Variability of the engineering properties V. Marinos of...

P. Marinos

E. Hoek

V. Marinos

Variability of the engineering propertiesof rock masses quantified by the geologicalstrength index: the case of ophioliteswith special emphasis on tunnelling

Received: 16 May 2005Accepted: 30 July 2005� Springer-Verlag 2005

Abstract The paper presents aquantitative description, using theGeological Strength Index (GSI), ofthe rock masses within an ophioliticcomplex including types with largevariability due to their range ofpetrography, tectonic deformationand alteration. This description al-lows the estimation of the range ofrock mass properties and theunderstanding of the dramaticchanges in behaviour which can oc-cur during tunnelling, from stableconditions to severe squeezing with-in the same formation at the samedepth. The paper presents the geo-logical model in which the ophioliticcomplexes develop, their variouspetrographic types and their tectonicdeformation, mainly due to over-thrusts. The structure of the variousrock masses includes all types frommassive strong to sheared weak,while the conditions of discontinu-ities are in most cases fair to poor orvery poor due to the fact that theyare affected by serpentinisation andshearing. Serpentinisation also af-fects the initial intact rock itself,reducing its strength. Associatedpillow lavas and tectonic melangesare also characterised. Based on theGSI, a classification of the behav-iour in terms of tunnelling ispresented, including stable condi-tions, structural instability, mildoverstressing, stress dependantinstability, squeezing and ravelling.

Keywords Ophiolites Æ Rock massclassification Æ Geological strengthindex Æ Tunnels

Resume Une description quantita-tive des massifs rocheux des com-plexes ophiolitiques est presentee parle moyen de l’index GSI. Les ophi-olites forment un cas particulier acause de leur variete petrographique,leur deformation tectonique et leuralteration. Cette description permetl’estimation des proprietes geotech-niques et la comprehension desdifferents types de comportementssouvent tres variables rencontres lorsdu creusement de tunnels. L’articlediscute brievement le modelegeologique de ces formations, leursvarietes petrographiques et leurdeformation a cause surtout descharriages. La structure des massifsrocheux ophiolitiques inclut tous lestypes, (du milieu continu au cisaille),tandis que l’etat des joints est touj-ours faible a cause de la serpentini-sation de leurs epontes. Laserpentinisation peut aussi affecter lamasse entiere de la roche saine. Uneclassification du comportement entunnel est presentee basee sur l’indexGSI: conditions stables, instabilitestructurale, instabilite due a desconvergences.

Mots cles ophiolites Æ massifrocheux Æ index GSI Æ tunnels

Bull Eng Geol Env (2005)DOI 10.1007/s10064-005-0018-x ORIGINAL PAPER

P. Marinos (&) Æ V. MarinosNTUA, School of Civil Engineering,9 Iroon Polytechniou str, 157 80 Athens,GreeceE-mail: [email protected]: [email protected]

E. HoekConsulting Engineer, Vancouver, CanadaE-mail: [email protected]

Introduction

Over the past decades, rock mass classification methodshave been developed in order to meet the needs fordesigning specific projects. The rock mass classificationsystems of Bieniawski (1973) and Barton et al. (1974)were originally developed to provide guidance on theselection of support for tunnels in blocky rock massesand they played an important role in the expansion ofthe tunnelling industry.

During the last decade, the development of ‘‘user-friendly’’ software has provided alternative design toolsthat are more appropriate in many cases. However, thishas brought with it the need for reliable input data,particularly that related to rock mass properties. In or-der to meet this requirement a different set of classifi-cation schemes for the characterisation of rock masseshas been developed. A system that is now widely used isthe Geological Strength Index (GSI) developed by Hoek(1994) and extended by Hoek et al. (1998), Marinos andHoek (2001) and Hoek et al. (2005) to incorporate weak,heterogeneous rock masses and rock masses with litho-logical variability. This classification system is used inconjunction with the Hoek and Brown failure criterionto estimate the geotechnical parameters of rock masseswhich fall within the range specified by the authors. Apresentation and a discussion on the use of GSI can befound in Marinos and Hoek (2000) and more recently inMarinos et al. (2005) hence there is no need to repeat thedetails here.

The GSI has considerable potential for use in rockengineering because it permits the manifold aspects ofrock to be quantified, enhancing geological logic andreducing engineering uncertainty. Its use allows theinfluence of the variables which make up a rock mass tobe assessed and hence the behaviour of rock masses hasto be explained more clearly. One of the advantages ofthe Index is that the geological reasoning that itembodies allows adjustments of its ratings to cover notonly a wide range of rock masses and conditions but alsovariations that may develop within the same rock type.It also allows the limits of its application to be under-stood.

This paper presents a quantitative description,through the GSI, of the rock masses of a particular typeof geological formation with both petrographic varietyand structural complexity due to tectonic deformationand alteration. This description allows an estimation ofthe variation in rock mass strength and how much thisrock mass strength can be reduced by shearing oralteration as well as an understanding of the dramaticchanges in behaviour where, in tunnelling, stable con-ditions and severe squeezing can occur within the sameformation at the same depth.

Ophiolites: geological model

Setting

The term ophiolite was initially given to a sequence ofmafic (basic) and ultramafic (ultrabasic) rocks, more orless serpentinised and metamorphosed, occurring in theAlpine chains. These complexes were considered, notlong ago, as enormous submarine volcanic effusionsinside which magmatic differentiations took place shel-tered by a cuirass of pillow lavas. Although this can bethe case in some regions, ophiolites are at present con-sidered as pieces of the oceanic crust generated at anoceanic ridge and the upper mantle of an ancient ocean,thrust up on the continental crust during mountainbuilding (e.g. collision between two continents or be-tween a continent and an insular arc; see Fig. 1). Theycan exhibit sections of more than 10 km in thicknesswhich leads to the conclusion that not only the oceaniccrust (6–7 km) but also part of the mantle are includedin the process (Debelmas and Mascle 1997).

The ophiolitic complex

The ophiolitic sequence (or complex in order toemphasise the diversity of materials) is fundamentallycharacterised by underlying peridotitic rocks which arecovered by gabbroic/peridotitic rocks which, in turn, arecovered by basalts or spilites. The basal peridotites arefoliated (‘‘tectonites’’). The subsequent alternations ofperidotites and gabbros often have a layered structure ofcumulates and are followed by massive gabbros, noritesor other basic rocks richer in SiO2. The overlying basaltsare either massive or in the form of pillow lavas. Inbetween these lavas sedimentary rocks may be depos-ited. In Fig. 2 a synthetic and theoretical column of anophiolitic complex is presented. The succession is idea-lised and in many cases some members may be absent, asfor instance the volcanic lava at the top.

This geometry is highly disturbed as the ophioliticcomplexes occur mainly in tectonic zones with super-position of numerous overthrusts. Metamorphism,which is also present, changes the original nature of thematerials. The high degree of serpentinisation and theintensity of shearing can make it difficult to identify anyinitial cumulate texture or fabric (Skemp and McCraig1984).

Serpentinisation

Serpentinisation is the transformation of ferromagne-sian minerals, olivine in particular, to serpentine—a

Engineering properties of rock masses: ophiolites

lattice mineral of either fibrous or laminar form. Thisunusual alteration is a phenomenon of autohydratationwhich takes place during the last phases of the crystal-lisation of magma where there is an excess of water.Thus it can be considered as a type of autometamor-phism. In other cases the serpentinisation corresponds toa low grade metamorphism of peridotites (Foucault andRault 1995). In all these cases the peridotites can betransformed into serpentinite. This new rock is origi-nally compact, relatively soft and more easily sheared bytectonic processes.

Serpentinisation can also be developed under exo-genic conditions with meteoric water under usualweathering processes. In this case the alteration disin-tegrates the parent peridotite to a clayey soil-like mass.The development at depth of weathered peridotites is

less generalised and obviously limited compared with theendogenic serpentinisation described previously.

Pillow lavas

Pillow lavas, usually of basaltic or andesitic composi-tion, have been extruded under water and consist of amass of more or less ellipsoidal bodies each with abillowy surface. The pillows range from 10 cm to a fewmetres in diameter and lie merged with one another,not unlike an irregular collection of ‘‘sofa pillows’’ (inVisser 1980). Radial joints are conspicuous in crosssections, forming radial columnar fragments (Pantazis1973).

Fig. 1 Tectonic model for theevolution of the Pindos andVourinos mountain ophiolitesin Northern Greece (modifiedfrom Jones et al. 1991, inPe-Piper and Piper 2002)

P. Marinos et al.

Melanges

As ophiolites are associated with large-scale overthrusts,tectonic melanges can be formed in the base and at thefront of such megastructures. These ophiolitic melangescontain ophiolitic rocks and other rocks of variouspaleogeographic origins; the whole entity being in con-siderable tectonic disorder with chaotic masses where

blocks and packages of various sizes of any kind of rock(sedimentary or volcanic) ‘‘float’’ inside a sheared soil-like mass.

Sketches of models around the world

Most of the ophiolites belong to the Alpine cycle (theirage ranges from 180 to 60 MA) but older deposits arealso known (e.g. in the Apalaches—Paleozoic, or inMaroc—Precambrian) (Debelmas and Mascle 1997).

Figures 3, 4 and 5 show schematically the generalgeological model of the ophiolitic formations in anumber of mountain chains of the world, where petro-graphic and tectonic complexities exist (Fig. 6).

Geological engineering characterisation of various rockmasses in the ophiolitic bodies

The geotechnical parameters

From the discussion on the geological model of ophio-lites it is clear that this formation contains a variety ofrock types with geotechnical qualities varying fromexcellent to fair, becoming poor to very poor whenserpentinisation is extensive and/or shearing present.These last processes are both very frequent in theophiolitic complexes.

The main fundamental types are peridotites, gabbros,peridotites more or less serpentinised, serpentinites,schisto-serpentinites, sheared serpentinites, pillow lavasand chaotic masses in ophiolitic melanges.

In this section the engineering geological characteri-sation of these various rock type areas are discussed.The discussion concludes with the assignment of therange of values of the (GSI) which are most likely tooccur for the fundamental types of rock masses occur-ring in the ophiolites. The field data are from outcrops,cuts in slopes, borehole cores and tunnel excavationsfrom various significant ophiolitic complexes andmelanges in northern and central Greece. It is hopedthat this assignment is not simply site specific but ofgeneral value and can be applied in a more universalway, given the similarities between the geological settingof numerous ophiolitic complexes of the Alpine cyclearound the world.

The GSI values characterising the various masses,together with the strength of the intact rock, rci, and thepetrographic parameter, mi, allow the geotechnicalparameters of strength and deformability of the variousrock masses to be estimated with a level of accuracywhich is generally adequate for engineering design. Forthis estimation the program RocLab can be used. Theprogram can be downloaded free from http://www.rocscience.com.

Fig. 2 Ophiolites: synthetic and theoretical column (from Foucaultand Raoult 1995 with simplified descriptions). 1 basal contact,overthrust; 2 basal body of peridotites with foliation (tectonites),generally harzburgites with chromites (ch) layered with dunites; 3tectonic cut or confused zone, 4 dykes, sills and layers of basic andultrabasic rocks; 5 layered peridotites (cumulates of dunites,lherzolites) and magmatic breccia; 6 alternations of peridotiteswith gabbros, 7 non layered gabbros over layered gabbros, 8variety of basic rocks, dolerites, diorites and granophyres (increaseof SiO2); 9 dykes, veins, sills (s) of basic rocks; 10 and 11 basalts(spilites) with compact flows (c) or pillow-lavas (lc), a tectoniccontact is often present at the base of 10, 12 Argilites rich in Fe andMn, 13 sedimentary rocks with siliceous beds (radiolarites) over orinterlayered with the lavas (volcano–sedimentary complex). Totalthickness: usually 4–5 km, can achieve 10–15 km (as an example:11 and 10=0, 5–1 km, 8=0, 5 km, 7 and 6=0, 5 km, 5=0, 2 km,2=2–3 km). These numbers can change dramatically due tooverthrusts


Peridotites

Peridotites (hartsbourgites, dunites etc.) are strong witha range of unconfined strength for the intact mass frommany tens of MPa to more than 100 MPa at which stagethey behave as typical brittle materials. Koumantakis(1982) gives mean values of about 90 MPa from tests on130 samples of peridotites, more or less serpentinised,

from various locations in Greece. The values of rci usedin tunnelling design approaches in Greece are at least50 MPa. Their tectonic disturbance is expressed in termsof intersecting joint sets distributed in accordance withthe state of stress under which they were developed.

Serpentinisation, as a result of a typical weatheringprocess or a process of alteration from endogenic causes,can be present on the surface of discontinuities. In such

Fig. 3 Schematic section ofnappes (overthrusts) in centralAlps. Nappes of Mesozoicophiolites in black (details inPollino et al. 1990 in Mercierand Vergely 1999)

Fig. 4 The ophiolites (oph) inthe context of Andes. Model inthe Colombian Andes. (detailsin Megard 1987)

Fig. 5 Simplified section in the Himalayas with ophiolites in black (details in Bassoulet et al. 1984)

Fig. 6 Geological sketch inOrthrys mountain, centralGreece with ophiolites (P–S) intectonic relation with sedimen-tary series of rocks (limestonesand siliciferous schists) (fromMarinos 1974). Scale of tens ofmeters

P. Marinos et al.

cases the initial rough conditions of the joints are dra-matically reduced to poor or very poor with coatings ofsmooth and slippery minerals such as serpentine or eventalc.

The range of GSI for peridotitic types of rock massesof the ophiolitic complex is shown in Fig. 7 (areas 1 and2). The rock mass can be almost massive, with only a fewwidely spaced discontinuities, even close to the surface intectonically quiet areas or in zones of ‘‘tectonic sha-dow’’. High values of GSI are to be attributed to thistype of peridotitic mass (GSI greater than 65—area 1 inFig. 7). Figure 8 shows representative cores of this good

quality peroditite. Figures 9 and 10 show outcrops ofsound and weathered peridotite.

When the rock mass is jointed or fractured the GSIvalues drop as low as 35, not only due to a disturbedstructure but also because of the conditions of the dis-continuities which become smooth and slippery due toserpentinisation. In a disturbed peridotitic mass, theserpentinisation process often affects and disintegratesparts of the rock, not only contributing to lower GSIvalues but also reducing the intact strength values. Such

Fig. 7 Ranges of GSI for various qualities of peridotite–serpent-inite rock masses in ophiolitic complexes

Fig. 8 Good quality peridotite from the mountain of Orthrys incentral Greece. The conditions of the widely spaced discontinuitiesare only mildly affected by serpentinisation. Tunnel stability iscontrolled by occasional structural failures. Depth of the cores inthe photograph about 380 m. GSI 80±5

Fig. 9 Good quality surface outcrop of blocky structure inperoditite with fair (smooth, moderately weathered and altered)surface conditions of discontinuities. Width of photograph about1.5 m. GSI�55


disturbed peridotites fall in the lower bound of area 2 ofthe GSI diagram of Fig. 7 and are shown in Fig. 11.

Gabbros

The gabbros follow the same principles in their engi-neering geological characterisation as all strong rocks.If they are sound their behaviour depends on theirdegree of fracturing. Their discontinuities can havebetter conditions than those of the peridotites as they

suffer less from alteration. When weathered, the dis-integrated materials contain clay which may be highlyexpansive.

Serpentinites

When the serpentinisation is due to weathering whichhas affected all of the mass, in addition to the reductionof the intact strength there is a dramatic disintegrationof the structure of the rock mass. If this process of ser-pentinisation is due to autometamorphism and/orassociated with tectonic thrust, the rock mass is poor,with a schistose disturbed structure which may reducethe GSI to values to 30 or less (area 3 in the GSI dia-gram of Fig. 7).

Measuring the strength of the intact rock, rci, fromsuch rock masses is always a problem. When testingschisto-serpentinites, the influence of ‘‘schistosity’’ re-sults in a significant reduction in the strength of alarge proportion of the specimens. Consequently, it isvery difficult to obtain reliable values for rci fromlaboratory tests and it is suggested that the uniaxialcompressive strength of the schisto-serpentinite shouldbe estimated from that of the normal serpentinite andreduced by about 30% to account for the schistosity.From cases in northern Greece it is considered that40 MPa may be a realistic value for the uniaxialcompressive strength (rci) of the serpentinite (Fig. 12)and 30 MPa for the schisto-serpentinite. Koumantakis(1982) gives values of 45 MPa from 12 samples ofserpentine.

It is essential to differentiate between intact rockstrength and rock mass strength as this has a significantimpact on the assessment of potential tunnelling condi-tions. Assigning the rock mass a low value of GSI and alow value of intact strength penalises the rock mass twiceand results in too low a value of the estimated rock massstrength.

Fig. 10 Weak weathered ophiolitic outcrop of serpentinised peri-dotite

Fig. 11 Fair quality peridotite, from the mountain of Orthrys incentral Greece, with discontinuities of low frictional properties dueto the presence of films of seprentinised material. Blocky-jointedwith short lengths of disintegrated or serpentinised sections. Tunnelstability will be controlled by structural stability of small blocks orby mild overstressing. Depth of the cores in photograph about200 m. GSI 35±5

Fig. 12 Compact serpentinite

P. Marinos et al.

Sheared serpentinites

In the sheared zones of serpentinites there is a lack ofblockiness, which allows the rock to disintegrate intoslippery laminar pieces and small flakes of centimetres ormillimetres in size. GSI values can drop to less than 20(Fig. 7, area 4). Such sheared serpentinite is shown inFigs. 13 and 14. The intact strength may vary from 20 to5 MPa or less.

In a recent paper, Glawe and Upreti (2004) illustrateand discuss the differences that occur in two serpenti-nites, one in Turkey and one in Indonesia. Varyingstrength values may result from differences in locallithological factors, micro- and macro- structures, min-eralogical compositions, variations in interlocking ofsmaller grains, sheared and angular rock fragments andre-cementation of matrix in serpentinite with bimrockfabric (Glawe and Upreti 2004).

Pillow lavas

Pillow lavas of basaltic nature exhibit exfoliated spher-ical zones. Friable or sheared material surrounds astronger main mass thus reducing the overall quality(Fig. 15). The condition is particularly likely to occur inareas of low overburden and in the tectonic zones of theophiolitic complex. Thus, the quality can be poor onlywhen weathering and tectonic shearing is generalised.The range of the geotechnical quality of the pillow lavagiven in terms of GSI may vary from 50 to 25. The GSIchart for these pillow lava structures is given in Fig. 16.Occasional shear planes occur (Fig. 15) and in tunnel-ling these could result in local structural failures unlessadequately supported. Under such conditions tunnellingin this mass may be fair to good and only poor when therock mass is at its lower bound (sheared and weathered).However, it will be essential to maintain confinementduring the excavation procedures in order to optimisethe temporary support.

Melanges

Low to very low GSI values can be attributed to massesin ophiolitic melanges where, as discussed earlier, rocksof the ophiolitic sequences are mixed in complete

Fig. 13 Poor quality sheared serpentinite from the mountain ofOrthrys in central Greece. Completely disintegrated peridotite withloss of blockiness and presence of clayey sections. Tunnel stabilitywill be controlled by stress dependent rock mass failure withsignificant squeezing at depth. Depth of samples in photographabout 175 m, GSI 15–20

Fig. 14 Piece of sheared ophiolite, which has been disintegratedinto flakes of weak serpentinite

Fig. 15 Basic volcanics in pillow lava structure. Blocky disturbedwith poor condition of discontinuities (highly weathered withcoatings or fillings). GSI approximately 30 in this particular site. Athin shear plane is also present (indicated by the notebook)


Fig. 16 Range of GSI ratings for basaltic pillow lavas. (Warning: the shaded area indicates the ranges of GSI values most likely to occur inthis type of rock. It may not be appropriate for a particular site specific case)

P. Marinos et al.

disorder with other rocks of various origins and aresituated at the base or in the front of great ophioliticoverthrust nappes. In soil-like material a GSI assign-ment is meaningless. However, it is possible that duringtunnelling inside these masses, extensive blocks of sedi-mentary rocks of good engineering quality (e.g. lime-stones or sandstones) can be encountered. Nevertheless,the transition to the surrounding sheared rock massof either ophiolitic or other clayey sedimentary rocks(flysch, siltstones, argillites) is sharp and unpredictable.Probe drilling ahead of the face is always prudent whentunnelling in such conditions.

The mi values

In the Hoek and Brown failure criterion the mi valuereflects the frictional characteristics of the componentminerals and grains of the intact rock. In the ophioliticrocks in the Greek Alpine context the mi values can be:peridotites more than 20; schistose serpentine: 12±2;altered material due to shearing: 8±2.

Behaviour in tunnelling

The great variety of the rock mass types, the irregularchanges and the alteration make the ophiolites a for-mation where extreme care is needed in the design of anyengineering structure founded on or crossing them. Thisis particularly true for tunnels as their linearity and theirdepth increase the possibility of encountering the ad-verse conditions and weak zones associated with theophiolites while the uncertainty as to their occurrenceand extent exacerbate the difficulty. In Fig. 17 the tun-nelling behaviour of peridotites—serpentinites is classi-fied following the characterization of their rock massesdiscussed in the previous section and shown in Fig. 7.

Peridotites

In good quality masses of peridotite, simple straightforward tunnelling conditions can be expected. Atten-tion has to be concentrated on avoiding structuralinstabilities from wedges. For these structurally con-trolled failures involving only a few discontinuities, theproblem is essentially one of three-dimensional geometryand stereographic techniques or numerical analyses suchas Unwedge (see http://www.rocscience.com) should beused for an analysis of failure processes and the designof reinforcement. However, compared with other rockmasses of similar structure, the peridotites generallyhave smoother discontinuities with low frictional prop-erties. As explained earlier this is due to the presence ofserpentinised material which is very often present even ifthe serpentinisation has not affected the fundamental

rock material. This makes the structurally dependantinstability more critical and generally demands heavierrock bolting patterns and/or thicker shotcrete (zone II inFig. 17). In very hard massive rock masses at greatdepths, spalling, slabbing and rockbursting are themodes of failure that may develop, controlled by brittlefracture propagation in the intact rock with the dis-continuities having only a minor influence (zone I ofFig. 17). In these cases the use of brittle rock failurecriterion should be considered, such as that proposed byKaiser et al. (2000).

Fractured peridotites or schistose serpentinites

In the case of a more fractured peridotite, schistose orweaker serpentinite (GSI values of 25–40), the behaviouris controlled by sliding and rotation on discontinuitysurfaces with relatively little failure of the intact rockpieces (zone II/III of Fig. 17). In this range of GSIvalues the RQD values can be very low. This is normal,given the structure of the rock masses, but some of thefrictional behaviour of the unaltered pieces of the mass isretained. Thus, the control of stability can be effectivelyimproved during excavation of the tunnel by keeping therock mass confined.

Sheared serpentinite—squeezing behaviour

In the poor quality serpentinite, due either to weatheringor shearing, blockiness may be almost completely lostand clayey sections with swelling materials may bepresent. Tunnel stability will then be controlled by stressdependant rock mass failure with significant squeezingat depths (Fig. 17, zone III). In these cases detailed de-sign has to be carried out using a numerical analysiswhich permits progressive failure and support interac-tion analysis to be modelled. However, it is veryinstructive to carry out a closed form analysis of thebehaviour of the tunnel in order to get some idea of thesignificance and magnitude of convergence and squeez-ing. The plot presented in Fig. 18 is taken from a paperby Hoek and Marinos (2000) in which it was shown that,for tunnels in weak rocks, the ‘‘strain’’ can be estimatedfrom the ratio of rock mass strength to in situ stress bymeans of the equation shown in the figure. This plot isfor single circular shaped tunnels. The strain for twintunnels which are reasonably close together is expectedto be higher than that indicated by the plot, which is forunsupported tunnels in a hydrostatic stress field.

Two cases from a tunnel in Greece in ophiolites in theform of more or less sheared serpentinites are plotted inpoints 7 and 8 in Fig. 18. Point 7 corresponds to a rockmass of a strength, rcm, of about 1.4 MPa and adeformation modulus, E=800 MPa, under a cover of


Fig. 17 Classification of thebehaviour in tunnelling forperidotite–serpentinite rockmasses in ophiolitic complexes(to be read in conjunction withFig. 7)

P. Marinos et al.

about 150 m. These values were derived from a backanalysis of the behaviour of a tunnel section where themeasured and computed displacements of about 1.5%were in close agreement. These properties correspond toa weak ophiolitic rock mass (sheared serpentinite) with aGSI=20, rci=16 MPa and mi=10.

For a depth of 500 m, the ratio of rock mass strengthto in situ stress is 0.11 and this gives a strain of 17% fora single tunnel. This means that there may be a closureof as much as 2 m in a 12 m span tunnel unless appro-priate steps are taken to control this deformation.

Squeezing and face stability can be controlled byforepoling, commonly used in many weak rock tunnelsin southern Europe. However, when severe squeezing isanticipated in very weak serpentinite in areas of thickcover, the use of yielding primary support (sliding jointsin steel sets or gaps in shotcrete) in conventional tun-nelling may be required. In these cases the ideal tunnelsection is circular. Where such difficult tunnelling con-ditions are encountered it is recommended that a robustdesign be provided with the possibility of varying theamount of yielding depending on the overburden andthe rock mass quality. In general, it is recommended thatthe chosen support types should be able to accommo-date changes without the need to change the principalelements of the support system.

It is unlikely that rockbolts will be effective in severesqueezing conditions as they are too stiff in relation tothe surrounding rock mass and the resulting strain dif-ferential causes shearing of the grout bond. Ravelling ofcompletely disintegrated serpentinite may also be aproblem and keeping confinement of the face is the keyaction to be undertaken.

Special provisions may be required to eliminate thepossibility of trapping the machine in the case of exca-vation by TBM. It is critically important that a TBMshould never be stopped in zones of severe squeezing.Experience has shown that this squeezing occurs rela-tively slowly and that a moving machine is seldomtrapped. Overboring devices are necessary and in somecases TBMs have been specially designed to permitreduction of the shield diameter. However, several ma-chines have been lost when they have been parked forone reason or another. An appropriately designed pre-cast segmental lining is generally installed directly be-hind the machine.

As the occurrence of weak or sheared rock masses israndomly distributed along the tunnel alignment due tothe particular geological model of the ophiolites, it isprudent to probe ahead of the tunnel face continuously,keeping a length of probe hole of at least 20 m ahead ofthe face at all times. In the case of TBM excavation,

Fig. 18 Plot of percentagestrain versus the ratio of rockmass strength to in situ stress(after Hoek and Marinos 2000).Calculated and predictedstrains for a tunnel in ophioliteare plotted as points 7 and 8 atratios of rock mass strength toin situ stress of 0.35 and 0.11,respectively


provision should be made for probe drilling through thecutter head or by means of inclined holes drilled over theshield.

Weathered peridotite. A case of ravelling

Tunnelling through weathered peridotite (serpentinite)close to the surface will require great care in order to

avoid subsidence and slope movement and a light fore-pole umbrella (75 or 100 mm diameter pipes) can beused. Pre-grouting an umbrella in the rock mass over theforepoles may be advisable in order to increase thecohesive strength of the rock mass. Figures 19 and 20illustrate a failure in weathered peridotite.

The tunnel face has been stabilised by the installationof a double forepole umbrella and by extensive groutingthrough the forepoles and also through horizontal holesdrilled through the muckpile. Figure 20 shows a situa-tion in which no material has been removed from themuckpile which has been covered with shotcrete.

Groundwater conditions

Groundwater can be present in fractured but otherwisesound peridotites. However, weak masses are usually oflow permeability; thus, water pressure has to be con-sidered in the design and systematic relief holes may berequired during construction.

Conclusions

Rock masses in an ophiolitic complex exhibit a widerange of engineering behaviour, particularly in tunnel-ling. This is due to their petrographic variety andstructural complexity. The GSI, enhanced by geologicallogic permits the characterisation of this wide range. Itsuse allows adjustments of the rating to cover not onlythe wide range of the ophiolitic rock masses and con-ditions but also the variations that may occur within thesame rock type. As a consequence a classification oftunnelling behaviour can be formulated for this widerange of ophiolitic rock masses from stable to severesqueezing conditions.

Acknowledgements We acknowledge the Operational Programmefor Educational and Vocational Training (EPEAEK) and partic-ularly the research programme ‘‘Pythagoras’’ for the financialsupport of this research; this project is co-funded by the EuropeanSocial Fund (75%) and National Resources (25%). We alsoacknowledge the opportunities provided by Ergose S.A. andEgnatia Odos S.A. to work on ophiolitic complexes. The assistanceof Professor E. Mposkos for the petrographic characterisation ofsamples of ophiolites is appreciated.

Fig. 19 Failure and interpretation of possible extent of rock massdisturbance resulting from the collapse due to ravelling ofweathered peridotite during tunnelling in a heavily disturbedophiolitic complex in Greece. The overburden is about 35 m.Forepoling and a grout umbrella for remediation are shown

Fig. 20 Appearance of stabilised muckpile at the face. No materialhas been removed since the collapse and stabilisation has beencarried out by the installation of a double forepole umbrella and bygrouting

P. Marinos et al.

References

Barton NR, Lien R, Lunde J (1974) Engi-neering classification of rock masses forthe design of tunnel support. RockMech 6(4):189–239

Bassoullet JP et al. (1984) L’ orogene Hi-malayen au Cretace superieur. MemSoc Geol Fr 147:9–20

Bieniawski ZT (1973) Engineering classifi-cation of jointed rock masses. Trans SAfr Inst Civ Eng 15:335–344

Debelmas A, Mascle G (1997) Les grandesstructures geologiques. Dunod, Paris,312 pp

Foucault A, Rault JF (1995) Dictionnairede geologie. Masson, Paris

Glawe U, Upreti B (2004) Better under-standing of the strengths of serpentinitebimrock and homogeneous serpentinite.Felsbau Rock Soil Eng 22(5):53–60

Hoek E (1994) Strength of rock and rockmasses. News J Int Soc Rock Mech2(2):4–16

Hoek E, Marinos P (2000) Predicting tun-nel squeezing problems for weak heter-ogeneous rock masses. Tunnels andTunnelling November and Decemberissues pp 45–51 and pp 33–36

Hoek E, Marinos P, Benissi M (1998)Applicability of the geological strengthindex (GSI). Classification for weak andsheared rock masses—The case of theAthens Schist formation. Bull Eng GeolEnv 57(2):151–160

Hoek E, Marinos P, Marinos V (2005)Characterization and engineering prop-erties of tectonically undisturbed butlithologically varied sedimentary rockmasses. Int Rock Mech Min Sci

Kaiser PK, Diederichs MS, Martin D,Sharp J, Steiner W (2000) Undergroundworks in hard rock tunnelling andmining. In: Proceedings of the Geo Eng2000, international conference on geo-technical and geological engineering,Melbourne, Technomic publishers,Lancaster, pp 841–926

Koumantakis J (1982) Comportement desperidotites et serpentinites de la Greceen travaux public. Leur propretes phy-siques et mechaniques. Bul IAEG25:53–60

Marinos G (1974) Geology of Orthrys andissues on its ophiolites. Ann Geol dPays Helleniques, 26, University ofAthens, pp 118–148

Marinos P, Hoek E (2000) GSI: a geologi-cally friendly tool for rock massstrength estimation. In: Proceedings ofthe GeoEng2000, international confer-ence on geotechnical and geologicalengineering, Melbourne, Technomicpublishers, Lancaster, pp 1422–1446

Marinos P, Hoek E (2001) Estimating thegeotechnical properties of heteroge-neous rock masses such as flysch. BullEng Geol Env 60:82–92

Marinos V, Marinos P, Hoek E (2005) Thegeological strength index—applicationsand limitations. Bull Eng Geol Environ64:55–65

Megard P (1987) Cordilleran Andes andMarginal Andes. In: Monger J, Fran-cheteau J (eds) Geodynamic series,American Geophysical Union, No. 18,pp 71–95

Mercier J, Vergely P (1999) Tectonique.Dunod, Paris

Pantazis TH (1973) Geology of Troodosmountain, Cyprus. Bull Geogr Assoc C,(3):5–6, 5–155

Pe-Piper, Piper D (2002) The igneous rocksof Greece. The anatomy of an orogeny.Gebrueder Borntraeger, Berlin

Skemp A, McCaig M (1984) Origins andsignificance of rocks in an imbricatethrust zone beneath the Pindos ophio-lite, northwestern Greece. In: Dixon JE,Robertson HF (eds) geological evolu-tion of eastern mediterranean, SpecialPublication 17, Geological SocietyLondon, pp 569–580

Visser W (1980) (ed) Geological nomen-clature. Royal Geological and MiningSociety of The Netherlands, Bohn,Scheltema and Holkema, Utrecht


Variability of the engineering properties V. Marinos of...

Documents

Transcript of Variability of the engineering properties V. Marinos of...