-like) tunnel project, but also to offer a warning information for
type with respect to two cutting forces: (1) thrust and (2) torque inadequate cutting strategy.
to evaluate their excavation/penetration rate. Furthermore, the
leading cuttability indices can be obtained to enable to assess
the underground excavation.
625
Li-Hsien Chen
Department of Civil Engineering, National Taipei University of Technology, Taipei
Yao-Chung Chen
Department of Construction Engineering, National Taiwan University of Science & Technology, Taipei
Keh-Jian Shou
Department of Civil Engineering, National Chung-Hsing University, Taichung
Gou-Lung Lin
Department of Civil Engineering, National Taipei University of Technology, Taipei
Keywords: tunnelling, rock failure, Mechanical excavation (TBM/RBM), case studies, site characterisation
In recent years, the construction projects increase the cases of situations for risk assessment. In addition, the in-situ data of
underground tunneling by mechanical cutting such as tunnel in trust in shield machine (Taoyuan tunneling project) is pre-
the mountain, rapid transit system in the city and sewer system, sented to confirm with. It depicts a favorable agreement for
etc. However, the complex geological conditions make exca- the estimation of thrust in this study.
vating difficulties in Taiwan. This paper presents an analytical The results shows that the total thrust for upper bounds
estimation to meet different mechanical cutting methods and lower bounds are: (401%, 37.8%) and (258.2%, 31.7%)
(tunnel boring machine, shield tunnel and pipe jacking), types compared with normal condition in gravel and weathered
(earth pressure balance, slurry pressure balance, thick-mud), sandstone cases respectively, which normalized boundary val-
and geological conditions (soil, gravel and rock) by gener- ues are (13 × 10−4 ∼82 × 10−4 ) and (0.97∼4.98) for cutting-
alizing their total thrust system. The straight-line thrust is head resistance respectively. It also found that the cutting-
calculated for either wedge- or conical-typed cutters of tun- head resistance take about 28% of the total trust resistance
neling machine. In this generalized work, the upper bound (F = 4773 kN) in the gravel case by taking cutters’ forces into
and lower bound of trust are highlighted for the warning account.
626
Experience on the underground storage for the crude oil and This paper focuses on the freezing temperature and penetra-
various types of hydrocarbons proved to be an economical tion length of groundwater investigated by a thermo-hydraulic
means for the storage of LNG. Capitalizing the knowledge coupled analysis and CFD analysis of the groundwater flow
from our past experience, attempts were made to develop the with in the joints at the time of ice-ring formation. The pre-
LNG underground storage system but were not successful due vious research related to the formation of ice-ring showed
to its failure in controlling the temperature and groundwater that the freezing temperature of water flowing through rock
without containment. From the past trials, it was observed that joints are affected by the latent heat generated during freezing,
the thermal stress and freezing of groundwater nearby to the aperture size of the joints and the flow rate of groundwater
cavern generated cracks in the host rock mass. However, these (Jeong et al, 2006).
thermal cracks contributed to induce gas leakage leading to In general as the flow rate of groundwater through rock
the increase in heat flux exchange between the stored LNG joints is not very fast and is controlled by Darcy’s lawexhibit-
and rock mass. Eventually, this causes excess of boil-off ratio ing a laminar flow with a Reynolds number less than 2000.
in the stored LNG. The laminar flow has the following relationship according to
A new technology has been developed for the storage of joint apertures.
LNG in a lined rock cavern by combining the well-proven
technologies and a new concept named “formation of ice-
ring”. The formation of ice-ring, one of the core technologies
in the LNG storage system can be described as a complex
mechanism underlying between the thermal characteristics of where d = joint aperture, µ = fluid viscosity, γ = unit weight,
rock and hydro-geological characteristics of groundwater. The P = difference in pressure, H = difference in hydraulic
process of ice-ring formation as a second barrier impervious head and L = travel length of groundwater through joints.
layer functioning against the leakage of contained LNG can From V = K · i,
be explained via each stage of the construction of a lined rock
cavern. Groundwater is temporarily removed from the rock
surrounding in the cavern during the first phase of the con-
struction. This preliminary de-saturation of the host rock mass
aims to prevent unacceptable hydrostatic pore pressure and ice where K is permeability and i is hydraulic gradient.
formation behind the cavern lining. After the completion of Considering equations (1) and (2), the flow rate of ground-
the excavation of the cavern and installation of the contain- water can be determined by joint aperture and hydraulic
ment system, the LNG is stored in the cavern, thereafter the gradient.
cold front from the cavern starts to propagate immediately. The analysis procedure using eqn (2) and equivalent per-
When the cold front advances far enough from the cavern meability were used for the calculation of the flow rate of the
wall, drainage can be stopped to allow groundwater progres- in-situ rock mass with infilling materials and joints. Table 1
sively to rise up and quickly form a thick ring of ice around the shows the range of the input values and the results of the anal-
cavern. Once the ice-ring formation is completed, the opera- ysis. The lowest freezing temperature of water was found to be
tion of the drainage system stops. The drainage period during −0.233◦ C in case of a fault or fracture zone (permeability of
LNG storage will last for several months or years depending 1.0 × 10−5 m/s and aperture of 5 mm). Therefore, the analysis
on the thermal properties of rock masses and hydro-geological result demonstrates that the ice-ring formation temperature
characteristics of the site. The key technologies of the LNG can be increased from −3◦ C depending on the safety issues.
storage system have been demonstrated through the design, As shown in the Figure 1, ice begins to form at the joint
construction and operation of the pilot plant in Daejeon, surface contacting with the groundwater. And it can be seen
Korea. that the completely frozen area is generated at the joint
627
Table 1. Analysis results of the large joint considering the real characteristics of rock mass.
surface first, and the water temperature above 0◦ C also exists factors were observed to be affecting each other too. In case of
in the center of the joint aperture. It can be explained as fol- laminar flow, it is possible to estimate the freezing temperature
lows: Because the groundwater flow is very slow, groundwater considering only two of these three variables as the recovery
contacting with the joint surface having lower temperature rate is affected by the hydraulic gradient and the joint aper-
than freezing one gradually turns into ice. In addition to the ture. The freezing temperature of groundwater in both narrow
increase of fluid viscosity, the flow velocity is much slow. and wide aperture of a joint can be applied with a value higher
However, groundwater may continue to rise through the center than −3◦ C, which was assumed as freezing temperature in
of the joint aperture. the past. However, these numerical analyses were conducted
In summary, the results of numerical modeling demon- in a limited range with several assumptions. Therefore, further
strated that when the ice-ring is formed, the factors that affect analysis with a wide range of varied conditions would be ben-
freezing temperature of the groundwater are the recovery rate eficial to determine the freezing temperature of groundwater
of groundwater and aperture of joints. In addition, these two in the joints of the rock.
628
A.M.B. Pereira
Department of Civil Engineering, Fluminense Federal University (UFF), Rio de Janeiro, Brazil
Keywords: rock caverns, mine design, numerical modeling, stability analysis, oil reservoir
629
REFERENCE
630
Keywords: back analysis, mechanical excavation (TBM/RBM), rock burst, rock stress, tunnelling
High in situ stresses strongly influence the stability of the The case of the Southern TBM-excavated part of the
excavation face in deep tunnels. When hard rock fails under Lötschberg Base Tunnel represents a very good example of
high acting stresses, it disintegrates and massive rock is such a condition. During construction a major phenomenon
transformed to blocky ground. Strong, instable rock blocks was observed, reported by the workers as “blocks formation”
may appear at the tunnel face and this condition may pre- in front of the TBM head. It consisted of blocks with a predom-
vent normal TBM operations When these instabilities appear, inant slab-like shape loosening from the tunnel face. These
instead of showing a flat shape with clear marks of cutters, blocks jammed conveyor intake, provoked frequent disc and
the tunnel face is irregular. This condition causes high fre- bearing failure, caused extensive wear of the cutterhead and
quency and great magnitude of variations on cutter force. As cutters.
a consequence, high loading and heavy dynamic impacts of This paper introduces the ongoing research on blocky
the cutters and the cutterhead against the hard rock blocks ground tunnelling at the Rock Mechanics Laboratory (LMR)
occur. Due to these effects abnormal cutter ring wear, such of the Ecole Polytechnique Fédérale de Lausanne (EPFL). The
as pitted, chipped and flatted wear appears, and the cutter aim is to show the problem by focusing on the principal aspects
bearing failure is enhanced. Furthermore, extensive cutter- involved in mechanized tunnelling in blocky rock masses and
head vibration and the cutter loading process may lead to to define the main effects of such a condition on machines per-
the instability of the rock blocks at the face. In return these formances. To this purpose, the TBM performances in some
falling blocks may destroy the cutters. To minimize the vibra- “representative” sections of the Lötschberg Base Tunnel are
tion of the cutterhead and the dynamic loading and impacts analyzed. The variation of the machine utilization factor and
on cutters and cutterhead, the thrust force and the cutter- total advance rate with rock mass conditions varying from
head rotational speed have to be decreased below the normal massive to very blocky is studied.
level. Finally, problems can be expected to the muck conveyor The results show a clear correlation between TBM perfor-
system which can be jammed or damaged by irregular rock mances and rock mass structure: the more blocky the rock
fragments. mass, the larger the downtimes, the lower the advance rate.
In this context, the required replacements of the cutters Therefore, retards in the construction of the tunnel and an
and frequent reparations of the cutterhead increase main- increasing of the construction costs must be expected when,
tenance requirements; the TBM utilization factor decreases due to high stresses, a massive to moderately jointed rock mass
while excavation times and costs increase. becomes blocky ground.
631
Keywords: Water problem; Tunnelling; Weak rock; Stability analysis; Rock support; Monitoring; Pressure cells; Tunnel closure
Jammu Udhampur Srinagar Baramulla Railway Link (JUS- wall. It is understood that once the rock pressure is stabilized,
BRL) project is one of the prestigious projects of India to these pressure cells can also provide information of the pres-
connect Kashmir valley of J&K state with rest of the country. sure exerted by the fluctuating water head during and after the
The railway link starts from Jammu and, when completed, will rainy season. Closure studs were also installed to monitor the
travel for 345 km including tunnel route of 100 km to the city tunnel roof and wall deformations. In addition to this, to mon-
of Baramulla on the northwestern edge of the Kashmir Valley. itor the ground water table fluctuation throughout the year,
The route crosses through very difficult Himalayan terrains number of bore holes were also drilled from ground surface
making it an extremely challenging engineering project. The to the tunnel invert level on both sides of the tunnel alignment
project envisages construction of a single broad gauge railway at various offsets. Data of three years have been analysed and
line with maximum speed potential of 100 km/h. The ruling following results were obtained.
gradient was limited to 1 in 100 whereas maximum degree of The maximum water discharge (421-1125 l/s) was recorded
curvature was restricted to 2.75◦ . in the month of August every year whereas minimum value
Jammu to Udhampur section (53 km) of the JUSBRL of discharge (below 25 l/s) was recorded in the month of June
project has been already completed in the year 2005. every year. Pressure cells installed in the problematic water
Udhampur – Katra section of this rail link is 25 km long charged zone, in general, showed increased support pressure in
and involves construction of 7 tunnels having total length of rainy season (July to September) only. However, no deforma-
10 km. The paper deals the water problem faced in 2.5 km tion in walls was recorded, which indicated that the pressures
long Tunnel No. 3 between Udhampur and Katra. Tunnels in are within the capacities of installed supports. The analysis
this section has the rock masses of the Siwalik group like further revealed that there is an increasing trend of support
clay and soft sandstone and Pleistocene to recent river borne pressure at chainage 12831 m and 13125.5 m every year. A
deposits. The D-shaped Tunnel has finished width of 5.2 m and pressure cell installed on the right wall at the chainage 12831 m
height 7.97 m. Tunnel depth or the overburden above tunnel indicated the pressure of 2 kg/cm2 in the month of August
roof varies from 30 m to 40 m throughout the tunnel alignment 2009 and the pressure decreased by about 0.5 kg/cm2 after
between chainage 12200 m and 14700 m. To support the weak the rain was over. It indicates that the pressure contributed by
tunnelling strata, steel rib supports ISHB150 and ISHB200 seepage of rain water is 0.5 kg/cm2 which is also indicated
have been used for roof and vertical legs respectively. by the analysis of water table data which gives the maximum
This tunnel was facing severe water inrush problems dur- water pressure at this chainage as 0.6 kg/cm2 . But, the pres-
ing construction in the stretch of 436 m i.e. between chainage sure cell at this chainage showed increasing trend of pressure
12831 m and 13267 m. To release the water pressure, weep every year. It was considered to be due to resettlement of the
holes were drilled on both the tunnel walls below springing side rock mass comprised of silt with cobles & pebbles and
level during the excavation of tunnel. Even after the construc- clay. It seems that the clay has washed away by running water
tion of the tunnel, water seepage problem remains continued. through drainage holes and the cobles and pebbles have filled
The zone affected by water inrush was supported by double the empty space created by clay behind the wall of the tun-
ribs system. Outer ribs were installed at the spacing of 1 m and nel resulting in increase of lateral pressure. It was therefore
inner ribs at the spacing of 0.5 m from centre to centre. The recommended to divert the water through another weep hole
rest part of the tunnel was supported by single rib supporting nearby and strengthen the area by grouting.
system. Final support of 30 cm thick PCC lining is yet to be The full length paper deals with detailed analysis of the
placed in this problematic zone of the tunnel. instrumented data like annual fluctuation in water table,
To observe the pressure on the steel rib supports by the water discharge through weep holes, pressure cell indications,
surrounding rock, vibrating wire type pressure cells were tunnel-closure data and suggestions for solving the water
installed on the steel ribs at various chainages on the tunnel inrush problem.
632
Y. Mitani
Department of Civil and Structural Engineering, Graduate School of Engineering, Kyushu University, Fukuoka, Japan
1 INTRODUCTION
An experimental test of this system was conducted by curing the concrete was compared to verify the effectiveness of the
a concrete lining for an actual road tunnel. As a result it was moisture curing for the lining, it clarified that when compared
possible to ensure that moisture curing of the concrete lining to cases when moisture curing was not used (air curing), the
was achieved. Furthermore, when the compression strength of compression strength was 2.0 to 2.5 times stronger (Figure 2).
633
B. Garitte
CIMNE-Universidad Politécnica de Cataluña, Barcelona, Spain
A. Gens
Universidad Politécnica de Cataluña, Barcelona, Spain
1 INTRODUCTION
634
REFERENCES
635
R. Jiménez
ETSI Caminos, C. y P., Technical University of Madrid, Spain
Keywords: tunneling, time-effects, rock rheology, tunnel support, back analysis, case studies
636
REFERENCES
Figure 3. Computed m parameters vs. support type. Gonzalez del Alamo J. A., 2010. Prediction of Convergence in Rock
Tunnels Considering the Rheological Behavior. MSc Dissertation,
et al. 1987). The fitting process is based on minimizing the ETSI Caminos, C. y P., Technical University of Madrid, Spain.
(squared) errors between observations and predictions of [In Spanish]
the model. Results indicate that the predictive capabilities of Kontogianni V. A., Psimoulis P. y Stiros S. 2006. What is the con-
the model are reasonable, and we presented an example of a tribution of time-dependent deformation in tunnel convergence?.
Engineering Geology 82 (2006) 264–267.
tunnel section in which the model closely reproduces the evo- Panet M., Guenot A., Sulem J. 1985. A new aspect in tunnel closure
lution of observed convergences (both with respect to time interpretation. 26th US Symposium on Rock Mechanics, Rapid
and with respect to distance to the face) for a case history in City.
Spain. Schubert, W., Grossauer, K. Lenz, G. 2007. Innovative approaches
In addition, to construct an increased database with real in monitoring data evaluation and interpretation. In: CD-ROM
tunneling case histories, we have back analyzed the model Ribeiro et. al. (eds). Proc. of the Congress of the 11th Interna-
parameters for a number of unpublished case histories in Spain tional Society for Rock Mechanics; specialized session. Lisbon,
and also for several case histories published in the literature. Portugal.
Such results have been employed to study the influence that Sellner, P., Grossauer, K., Schubert, W. 2002. Advaced analysis and
different factors have on the relative importance of rheolog- prediction of displacements and system behaviour in tunneling.
Proceedings of the ISRM, Eurorock 2002, Portugal, 25–28.
ical convergences with respect to convergences produced by Sulem, J. Panet, M. Guenot,A. 1987. Closure analysis in deep tunnels.
the advance of the face in real tunnels. Results are still not International Journal of Rock Mechanics and Mining Sciences &
completely conclusive in this case and, for instance, it was not Geomechanics Abstracts, Vol. 24, No. 3, pp 145–154.
possible to completely reproduce previous (reasonable) sug-
gestions (see e.g., Kontogianni et al., 2006) that the support
637
Reliable method for controlling tunnel convergence based on analytical results and
monitoring results in squeezing rocks
K. Yanagawa
Kanazawa Office of River and National Highway, Ministry of Land Infrastructure and Transport, Ishikawa, Japan
N. Kitou
Nanao Tunnel branch office, Nishimatsu Construction, Ishikawa, Japan
After excavating a tunnel in squeezing rocks, it is important to successfully by immediately closing the cross section of
close its cross section immediately to prevent the rocks from the tunnel. This result was verified by numerical analysis
being relaxed. However, because of the immediate tunnel ring (Fig. 2).
closure of a tunnel cross section, it is extremely difficult to 2) From the execution example, it was found that the immedi-
estimate an appropriate supporting scale and cross-section ate closure of the cross section is effective in securing the
closure timing corresponding to a ground condition. stability of the tunnel cross section in the rock mass whose
Currently, no guideline for the supporting scale and closure competence factor is below 1.0.
timing has been established. In addition, few studies describe
the mechanism and influences of difference in tunnel closure
timing on the tunnel surrounding.
Accordingly, this paper shows an example of the construc-
tion of the Nanao tunnel on the basis of information obtained
by feeding back measurement results about deformation dur-
ing the tunnel excavation to determine supporting scale and
closure timing at a subsequent period (Table 1). This paper
also considers the effect of deformation suppression during
implementation of early closure and the validity of the clo-
sure timing. Further, this paper considers influences of the
difference in tunnel closure timing on the tunnel surrounding
by using elastoplastic three-dimensional numerical analysis Figure 1. Excavation method.
using FLAC3D (Fig. 1).
Consequently, the following results are evident.
1) During the information-oriented execution of the Nanao
tunnel, the stability of the cross section was obtained
638
R. Jiménez
ETSI Caminos, C. y P. Technical University of Madrid, Spain.
D. Recio
Acciona Infraestructuras I+D+i, Madrid, Spain
Following Goel et al. (1995), the classifier is developed using Figure 1. Probabilistic squeezing predictions lines (log N –H
an input vector of observations composed of the Rock Mass space).
639
640
Shaft sinking and lining design for a deep potash shaft in squeezing ground
G. du T. Judeel
Mining & Engineering Technical Services, Shaft Sinkers (Pty) Ltd, Sandton, South Africa
G.J. Keyter
GeoStable SA cc, Randburg, South Africa
N.D. Harte
Mining & Engineering Technical Services, Shaft Sinkers (Pty) Ltd, Sandton, South Africa
1 INTRODUCTION The initial design of the shaft liner was based on a ground
pressure assuming hydrostatic in situ ground stresses that are
The main challenges of shaft sinking in weak sedimentary equal to the overburden pressure. Cast iron tubbing and con-
rocks in potash deposits are that of dealing with groundwa- crete backing thicknesses were selected to withstand such
ter when excavating through water bearing formations, creep external pressure with an adequate factor of safety.
or squeezing in weak rock, and the build up of significant A limit state design approach was later adopted as design
swelling pressures behind the shaft lining in anhydritic shales verification. The main limit state design load case considered
and marls when brought into contact with water. For these rea- included the rock load plus the maximum external hydrostatic
sons, thick cast iron tubbing linings with concrete backing are pressure with swelling pressures in swelling rocks added to
often required in certain sections of shaft excavated in such this load case. For this case study though, an extensive pro-
rock formations. gramme of grouting planned meant that external hydrostatic
Long term creep or squeezing in excavations in weak rock is pressures could be excluded from the liner design. Swelling
well known, the usual assumption for very weak rock masses pressures were also only introduced in the numerical models
is that they are incapable of sustaining significant differential to check whether further work will be required to quantify
stresses and that rock mass failure will occur around an exca- swelling potential of rock formations in which the shaft will
vation in such ground until the horizontal ground stress and be constructed.
vertical overburden stress have been equalized. For design, The main design criteria were as follows:
any external groundwater pressures acting on the shaft lining
• The minimum factor of safety against buckling of the cast
must also be taken into account. Significant swelling pres-
iron tubbing had to be more than 2
sures can furthermore develop in anhydritic shales and marl
• Working stresses in the lining had to remain below permis-
when confined during the swelling process, this also need to
sible stresses in compression and tension.
be considered during design.
• Working stresses in the concrete backing were similarly not
The approach adopted in designing the shaft lining in
allowed to exceed permissible stresses at any stage.
squeezing ground at a potash mine in Russia, including numer-
ical modelling verification of liner performance, is presented 3 NUMERICAL MODELLING
as case study.
A model calibration using unfactored material strengths was
2 LINER DESIGN carried out to ensure accurate modelling of the lining response
to ground loading. A typical calibration result is shown in
Parameters reviewed during liner design included shaft geol- Figure 1. From these results, limiting circumferential axial
ogy, intact rock and rock mass strength and modulus, ground stress σN to less than 80% of the characteristic compressive
squeeze expected, groundwater regime, the swelling poten- strength of the cast iron material was found to be critical to
tial of anhydritic shales and marls, horizontal stress, struc- ensure elastic stability of the cast iron material in compression.
tural properties of specified liner components, construction As next modelling step, characteristic ground reaction
sequence, and watertightness of the liner. curves were developed for use as input in 2D models of the
Geotechnical parameters and numerical modelling work various design sections. Ground reaction curves like Figure 2
described relate to specific shaft sections in some of the were obtained from models in which shaft closure as a func-
weaker sedimentary rock formations where severe squeeze tion of internally applied pressure as well as incremental shaft
was expected. Sections of shaft in halites, anhydrites and sinking, was studied at design sections.
sylvinites were specifically excluded from this part of the As final design verification, the cast iron tubbing with
design work. concrete backing was modelled as a 2D ring on a horizontal
641
642
K. Kadiri
Tuc Rail, Belgian Rail Engineering
C. Schroeder
Laboratoire J. Verdeyen, Building Architecture and Town Planning Dept (BATir), Université Libre de Bruxelles
J.C. Verbrugge
Laboratoire J. Verdeyen, Building Architecture and Town Planning Dept (BATir), Université Libre de Bruxelles
Service Mécanique et Construction, Gembloux Agro-Bio Tech (GxABT), Université de Liège
The economic progress that most industrialized countries continuous and not destructive. The principal objective of the
have experienced in the twentieth century as created a sig- work presented here is to test the feasibility to use resistivity
nificant problematic in both waste and water management to measurements to control the permeability and the resistivity
protect the environment. In order to satisfy the demand of in laboratory. Two clays of different mineralogy were tested
environment protection, a clay barrier is mostly used. This in a laboratory experiment using the adapted oedometer. The
barrier is designed to comply with permeability and thick- results demonstrate that under given conditions, a linear rela-
ness requirements. Nowadays, the permeability measurements tionship exists between permeability and resistivity for the
of the barrier are expensive, destructive and in particular two tested clays. An extension of Archie’s law to soils with
long. That is why these measurements and the earth works partly conductive matrix is also proposed and shows results in
progress in the site are incompatible. Meanwhile, geophysics agreement with the experimental data.
methods used for resistivity measurements of soil are fast,
643
Evaluation of the stability of a rock septum between two large derivation tunnels
under high hydrostatic pressure
M.A. Kanji
University São Paulo, Brazil
Keywords: case studies, stability analysis, tunneling, fluid flow, rock mass
1 INTRODUCTION
2 GEOLOGIC SETTING
The geologic characterization of the highly eroded stretch of Figure 2. Analysis with the program PHASE2, showing “Strength
tunnel was made by direct geologic mapping and by analysis Factors” at the rock septum lower than one.
of the geomechanical mapping during construction. The rock
is basaltic flow consisting of sound dense basalt except at the
upper one third of the tunnels where a layer of basaltic breccia The project designer run elastic FEM analysis with the
occurs. SAP and with the PHASE2 programs, utilizing the follow-
There are vertical joints occur at angles normally to the ing parameters: c = 1 MPa and φ = 35◦ for the breccia and
tunnel axis, plus horizontal joints allowing the formation of c = 2 MPa and φ = 50◦ for the dense basalt, and considering
blocks which could fail under the lateral pressure. The UCS of Ko = 1.
the rock is estimated as about 75 MPa for the Basaltic Breccia The result of the SAP analysis indicated that high compres-
and 100 MPa or more for the Dense Basalt. sion occurs at the walls of both tunnels and in the rock septum
in between, and tension occur at the roof of the eroded cavern.
3 ANALYSIS AND COMPUTER PROGRAMS USED The result of the PHASE2 analysis is presented in Figure 2,
showing that the Strength Factor (SF, representing the ration
Different analyses were run, with different programs and between the UCS of the rock and the maximum stresses) is 1
parameters, requiring judgment for their comparison. or less than 1 in the major part of the septum.
644
1 100 100 5 75 0 5
2 100 100 10 10 5 75
645
P. Vaskou
Geostock, Rueil-Malmaison, France
Subject: Underground storage of petroleum, gas, CO2 , and nuclear waste disposal
Keywords: rock caverns, geological model, BHTV, dyke alignment, back analysis, critical joints, support optimization
Engineering geology plays a key role in planning, design and The paper presents two case studies of underground storage
construction of underground cavern storage facilities. Dur- cavern sites:
ing investigation, geological setting of the site is assessed
through an integrated investigation approach such as satel- Case1:
lite image lineament studies, seismic and electrical resistivity During initial investigation of an underground rock cavern
surveys, geological surface mapping, core drilling, bore hole storage site, a major geological feature (a mafic dyke with
tele viewer etc. Based on the above assessment, a geological hydrothermal alteration zone on either side) of about 50 m
model of the site is developed highlighting major geological width was reported. Though, no specific surface exposure
features and discontinuities. The geological model along with was available, based on the interpreted orientation reported
other engineering geological parameters forms the basis for in two widely spaced boreholes, alignment of the dyke
design of the underground facilities such as access tunnel, was assumed for basic engineering purpose. However, dis-
water curtain tunnel, shafts and the main storage caverns. The position of the dyke along with its attitude was planned
layout is finalized in such a way that major geological features to be ascertained through a mandatory investigation cam-
are as far as possible avoided. However, limited unavoidable paign prior to undertaking the detailed engineering of the
geological features which are having the potential to inter- facilities.
sect the underground facilities are predicted in advance and During the mandatory investigation stage, the dyke was
identified as probable geological hotspots. studied in detail by undertaking two more core holes and
The geological assessment of the site is made not only usage of BHTV so as to estimate the thickness of hydro
during investigation but also corroborated and updated dur- thermal alteration zone. Thus, this stage wise investigation
ing construction. In case of cavern storage project, relatively approach helped in scheduling the design stages viz. initial
smaller dimensioned water curtain tunnels, which lie above the investigations for basic design and bid engineering purpose
main storage cavern, are always excavated in advance before and mandatory investigation for detailed design and execu-
excavation of the main large storage caverns. This water cur- tion purpose. Thereby, the geological model of the site was
tain tunnel acts as a pilot tunnel and helps assessment of the finalized along with all the major features/discontinuities and
likely geological setting in advance. As part of the water cur- the probable geological hotspots were identified.
tain system, large numbers of boreholes are drilled from the The site is presently under construction and excavation of
water curtain tunnel and are hydraulically charged to main- the water curtain tunnels is underway. During the excavation
tain a saturated rock mass surrounding the storage caverns. of water curtain tunnels the dyke zone was probed/cored in
Selected boreholes from the water curtain system are cored or advance. The requisite pre-grouting, applicable construction
logged using Bore Hole Tele Viewer (BHTV) or Optical Bore- sequence and special design support requirements of the water
hole Imaging (OBI) revealing additional information about curtain tunnel were finalized. Further based on above active
the geological and hydro-geological setting above the cav- design approach, the geological and geotechnical model is
ern. Thus a 3D geological (structural) model of the site is proposed to be updated and design and construction of the
updated through an active design process with focus on geo- Cavern will be updated and finalized.
logical hot spots intersecting the cavern. During construction
of caverns, a careful approach is adopted along these hot spot Case 2:
segments. These adverse rock mass conditions are probed, In case of another underground storage site, which is near-
assessed and supported in accordance with special design ing completion, a predictive geological model of the site was
supports. prepared based on investigation data and further updated dur-
The geological model is further updated during excavation ing construction of water curtain tunnel and water curtain
of cavern heading and becomes useful for excavation of bench- boreholes through an active design approach. The model was
ing, for verification of the rock support in heading and in further updated during excavation of the top heading of the
further optimization of rock support during bench excavations. cavern. Prior to undertaking the bench excavation, a detailed
646
647
A. Kizaki
Graduate School of Environmental Studies, Tohoku University, Japan
K. Yokoi
Aichi Prefectural Government, Japan
K. Sakaguchi & K. Matsuki
Graduate School of Environmental Studies, Tohoku University, Japan
Keywords: mechanical excavation (TBM/RBM), mining, tunneling, blasting, bolts and anchors
1 INTRODUCTION
4 CONCLUSION
3 RESULTS AND DISCUSSION
In this paper, a self-rotating waterjet nozzle system that has
Figure 2 shows the relation between the rotational torque and a vane pump for braking rotational speed of nozzle system
the rotational speed of the nozzle system for various choke was proposed, and a prototype of a nozzle system was devel-
valves. The rotational speed of the nozzle system obtained oped. The results of the measurement of the rotational speed of
for the nozzle system with the vane-pump type brake system nozzle system show that the developed vane-pump type brake
was much smaller than that obtained for the nozzle system system is effective for decreasing the rotational speed of the
without the brake system. In addition, the rotational speed self-rotating nozzle system.
of the nozzle system was decreased with the diameter of the
choke valve. Accordingly, it is concluded that the developed
vane-pump type brake system is effective for decreasing the
rotational speed of the self-rotating nozzle system.
648
C. Laughton
Laughton Associates, Austin, TX, US
During the latter part of the 20th century, a number of under- Mine located in Lead, South Dakota. The mine produced gold
ground research laboratories (URL’s) were built at depth. for over one hundred years and contains over five hundred
These facilities are accessed via deep shafts or through- kilometres of tunnel and vertical shafts that provide access to
mountain tunnels. As a part of these facilities, large-span depths in excess of two kilometres. At the DUSEL site various
caverns have been constructed. Within these caverns par- engineering and science communities interested in conduct-
ticle detectors, with multiple layers of detection systems ing research underground will be able to take advantage of
mounted around them,have been installed to study the behav- rehabilitated mine excavations (shafts, winzes and decline tun-
ior of sub-nucleic particles such, as the neutrino, proton, and nels) and installed infrastructure to support access to depth and
astrophysics phenomena. provide the requisite operational services.
These sophisticated detectors were necessarily built at depth The paper will review various excavation engineering con-
underground, where rock overburden provides the earth cover cepts being promoted by physics collaborations around the
needed to shield the sensitive detection systems from the dele- world and discuss the key rock engineering criteria by refer-
terious effects of cosmic radiation. Such facilities have been ence tocase history data sets, with an emphasis place on setting
built and operated successfully in Asia, Europe and North realistic design and performance expectations for laboratory
America. In particular, research performed at the Homestake construction at depth. In addition, the paper will review expe-
Mine in the United States of America and the Kamioka Mine rience gained in the access, design, construction, installation,
in Japan, have led to major advances in the fields of particle operation, and decommissioning of deep laboratory facilities
and astro-physics. The value of these experiments was recog- with a specific focus on key design aspects of building larger,
nized through the award of the 2003 Nobel Prize for Physics to deeper caverns.
Dr. Davis and Prof. Koshiba for their pioneering contributions The scale and sensitivity of this new generation of physics
Given the success of this recent suite of deep sited experi- apparatus will place demanding performance requirements on
ments, a great deal of interest has been generated in developing the cavern excavation, reinforcement, and liner systems. In
new experiments in the area of “underground physics”. The particular, experiment proposals for new neutrino experiments
international community is now developing plans for major call for the excavation of caverns ranging in span from 30 to
new particle physics experimental programs. These programs 60 m. The paper will promote the need for rock mechanics
will probe the fundamental behaviours and properties of sub- and engineers to actively support the physics program, and
nucleic particles. The experiments will be long-term,with data support their colleagues in the development of low-risk cav-
collection and analysis scheduled to continue fordecadesafter ern designs that can deliver stable and watertight structures.
commissioning of the detectors themselves. The detectors,and These caverns will need to provide for safe, long-term occu-
auxiliary equipment associated with them, will need to be built pancy of the underground facilities and protect the surface and
at depth underground within large, stable rock caverns. groundwater environments.
Experimental options currently under consideration call Construction of laboratory facilities is expensive, with cost
for the excavation of rock caverns ranging in span from 20 for creation of a cubic metre of laboratory space often run-
to over 60 meters. Some of the more sensitive experimental ning between five hundred to over a thousand dollars per cubic
equipment would need to be housed at depths approaching yard. For the large-scale experiments to be viable, it will also
two kilometres. The larger experiments call for the excavation be critical for the design to incorporate the use of the most
of bank rock volumes on the order of a half a million cubic cost-effective construction techniques. The paper proposes
metres. These dimensions of span and depth are at or beyond that the large, multiphase construction scope of the planned
the conventional limits of rock engineering practice and the new laboratory facilities justifies an early emphasis on state-
design, construction and operation of such facilities will be of-the-industry technologies and the identification of possible
a major undertaking. In the US, these large caverns would research opportunities, implemented project long, from site
be constructed as part of a major new initiative to build a investigation to decommissioning. An early focus on the opti-
multidisciplinary, underground laboratory complex within the mization of engineering practices and integration of research
boundaries of the US Deep Underground Science and Engi- tasking could be of particular benefit to the later phases of
neering Laboratory (DUSEL). DUSEL is being designed to construction most notably with regard to the construction of
fit within the old mine workings of the Homestake Mine Gold the larger caverns, sited at depth underground.
649
L.P. Li, S.C. Li, Z.H. Xu, B. Sui, S.H. Shi, Y. Zhou, C. Hu, Z.Q. Zhou & S.G. Song
Geotechnical and Structural Engineering Research Center, Shandong University, Jinan, Shandong, China
650
REFERENCE
5 ENGINEERING APPLICATION
651
With the high-speed development of infrastructure construc- construction. The challenging problems involved in are as
tion in China, more and more tunnels are and will be built follows: (i) identification and positioning of faults, frac-
in the fields of transportation, hydropower and mining. At tures, karst cavities and underground rivers; (ii) prediction
present, China has become the country in which the largest of water bearing stratum; and (iii) interpretation of prediction
number of tunnels has been built. With the development of result in the context of multiple explanations of geophysi-
infrastructure construction, more tunnels will be built in karst cal prospecting result (Yi-guo Xue, 2008, Qing-song Zhang,
areas in southwest of China, which have the characteristics of 2009).
long tunnel-line, great buried depth and complicated geologi- For the above problems, three measures as follows are taken
cal condition. In addition, the geological condition can not be in this paper. First, the method of risk evaluation for geologi-
detected clearly during engineering prospecting stage. There- cal hazards is presented, which lays a foundation for scheme
fore, the geological hazards such as groundwater inrush and establishing for geological prediction. Second, land-sonar
rock mass collapse are often encountered in construction of method is proposed as a new seismic prediction technique
tunnels in karst areas. in tunnel. the GPR signal response characteristics to water-
On one hand, a lot of tunnels in railway and highway under bearing structures are obtained by theoretical derivation. And
complicated geological and terrain condition is being in con- the judgment criterion to geological anomaly of TSP detection
struction. For example,Yiwan Railway is the most challenging is proposed, by which the level of single prediction technical
railway project in mountain area, which is known as the world- is improved. Third, a system for comprehensive prediction
famous puzzle in aspect of high risk of water inrush hazard of geological hazard sources is developed, which emphasizes
and high karstification level. Serious water inrush hazards on the role of geological analysis and the detection of water-
have taken place in Maluqing Tunnel and Yesangua Tunnel of bearing structures and provides an effective way for geological
Yiwan Railway with serious casualty and economic loss. In prediction during tunnel construction in karst areas.
addition, Qiyueshan tunnel and Wuchiba Tunnel in Hurongxi In this study, the concept that programming prediction is
Highway are constructed in the karst geo-environments. In the based on quantitative risk analysis is put forward. The risk
rock mass of the tunnel, karst cavities and underground rivers rank partitioning principle of karst tunnel should be oriented
are widely distributed. The excavation of tunnel induced the to save time, save costs, reasonable economic and controlling
groundwater inrush, along with mud and fractured rocks. All and humane construction. It don’t add unnecessary workload
of the above mentioned geological hazards can cause great and prevent the water and mud inrush (Yi-guo, 2009; Einstein,
loss of life and property to the stakeholders. 1996; Christian, 2002).
In the hydropower field, the auxiliary tunnel of Jinping According to petrofabric and lithology, geological struc-
hydropower station in Sichuan province is buried 2,375 m and ture, karst, groundwater and other factors, the risk of water
has a length of 17.5 km. The buried depth takes the construc- inrush are divided into 4 levels in karst tunnel construction,
tion of the tunnel into the condition of extremely high ground considering the rainfall and the positive and negative slope
stress level. In this circumstance, the construction of tunnel and other factors. Level I is the highest risk, while level IV is
can be influenced by the possible failure of rock mass, which the lowest risk. The purpose is to refine the range of the geo-
may be triggered by the relief of ground stress in unfavorable logical disasters and take effective prediction methods, which
geological conditions, such as the presence of faults, poor rock may occur in tunnel. The appropriate control systems and early
mass and groundwater (Shu-cai Li, 2008; Tian-bin Li, 2009). warning mechanisms are established, so construction safety is
One of the challenging issues in these hazardous conditions ensured.
is the prediction of the geological hazard sources in front Currently, the methods used in the prediction of geolog-
of tunnel face during tunnel construction. Extremely com- ical hazard sources can be classified into two categories:
plicated geological conditions and higher risk of geological geological investigation and geophysical detection. Geolog-
hazards put forward higher requirements to bad geological ical investigation includes engineering geology analysis, pilot
prediction and controlling technology in tunnel construction. excavation, and core drilling, while geophysical detection
The key technical problems of the theory, methods, tech- covers seismic, electromagnetic and geological radar methods.
niques and equipment of defective geological prediction and these prediction methods are compared in terms of prediction
controlling technology need to be solved urgently in tunnel objective, distance, advantage and disadvantage.
652
653
ABSTRACT:This paper simulated the injection-production cycles of 4-cavern reservoir group under different internal pressure
with small block masonry. The variation of the displacement and strain of the bedrock and the cavern group is automatically
recorded by the displacement meters and the strain gauges buried in the model. Cavern stability under different gas production
process, the failure pressure of the caverns and the impact of mudstone interlayer on cavern stability are analyzed.
654
655
4D seismic monitoring of small CO2 injection – Results from the Ketzin pilot site
(Germany)
C. Juhlin
Uppsala University, Department for Geosciences, Uppsala, Sweden
A. Kashubin
Uppsala University, now at Schlumberger Cambridge Research Centre, Cambridge, UK
N. Juhojuntti
Uppsala University, Department for Geosciences, Uppsala, Sweden
1 THE KETZIN SITE The storage site is located near the town of Ketzin in the
state of Brandenburg, about 20 km west of Germany’s capital
The geological storage of CO2 is debated as one option to Berlin. At this site, natural gas has been stored for thirty years
reduce greenhouse gas emissions and to mitigate the severe gas at a depth of 250–400 m until the operation was abandoned
consequences of climate change. Geophysical monitoring of due to commercial reasons. The Rupelian mudstone served as
storage operations is a key component assuring the safety of cap rock and has demonstrated its suitability for gas storage
the reservoir and providing essential observations enabling operation successfully. The former gas storage formation is
to simulate and predict reservoir behaviour during ongoing separated by an additional 200 m thick clay-rich caprock from
injection and storage. Commercial CO2 storage has been the CO2 storage formation at 600 m depth. The storage forma-
operational at gas field sites like Sleipner and In-Salah for tion is located in the sandstone layers of the Triassic Stuttgart
several years, where time-lapse seismic surveys were able to Formation which consists of sandy channel facies and muddy
image the stored CO2 , amounting to several million tons in flood-plain facies of highly variable reservoir quality. The top
the resepective reservoirs (Michael et al., 2010). seal of the storage formation consists mainly of mudstone,
At the Ketzin pilot site, since June 30, 2008, approximately clayey siltstone and anhydrite, including a persistent, ca. 20 m
18 ktons have been injected per year into a saline aquifer at thin anhydrite layer showing in the seismic measurements as
650 m depth (Würdemann et al., 2010). The Ketzin site is the a clear and bright reference horizon (K2 in Fig. 1). The site
first pilot project for on-shore geological CO2 storage into included industrial land and infrastructure and a gas storage
a saline aquifer in Europe. The site has been developed in permission that enabled an easy development of a testing site
order to enhance the understanding of the underlying physi- for underground injection of CO2 . The operation of the CO2
cal, chemical and biological processes involved in geological underground storage is regulated under mining legislation of
CO2 storage in saline aquifers, to gain practical experience the state of Brandenburg. A maximum of 60,000 t of food
with monitoring and verification of CO2 storage, to test the grade CO2 is targeted to be injected, with the total amount to
predictability of different coupled models, and to verify tech- be adjusted according to scientific and site-specific require-
nologies required for safe long-term underground storage ments. A project internal and an external evaluation of the
of CO2 . The main focus of the project is the development, storage activities will be carried out after injection of every
testing and benchmarking of monitoring techniques, using a 20,000 t CO2 as part of a risk assessment review.
broad range of geophysical, geochemical and microbiological
methods.
Geophysical monitoring tools are deployed to observe CO2 2 RESULTS OF 4D SEISMIC MONITORING
migration in the subsurface. We apply surface and down-hole
measurements to test and optimize the resolution of different Repeated 3D seismic surface measurements are performed
methods and to follow the developing CO2 -plume. in order to observe time-lapse signatures indicating reservoir
656
657
Efficient rock breakage in terms of energy consumption and The results were verified by relevant experiments on con-
dust generation has drawn significant attention. Extensive crete. Three different orientations were chosen. It was found
research has been done to find the optimized geometrical that the radius of the crushed zone along the direction of the
parameters of the cutting picks and to develop more efficient cracks and the maximum compressive stress in the crushed
mechanical fragmentation techniques. Edge chipping test has zone for Orientation (a) where an edge of the pyramid indenter
been regarded as a good fragmentation process because it is is perpendicular to the free surface of a specimen is the high-
highly efficient to generate fracture. Rock cutting can be con- est compared with those other indenter orientations. Therefore
sidered as a series of successive cycles of edge chipping, of this orientation has the worst energy efficiency. The best ori-
which each is composed of a crushing and chipping phase. entation is when a face of the indenter is in parallel to the
This research investigates the effect of pick orientation in the sample free surface.
holder on the efficiency of breaking rocks in terms of cutting
force and energy. The study was carried out by simulating a
series of edge chipping tests with a pyramidal indenter.
658
The soundness diagnosis system of plane concrete lining for railway tunnel in Japan
Railway tunnel inspections are classified into initial, general, the causes of deformation and deciding on countermeasures,
individual and occasional types. The standard maintenance which are general procedures in maintenance and control. To
and control procedure for railway tunnels is shown in Figure 1. do this, it uses deformation data on cracks, water leakages, etc.
Among these, general inspections are periodically executed and detailed data on tunnel interior displacement rates, land-
every two years to identify any deformation and the status of forms, geological conditions, design and construction details,
its progression by means of visual inspection and hammer test- etc. obtained during inspections.
ing. Therefore, such inspections are highly resource-intensive The overall system configuration is shown in Figure 2. After
in terms of time and effort. However, many railway companies inputting basic information such as the tunnel’s specifications,
are now facing a high rate of retirement among their skilled the system indicates the necessary positions for hammer test-
engineers and a reduction in the number of staff with suffi- ing based on an electronic deformation development diagram
cient knowledge of tunnel technologies. Accordingly, there is into which information on cracks and water leakages obtained
a strong need for a method to perform inspections effectively from visual inspection, etc. during a general inspection has
while also ensuring safety. been input. It then diagnoses soundness against flaking (rank
Accordingly, the authors have used past research achieve- α, β or γ) based on the results of hammer testing performed
ments to develop TUNOS (tunnel diagnosis system) – a diag- thereafter. Soundness diagnosis (rank A, B, C or S) for safety
nostic resource that allows automatic execution for most of considerations other than flaking is performed automatically
the work involved in diagnosing tunnel soundness, presuming where possible using the electronic deformation development
diagram, and the results of diagnosis are input manually for
other cases.
The system eliminates issues stemming from differences
between individual abilities, such as variations in the results
of soundness judgment (a common problem in tunnel inspec-
tions), and further improves inspection accuracy as well as the
efficiency of maintenance and control.
Figure 1. Standard maintenance and control procedure. Figure 2. Overall configuration of the system.
659
M. Sueoka
Nexco-Engineering Tohoku Company Limited, Japan
The Sakazuki-yama Tunnel on Yamagata Expressway which In order to understand above phenomena, we calculated
was built in 1991 is 1234 m long. The maximum overburden is deformation and lining stress using FLAC with several con-
about 100 m. The geology of this tunnel is rhyolitic tuff which ditions which represent the behavior.
received the hydrothermally alteration of the tertiary period. Consequently, we conclude the following result of this
At August 13, 2008, there occurred a sudden and huge heav- study.
ing of the road surface at the center of the tunnel, and the
1) The damage of the invert is caused by hygroscopic swelling
transportation of this tunnel was difficult. Uplifting height of
pressure of the surrounding rock.
the road surface reached to approximately 400 mm in ten days
2) Large displacement of the lining and invert heave was
from August 13.
occurred after the destruction of invert.
We investigated road surface leveling, convergence and
3) It is assumed that the cause of the above deformation of the
stress of lining, and evaluated the property of rock material
tunnel is stress-relief with squeezing after swelling.
around the tunnel.
4) It is possible to represent quantitatively using elastic-
From the results of investigation, the amount of reduced
perfect-plastic model that considered the swelling pressure
length of spring line of this tunnel is about 170 mm and lin-
and the reduction of shear strength around the tunnel.
ing compressive stress was −12 MPa and invert concrete was
damaged with open cracks.
660
C.S. Oteo
University of Coruña, Spain
J.M. Maya
Canary Regional Government, Spain
J. Vicente
Acciona Infraestructuras, Madrid, Spain
The San José tunnels (two parallel road tunnels which are both In the Project it was intended to use, in some cases,
19 m wide) were constructed as a road link of the third ring jet-grouting pre-crown (Fig. 2) which were swapped for
road of Las Palmas de Gran Canaria (Canary Island, Spain) micropiles (in view of the presence of pebbles in the Las
with the coastal Avenue. The tunnel (137 m2 section) is devel- Palmas Formation) and several levels of compensation
oped perpendicularly to the island’s coast. The excavation grouting under the houses.
involved (Fig. 1): Bench excavation was commenced, then leaving the arch
The land side mouth did not have any problem (conglom- supported on some small verges (Fig. 3). Slightly before
erate area of the Las Palmas Formation). Then the tunnel was the collapse of the right-hand tunnel, the tunnel construction
excavated using the Bernold method in the advance with a situation was the following:
central bench excavation and carrying out the side walls using
– Advance from the two mouths (after carrying out the sea
the alternate side wall system and concreting each excavated
side tunnel mouth (Fig. 1).
panel (Fig. 2).
– The central bench excavation was excavated along a very
– Antrophic filles and colluvials formations. extensive length.
– Las Palmas Formation: Quaternary conglomerates, – The arch support was being carried out with small benches
cemented, with some loose pebbles (which have average (Fig. 3).
diameters of 1.5 m). – The digging of the open bench excavation allowed the dry-
– Brown and whitish or volcanic tufa (with esmectites). ing of the site in the verges area where the earth was clayey
– Basalt rocks at some inverted arch points (ignimbites) (and expansive) tufs, opening up important cracks therein.
661
662
S. Sakurai
Construction Engineering Research Institute Foundation, Kobe, Japan
This project is named ‘Marmaray’which is a coined word com- To excavate tunnels while keeping safety for the buildings,
bining the sea of Marmara and railway. The Marmaray Project observational construction management is indispensable. It is
is the upgrading of approximately 76 kilometers of commuter important to grasp the occurred phenomena based on the mon-
rail to improve the railway system in Istanbul metropolitan itoring results. The phenomena are traced by means of back
area. Railway tracks will be connected between Asia and analysis and then the ground surface settlement and inclination
Europe through a tunnel under the Bosphorus Strait. Tun- are predicted by calculation for future tunnel excavation based
nels and stations have been and will be constructed by some on the result of back analysis. To reduce the settlement, some
methods such as immersed tunnel, shield tunnel, mountain countermeasures are considered. By analyzing the result, it is
tunnel, and cut and cover in 13.6 km length of this project. able to verify the effect of the countermeasure.
In this report, it is described that the observational construc- In this paper, the control of ground surface settlement
tion management by field measurement is performed in large was paid attention in observational construction management,
scale underground railway station (Sirkeci station) by means which is presumed of the two phenomena for the subsidence.
of NATM. One is the extension of shearing band to the ground surface.
Buildings above Sirkeci station are overcrowded as shown The other is the decrease in pore water pressure of the rock
in Figure 1. around the tunnel. Then, as the countermeasure for the sec-
Bird’s-eye view of Sirkeci station is shown in Figure 2. This ond phenomenon, recharge well was considered and it could
station has two platform tunnels that are abbreviated to PF reduce the decrease in pore water pressure. To confirm the
(about 270 m in length, about 90 m2 in cross-sectional area) effect of recharge well, site test was carried out. As a result,
and center walkway tunnel that is CE (about 152 m in length, recharge well was recognized to be adequate method for the
about 60 m2 in cross-sectional area) between two PF tunnels. control of ground surface settlement based on monitoring
The remaining width of rock between each PF and CE tunnel and analytical results. Therefore additional recharge wells
is just only 10 m. Moreover, platform tunnels connected with have been executed and can result in reducing ground surface
two shafts (about 510 m2 in cross-sectional area) are to be settlement.
equipped for the ventilating system and therefore they have The observational construction management will be contin-
large cross section, which called as PFV-L (about 190 m2 in ued in future.That means that the geological features condition
cross-sectional area). The excavation of this station is one of will be confirmed, the monitoring results will be analyzed,
the most difficult construction works in the world because and the adoption of an appropriate countermeasures will be
of its complicated structure in an urban with many old and executed according to the approrch that described on this
tumbledown buildings. paper.
Figure 1. Condition of ground surface above Sirkeci station. Figure 2. Bird’s-eye view of Sirkeci station.
663
Comparison of field and numerical modelling data for the steel pipe
umbrella arch applications
S.C. Ozer
Department of Mining Engineering, Istanbul Technical University, Istanbul, Turkey
2.1 Case 1
Within the scope of this studies, evaluation of numerical mod-
elling results and field measurements during 2. and 3. Phase
tunnels in İzmir Metro System. Although they have same tun-
nel cross-sections, rock mass properties vary. Plaxis 3DTunnel
V2 is used in numerical modelling. Figure 2. Mesh structure of numerical modelling and deformation
condition.
In the first section of the study, it is investigated whether
2 inches steel pipe umbrella arch application is sufficient in
Ýzmir Metro 2. Phase Construction on Üçyol-Üçkuyular route
at area between Poligon and Güzelyalı. Numerical Modelling
data is given in Tables 1 and 2. Mesh structure of numerical
664
REFERENCES
Aksoy, C.O. & Onargan, T. 2011a. The Role of Umbrella Arch and
Face Bolt as Deformation Preventing Support System in Pre-
venting Building Damages, Tunneling and Underground Space
Technology, doi:10.1016/j.tust.2010.03.004.
Aksoy, C.O. & Onargan, T. 2011b. Determining The Steel Pipe
Umbrella Arch Size by Numerical Modelling in Weak Rock
Conditions: İzmir Metro Case, Madencilik (in Turkish in Press)
Gibbs, P.W., Lowrie, J. Kieffer, D.S. & McQueen, L. 2002. Design of
a shallow soft ground shotcrete motorway tunnel. In Proceedings
of the 28th ITA-AITES World
Hoek, E. 2004. Numerical modelling for shallow tunnelling, Dis-
cussion Paper, p.10, www.rocscience.com/library/rocnews/Spring
2003/ShallowTunnels.pdf.
Kim, C.Y., Kim, K.Y., Hong, S.W., Bae, G.J. & Shin, H.S. 2004. Inter-
Figure 5. Vertical tunnel convergence values obtained from in-situ pretation of field measurements and numerical analyses on pipe
measurements and numerical models. umbrella method in weak ground tunnelling. In Proceedings of the
ISRM Regional Symp. EUROCK 2004 and 53rd Geomechanics
Colloquy (ed. W. Schubert), Salzburg, Austria, pp. 167–170.
models are shown in Figures 4 and 5 respectively. The proper- Miura, K. 2003. Design and construction of mountain tunnels in
ties of rock mass and structural elements are given in Tables Japan. Tunneling and Underground Space Technology, 18 (2–3),
3 and 4, respectively. pp. 115–126.Tunneling Congress, Sydney, Australia.
Onargan, T., Aksoy, C.O., Küçük, K. & Kun, M. 2008. Pre and Final
Support Project of 3rd Stage of Izmir Metro System, 2008.
3 CONCLUSIONS Volkmann, G.M. 2004. A contribution to the effect and behaviour
of pipe roof supports. In Proceedings of the ISRM Regional
In this research, we analysed the different pattern and dif- Symp. EUROCK 2004 and 53rd Geomechanics Colloquy (ed.
ferent size of pipes via numerical modelling technique for W. Schubert), Salzburg, Austria, pp. 161–166.
665
K.K. Panthi
Department of Geology and Mineral Resources Engineering, Norwegian University of Science
and Technology (NTNU), Trondheim, Norway
There are mainly three engineering geological uncertainties through four categories of rock sequences consisting of biotite
that directly influence on the stability of a tunnel or an under- schist, carbonaceous phyllite, Manikaran quartzite and schis-
ground cavern. These are rock mechanical properties, in-situ tose granite gneiss (Figure 1). The biotite schist is intercalated
stress conditions and groundwater inflow through fractures with small bands of quartzite schist. The carbonaceous phyl-
and weakness zones. The priorities given in dealing with these lite is deformed and folded. The schistose granite gneiss is
uncertainties may vary from project to project. In regards occasionally intercalated with small bands (1–10 meters in
with tunnelling in the Himalaya, due to high tectonic activ- thick) chlorite/talcose mica schist. However, the Manikaran
ity, the rock mass in the region are fractured, faulted, sheared quartzite is relatively fresh and massive in most of the places
and deeply weathered. Tunneling through numerous zones of excluding fractures and weakness zones. The headrace tunnel
weakness, fractures and faults is thus a matter of reality. Major- crosses several major and minor weakness zones.
ity of long tunnels meet several difficult stability problems According to the Norwegian rule of thumb (Selmer-Olsen,
and excavation challenges. Hence, selection of right types of 1965) rock spalling or rock burst is likely to occur once
excavation methodology based on sufficient engineering geo- rock cover above the tunnel exceeds approximate threshold
logical investigations during planning is crucial for successful of 500 m. The extent of this type of failure is even severe
completion of tunnel projects and also to reduce uncertainty if tunnel runs parallel to the valley side with a slope angle
associated with delay in completion and cost overruns. exceeding 25 degrees. In case of Parbati II, the rock cover
One of the major areas where instability problems are faced along 40 percent headrace tunnel length exceeds this thresh-
while tunneling through Himalayan rock mass is the instability old. Approximately 9.5 km headrace tunnel between chainage
related to rock stresses. One of such tunnel cases is the Parbati 4.5 km to 16 km exceeds 500 m rock cover. The main rock
II hydroelectric project under construction in the Kullu district formations along this tunnel reach are biotite schist, phyllite
of Himachal, India. This is a run-of-river scheme under con- and Manikaran quartzite.
struction since 2002 and has an installed capacity of 800 MW Even though headrace tunnel runs through steeply dipping
and utilizes a gross head of 862 m. The river water is being valley side slope and rock cover is exceeding the threshold
planned to be diverted from near village Pulga by construct- of 500 m in many locations up to chainage 7.5 km (Figure 1)
ing a 85 m high and 113 m long concrete gravity dam. The no noticeable rock spalling or rock burst occurred. The exca-
water will then be conveyed through a 31.37 km long head- vation went smoothly with no major stress related instability
race tunnel and through twin inclined penstock pressure shafts observed excluding minor squeezing. The headrace tunnel up
(1.54 km length each) to the semi-underground powerhouse. to this chainage mainly passes through biotite schist and a
Geologically, the project area is located in so called ‘Kulu small band of carbonaceous phyllite between chainage 7.4 km
Window’ that represents the Lesser Himalayan rock forma- to 7.5 km.
tion and the area is bounded by a major fault system in the However, after chainage 7.5 km brittle and massive
Himalaya called “Main Central Thrust (MCT)”. The Kulu Manikaran quartzite appeared. In this quartzite severe rock
Window mainly comprised by rocks from crystalline to meta- burst occurred continuously in valley side roof of the headrace
sedimentary sequence. The headrace tunnel passes mainly tunnel after almost every alternate blasting round.
Figure 1. Geological cross-section of the headrace tunnel giving two study sections (modified from Panthi, 2009).
666
667
668
669
670
671
Keywords: case studies, field measurements, numerical modelling, problem rocks, rock stress, tunnelling
672
the stresses in the steel reinforcement are far from the elastic
limit of the steel.
REFERENCES
673
Jonny Rutqvist
Earth Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, US
Deep underground injection of carbon dioxide (CO2 ) has At Krechba, satellite-based interferometry (InSAR) is used
emerged in recent years as an important option for seques- to monitor ground-surface deformations with remarkable
tering CO2 and thereby reducing the emission of greenhouse precision (a few millimeters), because of favorable ground-
gases to the atmosphere. Because CO2 is less dense than surface conditions, including hard desert sediments and bare
water, the targeted CO2 injection reservoir should be overlain rock. Inverse semi-analytical strain analysis as well as coupled
by low-permeability caprock formations that can sufficiently fluid flow and geomechanical numerical modeling are being
limit upward buoyancy-driven flow of the injected CO2 . Rock employed to interpret observed ground-surface deformations
mechanics plays an important role in the performance assess- in terms of underground fluid movements and structures, such
ment of such a CO2 injection site, including assessment of as faults and fracture zones. Significant modeling effort is cur-
caprock sealing performance, ground-surface deformations, rently being dedicated to analyze ground-surface uplift at one
and the potential for fault reactivation and induced seismicity. of the injection wells, where a double-lobe uplift pattern indi-
As a practical example of geomechanical aspect related to cates opening of a deep fault or fracture zone—a feature that
CO2 injection, this paper reviews recent results on model- has also been detected in a recent 3D seismic survey. Finally,
ing of geomechanical aspects associated with the In Salah coupled fluid flow and geomechanical modeling has been
CO2 storage project, at the Krechba gas field, Algeria. The In applied to estimate the potential for injection-induced micro-
Salah CO2 storage project can be considered an ideal case for seismicity. The analysis showed that the highest potential for
studies of rock mechanical aspects, because injection takes injection-induced microseismicity occurs along the horizontal
place at an injection pressure substantially higher than ambi- sections of the injection wells as a result of the stress changes
ent hydrostatic fluid pressure, and ground-surface conditions caused by the combined effects of injection induced cool-
are very favorable for monitoring of ground-surface defor- ing and pressure. However, for the best-estimated present-day
mations. Moreover, the targeted injection zone is a fractured strike-slip stress regime at Krechba, the analysis indicated a
sandstone layer intersected by minor faults and fracture zones. relatively low potential for injection-induced microseismicity.
This paper summarizes and reviews recently published work, The geomechanical models are being updated and recali-
including modeling of ground-surface deformations, potential brated as new field data becomes available, and the long-term
fault reactivation, and induced seismicity. geomechanical responses will be assessed.
674
The integrated prediction system for geological conditions ahead of tunnel faces
Y. Mito
Department of Urban Management, Graduate School of Engineering, Kyoto University, Japan
M. Ashraf
School of Civil Engineering, University Sains Malaysia, Malaysia
It is essential to employ the observational design and con- short-interval prediction subsystems for the several engineer-
struction method to manage tunnel projects efficiently. The ing purposes in the framework of observational design and
original tunnel design is immediately verified or modified construction in the tunnel project. Geostatistical method is
based on the in-situ data sequentially obtained in the con- employed to obtain not only the spatial distribution of geo-
struction stage. The authors develop the integrated prediction logical conditions but also the precision of prediction. The
system for geological conditions ahead of a tunnel face. The applicability of the proposed subsystems is verified through
system is composed of the long-interval, middle interval, and the field application.
675
Y.K. Lee
Department of Coastal Construction Engineering, Kunsan National University, South Korea
Keywords: compressed air storage, concrete plug, finite element analysis, contact pressure, factor of safety
676
677
Salt and rock mechanic challenges at the verification of drift seal systems at the
Morsleben repository, Germany
J. Stahlmann
TU Braunschweig, Germany
R. Mauke
Bundesamt für Strahlenschutz, Germany
M. Mohlfeld
Zerna Ingenieure, Germany
C. Missal
TU Braunschweig, Germany
Keywords: nuclear repository, case studies, field measurements, fluid flow, monitoring, problem rocks
1 ABSTRACT
678
TBM vs. D&B – a difficult choice in mountain terrain – some geotechnical guidelines
Keywords: blasting, mechanical excavation (TBM/RBM), risks and hazards, rock bursts, rock mass classification, site
characterisation, tunnelling
Tunnelling at depth through high mountainous terrain poses complex geology and stress conditions when mining at these
difficult challenges not just for tunnel boring machine (TBM) significant depths. Extremes of ground conditions present
drivages but also for application of traditional drill and blast major contrasts to tunnelling and they often demand use of
(D&B) methods. Dealing with adverse geology at any depth flexible rock engineering solutions in order for the tunnel to
can be problematic and can lead to significant tunnelling progress. This need to adopt flexible solutions is often seen as
delays if not adequately foreseen; but geological problem con- being at variance with the constraints imposed by the rigidity
ditions, which might be tractable at shallow depth, with either of design elements incorporated into the fabrication of a typ-
TBM’s or D&B approaches, when encountered at significant ical TBM. Traditionally, with some exceptions, TBM’s have
depth (>1000 m) can prove disastrous depending on stress only been utilized on tunnel projects where the risks associated
state, rock competence and prevailing groundwater inflows. with deep tunnelling are low, or expected to be low. This risk-
Mitigating delay problems associated with exceptionally bad aversion to use of machines in these conditions has generally
ground at depth requires considerable foresight and advanced been due to the perceived adverse consequences of entrap-
planning. The challenge is not just one of tackling the adverse ping or damaging a TBM under such conditions; in some part
ground condition, by modifying the excavation and support due to the relative difficulty of dealing with adverse ground
processes in order to deal with the specific problem zone, in the confined working area of a TBM, in comparison to a
stress state and groundwater conditions, it is also often about D&B/NATM heading.
logistics, as all too often for deep tunnels in mountainous Hard rock machine designs are, however, moving forward
regions the problem geologic zones are at significant dis- to more hybrid, universal concepts that encompass ability to
tance from the nearest portal, and at such significant depth execute full umbrella forepoling and soft rock machine face
that surface pre-treatment is generally impractical. conditioning, pre-grouting and ground treatment philosophies
Traversing faulted and disturbed ground at significant depth in an attempt to combat some of these problems by making
requires that tunnelling procedures be able to cope with a huge the machines sufficiently robust and at the same time flexible
range of difficult geological conditions. Investigating, evaluat- enough to be capable of safely and successfully excavating
ing and assessing anticipated geology ahead of tunnelling, and through extremely bad ground.
dealing with encountered difficult ground conditions requires Extremes of bad ground experience are cited from tun-
that better understanding be gained of the interaction between nelling in the Himalayas, Andes and Alps where a variety
of difficult ground conditions have had to be traversed, rang-
ing from rock-burst-prone conditions with hydrothermal water
inrushes through to coping with squeezing, very soft, soil-
like fault zone infills. Based on case records of tunnels,
geotechnical guidelines are presented for aiding the choice
of tunnelling method and for guiding the selection of appro-
priate equipment for dealing with the wide range of varying
ground support requirements expected at significant depth in
mountain terrain.
Geotechnical guidelines are presented for characterizing
and evaluating of the three main elements that limit our
ability to execute trouble free tunnels at significant depth –
stress state, groundwater conditions and rock itself (Figure
1). As explained with case examples in Carter et al, 2008,
based on Hoek and Marinos, 2000 for squeezing ground and
Diederichs, 2007 for spalling ground we now have methods
available to characterize each. Adverse characteristics of any
one of these elements can, on its own, compromise D&B or
TBM tunnelling but it usually takes a combination of all three
Figure 2. Primary Risk Elements – Water Jw, Stress SRF and Rock being adverse to trap a machine or halt a drill and blast. Rec-
Quality Q, GSI or RMR and Geotechnical Definition Measures. ommendations are thus included in the paper aimed towards
679
680
Peter H. Smeallie
American Rock Mechanics Association, Alexandria, VA, US
Xia-Ting Feng
Institute of Rock and Soil Mechanics, The Chinese Academy of Sciences, Wuhan, China
John A. Hudson
International Society for Rock Mechanics and Imperial College, London, UK
Keywords: case studies, field measurements, nuclear repository, rock caverns, site characterization, tunneling
Scientists and engineers in underground research laboratories in heterogeneity and fracture network characterizations, in
(URLs) have shared many common objectives and exchanged coupled process testing in controlled blocks, in quantification
information on experimental designs, sensor improvements, of induced deformation and seismicity potential, to coupling
field procedures, and research findings. The enhancement of underground signals to atmospheric and ionospheric pro-
and development of URL networks can further encourage cesses. While there are clearly site specific characterizations
multidisciplinary and cross-disciplinary research aimed at and experiment specific requirements unique for each URL,
technological invention and innovation to meet energy and there are common and new approaches that network exchanges
environmental challenges. Discussions in symposia of the can contribute. In addition to fundamental challenges, we
International Society for Rock Mechanics, American Rock also address practical aspects of organization, procedure, and
Mechanics Associations, and in sessions of the American incentive needed to develop and formulate URL networks.
Geophysical Union meetings, led to development of the URL In this paper, we first summarize the establishment and evo-
Workshop associated with the 2011 Beijing ISRM Congress. lution of URLs for radioactive waste research. Concurrently,
In this presentation, we summarize broad overviews and spe- there are underground laboratories with historical focus on
cific findings from lectures on URLs and panel discussions physics experiments with detectors in mines and along tun-
on what we have learned from existing URLs, what the main nels. In this paper, we use URL for both radioactive waste
problems are facing us today that can use existing URLs and and physics experiments. We then discuss examples of ongo-
require new URLs, and how valuable are virtual URLs linked ing efforts in URLs that have the potential for innovations
to real URLs. Lectures are planned for the URL workshop from interdisciplinary studies enhanced by URLs. The URL
on URLs in hard rock with varying degree of fracturing, in environments are quieter than laboratories on and above the
soft rock with plastic deformation, in shallow and in deep ground. We then discuss questions of URL networking. This
laboratories for physics and multidisciplinary research, in use paper presents an update of a review and evaluation on URLs,
of URLs for underground injection assessment, in search for and discusses trends and consensus from some recent papers
origins of life from microbes found in isolated water pockets, and presentations.
681
The state-of-the-art and research advances on the underground natural gas storage in
bedded salt rocks in China
C.H. Yang
State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics,
Chinese Academy of Sciences, Wuhan, China
Key Laboratory for the Exploitation of Southwestern Resources and the Environmental Disaster Control Engineering,
Chongqing University, Chongqing, China
Y.P. Li
State Key Laboratory of Geomechanics and Geotechnical Engineering, Institute of Rock and Soil Mechanics,
Chinese Academy of Sciences, Wuhan, China
D.A. Qu
West-east Gas Pipeline Company Gas Storage Project Department, Petro China Company Limited, Beijing, China
Key Laboratory for the Exploitation of Southwestern Resources and the Environmental
Disaster Control Engineering, Chongqing University, Chongqing, China
Keywords: rock caverns; stability analysis; rock properties; mining; numerical modeling
682
683
J.Y.K. Wong
Defence Science & Technology Agency, Singapore
An underground cavern complex is usually built in a com- of Universal Distinct Element Code (UDEC) in this paper.
plex geological environment and may be subjected to different The following stochastic variables are considered: the fric-
loading conditions. In view of this, the major factors influ- tion angle, the cohesion, the deformation modulus of the rock
encing the cavern performance such as the stiffness of the mass and the in-situ stress ratio. The overburden thickness,
surrounding rock, properties of the rock mass, overburden the unit weight of rock materials, the Poisson’s ratio and the
thickness, the magnitude and direction of the in-situ geo- joint strength are assumed as deterministic. The cavern width
stress, the size and shape of the underground cavern complex, and height are also assumed as uncertain variables in order
the type and magnitude of the loading as well as the sequence to optimize the shape of the rock cavern. Most cavern shapes
of excavation and support must be taken into consideration. are horse-shoe or bullet-head shaped. It is widely argued that
Modern design codes emanate from a probabilistic approach, the use of a flat-arch cavern would make the best use of cav-
for which either the probability of failure (or reliability index) ern space. Thus for this study, the initial cavern width and
can be used to quantify these uncertainties. Advanced numer- height are set as 30 m and 18 m, respectively. The influences
ical methods, which may be continuous or discontinuous, of the flattening process on cavern stability can be investi-
are commonly used in the design of rock caverns. Once the gated through incremental increases in the cavern width and
performance functions are obtained through numerical exper- reduction in the cavern height through six design levels to
imentations and distribution types and characteristic values assess the changes of safety factor and probability of failure.
are determined, several reliability-based design methods can For each configuration of the six levels, the probability of fail-
be used to assess the probability of satisfactory performance. ure is determined by Monte Carlo simulation incorporated by
Combining stochastic approaches with advanced numerical neural network results. The configuration satisfying the criti-
methods will not only provide a systematic evaluation of the cal safety factor and the expected performance level with the
uncertainties but also a means for assessing the impact of these flattest cavern roof can be termed as the optimal design. It
uncertainties on the likelihood of satisfactory performance of is also suggested that the critical factor of safety and the tar-
rock caverns. geted performance level be used together, as complementary
The global stability of underground rock cavern exca- measures of acceptable design.
vated in a Mohr-Coulomb material is investigated by means
684
X.P. Zhou
School of Civil Engineering, Chongqing University, Chongqing, China
Key Laboratory of New Technology for Construction of Cities in Mountain Area, Chongqing University,
Ministry of Education, Chongqing, PR China
Q.H. Qian
Engineering Institute of Engineering Crops, PLA University of Science and Technology, Nanjing, China
Keywords: stability analysis, rock support, tunneling, rock failure, numerical modelling
During the excavation, deformation and failure of surround- have not converged. From the current situation in excavation
ing rock masses around underground caverns of Jinping I site of surrounding rock masses, in the powerhouse, including
hydropower station are analyzed by the numerical simulation main transformer chamber lumbar arch, the sidewall of the
as well as the internal force of rock bolt and presstress anchor traffic tunnel and other parts, there exists the circumferential
cable. Effect of cracks and joints on deformation and failure of deformation and failure characteristics of surrounding rock
surrounding rock mass around underground caverns of Jinping masses after excavation from a high in-situ stress.
I hydropower station is investigated. The stress and displace- The deformation and failure mechanism of crack-
ment fields of surrounding rock mass around underground weakened rock masses around underground caverns in Jinping
caverns of Jinping hydropower station I is obtained as well hydropower station I, which located in a high in-situ stress
as the internal force of rock bolt and presstress anchor cable. area, are numerically simulated by the Extended Finite Ele-
Deformation and failure mechanism of surrounding rock mass ment Method in this paper. The deformation and failure
around underground caverns of Jinping hydropower station mechanism of surrounding rock mass around the under-
I are revealed. Comparison between the numerical results ground caverns is investigated. Finally, comparisons between
obtained from numerical results and monitoring data is carried results from numerical simulation and the monitoring data
out. It is shown that numerical results are in good agreement are made.
with the the monitoring data. Jinping hydropower station I underground powerhouse is
Jinping hydropower station I underground powerhouse located in right bank of the mountain with about 350 m far
mainly consists of the main and auxiliary powerhouse, main away from Dam downstream, in which there are complete
transformer chamber, tailrace surge chamber, penstock, traffic exposed bedrock and steep terrain near the river. Below the
tunnels and vertical shaft. These underground tunnels directly height above sea-level of about 1770 m, the slope angle is
impacts on project investment and construction period. Since about 70∼90◦ . However, above the elevation of about 1770 m,
late April 2008, some spray layered cracks have been detected the slope angle is about 40◦ . Underground caverns are char-
at the downstream side of the crown in the powerhouse, and acterized as intensive, huge and high walls of the main
these cracks continuously develop. In August 2008, crack- cavern, large span and complex construction conditions. The
weakened area has extended to the whole lower part of roof underground powerhouse is located between the exploration
arch. When the fourth layer of the main transformer chamber lines I and VI near PD27 exploration adit whose horizontal
is excavated, cracks occurred in the downstream lower arch depth is about 100∼380 m, and whose vertical depth is about
area, and the cracks developed rapidly. In December 2008, the 160∼420 m. Maximum principal stress near the underground
spray layered cracks appeared in almost all of the downstream powerhouse is 20∼35.7 MPa, the intermediate principal stress
side of main transformer chamber (including the exhaust fan near the underground powerhouse is 10∼20 MPa, the min-
room). Tension cracks are found at the downstream sidewall imum principal stress near the underground powerhouse is
of the traffic tunnel with the length of 3∼4 m. In March 2009, 4∼12 MPa. The direction of the maximum principal stress
the largest open displacement of cracks reaches 10 mm. In varing from N28.5◦ W to N71◦ W with an average value of
addition, in each hole with the vertical layout of underground N48.7◦ W is found. Dip angle of the maximum principal
powerhouse, such as the traffic tunnel, circumferential cracks stress is about 20∼50◦ with an average angle of 34.2◦ . The
are also found. According to the monitoring data and com- underground powerhouse is located in high stress area.
prehensive test data of acoustic emission, rock deformation Underground caverns of Jinping hydropower station I
and failure range were significantly greater than other similar passes through the second, third and fourth layer of mar-
depth and scale of the underground powerhouse. Moreover, ble. The second layer not only contains medium thickness
high proportion of rock bolts and anchor cables are overload- or thickness marble, but also contains thin marble and
ing. Till now, some important parts of the deformation of rock green schist. The main part of third layer contains mainly
685
686
Subject: Rock material and rock mass property testing (laboratory and in-situ)
Keywords: field measurements, anisotropy, deformability, dilatometer, modulus of deformation, joint sets
689
690
H. Dellero
Department of Geology, Faculty of Sciences, University Adbelmalek Essadi, Tetouan, Morocco
Keywords: risks and hazards, mining, case studies, rock slopes and foundations, site characterization
1 INTRODUCTION
691
4 CONCLUSIONS
692
Seismic stability analysis of rock slopes using the numerical manifold method
G.W. Ma
School of Civil and Resource Engineering, University of Western Australia, Crawley WA, Australia
Z.Y. Zhao
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
Seismic stability of rock slopes is numerically investigated φ (friction angle) > α (slope inclination angle), and unstable
using the numerical manifold method (NMM). A paral- for φ < α. This study investigates the horizontal ground accel-
lelepiped block resting on an inclined plane as shown in eration induced block instabilities based on these two cases.
Figure 1, which has been considered as a first approxima- Two different instability mechanisms are observed. The hor-
tion of the rock slope stability conditions, is employed in this izontal ground acceleration enlarges the destabilizing force,
study to find out the fundamentals of seismic instabilities of which makes the originally moving block (i.e., φ < α) move
rock slopes and to validate the NMM for seismic instability faster (e.g., Figure 4 for a block under a sinusoidal function)
analysis of rock slopes. and produces the originally stable block (i.e., φ > α) a finite
The block under gravitational loading only is first con- displacement (e.g., Figure 5 for a block under a sinusoidal
sidered to calibrate the NMM and to determine the control function & Figure 6 for a block under an arbitrary ground
parameters for the NMM calculation. A regularly-patterned acceleration record).
triangular mesh is used to generate the NMM models. The A vertical ground acceleration is lastly introduced and
proper mesh size is first determined by gradually refining its induced rock slope instability mechanism is discussed.
the mesh until the predicted displacement close to the ana- Consider a half-sine vertical ground pulse acceleration with
lytical solution. To this end, totally three different meshes are amplitude of a0 and duration of T. We assume φ = 30◦ and
investigated. Numerical results indicates that for such a sim- α = 25◦ . The block is initially stable since φ > α. However,
ple problem without high stress gradient, a very coarse mesh when a vertical pulse acceleration is introduced, the effec-
as shown in Figure 2 will produce an accurate result (with tive friction angle φ̃ vibrates around the static friction angle
an error of only 1.94%). The proper time step size t is then φ. For the case of a0 = 0.8 g, T = 0.8 ms, the φ̃ has a mini-
determined by being gradually reduced until the predicted dis- mum value of 6.71◦ (Figure 7), much smaller than φ, which
placement close to the analytical solution. To this end, totally is termed as ultra low friction phenomenon. The block moves
six different t, i.e., 0.01s, 0.008s, 0.005s, 0.002s, 0.001s when α > φ̃. Vertical pulse acceleration makes the originally
and 0.0001s, are considered. The errors to the analytical solu- stable block unstable and provides the block a finite displace-
tions of the displacement are 42.57%, 15.39%, 7.85%, 4.70%, ment. Its effect highly depends on the duration and amplitude
2.22%, 1.94%, respectively. Although when t = 0.002s, the of the pulse acceleration. Various vertical pulse accelerations
NMM predicted displacement is already acceptable within with the same amplitude but different pulse acceleration are
an engineering accuracy, we choose an even smaller t of
simulated. Figure 8 plots the variation of minimum φ̃ with
0.0001s throughout this study in order to achieve more accu-
the duration T. It is observed that there exists a critical dura-
rate results. Using the determined control parameters, five
tion, Tc . The minimum φ̃ decreases with the duration before
cases with different friction angles, i.e., 0◦ , 5◦ , 10◦ , 15◦ , 20◦ ,
are studied. The NMM predicted horizontal displacements
agree well with the analytical solutions for all the five cases,
as shown in Figure 3. Validity of the NMM for gravity induced
displacement is calibrated.
Horizontal ground acceleration is then introduced and its
induced rock slope instability mechanism is investigated.
When only under gravitational loading, the block is stable for
693
694
Keywords: numerical modeling, rock slopes and foundations, risk and hazard, rock support, site characterization, stability
analysis
695
696
Ya-Chu Chiu
Department of Civil Engineering, National Taiwan University, Taipei
Tsan-Hwei Huang
Department of Civil Engineering, National Taiwan University, Taipei
1 INTRODUCTION
2 STUDY AREA AND METHODOLOGY Figure 1. Topographic map of the study area.
Figure 1 shows the topography of the study slope. The rock of modern geodetic total station to acquire 3D global coor-
nearby the slope is mica schist with schistosity dipping to dinates of the points on tunnel walls, with a precision of
northwest. A single-lane highway tunnel built in 1972 suffered 2–3 mm. Displacement of each profile is then obtained. Also,
from significant squeezing after completion of construction, the surface condition of tunnel lining walls is recorded by
and a double-lane tunnel was built as a substitute for the col- image-mosaic technology through photography and image
lapsed one and completed in 1991. The slope west of the tunnel processing. Finally, discussion about site characteristics and
is active. Lining cracks developed approximately two years associated tunnel displacement are made.
after tunnel completion. Repair measures were implemented
in 1998. Surface geological survey has been done since early
1970’s. However, the instability of the slope was recognized 3 INVESTIGATION RESULTS AND DISCUSSION
until 1992 after the completion of the existing tunnel. Since
that a series of field investigations and monitoring have been Six aerial photos taken through 1981-2006 are used to identify
implemented. Rock formation is than revealed from results the multitemporal ground variations on study area. Aerial pho-
of 17 boreholes and geophysics investigation. And monitor- tos interpreting results show that the instable ambits were local
ing records of inclinometers and ground displacements are and associated sliding surfaces were not deep initially. The
available for the past two decades. construction of a forest road and reclamation activity resulted
First, multitemporal aerial photos are utilized to interpret in the enlargement of instable ambits. Bank erosions leaded
the possible ambits of movement masses. Field investiga- to an erosion gully at a corner of the slope toe, which wors-
tions focusing on phenomena of slope movement are carried ened the slope stability by means of headward erosion. The
out to confirm the ambits. The deformation of the exist- headward erosion is sped up by the steep gradient the water
ing tunnel is monitored using a micro-displacement survey flowing down as well as the schistosity dipping down, and
technology which integrates Differential Global Positioning reduced the stability of the slope above the highway. Addi-
System (DGPS) and Automatic Target Recognition (ATR) tionally, the affecting area caused by the slope movement near
697
698
A. Shahinuzzaman
Powerlink, Queensland, Australia
699
Field test and interpretation of the side shear resistance of a drilled shaft in rock
A static load test was performed on a drilled shaft in a flysch Considering those relations, their values should be as close
rock. The properties of the rock mass were extensively inves- as possible to a safety factor of 1.00, so it can be concluded that
tigated. The side shear resistance of the drilled shaft has been these theories are very conservative. This means that whenever
carefully studied during the load test. Two Osterberg cells were a pile is designed using these expressions, an additional Safety
embedded into the pile, along the 5 meter pile segment in the Factor (SF) is introduced in the analysis in addition to the one
good quality flysch rock, in order to introduce the load. Using considered for the pile. The synthesis of this comparison is:
laboratory test results, analyzing the field load test and mod-
eling the drilled shaft by the finite element method, the side
shear resistance transfer mechanism between the shaft and the Value
rock was reproduced. Author <MPa> SF
The test was carried out using two Osterberg cells embedded
in the pile, as have been said. The cells were separated from Rossenberg and Journeaux, 1976 2.45 1.43
each other. One was situated at half the tested height of the Horvath and Kenney, 1979 1.45 2.41
drilled rock shaft and another one near the base of the pile. Williams and Pells, 1980 1.77 1.97
The test was planned in two stages; during the first one only Rowe and Armitage, 1986 4.15 0.84
the bottom cell loaded the pile; while in the second stage only Carter and Kulhawy, 1988 1.38 2.53
the upper cell worked.
In addition to the results obtained from the load test by
analyzing settlements and mobilized shear forces, a review of A FEM analysis has been done, modeling the rock socket as
some of the existing empirical criteria to estimate the ultimate Horvath et al. (1983) proposed. This criteria causes the shaft-
side shear resistance of a pile has been presented. rock wall to be modeled as a toothed surface. According to
The following are conclusions from the load test results: the UCS of the rock and the excavation method, the shaft-
rock contact was modelled as a rough surface. This model has
– Neither the pile nor the rock failed during the test. allowed the load transfer mechanism to be represented from
– The analysis of the side shear resistance during the first the initial stages of loading to the failure. Analyzing FEM
stage of loading was complex due to the interaction between results it can be observed that the failure value obtained from
the base and the shaft of the pile. the analysis is very accurate to the one obtained in the load test.
– During the second stage of the test, only side shear resis-
tance was mobilized. Although the ultimate side shear
resistance could not be reached it was deduced that the
strength was near 3.5 MPa. The estimated value for ultimate
unit friction at failure obtained by the load test was com-
pared to the results obtained by the empirical relationships,
depending on the rock characteristics.
700
What modern rock mechanics owe to the Malpasset arch dam failure
Pierre Duffaut
French Committee on Rock Mechanics, Paris, France
Keywords: case studies; weak rock; field measurements; rock joints; stability analysis; risks and hazards
1 INTRODUCTION
2 THE FAILURE
Designed for water storage by famous Consulting Engineer Figure 1. Remains of Malpasset arch dam: a conspicuous crevice
A. Coyne, and built 1952–1954 in a narrow gneiss gorge, the appears along the concrete shell; the whole river discharge is flowing
60 m high dam has never been filled up before the occurrence below the dam (photo P. Duffaut, 1960).
of huge rains in November 1959. The failure was assigned to
water pressure through a unique combination of features:
3 MAIN LESSONS
i) high sensitivity of rock permeability to stress, creating a
deep impervious barrier below the dam; Two traps had mistaken the people in charge: the rock mass
ii) high deformability of rock, helping open a deep crevice “look”, as seen by many visitors, was treacherous, and the
behind the dam heel (Fig. 1); administrative status was rather low; nobody had realized this
iii) a fault dipping upstream ca.20 m downstream of the dam; dangerous object had been left without convenient control.
iv) lack of boreholes for foundation drainage. The main lesson from Malpasset failure was that, inside a
It must be added that the monitoring and supervision of the dam foundation, water forces may be the same order of mag-
dam were rather poor, as it was left unused by lack of any nitude that dead weight and dam thrust. All dams are only
water distribution network (nobody noticed a slight move of gravity dams under condition to include the weight of ground
the dam towards left side); lastly, the construction works of a abutments and engineers are the only ones, in the geotechni-
motorway bridge was glad no water could flow through. cal field, to be fully aware of water force, a true revolution
For having taken part himself in the investigations on many following the long time needed to understand uplift inside
dam sites in France at the same time (1948–1955), the author rock mass.
bears witness that the rules of art had been followed: As stated Before, only mining engineers had begun to formalize a
the technical literature, main dam site problems were bedrock meager corpus of rock mechanics. Since, civil engineering
depth below the valley floor, and bedrock imperviousness, Of and geological schools, labs and bureaus, as well public as
course all rules were quickly changed after the event: more private, developed an array of new concepts and methods,
site investigations became mandatory, how sound the rock which first appeared at the 8th ICOLD conference in 1964
may look; uplift drainage under gravity dams was extended to and later at ISRM ones, beginning in Lisbon, 1966. Civil engi-
arch dams; and the State control on all dams over 15 m height neers met petroleum and mining engineers to build together a
was reinforced. comprehensive Rock Mechanics.
701
Keywords: numerical modeling, rock failure, rock joints, rock mass, rock slopes and foundations
The ultimate bearing capacity is an important design consider- for determining the bearing pressure from the load-settlement
ation for heavy structures such as concrete dams, bridge piers, curves. The ultimate bearing capacities obtained in this man-
and multistory buildings in jointed rock masses. Because of ner are compared by the values obtained from an upper bound
existence of discontinuities within the matrix material, the limit analysis formulation developed in this paper. Results of
bearing capacity of rock masses is governed not only by the this comparison show that by increasing the degree of joint-
intact matrix material but also by the characteristics of the dis- ing, the ultimate bearing capacity of rock foundations will
continuities. According to the joint spacing in comparison to decrease. Also, the bearing pressure values obtained by the
the structure dimensions, two approaches called the direct and upper bound method are greater than the values obtained by
the homogenization approaches are applied for determining the distinct element using the Log-Log method.
ultimate bearing capacity. The effect of groundwater table on the ultimate bearing
In this paper, for the case of a non-homogeneous rock foun- capacity of rock foundations was also investigated. Using the
dation with large joint spacing, the ultimate bearing capacity distinct element method, different positions of water table were
is investigated using distinct element method. The distinct ele- analyzed and the reduction of ultimate bearing capacity was
ment modelling are performed for rock masses containing two investigated. It is found that in the cases analyzed, the maxi-
orthogonal joint sets in which, the orientation of the first joint mum reduction of bearing capacity caused by water pressure is
set was assumed to be 15◦ , 30◦ and 45◦ . Different spacing val- about 10 percent. By increasing the degree of jointing of rock
ues for joints were considered and load-settlement curves were foundation, the decrement of the ultimate bearing capacity
obtained in each case. Then, the Log-Log method was applied would increase.
702
Simulating the failure mechanism of rock slopes (due to kinked and secondary cracks
propagation) by a higher order displacement discontinuity method
Keywords: crack propagation mechanism, rock slope stability analysis, numerical methods, displacement discontinuity
method, LEFM
Numerical modeling of rock slopes is now used routinely in In this paper, the Fictitious Stress Method (FSM) is used to
the civil and mining engineering sectors as well as in academic discretize the boundary of the problem in two dimensions and
research of numerical applications available today. The stabil- the higher order Displacement Discontinuity Method (DDM)
ity of rock slope is mainly determined by its discontinuity and with kinked and special crack tip elements is used to dis-
Rock Bridge. However, the failure mechanism of discontinuity cretize the original cracks, the kinked cracks and the secondary
and Rock Bridge has not been studied comprehensively. In this cracks. A mixed mode fracture criterion known as maximum
paper, the stability analysis of jointed rock slope is carried out tensile stress criterion is used to predict the direction of crack
by kinked and secondary cracks propagation using the higher initiation and the crack propagation paths. This paper presents
order displacement discontinuity Method. This study presents a method for determining the effect of crack propagation in
a new approach for stability analysis of jointed rock slope. It rock slope stability analysis by the higher order Displace-
can also consider the complicated geological conditions and ment Discontinuity Method (DDM) which is a relatively newly
supported slopes. Fracture mechanics is the field of mechan- developed method for determining the effect of fracturing in
ics concerned with the study of the formation of cracks in the stability analysis of rock slopes. Here, some failure areas
materials. Previous researches have shown that Griffith’s brit- have been shown where the cracks can grow. The modified
tle fracture theory can be modified to account for the effects of DDM program can evaluate the effect of cracking of rock
crack closure in compression. They have also shown a useful slopes on the failure surfaces. Finally a real case study (the
basis for the study of the fracture of hard rocks based on the second tectonic block of Choghart iron mine) is solved numer-
modified Griffith theory. An analysis of the stress distribution ically and it is shown that there were some failure planes on
around a crack indicates the points of fracture initiation as which the joints may start to propagate. It has been concluded
well as the initial direction of crack propagation. As a result that the wing cracks can be produced in rocks under both
of the changes in stress distribution associated with fracture compressive and tensile loading conditions. Further more it
propagation it is, however, impossible to predict the final path may also be concluded that the secondary cracks which are
of the propagating crack. Consequently, a serious limitation produced due to shear stresses induced under compressive
of the Griffith theory lies in the fact that it can only be used loading may also be produced under tensile loadings (similar
to predict fracture initiation. Recently, several researches have to the wing cracks).
shown that the Mode II fracture toughness of rock material is
usually higher than the Mode I fracture toughness, especially
when the confining pressure increases.
703
Keywords: stability analysis, rock slopes and foundations, numerical modelling, case studies
1 DESCRIPTION OF GRAND DALLAGKAUS SLIDE mass movements which coincided with the contour of the
sliding mass.
The Grand Dallagkaus landslide is situated on the The Dallagkaus slide analyzed with the aid of a 3D model
Mamisondon River which is a left bank tributary of the Ardon (considering the width of the sliding body at all elevations),
River. The depth of the Dallagkaus landslide is up to 80 m and has shown that in the water saturated state (w = 28.5% at
its volume is about 13 mln m3 . The length of the slide body γR = 2.34 t/m3 ) it may pass in the limiting state of stability
is about 950 m, while its averaged width varies from 80 to at the angle of friction ϕ = 20.5◦ and cohesion C = 15 kPa.
400 m. The Grand Dallagkaus slide is devoid of wood cover According to the studies conducted on other slides situated
while the surrounding stable areas are covered with rather in this region, the following strength parameters for shear
a dense forest, which serves the evidence of the slide body ϕ = 24◦ and cohesion C = 25 kPa have been taken for the
movements. natural moisture conditions of the Grand Dallagkaus slide.
The landslide stays in the condition of limit equilibrium Analysis of the Quaternary deposits in the natural mois-
and its movements make 50 mm per year. And in spite of slow ture condition has yielded the safety factor of the slide mass
downward movement of the sliding mass these displacements, k = 1.20, while in the water saturated state, the safety factor of
in case of their re-activation, can constrict and dam the chan- the Quaternary deposits turned out to be k = 1.01. Thus, prac-
nel of the Mamisondon River. Especially this danger can be tically the same results have been obtained like in analyzing by
brought about by a strong earthquake shock effect. Consider- the limit equilibrium method the stability of the Dallagkaus
ing the volume of this slide body, the height of impoundment slide. These analyses confirm that excessive wetting of the
may be rather big and its potential breach may have catas- soil from intensive precipitation and snow melting may be the
trophic consequences for the dam and the downstream parts main triggering mechanisms of local slide processes. Besides,
of the Ardon valley. To preclude development of such a dan- the performed analyses permitted delineating of the least sta-
ger, the project provides for bypass spillway tunnel through ble zones of the mass where large movements were displayed
the opposite bank of the Mamisondon River which could be at low safety factors, hence stability.
capable of passing the river flow in case of the river channel The method of limit equilibrium was used to analyze the sta-
damming. bility of the sliding mass due to increase in its moisture content
(from precipitation and snow melt water) and due to seismic
shocks with intensity 7, 8 and 9, which has allowed the deci-
2 ANALYSIS OF STRESS-STRAIN STATE AND sion to be made on impounding of the reservoir with a fixed
STABILITY OF SLIDE BODY water stage. It has been established that in the water-saturated
state, the slide mass passes into the condition of incipient slip
A number of the following problems should have first been and any seismic shock effect will trigger its movement. At the
resolved to ascertain stability of the slide with due account for same time it should be stressed that the above movement will
the results of assessment of the slope stress-strain state and the not be of the avalanche nature as loose material of the slide
sliding mass contour: has been creeping (at the rate of about 50 mm/year) for a long
time and the displacements of the sliding mass most likely
1. First, to finalize the shear strength properties of the sliding
will be of the plastic nature.
mass understanding that they can not be defined exper-
To protect the landslide from excessive saturation with rain
imentally in the laboratory, considering the presence of
and snow melt water running down the slope, it would be
large size lumps in the slide mass,
practical to arrange gutters at upper elevations which would
2. Second, to define the cause of though-insignificant but on-
permit the rain and snow melt water running down the slope,
going movements of the sliding body,
to be channeled past the land slide.
3. Third, to define the conditions on implementation of which,
On displacement of the lower portion of the landslide
displacement of the sliding body will not present danger
for 10–15 m, the slide body becomes more stable and its
for functioning of the dam.
displacements are expected to cease.
The numerical model of the slope was used to analyze the In the natural moisture state, the loss of stability by the
stress-strain state and to determine the area of potential rock sliding mass will be triggered only by seismic shocks with
704
705
Slope stability analysis using a hybrid Finite-Discrete Element method code (FEMDEM)
Keywords: numerical modelling, rock failure, rock slopes and foundations, stability analysis, rock joints, rock mass
1 INTRODUCTION
3 NUMERICAL EXAMPLES
2 THE 2D COMBINED FINITE-DISCRETE ELEMENT 3.1 Rock fall
METHOD (FEMDEM)
The capability of FEMDEM to quantitatively simulate rock
FEMDEM is a numerical tool pioneered by Munjiza et al. falls was verified by modelling a real scale in-situ test carried
(1995) for the dynamic simulation of multiple deformable and out by Giacomini et al. (2010) in eastern Australia.
fracturable bodies. Within the framework of FEMDEM, dis- As described by Lisjak et al. (2010), the block was mod-
crete element method (DEM) principles are used to model elled as an elastic body having Young’s modulus and Poisson’s
interaction between different solids, whose deformation is ratio equal to 40 GPa and 0.3, respectively. The block-slope
analyzed by discrete element analysis (FEM). A unique interface was characterized by a contact stiffness of 4 MPa
feature of such a numerical tool is its capacity to model and a friction angle of 40 deg. The contact parameter B was
the transition from continuous to discontinuous behaviour by calibrated by comparing the numerical travel path with the
explicitly considering fracture and fragmentation processes. experimental trajectory in the initial part of the block motion,
Since an explicit time-marching scheme is used to integrate determined through digital video cameras. In general, the
Newton’s equations of motion, fully dynamic simulations can higher the B value, the higher the energy dissipation intro-
be performed. A number of different dissipative mechanisms duced is and the lower the bounce heights are. As can be
(e.g., friction, contact damping and fracturing) are imple- observed in Figure 1 a very good agreement between exper-
mented into the code while no artificial numerical damping imental and numerical travel path was obtained for B equal
or other numerical parameters, often needed by other DEM to 2.
codes, are used. For further details the reader should refer to Since FEMDEM can account for internal stress distribu-
Mahabadi et al. (2011). tion, interaction between discrete bodies, and intact material
The capability of FEMDEM to quantitatively simulate fragmentation, it has the potential to largely overcome the
rock slope instabilities has been tested and validated against drawbacks associated with limit equilibrium analyses and sim-
experimental results obtained during controlled laboratory ulate the entire dynamic process. Four idealized examples of
and field tests. rock slope instability were chosen to better illustrate the use
706
707
Evaluation of the stability of a bullnose within Coedmore Quarry, Durban, South Africa
Keywords: rock joints, rock mass classification, rock slopes and foundations, stability analysis
An assessment of the stability of a bullnose within Coedmore strength allowed for the back calculation of Joint Roughness
Quarry was conducted. The stability analysis was based on Coefficient (JRC). The rock mass classification concluded
detailed scanline surveys and rock strength testing. Kinematic that the slopes were stable. A slope stability analysis was
analysis of the discontinuity data was carried out and it can conducted using the deterministic approach, resulting in a
be concluded that a wedge failure was the most likely failure factor of safety >1. Analysis based on the worst case sce-
mechanism associated with the orientation of the discontinu- nario also rendered a stable slope but with a lower value of F.
ities. The shear strength of the discontinuities was analysed It must be noted that, in reality, the bullnose has been stable
and the estimated maximum normal stress acting beneath the for several decades apart from small localized failures, which
slope was determined. The empirical determination of shear is consistent with field observations.
708
Keywords: rock slopes and foundations, physical modelling, rock failure, dynamics, rock joints, stability analysis
The seismic stability of rock slopes has usually been evaluated As a result, the slope model collapsed when it was excited
analytically in terms of the sliding safety factors based on by a sine wave of 3.5 m/s2 , which was converted to a real field
limit equilibrium methods. However, because there are very scale. Figure 2a shows an aspect of the collapse of the slope
few model tests that focus on the dynamic failure behaviour model immediately after the collapse, and Figure 2b shows the
of rock slopes, quantitative studies on the applicability of the shape of the sliding surface, determined on removing the col-
seismic stability evaluation method based on these model tests lapsed block. From these figures, the artificial discontinuities
have also been few. considerably affected the collapse, and the type of collapse
We conducted a centrifuge model test on earthquake- was plane failure.
induced rock slope failure. Figure 1 shows the structure of a dip The test results were compared with safety factors related to
slope model for the centrifuge model test. The dip slope model the sliding surface, obtained by the limit equilibrium method;
was constructed from cement, sand, iron powder, and water, the stresses on the sliding surface were estimated by equiv-
and discontinuities were imitated by inserting Teflon sheets. alent linear analysis. Figure 3 shows the minimum sliding
In the test, the centrifugal acceleration was 30G, and the accel- safety factors at each excitation step. The trends of the com-
eration amplitudes of the input sine waves were increased puted safety factors showed that the threshold amplitude of the
gradually at every step. failure was lower than the actual amplitude that induced the
failure, implying that the seismic stability evaluation method
based on the safety factors was conservative.
709
Keywords: rock slopes and foundations, risks and hazards, case studies, back analysis
1 EXTENTED ABSTRACT
710
Seismic instability assessment of rock slopes in a large area based on planar sliding mode
Keywords: rock slopes and foundations, risks and hazards, stability analysis
A GIS-based approach for regional assessment of seismic in the area is shown Fig. 2. An artificial acceleration time
instability of rock slopes is developed. The rock slope land- history is used for displacement calculations. The distribution
slides induced by an earthquake is assumed to occur in planar of the permanent displacements obtained using the proposed
failure mode. A rigorous Newmark’s analysis is integrated method is also shown in Fig. 2. It is of interest to note that
with a GIS to predict seismically induced deformations of rock the seismic displacements only take place in slopes in which
slopes. The proposed method includes the following steps: the slope inclination is close to the dip of the geological dis-
continuity planes. No displacements are distributed in slopes
1) A Digital Elevation Model (DEM), a digital geological map
whose dip is opposite to the inclination of the slope surfaces.
and possible ground motion (i.e. acceleration-time history)
are first prepared for a specified region.
2) Shear strengths of rock slopes are assigned to each grid cell
in the DEM coverage by overlaying the DEM and digital
geological maps.
3) For a given gird cell, 3 × 3 cells consisting of the cell under
consideration and eight adjacent cells around it are consid-
ered, as shown in Fig. 1(a), so that four vertical sections of
the slope surface, A-A , B- B , C-C and D-D are obtained.
As the slope profile for the cell, a section, which is the clos-
est to the direction perpendicular to the strike of geological
discontinuities, is selected from the four vertical sections,
as illustrated in Fig. 1(b).
4) Judge if plane failure will occur by considering the strike
and dip of the geological discontinuities. If the conditions
for plane failure are satisfied, then go to the next step.
5) The critical acceleration is computed for the selected slope
section and then Newmark seismic displacement is deter-
mined based on the given ground motion using a double
integration procedure.
6) The above described procedure is in turn repeated for
each grid cell to obtain the distribution of permanent Figure 1b. Generation of slope section for analysis.
displacements over the whole region.
The proposed method is applied to a mountain area in
Tokushima prefecture, Japan. A 10 meter DEM coverage
is built from 1:2500 contour line map. Homogeneous soil
with c = 19.6 kPa, φ = 35◦ and γ = 17.64 kN/m3 and dry con-
ditions are assumed for the whole study area. In addition, the
strike and dip of the main geotechnical discontinuity planes
Figure 1a. An area defined by 3 × 3 cells around the grid cell under
consideration. Figure 2. Distribution of the calculated displacements.
711
Keywords: blasting, site characterization, stability analysis, deterministic approach, kinematic analysis
Coedmore Quarry’s on-going rehabilitation programme A pseudostatic analysis was conducted to account for the
involved the tipping of waste rock over final highwall’s. impact of blast induced acceleration. Results indicated that at
However, this was stopped as it was a concern that it might mean shear strength values with an angle of internal friction
prove to be a geotechnical hazard. The stability analysis was of 37◦ , the slope is stable. However the presence of a ground-
based on kinematic observation that failure would occur as water table, and occasional peak accelerations, of greater than
a translational slip of cohesionless waste rock material. This 0.060 g, reached during blasting could have an adverse impact
type of failure lends itself to the infinite slope stability model. on the stability of the slope.
712
713
A.J. Li
School of Engineering, Deakin University, Geelong, VIC, Australia
M.J. Cassidy
ARC CoE in Geotechnical Science and Engineering, The University of Western Australia, Perth, WA, Australia
A.V. Lyamin
ARC CoE in Geotechnical Science and Engineering, The University of Newcastle, Newcastle, NSW, Australia
1 INTRODUCTION
714
Control and evaluation for residual subgrade settlement of high speed railway
X.H. Li, Z. Chen, X.Y. Zhao, H.L. Sun, J.H. Guo & T. Zhao
China Railway SIYUAN Survey and Design Group CO., LTD., Wuhan, Hubei, China
At present, the total length of high-speed railway lines in oper- with bulk material pile (c such as rushed stones pile, sand
ation achieves about 7500 km in China, and about 8000 km pile, etc.) composite foundation, semi-rigid and rigid pile-
in Germany, Japanese, France, and other countries. The total net structure composite foundation;
length of high-speed railways under construction or to be con- • For the main line of high-speed railway with ballastless
structed all over the world is up to above 20,000 km. Subgrade track, the semi-rigid (such as CFG piles, etc.) composite
is an important part of high-speed railway, it’s therefore cru- foundation should be adopted, complemented with flexible
cial to control the subgrade settlement strictly so as to ensure pile (such as jet grouting pile, etc.), rigid pile-net structure
the long-term high-speed, smooth, comfortable and safe oper- composite foundation and sheet pile structure;
ation of railway. Base on a large number of research and tests • For stations, the combined treatment programs of rigid and
both in laboratory and field, China has formed a set of sophis- flexible piles composite foundation should be applied;
ticated theory, technology and standard for high-speed railway • For the transition of bridge-subgrade, culvert-subgrade,
subgrade foundation reinforcement and settlement control. and embankment-cutting and the uneven ground, combina-
In this paper, the concept of “settlement control” is put for- tion of different foundation reinforcement measures, pile
ward, and is required to be followed during the design and length & pile spacing transition, and pre-loading transition
construction of subgrade structures for high-speed railway etc. should be adopted, so as to achieve uniform subgrade
engineering; a brief introduction is made about the com- settlement and control the residual settlement.
prehensive survey methods for high-speed railway subgrade,
It’s indicated in the paper that the high-speed railway sub-
including exploratory boring, static Cone Penetration Test
grade settlement monitoring should mainly emphasizes the
(CPT), Vane Shear Test (VST), Stress Shovel Test (SST),
subgrade surface settlement and foundation settlement, with
Standard Penetration Test (SPT), Dynamic Penetration Test
adoption of settlement observation piles for subgrade surface
(DPT), large-scale loading tests, geophysical exploration,
settlement monitoring, settlement plate, single-point settle-
etc.; the concept, component and control conditions of post-
ment gauge & deep-layer settlement gauge for foundation
construction settlement on railway foundation are stated in
settlement monitoring, sectional settlement pipe for differ-
details; and the control standards of residual settlement and
ential settlement monitoring between subgrade in transverse
differential settlement with adjacent bridge, tunnel, and other
direction and transition segment, and inclinometer pipe for
structures for Chinese high speed (250 to 350 km/h) ballast
lateral deformation monitoring of subgrade; and specific
and Ballastless tracks are described.
requirements are specified for the setting, accuracy, obser-
It’s indicated in the paper that the high-speed railway
vation time, and frequency of settlement monitoring section.
subgrade settlement is closely related to the embankment
Acc. to the construction process, a curve showing relation-
filling height, ground soil properties, and the thickness of
ships between sectional settlement and embankment height &
compression layers; and then the calculation methods of sub-
time should be plotted, to carry out comprehensive analysis
grade settlement and post-construction settlement for Chinese
adopting the hyperbolic method, Logarithmic curve method,
high-speed railway is described in detail; base on field test,
settlement velocity method, Hoshino method and correction
the advice is given to the settlement calculation depth of
hyperbolic method, so as to predict the final settlement of
thick Quaternary stratum. In view of embankment height,
subgrade and the residual settlement after construction, for
ground conditions, construction schedule and engineering
which the specific assessment standards are illustrated. In
types, based on the systematic analysis of foundation soil layer
addition, acc. to the sectional measured settlement, predicted
distribution, thickness, depth, physical and mechanical and
overall settlement and post-construction settlement, longi-
deformation parameters, the assessment of post-construction
tudinal settlement curve incl. subgrade, bridge, tunnel and
settlement is made, and thereon the foundation treatment mea-
transition segment should be plotted, to analyze the unifor-
sures should be optimized in order of embankment setting,
mity of settlement along railway in longitudinal direction,
replacement, preloading, and foundation reinforcement.
so as to give comprehensive evaluation as weather the track
Foundation treatment measures should be adopted accord-
laying requirement for high speed railway is satisfied.
ing to project types:
In the course of high-speed railway construction in China,
• For the main line of high-speed railway with ballast track, to control the subgrade post-construction settlement, a series
the flexible pile (such as mixing pile, jet grouting pile, etc.) of theoretical studies and field tests have been carried out
composite foundation should be adopted, complemented on the reinforcement mechanism & effects, stress & strain,
715
716
Keywords: lab testing, rock mass, rock mass classification, rock properties, rock stress and site characterization.
Bakhtiary Dam & Hydropower Project is located on Bakhtiary from inside the exploratory galleries. The boreholes are sit-
River (the main branch of Dez River), about 70 km North-East uated at or around the arch dam, underground powerhouse
of the city of Andimeshk in South-West of Iran. The main dam complex, other main appurtenant structures and also at the
is a double-curvature concrete arch type with maximum height quarry areas. Geotechnical investigations performed in these
of 325 m and crest length of 509 m at El. 840 masl. It will be boreholes included RQD measurements, lithological survey,
the highest arch dam in the world now at the design stage. joint survey and analysis of permeability tests’ results.
The hydropower plant consists of an underground powerhouse Bakhtiary dam and hydropower plant project is unique for
complex with a total installed capacity of 1’500 MW which is its outstanding characteristics. One of them is the extensive
designed for annual energy output of 3’000 GWH. engineering geological and rock mechanics investigations car-
At present the dam site is only accessible by train through ried out during its design studies. These investigations were
Tehran-Ahwaz railroad. It is nearby a small railway station planned to fulfill the high technical requirements of Bakhtiary
called Tang-e-Panj Station. project. As a part of this extensive program, several laboratory
The updated feasibility studies of the project which was and in-situ rock mechanics tests were planned and executed
performed by Bakhtiary Joint Venture Consultants, BJVC during 2003 to 2010.
(2009a), confirmed that the higher the dam the more eco- A large number of laboratory rock mechanics tests were
nomical scheme can be planned for the project and the main performed on rock samples selected from the drill cores of
limitation is topography. Therefore, the optimum reservoir the exploratory boreholes including index tests, ultrasonic
normal water level was selected at 830 masl and the dam crest wave velocity measurements, unconfined compression tests,
elevation was designed at 840 masl, considering 10 m free triaxial tests and direct shear tests. The results of labora-
board for flood routing. Considering the narrow valley at the tory rock mechanics tests indicated that the rock material
dam location (width/height ratio of about 1.5) and suitable of siliceous limestone of Sarvak formation at Bakhtiary dam
condition of rock foundation, a concrete double curvature arch site is characterized by its very low porosity, very low water
dam was designed for Bakhtiary Project. absorption, relatively high compressive strength (UCSAve.
Bakhtiary dam site and its reservoir are located in the Folded ≥100 MPa) and high modulus of Elasticity (EAve. ≥60 GPa).
Zagross tectono–sedimentary province. The dam site is mostly Meanwhile, the average shear strength parameters along the
placed on siliceous limestone of Sarvak formation (Cretaceous bedding planes and discontinuities with planner-smooth sur-
period) which has been divided by BJVC (2009b) to seven faces (JRC <6) are C = 0–20 kPa and ϕ = 32◦ and for the
units from Sv1 to Sv7, based on the thickness of bedding discontinuities with planner-rough surfaces (JRC >6) the
planes and the state of siliceous combination outcrop. The average shear strength parameters are C = 80 kPa and ϕ = 36◦ .
most important geological structures at the dam site consist of In order to make a realistic estimate on the geomechani-
Griveh and Siah Kouh anticlines (dam axis will be laid almost cal parameters of the existing rock masses at the dam site, an
parallel to the axis of these structures), F1–F3 thrust and F2 extensive in-situ rock mechanics test program was performed
fault, kinkbands, chevron folds and a joint system (including in the exploratory galleries excavated in the dam foundation
bedding and 3 joints set) at the dam site. and underground powerhouse areas. This test program con-
Subsurface exploration at Bakhtiary dam site consisted sisted of deformability tests including Plate Load, Large Flat
of the core drilling of more than 178 boreholes, with the Jack and Dilatometer tests, in-situ stress measurement includ-
total length of about 12’500 m, in which water pressure ing Borehole Slotter and Hydraulic Fracturing tests and Direct
tests were also carried out. In addition, 11 exploratory gal- Shear test for determining the shear strength parameters of the
leries were excavated with the total length of 2’850 m for bedding planes as the most frequent and dominant discontinu-
engineering geological mapping, discontinuity survey and ity at the dam site. Since the Plate Load Tests provided the most
in-situ rock mechanics tests. Among these boreholes, 70 bore- reliable results for estimating the deformability characteris-
holes were drilled from the ground and 108 boreholes drilled tics of the rock masses therefore; a correlation was defined
717
718
1 BEARING CAPACITY OF A ROCK FOUNDATION considering additional information related with the uncon-
ACCORDING TO THE EC7 fined compressive strength of the rock and the spacing of the
main joint set.
For the design of spread foundations on rock masses the EC7
provides some considerations. It points out to the necessity of 1.1 Exampe 1
taking into account the following features in the design:
This example is concerned with the computation of the allow-
• the deformability and strength of the rock mass and the able bearing capacity of a 2 m width square footing built in
permissible settlement of the supported structure; a jointed and weakly cemented sandstone. The weight den-
• the presence of any weak layers, for example solution sity (γ) of the rock was set to 22.5 kN/m3 and the spacing of
features or fault zones, beneath the foundations; the main joint set (s) of 0.2 mm. The following parameters
• the presence of bedding joints and other discontinuities and were also considered: Geological Strength Index (GSI) = 20;
their characteristics (for example filling, continuity, width Uniaxial compressive strength (σ c ) = 10 MPa and SF = 3. The
and spacing); calculation was carried out for the case of the footing built at
• the state of weathering, decomposition and fracturing of the
rock;
• disturbance of the natural state of the rock caused by
construction activities, such as, for example, underground
works or slope excavation, being near the concrete founda-
tion.
It states also that for intact igneous and gneissic rocks,
limestones and sandstones, the presumed bearing pressure is
limited by the compressive strength of the concrete founda-
tion. Finally, a remark for the necessity of assessment of the
foundation settlement by means of comparable experience.
The method presented in the annex G of the EC7 can be
used for the calculation of the bearing resistance of spread
foundations in weak and broken rocks with tight joints. For
weak and broken rocks with open or infilled joints, reduced
values of the bearing capacity should be used.
The application of the method starts with the choice of the
group of the rock considering the classification provided in
Table 1. After the definition of the group, the bearing resis-
tance of the foundation can be estimated using Figure 1,
719
qa (MPa) qa (MPa)
Method depth = 0 m depth = 2 m
the surface and for a depth of 2 m using the three methods Figure 2. Variation of qa with σ c .
presented before.
The results of this calculation are presented in Table 2:
The results show considerable dispersion between the 2 CONCLUSIONS
results of the different methods. The method of Wyllie (1992)
provides the lowest values and the one suggested by the EC7 For the design of spread foundations in rock masses the EC7
the highest. One important issue is the significant increase of provide some aspects that should be taken into account and
qa with depth for two of the methods which were an expected proposes a method for the calculation of the bearing capacity.
fact due to the increasing vertical stress. For a 2 m depth of The comparison of three different methods based on two cases
the foundation base the method of Wyllie (1992) point out to allowed concluding that considerable different results can be
a double value of qa and the one by Serrano et al. (2000) to an found even for simple examples. The method from Wyllie
increase of almost 50%. The EC7 method (CEN, 2004) does (1992) seems to provide a lower boundary for the values of
not take into account with the increase of the bearing capacity the allowable bearing capacity.
due to this fact. The method proposed by the EC7 is rather simplistic and
does not take into account important aspects that influence the
bearing capacity like the depth of the foundation, its shape,
1.2 Example 2 the eccentricity of the loads, water, etc.
A pillar transmits to a square footing a load of 4 MN. The In spite of the suggested EC7 method has been based on
geotechnical survey identified a layer of soil up to a depth conservative assumptions, it gave the less conservative results
of about 3 m (γ = 20 kN/m3) and below a jointed weathered in example 1. However, the results obtained with this method
granite where the foundation will rest. The design parameters in example 2 are similar to the lower values obtained using
for this granite were the following: GSI = 40; γ = 24 kN/m3; Wyllie method.
σc = [20–50] MPa; s = 0.2 m and SF = 3. The objective is In conclusion the EC7 method for the design of spread foun-
to compute the allowable bearing capacity using the three dations should be used with caution and to provide only an
methods. Given the uncertainty related to the value of σc, approximate idea of the bearing capacity. The results should
it was decided to carry out a parametric study to analyse the be validated using other analytical procedures or even using
importance of this parameter. numerical methods.
The main results of this calculation can be observed in
Figure 2 where the relation between qa and σ c is presented.
In this example the influence of the variation of σ c was REFERENCES
investigated. For all the tested methods, qa increases almost
linearly with σ c with a small exception for the Wyllie (1992) CEN 2004. Eurocode 7 Geotechnical Design – Part 1: General Rules.
Hoek, E. and Brown, E. 1980. Underground excavations in rock. Inst.
method where this increase is slightly non-linear. This method
of Mining and Metallurgy, London, 627p.
and the one from the EC7 provide similar results in this case Wyllie, D. 1992. Foundations on rock. E & FN Spon.
pointing out for a value of qa approximately equal to 25% of Serrano, A., Olalla, C. and González, J. 2000. Ultimate bearing
σ c . On the other hand, the method of Serrano et al. (2000) capacity of rock masses based on the modified Hoek-Brown cri-
provides values of qa which are about 74% of σ c , almost three terion. Int. Journal of Rock Mechanics and Mining Sciences 37:
times higher than the previous ones. 1013–1018.
720
Unloading scheme to control sliding mass at Angouran open pit mine, Iran
P. Moarefvand
Department of Mining and Metallurgy, Amirkabir University of Technology, Iran
M. Ahmadi
Department of Mining Engineering, Tarbiat Modares University, Tehran, Iran
M. Afifipour
Department of Mining and Metallurgy, Amirkabir University of Technology, Iran
Keywords: rock slope, site characterization, rock failure, numerical modeling, stability analysis.
Angouran zinc-lead open pit mine, located in the northwest right sides, without significant disturbance and overturning.
of Iran, has experienced a large-scale slope failure in the Finally the crushed zone dropped on lower benches with
order of 12 Mm3 volume. The whole failed mass moved about respect to the slip surface geometry and fragmented. A
100 m horizontally and 65 m vertically. The failure region was three-dimensional discrete element modeling was executed
mainly comprised of a jointed limestone rock mass as a hang- to evaluate the slip surface and predict responses of the slid-
ing wall at the northwestern wall of Angouran pit. The hanging ing mass. The slip surface geometry estimated through the
wall contact toward the thick schist bedrock might be consid- numerical simulation was in close agreement with the site
ered as the triggering zone. The interface between these two investigation results, and proposed the failure scenario. The
formations (hanging wall limestone and footwall schist) was slip surface and the sliding zone geometries implied that
estimated to be the slip surface, owing to the poor geomechan- unloading below this level would lead to lightening the toe
ical properties. The main challenges due to this giant failure of slope and consequently, causing instability in the whole
are subsequent responses of the sliding mass, geometry of volume. The gradient of the slip surface between level 2970
the slip surface, the regions of critical movements, and where and level 2960 approximately reached to zero and the surface
unloading procedures should be performed to stabilize the became almost horizontal. This fact was the main reason for
slope and restart mining operation. To find solutions for the the temporary stability of the sliding zone. The sheared zone
challenges, a comprehensive site investigation and monitor- does not need to be unloaded and disturbance in this zone
ing was conducted to compare the primary and secondary might decrease its strength and stability. Based on the findings
discontinuity patterns in order to identify the failure scenario from site investigation and numerical analysis and according
and determine the governing mechanism. Using the results of to the slip surface geometry and the relative position of the
field investigations and analyzing the primary and secondary mentioned main parts in the failure region, a decision was
discontinuity patterns, the failure area was divided into three made to stabilize the slide by unloading the active part of the
main parts: the sheared zone, the sliding zone, and the crushed complex unstable mass from the levels upper than 3050 to
zone. The sheared zone is the upper section that is restricted restart mining operation safely in the adjacent sections of the
by two lateral faults. These faults that dipped into the failure pit. It is estimated that over 4 Mm3 , i.e. approximately 35 per-
area were considered as two lateral slip surfaces. The slid- cent of the whole volume, should be unloaded to restart mining
ing zone moved and created the sheared zone in both left and operation.
721
Yu-Chang Chan
Academia Sinica, Institute of Earth Sciences, Taipei
Kuo-Jen Chang
Department of Civil Engineering, National Taipei University of Technology, Taipei
Recent in Taiwan the dip-slope landslide failure frequently first obtained the basic engineering parameters, c, φ, and τ,
caused huge casualties. Except for rainfall or earthquake by numerical modeling method, 2-D discrete element method,
directly induced slope failure, the man-made interference such in one case study. At same time, we considered groundwater,
as cutting the toe of slope had been one of the crucial factors of retaining wall, and geological factors so that to simulate real
losing slope stability. One representative case of dip-slope fail- field condition. Finally, we are going to look for all of the
ure occurred on freeway locates Keelung river northern basin, potential slopes using LiDAR DTM, and evaluate the possi-
which consist of dip-slope strata. Noteworthy issue around bility of sliding of slope failure in those slopes through above
this area is that there is still a lot of potential slopes may slide case study.
because of influences of artificial building. In this study we
740
C.S. Oteo
University of Coruña. Spain
724
725
Phyllite is a low grade metamorphosed rock. In Korea, phyl- are well developed along the lithologic boundary of phyllite
lite is mainly distributed in the Chungju and Boeun areas, and granite porphyry.
which belong to the Ogcheon Supergroup. Phyllite is mainly In the study site, slope failures occurred in four (4) cut
composed of platy minerals such as fine-grained micas and slopes as shown in the Fig. 2 and Table 1. The heights of the
chlorite, and it has foliation developed in metamorphic pro- cut slopes varied from 30 m to 70 m approximately. Plane and
cess. Phyllite is an anisotropic rock in terms of engineering wedge failures occurred by foliation and joint in the site A.
properties because strength and deformation modulus vary in In the site B, the 3.0 m thick intruding dyke of granite por-
different directions. It shows fissility developed along long- phyry, intruded into phyllite, is highly affected by shearing
persistent foliations, resulting in low shear strength. Also, and weathering, and exists under completely weathered soil
phyllite is easily weathered when exposed to the air since it condition, containing small rock fragments and fault breccia.
is basically formed with clay minerals. Those characteristics, Plane failures arose along foliation of phyllite in the site B. In
such as anisotropy, fissility and easiness for weathering are the site C, a large-scale wedge failure occurred by intersec-
considered as major causes for slope failures, and it is proven tion of joints. In the site D, the 25 ∼ 30 m thick dyke of granite
by the fact that many cases of small to large scale rock slope porphyry, intruded into phyllite, shows well developed fault
failure in phyllite have been reported in Korea, resulting in shear zones along the lithologic boundary, and plane failures
increase of construction cost and delayed construction sched- arose along the lithologic boundary and foliation in phyllite.
ule. Therefore, due considerations are required for design or
construction of rock slopes in phyllite. Even it becomes more
important when we consider that construction cases of national
roads and highways across inland areas in Korea have been
increasing rapidly, and as a consequence, many cut slopes
have been constructed within geologically poor areas in Korea
such as the Ogcheon Supergroup where shows complex geo-
logical structures within metasedimentary rocks. Considering
the fact, this case study aims to introduce failure modes of
rock slopes which consist of phyllite and granite porphyry,
intruding along foliations of phyllite. Figure 2. Location of the study sites.
The study site consists of low grade metamorphosed rock
Table 1. Failure modes.
(phyllite) and granite porphyry which belong to the age-
unknown Hwanggangri Formation of the Ogcheon Super-
group and Cretaceous dyke, respectively.And alluvial deposits
of Quaternary Period are unconformably overlaying them.
Most sites belong to the Hwanggangri Formation of the
Ogcheon System and major component of rocks is dark grey
or black pebble-bearing phyllite. On the slope surface, phyllite
is highly weathered and it has become platy rock fragments
as shown in the Fig. 1. The granite porphyry, which intruded
phyllite, shows differential weathering, and fractured zones
726
Figure 4. Results of boring survey and electric resistivity survey (Site D).
727
Manuel Romana
Technical University of Valencia, Spain
The requirements for the foundation of concrete (either CVC C (Km/s) Terrain class
or RCC) dams traditionally have been expressed in qualitative
∼5 Very hard rock (good concrete)
terms: “sound rock”, “fresh rock”, “below the decompressed 5–3.5 Hard rock (concrete)
surface strata” . . . Fraser (2001) describes the “different 3.5–2 Medium/Soft rock
approaches for defining a (dam) foundation objective”: “1 2–1.5 Very soft rock/indurated soil
attain a specific geologic unit; 2 excavate to a grade based on <1.5 Soil
field testing results; 3 attain a specific rock quality; 4 achieve
a surface that meets a construction control test; 5 excavate to
a surface based on the ability of the excavation equipment;
6 excavate to a depth indicated by the design; 7 achieve a Table 2. Geotechnical zoning conditions used in Portugal.
material judged adequate based on visual observation”.Those
approaches are subjective. Fraser ends: “ultimate adequacy of Zone Weathering degree Celerity (Km/s)
the foundation must be confirmed . . . to insure the foundation
will perform” GZ 1 W1–W2 >42.5
The most usual requirement for rock foundation quality of GZ 2 W3 1.2–2.5
GZ 3 W4–W5 <41.2
gravity dams is “good quality, sound, fresh and unweathered
rock”. Sharma (1998): “the entire (foundation) area should
be stripped to firm rock”. Fell et al. (2005): “after normal
treatment . . . the rock foundation must be adequate to support
Table 3. Effect of Ec/Em on gravity dams behaviour
dam loads”.
Actually an increasing number of CVC and RCC dams have Ec/Em* Influence on dam Problems
to be founded in less sound rock, border-line terrains.
The problem of dams’ foundation in “low strength mate- <4 Negligible None
rials” was addressed in a quantitative way, by Rocha (1975, 4–8 Low importance Minor
1976). He had started to work on the problem of rock mass 8–16 Important Some
deformability for double arch dams (Rocha, 1964, 1974). >16 Very important Moderate to big
Sadly he died and his last papers for gravity dams, in
* Ec/Em concrete/foundation deformation modulus
Portuguese, have not been published in English.
Dam stability against horizontal sliding must be checked by
the designer taking account of the distribution of water pore
pressure across the dam foundation and of the mobilized shear Portuguese authors use the concept of “geotechnical zon-
strength of the governing joints. The calculation must be based ing”, with GZ 1 adequate fro dam foundations
in a good knowledge of geology below the dam. There are no According to Marulanda (2003) “foundation levels are
approximate solutions. acceptable when a minimum compressive seismic velocity
(celerity C) between 2.5 and 3.5 Km/s are recorded. Oth-
ers . . . define the rock foundation level on the basis of
2 ISRM SCALE OF WEATHERING GRADES
ripability . . . by D8 dozers”. The generally stated limit for
rippability is terrain celerity of 2 Km/s, (Caterpillar, 2008)
Weathered material must be striped away from the foundation.
which could be accepted as the lower limit for gravity dams’
In most cases the foundation is excavated to grade II rock in the
foundation, with 2.5 Km/s as desirable value
valley central part and to grade II–III rock in abutments. Spill-
ways should be founded on grade I rock. These requirements
could be minored for RCC dams. 4 ROCK MASS DEFORMABILITY MODULUS
3 LONGITUDINAL VELOCITY (CELERITY) Marulanda (2003) says that “a modulus of elasticity ratio
above 1/4 is commonly considered for compatibility between
A parameter frequently quantified to check the rock quality foundation rock and RCC”. Rocha’s work for arch dams was
has been C, celerity of P-waves, measured by geophysical endorsed by Serafim (1988) and Oliveira (1990) Rocha’s
methods (refraction). (1975, 1976) results for gravity dams are in table 3,
728
6 FOUNDATIONS OF HARDFILL DAMS The author will be very thankful for any comments which
could confirm, discuss, or deny, the limit numbers contained
A new class of gravity dams, built with hardfill, tries to avoid in this paper for the several requirements. The e-mail address
the problem of poor strength foundations. (Londe & Lino, of the author is mromana@stmr.es
1992; ICOLD, 2000). As Dunstan (1995) explained: “if the
729
L. Leoni
IDS Ingeneria Dei Sistemi, Pisa, Toscana, Italy
S. Fortin
Highland Valley Copper Mine, Logan Lake, British Columbia, Canada
1 INTRODUCTION
730
731
The stability of fractured rock masses is one of the classical method is higher than that by the DDA. Sarma’s method car-
problems in rock mechanics and still poses a major chal- ries out analysis based on the initial configuration of the rock
lenge to designers. The limit equilibrium method and the slope, which means that a bock slides only along its own base
continuum methods are best suitable to evaluate the factor segment. However, the DDA considers the initial configura-
of safety against failure for the rock slopes which are com- tion as well as the block kinematics. In the DDA analysis,
prised of massive intact rocks, weak rocks, or heavily fractured at the verge of the potential failure initiation, the small dis-
rock masses. However, when rock mass behavior is con- placement taking place to a block may carry it to contact with
trolled by the discontinuities, such as fractures and joints, the another base segment adjacent to its original base segment.
interaction between individual rock blocks defined by the dis- As the adjacent base segment may have different base angle,
continuities must be taken into account, the limit equilibrium friction angle and cohesion, the stability of the corresponding
method and the continuum methods may not be appropri- block and the whole slope will change accordingly. Figure 4
ate for the analysis of such block system. The Discontinuous
Deformation Analysis (DDA) is highly applicable to the mod-
eling of discontinuity-controlled rock block systems, allowing
two-dimensional analysis of slope failure mechanism and is
capable of simulating large displacements due to slip, or open-
ing along discontinuities. The advantage of the DDA is that
the mode of failure is one of the results of the analysis, rather
than one of the underlying assumptions. In addition, the DDA
reproduces the real displacements and interaction forces along
discontinuities within the sliding mass, as well as along the
failure surface.
This paper presents a detail study of the DDA for rock slope
stability analysis. The study concerns on the stability analysis
of a classic rock slope (Figure 1). A fictitious horizontal accel-
eration Kc is used in the DDA analysis as the safety measure
of the two-dimensional slope. For a given factor of safety, F, a
horizontal volume force, Kc Wi , will be acted on all the blocks
simultaneously. With the variety of Kc , the block displace- Figure 2. Displacement versus time at point A for a given factor of
ment will vary accordingly and a critical value of Kc can be safety.
obtained when the slope is at the verge of potential failure initi-
ation (Figure 2). The comparison of the F − Kc curves derived
by the DDA and Sarma’s methods (Figure 3) indicates that for
a given factor of safety, the value of Kc obtained by Sarma’s
Figure 1. Block partition of a rock slope. Figure 3. Factor of safety F versus horizontal acceleration Kc .
732
relative to the right one in reality. The DDA can identify the
directions of the block movement and the shear thrust forces
automatically based on the block kinematics. In the DDA anal-
ysis as shown in Figure 6, the ratios of the shear thrust forces
to the corresponding resistant shear strength are all smaller
than 1, the limit equilibrium condition will not occur along
the interfaces between blocks at the initiation of slope failure
in this example. Meanwhile, since the ratio between the shear
thrust force and the shear strength varies greatly at different
interfaces, it can also be deduced that when the failure occurs
along the interfaces in the slope failure process, the failure
will not take place simultaneously among all the interfaces,
which is totally different from the corresponding assumption
in Sarma’s method. Sarma’s method is revised so that the rel-
Figure 5. Distribution of shear thrust force between blocks. ative movement directions of adjacent blocks are determined
based on DDA’s results and the failures don’t take place along
shows the distribution of normal thrust forces between blocks the interfaces. The results show that the factor of safety form
at the verge of the potential failure. Because different Kc is the revised Sarma’s method is closer to that from the DDA,
applied, the normal thrust forces obtained from the two meth- but there are still big differences (Figure 3). The reason is that
ods are different but with similar distribution patterns. Figure Sarma’s method carries out analysis based on the initial con-
5 shows that the distribution of shear thrust forces between figuration of the rock slope and the DDA considers the initial
blocks obtained from Sarma’s method is more regular than configuration as well as the block kinematics. The results illus-
that from the DDA. The distribution of shear thrust forces trate that the normal thrust forces and the shear thrust forces
from the DDA is very complex with smaller absolute values obtained from the two methods (DDA and revised Sarma’s
and shows that the direction of possible relative movement of method) are different but with similar distribution patterns
two adjacent blocks may not be the same as Sarma’s assump- (Figures 4 and 5). As a numerical modeling method, the DDA
tion. Sarma’s method only considers the condition that the is much more complicated both in theory and application;
left block moves upward relative to the right one. However, nevertheless, it gives a more realistic solution to a problem
it is sometimes possible that the left block moves downward and can be used to reproduce the whole rock failure process.
733
Xinbin Tu
Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, P. R. China
Gaofeng Zhao
School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW, Australia
Fuchu Dai
Institute of Geology and Geophysics, Chinese Academy of Sciences, Beijing, P. R. China
Jian Zhao
Ecole Polytechnique Federale de Lausanne (EPFL), Laboratory for Rock Mechanics (LMR),
EPFL-ENAC-IIC-LMR, Lausanne, Switzerland
Keywords: numerical modeling; rock mass; rock slopes and foundations; stability analysis
1 INTRODUCTION
734
4 CONCLUSIONS
REFERENCES
735
Alex Varughese
Central Soil and Materials Research Station, New Delhi, India
1 THE PROJECT
736
737
Erik Wunder
Estelar Engenheiros Associados Ltd.
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: weak rock, lab testing, site characterisation, rock mass, rock properties
1 INTRODUCTION Table 2.1. Average values from the physical characterization tests.
738
Mean Value 0.79 ± 0.22 19.07 ± 5.31 confined(1) unconfined(2) confined unconfined
739
Yu-Chang Chan
Academia Sinica, Institute of Earth Sciences, Taipei
Kuo-Jen Chang
Department of Civil Engineering, National Taipei University of Technology, Taipei
Recent in Taiwan the dip-slope landslide failure frequently first obtained the basic engineering parameters, c, φ, and τ,
caused huge casualties. Except for rainfall or earthquake by numerical modeling method, 2-D discrete element method,
directly induced slope failure, the man-made interference such in one case study. At same time, we considered groundwater,
as cutting the toe of slope had been one of the crucial factors of retaining wall, and geological factors so that to simulate real
losing slope stability. One representative case of dip-slope fail- field condition. Finally, we are going to look for all of the
ure occurred on freeway locates Keelung river northern basin, potential slopes using LiDAR DTM, and evaluate the possi-
which consist of dip-slope strata. Noteworthy issue around bility of sliding of slope failure in those slopes through above
this area is that there is still a lot of potential slopes may slide case study.
because of influences of artificial building. In this study we
740
In this paper, a comprehensive geomechanical approach is When the core samples from the field are secured, experimen-
proposed to predict the excessive sand production in an effec- tal testing is an effective way to simulate the sanding scenarios
tive way that includes: a) employing analytical method with under realistic, simulated down-hole conditions. The results
the combination of experience-based empirical rules and ana- can be used to validate the studies obtained from analytical
lytical expressions from the simplified sanding model (e.g., and empirical methods.
determining the correlation between critical drawdown rate The benefits of adopting the unified approach are partially
and formation rock strength, delineating the sand-free enve- explained in a case study of this paper. In conclusion, the
lope based on the reservoir depletion history and desired ultimate purposes of employing the unified approach can be
critical bottomhole flowing pressure as well as formation com- given as follows:
pressive strength, and assessing the suitability of each method
• To avoid applying one particular method that can be biased
under a particular case); b) performing a numerical method
and can lead to an insensible decision for sand control.
using the complex constitutive relationships and realistic com-
• To generate the self-calibrated results of sanding predic-
pletion configurations in three-dimensional field environment
tion that can result in making an effective and economical
(e.g., recording the transition of formation from brittle to
decision for sand control.
ductile states with increasing drawdown pressure until the
catastrophic failure occurs, i.e., the rock plastic deformation With respect to the application of the presented technology,
exceeds the critical value; employing the finite element tech- the presented method can be used to predict excessive sanding
niques to represent the behavior of heterogeneous rocks, such potential and assess completion-device stability near perfora-
as faults; and simulating the production scenario under desired tion tunnels to aid a good decision on the type of sand control
sand prevention devices); and c) conducting an experimental method to be employed. In a particular case study, the weak
method using laboratory rock mechanical testing techniques spots in the reservoir perforation intervals were identified.
(e.g., determining the correlation between wellbore strength The formation strength at the weak spots was determined from
and maximum differential pressure via thick-wall cylinder sonic measurement and rock mechanical testing data. The crit-
tests, calculating the maximum drawdown pressure using ical values of sanding rate, unconfined compressive strength,
simulated depletion tests, and determining the impact of critical plastic strain, critical bottomhole flowing rate, maxi-
stress path on the maximum depletion from pore-volume mum wellbore strength, and critical depletion pressure were
compressibility tests). assessed using the analytical, empirical, numerical, and exper-
Using one particular method (e.g., either analytical, or imental methods under the field conditions (e.g., reservoir
empirical, or numerical, or experimental) may be convenient, pressure, in-situ stress, stress path, rock strength, etc.).
but it can lead to an incomplete or a biased analysis of sanding Based on the conclusions drawn from the paper, it is noted
prediction. To avoid this shortcoming, the presented technique that the presented approach on the sand prediction differs
provides a unified approach that employs multiple methods from the conventional approach, where the latter is often
with complementary objectives. In the present study, atten- biased toward a particular method (e.g., empirical, analyt-
tion is focused on predicting the critical drawdown pressure ical, numerical, or experimental). The presented technique
(CDP) under tolerable sand production. provides a unified approach that uses multiple methods with
The primary purpose of the analytical and empirical meth- complementary objectives: a) analytical and empirical meth-
ods in this paper focuses on determining the critical drawdown ods initially assess the sand production risk under the simpli-
pressure constrained by the sand-free envelope. The sand-free fied conditions; b) the numerical method examines wellbore
envelope is defined by the border lines of: a) an equal line stability issues under the detailed configuration of sand pre-
between the critical bottom-hole flowing pressure and reser- vention devices; and c) the experimental method calibrates the
voir pressure, b) the critical drawdown pressure at the current sand prediction under the simulated downhole conditions.
reservoir condition, and c) the minimum critical bottom-hole There are number of significant results being derived from
flowing pressure. the paper. Noticeably, classical sand control techniques are
For complex perforation geometry, numerical methods, primarily based on installing the sand exclusion devices, such
such as the finite-element method, are more amenable. Using as gravel packing, expandable screens, etc., which dramati-
critical plastic strain to assess the sanding-induced perfora- cally increase the completion costs and detrimentally reduce
tion tunnel stability via the finite-element method, numerical the efficiency and potential for the hydrocarbon production.
methods can also evaluate the critical drawdown pressure The presented approach uses the self-calibrated predictions
under the specific production scenario. The process parallel on sand production via multiple methods, which can lead to
to the analytical and empirical methods can provide a sensible a sensible decision for selecting the effective and economical
validation for the determined critical drawdown pressure. sand control measures.
Numerous experimental methods involving laboratory-
rock mechanical testing are available for the sanding analysis.
743
ABSTRACT: The theory of analytical methods for underground structures design was born in the middle of twentieth century
due to professor K.V. Ruppeneyt, who suggested to consider mine support and rock mass as elements of a united deformable
system. Small values of contact displacements make possible to use linearly deformable medium as a model of rock mass.
Today modern design methods based on rigorous analytical solutions of the elasticity theory plane problems are developed for
stress state determination of shaft supports and tunnels linings subjected to different loads.
1 INTRODUCTION methods. The first analytical solution, which can be used for
lining design, had been obtained by Gabriel Lame in the mid-
The first attempts to obtain underground constructions stress dle of nineteenth century. The problem was quite simple and
state are dated by the end of nineteenth century. Early design consists in stress strain state determination of a ring sub-
methods are based on structural mechanic solutions of the jected to inner and outer pressures. Some developments of
problems consist in internal forces determination for a bent this problem led to appearance of an analytical method for
axis beam subjected to the vertical and horizontal pressures. shaft lining design.
Later these methods were developed by external load dividing Muir Wood (1975) and Curtis (1976) had obtained analyt-
into ‘active’and ‘passive’components. Passive part of the load ical solutions for a circular ring supporting an opening in the
consisted in the reaction of elastic foundation to the beam linearly deformable medium, which later has been used for
displacements. The exclusion of rock mass from the design the design of Channel Tunnel Rail Link in UK.
scheme turned internal load determination to a very difficult The analytical method for design of multilayer tunnel and
problem. shaft linings has been developed in the same time in Russia by
The conception of joint contact interaction proposed by Bulychev (1974). This method was used for design of some
Fenner and Ruppeneyt (1954) had led to the basis of a new tunnels with total length about 60 km along Baikal-Amur
theory of underground structures design. The new approach railway, and numerous shafts. Further development of this
does not demand rock pressure determination, because nor- method made possible to solve the inverse problem connected
mal and shear stresses along the contact line are obtained with initial stress field determination on the base of stress or
as intermediate values during calculations. The input data displacement measurements in some points of a tunnel lining.
for calculations are deformation properties of the rock mass Design of non-circular tunnel linings needs solution of
and a support material as well as parameters of the intact rock more complex problems of the elasticity theory. The general
mass stress field. approach to the wide class of elasticity theory problem was
Structural mechanics methods were not able taking into made by Muskhelishvily (1953). This approach was applied
account ‘rock-support’ interaction, so analytical design meth- for non-circular supports design (Fotieva, 1980). Analytical
ods based on solutions of elasticity theory problems have been design methods for non-circular tunnel linings subjected the
developed for underground construction design. The main action of rock weight, tectonic forces, underground water
features of underground structure analytical design theory are: pressure and internal pressure applied to the internal outline
of a lining cross-section have been developed.
– the rock mass and a support are considered as elements of
Another problem difficult to solve was seismic design of
a united deformable system subjected to external loads and
underground structures. The characteristic feature of seismic
actions;
loads is the fact, that seismic waves direction and combination
– rock mass model is linearly deformable medium;
are not known in advance. So, the design of an underground
– construction technology strongly influences upon under-
structure subjected to the action of prescribe directed seismic
ground structure stress state;
wave is not effective, because this direction hardly corre-
– mathematical apparatus of analytical design methods is
sponds to the real direction of incoming waves propagating
mathematical apparatus of the elasticity theory.
in the rock mass during the Earthquake.
Now there are various analytical design methods developed In the work (Fotieva, 1980) an original method of seis-
in the frame of general theory for stress state determination of mic design is proposed. The design consists in determining
shaft supports and tunnels linings subjected to different loads. the most unfavourable lining stress state at any combinations
and directions of long longitudinal (compressive-tensile) and
shear waves propagating in the plane of the tunnel cross-
2 MODERN ANALYTICAL DESIGN METHODS section during the Earthquake of certain magnitude. With the
aim of designing tunnel linings of an arbitrary cross-section
Analytical methods used for the design of circular tunnel shape two plane quasi-static contact problems of the elasticity
and shaft linings are the simplest among modern analytical theory are considered.
744
745
Many fracture mechanisms models have been suggested for particularly flexural strength and tensile strength of SFRS
concrete behavior, such as Hillerborg’s fictitious crack model specimens have increased significantly, the tensile strength
(Hilleborg et al. 1976), Bazant and Oh’s crack band theory corresponding to 4.5 percent steel fiber is two times of
(Bazant & Oh, 1983), Jenq and Shah’s two parameter fracture plain concrete. Also the post-peak behavior of this material
model (Jenq & Shah, 1985). improves with steel fiber content significantly.
In this paper, application of Hillerborg’s theory in Steel The experimental results obtained through the SFRS spec-
Fiber Reinforced Shotcrete (SFRS) is assessed, For this pur- imens were applied in the concrete damaged plasticity model
pose a series of laboratory tests were performed including, based on Hillerborg’s fracture theory using ABAQUS pro-
uniaxial compression, direct tension and flexural tests on gram, for different percentages of steel fiber.
SFRS specimens, then simulate the tests with concrete dam- This model named Concrete damaged plasticity uses two
aged plasticity based on Hillerborg’s theory for behavior of compressive and tensile mechanism of failure. Compressive
shotcrete and finally the results were compared. behavior, elastic modulus and poison’s ratio were extracted
The fracture energy required for extending the crack pro- from uniaxial compression test. Tensile behavior was pre-
posed by Hillerborg’s theory, for using this concept in the dicted in this simulation. The tensile behavior was estimated
present experimental investigation three parameters of uni- using Hillerborg’s fracture energy iteratively until the labo-
axial tension strength ( ft ), fracture energy (GF ) and crack ratory and simulating results were converged. In this way, a
cohesion stress curve (σ(ω)) are needed. tensile behavior for each steel fiber reinforced specimen was
The 28-days compressive strength of shotcrete was deter- suggested.
mined using 15 cm cubes according to ASTM C39 standard. Geometery of the model and loading is exactly similar
Direct tensile strength was performed in a particular man- to four-point flextural test, a three dimensional deformable
ner using cilynrical specimens with 5 cm in diameter and element and fine and middle-size meshes were applied in
10 cm length. In order to control the mode of failure in pure the model. The estimated curves suggest a two-line curve for
tension a weakend section was produced in the middle of tensile post-peak behavior.
the specimens (B. Mehrzad, 2010). Four-point flexural tests The results show a good agreement between laboratory and
on 10 * 10 * 35 cm specimens based on ASTM C1018 for concrete damaged plasticity based on Hillerborg’s theory. A
evaluating the flexural strength were used. two-line post-peak behavior for SFRS is appropriate. For lower
All the tests were carried out using a servo-controlled test- steel fiber contents the model curve lie below the laboratory
ing machine through strain control as the variable parameter curves and for higher steel fiber content it show higher results,
and the complete curves was obtained for each specimen. that means the curves stand above the laboratory results.
The results show that, there is a dramatic change in the
stress-strain curves of specimens containing steel fibers,
746
D.A. Curry
Baker Hughes Incorporated, London, England
A clear understanding of the mechanics of the rock-cutting the interaction of a single PDC cutter with rock and hence
process is necessary to solve many problems of reduced to evaluate the MSE. In our modelling of the rock cutting
drilling efficiency. Assessing penetration rate performance process two type of analysis were carried out: 1) 2-D analy-
objectively is a difficult task. Numerical simulation can be sis of linear cutting and 2) 3-D analysis with curved cutting
a valuable instrument for analysing the particular conditions trajectory. For each model, the cutting simulation was done
of each borehole drilling operation without the often costly under atmospheric pressure (0 MPa) and at hydrostatic pres-
investment required for a full scale experimental investigation. sure (10 MPa), the latter simulating the drilling fluid pressure
In the petroleum industry, the Polycrystalline Diamond acting on the hole bottom when drilling at depth.
Compact (PDC) bit is a popular drilling tool that employs the The aim of this study was to evaluate the rock behaviour
cutting mechanism to shear a rock. It accounts for more than during cutting under pressure. In this condition, the rock
one third of the total footage drilled worldwide. It widespread behaviour is predominantly inelastic, so, the extended
use makes it important to understand the interaction of this Drucker & Prager model was used because it considers
tool with different rock environments and diverse conditions consider materials that exhibit pressure-dependent yield.
of drilling. The Drucker & Prager model provides three different yield
The single cutter test allows understanding the interac- criteria. The differences between these criteria relate to the
tion rock-cutting tool by determining the penetration and shape of the yield surface in the meridional plane, which can
drag forces acting on each cutter located on the bit face. be a linear form, a hyperbolic form, or a general exponent
Furthermore, this laboratory test enables to duplicate many form. The linear form assumes linear dependence of devia-
parameters that affect the forces including the rock type, cut- toric stress on hydrostatic pressure and the hyperbolic model
ter design and wear state, position on the bit, cutter interaction, makes a similar assumption at high confining pressure. There-
cutting speed, rock stress state, and fluid environment. In this fore, these forms of the model do not provide a particular
paper, a numerical model of single PDC cutter is developed good match to triaxial test data. Because the exponent form
to assess the forces of the cutting rock process. of the yield criterion provides the most flexibility in matching
The concept of Mechanical Specific Energy (MSE) has triaxial test data, it was selected for these simulations.
been used effectively to evaluate drilling efficiency, Pessier & For this study, we chose to model Carthage limestone, a rock
Fear (1992). The MSE is defined as the work done per unit whose deformation, fracture and drilling properties have been
volume of rock cut; the higher the MSE for a given drilling extensively studied and for which we had triaxial test data at
environment, the lower system efficiency. Currently, MSE several different confining pressure.
measurement provides the ability to detect changes in the First, we selected the point in each stress-strain curve
efficiency of the drilling system while drilling, Dupriest & corresponding to initial deviation from elastic behaviour for
Koederitz (2005). However, effective numerical simulation of calibrating the yield parameters. Thus, one stress data point
the drilling operation would allow an early estimate of the effi- from each stress-strain curve at a different level of confining
ciency of drilling, making it easier to propose and to improve pressure is plotted in the meridional stress plane. Finally, a
solutions for better drilling performance. least-square fit that minimizes the relative error in stress is
By definition, for a single cutter the MSE is the rock cutting used by ABAQUS to set the yield surfaces in the meridional
force integrated over the distance cut, divided by the volume plane and to determine the model parameters.
of rock cut. Hence, this paper focuses on the variation of the The model used the ductile damage initiation criterion.
cutting force (Fx), i.e, the component of the force parallel to According to this criterion the material starts to deform once
the direction of the movement of the cutter, with the position the shear strain reaches a certain critical plastic strain value.
of the cutter as that is directly related to the variation of the Furthermore, the Drucker & Prager compression type hard-
specific energy. ening option was used to define the hardening behaviour of
Our numerical simulation of cutting process used the com- the uniaxial compression yield stress, as a function of uniaxial
mercial Finite Element Analysis software ABAQUS to model compression plastic strain. The Displacement type option was
747
748
N. Ablyamitov
Odessa Port Plant, Odessa, Ukraine
G. Bich
Chernomorniiproject, Odessa, Ukraine
749
750
M.A. Kanji
University São Paulo, Brazil
Subject: Rock material and rock mass property testing (laboratory and in situ)
1 INTRODUCTION
751
752
L.Q. Guo
College of Manufacturing Science and Engineering, Sichuan University, Chengdu, China
Northwest Institute of Nuclear Technology, Xi’an, China
J.B. Liao
College of Manufacturing Science and Engineering, Sichuan University, Chengdu, China
753
754
755
756
T.H. Kim
Posco Engineering & Construction Co. Ltd., Seoul, South Korea
These days, cities have dramatically expanded due to over- weighted values by core element technology for construction
population, the concentration and extension of urban func- of the infrastructure.
tions. As a result, multiple urban renewal projects have been Lately, Gyeonggi-do announced the Great Train eXpress
planned and executed to strengthen the efficiency of urban (GTX) plan which connects Gyeonggi-do with Seoul while
functions and improve living environment. Because of the the City of Seoul unveiled the U-smart Way project. In con-
constraints of urban ground space, there have been limi- sideration of current social and economic situations, these
tations in the improvement of urban functions through the urban underground space development projects are targeted
horizontal expansion of urban spaces. Therefore, it is nec- to overcome the following problems; increase in the ratio
essary to set underground express ring roads which can be of intercity buses and traffic congestion due to large hous-
highly applicable to urban development and improvement ing development in the outskirts of capital region, difficulties
as target infrastructure, examine the location conditions and in securing spaces and decent ground traffic infrastructure,
ripple effects of the development, collect necessary informa- huge land compensation cost and civil petitions by civil and
tion for conceptual design and estimate weighted values by environmental groups.
core element technology for construction of the infrastructure. The development of underground transport infrastructure
When underground transport infrastructure is constructed in is one solution to the problem of limited ground space. It
the downtown, there are many technical problems to be solved appears that the development project would generate various
such as interaction between underground facilities and the positive effects such as decrease in ground traffic congestion
existing facilities on the ground, tunnel stability, ventilation and preservation of current urban living environment. When
and environmental problems. Hence, this paper has inves- underground transport infrastructure is constructed down-
tigated the matters and technical problems that should be town, there are many technical problems to be solved such
considered during the construction of underground space and as interaction between underground facilities and the exist-
analyzed the technologies that are necessary in developing ing facilities on the ground, tunnel stability, ventilation and
urban spaces based on patent trend analysis in construction of environmental problems. Hence, this paper has investigated
underground spaces. the matters and technical problems that should be considered
There have been limitations in urban regeneration and during the construction of underground space.
expansion of urban functions with the horizontal expansion The prerequisite technologies for urban underground space
of urban spaces only. Therefore, it has been suggested to pro- development have targeted to investigate the current status of
mote urban development and improvement in 3-dimensional conventional technologies based on patent trend analysis, ana-
underground structure. In Korea, there have been efforts to lyze the technical directions of patent applicants and inventors
achieve qualitative urban growth and develop a pleasant urban and figure out the current status of core patents. On the con-
space by overcoming the limitations of ground spaces since trary, the patent trend analysis in this paper has dealt with
early 1990s. technologies associated with the construction of main tun-
Unlike ground facilities, it is difficult to expand or improve nel and shaft, tunnel ventilation and disaster prevention and
underground facilities once they are built. Hence, it is impor- measurement of groundwater for maintenance.
tant to come up with a reasonable and detailed development Figure 1 reveals the possible three GTX lines; KINTEX,
plan based on the precise forecast on the overall conditions of a Goyang-Dongtan New Town (74.8 km), Cheongyangri-
future city. For this, it is necessary to set underground express Songdo, Incheon (49.9 km) and Uijeongbu-Geumjeong,
ring roads which can be highly applicable to urban develop- Gunpo (49.3 km). Gyeonggi-do estimates that the GTX
ment and improvement as target infrastructure, examine the Project may need KRW 13.9 trillion if promoted with gov-
location conditions and ripple effects of the development, col- ernment money and KRW 11.1 trillion if promoted with
lect necessary information for conceptual design and estimate private funds.
757
758
P. Radziszewski
Department of Mechanical Engineering, McGill University, Montreal, Canada
K.E. Waters
Department of Mining and Materials Engineering, McGill University, Montreal, Canada
Mining Engineering, Montreal, QC, Canada
1 INTRODUCTION
759
Figure 2. Norite being cracked at 120 sec in 5 kW. 5 kW, and Figure 3 shows the reduction in tensile strength as
a function of power level and exposure time.
By comparing the power level and the tensile strength value
in 10 and 20 s of exposure of Basalt samples, it can be observed
that in 10 seconds of exposure time, the tensile strength value
is reduced very little. At 20 s of exposure, the tensile strength
was reduced significantly as the power level increased (Fig. 4).
This work shows that the exposure to microwave radiation
can significantly reduce the tensile strength of rocks, which
would lead to reduced wear on breakage equipment. Sig-
nificantly, some samples melted after long exposure times,
therefore optimal exposure times and power levels must be
found for each rock that is being investigated.
760
J.C. Hu
State Key Laboratory for GeoMechanics and Deep Underground Engineering,
China University of Mining & Technology, Beijing, China
School of Civil & Architecture Engineering, Zhongyuan University of Technology, Zhengzhou, China
F. Wang
State Key Laboratory for GeoMechanics and Deep Underground Engineering,
China University of Mining & Technology, Beijing, China
School of Materials & Chemical Engineering, Zhongyuan University of Technology, Zhengzhou, China
J.J. Cui
School of Civil & Architecture Engineering, Zhongyuan University of Technology, Zhengzhou, China
M.C. He
State Key Laboratory for GeoMechanics and Deep Underground Engineering,
China University of Mining & Technology, Beijing, China
Subject: Rock material and rock mass property testing (laboratory and in situ)
Electromagnetic properties of rock and mineral are important Based on the microstructure analysis of electrical signal pass-
in resources engineering, civil engineering and environmental ing rock, we know that when alternating current acts upon rock
engineering. The paper includes three parts as follows. specimen, there are different conductive paths, and called as
SCP, CCP, DCP, ICP.
The simplified microstructure model of conductive path of
2 EIS PHYSICAL MODEL OF ROCK rock is established.
Electrical model is drawn out base on the basis of physical
Based on the experimental device, the microstructure of model, such as figure 2.
electrical signal passing rock is established as figure 1. The total impedance of rock can be drawn from figure 2:
Figure 1. Microstructure of electrical signal passing rock. Figure 2. Electric conduction model of rock.
761
762
Workflow design and software customization to enable integration of lidar data with active
drill-and-blast tunnel construction projects
Lidar imaging is a powerful technology that has recently been resources over the past three years to develop an operational
adopted by geotechnical engineers as an operational tool. The workflow, as illustrated in Figure 1. The workflow is designed
remotely collected data has the ability to enable geotechni- to enable accurate and rapid processing, including feature
cal engineers to perform advanced geometric, mathematic, extraction and interpretation of the lidar data. Through the
and spatial analyses without direct exposure to the rockmass. development of in-house macros and processing standards, the
Often working environments, such as drill and blast tunneling, use of lidar data has become efficient, reliable, and extremely
require immediate results and interpretation while the time robust. This paper outlines the approach taken by NGI to
allotted for on-site data collection is limited by a demanding develop the workflow as well as demonstrate numerous tools
production/advancement schedule. The Norwegian Geotech- currently used on-site in active drill and blast environments.
nical Institute has invested a significant amount of time and NGI has adopted the software suite PolyWorks as the prin-
ciple data processing option for all lidar data. The selection
of PolyWorks was made for two distinct reasons. The first is
its ability to manage and manipulate raw lidar data through
advanced computational algorithms. Secondly, and the focus
of this paper, is the ability to customize the software, create
macros, and expand the capability of the modules within the
PolyWorks framework.
The workflow for processing lidar data, in place at NGI,
is optimized for efficient and reliable results. The steps taken
to process the data (Figure 1) involve the user to combine
datasets, align the data with scan images collected from the
previous blast round, convert the lidar points to surface mod-
els, and extract geotechnically relevant information from the
surface models. The development of the surface model from
the lidar points is a critical process that is computationally
demanding. Through extensive sensitivity testing of different
surface input parameters, optimal values have been deter-
mined and set as user defaults. The setting of input and surface
meshing variables reduces the potential for user error and
variability between individuals processing lidar data.
Figure 1. Lidar data processing workflow developed at, and in This workflow, and the use of internally established param-
place, at NGI. eters, prepares the data for detailed and accurate geotechnical
763
analysis. Upon completion of the conversion from raw lidar blasting, and prior to the application of shotcrete. This struc-
points to a true 3-dimensional surface model, the IMInspect tural evaluation (Figure 2), as well as assessment of shotcrete
module within PolyWorks is used to analyze the model and thickness (Figure 3), shotcrete tolerances, and volume of
extract pertinent features. blasted rock is calculated by the user through semi-automated
Construction of a drill-and-blast tunnelling requires a geo- processes.
logical investigation of the face and back of the tunnel post
764
G. Li & M. Bai
Halliburton
Wellbore-stability problems are frequently encountered when were performed to determine the relative effects of various
drilling deviated wellbores near salt or in tectonically active factors in maintaining wellbore stability. This study focuses
regions. When these unplanned geomechanical issues go on the most-sensitive parameters to develop a better under-
unresolved, it can result in wellbore collapse, lost circulation, standing of the controlling parameters in determination of the
stuck pipe, and casing collapse. The quality of wellbore- wellbore stability. Consequently, the less-sensitive parame-
stability predictions in drilling operations strongly depends ters were filtered out to minimize both the requirement for
on the quality of input parameters. The Quality Control (QC) necessary data entry and computational time.
of input parameters described in this paper includes perform- Case studies using the proposed methods are provided to
ing a comprehensive sensitivity study to identify the key demonstrate the workflow of the sensitivity investigation.
parameters that dictate wellbore stability and eliminating the After providing each input parameter with the designated func-
less-sensitive parameters that are insignificant to wellbore sta- tional distribution (e.g., normal or log normal distributions),
bility. Case studies using the proposed methods are provided the Monte Carlo simulation was launched. The probabilities
to demonstrate the workflow of the sensitivity investigation. of a stable drilling process were computed from the model-
This study explores many factors that affect the stability ing, using the shear-failure gradient as a lower bound and the
of a wellbore, such as the in-situ state of stress, rock proper- fracture gradient as an upper bound to guide the selection
ties from core and logs, and the failure criteria selected. In this of the desired mud-weight windows. The advantage of using
study, methods for determining various input factors using the the Monte Carlo simulation rests on the freedom to select
Wellbore Risk-Analysis (WRA) method with the wellbore- the uncertainty range for each parameter and to indicate the
stability probability assessment by means of Monte Carlo confidence level of each input parameter. Consequently, the
simulations to understand the influence of various wellbore- determined results can be more sensitive and realistic than
stability factors were analyzed first. The sensitivity analyses those calculated from a deterministic model.
765
Figure 1. Illustration of the concept of visual physical model test. Figure 3. Models with single cavity and several cavities.
766
767
X. Luo
CSIRO Earth Science and Resource Engineering, Australia
Directional drilling of a CBM hole can be difficult, risky and An increase of seismic amplitude with the approach of the
costly. One of the key issues in guaranteeing the success of drill bit to the borehole geophones was evident. Good seismic
the drilling is to ensure the drill bit is steered in the correct signal to noise ratio was observed when the drill bit was within
direction. Using drilling noise to locate the seismic source 150 m from the geophones (Figure 2).
(drill bit) is proposed (Figure 1) and trialed in this study. The results have shown that the seismic signals generated
This field trial was carried out at a CBM site in Bowen from drilling inside a coal seam can be efficiently observed
Basin, Australia, where directional drilling in the 5 m thick using triaxial geophones when the drill bit was within 150 m
Pleiades seam (P-seam) at 340 m deep was conducted to from the geophones. Consistent azimuth values of the drill
extract coal gas for commercial use. The distance from the bit were determined when the bit was within 120 m from the
drill rig to the target well is 1,344 m. geophones. Through a combination of the drill bit azimuth
and the drill bit distance from the drill rig, the trajectory of
the drill bit can be determined in real-time.
768
1 INTRODUCTION
where η(x; ω) is a local random medium property, being a where ηj is a structural function corresponding to the
function of a space position and a realization ω, p(ω) is a j-realization of RVE and n is the size of the sample.
probability density function. The crucial role in this approach plays a proper determina-
The definition (1) implies a necessity of generation of tion of the RVE size, NRVE , as well as the size, n, of the sample.
all realizations forming ensemble in order to determine the The both values are related to each other, i.e. larger size of RVE
ensemble average. This exceptionally complex procedure is then a smaller sample size n can be used or inversely smaller
usually overcome using an ergodic hypothesis which allows size of RVE requires necessity of larger size of the sample use.
replacing the ensemble average with a volume average, provid-
ing that a volume of the medium considered tends to infinity,
i.e.: 2 MICROSTRUCTURE DESCRIPTORS
769
REFERENCES
770
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: deformation modulus, rock mass properties, plate loading test; empirical equation. GSI-System
Rapid growth of engineering rock mechanics and rock engi- The source of data used in this paper were obtained from
neering makes the determination of rock mass properties as six different Iranian dam sites and powerhouse locations,
one of the most important requirements in design and con- namely, Bakhteyari, Gotvand, Khersan-3, Chamshir, Parsian
struction of rock engineering projects such as underground and Siahbishe.
excavations, tunnels, dams, etc. Among the rock mass prop- The data used in this study include; the deformation modu-
erties, modulus of deformation is of paramount practical and lus of rock masses measured by plate loading tests, laboratory
economical importance. measured modulus of Elasticity (Ei), Uniaxial Compressive
Though, the in-situ determination of rock mass deforma- Strength (UCS), along with the information obtained during
tion modulus is of vital importance, however, the complex rock core recovery such as; Rock Quality Designation (RQD),
and time consuming procedures makes the method uneco- ground water condition, and all the other parameters required
nomic and unfavorable for some projects. Hence, in the recent to determine GSI and RMR indices. A summary of the
years, predictive models have been employed for the indirect employed data is given in Table 1. Finally, a database includ-
estimation of the deformation modulus. Nonetheless, the pre- ing a total of 150 cases was acquired for further statistical
dictive models results may not represent the realistic values analysis.
of deformation modulus as expected. The frequency histograms of GSI and Em to determine the
The empirical equations for indirect estimation of the defor- number of data are given in Figure 1. The plots help us to
mation modulus are simple; hence, they are unable to include signify density of samples in each parameter. As it can be
some uncertainties with regards to the limited data availability, seen, distributions of the data include a wide range of the
variability of rock type and complex heterogeneous nature of parameters. In other words, the GSI parameter ranged between
the rock masses. Hence, comprehensive investigations on the 34 and 76 (Figure 1a) and modulus of deformation is ranged
integrity of the results obtained by these empirical equations from 1.0 GPa to 40 GPa in which samples less than 20 GPa are
are crucial for rock engineering projects. more grouped than other intervals.
Consequently, several researchers have proposed some A plot of deformation modulus, measured by plate loading
empirical equations to estimate deformation modulus of test, against GSI is shown in Figure 2. As it can be seen,
isotropic rock mass on the basis of classification schemes there is a good and reasonable agreement between these two
such as the Rock Mass Rating, RMR, the Tunneling Quality parameters. The Root Mean Square Error (RMSE) index was
Index, Q, and the Geological Strength Index, GS. also calculated in order to check and compare the performance
In this paper, at first some statistical evaluations of the capability of predictive models developed in this study.
gathered information including plate-loading test data and The Comparative results obtained based on prediction per-
the corresponding rock mass properties have been presented. formances of the proposed models with the existing empirical
Then, two new empirical equations, based on simple and equations leads to the following concluding remarks;
multiple regression analyses, are proposed. Finally, a compar- Considering the RMSE values, the two equations proposed
ison of prediction performance between the existing empirical in this paper can be used as highly reliable predictive models
equations and the new proposed equations is made. as compared with the existing empirical equations.
Statistics
No. of data Valid 130 95 139 150 150 150 101
(PLT2 ) Missing 20 55 11 0 0 0 49
Mean 58.2 2.1 53.3 9.9 89.5 46.2 58.9
Median 60.0 2.0 51.0 7.9 100.0 60.0 56.0
Std. Deviation 21.5 0.7 11.2 8.0 48.6 27.2 11.6
Variance 463.9 0.5 125.0 64.3 2360.6 741.1 135.5
Minimum 0.0 1.0 34.0 1.0 10.0 2.5 39.0
Maximum 98.0 4.0 76.0 39.0 180.0 85.0 81.0
1
Weathering Degree
2
Plate Load Test
771
772
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: rock slope, flexural toppling failure, tilting method, theoretical approaches
Values 600 50–110 15–35 17–22.3 35–40 Figure 2. Comparison between theoretical and experimental results
for a single column.
773
by a small screw jack. Next the frame was fixed on the tilt-
ing machine and then the monitoring equipments were setup.
Finally, the angle of main plate of the tilting machine was
increased gently and slowly till failure occurs or the angle
of tilting reaches at 45 degrees (Figure 2). In the process of
this modeling the dimensions of the models are of paramount
importance. If the dimensions are too small, then the failure
will not occur. On the other hand, as was mentioned earlier
in this paper, the samples are so brittle, hence, construction
and handling of large samples are so difficult. Thus, the max-
imum height of slope models, in this study, was determined
to be as of 42 centimeters. For this limitation, most of our
models were stable during the test and only one model was
unsuccessful (Figure 3).
To assess the results of the existing theoretical approaches,
the stabilities of these models have been analyzed with Majdi
and Amini, and Amini et al methods. The comparative results
are presented in Figure 4. As it can be seen from the figure,
the results of the experimental and the theoretical models are Figure 4. Comparison between the theoretical and experimental
fairly close to and are in a good agreement with each other. results on the basis of Majdi and Amini, and Amini et al. method.
774
775
Keywords: numerical modelling, rock failure, rock properties, rock stress, stability analysis
776
R.A.O. Pedroza
Geotechnical and Mining Department, Faculty of Engineering, Francisco de Paula Santander University,
Cucuta, Colombia
J.C.A. Vargas
Geosciences Department, National University of Colombia, Bogotá, Colombia
P.C.E. Rodríguez
Faculty of Engineering, Javeriana University, Bogotá, Colombia
Subject: Rock material and rock mass property testing (laboratory and in situ)
777
778
Rock mass classification based on seismic measurements using Support Vector Machines
A.J. Rechlin
Scientific Drilling, GFZ German Research Centre for Geosciences, Germany
S. Lüth
Centre for CO2 Storage, GFZ German Research Centre for Geosciences, Germany
R. Giese
Scientific Drilling, GFZ German Research Centre for Geosciences, Germany
779
780
A rock mass classification to estimate excavation rate, support and rock properties
in a borehole survey
T. Shidahara
Newjec Inc., Tokyo, Japan
T. Hatano
Electric Power Development Co. Ltd., Tokyo, Japan
T. Miwa
Tohoku Electric Power Co. Inc., Sendai, Japan (formerly at Nuclear Waste Management Organization of Japan)
Keywords: nuclear repository, site characterization, rock mass classification, tunnelling, rock properties
A method of evaluating engineering rock mass characteristics level. The distribution of excavation rate in the qu-RQD-Jc
has been proposed for the purpose to help the adequate selec- space has been also analyzed in the same way.
tion of the detailed investigation areas, which is an important Relation between 3 parameter-rock mass description and
decision making for the selection of the final disposal site of rock mass properties of strength and Young’s modulus has
high level radioactive waste in Japan. been analyzed based on published in situ rock shear tests and
The method requires the estimation of excavation rate, plate jack tests in Japan. Figure 2 shows the contour lines
support and mechanical properties of rock mass using the (red) for cohesion on qu-RQD graph at Jc = 3. The values in
information from the rock core, which is qu: uniaxial strength the small squares are the cohesion from the in-situ rock shear
of rock core, RQD and Jc: joint condition. tests at dam sites in Japan. The contour lines are drawn based
Relation between the 3 parameters and advance rate of on the values in the squares. It is obvious that higher RQD and
excavation and support has been analyzed based on published qu gives higher cohesion. This qu-RQD graph has also been
records of tunnel excavation. Each of qu, RQD and Jc is cat- made for other values of Jc. Lower Jc, which means better
egorized into 5 groups. Jc is categorized in the same way joint condition, gives higher cohesion. The contour lines for
as in Rock Mass Rating system, RMR. Figure 1 shows the Young’s modulus on qu-RQD graph have also been obtained
distribution of support classes for tunnels excavated by drill in the same way.
and blast in the qu-RQD graph at Jc = 3. The average numbers We can roughly estimate excavation rate, support, shear
of rockbolt are written at corresponding positions in the field strength and Young’s modulus by the contour lines obtained
with colored ellipses. The degree of support has been catego- in this study like Figure 1 and 2, if qu, RQD and Jc are deter-
rized into 3 classes of A, B and C. A is light support level, B mined. All of these 3 parameters are able to be obtained in a
is medium support level, and C is heavy support level. This borehole survey. The feature of this method is that it uses the
qu-RQD graph has been made for each Jc value. It is found 3 dimensional information of qu, RQD and Jc without con-
that higher RQD and Jc, and lower Jc require lighter support verting the information to one dimensional value as in RMR.
781
A. Serrano
Universidad Politécnica deMadrid, Spain
1 INTRODUCTION
2 BASIC HYPOTHESES
782
Parameter Unit
4 CONCLUSIONS
783
H. Soroush
Well Engineering Centre of Excellence (WECOE), Weatherford Oil Tool M.E., Dubai, UAE
A. Fahimifar
Department of Civil and Environmental Engineering, Amirkabir University of Technology, Tehran, Iran
Subject: Rock material and rock mass property testing (laboratory and in-situ)
Keywords: rock mass classification, rock properties, lab testing, rock failure
784
The life-cycle information management practice of Shanghai Yangtze River Tunnel and
Bridge Project
Keywords: tunneling, Shanghai Yangtze River Tunnel, life-cycle information management, project information management,
construction information systems, operation information management
Shanghai Yangtze River Tunnel & Bridge Project is an extra- introduced. The difficulties of the overall life-cycle informa-
large transportation infrastructure project that is an important tion management for ShanghaiYangtze River Tunnel & Bridge
part of the express way from Shanghai to Xi’an. The com- Project such as large site location, multi-parties involved, huge
pletion of this project will improve the structure and layout amount of information data and new technologies, equipments
of transportation system in Shanghai, speed up the regional and materials adopted were illustrated.
economic integration of the Yangtze River Delta, better pro- Second, in order to better carry out information manage-
mote the economic developments of the Yangtze River area ment for such an extra-large river-crossing project and to
and even China overall, and enhance overall economic com- improve the efficiencies for management, the engineering
petitiveness of Shanghai within China. The project starts from information management system was established to achieve
wuhaogou in Pudong, crossing Yangtze River South Channel the project progress control, investment control, contract man-
waterway by tunnel to Changxing Island and spanningYangtze agement, quality control, construction management, design
River North Channel waterway by bridge to the Chongming management, document management, security management,
Island. The total length of the project is 25.5 km long, within materials management and equipment management.
which the length of the tunnel is 8.95 km (the length of river- Third, for such a project, all kinds of information and
crossing is 7.5 km), and the length of the bridge is 16.63 km data from survey, design and construction should be received,
(the length of river-crossing is 9.97 km). There are many
difficulties of a huge river-crossing project, such as large
scale, large investment, long construction period, complex
technologies, multi-parties involved, high risk and long site
locations.,etc. Then information management has become the
effective method for construction and operation management
for such a project. For this, overall life-cycle information
management was set as the target, the project information
management system, construction information management
system, operation information management system are estab-
lished to effectively manage all the in formations from project
management, design, construction and operation of such an
extra-large project, that have effectively improved the effi-
ciency and quality of the construction and operation of the Figure 2. Universal fitting shield tunnel segment arrangement.
Shanghai Yangtze River Tunnel & Bridge Project.
This article first described the trend of information sys-
tem applications for civil engineering project. The details of
the Shanghai Yangtze River Tunnel and Bridge Project were
785
786
S. Utsuki
Hazama Corporation, Japan
Subject: Rock material and rock mass property testing (laboratory and in situ)
Keywords: case studies, rock mass, rock properties, rock mass classification
1 INTRODUCTION
2 CONTENT OF IN-SITU EXPERIMENTS is improved to CM class at all the measurement points. More-
over, about the comparison of rock types, it is clear that the
The borehole load test is comparatively simple and used to igneous rocks are made harder than the sedimentary rocks.
measure the deformability of rock masses in the borehole.
So, in order to confirm the effects of improvement about the 3.2 Examinations of making rock masses homogenized
mechanical properties by grouting quantitatively, the borehole To make rock masses homogenized is to make variety of the
load tests were executed before and after grouting at the same mechanical properties small after grouting. It is shown in
position of boreholes. Figure 5 that the ratio of the modulus of deformation before
and after grouting is small as the modulus before grouting is
3 EXAMINATIONS OF EFFECT OF IMPROVEMENT large originally, on the other hand, it is large as the modulus
ABOUT DEFORMABILITY OF ROCK MASSES before grouting is small. As a result, it is clear that rock masses
are made homogenized by grouting.
3.1 Examinations of making rock masses hard
3.3 Examinations of making rock masses unified
Here, the result of tests is shown in Figure 2 that indicates
the relation between the modulus of deformation obtained at To make rock masses unified is to make joints fulfilled with
the same point in the boreholes before and after grouting. This cement, and the difference between the modulus of deforma-
figure shows that the rise of the modulus is admitted in almost tion and the modulus of elasticity is made small after grouting.
all the points, so it is clear that the rock masses are made hard In Figure 7, at many measuring points, D/E ratio after grouting
by grouting. And it is also seen the tendency roughly that is larger than before grouting, so it is thought that, as the joints
the effect of improvement is small when the modulus before are fulfilled with cement, the difference between the modulus
the grouting is large originally, on the other hand, it is large of deformation (D) and the modulus of elasticity (E) is made
when the modulus before grouting is small originally. And it is decreased and as a result, it is clear that rock masses are made
remarkable in CL-1 class so that the modulus of deformation unified after grouting.
787
B.R. Chen
Institute of Rock and Soil Mechanics, the Chinese Academy of Sciences, Wuhan, China
Keywords: neural network; numerical modeling; rock properties; back analysis; rock failure
During recent several decades, different theories and tech- This kind of methods has been successfully used in mod-
niques from the fields of artificial intelligence, system science eling the creep behavior of some engineering softrock and
and nonlinear science have been systematically investigated the mechanical behavior of limestone under different water-
for different purpose of modeling tasks in rock mechanics chemical environment (Yang et al. 2008; Chen et al. 2010).
and rock engineering. In these methods, the rock mechanics Modeling based on global optimal search of nonlinear
models are established through nonlinear relationship map- parameters. By adding some nonlinear items to simple mod-
ping, global optimal search, self-learning, self-organization els (e.g., linear elastic), we can also find another way to build
and data mining in a data-based modeling manner. The results up nonlinear models. In these cases, the nonlinear items should
are encouraging and present us another alternative method- be carefully derived to obtain a reasonable level of nonlinear-
ology for rock mechanics modeling (Feng et al. 1997; Feng ity. Usually, the nonlinear levels are dependent on a series of
2000). This paper gives some brief review of these works and parameters, for example, the coefficients of polynomial, the
makes some summery issues concerning with the application exponential parameters of the time dependent item, et al. The
aspects. purpose of the optimization problem is to find these struc-
Modeling based on nonparametric regression. Because ture parameters of rocks that will result in good agreement of
the mechanisms of most of the constitutive factors are farm simulating results with lab or field tests. For nonlinear rock
form well-known. Sometimes, it is impossible to derive a materials, the objective function is a highly nonlinear function
mathematic expression incorporating all those parameters to of parameters. It is very difficult to obtain an analytical evalua-
establish a reasonable model. For these cases, nonparame- tion of the function gradients. Those optimization procedures,
ter modeling techniques can provide potential alternative. Of such as Genetic Algorithms (GAs), Particle Swarm Optimiza-
these methods, Neural Networks (NN) provided rich, powerful tion (PSO), requiring no knowledge of gradient information
and robust nonparametric modeling framework with proven of the objective function, provide powerful solving tools. We
and potential applications across sciences. By selecting differ- have made systematically investigation of these methods and
ent input and output variables, we can build up different styles have proposed a series of algorithms for the optimization prob-
of models. It is worth to note that the new-found constitutive lem. Here, a GA procedure is employed to explain the solving
affects, such as the time-dependent (t), hydraulic, chemical process.
parameters, et al., can be conveniently incorporated into the The solution procedure starts from a population of candi-
model. If provided relative testing data, the NN model can date solutions. Each possible solution is used as real structure
learn the unknown mechanism of these factors. parameters of rock mechanics models for output analysis
The underlying relationship can be learned directly from to predict the response of ui at the observation points. The
testing data by train process of the NN model. Of these kinds response obtained from the output computation is then com-
of method, efficient learning algorithms and sufficient test- pared with those obtained from the lab or field tests to evaluate
ing data containing enough response information are of great the fitness of the population of the candidate solutions. The
importance. For the former, we have evolved a series of opti- estimated fitness values are then used by the GA operations to
mal algorithm, e.g., Genetic Algorithm (GA) optimized NN, obtain a new population of the parameter sets and thus evolve
to approve the conventional gradient search procedures and into a new generation. The population of candidate parameter
have been proved with powerful self-learning capacity. As to sets is updated until the prescribed stopping criterion is met.
the sufficiency of the testing data, different lab experiments For practical application, the output analysis procedure can
have to be designed to investigate each constitutive factor be an analytical or numerical computation process, depend-
under consideration. On the other hand, the macro response of ing on the testing data. Usually, if global load-deformation
rock structure, such as the load-deformation and excavation- data is used, a numerical procedure should be employed to
deformation responses, may contain abundant constitutive link the testing data to the stress-strain response.
information. This information can also be extracted through Using these kinds of modeling methods, we established the
numerical computation and be used for NN learning. nonlinear elastic constitutive model of laminated composites
788
789
Keywords: bolts and anchors, lab testing, numerical modeling, rock support, tunneling, weak rock
790
Author index
791
792
793
794
795
796