32

Tunnelling A M Muir Wood FRS, FEng, FICE

Consultant, Sir William Halcrow & Partners

Contents 32.1

The options for a tunnel route

32/3

32.2

Costs of tunneling 32.2.1 Principal factors 32.2.2 Effect of tunnel size

32/3 32/3 32/3

32.3

Systematic site investigation 32.3.1 Geological data 32.3.2 Objects 32.3.3 Means

32/4 32/4 32/4 32/4

Tunnelling methods related to the ground 32.4.1 Historical background 32.4.2 Shield tunneling 32.4.3 The bentonite shield 32.4.4 Rock-tunnelling machines

32/4 32/4 32/4 32/6 32/6

32.4

32.5

Tunnel construction 32.5.1 Drilling and blasting 32.5.2 Spoil handling 32.5.3 Tunnel lining 32.5.4 Thrust boring 32.5.5 Waterproofing 32.5.6 Temporary support 32.5.7 Advance by full face or by heading

32.6

Aids to tunnelling 32.6.1 Compressed air 32.6.2 Ground treatments 32.6.3 Freezing 32.6.4 Dewatering

32/14 32/14 32/14 32/15 32/15

32.7

Ground movements

32/15

32.8

Tunnel design 32.8.1 Stresses around a tunnel 32.8.2 Stability ratio, rock competence and classification 32.8.3 Stiffness of the tunnel lining 32.8.4 Towards a better understanding

32/16 32/16

References

32/9 32/9 32/9 32/9 32/11 32/12 32/13 32/14

This page has been reformatted by Knovel to provide easier navigation.

32/18 32/18 32/19 32/19

Many concerned with tunnelling continue, at their own peril and that of others, to underestimate the need for practical understanding of the behaviour of the ground, the essence of good tunnelling. No two tunnels are the same; experience, and real insight of the value of that experience, are necessary to transmute particular experience to more general understanding and thus to transmit the experience of one tunnel appropriately to another. Advances in tunnelling usually arise not from research so much as from innovations in methods of design and construction. Monitoring of the results in the field may then follow, supported by research where existing knowledge fails to explain the findings. Two essential elements to economic tunnelling are: (1) The tunnel (unless permanently unlined) must be considered as a composite ground-lining structure. Not only does the lining support the ground but the ground in its turn supports the lining. (2) The design of the permanent tunnel must be considered in association with the methods of construction. The overall cost of the process requires to be minimized and the finished geometry is only one of many factors. There are many barriers to a full understanding of the behaviour of the ground around a tunnel: (1) the three-dimensional, time-dependent nature of the problem; (2) the complexity of the stress-strain relations in soft ground; (3) the effects of the initial state of stress, discontinuities and joints upon the behaviour of a rock; (4) the dependence upon the method of excavation; (5) the standard of workmanship and (6) inhomogeneity of the ground. Full-scale tunnels provide the one reliable laboratory for testing theory against practice.

32.1 The options for a tunnel route The ground is the principal determinant of the cost of a tunnel of a given size. For this reason great economic benefits may derive from the capability of selecting a favourable and relatively consistent type of ground for tunnelling. Until the geological structure is known, the object should be to keep the options for a tunnel route as open as possible. For each type of tunnel there are certain geometrical constraints and other specific factors affecting cost. For a road tunnel, for example, acceptable gradients and curves will be related to the design speed and, hence, to traffic costs.1 For a pressure tunnel, on the other hand, there is little direct geometrical constraint and the differential cost of construction in relation to the ground would need to be considered against the capitalized head losses. A general knowledge of the geological structure will indicate whether or not the most direct route conforms to a favourable geological horizon or whether, on the contrary, it may encounter unstable ground such as squeezing rock, running sand, major fault zones, decomposed rock, karstic limestone or similar hazards which may only be penetrated at great expense. Where there is a possibility of adopting an economic method of tunnelling, related specifically to a type of ground with limited variation, there may be the greatest benefit from diverging from the most direct route, in order to situate the tunnel throughout in such ground. At the earliest stage in planning, such factors should be considered so that the options may be described, systematically tested and reduced as information arises from the first stage of site investigation.

32.2 Costs of tunnelling 32.2.1 Principal factors Attempts are made periodically to set out tunnelling costs in a systematized form, with costs per unit length of a certain size of tunnel related to a few generalized ground types and to a few other simplified categories of accessibility and tunnel length. Except for specific areas in which the ground can be reliably depended upon, there is no valid way of expressing tunnelling costs on a simple unit cost basis. From a knowledge of the ground a system of tunnelling may be selected and the costs evaluated on an assumed average rate of progress. The rate of progress may be assessed from experience in similar ground elsewhere, taking account of any innovation in the tunnelling method, and not forgetting the costs of ground treatments or similar ancillary operations. In general, the extent of variability in the cost of tunnelling is increasing for these principal reasons: (1) Tailor-made tunnel systems to suit a particular type of ground permit increasing economies in construction. (2) The cost of labour-intensive tunnelling systems adopted for difficult ground or in congested circumstances will naturally reflect the trend of labour costs including incentive payments. (3) The demand for tunnels in urban development tends to reduce the options available for a tunnel route. As the result of these factors, at the present time there is at least an order of magnitude between the unit cost of constructing the cheapest and the most expensive tunnel of the same size. Hence, there is an increasing benefit to be derived from undertaking studies appropriate to choosing the most economic expedient in each situation. A feature that may be overlooked in comparing the costs of tunnels concerns the means of access during construction. While a shallow urban tunnel or a short tunnel through a hill may be approached directly from the ground surface, long and subaqueous tunnels usually require working shafts and access headings, adding not only to the direct cost of the project but also to the cost of all the consequent tunnelling operations. 32.2.2 Effect of tunnel size The cost of a given tunnel is specific to its situation and its timing, on account of the varying differences in prices, varying local skills and technical capabilities. There is therefore no simple factor to be applied to the cost of a tunnel in order to determine its hypothetical cost at a different place or time. Neither is there a simple formula to determine the cost of a tunnel by consideration of another tunnel in the same ground and conditions but of different size. As a simplification, where variation in size does not entail a change in basic techniques, we may consider each factor in construction as entailing a unit cost U expressed as: U= A +Bd+Cd*

(32.1)

where A, B and C are constants and d is the finished diameter For a highly mechanized system, A will be high, while for a labour-intensive system C will be high. For excavation there will be an appreciable element in spoil disposal costs for which C will predominate while for temporary tunnel supports A and B will be the principal factors. As the size of tunnel is reduced, the increasing congestion leads to reduced efficiency in working. In consequence, there is a

size of tunnel for which the costs will be a minimum; the greater the degree of mechanization, the greater will be the size dmin for minimum cost (i.e. B and C->0 as d-+dmin). For a long length of tunnel in the London Clay the minimum cost is obtained for a tunnel diameter of about 2.5 m while for certain machine-driven tunnels in soft rock the optimum diameter has been found to be about 3 m, and about 2 m for a hand-driven tunnel in hard rock.

32.3 Systematic site investigation 32.3.1 Geological data The scheme for determining the geological conditions should work from the general towards the particular. This will entail a study of geological maps and papers, first on a regional and then on a local basis. In the UK there are normally available sheets at scales of 1:50000 and 1:10000 with explanatory memoirs, produced by the Institute of Geological Sciences. Where geological maps do not exist, aerial photographs often provide useful information on the geological structure. 32.3.2 Objects According to the apparent options for the tunnel the scheme of site investigation may then be designed with these main objects: (1) To test geological data at doubtful points. (2) To explore particular areas of tunnelling difficulty. (3) To obtain information necessary to complement available data on important aspects of geology and geohydrology. (4) To obtain samples for testing and to undertake in situ tests in order to establish the suitability of ground for alternative methods of tunnelling. (5) To determine design and construction parameters. Far too often a site investigation is undertaken without adequate thought to its purpose; in consequence, information vital for good tunnelling is overlooked at the expense of acquiring much irrelevant material. The site investigation should be supervised by those with a direct practical understanding of the associated techniques of tunnel design and construction. 32.3.3 Means A few large boreholes or adits may be justified for direct examination, in situ testing and for subsequent inspection by tendering contractors and others. There is no general rule on the spacing between boreholes. At one extreme, for sedimentary rocks of a uniform character it may only be necessary to be able to establish general continuity of the geological sequence by identification of marker beds or horizons. At the other extreme, igneous intrusions and metamorphosed rocks may present so complex a pattern as to necessitate a tunnelling method highly tolerant to change, however well the ground may be investigated. A good general rule is to establish during site investigation a set of hypotheses, concerning the geological structure and the properties of the ground to be tested so that when a conflicting anomaly is indicated by a new borehole its significance is appreciated, i.e. is the benefit of an additional borehole likely to exceed its cost? Where there is doubt concerning the practicability of adopting a mechanical system of tunnelling, special care is required to ensure exploration of the ground in sufficient detail to determine the feasibility of the scheme.2 Geophysical methods of exploration may serve not only to extend the data from individual boreholes in the second and third dimension but also to reveal specific features such as faults

and igneous intrusions. Without adequate 'fixes' geophysical results may permit widely different possible interpretations. Benefits are usually to be found in undertaking a site investigation in two or more stages, depending on the initial knowledge of the terrain, the magnitude of the project and the diversity of possible options. The investigation should be designed initially to investigate those features most likely to determine the tunnel location; otherwise money and time are wasted on investigation too far from the selected line to be of great value. However, in ground variable to a common pattern, information obtained away from the tunnel route may yet be relevant; the validity of such transference needs careful assessment. Water constitutes a hazard encountered in many forms. The site investigation should, as appropriate to the circumstances, be designed to provide information about water-bearing fault zones, fault zones with a weak filling, open joints and the effect of tunnelling upon aquicludes whose rupture may expose the tunnel to water from aquifers. The geological structure and the possible head of water will control the zone of ground around the tunnel which calls for investigation.

32.4 Tunnelling methods related to the ground 32.4.1 Historical background The history of tunnelling is one of increasing diversification of methods with an increasing capability to explore and to understand the ground. While Brunei used the first tunnelling shield for the Thames Tunnel in 1825-28, tunnels throughout the nineteenth century continued generally to be constructed by means of one of the traditional methods of excavation and timbered support.3 Although these are now largely of historical interest only, the English method, widely and successfully used, sometimes in soft ground and in broken jointed rocks where other methods had failed, merits mention. An essential feature of the English method concerned the use of longitudinal crown bars, supported at the forward end on props and sill and at the rearward end on the last section of completed permanent lining, which might be brickwork or masonry. In this way continuous support was provided to the ground over the tunnel from the time of first excavation and, in principle, the method may be considered as the forerunner of the tunnelling shield. 32.4.2 Shield tunnelling Shield tunnelling is strongly associated with the name of Greathead. He worked with the first circular shield designed by Barlow for the Tower Subway beneath the River Thames in 1869. Greathead designed a shield (Figure 32.1) for the South London Railway in 1886-90 incorporating most of the essential features which have survived to the present day.4 Greathead not only recognized that a shield reduced the risks in tunnelling in water-bearing ground but he was also one of the few of his time to appreciate that it permitted faster and cheaper tunnelling in good ground. The first shield with a mechanical cutting head was the Price excavator used in 1897 for the Central London Railway. Since this time there have been rapid developments, predominantly in Japan, followed by Western Germany, the US and Canada, of mechanical shields provided with means for (partially) balancing soil and groundwater pressures.5 For the most open-textured grounds, sands and gravels, the choice may be between a bentonite (Figure 32.2) or a type of earth-balance shield. This latter may use a fully plated head with controllable

Hydraulic main

Half elevations

External diameter of lining plus 60mm

To hydraulic manifold

Section A-A

Figure 32.1 Hooded Greathead shield with platform rams suitable for 3.5-m diameter tunnel slots or (Figure 32.3), alternatively, the shield may be openfaced and the soil extracted at a controlled rate by an archimedean screw conveyor as the shield advances. For finer-textured ground, a slurry shield may be used with natural clay mixed with water and used for pumping the spoil to the surface. The hydro shield for yet finer ground maintains the face of the shield under hydraulic pressure with a supply of water mixed with the spoil which, again, is transported by pumping. Developments in these directions have incorporated a number of novel features. One particular contribution has been the perfection of seals between

Stones Cyclone Sand

the shield tail and the enclosed lining; another, for the pressure balancing shield, has been the use of an air-vessel in the pressurized face of the shield to dampen pressure fluctuations. All such special shields are designed for limited variability of the ground and contingency measures may need to be incorporated to deal with departures, such as the presence of large boulders. In Japan, developments are proceeding towards full automation so that all operation of advancing and steering the shield, excavating the ground and transporting spoil are controlled from the surface.

Vibrating screen

Clean bentonite Pump Tank

Slurry extraction device Shield

Wet sump

Pump

Pump

Figure 32.2 Bentonite tunnelling machine. (After National Research Development Council)

32.4.3 The bentonite shield It is possible, where frequent access is not required to the face of a shield, to provide for ground support by compressed air, water or mud confined to the face of the shield. Air is not, however, recommended for this purpose. The use of mud in this application offers considerable benefits in permitting an approximately balanced pressure over the full height of the face and in providing a suitable medium for the pumping away of spoil. The first such shield was used in Mexico City,5 utilizing the mud formed from the natural montmorillonitic clay spoil of the area. A true bentonite tunnelling machine was first used successfully in London (Figure 32.2) in sands and gravels.6 32.4.4 Rock-tunnelling machines Many tunnelling machines for rock have been evolved since 1956, although here again the prototype machine belongs to the last century, usually attributed to Beaumont,7 used for Channel Tunnel heading in 1881-82 and subsequently for the Mersey Railway Tunnel.8 There are several features of such machines9 which merit differentiation as shown below. (1) Cutters (see Figure 32.4). For the softest rock the cutters are fixed picks which chisel the rock out as a succession of grooves. For harder rocks, generally in ascending order of hardness, machines make use of single or multiple-disc cutters, toothed cutters, roller cutters or cutters with tungsten carbide insert buttons.

Figure 32.3 Shield with fully-plated head with controllable slots

(2) Cutter heads. For the smaller machines a single full-diameter rotary head is adopted (Figure 32.5). As the machine size increases so there is a tendency to introduce planetary cutters to share the work between cutters and reduce the range between minimum and maximum cutter speeds. (3) Thrust of machine. For the softest rock, machines receive purchase by rams thrusting against a gripper ring expanded against the periphery. With a few exceptions the remainder of the machines obtain their forward thrust by means of diametrically opposed thrust pads jacked against the ground. All such machines advance by periodically withdrawing and repositioning the thrust ring or pads. One machine uses a central pilot drill which is firmly anchored into a hole ahead of the face. Not only does this provide a means for pulling the machine forward but it also establishes a firm forward bearing for the cutter head which may be a valuable feature where rock variation in the face causes uneven loading on the head. The road header, a machine developed originally for mining, has a rotary milling head on a telescopic boom, attached to the body of the machine by a universal joint (Figure 32.6). Thus a typical machine may excavate a gallery up to 4.5 m high and 5.8 m wide. The cutter head usually mounts picks in the pattern of a conical scroll. Operation is by means of a 'sump' formed in the face, extended by lateral pressure on the rotating head. Generally the loading on each pick will be less, and less controllable, than for a full-face machine and hence for the limitations on rock strength for effective application.

Figure 32.4 Types of cutters for tunnelling machines 32.4a 'Series 12' tooth cutter-MNX 32.4b 'Series 12' HHIX cutter 32.4c Bolt-on-disc cutter, type DGX (Courtesty: Hughes Tool Co.)

The selection of a tunnelling machine must take account of the rock types to be encountered along the length of drive. The efficiency of the machine is related to the inherent properties of the rock, principally to strength, the extent of jointing, strain modulus and abrasion. The aim is to minimize the specific energy needed to fragment a rock by keeping the size of particle high and by provoking brittle fracture. The cutter action aims to set up a high local difference in principal stresses and highenough tensile stresses to induce cracking. Each type of cutter and pattern of cutters operates most efficiently in a rock whose properties lie within a limited range. Difficulties may arise from several causes, e.g.: (1) from excessive wear of the cutters in hard rock which grooves instead of fragmenting; (2) from inefficient fracture of rock too soft or plastic for the type of cutter; (3) from excessive wear due to overheating in the presence of high content of silica; (4) from excessive bearing loads where rock varies appreciably in the face; and (5) from jamming of the head where hard rock is heavily jointed and tends to collapse on to the machine or at the face.10 A simple criterion for the economics of machine excavation concerns the cost of repairing and replacing cutters. In sound soft rock this will be found to be a trivial sum in relation to other costs of excavation. For the hardest rock, the cost will be found to climb to a figure of several pounds per cubic metre, with stoppages every few metres for replacement, and this stage represents the present economic limit. The tolerance of a machine to the full range of rock types to be encountered should be considered in weighing the overall merits and costs of its introduction. Another essential question concerns prediction of the need for temporary support close to the face, a process that presents greater difficulty for the full face machines and for which object certain machines make special provision in their design."

Figure 32.5 Cutter head of hard-rock machine (Courtesy: The Robbins Co.)

Figure 32.6 Road header in iron mine (Courtesy. Anderson Mavor Ltd)

32.5 Tunnel construction 32.5.1 Drilling and blasting The traditional scheme of advancing rock tunnels has been by drilling and blasting and this method continues to be generally adopted for short tunnels, hard rock tunnels and for tunnels in variable ground. At the present date, for example, machine tunnelling is unlikely to be economic in shattered rock or in rock of strengths greater than 200 MN/m2. The principle behind blasting in a tunnel is to obtain the greatest 'pull' for the minimum explosive charge and for the minimum damage of the rock around the tunnel. Secondary objectives are: (1) to fragment the rock adequately; and (2) to form a compact stock pile against the face. The pattern of drill holes is designed to suit the rock and the explosive. Cut holes are arranged towards the centre of the face, usually inclined towards each other in order to remove a cone or wedge. One or more central unloaded holes of larger diameter may be used to assist the cut. The remainder of the holes are drilled parallel to the tunnel axis. Delays of a few milliseconds are used between groups of drill holes, from the cut outwards, so that the excavation is enlarged with the travel of the shock wave. Considerable effort has been applied to establishing the neatest periphery to the excavation by the trimming holes, which may be charged or uncharged. In the technique of presplitting, the trimming holes are fired before the remainder with distributed charges to cause cracking around the periphery between adjacent holes. Another technique which has been used in tunnelling is termed smooth-blasting, whereby the line of trimmer holes is required to coincide with the periphery of the excavation, each being loaded with a reduced distributed charge and fired with a short delay after the remainder. It may be well worth considering means for reducing overbreak by careful control of the spacing, line and charging of the trimmer holes. The geometry of the drills or the drill carriage should be designed to permit the trimming holes to be drilled as parallel as possible to the tunnel axis. Care in these respects may show considerable benefit not only in reduction of direct overbreak but also in the reduced extent of the surrounding zone of cracking and displacement of the rock, with consequent savings in the extent of temporary support. Sectional drawings of tunnels have often indicated the periphery of the 'minimum section' and the 'payment line' which allows payment for overbreak to be assessed in relation to the volume of excavation and the volume of concrete lining. Occasionally a 'limit line' is also shown, beyond which a leaner concrete mix may be used for filling. Overbreak in a tunnel is frequently expressed as a percentage of sectional area but, without knowledge of the size of tunnel, this designation has little significance. The present tendency12 is to indicate surface areas of different sizes of tunnel, possibly subdivided into different types of ground, in order to facilitate translation of overbreak into the corresponding additional volumes of ground to be excavated. For small tunnels, hand-operated drills are used on telescopic air-legs. For larger tunnels there is usually a wider choice, including ladder drills, light mobile boom-mounted drills or heavier drills mounted on a jumbo. The latter may provide advantage in controlling the drill pattern and with the speed of drilling, also in protection close to the face for other operations; the main disadvantage arises from inflexibility in the event of departure from full-face driving. For the Mont Blanc Tunnel, for example, it was fortunate that a jumbo was used only from the French end, since difficulties encountered along the Italian drive compelled the enlargement from headings over a considerable length of tunnel.13 A more recent development has been the introduction of the

hydraulic drill, offering a rate of drilling some 50 to 100% greater than the corresponding pneumatic rotary percussion drill, at a considerable reduction in noise level of 10 to 15dB. 32.5.2 Spoil handling The handling of spoil from the face cannot be considered separately from the method of excavation. Mechanical shields and tunnelling machines have built-in chain or belt conveyors loading to a hopper or to another conveyor. The same operation is achieved in a drill-and-blast tunnel by means of a mechanical loader, often with composite face shovel and conveyor. The general trend is to use rail wagons for transport for tunnels up to about 7 m diameter and for tunnels worked from vertical shafts and to use dump trucks for large tunnels directly accessible from the surface or for tunnels at a gradient of more than about 2.5%. Many solutions to the problem of loading rail cars at the face have been adopted. One currently used in relatively small tunnels (say 3 m diameter) is to use a long transit car with an armoured conveyor floor so that spoil loaded at one end may be evenly distributed. Another system for rather larger tunnels (say 4 to 5 m diameter) uses an overhead conveyor capable of loading in turn each of a train of six or more (or fewer) rail cars, preferably to contain the spoil from a complete round. An alternative uses a long sliding platform with rail track and turnouts in consequence maintained close to the working face. Conveyors are also used for dry materials and where access is by inclined shaft. The pulverizing of spoil and its discharge by pipe as a slurry has been adopted for suitable soft rock. Frequently the bottle-neck in materials handling is found to occur at the foot of a working shaft and here mining practice has introduced the use of automatic tipping of tunnel wagons into large hoppers from which shaft skips are rapidly loaded. The entire process of excavation and removal of spoil merits considerable study at an early stage as to its adequacy, with contingency plans to overcome foreseeable causes of breakdown. 32.5.3 Tunnel lining The method of tunnel lining is essentially related to the nature of the ground and to the scheme of excavation. General-purpose tunnel lining has economic application to small tunnels in variable ground. Recent progress and attendant economy have been demonstrated to result from the capability for designing the lining specifically to the condition of the ground and the overall tunnelling system. The first subdivision in type of lining results from whether or not the need exists for an immediate support at the face. In North America the common practice in tunnelling in soft ground has been to tunnel by hand, to erect continuous support in timber sets or steel liner plates and subsequently to place an in situ concrete lining. In the UK, and generally throughout Europe, shields have been more widely used together with permanent primary segmental linings. The traditional lining over more than 100 yr has been the ring of bolted cast-iron segments built within the protection of the tail of the shield, with the external annulus often grouted with lime or cement. Improvements in site investigation procedure have allowed the development of alternative types of lining which can be adopted in certain restricted types of soft ground. Reinforced concrete segments14 have been preferred to castiron segments for reasons of cost since 1938 except where loading is heavy or where watertightness is an essential object. Another general type of tunnel lining is built in rings of segments immediately behind the shield. Each ring is then expanded directly against the ground with elimination of the procedures of bolting and grouting. Evidently the system can

only be used where the ground around the tunnel is selfsupporting over the width of a ring for a short period and thus a certain minimum apparent cohesion of the ground is necessary. Such techniques were developed predominantly in London Clay, sufficiently stiff (i.e. with a low enough stability ratio - see page 32/18) and homogeneous for a specifically designed system.15 Two types of lining based on this principle merit mention. The Donseg lining is created from rings of tapered segments, expansion against the ground being achieved by the process of inserting alternate segments, as longitudinally tapered keys, into the ring by the shield rams (Figure 32.7). This is a highly economic method, limited to tunnels of diameter not exceeding about 3 m, because of the geometry of the lining. For larger tunnels, the Halcrow lining provides for articulating joints between segments. In this way, a part ring of segments may be assembled clear of the extrados. The insertion and expansion of jacks between special segments cause the ring to expand against the ground, accompanied by relative rotation between adjacent segments (Figure 32.8). A special feature of a lining of this type is that secondary stresses are limited to a low level with consequent savings in the structural thickness of the lining. For the Cargo Tunnel at Heathrow11 a lining 300mm thick has been used for a 10.3-m diameter tunnel. One of the most highly developed linings of this type has been developed by Holzmann and used inter alia by Wayss and Freytag for metro tunnels in Antwerp in open water-bearing ground, using a slurry shield and necessitating high standards of water tightness.16 Each ring (Figure 32.9) comprises eight longitudinally tapered segments, the width of the ring itself being tapered so that all segments are built without packings, corrections to line and level being achieved by relative rolling of the ring. This is one example of recent concrete and (ductile) iron

'Don - Seg' segment

152 mm normally 191 mm where cover is greater than about 40 m

Ring of segments as placed by hand Segments after being pushed home by rams Figure 32.7 Donseg tunnel lining rings which depends on extruded plastic seals compressed into recesses to achieve watertightness.17 Steel linings of two basic types are used for soft-ground tunnels. Pressed liner plates with a maximum sheet thickness of about 8 mm serve as a primary lining for hand-driven tunnels.18 Such a lining is inadequate for accepting the thrust from a shield but, for particularly arduous conditions, fabricated steel linings may be used here. These conditions may arise from excessive variation in loading around the lining, on account of the nature

Taplow terrace gravels and brickearth 317-5cm radius Water table

Lifting holes 25x25mm chambers

Jacking spaces packed with earth dry concrete in two stages after stressing, firstly between horns and secondly in recesses following removal of jacks

Top of London clay

Segment Mk 1

317-5cm radius Lifting holes 25x25mm chambers Jacking recess

Jacking recesses Jacking recesses

Three-dimensional view of jacking space

Segment Mk 3 Arrangement of segments in ring (arrangement symmetrical aboul vertical centreline) Figure 32.8 Lining for cargo tunnel at Heathrow Airport, London. (After Muir Wood and Gibb (1971) 'Design and construction of the cargo tunnel at Heathrow Airport, London.' Proc. lnstn. Civ. Engrs, 48,11-34)

Detail of neoprene gasket

Fixing and lifting details omitted

Extrados Recess for neoprene gasket lntrados Longitudinal section of segment on A -A Figure 32.9 Precast lining with neoprene gasket seal of the ground, low top cover, confined side clearance or proximity to foundations. Several types of flush lining have been designed for initial erection around a central spider but these are suitable only for small-diameter tunnels. For the Mersey Tunnels 3A and 3B, lining segments (Figure 32.10) were made in mass concrete with an internal steel face.10 Each ring is attached to the previous ring by means of long bolts inserted into threaded sleeves. Waterproofing of the lining is achieved by welding cover plates across the joints. The concrete expanded linings result in a flush interior surface, which may be beneficial for tunnels serving as conduits. Where a lining is built in any but very weak ground, bolting between segments has no permanent structural significance. Fastenings are therefore required primarily to control shape during erection prior to filling the extrados, the space between ground and lining. One effective cheap system uses tapered elm dowels to achieve alignment of adjacent circumferential joints while each radial joint is located by a longitudinal tube in semicircular channels; the tube collapses under load to avoid excessive local pressures. Ductile (spheroidal graphite) iron has been used for tunnel linings. While this material allows a considerable saving in weight by comparison with grey iron, the reduced depth of segment is a disadvantage for obtaining purchase for the thrust rams but generally such linings are found to offer economic benefits as alternatives to steel linings where high loading and appreciable tensile stresses are expected. Tunnel segments are erected in rings and the width of the ring determines the stroke of the propelling ram and hence the length of the shield. In the UK the tendency has been, in soft ground,

to maintain tunnel linings to a width of no more than 70 cm while on the Continent segment width is generally greater, with 1 m as a common standard and this trend is generally extended as new shields are built. Rock tunnels are usually lined in situ with concrete placed behind shutters. The lining may be cast in discrete lengths or continuously behind shutters travelled forward in a retracted mode. Concrete is usually pumped, with placers used with decreasing frequency for filling the crown. Subsequent contact grouting is usually necessary to fill shrinkage cracks and voids between lining and rock. For many years, attempts have been made to form a continuous in situ lining immediately behind a shield in soft ground. Success requires synchronization of advancing the shield and filling the concrete annulus. This has been achieved by Hochtief, Holzmann and Wayss and Freytag in Hamburg for the metro. The shield thrust is transmitted through internal shutters to avoid pressure on newly placed concrete.19

32.5.4 Thrust boring Thrust boring of tunnels20 has developed from pipe jacking, whereby lengths of steel pipe are pushed through the ground, from a jacking pit, with the addition of a new length of pipe at the rearward end after each extension of the jack. Thrust-bored tunnels are frequently in the form of reinforced concrete elements or layered materials incorporating fibre reinforced plastic. The limiting distance of thrust boring depends upon the ground, the geometry of the tunnel and the capacity of the jacks. This may be extended by the use of an external lubricant such as

TV camera Wig-wag sign Segments above deck level break joint by half a segment as indicated by dotted lines

Grouting space

Tunnel lighting Wig-wag sign All joints made watertight by continuous welding of mild steel cover strips C02 detector Visibility detectoi

Caulking groove Secondary lining Fireproof or electrical distribution cabinet Horizontal axis

Grouting space Holes for longitudinal i tie rods

Secondary lining Inspection walkway

Fire main valve 152 mm fire main Extent of mild steel skin Side entry gulley Fresh-air duct 152mm dia. road drain

305mm thick precast concrete segments 1-22 m wide with 6 mm thick mild steel skin anchored to inner face with 13mm dia. hoop studs

7-32 m carriageway Crossfall Prestressed precast concrete beams filled with in situ concrete

203 mm dia. pumping main from Wallasey portal to mid-river sump

•Extent of mild steel steel skin Fresh-air duct

Figure 32.10 Cross-section of Mersey Kingsway Tunnel. (After McKenzie and Dodds (1972) 'Mersey Kingsway Tunnel: construction' Proc. lnstn. Civ. Engrs, 51, 503-533) bentonite or by using intermediate jacking points to control the maximum length to be advanced at a time. For small pipes excavation is often by continuous-flight auger; for larger tunnels, excavation may be by hand or by small mechanical excavator. Various types of cutting head or shield are used with jacked tunnels which, for longer drives, may be designed to help correct errors in alignment. The extent of support to the face will compare with that necessary for shield tunnelling in similar ground. Thus, very weak silt may be extruded into the tunnel through a ported head. The most variable features of jacked tunnels concerns the joint between elements. Traditionally, for a concrete lining, a spigot was used with clearance between the internal parts of the joint which would subsequently be sealed. Such an arrangement reduces the surface area available to transmit thrust. Several types of flush external sleeve are now used in association with a butt joint, usually associated with an external annular seal to exclude the ground. For relatively short lengths of tunnel through soft ground thrust-boring offers the benefit of erecting all lining at the thrust pit directly accessible from the ground surface, in lengths of 2 m or more, thus reducing manufacturing costs and the aggregate lengths of joints to be sealed. Furthermore, in weak ground the pipe form provides improved circumferential strength. Tunnels may be jacked in continuous easy curves using tapered pipes (or tapered packings) at the expense of increased thrust. Experience shows that a well designed and engineered jacked tunnel, built

to fine tolerances, considerably reduces thrust loads and in consequence extends the total length of tunnel, or spacing between intermediate jacks. Measurement of the build up of thrusts in the initial period of jacking will help to establish appropriate spacing between jacking stations.

32.5.5 Waterproofing The availability of new sealing materials provides a wide choice of waterproofing systems for the joints between preformed tunnel elements.23 Selection will normally be on the basis of cost and durability, to meet particular criteria concerning: (1) Capacity to tolerate relative movement between elements. (2) Hydraulic pressure. (3) Application to wet surfaces and under pressure. The first barrier is normally provided by annular or contact grouting. Thereafter there are fundamentally three choices: (1) a sealant provided in a liquid or plastic state; (2) a material caulked into the joint space; and (3) a preformed gasket compressed between elements. The latter has found wide acceptance as a high-performance seal used with segments cast to fine dimensional tolerances. Where practicable, seals should be formed at a radius beyond that occupied by bolts or other fastenings so that these do not require separate treatment for the exclusion of water.

32.5.6 Temporary support In rock tunnelling, the permanent lining cannot be considered separately from the scheme of excavation and temporary support. The initial stability of the excavated ground depends not only upon the inherent quality of the rock but also on the method and quality of the excavation process. Generally, mechanical excavation will not only provide a better shaped arch around the tunnel but, more important, also much less disturbance of the surrounding rock. Recent studies have indicated that blasting may cause cracking of the rock up to a diameter outside the tunnel. The essence of good tunnelling in jointed rock is to provide adequate support to incipiently collapsing rock as soon as possible. The means for achieving this end are directly related to the nature of rock and its jointing. The situation may be summarized thus: (1) Where the rock is highly shattered or with frequent open joints, effective support may require the use of heavy arches. These must be provided with adequate foot supports to avoid punching into the invert and must be blocked off the rock sufficiently frequently to avoid excessive bending stresses.22 One means of achieving an even or virtually continuous blocking is by the use of porous bolsters placed behind the arch into which a weak element/flyash grout is pumped.23 Arches of the yielding type,23 designed originally for colliery support, are now widely used in tunnels in recognition of their ease in erection and the virtual equivalence of their major and minor second moments of area, and hence greatly reduced tendency to distort, in conditions in which their higher cost may be justified. (2) Where the rock is subject to progressive deterioration or to surface weathering, an immediate application of concrete or mortar may provide great benefit. A thin application of pneumatically applied mortar (gunite) or fine concrete (shotcrete)24 will often serve in this respect, applied preferably to enter open crevices between blocks so that an adequate arch is provided around the tunnel. Shotcrete is frequently reinforced with steel mesh attached to the rock face by rockbolts or pins. Alternatively, the shotcrete may be applied with a wire staple or fibre reinforced content. A somewhat heavier and more expensive version with the same general object may be provided by an initial concrete lining placed against the newly exposed rock, possibly behind perforated steel sheeting supported by arches.26 There are often great advantages in the reduction of overbreak if support of this nature can be applied so close to the face as to receive benefit of the three-dimensional dome that occurs here. There is also a certain time dependence of the tendency for collapse from a tunnel roof; thus a great deal of the barring down of an unstable tunnel roof can frequently be avoided by immediate support. (3) The action of rock bolts in supporting the ground around a tunnel depends upon the nature of the jointing.27"29 For a regular pattern of sets of joints, the areas around a tunnel arch may be identified from which unsupported blocks may tend to fall or slide. Rock bolting will be designed in a regular pattern to create a reinforced rock arch, taking account of the strength characteristics of the joints or the stress-strain behaviour of the rock mass, where unacceptable rock convergence may otherwise develop. Special circumstances for rock support may arise from: (1) High horizontal ground stresses, recognizing the need for appropriate disposition of support. (2) Strongly laminated rock, for which rock bolts may serve to tie together the laminations.

(3) Presence of dominant pattern of open joints, necessitating great care in design of support. (4) Weak filling of joints which may further weaken or be eroded as a result of tunnelling. Large blocks of rock adjacent to rock caverns, bounded by joints of low strength, have called for special measures of anchorage in order to ensure stability, by means of anchored tendons and cables. There are many types of rock bolt but these may conveniently be considered in two groups: (1) those which rely upon end anchorage, usually by some method of mechanical expansion of the end of the bolt, and (2) those which are keyed along their length. The latter type may be deformed bolts, set in cement or in epoxy resin or similar adhesive. The cement is introduced either as a mortar introduced in a split expanded metal cage, or as a grout through a perforated bolt or a separate tube. The epoxy resin is usually in the form of cartridges inserted ahead of the bolt, with twisting of the bolt used to burst the cartridges and mix the two-part resin. The form of keying depends, inter alia, on the ability to drill true regular holes in the rock. Where this is possible a bolt in the form of a hollow split sleeve may be driven into the rock. The end anchorage bolt is generally the cheaper expedient and is more readily stressed but, in soft or weathered rock the anchoring should be achieved by a resin bonding. The head of the bolt should be fitted with plate washers or a short length of channel to spread the load adequately over the surface of a soft rock. Progressive failure of a jointed rock may be controlled by a wire mesh between bolts acting as a containing cage which is also a useful safety measure. Evidently the effective depth of a bolted rock arch or slab depends upon the bolt size, length and spacing, the length usually requiring, for overall economy, to be twice the spacing or more.28 Rock bolting and shotcreting are often used in association, the former providing the major support, the latter controlling surface deterioration without which aid the bolts would be effective for a short time only. Surface cracking of shotcrete provides early warning of continuing movement of the ground. Dowels generally represent reinforcement placed in the rock without tensioning. They may be preferred to bolts in the following circumstances: (1) where subsequent movement of the rock will suffice to stress the dowel but would be otherwise liable to overstress or dislodge a prestressed bolt; (2) where light (possibly temporary) support only is required; and (3) where subsequent tunnelling or mining will excavate through the supported area, favouring the use of wooden (bamboo) or reinforced plastic dowels. Spiles are driven into the ground, ahead of the face or around the tunnel periphery, providing support by shear stress mobilized along the spile. In consequence, spiles are frequently of rolled steel sections providing a high superficial area per unit weight. They may be driven obliquely ahead of the tunnel face, first to support the face itself and, subsequently, by successive redriving as the face advances, to support the periphery of the tunnel. There has been much development in recent years in tunnelling with support designed to reinforce the rock (or stiff soil) such that minimum applied support is required to achieve stability. Such support is often in the form of rock bolts and projected concrete (shotcrete) but may also include the use of grouting, compressed air and similar expedients. The principle of such a method is that observations should be undertaken to confirm that the support is adequately stabilizing the surrounding grourld, so that the rate of convergence towards the tunnel is perceived to be approaching an asymptotic value. Where, after initial support, this is not assured, further support may be applied incrementally. Thus, the principle of such a system may be described as that of 'incremental support5. The most widely

known application of this approach is the New Austrian Tunnelling Method (NATM). Confusion has arisen by the incorrect application of the term NATM to all forms of support by rock bolts and shotcrete or, more widely yet, to forms of tunnelling which do not adopt formal linings. The essence of design of a system of incremental support is to understand the stress-strain properties of the ground and of the tunnel support since, essentially, controlled strain of the ground has to occur in order to develop a changed stress field in the ground around the tunnel compatible with the degree of support. 32.5.7 Advance by full face or by heading A first consideration in excavating a large tunnel concerns the practicability of full-face excavation. This will depend upon the stability of the rock in relation to the tunnel size and upon the need for any advance heading to explore the ground and to provide an opportunity for undertaking ground treatment ahead of the main excavation. In very large tunnels it may be economic to excavate a top heading first in order to insert supports for the crown and then subsequently to work the invert section as a vertical bench; a variation to such a method may utilize a bottom heading in addition, serving for drainage and for removal of spoil.30 In swelling ground (i.e. in rock containing a montmorillonitic clay) the difficulty in support may be roughly expressed as proportional to the area of the tunnel. In consequence there may be considerable benefits in utilizing a series of small headings or drifts around the periphery of the tunnel in which the permanent lining for the full arch and invert is cast section by section.

32.6 Aids to tunnelling 32.6.1 Compressed air As a tunnel advances, relaxation of the ground in the vicinity of the face will induce dilation. In fine-grained soils this can occur only at the rate at which water can be drawn into the soil. As the soil dilates, effective stress between the grains is reduced and the soil may flow or ravel. The period during which the face remains stable is known as the stand-up time and, in any particular circumstances, the dominant controlling features are the soil permeability and swelling modulus. A first aim of any one of the several aids to tunnelling will be to extend the stand-up time. The application of the use of compressed air to soft ground tunnelling is another development associated with Greathead and the South London Railway (1886-90).4 In soft clays, compressed air will provide direct support to the ground. In silts and sands the compressed air displaces the greater part of the pore water and causes cohesion between grains of the soil by surface tension. The effect allows running sands to be treated in excavation as a soft rock. Another sideeffect of compressed air in the ground is to reduce its permeability to the flow of water (by as much as an order of magnitude for silts). The use of compressed air necessitates a considerable outlay in low-pressure compressors, air coolers, air locks (including a medical lock) and the associated control and monitoring system. Even a momentary loss in air pressure might have fatal consequences and, hence, the need for a high degree of duplication and standby equipment. The working conditions in compressed air owe a great deal to pioneering studies by Professor J. S. Haldane, leading to a set of recommendations by the Institution of Civil Engineers later revised and issued as a set of regulations under the Factory Inspectorate.31 Comparable standards have

been evolved in other countries which undertake tunnelling or caisson work in compressed air. Compressed air introduces increased direct and indirect costs, the latter arising from the reduction in effective working time and the period spent in 'locking out' which may for instance increase from about 25 to 45 min for a 6-h shift as the working pressure (measured above atmospheric) rises from 1 to 2 atmospheres. The upper limit, without special air mixtures, is about 3 atmospheres. It is recognized that it is necessary for strict medical supervision to be provided for workmen in compressed air.32 For many years it has been known that the amount of nitrogen dissolved in the blood is related to the period of exposure to a given pressure so that if the pressure is lowered too rapidly bubbles are formed, particularly at the joints, leading to the condition known as 'the bends'. More recently, compressed air has become associated with a more serious complaint, that of bone necrosis, which may leave the victim crippled. One reaction to this discovery has been to resort to other forms of aid in order to dispense totally with the use of compressed air. A more reasonable attitude appears to be to discover the causative process and to eliminate the offending factor, since alternatives for compressed air may not only entail high cost but also introduce new hazards. In the past, before bone necrosis was associated with compressed air, medical inspection of workmen passed them fit to work in compressed air without attention being given to any latent defect of the bones or joints, an oversight that should not recur for future compressed-air working. Compressed air has been used for many subaqueous tunnels in soft ground. The problem of balancing the external water pressure increases with the depth, as well as the size, of the tunnel. The depth below the water surface and the texture of the ground will determine the quantity of air required. In coarse sand a rule of thumb for determining the maximum demand in relation to losses through the face has been stated as 7.5Z)2m3/ min where D is the tunnel32 diameter in metres. To face losses need to be added losses through the lining and through airlocks and bulkheads. Lining losses, in the absence of special care in sealing and caulking, can for a long length of tunnel represent a high fraction of total losses. Where the ground comprises clay interbedded with thin layers of sand or silt it has frequently been found that a relatively low ratio between the pressure of air and the external head of water is adequate to provide greatly improved stability to a tunnel face. In open ground, air losses may be reduced by locally sealing the exposed face, for which purpose bentonite dust has been used. A further problem area arises, in the construction of a segmental tunnel lining behind a tunnelling shield, in the avoidance of collapse of the ground on to the lining immediately behind the tail of the shield. This has been countered by grouting with bentonite through the skin of the shield in order to increase the capability of supporting the ground by compressed air. An alternative method has been to fill the annular space with pea gravel as the shield advances, by no means easy to perform satisfactorily.

32.6.2 Ground treatments A wide choice of grouting media33 is now available for consolidating weak or water-bearing ground: (1) Setting grouts containing cement, bentonite, fly ash and other materials may be selected, at the lowest cost compatible with adequate travelling capability for the dimensions of pores and joints to be filled. Bentonite may also be used on its own as a lubricant for the extrados of the shield skin,

for thrust-bored tunnelling and for shaft sinking. Bentonite mixtures are thixotropic, i.e. they form a gel in the absence of shearing motion. (2) Chemical grouts are used in medium to fine sands, single chemical systems having a time-dependent control of setting and two-chemical systems, of which the Joosten is the most familiar process, depending on contact between the two components. (3) For silty sands, resin grouts may be used, low viscosity grouts being available for permeabilities down to about 10~5 m/s. Generally, the finer the ground and the lower its permeability the more expensive the grouting process. The principle in grouting variable ground is therefore one of working through the available grouts from cements, clays, chemicals and resins as appropriate, so that the cheaper grouts are used to confine the travel and hence the 'take' of the more expensive grouts (Figure 32.11). In fine material, electrochemical grouting may be used in tunnels in the future. It has already been used for foundations.34 In this process, electro-osmosis accelerates the rate of penetration of the chemical agent through the ground.

Clay cement and silica based grouts

2 13m i.d. Pilot tunnel

Silica based grouts Resin grout

Gravel

5-6£m pilot tunnel Clay

858m main tunnel Silty fine sands of Woolwich and Reading beds

Figure 32.11 Pattern of ground treatment for north end of second Blackwall Tunnel, London (J. lnstn Civ. Engrs, 35, 19, October 1966-with acknowledgement to the Council of the Institution of Civil Engineers) 32.6.3 Freezing Freezing has had a longer history as an aid to sinking mine shafts than for civil engineering applications.35 It has been used in civil engineering works predominantly for situations of unusual difficulty and for installations using vertical freezingholes sunk from the surface. In each hole is inserted a U-tube or, alternatively, a composite freezing tube with concentric inner and outer tube through which cold brine is circulated, usually down the inner and up the outer tube. The brine is usually used at a temperature of about - 2O0C but this may be reduced to -350C. Freezing was widely used for construction of the Moscow Underground in the early 1930s, for vertical shafts and for inclined escalator tunnels. For several lengths of recent sewer tunnel in Germany freezing has been adopted with horizontal freezing holes. A freezing operation with the use of brine usually occupies several weeks after the installation of the tubes and equipment. Freezing has been used for six shafts for the Ely-Ouse water

scheme,36 each requiring control of groundwater to a depth of 25-65 m below its surface. The cost of freezing (1969) was about £560/m for a 4.5-m internal diameter shaft and £740/m for a 7.5-m internal diameter shaft. A new development has entailed the use of liquid nitrogen as the freezing agent. Since the operating temperature may then be lowered to — 15O0C the freezing operation occurs rapidly and the process has frequently been used for penetrating relatively thin bands of water-bearing ground during the sinking of shafts. Freezing by liquid nitrogen has been used in tunnels in Switzerland and South Africa in conjunction with shotcrete, to provide a secure, if expensive, temporary support with low subsidence in weak ground. 32.6.4 Dewatering Control of water for tunnelling may be achieved by lowering the water table by pumping or by diverting the water as a tunnel is lined. The first requires no further explanation here beyond the observation that pumping continues to be adopted in association with urban tunnelling with inadequate appreciation of the risks of settlement to adjacent buildings, particularly where organic soils are concerned. In certain circumstances recharging wells may be used to control the extent of the depression of the water table. If the water is permitted to flow freely into a tunnel there may be a risk of ground settlement but this is not generally an important consideration in rock tunnels. Exceptions to this general rule occur in crushed or altered fault zones, where weak joint filling may be softened or washed out or where the rock is incompetent in relation to the pressure of groundwater. Particular care in controlling water is demanded where weak, jointed, rock is associated with stronger rock serving as aquifers. Gypsiferous rocks in the presence of water may continue to swell over a long period. Provisions may be made to permit continued swelling without excessive pressure on the tunnel; another expedient may be to exclude water from the area by sealing or drainage. In the Seelisberg Tunnel, Switzerland,36 such expedients were employed conjointly. While major flow of water will require to be controlled in consideration of pumping capacity and deterioration of the ground, minor flow will only present problems for the lining operation. Over a period of many years several expedients have been devised for the diversion and control of water to allow placing of the lining. One of the successful methods has been to provide a continuous protection of plastic sheeting around the tunnel supported on panels of steel mesh with longitudinal french drains along each side of the invert which are grouted up as a final operation. An alternative arrangement, where water flow is general but not great, will use a quick-set mortar pneumatically applied on to steel mesh with pipes inserted at intervals to concentrate the water flow. The pipes are stopped off on completion of lining or, occasionally, allowed to flow where permanent drainage and pressure relief are intended. Where water is confined locally to joints it may be adequate to form a stopping of flash-set mortar around a flexible tubular former to provide a drainage path.

32.7 Ground movements Excavation for a tunnel may give rise to associated ground movements for two principal reasons. These may either be caused by over-excavation, leaving cavities beyond the space occupied by the lined tunnel, or by release of original stresses in the ground, giving rise to elastic or plastic deformation towards the tunnel.

In rock, over-excavation may occur from roof collapses or from failure to line solidly against the ground. Rock falls may develop domes, arches or chimneys depending upon the nature of the ground and the pattern of initial stresses. High horizontal stresses, for example, will normally tend to limit the extent of the cavity provided the rock is sufficiently competent in relation to the maximum resultant stress around the periphery of the tunnel. Crush zones in a homogeneous rock around a tunnel may indicate high stress; the same phenomenon occurs in the release of strain energy by rock bursts when thin slabs of rock become violently detached from the periphery, in strong rocks at depth and in weaker rocks nearer the surface. As rock fractures it increases in bulk; in consequence, once a plug is provided at the base of a cavity its upward extent will be limited and may be approximately calculated. Even small cavities immediately behind the lining are serious in that they may lead to uneven loading on the lining and consequential failure; hence the need for systematic contact grouting. In soft incompetent ground, over-excavation will usually be transmitted in full to the surface approximately to equate to the volume of surface settlement. However, in dense sand the total settlement at the surface will be reduced; in loose sand, settlement may occur as a result of disturbance by tunnelling even in the absence of any over-excavation. It is often impossible in soft ground to subdivide the effects of overexcavation and of changes in the stress pattern, the latter tending to give rise to loss of ground towards the exposed face. The shape of the 'trough' of settlement at the surface will usually be influenced by loss of ground along a length of tunnel somewhat greater than its depth below surface, the influence factors being highly dependent upon the geological structure. In homogeneous soft ground and for a tunnel advanced with consistent standards of design, workmanship and progress, a characteristic depression will develop over the tunnel which may be described approximately in terms of the shape of statistical normal distribution curves.37 Approximately half the total settlement will have occurred immediately above the advancing tunnel face, for tunnels at no great depth. Tests undertaken during the construction of tunnels in London Clay indicate that with increasing depth, there is a greater tendency for loss of ground arising from deformation towards the advancing face. In general, the contribution to loss of ground may be as set out in Table 32.1. A special cause for settlement over a tunnel may occur where a shield in soft ground can only be kept to correct line by means

of maintaining an appreciable 'look up' on account of a tendency to settle at the cutting edge. This loss may be countered to some extent by grouting above the shield as it advances, with fly ash or similar material.

32.8 Tunnel design 32.8.1 Stresses around a tunnel The state of stress in real ground around a full-size tunnel during the course of construction is too complex to analyse fully. A more rewarding process is to idealize the problem to a certain degree and then, by inference and judgement, determine the significance of inadequacies of the conceptual model. We start by considering the two-dimensional problem of a long unlined circular tunnel pierced instantaneously at great depth in perfectly elastic ground. We can in this instance build up the overall stress pattern around the tunnel by superposition of its constituents.38 The initial vertical loading will be redistributed and will set up the tangential and radial principal stresses ae and ar shown in Figure 32.12 for the vertical and horizontal axes. At the periphery,