REVIEW ON TBM PERFORMANCE IN SQUEEZING GROUND CONDITIONS

Rafael Rojas, Jian Zhao

EPFL-ENAC-ICARE-LMR, Lausanne, 1015, Switzerland

Keywords: Squeezing,ground,TBM

INTRODUCTION TBM excavation represents a big investment in an inflexible but potentially very fast method of excavating and supporting a rock tunnel. However, when unfavourable conditions are encountered without warning, time schedule and practical consequences are often far greater in a TBM driven tunnel than in a tunnel excavated by means of traditional techniques. Such is the case of excavation through squeezing ground conditions, which might cause difficulties whenever tunnel convergences occur with considerable magnitude at a short distance from the face within short periods of time, having a serious impact on advancement rates and leading to TBM jamming as a worst scenario. MAIN SUBJECT MATTER Squeezing ground definition It is possible to find common arguments amongst the different definitions of squeezing rock behaviour that have been collected from different authors (Barla,1995), (Terzaghi,1946), (Deere,1981), (Jethwa,1981-1996), (Panet,1996), (Kovari,1988), (Einstein,1989), (Aydan,1993). The main aspects can be shortlisted as follows: i. Squeezing response of the rock to excavation implies large ground deformation around a tunnel that can take place during and well after its excavation. ii. The deformation is first of all produced by the disturbance of the primitive stress field as a consequence of the excavation of the tunnel. iii. The real particularity of squeezing phenomena is the aspect of the time dependency of the rock mass behaviour. This time dependency is explained in the literature by two different factors: a) Creep caused by exceeding a limiting shear stress Creep in the particles of the intact material (viscous behaviour or unstable crack propagation) Creep along the interfaces between particles, due to a complete shear failure around the tunnel Creep along larger scale discontinuities such as bedding and foliation surfaces, joints and faults. These creep mechanisms involve the three components (primary, secondary and tertiary) and typical combinations thereof. Usually, the creep mechanisms underlying squeezing is of visco-plastic nature but, particularly at low stresses, some of the strains may be recoverable i.e. visco-elastic behaviour occurs.

Figure 1 Different components of creep in rocks (Dusseault & Fordham, 1993)

b) Consolidation dissipation of pore water pressure in low permeability rock masses In the vicinity of the working face, creep and consolidation are in general superimposed on the spatial stress redistribution. The mechanism of consolidation is relevant for tunnelling through water-bearing, low permeability ground. iv. Squeezing might be accompanied but not mistaken with swelling phenomena v. It can occur without volume change, however, it might be associated with volume increase in dilatant materials Mechanisms that explain squeezing ground Different explanations of mechanisms that can trigger squeezing ground behaviour can be found in literature: i) Complete shear failure This mechanism can be observed in rock masses of low strength and high deformability, as long as the particular combination of induced stresses and material properties pushes some zones around the tunnel beyond the limiting shear stresses at which creep starts (Einstein,1989), (Peck et al., 1972), (Jethwa et al.,1984), (Gioda & Swoboda,1999), (Barla,1995). This mechanism often leads to very large convergences that occur in the short term, while the TBM is still excavating, and hence becomes the most critical for TBM operation. ii) Buckling failure It is due to the flexural tensile buckling of the inter-bedded formation, generally in metamorphic rocks (phyllite, mica-schists) or thinly bedded ductile sedimentary rocks (mudstone, shale, siltstone, sandstone, evaporitic rocks). (Aydan et al.,1996). iii) Shearing and sliding failure Generally observed in relatively thickly bedded sedimentary rocks, it involves sliding along bedding planes and shearing of intact rock (Aydan et al.,1996), (Hsu S.C.,2004). iv) Squeezing due to stress relaxation in foliated rock Due to creep and relaxation of metamorphic rocks containing phyllosilicates, as in the long term only small or vanishing shear stress can be sustained in the foliation planes (Kolymbas,2006). v) Squeezing due to consolidation processes Tunnelling through water bearing and low permeability ground could lead to time dependency due to consolidation effects. vi) Time dependent microcraking Brittle and quasi-brittle hard rocks can display squeezing behaviour due to time-dependent microcracking, triggered by a stress corrosion process (Malan,1999-2002), (Boukharov
2

and Chanda,1995), (Gioda and Cividini,1996), (Barla,2000), (Shao et al.,2003), (Shao and Chau,2005). TBM excavation through squeezing ground conditions: main problems The hazards associated with squeezing ground concern both the TBM machine and the back-up area. Due to the fixed geometry and the limited flexibility of the TBM the room to be allowed for ground deformations is restricted. Convergences which exceed 5 % of the tunnel radius are to be considered problematical (Kovri,1996). The consequences of squeezing can range from large tunnel closures and high pressures exerted by the rock mass on the shield of the TBM to more extreme conditions, when the friction produced by the ground in contact with the machine cannot be counteracted by the available thrust and the TBM becomes jammed (Steiner,1996). Therefore, in severe squeezing conditions when face extrusion may become important, problems might also be experienced at the cutting face. As the TBM types are different with respect to the thrusting system, the type of support and the existence or not of a shield, different hazard scenarios have to be considered, that will depend on the machine type (Ramoni and Anagnostou, 2006): i) Gripper TBMs: These TBMs are today generally equipped with a short shield (canopy, cutter head shield). Depending on the rheological behaviour of the ground, a high radial ground pressure acting upon the cutter head or the canopy as well as an extremely high extrusion rate of the core can also immobilise the machine. Normally, the excavation speed is high enough to avoid such problems. The short length of the shield has a positive influence. If the TBM is moving the risk of a shield jamming is lower (deformations occur mostly only after the passage of the machine). Maintaining a high advance rate may nevertheless, be difficult in poor ground, because support installation needs more time and squeezing may also reduce the performance of the back-up system (e.g. re-profiling works, differential heave or twisting of the tracks). In the extreme case of a standstill, the TBM can be freed if the installed thrust force and torque are high enough and the ground can provide a sufficient reaction to the gripper forces. ii) Single shielded TBMs: These TBMs are longer than gripper machines. The bigger length increases the risk of becoming trapped in squeezing ground. On the other hand, single shielded TBMs have the advantage of a higher advance rate in poor ground, although feedback effects are possible for these machines too. For example, high water inflows or unstable tunnel walls may make installation or backfilling of the lining difficult and therefore slowdown advance. A single shielded TBM is jacked against the segmental lining. The possible thrust force and torque depend not only on the design of the machine (installed thrust force and torque) but also on the structural design of the segmental lining and the quality of annulus grouting. The lining has to be designed for the combined action of ground pressure and maximum jacking forces, in order to avoid overstressing or inadmissible ovalisation. iii) Double shielded TBMs: These machines are longer, particularly in small diameter tunnels. In weak ground prone to squeezing, the bracing by the gripper may be impossible and, furthermore, additional problems may occur with the extension and compression of the telescopic joint. The machine is then operated in single shield mode with jacking against the segmental lining. In this case, the same remarks as for the single shielded TBM apply, but due to its longer shield, friction forces and risks of TBM jamming will be larger. The design of the TBM also plays an important role. The improving TBM technology allows the installation of higher thrust force and torque and the reduction of the machine length also for double shielded TBMs (easier to realize for bigger diameters). The shield can also be slightly "conical" (23 cm in radius). The friction between shield skin and ground can be reduced (up to 50 %) by
3

lubricants such as bentonite. A moderate amount of squeezing can be accommodated by using extendable gauge cutters when such ground is encountered. This solution allows an increasing of the boring diameter up to 30 cm and can be easily handled by gripper TBMs; for shielded TBMs, lifting of the centerline of the shield is necessary.

Figure 2 Solution for increasing the excavation diameter by radial overcut (Voerckel,2001)

Another example of TBM adaptation for squeezing ground conditions is the case of parallel blade shield TBMs, machines where the outer shield consists of a number of blades, each one of which is supported on hydraulic rams so that the blade can move independently in both axial and radial directions. The machine advances by a shuffle-shoe process that is capable of accommodating squeezing (Hoek,2000). Need of a reliable approach for the assessment of TBM excavation in squeezing ground Particularly due to the inflexibility of TBMs, it is important to adopt technological measures on the TBM or the excavation procedure to cope with the problems expected. However, a reliable method to assess the squeezing behaviour and its impact on TBM excavation is paramount. Nowadays, there is still need for such a method that might be able to provide to the tunnel constructor, TBM operator and TBM designer an estimation of the following aspects: i. Quantification of squeezing rates of the ground around the TBM ii. Quantification of squeezing pressures that build up against the TBM shield when the rock gets in contact with the machine iii. Quantification of risks of TBM jamming Assessment of technological measures to cope with squeezing conditions iv. a) Need of fitting out the TBM with skin lubrication systems around the shield b) Need of fitting out the TBM with over-coring systems in the cutter-head c) Need of increasing the installed thrust of the TBM d) Optimum length of the TBM shield v. Assessment of construction procedures to cope with squeezing conditions a) Minimum advancement rates to be maintained b) Maximum allotted stand-still durations to avoid risks of jamming

This assessment method should take into account the different factors that play a role in the display of squeezing behaviour and in the process of TBM excavation through such conditions. These factors, according to existing literature, are multiple: i) Factors related to the nature of the ground: Strength and deformability of the rock mass Geology and lithography Weathering Content of certain minerals like mica, chlorite, serpentine and clay Permeability (when consolidation might take place) Presence of groundwater, pore pressure distribution and piezometric head
4

Macroscopic heterogeneities along the longitudinal profile Overburden in situ stresses Orientation of schistosities Orientation, disposition and thickness of discontinuities and bedding planes Thickness of squeezing prone fault zones (silo effect)

ii) Factors related to the TBM and construction procedure: Characteristics of the TBM used for excavation: Type of TBM, length of shield, available thrust. Stiffness of the backfill used to fill the gap between rock and segmental lining Gap between the shield skin and the tunnel face (overcutting) Skin friction between the shield skin and the rock mass Advancement rates of excavation Operation Standstills Distance from the face to the section where the segmented rings are installed As it can be understood, the difficulty to obtain a reliable method of assessing the performance of TBMs in squeezing ground is considerable, due to the large number of factors involved, different mechanisms that can trigger squeezing behaviour and the complications to model time dependency in the ground. This method, however, might be built on existing approaches that attempt to model the behaviour of squeezing ground in general tunnelling activities. EXISTING APPROACHES FOR THE ASSESSMENT OF SQUEEZING GROUND BEHAVIOR Different approaches exist nowadays to assess the squeezing potential, however, there are still limitations and their applicability to the case of TBM excavation is not straightforward. The different approaches that can be found in the literature to account for squeezing behaviour are: Qualitative Empirical methods: These empirical approaches are essentially based on i) classification schemes, supported by the study of different case histories of tunnelling through squeezing ground (Singh et al. 1992, Goel et al. 1995) that enable to give demarcation lines to differentiate squeezing cases from non-squeezing cases by matching data on rock mass geomechanical classification and overburden. The time variable is not taken into account and the results are only qualitative. ii) Semi-Empirical methods: These methods offer indicators for predicting squeezing, but also providing some tools for the estimation of the expected deformation around the tunnel and/or the support pressure by using closed form analytical solutions for a circular tunnel in a hydrostatic stress field. The common starting point of all these methods for quantifying the squeezing potential of rock is the use of the competency factor, which is defined as the ratio of uniaxial compressive strength of rock/rock mass to overburden stress (Aydan et al., 1993), (Hoek and Marinos,2000). No time variable is taken into account in these methods. iii) Analytical Elasto-Plastic models - closed form solutions: These are methods for analysis of the onset of yielding within the rock mass. This is determined by the shear strength parameters relative to the induced stress, as well as the extent of the total yielded or plastified zone around the tunnel, deformations around the tunnel and stress exerted on the tunnel support. These models do not consider the time-dependency of the squeezing phenomena, the tunnel is assumed to be circular and the rock mass elasto-plastic and isotropic, subjected to a hydrostatic in situ state of stress.

iv)

v)

vi)

Rheological models: In order to describe rock creep and take into account the time dependency of the rock mass behaviour, many rheological models have been developed: Visco elastic: Specially used is the Burgers model, derived from combinations of the Kelvin and Maxwell models, which can describe the elastic strain, the primary creep and the secondary creep. No plasticity is taken into account. These models present limitations, but according to some authors (Malan 2002, Gioda and Cividini 1996) this does not detract from its usefulness as a concept in creep analysis as long as its limitations are clearly understood. Visco-elastic-plastic: A typical example is the Burger-MC model, whose constitutive laws are characterized by an elastoplastic volumetric behaviour and a visco-elasticplastic deviatoric behaviour. All viscous deformations are supposed to be elastic, and so, recoverable. An example is the CVISC model, adopted by Itasca and integrated within the Flac and Flac3D finite difference codes. Elasto-visco-plastic: Elasto-viscoplasticity is essentially a modification of classical plasticity theory by the introduction of a time-rate rule in which the yield function and plastic potential function of classical plasticity are incorporated. These models can account for the primary and secondary creep, but still fail to model the tertiary creep. An example is the VIPLA model, adapted for rocks by Pellet (2005). Elasto-visco-plastic-damaged: The viscoplastic-damaged models are first proposed to account for the tertiary creep phase phenomenon, where the classic viscoplastic models failed to do so. To account for the tertiary creep, a damage module is incorporated. For this, a suitable law governing a decrease of the strength and viscoplastic parameters with the increasing strains is provided (Gioda and Cividini,2007), (Malan,2002). Models based on physical mechanisms (relaxation along foliation planes, material degradation and damage): As opposed to rheological models, which could be considered as bare mathematical frameworks for the modelling of creep deformations, these models attempt to explain the rheological behaviour of the rock by taking into account physical mechanisms related to these deformations. Some examples are given by some authors. Kolymbas (2006) proposes a model to account for rheological behaviour of metamorphic foliated rocks containing phyllosilicates, by taking into account an expression of shear stress decay for the schistosity planes that will depend on two viscosity parameters. Shao et al. (2003 and 2005), Challamel et al. (2005) propose continuum models based on damage mechanics to account for time dependency in brittle and quasibrittle rocks. The time dependent deformation is considered as a macroscopic consequence of progressive degradation of the material structure at a microscopic scale. Napier and Malan (1997) propose a discontinuum viscoplastic formulation to relate the rate of slip on a crack to the shear stress acting on it, and by taking into account the time dependent interaction of cracks. Potyondy (2006) presents a numerical model for rock that extends the formulation of the bonded-particle model (BPM) to include time-dependent behaviour by adding a damage-rate law to the formulation of the parallel bonds between the particles. Quantitative empirical models: The main drawbacks of the above mentioned models are the large number of parameters needed and their limitations to overcome the scale effect normally present in tunnels (Sulem,1987). Empirical creep models, however, are derived directly from the observed relationship of time, stress, and strain or strain rate, by using directly the tunnel monitoring results, which represent the real behaviour of the rock mass at the scale of the tunnel.

vii)

viii)

The empirical models are usually expressed in simple mathematical forms with a small number of parameters. The commonly used empirical creep models are the power law (Obert,1965), (Phienwej,2007), the exponential law (Singh and Mitchell,1968), (Semple,1973), and the hyperbolic law (Mesri et al.,1981), (Phienwej,2007). These empirical laws have been used in developing solutions for prediction of tunnel closure and ground pressure on supports by numerous authors (Aiyer,1969), (Semple,1973), (Hanafy and Emery,1979), (Phienwej,1987), (Sulem et al.,1987), (Schubert et al.,2003). Convergence empirical laws (Sulem,1987): Instead of using convergence measurements to check a constitutive behaviour law and to determine the ground parameters, Sulem (1987) proposes to analyze convergence measurements themselves, to determine a law for the tunnel closure and to propose a quantitative method to predict the final value of wall displacements. In the application of this method, it is essential that the time dependent face advance which appears as a delayed loading is recognized and separated from the time-dependent rheological behaviour of the rock mass. Models for the consolidation component of squeezing behavior: Different authors have tackled the consolidation process taking place around the tunnel by using the classical consolidation theory of Terzaghi. Even if the number of approaches to model squeezing behaviour is very large, all present limitations when it comes to assessing the performance of TBMs in squeezing ground. Further research is needed to produce a model that would take into account the particularities of TBM excavation, the time dependency due to creep and the scale effect of the rock mass at the tunnel level. This model should be able to produce good quantitative estimations of the risk of TBM jamming at the tunnel scale and should ideally be based on a reduced number of common or easy to obtain parameters.

CONCLUSIONS Squeezing ground conditions might cause difficulties in TBM excavation, having a serious impact on advancement rates and leading to TBM jamming as a worst scenario. This is the reason why a good understanding of the different mechanisms that might lead to squeezing ground behaviour as well as of correct methods of qualitatively and quantitatively assessing its magnitude are of paramount importance. This understanding might, therefore, enable us to adopt technological measures on the TBM or the excavation procedure in order to cope with the problems expected. However, nowadays the current state of the art is still not enough to solve the problem satisfactorily. Further research is needed to provide a method that takes into account the particularities of TBM excavation, the time dependency due to creep and the scale effect of the rock mass at the tunnel level.

REFERENCES
Aiyer, A. K. (1969), An analytical study of the time-dependent behavior of underground opening. Ph.D. thesis, Univ. of Illinois at Urbana-Champaign, Urbana, Ill. Anagnostou G. and Kovari K. (2005), Tunnelling through Geological Fault Zones, Institute for Geotechnical Engineering, ETH Zurich Aydan . And al. (1993), The Squeezing Potential of Rocks around Tunnels; Theory and Prediction, Rock Mechanics and Rock Engineering (1993) 26 (2), 137 163 Aydan . And al. (1996), The Squeezing Potential of Rocks around Tunnels; Theory and Prediction with examples taken from Japan, Rock Mechanics and Rock Engineering 29 (3), pp. 125 143 Barla G. (1995), Squeezing Rocks in Tunnels, ISRM News Journal, 3/4, pp. 44-49 Boukharov G.N. and al (1995), The three Processes of Brittle Crystalline Rock Creep, Rock Mechanics and Mining Sciences, Vol. 32, No 4, pp 325-335 Chalamel N., Lanos C., Casandjian Ch. (2005), Creep damage modeling for quasi-brittle materials, European Journal of Mechanics A/Solids, 24 (2005) pp. 593-613 Deere D. (1981), Adverse Geology and TBM Tunnelling Problems, Proc. RETS, Society of Mining Engineers 1, pp. 574586 Einstein H.H. (1989), Design and analysis of underground structures in swelling and squeezing rocks, Underground Structures Design and Instrumentation Developments in Geotechnical Engineering (1989), pp. 202262. Gioda G. and Cividini A. (1996), Numerical Methods for the analysis of Tunnel Performance in Squeezing Rocks Rock Mechanics and Rock Engineering, 29 (4) 171-193 Gioda G. and Swoboda G. (1999), Developments and applications of the numerical analysis of tunnels in continuous media, International journal for numerical and analytical methods in geomechanics, 23, 1393-1405. Gioda. G and Sterpi D. (2007), Visco-plastic Behaviour around advancing Tunnels in Squeezing Rock, Rock Mechanics and Rock Engineering, DOI 10.1007/s00603-007-0137-8 Goel, R.K., Jethwa, J.L., Paithakan, A.G. (1995), Tunnelling through the young Himalayasa case history of the Maneri-Uttarkashi power tunnel Journal of Engineering Geology, 39, 31-44. Guan Z. Jiang Y., Tanabash Y.and Huang H. (2007), A new rheological model and its application in mountain tunnelling, Tunnelling and Underground Space Technology, 23 (2008) 292299 Hoek, E., Marinos, P. 2000, Predicting tunnel squeezing problems in weak heterogeneous rock masses Tunnels & Tunnelling International, 45-51:part one, 33-36: part two. Hsu S.C. (2004), Failure Mechanisms of Tunnels in weak Rock with inter-bedded Structures International Journal of Rock Mechanics and Mining Sciences, 41, Pages 670-675 Jethwa J.L. (1981), Evaluation of Rock Pressures in Tunnels through Squeezing Ground in the lower Himalayas, PhD thesis, Department of Civil engineering, IIT Roorkee, India, 272 Jethwa J.L. and Dhar B.B. (1996), Tunnelling under Squeezing Ground Condition Proc. Recent Advances in Tunnelling Technology, New Delhi, 209 214 Jethwa J.L., Singh B and Singh B. (1984), Estimation of ultimate rock pressure for tunnel linings under squeezing rock conditionsa new approach, Design and Performance of Underground Excavations, ISRM Symposium, Cambridge, E.T. Brown and J.A.Hudson eds., pp. 231-238. Kolymbas D., Fellin W., Kirsch A. (2006), Squeezing due to Stress Relaxation in foliated Rock, International journal for numerical and analytical methods in geomechanics, 2006, 30, 13, pp.1357-1367 Kovari K., Amstad C. and Anagnostou G. (1988), Design/construction Methods Tunnelling in Swelling rocks, Key Questions in Rock Mechanics, Balkema, Rottedam (1988), pp. 1732.

Kovari K. (1996), Basic considerations on Tunnelling in Squeezing Rock, International Journal of Rock Mechanics and Mining Science 29, pp. 203-210. Malan, D.F. (1999), Time-Dependent behaviour of Deep Level Tabular Excavations in Hard Rock, Rock Mechanics and Rock Engineering (1999) 32 (2), 123-155 Malan,D. F. (2002), Simulating the time-dependent Behaviour of Excavations in Hard Rock, Rock Mechanics and Rock Engineeering (2002) 35 (4), 225-254 Mesri, G., Febres-Cordero., E., Shield, D. R., and Castro, A., (1981), Shear-stress-strain behavior of clays Geotechnique, 31,4, 537552. Napier J.A.L., Malan D.F. (1997), A viscoplastic discontinuum model of time-dependent fracture and seismicity effects in brittle rock, Rock Mechanics and Mining Science, 34, 7, 1075-1089 Obert, L. (1965), Creep in mine pillars: Report of investigation. Rep.No. 6703, U.S. Bureau of Mines. Panet M. (1996), Two Case Histories of Tunnels through Squeezing Rocks, Rock Mechanics and Rocks Engineering 29 (3), 155-164 Peck R.B., Hendron, A.J., Mohraz, B., (1972), State of the Art of soft-ground Tunnelling, Proceedings of the North American Rapid Excavation and Tunnelling Conference, Chicago, IL, pp. 259286. Pellet F., Hajdu A., Deleruyelle F., Besnus F. (2005), A viscoplastic model including anisotropy damage for the time dependent behaviour of rock, International journal for numerical and analytical methods in geomechanics, 29, pp. 941-970. Phienwej, N. (2007).Time dependent response of tunnels considering creep effect., International Journal of Geomechanics, 7, 4, 296-306 Phienwej, N. (1987), Squeezing ground response of sheared shale of Steelwater Tunnel. Ph.D. thesis, Univ. of Illinois at Urbana-Champaign, Urbana, Ill. Potyondy D. (2006), Simulating stress corrosion with a bonded-particle model for rock, Rock Mechanics and Mining Science, 44, 677-691 Ramoni M., Anagnostou G. (2006) , On the feasibility of TBM drives in Squeezing Ground, Tunnelling and Underground Space Technology , 21, 3-4, 262 Semple, R. M. (1973), Effect of time-dependent properties of altered rocks on tunnel Support requirements Ph.D. thesis, Univ. of Illinois at Urbana Champaign, Urbana, Ill. Shao and al (2003), Modelling of creep in Rock Materials in term of Material Degradation, Computer and Geotechnics 30 (2003) 549-555 Shao J.F. et al (2005), Modelling of anisotropic damage and creep deformation in brittle rocks, Rock Mechanics and Mining Sciences, 43, 582-592. Schubert, W.,Button, E. A., Sellner, P. J., and Solak, T. (2003), Analysis of time dependent displacements of tunnels. FELSBAU, 21_5_,96103. Singh, A.,and Mitchell, J. K. (1968), General stress-strain time function for soils. ASCE Journal of Soil Mechanics & Foundations, Vol 95, 1526-1527 Singh, B., Jethwa, J.L., Dube, A.K., Singh, B. (1992), Correlation between observed support pressure and rock mass quality, Tunneling and Underground Space Technology., 7, 59-74. Steiner W. (1996), Tunnelling in squeezing rocks: Case histories, Rock Mechanics and Rock Engineering, 4, 211-246. Sterpi D., Gioda G. (2007), Visco-plastic behaviour around advancing tunnels in squeezing rock, Rock Mechanics and Rock Engineering, doi 10.1007/s00603-007-0137-8 Sulem J., Panet M. and Guenot A. (1987), Closure analysis in deep tunnels, Rock Mechanics and Mining Science, 24, 3, 145-154.