DESIGN OF SPRAYED CONCRETE FOR UNDERGROUND SUPPORT · Recommendations for the design of sprayed...

AFTES RECOMMENDATIONS FOR THE

DESIGN OF SPRAYED CONCRETE

FOR UNDERGROUND SUPPORT

2

Document presented by Eric LECA (SCETAUROUTE) - Chairman Work Group 20

with the participation of :Jean LAUNAY (GTM-DUMEZ) co-chairman - Philippe AUTUORI (BOUYGUES) -

Claude BASCOULERGUE (CAMPENON-BERNARD) - Alain BERNARDET (EEG - SIMECSOL) - Daniel DUPUCH (SIKA) - Gilles GASPAR (SNCF) - Alain GUILLOUX (TERRASOL) - Jean PIRAUD (ANTEA) -Michel PRE (SETEC) - Adrien SAITTA (CETU) - Laurent SAMAMA (SCETAUROUTE) - Joël SOLARD (RATP) -

Christophe VIBERT (COYNE & BELLIER) - Jacques WILLOCQ (LAFARGE)

Technical committee reviewers:Guy COLOMBET (COYNE & BELLIER) - Pascal DUBOIS (CETU) - Bernard GODINOT (GTM) and Yann LEBLAIS (EEG - SIMECSOL)

AFTES welcomes comments on this paper

Version 1 - Approved by Technical Committee 9 november 2000

PagesPages

1 - INTRODUCTION - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 31.1 - THE THREE PRINCIPAL ACTION MODES OF SPRAYED

CONCRETE - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 31.2 - PRESENTATION OF THE RECOMMANDATIONS - - - - - - - - - 4

2 - THE THREE TYPES OF SPRAYED CONCRETE SUPPORT- - - - 42.1 - TYPE 1 : SPRAYED CONCRETE AS A PROTECTIVE LAYER - - 42.2 - TYPE 2 : SPRAYED CONCRETE AS A STRUCTURAL LAYER

ASSOCIATED TO REINFORCEMENT - - - - - - - - - - - - - - - - - 52.3 - TYPE 3 : SPRAYED CONCRETE AS A STRUCTURAL RING - - 6

2.3.1 - Limitation of the ground’s elasto-plastic deformations- - 62.3.2 - Confinement in grounds that cannot be bolted- - - - - - - 62.3.3 - Confinement of swelling grounds (clays anhydrites) - - - 62.3.4 - Repair of older tunnels- - - - - - - - - - - - - - - - - - - - - - - - - 6

3 - RHEOLOGY OF SPRAYED CONCRETE - - - - - - - - - - - - - - - - - 63.1 - SPRAY CONCRETE: REQUIRED COMPOSITION - - - - - - - - - 6

3.1.1 - Aggregates- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 63.1.2 - Cement- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 63.1.3 - Water - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 63.1.4 - Fibres - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 73.1.5 - Admixtures - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 73.1.6 - Additives- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 7

3.2 - REQUIRED MECHANICAL CHARACTERISTICS FOR THEDIFFERENT TYPES OF SPRAYED CONCRETE - - - - - - - - - - - 7

3.2.1 - Type 1 sprayed concrete (protective layer) - - - - - - - - - - 73.2.2 - Type 2 sprayed concrete (structural layer) - - - - - - - - - - - 73.2.3 - Type 3 sprayed concrete (protective ring) - - - - - - - - - - - 7

3.3 - LABORATORY TESTS - METHODOLOGY- - - - - - - - - - - - - - - 8

CONTENTSCONTENTS

PREFACE

T his document presents the recommendations prepared by AFTES Work Group #20 for the design of sprayed concrete used in tun-nels and underground openings. The proposed approach accounts for the different categories of sprayed concrete used in tunne-ling and underground works. The document presents the state-of-the-art and practice at the moment of its publication. It shall be

revised in time to reflect the evolution of technologies and design concepts.

The following three types of sprayed concrete support are considered in this recommendation: protective layer, structural layer, and structuralring. Cases where a structural layer is used in association with other support elements (mesh, ribs, bolts,…) are also addressed, but solely froma sprayed concrete design perspective; design procedures for support systems primarily consisting of bolts, whether these are associatedwith a protective layer of shotcrete or not, will be dealt with in a forthcoming specific recommendation.

This document is based, amongst other things, on material published in literature or derived from experimental works on the characterizationof the mechanical properties of concrete, as well as in situ monitoring and design case histories. This material is documented in articles publi-shed in the present issue of the AFTES magazine, “Tunnels et Ouvrages Souterrains”.

3

Recommendations for the design of sprayed concrete for underground support

1 - INTRODUCTIONThe use of sprayed concrete as a supportelement in tunnels became popular inEurope in the early 1960s. Meanwhile, cal-culation methods aimed at a better repre-sentation of the ground - structure interac-tion were being developed (the idea ofusing a combination of sprayed concreteand rock bolts was introduced in 1956 bythe Austrian engineer L. von Rabcewiczduring the construction of highway tunnelsnear Caracas). These approaches, which aimat making the best use of the ability of theground to “support itself”, assume that the

primary role of the support system is to pro-vide ground confinement immediately afterexcavation, thus helping the ground to sta-bilize itself around the opening.

1.1 - THE THREE PRINCIPALACTION MODES OFSPRAYED CONCRETE

Depending on the mechanical properties ofthe ground, and on the dimension anddepth of the opening, sprayed concrete canserve one of the following purposes (Heuer,1974):

• (1) For lightly fractured grounds, of highstrength in comparison to the naturalstresses they are subject to before excava-tion, the role will be that of a simple “pro-tective skin”, thus preventing physical,hydrological, or chemical alteration pheno-mena to occur within the exposed ground;

• (2) For relatively less resistant grounds, anadditional level of reinforcement will berequired to support the ground around theopening, with this being generally achievedthrough systematic bolting of the openingwalls: this will result in a “structural layer”;

PagesPages

3.4- CONSEQUENCES OF THE EVOLUTION OF YOUNG SPRAYED CONCRETE PROPERTIES - - - - - - - - - - - - - - - - - - 8

3.4.1 - Studies on fresh sprayed concrete behavior - - - - - - - - - 83.4.2 - Other studies dedicated to the creep capacity of young

sprayed concrete- - - - - - - - - - - - - - - - - - - - - - - - - - - - - 93.5 - EFFECTS OF SHRINKAGE- - - - - - - - - - - - - - - - - - - - - - - - - - 9

4 - DESIGN OF SPRAYED CONCRETE FOR UNDERGROUNDSUPPORT- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 104.1 - REVIEW OF AVAILABLE DESIGN METHODS - - - - - - - - - - - - 10

4.1.1 - GENERAL Considérations - - - - - - - - - - - - - - - - - - - - - - 104.1.2 - Methods for load evaluation - - - - - - - - - - - - - - - - - - - - 104.1.3 - Verification of the structural stability - - - - - - - - - - - - - - - 104.1.4 - Main characteristics of standard design methods - - - - - 10

4.2 - TYPE 1 SPRAYED CONCRETE - PROTECTIVE LAYER - - - - - - 114.2.1 - Description - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 114.2.2 - Technical recommendations - - - - - - - - - - - - - - - - - - - - 124.2.3 - Objectives- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 124.2.4 - Design - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 12

4.3 - TYPE 2 SPRAYED CONCRETE - STRUCTURAL LAYER - - - - - 124.3.1 - Description - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 124.3.2 - Technical recommendations - - - - - - - - - - - - - - - - - - - - 124.3.3 - Objectives- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 124.3.4 - Design - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 12

4.4 - TYPE 3 SPRAYED CONCRETE - STRUCTURAL RING - - - - - - 134.4.1 - Description - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 134.4.2 - Technical recommendations - - - - - - - - - - - - - - - - - - - - - 134.4.3 - Objectives- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 134.4.4 - Design - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 13

4.5 - CONSIDERATION OF SPRAYED CONCRETE INTHE FINAL LINER DESIGN - - - - - - - - - - - - - - - - - - - - - - - - - 15

4.5.1 - Introduction - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 154.5.2 - Sprayed concrete not included in the final liner design - 154.5.3 - Recommendations for the consideration of sprayed

concrete in the final liner design- - - - - - - - - - - - - - - - - - 154.5.4 - Recommendations for the evaluation of stresses in the

structures.- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 164.5.5 - Plain or reinforced shotcrete - - - - - - - - - - - - - - - - - - - - 16

5 - MONITORING- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 16

5.1 - GENERAL BACKGROUND - - - - - - - - - - - - - - - - - - - - - - - - - 16

5.2 - MONITORING OBJECTIVES AND MEANS - - - - - - - - - - - - - 16

5.2.1 - Deformation measurements- - - - - - - - - - - - - - - - - - - - - 165.2.2 - Stress measurements- - - - - - - - - - - - - - - - - - - - - - - - - - 17

5.3 - MONITORING SECTIONS- - - - - - - - - - - - - - - - - - - - - - - - - - 17

5.3.1 - Standard monitoring sections - - - - - - - - - - - - - - - - - - - 185.3.2 - Enhanced monitoring sections - - - - - - - - - - - - - - - - - - - 185.3.3 - Spacing of the sections and frequency of

measurements - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 18

5.4 - INTERPRETATION AND BACK CALCULATION - - - - - - - - - - 18

5.4.1 - General background - - - - - - - - - - - - - - - - - - - - - - - - - - 185.4.2 - Convergence measurements - - - - - - - - - - - - - - - - - - - - 185.4.3 - Extensometers placed within the ground - - - - - - - - - - - 195.4.4 - Stress measurements- - - - - - - - - - - - - - - - - - - - - - - - - - 195.4.5 - Back-analyses - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 19

6 - STRUCTURES NOT LINED WITH CAST-IN-PLACE CONCRETE 19

6.1 - INTRODUCTION - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 19

6.2 - AREA OF APPLICATION - - - - - - - - - - - - - - - - - - - - - - - - - - - 20

6.3 - SPECIFIC LIMITAITONS - - - - - - - - - - - - - - - - - - - - - - - - - - - 20

6.3.1 - Water-tightness - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 206.3.2 - Surface roughness - - - - - - - - - - - - - - - - - - - - - - - - - - - - 206.3.3 - Heterogeneity and variability of concrete properties- - - 206.3.4 - Durability - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 21

6.4 - CONCRETE COMPOSITION - - - - - - - - - - - - - - - - - - - - - - - 21

6.5 - DESIGN OF THE FINAL SPRAYED CONCRETE LINER - - - - - 21

CONTENTSCONTENTS

4


• (3) Finally, in some cases (e.g. shallowtunnels), sprayed concrete, more oftenreinforced, will be required to play a moresignificant supporting role, and will in suchcase need to be designed as a true “struc-tural ring”.

Each of these roles corresponds to a diffe-rent mode of action. This leads to the consi-deration of three types of sprayed concretesupport systems for use in undergroundconstruction:

• Type 1: sprayed concrete as a protectivelayer; in this case, sprayed concrete acts asa cement, thus ensuring cohesion to deve-lop between ground particles and/or rockfragments and preventing the ground de-structuring to occur. This support system islimited to surficial action, a few millimetersto a decimeter in depth, and is not meantto carry any load.

• Type 2: sprayed concrete as a structurallayer; this second type of support must bedesigned as a composite structure, usingthe combined action of a ring of reinfor-ced ground and a layer of sprayedconcrete. In these conditions, the sprayedconcrete supports the ground, maintainsits cohesion over a limited depth (rangingfrom a few decimeters to a meter), andacts as a “bridge” between successivesupport elements (such as rock bolts). It ismainly subjected to shear and has to bereinforced with welded wire mesh, fibersor steel girders.

• Type 3: sprayed concrete as a structuralring; in this last case, sprayed concrete hasto be designed as a structure per say, andhas to be able to resist normal forces andbending moments. As with the sprayedconcrete of Type 2, reinforcement isnecessary.

These three modes of action are not exclu-sives nor fully delimited: in particular,sprayed concrete which is meant to serve as“structural layer” always starts to act as a“protective layer”. Likewise, a “structurallayer” can act locally as a “structural ring”,and provide some confining pressure,although the ground is not in a post failurestate over the entire periphery of the exca-vation. At shallow depth, one cannot fullyrely on the designer’s ability to predict if andwhere localized failure will take place. Onefundamental benefit of sprayed concrete isprecisely that it is uniformly applied to theentire tunnel surface, and therefore able to resist at an early stage to defects, hetero-geneities or failures within the ground,

which are randomly distributed and can playa determinant role in the overall stability ofthe opening.

1.2- PRESENTATION OF THERECOMMENDATIONS

The objective of the present document is toput forward some elements of design foreach of the three types of sprayed concreteused in underground support systems. It fol-lows three recommendations of AFTESWork Group #6 which contain backgroundinformation on technological and construc-tion aspects: "Mise en oeuvre du béton pro-jeté dans les travaux souterrains" (Tunnelset Ouvrages Souterrains no1, 1974),"Présentation de la méthode de construc-tion des tunnels avec soutènement immé-diat par béton projeté" (Tunnels etOuvrages Souterrains no31, 1979), and"Technologie et mise en oeuvre du bétonprojeté renforcé de fibres" (Tunnels etOuvrages Souterrains no126, 1994).Additional material can also be found in twomore recent publications in the UK by theInstitution of Civil Engineers (1996) and theHealth and Safety Executive (1996), whichprovide a description of sprayed concreteapplications as a support element in tunnelsand recommendations on the design andconstruction of these structures, with focuson shallow tunnels.

This document covers all uses of sprayedconcrete in underground works: support ofgallery or shaft walls, temporary support atthe tunnel face during excavation, eitheralone or in conjunction with wiremesh, boltsor fibers. However, the design of steel ribsand bolts – which represents, in some appli-cations, an essential element of support –will not be addressed in this document. Onthe other hand, information will be providedon the use of sprayed concrete as final tun-nel liner.

Besides the present introduction, theserecommendations consist of five chapters.The three modes of action of sprayedconcrete presented in §1.1 are described inChapter 2. Chapter 3 reviews the principalaspects of sprayed concrete behavior; ele-ments of concrete composition and requi-red mechanical characteristics for each typeof sprayed concrete are also considered.The recommendations for the design ofsprayed concrete support systems are deve-loped in Chapter 4, with due considerationof all three usage types described above.Aspects related to the inclusion of sprayedconcrete into the design of final liners are

also treated. Given the role played by instru-mentation (especially convergence monito-ring) in the design of sprayed concrete sup-port systems, it was felt necessary to reviewthe principles of monitoring; this is discus-sed in Chapter 5. Finally, Chapter 6 coversthe situations where sprayed concrete isused as final liner.

2- THE THREE TYPES OFSPRAYED CONCRETE SUPPORTThis Chapter describes in detail the threetypes of spayed concrete support introdu-ced in §1.1. It is primarily intended to clarifythe field of application of each of the threetypes of support previously identified, aswell as their acting mechanisms.

2.1- TYPE 1: SPRAYEDCONCRETE AS A PROTECTIVE LAYER

The first layer of shotcrete placed immedia-tely after excavation, with a thickness of 2 to5 cm, is essentially meant to protect theground against any surface alteration.Indeed, this weathering can have detrimen-tal effects on the geotechnical properties ofthe ground:

• Dessication as initially saturated groundsbecome exposed to ambient air, thus pro-ducing a loss of cohesion (e.g. of capillarytype) within the first centimeters of ground;

• Washout of finer materials at discontinui-ties which presence controls the shear resis-tance of joints;

• Degradation of the hydromechanical pro-perties of the ground at the opening;

• Initiation of swelling phenomena associa-ted with the migration of pore water.

This first layer of sprayed concrete also has avery local (centimeter scale) mechanicaleffect. This role is vital because it restrainsmicro displacements, which may in turn tolead to micro failures and subsequent in-depth degradation of the rock characteris-tics (loss of cohesion) through some "chainreaction" type of mechanism: sand particlesleft free to move will free the stones thatlock the block, and so on (cf. figure 2.1).

Obviously, this weathering mechanism willstart developing as soon as the ground isexposed to air and/or water, with moremechanical degradation occurring as thevolume of excavated material increases. Forthis reason, spraying of concrete must startas early as possible, particularly in grounds

5


sensitive to such weathering: heteroge-neous and non cohesive grounds, clays andshales, highly fractured rocks, and swellinggrounds. It is often recommended to startthe spraying before the completion of theexcavation cycle, including over the facethat will experience similar weatheringeffects as the tunnel walls once it is expo-sed. It must be noted that the mechanicaleffect expected from the shotcrete layer isnot structural resistance but surface cemen-tation.

2.2- TYPE 2: SPRAYEDCONCRETE AS A STRUCTURALLAYER ASSOCIATED WITHREINFORCEMENT

In reality, the ground has already experien-ced some degree of stress release when theconcrete is sprayed: as the ground startsloosening at the opening, stresses are trans-ferred to the underlying grounds, thus crea-ting an area of loosened material thatrequires support. Even in the absence ofsurface weathering, grounds such as fractu-red or stratified hard rocks may lead to simi-lar mechanisms due to unfavorably orientedcracks allowing blocks or wedges of unsup-ported material to separate from theground mass.

Similarly to the first case (Type 1), the struc-tural layer of sprayed concrete has the effectof cementing opened cracks and lockingunstable blocks. This role is similar to thatplayed by mortar in a masonry arch: it isessential for stability, unless the basic buil-ding blocks fit perfectly, which is rarely thecase in practice. Besides this stabilizing

effect, sprayed concrete acts locally as astructural layer to prevent differential move-ments of the blocks from occurring:

• Either as a protection net, for large spans(which requires the sprayed concrete to bereinforced). This is the case of the vaults oflarge underground openings supportedwith massive rock bolting;

• Or as a compressive shell, when the curva-ture of the opening is small enough for theconcrete to work in compression. This is thecase of small arches of sprayed concretethat are placed between closely spaced lat-tice girders to transfer stresses to the steelelements (using a principle similar to that ofsoldier piles).

Thus, the main role of a structural layer ofsprayed concrete is to resist local failuresand small displacements of blocks thatcould propagate to the surroundingground. This type of support starts actingon the ground located between the faceand the last row of installed rock bolts,then between two consecutive rows ofrock bolts and, when necessaryon the excavation face.

The second type ofsprayed concrete sup-port must thereforeoffer shear strengthand even tensilestrength when theexcavation profileis irregular or flat(e.g. at the face),or even locallyconvex. This justi-fies the use of fibers

or wire mesh. The same applies when theshell of sprayed concrete is designed toact as a small arch between lattice girders.The area of influence of the shotcrete layer(at the surface and in depth) ranges bet-ween a few decimeters to a meter, i.e. afraction of the distance between bolts orlattice girders.

Beyond such depth, ground support is pri-marily achieved by the effect of bolts whichprevent plastic deformations to developwithin the ground mass as a result of exca-vation induced stress release. Such defor-mations, if permitted, could result into pro-gressive failure along weak planes presentwithin the ground mass. Rock bolts canhowever not act on the area of loosenedmaterial that is left unsupported at the sur-face of the opening, as a result of archingeffects developing between the bolts. Therole of Type 2 sprayed concrete is exactly tosupport this area. Because of its creep cha-racteristics, this concrete is flexible enoughat the early age to allow the ground tohomogeneously deform, without creatingany permanent damage in weaker areas.

However, given its limited thickness andirregular distribution Type 2 sprayedconcrete cannot be relied upon for transfer-ring stresses over a larger area of the ope-ning. Higher stresses will result in the forma-tion of hinges in the fresh concrete layer. Infact, despite its apparent continuity, the“structural layer” of sprayed concrete canbe expected to act as a set of shell ele-ments, bound by hinges, and held by rockbolts. In this respect, it is often compared tothe coating plates used with reinforcedearth, to contain the ground in the imme-diate vicinity of the vertical face.

Structural layer types of usage can also befound in the form of grooved shotcreteliners, in which longitudinal recesses, 20 to30 cm wide, are cut along the shotcreteshell, with the wire mesh being left uncove-

red at the recesses(figure 2.2). Thistechnique has beenused in Austrian deeptunnels in order toallow large conver-

Figure 2.1 - Wheathering mecanisms at excavation faces

Figure 2.2 - Longitudinal recesses

Fault

Wateringress

2 to 5

Resistance against micro displacement of blocks in zonesunreinforced by rock botls

Resistance against shearingaction along weak planes

Sprayed concrete preventingfall-out of small particles

Prevention against dessicationand washout of surface

20 to 30 cm wide recesses not

fillled with concrete

Sprayed concreteAppearance of the recess after convergence

Welded wice mesh

3 to 4 cm

6


gences of the tunnel walls to occur (severaltens of centimeters) without incurring anydamage to the elementary shells of sprayedconcrete.

2.3- TYPE 3: SPRAYEDCONCRETE AS A STRUCTURAL RING

The layer of sprayed concrete used as astructural ring must be continuous andconcave, which is not necessarily the casewith Type 2 supports. The layer's curvaturemust also be as regular as possible: it shouldbe designed as an arch with sufficient resis-tance to sustain all applied loads. This ringmay be closed or opened at the invert,depending on the level of in situ stresses. Itsability to maintain the stability of the wholeopening of course relies on the tunnel beingstabilized, if required, by appropriate meansof support.

Sprayed concrete as a structural ring is pri-marily used in four situations which are sum-marized in the following sections.

2.3.1- Limitation of the ground’selastic-plastic deformations

In some heterogeneous grounds and highlyfractured rocks, the stability of the openingafter excavation cannot be reached withoutthe rapid placement of confinement that willprevent excessive deformations to occur.The main objective of the structural ring ofsprayed concrete is to contribute to therequired confinement level through itscontinuous contact with the ground.

The required level of confining pressure canbe reduced if a certain degree of conver-gence is tolerated, provided the correspon-ding deformations will not damage theground’s characteristics. Due to the highlevel of creep that develops in youngsprayed concrete, these deformations cantake place without damaging the concrete,while maintaining the integrity of the sur-rounding ground and preventing failure tooccur in the weaker areas.

The structural ring of sprayed concrete can,in this case, be compared to a structurallayer of Type 2, but of higher and moreregular thickness, and the ability to resisthigher thrusts. It can be considered, in someways, as a thin shell element acting over theentire vault, or the entire section in the caseof a circular tunnel. It should be noted thatthe larger the radius of the tunnel, the largerthe shotcrete thickness required to achievethe desired level of confinement.

A particular case of a structural ring applica-tion is that of shallow tunnels in weakgrounds where it is imperative to limitground deformations. The role of the shot-crete ring in that case is to block the ground,even if it means resisting to higher stresses.

The longitudinal stiffness brought by thesprayed concrete shell – especially when it isreinforced by wire mesh, or associated togirders (lattice girder, light or heavy ribs) orrock bolts - can also contribute to controlground deformations in cases where stagedexcavation is used or where weak groundsare encountered.

2.3.2- CONFINEMENT IN GROUNDSTHAT CANNOT BE BOLTED

Most of the time, sprayed concrete rings areassociated with rock bolts that limit theconvergence, so that ring thrusts are nothigher than the level that can be withstoodby the structure. However, the ground canin some cases not be bolted, either becauseof geometrical reasons or of the nature ofthe ground. In those cases, the concretering is left alone to support the ground, andone must ensure that the loads carried bythe structure do not exceed its capacity,particularly if the ground pressure tends, inthe long run, to reach the initial geostaticpressure, therefore limiting the use of thistype of support to shallower tunnels.

2.3.3- CONFINEMENT OF SWELLINGGROUNDS (CLAYS, ANHYDRITES)

Due to its continuous contact with theground, spayed concrete allows to limitwater seepage towards the tunnel walls,which is one of the main factors in the deve-lopment of swelling phenomena.

2.3.4- REPAIR OF OLDER TUNNELS

A continuous ring of reinforced sprayedconcrete, placed against a regular masonryvault, can in some cases present a viableeconomical refurbishment solution evenwith a small thickness, due to the regularityof its profile (provided the reinforcing wiremesh is anchored).

3- RHEOLOGY OF SPRAYEDCONCRETEDepending on the expected effect and jobsite conditions, sprayed concrete will playdifferent roles (cf. §1.1).

Some characteristics will require specificattention in order to ensure that the expec-

ted effect is achieved. Particular composi-tions (choice of cement…) should be used incases where sprayed concrete is taken intoaccount in the final liner design, or if it isused as a long term support element. Thedifferent concrete constituents should res-pect a number of precise guidelines in orderto achieve mechanical characteristics thatare sufficiently reliable to be taken intoaccount in the design.

3.1- SPRAYED CONCRETE:REQUIRED COMPOSITION

The most recent guidelines concerningsprayed concrete composition are descri-bed in the recommendations on the techno-logy and the application of fiber reinforcedsprayed concrete – prepared by the AFTESWork Group #6 (Tunnels et OuvragesSouterrains no 126, 1994) – which should bereferred to for more information. The follo-wing sections review the main requirementsfor achieving the minimal characteristicsrequired for sprayed concrete in under-ground works.

3.1.1- AGGREGATES

The aggregates must comply with currentstandards. The aggregate distribution mustlie within an accepted grading range (cf.Recommendations of the AFTES WorkGroup #6, 1974 & 1994). This requirementmust absolutely be respected (no tolerance)to ensure the characteristics introduced inthe design calculations will be as accurate aspossible.

The flatness coefficient of the aggregates,measured using the NF P 18-561 standard,must be smaller than 0.25 for aggregatesranging from 5 to 16 mm. This size range islimited to 10 mm in the case of fiber reinfor-ced concrete.

The aggregates must be non-sensitive toalkali reaction.

3.1.2- CEMENT

The cements must comply with current stan-dards, and must be part of the C.E.N. list.The quality and quantity of cement will alsobe selected based on the aggressiveness ofthe environment (documentation P 18-011 –June 1992).

3.1.3- WATER

The water used for concrete hydration mustmeet current standards.

7


3.1.4- FIBERS

There is no current standard for fibers thatare mostly of metallic type. The other cate-gories of fibers (synthetic, glass, carbon) canbe used, but they must be tested to provetheir efficiency.

The fiber length must not exceed 0.7 timesthe diameter of the nozzle or else testsshould be conducted to assess the risk ofpipe blockage.

3.1.5- ADMIXTURES

It is useful to differentiate between twocategories of admixtures:

• admixtures included in the concrete com-position before spraying (superplastisizer,plastisizer, air entraining admixture, settingdelayer and hydration controller);

• Setting and/or hardening accelerators.

The admixtures included in the concretecomposition must comply with current stan-dards.

Setting accelerators must be adapted to thetype of cement used. One must be carefulwith some admixtures that can lead to redu-ced strengths in the medium or long term.

3.1.6- ADDITIVES

Additives (silica fume, fly ash and fillers),when used in the concrete composition,shall conform to the current relevant stan-dards.

3.2- REQUIRED MECHANICALCHARACTERISTICS FOR THEDIFFERENT TYPES OFSPRAYED CONCRETE

Sprayed concrete will be required to servedifferent purposes in different situations; itscharacteristics will need to be definedaccordingly.

3.2.1- TYPE 1 SPRAYED CONCRETE(PROTECTIVE LAYER)

This thin protective layer (≤ 50mm) is notintended to play a structural role, and as aresult the mechanical characteristics requi-rements for the concrete are relatively limi-ted. Nonetheless, the concrete will need to(1) adhere to the support, (2) sustain its ownweight, and (3) offer immediate protectionof the ground.

The spraying method (wet or dry-mix pro-cess) will have to be carefully selected in

order to ensure that these requirements aremet. Concrete bonding to the support canbe enhanced through the introduction ofadditives, such as silica fume (Beaupré). Theaddition of fibers increases concrete ducti-lity. The mechanical strength of youngconcrete can be relatively small if the layeronly acts as a protective layer. In the specificcase where only a protective skin effect isrequired, a dry-mix sprayed mortar may beused.

3.2.2- TYPE 2 SPRAYED CONCRETE(STRUCTURAL LAYER)

In this case, the objective is to obtain a mini-mum level of strength in the youngconcrete. Contrary to Type 1 sprayedconcrete, it is also required that thisconcrete be reinforced with welded wiremesh or fibers in order to meet minimalsafety criteria with respect to shear failure(including bending failure) or spalling due tostrong convergences (sprayed concretealone is not capable of playing this role).This reinforcement allows limiting crackingdue to restrained shrinkage of concrete(Abdul-Wahads, 1992; Ong & Paramasivam,1989).

High early strength will be obtained with theaddition of setting accelerators during thespraying process. Compressive strengthscould in that case reach values around 10MPa after 24 hours, and even 3 MPa after 3hours (see recommendations by the AFTESWork Group #6, 1994). When using wetspraying process, concrete compositionshould be optimized in order to achieve anas low as possible W/C ratio.Pumping can be achieved using asmall amount of water provi-ded appropriate admix-tures are used.

3.2.3- TYPE 3SPRAYED CONCRETE (STRUCTURAL RING)

The main concrete parame-ters for this type of supportelement are:

• The mechanical strength(short and long term);

• The deformation capacity.

It should be reminded thatdisplacement based calcu-lation methods require thedetermination of a modu-lus value to characterize

the sprayed concrete behavior. The diffi-culty, in this case, is to define a simpledesign approach allowing to account for thesimultaneous evolution of the concrete cha-racteristics and the loads transferred by thesurrounding grounds during construction.For example, the approach proposed byPöttler (1990) is based on an overall modu-lus value of 7,000 MPa which is used to eva-luate the short-term state of equilibriumreached by the surrounding ground and itsconcrete support (cf. § 3.4). Typical modulusvalues of 10,000 to 15,000 MPa are oftenused for the evaluation of the support cha-racteristics. It must however be kept in mindthat requirements for high early strengthvalues may result in higher modulus values.

Figure 3.1 (after Laplante, 1993) shows theevolution of the elastic modulus Eb of nor-mal concrete (NC) and high performanceconcrete (HPC) as a function of the com-pressive strength, σb.

This relationship was derived from a modeldeveloped by de Larrard and Leroy (1992),using an approach inspired by Hashin’sworks (1962). It shows a quasi-linear rela-tionship between σb and Eb in the shortterm, with an Eb/σb ratio of the order of2,500 for σb < 10 MPa.

Recent tests on sprayed concrete used forrailway tunnel structural refurbishmentworks tend to support the relationshipshown in Figure 3.1, both in qualitative andquantitative terms, although some precau-tions are required for its application tospayed concretes.

Figure 3.1 - Modulus of elasticity against compressive strength of concrete (from Laplante, 1993)

Mod

ule

of e

last

icity

Eb

(in G

Pa)

Compressive strength σb (in MPa)

Model

Eb NC

Eb HPC

8


3.3- LABORATORY TESTS -METHODOLOGY

It is necessary to perform a number ofcontrol tests on the sprayed concretesdepending on the objectives that arepursued.

The test methods to adopt, for pre-construction trials or for control purposes,are described in the recommendations ofthe AFTES Work Group #6 "Mise en œuvredu béton projeté dans les travaux souter-rains" (Tunnels et Ouvrages Souterrains n°1, 1974) and "Technologie et mise en œuvredu béton projeté renforcé de fibres"(Tunnels et Ouvrages Souterrains n° 126,1994). One can also refer to the documentprepared by the AFREM, 1995 (AssociationFrançaise de Recherche et d’Essais sur lesMatériaux et les Constructions).

A major difficulty resides in the evaluationof the deformation characteristics of theconcrete, for which no standard test is yetavailable. The analysis of numerous recentexperimental results show that the valuesof Eb /σb are relatively scattered and thatthe evaluation of this parameter is closelyrelated to the method used to measure theconcrete modulus. Given the absence ofstandards in this respect, the approachproposed by the LCPC (Torrenti et al.,1999; Boulay et al., 1999) can be conside-red to limit the number of uncertaintiesassociated with the testing equipment.This testing method leads to results thatare relatively consistent and somehowagree with Figure 3.1. We recommend theuse of this test method using 50 mm dia-meter samples with a length/diameterratio of 2.

3.4- CONSEQUENCES OFTHE EVOLUTION OF YOUNGSPRAYED CONCRETE PROPERTIES

Experience shows that support systemsmade of young sprayed concrete exhibit aninteresting ability to adjust to ground move-ment and sustain, without breaking, impor-tant deformations. In fact, calculations runwith modulus values of 20,000 to 30,000MPa (corresponding to hardened concrete)lead to stresses in the sprayed concrete thatlargely exceed its strength.

3.4.1- STUDIES ON FRESH SPRAYEDCONCRETE BEHAVIOR

To clarify this issues, several studies wereundertaken in German speaking countries inthe late 1980s, and more recently in Sweden(Chang, 1994), based on intensive labora-tory testing programs using young concretesamples. The works of Peterson (1989) andRokahr & Lux (1987) in Hanovre, those ofAldrian (1991) and Schubert (1988) inLeoben and those of Huber (1991) andFischnaller (1992) in Innsbruck are worthmentioning. In particular, Peterson (1989)formulated a three-dimensional constitutivelaw similar to that of salt, which incorporatesthe three essential characteristics of youngsprayed concrete:

• An instant elastic modulus that increaseswith time during the hardening process;

• A short term creep capacity that increaseswith the applied stress, and is even moreapparent when the concrete is young(Figure 3.2);

• A deformation generated by hydrationand shrinkage.

These laws allow simulating with precisionthe behavior of concrete measured duringrepeated tests in the laboratory, using atime-step calculation method.

The consequences of this particular featureof young sprayed concrete behavior, andespecially of its creep capacity, have beenformulated by Pöttler (1990). Using theabove described law, Pöttler (1990) firstmodeled the construction of a tunnel oversuccessive passes, with sprayed concreteliner support, and different values of theconcrete modulus within the usual range.He showed that stresses in the concretereached a maximum value at a distance ofone diameter from the front (Figure 3.3),and then decreased due to the importanceof creep, before stabilizing around a valueof 4 MPa, and this, independent of thegeometrical and geotechnical conditionsconsidered.

This parametric study also demonstratedthat the maximum normal stress in theconcrete, at distance of one diameter fromthe front, could be approached with goodaccuracy by selecting a fictitious modulusfor the concrete of 7,000 MPa. These resultsled the author to propose a simplified two-

dimensional calculation approach, using thismodulus as an overall parameter allowing toaccount for both the three-dimensionaleffect of the excavation at the front and thecreep capacity of concrete.

It is important to note that these constitu-tive laws were established for dry sprayedconcrete (usually used in German speakingcountries) for which it is known that thecreep rate is three to four times higher thanfor wet sprayed concrete, with this pheno-menon being due to the higher proportionof cement paste present in dry sprayedconcrete. From a more general standpoint,one should also recognize that the resultsobtained by Pöttler (1990) refer to a particu-lar work configuration, and it that some cau-tion would be advised when applying theseconclusions to other tunnels supported withsprayed concrete.

Observations made in the Langen tunnel inthe Alberg (Schubert, 1988) offer a goodexample of the stress evolution in theconcrete. Figure 3.4 clearly shows the effectof each face advance during the first threedays after installation: creep of concrete (intheory, assumed to lower stresses betweeneach excavation stage) which, in that case,were largely compensated by the creepbehaviour of the surrounding ground.

Figure 3.5, taken from the same study,shows the deformations measured with apair of sensors located on the inner andouter portions of the sprayed concrete ring.The corresponding normal stresses show areduction after a week, as predicted in

Figure 3.2 - Creep capacities of sprayedconcrete of different ages as a fonction of the

applied stress (from Pötler, 1990a)

day

day

day

day

Creep (%/day)

Experimentalresults

9


Figure 3.3. Radial and normal stresses bothconverge to a value around 5 MPa after aperiod of approximately 15 days, whichtranslates into the disappearance of ben-ding moments at the end of the monitoringphase. This result is in accordance with theobservation that failure of sprayed concreteshells occur through shearing – not bending– as described by Rabcewickz and Sattler(1965).

3.4.2- OTHER STUDIES DEDICATEDTO THE CREEP CAPACITY OFYOUNG SPRAYED CONCRETE

Different studies conducted by Neville et al.(1983) showed that creep of young concreteis dependent upon several concrete relatedfactors:

• Age of the concrete;

• Composition of the concrete;

• Amount of water;

as well as external parameters such as:

• Duration of loading and load patterns;

• Ambient temperature and humidity.

It should be noted that concrete age is notthe only criterion. Test results show, in fact,that a majority of the variations observed inthe early creep capacity can be attributed tocements which do not react at the samespeed. Likewise, for a given aging level, thelower the ambient humidity, the higher theconcrete creep capacity. The use of settingaccelerators will cause a significant increasein internal temperature of the concrete, andconsequently, of the amount of creep. Onthe other hand, creep is less important withhigh performance concrete than it is withstandard concrete.

For a given spraying technique (wet or dry),the amount of creep will be higher forhigher short-term concrete modulus values.Of course, the sooner the loads are applied(i.e. for rapid tunnel excavation), the moreimportant the final deformations due tocreep

3.5- EFFECTS OF SHRINKAGE

Shrinkage phenomena can lead to residualstresses in the concrete, which will even-tually translate into the apparition of cracksin the structure.

Figure 3.3 - Theoretical evolution of normal stress in sprayed concrete

Figure 3.4 - Langen’s Tunnel : tangent strains in concrete with corresponding stresses

after cach advance stage(after Schubert, 1988).

times(day)

Calculated stress(MPa) Measured stain

(mm/m)

Figure 3.5 - Langen ‘s tunnel : strains measured at the inner and outer face of the concrete with corresponding stresses (after Schubert, 1988).

σe (MPa)Maximum stress

Increase of stress due to face advance

Relaxation between face advance stages

Effect of excavation front

three dimensionnalstress state

Load(Phase I)

Relaxation(Phase II)

Constant stress(Phase III)

No effect ofexcavation front

Two dimensionnalstress state

t (day)

StressStrain

Advances

Strain (measured)

Stress (calculated)

Time (day)

Dominating effect of creep over the effectof the excavation front

Inside

Inside

outside

outside

10


4- DESIGN OF SPRAYEDCONCRETE FOR UNDER-GROUND SUPPORT The objective of this chapter is to describethe principal technical recommendations forthe design of underground support sys-tems. A brief overview of designapproaches for underground openings isprovided in the following sections beforeaddressing the three types of sprayedconcrete described in Chapter 2.

In every case, and based on good practice,it will be assumed that the excavation sur-faces have been adequately drained so that:

• Washouts of the first few centimeters ofconcrete is prevented;

• No hydraulic pressure is present duringthe excavation and shotcreting phases.

It should however be noted that sprayedconcrete can contribute to the stability ofthe opening only if the cohesion of theexposed ground is sufficient to maintain theoverall stability for a few hours. Therefore,sprayed concrete cannot in any case beused as a substitute to preconstructiongrouting, pre-support systems or sheeting,which are necessary for excavating in softground presenting no short term cohesion.For more information on the conditions forwhich sprayed concrete is appropriate, thereader can refer to the table presented inthe documents of the AFTES Work Group#7, "Choix d’un type de soutènement engalerie" (Tunnels et Ouvrages Souterrainsn° 1, 1974).

4.1- REVIEW OF AVAILABLEDESIGN METHODS

4.1.1- GENERAL CONSIDERATIONS

The present recommendations come inconjunction with the existing document tit-led "Réflexions sur les méthodes usuellesde calcul du revêtement des souterrains "prepared by the AFTES Working Group #7(Tunnels et Ouvrages Souterrains n° 14,1976), which they complement on specificaspects related sprayed concrete construc-tion.

The design of underground support sys-tems is typically performed using one of thefollowing approaches:

• The load acting on the liner is first deter-mined independently of the support sys-tem, and then applied to the liner accor-ding to the actual ground-liner contactconditions or,

• The ground/ structure interactions aredirectly taken into account in the definitionof both the loads applied to the structureand the support reaction. Due to the inti-mate contact conditions existing betweenthe ground and sprayed concrete, the latteris obviously more realistic.

It is important to point out that the designmethod should account for the behavior ofthe support system, and more specificallythe Type of sprayed concrete to be used (asdescribed in Chapter 2: Type 1, Type 2 andType 3), and be adjusted in accordance withthe requirements of each project phase thatis considered (feasibility study, conceptdesign, detailed design or shop drawings).

Finally, for openings where the magnitudeof ground movements is an importantdesign criterion, the design method shouldnot address the required of structural capa-city, but also provide a quantitative evalua-tion of the ground deformations. In thiscase, the design methods must make use ofground deformation parameters.

4.1.2- METHODS FOR LOAD EVALUATION

Two approaches are commonly used for thedetermination of the loads applied to thesupport structure: the limit state methodand the displacement approach based ontechniques such as the convergence/confi-nement method.

With the limit state method, the “load”applied to the support structure is obtainedby using well established formula, as propo-sed by Terzaghi (1946), Protodiakonov, andCaquot, which all involve the analysis of afailure mechanism in the surroundingground, generally independently of the pre-sence of any support structure. Care musthowever be taken in using these methods tothe fact that the support system must sus-tain the displacements induced in theground before a state of limit equilibrium isreached.

On the other hand, the convergence/confi-nement method was introduced to consider,in plane strain 2D calculations, the three-dimensional effects pertaining to the inter-action that develops between the groundand the support system during the excava-tion phase (Panet, 1995). It is worth mentio-ning that this method should not be viewedas a design method for sprayed concreteper se, but rather a method for evaluatingthe ground-structure contact stresses atequilibrium. With this in mind, this methodcan be used in conjunction with some of thetechniques developed for the design of

sprayed concrete support systems that arepresented hereafter.

4.1.3- VERIFICATION OF THESTRUCTURAL STABILITY

The methods available for verifying the sta-bility of the support structure are fully pre-sented in Sections 4.2 to 4.4. They apply toall design methods, including:

• Punching of the sprayed concrete where itis used as a membrane between rock bolts,in which case the strength of the membranein between the rock bolts should be chec-ked;

• Buckling of the concrete: in this case, theverification consists in making sure that thestructure is stable under the induced linerthrust, as well as eccentrically when appli-cable;

• Strength of materials types of calculations,allowing to verify that the values of the inerthrust, N and bending moment, M (oreccentricity e) generate acceptable stresses.

4.1.4- MAIN CHARACTERISTICS OFSTANDARD DESIGN METHODS

One can differentiate between five majorfamilies of underground opening designmethods: empirical methods (qualitative),semi-empirical methods, the subgrade reac-tion modulus method and the solid compo-site method.

4.1.4.1- Empirical methods

Empirical methods (qualitative) allow todetermine the overal dimensions of the sup-port system to be put in place. Thesemethods are essentially based on a qualita-tive description of the ground mass and theconditions in which the opening must beexcavated, which often translates into theuse of index classifications (AFTES, 1978;Bieniawski, 1974 & 1989; Barton et al., 1974& 1975, Barton, 1983). No calculation per seis performed; the approach is essentiallyempirical and based on a large number ofobservations which do not explicitly accountfor the actual ground-support interactionsor any deformation related issues. Thesemethods should therefore be restricted toType 1 sprayed concrete (all phases of thepreconstruction study), and may be used inthe preliminary design of Type 2 and 3sprayed concrete liners. Considerationshould also be made to the fact that theyare generally limited to specific types ofgrounds and completely rely on the qualityof the ground characterization parametersused in their implementation.

11


4.1.4.2- Semi-empirical methods

A number of design approaches fall underthis category. They are based on well-identi-fied failure mechanisms of the groundand/or the support system. In general, thesemethods do not take into account the inter-actions between the ground and the sup-port structure, or allow an evaluation of theground deformations to be made. Theapproach consists in first identifying the“load” applied onto the structure and thenanalyzing the stability of the “structure”subject to this external load, irrespective ofany interaction that may develop with thesurrounding ground. These design methodsare mainly adapted for Type 1 and 2 sprayedconcrete at every stages of a project (parti-cularly if the method convergence/confine-ment is used for estimating the “load”).They should not be used for Type 3 sprayedconcrete, except at preliminary designstages.

Amongst these methods, the one proposedby Rabcewicz (1973) is of particular interestfor historical reasons, as well as for the role ithas played in the early underground appli-cations of sprayed concrete support tech-niques (New Austrian Tunneling Method).

The Rabcewicz method considers the combi-ned ground and sprayed concrete (and even-tually rock bolts) components as a whole. Itinvolves the equilibrium analysis of groundvolumes, delineated by logarithmic shapedfailure surfaces, with due account beingmade of all applied loads. This method isadapted to Type 2 sprayed concrete, forwhich it was originally developed. However, ithas become rarely used in practice.

4.1.4.3- Subgrade reaction modulusmethod

This method consists in modeling the sup-port system by means of “bar” elementsand the ground with “springs”. As a result, itallows to directly take into account the inter-actions between the ground and the sup-port structure. Its practical implementationis performed in two steps:

• Evaluation of the loads: this can be achie-ved by means of semi-empirical methods(Terzaghi’s formulation for example), theconvergence/confinement method in thecase of continuous media (including theeffect of possible rock bolts), or using ablock equilibrium approach in the case ofrock masses; load evaluation is a delicateexercise which result will determine the finaldesign of the support system;

• Evaluation of the stiffness of the springsused to model the ground stiffness reaction:in the simple cases of closed support systemswhich are sufficiently close to a circular shape,the evaluation of the subgrade raction modu-lus is derived from the deformation modulusof the ground by applying the simple equa-tion k = E / (1+ν).R (with E being the ground’sYoung modulus, n its Poisson’s ratio and Rthe radius of the opening). However, in manycases (horseshoe shape, no invert or invertwith a large radius of curvature), the selectionof a modulus values can be delicate in areassuch as the tunnel walls, the invert or thevault footings. For linear sections of the sup-port system, reference can be made tomethods developed for the determination ofthe reaction modulus used in the design offoundations or retaining walls (elastic or pres-suremeter method).

Provided these limitations are well unders-tood, , this approach is well suited for Types2 and 3 sprayed concrete for every phase ofa project. However, it is unsuited for tunnelsconstructed with staged face excavation, asit is not possible in that case to model theeffects of some intermediate excavationphases on the already installed support ele-ments. It must be kept in mind that the cal-culated deformations may not be represen-tative of the actual ground response.Another limitation is that the method onlyallows structural deformations to be evalua-ted, which means that an extrapolation isrequired if other parameters such as surfacesettlements need to be computed (e.g.shallow structures).

4.1.4.4- Solid composite method

The solid composite method models thesurrounding grounds as a continuousmedia; it can be applied in different ways:finite elements, finite differences or equiva-lent, and even analytical methods for thesimpler cases. It allows modeling of boththe support structures and surroundingground with due account of all constructionphases (staged face excavation), as well asindividual interaction mechanisms. Besides,it can also be used in its analytical form forthe evaluation of the modulus used in thesubgrade reaction modulus method.

It should however be emphasized that thisapproach which allows a high degree ofsophistication (2D or 3D models, elaborateconstitutive models, allowance for rock boltinclusions and rock mass discontinuities) canbecome extremely cumbersome, as soon asa high degree of refinement is introduced

and require the determination of numerousparameters that are difficult to accurately eva-luate. In most cases, the approach is limited tosimple models (2D models with allowance fortunnel face effects by means of the conver-gence/confinement method, simplifiedconstitutive models involving a limited num-ber of material parameters). Such modelsremain easy to implement and are usuallyfound to produce satisfactory results in eva-luating the overall response of the structureand the ground, both in terms of loads anddeformations. The degree of sophisticationshould not overshadow the accuracy of thenumerical predictions, especially when itcomes to estimating surface settlements.

In this perspective, the solid compositemethod is well suited to the design of the“structural” types of sprayed concrete (Type3 sprayed concrete), at the more advancedphases of the project (detailed design stu-dies, especially when design is controlled bydeformation criteria; shop drawings). Onthe other hand, the use of this method forType 1 and even Type 2 sprayed concrete isnot recommended (unless the control of thedeformations is an important parameter). Infact, for such categories, the role of sprayedconcrete is implicitly taken into account inthe design of the support system throughthe introduction of improved (or at leastnon-reduced) mechanical ground characte-ristics (associated by the presence ofsprayed concrete). These considerationsmay lead, in some cases, to neglecting thestructural effect of sprayed concrete. As anexample, the large caverns of the CERN(Boymond and Laigle, 1999) excvated at 60meter depth through Geneva’s molasses,were designed with no consideration of thesprayed concrete contribution to the sup-port system, despite its relatively largethickness (20 cm).

4.2- TYPE 1 SPRAYEDCONCRETE– PROTECTIVELAYER

4.2.1- DESCRIPTION

This type of support is made of a relativelythin sprayed concrete layer directlyapplied at the excavation surface. Thesprayed concrete may be reinforced withwire mesh or fibers. This relatively thinlayer is not accounted for in the design ofthe support system, and stability is achie-ved by means of:

• The ground itself, protected by thesprayed concrete layer;

12


• The rock bolted ground;

• A sprayed concrete shell applied later(Types 2 or 3 sprayed concrete, as describedin the following sections). In this case, theType 1 sprayed concrete layer can be takeninto account in the thickness of the finalliner, provided the characteristics of the ini-tial and final sprayed concretes are suffi-ciently close;

• Any other support system.

4.2.2- TECHNICAL RECOMMENDATIONS

This sprayed concrete layer is purposelythin, as a thicker layer could attract unwi-shed loading and generate structuraldamage. The thickness is generally limitedto 50 mm for tunnel sections over 30m2.

This thin layer is placed immediately after (ifnot during) excavation and generally beforethe installation of any other type of support.

4.2.3- OBJECTIVES

The objective of this shotcrete membrane is(cf. § 2.1):

• To protect the exposed ground surfacesagainst rapid alteration by air or water;

• To provide a primary means of protectionagainst limited ground fall-outs.

It can also be used as support for the instal-lation of water barriers.

4.2.4- DESIGN

This type of ground support is mainly desi-gned as a protective layer. It does not playany structural role and, as such, does notrequire any specific calculation. Design issolely based on empirical considerations.

4.3-TYPE 2 SPRAYEDCONCRETE –STRUCTURALLAYER

4.3.1- DESCRIPTION

This type of ground support is composed ofa sprayed concrete layer combined withrock bolts and/or steel ribs. Rock bolts canbe made of fiberglass or steel. Steel ribs canbe of the H, lattice girder or TH types.

The overal stability of the excavation is notprovided by the shell action of this sprayedconcrete layer (cf. Type 3 sprayed concrete),but is rather based on the ability for theground itself to mobilize its resistance withthe assistance of the rock bolts or steel ribs.

This type of sprayed concrete can even begrooved to release structural stresses in thecase of deep excavations (which precludesthe use of heavy steel ribs and leads to theinstallation of sliding TH types of ribs).

4.3.2- TECHNICAL RECOMMENDA-TIONS

This type of sprayed concrete must be rein-forced either by the introduction of wiremesh or fibers, in order to allow a minimaldegree of ductility. Moreover, it would betoo risky to rely on the early strength of theconcrete alone.

In any case, design parameters should bevalidated through suitable testing. Forexample, the plate test described in theAFTES Work Group #6 document"Technologie et mise en œuvre du bétonprojeté renforcé de fibres" (Tunnels etOuvrages Souterrains n° 126, 1994) is welladapted to the evaluation of the ductility ofthis type of support.

In order to be able to rely on the combinedaction of the sprayed concrete and rockbolts, the construction methods that areused must be provide some guaranty thatload transfer can occur between thesprayed concrete and the bolts.

In situations where the sprayed concretelayer is cut to release stresses, it should bereinforced with wire mesh. Fibers are notconsidered to provide sufficient continuityfor the stability of the concrete support pla-ced between bolts.

4.3.3- OBJECTIVES

This type of support is intended to ensurethe local stability of the excavation and toprevent ground fall-outs to (cf. § 2.2).

It is intended to prevent any local failurethat could be a source of reduction of theself-stabilizing capacity of the surroundingground. It also contributes to the appro-priate response of rock bolts or steel ribs,particularly by preventing local decompres-sions.

4.3.4- DESIGN

This type of support must be capable of sus-taining local loads induced by the surroun-ding ground and transfer them to the sup-port system installed to ensure the overallstability of the opening (rock bolts or archgirders).

The loads carried by the sprayed concretelayer depend on the ground conditions andshould be evaluated on a case-by-case

basis, with due consideration of the follo-wing factors:

• geo-mechanical properties of the ground;

• location of the support elements;

• orientation of the joints (in the caseof rock);

• properties and frequency of the joints(with particular attention to the shearstrength that can be mobilized along slipplanes);

• local effects induced by rock bolts orsteel ribs;

• influence of ground stresses.

Once the principal orientation of cracksand the properties of the ground alongthese planes have been established, onecan estimate the volume and therefore theweight of ground to be supported. In themore frequent case however where nocrack systems can be identified - or wherethe principal orientation cannot be easilyestablished - the weight of ground to besupported shall be evaluated using someappreciation of the overall probableground response. Figure 4.1 shows a typi-cal ground response mechanism where abell shape mass of ground is acting on thesprayed concrete layer.

Based on experimental works, Fernandez-Delgado et al. (1981), Holmgren (1987) andVandewalle (1992) proposed several pos-sible failure modes for a sprayed concretelayer subject to a block weight loading:

• Shear failure, either on the perimeter ofthe supported block or by punching in theanchoring areas (rock bolts or steel ribs);

• Bending failure.

In the former (Figure 4.2), the shear resis-tance of the sprayed concrete is calculated

Figure 4.1 - Exemple schematic of liner loading(after Barett & McCreath, 1995).

(1) ground zone reinforced by bottin(2) ground zone which might push on the liner

13


using the simple equation: R = u.t.τs, whereu is the perimeter of the failing area, t is thethickness of the sprayed concrete layer andτs is the shear strength of the sprayedconcrete. This last property must be evalua-ted with care; a value of 0.2 σb (σb being theconcrete compressive strength) can gene-rally be used as a rule of thumb.

For compressive strengths below 5 MPa –which is typically reached in a few hours- thisequation is not applicable and the resis-tance of the sprayed concrete layer is achie-ved through other phenomena, such as wiremesh web action. This type of action pro-vides some safety against loading of youngsprayed concrete as could be caused, forexample, by unidentified ground heteroge-neities, although this should not be taken asa reliable means of ground support. In fact,sprayed concrete alone is not suitable forgrounds where every square meter of expo-sed surface requires an immediate mechani-cal support or in cases where the loadingrate would exceed the rate of strengthe-ning. For this reason, sprayed concrete mustbe associated to other types of supports. Incases where a significant amount of confi-ning pressure is required, the resistanceagainst potential bending failure must bechecked (Figure 4.3).

This verification can be achieved usingconventional reinforced concrete designmethods (refer to the Section 4.4 below for

Type 3 sprayed concrete). However, whenbending moments are not associated withcompressive loads, these verificationsrapidly reach their limits due to the lowstrength properties of young concrete. Inthis case, a minimum level of safety shouldbe provided with respect to the post-failurebehavior of the concrete by using:

• wire mesh, which will act as a “net” inten-ded to capture falling blocks;

• fibers, which induce a more ductile type offailure.

4.4- TYPE 3 SPRAYEDCONCRETE –STRUCTURALRING

4.4.1- DESCRIPTION

This type of ground support is made of athick sprayed concrete shell (hundreds ofmillimeters thick), which is capable to alonecontribute to the overall stability of the ope-ning. The concrete may be reinforced, withor without fibers. This shell may be associa-ted to rock bolts or steel ribs with directeffect on the mechanical performance ofthe structure.

4.4.2- TECHNICAL RECOMMENDA-TIONS

This shell must have a relatively importantthickness, in order to guaranty an overallresponse similar to that of an arch, as oppo-sed to Type 2 sprayed concrete, where onlylocal stability between rock bolts is sought.

The minimal thickness of this shell will beimposed by construction considerations(the excavation profile may be more or lessirregular, depending on the ground condi-tions and the excavation method), with theobjective being to guaranty at minimum theshell thickness is equal to the theoreticalvalue used for design. the shell dimensions

will also be function of the opening size andof planned construction phasing.

As soon as its thickness reaches 1 or 2% ofthe radius of the opening, the sprayedconcrete doesn’t act as a simple skin any-more and shell effects must be examined tocontrol disorders that may result from unde-restimating rigidity of the structure. Finally,the shell action can only be obtained if theconstruction method allows to ensure thecontinuity of the concrete shell in the trans-verse direction.

4.4.3- OBJECTIVES

The main objective of this structural shell isto guaranty the overall stability of the entireexcavation (cf. § 2.3). It can also offer pro-tection against local disorders (similar to therole of Type 2 sprayed concrete) and thefirst layers of sprayed concrete can be put inplace immediately after excavation in orderto protect the surrounding ground againstrapid alteration (similar to the role of Type 1sprayed concrete).

The role of the sprayed concrete shell is tolimit the convergence of the excavation aswell as to avoid any excessive groundrelaxation, which would reduce its strength(loss of cohesion caused by excessive defor-mations along weak planes for example),and therefore limit its contribution to thestability of the opening. It also allows limi-ting surface deformations, which is impor-tant when construction takes place in anurban area or at shallow depths, or moregenerally, when excavation takes place inthe vicinity of structures that are sensitive tosettlements.

4.4.4- DESIGN

The design and verification methods of asprayed concrete shell can be divided in twocategories:

• Impact on ground movements, as estima-ted with displacement based methods (solidcomposite method for example) and/orcontribution to the stability of the opening;

• Verification of the shell’s resistance.

Each category relies on different methods ofanalysis and different material propertyinput.

4.4.4.1- Displacement based calculationand stability analyses

For displacement based calculations andthe verification of the contribution of thesupport system, different methods can beused (cf. Chapter 4.1), depending on the

Shear failure at support Shear failure under block movement

Figure 4.2 - Shear failure (after Barett & McCreath, 1995).

Figure 4.3 - flexural faulue(dafter Barett & McCreath, 1995).

14


actual project stage (concept design,detailed design, shop drawings), as well ason the complexity of the project and thesophistication of the construction methodemployed. The implementation of theseapproaches generally rely on the determi-nation of a suitable stress release coeffi-cient to reflect the ground response whenthe liner support is installated and of thesupport stiffness.

Regarding the former, current calculationmethods allow the influence of the alreadyplaced sprayed concrete to be accountedfor in the determination of the stressrelease coeficient at the front (new implicitmethod) : this effect which typically appliesto very stiff support systems can howeverbe neglected in the case of sprayedconcrete shells.

The stiffness of bar or shell elements usedin the model to account for the supportsystem depends on two parameters: thethickness and deformation modulus of theshell. A suitable representation of thebehavior of sprayed concrete can be obtainedby selecting nominal values for these twoparameters, given that both an upper orlower estimate would present some short-comings. The value of the nominal thicknessshould be based on an evaluation of theoverbreak volume and placement procedures.

A major issue with the determination ofthe deformation modulus is that it is crucialand critical, given that one can in principlenot rely on one single value of this parame-ter, which would correspond to a givenage of the shell and loading stage. In fact,the response of sprayed concrete shells isdetermined by the concurrent evolution ofthe loading (which magnitude increasesprogressively with the distance from theface) and concrete hardening processes,and subsequently by the deformability andcreep properties of the concrete.Moreover, the evaluation of the concretestiffness is made difficult by the fact thatsprayed concrete is applied in successivelayers.

Unfortunately, few studies have addressedthis issue to date. On this basis, it is propo-sed to use the following approach, which isderived from the work of Pöttler (1990),and allows to account in a relatively simplemanner for the specificities of sprayedconcrete support systems:

• For the excavation and sprayed concreteplacement phases, until the distance fromthe face becomes 2 to 3 times the diameterof the opening, a fictitious modulus shall be

adopted for the calculations, whatever theactual composition of the sprayed concrete(provided it complies with current andaccepted practice). This fictitious modulus isin the range 7,000 to 15,000 MPa, with thevalue 7 000 MPa being more appropriatefor rapid excavation progress cases (sec-tions smaller than 50 m2);

• When the construction phase takes placeafter hardening of the concrete (e.g. whenresuming excavation in the case of stagedface excavation), one shall use conventionalformulas (derived from standards or othersources) for calculating the instantaneous orlong-term deformation modulus.

It should also be reminded that the modulusvalue should be selected according to thepurpose of the calculation. In particular, fora given project, lower estimates of themodulus should be used in the evaluation oftunneling induced ground settlements (e.g.for shallow tunnels), whereas upper esti-mates should be used in the evaluation ofliner loads to ensure that these remain onthe conservative side.

This approach aims at determining themaximum loads carried by the sprayedconcrete support system. It should be notedthat the portion of the load transferred atthe early stage (before 3 days) could inducesignificant creep in the concrete. This phe-nomenon might be worth considering whenevaluating long-term ground deformations,if the early share of loading is significant.

In the case where reinforcement with heavysteel ribs is contemplated, the thickness andmodulus of the combined (concrete and rib)structure can be adjusted by introducingequivalent values where the contributions ofthe two components are in proportion totheir respective inertia and area. On theother hand, wire mesh or lattice girdershave little effect on the stiffness of the sup-port system, and shall be neglected in theestimation of the mechanical characteristicsof the support system.

In any case, the steps taken for the selectionof the modulus value of sprayed concretewill have to be substantiated and clearlydocumented in the design documents.

A few additional points shall require particu-lar attention:

• First, it is important to make sure that theassumed continuity of the shell is actuallyachieved on the job site, especially in sensi-tive areas such as the sections wherespraying operations resumes (staged face

excavation) or at the junction of the wall andinvert;

• It should be kept in mind that because ofits application method and load transfermode, the sprayed concrete support systempossesses some ability to adjust to the loadsit is subject to. Particularly, the occurrenceof bending moments, capable of exceedingthe flexural capacity of the support system,leads to the formation of cracks, which inturn allow for stress redistribution to takeplace without affecting the overall stability.This is the case of areas where the shell canpresent abrupt changes in curvature, suchas the crown of ogive shaped sections. Inthese areas, calculations will not reflect rea-lity. Therefore, based on the verificationprinciples that will be developed in the fol-lowing sections and on existing standards, itis proposed to use the following approach:

- Calculate the stresses in the shell assu-ming an elastic behavior with no occurrenceof cracking;

- Allow for bending moment redistribu-tion in those points where the ultimatecapacity of concrete is reached, either byrecalculating the stresses after introducingplastic hinges at these points or by carefullycalculating fixed redistribution values.Bending moment redistribution must beperformed carefully, especially if bendingcan lead to a loss of contact between theshell and the ground (as in the case of a rigidshell not tied with rock bolts);

• When the finite element method is used,the attention of the reader must be drawnto the sensitivity of the results to theassumptions used to characterize theground/concrete interface (risk of slippageat the the ground - concrete interface,which can lead to increased stresses in thelower part of the shell, and subsequent pun-ching and/or settlement during bench exca-vation if a bench and heading excavationtechnique is used). Except in such specificsituations, the appropriate contact condi-tion to be used for the interface betweenthe sprayed concrete shell and the ground isthat of non-slippage, because of the place-ment method which allows an intimatecontact to be achieved between the groundand sprayed concrete.

4.4.4.2- Verification of the shellresistance

The verification of the shell resistanceconsists in checking that the sprayedconcrete sections possess sufficient resis-tance to support the stresses calculatedusing the above described approaches.

15


For distances to the front ranging from 2 to3 times the diameter of the opening, theconcrete strength to be used in the verifica-tion during the excavation and sprayedconcrete placement stages shall correspondto the age of concrete found at 2 diametersfrom the face, which depends on theconstruction rate. For the other stages, thecharacteristic strength shall be used, sincethe stability of the section must be verifiedfor the loads carried once hardening ofconcrete has taken place. In both cases,attention shall be paid to the scattering inmechanical properties of the sprayedconcrete, which result from the placementmethod.

Young concrete strength shall be achievedusing the principles presented for Type 2sprayed concrete (sprayed concrete withguaranteed initial strength, use of fibers andwire mesh). The thickness value to be usedis the same as that taken in evaluation of theliner loads.

The following principles apply for the verifi-cation of the sprayed concrete shellstrength:

• For plain concrete, the method describedin the text "Recommandations relatives àl’utilisation du béton non armé en tunnel",prepared by the AFTES Work Group #7(Tunnels et Ouvrages Souterrains, 1998)shall be used;

• For reinforced concrete, the BAEL stan-dard (French standards for Limit StateDesign of reinforced concrete) shall beapplied with the following safety coefficientfor the loads: 1.35 x 1.20 (BAEL’s coefficientfor dead load x a safety coefficient associa-ted with uncertainties on the reinforcementencasement and the actual position of thereinforcement within the section).

Given that sprayed concrete shells tend tobe of constant thickness, the most effectiveverification approach is to develop an inter-action diagram (N, M) allowing the combi-ned moment (M) and axial load (N) values tobe checked against the envelope of allo-wable loads. Moreover, this type of diagramprovides some indication of the potentialeffect of increasing the amount of steeland/or the thickness of the shell. In the lat-ter, as the stiffness of the shell is changed,new calculations are needed to re-evaluatethe stresses to be accounted for in the verifi-cation. Finally, a verification must be madefor the shear load, V.

For sections including steel ribs, the loadsobtained from the calculations must beredistributed amongst the ribs and the

sprayed concrete; once this distribution isdone, each element (steel ribs and sprayedconcrete) must be verified separately, usingexisting standards for steel ribs, and follo-wing the approach developed above for thesprayed concrete component. The alloca-tion of loads between the ribs and sprayedconcrete can be obtained using one of thefollowing approaches:

• Distribution of the axial loads and the ben-ding moments in proportion to the stiff-nesses and the inertias;

• Introduction of a homogenized compositestructure subject to the applied axial loadsand flexural moments, with the tensile por-tion of the rib acting as reinforcement.

4.5- CONSIDERATION OFSPRAYED CONCRETE IN THEFINAL LINER DESIGN

4.5.1- INTRODUCTION

This chapter addresses the case where acast in-place concrete liner is placed inside alayer of sprayed concrete support system.Cases where sprayed concrete is used asfinal liner are dealt with in Chapter 6.

The quality and durability of initial linersused nowadays allow for the possibility toincorporate their effect in the design of thefinal liner, with obvious economical conse-quences. It is of course the Owner’sRepresentative’s choice to decide to do so,and as such it should be explicitly formula-ted in the tender and contract documents.

However, up until recently, the contributionof the initial support system to the designcapacity of the final liner has been ignored,due to the following reasons:

• Lack of experience with the long termbehavior of concrete shells, especially thosemade of guaranteed initial strengthconcrete, for which the admixtures used toachieve high early strength might affectlong term characteristics;

• Risk of heterogeneity between the diffe-rent layers;

• Risk of heterogeneity when resumingshotcreting, in the case where staged faceexcavation is used;

• Variability of mechanical characteristicswhich are affected by the quality of work-manship;

• Impossibility (contrary to sprayed concretefinal liners) to easily monitor the response ofthe shell (visual inspection), as this wouldrequire the installation of expensive andcomplicated monitoring devices.

On the other hand, the quality of sprayedconcrete support has been found to consis-tently improve over the years, due in parti-cular to the following reasons:

• Its composition can be adjusted to matchthe contractual requirements;

• Preconstruction trials are systematicallyconducted to confirm that these require-ments can be met with the proposed equip-ment and installation method;

• The quality of the concrete is checkedthrough trials conducted throughout theconstruction period;

• The concrete mechanical characteristicscan be improved by incorporating admix-tures or fibers without adversely affectinglong-term properties.

Moreover, sprayed concrete shells can beused along with steel ribs or lattice girders,rebar or rock bolts, with thicknesses oftenreaching 20 to 30 cm, which all contribute totheir mechanical resistance.

4.5.2- SPRAYED CONCRETE NOTINCLUDED IN THE FINAL LINERDESIGN

Sprayed concrete support should not beaccounted for in the design of the final linerin the following cases:

• Type 1 sprayed concrete used alone;

• Type 2 sprayed concrete, unless it is desi-gned as a shell to form part of the final liner;

• Sprayed concrete placed in the invertduring construction.

4.5.3- RECOMMENDATIONS FORTHE CONSIDERATION OF SPRAYEDCONCRETE IN THE FINAL LINERDESIGN

For practical purposes, it is reasonable toconsider that a minimum thickness of 30to 50 cm of cast in-place concrete shouldbe used in all sections of tunnels of stan-dard shape which are more than 10 m indiameter.

Type 2 sprayed concrete, if designed as a"shell", and Type 3 sprayed concrete can beallowed for in the calculations. In this pro-cess, the following recommendationsshould be observed:

• The guidelines described in the documentof the AFTES Work Group #6 (1994) on theapplication of sprayed concrete shall beused;

• The sprayed concrete composition shallbe defined such that contractual criteriapertaining to structural durability are met,

16


especially with respect to the aggressive-ness of the environment and the compatibi-lity between the aggregates, the cementand the admixtures (mainly with the alkali-aggregate reaction);

• The sprayed and cast-in-place concrete,where put in contact shall have compatiblecompositions;

• In cases where staged face excavation isused, the quality of concrete at the interfacebetween subsequent sprayed layers mustbe checked;

• Depending on the ground type and exca-vation method, a nominal guaranteed thick-ness of sprayed concrete shall be defined inthe contract and verified on site;

• Depending on the long-term risk of degra-dation of the sprayed concrete mechanicalproperties (related to admixtures used toincrease the strength of young concrete), itmay be appropriate to perform tests duringthe entire life cycle of the structure.

Generally, Types 2 and 3 supports areapplied on top of a protective layer ofconcrete (Type 1 support). This initial pro-tective layer shall not be included in thedesign concrete thickness, given the risk ofalteration and cracking that may affect itsmechanical properties.

Steel ribs or lattice girders, rock bolts andrebars incorporated within the sprayedconcrete shell can be allowed for in thedesign calculations, provided their effectiveparticipation to the structural strength canbe adequately appreciated and all neces-sary precautions against corrosion areundertaken.

4.5.4 – RECOMMENDATIONS FORTHE EVALUATION OF STRESSES INTHE STRUCTURE

Whatever the method used to evaluate thestresses in the structure (see Chapter 4.1), itis necessary to ensure that design results areconsistent with the assumptions used in thecalculations. Indeed, any stress evaluationthat would not account for the long-termthickness of the sprayed concrete layercould not be used to evaluate its contribu-tion to the resistance of the final structure,as this approach would cause the liner stiff-ness and subsequently the stresses to beunderestimated.

Similarly, the thickness of the initial liner anddesign of the final liner should be checkedseparately.

Some degree of load redistribution shallbe allowed for within the shell as defined in

§ 4.4. The deformation modulus of sprayedconcrete to be included in the calculationsshall be estimated using conventional for-mulas (specified or other) for the determina-tion of the long term deformation modulus.When no water-tightness membrane isused, the sprayed concrete and cast–in-place concrete liners can be assumed to bein perfect contact, with water pressures (ifany) being applied at the extrados of thesprayed concrete liner. If a water-tight mem-brane is used at the interface between theinitial and final liners, it is recommendedthat water pressures be assumed to act onto the final liner. It is important, in that caseto carefully analyze the structural responseat the interface.

4.5.5- PLAIN OR REINFORCEDSHOTCRETE

Welded wire mesh present in the sprayedconcrete is usually not accounted in the cal-culations, particularly because of uncertain-ties related to its location. If reinforcement isrequired inside the liner, rebar shall be incor-porated within the formwork. In caseswhere perfect bond is assumed betweenthe two concrete layers, the amount (or lackthereof) of reinforcement shall be determi-ned on the basis of a sole concrete layer ofoverall thickness. Design verifications usingthe estimated design loads shall be conduc-ted according to § 4.4.

It is also desirable to ensure that a sufficientthickness of cast-in-place concrete is in com-pression (in any section). The magnitude ofthis minimal thickness may be specified bythe Owner’s Representative.

5- MONITORING

5.1- GENERAL BACKGROUND

Modern design and construction methodsof underground structures regard monito-ring as an essential element for safety andlongevity.

It is particularly the case of undergroundstructures where stability is primarily achie-ved through the self-sustaining capacity ofthe surrounding soils and rocks. These are,by nature, heterogeneous and anisotropic,which can only be modeled approximatelyin the analysis using simplified designmethods. It is therefore essential to incorpo-rate in addition to with the design processsome monitoring approach that would allowto check the design and control the impactof the structure on the environment.Monitoring should therefore be viewed as

an integral part of the design process, in asimilar manner as the design approachespresented in Chapter 4. For this reason, themonitoring approach is detailed in the follo-wing sections. The reader may also refer tothe document prepared by the AFTES"L’organisation de l’auscultation des tun-nels" (Tunnels et Ouvrages Souterrains, n°149, 1998) for more information. The mainfactors involved in the adjustment of thetunnel liner design due to monitoring obser-vations are:• The density and length of rock bolting;• The thickness of the sprayed concretelayer;• The length of each excavation stope;• The degree of staging (provided suchchanges can be implemented at the wor-king face);• And lastly, the sprayed concrete thickness.

In this respect, the design of a sprayedconcrete tunnel support can typically beconsidered part of the observationalmethod.

5.2- MONITORING OBJECTIVES AND MEANS

The monitoring approach should embraceall elements of the supporting structurewhich is composed of the surroundingground and sprayed concrete shell, and mayinclude anchor bars, steel ribs and cast-in-place concrete. The instrumentation put inplace is intended to monitor the response ofthe structure so that design parameters canbe adjusted, as required, throughout theconstruction process. It should allow com-parisons to be made between recordedvalues and those predicted at the designstages, as well as provide some comfort thatobserved measurements and variation ratesremain acceptable.

A variety of methods are available to theengineer in this respect; they related to:• Measurements of relative and total displa-cements;• Measurements of stresses within theground mass and in the different liner ele-ments (sprayed concrete, steel ribs, activeor passive anchors);• Measurements of water pressures andflows.

The most widely used and reliable measure-ments are those of relative displacements.They are easy to take and their interpreta-tion allows immediate adjustment to bemade to the support system. Other measu-rements may be prone to errors that mayresult from complex monitoring procedures

17


or difficulties found in installing the equip-ment or the benchmarks.

5.2.1- DEFORMATION MEASUREMENTS

One can distinguish three categories ofmethods used for measuring deformations:

• Topographic markers placed at the innersurface of the sprayed concrete shell;

• Measurement points placed in boreholesdrilled from the tunnel;

• Measurement points placed in the groundfrom the surface or from an adjacent tunnel.

5.2.1.1- Inner Markers

Inner Markers most often consist of opticaltargets or anchor points for Invar wiremeasurements. These devices must beanchored at sufficient depth so that theactual displacements of the sprayedconcrete and/or ground can be monitored.The accuracy of modern topographicdevices allows a simultaneous interpreta-tion of displacement records to be made,both in terms of convergence and settle-ment. The expected error is of the order of3 to 5 tenths of a millimeter.

5.2.1.2- Ground measurement pointsinstalled from the tunnel

Displacements of the surrounding ground(with or without reinforcement) around thetunnel are measured using radial bore-holes equipped with anchor points whichdisplacements are monitored by means ofextensometers. The accuracy of the exten-someters is sufficient to monitor the dis-placements that can be expected, provi-ded the anchors are properly sealed intothe ground.

5.2.1.3- Ground measurement pointsinstalled from the outside

Measurement points placed from outsidethe tunnel are generally used for two parti-cular types of tunnels:

• Shallow tunnels in urban areas;

• Large excavations such as hydroelectricpower plant caverns.

In the case of urban tunnels, measure-ments are intended to monito the evolu-tion of settlements at the surface or nextto existing buildings. These measurementpoints consist of topographic surface markers, associated, in sensitive areas,with deep settlement markers or extenso-meters.

This apparatus (settlement markers orextensometers) are also used to monitor theevolution of ground deformations aroundlarge underground openings. In that case,the equipment is installed above the crownof the opening from a smaller (15 m2) ancil-lary gallery excavated parallel to the ope-ning. It is also worth mentioning an increa-sing use of electro-levels, in which measuredrotations are used to derive ground settle-ments.

5.2.2- STRESS MEASUREMENTS

These measurements are primarily intendedto evaluate the state of stress within thesprayed concrete layer and at the contactwith the ground. They make use of pressurecells, installed as shown in Figure 5.1, orvibrating wire extensometers placed withinthe concrete. Each of these devices pre-sents a number of advantages and shortco-mings that should be taken into accountwhen interpreting the results.

In the case of pressure cells, measurementsare mainly influenced by the type of surfacethey are placed against, as well as the instal-lation and measurement procedures. This isessentially true of radial cells ; the tangentialcells are considered more reliable becausethey are better encapsulated in the sprayedconcrete. Another feature of these devicesis that they lead to measurements that are

often erratic, which is the result of the actualdisplacements of the structure, as well ascracking and hardening phenomena withinthe young sprayed concrete (Golser et al.,1990).

Deformation measurements taken with thevibrating wire extensometers can be expec-ted to be very accurate and reliable, provi-ded the extensometer has been properlyencased and is perfectly fit for purpose.With thin sprayed concrete layers (10 to 15cm), the location of the device has littleinfluence on the measurements. However,when the thickness of the concrete reaches30 to 40 cm, bending moments may appearin the support. In that case, the location ofthe device has direct influence on the resultand must be taken into account in the inter-pretation of the monitoring data. Moreover,the exothermic reaction associated with thesetting of the young concrete affects thesensors’ response; it is essential that thisphenomenon is properly evaluated toachieve a satisfactory interpretation of theobserved deformations.

The interpretation of the data also requiresan accurate evaluation of the concretemodulus at the time the measurement istaken. Errors in stress estimates obtainedwith this device are mainly a result of inaccuracies in tests performed on cores ofthe same concrete at the same age, whichare assumed to be homogeneous andreflect the actual state of the concrete insidethe structure. Recent research on the rheo-logy of concrete has allowed significant pro-gress in this respect.

It should also be mentioned that straingauges are sometimes installed on steel ribsor rock bolts. These gauges are extremelyfragile, and their placement, as well as main-tenance, require special attention. Providedthese conditions are met, they can providevaluable information on the state of stress ofthe structure.

Globally, stress estimates derived fromdeformation measurements are nowadaysconsidered more reliable than pressure cellmeasurements, in spite of the assumptionsthat they require on the value of theconcrete modulus.

5.3- MONITORING SECTIONS

Monitoring sections are intended to pro-vide, all along the tunnel, the data neededto achieve a good appreciation of the geo-mechanical response of the ground and of

Figure 5.1 - Working principle of pressure cell

Sprayed concreteWelded wice mesh

Leveling mortar

Pressure cell

18


the behavior of the support structure. Oneusually differentiates between standardmonitoring sections and enhanced monito-ring sections.

5.3.1- STANDARD MONITORINGSECTIONS

These sections are easy to install, and, as aresult, used at several locations along thetunnel. They consist ofmarkers placed onthe inner liner surface to record conver-gences and settlements. The number ofmarkers depends on the tunnel geometryand construction phasing. It should be atleast three, inasmuch as this does notimpede construction activities. This type ofmonitoring section is appropriate for theexcavation stage as well as after completionof the works.

5.3.2- ENHANCED MONITORINGSECTIONS

These sections (Figure 5.2) are more heavilyequipped with sensors, as they are meant toprovide information on both deformationsand stresses. They consist of:

• Pressure cells (placed radial or tangent tothe section) or extensometers placed withinthe concrete;

• Radial extensometers, which length andnumber depend on the size of the tunnel: itis recommended that a minimum of twoanchor points be used with each extenso-

meter; the length of this apparatus must bedefined in accordance with the expectedmagnitude of displacements and should notbe less than one diameter of the tunnel;

• Convergence and settlement markers pla-ced on the inner surface of the liner next tothe tips of the extensometers.

In the case of urban tunnels or large under-ground openings, these sections can be valua-bly complemented with settlement indicatorsinstalled from the surface or with extensome-ters installed using ancillary galleries.

5.3.3- SPACING OF THE SECTIONSAND FREQUENCY OF MEASUREMENTS

The spacing and type of monitoring sec-tions should be selected in view of the typeof anticipated geology. The monitoring sys-tem should be defined in accordance withthe geomechanical properties of the groundand the depth of cover, as well as the envi-ronment of the tunnel or the undergroundopening.

For linear structures, standard monitoringsections can be placed at approximately 30 m intervals where geotechnical condi-tions are relatively homogeneous. In thecase of large openings, their location shouldalso account for geometrical changes incross-sections, as well as the environment ofthe opening (intersection with a tunnel oradjacent tunnel).

These sections are more difficult to install,therefore more costly. For this reason, theywill be preferably located, in the case oflinear structures, in areas where importantconvergences are expected and at refe-rence design sections. In the case of largeunderground openings, they will be distri-buted along the structure with a maximumspacing in the order of 50 m.

Whatever the type of monitoring section,the frequency of measurements should beleft to the appreciation of the engineer res-ponsible for the monitoring program or thedesign. This frequency should take accountfor the different phases of construction andthe rate of observed deformations (in theparticular case of surface settlements thefrequency would need to be significantlyhigher, up to several per day, for monitoringsections located in the vicinity of the advan-cing tunnel face).

5.4- INTERPRETATION ANDBACK CALCULATION

5.4.1- GENERAL BACKGROUND

The main purpose of the monitoring pro-gram is to verify the adequacy of the sup-port structure with the geomechanicalconditions that are actually encountered,and as such the measured stresses anddeformations can be expected to stabilizewith time. It is therefore essential to focuson the relative variations in measurements.The magnitude of deformations required toreach equilibrium depends on the depth ofthe tunnel and the characteristics of theground. Moreover, in the case of urbanunderground structures at shallow depth,the deformations must be compatible withthe sensitivity of adjacent buildings. In ana-lyzing the results, one should therefore takeaccount of the depth of the tunnel and itslocation, as well as its urban or non-urbannature. The interpretation of the results interms of structural response therefore dif-fers from a mere compilation and reductionof the monitoring data. It requires enginee-ring know-how (refer to the document fromthe AFTES on ”Organisation de l’ausculta-tion“, Tunnels et Ouvrages Souterrains, n°149, 1998).

5.4.2- CONVERGENCE MEASUREMENTS

5.4.2.1- General considerations

Convergence markers or extensometersshould be placed as close as possible to theFigure 5.2 - Entranced monitoring section

Bored extensometers (L=12m)

Convergence markerPressure cellVibrating wire extensometer

19


advancing front, and the first readings takenimmediately after installation. The distanceof installation behind the face should be lessthan that equivalent to one stope or one day.

Design calculations based on the solid com-posite method allow to account for thedeformations induced by each constructionphase, as well as long term effects; it isimportant that these aspects be used in theinterpretation of the convergence or settle-ment measurements. The methods propo-sed by Panet & Guénot (1982) and Panet(1995) can be used in order to comparelong-term predictions with observed data.These methods provide some appreciationof the ultimate convergence, based on mea-surements taken over a period of severalweeks to a few months.

Whatever the structure (tunnel or opening),the observed convergence rates, outsidethe zone of influence of construction,should always decrease with time. If this isnot the case, some kind of reinforcement ofthe support system must be providedimmediately. In sensitive urban areas, theallowable convergence rate may be redu-ced in order to achieve better control ofsurface settlements. Specific settlementmeasurements may be required in this casein order to adjust the construction phasingas needed. In all cases, the uncertaintiespertaining to the design and constructionof sprayed concrete structures requirecontinuous monitoring, with real time useof monitoring data.

5.4.2.2 – Real time response of instru-mentation data

Relative convergence measurements shouldbe analyzed in real time in order to allow thedesigner to take corrective action on thesupport system when this becomes neces-sary. The concept of warning levels can beintroduced to determine when such actionshould be triggered. It is however extre-mely difficult to assign any pre-establishedfixed value to this parameter, as it must beadjusted on the basis of the environment,depth of tunnel and behavior of the groundat "failure" (brittle failure, hardening, dila-tancy, etc…). It is mostly the rate of defor-mation (velocity) or its variation over timethat must be carefully monitored. Any sud-den change in the rate of deformation mustbe interpreted as a potential risk and betreated consequently.

Computational results, given the numberof assumptions required, must always be

checked with observed data. Comparisonsshould of course be made with dueaccount of the accuracy of measurements,especially with small magnitudes. However,the impact of these measurement errorstends to decrease when the rates and theirvariations are considered over a sufficientperiod of time.

5.4.3- EXTENSOMETERS PLACEDWITHIN THE GROUND

Extensometers placed in the groundthrough boreholes allow to measure relativeground deformations and compare them tovalues capable of causing yield within theground mass.

5.4.4- STRESS MEASUREMENTS

The difficulties of achieving reliable stressmeasurements in sprayed concrete and atthe ground-concrete interface are descri-bed in section 5.2.2. For this reason, it ispreferable to focus on the variations ratherthan total values of measured data. Factorssuch as the concentration and length of rockbolts have some influence on the stressescarried by the sprayed concrete layer. Forexample, when 4 to 5 m long rock bolts areused with a density of one or two per squaremeter, the ground pressure induced on thesprayed concrete liner is significantly reduced as a result of arching that developsbetween the bolts. Similarly, when no bol-ting is used, the heterogeneous nature ofthe rock or lack of contact between thepressure cell and the ground may causesome arching to develop over the pressurecell.

The quality of data obtained from tangentialcells is dependent on the response of thestructure: in the case of strong conver-gences, the sprayed concrete shell will tendto be subject to shear, and the pressure cellreadings will in no case reflect predictedvalues. In such case, convergence measure-ments will provide the only reliable indica-tions on structural safety.

5.4.5- BACK-ANALYSES

Back-analyses shall be based based on (1)total or relative deformations measured atenhanced monitoring sections (that aremore heavily instrumented), and (2) geologi-cal and geomechanical surveying conduc-ted at the face. Parameters for these ana-lyses may be obtained through in situtesting using devices such as the dilatome-ter, flat jack or pressuremeter.

It is also suggested that analyses be bestconducted in two phases:

• The first phase is based on simplified com-putational methods (e.g. analytical) and isaimed at evaluating the state of deforma-tion of the support structure (purely elasticor elasto-plastic) and at identifying the mostcritical parameters. This phase should helpverify if the basic design assumptions arecorrect. However, one must be aware of theinaccuracies that are inherent to theassumptions used in such approach andsubsequent computational results. It wouldbe unrealistic to try to reach a high level ofaccuracy in such back-analyses.

• Once the response mechanism has beenidentified, a finite element analysis can beconsidered. This would tend to be standardpractice for large underground openings(hydroelectric plant, storage). The sameconsiderations as indicated for the fristphase obviously apply as regards the accu-racy of computational results.

6- STRUCTURES NOT LINEDWITH CAST-IN-PLACECONCRETE

6.1- INTRODUCTION

Sprayed concrete was used for the first timeas final liner of an underground structure inthe late 1960's. This practice has developedsince then due to advances made in theunderstanding of the mechanical behaviorof the concrete and the ground. It stillremains uncommon however both becauseit is not suited to certain projects (withissues related to water tightness and surfaceroughness) and because of the reluctance ofsome stakeholders and lack of real case evi-dence.

On the other hand, sprayed concrete usedas part of non structural liners has becomerelatively popular (Type 1 and 2 sprayedconcrete), with applications in road tunnels(usually with low traffic), railway tunnels andhydraulic tunnels (particularly in Norway).

When used in a structural role (Type 3sprayed concrete), it is generally comple-mented with an inner liner of cast-in-placeconcrete. Nevertheless, a few structures ofvarious sizes and shapes have beenconstructed with sprayed concrete as finalliner. The most remarkable applicationsinclude the Furka (14 km) and Vereina(19 km) railway tunnels in Switzerland andthe Sèvres-Achères water treatment cham-

20


ber (61 m long with 160 m2 cross-section) inFrance.

Finally, it is worth noting that sprayedconcrete can also be used as inner liner inthe the refurbishment of older tunnels(S.N.C.F., S.I.A.A.P.).

6.2- AREA OF APPLICATION

Sprayed concrete final liners differ fromcast-in-place concrete liners on the follo-wing aspects:

• The placement technique (no form workrequired, need for an appropriate construc-tion ventilation system, ability to spray thinlayers of concrete, ability to adjust theshape of the concrete shell to the actualexcavated profile);

• The type of surface (higher surface rough-ness).

On the other hand, with the exception ofyoung concrete - which is of lesser impor-tance when it comes to final liner (in compa-rison with initial support systems) - sprayedconcrete presents mechanical propertiesthat are similar to those of cast-in-placeconcrete (with typical 28 days strengths ofσb = 25 MPa to 30 MPa).

Consequently, the situations where the useof sprayed concrete as final liner could be ofparticular interest are:

• Short tunnels, tunnels with varying crosssections, underground intersections, forwhich it would not be economical to recurto prefabricated form works (e.g. theChauderon railway station in Lausanne);

• Caverns;

• Structures where the introduction of formworks is difficult (sewers tunnels, etc.);

• Tunnels with low operational require-ments.

In these conditions, the main advantage ofsprayed concrete used as final liner is costreduction

• Reduction in the volume of excavatedmaterial, which can be equivalent to thethickness of the cast-in-place concrete ring(at least 30 cm thick) when its installationcan be avoided;

• Reduction in costs and construction time,if the installation of a cast-in-place concretering can be prevented, even though thismay lead to a slightly thicker layer ofsprayed concrete. In the case of heteroge-neous or fractured grounds where theamount of over-cutting is significant, sub-stantial savings can be expected in compari-

son with the costs associated with the instal-lation of a full ring of cast-in-place concrete.

6.3- SPECIFIC LIMITATIONS

Reservations are often put forward by desi-gners, contractors and operators when itcomes to sprayed concrete being used asfinal liner, for the following reasons:

• Difficulty to make the tunnel watertight;

• Surface roughness, air friction and sensiti-vity to dust;

• Heterogeneity and variability of theconcrete characteristics;

• Difficulty to guarantee its durability.

Difficulty to attach inserts to the liner is alsoquoted as one shortcoming for this type ofstructure. These aspects are further develo-ped in the following sections along withpossible remedial steps that can be taken tocope with these limitations.

6.3.1- WATER-TIGHTNESS

In many cases, a structure entirely made ofsprayed concrete will provide a level ofwater-tightness which is satisfactory to meetthe operation requirements of the tunnel.Water ingress can be controlled using drainsor a secondary non-structural shell (that canincorporate an architectural finish).

on the other hand, the use of sprayedconcrete as inner liner is not recommendedin the following cases:

• High water head;

• Highly pervious grounds (AFTESClassification, Tunnels et OuvragesSouterrains, n° 28, 1978);

• Environmental conditions prohibitingwater table drawdown;

• Functional requirement to achieve fullwater-tightness.

Some practical evolution is however fore-seeable in this field in the near future. Ofparticular interest in this respect is the caseof the railway station of Chauderon inLausanne, where the water-tightness sys-tem, made of a geomembrane, was succes-sively coated with a final liner made of a finewire mesh, a dry-sprayed mortar (a few milli-meters thick) and wet sprayed concrete.Feedback from this case history in thecoming years will certainly contribute toexpanding the scope of application ofsprayed concrete used as final liner.

6.3.2- SURFACE ROUGHNESS

The use of sprayed concrete as final linermay result in specific provisions being requi-red as regards operation and safety equip-ment, and possibly with the architecturalcoating; such provisions must be introducedat the design stage.

For road and motorway tunnels, the higherroughness of sprayed concrete walls results,on the one hand, in faster dirtying by vitia-ted air and, on the other hand, in increasedair resistance, with subsequent incidence interms of required ventilation capacity andassociated equipment installation and ope-ration costs.

It may also be necessary, in order to guaran-tee sufficient brightness and sensation ofcomfort for the user, to cover the verticalsurfaces of sprayed concrete walls with clad-ding (in which case provisions shall be madeto prevent the cladding to be torn off byvehicles).

For hydraulic structures, the poor hydraulicproperties of the sprayed concrete surfacescan be appropriately overcome by the intro-duction of plastic coating (e.g. high densitypolyethylene). It can however be observedthat a number of hydraulic transfer struc-tures are lined with sprayed concrete, whichtends to balance arguments on this factor oflimitation. The effective roughness due tocross-section changes is in that case muchmore penalizing in terms of head loss thanthat produced by the roughness of sprayedconcrete walls.

6.3.3- HETEROGENEITY ANDVARIABILITY OF CONCRETE PROPERTIES

Current technology in terms of placementof sprayed concrete leads to:

• Some degree of material heterogeneity,especially within the first layer sprayed ontothe ground, which water-cement ratio is dif-ficult to control and contains acceleratorsthat may result in reduced long-termstrength;

• A variability of concrete properties that isprobably higher than for cast-in-placeconcrete.

Moreover, the geometry of the sprayedconcrete shell is necessarily either less regu-lar than for cast-in-place concrete or ofvariable thickness.

Given these considerations, the followingprovisions should be made in the designprocess:

• strength characteristics should be taken

21

lower than those obtained from cores takenin situ ;

• a nominal thickness equal to the minimumshell thickness should be used.

6.3.4- DURABILITYSprayed concrete has been used as finalliner for some 30 years, which is less thanthe expected lifetime of tunnels. This lack ofreal case record on the long-term perfor-mance of such structures justifies somedegree of caution. In practice, the Owner’sRepresentative may elect to suggest theintroduction a sacrificial thickness ofsprayed concrete thickness in the design, inorder to cope with the risk for degradationof the sprayed concrete mechanical proper-ties in the very long term.

6.4- CONCRETE COMPOSITION

The composition and placement of sprayedconcrete used as final liner must be analy-zed and adjusted in order to guarantee itsdurability over the entire life of the struc-ture. From this standpoint, the recommen-dations of section 4.5 relative to the condi-tions for accounting for the sprayedconcrete layer in the design of the final linerremain applicable. In particular:

• Admixtures shall be selected such thatthey shall not deteriorate the long-term pro-perties of the concrete. provisions shall alsobe made to verify, through very long-termtests, that the mechanical properties - nota-bly strength - used for design are met ;

• Longitudinal joints, if required with multi-layer placement, should be checked forquality and location (so that the presence ofa discontinuity through the entire shell thick-ness can be prevented) ;

• Potential corrosion issues related to theuse of metallic fibers should be addressed ;special steps should be taken (application ofan additional non-fiber reinforced layer orallowance for a 2 to 3 cm sacrificial thicknessin the design computations) to ensure therequired resisting shell thickness isachieved ;

• The final geometry (thickness) of the shellshall be controlled in situ.

6.5- DESIGN OF THE FINALSPRAYED CONCRETE LINER

From a computational point of view, twosituations can arise:

• A Type 3 sprayed concrete support is usedas final liner without any additional support.

In this case, the principles presented in sec-tions 4.4 and 4.5 apply, with the restrictionthat the thickness of the sprayed concreteliner should be checked for long-termconditions. The deformation modulus ofsprayed concrete to be included in the cal-culation should be evaluated on the basis ofconventional formulas (regulatory or other)for the determination of the long termdeformation modulus (cf. Chapter 4) ;

• The final liner is made of one or morelayers of sprayed concrete placed in lieu ofcast-in-place concrete. In this case, the veri-fication of the liner should be based onmethods pertaining to cast-in-placeconcrete, with particular attention beingpaid to the evaluation of the concrete cha-racteristic strength, σb which should allowfor some degree of dispersion in relation tothe spraying process. The loads applied tothe structure should be evaluated usingavailable conventional methods; Short andlong term modulus values should be evalua-ted using conventional formulas (regulatoryor other) for the determination of long termmodulus (cf. Chapter 4). It must however bekept in mind that some degree of creepbehavior must be assumed for the groundand/or liner for the loads to be transferredto the inner sprayed concrete layers.

REFERENCESREFERENCES ••••••••••••••••CHAPTER 1

AFTES - Work Group n°6, 1974, Recommendations on the "Mise en oeuvre du béton projeté dans les travaux souterrains", Tunnels etOuvrages Souterrains n° 1, pp. 40-47.

AFTES - Work Group n° 6, 1979, "Présentation de la méthode de construction des tunnels avec soutènement immédiat par béton pro-jeté", Tunnels et Ouvrages Souterrains n° 31, pp. 13-21.

AFTES - Work Group n° 6, 1994, Recommendatrions on "La Technologie et la mise en oeuvre du béton projeté renforcé de fibres",Tunnels et Ouvrages Souterrains, n° 126, pp. 307-317.

Health & Safety Executive, 1996, Safety of New Austrian Tunnelling Method (NATM) Tunnels : "A review of sprayed concrete lined tunnelswith particular reference to London clay", HSE Books, 86 p.

Heuer, R.E., 1974, "Selection / Design of Shotcrete For Temporary Support", Use of Shotcrete For Underground Structural Support, ACI-ASCE, pp. 160-174.

Institution of Civil Engineers, 1996, "Sprayed concrete linings (NATM) for tunnels in soft ground", ICE design and practice guides, ThomasTelford Publishing, 88 p.

Pöttler, R., 1990a, "Time-dependent rock-shotcrete interaction : a numerical shortcut", Computers and Geotechnics, Vol. 9, pp. 149-169.

Pöttler, R., 1990b, "Konsequenzen für die Tunnelstatik aufgrund des nichtlinearen Materialverhaltens von jungem Spritzbeton", Felsbau 8,Nr. 3, pp. 121-128.


22


CHAPTER 3

Abdul-Wahab, H.M.S., Ahmad, H.K., 1992, "Fibre effect on cracking of concrete due to shrinkage", 4th RILEM International Symposiumon Fibre Reinforced Cement and Concrete, R.N. Swamy Ed., E & F.N. Spon, pp. 102-113.

AFREM, 1995, Workshop "Les bétons de fibres métalliques – Méthodes de dimensionnement ; essais de caractérisation, de convenance etde contrôle", ed. P. Rossi, 73 p.

AFTES - Work Group n° 6, 1979, "Présentation de la méthode de construction des tunnels avec soutènement immédiat par béton pro-jeté", Tunnels et Ouvrages Souterrains n° 31, pp. 13-21.

AFTES - Work Group n° 6, 1994, Recommendations on "La Technologie et la mise en oeuvre du béton projeté renforcé de fibres", Tunnelset Ouvrages Souterrains, n° 126, pp. 307-317.

Aldrian, W., 1991, "Beitrag zum Materialverhalten von früh belastetem Spritzbeton", Thesis, Leoben.

Chang, Y., 1994 "Tunnel Support with Shotcrete in Weak Rock - A Rock Mechanics Study", Ph.D. Thesis, Royal Institute of Technology,Stockholm, 166 p.

Fischnaller, G., 1992, Diplomarbeit, Fakultät für Bauingenieurwesen und Architektur, Innsbruck University.

Golser, J., Schubert, P., Rabensteiner, K., 1989, "A new concept for evaluation of loading in shotcrete lining", Congress on ‘Progress andInnovation in Tunneling’.

Golser, J., Rabensteiner, K., Sigl, O., Aldrian, W., 1990, "Kontrolle der Spritzbetonbeanspruchung im Tunnelbau". In Berg-undHüttenmännische Monatshefte, Leoben, vol. 135, Heft 10, pp. 376-383.

Hashin, Z., 1962, "Elastic moduli of heterogeneous material", Journal of Applied Mechanics, 29, pp. 143-150.

Huber, H.G, 1991, "Untersuchungen zum Verformungsverhalten von jungem Spritzbeton im Tunnelbau", Thesis, Innsbruck.

De Larrard, F., Leroy, R., 1992, "Relation entre formulation et quelques propriétés mécaniques des bétons à haute performance",Materials and Structures, Vol. 5, pp. 463-475.

Laplante, P., 1993, "Propriétés mécaniques des bétons durcissants : analyse comparée des bétons classiques et à très hautes perfor-mances", Etudes et Recherches des LPC, OA13, 299 p.

Neville, A.M., Dilger, W.H., Brooks, J.J., 1983, "Creep of plain and structural concrete", Construction Press.

Ong, K.C.G., Paramasivam, P., 1989, "Cracking of steel fibre-reinforced mortar due to restrained shrinkage", Fibre Reinforced Cementsand Concretes. Recent Developments, R.N. Swamy Ed., Elsevier Applied Sciences, pp. 179-187.

Petersen, A., 1989, "Geostatische Untersuchungen für tiefliegende Regionalbahnen am Beispiel", Hannover. Thesis, TU Hanovre.Forschungsergebnisse aus dem Tunnelbau, Nr. 12.

Pöttler, R., 1990a, "Time-dependent rock-shotcrete interaction : a numerical shortcut", Computers and Geotechnics, Vol. 9, pp. 149-169.

Pöttler, R., 1990b, "Konsequenzen für die Tunnelstatik aufgrund des nichtlinearen Materialverhaltens von jungem Spritzbeton", Felsbau 8,Nr.3, pp. 121-128.

Rabcewicz, L. von, Sattler, K., 1965, "Die neue österreichische Tunnelbauweise", Bauingenieur, Ann.40, Vol. 8.

Rokahr, R., Lux, K.H., 1987, "Einfluss des rheologischen Verhaltens des Spritzbetons auf den Ausbauwiderstand". Felsbau, vol. 5 (1987),Nr. 1, pp.11-18.

Schubert, P., 1988, "Beitrag zum rheologischen Verhalten von Spritzbeton", Felsbau, vol 6, Nr 3, pp.150-153.

CHAPTER 4

AFREM, 1995, Seminar "Les bétons de fibres métalliques – Méthodes de dimensionnement ; essais de caractérisation, de convenance etde contrôle", ed. P. Rossi, 73 p.

AFTES - Work Group n° 1, 1978, Recommendations for a "Description des massifs rocheux utiles à l’étude de la stabilité des ouvrages sou-terrains", Tunnels et Ouvrages Souterrains, n° 28, pp. 176-185.

AFTES - Work Group n° 6, 1994, Recommendations on "La Technologie et la mise en oeuvre du béton projeté renforcé de fibres", Tunnelset Ouvrages Souterrains, n° 126, pp. 307-317.

AFTES - Work Group n° 7, 1974, Recommendation for the "Choix d’un type de soutènement en galerie", Tunnels et Ouvrages Souterrainsn° 1, pp. 31-39

AFTES - Groupe des Travail n° 7, 1976, "Réflexions sur les méthodes usuelles de calcul du revêtement des souterrains", Tunnels etOuvrages Souterrains n° 14, pp 50-76.

AFTES - Work Group n° 7, 1998, "Recommandations relatives à l’utilisation du béton non armé en tunnel", Tunnels et OuvragesSouterrains, pp.387-395.

Barett, S.V.L., McCreath, D.R., 1995, "Shotcrete Support Design in Blocky Ground : Towards A Deterministic Approach", Tunneling andUnderground Space Technology, vol. 10, n° 1, pp. 79-89.

23


Barton, N., Lien, R., Lunde, J., 1974, "Engineering Classification of Rock Masses for the Design of Tunnel Support", Rock Mechanics, vol.6, n° 4, pp. 189-236.

Bieniawski, Z.T., 1974, "Geomechanics Classification of Rock Masses and its Application in Tunneling", Proceedings of the 3rdInternational Congress on Rock Mechanics, Denver.

Bieniawski, , Z.T., 1989, "Engineering Rock Mass Classifications – A Complete Manual For Engineers and Geologists in Mining, Civil andPetroleun Engineering", John Wiley & Sons, 251 p.

Boymond, B., Laigle, F., 1999, "Conception, modélisation et réalisation des ouvrages souterrains du projet CERN - LHC1", Proceeding ofthe International Study Days, AFTES, Paris, pp. 57-64.

Fernandez-Delgado, G., Cordin, E.J., Mahar, J.W., Van Sint Jan, M.L., 1981, "Thin Shotcrete Linings in Loosening Rock, The AtlantaResearch Chamber : Applied Research for Tunnels", U.S. Department of Transportation, UMTA-GA-06-0007-81-1.

Homgren, J., 1987, "Bolt-anchored steel reinforced shotcrete linings", Tunneling and Underground Space Technology 2, pp. 319-333.

Panet, 1995, "Le calcul des tunnels par la méthode convergence-confinement", Presses des Ponts et Chaussées, 177 p.

Panet, M., Guénot, A., 1982, "Analysis of convergence behind the face of a tunnel", Proceedings of the Symposium InternationalTunnelling ’82, Brighton, pp. 197-204.

Protodiakonov, M.M., 1965, Proceedings of the 4th International Congress "Strata Control and Rock Mechanics", Columbia University,New-York.

Terzaghi, K., "Rock Defects and Load on Tunnel Supports", Rock Tunneling With Steel Supports, R.V. Proctor & T. White Eds., CommercialShearing & Stamping Co., Youngstown, pp. 15-99.

Wandewalle, M., 1992, Dramix Tunnelling the World, 2nd edition, N.V. Bekaert S.A., Zwevegem, Belgique.

CHAPTER 5

AFTES - Work Group n° 1, 1978, Recommendations for a "Description des massifs rocheux utiles à l’étude de la stabilité des ouvrages sou-terrains", Tunnels et Ouvrages Souterrains, n° 28, pp. 176-185.

AFTES - Work Group n° 19, 1998, Recommendations for the "L’organisation de l’auscultation des tunnels", Tunnels et OuvragesSouterrains, pp.397-418.

ADDENDAADDENDA

PagesPages

SOMMAIRESOMMAIRE

I - MECHANICAL PROPERTIES OF CONCRETE - - - - - - - - - - - - 241 - TESTS ON SPRAYED CONCRETE - - - - - - - - - - - - - - - - - - - - - - 24

1.1 - Study 1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 241.2 - Study 2 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 24

2 - RESULTS OF THE E/RC RATION (MODULUS OVER COMPRESSIVE STRENGTH) - - - - - - - - - - - - - - - - - - - - - - - - - - 24

II - OBSERVATIONS AND IN-SITU MEASUREMENTS- - - - - - - - - 251 - BOIS DES CHENES TUNNELS - - - - - - - - - - - - - - - - - - - - - - - - 252 - THIAIS GALLERY- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 253 - GALLERY OF THE MONACO TUNNEL RAMP- - - - - - - - - - - - - 254 - QUATRE CHEMINS GALLERY - - - - - - - - - - - - - - - - - - - - - - - - - 255 - LAS PLANAS TUNNEL - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 266 - PESSICART TUNNEL - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 267 - FURKA BASE TUNNEL (SWITZERLAND) - - - - - - - - - - - - - - - - - 268 - CADI TUNNEL (SPAIN) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 27

III - DESIGN OF SUPPORT STRUCTURES - - - - - - - - - - - - - - - - - - 271 - UNDERGROUND BYPASS THE MONACO RAILWAY - - - - - - - 272 - CHANNEL TUNNEL CROSS OVER BRTISH SIDE - - - - - - - - - - 283 - CHAUDERON RAILWAY STATION (SWITZERLAND) - - - - - - - - 294 - DOMBES AND CÔTIERE TUNNELS - - - - - - - - - - - - - - - - - - - - 295 - MEYSSIEZ TUNNEL - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 30

IV - LITERATURE REVIEW - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 31• Summary of papers dealing with the design of sprayed

concrete used in underground works - - - - - - - - - - - - - - - - - - - - 32• On the rheology of sprayed concrete - - - - - - - - - - - - - - - - - - - - 33• Lattice girdes - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 34• Design of the sprayed concrete shell - - - - - - - - - - - - - - - - - - - - - 34• Effect of creep in young sprayed concrete used for tunnel

confinement - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 34• Stresses in sprayed concrete in tunnels - - - - - - - - - - - - - - - - - - - 35• Sprayed concrete final liner on a test section in the

Munich subway - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 35

24


II- MECHANICAL PROPERTIESOF CONCRETE

1- TESTS ON SPRAYEDCONCRETE

The following paragraphs present an over-view of two studies on sprayed concrete.

1.1- STUDY 1

This first study was conducted by Solens-Alpes during the repair of the St-Martin-La-Porte's tunnel in collaboration with theSNCF, the CETU and the LCPC. This studywas part of the BEFIM national project. Theobjective was to evaluate the elastic proper-ties of young sprayed concrete, as early as afew days for standard concrete (sprayedconcrete normally used in ground support)and as early as a few hours for the GISconcretes (Guaranteed Initial Strength).Four categories of sprayed concretes weretested:

• Classic concrete 0/8 premix S533 fromTECHNIA;

• "GIS" concrete 0/8 premix S555 fromTECHNIA;

Two FRSC (Fiber reinforced sprayedconcrete):

• Classic concrete with an addition of DRA-MIX ZP 30/50 from BEKEART to a dosage ofapproximately 15 kg/m3 of fibers in the testpanels and the slabs;

• "GIS" concrete with an addition of DRA-MIX ZP 30/50 from BEKEART to a dosage ofapproximately 25 kg/m3 of fibers in the testpanels and the slabs.

The experimental program included com-pressive strength tests at 5 ages (3 speci-mens per age) with an evaluation of theelasticity modulus. Also, punching-flexuraltests were performed at low speed on slabswhere the load-displacement evolution wasrecorded until failure or apparent cracking(2 slabs per age) occurred.

Compressive strength tests were performedfollowing the NF P 18.406 standard on spe-cimens with a 7.4 cm diameter and a slen-derness ratio of 2. The elastic modulus wasevaluated using a LCPC J2P extensometer.

1.2- STUDY 2

These tests were conducted in the SNCFlaboratory in St-Ouen by SIMECSOL.

The objective of this study was to identify asimple relationship between the compres-sive strength of young sprayed concreteand their modulus of elasticity.

Three types of wet and dry sprayedconcrete were considered:

• Classic concrete B25 – 350 and 400 kg ofCLK 45, 425 kg of CPA 55 PMES;

• FRSC (Fiber reinforced sprayed concrete)B25 – 350 kg of CLK 45 – 45 and 60 kg ofmetallic fibers;

• GISSC (Guaranteed Initial StrengthSprayed Concrete) B25 – 360 and 380 kg ofCPA 55 PMES.

The experimental program included com-pressive strength measurements along withelasticity modulus.

Compressive strength tests were performedfollowing the NF P 18.406 standard on spe-cimens with a 6 or 6.4 cm diameter and aslenderness ratio of 2. The elastic moduluswas evaluated by measuring the distancebetween the plates of the test apparatus.

2- RESULTS OF THE E/RC

RATIO (MODULUS OVERCOMPRESSIVE STRENGTH)

Figure 1 presents the results obtained forthe elasticity modulus and compressivestrength for the two previously describedstudies.

The results presented by Laplante (1993) arealso included in Figure 1 for comparisonpurposes. Tests were performed on B40 and

B80 concretes. Compressive strengths ofyoung concrete were evaluated, along withelasticity modulus on specimens having an11 or 16 cm diameter and a slendernessratio of 2. The specimen’s deformation wasmeasured with the LCPC J2P extensometer,identical to the one used in the first study.

Comments on Figure 1 are:

• The tests performed in the second studyshow that for any type of sprayed concreteused, the ratio of the modulus to the com-pressive strength is between 160 and 460,the average being 310;

• The tests performed in the first study showthat this ratio is, on average, three timeshigher than that of study 2. Incidentally,these results are much closer to those repor-ted by Laplante (1993), even if the latterwere obtained on samples of ordinaryconcretes.

The type of device used for measuring thespecimen’s deformation is a key parameterin the evaluation of the elasticity modulus.Namely, the LCPC method, which reducesthe number of unknowns by measuringdirectly the specimen’s deformation leads toresults that are less scattered and presuma-bly closer to reality.

These results tend to show that the relation-ship proposed by Laplante (1993), based onthe model developed by de Larrard andLeroy (1992), can yield a satisfactory esti-mate of the concrete's elasticity modulus bycorrelation with its compressive strength.The reader will also notice that this relation-ship provides a reasonable estimate of the

Figure 1 - Modulus of elasticity against uniaxial compressive strength

Mod

ule

of e

last

icity

(MPa

)

Compressive strength (MPa)

Laplante’s reportConcrete:Ordinary concrete: B40HPC: B80

St-Martin-La-PorteConcretes:RIG S.55: B38Normal S.533: B43Fibers:

- no fibers-25 kg/m3 Dramix

SNCF - VOSNConcretes:B25 - 350 kg of CLK45B25 - 400 kg of CLK45B25 - 425 kg of CPA55PMESB25 - 350 kg of CLK45 with 60 kg

of metallic fibusB25 - 350 kg of CLK45 with 45 kg

of metallic fibus

25


experimental values obtained on sprayedconcrete specimens.

I- OBSERVATIONS AND IN-SITU MEASUREMENTS

1- BOIS DES CHÊNES TUNNEL

Characteristics:

Highway Tunnel (Highway A30) with a130 m2 section;

Cover: 30 m maximum;

Length: 300 m.

Other information:

Ground conditions:

• Toarcic marl;

• Sandy clay with gravely zones

• (c' = 100 kPa and ϕ' = 40o);

Staged face excavation (bench and hea-ding);

Support: rock bolting and sprayed concrete,with lattice girder as necessary.

Measurements:

• Sections monitored with pressure cells, inareas where lattice girders were used;

• Convergence measurements in top half-section and in vertical sections;

• Stabilization of convergence at 2 to 3 cm.

Information on sprayed concrete struc-ture:

Monitoring: a section of the tunnel wasequipped with 22 pressure cells placed onthe entire perimeter at the interface bet-ween the ground and the sprayed concrete.

Results:

• During excavation of the upper half:increased pressure and stabilization at 200-300 kPa at the crown and 100-200 kPa atthe base;

• During the excavation of the stross: thehorizontal convergence resulted in a reduc-tion in lateral pressure and a slight increasein the crown;

• After spraying the concrete on the verticalsections, the pressure increased by approxi-mately 50 kPa;

• In the long run (1000 days): the pressurestended to stabilize at 100-200 kPa. Someredistribution took place.

Note: dissymmetry between the right andleft side of the tunnel made it more difficultto conclude.

2- THIAIS GALLERYCharacteristics:

Round gallery with a 9 m2 section;

Depth: 60 m;

Length: 60 m.

Other information:

• Ground: Pantin marl;

• Shear strength: c' = 100 to 350 kPa and ϕ' = 20°);

• Modulus (pressuremeter): Ep = 80- to 100MPa;

• Hand mining: 1 m/day;

• Support: 10 cm of sprayed concreteapplied next to the face;

• Measurements:

– 12 relative convergence profiles and mea-surements by extensofor;

– Convergences remained very small (mar-kers placed at 50 cm from the working face):from 1 to 4 mm on vertical wires and from 2to 3.5 mm on the inclined wires;

– The extent of deformation zone was verylimited around the excavation.

Information on the sprayed concretestructure:

Monitoring: a monitoring section was instal-led in the middle of the gallery (PM 32,70),with 8 Gloetzl cells placed behind thesprayed concrete;

Results:

• Results extremely different between theright side (North) of the gallery and the leftside (impossible to explain based on groundconditions);

• Pressures measured shortly after theconstruction, and three years later, werevery small and below 50 kPa.

Conclusions:

• Convergence as well as pressure measure-ments show very good stability around theopening;

• Sprayed concrete is barely loaded.

3- GALLERY OF THEMONACO TUNNEL RAMP

Characteristics:

Circular testing gallery with a 9 m2 section,perpendicular to the investigation gallery;

Depth: 200 m;

Length of gallery: 16.7 m.

Other information:

Ground: Trias clay:

• Black plastic clay with slickenside andmarl-limestone with sandstone inclusions;

• Inclusions of clay or marl sections;

Excavation: by hand, approximately 1m/day;

Support: 5 to 10 cm of sprayed concrete.

Measurements:

• 11 relative convergence sections andextensofors;

• Convergences in the gallery: 2 to 4 cm;

• Localized displacements; 4 to 5 m deepzones around the gallery.


Monitoring: 2 monitoring sections eachequipped with 12 pressure cells placed bet-ween the sprayed concrete and the ground.

Results:

• Around the gallery, pressures remained inthe 200 to 500 kPa range;

• Some individual pressure cell measure-ments yielded higher values.

4- QUATRE CHEMINS GALLERYCharacteristics:

Experimental gallery:

Section: 12 m2;

Length: 38 m;

Excavation: September 1977-November1978.

Other information:

Ground:

• PM 0 to 30: superior cretaceous schistifiedmarl;

• PM 30 to 38: crushed Cretaceous marl-limestone.

Excavation: roadheader.

Support:

• PM 0 to 16.5: ribs and lagging, followedwith sprayed concrete;

• PM 16.5 to 22.5 (ring 1): 5 cm of sprayedconcrete + wire mesh + HA 32 bolts;

• PM 22.5 to 28.5 (ring 2): 5 cm of sprayedconcrete + wire mesh.;

• PM 28.5 to 33.5 (ring 3): sprayed concrete+ wire mesh followed by a cast in placeconcrete liner;

• PM 33.5 to 38.0 (ring 4): fiber glassanchors; fiber glass and resin liner; collapseafter 4-25-79.

Measurements:

• 13 convergence profiles

• 2 profiles with 3 extensometers, each

26


equipped with 3 rods;

• 3 profiles with glötzl cells.


Monitoring:

Results:

• The convergence measurements allowedto clearly identify the displacement pattern.They follow typical power time laws;

• Maximum pressure: 1 MPa; most often at0.2 MPa – irregular distribution;

• No differences between the responses ofrings #2 and #3 (limited anchor role);

• Experimental convergence-confinementrelationship: 5 cm of sprayed concrete +wire mesh; E = 12 000 MPa.

5- LAS PLANAS TUNNELCharacteristics:

Highway tunnel – Highway A8 – bypassaround Nice;

Section: 100 m2 of interior section, 125 m2

excavated in marl;

Maximum cover: 50 and 90 m;

Lengths: 140 and 220 m in marl.

Other information:

Ground:

• Plastic marl: σi = 4 to 7 MPa;

• c' = 100 kPa; ϕ' =27o.

Excavation:

• Heading: 50 m2

• Stross: 50 m2

• Sole: 25 m2

Support:

• Sprayed concrete: 10 cm with 150 x 150 x10 mesh

• Steel bars anchored with grout, L = 5.0 m,spaced at 1.5x3.0 or 1.5x1.5 adjusted onthe basis of convergence measurements

• TH 21/48 arch girders at 3 m in some areas

• Temporary footings of 20 cm placed everyweek-end in the top half of section.

Convergence measurements: maximum 80mm with an average of 30 mm.


Initial section with 50 m cover:

• Radial pressure cells: uniform pressure of150 kPa, which increases to 200 kPa duringthe excavation of the stross.

• Tangential pressure cells: they show largevariations over time (stresses redistribution);

rapid increases to 800-900 kPa and reduc-tion to 400-500 kPa on the lateral walls and200 kPa at the crown. Little variation duringthe construction of the stross.


• Radial pressure cells: crown cells at 350kPa, the vertical wall cells show identicalpressures of 180-200 kPa.

• Tangential pressure cells: only one cellworked on the vertical wall: it reached 650kPa before dropping to 300-400 kPa.

There are no long-term measurements avai-lable after placement of the inner plainconcrete liner.

6- PESSICART TUNNELCharacteristics:

Highway tunnel – Highway A8 – bypassaround Nice;

Section: 100 m2 of interior section, 125 m2

excavated in marl;

Maximum cover: 65 m;

Length: 160 m in marl.

Other information:

Ground:

• Plastic marl: σi = 4 to 7 MPa;

• c' = 100 kPa; ϕ' =27o.

Excavation:

• Heading: 50 m2

• Stross: 50 m2

• Sole: 25 m2

Support:

• Sprayed concrete: 10 cm with 150 x 150 x10 mesh

• Steel bars anchored with grout, L = 5.0 m,spaced at 1.5 x 3.0 or 1.5 x 1.5 dependingon observed convergence measurements

• TH 21/48 arch girders at 3 m in some areas

• Temporary footings of 20 cm placed everyweek-end in the top half section.

Convergence measurements: maximum 20mm with an average of 5 to 10 mm.



Radial pressure cells: uniform pressures of150 kPa, which increases to 200 kPa duringthe excavation of the stross.

Tangential pressure cells: 200 kPa for thecells in the tunnel’s axis at the crown duringthe excavation, increasing to 450 kPa duringthe construction of the stross.

The pressure on the lateral wall increased to

900 kPa during the excavation of the strossand dropped to 500 kPa in the long-term.

There are no long-term measurements avai-lable after placement of the inner unreinfor-ced concrete liner.

7- FURKA BASE TUNNEL(SWITZERLAND)Characteristics:

Electrical railway tunnel, 1 narrow lane.

Length: 15.4 km

Section excavated: 26 to 42 m2, most oftenin a horseshoe shape, otherwise elliptical.

Maximum cover: 1520 m; intermediateaccess point in the middle of the tunnel

Other information:

Ground:

• 11 km of St-Gothard's gneiss;

• Very strong granite on 3.5 km;

• A few fractured zones;

• Water ingress that could reach, locally, 100to 200 l/s.

Excavation:

• Drill and blast: from 1972 to 1982;

• Sealing of many aquifer zones by polyure-thane injection; collection of surface runoffswith half PVC pipes fixed by spraying resin.

Support: 16 types of supports (differentones every 20 m on average):

• On 55% of its length: 5 cm of sprayedconcrete (not on the total section), withoutbolts or mesh;

• On 26% of its length: 15 to 40 cm ofsprayed concrete with mesh and 4 to 11bolts per meter (fiber glass bolts, anchoredwith resin);

• After carefully draining and cleaning of therock with water jets, spraying of concrete;

• Little amounts of setting accelerator,hence the high ultimate strengths (40 MPa);

• Sprayed concrete final liner, put in placefar behind the support.

Measurements:

• Numerous convergence sections withIseth extensometers and borehole extenso-meters with automatic logging;

• Long-term monitoring on critical sections(1 measurement/month);

• Continuous implementation of monitoringresults to determine the type of final linerand its moment of application, as for theAlberg tunnel (in such long and narrow tun-nels, only sprayed concrete final liner offersthat kind of operational flexibility);

27


• Local disorders (swelling of the face, spal-ling of the granite) well controlled, either byadditional rock bolting or reinforcement ofthe section, until an elliptical shape is rea-ched.

ReferencesAmberg, R., 1983, "Design andConstruction of the Furka Base Tunnel"Rock Mechanics and Rock Eng., vol 16, N°4p. 215-231.

Amberg, R. & Sala, A., "Shotcrete asPermanent Lining for the Furka BaseTunnel", Rock Mechanics and Rock Eng., vol17, N°1, 1984, p. 1-14.

8- CADI TUNNEL (SPAIN)

CHARACTERISTICS:

Two-way tunnel, on the Toulouse-Barcelonaaxis, through Puymorens;

Section: 82 m2 on average (width 10.5 m atbase);

Maximum cover: 980 m;

Length: 5 km.

OTHER INFORMATION

Ground:

• Paleozoic era with very complex tectonic,with faults and erosion thrust, made mainlyof limestone and shale; graphitic shale ofthe Silurian period on 6% of its length.

• Ground is probably unpervious (no hydrogeological comments).

Excavation:

• Drill and blast from 1982 to 1984, with 3.5to 0.8 m passes depending on the ground,with top, heading and sole.

• Four excavation fronts: North, South and 2others from the investigation gallery(length: 1.7 km)

• Construction time and cost within 1% ofdesign values.

Support:

• 4 typical profiles consisting of 10 to 45 cmof sprayed concrete applied in successivelayers (1 to 3) depending on the monitoringresults, sometimes many months after exca-vation (Rc = 30 MPa); immediate rock bol-ting (3 to 4 m long bolts), anchored withresin (1 bolt/4 m2 to 1 bolt/m2 density); THarch girders and rapid closing of the sectionin graphitic shale.

• No cast in place liner; water tightness pro-vided by a polyethylene membrane placed

on the crown and precast concrete walls(H = 4 m) on each vertical wall.

Measurements:

• Convergence measurements made withINTERFELS wires: monitored sections 20-25 mapart on average and as close as 10 m inpoor ground conditions; 2 horizontal wires(1 reading/day to 1 reading/month);

• Monitoring was an essential element inchecking and adapting the support system(thickness, bolting, time to first layer ofsprayed concrete and between successivelayers);

• One special monitoring section in the gra-phitic shale, with multiple extensometersand Glötzl radial cells that measured confi-nement pressures of 100 to 240 kPa.


Variable convergence velocity, from severalcm/day to 10-2 mm/day in the stabilizingphase;

In the Silurian, very strong convergence (upto 50 cm) were expected and dealt withthanks to a gradual stiffening of the crown,with the last layers of sprayed concretebeing less and less loaded.

Monitoring of the long-term convergenceduring at least 10 years on 11 sections of2 wires each; stabilization was reachedeverywhere (residual velocities of 0.1 to0.5 mm/year).

ReferencesVidal Font, J. 1985, "Auscultation du tunneldu Cadi" Tunnels et Ouvrages Souterrains,N° 70, July-August 1985, pp. 169-172.

Serratosa, A., 1986, "Un atout pourl’Europe : le tunnel du Cadi (Espagne)",Tunnels et Ouvrages Souterrains, N° 73,Jan.-Feb. 1986, pp. 11-17.

Monclus Jurado, B., 1992, "Le tunnel duCadi", CEIFICI Workshop "Des tunnels pourl’Europe", Paris, Dec. 1992., 37p.

III- DESIGN OF SUPPORTSTRUCTURES

1- UNDERGROUND BYPASSTHE MONACO RAILWAYGeneral characteristics:

• Design: SNCF

• Construction: Group SOGEA-COGEFAR-BORIE- NICOLETTI-SPADA-GTM

• Construction period : 1994-1997

Geometrical characteristics :

• Width: 11 m to 25 m (undergroundrail station)

• Height: 11 m to 13 m

• Lenght: 2250 m

• Cover at crown: 70 to 180 m

• Water above the crown: 0 m

Support characteristics: variable depen-ding on the ground conditions (see Table 1)and the selected profiles:

• Ground A: staged excavation, 5 cm ofsprayed concrete + bolts

• Ground C: full face excavation: 27 cm ofsprayed concrete and HEB

• Grounds B, C, or D: full face exvavation:22 to 32 cm of sprayed concrete + steel ribsand bolts

Geomecanical characteristics(see Table 1)

Design model for sprayed concrete:

• Elastic behavior with constant elasticitymodulus equal to 10 000 MPa (short term)

• Allowable stress in concrete: 15 MPa incompression and 1.25 MPa in tension

• Shell reinforcement by the arch girders:modeled by homogenization

• Bolts modeled using the “equivalent pres-sure” approach.

N.B.: Once the cast in place concrete is inplace and active, the bolts and steel ribs are”removed“ in the model and the supportsprayed concrete is kept with a long-termmodulus of 5 000 MPa.

Table 1

28


General characteristics:• Design: ILF (R Pöttler)• Construction: TML UK• Construction period : 1989-1991

Geometrical characteristics :• Width: 21.20 m• Height: 15.40 m• Lenght: 164 m• Cover at crown: 71.30 m

Support characteristics: • Staged excavation• Sprayed concrete shell: thickness rangingfrom 15 to 20 cm (1st layer) + 10 cm (2nd layer)• Bolting: 1 bolt per 3 to 4 m2 density.

2- CHANNEL TUNNEL CROSS OVER BRITISH SIDE

29


Geomecanical characteristics(see Table 1)

Design model for sprayed concrete:

Elastic-plastic behavior (see Table)

• Variable elastic modulus adjusted on thebasis of the duration of the phase;

• Yield value: 21.25 MPa;

• For the initial phase (excavation and shot-creting) a "hypothetic elasticity modulus" isconsidered (HME).

The reinforcement of the shell brought bylattice girders and of the ground by thebolts are not considered in the calculations.

Main results expected:

• Settlement at crown: 40 to 50 mm

• Horizontal convergence: 30 to 40 mm

• Stresses in the sprayed concrete:

– Lateral galleries before crown excavation:2 to 7 MPa

– Lateral galleries after crown excavation: 13to 16 MPa

– Crown short term: 6 to 7 MPa

– Crown long term: 12 MPa

Monitoring equipment:

• 16 profiles with a total of 200 instruments:

– Tangential pressure cells (stress in thesprayed concrete): 36 placed on 2 profiles

– Triple extensometers (3, 6, 9 m) in theground: 19 out of which 13 are on 1 profile

• Bolts equipped with strain gauges: 28

RESULTS OF SPRAYED CONCRETE MONI-TORING:

• Tangential pressure cells: compressivestress in sprayed concrete between 4 and 6MPa (a few readings at 11 MPa)

• Radial pressure cells: stresses induced bythe surrounding ground between 0.5 and0.6 MPa

N.B.: the readings are insufficient beforecrown excavation to reliably evaluate thepressure on the inner sides of lateral galle-ries.

3- CHAUDERON RAILWAYSTATION (SWITZERLAND)

General characteristics:

• General contractor: Compagnie duChemin de Fer Lausanne-Echallens-Bercher

• Underground Construction, shafts, stationand tunnel: Consortium Losinger, Deneriaz,

Reymond S.A., Locher

• Construction period : 1992-1995

• Ground conditions: aquitanian molassetopped with moraines.

Geometrical characteristics :

• Lenght: 710 m

• Width of the station: 12 to 20 m

Support characteristics:

• Staged face excavation (station);

• B30-40 concrete according to Swiss stan-dards (B30 in France);

• Modulus: E = 20 000 MPa (which takesinto account cracking of the sprayedconcrete)

Design model:

• Given the size of the train station, initialdesign and phasing analyses were perfor-med using the finite element method (forinitial and final liners):

– 30% of the loads are redistributed throughface deformation;

– 40% applied in the short term to thesprayed concrete and the lattice girderswhen using E = 20 000/2.5;

– 30% remainder of loads are applied to thesupport with a modulus of E = 20 000 MPa.

• Calculations are completed, for all sec-tions, based on a ground reaction curvemodel calibrated against the finite elementresults at a few sections.

• Design of the sprayed concrete/lattice gir-der based on the BAEL rules (FrenchStandard): iterative calculations where anew cracked section is determined for each(M,N) values.

• Thickness of the first sprayed concrete (30to 40 cm) considered without any steel ribsor wire mesh, these are supposedly corro-ded in the long-term;

• Interior concrete for water tightness consi-dered with its reinforcement;

• The inner and outer concretes form a

monolithic layer;

• For safety, in the long-term,: verificationwith a resistance factor F = 1 for the soleinner concrete (and rebar);

• No creep analysis were made.

Main results:

• Finite element calculations:

– Uniform settlement of 3 cm, confirmed onsite;

– Long-term: the long-term stresses suppor-ted by the sprayed concrete are evaluatedassuming that the entire load of the groundabove is applied.

• The Gloetlz pressure cells placed in theliners did not yield any result.

References

Presentation O. Tappy, Engineer SIA EPFLdiploma, June 1995 – AFTES Work Group#20

4- DOMBES AND COTIERETUNNELS


• Tunnels are 500 m (La Dombes) and 300 m(La Côtière) long;

• Supported by steel ribs and sprayedconcrete (classic B25);

• The crown of the La Dombes tunnel is 24m deep and 14 m deep for the La Côtièretunnel;

• The entire excavation took place in yellowalluvium.

Finite element analysis:

• Calculations were conducted using theCESAR-LCPC FEM software;

• Phases of construction accounted for inthe analysis of both tunnels: staged faceexcavation with subsequent placement ofsupport and full section spraying, includingthe footing of the final liner;

• Phases of construction different than what

30


was executed for the lower half section: cal-culations assumed full face excavation, whe-reas staged excavation was used with cen-tral stross and lateral support walls; inaddition, support installation was delayeduntil excavation of the side drifts and rockbolts were abandoned;

• Support system considered: 200 HEBevery meter with 25 cm of sprayed concreteand 4 HA 25 3.80 m long bolts on the lateralsupport walls;

• Support by steel ribs and sprayedconcrete is included as an homogenizedequivalent section; redistribution of calcula-ted stresses in the ribs and the sprayedconcrete according to their relative stiffness;

• Modulus of deformation taken into themodel:

– For sprayed concrete:10 000 MPa for theshort term and 15 000 MPa for the longterm;

– For the liner: 30 000 MPa for the shortterm and 15 000 MPa for the long term;

Monitoring:

• Two profiles were instrumented:

– At PM 190 for the La Dombes tunnel;

– At the PM 44.5 for the La Côtière tunnel.

• Vibrating wire extensometers were fixedto the rib of one section, to monitor the ribdeformations; it allowed the evaluation ofthe stresses in the arch rib and derive thestresses in the sprayed concrete;

• Results were interpreted using a modulusof 10 000 MPa for the short term and 7 500MPa for the long term for the sprayedconcrete and 22 500 MPa for the final liner.

Analysis of results and comparison bet-ween prediction and monitoring / LaDombes Tunnel:

• Stresses in the support at monitored sec-tion:

– The measured stresses during the excava-tion of the upper half section (phase 2 of thefinite element calculations) are higher in therib and the sprayed concrete than predictedby the FE analysis (twice as much) with aslight dissymmetry;

– The measured stresses during the installa-tion of the slab (phase 6 of the finite ele-ment calculations) in the arch girder and thesprayed concrete are close to those predic-ted ones by the FE analysis, but remainshigher than the predicted values for theupper half section of the rib.

• Deformations in the support at monitoredsection:

– The convergence measured in situ arealmost identical to those predicted inPhases 2 and 6;

– In situ survey measurements are slightlyhigher than the predicted values for phase2; they are much more important than thepredicted values for phase 6 (4 to 5 timeshigher).

• Deformations in the support:

– The comparison is limited to the upperhalf section, the phases used to model thelower half section being completely diffe-rent from what was finally done;

– The settlement associated to the excava-tion of the upper half is 1 to 2 times moreimportant than predicted before the exca-vation of the lower section;

– The convergences are globally of thesame order of magnitude as those predic-ted.

• Pressure cells at the monitored section:

– The various pressure cells placed at thesupport/ground contact indicated relativelysmall pressures;

– Only cells located on the East side arehighly loaded, with a pressure in the orderof 350 to 500 kPa.

5- MEYSSIEZ TUNNEL


1800 m long tunnel;

Support made of a sprayed concrete shellalong with steel ribs in the upper half sec-tion and bolts for the lower half section;

Classic sprayed concrete (B25 with 425 kgof CPA 55 cement);

Tunnel entirely excavated in gravel-sandmolasses and marl-sand molasses;

Coring was done through the shell to eva-luate the modulus of elasticity in the labora-tory.

Finite element analysis:

Calculations made using the CESAR-LCPCFEM software;

• A number of finite element analyses wereperformed:

– At PM 280 (from PM 0 to PM 450, crownat 54 m deep),

– At PM 650 (from PM 450 to PM 1000,crown at 84 m deep),

– At PM 1210 (from PM 1000 to PM 1400,crown at 74 m deep),

• Phases considered in the calculations:

– Staged excavation with placement of sup-port (lower half section completed in 2steps);

– Full section spraying, including the footingof the final liner.

In the upper half section, support by steelribs and sprayed concrete is modeled as ahomogenized equivalent section.

Tests and monitoring:

• In situ testing with flat jack:

– Sprayed concrete shell equipped with flatjacks at PM 35, 85, 239, 475, 651 and 1015;

– The modulus of the sprayed concrete isestimated from the stress-deformationcurves obtained with the flat jack test: a 3Dfinite element analysis was made to modelthe effect of a recess in the sprayedconcrete shell; the following relationshipwas established between the deformationmodulus of the concrete and the slope ofthe curves found by calculation: E =0.34*P/e where E = modulus (MPa) and e =displacement between the markers for apressure P relative to the initial state whereno pressure is applied to the jack (m).

• Sprayed concrete testing:

– Coring was performed at the same PMwhere the flat jack tests were performed.

Results:

• Pressure in sprayed concrete:

– For any section, the East side of the tunnel(left) underwent a more important pressurethan the other side (West-right);

– The pressures were relatively small (in thevicinity of 2 MPa) except for PM 239 whereit reached a value of 4.8 MPa;

– The evolution of the stresses over timeappeared regular.

• Deformation modulus of in situ sprayedconcrete:

– For any location of the test (except for PM475), the value of the modulus increasedover time;

– On the last testing campaign (February 25and 26, 1992), the modulus varied from 5470 MPa to 18 760 MPa (average of 11 545MPa and a standard deviation of 4490 MPa);these results confirm the high heterogeneityof the sprayed concrete.

31


• Deformation modulus of sprayed concretein the laboratory:

– The modulus measured in the laboratoryare higher for the PM 239, 650 and 1015(measurements made from August 20 toSeptember 3, 1991) than those obtainedwith the fat jacks; they are lower for PM 35,85 and 475 (February 25-26, 1992);

– The average values measured betweenAugust 20 and September 3, 1991 is 5921MPa, with a standard deviation of 1449MPa; for the tests of February 25-26, 1992,it is of 8578 MPa with a standard deviationof 3542 MPa;

Comparison prediction/monitoring:The phases of construction correspond tothe phases of the finite element analysis;

• Pressure:– The long-term pressure measured in situare always smaller (one fourth) than theones predicted, except for PM 239 (leftside) where the pressures are closer to thosepredicted, but still smaller (half);

• Deformation modulus:

– The modulus introduced in the finite ele-ment calculations are of 10 000 MPa for theshort term, 7 500 MPa for the medium termand 5 000 MPa for the long-term;

– The modulus measured in situ are veryscattered, but show a definite increase overtime (except for PM 1475);

– The average of the last phase is 11 545MPa, which is close to the value of 10 000

MPa taken in the short-term calculationsand close to the value of 8 570 MPa obtai-ned in the laboratory at the same date.

Notes:The correlation between the flat jack mea-surements and the deformation modulus islinked directly to the hypotheses made forthe 3D calculations; no parametric studieswere made.

It is possible that the modulus measured insitu and in the laboratory are close to thevalue taken in the finite element analysis,but that the pressures found in situ aregenerally smaller than the stresses found inthe finite element calculations.

Summary of papers dealing with thedesign of sprayed concrete used in under-ground works (by E. Leca, SCETAUROUTE)

Heuer, R.E., 1974, " Selection / Design ofShotcrete for Temporary Support ", Use ofShotcrete For Underground StructuralSupport, ACI-ASCE, pp. 160-174

(1) Three types of sprayed concrete can beconsidered for design: placement of a pro-tective skin preventing air or water weathe-ring of the ground; support action to helpthe ground support itself, in the case ofloads associated with fracture blocks forexample; placement of a support structure,aimed at supporting all or a part of theloads induced by the excavation (forgrounds with low resistance, function of thestress levels, or in the case of swellinggrounds).

(2) The use of sprayed concrete is particu-larly adapted to the second type, as long as

the support can only be effective after a cer-tain amount of time, corresponding to thedecompression of the crown; the third typeis a little more delicate since there is a risk offailure if the sprayed concrete does not har-den fast enough with respect to the periodover which the ground remains stable andalso if the placement (and setting) of theconcrete is to rapid, which could lead tostresses exceeding its capacity.

(3)For the first type (protective skin), therecommended thickness of concrete is 2 in.(with local tolerance of 0.5 to 1 in.) for wea-thering of the ground by air; when water ispresent, it is recommended to rapidlyconstruct a closed ring of sprayed concreteor reinforce the sprayed concrete skin withmesh.

(4) The following rules can be applied in thecase of the second type of sprayedconcrete. For tunnels with a diameter of 4 to

LETERATURE REVIEW

T he following text is a compen-

dium of articles related to the use

of sprayed concrete in under-

ground works. These summaries were pre-

pared by the AFTES Work Group #20 and

represent part of the literature used to

prepare this document, also presented by

this working group. The names of the

members of this working group and

authors of each summary, are mentioned

at the beginning of each text, along with

the full reference of the original paper.


EXEMPLES DE TUNNELS ET OUVRAGES SOUTERRAINS DONT LE REVÊTEMENT INTÉRIEUR EST CONSTITUÉ DE BÉTON PROJETÉSTRUCTURE

HEHLRATH Tunnels (R.F.A.) Road 1958/59 14,5 440 20 S.C. + Mesh Brown coal. Ligh hydrostatic

+ Arch girders GI 110 pressure

Luxembourg Outlet 1961/62 14,8 880 35 m S.C. + bolts l = 1,5 m in More or less compact sous la ville lsandstone sandstone

MEXICO Drainage 1968 60 à 140 90 000 50 S.C. 10 cm + mesh + bolts = 3 m Half consolidated rocks,loose and watery, Rc = 200 kPa, clay

FURKA ( Switzerland) Train 1972 /82 26 à 42 1 368 1 000 Dry S.C. + mesh + passive Good to weathered rock(profil S4 + S5) anchors

Nuremberg Metro Subway 1972 50 500 4,5 to 10 S.C. 15 cm + mesh + 3,5 m bolt + Sandstone with clay lens(R.F.A.) 15 cm wet S.C. final liner

Lehrental East Highway 1973 80 500 50 S.C. 20 cm + mesh + arch girders Molasses contact at crown, Deviation - ULM (R.F.A.) + 3 m bolts at 1,5 to 2,5 m compact and

cracked limestone at slab

LLORET Road 1974 70 150(Spain)

BOURGET lake Hydraulic 1977 6 to 7 12 000 1 000 S.C. + mesh according to ground Molasses and sandstonegallery – Rhône and going conditions

CADI (Spain) Road 1987 82 5 000 S.C. (10 cm to 50 cm per successive passes) + mesh + passive anchor

BIELEFIELD Metro (R.F.A.) Train 1987 35 104 5 to 10 S.C. support (15 cm) black fiber Compact clay+ lattice girders; liner, 10 cm wet S.C.

FRASDORF (R.F.A.) Sanitation 1989 3 200 100 Dry S.C. (2 cm) + microsilica Gravel, sand, salt + water head of 80 m

MUNICH Metro (R.F.A.) Train 1990 38 to 52 60 8 Dry S.C. (15 cm + 10 cm Sand and gravel on top of compact+ 10 cm) + mesh marls; water table at mid-section

BRASILIA Metro (Brazil) Train 1996 72 6 500 6 à 10 S.C. (21 cm + 20 cm) + lattice Highly shrinking soft clay, girder + mesh water table at mid-section

VEREINA (Grisons Train 1994-99 39 to 46 19 050 < 1500 S.C. (15 to 30 cm) + bolts + ligh Amphibolite and paragneissSwitzerland arch girders

UNDERGROUND OPENINGSACKINGEN Black Forest Plant 1962/64 620 161 400 S.C. 30 cm + mesh + bolts = Solid paragneiss with marked (R.F.A.) 3 to 4 m, 100 to 200 kN schistosity

Veytaux - Leman Pumping 1967 645 100 S.C. 15 cm + mesh + 4 m resin Cracked limestone shale, Lake near Montreux plant bolts 160 kN + ties 1400 kN permeable and subhorizontal

EFRINGEN KIRCHEN Warehouse 1967 16 1 080 70 S.C. 15 to 20 cm + steel ribs Limestone, clay pocket(R.F.A.) 40 350 + 3m bolts,

73 1 400 meshed at 1,4 m2 post-tensioning, 122 150 wet mix

WALDECK II (R.F.A.) Hydroelectric 1971/72 1 540 250 Resin bolts l = 4 to 6 m, 12 t., Clayey shale and graywacke plant meshed at 1 to 3 m sandstone with faults

+ ties l = 20 to 25 m, 1700 kN + 5 cm S.C. and many meshes

Hermillon - d'Echaillon - Hydroelectric 1974 200 to 450 50 300 S.C. 8 to 10 cm + mesh + bolts, GneissArc Valley plant l = 2,50 m – Ø 25, meshed at 4 m2

SEVRES-ACHERES Lot 6 Sanitation 1987 160 61 47 Wet S.C. (10 cm + 10 cm + 20 cm Crown: plastic clays, (France) + mesh + TH girders + Walls:

passive anchor Meudon marl, Slab: chalk

CHAUDERON Train Train 1993 73 à 200 146 20 S.C. (30 to 40 cm) + mesh Molassestation (Switzerland) station + lattice girders.

Crown liner: on membrane, dry + wet mix S.C. 30 cm

HECKARZIMMER Warehouse 300 150 S.C. 20 cm + 2 meshes + bolts Muschelkalk, dolomite, Heidelberg (R.F.A.) l = 25 m Ø 26, l = 5 m, gypsum,

h = 10 m meshes at 1/5 m2 sulphates, anhydrite clay

LORCH (R.F.A.) Warehouse 26,5 to 65,4 S.C. 15 to 20 cm Rhenan shale rock

MECHERNICH Warehouse

WEHR (R.F.A.) Idem Sackingen

SINGKARAK Hydroelectric plant Limestone, freestone (tuff)

DUL HASTI “ 1991-93

ERTAN “ 1991-96 Diorite, gabbro, granodiorite

XIAOLANGDI “ 1995-98 Sandstone and marl

Structure Usage Year of Section Lenght Cover Suppor system Geologyconstruction (m2) (m) (m)

33


6 m: (a) for RQD > 75%, a thickness e = 2 in.of sprayed concrete applied at the crown issufficient; (b) for RQD between 50% and75%, a 3 in. sprayed concrete thickness isrequired; (c) for RQD between 25% and50%, 3 to 4 in. is required at the crown andapproximately 3 in. on the walls, and therest of the opening; (d) for lower rock cha-racteristics, an extra 1 in. is required on topof the previous case, except for unstablegrounds (Type 3).

(5) These rules can apply to larger tunnels ifthe thicknesses are modified accordingly,proportionally to the power of the diame-ter’s ratio (order of 1.25 to 1.5).

(6) In bad ground conditions requiring Type3 sprayed concrete, it is possible to applythe empirical recommendations of the pre-vious case (Type 2) if the ground is reinfor-ced with bolts; however, the additional rein-forcement brought by the wire mesh is notproven in this case, and it is advisable toincrease, when necessary, the thickness ofthe sprayed concrete.

(7) The third configuration (Type 3) corres-ponds to either situation where the loadsbrought by the excavation exceed the capa-city of the ground or to swelling grounds. Inthese conditions, the sprayed concrete mustbe designed as a ring of reinforced concreteusing the limit state design rules; this situa-tion corresponds to a thickness of sprayedconcrete over 6 in. A first estimate of thesupport thickness is given by the followingrelationship:

e = 2(pR / ƒ'c) + 2 to 4 in.

where R is the radius of the tunnel, ƒ'c is thecompressive strength of the concrete and pthe pressure induced on the structure by thesurrounding ground; this last parametermust be evaluated separately; the extra 2 to4 in. are for additional safety, mainly tocover for the potential placement problemsof sprayed concrete. This formula includes aglobal safety factor of 2, adapted for sup-port design; the value of this global safetyfactor must be increased to 2.5 to 3 for thedesign of the final liner.

Morris, J.W., " Bureau of ReclamationShotcrete Design Practices ", pp. 153-159.

This paper mainly insists on the economicaland practical aspects related to the use ofsprayed concrete:

(1) Sprayed concrete can be an economicalsupport method for some tunnels

excavated with drill and blast, assuming itcan also serve as a final liner;

(2) Small amounts of sprayed concrete(thicknesses of about 1.5 in.) can also beused in combination with steel ribs to gua-rantee stability in soft sensitive rocks; froman economical point of view however, itcould be more interesting to increase thesprayed concrete thickness instead ofadding arch girders;

(3) The use of sprayed concrete has a limitedcompatibility with the contractual contextwhere the designer and the contractor arehired independently.

Barret, S.L.V., Mc Creath, D.R., 1995, " ShotcreteSupport Design in Blocky Ground : Towards aDeterministic Approach ", Tunnels andDeep Space, Tunnelling and UndergroundSpace Technology, Vol. 10, n° 1, pp. 79-89.

(1) This paper deals with the design ofsprayed concrete used in combination withrock bolts for the support of undergroundopenings in fractured rock. Rock boltsshould be designed to hold blocks aroundthe opening; the size of the opening andthe level of fracture in the rock itself deter-mine the size of those blocks. Even thoughsprayed concrete is only present to comple-ment the rock bolt action, it is still essentialto the stability of the opening.

(2) The paper also reviews some of the prin-cipals behind empirical design methods forrock support by sprayed concrete (Albert,1965; Kobler, 1966; Cecil, 1970; Heuer,1974) and proposes an analytical designapproach.

(3) This design approach considers four fai-lure modes: loss of adhesion between thesprayed concrete and the rock surface,shearing of the concrete liner by a fallingblock, flexural failure and punching shear ofthe concrete liners by the rock bolt head.The last two failure modes are possible onlyif there is a loss of adhesion between theconcrete and the rock surface. Analyticalrelationships are proposed for each failuremode.

"Gebauer, B., Lukas, W., Kusterle, W., 1991,Monocoque Shotcrete Lining", WorldTunnelling, October, pp. 357-360."

This article describes a support/liningmethod for tunnels based on sequentialsprayed concrete layer applications:

(1) The support system is made of a combi-nation of bolts, steel ribs and wire meshassociated with a layer of sprayed concrete.

(2) The other layers of sprayed concrete areapplied by successive thin layers once thedeformation rate of the temporary support

is stable. The tunnel liner is therefore madeof a homogeneous shell, including the tem-porary support sprayed concrete layer andthe later applied sprayed concrete layers.

(3) It is generally accepted that the use ofsetting accelerating admixtures for the sup-port concrete modifies the material andmakes it more permeable in the long term.The layers sprayed after stabilization of theopening must, however, be watertight.Applying the sprayed concrete in thin layerscan reduce cracking of the liner; the use ofmicro silica will allow for non acceleratedmixtures.

(4) The use of a temporary support made ofsprayed concrete allows for a redistributionof stresses generated during excavation inthe surrounding ground. The constructionphases can be optimized by monitoring thedeformations of the structure during thework process.

(5) This process was tested on two sites inGermany: one 3200 m long sanitation tun-nel in Frasdorf (Bavière), excavated at 100 mdeep and under 80 m of water and a sub-way tunnel in Munich excavated 5 m indepth in granular ground under the watertable. In the second case, the temporarysprayed concrete support is 15 cm thick andthe two other layers placed to guaranteethe long-term stability of the opening are 10cm thick. The use of this method generatedimportant savings to the project (especiallywith the thickness of the sprayed concretelimited to 35 cm instead of 50 cm forconventional approaches).

On the rheology of sprayed concrete (byJ. Piraud, ANTEA)

"Schubert, P. (1988) - Beitrag zum rheologi-schen Verhalten von Spritzbeton. Felsbau,vol. 6, n° 3, pp. 150-153"

If one wants to evaluate the stresses in asprayed concrete shell based on the full his-tory of the measured deformations, one willneed a constitutive law, function of time, forthis material. The author used the work,now old, of England and Illston (1965), whoproposed a numerical solution based on atime step analysis. In this approach, thedeformation is composed of 4 contribu-tions:

• An instantaneous elastic deformation,

• A delayed elastic deformation,

• A permanent creep deformation, stronglytime dependent,

• A thermal contribution.

34


This constitutive law was accurately calibra-ted against creep and relaxation testsconducted in the laboratory of the "Écoledes Mines de Loeben" (Austria). Notehowever that this constitutive law was adjus-ted for dry sprayed concrete; we know thatfor wet sprayed concrete, the creep valuesare only 25 to 30% of those obtained fromthe dry process, because of the larger pro-portion of large aggregate in the wet pro-cess concrete. The author then explains apractical application of stress calculations ina tunnel liner in the Langen tunnel in theAlberg. Numerous deformation measure-ments were made in the first days in thesprayed concrete shell, allowing for the tan-gent stresses to be estimated. This exampleis developed in Chapter 3.4.1 of this docu-ment (Figures 3.4 and 3.5).

Spikes associated with the advance of thetunnel front are easily visible on thosefigures, as well as the stress stabilization –tangent and radial – around a value of 5MPa after 20 days. (cf. Pöttler, 1990, furtherin the text).

Lattice girders (by C. Bascoulergue, CAM-PENONBERNARD)

"Dr. Betzel, 1988, Analyse statique et appli-cation de cadres réticulés utilisés en chan-tiers de tunnels (Static analysis and applica-tion of lattice girders in tunnel construction),Tunnels et Ouvrages Souterrains, n° 86, pp.93-104."

The use of lattice girders in undergroundsupports has recently developed due to theincrease use of sprayed concrete in tunnelconstruction. Indeed, when ribs are requi-red, it appears more interesting to use, ins-tead of a beam shaped section, a latticestructure for which an intimate interconnec-tion with sprayed concrete is achieved, allo-wing for a real reinforced concrete compo-site.

This will lead, for an equal support capacity,to a reduced sprayed concrete thicknessand consequently a reduction of overallcosts.

This article is divided into four sections:

• General specifications for ribs used inconjunction with sprayed concrete for tun-nel support;

• Applications;

• Justification for immediate support;

• Justification of the sprayed concrete shellby considering the association of lattice gir-ders – sprayed concrete.

General specifications for lattice girdersused with sprayed concrete for support

Lattice girders are well suited to tunnelconstruction through the NATM (NewAustrian Tunnelling Method) since they canbe totally encased in sprayed concreteduring construction. They allow for an easyaccess to fill the over-breaks. They can alsobe included as reinforcement in the evalua-tion of the support strength. The quality ofthe bond between the concrete and thesteel is a function of the characteristics ofthe concrete, of the geometrical characte-ristics of the girder and of the sprayingdirection.

Applications

Lattice girders can be used for immediatesupport. Their isotropic stiffness offers agood resistance to buckling failure, either inthe frame’s direction or perpendicular to it.

The good bond between the lattice girdersand the sprayed concrete makes for a sys-tem where both materials share the loads,which offers an additional support duringthe hardening of the sprayed concrete.

Lattice girders allow a longitudinal conti-nuity in the sprayed concrete shell as oppo-sed to standard beam shaped sections thatcreate a discontinuity at each pass. This lon-gitudinal continuity increases the stabilityand strength of the shell during excavationin poor quality ground.

Justification of immediate support

Immediate support at the excavation face isextremely important. An arch girder can besolicited as soon as an important load deve-lops at the crown even if it is not encased insprayed concrete. In this section, Dr Betzeloffers a verification example under suchloads. His verification takes into accountsecond order effects, conforming to DIN4114.

Justification of the sprayed concrete shell byconsidering the association of lattice girders– sprayed concrete

Once the lattice girder is encased in sprayedconcrete and that it has set, the various barsof the web act as reinforcing bars for theconcrete.

First, the author evaluates the strength ofplain sprayed concrete section and,secondly, evaluates the increase in capacityof the sprayed concrete brought by the lat-tice girders. The calculations are function ofthe strength of the concrete at differentages. The strength of the reinforcedsprayed concrete shell is evaluated accor-

ding to the DIN 1045 standard and the spe-cial addendum 220.

The strength of the sprayed concrete shellreinforced with lattice girders is obtained byadding the admissible loads of the sprayedconcrete and in steel. The capacity of thereinforced concrete obtained using a para-bolic diagram method with the followingdeformation values:

εb = 2 0‰

εs = -2 0‰

Different curves show the increased capa-city from the lattice girders. This improve-ment can reach 50% in some cases.

Design of the sprayed concrete shell (by J.Launay, DUMEZ-GTM)

Rabcewicz, L. v. Principles of dimensioningthe supporting system for the ‘New AustrianTunnelling Method’, Water Power, June1969, July 1969, March 1973

The three papers deal with a tunnel liner fai-lure and present a method to evaluate therequired support.

The first observation of M. Rabcewicz is thatfailure occurred by shearing, and not bybending. This confirms Sattler’s theory andthe tests conducted by the author.

The second observation concerns theground confinement brought by thesprayed concrete which allowed retainingthe high shear capacity of the ground (E, cand ϕ) and also mobilize the strength of theground created by this confinement pres-sure. This pressure is calculated using theshear strength of the sprayed concrete, thebolts and the ribs.

The third and most important observation isthat the liner’s strength is mainly related tothe ground’s capacity itself reinforced withbolts, and not the sprayed concrete.Sprayed concrete is a means to mobilize thiscapacity, not the principal player in the sup-port of the opening per say. According tothe author, one should not refer to sprayedconcrete as a support method.

Finally, in the case of highly stressedgrounds, it is essential to close the supportshell in order to obtain a self-supportingstructure. The stabilization of movementsmust be monitored through instrumentationof the tunnel. Validation of calculations canonly be done by verifying the stabilization ofthe movements. This aspect is essentialaccording to Rabcewicz.

Effect of creep in young sprayed concreteused for tunnel confinement (by J. Piraud,ANTEA)

35


Pöttler, R. (1990) - Konsequenzen für dieTunnelstatik aufgrund des nichtlinearenMaterialverhaltens von jungem Spritzbeton.Felsbau, vol 8, N° 3, pp. 121-128.

Pöttler, R. (1990) - Time-dependent rock-shotcrete interaction: a numerical short-cut.Computers and Geotechnics, vol. 9, N° 3,pp. 149-169.

A realistic evaluation of a tunnel’s supportby sprayed concrete should take intoaccount the evolution over time of theapplied loads (consequent to the advance-ment of the excavation face) and that of thesprayed concrete properties (stiffening andcreep). After analyzing the problem in a 3Dfinite element model with a complex timedependent constitutive law, the author sho-wed that it can be replaced with an explicit2D model with a linear elastic concretebehavior:

In the short term, the maximum stress foundin sprayed concrete at approximately onediameter behind the front can be evaluatedwith good precision with a fictitious equiva-lent modulus Ei = 7 000 MPa;

After a few days, creep in concrete leads toa relaxation of the tangent stress in the sup-port, which stabilizes after 2 weeks around avalue close to σLT = 4 MPa.

The parametric study conducted by Pöttlershows that, these two values are acceptablefor any depth, the diameter, the concretethickness and the rock modulus. Thisconclusion – which corroborates experimen-tal observations concerning the universalityof sprayed concrete support – allows torapidly estimate the stresses in a temporarysupport by using the characteristic curvemethod, and to verify them by measuringthe deformation in the concrete shell.

Stresses in sprayed concrete in tunnels (byJ. Piraud, ANTEA)

Golser, J., Rabensteiner, K., Sigl, O.,Aladrian, W. (1990) - Kontrolle derSpritzbetonbeanspruchung im Tunnelbau –Berg- und hüttenmännische Monatshefte,Leoben, Vol. 135, n° 10, pp. 376-383.

This paper presents the results of a researchconducted at the Mining School of Leoben(Austria) which included several laboratoryexperiments and in situ measurements, andmodeling efforts aimed at determining aconstitutive law for the behavior of sprayedconcrete. Particularly, a time step methodfor the behavior of sprayed concrete wassuccessfully calibrated against experimentalresults (cf. Schubert, 1988, above in thetext). The objective was to evaluate the

extent of a sprayed concrete shell unloadingby using deformation values taken in situ.

On this point, Golser et al. (1990) considerthat creep of young sprayed concrete is atthe same time an essential property of thismaterial (capacity to undergo large defor-mations without failure) and an essentialcharacteristic to be determined.

Critics on direct stress measurements usinghydraulic cells

Inherent defects of this monitoring tech-nique were brought to light both throughnumerical simulation and laboratory experi-ments (cell encased in concrete placedunder a load frame). These studies showedthat such cells always yield short term pres-sure values that are too low and long-termvalues that are too high, without any indica-tion of when the measured value is exact.Moreover, they do not return to zero afterunloading. Their reliability is however betterat high stress levels, especially if they are ini-tially “pre-laoded” (up to 400 kPa) toimprove contact.

Measuring deformations in a sprayedconcrete shell

Golser et al. (1990) recommend, for futureprojects, to determine the stresses in thesprayed concrete using deformation measu-rements, much more accurate. This assumesthat the placement of deformation monito-ring devices are on both sides of the shell;using a few hypothesis (no traction, limit tocompressive stress, etc.). One can then cal-culate the normal load and the bendingmoment in the sprayed concrete shell, andcompare that to the allowable values forunreinforced concrete according the DIN1045 standard.

A real calculation example of load N andbending moment M from deformation mea-surements taken from sensors located onboth sides of the shell in a tunnel showedthat even after reaching very high valuesafter a few days (near 20 MPa), the stressesgo down to approximately 1/3 of the maxi-mum value. This corroborates Rabcewicz’sold principle stating that the bendingmoments could be neglected for the designof sprayed concrete supports, eitherbecause they decrease over time or lead tocracking and subsequent formation ofhinges.

Sprayed concrete final liner on a test sec-tion in the Munich subway (by J. Piraud)

Honnefelder, N., Theimer, G.U. (1992)–Einschalige Spritzbetonbauweise im

Münchner U-Bahn-Bau. Der Bauingenieur,n° 67, pp. 393-399.

In a tunnel constructed using a conventionalmethod, with temporary supports made ofsprayed concrete possibly covered with acast in place liner, we consider the cast inplace concrete to take all of the loads at itsfinal state. However, long term monitoringshowed that the sprayed concrete layersretained some long-term support capacitywhich represented a safety factor not consi-dered in the design.

Location and geology

It is this economical aspect that led theDYWIDAG Company to test sprayedconcrete “single shell” final liner first in1989 in the Frasdorf outlet (L = 3,2 km), andthen on a test portion of the Munich subway(lot "6 West 5").

This test portion has a variable section of 38to 52 m2 and is located 8 m underground.The tunnel is at an intersection betweencompact ternary marls (bottom), sand andgravel (top). The water table is located atmid-height but can exceptionally reach thecrown; during construction, surface pumpswere used to lower the water table.

Construction method

The following method was used:

• Staged excavation excavation, bench andheading,

• Placement of a 5 cm layer of an accelera-ted dry process sprayed concrete to imme-diately seal the surface,

• High quality micro silica sprayed concrete(no accelerators) is sprayed to reach a totalthickness of 15 cm,

• For planning reasons, a thick slab is cast inplace,

• 6 months after the excavation, the sprayedconcrete surface is water jetted (600 kPa), inorder to guarantee an optimal bond withthe next layer of concrete,

• Spraying of two layers of micro silicaconcrete, each 10 cm, 14 days apart; a totalthickness of 35 cm of “active” sprayedconcrete is thus obtained (used for calcula-tions as well).

Interior sprayed concrete characteristics

Composition of 1 m3 of fresh sprayedconcrete (total mass of 2 370 kg/m3):

• 190 l of water,

• 60 kg of micro silica slurry (including 50%water)

36


• 380 kg of 45F Portland cement,

• 1740 kg of aggregates, of which:

o 43% are 0-2 mm

o 23% are 2-4 mm

o 34% are 4-8 mm

The dry mixture is plant produced and ovendried. It is pre-wet in the tunnel and trans-ported through thin stream method. Themicro silica, which replaces the accelerators,is precisely added to the mixing water thatfeed the manually operated nozzle. Thismicro silica is characterised by its extremefineness (specific surface of 200 000 cm2/g);it has the advantage of cutting by half theamount of rebound while increasing thecompaction and the overall strength of theconcrete.

Particular care is provided in the presence ofcold joints, for setting of the layer thick-nesses, for securing the mesh reinforcingthe second layer and for the curing of theconcrete (regular spraying of each layer with200C water for 7 days in order to reducecracking).

QA/QC testing program

A rigorous quality control program was setforth for the materials, equipment, proce-dures and concrete thickness applied. The

quality of the surface was an item that recei-ved particular attention: 50 pull out testswere conducted, yielding an average adhe-sion of 1.87 MPa (following the STV-SIB 878procedure). The 28 days compressivestrength is clearly above 50 MPa, both onsamples taken from test panels and onsamples cored in situ.

Water penetration depth measured accor-ding to the DIN 1045 standard did not gobeyond 24 mm. Concrete temperatureincreased from 10 to 180C after setting.

Adhesion tests of the 2 interfaces of the 3sprayed concrete layers were conductedboth on samples taken from test panels andon samples cored in situ; they yieldedresults clearly above the minimum requiredvalue of 0.6 MPa. Indeed, the cold jointswere difficult to identify on the samples,which clearly shows that a homogeneousshell was obtained even though sprayingwas done in 3 passes. Finally, shrinkagemeasurements have shown the positiveeffect of warm water.

We also noticed that the tunnel was water-tight, almost no cracks were observed.Spraying on test panels containing reinfor-cing bars showed, after they were saw cut,an excellent encasement of the bars andjoints, without shadowing effects.

Conditions for success and conclusions

A homogeneous sprayed concrete shell canonly be obtained through excellent know-ledge of the technology and a strong qualitycontrol program. The key elements respon-sible for the success of the projects are:• Meticulous cleaning of all surfaces to spray(in order to guarantee adhesion),• Prewetting of the surfaces prior to theapplication of the sprayed concrete (inorder to limit concrete’s water to be absor-bed by capillarity),• Careful watering of the surfaces after set-ting (curing),• Reinforcement was tight and rigid, whileavoiding high steel concentration, • Vigilant control of reinforcement positionand sprayed concrete thickness.

This single shell liner ended up 10 to 15 %less expensive than the usual solution, whichplanned for 15 cm of immediate "inactive"sprayed concrete + 35 cm of cast in placeconcrete. The economy comes essentiallyfrom the reduction in the volume excavatedand the volume of concrete, as well as thesuppression of the forms. This solution isespecially indicated in the case of short tun-nels, with variable sections or short by-passes, particularly for tunnels above thewater table.

DESIGN OF SPRAYED CONCRETE FOR UNDERGROUND SUPPORT · Recommendations for the design of sprayed...

Documents

Transcript of DESIGN OF SPRAYED CONCRETE FOR UNDERGROUND SUPPORT · Recommendations for the design of sprayed...