TBM TUNNELLING IN DIFFICULT GROUND CONDITIONS Giovanni Barla (1) and Sebastiano Pelizza (2)
ABSTRACT This paper is to discuss TBM tunnelling in difficult ground conditions, when problems are met which may reduce dramatically the average progress rates and practical consequences may be such as to pose serious questions on the use of mechanised TBM tunnelling versus drill and blast and other so-called traditional excavation methods. Following a few remarks on rock TBM tunnelling in relation to the selection and dimensioning of the machine, the attention is posed on the limiting geological conditions which may be envisaged with respect to the use of TBM tunnelling and on the importance of geological and geotechnical investigations, in order to derive an appropriate understanding of the rock mass conditions along the line of the tunnel. The discussion is centered upon the relatively more important or difficult ground conditions inclu ding borability limits, instability of excavation walls, instability of excavation face, fault zones and squeezing. Whenever available to the authors and based on project experience, the point of view is illustrated by case examples, which give the opportunity to underline specific difficulties encountered and recommendations. INTRODUCTION TBM excavation represents a big investment in an unflexible but potentially very fast method of excavating and supporting a rock tunnel (Barton, 1996). When unfavourable conditions are encountered without warning, time schedule and practical consequences are often far greater in a TBM driven tunnel than in a drill and blast tunnel. The unfavourable conditions can be produced by either a rock mass of very poor quality causing instability of the tunnel or a rock mass of very good quality (i.e. strong and massive rock mass) determining very low penetration rates. However, it is to be observed that when using the full face mechanized excavation method, the influence of the rock mass quality on the machine performance has not an absolute value: the influence is in fact to be referred to both the TBM type used and the tunnel diameter. Right from the beginning of its earliest applications, the use of full face mechanised excavation was to overcome the limits imposed by the local geology, the economic challenges and schedule competitions of the drill and blast method and other so-called traditional excavation methods. A prominent example is given by the recent (from 1995 to 2000) construction of the one tube 24.5 km long Laerdal Tunnel in Norway, the world’s longest road tunnel. This 100 m2 cross section tunnel is being excavated in a precambrian gneiss, a very good and stable rock mass: the supports are on average only 7-8 rock bolts plus a 7 cm thick shotcrete lining per meter of tunnel. The excavation is carried out by the drill and blast method, which has been evaluated to be less expensive and more reliable than the use of a large diameter TBM. The average progress rate is 4.8 – 5.0 km per year with two faces, against the 2.3 - 4.8 km per year, estimated for a large-diameter TBM (Kovari et al., 1993). With this background in mind, this paper is intended to address the problem of TBM tunnelling in difficult ground conditions. Based on a few selected case examples, the discussion is centered upon the relatively more important or difficult ground conditions which can be listed as follows: borability limits; instability of excavation walls; instability of excavation face; fault zones; squeezing.
(1) Politecnico di Torino, Dipartimento di Ingegneria Strutturale e Geotecnica, corso Duca degli Abruzzi, 24 – 10129 Torino, [email protected]
(2) Politecnico di Torino, Dipartimento di Georisorse e Territorio, corso Duca degli Abruzzi, 24 – 10129 Torino, [email protected]
ROCK TBM TUNNELLING The practically infinite number of combinations of rock, soil and environmental conditions which may be encountered during tunnel excavation has determined a great difference in the types and characteristics of the available TBM’s. There are many different schemes for the classification of tunnelling machines. For example the AITES/ITA Working Group No. 14 (Mechanisation of Excavation) is currently working on the definition of an internationally acceptable classification of TBM’s with the purpose of establishing terminology and “guidelines” for the optimum choice of the machine (Table 1). Table 1 : General classification scheme for tunnelling machines (AITES / ITA, Working Group No.14). Method
Partial Face Excavating Machines (PFM)
None or Grippers
Full Face Rotating Cutting Head (TBM)
Cutting disk/ Cutting bits/ Cutting knives & teeth
Grippers and Thrust Jacks
Double Shielded TBM (DS-TBM)
PFM TBM Mechanical Shield
Face and cavity
None or slurry or Earth Press. Balance
Road header/Back hoe Cutting bits/Cutting knives & teeth Road header/ Back hoe/ Manual excavation Cutting disk/ Cutting bits/ Cutting knives & teeth
Road header/Back hoe
Cutting disk/ Cutting bits/ Cutting knives & teeth
Earth Pressure Balance None or fluid
Rod header/ Back hoe/ Manual excavation Cutting bits/ Cutting knives & teeth
Machine Reaction Force
Soft Ground Machines
Support System Location Cavity Face
Type Special Rock Tunnelling Machines - Mobile Miner Contin uous Miner - Other Unshielded TBM Special Unshielded TBM Single Shielded TBM (DS-TBM)
Mechanical Supported Closed Shield Mechanical Supported Open Shield Compressed Air Closed Shield Compressed Air Open Shield Close Slurry Shield – Slurry Shield – SS-Hydroshield Open Slurry Shield – Special Open - Slurry Shields Earth Pressure Balance Shield - EPBS Special EPBS Combined Shield - Mix Shield - Polishield
Rock tunnelling machines can be grouped into three main categories (Table 2): Unshielded TBM (i.e. Open TBM), Single Shielded TBM and Double Shielded TBM which is the way of creating new types of TBM’s that are suitable for application over a wider range of geological conditions, even though the distinction between TBM for rock and TBM for soft ground remains. From the point of view of rock TBM dimensioning, it is necessary to point out that, although TBM’s of more than 10 m in excavation diameter have been constructed, it is always advisable to try to limit the maximum size of the tunnel and therefore that of the TBM.
Table 2 : Schematic comparison among various types of large diameter TBM’s. Reamer TBM Ventilation in the pilot tunnel
Open TBM Easy operatively
Yieldable Single or Double Shield TBM
Single Shield TBM
Double Shield TBM
Advantages Large application range
Large application range
Easy variation of diameter It can be used in hard Safety enlargement rock Precasted lining installation High performance
Support and lining flexibility
Flexibility of supports High performance Construction cost Drive in difficult ground condition
Limited investment Gripper
Disadvantages Gripping in soft or un- Two work phases stable rock mass Drive in weak ground
Cost of investment
Drive in difficult ground
Behaviour in unstable rock mass
Adherence Need of precast lining Behaviour in hard rock mass Behaviour in soft or unstable rock mass Time of construction Costs of construction Range of application
Support installation in unstable rock mass
Cost of investment Complex operatively
Need of cleaning the tele- Cost of investment scopic joint Complex operatively The advantages and disadvantages are mainly theoretical since this machine is under develo pment
As easily perceived, the reasons for limiting the tunnel diameter are: − the potential of a TBM in hard rock decreases with increasing diameter (Kovari et al., 1993; Bruland, 1998); − there are technological limits for the maximum dimensions of some major TBM components, for example, the bearing and the head; − the intensities of both the instability phenomena and the induced convergence also increase with increasing diameter of excavation (Tseng et al., 1998; Barla G. and Barla M., 1998). There already exists a positive and consolidated experience in the use of TBM’s in rocks of different qualities and strengths for excavation diameters up to 12-12.5 m. Beyond 13.5-14 m excavation diameter, the present technology is probably not up to the level of guaranteeing a good performance of TBM’s in hard rock. Designers should take into account these limits during the tunnel design phase, making use whenever possible of the advantages offered by the reduced sections of the tunnel or even by considering the possibility of having other tunnels running in parallel. This is particularly true for motorway tunnels where, in some cases, it is preferable to make triple tunnels each with two lanes for the traffic flow rather than twin tunnels each with three lanes. In the case of railway tunnels, it is better to have two relatively small, single-track tunnels, rather than a large, double-track tunnel. A great help in the use of TBM’s could possibly be achieved through the standardization of the section types for road, motorway, and railway tunnels. This could favour the re-use of TBM’s by obtaining, at the same time, a constancy in the typology and quality of homogeneous construction works, in addition to gain ing considerable advantages in construction times and costs. As well known, the value of the TBM, in terms of direct project costs, is relatively insignificant. Failure to achieve the desired results and maintain the time schedule, however, significantly affects the project. From the outset it is therefore important to adopt the approach of utilizing the best possible equipment, as far as all the aspects pertaining to the TBM and the supporting services are concerned. Generally speaking, the most reliable machines are the simple ones as they have the least amount of equipment that can break down (Foster, 1997). In fact, the TBM that is designed to cover all eventualities has, too frequently in the past, tended to be problematic in service and produced performances below expectations. On the basis of the above considerations and recent and past experience, it is possible to establish the following points regarding the selection of the type of TBM: ⇒ the shielded TBM’s have a wider range of application than the open TBM’s;
⇒ this difference in the range of application increases with increasing diameter of excavation; ⇒ open TBM’s with a double system of grippers are more sensitive to unstable ground than shields with only a single system of grippers; ⇒ the wider or narrower range of application of a single shielded TBM, with respect to the double shielded one, depends on the design and dimensioning of the TBM and on the type of limiting situations to be faced, rather than the TBM type; ⇒ the choice between a single shielded and a double shielded TBM depends on the design of the tunnel section and whether it is necessary to install a precast lining along the entire length of the tunnel to be constructed. It should be recalled that in a tunnelling project there are other problems which need to be solved in addition to the technical ones, i.e.: ⇒ one should be reminded that today there are TBM’s which can reduce significantly the number of geological situations which cause important problems to face advance assuming that these TBM’s are correctly designed and utilized; ⇒ there still exist some limiting situations which can only be overcome by special interventions with unavoidable consequences on the construction time and cost of the project; ⇒ the use of mixed shields is not a solution for overcoming the limits of TBM application in rock, except for some very special cases; in this area there is, however, a great possibility for further development in the design of TBM’s for rock and in the definition of special interventions for application to a given situation; ⇒ the importance of the contractor’s expertise and, above all, the personnel director and technical staff on site, is often not given the right attention, while in reality it plays a primary role in the functioning of a TBM, particularly under limiting conditions; ⇒ the time and cost for overcoming limiting conditions, in a tunnelling project acquired through a competitive bidding, should be supported by the client. He is to account for adequate margins in programming and budgeting whenever a tunnel is to be excavated in difficult ground conditions, according to a risk assessment evaluation; this point is valid also for tunnelling projects to be constructed by the drill and blast method. Given that a TBM capable of advancing under whatever geological condition does not exist, it is also true that the overall result of a project depends on: ⇒ the type of the TBM used, and ⇒ the design and special construction characteristics of the TBM adopted. In fact, it is not sufficient to just order from a qualified manufacturer a particular type of TBM; instead, a continuous collaboration and control of all design and construction details are essential by its intended user, the contractor. This is particularly true as far as there are still no “Accepted Standards” for the design and construction of a type of TBM, and each TBM to be constructed is to be considered as a prototype, one different from another one, given that: ⇒ the design and manufacturing of TBM’s is a continuous, technologically innovative process, ⇒ each tunnelling project has its own characteristics, and each specialised contractor has his own traditions and opinions. LIMITING GEOLOGICAL CONDITIONS FOR TBM’s APPLICATION A limiting condition for the use of TBM excavation can be defined where the geological conditions are such that the same TBM cannot work in the execution modes for which it was designed and manufactured. For this reason the advance of the TBM is significantly slowed down or even obstructed. A geological condition is intended to be a limiting one only in relation to the type of TBM used, its design and special characteristics, and eventually any operating errors. A particular geological condition becomes a limiting one only when it is beyond a certain importance, or when the associated problems are beyond a certain level of severity, or else due to combination of events each being by itself not critical. The Importance of Geological and Geotechnical Investigations Despite the excellent performance of TBM’s in favourable ground conditions, as reported in recent years (e.g., more than 1 km advancement per month for some hydraulic tunnels), in many cases the actual advancement rates have been below expectations and certainly less than claimed by TBM manufacturers.
It would therefore be legitimate to think that, besides the unforeseen events, such as breakdown or failure of the TBM components, the rock mechanics problems are often under-evaluated or neglected. It should be noted that the purpose of construction is to achieve the objective of the design and that the work must be manlike (as defined in design), according to the specified safety factors and the expected time and cost. The design has always been carried out by using a deterministic approach. Reality of construction however has never been so. This is due to the large number of uncertainties that cannot be avoided at the design stage: geological, geotechnical, hydrogeological uncertainties, different types of machines available (new or used), and different construction techniques (Pelizza, 1998). Hence, at the design stage, it is impossible to know every aspect of the geological profile. It is therefore necessary to decide whether to optimize the choice of the construction method or the selection of the machine for a given tunnel, on the basis of the understanding of site geology and geotechnical conditions or of the level of prediction about these conditions (up to which point are these predictions optimistic or pessimistic?). On the other hand, the problem of global optimization is very complex, given the large number of geological, technological, environmental, and economic-financial variables involved. At the present time, it is becoming possible to manage, in probabilistic terms, the decision strategies for tunnelling under uncertainty conditions involving various levels of variability (Einstein, 1996; Xu et al., 1996). The fundamental problem is always determined by the physical and geotechnical heterogeneity of the rock mass in which the tunnel is to be excavated. For a full face mechanised excavation, which is a rather rigid system, the strength heterogeneity of the material to be excavated is even more important, be it a rock or soil. Prior understanding, obtained in a correct manner, of the geological and geotechnical conditions of the site is fundamental for the development of underground works. Up to now, too little money has in general been spent on preliminary investigations. It has in fact been demonstrated that money spent on such investigations is greatly compensated by the savings made in terms of construction cost and time. Forward probing from a TBM driven pilot tunnel or a main tunnel is not an alternative to an adequate pre-investigation. A considerable example of the positive effects on the performance of a TBM purposely constructed by accounting for a good understanding of geology obtained on the basis of a detailed and preventive geological investigation, is offered by the recent (1995-1997) construction in Spain (Trasvase Guadiaro-Majaceite) of a 12.185 kilometer long water conveyance tunnel, 4.2 m inside diameter (Castello et al., 1999). The tunnel runs under a maximum overburden of 500 m, through heterogeneous and complex ground, composed of sedimentary and tectonized rocks (from hard limestone, marbly limestone and soft and swelling argillo-arenaceous and clayey flysch). The preliminary investigations comprised 29 probe holes for a total length of approximately 6000 m (0.5 m of investigation hole per meter of tunnel). Significant convergences with a squeezing behaviour were observed to occur in flysch along a 3940 m length in the northern section of the tunnel, where a lining formed by precast segments was installed. In the remaining tunnel length, in a predominantly hard limestone, where stable rock conditions were expected, shotcrete and rock bolts were installed followed by a cast in situ concrete lining. On the basis of a detailed understanding of both the geological conditions and the construction requirements, the TBM used has been constructed with a continuous interaction between the TBM manufacturer and the contractor. The TBM had to be designed to bore in both the soft argillaceous flysch and in the hard limestone, with the purpose to avoid it from being trapped especially in the clayey sections under a thick overburden. The important characteristics of the TBM are the following: − double shielded TBM with four cylindrical shells of 4794 mm diameter, so as to allow the lining installation in the argillaceous sections and to advance with grippers in the hard rock mass section; − cutter head diameter of 4.88 m with the opportunity to overcut up to 50 mm and to be used in the clayey sections with squeezing behaviour; − variable speed rotation of the cutter head from 0 to 9 rpm versus the ground type, with a maximum torque of 3900 kNm; − the standard upward thrust necessary to the TBM was about 3300 ton and up to 3700 ton when the TBM staggered and set off again. The recommended value of the total upward thrust for the design of the TBM was 4000 ton in squeezing ground (Lombardi and Panciera, 1997). This 10% margin has been very useful to prevent blocking and has revealed to be insufficient only in very few situations.
The TBM exhibited a good performance in both hard and soft ground. The overall average penetration of 1.82 cm/min (calculated for all the tunnel sections) corresponds to a TBM average daily advance of 26 m. The best performances of the TBM have been: − 78 m/day; − 342 m/week; − 1335 m/month. The tunnel has been completed in 17 months with an average monthly progress of 747.73 m. TYPICAL CASES In the following, the relatively more important or frequent difficult ground conditions, which can affect TBM performance, will be considered: borability limits; instability of the excavation walls; instability of the excavation face; fault zones; squeezing. It should be pointed out that tunnel excavation by a TBM may encounter other difficult ground conditions due to the presence of: clayey soil; soft ground resulting in settlement of the TBM; strong inflow of groundwater and gas; rock bursting, rock and water at high temperature; and karstic cavities. Borability Limits A rock is said to be not borable if the TBM cannot penetrate the face to a sufficient rate and/or the wear of the cutting tools exceeds an acceptable limit. The borability of a rock should not be established in an absolute manner, but only relative to an alternative drill and blast method, comparing the cost and scheduling aspects of both the methods. The main index describing the capacity of a TBM to excavate a given rock is the penetration rate per revolution of the cutter head which the TBM is able to achieve under the maximum thrust. It is not possible to establish a limit of penetration per revolution below which a rock shall be considered non borable; such a limit is also influenced by the abrasivity of the rock, the diameter of the tunnel, and the thickness of the rock formation in question. The high abrasivity, associated with low penetration, dictates frequent changes of cutters, increasing the cost for each cubic meter of rock excavated, in addition to the time lost in substituting the cutters. With increase in the tunnel diameter there are three different effects which make the situation worse: ⇒ the rotational speed of the cutter head should decrease for an equal penetration per revolution, because the bearings and seals of the disc cutters permit only a maximum speed equivalent to 150 m/min; ⇒ the number of cutters to be changed per meter of tunnel advance increases, increasing, therefore, the stoppage time required for such operations; ⇒ the state of average wear of the cutters mounted on the head increases, thus decreasing the penetration per revolution. Under extreme conditions, each one of the above three factors excites the other one bringing the progress rate down to unacceptable values. For these reasons, a rock type may be borable for a TBM of small diameter, but not for a TBM of large diameter (Figure 1).
Figure 1 : Comparison of penetration rates.
To give some numbers for reference, one may consider that penetration rates below 2~2.5 mm/rev of the cutterhead are indicators of borability problems, whereas an excavation starts to be efficient if the penetration rates are above 3~4 mm/rev. Naturally, the theoretical performance of a TBM is affected by various activities strictly related to the functioning of the machine. It happens rather often that the cutting wheel is pushed to a maximum in order to maintain an adequate penetration rate, even in high strength rocks. If each part of the TBM has not been planned and built to work under these conditions, the machine will vibrate in an anomalous manner and cracks will gradually appear in the cutting wheel and gripper structures. As it is not easy to repair or substitute the cutting wheel in a tunnel, the consequent loss in production of the machine can be very serious. Instability of the Excavation Walls The instability of the excavation walls is a limiting characteristics for open TBM’s. The problem arises when the instability occurs immediately behind the cutter head, making it difficult to install important supporting elements and position the grippers. The consequences on the progress rate and on the methods employed to overcome the instabilities vary significantly depending on: − magnitude and type of the instability phenomena; − type of TBM used (a simple or a double system of grippers); − design and characteristics of the TBM; − tunnel diameter; − system built in the TBM for installation of tunnel support, and the type of support used. In very serious situations, the daily advance may reduce to 1~2 m, or even to zero. A case example is briefly described in the following. Case Example The design of the new Highway A1 between Sasso Marconi and Barberino del Mugello (Italy) called for the excavation of four exploration pilot tunnels on the line of the Appennines base tunnel. This is to take place before the excavation of two parallel twin tunnels, each 8500 m long under an overburden ranging from a few metres to 400 m. Four open TBM’s, each 3.9 m in diameter, were selected for the excavation of the pilot tunnels under the condition to start from the South Portal (Florence Left and Right Pilot Tunnels) and North Portal (Bologna Left and Right Pilot Tunnels) respectively. For the purpose of the present paper and the interest to pay attention in this section to the instability of the excavation walls, the difficulties met by the two open TBM’s (Atlas Copco), excavating from the North Portal in Mudstone and in Flysch (composed of sandstone and marl layers), can be mentioned. Here the average progress rate resulted to be equal to 13.3 m/day and 12.5 m/day respectively for the right and left tunnels against the expected 25 m/day. The total length excavated with the above poor performance summed up to nearly 3.5 km. During the pilot tunnel excavation the rock mass conditions were assessed systematically in terms of both the RMR index (Bieniawski, 1989; 1997) and the Önorm 2203 system (Lauffer, 1997). A simplified plot of the rock mass conditions in Flysch between chainage 1800 m and 3600 m along the left pilot tunnel from the North Portal is shown in Figure 2 (the plot has been simplified for the purpose of this paper). It is apparent that the rock mass conditions are changing continuously at least up to chainage 3000 m, thereafter attaining a RMR value of 35 approximately. Over a total length excavated of 2400 m the mean RMR is 42 with a standard deviation of 12.7. In terms of the normal distribution curve obtained for the overall data, it is shown that the probability that the rock mass is poor or very poor is greater than 88 per cent.
Figure 2 : Rock mass conditions along the pilot tunnel length. As shown in Figure 3 a, with the worsening of the rock mass conditions and the reduced stand-up time, instability of the tunnel circumference started to occur immediately after excavation, so as to require installation of a nearly continuous form of support close to the face. The instability was observed to take place with rock blocks (with dimensions of several centimeters to decimeters) separating from the crown and the walls along bedding planes and jointing (Figure 3 b), thus determining overbreaks reaching a depth in the rock mass surround up to 0.5-1.0 m in 30 min approximately. In cases, the presence of water was such as to favour this instability. With the purpose to underline the type of instability observed, we would like to discuss the results of computations of the disturbed zone around the tunnel in terms of numerical methods (by the use of a 2D distinct element code, UDEC) (Shen and Barton, 1997). The rock mass is simulated by using two sets of persistent joints nearly perpendicular with each other, with the bedding (B) and jointing (J) as shown in Figure 4, having a mean spacing of 30-40 cm, with a total of 6500 blocks in the model. The discontinuities are assumed to be Mohr-Coulomb joints, i.e. elasto-perfectly plastic joints. The blocks are treated as elastic blocks. The mechanical properties are as summarized in Table 3 and have been set in order to attain a Q value of 0.43, equivalent to RMR equal to 40 m.
Figure 3 : Photographs showing: (a) the installation of continuous support just behind the TBM head and (b) the flysch as observed at the sidewalls of the pilot tunnel.
Table 3 : Mechanical properties. Block Young’s modulus, Ed (GPa) 20.0 0.2 Block Poisson’s ratio, νd (-) Joint normal stiffness, kn (GPa/m) 24.0 12.0 Joint shear stiffness, ks (GPa/m) 2.4 1.2 30 20 Friction angle (peak value), φ p (°) 20 15 Friction angle (residual value), φ r (°) Cohesion (peak value), cp 0.0 0.0 Cohesion (residual value), cr 0.0 0.0 Note: Joint and bedding are given zero dilation angle. As shown in the plot of Figure 4 a, the local instability at the crown and sidewalls, illustrated by the overbreak with blocks falling into the tunnel, gives a representation of the instability mode which is similar to that observed in situ. The stress condition assumed is σy = 8.0 MPa (vertical stress), σx = 4.0 MPa (horizontal stress), which represent gravity induced stresses at a depth of 320 m, with a stress ratio of 0.5. It is interesting to compare the overbreak shape given by the discontinuum UDEC model with the shear zones around the tunnel computed by an analytical method (Shen ad Barton, 1997) where the ubiquitous joint concept has been applied.
(a) discontinuum model – loose blocks
(b) continuum model – yielded zones
Figure 4 : Continuum and discontinuum modelling by back analysis of the wall instability in flysch, just behind the TBM head. Remarks on the use of Shielded TBM’s The shielded TBM’s, either single shielded or double shielded, are not as sensitive as open TBM’s to the instability phenomena of the excavation walls since a precast concrete or steel lining (or pre-lining) inside and under the protection of the shield can be installed. By pushing against the lining it is thus possible for the TBM to advance independent of the instabilities. In the case of medium to large diameter tunnels (from 6 to 12 m) the difference in behavior and progress rate between open TBM’s and shielded TBM’s, under the condition of wall instability, increases considerably even of orders of magnitude, the advantages being naturally on the side of the shielded TBM’s. With open TBM’s the possibility of counteracting effectively the instability phenomena at the excavation walls depends on the following: − stabilisation and reconstruction of the walls, executed immediately behind the support of the cutter head using steel arches, wood lagging and shotcrete. The installation of these supporting elements, particularly shotcrete, in this delicate zone of the machine requires long times and also risks to damage the excavation equipment; − traditional excavation in front of the TBM, often by the top heading method (Tseng et al., 1998);
− pre-treatment of the rock ahead of the excavation face through borings and injection holes, or by installing an umbrella arch above the TBM. The main problem here is that often, being different from other limiting situations of localised character, the phenomena of strong instability of the excavation walls may involve important sections of a tunnel, especially if the tunnel is of large diameter, where the quality of the rock is technically speaking not necessarily very poor. Under these circumstances the choice is either to withdraw the TBM from the tunnel, or to accept serious delays. Instability of the Excavation Face When the state of fracturing and/or weathering of the rock mass to be excavated is such that major instabilities occur at the excavation face with falling down of blocks and fine materials, which does not stop until equilibrium is attained causing large over-excavations, it is possible to arrive at a limiting situation for the functioning of shielded TBM’s for rock. In this situation, the advance of the machine may be hampered for two fundamental reasons: • the cutter head can no longer rotate because the accumulated, failed materials act against or block the head • the over-excavation caused by the instability is such that cavities are formed in front of the TBM, which suggest to stop the advance and treat the problem before the situation self excites and eventually becomes uncontrollable. This is a typical limiting situation which also affects shielded TBM’s of any type as briefly discussed in the case example which follows. Case Example The Pinglin Highway tunnel in Taiwan is a case of great interest and very instructive one, when dealing with instability conditions at the face (Tseng et al., 1998; Wallis, 1998). This tunnel is to be constructed by using three shielded rock TBM’s: a 4.8 m diameter Robbins machine for the pilot tunnel; two 11.74 m diameter Wirth machines for the main tunnel (Figure 5). For the main tunnel average progress rates of up to 360 m/month/machine were anticipated using TBM’s as opposed to the 50 m/month/heading forecast for drill and blast. A 4.5 TBM programme for the 12.9 km long tunnel outstripped the estimated 15 years drill and blast alternative.
Figure 5 : Photographs showing the east end portal of the Pinglin tunnel with: (a) the three tunnel tubes with the two large diameter TBM’s in assembly; (b) the pilot tunnel. The many problems met up to now during excavation of the Pinglin tunnel are connected with partic ularly complex geological conditions characterizing the first 2-3 km on the eastern side of the tunnel, which were not adequately evaluated. These conditions can be summarized as follows:
− the rock mass is a very hard sandstone (σc up to 350 MPa; quartz content up to 98%), layered and tectonized with thin clay fillings; it behaves sometimes as ravelling rock and sometimes as squeezing ground, when clay is present; − sudden inrush of great quantities of groundwater took place in the tunnel through zones of faulted and fractured rock; − the small diameter TBM, which was specifically constructed for the excavation of the pilot tunnel in a good rock mass and expected good progress, encountered instead severe problems in difficult geological conditions (in little more than 1 km, face collapses have on ten occasions caused the TBM to jam); − exploration and preventive drainage holes during face advance revealed to be very difficult, time consuming and extremely costly to drill, due to the rock high strength and abrasiveness; the same is true if holes are to be driven around the circumference of the tunnel in order to improve the rock mass properties by grout injection, thus facilitating the excavation by the large diameter TBM’s; − the two large diameter TBM’s were constructed with the most advanced technological characteristics available at the time, in consideration of the difficult conditions expected; however, these are to be used in a rock mass of very unfavourable characteristics which cannot be improved from inside the pilot tunnel, as initially expected. Considering the above, the following problematic conditions occur: • instability of the face, which is difficult to control, considering that a thrust adequately high cannot be applied due to the extremely high rock abrasivity that can cause a rapid wear of the cutter head; • instability of the tunnel circumference (crown and walls), which makes it difficult to grout the annular gap between the concrete lining and the tunnel profile; • tightening of the shield due to the likely interaction with the rock mass, associated with heavy loading of the thin concrete lining, which as such is not able to react adequately to the high axial thrust required. As a consequence, the use of systematic and continuous pre-treatment ahead of the face, in order to overcome any difficult conditions along the tunnel, was ruled out, due to the high costs expected and the long time required. Rather, the decision has been to tunnel through the quartzite sandstone formation by drill and blast, in the hope to be able to come back to mechanized excavation by the two TBM’s, where more favourable rock mass conditions will be encountered, allowing this to be effectively done. Considerations According to recent experiences (gained from Evinos, Yellow River, EOS, Pinglin), in order to avoid situations like those described above it is necessary to design the TBM with the cutter head protruding outside the shield as less as possible, allowing the shield itself to support the excavation as close to the face as possible. In addition, it is also true that under these limiting conditions, the capability of yielding a high cutter head moment, both during starting up and during excavation, and of adjusting the rotational speed of the head itself, are of great help. Furthermore, the cutter head should be designed without protrusion and without adjustable buckets. With these improvements in machine construction, it is possible to move the limits toward those situations which are truly exceptional, without avoiding them. In these cases, the frequently used interventions are (Figure 6): • grouting and backfilling of the cavity with resins and foams, in order to create a kind of artificial conglomerate; drilling and grouting are usually performed through special holes in the cutter head, which should have an accessible, internal cavity with adequate dimensions to permit such operations • excavation of a by-pass tunnel (preferably in the roof) to free the cutter head blocked by rock blocks, to stabilize the opening, and to excavate a section of the tunnel using conventional methods, or improve the ground conditions through grouting and forepoling.
Figure 6 : Examples of instability problems at the tunnel face, as met during excavation of the Pinglin tunnel. The execution of an umbrella by drilling and injecting through special holes in the shield has not been demonstrated to be of great help in overcoming such exceptional situations, because the holes arranged in the shield are widely spaced (for the functioning of other components of the TBM) and diverge towards the face, thus not permitting a sufficiently effective ground treatment. Nevertheless, a complete set of equipment and treatment material (grouting pumps, drilling machines, drill bits and rods, resins, foams, etc.) should be maintained at the construction site for the special remedial operations, in order to limit the duration of ground treatment (except for exceptional events) to a few days; otherwise, TBM blockages may last weeks or even months. Fault zones The crossing of fault zones in TBM tunnelling represents in general a problematic event and is often associated with a slow-down of progress rate, if not big delays in time schedule, when a blockage of the TBM head occurs. Although localized along the tunnel length, this may represent an unexpected event, due to either a lack of warning during excavation or to a difficulty which has been greatly under estimated or not understood well ahead of time. The types and extent of instabilities which occur in such cases may involve face collapses as previously described. Major high pressure inflows, outwash of fines, formation of cavities at the crown (“church roof” or “natural shafts”) or ahead of the TBM take place at the same time. The most serious situation is when the tunnel is flooded and the fines are pushed into the tunnel, while rock blocks are forced against the TBM head. If an open TBM encounters this type of situation without pre-identifying it by probe drilling, the condition may become dramatic and very difficult to deal with. When the same fault is met by a shielded TBM, be it a single shielded or a double shielded machine, although it is certainly not possible to continue the excavation, the treatment of the fault from inside the shield becomes possible and at the same time the tunnel may be kept from being completely filled up.
Before presenting two cases of fault zones crossing, it is relevant to point out that, unfortunately, with a certain frequency the encounter of a fault zone may become a catastrophic event due to an incorrect excavation procedure: the TBM operator, by stopping the machine and rotating the cutter head, induces the fault to “emptying over the TBM”. Moreover, when a fault is encountered, a conflict usually arises between (1) people who would like to adequately treat the ground and (2) people who are in favour of resuming excavation as quickly as possible, thus using only light and fast ground injections, without retreating the TBM. In the latter case, an argument which is generally invoked is the fear to stuck the head by cement grout. The catastrophic consequences of the latter procedure is demonstrated by the example of Figure 7 which shows the case of a pilot tunnel TBM excavation. The second collapse, as illustrated in Figure 7 b, was the cause of two fatalities and the permanent stoppage of the TBM excavation, which was intended to be an exploratory tunnel, before excavating to full size. fault gouge
(a) TBM blocking: the TBM operator tries to “empty the fault”: − the cutting wheel is turning; − the machine is stopped; − a large block of gneiss enters the gap behind the cutting wheel and jams it.
(b) Tunnel Collapse: − the TBM is pulled back to remove the gneiss block; − the TBM starts and a first collapse occurs; − fast ground injections (chemical) are carried out; − the TBM restarts and a second collapse occurs with a large water inrush; − the TBM is permanently stopped.
Figure 7 : Schematic representation of erroneous procedures sometimes adopted when crossing fault zones. The sketch illustrates a real case in a gneiss rock mass. Case Example 1 Typical instability problems when tunnelling in jointed and faulted rock are being experienced to a different degree of severity in the F2 tunnel, one of the major element of the Pont Ventoux - Susa Hydropower System in the Susa valley, near Torino (Italy). This tunnel (4.75 m diameter) is being excavated by an open TBM configuration (Robbins) through quartzitic michaschists under a cover which is to reach 800 m maximum. Following the first 1800 m approximately where the rock mass conditions were generally good, with RMR values ranging from 65 to 75 (Figure 8 a), TBM tunnelling in the F2 became progressively more difficult as fault zones crossing the tunnel in a subparallel manner were met which had disastrous results on progress (Figure 8 b). Two of such zones are briefly described below.
Figure 8 : Photographs showing: (a) the tunnel in good rock mass conditions (RMR: 65 to 75) compared to (b) the difficulties being met in the fault zones following chainage 2800 m approximately, well stigmatized by the heavy support being installed immediately behind the TBM head. • The sketch shown in Figure 9 a gives a simplified illustration of the conditions at the tunnel face at chainage 2360 m approximately, where the TBM became temporarily stuck as a consequence of overstressing and a 25 cm block movement of the right sidewall (Figure 9 a). The quartzitic micaschist is characterized by the presence of three to four joint systems including foliation. Based on geological mapping, which was carried out by the contractor’s geologist (Pont Ventoux, 1997) once the TBM could drill a few meters ahead of the section where it jammed, at least two sub-parallel discontinuities could be evidenced, the second of which (a fault with strike N66E and dip 83 to the S, which intercepts the tunnel axis) has a clay filling and gouge with aperture ranging from a few centimeters to more than a decimeter.
Figure 9 : Typical fault condition in quartzitic michaschist: (a) sketch showing the machine jammed; (b) photograph taken just behind the TBM head, illustrating the block movement at the right sidewall.
The rock mass conditions were estimated on a 7 m tunnel length, from ch 2349 to ch 2356 m , with RMR index equal to 31. According to a more complete Q-logging estimate due to Barton (1997), from ch 2350 to ch 2360 m, an extreme range of Q-values of about 0.007 (“exceptionally poor” – locally) to 0.3 (“poor”) showed a weighted mean of all recordings of about 0.05. Water is flowing through the joints. No quantitative data have been recorded of the water pressure, which seems unlikely to be exceeding, with an overburden of 650 m, a maximum of 6 to 7 MPa (outside the tunnel drainage area). Although, this fault zone does not appear to be particularly severe, at least when compared to the conditions which are presently experienced, the progress rate in the tunnel was seriously affected, for more than 50 meters. • The sketch shown in Figure 10 give the illustration of a fault zone which was met along the same F2 tunnel from chainage 2727 m onwards. The consequence of the TBM performance were extremely serious and the face advance had to stop for several months in order to: (1) probe ahead of the tunnel face by means of 3 m long boreholes, drilled horizontally; (2) map the extent of the fracture zones to within 30 m ahead of the tunnel face by means of tomographic imaging; (3) decide the stabilization measures to adopt in order to secure the TBM advancement. The photographs shown in Figure 11 a, b give a fair illustration of the fault as observed when standing on the TBM head, at chainage 2742 m approximately. The fault is nearly vertical (120-125/70-75 estimated orientation). The nature of the fault material is clay with crushed rock and the width of the fault zone varies from a minimum of 0.5 m to more than 1.0 m. The right wall is formed by a sound rock mass, whereas on the left wall a number of unstable blocks were observed, posing stability problems. An abundant inflow of water has outwashed the clay thus creating an open cavity upwards, extending up to 5 m as shown in the sketch of Figure 10. From the decision to resume the TBM excavation, from chainage 2727 m up to reaching chainage 2750 m approximately, at the end of the fault zone intersecting the tunnel, it required a total of 10 weeks, with weekly progress ranging from 0-0.4 m up to 4.6 m. The provisions taken in order to excavate through the fault zone called for the concurrent placement of steel sets one close to the other one, while back filling the cavity at the crown and on the left wall by concrete. As the excavation was slowly proceeding ahead, the steel sets were placed immediately behind the head (as shown in Figure 10) and 2.4 m long bolts were installed at the crown, one on each side of the fault, with a spacing of 0.5 m.
Figure 10 : Sketch of the fault, causing a stoppage of the TBM (from the Geologist report).
Figure 11 : Photographs of the fault taken at chainage 2742 m looking along the fault: (a) left wall; (b) right wall. Case Example 2 This case example considers the instability which occurred during the excavation of the Frasnadello tunnel, near the town of San Pellegrino Terme, in Italy (Barla, 2000). The tunnel was excavated by a 11.8 m diameter shielded TBM (Figure 12), which was used previously to excavate the Mont Russelin tunnel (predominantly in shales and marls), in Switzerland (Steiner, 2000). When going through the thrust zone illustrated in Figure 13, the instability developed blocking the TBM head, thus causing a serious delay to project completion.
Figure 12 : Photograph showing the 11.8 m diameter TBM used to excavate the Frasnadello tunnel.
Figure 13 : Geological conditions at the thrust zone. Based on preliminary investigations and pilot tunnel mapping, in conjunction with drilling of a number of exploratory holes following the machine blockage, the geological conditions in the thrust zone could be defined in detail as shown in Figure 13. The following remarks are possible on the type of instability: • the instability occurred with a sudden inflow of rock blocks, clay and water into the pilot tunnel, which was excavated two years in advance of the main tunnel excavation (the 3.9 m diameter pilot tunnel was excavated by an open TBM) • the TBM was stuck by the ground above, making it impossible to continue with face advance, independent of the many attempts made to free the TBM head • water was percolating through the thrust zone with flow rate ranging from 6.6 l/s minimum to 10 l/s maximum: as a consequence of the ground collapse, the water flow rate increased from the initial 1-2 l/s in the pilot tunnel. As usual in similar circumstances, once it was realized that any attempt to free the TBM was to be unsuccessful, lengthy discussions took place in order to decide the measures to adopt. Finally, ground freezing was chosen as the most reliable measure to be carried out from the pilot tunnel. As shown in Figure 14 a, a working access chamber was created, starting from the pilot tunnel, with the intent to reach the TBM head. The main working stages were as follows: • stage 1 – creation of a consolidated, nearly impervious ground arch around the tunnel perimeter to 5-8 m depth, with drainage holes above: this was performed from the back of the TBM, just behind the shield, and intended to diminish the water flow down to 3 l/s; foam resin and silicate injections were carried out over a 140-150° arch • stage 2 – creation of a working access chamber (length 8 m approximately) starting form the pilot tunnel, in order to allow for the launching of pipe spiles (length 22 m) ahead: the spiles (with spacing 60 cm) of the MPSP (Multiple Packer Sleeved Pipes) type, to avoid any cement from reaching the TBM, were anchored into the sound limestone, above the shield; also this preliminary ground treatment was carried out in order to compact the loose zones, and reduce the risk of water percolation during freezing • stage 3 – execution of ground freezing by using liquid nitrogen: a frozen vault was formed having a minimum thickness of 80 cm at the crown and 100 cm at the footwall; the ground freezing system was designed so as to allow for continuous temperature monitoring (by means of 34 thermocouples well distributed along the vault) • stage 4 – excavation of the access chamber to full length, to reach the TBM head: this was done with a 0.9 m face advance each time, by using a primary lining formed with heavy steel ribs and shotcrete (20 cm thick); provisions were taken to allow for placing of a provisional shotcrete invert, if needed (see
Figure 14 b, which shows the chamber completed); during the excavation period, intermittent freezing was used to limit the expansion of the frozen structure beyond the design limits • stage 5 – driving of the TBM through the thrust zone and placement of the precast reinforced concrete segments, followed by filling the gap with pea-gravel: as shown in Figure 14 a, a fiber shotcrete and reinforced concrete pre-lining was placed along the access chamber perimeter, leaving 10-20 cm spacing between the same pre-lining and the TBM.
Figure 14 : (a) Schematic illustration of the stabilization measures adopted to free the TBM head; (b) photograph showing the working chamber completed with the TBM in the background. Squeezing Ground Conditions A TBM is definitely under difficulty whenever tunnel convergences (it does not matter its cause or origin) occur with considerable magnitude at a short distance from the face (a few metres) within periods of time shorter than 4-8 hours. This is a limiting condition which is of concern for both designers and contractors. In reality, jamming of rock TBM’s for long periods of time due to convergence problems has not been heard of, at least recently, whereas problems related to the need to overcome face instabilities are reported more frequently. On the other hand, medium to long term convergence related problems are perhaps more frequent in relation to floor heave and support failures, once the tunnel has been supported and/or completed (Figure 15 a, b).
Figure 15 : Photographs showing typical floor heave and squeezing and/or swelling phenomena in tunnels (a, tunnel in claystone – South Italy: courtesy of Prof. V. Cotecchia; b, tunnel in marl and marly chalk – Cyprus: courtesy of Dr. J. Sharp). Shielded TBM’s are notoriously sensitive to rapid convergences and to the risk of blockage by converging rocks, if special technical precautions are not taken. For the open TBM’s, whenever large convergences take place in a short time and if these are associated with instabilities, as observed in situ in a number of cases, problems of support installation and gripping may occur, hampering the progress of excavation. In order to cope with these problems, for most TBM’s one foresees the possibility of increasing the diameter of the cutter head (overcutting), with the aim to be able to adjust the gap between the shield and the excavation contour from the usual value of 6-8 cm to 14-20 cm. Nevertheless, the major difficulty for open TBM’s advancing in strongly converging and unstable ground derives from the difficulty of placing supports like steel sets, wire mesh and shotcrete, and the inability of these to counteract immediately the tendency of the ground to deform and squeeze. With shielded TBM’s the maximum thrust of the longitudinal jacks can be increased to reach such a high level to permit the machine to advance also in the presence of high ground pressures (2-5 MPa), provided that the segmental lining is sufficiently robust and capable of supplying the necessary reaction to the thrust jack, otherwise the same lining will collapse. These tactics and measures, together with overcutting, allow the shielded TBM’s to progress in nearly every type of situation including the so-called exceptional situations. However, in cases where a TBM is forced to stop for a long period of time due to mechanical breakdowns, in zones of squeezing ground, the risk for the TBM being blocked becomes very high. The operation of freeing a TBM is relatively easy with a double shield TBM, where it is possible to interfere at 4-5 m from the face through opening of the telescopic zone. Instead, for a single shielded TBM, the freeing operation has to start from behind the tail, by demolishing one or two rings of the precast lining at about 8-9 m from the face. REFERENCES Barla G. and Barla M. (1998). “Tunnelling in difficult conditions”. Int. Conf. on Hydro Power Development in Himalayas, Shimla (India), 20-22 April, 19 pages. Barla G. (2000). “Lessons learnt from the excavation of a large diameter TBM tunnel in complex hydrogeological conditions”. GeoEng 2000, International Conference on Geotechnical & Geological Engineering, Melbourne, Australia, 19-24 November 2000. Barton N. (1996). “Rock mass characterisation and seismic measurements to assist in the design and execution of TBM projects”. Taiwan Rock Engineering Symposium, December 12-13, pp.1-16. Barton N. (1997). “Pont Ventoux Hydropower Project. Tunnel F2 – Pont Ventoux. Assessment of geological conditions and required support for ch 2350-4000”. Report to Nocon. October 1997. Pont Ventoux (1997). “Nota relativa al sopralluogo effettuato in data 8/10/1997 in corrispondenza del fronte di scavo della galleria di derivazione nel tratto Pont Ventoux – F2 (Dott. Geol. G. Venturini).
Bieniawski Z.T. (1987). “Quo vadis rock mass classifications? - Some thoughts on rating methods for tunnel design”. Felsbau – Rock and Soil Engineering 3/97, pp. 177-178. Bieniawski Z.T. (1989). “Engineering rock mass classifications”. John Wiley and Sons, New York. Bruland A. (1998). “Future Demands and Development Trends”. Norvegian TBM Tunnelling, Publication No. 11, NFF, Oslo, pp. 99-103. Castello G., Chantron L., Fabre D. and Sem M. (1999). “Analysis of the TBM performances for the excavation of the Trasvase Guadiaro-Majaceite, Province of Cadiz, Spain”. Proc. Of ITA - World Tunnel Congress, Oslo 1999, Balkema, pp.829-838. Einstein H.H. (1996). “Risk and Risk Analysis in Rock Engineering”. Tunneling and Underground Space Technology, Vol.11, N.2, 141-155. Foster J.R. (1997). “Gibraltar Strait Crossing. Characterization of TBM”. UN/ITA Workshop on Characterization for tunneling flysches, Tarifa, 20-22 February, 16 pages. Kovari K. Fechtig R., Amstad Ch. (1993). “Experience with large diameter Tunnel Boring Machines in Switzerland”. Int. Congr. Option for Tunnelling, ITA, Amsterdam, 14 pages. Lauffer H. (1997). “Rock classification methods based on the excavation response”. Felsbau – Rock and Soil Engineering 3/97, pp. 179-182. Lombardi G., Panciera A. (1997). “Problems with TBM and linings in squeezing ground”. Tunnel and Tu nneling International, June 97, pp. 54-57. Pelizza S. (1998). “Selection of TBM’s”. Workshop on Selection of Tunnelling Methods, ITA World Tunnel Congress ’98, Sao Paulo (Brazil), 15 pages. Shen B. and Barton N. (1997). “The disturbed zone around tunnels in jointed rock masses”. Technical Note. Int. J. Rock Mech. & Min. Sci., 34 (1), pp. 117-125. Steiner W. (2000). “Mechanised construction of large tunnels in rock”. Eurock 2000 Symposium. Verlag Glückauf GmbH. Essen, pp. 235-242. Tseng Y.Y., Wong S.L., Chu B., Wong C.H. (1998). “The Pinglin Mechanized Tunneling in difficult Ground”. St 8th Congr. of IAEG, Vancouver (Canada), 21-25 September, 6 pages. Wallis S. (1998). “Pinglin perseverance in Taiwan”. Tunnel 7/1998, pp. 10-24. Xu S., Mahtab A., Grasso P. (1996). “The use of some decision-making tools in Tunneling”. Gallerie e Grandi Opere Sotterranee, No.49, Luglio 96, pp. 38-49.