ISRM newsjournal 2019 final

NEWS J URNAL ISRM 12 Representative members of ISRM NGs of Bosnia and Herzegovina, Bulgaria, Croatia, Hungary, Macedonia and Slovenia (Omis, Croatia). Concerning international cooperation and exchange, Prof. Shimizu and Prof. Mato Uljarević (President, Prof. He carried out several activities. He visited the Vietnam ISRM NG of Boznia and Herzegovina) Institute of Mining Science and Technology (VIMSAT) on February 26-28, where the collaboration plan on the 110 Prof. Shimizu attended the ISRM Specialized Conference mining technology was discussed. held in Omis, Croatia from 11-13 April 2019 and joined a Prof. He attended the First International Conference on meeting at the conference to discuss the promotion of the High-stress Rock Mass Mechanics, in Russia, from July 15 activities of ISRM NGs in the Balkan region. to 18, in order to debate and reach a consensus on brics He was invited to a regional conference held in Bosnia international cooperation projects. In this meeting, he also and Herzegovina to deliver a lecture and to be recognized presented a keynote lecture entitled “Latest Progress of as an Honorary Member by the Geotechnical Society of Research on Rockburst Mechanism & Its Control”. Bosnia and Herzegovina (ISRM NG) for his contribution CSRME organized the first scientific expedition to the Arctic to the Society. Arctic region, Svalbard, on August 6-18. Prof. He led the In 2019, he participated in three other ISRM Specialized team as chief scientist, and Prof. Charlie Li, ISRM VP for Conferences held in China, Japan and Vietnam and gave Europe, and Prof. Fawu Wang, President of the International them his support. Consortium on Geodisaster Reduction ( ICGdR), were invited From 2015-2019, during his term as Vice-President at Large, to join the expedition. Prof. Shimizu visited 18 countries, attended more than 30 In the scope of the collaboration with ICGdR, Prof. He conferences, symposia and seminars, delivered 25 lectures, attended the 17th International Symposium on Geo-disaster and collaborated with 10 ISRM NGs in order to encourage Reduction, in Kyrgyzstan, from 19 to 23 August, and made a their activities on behalf of ISRM. keynote lecture entitled “An Accurate Prediction Method for Geo-disasters”. Besides his routine tasks as ISRM Board member, Porf. He organized the 4th Early Career Forum as a special session in the 14th ISRM International Congress, in Foz do Iguassu, Brazil where 10 junior delegates from developing countries in South America were selected and sponsored to participate and deliver presentations. Prof. Norikazu Shimizu, as a member of the Technical Oversight Committee, cooperated in the analysis of the commissions’ reports and in the preparation of the Annual Report 2019 of TOC. As a member of the Young Members’ Committee, he supported the organization of the 40th West Japan Rock Engineering Symposium held in Kumamoto (Japan) with 49 student participants from China, Egypt, Indonesia, Japan, Korea and Thailand. Prof. Shimizu chaired the Organizing Committee of YSRM2019&REIF2019 (the 5th Young Scholars’ Symposium on Rock Mechanics and the International Symposium on Rock Engineering for Innovative Future) held as an ISRM Specialized Conference in Okinawa (Japan) from 1-4 December 2019. 274 participants from 20 countries and regions, including 157 students, got together to share their knowledge and experiences and to form friendships. 49

December 2019 13 2019-2023 BOARD ISRM President Reşat Ulusay presenting the 50th Anniversary Commemorative Book of ISRM (1962-2012) to Mr Nenad Šušić, ACTIVITIES the President of Serbian Geotechnical Society Finish NG Rock Mechanics Day opening session of the Conference. More than 150 delegates The Finish National Group of the ISRM organized on 17th attended to this event which is the largest gathering of this September 2019 his Rock Mechanics Day (Kalliomekaniikan kind in Serbia. Päivä 2019), which takes place every two years. The day was Prior to the Vrnjačka banja conference, the ISRM delegation held at the House of Sciences (Tieteiden talo) in the center visited the Faculty of Civil Engineering in Belgrade. The of Helsinki on October 17, 2019 and was funded by the purpose of the visit was to meet with the President of the Finnish companies FinMeas and Fractuscan. Serbian Society for Rock Mechanics, Mr Milan Tričković, It included talks by several international experts. Leandro with the aim of aiding the activities of the society. It was Alejano, ISRM Vice-President for Europe was invited to give concluded that the Serbian Society for Rock Mechanics, a talk about the post-failure behavior of rocks and rock which was relatively inactive in the recent years, can masses and its application to the mine pillar design. Johan take part in more of the ISRM activities in particular to Spross from KTH Stockholm talked about risk management support the Early Carrier Forum (ECF) activities by sending principles in rock engineering works and Noemi Boldrini a young researcher to the next EUROCK 2020 conference from IDS Georadar talked about slope monitoring. Among in Trondheim. Finnish experts, topics such as seasonal underground As part of the visit, the delegation was also received by storage of thermal solar energy in hard crystalline rocks the Dean of the Faculty of Civil Engineering, Dr Vladan were discussed by Mateusz Janiszewski, a PhD student Kuzmanović. Here, the discussion emphasised the need at the University of Aalto (Helsinki) or the numerical to introduce the subject of Rock Mechanics to the modeling of the effect of thermal shock weakening on undergraduate course in Civil Engineering. resistance to rock compression by Martina Pressacco of the University of Tampere. From left to right, Lauri Uotinen, President of the Finnish Rock From left to right: Prof. Ivan Vrkljan, Past ISRM Vice-President at Large, Mechanics Society, Johan Spross Professor at KTH (Stockholm) and Prof. Vladan Kuzmanović, Dean of the Faculty of Civil Engineering at Leandro Alejano, ISRM VP for Europe on the day of the Finnish the University of Belgrade, Prof. ReŞat Ulusay, ISRM President; Dr Vojkan Society of Rock Mechanics Jovičić, ISRM Vice-President at Large; Mr Milan Tričković, President of Serbian Society for Rock Mechanics, and Dr Mirjana Vukičević, Chair for Visit of ISRM delegation to Serbia Geotechnics at the Faculty of Civil Engineering. A delegation of ISRM, including the President Reşat Ulusay, Vice-President at Large Vojkan Jovičić and Past Vice- President at Large Ivan Vrkljan, took part in the national Conference organized by Serbian Geotechnical Society at Vrnjačka banja. The conference entitled Geotechnics in Civil Engineering is held bi-annually in Serbia and gathers mainly civil and mining engineers with a strong presence of engineering geologists. All the members of the ISRM delegation were invited to deliver Keynote Lectures: Reşat Ulusay on the topic of “Present and Future Testing, Rock Characterisation and Monitoring with Emphasis on ISRM Suggested Methods”, Vojkan Jovičić on the “Particular Conditions for Tunnel Construction in Karst” and Ivan Vrkljan on “Contribution of Karl Terzaghi to Engineering Geology, Geomorphology, Rock Mechanics and Tunnelling”. In addition, Reşat Ulusay gave an opening address speech and also performed a short presentation about ISRM in the 50

NEWS J URNAL ISRM XVI Panamerican Conference on Soil Mechanics and 13 Geotechnical Engineering The XVI Panamerican Conference on Soil Mechanics and Geotechnical Engineering was held in Cancun, Mexico, from November 17th to 20th, 2019, organized by the Mexican Society of Geotechnical Engineering. The Conference was very successful with more than 600 participants from several countries. The objective was to present the experiences of projects and discuss the advances in technology, integrating the main branches of geotechnics. Initially there was a pre-conference course on \"Rock Excavations\", and in Session 17 “Rock Mechanics”, the Co-Chairs Carlos Carranza Torres (USA) and Valentin Castellanos Pedroza (Mexico), introduced the Keynote Lecturer and 8 papers, as well as posters. The technical presentations included rock excavations, site characterization, mines, tailing dams and hydropower. In the occasion, the ISRM Vice-President for Prof. Eduardo Alonso delivering the 20th Šuklje Lecture LatinAmerica, Prof. José Pavón, exposed the The audience was also addressed by the ISRM Vice-President main objectives and purposes of the Society and the main at Large, Dr Vojkan Jovičić, who introduced the ISRM activities in order to achieve this goals, and revealed how to mission and activities to the members of the Slovenian join the ISRM, and the benefits to members. society. Dr Vojkan Jovičić took the opportunity to announce the forthcoming ISRM workshop “Recent Trends in Rock Mechanics”, which is jointly organized by the Slovenian Society for Research in Road and Traffic Infrastructures (DRC), the Slovenian Geotechnical Society (SLOGED) and the Institute for Mining, Geotechnology and Environment (IRGO). The workshop will take place at Radisson Blu Plaza Hotel in Ljubljana on 31st of January 2020 as a parallel event to the meeting of the newly elected ISRM Board. The workshop presents a unique opportunity to listen to the contributions of the leading experts in the field, as each member of the ISRM Board is going to deliver a lecture at Participants in the Rock Mechanics session, from left the workshop. Additionally, representatives of the regional to right, Carlos Carranza-Torres, Valentin Castellanos, national societies of Croatia, Bulgaria, Slovenia, North Melissa Miuzzi, José Pavón and Natalia Giraldo Beltrán Macedonia, Bosnia and Herzegovina, and Serbia are going to contribute to the workshop with their lectures. Šuklje Day Šuklje day is an annual event of Slovenian Geotechnical Society, held this year on November 22 in Bled, which celebrates the memory of late Prof. Lujo Šuklje, who was one of the pioneers of Geotechnical Engineering in Eastern Europe. The highlight of the event is the Šuklje Lecture, which was delivered in 2017 by the current ISRM President Prof. Reşat Ulusay. This year, the 20th Šuklje lecturer was Prof. Eduardo Alonso, from the Universitat Politecnica de Catalunya, who delivered a lecture under the title “Slow- Fast Landslide Interaction”. The event was attended by 150 participants, including delegates from Austria, Croatia, Hungary and Northern Macedonia. Dr Vojkan Jovičić presents the mission and the main activities of ISRM 51

December 2019 14 FORTHCOMING ISRM SPONSORED CONFERENCES 28 - 30 2020 5th Symposium of the Macedonian Association for Geotechnics - an ISRM Specialized Conference May Ohrid, Macedonia 13 - 19 Junho EUROCK 2020, Hard Rock Excavation and Support - the 2020 ISRM International Symposium 15 - 19 Junho Trondheim, Norway 15 - 18 September 23 - 27 October XIII International Symposium on Landslides – a Joint Technical Committee on Natural Slopes and 11 - 13 November Landslides (JTC1) from FedIGS sponsored conference Cartagena, Colombia 21 - 25 9 - 11 SBMR 2020 - 9th Brazilian Rock Mechanics Symposium - an ISRM Specialized Conference Campinas, Brazil ARMS11 - 11th Asian Rock Mechanics Symposium, Challenges and Opportunities in Rock Mechanics - an ISRM Regional Symposium Beijing, China CouFrac2020 - the 2nd International Conference on Coupled Processes in Fractured Geological Media: Observation, Modeling, and Application Seoul, Korea 2021 EUROCK 2021, Mechanics and Rock Engineering from theory to practice - the 2021 ISRM June International Symposium September Torino, Italy 5th International Workshop on Rock Mechanics and Engineering Geology in Volcanic Fields an ISRM Specialized Conference Fukuoka, Japan 13-17 2022 EUROCK 2022 - Rock and Fracture Mechanics in Rock Engineering and Mining 20 - 22 June an ISRM Regional Symposium 9-14 Helsinki, Finland September IX LARMS - Latin American Congress on Rock Mechanics, Rock Testing and Site Characterization an ISRM Regional Symposium Asunción, Paraguay 2023 15th ISRM International Congress in Rock Mechanics October Salzburg, Austria The ISRM holds International Congresses on Rock Mechanics National Groups seeking to host an ISRM Regional and Rock Engineering, at four year intervals, on themes of Symposium or Specialized Conference shall submit a written general interest to the majority of the membership, and proposal to the Secretariat, at least one, but preferably sponsors a co-ordinated program of International Symposia, two to three years before the date of the event. Their Regional Symposia and Specialized Conferences organised organization is ruled by By-law No. 5, and application by National Groups of the Society. forms are included in specific Guidelines prepared by the The annual ISRM International Symposium is chosen from Board, and available on the ISRM website (https://www. the ISRM Regional Symposia that take place in that year isrm.net/conferencias/submit.php?show=conf). Since 2018 and is the venue for the annual meetings of the Council, no financial contribution to the ISRM is due from ISRM Board, and Commissions of the Society. ISRM Specialized Specialized Conferences. Conferences are events of a smaller nature, usually focused Proceedings of ISRM conferences are stored in the ISRM digital on a specific theme. library available in the OnePetro platform (onepetro.org). 52

NEWS J URNAL ISRM MÜLLER LECTURE 15 Peter Kaiser | Laurentian University, Canada FROM COMMON TO BEST PRACTICES IN UNDERGROUND ROCK ENGINEERING This article provides a summary of the 8th Müller lecture delivered by the author in memory of Prof. Leopold Müller at the 14th ISRM Congress in Brazil. For the detailed reasoning on why and how to move from common to best practices in underground engineering, the reader is referred to the conference proceedings (pp 141-179; or www.mirarco.org/category/news_events/) ABSTRACT: Common practices are not necessarily best practices when judged from an economic or workplace safety perspective. As in other engineering disciplines, it is necessary to systematically improve engineering design practices. This lecture addresses some deficiencies in common practice that may lead to flawed or ineffective rock engineering solutions. In the past, common practices that worked well at shallow depth may now need to be replaced as the rock mass behavior has changed and poses new hazards at depth. This lecture focuses specifically on opportunities resulting from better means to assess the vulnerability of excavations, to characterize the rock mass, for ground control, and rockburst damage mitigation. Theoretical considerations and field observations are used to justify the proposed changes and to highlight practical implications and benefits. 1. THE STORYLINE Enormous progress has been made over the last decades Müller (1963), in the preamble to ‘Der Felsbau’, stated “Der and the new knowledge has found application through Felsbau ist auf dem Wege, eine wissenschaftliche Disziplin engineering standards, ‘ISRM suggested methods’, or zu werden. (Construction in rock ... is on the way to become common engineering practices. This is also reflected in the a scientific discipline.)” Over the last 50+ years, the science recent change of ISRM’s name to ‘International Society of rock mechanics and its implementation through rock of Rock Mechanics and Rock Engineering’. Today, as in engineering has evolved into a mature discipline, and 1963, new challenges force us to leave our comfort zone new challenges emerge as we tackle larger and deeper of common practices and develop more sophisticated and, excavations in civil construction and in mining. at the same time, simple, more efficient and effective rock He elaborated “Mit der Größe von Projekten wuchs die engineering practices. Verantwortung des Ingenieurs und des Baugeologen “Everything should be made as simple as possible, but not ganz bedeutend und zwang dazu, die bis anhin geübte simpler” (attributed to A. Einstein). Common practices gefühlsmäßige Behandlung dieser Aufgabe mehr und mehr are often too simple and have frequently led to failures zu verlassen und eine Theorie des Felsbaues zu erarbeiten.” and costly mistakes, largely because the fundamental (With the size of projects, the responsibility of the engineer understanding of rock behavior and response to and the construction geologist grew quite significantly and construction or mining are not fully reflected in experience- forced us to leave intuitive treatment practices behind to based approaches. Best practices make it simple, reflecting develop a theory of rock engineering). In the spirit of his all essential or dominant Engineering Design Parameters vision, it follows that rock engineering evolves over time (EDPs) – no more, no less. and common practices need to be questioned and improved During the MTS lecture at the 50th US Rock Mechanics to arrive at best practices that lead to safer and more Symposium in 2016 on ‘Underground rock engineering to economic solutions. match the rock’s behavior – Challenges of managing highly stressed ground in civil and mining projects’, the author suggested that dichotomies exist and gaps between reality and current practices have to be closed by the application of recent advances in rock mechanics research to arrive at sound rock engineering solutions. 53

December 2019 15 “A robust rock engineering solution in underground mining recent experience from major mining (caving) operations, or construction must respect the complexity and variability this lecture builds on lessons learned from moderate to of the geology, consider the practicality and efficiency of severe excavation instabilities in mining. construction, and provide safe and effective rock support. The vulnerability to damage or the proximity of an For this purpose, it is essential to anticipate the rock mass excavation to failure is assessed by comparing load, and excavation behavior early in the design process, i.e. displacement and energy demands to the capacity of rock at the tender stage before excavation techniques are support systems. The probability of failure or the factor chosen and designs are locked-in in construction contracts. of safety can then be assessed and appropriate support Whereas it is possible in most engineering disciplines to systems can be selected. select the most appropriate material for a given engineering Once the vulnerability is established, questions such as ‘How problem, in rock engineering, a design must be made to severe or violent will the failure be?’, i.e. ‘How fragile is the fit the rock, not vice versa. excavation?’ arise. The fragility of an excavation describes the likelihood of damage at a defined severity, e.g. in terms of Lessons learned from volume of displaced rock, cumulative displacement imposed excavation failures tell on the support, or the violence in terms of energy release. us that stressed rock at In other words, critical Engineering Demand (or Design) depth is less forgiving Parameters (EDP) have to be selected to establish the and that advances in vulnerability and fragility of an excavation. The development rock mechanics demand of best practices therefore starts with the identification of a sound comprehension all dominant EDPs (Kaiser & Cai, 2020). of the behavior of stress- From common to best practice damaged rock near Identification of EDPs is a prerequisite to determine whether excavations. Comprehension common practices are applicable or flawed. For excavation in this context means design and support selection, EDPs characterizing the explaining all observations vulnerability and fragility must be clearly identified and then such that fiction can be used to recognize potential failure modes. separated from reality and Under static loading, the excavation behavior can be engineering models and characterized by two dominant EDPs defining the 3 methods become congruent [Symbol] 3 excavation behavior matrix (Kaiser & Cai (2020). with the actual behavior of The horizontal axis represents the Rock Mass Quality (RMQ) a rock mass.” (Kaiser 2016b). or rock mass strength and the vertical axis the stress level SL = σmax/UCS = (3σ1- σ3)/UCS. Because many engineering The rock mass quality is grouped from RMQ1, for massive approaches and common to discontinuously jointed rock, to RMQ2, for fractured and practices are flawed or blocky to disintegrated ground, and RMQ3, for weak and ineffective, and may soft, highly fractured or ,sheared rock. The boundaries have even be obsolete, best been slightly adjusted from Kaiser et al. (2000) based on practices have to be Fig. 1 - Path of discovery to arrive at developed based on best practices .experiences with excavation failures at depth to: Q’ >40, 40 to 0.4, <0.4 with Q’ = modified rock hypotheses of the underlying causes for the deficiencies. tunneling quality index Q for Jw/SRF = 1; Once these hypotheses have been verified by testing and . RMR (Rock Mass Rating) >75, 75 to 35, <35; or field observations, improved solutions can be found and . GSI (Geological Strength Index) >70, 70 to 30, or <30. implemented. Eventually, new knowledge and experiences The stress intensity is grouped into three ranges: SL <0.4, 0.4 may be acquired and the innovation cycle may restart as to 1.15 and < 1.15. indicated in Figure 1. Using the stress level as an EDP is helpful as the impact of the in-situ stress ratio k= σ1/σ3 or mining-induced This lecture focuses on how to move to best practices in stress ratio km= σ1m/σ3m is accounted for. This distinction three key areas of rock engineering: is particularly relevant when mining changes the mining- induced stress field, and failure modes may change, e.g. due . excavation vulnerability and fragility assessment; to relaxation. . rock mass behavior and characterization for rock mass As indicated earlier, best practices must be simple but . strength determination; and reflect all essential EDPs. For the identification of potential excavation failure modes, unnecessary complexities can be ground control and support selection. eliminated and the rock mass can be, as suggested by Terzaghi (1946), characterized by block size (with boundaries at 10 2. EXCAVATION BEHAVIOR cm and 1 m; Fig. 2) and joint condition (with boundaries at Understanding excavation and rock mass behavior is a Jc = 3 and 0.25 in GSI chart (Hoek et al. 1995)). prerequisite for successful rock engineering and therefore for the development of best practices. Only if all potential excavation failure mechanisms are understood can the vulnerability of an excavation to failure and the potential severity (extent or violence) of damage be anticipated, and the resulting critical load, displacement, and energy demands be established. Because of the author’s 54

NEWS J URNAL ISRM For the selection of appropriate rock engineering 3. ROCK AND ROCK MASS BEHAVIOR 15 Once the excavation behavior and failure processes are .approaches, it is, in order of priority, necessary to: understood and relevant EDPs identified, the designer has Establish whether the intact rock strength dominates: to obtain representative rock mass strength envelopes RMQ1 versus RMQ2 or 3. In ‘good’ rock masses the key and then capture the rock and rock mass behaviors using EDPs relate to the intact rock and rock block strength. representative yield or failure criteria. Kaiser (2016b) The block size does not matter and engineering methods elaborates on common practices that often lead to . applicable to blocky rock mass models are not suitable. unrealistic rock and rock mass peak strength envelopes. Establish whether inter-block characteristics dominate: S-shaped or tri-linear failure envelopes are often required . distinguish RMQ2 from RMQ3. to properly describe brittle rock failure (Kaiser & Kim, 2014). In RMQ2, interlock contributes to the rock mass Furthermore, it is discussed that the ‘residual strength’ strength and the key EDP describes the block size as it introduced in soil mechanics often leads to unrealistically . dominates the excavation behavior; and low post-peak rock mass strength because the strain levels In RMQ3, the characteristics of infilling dominate and near underground excavations are much lower than those the joint condition becomes the key EDP. needed to reach the residual strength. Recognizing this sensitivity led to the development of the GSI with particular focus on weak and soft rock. From common to best practice From common to best practice Excessive effort is often expended in common practice to Ample opportunities and economic as well as safety benefits collect and rate rock mass details that have little impact can be derived by moving from common to best practices on the excavation behavior and consequently on a design. in rock strength determination. Common practices of fitting Meanwhile, dominant factors, e.g. persistence, veining, and peak and residual strength data are clearly not best practice. rock alteration, are either ignored or underrepresented. Best practices must ensure that rock mass quality classes .It is concluded that: RMQ1 to 3 and the stress level as well as their evolution Much care has to be taken to obtain representative over the duration of a project are properly defined. UCS-values by identifying failure modes and separating In mining, it is important to reflect on the stress path homogeneous from heterogeneous specimen (Bewick et causing changes in stress level that may lead to changes . al. 2015). in failure mode. Failure envelopes are often s-shaped or tri-linear when cohesion and frictional strength components depend on plastic straining. This is valid for the peak . and post-peak strength. Unless numerical models are calibrated to capture post- peak strength degradation, mobilized post-peak strength limits should be defined for anticipated plastic strain limits. Fig. 2 - Rock mass quality grouping for excavation failure and EDP selection. Left: GSI block size description. Right: simplified grouping of rock mass quality RMQ1 to 3 superimposed on chart from Hutchinson & Diederichs (1996) 55

December 2019 15 As explained by Kaiser (2016a, b) based on work by Diederichs (2007), tensile stresses induced during deviatoric loading of heterogeneous rock lead to Griffith-type extension fracturing with the consequence of a depressed failure envelope in the low confinement zone where fracture propagation causing spalling is not suppressed by the available confining pressure. The resulting failure envelope for a rock mass is therefore s-shaped and the stress space can be divided into two distinct behavioral zones: spalling dominated stress fracturing at low confinement (to the left of the spalling limit), and shear rupture dominated behavior at high confinement. This divides the rock mass surrounding an excavation into two zones, an ‘inner’ and an ‘outer’ shell as shown in Figure 3. Fig. 3 - S-shaped failure criterion for brittle rock masses (center) and zoning of stress space for inner (left) and outer (right) shell behaviors in underground rock engineering (red contour = σ3 = 12 MPa) Inner shell engineering problems are those (a) (b) dominated by the behavior of the rock mass Fig. 4 - (a) Unidirectional bulking due to stress-fracturing (photo) reflected in cartoon-like Voronoi in the zone immediately surrounding an model; and (b) semi-empirical bulking factor charts (modified from Kaiser 2016c) for excavation and excavation where the confinement is low, mining-induced bulking. i.e., where stress-fracturing can occur and blocks or fragment can rotate. Engineering challenges of support design, strainbursting, etc. fall into the class of inner shell problems. On the other hand, engineering problems related to pillar instability, including pillar bursting, fall in the outer shell class where shear rupture dominates because spalling is partially or fully suppressed. In the inner shell, stress-fractured and bursting ground bulks when deformed past the peak strength of the rock mass. This leads to unidirectional bulking deformations that are controlled by the excavation geometry and the imposed tangential strain (Fig. 4a). This directional bulking process is not captured by dilation models that relate the strength to the volumetric strain (bottom of Fig. 4a). For preliminary design purposes, the bulking factor can be estimated from the chart in Figure 4b (Kaiser 2016c). 56

NEWS J URNAL ISRM From common to best practice 15 Economic as well as safety benefits that can be derived from differentiating between near wall (inner shell) and outer shell behaviors. This is particularly valid for support designs where bulking deformations from statically and dynamically deformed stress-fractured rock dominate the rock mass and support behavior. 4. ROCK MASS CHARACTERIZATION “Den Werkstoff Fels in seinem Zustande und seinem Verhalten zu beschreiben, ist die erste Aufgabe dessen, der seine Felsbauwerke sicher anlegen, zweckmäßig konstruieren, schön gestalten und wirtschaftlich ausführen Fig. 5 - Evolution of data collection to eliminate or confirm geomechanics assumptions at möchte.” or “Describing the construction different stages of a rock mass characterization program material, the rock mass, in its condition and its behavior is the first task … to safely lay out, sensibly construct, …, and economically execute works in rock.” (Müller Assumptions have to be made in the early stages of a 1963). As discussed previously, the development of best characterization program, assumptions that are inferred practices in rock mass characterization therefore starts with from comparable rock formations and empirical rules the identification of critical EDPs including related rock mass from similar rock masses. Next, uncertainty has to be characteristics and properties. incrementally removed by reducing the variability and move toward ‘probable’ data. For this purpose, targeted data collection focused on specific assumptions is to be used and In underground construction, the rock mass has to be documented to ensure the credibility of information. Finally, described long before access to observe its behavior is possible. This demands a systematic approach of rock mass assumptions can be removed and replaced by ‘measured’ data or ‘proven’ quality ratings when sufficient factual data quality quantification moving from ‘inferred’ to ‘proven’ is available. Some assumptions may never reach the status rock mass quality (Section 4.1), a practice that is often ignored or rarely systematically executed. of ‘measured’, e.g. the in-situ stress, and back-analyses of other monitoring data may be required to reach the ‘proven’ class. An effective rock mass characterization program, including This approach is illustrated by the opposing wedges in Figure 5. logging, mapping and laboratory testing, needs to collect and interpret features that are relevant for clearly defined From common to best practice purposes. How to overcome challenges of rock mass Best practice in rock mass characterization includes characterization for underground construction in deep a registry of assumed EDPs and tracks the path from mining is covered by Kaiser et al. (2015). The recommended assumption to fact. While EDP-values have to be inferred methodology is not repeated here. Two aspects are however during the preliminary design phase, good engineering addressed in this lecture to assist in the application of practice demands that they are raised to probable and then best practices in rock mass characterization: (1) stages of to a proven status with minimal remnant uncertainty. characterization to move from ‘inferred’ to ‘proven’ quality descriptors, and (2) the suitability of classifications for rock Many ‘unexpected’ problems can be anticipated if the full spectrum of possible ground conditions was properly mass characterization. described. In this manner, claims for ‘changed conditions’ can be prevented because construction techniques suitable for 4.1 Stages of rock mass characterization – from assumption to fact uncertain conditions can be selected. Rock mass characterization involves many model-building 4.2 Characterization versus classification components: geological, structural, rock mass, and hydro- Ideally, a rock mass is characterized in a comprehensive geological models. These models are being developed in manner such that all engineering questions can be answered. an incremental and iterative manner with data initially However, because different engineering tasks require collected during scoping studies from boreholes. The data different EDPs, most rock mass classification systems focus is then gradually refined during follow-up studies and on one or two applications. Strictly speaking only part of the eventually during the construction phases by mapping classification deals with the characterization of the rock mass and back-analysing monitoring results. What is often (block size and interlock, condition and persistence of joints, ignored in common practice is to clearly define and register etc.), the remainder deals with the application for a defined assumptions such that the characterization program can purpose (support selection, rock mass strength determination, focus on eliminating one assumption after another to etc.). For this reason, and because each classification system replace them with factual information. considers different parameters and assigns different weights, As for mineral resource definition increased knowledge and specific classification indices (Q, RMR, MRMR, GSI, etc.) should confidence need to be built from ‘inferred’ to ‘indicated’ to not be correlated. They should be used independently ‘measured’ or from ‘inferred’ to ‘probable’ to ‘proven’ (on for the purposes defined by the developers and for the left of Figure 5). ground conditions that form the foundation for the classification system. 57

December 2019 15 Even the GSI (Geological Strength Index), while not “… we tend to underestimate the rock mass strength for providing guidance for support selection, was developed underground construction at depth” as discussed in the with a bias toward the characterization of a rock mass in 13th ISRM lecture. Common practices fail to capture the the low confinement range near excavations (slopes or strengthening effect of interlock in non-persistently jointed inner shell of tunnels). Furthermore, the GSI was introduced rock where failure through intact rock adds strength, and for rock mass strength determination and “to address two geometric bulking leads to a more rapid strength gain at principal factors considered to have important influences on elevated confining stress. the mechanical properties of a rock mass, i.e., the structure (or For situations where blocks exist and block rotation is blockiness) and the condition of the joints” (Hoek & Brown, possible, Hoek & Brown (2019) present GSI-based strength 2019). Strictly speaking, the GSI-system was developed for equations for isotropic rock masses. The authors indicate that rock mass strength estimation for blocky to weak ground. the underlying GSI-experience stems from excavations in rock While each classification system stands on its own masses where block rotation contributes to the failure process. merit, the indiscriminate use of rock mass classification As mentioned previously, Bewick et al. (2019) provide without consideration of the limit of applicability is guidance for rock mass strength estimation when the strongly discouraged. limitations of the GSI-strength equations are reached. In For example, conventional rating systems such as RMR, Q other words, it is not the GSI that is not applicable, it is the and GSI were developed and calibrated for conditions not applicability of the GSI-strength equations that is limited. dominated by large mining-induced stress changes and From common to best practice stress-fracturing of strong rock blocks. Hence, they are often The common practice of indiscriminate use of the GSI- not applicable, for example, for defected rock and in large strength equations for rock masses of RMQ1 and part of strain environments. If the GSI is indiscriminately applied to RMQ2 tends to underestimate the rock mass strength. For conditions other than those used to develop the GSI-based massive to moderately jointed rock masses with GSI > 65, strength equations, the resulting rock mass strength tends the systematic methodology for estimating equivalent rock to be underestimated. mass strength parameters, outlined by Bewick et al. (2019), Other deficiencies of rock mass rating systems are addressed should be adopted. This methodology compliments the HB- in more detail in the ISRM on-line lecture on “Challenges GSI approach for rock mass strength estimation of a massive of Rock Mass Strength Determination” (Kaiser 2016a) and to moderately jointed rock mass. the impact of rock mass heterogeneity on in-situ stress variability is discussed by Kaiser (2016b). 5. SUPPORT SELECTION From common to best practice Various rock mass classification systems (e.g., Q, RMR, etc.) Benefits can be derived by moving from common to have found wide application for support selection. These best practices in rock mass characterization. Too many approaches are suitable if conditions at a project match parameters are frequently collected that, in the end, do those that form the underlying classification databases, i.e. not matter, and essential parameters are ignored or only civil tunneling data. They may however not be suitable when collected late in a project when unexpected rock mass conditions change as they do in mining where stresses and qualities (changed conditions) are encountered. failure modes change. Targeted rock mass characterization should be guided by Whether a support is selected based on risk- or factor of identified EDPs and by the eventual intent (engineering tasks). safety-based design approaches, load, displacement, and For example, the Q-system was originally intended for support energy demands are compared with respective support selection and then expanded for TBM and other applications, capacities. For static conditions, the support design whereas the GSI-system was developed for rock mass is commonly dominated by load equilibrium (wedge strength determination. Both should be used with discretion stability) or displacement compatibility (squeezing ground) for the intended purposes. They can be used to identify considerations. For dynamic earthquake or rockburst RMQ-classes that matter for the performance of underground loading, it may be necessary to also consider energy excavations (RMQ1 to 3). The influence of stress (and water) equilibrium. It is therefore mandatory to estimate three should be treated separately at the characterization stage and demands (load, displacement and energy) and respective only considered when used for engineering designs. Limits of capacities of the integrated support system. applicability need to be respected. 5.1.1 Capacity estimation The load, displacement, and energy capacities of individual 4.3 Rock mass strength estimation support components are obtained from pull-out (direct Some of the challenges in rock mass strength estimation for loading) or split tube (indirect loading) tests in the the design of deep underground excavations are covered by laboratory or in the field. Respective component capacity Kaiser (2016a). More recently, the GSI-approach has been data are available from the literature (e.g. Cai & Kaiser, 2018) ‘updated’, not because fundamental changes were required, or site-specific values are obtained from field tests. However, but to “… discuss many of the issues of its utilization a methodology to establish the capacity of integrated and to present case histories to demonstrate practical support systems utilizing all individual support components applications ...” (Hoek & Brown, 2019). In a companion is required and this is discussed in the lecture notes. paper Bewick et al. (2019) provide guidance for rock mass strength estimation when the limitations of the GSI- strength equations are reached. This was required because 58

NEWS J URNAL ISRM 5.1.2 Demand estimation 15 Load demands are obtained by estimating the volume of anticipated unstable ground and by assessing the remnant load capacity after sequential support mobilization. Displacement demands are commonly obtained Fig. 6 - Energy–displacement characteristics of a support system (black) loaded by analytical or numerical models with non-linear via a mesh-reinforced shotcrete surface support; and remnant capacity (red) as a constitutive models. Both models are deficient function of applied displacement (central deflection between bolts) when geometric bulking of stress-fractured ground in the inner shell of the excavation dominates the displacement demand. As a consequence, commonly adopted capacity models to represent individual support components are often limited, particularly in stress-fractured rock where the displacement demand by the fractured rock is underestimated. Energy considerations are adopted when kinetic energy been imposed on the support. This is called ‘proactive or demands from earthquakes or rockbursts are anticipated. preventive support maintenance’ (PSM). Unfortunately, energy cannot be measured rendering energy- based designs unverifiable. However, because the product 5.2.2 Support capacity restoration by PSM of displacements times bolt forces defines the energy Once some support capacity has been consumed, the consumption of a support component and the integrated support capacity can be restored by adding bolts offering an support system, displacements provide an indirect measure extra displacement capacity. An example is presented in the of the energy consumed by the support system. It is for this lecture notes to illustrate that the energy capacity increases reason that a deformation-based support design approach by 35% when adding a row of cablebolts after a bolt head is introduced by Kaiser (2014) and discussed in the lecture displacement of 70 mm. notes. Displacement-based designs have a major advantage Proactive support maintenance is a practical and often in that both displacement demands and capacities can be economical means to increase workplace safety and reduce measured and compared with model outputs. the potential severity of excavation damage. From common to best practice 5.2 Support system capacity (SSC) estimation While single pass support systems are cost-efficient, relying An integrated support system is made-up of compatible on the capacity of the original installed support system support components with load–displacement characteristics can lead to uneconomic and potentially unsafe conditions of individual support components obtained from static or because mining-induced deformations consume the dynamic laboratory or field tests. These components have support capacity. As mining proceeds, both the remnant to work together (in parallel) to provide rock retention, displacement and energy dissipation capacity decrease and reinforcement, and holding functions (Cai & Kaiser, 2018). the vulnerability of an excavation increases. The capacity of each component is first mobilized and then When designing a support system best practices consider consumed until it fails. The methodology of SSC estimation the potential loss of the support capacity over the life is discussed in the lecture notes. of a support system. A proper design focuses on the remnant capacity at the time when loading conditions are 5.2.1 Support system capacity consumption (SSCC) most critical, e.g. at the time of dynamic loading by an The effectiveness of support systems can be compromised anticipated seismic event. by quality deterioration and by the consumption of a If support capacity consumption is identified as a critical support system’s displacement and energy capacity. design criteria the best or most economic practice may Mining not only causes stress changes but also produces involve proactive support maintenance. associated displacements which strain the support. As these displacements increase, part of a support’s displacement and 5.3 Self-supporting capacity of a support system – energy dissipation capacity is consumed. This is called support Gabion concept capacity consumption and is reflected in the support energy The most fundamental principle of rock support is to make consumption plot (shown in red in Fig. 6). The support system the reinforced ground self-supporting by creating a stable capacity (black) for this illustrative example is gradually lost rock/support arch. until all of its capacity is consumed at 200 mm imposed central displacement. How to recognize support consumption in the field is discussed in the lecture. The practical implication of support system capacity consumption is that a support system rarely exhibits the full capacity available at the time of installation. By analogue, if the capacity of a support system can be consumed, it must be possible to restore its capacity by installing additional support after displacements have 59

15 (a) December 2019 (d) (b) (c) Fig. 7 - (a) Supported rock arch principle (Hoek et al. 1995); Gabion concept: (b) support of slope; (c) representation of self-supporting wall rock arch, and (d) flat wall equivalent showing tangential resistance forces (yellow, vertical) and radial confining forces (orange, horizontal) For wall support in cave mining, the equivalent model to 2016c). For detailed discussions of the support selection Lang’s self-supporting arch model (Fig.7a by Lang (1961)) process, the reader is also referred to Kaiser (2014). is to create ‘gabions’. Just like for slope stability (Fig. 7b), A deformation-based support design to manage bulking gabions have to provide: .aims at two fundamental support design axioms: . immediate retention of broken or fractured rock; Control of the cause for bulking by minimizing the . reinforcement of broken rock in the inner shell; tangential straining of the rock in the immediate . surface pressure increasing the self-supporting capacity vicinity of an excavation (resist HW/FW closure as . of broken rock; and . illustrated by the yellow arrows in Fig. 7d); and Control of the geometric bulking of stress-fractured bulking restraint by reinforcement. ground by rock reinforcement inside the gabion and the In this manner, a support system is created that behaves like application of confining pressure by the gabion to the surrounding ground (Fig. 7d). .a gabion or a wall arch (Fig. 7c) and Means for estimating static and dynamic support system provides a radial resistance (orange arrows in Fig. 7d) to demands are presented in the lecture. . the rock mass behind the gabion; and From common to best practice provides tangential resistance (yellow arrows in Fig. 7d) to Ample benefits can be derived by moving from common resist the tangential strain driver (HW/FW convergence). support selection to deformation-based support system Why and how a gabion, consisting of a robust surface design practices. Common practices of support design support system, works to ensure full utilization of the without giving due consideration to the displacement capacity of the rock mass reinforcement is outlined in the demand and consumption are flawed and common practice lecture. In an nutshell, both the tangential load and energy of pure energy-based design for burstprone ground is highly dissipation capacity as well as the radial support capacity is flawed. Both can lead to serious workplace hazards. superior when an integrated support system mobilizes the Deformation-based support selection constitutes best collective capacity of the rock mass and the support. practice for ground control in highly stressed, brittle failing As is illustrated in the lecture notes, the gabion concept ground. In burstprone ground, the displacement and with varying support system components finds applications energy demand from strainbursts and from remote seismic in a wide spectrum of rock mass behavior; i.e. from bursting events must be simultaneously considered. It is essential to to blocky to yielding and even squeezing ground. compare the demands to the remnant capacity and not to the capacity of the originally installed support system. 5.4 Deformation-based support selection for stress- fractured rock 6. MOVE TO BEST PRACTICES! In stress-fractured ground, two mechanisms affect the Common practices are often not best practices when judged excavation performance during construction: (a) raveling from an economic or workplace safety perspective, and of broken rock resulting in short stand-up times, and (b) common practices that seem to work well at shallow depth large deformations caused by geometric bulking imposing may need to be replaced because the rock mass behavior large radial deformations on the support. The first is met by has changed and poses new hazards at depth. ensuring robust rock retention and the second by providing This lecture specifically addresses opportunities resulting a deformable bolting system. from better means to assess the vulnerability of excavations, The challenge of controlling highly stressed brittle rock to characterize the rock mass, for ground control, and in civil and mining projects with deformation compatible rockburst damage mitigation. Benefits from moving from support was addressed in the written version of the Sir common to best practices are identified in the following areas: Allan Muir Wood lecture entitled “Ground Support for Constructability of Deep Underground Excavations” (Kaiser 60

NEWS J URNAL ISRM - Identification of engineering design parameters EDPs that Diederichs, M. S. 2007. Mechanistic interpretation and practical 15 characterize the vulnerability and fragility of underground application of damage and spalling prediction criteria for excavations. deep tunneling, Canadian Geotechnical Journal, 44(9): 1082–1116. DOI: 10.1139/T07–033. - Rock mass characterization that follows a systematic process of moving from inferred to proven rock mass Hoek, E. & Brown, E.T. 2019. The Hoek-Brown failure criterion quality designations. and GSI -218 edition. Journal of Rock Mechanics and Geotechnical Engineering, 11(3): 445-463. - Grouping of rock mass qualities into three classes (RMQ1 to 3) that reflect three characteristic rock mass and Hoek, E., Kaiser, P.K. & Bawden, W.F. 1995. Rock Support for excavation behavior modes. Underground Excavations in Hard Rock. A.A. Balkema, Rotterdam, 215 p. - Methods to obtain appropriate rock and rock mass strength envelopes for peak, post-peak and residual strength. Hutchinson, D.J. & Diederichs, M.S. 1996. Cablebolting in Underground Hard Rock Mines, Bitech Publishers Ltd., - Practices that respect the limitations of classification and Richmond, BC, Canada, 406 p. characterization systems, in particular, in the use of GSI- strength equations for rock mass strength determination Kaiser, P.K. & Cai, M. 2020. Rock support to mitigate rockburst for good rock. damage caused by dynamic excavation failure, Volume II in Rockburst Support Reference Book. MIRARCO Laurentian - Differentiation between near wall (inner and outer shell) University, preliminary ISBN: 978-0-88667-097-9 (to be behaviors for support design and pillar sizing. released late 2020). - Deformation-based support selection for ground control Kaiser, P.K. 2017. Ground control in strainbursting ground - A in highly stressed, brittle failing ground. critical review and path forward on design principles. 9TH Int. Symp. on Rockbursts and Seismicity in Mines, Santiago, - Utilization of the self-stabilizing capacity of well- Chile, 146-158. constrained broken rock by adopting the gabion concept to provide effective support to resist static and dynamic Kaiser, P.K. 2016a. Challenges in rock mass strength load demands. determination for design of deep underground excavations. ISRM on-line lecture; https://www.isrm.net/gca/?id=1227 - Quantification of support capacity consumption as a critical design criteria. Kaiser, P.K. 2016b. Underground rock engineering to match the rock’s behaviour - a fresh look at old problems. MTS lecture - Use of proactive support maintenance (PSM) based on at 50th US Rock Mechanics Symp., 6., abridged summary in support deformation monitoring to restore consumed ISRM news 2017. support capacity. Kaiser, P.K. 2016c. Ground Support for Constructability of - Replacement of energy-based by displacement-based Deep Underground Excavations - Challenge of managing support designs for burst-prone ground. highly stressed brittle rock in civil and mining projects. ITA Sir Muir Wood lecture at World Tunneling Congress, San Best practices take, at the design stage, the evolution of Francisco, 33 p. www.ita-aites.org or http://www.mirarco. deformation demands over the life of an excavation into org/grc/#ert_pane1-4 account. Proper designs focuses on the remnant support capacity at the time when loading conditions are most critical. Kaiser, P.K. 2014. Deformation-based support selection for tunnels in strain-burstprone ground. DeepMining’14, ACG ACKNOWLEDGEMENTS (eds. Hudyma and Potvin), 227-240. The author wish to acknowledge the many contributions of Kaiser, P.K., Diederichs, P.K., Martin, C.D., Sharp, J. & Steiner industrial sponsors, Rio Tinto through the Rio Tinto Centre W. 2000. Underground works in hard rock tunneling and for Underground Mine Construction, Freeport McMoran, mining. GeoEng 2000, Melbourne, Australia, Technomic LKAB, Vale, Glencore, as well as the financial contributions Publ. Co., 1: 841-926. of NSERC (Natural Sciences and Engineering Research Council of Canada). Kaiser, P.K. & Kim, B-H. 2014. Characterization of strength of intact brittle rock considering confinement dependent Much of the presented material is a result of long-term failure processes. Rock Mech Rock Eng.; DOI 10.1007/ collaborations and efforts of graduate students. They are s00603-014-0545-5. acknowledged in the written text of the lecture. Kaiser, P.K., McCreath, D.R. & Tannant, D.D. 1996. Rockburst REFERENCES Support. In Canadian Rockburst Research Program 1990- Bewick, R.P., Kaiser, P.K. & Amann, F. 2019. Strength of massive 95. Vol.2, 324 p. to moderately jointed hard rock masses. Rock Mechanics Lang, T.A. 1961. Theory and practice of rockbolting. Trans. Amer. and Geotechnical Engineering, 11(3): 562-575. https://doi. Inst. Min. Engrs, 220: 333-348. org/10.1016/j.jrmge.2018.10.003 Bewick, R. P., Amann, F., Kaiser, P.K. & Martin, C.D. 2015. Müller, L. 1963. Der Felsbau. F. Enke Verlag, Stuttgart, Vol.1, 624 p. Interpretation of UCS test results for engineering design. 13th Terzaghi, K. 1946. Rock defects and loads on tunnel supports. ISRM Congress of Rock Mechanics, Montreal, Canada, 14 p. Cai, M. & Kaiser, P.K. 2018. Rockburst phenomena and support In Rock tunneling with steel supports, (eds R. V. Proctor characteristics, Volume I in Rockburst Support Reference and T. L. White) 1, Youngstown, Commercial Shearing and Book. MIRARCO Laurentian University, preliminary ISBN: Stamping Company, 17-99. 978-0-88667-096-2, 191 p. 61

December 2019 16 ROCHA MEDAL LECTURE Qinghua Lei | ETH Zürich & Imperial College London CHARACTERISATION AND MODELLING OF NATURAL FRACTURE NETWORKS: GEOMETRY, GEOMECHANICS AND FLUID FLOW Natural fractures are ubiquitous in crustal rocks and often dominate the bulk properties of geological media. The understanding of their geometrical, geomechanical and hydrological properties is a challenging issue which is relevant to many rock engineering problems. In this paper, I first present a study of the statistics and tectonism of a multiscale fracture system and propose an interpretation to the underlying mechanism driving fracture network evolution. To model the geomechanical behaviour of natural fractures in rock, a joint constitutive model is implemented in the context of a hybrid finite-discrete element method. This geomechanical model is used to simulate the damage evolution around an underground excavation in a crystalline rock embedded with pre-existing fractures. The model is also applied to derive the aperture distribution of various metre-scale fracture networks under in-situ stress conditions, based on which stress-dependent fluid flows are analysed. A novel upscaling approach employing discrete-time random walks is then developed to extrapolate fracture network geometries together with their variable apertures into larger scales for permeability prediction. The research findings demonstrate the importance of realistic fracture network representation and systematic geomechanical simulation for understanding the hydromechanical behaviour of fractured rocks. 1. INTRODUCTION 2. GEOMETRY OF NATURAL FRACTURE NETWORKS Fractures such as faults and joints often form complex The growth and interaction of natural fractures create networks in subsurface, resulting in highly disordered a hierarchical geometry that may exhibit long-range geological conditions. The widespread presence of natural correlations from macroscale frameworks to microscale fractures raises a fundamental question about the underlying fabrics. The proximity of the connectivity state of natural mechanisms that drive such complicated evolutionary fracture networks to the percolation threshold remains an and collective phenomena. These fractures produced by unresolved debate. It was argued earlier that natural fracture mechanically-controlled processes often exhibit self- systems are close to the percolation threshold (Renshaw, organised (i.e. non-random) population statistics (Lei et al., 1997), because the driving force (tectonic stress or hydraulic 2017a). Fractures, along which rupture has caused cohesion pressure) is abruptly released once the system is connected, loss and mechanical weakness in the rock, often dominate and a diminished mechanical strength and an enhanced the strength and deformation of geological formations; they hydraulic conductivity are likely to occur. However, extensive may also serve as conduits or barriers for fluid and chemical field observations suggest that natural fractures may be migration in subsurface. The characterisation and simulation well-connected and significantly above the threshold (Barton, of the effects of natural fractures on the hydromechanical 1995). The geometrical scaling of a fracture population behaviour of geological media is of central importance for provides clues for a better understanding of the geology and a variety of engineering applications such as hydrocarbon physics behind the statistics. extraction, geothermal production, groundwater remediation The fracture networks located in the Languedoc region of and geological disposal of radioactive waste (Rutqvist and Southern France are studied. This area has been affected Stephansson, 2003). by multiple tectonic events: normal faulting in the Jurassic, strike-slip faulting and thrusting faulting in the Late 62 Cretaceous to Eocene, and normal faulting in the Oligocene. A series of outcrop patterns exposed at the Earth’s surface over various scales is mapped (Fig. 1a) to measure the statistics of the fracture system (Figs. 1b-d). The fracture spatial organisation is governed by a fractal dimension D ≈ 1.65, while the length distribution is characterised by a power law exponent a ≈ 2.65. The relationship between a

NEWS J URNAL ISRM 16 Fig. 1 - (a) A compilation of multiscale fracture patterns from the Languedoc region in Southern France. Statistical analysis of (b) the spatial organisation, (c) length distribution and (d) distance-length correlation of the fracture system and D, i.e. a ≈ D+1, indicates that this multiscale fracture connectivity considerably lower than pc. Thus, in response to system may be self-similar (Davy et al., 2010). Mathematically, later tectonic events, further cracking may occur within the the connectivity of a self-similar fractal population is scale network until the system once again becomes connected. invariant. However, the percolation parameter p of these 3. GEOMECHANICS OF NATURAL FRACTURE fracture patterns varies significantly at different scales, NETWORKS ranging from 4.60 (close to the percolation threshold pc ≈ A geomechanical model for simulating natural fractures 3.6-3.8) to 14.69 (much larger than pc). is developed based on a hybrid finite-discrete element An understanding of the process by which the natural method (FEMDEM) (Munjiza, 2004) with an empirical joint fracture networks evolve may offer an explanation to this constitutive model (Bandis et al., 1983; Barton et al., 1985) paradox. Fracture networks in rock develop over geological implemented (Lei et al., 2016), such that the complex non- time by the superposition of successive fracture sets linear, scale-dependent strength and deformation behaviour each linked to a different stress regime and set of crustal of natural fractures can be captured. The model is verified by conditions. Thus, there is a strong possibility that early the consistency between the numerical results and empirical fracture sets may become partially or totally cemented as the solutions for a direct shear test of different sized fracture network evolves and fluids move through it. These sealed or samples (Fig. 2). partially sealed early fracture sets may act as barriers to fluid The geomechanical model that can simulate both the flow and the integrity of the rock has been to some extent reactivation of pre-existing fractures and the propagation recovered. Although the network geometrically remains of new cracks is used to study the excavation damaged almost the same, its “effective” connectivity has been reduced zone evolution around tunnel excavation in crystalline rocks well below the percolation threshold. As a result, subsequent (Lei et al., 2017b). As shown in Fig. 3, after the excavation, stress fields could continue to propagate new fractures until an interior low stress zone (stress loosing zone) is formed the critical state is re-established. However, if the “apparent” surrounding the tunnel boundary, where intensive rock connectivity of trace patterns is measured without taking mass failure develops as a result of structurally-controlled into account their internal sealing conditions, it is likely to kinematic instability (e.g. key blocks) and stress-driven derive a percolation state significantly above the threshold. In brittle fracturing (e.g. wing cracks). The stress loosing zone addition, the intrinsic anisotropy of the fracture network may seems to have a long axis along the direction of the in- also permit tectonic energy to accumulate in other directions plane minimum principal stress. An exterior high stress zone which have a higher mechanical strength/stiffness and can (stress arching zone) is promoted at a certain distance to accommodate more new cracks. the tunnel periphery, where compression arches seem to To test this concept, the percolation parameter of the evolve along the direction of the in-plane maximum principal progressively developed fracture networks at the end of each stress. More interestingly, the high-stress contours of these different formation stage is calculated (Lei and Wang, 2016). arching zones tend to be microscopically constrained by the Generally, the first stage fracture set exhibits a connectivity structures of pre-existing fractures. Furthermore, extensive state close to pc, consistent with the postulation of energy tension-dominated new cracks accompanied by a few shear- relief at the connecting moment (Renshaw, 1997). However, dominated ones are created around the excavation. because of the possibility of early fractures becoming cemented as has been observed in the Languedoc area (Petit 63 et al., 1999), a fracture network which at the time of its formation was at pc may subsequently have an “effective”

December 2019 16 Fig. 2 - Verification of the numerical model based on a direct shear test of different sized fracture samples under a constant normal stress: (a) shear stress-shear displacement curves, and (b) dilational displacement-shear displacement curves Fig. 3 - Stress redistribution and fracture development around tunnel excavation in fractured rocks 64

4. HYDROMECHANICS OF NATURAL FRACTURE NEWS J URNAL NETWORKS ISRM The numerical model is employed to investigate the effects 16 of in-situ stresses on fluid flow in 2D fracture networks (Lei et al., 2014), including an analogue fracture network from single-phase steady state flow simulations (Fig. 5). The (AFN) mapped based on a limestone outcrop at the Bristol mean permeability of the multiple DFNs shows a reasonably Channel Basin and its random discrete fracture network (DFN) good match with that of the AFN in the x direction equivalents. Ten DFN realisations were generated based on associated with an initially good connectivity state, whereas the measured statistics of the AFN. A series of plane strain a significant discrepancy was observed in the y direction with numerical experiments is designed with the far-field effective a connectivity close to the critical percolating state (more stresses applied at a range of angles to the AFN and DFNs sensitive to geomechanical effects). The high variability in the (Fig. 4). A comprehensive comparison between the AFN DFN simulation results suggests that large uncertainty may and DFNs is made with respect to various geomechanical exist in the random DFN approach, especially when the matrix responses such as shear displacement, crack propagation, is less permeable and flow is more dominated by fractures. hydraulic aperture and network connectivity. The equivalent The quality of DFNs for hydromechanical modelling of permeability of the stressed 2D fractured rocks is derived fractured rocks is considered to be governed by the accuracy in representing geologically-formed network geometries and geomechanically-controlled fracture apertures. Fig. 4 - Stress distribution in the AFN and one of its DFN equivalent under in-situ stresses applied at different angles 65

December 2019 16 Fig. 5 - The equivalent permeability components kxx in the x direction and kyy in the y direction, and the permeability anisotropy ratio kxx/kyy of the AFN and DFNs with the matrix permeability km assumed to be 0.1, 1 and 10 mD The geomechanical model is also applied to investigate the with respect to the change of polyaxial stress orientation, as flow heterogeneity in an idealised 3D persistent fracture revealed by the simulation results of a 3D fractured limestone network caused by both the fracture-scale roughness effect layer embedded with realistic joint sets (Lei et al., 2017c). and the network-scale fracture interactions under polyaxial Geomechanical modelling of the development of fracture in-situ stresses (Lei et al., 2015b). The system involves three patterns and apertures on a scale spanning the laboratory orthogonal sets of persistent discontinuities with one specimen to a few metres has been achieved to high horizontal set of bedding planes and two vertical sets of accuracy. However, due to the limits of processing power, fractures. Significant heterogeneity in stress and displacement it is currently impossible to directly extend this accuracy distributions is generated when the stress ratio is high (Fig. to macroscale computations. Hence, upscaling is required 6a). The deformation of the fractured rock under a high to estimate important subsurface properties of naturally stress ratio results in a highly variable aperture field in single fractured rocks at larger scales based on models established fractures (Fig. 6b). Very large apertures are clustered in some at a smaller scale. Natural fracture systems often exhibit local areas, which seem to be connected and form a slightly significant self-similarity and scaling behaviour (Bonnet et al., diverted vertical channel from the top to the bottom of the 2001), the understanding of which opens the possibility that domain. The increased stress ratio leads to a considerable hydromechanical properties of a macroscale fractured rock increase in rock mass permeability over several orders of may be estimated based on the characterisation of its crucial magnitude (Fig. 6c). The permeability tensor also changes features from a relatively smaller sample (Lei et al., 2015a). from isotropic to highly anisotropic with the increase of stress The scaling properties of a 6 m × 6 m natural fracture ratio due to the deviatoric stress acting with respect to the system (Fig. 7) are first examined in terms of spatial favourably oriented vertical fractures, resulting in zigzag- organisation, lengths and connectivity using fractal shaped localised pathways and very high permeability in the geometry and power law relations. The fracture pattern subvertical direction (Fig. 6d). In addition to the effect of is observed to have a fractal dimension D ≈ 2 (i.e. stress ratio, the rock mass permeability also varies significantly 66

NEWS J URNAL ISRM 16 Fig. 6 - (a) Shear displacements and (b) hydraulic apertures of a persistent 3D fracture network under polyaxial stress loading. (c) Variation of the rock mass permeability as a function of the stress ratio. (d) Permeability tensor and flow pathways of the fracture network under different stress conditions homogeneous space filling), while its length distribution growth scheme with the important natural fracture tends to follow a power law with an exponent a ≈ 2.37. A characteristics (e.g. non-planarity, segmentation, local smaller domain of size L = 2 m (Fig. 7b) is used for FEMDEM clustering and length scaling) preserved (Fig. 8). The simulation to derive the distribution of fracture apertures equivalent permeability of the growth networks is derived and shear displacements under a hydrostatic or deviatoric from single-phase flow simulations with the matrix stress condition. This smaller pattern also serves as the permeability assumed to be 1 × 10-15 m2. As shown in Fig. source for network upscaling. 9, with the increase of the scale, the permeability of the By assuming the fracture pattern repeats itself in fractured rock in the deviatoric stress scenario displays an progressively larger and larger Euclidean space, a novel upward trend at the small and intermediate scales (<10-20 upscaling scheme is developed to extrapolate the m) and a continued downward trend at larger scales (>20 geologically-mapped fracture geometry together with its m), whereas the permeability in the hydrostatic case mainly stress-dependent, spatially-variable displacement attributes shows a downward trend except a slight increase in the y into larger scales using a recursive growth lattice (Figs. direction at the small scale (<10 m). Fracture networks under 7c-d). There are two types of cells in a growth lattice: the the deviatoric stress condition appear to be more permeable source cell that is the reference for network repetition, and than those under the hydrostatic condition due to the effects the growth cell that is a clone of the source cell sharing of shear dilations in response to differential stresses. common geostatistics. The fractures are classified into Two factors may dominate the permeability scaling trend: censored (partially sampled) and uncensored (completely (i) the length exponent a that governs the connectivity observed) types (Fig. 7b). The growth of censored and scaling of a fracture population (Berkowitz et al., 2000), and uncensored fractures in a growth cell is implemented in (ii) the scaling exponents of fracture apertures and shear different ways due to their distinct geostatistical features. displacements which control the transmissivity scaling of A censored fracture in a growth cell evolves from a nucleus each individual fracture (de Dreuzy et al., 2002). For the located on the lattice edge and propagates following a studied case of 2 < a < D+1, with the increase of domain random walker, while an uncensored crack hatches from size L, the number of fractures larger than L (i.e. traversing the barycentre randomly seeded inside the cell and fractures) increases as ~L-a+D+1 (Davy et al., 2006), propagates as two synchronised walkers jogging towards whereas the relative percentage of such fractures decreases opposite directions. as ~L-a+D-1. Thus, a global downward trend might be Multiscale growth networks with stress- and scale- expected for rock permeability at large scales. The flow dependent apertures are constructed using the recursive behaviour is also significantly affected by the distribution of variable apertures, which leads to various fluid flow 67

December 2019 16 structures and permeability scaling trends (de Dreuzy et al., 2002, 2001). Under a higher in-situ stress ratio, longer fractures play a more important role for fluid migration due to their lower resistance (Tsang and Neretnieks, 1998) in association with wider apertures that are correlated with fracture length. Hence, at smaller scales, an increased permeability occurs in the deviatoric stress case attributed to the considerable contribution from long fractures. However, a global decreasing trend is inevitable due to the decreasing proportion of traversing fractures at larger scales, where shorter fractures tend to take a heavier role in fluid flow. Fig. 7 - (a) The 6 m × 6 m outcrop pattern for measuring scaling properties, (b) the 2 m × 2 m source cell pattern including censored and uncensored fractures for geomechanical modelling and network growth, and (c) the growth lattice for extrapolating fracture networks progressively into larger domains based on (d) a recursive scheme Fig. 8 - Multiscale growth realisations achieved by the recursive cell culture scheme In the hydrostatic stress case, the Fig. 9 - Equivalent permeability kxx and kyy derived from flow simulations and analytical permeability equivalent permeability mainly kharm, karithm, kgeom calculated using the harmonic, arithmetic and geometric mean apertures declines with the increased scale, under different stress conditions because the slightly scaled apertures with no shear-induced dilation do not endow long fractures with highly conductive capability compared to the decreased relative frequency of long fractures whose length follows the power law. A comparison with the analytical solution of the equivalent permeability kharm, karithm, kgeom based on harmonic, arithmetic or geometric mean apertures, respectively, reveals that the numerically derived permeability is well bounded by the harmonic and arithmetic values, while the median trend is better tracked by the geometric one (Fig. 9). 68

NEWS J URNAL ISRM 5. CONCLUSIONS 16 To conclude, a systematic study was conducted on the Lei Q, Latham J-P, Tsang C-F, Xiang J, Lang P. 2015a. A new geometrical, geomechanical and hydrological properties approach to upscaling fracture network models while of natural fracture networks. The complexity of natural preserving geostatistical and geomechanical characteristics. fractures with respect to hierarchical topologies and Journal of Geophysical Research: Solid Earth 120, 4784–4807. underlying formation mechanisms was investigated through a study of the statistics and tectonism of a multiscale fracture Lei Q, Latham J-P, Xiang J. 2016. Implementation of an system. To simulate the complex geomechanical behaviour of empirical joint constitutive model into finite-discrete natural fractures associated with intrinsic surface roughness, element analysis of the geomechanical behaviour of a joint constitutive model was implemented into the hybrid fractured rocks. Rock Mechanics and Rock Engineering 49, finite-discrete element model. This model was applied to 4799–4816. simulate the geomechanical behaviour of various fracture networks with the consequences on fluid flow further Lei Q, Latham J-P, Xiang J, Tsang C-F. 2017b. Role of natural analysed. To estimate the hydromechanical properties of a fractures in damage evolution around tunnel excavation in natural fracture network at larger scales, a novel upscaling fractured rocks. Engineering Geology 231, 100–113. approach employing discrete-time random walks in a recursive self-referencing lattice was developed. The research Lei Q, Latham J-P, Xiang J, Tsang C-F, Lang P, Guo L. 2014. findings highlight the importance of realistic fracture network Effects of geomechanical changes on the validity of a representation and systematic geomechanical simulation for discrete fracture network representation of a realistic two- studying the hydromechanical behaviour of fractured rocks. dimensional fractured rock. International Journal of Rock Mechanics and Mining Sciences 70, 507–523. REFERENCES Bandis SC, Lumsden AC, Barton NR. 1983. Fundamentals of Lei Q, Latham J, Xiang J, Tsang C. 2015b. Polyaxial stress-induced variable aperture model for persistent 3D fracture networks. rock joint deformation. International Journal of Rock Geomechanics for Energy and the Environment 1, 34–47. Mechanics and Mining Sciences & Geomechanics Abstracts 20, 249–268. Lei Q, Wang X. 2016. Tectonic interpretation of the connectivity Barton CC. 1995. Fractal analysis of scaling and spatial of a multiscale fracture system in limestone. Geophysical clustering of fractures, in: Barton, CC, La Pointe, PR (Eds.), Research Letters 43, 1551–1558. Fractals in the Earth Sciences. Plenum Press, New York, pp. 141–178. Lei Q, Wang X, Xiang J, Latham J-P. 2017c. Polyaxial stress- Barton N, Bandis S, Bakhtar K. 1985. Strength, deformation and dependent permeability of a three-dimensional fractured conductivity coupling of rock joints. International Journal rock layer. Hydrogeology Journal 25, 2251–2262. of Rock Mechanics and Mining Sciences & Geomechanics Abstracts 22, 121–140. Munjiza A. 2004. The Combined Finite-Discrete Element Berkowitz B, Bour O, Davy P, Odling N. 2000. Scaling of fracture Method. Wiley, London. connectivity in geological formations. Geophysical Research Letters 27, 2061–2064. Petit JP, Wibberley C a J, Ruiz G. 1999. “Crack-seal”, slip: A new Bonnet E, Bour O, Odling NE, Davy P, Main I, Cowie P, Berkowitz fault valve mechanism? Journal of Structural Geology 21, B. 2001. Scaling of fracture systems in geological media. 1199–1207. Reviews of Geophysics 39, 347–383. Davy P, Bour O, de Dreuzy J-R, Darcel C. 2006. Flow in Renshaw CE. 1997. Mechanical controls on the spatial density multiscale fractal fracture networks, in: Cello, G, Malamud, of opening-mode fracture networks. Geology 25, 923. BD (Eds.), Fractal Analysis for Natural Hazards. Geological Society, London, Special Publications, London, pp. 31–45. Rutqvist J, Stephansson O. 2003. The role of hydromechanical Davy P, Le Goc R, Darcel C, Bour O, de Dreuzy J-R, Munier coupling in fractured rock engineering. Hydrogeology R. 2010. A likely universal model of fracture scaling and Journal 11, 7–40. its consequence for crustal hydromechanics. Journal of Geophysical Research 115, B10411. Tsang C-F, Neretnieks I. 1998. Flow channeling in heterogeneous de Dreuzy J-R, Davy P, Bour O. 2002. Hydraulic properties of fractured rocks. Reviews of Geophysics 36, 275–298. two-dimensional random fracture networks following power law distributions of length and aperture. Water 69 Resources Research 38, 1276. de Dreuzy J-R, Davy P, Bour O. 2001. Hydraulic properties of two-dimensional random fracture networks following a power law length distribution: 2. Permeability of networks based on lognormal distribution of apertures. Water Resources Research 37, 2079–2095. Lei Q, Latham J-P, Tsang C-F. 2017a. The use of discrete fracture networks for modelling coupled geomechanical and hydrological behaviour of fractured rocks. Computers and Geotechnics 85, 151–176.

December 2019 17 ISRM CORPORATE LECM - Civil Engineering Laboratory of Macau, Macau, China MEMBERS IN 2018 Lehrstuhl für Ingenieurgeologie und Hydrogeologie, A SUPPLIERS OF ROCK MECHANICS EQUIPMENT AND Aachen, Germany MATERIALS Cerro Vanguardia, Salta, Argentina LNEC - Laboratório Nacional de Engenharia Civil, ESS Earth Sciences, Richmong, Australia Lisbon, Portugal Fondasol, Avignon, France GCTS Testing Systems, Tempe AZ, USA National Engineering Research Center for Coal Control, Geobrugg, Romanshorn, Switzerland Huainan City, China GEOS Ingénieurs Conseils, Courbevoie, France Glötzl, Rheinstetten, Germany Nick Barton and Associates, Høvik, Norway Golder Associates, Montreal, Canada Nitro Consult Ab., Stockholm, Sweden Guangdong Hongda Blasting Engineering Co. Ltd., China Nitro Consult AS., Lierstranda, Norway Lehrstuhl für Ingenieurgeologie und Hydrogeologie, Norconsult AS, Sandvika, Norway Orica Mining Services Portugal, SA., Lisboa, Portugal . Aachen, Germany Oyo Corporation, Tokyo, Japan LNEC - Laboratório Nacional de Engenharia Civil, Pöyry SwedPower, Lulea, Sweden Lisbon, Portugal RISE Research Institute of Sweden, Borås, Sweden MTS System Corporation, Éden Prairie, USA Royal Institute of Technology - KTH, Stokholm, Sweden New Concept Mining, Johannesburg, South Africa Skanska Norge AS, Oslo, Norway Orica Mining Services Portugal, SA., Lisboa, Portugal Solexperts Ag., Mönchaltorf, Switzerland Oyo Corporation, Tokyo, Japan State Key Lab. For Geomechanics and Deep Underground Pagani Geotechnical Equipment, Calendasco, Italy Shanghai Urban Construction (Group) Corporation, Engineering, China University of Mining and Technology, Beijing, China Shanghai, China Sweco AB, Stockholm, Sweden Sibelco NordicAS Avd Stjernøy, Alta, Norway Swedish Nuclear Fuel and Waste Management, SKB AB, Solexperts Ag., Mönchaltorf, Switzerland Figeholm, Sweden B SUPPLIERS OF ROCK MECHANICS SERVICES Tractebel Engineering, Gennevilliers, France AF Grupen Norge AS, Oslo, Norway Trilab Pty Ltd, Geebung, Australia Alliance Geotechnical Pty Ltd, Seven Hills, Australia Vik Ørsta AB, Ørsta, Norway ALS Group LLC, Ulaanbaatar, Mongolia WSP Sverige AB, Stockholm-Globen, Sweden Arcadis ESG, Le Plessis Robinson, France Beck Engineering Pty Ltd, Chatswood, Australia C CONSULTANTS Bergab AB, Solna, Sweden 3D Geoscience, Inc., Tokyo, Japan Besab AB Hisings Backa, Sweden Abeis Konsult AB, Stenhamra, Sweden Bleikvassli Gruber AS, Bleikvasslia, Norway AF Grupen Norge AS, Oslo, Norway BRGM, Orleans, France AGL Consulting, Dublin, Ireland Chalmers University of Technology, Göteborg, Sweden Alliance Geotechnical Pty Ltd, Seven Hills, Australia China Three Gorges University, Yichang, China Arcadis ESG, Le Plessis Robinson, France Dalian Mechsoft Co.,Ltd, Dalian City, China Asplan Viak AS, Molde, Norway Entreprenørservice AS, Rud, Norway Bauer Spezialtiefbau Schweiz AG, Switzerland ESS Earth Sciences, Richmong, Australia Beck Engineering Pty Ltd, Chatswood, Australia F-Infrastructure, Stockholm, Sweden Bergab AB, Solna, Sweden Fondasol, Avignon, France Besab AB Hisings Backa, Sweden Geobrugg, Romanshorn, Switzerland BG Ingénieurs Conseils SA, Switzerland GEOS Ingénieurs Conseils, Courbevoie, France Bleikvassli Gruber AS, Bleikvasslia, Norway Geoscience Ltd., Falmouth, United Kingdom BRGM, Orleans, France Geosigma AB, Stockholm, Sweden Cartledge Mining & Geotechnics, Brisbane, Australia Golder Associates, Montreal, Canada CETU (Centre d’Études des Tunnels), Lyon, France Hydrochina Kunming Engineering Corporation, Chalmers University of Technology, Göteborg, Sweden Changjiang River Scientific Research Institute, Kun'ming, China Wuhan, China Ineris, Verneuil en Halatte, France China Coal Technology & Engineering Group (CCTEG), Ingenieursozietät Prof. Dr-Ing. Katzenbach Gmbh, Beijing, China Frankfurt am Main, Germany Institute of Rock and Soil Mechanics, The Chinese Academy of Science, Wuhan City, China 70

NEWS J URNAL ISRM 17 China Institute of Water Resources and Hydropower LNEC - Laboratório Nacional de Engenharia Civil, Research, Beijing, China Lisbon, Portugal Chinese Nonferrous Metal Survey and Design of Changsha, Locher Ingenieure AG, Zurich, Switzerland Changsha, China Lombardi Engineering Ltd., Minusio, Switzerland Lombardi SA, Minusio, Switzerland Chuo Kaihatsu Corporation, Tokyo, Japan LOTTE E&C, Seoul, Korea CNNC Beijing Research Institute of Uranium Geology, Multiconsult AS, Oslo, Norway National Engineering Research Center for Coal Control, Beijing, China Coffey Geotechnics, North Ryde, Australia Huainan City, China Daelim Industrial Co., Ltd, Seoul, Korea Newjec Inc., Osaka, Japan Datong Coal Mine Group, Datong City, China Nick Barton and Associates, Høvik, Norway Dia Consultants Co., Ltd., Tokyo, Japan Nishimatsu Construction CO., Ltd., Tokyo, Japan Distruct Solutions SARL, Zalka, Nabanon Nitro Consult Ab., Stockholm, Sweden Docon Corporation, Hokkaido, Japan Nitro Consult AS., Lierstranda, Norway Dongbu Corporation, Seoul, Korea Norconsult AS, Sandvika, Norway Electric Power Dev. Co., Ltd., Tokyo, Japan NTNU Inst for Geologi og Bergteknikk, Trondheim, Norway Entreprenørservice AS, Rud, Norway Pöyry SwedPower AB, Luleå, Sweden ESS Earth Sciences, Richmong, Australia Prof. Dipl.-Ing. H. Quick, Darmstadt, Germany F-Infrastructure, Stockholm, Sweden Ramböll Sverige, Stockholm, Sweden Fondasol, Avignon, France Ramboll UK Ltd., London, United Kingdom GEOS Ingénieurs Conseils, Courbevoie, France RCC-Group, Moscow, Russia Geoscience Ltd., Falmouth, United Kingdom RISE Research Institute of Sweden, Borås, Sweden Geosigma AB, Stockholm, Sweden Royal Institute of Technology - KTH, Stokholm, Sweden GMI S.A. Ingenieros Consultores, Lima, Peru Seohyun Eng, Anyang-si, Korea Golder Associates, Montreal, Canada Shanghai Urban Construction (Group) Corporation, GS E&C, Seoul, Korea Guangdong Hongda Blasting Engineering Co. Ltd., China Shanghai, China Hazama Corporation, Ibaraki, Japan SK E&C, Seoul, Korea Høgskolen i Narvik, Narvik, Norway Skanska Norge AS, Oslo, Norway Hokkaido EL.P.CO., Hokkaido, Japan Solexperts Ag., Mönchaltorf, Switzerland Hokuriku Electric Power Co., Toyama City, Japan Sumiko Resources Exploration & Development Co Ltd., Hydrochina Guiyang Engineering Corporation, Guiyang Tokyo, Japan City, Guizhou Province, China Suncoh Consultants Co., Ltd., Tokyo, Japan Hydrochina Kunming Engineering Corporation, Sweco AB, Stockholm, Sweden Sweco Norge AS, Oslo, Norway Kun'ming, China Swedish Nuclear Fuel and Waste Management Co. – SKB, Hydrogeotechnique, Fontaines, France Hyundai E&C, Yongin-si, Korea Stockholm, Sweden Implenia Suisse SA, Switzerland Swedish Rock Engineering Association, Sweden Ineris, Verneuil en Halatte, France The Institute of Crustal Dynamics, China Earthquake Ingenieursozietät Prof. Dr-Ing. Katzenbach Gmbh, Administration, Beijing, China Frankfurt am Main, Germany Three Gorges Geotechnical Consultants Co., Ltd., Itasca Consulting China Ltd., Wuhan City, China Itasca Consulting Group Inc., Minneapolis, USA Wuhan City, China Itasca Consulting Ltd, Shrewsbury, United Kingdom Tongji University, Shaghai, China ITOCHU Techno – Solutions Corporation, Tokyo, Japan Tractebel Engineering, Gennevilliers, France Japan Underground Oil Storage Group Comp., Tokyo, Japan Tractebel Engineering GmbH, Bad Vibel, Germany Ji’nan Rail Transit Group Co., LED., China WSP Sverige AB, Stockholm-Globen, Sweden Kajima Co., Tokyo,Japan Yachiyo Engineering Co., Ltd., Tokyo, Japan Kawasaki Geological Engng. Co., Tokyo, Japan KDC Engineering, Tokyo, Japan D CONTRACTORS Kiso-Jiban Consultants Co., Ltd., Tokyo, Japan Changjiang River Scientific Research Institute, Wuhan, LCW Consult , S.A., Algés, Portugal China Lehrstuhl für Ingenieurgeologie und Hydrogeologie, China Coal Technology & Engineering Group (CCTEG), Beijing, China Aachen, Germany 71

December 2019 17 China Three Gorges Project Corporation, Yichang, China G RESEARCH ORGANIZATIONS Chinese Nonferrous Metal Survey and Design of Changsha, Bergab AB, Solna, Sweden Changsha, China Besab AB Hisings Backa, Sweden Docon Corporation, Hokkaido, Japan Central Research Institute of Electric Power Industry, ESS Earth Sciences, Richmong, Australia Chiba, Japan Geoscience Ltd., Falmouth, United Kingdom CETU (Centre d’Études des Tunnels), Lyon, France Jan De Nul N.V., Hofstade-Aalst, Belgium Chalmers University of Technology, Göteborg, Sweden Japan Underground Oil Storage Comp., Tokyo, Japan F-Infrastructure, Stockholm, Sweden Kajima Co., Tokyo, Japan Fachhochschule Nordwestschweiz, Switzerland Lombardi SA, Minusio, Switzerland Geoscience Research Laboratory, Co. Ltd, Yamato, Japan Nishimatsu Construction CO., Ltd., Tokyo, Japan Geosigma AB, Stockholm, Sweden Obayashi Corporation, Tokyo, Japan Høgskolen i Narvik, Narvik, Norway Shanghai Urban Construction (Group) Corporation, Kajima Co., Tokyo,Japan Shanghai, China Korea Expressway Corporation Research Institute (KECRI), Shimizu Corporation, Tokyo, Japan Hwaseong-si, Korea Skanska Norge AS, Oslo, Norway LECM - Civil Engineering Laboratory of Macau, Solexperts Ag., Mönchaltorf, Switzerland Macau, China Sweco Norge AS, Oslo, Norway Lehrstuhl für Ingenieurgeologie und Hydrogeologie, Taisei Corporation, Tokyo, Japan Aachen, Germany Tekken Corporation, Tokyo, Japan LNEC - Laboratório Nacional de Engenharia Civil, The Institute of Crustal Dynamics, China Earthquake Lisbon, Portugal Administration, Beijing, China Nitro Consult Ab., Stockholm, Sweden Tobishima Corp., Chiba, Japan Norges Geologiske Undersøkelse, Trondheim, Norway Toda Corporation, Tokyo, Japan Norges Geotekniske Institutt, Oslo, Norway Tongji University, Shaghai, China Norsk Forening for Fjellsprengn Teknikk, Oslo, Norway Vik Ørsta AS, Ørsta, Norway Norsk Geoteknisk Forening, Oslo, Norway NTNU Inst for Geologi og Bergteknikk, Trondheim, Norway E ELECTRICITY SUPPLY COMPANIES Pöyry SwedPower AB, Luleå, Sweden China Three Gorges Project Corporation, Yichang, China RCC-Group, Moscow, Russia Chugoku Electric Power Co., Inc., Hiroshima, Japan RISE Research Institute of Sweden, Borås, Sweden Electric Power Dev. Co., Ltd., Tokyo, Japan Royal Institute of Technology - KTH, Stokholm, Sweden Hokkaido Electric Power Co. Inc., Hokkaido, Japan Saint-Petersburg Mining University, St. Petersburg, Russia Hokuriku Electric Power Co., Toyama City, Japan Shaoxing University, Shaoxing, China Hydrogeotechnique, Fontaines, France SINTEF Teknologi og samfunn, Trondhein, Norway Kyushu Electric Power Company, Fukuoka, Japan Solexperts Ag., Mönchaltorf, Switzerland Shikoku Electric Power Co., Kagawa, Japan Sweco AB, Stockholm, Sweden Swedish Nuclear Fuel and Waste Management Co. – SKB, F MINING COMPANIES Stockholm, Sweden Datong Coal Mine Group, Datong City, China WSP Sverige AB, Stockholm-Globen, Sweden Mireco - Mine Reclamation Corp., Wonju-si, Korea Nittetsu Mining Company, Ltd., Tokyo, Japan H GOVERNMENT DEPARTMENTS Orica Mining Services Portugal, SA., Lisboa, Portugal CETU (Centre d’Études des Tunnels), Lyon, France Polymetal, St. Petersburg, Russia LECM - Civil Engineering Laboratory of Macau, RСС-Group, Ekaterinburg, Russia Macau, China Shanghai Urban Construction (Group) Corporation, LKAB - Luossavaara-Kiirunavaara AB, Sweden Shanghai, China LNEC - Laboratório Nacional de Engenharia Civil, Somincor - Sociedade Mineira de Neves Corvo, S.A., Lisbon, Portugal Castro Verde, Portugal Okumura Corporation, Ibaraki, Japan I OTHER CORPORATE MEMBERS Itasca Consulting China Ltd., Wuhan City, China ITOCHU Techno – Solutions Corporation, Tokyo, Japan Kumagai Gumi Co., Ltd., Tokyo, Japan Okumura Corporation, Ibaraki, Japan Orica Mining Services Portugal, SA., Lisboa, Portugal WSP Finland Oy, Helsinki, Finland 72

HOW TO JOIN THE ISRM Individual members: Individuals shall normally apply for membership of the ISRM through their National Group. Corresponding members: As some countries do not have a National Group, individuals may also apply for ISRM membership directly to the Secretariat. Corporate members: Companies shall apply for membership of the ISRM as corporate members directly to the Secretariat. In some cases companies have also joined the ISRM as corporate members through their National Group. National Groups: For a national organisation to be recognised as a National Group, it is required to formally apply to the President through the Secretary-General for recognition according to the ISRM statutes. This shall be an organisation such as a society or a committee that represents Rock Mechanics in that country, either solely concerned with Rock Mechanics, or as part of a broader field of scientific or engineering interest. Each country shall have no more than one National Group. BENEFITS TO MEMBERS Current benefits given to ISRM Individual and corresponding members are: • ISRM Newsletter • Access to the members area in the website (download of Suggested Methods and Reports, Rock Mechanics lectures, videos, slide collection, etc.) • Right to participate in the ISRM Commissions • Registration with a 20% discount in the ISRM Congress, International and Regional Symposia and Specialized Conferences • Personal subscription to the International Journal of Rock Mechanics and Mining Sciences at a discounted price (see details). • Personal subscription to the Journal Rock Mechanics and Rock Engineering at a discounted price. • Free download of up to 100 papers per year from the ISRM Digital Library at OnePetro: www. onepetro.org • Discount on purchases of CRC Press books: 40% discount on books in the ISRM Book Series; 30% discount on all other CRC Press books. Current benefits given to ISRM corporate members are: • Being listed in the ISRM website, with a link to the company’s website • Being listed in the ISRM News Journal • Access to the members area in the ISRM website • ISRM Newsletter • 1 copy of the ISRM News Journal • 1 registration with a 20% discount as ISRM member in the ISRM Congress, International and Regional Symposia and Specialized Conferences • Free download of up to 250 papers per year from the ISRM Digital Library at OnePetro: www.onepetro.org.

ISRM International Society for Rock Mechanics and Rock Engineering ISRM Board Itaipu dam, photo by Itaipu Binacional President Prof. ReŞat Ulusay Vice-President for Africa Dr Michael du Plessis Vice-President for Asia Dr Suseno Kramadibrata Vice-President for Australasia Dr Sevda Dehkhoda Vice-President for Europe Prof. Leandro R. Alejano Vice-President for Latin America Prof. José Pavón Mendoza Vice-President for North America Prof. Laura J. Pyrak-Nolte Vice-President at Large Prof. Yang Qiang Vice-President at Large Prof. Ömer Aydan Vice-President at Large Dr Vojkan Jovičić Secretary-General Dr Luís Lamas ISRM Secretariat LNEC | Av. Brasil, 101 | 1700-066 Lisboa | PORTUGAL | Tel.: (+351) 218443419 | E-mail: [email protected]