REVUE FRANCAISE DE GEOTECHNIOUE NO 23 PARIS 1983

General report topic 6 / Rapport général thème 6

LOADING, UNLOADING AND SHEARING TESTS CARRIED OUT IN UNDERGROUND WORKSINCLUDING STRESS MEASUREMENTS

ESSAIS DE CHARGEMENT, DE DECHARGEMENT ET DE CISAILLEMENT EN TRAVAUXSOI]TERRAINS Y COMPRIS LES MESURES DES CONTRAINTES

SHARP John C.,*

1. Introduction

The title of this topic includes a number of important ele-ments that can be classed as follows:

- The in situ nature of the tests implies that they areessentially direct measurements of rock moss response**.

- The scale and number of the tests must bê directlyrelated to the nature of the rock mass such that a repre-sentative sample is involved.

- The loading characteristics including both force andstiffness considerations must be directly related to theproposed engineering structure.

- The essence of the tests is to provide direct observa-tions of rock response that can be used as design paru'meters.

The type of tests that are commonly employed include thefollowing:

- Load-displacementmass deforrnability,

shear- displacerne ntmass shear strength,

stress measurementsstress in the rock,

- thermal heating testsresponse of the rock.

In addition to individual parameter assessment, tests are

often carried out on prototype structures such as under-ground openings and pressure tunnels where rock mass

conditions in relation to the applied loadings might be critical or where complex interactive loading mechanismsare involved.

For the planning and assessment of in situ testing pro-grammes, three criteria need to be carefully considered as

follows:

- Test scale.

- Stress state of the test sample.

- Stiffness or confinement of the test sample.

Before proceeding into specific test areas, it is perhapsuseful to reflect on the value of the testing process inrelation to the final design product. Throughout the de-velopment of soil and rock mechanics there have beentwo poles of thought on testing as follows:

- C"^rlt*t in Rock E,ngineering, San Petet, Jetsey, ChannelIslands (U.f.1

x x The term rock mass is used in this report to emphasise theassemblage of discrete material pieces. Many of the remarksare equally applicable to soils.

measurements to determine rock

measurements to determine rock

to determine the initial state ofF

to determine the thermomechanical

- undertake a larye numbei; of relatively quick, cheaptests to provide an overall coverage of the site

- perform THE UNIQUE large scale in situ test.

Those that live exclusively at the poles do so at their peril.Those that combine both approaches usually find success.

In planning in situ test progranrmes it is therefore essentialto have a hierarchy of tests that provide a complete spectrumof answers to cover all aspects of the design process. It iswell known that large scale in situ tests provide extremelyvaluable DIRECT observations of rock mass behaviourwhilst other smaller scale tests (typically performed inboreholes) and geophysical tests generally provide INDI-RECT values that must be calibrated by overall CONTROLmeasurements from the DIRECT observations.

In the evolution of rock mechanics great pains were takento try and relate laboratory values to in situ properties. Inrock masses that not only contain material but also gaps

fioints) with different or no material, relationships couldoften not be found. Laboratories were than deserted andpeople left for the field. Unfortunately it is my belief thatin many cases people took their laboratory training into thefield and continued in their former ways by trying toexecute small scale borehole tests and again relate this tothe rock mass in an attempt to produce design properties.In some ways they were simply testing outwards from theircore sample instead of inwards.

Whilst all this activity was proceeding a group of peoplecalled Engineers (who were somewhat sceptical about theborehole people and their gadgets or 'patentes') continuedto build projects including dams on difficfrlt foundationsand large underground structures. Fortunately for us todaymeasurements were made on the overall response of therock mass in relation to the construction activity that haveprovided us with an empirical basis for design.

It is from this empirical understanding of rock mass beha-viour that rock mass test programmes should be planned.Every site has its specific features and every engineeringstructure imposes different loading conditions. Thustesting will always be required if meaningful rock mass pro-perties for design are to be obtained. The scope of the test-ing will be dictated by both the rock mass conditions at thesite and an estimate of the likely effects of the rock beha-viour on the final engineering product. At this point it isworth mentioning that we are experiencing an ever increas-ing participation of the rock in the final engineering productand this can be well demonstrated for underground structuresin the field of rock reinforced cavern arches (as a replace-ment for thick concrete arches) and in pressure tunnels

totlt /ê'1ke

78

PRESSI.RE SHAFTS ,TUVNELS AND PENSTOCKSHFADU'ICRKS

tpær, alcorpment rockt

Lorræctpmml rocks

| . foull

POWERSTATIONcorvtPLEx

ELEVATÊTI(m;:8col

I

rsoc iII

t4oo iII

t?oo iI

rcoo jti't

TAILWORKS

) O.5 kml=ir{==l!J_l

Scole

Ëf ,ooo

['.oo

[,*o

8El3N o=rgn

OVERALL ALIGNMENT

MACHINE TRA}JSFCRMERHALL HALL

MAiIJ DRAINAGETUNilgL

POWER STATION COMPLEX

where reliance on the rock for long term support purposeshas reduced the need for expensive, thick steel linings.

A fairly comprehensive example of underground testingcan be found at the Drakensberg Project (Mellors andSharp 1982) the overall layout for which is shown onFigure 1 and the exploratory works on Figure 2.

The scheme- proposals involved two major features forwhich little precedence was available as follows:

Excavation and permanent stabilisation of an under-ground power station complex in relatively weak argillaceousrocks that were subject to deterioration on exposure.

Use of concrete-lined pressure tunnels in the relativelydeformable argillaceous rocks with heads in excess of500 m. (Sharp & Gonano 1982).

Because of the nature of the potential engineering problems,an overall testing strategy that incluâed tht foilowingelements was developed:

Full scale prototype structures.Large scale in situ design tests at selected sample loca-

tions.Small scale in siru classification tests over a widespread

area.

alignment.

A breakdown of the principal tests in relation to the mainscheme elements is given in Table l.In reviewing the papers submitted to the Symposium anattempt has been made to evaluate the purpose and scopeof the test measurements and their justification or other-wise in relation to our current understanding of rock mass

VALVEHAII o

@@

E@o@@@@@@@

r

T

DETAILEDN,/ESTIGANON /OESIGN

GENERALIÎWESTIGATION /DESIGN

HEADRACE INTAKE

TWIN HEADRACE TUNNELS

TWIN SURGE SHAFTS

TWIN PRESSURE SHAFTSTWIN PRESSURE TUNNELS / PENSTOCKS

VALVE HALL

MACHT€ HA,11

TRANSFORMER HALL

TWIN SIJRGE CT{AIVIBERS

AI.'XILARY ACCESS TUrurugÈ

EXPLORATORY ADIT

MAN ACCESS TUNNEL

SINGI_E TAII.JÙTC€ TUNNEL

TAILRACE TU.ÏNEL PORTAL

Fig. 1: Drakensberg pumped storage scheme - Longitudinal section

behaviour as well as the engineering structure being inves-tigated. The tests in Topic 6 should fall into the followingcategories and the papers have been grouped accordingly:

I In situ Design Property TestsRock Mass Deformability

- Shear Strength

- Thermal Properties

I[ In Situ Stress Measurements

III Full Scale In Situ Observations

The distribution of papers received in relation to the abovegroups is indicated on Table 2. It is particularly encouragingto see a significant number of papels dealing witn fuil séaléobservations. Most of the papers dealt either with damfoundations or pressure tunnel design and have beengrouped accordingly.

Tests at a smaller scale in boreholes which can usually onlybe used indirectly to determine rock mass property valueshave already been addressed in Topics I and 5. The asso-ciated and important range of geophysical tests has beencovered in Topic 2.

In terms of test methods, the International Society forRock Mechanics has published a number of standardscovering most test types and therefore the subject of basictest methodology wiil-not be addressed here.

2. In Situ design property tests

overail, twelve contributions were received all of whichwere addressed wholly or in part to the measurement of

79

ELEVATTON ( m )

r240

t2 30

,22U^

12io

i?co

ti90

il80

f l70

t!60

tf5 0

ELEVATfON ( m )

EXPLORATORYADIT

Ot TUll€ OF l,lACHlNE HALL t240

t2 30

t22o

t2 to

t200

il90If80

I t70

I t60

| 150

TTT

MACHINE HALLEXPLORATORY

HEADING

MACHINE HALL TESTENLARGEMENT

PLATEEEARING

TEST ADIT

rock mass deformability. Only one paper contained someinformation on shear strength measurement which may bean indication of the difficulties of representative determi-nation of this parameter in situ as well as a recognition ofthe advances made in the empirical determination of shearstrength parameters (Barton 1977). No papers on thermalproperty determination were forthcoming.

2.1. Rock mass deformability

The most cornmonly accepted large scale deformabilitymeasurement technique is the plate loading test and loadsup to 1000 tons on aî area of the order of 1 m2 are nowaccepted practice. Between 1957 and 1982, some twenty-five progranrmes of rock mass deformability measurementusing the plate loading test method have been carried out.Some of the published results are given in Table 3.

A second less popular method is the flat jack methodthat was developed largely by the late Dr. Rocha and hiscolleagues at LNEC in Portugal. Many sites have also beeninvestigated using this method.

Deformability M easurement s : Dam Foundations

Okamoto and his colleagues from Japn have reported onboth plate bearing and flat jack tests carried out at fourdifferent dam sites located in various rock types, princi-

ELEVATION

MACHINE HALLEXPLORATORY

HEADING

LEGE N D

PLATE SEARING TEST

r*l $TU STRESS PIEASUREiiENTS

DETAILEO CORÉ LOGGII'TG

BOREHOLE MODI,LJS TESTS

#Scote

pally granite, diorite and tuff. Valuable results are prcvided from their measurements. The flat jack type usedis semicircular and has limited penetration into the rockmass. The effective loaded area is approximately 0,25 m2 .

In comparing plate bearing and flat jack tests, the latterare preferred by the authors as they believe that they areless affected by blasting damage. Details of the site prepa-ration for the plate bearing tests are not given but clearlyit is necessary to produce a representative surface bycontrolled blasting and if necessary manual excavationas they in fact did for their flat jack tests. Whilst theydevelop an expected correlation between deformationand elastic moduli, their correlations between modulusand rock type I weathering class show a significant degreeof scatter indicating the influence of variability on theresults particularly at the test scale selected.

Additional general information on dam foundation testsin Japan is given by B aba. The need for overall geologicalclassification in relation to the selection of detailed testsites is emphasised but few test details or results areincluded.

Price Jones and Hobbs describe a series of 800 ton capacityplate load tests carried out in adits for the assessment ofabutment conditions for the arch dam at the VictoriaProject in Sri Lanka situated in high grade Precambrianmetamorphic rocks. Useful practical details are given

nvz

s o@

lTr-

oPENSTOCK

TESTCHAMBER

PLAN

Fig. 2 : Drakensberg pumped storage scheme - Layout of exploratory works and tn situ test locations.

BO

Tabl. 1: Example of in situ test programmes for undergroundcavern and pressure tunnel design (drakensberg

pumped storage scheme)

Power station complex cavern design

Full scale observations in situ

Machine Hall Test Enlargement

- Displacement measurements (MPBX, Convergence, Levelling,Rock Bolts)

- Load measurements (Hydraulic support props, Rock bolt loadcells)

- Stress change measurements (Embedded stress cells)

- Piezometric measurements (Hydraulic type piezometers)- Rock determination observations (High resolution MPBX

system direct observation)

In situ design property tests

Rock Mass D eformo bility

- Plate bearing tests (500 ton; horizontal and vertical on roofstrata)

- Borehole modulus tests (Goodman jack tests [limited value])- Geophysical measurements

Shear Strength

- Geometrical characterisation of significant planes of weakness(with laboratory testing)

In sttu stress measurements

- Triaxial strain cell overcore tests from exploratory headings

High Head concrete lined pressure tunnel design

Full scale observations in situ

Test Chamber

- Displacement, stress and water pressure monitoring in rock- Displacement, stress and strain measurements in lining

In situ design property tests

- Plate bearing tests (5 00 ton around test chamber periphery[H, V, 45" l) Permeability testing at ambient and elevatedpressures..r

trn siru stress measurements

- Triaxial strain cell overcore tests from pilot tunnels- Stress measurements in chamber lining- Hydrofracture stress measurements from pilot tunnels

highlighting the need for careful test site preparation. Theloading affaîgement was conventional except that loadswere measured in each loading column in addition to theloading plate flat jacks. No reasons are given as to the neces-sity of the extra load cells. Although the re$lts were highlysatisfactory from the dam designer's viewpoint giving insitu deformation moduli (close to elastic) in excess of35GPa, the small displacement, as measured by the exten-someters, required particular care in their evaluation. Inspite of a high resolution measurement system, displace-ments measured more than 2,5 m below the loading surfacecould not be interpreted.

ln comparing the results from Victoria with those reportedfrom Japan, a lesson on the need for thorough site prepa-

Tab. 2: Distribution of papers: Topic 6

Number of papers

Dam Undergroundfoundations works

l. In situ design property tests

- Rock Mass Deformability

o Flat jack testso Plate bearing testso Radial jacking testsr Pressure chamber testso Seismic methods

- Shear Strength

- Thermal Properties

lL In situ stress measurements

- Flat jack method

- Borehole overcore methods

- Hydraulic fracturing andassociated methods

III. Full scale observations in situ

- Underground Storage

- Tunnels

1

2

000

1

0

2

01

3

1

0

0

ration can be learned. It is however important to create arepresentative surface either of the rock mass or of the sur-face that may form the actual dam abutment when exca-vation methodology may require to be taken into account.

Fernandez Bollo and colleagues from Spain have reportedon in situ flat jack tests using one or two (orthogonal)jacks typically 2 mz in area. The paper provides a criticalreview of the measurement method and then gives outlinedetails of time dependant tests for a damsite study.

D eformability Measurement s : Pressure Tunnels

Within Topic 6, this class of measurements has receivedconsiderable attention with nearly all data being derivedfrom chamber tests designed to act as a prototype for theactual strucfure. The use of large scale tests in this manneroften coupled with seismic investigations to extend thetest data to other geological conditions shows a promisingtrend in investigation methods. The use of such simulatedtests recognises that the problem of rock support to pres-sure tunnel / penstock linings involves a cômplex inter-action of geological, construction and stress variables thatcannot necessarily be tested individually and the synthe-sised into a design model.

In the planning of chamber tests it is necessary to use anexcavation technique that will be representative of the maintunnel excavation method and carry out tests in a represen-tative geological and stress environment. Because of thecomplexity of the tests, it is normally only possible tocaffy out one or two tests at a given site and thus theresults must be applied to the overall tunnel alignment bymeans of more general geological observations often em-ploying ge ophysical methods.

Chamber tests that utilise water as their loading mediumprovide the best simulation of the pressure tunnel loading /

Tab. 3lt: Selected plate bearing test data

TEST APPARATUS TEST METHODPROJECT

t ocertoNDATE

-

ROCK

TYPENO. OF

TESTS

257

20

PLATE

orlrgNsIoNS@)--

4576

25 x 25

LOADINGMETHOD

ORIENT-AT ION

Hor i zontal

HorizonÈalVertical

Hor i zonta I

HorizontalVert ical

Hor i zonÈalVerË ical

Hor i zonËa1Vert ical

ROCK-

PLATEfficr

DEFORMATION TESTING TIME MÆ(II.,IU}'l MODT]LUS REFERENCESTRESS @)-ïf.P:-

m mo--CEDURE

(DAYS )

IGRADJ ARCHI RAI{

DEZ DAI'1

IRAI{

Dior ite

Doleri te

Muds toneSiltstoneSand s tone

Gne i ssAmphibol ite

Mortar

Sulphur HydraulicIack

Hydraulic InclinedJack

Plate andRock SurfaceDial Gauges

Plate SurfaceDial Gauges

Plate andRock SurfaceDial Gauges

Center MPBX

TDG I

Rock SurfaceDial Gauges

Rock SurfaceDiaI Gauges

Rock SurfaceDial Gauges

DA}1

19573l

Horizontal Rock and 3 Inc.Vertical Plate Surface,

Dial Gauges

4 Inc.

6 Inc.

Er = 1.8 Drozd andtô 9.8 Louma,E - 1.0 1966rE tg.l

Wa ldor fVel trâp,Curtis I 963

Dodd s ,| 966

B 1.0 Timur,12,3 lg70r 0.33'3

a

62.7 Bieniawski ,77,2 1978

18.6 Bieniawski,40.6 1978

6. 5 Bieniawski ,26.2 1978

Serafim andGuerreiro,1 966

= 1.4 Sapegin and53.5 Shiryaev,

| 966

= .08 Sapegic and1 .6 Shiryaev,

1 966= 2.9I 5.4 Sapegin and

Shi ryaev,1 966

1 Inc.

4 Inc.Cre ep

4 Inc.Creep

6 Inc.

6 Inc.

1.0t6

1.2

22 .'4

13.8

10. 9

6.9

8.0

3.9

E = 14.lr8 lB.o

E = 6.9r8 lz.4E = 10.3.È5 26.z

Euto

E.to

E=to

Conglo-I 960 meraÈe

TI'MUT 2

AUSTRALIA I 960 Chert

OROVILLE POWER- Amphi-HOUSE boliteCALIFORNIA 1 96 I

AD IGUZEL DAI'! Ma rb 1 e ,TURKEY | 963 Schist

HENDRIK VERI^IOERD DoleriteDAI"l

S. AFRICA 1965

P. K. LEROUX DAI'IS. AFRICA I 965

ORAI.IGE FISHTUNNELS. AFRICA 1965

ARCH DAI'ICZECHOSLOVAKIA

I 965

30

53

Rubber HydraulicConcrete Jack

Concrete HydraulicJack

Concrete HydraulicIack

Hydraul icConcrete Jack

Rubbe rltorcar

Rubb e rMortar

E = 1.8 Alexander,to 51.7 | 960

E = 8.3 Kruse,Èo 12.4 I 970

E=fo

E=Èo

r5

30

7l x 7l

t60

80

2l

GRAN SUARNA DAI'{ QuartziteSPAIN 1966

HYDROELECTRIC DAl.{ LimesroneussR 1966

HYDROELECTRIC DAI"I

USSR 1966 Gneiss

HYDROELECTRIC DAI'I BituminousUSSR 1966 Limestone

Neoprene Hydraulic Horizontal Rock SurfaceMortar Jack VerÈical Dial Gauges

Inc I ined

Rubber Hydraulic Vertical Rock SurfaceMortar Jack Dial Gauges

Esto

Esto

Esto

7.1480

880

Hyd raul icJack Vertical

Hydraulic VerÈicalJack

7.1

Tab. 312: Selected plate bearing test data (continued)

TEST APPARATUS TEST }IETHODPROJECTIffinDATE

CRESTMORE MINECALIFORNIA 1967

TEHACHAPI TUNNEL

CALIFORNIA 1967

AUBURN DAM

CALIFORNIA 1969

LAGO DELIO POWER-HOUSE ITALY I97O

CHURCHILL FALLSPOWERHOUSE

CANADA I 970

MICA PROJECTCAI{ADA 197 O

GORDON DAM

TAS}IANIA 197 1

ELANDSBERG SCHEI'IE

s. AFRTCA 197A-77

LG-z POI.IERHOUSE

CANADA 1976

DRAKENSBERG SCHEME

S . AFRICA 197 6

ROCKm

Marb 1 e

Diori teGnei ss

Amph i-bol i te

Gneiss

DioriteGneiss

GraniteGne is s

Schi s t

Quart z i te

Greywacke

Grani te

Mud s ÈoneSiltsÈoneSandstone

NO. OF

TEffiPLATESrTqE-usroNs@)-

ROCK LOADINGPI-ATE METHOD

CONTACT

concrete TTll""t"

Concrete Flat Jack

Mortar FlaÈ Jack

ConcreEe Hydraul icJack

I'torÈar Flat Jack

Surphur TTI;""t"

concreÈe }T:i""t"

ConcreÈe Flac Jack

ConcreÈe Hydraul iclack

MEASUREMENTSDEFOR}IATION TESTING TIME

PnoceounE GAFS )MODI.JLUSlry REFERENCE

E = 12.0 Heuze andto 16.9 Goodman,

r 968

E = 0.9 Kruse , 1970to 15.2

Wallace Slebirand Ander-son,1970

E E 7.0 EnricorPaoloto 21 .O & Armando ' 1970

E _*= 4. I Benson,tE 134 l'lurphy andE *=33.8 McCreath,rB 224 tgTo

E = 1.3 Meidal andÈo 48.8 Dodds, 1973

'E- = 3.2 Giudici, o

18 38.7 1g7gE = 7.515 43 .6

E = 27.1 Bieniawski,Èo 59.4 1978

E = 38 .0 Murphy , Lewayto 60.9 &Rancourtrl9T6

E rt= 1.6 Bowcock etal. 197618 27 .4 Golder Ass. 1g7g

ORIENT-Itto-N-

Ve rt ica 1

HorizonÈalVert ical

Incl ined

Hor i zontalVert ical

Hor i zonta 1

Vert ica 1

HorizonÈalVertical

Hori zonÈal

HorizontalVerÈical

HorizontalVerÈ icalIncl ined

Rock andPlace SurfaceDial Gauges

Center l'lPBXTDG

Center MPBX

TDG

Rock SurfaceDial Gauges

Center MPBXEdge MPBX

TDG

Plate SurfaceDial Gauges

Plate SurfaceDial Gauges

CenÈer MPBXTDG

CenÈer MPBX

Rock SurfaceDial Gauges

3 Inc.

3 Inc.

5 Inc,Creep

3 Inc.Creep

4 Inc.Creep

4 Inc.

5 Inc.Creep

MAXIMI'MSTRES S

]EFâ_

20.7

l7 .2

6.9

9.8

9.3

27 .5

7 .0

4.59.. 0

16 53

86

86

28

50

| 13

56

12

5l

r0

t2 o:t

æN)

78

t4

Plate D imens ions (cm)

100 = Circular plate diameters25 x 25 = Rectangular plate, lengÈh x width

Def orura t ion Measurement

MPBX = Multiple position borehole extensometerTDG = Tunnel diameter gauge

Testing Procedure

4 Inc. = Number of increasing pressure incremenËs and decrementsCreeP = CreeP tesÈ

Modul.us

E = Modulus of deformation (undefined)

E, = Initial loading modulus.1

E^ = Tangent ModulustE = SecanÈ Modulus (1st Pressure increment).

s

E = SecanÈ llodulus (maximum Pressure).m

E = Unloading Modulusu

E = Reloading ÈlodulusrE* = Modulus ac depth

in one cycle

83

deformation state as they are able to take into accountthe effects of water pressure within the rock mass. Thisis particularly important for concrete-lined tunnels whereleakage into the rock will inevitably occur unless specialwater-proofing provisions are provided.

Other simulation tests such as the radial jacking test do notallow for the effect of water pressure in the rock mass andthus are better directed to the determination of the poten-tial contribution of rock support to the dimensioning ofsteel linings. One major benefit of the radial jacking testsas opposed to chamber tests in the use of the test at mul-tiple locations along the completed tunnel alignment. Thisdoes however assume that testing and lining dimensioningcan be carried out after tunnel excavation which may beat a late stage in the overall construction programme. Theeffect of grouting is obviously difficult to measure withthis method.

A summary of past chamber test experience together withother modulus determination methods for pressure tunneldesign is given in Table 3.

The use of the radial jack test in China is covered in acomprehensive paper on pressure tunnel investigations byYang Zi W en. Twenty-one tests on ten different projects'are reported covering tunnel sections in granite, grano-diorite, basalt, conglomerates, limestones and sedimentaryrocks. The test techniques appear similar to those deve-loped by TIWAG of Austria (Lauffer and Seeber 1961). Inaddition to providing a considerable amount of test data,studies covering the influence of cylinder length and thedepth of influence of the jacking method are described.An interesting section of the paper deals with interpretationof rock mass anisotropy. For a tunnel section in alteredbasalt a ratio of maximum to minimum diametral deflec-tions of 3.9 was recorded (vertical to horizontal). The ani-sotropy is attributed to the rock mass but it is perhpasbetter explained by the potentially loosened zone abovethe crown giving exaggerated vertical deflections. Never-theless such results are applicable to the resulting liningdesign unless the apparent deflection anisotropy can bereduced by grouting.

The use of chamber tests in conjunction with seismicmethods to provide input to an analytical model simulatingthe various zones around a pressure tunnel are well illus-trated by Doucerain from EDF, France. His paper describestests conducted in granite at the Super Bissorte project todetermine the participation of rock support in the steelliner design. The test chamber with a 2 m steel lining dia-meter had an excavated diameter of some 3,20 ffi, â lengthof 30 m and was placed close to a fault zone some 300 mbelow surface. Extensometers were used to measure theresponse of tunnel loading into the rock which coveredan internal pressure range up to 25 MPa which is an extre-mely high loading in this type of test. The steel lining itselfreached its elastic limit at about l0 MPa internal pressureand was thus subjected to permanent deformations abovethis value. Seismic methods were used to define three dis-tinct zones around the excavation corresponding to a dis-turbed zone, a transition zone and the undisturbed rockmass. The pressurisation tests together with the numericalmodel derived in conjunction with the radial deformabilitydistribution taken frôm the seismic tests allowed the loaôtaken by the steel and rock to be determined together withdeformability values for the rock mass. Values of defor-mability between 5 0 and 100 GPa were obtained afterthe first loading. Like other workers in this field theauthors found that the values obtained from a confinedchamber test were high in relation to values predictedfrom small scale test measurements. The overall test

arrangement, measurement programmed and analyticalinterpretation have provided a good example of the use ofin siru tests to provide representative design data for highhead penstocks.

In situ tests at the Nurek Hydro Project (USSR) reportedby Rukin and colleagues have been directed towar-ds theys-e - of 'polymercement' concrete linings developed toinhibit cracking where the stiffness of the rock surroundpermits excessive deformations. The use of latex additivesin the concrete linings have reduced the overall concretemodulus two fold and allowed water tightness to be im-proved under enhanced deformation. A test tunnel com-prising sections of ordinary and polymer concrete liningswas constructed in the sandstone silsttone rock mass.Limited contact grouting was undertaken. Using seismicmeasurements, it was established that a blast damagedzone up to 1,2 m thick was present compared with an exla-vation diameter of 3,1 m which appears to indicate limitedcontrol on the excavation method. Moduli of the surround-ing rock mass were measured prior to and following grout-ing, the grouting resulting in a 50 % increase in modulusto about 18 GPa. (This value appears very representative forthe type of rock mass at the tett site). lests were conduct-ed using both an internal loading frame (radial flat jackarrangement) and pressurisation of the entire chamber.Typical tests pressures were up to 1,4 MPa. Significant ani-sotropy in diametral deflections was observed with verticaldeflections being some 2 - 3 times horizontal. This againcould be related to disturbance to the bedded strata in thefunnel crown as such values in the undisfurbed rock massdo not seem plausible. The tests demonstrated the superio-rity of the polymer concrete linings over the ordinary con-crete the latter exhibiting longitudinal cracks followingpressurisation. Leakage rates were also comparatively less-.

In relation to the internal pressures used, cracking in theordinary concrete appears to have taken place at a relati-vely low pressure. This could have possibly been preventedor suppressed by more extensive grouting measures. Thepaper again is a good example of the use of representativetest measurements to provide reliable data for tunnellining design.

Another paper from the USSR by Lavrov describes the useof 'seismoacoustic' measurements along pressure tunnelalignments together with borehole ultrasonic methods todetermine the effective deformability of the rock surround.He emphasises the important need to measure the defor-mation properties of the rock in the immediate vicinity ofthe excavation which has been disturbed by the excavationprocess and by stress redistribution. The methods discussedhave been used or] four major projects including the Inguritunnel with a length of 15 km and a diameter of 9,5 m.Some 25 km of seismic profiling and 6,4 km of boreholelogging have been executed over the past 16 years. Theuse of radial boreholes allows detailed cross sections to beestablished which can then be extrapolated along the tunnelalignment by the longifudinal profiling. Useful descriptionsare given concerning the results from a number of differentrock types and using a three zone radial classification (nearsurface fractured zone I zone of relaxation / undisturbedrock) typical average values for depth and deformabilityare given. The theory used has been checked against staticdeformability tests (presumably plate bearing or chambertests) and interesting data are provide d on correlationfactors. The method has also been extended to checking theadequacy of grouting behind the concrete lining of theInguri tunnel using a total of 800 boreholes to depths of8 m. Although the techniques described are in use in anumber of countries, the Russian work provides useful new

84

data and may encourage other engineers to adopt ordevelop this approach.

Additional summary data from Japan has been providedby Baba and quotes results from the Shintoyone schemewhere both steel lined and concrete lined chambers wereconstructed in granitic rocks.

DeformabiliQ Measurements: Underground Excavations

Unlike the case of pressure tunnels where one is trying todeterrnine the 'reaction modulus' at the surface of a rockexcavation, the design of large underground openingsrequires a knowledge of rock mass deformability.

The combined resources of ENEL and ISMES in ttaly haveover a number of years been studying a number of hydroelectric and pumped storage sites in various rock types inan attempt to develop general trends for undergroundexcavation response. Their report to this conferenceconcerns the interpretation of flat jack measurements ingranitic metamorphic and sedime ntary rocks. The flat jackarrangement used is the established method of drilling a

series of overlapping holes. Like many researchers theyhave opted for relatively shallow jacks (0,5 m) in orderpresumably to overcome the problem of deflection measu-rements which they carry out using rigid rods anchored atthe flat jack mid-plane. The measurements made (whichwere compared with other borehole stress data) allowed a

cornparison between the theoretical stress state at the jacklocation and that acfually measured. A good agreementbetween theoretical (elastic) stresses and measured valueswas only obtained in the soft sedimentary rocks. In othercases the stress concentration was either less or greaterthan 2. For values less than 2, an attempt was made toobtain results on the residual strength parameters of therock on the basis that yielding had occurred around thetunnel. As pointed out in the paper the results show thatthe flat jack measurements are not conclusive either interrns of the mechanical properties of the rock or the stressstate. It would appear as others have found that the over-lying assumption of elastic or at best a uniform but yieldingrock is rarely applicable in rock masses particularly ifdisturbance due to the tunnel excavation has occurred. Bycorrelating the flat jack data with other plate bearing testinforrnation and microseismic studies, a layer of loosenedrock up to 0,8 m from the excavation surface was deduced.A further problem that arises is in the rotation of themeasurement rods. This will be influenced in particularby unfavourably oriented rock structure and other inhomo-geneities. The paper mentions that further study of thedata will be carried out using a 3-D elastic, isotropic andhomogeneous model. For the geometrical arrangementused (shallow jack emplacement) it is doubtful, given theuncertainties arising from geological and excavation conditions, whether meaningful values will ever be deduced.Furthermore the use of the data from flat jack tests inpressure tunnel design is considerably less direct thant theplate bearing or radial jacking test which loads in the samesense as the tunnel lining.

Additional laboratory studies have been conducted byBorsetto and colleagues from ISMES and ENEC in anattemp t to further rationalise the in situ observations.The tests which were made on a 1 m concrete block showan expected agreement between measured and theoreticalvalues. Problems of measurements of internal displacementsarc addressed and the rotation of rods (as used in thet insitu experiments) was found to be quite signific ant. Theirstudies on a rather idealised 'massive' specimen tend toreinforce earlier doubts as to the applicability of shallow

flat jack measurements for the determination of the mecha-nical properties of real rock masses.

2.2. Rock mass shear strength

The execution of shear tests for determining the strengthof dam foundations is briefly reported by Baba frômJapan. The test affangement is fairly classical and nofield results are quoted.

For both dam foundations and underground openings, thein situ measurement of joint shear strength remain a problem. With the standard ISRM test affangement, the norrnalstiffness of the test specimen cannot be controlled and thusthe major parameter controlling the strength of joints ininsitu rock masses cannot be properly taken into account.

3. In situ stress nreasurements

The in situ state of stress has a well known influence onthe stability and deformation properties of undergroundopenings. Before dealing with actual measurements it isperhaps useful to mention that a great deal of informationcan be obtained from regional geological observationsincluding the nature of deposition and erosion processesand tectonic events. In planning stress measurements it isthus useful to assess the likely stress field based on geological evidence.

A further general comment is concerned with the under-lying assumptions of linear elasticity for the interpretationof many of the test methods. Whilst competent graniticrock masses may respond in this manner stresses are oftennot relevant in such rock types. [n the weaker rocks wherestress to strength ratios approaching unity may be encoun-tered in common underground construction, the behaviourof the rock to an unloading stress path (common to allmethods except hydrofracturing) may be quite non-lineat.For such test methods it is very important to determine therelationship between stress and strain in the unloadingmode on the same material sample as that used for theactual in situ measurements. (Gonano and Sharp 1983).

Stress measurement programmes should not be embarkedupon lightly. No measurements are better than the oddstress measurement. For many projects they may not benecessary and an estimation of stress ranges may be allthat is required. Critical conditions in underground cons-truction can be summarised as follows:o excavation resulting in induced stress to strength ratiosgreater than 0,5o geological environments where stresses may be highlyanisotropic (near surface excavations in sedimentaryrock)o larye shallow excavations in low stress environments. pressure tunnels in regions where the internal pressureis of the same order as the minimum principal stress.

The criteria governing the choice of method are complexand relate particularly to the nature of the rock. Becauseof uncertainties associated with all current methods, mostwell planned investigations will involve more than onemethod to provide a realistic degree of measurement redun-dancy and a cross check between results. In many rockmasses significant variability in data should be expectedand thus the test programme must be sufficiently compre-hensive to cater for this.

Tab. 4lL: Review of modulus determinations for various hydroelectric projects

SCHEME

DEZ DA},!

DMKENSBERG

FFESTINIOG

I(AUNERTAL

KET'IANO

LAGO DELIO

ORAT.IGE FI SH

ROCK CONDITIONS

poor conglonerates ofvariable compositionand cementat ion

impermeable inter-bedded sandstones/siltstones/mudstones

DESIGN REQUIREMENTS METHODS

Laboratory TesÈsHydraul ic j acksPlate Bearing Testsl\lo hydros tat ic

chambers

Laboratory TestsPlate Bearing TesÈs

Chamber Tes t

rock support condit-ions for design ofprestressed concretepens tocks

assessment of rocksupport and over-burden criteria

D

DSs

RESULTS REI'IARKS REFERENCE

Special ar,renrion ro creep be- Dodds (1966)haviour. Moduli are quoted as

= 7-12 GPa secant values. Behaviour of= 8-10 GPa rock was very non-linear.

verify support cap- Seismic Testsacity of grouËed roik Pressure Galleryand determine over- TesËs (t, off )burden requirements

design of pressuretunnel using pre-stressed rocksuPPort

TIWAG radial jack

rock modulus required Instrumented sÈeelfor optimization of sphere groutedsteel lining thickness into cavity

Plate Bearing lestsPressure chamberBack analysis of

Machine Ha11

Lab Tests mudstonesiltsÈonesand s tone

PBT muds tonesiltsÈonesands tone

Pressure GalleryTests (in silt/sand s tone

E=Es = 18-24 GPad8pending ond irect ionD = 16-22 GPad8pend ing ond irect ion

E. = 46-58 GPaDo = 2l-53 GPapât"11e1 tobedd ingD- = 19 GPa per-pSndicular tobedd ing

D_ = 6 GPa bef- Design method using radialq

oie grouting jack testing and deformationD = 12 GPa af- moduli is well established.

stër grouting (forfissured line schist)

D- = 15-30 GPa 15 GPa used in design. At rhealloring for time few designs relied oncreep rock supporc Ëo minimize steel

thickness.

7-21 GPa| 0-27GPa7-20 GPa

Plate Bearing TesÈ results do Project ReporÈnot include disrurbed zone. (1979)VerÈical urodulus is signif i- Sharp & GonanocanÈly less than horizonÈal. (tggZ)

SubsÈanÈia'l saving in sÈeeleffected by investigating rocksupport conditions under lowcove r

Chapman (1961)good siltstones andgreywacke, strongcompetenÈ bedding

schists, gneiss,phyllite

s imi lar toDraken sbe rg

æ(tl

Lauffer & Seeber(1 e62)

D=DsËDs=Ds :r

Ds=Ds E

s

Brewer (1952)

Jaeger (t 97t )

On average, f ield val.ues are Kidd et al.about 016 times the laboratory (1976)values, PBT and pressuregallery tests are comparable.

SGPa35 GPa25 GPa7 GPa25 GPa25 GPa

= | 5-35 GPas

Tzbl. 412: Review of modulus determinations for various hydroelectric projects (continued)

SCHEME ROCK CONDITIONS

RAMA competent rocks

SIiOWY (EUCEMBENE- f ine grained quart z-TLIMUT TUNNEL) ite tightly jointed

gneissic granite

Various projects

Various projects variety of rocksnone sedimentary

DES IGN REQUIREMEI.ITS

investigation ofgrouting to improverock support capacity

to test load sharingcapacity of rock fivesites tested

pressure tunnel des-igt, and to studyeffect of grouting

METHODS

Micro-se i smicintact specimensin situ

Flat Jàck

Pressure Gallery

Lab specimens

Water Pressureloading of tunnelwi th measurementof def ormat ion

Plate Bearing lestsPressure Chamber

Tests

RESULTS

E 67 GPa

= 38 GPa

= | 3-33 GPa

= t0-16 GPa

= 45 GPa

= 24 GPa

= 37-62 GPa

= 38-63 GPa

F

la:dE

DS

ES

DS

DS

ESs

REMARKS REFERENCE

Kujundzic etal (1970)

Concluded that design could relyon supporÈ from rock and thatPBT an adequate indication of Belin (tg0O)modulus.

D = GPa AV.s

determination of TIWAG radial jackmoduli of reaction forPressure tunnel andshaft design

Elastic Modulus (static)Elas t ic Modulus (dynamic )Deformability Modulus (static)

Comparable res- Notes that grouting improves Jaeger (1955)ults obtained modulus considerably by up tofrom the two a factor of 2 or more.me thod s

Correct design values of Seeber (1970)deformability believed to becloser to bottom of experim-ental range of values.

E=-sL. =Do=

s

87

3. I . Measurernent methods

Bertrand and Durand from BRGM in France have provideda survey of test methods dividing them into three mainclasses:

o flat jack relatation methodo borehole overcore technique. hydraulic fracturing.

In particular they point out the limitations of each methodin the light of simplifying assymptions that arc oftenrequired (linear eleasticity, undisturbed rock mass zoneetc.). The main contributions of this paper are the resultsprovided frorn the Sain Sylvestre granite massif. Thegeological history of the site has been carefully studied toprovide geomorphological and tectonic evidence for regio-nal stress orientations.

Tests were carried out broadly in the same area using allthree methods. As usual interpretation of the flat jackmeasurements was influenced by assumptions regarding thestress concentrations around the adit periphery in whichthe tests were executed, the adit having been adverselyinfluenced by blasting damage. More useful data wereobtained frorn the borehole overcore tests which provideda basic indication of the governing stress field. Hydraulicfracturing tests were conducted both in a vertical boreholefrom surface and from horizontal boreholes underground.The data frorn the vertical hole gave a good correlationwith the overcore test data for the minimum principalstress. These data in turn correlated well with the major tec-tonic activity associated with the alpine thrust. The authorshave demonstrated the benefit of using overcoring teststogether with hydraulic fracturing to provide a compre-hensive understanding of stress conditions in relation tothe geological environment. A similar test programmealthough at considerably greater depths was recentlycarried out in the granites in Cornwall, U.K. Again goodcorrelation between overcoring, hydraulic fracturing andthe major alpine thrust direction was obtained (Pine et al.I e83).

A contribution from Charru a - Graéa from LNEC, Portugaldescribes a test progranrme involving both flat jack andovercoring measurements for a damsite foundation in ananisotropic schistose rock mass. The results of the flat jackprogramme were aS usual adversely influenced by exca-vation disturbance and the validity of the gallery stressconcentration model used was obviously questionable.Ten overcoring tests were made at three separate locationsand biaxial tests were subsequently carried out on theovercore retrieved. The isotropic interpretation of the datawas correctly viewed by the author as a limitation of themethod. In spite of reservations about various aspects ofthe tests, a reasonable correlation was obtained between thetwo methods as shown by stereographic plotting of theprincipal stresses.

An interesting contribution on a new development usingthe sleeve fracturing technique has beep provided byStephansson from the University of Luleâ, Sweden. Thetechnique in principal is similar to the hydrofracture tech-nique although problems regarding pore pressure assump-tiohs and effective stress conditions are avoided by isolatingthe pressurising fluid from the rock. The use of the cell todeteimine other properties such as modulus suffers fromthe same limitations as the Goodman Jack and the CSM cellin that they can only be rcgarded as point measurements as

far as most rock masses are concerned. Thus it cannot be

considered as an adequate loading mechanism for the rock

mass. [n terms of stress measurement the method is a usefuldevelopment but should it is suggested be used in conjunc-tion with the hydrofracture technique. If used solely as aborehole technique, the principal advantage of hydro-fracturing, namely the stressing of a representative volumeof the rock mass is lost. (It is well know that all boreholemethods require assumptions regarding the stress fieldaround the borehole and that this can be modified byproximal fractures.) Although overcoming the uncertaintyof pore fluid pressure conditions in the rock prior tofracfuring, the method still requires assumptions on thetensile strength of the rock in order to obtain the maximufnprincipal stress, a major difficulty that is inherent in thehy drofrac ture techniqu e.

An exemple of the use of structural geological observationsin predicting potential stress conditions followed up byextensive confirmatory measurements is given by Kazikaevfrom the USSR. The approach used is a good example ofhow stress measurements should be used to confirm thefindings of prior geological evaluation. Too many stressmeasurement studies are carried out without prior thoughtfor the geological environment and end up as a critique ofthe method rathçr than as an applied engineering study.The Russian data gives stress measurements over a depthrange from 150 to nearly 400 m and interestingly showincreasing ratios of horizontal to vertical stress with depth.Principal stress ellipses in relation to rock structure arepresented showing how the stress conditions vary over thestudy region (3 km2). A further interesting finding relatesto vertical stresses in a 100 m wide shale block boundedby quartzites. The vertical stresses in the shale corres-ponded to only 50 % of those in the qvaftzites as a resultof interactive support between the quartzites and the shale.

Most of the contributions reviewed so far have involvedmeasurements in rocks where linear elasticity can be reasonably assumed and where time dependancy effects aresmall. Unfortunately not all engineering projects can belocated in such easily understood materials and it is oftennecessary to understand stress conditions in weaker sedimentary rocks and salt deposits. (Particular attention isbeing focussed on salty rock formations for the disposal ofradioactive waste products).

A contribution from Pahl and colleagues from Germanydescribes overcoring and flat jack tests in anhydrite androck salt formations. They report initially on stresses withinboth intact and overloaded pillars and present interestingdata on both absolute stress levels and horizontal to verticalstress ratios in anhydrite rock. Measurements in rock saltwere also conducted using both methods. The rock stresseswere calculated assuming linear elasticity, the effect of rockcreep being compensated by using a dilatometer loadingperiod equal to the overcore unloading period. The authorsdo not comment on the differences between the unloadingand loading stress paths which could be significant in thistype of rock.

Hydraulic fracturing stress measurements in a bedded saltydèposit at a depth ôf some 700 m are described by Cornetand Thomas from France. Their paper presents a criticalreview of the hydrofracture method and the assumptionsused. The use of properties derived from fracture propa-gation (second stage injection) to overcome assumptionsregarding the tensile strength of the rock is proposed andexamples are given of the results obtained from bothmethods. Results are provided from two sites in the sameregion. At the first siie, three tests were carried out at adepth of some 700 m in a vertical borehole and attemptsto obtain initial breakdown and repressurisation charac-teristics were made. At the second site two subhornontal

88

boreholes from galleries at a similar depth to the first siterwere tested. A good correspondence between the varioustests was obtained indicating the validity of the method andassumptions used. In particular the results obtained fromholes in different directions allow an improved under-standing of the interrelationship of the principal stresses.

4. Full scale in situ observations

An important group of full scale tests has already beendiscussed for establishing design parameters for pressuretunnels. Other papers received have cov ered observationsin salt cavities, low temperature caverns in argillites, measu-rernents in coal mine roadways and settlement observationsassociated with road tunnels.

A paper from Hugout and Dussaud of France discusses theperformance of gas storage in leached out salt cavities inwhich large pressure variations occur as a result of theoperating conditions. Particular attention is focussed onvolume losses amounting to some 30% over 10 years.A sonar survey after l0 years operation showed a significantheave of the floor of the cavity with little distress to thecavity roof. The changes in cavity volume are related tonumber of other factors besides creep such as thermarexpansion, complementary dissolution and apparentcompressibility. After devising tests with different liquidand gaseous fluids, it was found possible to isolate thecreep component and produce reliable simulations of thisfactor using established creep laws and measured materialpropertie s.

The verification of creep behaviour on this overall scaleand over a prolonged time period is a valuable developmentin the undeistanding of salt-cavern behaviour.

Berest and colleagues from France have presented a newtechnique for determining the volume of underground storage caverns. By inducing a pressure drop at the well headand observing the subsequent pressure oscillations it is pos-sible to relate the level response to the cavity volumeknowing the compressibility characteristics of the fluidand the cavity walls. The theoretical studies which involvedynamic simulation of the overall system are applied to anactual cavern in salt filled with brine and fuel oil. Compa-risons are made with a static compressibility method whichis regarded as complementary to the dynamic method pro-posed. Data are given for a 7 500 m3 cavern at variousarnbient pressures. The method proposed involves anelegant dynamic simulation approach and overcomescertain limitations of established static methods.

An interesting pilot study for a low temperature under-ground LNG storage cavern in clay is reported by DeSloovere from France. The need to study problems involv-ing both high and low temperatures in underground exca-vation design is becoming increasingly common and it isfitting that at least one paper in this Topic should be con-cerned with thermal effects. The excavation has been creat-ed in a stiff eocene clay with a diameter of 3,7 m and alength of 100 m by means of a roadheader followed by a

segnental concrete lining. The depth below surface is some23 m. A test section some 30 m long was cooled to

196 " C over the period November 1981 to June 1982.The temperature was maintained at that value until Septem-ber 1982 and then allowed to recover. The response of thecavity to cooling was measured by multipoint extensometers on horizontal and vertical arrays and by fïxed clino-meters in vertical boreholes. In addition monitoring of thesegmental concrete lining was undertaken together withmonitoring of the excavated clay surface. Extensive use was

made of vibrating wire instrumentation systems which werefound to be the only reliable measurement method atthe temperatures involved. In general the instrumentationappears to have behaved extremely well. The field resultshave been analysed by a study of displacement fields andinduced stresses on the lining structure. Apart from beingable to determine the interactive effects of the clay andchamber lining from the measured values, this type of pilotfacility provides an extrernely good overall dembnstràtionof the feasibility of operating such underground structuresunder adverse thermal conditions. Jt is perhaps worthnoting that similar studies to + 200 "C are in the advanceplanning stages for the underground storage of nuclearwaste materials.

A paper from China by Zhenxi and Guoliang provides inte-resting observed performance from underground openingsin coal mines. Perhaps in the mining industry one has thegreatest opportunity for developing designs through observ-ed performance owing to the ongoing nature of the ope-ration. Clearly in China where some I 000 km of shaftsand tunnêls per year arc supported by bolting and shot-creting there is tremendous opportunity for optimisation.The instrumentation used is described and comprisesconvergence, extensometer and support load monitoringas well as acoustic investigation of the excavation srround.Results are given from the sites at depth in the range 27U360 m. As might be expected in coal measure rocks, closures of the order of 20 mm have been typically observed.The extent of rock loosening around the opening is of theorder of 1,0 m for a 4 m wide roadway. The performanceof the roadways subsequently subjected to coal seam ex-traction has also been monitored and additional closuresof some 20 mm were measured when the face was withint 50 m of the roadway location. From the evidence givenit is apparent that the shotcrete lining in conjunction withbolting can in the softer rocks withstand up to 100 mmof closure without general distress. It is reported thatsome spalling of shotcrete in the crown has been observed(a rather typical occurrence in this type of rock) whichrequires remedial treatment. Use of mesh reinforcementin such areas generally improves the performance of theshotcrete lining. The paper reports a well planned andexecuted observations program which has allowed modernsupport techniques to be applied with confidence. In a

world where we are still so often bound by traditionnecessitating the continuing use of inefficient steel archsystems, this alternative, systematic approach is veryencouraging.

The final paper on observed conditions concerns shallowtunnelling in Japan through sand and silt with an excavatedcross section of some 85 mz. Yokoyarna and colleagueshave relied heavily on observations of tunnel performancetogether with borehole loading tests to estimate the magnitude of ground settlement and the stability of the tunnelface. Shotcrete, mesh and steel ribs together with anchorbars are used to stabilise the tunnel periphery. The groundcharacteristics in terms of stiffness are determined ahead oftunnelling using pressuremeter and borehole jacking tests.With the type of support used (relatively stiff) and thelimited overburden depth, the main settlement problemsarc related to the behaviour of the tunnel face and theoverlying ground ahead of the tunnel. Face collapse andcaving are predicted from a study of ground settlementsrelated to the material characteristics. The authors give a

collapse criterion based on field measurements inclûdingstudies of conditions before and after an actual failure ofthe face resulting in a ground settlement of some 40cm.The benefit of spiling bolts is also demonstrated in limitingground settlements by up to 50 Vo. Such reports are always

89

most valuable to the practising engineer particularly whenthey include observations of the actual failure state. Unfor-tunately without the occasional full scale failure to learnfrom it is extremely difficult to know the degree of conser-vatism in designs which are based solely on measurementof the material properties.

5. Summary and conclusions

The papers addressed to this Topic have covered a widearca. In particular valuable contributions have been made offull scale observations both under difticult ground conditions and for new types of engineering developments.

The current status of the in situ test methods covered canbe summarised as follows:

In situ design property tests

Ro c k Ma ss D eforrnability

The test rnu st be related to the engineering product forwhich this data will be used. The deformability characte-ristics of both disturbed and undisturbed zones can bemeasured.

- Plate bearing tests are still the most useful current test.High loadings (up to 1000 tons) may be required in stiffrock masses to induce adequate deformations at depth forrock mass property determination. The tests can be extend-ed over a wider site scale using geophysical methods.

Flat jack testing is only meaningful in most fracturedrock masses if a sufficient depth of jack emplacement canbe achieved together with a measurement system within therock mass- If internal jack deformation measurements aremade (Rocha system) iawn slots are preferred. Shallow flatjack tests close to the rock surface using surface or boreholeanchored systems do not yield representative data.

For pressure tunnel design, full scale chamber teststogether with geophysical classification along the tunnelalignment still provide the most reliable method for assess-

ing long term support that may be derived from the rock.For concrete linings that may leak, water is the mosteffective loading medium as effective stress conditions inthe rock can be observed directly.

Shear Strength

Larye scale in situ shear tests ate no longer consideredjustifiable except perhaps for major and unique planes ofweakness affecting say the stability of a dam abutment.Problems remain in the execution of the test in relationto the correct representation of normal stiffness.

In situ stress measurements

The use of more than one test method to determine stressconditions is now common practice. Particularly valuableare combinations of the overcoring and hydrofracture rne-thods. The overcore methods permit determination of thecomplete stress tensor although the absolute magnitude ofthe stresses measured is often influenced by rock propertyassumptions. They iue also limited in relation to depthfrom access locations (surface or underground). The hydro-fracture test is valuable in that it allows reliable deterrnina-tion of the minimum principal stress and can be conductedto great depth.

With the increasing use of deep bored shafts and tunnels,the overcored rosette rnethod is also worthy of furtherconsideration. (Brady et al., 1976).

Th9 flat jack method as currently used in proximity tounderground excavations is not recommended-owing to theuncertainty and variability associated with stresses aroundunderground openings.

Full scale in situ observations

The most valuable design basis is without doubt derivedfrom full scale in siru performance observations. [n tunnell-ing the approach is quite routine and used to determinesupport requirements as the proj ect proceeds. For per-manent underground caverns, trial or prototype openingshave been constructed or are planned for many differentuses including hydroelectric installations, gas and radio-active waste storage.

In spite of their apparently high cost, the potential designrefinements that can be made and in particular the reductionin uncertainty during construction usually result in signifi-cant overall project savings. In many cases such trial open-ings may be essential to confirm the feasibility of a particular development.

5.1. Future developments of in situ testing techniques

From the papers presented it is encouraging to note thatconsiderable emphasis is now being placed on the testingof soil and rock masses in sttu at a scale that provides botha representative sample and a direct relationship to theengineering structure under consideration. Further testingdevelopments should be aimed at providing representativemass information both for specific project sites and forrock masses in general.

Specific areas of development are suggested as follows:

Rock mass deformability

Plate Bearing Test

- Development of more sensitive displacement measure-ment systems together with increased load capacities fortesting in stiff, undisturbed rock masses including long termcreep evaluations.

- Determination of the state of stress under the loadedarea as a function of the rock mass characteristics and thetest excavation geometry.

- Development of improved correlation techniques betweenthe plate bearing test and geophysical measurement methods.

Flat Jack Test

- Development of a deep slotting technique to allowemplacement of jacks within the undisturbed rock mass.Associated development of transverse MPBX system tomeasure deformations on a representative scale. Test layoutsmust incorporate facilities for drilling orthogonally tojacking plane.(Kim and Sharp, 1983).

- Development of improved correlation techniques betweenthe flat jack test and geophysical measurement methods.

In situ Block Tests

- Loading of in situ rock samples in the form of blocksisolated in the walls or inverts of underground openings isbecoming a relatively common form of test. Flat jacksemplaced around in slots are used for biaxial loading of thesample. Loading in the third direction can be carried outusing stressing tendons which pass through the block itself.(Kim and Sharp 1983).

90

Further development of this test to determine rock massdefonnability under a controlled state of stress (knownboundary conditions) is advocated.

G e ophy sic aI M e asu rem en t s

Further development and increased usage of thesemeasurements are proposed to provide a reliable means ofsampling rock masses over a considerable extent.

Shear strength

Continued development of empirical data and compa-rison with observed performance in situ.

- D evelopment of a stiff in situ testing arrangement suchas block shearing between parallel joint planes in a tunnelsidewall including measurements of stress and stiffnessnormal to the shear plane.

In sîtu stress measurement

- Extended use of test prografirmes incorporating measure-ments with more than one technique (combined overcoreand hydrofracture tests).

- Development of existing techniques to overcome prcblerns associated with rock mass structure.

- D evelopment of the overcoring technique to permit itsusage at depth in conjunction with hydrofracture measu-rements.

Full scale in situ observations

Continue emphasis on the value of such observations todevelop a comprehensive understanding of rock massresponse to the loading conditions imposed by the engineer-ing structure. Particular attention should be focussed on thefollowing areas:

- Non-watertight pressure tunnels with interaction betweenthe water, lining and rock to evaluate long term supportcapacity and lining behaviour.

- Underground storage particularly at elevated pressures orwhere high thermal gradients are induced (gas, liquid andwaste storage facilities).

- Tunnels in general to permit an improved understandingof rock behaviour and level of support requirements.

- Lal$e -underground openings in weak ground requiringspecific developments in support techniques.

References

BARTON N., and CHOUBEY v. (1977): Shear Strength of RockJoints in Theory and Practice. Rock Mechanics 10, r-s4(197 7).

BRADY B.H.B., et al. (r97 6): Stress Measurement in a BoredRaise at the Mount Isa Mine. I.S.R.M. Symposium on Inves-tigation of Stress in Rock, Advances in Stress Measurement,Sydney.

GONANO L.P. and SHARP J.C., (1983): Critical Evaluation ofRock Behaviour for In situ Stress Determination using Over-coring Methods. Proc. Sth Congress of Int. Soc. Rock Mech.,Melbourne.

HARDIN E. et al., (1982): Measuring the Thermomechanical andTransport Properties of a Rockmass Using the Heated BlockTest. Proc. 22nd U.S. Symposium on Rock Mechanics,Cambridge, Mass.

KIM K. and SH.ARP J.C., (1983): Design and Execution of a LargeScale Heated Block Test in Basalt to Determine the Thermo-mechanical Properties of Basalt under Triaxial LoadingConditions. [nt. Symp. on Field Measurements in Geo-mechanics, Zurich, (to be presented).

LAUFFER H. and SEEBER G., (1961): Design and Control ofLinings of Pressure Tunnels and Shafts, Based on Measure-ments of the Deformability of the Rock. Proc. 7th Congressof Large Dams, Rome, Report 9L.

MELIORS T.w. and SHARP J.C., (1982): Investigations for theDrakensberg Power Stations Complex in Weak SedimentaryRock. Proc. ISRM Symp. Rock Mech.: Caverns and PressureShafts, Aachen.

PINE R.J. et aL, (1983): In sifn Stress Measurement in the Camme-nellis Granite; Part 1, Overcoring Tests at 790m; Part 2,Hydrofracture Tests to 2000m. Camborne Shool of Mines -Geothermal Energy Project, I.J.R.M.M.S. (in press).

SHARP J.C. and GONANO L.P. (1982): Rock Engineering Aspectsof the Concrete Lined Pressure Tunnels of the DrakensbergPumped storage Scheme. Proc. ISRM Symp.o Rock Mech.:Caverns and Pressure Shafts, Aachen.