Nick Bartonse a Back cover ilusration: Cutter change under a dangerous working environment (Shen et al. 1999). Authorization to photocopy items for intemal or personal us, or the internal or personal use of spe- cific clint, is granted by A.A.Balkema, Rotterdam, provided that the base fee of USSI.50 per copy, plus USS0.10 per page is paid directly to Copyright Clearance Center, 222 Rosewood Drive Danvers, MA 01923, USA. For those organizations that have been granted a photocopy license by CCC, a separate system of payment has been arranged. The fee code for users of the Transactionat Reporting Service is: 90 $809 341 7/00 US$1.50 + USSO.10. \ ‘Second print Published by A.A Balkema, P.0.Box 1675, 3000 BR Rotterdam, Netherlands Fax; +31.10.4135947; E-mail: [email protected]; Intemet site: htp://www-balkema.n! } AA.A.Balkema Publishers, Old Post Road, Brookfield, VT 05036-9704, USA. Pax: 802.276.3837; E-mail: [email protected] ISBN 90 5809 341 7 (© 2000 A.A.Balkema, Rotterdam | Printed in the Netherlandsal 1999). use of spe. $51.50 per ‘ood Drive, license by ansactional Contents PREFACE, ACKNOWLEDGEMENTS PART I BASIC INTERACTIONS BETWEEN THE ROCK MASS AND THE TBM 1. Introduction 2. Some basic TBM designs 3. Summary of common geotechnical problems 4, The TBM excavation disturbed zone 5. Basic factors affecting penetration rate 6. Penetration rate and thrust per cutter 7. The possible influence of stress and strength ratios 8. Basic mechanism of chip formation with roller cutters 9. Tensile strength and its anisotropy (0. Penetration rate and fabric anisotropy 1. Penetration rate, joint spacing and joint character PART 2 Q, Oram, AND ROCK MASS VARIABILITY 12. TBM performance and rock mass classification 13. TBM performance and Q-system parameters 14, TBM performance and the initial requirements for ‘Qraw’ 15, The law of decelerating advance rates 16. Utilisation and its decay with time 17. Unexpected events and their Q-values 18, Water inflows in TBM driven tunnels 19. Consequences of limited stand-up time in TBM tunnels 20. The relationship between PR, AR and Ore v vu 39 45 51 56 37 63 65 70vi 21 22. 23, 24, 25. 26. 21. 28. Contents Rock masé variability and its effect on predicted performance Fine-tuning Qrau for anisotropy Effect of porosity and quartz content on gradient m and PR Tunnel size effects Boring in exceptionally tough, high-stress con« Revisiting cutter force effects Predicting advance rates in faulted rock PART 3 LOGGING, TUNNEL SUPPORT, PROBING AND DESIGN VERIFICATION 29. 30. 31 32. 33. 34, 35. 36. 37. ‘TBM Q-logging and tunnel scale effects Rock support methods commonly used in TBM tunnels Some support design details for TBM tunnels Probing and convergence measurement Probing and seismic or sonic logging Verifying TBM support classes with numerical models Logging rock quality and support needs ‘TBM or drill-and-blast excavation? Conclusions APPENDIX AL AQ A3 Ad ‘Q-method of rock mass classification ray ~ the final version of Figure 44 Input data summary for estimating PR and AR using Orem Worked example REFERENCES INDEX 14 1 81 86 88 1 95 99 107 114 118 123 129 133 138 142 147 151 151 154 154 156 161 1694 7 81 86 88 91 95 ATION 107 114 118 123, 129 133 138 142 147 151 151 154 154 156 161 169 Preface During a long intemational involvement with drill-and-blast excavations — en- gagements that started with the development of the Q-sysiem in the early seven- ties ~ the writer has also been drawn into smooth bored TBM tunnels — at first with some reluctance it must be admitted. How could one classify the jointing properly when there was almost no overbreak? Why did low quality rock masses — where there was too much overbreak — cause such problems for the machines? The writer’s engagements with TBM projects has generally focussed on con- flicts between Owner and Contractor, sometimes Contractor and sub-Contractor. Inevitably, these conflicts have related to the geology and hydrogeology — and their characteristic lack of concer for the TBM and its operator's wellbeing, not to mention that of all other parties involved. ‘The two geotechnical difficulties referred to above are specific to TBM tunnels. Lack of overbreak can easily mask the true rock mass quality and is actually a source of risk. When tackling ground that is of obvious bad quality, it appeared to the writer that the canopy or shield were actually hindering efficient pre-treat- ‘ment ~ an opinion there is much support for in the literature these days. However, and this is a very large however, the fact that TBM might advance ‘more than 150 m per day, 500 m per week, 2 km per month or even 15 km in a year guarantees the admiration ~ almost incredulity — of any tunnelling engineer, and forces our acceptance of some TBM limitations. This book is an attempt to quantify and understand both the records and the limitations in poor ground, whether from hard, abrasive and massive rock, or from faulted regions with erosion of fines and chimney formation. The tunnelling pro: fession needs to explain the records and the occasional failures, and all the vari- able ground in between, which has most impact on prognoses. vilAcknowledgements ‘The opportunity to work professionally in widely varying TBM tunnelling condi- tions with contractors, owners and consultants has been an essential ingredient in the development of the Qrew method. For many reasons, including these widely different tunnelling conditions, the writer would like to acknowledge EuroTunnel, UK Nirex and GeoEngineering in the UK, Statkraft in Norway and Kashmir, Kraftbyggama in Sweden, NOCON in Norway and Italy and Fuji RIC in Japan., not forgetting nature herself, and all the devious ways in which geology, hydrogeology and tunnel depth interact. Discussions with NGI colleagues Fredrik Laset and Eystein Grimstad and di- rect application of some of their richly varied tunnelling experience, is also sin- cerely acknowledged. The committed and professional expertise of Pat Coughlin and Marcelo Medina Abrahdo in production of the text and graphics will facilitate the trouble- free application of Orey for which I thank them sincerely. Finally, I dedicate this book to Eda Quadros, who was instrumental in creating both the necessary atmosphere and opportunity to concentrate on only ane project at a time, after many years of varied but enjoyable professional work at NGI. Both Professor Lineu Ayres da Silva and Professor Carlos Maffei of the Univer- sity of Sao Paulo, played an indirect but generous part in this project, including the loan of an office without telephone or e-mail, for which I am thankful.Part 1. Basic interactions between the rock mass and the TBM— — CHAPTER | Introduction The frontispiece shows a photograph from the world’s first bored tunnel. It is also a good example of the influence of jointing on overbreak. The exploratory Beau- mont Tunnel was driven in chalk marl in 1881-1882. The classic wedge-shaped fall-out caused by three joint sets intersecting the tunnel wall was also a source of major difficulties when the Channel Tunnel was driven by much larger machines some 110 years later. The Beaumont TBM of 2.1 m diameter achieved weekly advance rates of between 30 and 60 m (Varley & Warren 1996), a highly respect- able effort but later overshadowed by much higher performances and a best week of 426 m, by the time the 8.7 m diameter twin-tube Channel tunnels were com- pleted in 1990-1991 (Warren et al. 1996). 33'thy The fact that penetration rate (PR) can reach extremes of 10 mvhr in certain TBM projects, while advance rates (AR) as low as 0.005 nvhr (or even zero) are experienced on occasion, suggests the need for predictive models and a good un- derstanding of the important variables. Perhaps in no other rock engineering ac- tivity is the need for rock classification so important. The fraction of total construction time (U) that the TBM can be utilised for boring, where: AR=UxPR a is strongly related to rock conditions, but also to many other factors. Nelson (1993) suggested that penetration rate (PR) and utilisation (U) could be separated and each related to a rock mass classification. This book is an attempt to do exactly this, but with the difference that PR and U are partly related to the same basic classification system (the Q-system) with necessary ‘modification’ of RQD (it must be oriented in the tunnelling direction). A new stress-strength term is used to capture some extra features of joint aniso- tropy and its interaction with machine capacity (cutter thrust F). Anew term Ora is formulated in stages, starting from the Q-value (since rock mass conditions are so important) and finishing with rock-machine and rock ‘mass-machine interaction parameters. Qray is designed to allow PR to be esti- mated, or Qrsq can be back-calculated from actual performance, when tunnelling begins. Advance rate AR is estimated using Equation 1, but with the important difference that U is recast as a time-dependent and rock quality dependent vari- able. 3