7.9 Lining of pressure tunnels and hydrofracturing potential. D.U. Deere! / G. Lombardi? ul Summary Pressure tunnels as clements of a hydro- power plant represent an important share of the total investment. ‘A.trend to cheaper solutions is therefore de- veloping. It consists in simplifying, reducing or eliminating the lining of the tunnels. In doing 0, important desiga tules were disregarded in some occasions leading to unpleasant events. In the following, these rules are recalled, thus putting special emphasis on the question of hydrofraccuring. Introduction A ctitical clement of hydroelectric projects with remotely located powerhouses is the pres- sure tunnel that conducts the water from the in- take structure at the reservoir to the turbines in the powerhouse. The construction of the tun- nicl and its lining ate major cost items of the proj- ect and any delays in the construction can place the tunnel lining on the critical path for power generation. Thus, for economic and time reasons there are incentives to reduce, of to eliminate alto- gcther, the lining from some lengths of the tun- nel, Such economy may be false, however, be- cause the cost of providing originally and ade- quate lining may be much less than the cost of, 1 Consulting Engineering Geologist, Gainesville, Horida, usa’ 2 Consulting Engineer. Locarno, Switzerland. repairing a lining that has failed during opera- tion considering the value of the lost power and the social and political costs (1). Moreover, avail- able sites in recent years often have presented poor geological conditions and, in some cases, with increasingly higher hydraulic heads in the tunnel, ‘A number of failures have occurred over the past decade with severe cracking of concrete Tinings and wich escape of tunnel water (1.2). Some of these appear to have been associated with hydrofraccuring of the adjacent rock (hydro- splitting or hydrojacking along a bedding plane ‘or joint). In the present paper this phenomenon is examined both from the theoretical and practi- cal viewpoints. Preceding that discussion is a brief review of types of linings, and their roles in pressure tunnels. Types of Linings Classical Solutions No lining. Many existing tunnels have no lining at all although they may be locally sup- ported by rock bolts. Weak zones will be sup- ported by bolts, mesh, and shoterete, or by steel ribs and concrete. Deere (1) notes the desirable ‘gcotechnical characteristics for an unlined tunnel: — Massive, good quality rock — Absence of erodible or soluble material — Substantial rock cover = High groundwater level — Low rockmass permeability122 Buen and Palmstrom (3) in their 1982 pa per state, ‘The solution with unlined pressure shafts and tunnels is both cost — and time saving. Unexpected adverse geological conditions have, however, in the past been the cause of ex- tensive repairs and loss in production in several hydro power schemes. This proves the necessity of thorough planning and control during all phases of construct: Reinforced shotcrete lining. Shotcrete with or without steel mesh reinforcement, and. with of without rock bolt support, are commonly em- ployed as initial support duting tunnel excava- tion, For use as the lining of a pressure tunnel, however, both the rock bolts and the mesh rein- forcement of the shotcrete should be manda- tory. Shotcrete is a semi-permeable lining with the permeability dependent upon the number of shotcrete layers, the quality of the mix, and the workmanship ‘of the application. Plain concrete lining. Without doubt the most common type of lining for pressure tunnels is the unteinforced concrete lining. This does not imply that it is necessarily the best type for all conditions. Such lining is subject to /ongttu- inal cracking (usually a single crack along either the horizontal of vertical axis) due to radial dis- placement when the tunnel is fitst pressurized if the rock modulus is too low for the applied inter- nal head. Concrete linings may also develop sransverse (circumferential) cracks on 3-m to 9-m spacing, and occasionally longitudinal cracking as well, due to thermal shrinkage as the fieshly placed concrete cools down to ambient from its ‘maximum setting temperature. Thermal cracking can be reduced or eliminated by using a cement with low heat of hydration, a low cement con- tent (225-275 kg per m?), cool aggregates and cool mixing water (or ice), and reasonable lining thickness (30-35 cm). Backfill grouting of the crown voids and high pressure consolidation grouting are both required. All honeycombed concrete, cold-joints, shrinkage cracks, and construction joints should be repaired with mor- tat ot epoxy resin before first filling. While a concrete lining is considered a semi-permeable lining, its permeability can range from low to high depending on the care taken in design and con- struction. The tensile strength of a lining should not be counted on, in taking part of the internal pressure, because much of it has already been used in the thermal cooling: the concrete lining ‘may contain many incipient longitudinal cracks D.U. DEERE/G. LOMBARDI (e.g. along construction joints) even before the tunnel is pressurized. Also, cold reservoir water can induce some additional thermal shrinkage. Plain concrete is in fact a brittle material. Reinforced concrete lining. A reinforced concrete lining is more ductile than one of plain concrete, the ductility increasing with the per- centage of steel. The steel reinforcing distributes the crack and they are much thinner than che single crack of a plain concrete lining. Because the water loss is proportional to the second or third power of the crack width, the leakage through numerous thin cracks in the reinforced concrete lining (of pethaps 0.1.0.3 mm widths) could be one to two orders of magnitude less than through a single 5$-mm crack in a plain con- crete lining. All of the cautions given for design and construction of the plain conerete lining also apply for the reinforced lining. With these pre cautions, the reinforced concrete lining can be considered a slightly permeable lining, appro- priate for use in soil and weak tock tunnels, in zones of low rock cover, and as transitions be- tween sections of steel lining and plain concrete lining. For high heads in low quality, permeable rock, very heavy reinforcement could be required. Steel lining. A steel inner-lining with 30cm or s0 of backfill concrete between the steel lining and the rock walls of the tunnel can be considered an impermeable lining (providing that the grout-hole caps are sealed tightly). Steel thick- ‘esses can vary greatly commonly 12 mm-30 mm) depending upon internal head, rock modulus, tunnel diameter, and reinforcing bands. Stee! lin ers are used at the powerhouse end of pressure tunnels where hydraulic pressures are high and the rock cover often low. Steel linings ate frequently required along the tunnel alignment in arcas of permeable ground, water-sensitive rocks or low cover where water losses are unacceptable. Freestanding steel pipes may be used in some cases instead of a steel lining. Novel Solutions Prestressed concrete lining. While not com- mon, prestressed concrete linings are gaining some popularity, especially in Europe, as they are only slightly permeable and can be thirty percent cheaper than steel. The tangential pre- stressing develops compressive stresses in the lining that offsets the thermal stresses and theLINING OF PRESSURE TUNNELS AND HYDROFRACTURING POTENTIAL 123 tendency for longitudinal cracking induced by the radial expansion of the lining under internal tunnel water pressure. Various prestressing tech- niques have been used: tensioning by hydraulic jacks of embedded, anchored cables such as the ‘VSL method (4); high-pressure grouting, simul- taneously, of several adjacent rings of radial grout holes (5); and gap grouting of perforated tubes uniformly placed around che circumference between rock wall and concrete lining (TIW AG) (9). The latter method was employed in the Drakensberg pumped storage scheme recently constructed in South Africa (10,11). Composite membrane linings. Secber (9) describes the use on several projectS“of imper- meable seals (sheets) attached to the rock or to the shorcreted surface, that are held in place by a cast-in-place concrete lining. The impermeable seals may be of plastic (2-5 mm PVC) or pol- yethylene with a fabric o¢ a perforated steel sheet backing) on of thin steel lining (4-5 mm) for higher pressures. The smooth walls of machine- bored tunnels are ideal for this procedure. Prestressing of the concrete lining is done by high-pressure grouting through circumferential ‘grout tubes between the rock and the sealing material (TIWAG gap grouting). Seeber (9) be- lieves this syseem can replace part of the thick- walled steel lining and serve as a transition to the concrete-lined section. Also, precast concrete pipes with an external thin steel sheet and a backdilling of plain concrete have been used. Role of the Lining of a Pressure Tunnel Hydraulic Requirements The hydraulic requirements for the lining are well understood: 1) To conduct the water from the reservoir to the powerhouse with a reduced head loss; 2) To prevent, or strongly limit, seepage loss of the water being conducted. Both concerns involve loss of energy and revenue, Engineering studies involving the geo- logical conditions, groundwater level, construc- ‘ion costs, and eneigy losses determine the optimal diameter, type of lining, and required rock treatment. The possible consequences of water seepage on potential piping, saturation, and landsliding of adjacent slopes must be evalu- ated. Structural Requirements ‘The structural requirements are to support, to protect, and to resist. The lining must aid the initial rock reinforcement in supporting the longterm rock loads and external watet pressure. ‘The lining must protect the initial support from cortosion and the rock from erosion, piping, and solution. And, when the tunnel is pressurized, the lining must help the rock in resisting che dif- ferential pressure between the pressurized cun- rel water and the external groundwater without appreciable cracking and leakage. Desiga Philosophy ‘The desiga must take care of all aspects of, the hydraulic requirements and the structural re- quirements. There are a variety of lining types to be considered. Each solution has proper meth- ods of computation, design, and construction 10 be followed, Key geotechnical aspects are the position of the groundwater level (which may have become temporarily lowered during tunnel excavation); the modulus of the rock and the distressed zone following grouting; and the per- meability of the rock mass. Each solution has, therefore, its own field of applicability and sometimes various solutions ‘may be equivalently well suited to a given situa- tion. In the next section attention is given to a particular type of failure mechanism chat has not always been considered in the past. The Problem of Hydrofracturing General As already mentioned, 2 number of acci- dents have occutted to pressure tunnels that may be tclated to a hydrofracturing of the rock mass. ‘The escape of large water flow has caused secon- dary damages in addition to the economical loss of water or energy. To eliminate the risk of hydro- fracturing has to be considered a very fundamen- tal task for the engineer involved in hydropower desiga. ‘According to some authors, hydrofracturing should be the event that produces fractures in a sound rock while hydrojacking should be just124 the opening of existing crack or joints due to high pressure water. Asa conservative rule, one should consider that cracks will or can exist in ‘any rock mass so that only the case of the open- ing of existing discontinuities may in fact be of concern in the construction practice. Hercin, this phenomenon is also simply called “‘hydro- fracturing”. Jointed Rock Mass ‘As a rule any rock mass has to be consid- ‘red, at least from a practical viewpoint, as more or less pervious. The permeability is obviously due to the existing cracks, fissufes and joines while the rock itself often can be considered im- pervious. Any water pressure gradient will there fore produce a flow through the rock mass. This flow is harmless as long as the water pressure is relatively low, and no opening of the joints is produced. But as soon as the pressure is high ‘enough to open the joints, the permeability and therefore the flow of water increases. Depending on the configuration of the problem, the in teased flow may lead to an increase of the inverstitial water pressure. In this case the process is self-feeding and leads to a hydrofracturing o a hydrojacking of the rock mass with a large unacceptable escape of water, Theoretically, in the case of unfissured rock, the rock's tensile strength could resist the Figure 1 — Non-linear elastic opening of the joints due to internal water pressure according to E-E.S. - Model (Fissured, Elastic, Saturated Rock Mass-Model). ot = total stress in the rock mass ™ = intornal water pressure © = opening of the joints % = opening at last contact between the two rock blocks. D.U.DEERE/G. LOMBARDI water pressure of part of it. The question is how wise ic is to trust the tensile strength of a large rock mass, in the case of an important structure like a pressure tunnel or shaft. ‘We will therefore in the following consider only the conditions of a naturally jointed (or fis sured) rock mass. Figure 1 shows the working line ofa joint in a tock mass submitted to a total stress and to an internal water pressure accordingly to the results of the BLE.S. (Fissured, Elastic, Saturated Rock Mass) -Model.(6,7,8). From this figure, we may conclude that: — any increase of the water pressure in the joint will open ic more or less according to its working line OA (unless obviously the rock mass is absolutely tight due to extremely high total stresses as at sec- tion AB); — as soon as the water pressure equals the total stress no limit exists any more to the widening of the joints, section DO); — obviously, the water pressure can never be higher than the cotal stcess acting in the rock mass at the same point Should the water pressure tend to overpass the existing total stress — as in the case of a hydrojacking — a rearrangement of the stress pattern would take place in the rock mass and a ‘new equilibrium will build up where the total stress in the rock will be at any point at least equal to the water pressure. This new equili- brium, as far as it can exist, may be understood asa kind of “overbridging” of the excceding water pressure. Natural Stress Field Before the excavation of the tunnel, a natu tal stress field existsin the mountain. The knowl- edge of the same is obviously a basic necessity 10 propetly design a pressure tunnel so as co avoid any risk of hydrofracturing, At least a prudent estimate of the minimum stress component at any point along the tunnel is mandatory. Ie is evident that the fracturing will likely occur along. the weakest surface, that is along the joints which are submitted to the smallest total stress. These would be oriented more or less perpen- dicular to the axis of the minimum stress compo- nent ‘An older ctitetion used in designing pres- sure tunnels referred only to the overburden onLINING OF PRESSURE TUNNELS AND HYDROFRACTURING POTENTIAL 125 the tunnel. Implicitly only a hydrofracturing along a horizoncal plane through the tunnel was taken into consideration at that time, as if the vertical stress component were automaticaly the minimal one. As later on cases of hydrofracturing were al- so observed along planes parallel to the valley slope, the stress component perpendicular to it was then considered as the probable minimum and used as criterion for the dimensioning of the tunnel. Additionally an Australian criterion was used which requites the lateral extension of the ‘mountain to be double of the overburden needed to resist the internal pressure, i All these criteria, shown at Figure 2, may furnish in the best hypothesis only a first rough estimate for the minimum stress component. To improve them, but just slightly, they are based ‘not simply on the actual topography, but on a topography from which protcuding ridges were removed. ‘Obviously one must take into consideration planes of weakness which may cross the tunnel axis in any direction in space. A further step in improving the estimate of the prevailing stresses consists in computing the stress pattern in the mountain due to the dead weight of the rock on the basis of an elastic mod- el using any kind of E.E.-computations. This method overlooks however the fact that, due to tectonic stresses (or strains), the actual stress pat- tern needs not t0 correspond co the assumption made of an elastic steess law. In fact very impor- mel Oh Figure 2 — Simple criteria to define the risk of hydro fracturing, Tunnel ‘Overburden Australian criterion ‘Stress component perpendicular to slope Natural ground water Terrain ine onan tant deviations were observed in the field from the stresses computed in this way In any case an adequate factor of safety must be used in accepting the computed stresses ‘on the base of these criteria. In certain circumstances the observation of the natural ground water table may be of great help. Should, e.g., the ground water table be higher than the future piezometric level in the tunnel, obviously no risk of hydrofracturing can exist, ‘The most reliable method to define the mini- ‘mum sttess component, when the risk of hydro- fracturing cannot be excluded by means of other considerations, is the use of hydroftacturing tests. The many questions related to such tests can of course nor be discussed here. Suess Field around the Tunnel ‘The knowledge of the natural or primary stress field in the rock mass is however just 2 first step in the evaluation of the risk of hydro- fracturing. The next step is the definition of the secondary stress pattern around the tunnel after the excavation of the tunnel and also on a long term basis, especially before the first filling of the pressure tunnel. At this moment the rock mass around the tunnel has to be considered as being drained, Figure 3 shows the well known secondary stress distribution due to the cavity. For the sake of simplicity of the representation we will con- sider the horizontal plane to be the critical one even if this is generally not the case. Grouting of the rock mass around the cun- nel will induce a favourable change of the stress distribution, An example of such a change may also be seen on figure 3. Water Pressure in the Rock Mass after Filling Unless the lining is absolutely tighe, filling the cunnel will cause a certain water flow from the tunnel to the exterior. We may consider that after a certain transient state, which may be of long duration, a steady flow will take place, an example of which being shown on Figuce 4. Even this steady flow is the result of a quite complex combination of hydraulics and rock mechanics phenomena ina partially natural and partially artificial eavionment, Starting from the internal pressure in the cunnel (where dynamic variations of short dura-126 Figure 3 — Empty tunnel in the drained rock mass; stresses in horizontal olane (assumed to be the critical cone). 1 2 3 4 5 6 7 8 Excavated and lined tunnel Plastic zone around the tunnel Elastic zone . Natural compressive stress befére excevation Vertical (tangential) stress, component after ex- cavation ‘Vertical component of the rock pressure on the fining Stresses induced by @ grouting around the tun- nel Grouted zone Figure 4 — Example of distribution of water pressure round the tunnel. Excevated tunnel Grouted zone Ungtouted zone ain Internat water pressure Pressure drop through the lining Pressure drop in the grouted zone Pressure drop in the ungrouted zone (decreasing permeability with increasing distance ‘due to the closing ofthe joints by decreasing wa- ter pressure) D.U.DEFRE/G. LOMBARDI tion may be disregarded), there will be a pres- sure drop due to the more or less pervious lining The eventually grouted zone around the tunnel will be highly beneficial in reducing the perme- ability of the rock (which may have been loosened by blasting and by its inward displacement) and in producing an additional pressure drop. From this point on, the permeability of the rock mass will define the further drop in pres- sute, The boundary conditions may be defined by natural factors like the groundwater level o the terrain surface, or by artificial means like drains In the determination of the flow pattern we must consider thac the water pressure will open the joints and thus change the permeability of the rock mass which in turn will modify the pres- sure distribution in a somewhat complex way Furthermore it has to be considered that the problem is not simply 2 one-dimensional, as shown in the picture, but a ui-dimensional one. Most likely a longitudinal flow along the tunnel will take place which will tend to equal- ize the pressures in the vatious cross sections. Of course, the results will be strongly influenced by anisotropy of the rock mass and its permeability. Te is immediately clear that the ratio of the permeabilities of the lining to that of the rock ‘mass at various places will play 2 determinant tole in the distribution of the water pressures around the tunnel. Equilibrium around the Tunnel To avoid a hydroftacturing of the rock mass an equilibrium between the existing stresses and the pressures due to the water flowing out from the tunnel must be obtained. In Figure 5 the stresses from figure 3 are compated with the pressures from Figure 4 ‘As a rule it may happen that the equi- librium beeween the toral stress and the water pressure will not be obtained at every point around the tunnel, especially close-in. ‘As the internal pressure can never be highet than the total stress, a rearrangement of the siress field must take place which may be inter- preted as an overbridging of the excess water pressure in the vicinity of the wnnel. In fact, a kind of local hydroftacturing will take place al- ‘most in any case in a small zone near the lining. Its extension will be limited by the decrease of the water pressure and by the increase of the rock stresses with distance from the tunnel.LINING OF PRESSURE TUNNELS AND HYDROFRACTURING POTENTIAL ur be Figure’ — Equilibrium between water pressure and vor- tical stresses in the horizontal plane as afvexample. 1. Stross in the rock mass (after tunnel excavation) from figure 3 2. Water pressure around the tunnel, from figure 4 3. Excess of water pressure to be supported {ov bridged) by stress rearrangement in the rock Note: stress changes will modify permeability of the rock joints and thus the water pres- sure distribution and the excess af water pressure to be taken by the rock, leading possibly to a hydrofracturing F Tensile force in the steel ‘An actual hydrofracturing will be con- sidered to occur only if the desetibed extension will exceed an acceptable limit, and therefore a large flow will escape from the tunnel. In fact the conditions of equilibrium need not be satisfied at any point, but the coral forces have to comply with this condition. In a simple manner one may say that the weight of the over- burden must be higher chan che total hydraulic uplifc However it must be kept in mind that with the opening of fissures the uplift force can in- crease dramatically, leading to a kind of elastic instability: a hydeofracturing. Practical Procedure The just described procedure to define the accual stresses in the rock masses, the flow pat- tern and the complex manifold interaction be- ‘ween stresses, pressures, joint opening and per meability is considered correct from a theoretical viewpoint. However, sufficiently exact information on the actual permeability of the lining and of the rock mass, as well as on the stress distribution around the tunnel, will seldom be availabe, Also in many cases such complicared computations are not required, e.g. if the conditions of equi- librium can be satisfied with a high enough fac- corof safety. In such cases a simplified procedure can be followed (even if it is not as simple as the criteria mentioned at the beginning). In many cases it could be sufficient to prove that the minimum stress measured by means of a number of hydrofraccuring tests is higher than the future internal water pressure in any section of the tunnel, using however an adequate factor of safety. Itis believed anyhow, e.g. when drains are necessary to ensure an equilibrium, that investi- gations of the nature of the above described ‘could be very useful Conclusions The problem of hydroftacturing of a pres- sure tunnel can be of great concern when the is ternal pressure approaches or overpasses the val- ue of the minimum natural stress component at any section of the tunnel stretch. A number of measures can be taken to im- prove the situation and to increase the factor of safety. Among others the following procedures can be mentioned. — make the lining completely tight using a sheet of meral or plastic; — make the lining less pervious in avoiding cracks and defects in the concrete; — reinforce the lining co get more, but finer, cracks instead of few wide ones; — pre-stress the lining using one of the many available technical solutions; — grout the rock mass around the tunnel to reduce its permeabilicy and also to in- crease its effective stcesses; — use drains which may additionally con- tribute (0 reduce the water flow es- caping from the tunnel, as the reduc- tion of che water pressure will permit the fissues to close and will reduce the permeability of the rock mass; — set the tunnel at a lower elevation, the simplest solution if possible. In many cases a more detailed analysis of the stress and pressure distribution around the tunnel can be useful. A good precaution is to fill128 DU, DEERE/G. LOMBARDI the tunnel slowly and ro monitor the procedure natural stress ficld as well as the rig’ : choice of with leakage tests and observation for surface the lining type, are essential requisites for a suc- leakage. cessful design of a pressure tunnel, Where hydro- In any case, a good knowledge of the gco- fracturing potential exists, the lining rather than logical and geomechanical conditions and of the the rock must be counted on to imic the leakage. References DEERE, D.U. — Unique Geotechnical Problems at Some Hydroelectric Projects, Seventh Panamerican Conf. on Soil Mech. and Found. Eng., Vancouver. Vol. Ill, 1983, pp. 865- 888. BROCH, E. — The Development of Unlined Pressure Shafis and Tunnels in Norway, Proc. .S.R.M. Symp. on Rock Mech.: Caveins and Pressure Shafts, Aachen, Vol. 2, 345- 554 BUEN, B. AND PALMSTROM, A. — Design and Supervision of Unlined Hydro Power Shafts and Tunnels with Head up t0 590 Meters, Proc. 1.S.R.M. Symp. on Rock Mech. Caverns and Pressure Tunnels, Aachen, Vol. 2, 1982. pp. 567-574. SWANSON, D.T. — Prestressed Concrete Pressure Tunnels, Proc. Rapid Excavation and Tunneling Conf., San Francisco. Vol. 1, 1981, pp. 519-532. KUJUNDZIC, B., IVANOVIC, K., MARKOVIC, O.., and NIKOLIC, Z. — Some Resuks of Research on Lining Prestressing for the Pressure Tunnel of a Pumped Storage Hydroelectric Plant, Proc. 1.S.R.M. Symp. on Rock Mech.: Caverns and Pressure Shafts, Aachen, Vol. 2, 1982, pp. 639-652. LOMBARDI, G. — L'injection des massifs rocheux, Sté Suisse de mécanique des sols t des roches, Journée de printemps 1987 4 Bienne, Publication No. 115 de la Société Suisse de Mécanique des Sols et des Roches. pp. 29-41 LOMBARDI, G. — La roca y el macizo rocoso, leyes constitutivas, Academia Nacional de Ciencias Exactas, Fisicas y Navurales, Buenos Aires, December 1987. LOMBARDI, G. — The F.E.S.-Model and Foundations for Concrete Dams, Victor de Mello Volume 1989, Brazilia. SEEBER, G. — Power Conduits for High-Head Plants, Water Power and Dam Construction, June-July, 1985, pp. 50-54, 95-98. SEEBER. G. — Neue Entwicklungen im Druckstollenbau, Proc. 1.S.R.M. Symp. in Rock Mech.: ‘Caverns and Pressure Shafts, Aachen, Vol. 2, 1982, pp. 705-716. SHARP, J.C., and GONANO, L.P. — Rock Engineering Aspects of the Concrete Lined Pressure ‘Tunnels of ehe Drakensberg Pumped Storage Scheme, Proc. 1.S.R.M. Symp. on Rock Mech.: Caverns and Pressure Shafts, Aachen, Vol. 2, 1982, pp. 717-733. DE MELLO © EDITORA EDGARD BLECHER LTDA VOLUME Fo inquires plate write 1 EDITORA EDGARD BLUCHER LTDA. A teibure CAIXA POSTAL 5450 jctor F. B. de Mello 01051 SAO PAULO - SP. BRASIL 0 Prof. Dr.