UNDERGROUND ROCK ENGINEERING METHODOLOGY

In the years since the construction of the Snowy Scheme, a reasonably well defined underground rock engineering methodology has been developed and formalised. The rock engineering process may involve feasibility and location studies, site investigation, design analyses, cost and optimisation studies, detailed design, environmental impact studies, construction planning, construction including excavation, reinforcement and other ground improvement measures, and performance monitoring during and after construction. It will be shown that SMHEA engineers made important original contributions in several of these areas.

Figure 3 shows the now widely accepted components and logic of a modern rock mechanics program in generalised and simplified form. More detailed versions of this general approach applied to underground excavations are given by Bieniawski (1984) and by Hoek and Brown (1980). In practice, the rock engineering process is not as linear as, and involves more complex interactions and in-parallel studies than, the simple depiction of Figure 3 might suggest (Hudson 1992). Nevertheless, it is convenient to analyse the investigation, design and construction of the Tumut 1 and 2 underground power stations within this overall framework.

Figure 3: Components of a generalised rock mechanics program (from Brown 1993).



SITE CHARACTERISATION AND GEOTECHNICAL MODEL FORMULATION

Possible sites for the underground power stations were investigated using air-photo interpretation, surface geological mapping and diamond drilling. Electrical resistivity logging and seismic refraction surveys were also used in site investigations for other structures in the Scheme (Moye 1955). In the case of Tumut 1, four diamond drill holes ranging from 300 to 600m in length were first drilled from the surface into the rock at the proposed power station site. As the surface drilling indicated favourable conditions at the proposed site, an exploratory tunnel was driven to confirm and augment the data obtained from the surface drill holes and to enable direct examination of the rock in which the excavations would be made. From a chamber located at the end of the exploratory tunnel (see Figure 1), six diamond drill holes with a total length of 420m, were drilled across the proposed excavations for the machine and transformer halls and draft tubes (Moye 1959, Pinkerton et al 1961).

A similar process was followed for Tumut 2 (Figure 2). However, as a result of the identification of a crushed and sheared zone in the exploratory tunnel and the underground diamond drill holes, the initially proposed location of the power station and much of the associated works was moved by approximately30m along the longitudinal axis of the machine hall (Stapledon 1961, Pinkerton & Gibson 1964).

The core logs produced for diamond drill holes on the Snowy Scheme (eg Moye 1967) contain the same essential features as modern core logs although the details differ. The main rock mass classification scheme available at the time of the investigations for Tumut 1 and 2 was Terzaghi's scheme for estimating loads on tunnel supports (Terzaghi 1946). The concept of Rock Quality Designation (RQD) was not introduced until about 1963 (Deere 1964). The NGI Q system (Barton, Lien & Lunde 1974) and Bieniawski's Rock Mass Rating (RMR) (Bieniawski 1973) were not developed until the early 1970s. A simple rock classification scheme was developed for use on the Snowy and correlated with the support typically used in the underground works (Pender et al 1963). Weathering of the essentially granitic rock masses at these and other sites was a major issue on the Snowy Scheme. Moye (1955) made a major and lasting contribution to engineering geology by developing a weathering classification scheme for granitic rocks. Water pressure testing was also carried out in diamond drill holes from an early stage in the site investigation process on the Snowy (Moye 1955, Stapledon 1961).

An especially important feature of the site investigation work carried out in rock on the Snowy Scheme was the attention paid to high quality diamond drilling and core recovery (Moye 1955, 1959, 1967). The desirability of core orientation was recognised but was not then able to be achieved (Stapledon 1961). Many of the issues involved were developed further by Rosengren (1970) and re-emphasised by Hoek and Brown (1980). In the author's opinion they have not since received the attention that they merit despite the recent introduction of drill-rig monitoring systems and a wide range of down-hole logging systems. In this area, as in so many others, the Snowy's geological and engineering staff were "ahead of their time".

The importance of jointing and faulting was well recognised and was taken into account in the design, including the orientations, of the underground excavations. It appears that not all discontinuities encountered in a borehole or along an exposure were mapped and plotted on a stereographic projection as they are today and were in the latter stages of the Snowy Scheme. However, the major joint sets and their orientations were identified (but not expressed in the now standard dip and dip direction terms) and presented as rosette diagrams (see Figure 2). The conditions of joint surfaces were also recorded.

Laboratory tests were carried out on unjointed cores of the rocks to measure their uniaxial compressive and tensile strengths, Young's moduli and Poisson's ratios. No published evidence has been found of the systematic shear strength testing of joints at the time of the design of Tumut 1 and 2 power stations. This was not then a well-established practice. Pinkerton et al (1961) make a revealing comment about the rock property measurements:

These are not necessarily the properties of the in situ rock mass, intersected as it is by such a large number of joints. It is considered that the compressive strength of the rock mass should, however, be very similar to the values measured in the laboratory, particularly where joints are tight and the intervening rock is interlocked in place. The tensile properties, on the other hand, are likely to be reduced almost to zero on account of the joints.

This quotation reflects a developing, but as then incomplete (as it remains), understanding of the complex problem of establishing in situ rock mass properties.

No in situ stress measurements were made at the Tumut 1 site in the investigation stages in the early to mid-1950s. The science and art of in situ stress measurement in rock was in its embryonic stages at the time. Hast's pioneering study showing that the horizontal component(s) of in situ stress could be several times the vertical or overburden stress was published in Sweden only in 1958 (Hast 1958). In accordance with the received wisdom of the day, it was assumed for design purposes that the ratio of horizontal to vertical stress was 0.25, a figure arrived at using the well-known assumption of complete lateral restraint during the gravity loading of an elastic medium. Measurements and observations made during construction (see below) soon led to the conclusion that high horizontal in situ stresses existed at the Tumut 1 underground power station site (Alexander et al 1963, Pinkerton et al 1961). Rock stress measurements using the then recently developed flat jack method (Talobre 1957) were made at Tumut 1 when construction was well advanced. The flat jack method of stress measurement had been first used on the Snowy Scheme at the proposed site for the Tumut 2 power station in the exploratory tunnel at approximately the mid-depth of the machine hall (Alexander 1960). The primary stress measurement specialist, L G Alexander, became widely recognised for, among other qualities, his expertise and deep understanding of the complex issues involved in making in situ stress measurements in rock.

DESIGN ANALYSIS

Stress Analysis

At the time of the design of the major underground excavations on the Snowy Scheme, there existed a number of closed-form solutions for the stresses around excavations with simple shapes (eg Terzaghi & Richart 1952). In Europe, methods of calculation for tunnel support and lining design had been developed (eg Szechy 1973). There were no digital computers or numerical methods of the types that are now used in carrying out stress analyses of underground excavation complexes.

The design stress analyses for the Tumut 1 and Tumut 2 power stations were carried out using the photo-elastic method. This was an undertaking requiring considerable ingenuity and skill. The state-of-the-art at the time was to use fully concrete-lined arched roofs with load bearing ribs. In the case of the underground power stations on the Snowy, it was planned to use systematic rock bolting "to a fairly close pattern so as to tie the rock together at the joints, thus maintaining structural continuity of the rock immediately surrounding the excavations" (Pinkerton et al 1961 The roof of Tumut 1 was supported by concrete ribs designed to support a nominal load of 12ft (3.66m) of loosened rock as suggested by Terzaghi (1946). In the design, it was assumed that the roof section would be excavated first and that the concrete ribs would be placed before the excavation was carried further down.

Figure 4: Stress concentration factors in the Tumut 1 machine hall excavation


(from Pinkerton et al 1961).

The results of the two-dimensional photo-elastic stress analysis for this stage, assuming a horizontal to vertical in situ stress ratio of 0.25, are shown in the upper left-hand part of Figure 4. These results indicated that the central part of the roof would be in tension and that high compressive stresses would exist at the abutments. These studies also showed, as illustrated in the lower left-hand side diagram of Figure 4, that the value of the roof tension would decrease and that tensionwould develop in the rock walls immediately below the roof abutments as excavation proceeded downwards (Alexander et al 1963, Pinkerton et al 1961). It was recognised that the latter tensile zones were a function of the design with a re-entrant angle then used to support the roof ribs. For this essential reason, underground power station designs were subsequently changed to use smoother, more regular shapes (Hoek & Brown 1980). In the remaining sections of the walls, the compressive stresses indicated by the photo-elastic analyses were considerably less than the compressive strength of the rock as determined in the laboratory. The designers "concluded that, if the rock in the roof was stable following the initial excavation, no troubles should be encountered as the excavation progressed further" (Pinkerton et al 1961).

The excavations were thoroughly monitored during and after construction. Measurements made in the roof ribs showed that the compressive stress in the concrete increased very sharply as the excavation proceeded down from the roof abutments. It was recognised that the likely explanation of this was the existence of much higher horizontal stresses than those assumed on the basis of classical "theory". It was decided to immediately make stress measurements in the power station excavations using the flat-jack method.

Sites for the flat-jack tests were carefully selected at locations where fairly massive rock occurred. It was recognised that in order to determine from these results the stresses in the rock mass before excavation, it was necessary to establish the local stress concentration factors taking account of the shapes of the excavations in which the measurements were made. The final results showed an average equivalent vertical stress of approximately 12.4 MPa and a horizontal stress transverse to the machine hall of 10.3MPa (Alexander et al 1963, Pinkerton et al 1961).

Further photo-elastic stress analyses were then carried out for Tumut 1 with a ratio of horizontal to vertical in situ stress of one. The results are shown on the right hand side of Figure 4. The predictions made using these models showed very good correlation with the monitoring data.

The design investigations for Tumut 2 underway at the time made use of this experience in Tumut 1. Flat-jack tests were carried out as a result of which a ratio of horizontal to vertical stress of 1.2 was adopted in the design. Based on the experience of Tumut 1, the roof abutment design was modified for Tumut 2. The stresses around the periphery of the excavation calculated from the photo-elastic stress analyses showed remarkably good agreement with those measured during and after excavation (Alexander et al 1963, Pinkerton & Gibson 1964).

Although the techniques used in these studies may seem primitive in today's computer age, at the time they were at the forefront of developments in the field and demonstrated engineering skills of a high order. The author knows of no more sophisticated and effective stress analyses being carried out at the time (or for a long time afterwards) than the photo-elastic analyses of G Worotnicki on the Snowy. The skills of L G Alexander in carrying out and interpreting early flat-jack and monitoring measurements have been referred to above. The approaches and skills developed in the studies for Tumut 1 and 2 were subsequently used and extended in the Poatina underground power station in Tasmania which was notable for its daring and innovative de-stressing technique (Endersbee & Hofto 1963).

Rock Bolting

Although it had been used on a few projects in North America and France in the early 1950s, at the time of the design and construction of the early tunnels on the Snowy Scheme, rock bolting was not an established method of rock support in civil engineering either in Australia or elsewhere. On the Snowy Scheme, some rock bolts were used in the Guthega Project excavated in the period 1952-54, but the first major use of rock bolts in the Authority's works was in Tumut 1 power station excavated in 1956-57 (Pender et al 1963). This represented quite a major departure from the then traditional steel set and concrete support. Here again, the SMHEA was at the forefront of the application of this new technology in the design and construction of its underground excavations.

In the early to mid-1950s, it was generally held that the purpose of rock bolts was to pin surface rock (either individual blocks or bedded strata as encountered in underground coal mining) to more stable rock some distance from the surface of the excavation. At the time, mechanically anchored (slot and wedge or expansion shell) bolts were used. Rabcewicz (1957) carried out model tests which suggested that a rock mass consisting entirely of blocks or fragments could be held stable by systematic bolting. The team of investigation and design engineers working on the Snowy with the Assistant Commissioner, T A Lang, as the driving force proved, developed and applied this concept with remarkable effect. Indeed, it has been suggested that the development and use of rock bolting for permanent support of underground excavations was probably the most significant engineering development made on the Snowy Scheme.

At this point it is interesting to note that, at the time, and for some years later (eg Hoek & Brown 1980), rock bolts were referred to as providing "support". Today, it is usual to distinguish between support and reinforcement on the basis of the method by which the rock adjacent to an excavation is stabilised. Support is the application of a reactive force at the face of the excavation. Reinforcement is the improvement of the overall rock mass performance from within the rock mass by techniques such as rock bolts, cable bolts and ground anchors (Windsor & Thompson 1993). At the time of the construction of the Snowy Scheme, support was most often described as being "temporary" or "permanent". In their important paper on the use of grouted rock bolts on the Snowy, Pender et al (1963) refer, more helpfully and correctly, to "construction" and "permanent" support, although the term "temporary" was used in other Snowy publications.

In the mid- to late 1950s, a detailed series of laboratory experiments was carried out by the SMHEA to investigate the action and effect of rock bolts. Simultaneously, experience was gained with the practical application of rock bolting, initially on the Tumut 1 project. The laboratory experiments described by Lang (1961) and by Alexander and Hosking (1971) included model tests of the arched roof of the Tumut 1 machine hall using gravel and regularly shaped perspex blocks to simulate the rock mass; cubical box models of various sizes in which the boxes were filled with crushed rock, provided with model or prototype rock bolts and turned upside down with the open side facing downwards and often carrying an applied load; rod models; photo-elastic models using a solid plate of photo-elastic material or an assembly of blocks; and the famous inverted bucket model. In the latter case, "a household bucket was filled with gravel and the surface layer was bolted with model rock bolts with 1 inch (25.4mm) square bearing plates at both ends of the bolt. When the bucket was inverted, the lateral pressure developed on the sloping sides of the bucket was sufficient to support not only the gravel mass but also a central load of 50 lb (0.22 kN)" (Alexander & Hosking 1971).

Figure 5: Action of pattern rock bolting to form a self-supporting ring or arch


(from Pender et al 1963).

On the basis of these experiments and through field experience, the Snowy team developed an understanding of the way in which systematic rock bolting in a jointed rock mass forms a self-supporting compression zone within the rock mass. This effect is illustrated in Figure 5. In addition, a set of design rules was developed for pattern rock bolting which related bolt length and spacing to block size (Lang 1961). These rules represented the state-of-the-art for many years subsequently (Hoek & Brown 1980). Lang (1961) also published pioneering analyses of the ways in which single rock bolts may prevent slip on single joints and single blocks of rock may be stabilised.

In the Tumut 1 machine hall, generally 4.6m long ungrouted mechanically anchored bolts on 1.4m centres were used in the walls. In the roof, 3.7m long bolts on 1.2m centres were used as "construction support" between the 1.2m square concrete ribs (Pinkerton et al 1963). The use of grouted rock bolts for permanent support was pioneered at Tumut 2 built a few years later. Generally, 4.3m long cement grouted bolts on 1.2m centres were used. Some 2.4m and 3.7m long grouted bolts were also used (Pinkerton & Gibson 1963). Following its successful use on these large underground excavations, pattern rock bolting, often with grouting, became an important feature of the several large tunnels constructed subsequently on the Snowy Scheme (Andrews et al 1964, Pender et al 1963).

MONITORING AND RETROSPECTIVE ANALYSIS

The final elements of the generalised rock mechanics program show in Figure 3 are monitoring rock mass performance and retrospective analysis. Pinkerton et al (1961) indicate that the following monitoring measurements were made in the Tumut 1 underground power station excavation:

(a) measurement of strain in the reinforced concrete arch ribs;
(b) measurements of rock and concrete movements were made by means of precise survey methods, and by clinometers which had a very sensitive level;
(c) measurements of variations in the tension of selected rock bolts were made by means of calibrated rubber compression pads;
(d) rock noise measurements were taken."

1. Alexander, L G 1960. Field and laboratory tests in rock mechanics. Proc. 3rd A.-N.Z. Conf. Soil Mech. Found. Engng, Sydney, pp.161-168.
   
2. Alexander, L G, Worotnicki, G & Aubrey, K 1963. Stress and deformation in the roof and roof support, Tumut 1 and Tumut 2 Underground Power Stations. Proc. 4th A.N.Z. Conf. Soil Mech. Found. Engng, Adelaide, pp.165-178.
   
3. Alexander, L G & Hosking, A D 1971. Principles of rock bolting - formation of a support medium. Proc. Symp. Rock Bolting, Illawarra Branch, Australas. Inst. Min. Metall. Feb 1971, pp.1.1-1.21.
   
4. Andrews, K E, McIntyre, A R & Mattner, R H 1964. Some aspects of high speed hard rock tunnelling in the Snowy Mountains. Civ. Eng. Trans., Instn Engrs, Aust. Vol.CE6, No.2, pp.51-69.
   
5. Barton, N, Lien, R & Lunde, J 1974. Engineering classification of rock masses for the design of tunnel support. Rock Mechanics, Vol.6, No.4, pp.189-236.
   
6. Bieniawski, Z T 1973. Engineering classification of jointed rock masses. Trans. S. Afr. Instn Civ. Engrs, Vol.15, No.12, pp.335-344.
   
7. Bieniawski, Z T 1984. Rock Mechanics Design in Mining and Tunnelling. Balkema, Rotterdam.
   
8. Brown, E T 1991. Australian advances in geomechanics. In Beer, G, Booker, J R and Carter, J P (eds.), Proc. 7th Int. Conf. Computer Methods & Advances in Geomechanics, Vol.1, pp.15-29. Balkema, Rotterdam.
   
9. Deere, D U 1964. Technical description of rock cores for engineering purposes. Rock Mech. & Engrg Geol., Vol.1, pp.16-22.
   
10. Endersbee, L A & Hofto, E O 1963. Civil engineering design and studies in rock mechanics for Poatina Underground Power Station, Tasmania. J. Instn. Engrs, Aust., Vol.35, No.9, pp.187-206.
    
11. Hast, N 1958. The measurement of rock pressure in mines. Årsb. Sver. Geol. Unders. Vol.52, No.3.
    
12. Hoek, E & Brown, E T 1980. Underground Excavations in Rock. Institution of Mining & Metallurgy, London.
    
13. Hood, M & Brown, E T 1999. Mining rock mechanics, yesterday, today and tomorrow. Proc. 9th Int. Congr. Rock Mech., Paris (in press).
    
14. Hudson, J A 1992. Rock Engineering Systems - Theory and Practice. Ellis Horwood, London.
    
15. Jaeger, C 1972. Rock Mechanics and Engineering. Cambridge Univ. Press, Cambridge.
    
16. Lang, T A 1961. Theory and practice of rockbolting. Trans. Soc. Min. Engrs AIME, Vol.220, pp.333-348.
    
17. Maddox, J M, Kinstler, F L & Mather, R P 1967. Meadowbank Dam - foundations. Civ. Eng. Trans. Instn. Engrs, Aust., Vol.CE9, pp.321-329.
    
18. May, J R 1980. Industry-sponsored rock mechanics research in Australia. Proc. 13th Can. Rock Mech. Symp., Toronto, pp.219-255.
    
19. Moye, D G 1955. Engineering geology for the Snowy Mountains scheme. J. Instn Engrs, Aust., Vol.27, pp.287-298.
    
20. Moye, D G 1959. Rock mechanics in the investigation and construction of T.1 underground power station, Snowy Mountains, Australia. Geol. Soc. Am. Engrg Geol. Case Histories, No.3, pp.13-44.
    
21. Moye, D G 1967. Diamond drilling for foundation exploration. Civ. Engrg Trans., Instn Engrs, Aust., Vol.CE9, pp.95-100.
    
22. Pender, E B, Hosking, A D & Mattner, R H 1963. Grouted rock bolts for permanent support of major underground works. J. Instn Engrs Aust., Vol.35, No.7-8, pp.129-145.
    
23. Pinkerton, I L, Andrews, K E, Bray, A N G & Frost, A C H 1961. The design, construction and commissioning of Tumut 1 Power Station. J. Instn Engrs, Aust., Vol.33, No.7-8, pp.235-252.
    
24. Pinkerton, I L & Gibson, E J 1964. Tumut 2 underground power plant. J. Power Div., Am. Soc. Civ. Engrs, Vol.90, No.PO1, pp.33-58.
    
25. Potts, E L J 1957. Underground instrumentation. Q. Colo. Sch. Mines, Vol.52, No.3, pp.135-182.
    
26. Rabcewicz, L 1957. Modellversuche mit Ankerung in kohäsionslosem Material. Die Bautechnik, Vol.34, No.5, pp.171-173.
    
27. Rosengren, K J 1970. Diamond drilling for structural purposes at Mount Isa. Ind. Diamond Rev., Vol.30, pp.388-395.
    
28. Sandström, G E 1963. The History of Tunnelling. Barrie & Rockliff, London.
    
29. Snowy Mountains Hydro-electric Authority, 1993. Engineering Features of the Snowy Mountains Scheme, 3rd ed. SMHEA, Cooma.
    
30. Stapledon, D H 1961. Geological studies for the planning and construction of Tumut 2 underground power station. Engrg Geol. Rep. S.G. 86, SMHEA. M.Sc. thesis, Univ. Adelaide.
    
31. Széchy, K. 1973. The Art of Tunnelling, 2nd edn. Akadémiai Kiadó, Budapest.
    
32. Talobre, J 1957. La Méchanique des Roches. Dunod, Paris.
    
33. Terzaghi, K 1946. Rock defects and loads on tunnel supports. In Proctor, R J & White, T L (eds.), Rock Tunnelling with Steel Supports, pp.15-19. The Commercial Shearing & Stamping Co., Youngstown, Ohio.
    
34. Terzaghi, K & Richart, F E 1952. Stresses in rock about cavities. Géotechnique, Vol.3, pp.57-90.
    
35. Windsor, C R & Thompson, A G 1993. Rock reinforcement - technology, testing, design and evaluation. In Hudson, J A (ed.), Comprehensive Rock Engineering, Vol.4, pp.451-484. Pergamon Press, Oxford.