Goodman R E - Introduction To Rock Mechanics 2nd Edition
Goodman R E - Introduction To Rock Mechanics 2nd Edition
Goodman R E - Introduction To Rock Mechanics 2nd Edition
was constructed by le o o o' 111 -
completion of the pow h avmg a core of sohd gramte unexcavated until the
plished by using ouse sorne years later; this was accom-
limits of the bl t dO blastmg techmque on the upstream and downstream
as a to the cofferdam
D oTransportation engineering also eaUs roek mechanics in many ways
of cut for highwoays, railways, canals, pipelines, and penstock;
may mvolve testmg and analys1s of the system of discontinuitieso Considerable
1.1 Fields of Application ofHock Mechanics 9
cost savings are possible if the orientation of the right of way can be adjusted
based on the rock mechanics studies, but this is not always practical. The
decision to place portions of such routes underground is partly determined by
judgments about the rock conditions and relative costs of open cuts and tun-
neIso Savings can be realized in penstock steeI by assigning a portion of the
stress to the rock if the penstock is placed in a tunnel; in that case rock tests can
determine rock properties for the designo Sometimes penstocks can be left
unlined; rock stress measurements may then be required to assure that leakage
will not be disastrouso In urban areas, transportation routes at the surface may
have to accept subvertical sIopes because of the high values of Iand, and,
accordingly, permanentIy stable slopes will have to be maintained by artificial
supportso Considerable testing and analysis of the rock may be justified to
provide an interpretational framework for instruments provided to monitor
long-term safetyo
Surface excavations for other purposes may also demand rock mechanics
input in control of blasting, selection of cut slopes and Iocation of safety
benches, and provision for support. In the case of open pit mines, which reIy on
economical excavation for profitabIe operation, considerable study may be
warranted in choosing appropriate rock slopeso Statistical methods of dealing
with the many variables are being developed to enable the mine planner to
determine mining costs in the most useful termso Since these mines cannot
afford generous factors of safety, they often support thorough monitoring of
rock deformation and stress. Normally, artificial supports are not provided
because the costs would be prohibitive, but rock boJts, retaining structures,
drains, and other measures are sometimes required at the sites of power struc-
tures and at crushers or conveyor belts within th pit. Spillway cuts for dams
also can attain impressive dimensions and demand rock mechanics attention
(Figo 1.3)0 Such cuts as sume a value far greater than their cost since failure at
an unfortunate time could allow overtopping of the dam; even so, the costs of
major spillway cuts can rival the cost of even a large dam and thus such
excavations can be considered engineering structures in their own righi. Rock
mechanics affects the decision on whether to locate spillways in open cut or in
tunnelso
Underground excavations call upon the discipline of rock mechanics in
many wayso In mining, the design of cutters and drills can be tailored to the
rock conditions, which are determined by suitable laboratory tests. This also
applied to tunneling with moles or tunneling machineso A major decision of
mining is whether to attempt to maintain the openings while removing the ore,
or instead to let the rock deformo The rock condition and state of stress is
fundamentally important in reaching this decision correctlyo In stable mining
methods, the dimensions of pillars, rooms, and other rock components are
based upon rock mechanics studies using numerical analysis or applicable the-
ory, and calling into play thorough rock testing programs. In the case of unsta-
10 Infroduction
Figure 1.3 The ftip bucket for the side-hill spillway
for Chivor rock-fill dam, Colombia. Note the road-
way and access tunnel in the lower left and the
drainage tunnel under the ftip bucket. (Owner,
I.S.A.; Engineer, Ingetec, Ltda.)
ble the layout of haulageways and "draw points" is based
studles aImmg to minimize dilution of ore with waste rock and to optimize
efficlency.
Underground chambers are now being used for a variety of purposes other
transportation mining. Sorne of these applications are demanding new
kmds of data and specIaI technology. Storage of liquefied natural gas in under-
ground chambers requires determination of rack praperties under conditions of
extreme cold and analysis of heat transfer in the rack. Storage of oil and gas in
mined chambers (Figure 1.4) requires a leakproof underground environment.
Any large undergraund chamber, regardless ofits special requirements, should
1.1 Fields of Application of Rock Mechanics 11
Figure 1.4 An underground chamber for st?:age of
troleum products in Norway. A storage facIllty conslsts
of a number of such chambers. (Photo courtesy of Tor
Brekke.)
be stable essentially without support and this depends upon the state of
and the pattern and properties of discontinuities. Underground
power plants which offer advantages over surface power plants m mountam-
, ) d ous
ous terrain feature very large machine halls (e.g., 25-m span an numer
, . . p. 7 1) The
other openings in a complex three-dlmenslOnal arrangement 19. . .
orientation and layout of these openings depend almost entIrely upon rock
mechanics and geological considerations. Blasting, design of supports, and
most other engineering aspects of such schemes depend rack
conditions therefore rack mechanics is a basic too1. The mibtary IS mterested
in openings to create invulnerable facilities. dynamics has
figured prominentIy in design of such schemes, since the secunty of the open-
ings must be maintained in the face of enormous graund shock pressures. The
1: lntroduction
military has sponsored special prototype tests to failure that have advanced the
knowledge of rock properties and behavior and of rock/structure interactions.
mechanics is important in the field of energy deuelopment (in
addlhon to the hydroelectnc works aIread y mentioned). In petroleum engineer-
ing, design of drilling bits depends upon rock properties; bit wear is one of the
major elements of cost. Rock mechanics studies are being directed toward
solving the problems associated with deep drilling, to allow recovery from
greater depths. In shales, salts, and certain other rocks, depth limitations are
created by flowage of the rock and rapid c10sure of the holeo A laboratory has
been built in Salt Lake City (Terra Tek Drilling Laboratory) to allow full-scale
simulation of drilling at depths up to 20,000 feet and at temperatures up to
340C. The petroleum industry pioneered the use ofhydraulically induced frac-
tures to increase reservoir yield. Hydraulic fracturing is now a standard re ser-
voir operation. It is also being investigated as a mechanism for exchanging the
earth's heat as a source of geothermal energy in dry, hot rocks. In the Los
Alamos. Scientific Laboratory scheme, under full-scale field investigation, a
hydraulIc fracture circulates cold water into hot rock; the heated water is
returned to the surface through a second drill hole intersecting the top of the
fracture. In the nuclear energy field, in addition to the problems of constructing
the surface and/or underground facilities in rock and the elaborate precautions
required by licensing agencies to insure that there are no active faults or other
hazards. on site, the industry is burdened with large quantities of
hlgbly tOXIC, 10ng-lIved radioactiue wastes. The current plan is to isolate these
in stainless steel canisters by emplacement in specially mined cavities in
deposlts ofrock salt and perhaps in granite, basalt, tuff or other rock types. Salt
was selected because of its relatively high heat conductivity together with
water tightness since fractures tend to be absent or healed. The rock
wIlI. as sume temperatures of approximately 200C after emplacement of the
camsters.
New. applications for rock mechanics are appearing with great rapidity.
ExploratlOn of extraterrestrial space, prediction of earth-
quakes, mmmg, compressed air storage in underground chambers, and
other eX?tlc fields calling on further development of rock technology.
we.are. still not completely in command ofthe essential ingredients
for m sorne of the more mundane applications mentioned previ-
ously. Thls IS because of the special nature of rock, which renders it different
and perhaps more difficult to deal with than other engineering materials.
1.2 The Nalure ofRocks
When attempting to formulate mechanical behavior of solids, it is common to
as sume they are ideally homogeneous, continuous, isotropic (nondirectional in
1': The Nature of Rocks 13
rt
.e) llnear and elastic. Rocks can be nonideal in a number of ways.
prope 1 s , ' 1
. t they are seldom truly continuous, because pores or fissures are usua .
Flrs , d . 1 t ., d b
ected P
ores approximately eqUl ImenslOna caViles, are loun e-
Interconn , . . ., d
th g
rains of sedimentary rocks. Isolated vugs of other ongms are loun
tween e .. f k t
. 1 a IC rocks and soluble carbonate rocks. Smce the capaclty o roc s o
10 VO C n . f h d
d t
ransmit fluid s is largely dependent upon the behavlOr o t ese VOl s,
store an .. k 1 .
. 1 theory has been developed pnmanly by wor ers m petro eum engl-
a specta ' .
. to deal with the deformations, stresses, and water pressures m porous
neenng, . h d k h t h
k
Ml
crohssures are small planar cracks common m ar roc s t a ave
roc s. 'J< 11 d t 1
undergone internal deformation; they occur as me an crys a
boundary cracks. A fissured rock is like a test speclmen that has loaded
into the cracking region (i.e., that has been behavlOr of the
network of fissures is as important or even .more Vital regard to rock
properties than the mineralogic Itself. fissures and
pores do the following: they create nonlInear re.sponse, espe-
cially at low stress levels; they reduce t?e tensIle (espectally
they create stress dependency in propertles; they
and scatter in test results; and they mtroduce a scale effect lOto predlctlOns of
behavior. . .
A related nonideality of most rocks is the presence of macrodlscontl-
nuities. Regular cracks and fractures are usual at shallow depth.s beneath. the
surface and sorne persist to depths of thousands of meters. Jomts: beddmg-
plane partings, minor faults, and other recurrent planar fractu!es radlc.ally. alter
the behavior of rock in place from that predictable on the basls of .testm
g
.mtact
samples, even though the latter may possess fissures. The mechamcs of dlscon-
tinuous rocks is especially relevant to engineers of surface structures, surface
excavations and shallow underground excavations. Indeed, it was the move-
ment of a biock bounded by faults and joints that undermined the Malpasset
Arch Dam in 1959 (Figure 1.5). .-
The effect of a single fracture in a rock mass is to lower the tensIle strength
nearly to zero in the direction perpendicular to the fracture plane, and to
restrict the shear strength in the direction parallel to the fracture planeo If
joints are not randomly distributed (and they almost never are) then the .IS
to create pronounced anisotropy of strength, as weH as of aH other propertles m
the rock mass. For example, the strength of a foundation loaded to
the bedding may be less than one-half of the strength when the load IS apphed
perpendicular or paraHel to the bedding. Anisotropy is comm?n in
even without discontinuous structure because of preferred onentatlOns of mm-
eral grains or directional stress history. Foliation and schistosity make schists,
slates, and many other metamorphic rocks highly directional their deforma-
bility, strength, and other properties. Bedding makes shales,
stones and limestones and other common sedimentary rocks hIghly amsO-
tropic. AIso, even rock specimens apparently free from bedding
such as thick-bedded sandstones and limestones, may prove to have dlrechonal
14 Introduction
Figure 1.5 A view of the left abutment of Malpas-
set arch after its failure. The movement of a
dehmlted by discontinuity surfaces, one of
whlch forms the newly exposed rock surface on
the abutment, brought on the rupture of the con-
crete arch.
properties because they were subjected to unequal principal stresses as the
were transformed from sediment into rock. FinalIy fi d
unequal initial stresses wilI be anisotropic
y In uenced by the state of stress across the fissures' the are one
matehnal wdhen the fissures are cIosed, and another when the fissu;es a;e opened
or s eare .
d
. We. can discuss a "mechanics of rocks" in these chapters but such a
lSCUSSlon must be b d . 'f' .
" k'" roa In 1 It lS to have general value because the term
roc Includes a great vanety of material types. Granite can behave in a
Sources of Information in Rack Mechanics 15
brittle, elastic manner, Up to confining pressures of hundreds of megapascals
l
(MPa), while carbonate rocks become plastic at moderate pressures and ftow
like cIay. Compaction s hale s and friable sandstones are weakened by immer-
sion in water. Gypsum and rock salt are incIined to behave plasticalIy at rela-
tively low confining pressures and are highly soluble.
Despite all these problems with rock as an engineering material, it is possi-
ble to support engineering decisions with meaningful tests, calculations, and
observations. This is the subject of our study.
Sources of Information in Rock Mechanics
BIBLIOGRAPHIES
KWIC Index of Rock Mechanics Literature published before 1969, in two volumes, E.
Hoek (Ed.). Produced by Rock Mechanics Information Service, Imperial College,
London. Published by AIME, 345 E. 47th Street, New York, NY 10017. A compan-
ion volume, Part 2, carrying the bibliography forward from 1969 to 1976 was pub-
lished by Pergamon Press Ltd, Oxford (1979); J. P. Jenkins and E. T. Brown (Eds.).
Geomechanics Abstracts: see International Journal of Rock Mechanics and Mining
Science. These are key-worded abstracts of articles published worldwide; issued
and bound with the journal.
BOOKS
Attewell, P. B. and Farmer, 1. W. (1976) Principies of Engineering Geology, Chapman
& Hall, London.
Bieniawski, Z. T. (1984) Rock Mechanics Design in Mining and Tunneling, Balkema,
Rotterdam.
Brady, B. H. G. and Brown, E. T. (1985) Rock Mechanics for Underground Mining,
Allen & Unwin, London.
Brown, E. T. (Ed.) (1981) Rock Characterization. Testing, and Monitoring: ISRM
Suggested Methods, Pergamon, Oxford.
Brown, E. T. (Ed.) (1987) Analytical and Computational Methods in Engineering Rock
Mechanics, Allen & Unwin, London.
Budavari, S. (Ed.) (1983) Rock Mechanics in Mining Practise, South African Institute
of Mining and Metallurgy, Johannesburg.
Coates, R. E. (1970) Rock Mechanics Principies, Mines Branch Monograph 874, re-
vised, CANMET (Canadian Dept. of Energy, Mines and Resources), Ottawa.
DOWding, C. H. (1985) Blast Vibration Monitoring and Control, Prentice-Hall, Engle-
wood Cliffs, NJ.
Farmer, 1. W. (1983) Engineering Behaviour of Rocks, 2d ed., Chapman & Hall, Lon-
don.
I One megapascal equals 145 psi.
16 Introduction
Goodman, R. E. (1976) Methods of Geological Engineering in Discontinuous Rocks,
West, St. Paul, MN.
Goodman, R. E. and Shi, G. H. (1985) Block Theory and lts Application to Rock
Engineering, Prentice-Hall, Englewood Cliffs, NJ.
Hoek, E. and Bray, J. (1981) Rock Slope Engineering, 3d ed., Institute of Mining and
Metallurgy, London.
Hoek, E. and Brown, E. T. (1980) Underground Excavations in Rock, Institute of
Mining and Metallurgy, London.
Jaeger, C. (1972) Rock Mechanics and Engineering, Cambridge Univ. Press, London.
Jaeger, J. C. and Cook, N. G. W. (1979) Fundamentals of Rock Mechanics, 3d ed.,
Chapman & Hall, London.
Krynine, D. and Judd, W. (1959) Principies of Engineering Geology and Geotechnics,
McGraw-Hill, New York.
Lama, R. D. and Vutukuri, V. S., with Saluja, S. S. (1974, 1978) Handbook on Mechan-
ical Properties of Rocks (in four volumes), Trans Tech Publications, Rockport,
MA. Vol. 1 (1974) by Vutukuri, Lama, and Saluja; Vols. 2-4 (1978) by Lama and
Vutukuri.
Obert, L. and Duvall, W. (1967) Rock Mechanics and the Design of Structures in Rocks,
Wiley, New York.
Priest, S. D. (1985) Hemispherical Projection Methods in Rock Mechanics, Allen &
Unwin, London.
Roberts, A. (1976) Geotechnology, Pergamon, Oxford.
Turchaninov, 1. A., Iofis, M. A., and Kasparyan, E. V. (1979) Principies of Rock
Mechanics, Terraspace, Rockville, MD.
Zaruba, Q. and Mencl, V. (1976) Engineering Geology, EIsevier, New York.
JOURNALS
Canadian Geotechnical Journal, Canadian National Research Council, Toronto,
Canada.
lnternational Journal of Rock Mechanics and Mining Sciences & Geomechanics Ab-
stracts, Pergamon Press, Ltd., Oxford.
Geotechnical Testing Journal, American Society for Testing Materials.
Journal of the Geotechnical Division, Proceedings of the American Society of Civil
Engineering (ASCE), New York.
Rock Mechanics, Springer-Verlag, Vienna.
Underground Space, American Underground Association, Pergamon Press, Ltd.,
Oxford.
PROCEEDINGS
Canadian Rock Mechanics Symposia, Annual; various publishers. Sponsored by the
Canadian Advisory Committee on Rock Mec;hanics.
Congresses of the Intemational Society of Rock Mechanics (ISRM), First-Lisbon
(1966); Second-Belgrade (1970); Third-Denver (1974); Fourth-Montreux
(1979); Fifth-Melboume (1983); Sixth-Montreal (1987).
Sources of Information in Rack Mechanics 17
al Conferences and Symposia sponsored by ISRM, Institute of Civil Engineers
Specl ty .. . al C L
(L don) British Geotechmcal Soclety, AIME, IntematlOn ongress on arge
n:;s and other organizations as cited in the references after each
chapter. . bl h S
sosia on Rock Mechanics, Annual U. S. Conference; pu IS ers. pon-
by the U. S. National Committee on Rock Mechamcs.
STANDARD S AND SUGGESTED METHODS
Rock mechanics has not yet advanced the stage where teIstin
g
techniques can be rigorously standardlzed. the oCIe y. or
Rock Mechanics (ISRM) and the American Soclety for Testmg and Matenals
(ASTM) have published "designations" and methods" for
and field testing and for description of rock .matenals. Several of these are bsted
with the references at the ends ofthe appropnate chapters. See (.1981) under
"BOOKS" aboye. For up-to-date information about m
h ics communicate directly with ISRM, Commlsslon on Standardizatlon,
Nacional de Engenharia Civil, Avenida do Brasil, P-1799,
Portugal; and with ASTM, Committee D-18 on Soil and Rock for Engmeenng
Purposes, 1916 Race Street, Philadelphia, PA 19103.
Classification and
Index Properties of
Rocks
2.1 Geological Classification of Rocks
Chap
ter
2
Although they were not developed to satisfy the needs of civil engineers, the
names geologists are able to attach to rock specimens on the basis of limited
observations with a hand lens, or with the eye alone, do often reveal something
about rock properties. If you are unfamiliar with the common rock names and
how to assign them to an unknown rock, a review of geology is highly recom-
mended. A good way to begin is to study Appendix 3, which explains simplified
schemes for classifying and naming the principal rocks and minerals. Appendix
3 also lists the periods of the earth's history, the names of which indicate the
age of a rock. A rock's age often, but not infallibly, correlates with its hardness,
strength, durability, and other properties.
From a gene tic point of view, rocks are usually divided into the three
groups: igneous, metamorphic, and sedimentary. Yet these names are the
~ e s u l t s , not the starting point of classification. Since we are interested in behav-
loral rather than genetic attributes of rocks, it makes more sense to divide the
rocks into the following classes and subclasses:
19
20 Classification and n d e ~ Properties of Rocks
l. CrystaUine Texture
A. Soluble carbonates and salts
B. Mica or other planar mineral s
in continuous bands
C. Banded silicate minerals with-
out continuous mica sheets
D. RandomIy oriented and distrib-
uted silicate minerals of uni-
form grain size
E. Randomly oriented and distrib-
uted silicate minerals in a back-
ground of very fine grain and
with vugs
F. Highly sheared rocks
11. Clastic Texture
A. Stably cemented
B. With slightly soluble cement
C. With highly soluble cement
D. Incompletely or weakly ce-
mented
E. Uncemented
111. Very Fine-Grained Rocks
A. Isotropic, hard rocks
B. Anisotropic on a macro scale
but microscopically isotropic
hard rocks
C. Microscopically anisotropic
hard rocks
D. Soft, soil-like rocks
Examples
Limestone, dolomite, marble, rock
salt, trona, gypsum
Mica schist, chlorite schist, graph-
ite schist
Gneiss
Granite, diorite, gabbro, syenite
Basalt, rhyolite, other volcanic
rocks
Serpentinite, mylonite
Examples
Silica-cemented sandstone and
limonite sandstones
Calcite-cemented sandstone and
conglomerate
Gypsum-cemented sandstones and
conglomerates
Friable sandstones, tuff
Clay-bound sandstones
Examples
Hornfels, sorne basalts
Cemented shales, flagstones
Slate, phyllite
Compaction shale, chalk, marl
2.1 Geological Classification of Rocks 21
IV. Organic Rocks
Examples
A. Soft coal Lignite and bituminous coal
B. Hard coal
C. "Oil shale"
D. Bituminous shale
E. Tar sand
Crystalline rocks are constructed of tightly interlocked crystals of silicate
mineral s or carbonate, sulfate, or other salts (Figure 2.1a). Unweathered crys-
talline silicates like fresh granite are usually elastic and strong with brittle
failure characteristics at pressures throughout the usual range for civil engi-
neering works. However, ifthe crystals are separated by grain boundary cracks
(fissures), such rocks may deform nonlinearly and "plastically" (irreversibly).
Carbonates and crystalline salt rocks may also be strong and brittle but will
beco me plastic at modest confining pressures due to intracrystalline gliding.
AIso, they are soluble in water. Mica and other sheet minerals like serpentine,
tale, chlorite, and graphite reduce the strength of rocks due to easy sliding
along the cleavage surfaces. Mica schists and related rocks are highly aniso-
tropic rocks with low strength in directions along the schistosity (Figure 2.tb)
except when the schistosity has been deformed through refolding. V.olcanic
rocks like basalts may present numerous small holes (vugs); otherwise, they'
behave similarly to granitic rocks (Figure 2.2c). Serpentinites, because they
tend to be pervasively sheared on hidden surfaces within almost any hand
specimen, are highly variable and often poor in their engineering properties.
The clastic rocks, composed of pieces of various rock types and assorted
mineral grains, owe their properties chiefly to the cement or binder that holds
the fragments together. Sorne are stably and tightly cemented and behave in a
brittle, elastic manner. Others are reduced to sediment upon more soaking in
water. In the clastic rock group, the geological names are not very useful for
rock mechanics because the name doesn't indicate the nature of the cement.
However, a full geological description can often suggest the properties of the
cement; for example, a friable sandstone, where grains can be liberated by
rubbing, is obviously incompletely or weakly cemented at best.
Shales are a group of rocks primarily composed of silt and clay that vary
Widely in durability, strength, deformability, and toughness. Cemented shales
can be hard and strong. Many so-called "compaction shales" and "mud-
stones," however, are just compacted clay soils without durable binder, and
have the attributes of hard soils rather than of rocks: they may exhibit volume
change upon wetting or drying together with extreme variation in properties
with variations in moisture content. Unlike soils, which quickly lose strength
when kept moist at their natural water content, compaction shales remain
Figure 2.1 Photornicrographs of thin sections of rocks, viewed in polarized, trans-
rnitted light (courtesy of Professor H. R. Wenk). (a) Tightly interlocked fabric of a
crystalline rock-diabase (x27).
22
Figure 2.1 Photornicrographs of thin sections or rocks, viewed in polarized,
transrnitted light (courtesy of Professor H. R. Wenk). (b) Highly anisotrophic
fabric of a quartz rnylonite (x20).
23
24 Classification and n d e ~ Properties of Rocks
Figure 2.2 Photomicrographs of thin sections of fissured rocks, photographed in
transmitted, polarized light (courtesy of H. R. Wenk). (a) Anorthosite with many
intracrvstalline and sorne nterervt"lIine fractures (x6.5).
2.1 Geological Classification of Rocks 25
Figure 2.2 Photomicrographs of thin sections of fissured rocks, photographed in
transmitted, polarized light (courtesy of H. R. Wenk). (b) Gabbro with regular fis-
sures oriented across the cleavage (x7).
26
Classification and Properties of Rocks
Figure .2.2 of thin sections of fissured rocks, photographed in
tr?nsmltted, polanzed hght (courtesy of H. R. Wenk). (e) Volcanic rock (trachyte)
wlth fissured sanidine phenocrysts (x30).
2.2 Properties of Rock Systems 27
intact for sorne time. However, when dried and then immersed in water, they
gradually decrease in density and strength over days, weeks, or longer. Chalk
is a highly porous elastic Crbonate rock that is elastic and brittle at low pres-
sures, but plastic at moderate pressures.
Organic rocks inelude viscous, plastic, and elastic types. Hard coal and oil
shale are strong, elastic rocks; however, the former may be fissured. Soft coal
is highly fissued and may contain hydrocarbon gases under pressure in the
pores. Tar sand may behave like a viscous liquid at high pressure or tempera-
ture; it also may contain gas under pressure.
We see that the rock family is large and "nonexelusive." Sorne of the
simple laboratory tests and measurements enumerated below will help to de-
cide what kind of material you are dealing with in any specific case.
2.2 Properties of Rock Systems
Because of the vast range in properties of rocks, which reflects varie1es of
structures, fabrics, and components, we rely on a number of basic measure-
ments to describe rocks quantitatively. Certain properties that are relatively
easy to measure are valuable in this regard and may be designated index proper-
ties for rock specimens. Porosity identifies the relative proportion of solids and
voids; density adds information about the mineralogic or grain constituents.
The sonic uelocity together with a petrographic description evaluate the degree
of fissuring. Permeability evaluates the relative interconnection of the pores;
durability indicates the tendency for eventual breakdown of components or
structures, with degradation of rock quality. FinalIy, strength determines the
present competency of the rock fabric to bind the components together. These
attributes need to be evaluated for engineering classification of rock, and to-
gether they permit one to draw useful correlations with experience for practical
applications. However, the behavior of rock specimens under changing stress,
ternperature, fluid pressure, and time includes many other facets that are not
represented by the aboye list of index properties. Therefore, characterization
of a series of indexes in the laboratory is not a substitute for careful and detailed
testing in other areas of special concern.
A list of index properties related to laboratory specimens of rock can help
classify it for applications related primarily to the behavior of the rock itself as
to the rock mass with the interactions among its system of discontinui-
A little reflection on the spectrum of applications of rock mechanics will
sorne that do involve mainly rock specimen characteristics, for example,
dnl.lability, cuttability, aggregate selection, and rip-rap evaluation. Most appli-
cahons involving excavation at the surface or underground, on the other hand,
test the system of discontinuities as much as or more than the nature ofthe rock
2S Classification and n d e ~ Propernes of Rocks
itself. In these instances, the classification of the rock mas s for engineering
purposes reftects not only laboratory tests but structural and environrnental
characteristics ofthe rock mas s in the field. We consider engineering classifica-
tion of rock mas ses later in this chapter.
2.3 Porosity
The porosity of a rock, indicated by the dirnensionless quantity n, is a fraction
expressing the proportion of void space to total space in the rock.
(2.1)
where v
p
is the volurne of pores in total volurne VI' In sedirnentary rocks,
forrned by the accurnulation of grains, rock fragrnents, or shells, the porosity
varies frorn close to O to as rnuch as 90% (n = 0.90) with 15% as a typical value
for an average sandstone. In these rocks, porosity generally decreases with
age, and with depth below the surface, other things being equal. Table 2.1
illustrates these tendencies for a nurnber of sedirnentary rocks: a typical Carn-
brian sandstone had a porosity of 11 % while a Cretaceous sandstone contained
34% pores. The effect of depth is' rnost striking in the rocks derived frorn
cornpaction
l
of clay as shown in Table 2.1. A Pennsylvanian age shale frorn
Oklahorna encountered at depth of 1000, 3000, and 5000 feet had poro sities of
16%, 7%, and 4%, respectively. Chalk is arnong the rnost porous of all rocks
with porosities in sorne instances of more than 50%. These rocks are forrned of
the hollow skeletons of rnicroscopic anirnals-coccoliths. Sorne volcanic rocks
(e.g., purnice) can also present very high porosity due to the preservation ofthe
sites of volcanic gas bubbles; in volcanic rocks, the systern of pores is not
always well connected.
In crystalline lirnestones and evaporites, and rnost igneous and rnetarnor-
phic rocks, a large proportion ofthe pore space belongs to planar cracks terrned
fissures (Figure 2.2). A relatively srnall porosity due to fissures affects the
properties of the rock to the sarne degree as a rnuch larger percentage of
subspherical pore space and, as noted in the previous chapter, creates stress
dependency in a nurnber of physical properties. In the igneous rocks, porosity
is usually less than 1 or 2% unless weathering has taken hold. As weathering
1 Compaction is a term used by geologists and petroleum engineers to describe processes by
which a sediment is densified. Soils engineers reserve this term for processes of densification
involving the expulsion of air from the voids. Consolidation refers to the expulsion of water from
the voids of a clay, in soil mechanics usage, whereas geologists and petroleum engineers use
consolidation for processes of lithification.
2.3 Porosity 29
Table 2.1 porosities of Some Typical Rocks
Showing Effects of Age and Depth
a
Rock
Age Depth Porosity (%)
Mount Simon sandstone
Cambrian 13,000 ft 0.7
Nugget sandstone (Utah)
Juras sic 1.9
potsdam sandstone
Cambrian Surface 1l.0
pottsville sandstone
Pennsylvanian 2.9
Berea sandstone
Mississippian 0-2000 ft 14.0
Keuper sandstone (England) Triassic Surface 22.0
Navajo sandstone
Juras sic Surface 15.5
Sandstone, Montana
Cretaceous Surface 34.0
Beekmantown dolomite
Ordovician 10,500 ft 0.4
Black River limestone
Ordovician Surface 0.46
Niagara dolomite
Silurian Surface 2.9
Limestone, Great Britain Carboniferous Surface 5.7
Chalk, Great Britain Cretaceous Surface 28.8
Solenhofen limestone
Surface 4.8
Salem limestone Mississippian Surface 13.2
Bedford limestone Mississippian Surface 12.0
Bermuda limestone Recent Surface 43.0
Shale Pre-Cambrian Surface 1.6
Shale, Oklahoma Pennsylvanian 1000 ft 17.0
Shale, Oklahoma Pennsylvanian 3000 ft 7.0
Shale, Oklahoma Pennsylvanian 5000 ft 4.0
Shale Cretaceous 600 ft 33.5
Shale Cretaceous 2500 ft 25.4
Shale Cretaceous 3500 ft 21.1
Shale Cretaceous 6100 ft 7.6
Mudstone, Japan Upper Tertiary Near surface 22-32
Granite, fresh Surface Oto 1
Granite, weathered 1-5
Decomposed granite (Sapro-
20.0
lyte)
Marble 0.3
Marble 1.1
Bedded tuff 40.0
Welded tuff 14.0
Cedar City tonalite
7.0
Frederick diabase 0.1
~ n Marcos gabbro
0.2
Data selected from Clark (1966) and Brace and Riley (1972).
30 Classification and Properties of Rocks
progresses, the porosity tends to increase to 20% or more. As a result, mea-
surement of porosity can serve as an accurate index to rock quality in such
rocks. In several projects in granitic rocks the National Civil Engineering Lab-
oratory of Portugal was able to classify the rock for the purposes of engineering
design mainly on the basis of a quick porosity measurement, obtained from the
water content of the rock after immersion for 24 hours at a standard tempera-
ture and pressure (Hamrol, 1961). Among unweathered rocks, there is also a
general correlation between porosity and mechanical properties such as uncon-
fined compressive strength and modulus of elasticity; but such relationships are
usually marked by enormous scatter. In the case of weak sandstones (having
saturated compressive strength less than 20 MPa) Dobereiner and de Freitas
(1986) have demonstrated good correlations of density, modulus of elasticity,
and compressive strength with the saturated moisture contento The moisture
content of a saturated specimen is linked with its porosity by Equation 2.5.
Saturation can be approached by soaking a specimen in water while it is sub-
jected to a laboratory vacuum.
Porosity can be measured in rock specimens by a variety of techniques.
Since it is the pore space that governs the quantity of oil contained in a satu-
rated petroleum reservoir, accurate methods for porosity determination in
sandstones have been developed by the oil industry. However, these methods
are not always suitable for measurements in hard rocks with porosities of less
than several percent. Porosities can be determined from the following calcula-
tions.
1. Measured density.
2. Measured water content after saturation in water.
3. Mercury content after saturationwith mercury using a pressure injector.
4. Measured solid volume and pore air volume using Boyle's law.
These are considered further below.
2.4 Density
The density or "unit weight" of a rock, ')1, is its specific weight (FL -3),2 for
example, pounds per cubic foot or kilonewtons per cubic meter. The specific
grauity of a solid, G, is the ratio between its density and the unit weight of
water ')Iw; the latter is approximately equal to 1 g-force/cm
3
(9.8 kN/m
3
or
approximately 0.01 MN/m
3
).3 Rock with a specific gravity of 2.6 has a density
2 The terms in parenthesis indicate the dimensions ofthe preceding quantity. F, L, T indicate
force, length, and time, respectively.
3 At 20C, the unit weight of water is 0.998 g/cm
3
x 980 cm/s
2
= 978 dynes/cm
3
or = 0.998
g-force/cm
3
2.4 Density 31
of approximately 26 kN 1m
3
In the system, the density of water is 62.4
ounds per cubic foot. (Mass denslty p equals ')I/g.)
p It was stated previously that the porosity of a rock can from
knowledge of its weight density. This that the speclfi.c gravlty ?f
ins or crystals is known; grain specific gravlty can be determmed by gnndmg
rock and adapting methods used in soils .. If the percentages of
different minerals can be estimated under a bmocular mlcroscope, or from a
thin section, the specific gravity of the part. a rock can then be
lated as the weighted average of the speclfic gravltles of the component grams
and crystals:
n
G = 2: GNi
(2.2)
i=l
where G
i
is the specific gravity of component i, and Vi is its volume percentage
in the solid part of the rock. The specific gravities of a number of common rock-
forming mineral s are listed in Table 2.2. The relation between porosity and dry
density ')Idry is
')Idry = GywO - n)
Table 2.2 Specific Gravities
oi Common Minerals
a
Mineral
Halite
Gypsum
Serpentine
Orthoclase
Chalcedony
Quartz
Plagioclase
Chlorite and illite
Calcite
Muscovite
Biotite
Dolomite
Anhydrite
Pyroxene
Olivine
Barite
Magnetite
Pyrite
Galena
A. N. Winchell (942).
G
2.1-2.6
2.3-2.4
2.3-2.6
2.5-2.6
2.6-2.64
2.65
2.6-2.8
2.6-3.0
2.7
2.7-3.0
2.8-3.1
2.8-3.1
2.9-3.0
3.2-3.6
3.2-3.6
4.3-4.6
4.4-5.2
4.9-5.2
7.4-7.6
(2.3)
32 Classification and n d e ~ Properties of Rocks
The dry density is related to the wet density by the relationship
'rwet
'rdry = 1 + W
where w is the water content of the rock (dry weight basis).
Water content and porosity are related by
wG
n=
1 + W G
(2.4)
(2.5)
If the pores of the rack are filled with mercury, and the mercury content is
determined to be WHg (as a proportion of the dry weight of the rock before
mercury injection), the porosity can be calculated more accurately as follows:
n = WHg' GIGHg
1 + (WHg' GIGHg)
(2.6)
The specific gravity of mercury (G
Hg
) equals 13.546.
The densities of sorne common rocks are given in Table 2.3. These figures
are only sample values, of course, since special factors can cause wide varia-
tions in individual formations.
Rocks exhibit a far greater range in density values than do soils. Knowl-
edge of rock density can be important to engineering and mining practice. For
example, the density of a rock governs the stresses it will experience when
acting as a beam spanning an underground opening; unusually high density in a
roof rock implies a shortened limiting safe span. A concrete aggregate with
higher than average density can mean a smaller volume of concrete required for
a gravity retaining wall or dam. Lighter than average aggregate can mean lower
stresses in a concrete roof structure. In oil shale deposits, the density indicates
the value of the mineral commodity because the oil yield correlates directIy
with the unit weight; this is true because oil shale is a mixture of a relatively
light constituent (keragen) and a relatively heavy constituent (dolomite). In
coal deposits, the density correlates with the ash content and with the previous
depth of cover, accordingly with the strength and elasticity of the rock. It is
easy to measure the density of a rock; simply saw off the ends of a dried drill
core, calculate its volume from the dimensions, and weight it. In view of the
possible significance of variations from the norm, density should therefore be
measured routinely in rock investigations.
2.5 Hydraulic Permeabl1ity and Conductivity
Measurement of the permeability of a rock sample may have direct bearing on a
practical problem, for example, pumping water, oil, or gas into or out of a
2.5
Hydraulic Permeability and Conductivity
Table 2.3 Dry Densities of Sorne Typical Rocks
a
Dry Dry Dry
Rock
(g/cm
3
) (kN/m
3
) (lb/ft
3
)
Nepheline syenite
2.7
26.5 169
Syenite
2.6
25.5 162
Granite
2.65 26.0 165
Diorite
2.85
27.9 178
Gabbro
3.0 29.4 187
Gypsum
2.3
22.5 144
Rock salt
2.1 20.6 131
Coal
0.7-2.0
(density varies with the ash content)
Oil shale
1.6-2.7
(density varies with the kerogen content, and
therefore with the oil yield in gallons per ton)
30 gal/ton rock
2.13
21.0 133
Dense limestone
2.7
20.9 168
Marble
2.75
27.0 172
Shale, Oklahoma
b
1000 ft depth
2.25
22.1 ' 140
3000 ft depth
2.52
24.7 157
5000 ft depth
2.62 25.7
163
Quartz, mica schist
2.82
27.6 176
Amphibolite
2.99
29.3 187
Rhyolite
2.37
23.2
148
Basalt
2.77
27.1 173
a Data from Clark (1966), Davis and De Weist (1966), and other sources.
b This is the Pennsylvanian age shale listed in Table 2.1.
33
porous formation, disposing ofbrine wastes in porous formations, storing ftuids
in mined caverns for energy conversion, assessing the water tightness of a
reservoir, dewatering a deep chamber, or predicting water inftows into a tun-
nel. In many instances the system of discontinuities will radically modify the
permeability values of the rock in the field as compared to that in the lab, so
that sorne sort of in situ pumping test will be required for an acceptable forecast
of formation permeabilities. Our motivation for selecting permeability as an
index property of rock is that it conveys information about the degree of inter-
connection between the pores or fissures-a basic part of the rock framework.
Furthermore, the variation of permeability with change in normal stress, espe-
cially as the sense ofthe stress is varied from compression to tension, evaluates
the degree of fissuring of the rock, since ftat crack s are greatly affected by
normal stress whereas spherical pores are not. Also, the degree to which the
permeability changes by changing the permeant from air to water expresses
34 Classification Properties of Rocks
interaction between the water and the mineral s or binder of the rock and can
detect subtle but fundamental flaws in the integrity of the rock; this promising
aspect of permeability as an index has not been fully researched.
Most rocks obey Darcy's law. For many applications in civil engineering
practice, which may involve water at about 20C, it is common to write Darcy's
law in the form
dh
qx = k dx A
where qx is the flow rate (VT-I) in the x direction
h is the hydraulic head with dimension L
A is the cross-sectional area normal to x (dimension L2)
(2.7)
The coefficient k is termed the hydraulic conductivity; it has dimensions of
velocity (e.g., centimeters per second or feet per minute). When temperature
will vary considerably from 20C or when other fluid s are to be considered a
more useful form of Darcy's law is '
Kdp
q=--A
x JL dx
(2.8)
in which pis the fluid pressure (equal to 'Ywh) with dimensions of FL -2 and JL is
the viscosity of the permeant with dimensions FL -2T. For water at 20C, JL =
2.098 x 10-
5
lb S/ft2 = 1.005 X 10-
3
N s/m
2
and 'Y = 62.4lb/ft3 = 9.80 kN/m3.
When Darcy's law is written this way, the coefficient K is independent of
the properties of the fluid. Its dimensions are those of area (e.g., square centi-
meters). K is termed the hydraulic permeability.
A common permeability unit is the darcy: 1 darcy equals 9.86 x 10-9 cm2.
Table 2.4 gives typical values of conductivities calculated for the properties of
water at 20C; 1 darcy corresponds approximately to a conductivity value of
10-
3
cm/s.
Permeability can be determined in the laboratory by measuring the time for
a calibrated volume of fluid to pass through the specimen when a constant air
pressure acts over the surface of the fluid. An alternative method is to generate
radial flow in a hollow cylindrical specimen, prepared by drilling a coaxial
central hole in a drill coreo When the flow is from the outer circumference
toward the center, a compressive body force is generated, whereas when the
flow is from the central hole toward the outside, a tensile body force is set up.
Consequently, rocks that owe their permeability partly to the presence of a
network of fissures demonstrate a profound difference in permeability values
according to the direction of flow. A radial permeability test was devised by
Bernaix (1969) in testing the foundation rock of the Malpasset Dam after the
failure. The permeability of the mica schist from that site varied over as much
as 50,000 times as the conditions were changed from radially outward flow with
2.5 Hydraulic Permeability and Conductivity 35
Table 2.4 Conductivities of Typical Rocks
a
Rock Lab
k (cm/s) for Rock with
Water (20C) as Permeant
Field
Sandstone
3 X 10-
3
to 8 X 10-
8
2 X 10-
3
1 X 10-
3
to 3 X 10-
8
Navajo sandstone
Berea sandstone
Greywacke
Shale
Pierre shale
Limestone, dolomite
Salem limestone
Basalt
Granite
Schist
Fissured schist
4 X 10-
5
3.2 X 10-
8
10-
9
to 5 X 10-
13
5 X 10-
12
10-
5
to 10-
13
2 X 10-
6
10-
12
10-
7
to 10-
11
10-
8
1 X 10-
4
to 3 X 10-
4
10-
8
to 10-
11
2 X 10-
9
to 5 X 10-
11
10-
3
to 10-
7
10-
2
to 10-
7
10-
4
to 10-
9
2 X 10-
7
a Data from Brace (1978), Davis and De Wiest (1966), and Serafim (1968).
t' of 1 bar, to radially inward flow with I:l.P of 50 bars. The hydraulic conduc-
tivity (velocity units) from a radial flow test can be approximated by
q In(R
2
/R
1
)
k = 27/'Ll:l.h
where q is the volume rate of flow
L is the length of the specimen
R2 and RI are the outer and inner radii of the specimen
(2.9)
I:l.h is the head difference across the flow region corresponding to I:l.P
An advantage of the radial permeability test, in addition to its capability to
distinguish flow in fissures from flow in pores, is the fact that very large flow
gradients can be generated, allowing permeability measuremejlt in the milli-
darcy region. For rocks considerably less permeable than that, for example,
granites with permeability in the region 10-
9
darcy and below, Brace et al.
(1968) devised a transient flow test.
Dense rocks like granite, basalt, schist, and crystalline limestone usually
exhibit very small permeability as laboratory specimens, yet field tests in such
rocks may show significant permeability as observed in Table 2.4. The reason
for this discrepancy is usually attributed to regular sets of open joints and
throughout the rock mass. Snow (1965) showed that it is usefuI to
ldealize the rock mass as a system of parallel smooth plates, all flow running
between the plates. When there are three mutually perpendieular sets of frac-
36 Classiflcation and Properties of Rocks
tures with parallel walls, all with identical aperture and spacing and ideally
smooth, the conductivity of the rock mass is theoretically expressed by
k =
(2.10)
where S is the spacing between fractures and e is the fracture aperture (in-
terwall separation). It is seldom feasible to calculate the rock permeability from
a description of the fractures, although Rocha and Franciss (1977) have shown
how this can be done by using oriented, continuous core samples and correct-
ing the data with results from a few pumping tests. Equation 2.10 is useful,
however, for calculating the hypothetical fracture aperture e, that gives the
same permeability value as measured in the field (corresponding to an assigned
fracture spacing S). The aperture and spacing of the fractures then pro vide
quantitative indexes of rock mass quality.
2.6 Strength
The value of having an index to rock strength is self-evident. The problem is
that strength determinations on rock usually require careful test setup and
specimen preparation, and the results are highly sensitive to the method and
style of loading. An index is useful only if the properties are reproducible from
one laboratory to another and can be measured inexpensively. Such a strength
index is now available using the point load test, described by Broch and Frank-
lin (1972). In this test, a rock is loaded between hardened steel cones, causing
failure by the development of tensile cracks parallel to the axis of loading. The
test is an outgrowth of experiments with compression of irregular pieces of
rock in which it was found that the shape and size effects were relatively small
and could be accounted for, and in which the failure was usually by induced
tension. In the Broch and Franklin apparatus, which is commercially available,
the point load strength is
P
Is = D2 (2.11)
where P is the load at rupture, and D is the distance between the point loads.
Tests are done on pieces of drill core at least 1.4 times as long as the diameter.
In practice there is a strength/size effect so a correction must be made to
reduce results to a common size. Point load strength is found to fall by a factor
of 2 to 3 as one proceeds from cores with diameter of 10 mm to diameters of 70
mm; therefore, size standardization is required. The point load index is re-
ported as the point load strength of a 50-mm coreo (Size correction charts are
2.7 Slaking and Durability 37
Table 2.5 Typical Point Load Values
a
Material Point Load Strength Index (MPa)
Tertiary sandstone and claystone
Coal
Limestone
Mudstone, shale
Volcanic fiow rocks
Dolomite
0.05-1
0.2-2
0.25-8
0.2-8
3.0-15
6.0-11
a Data from Broch and Franklin (1972) and other sources.
given by Broch and Franklin.) A cite.d correlation between point
load index and unconfined compreSSlOn strength IS
qu = 24I
s
(50) (2.12)
where qu is the unconfined compressive strength of cylinders with a length to
diameter ratio of 2 to 1, and I
s
(50) is the point load strength corrected to a
diameter of 50 mm. However, as shown in Table 3.1, this relations?ip be
severely inaccurate for weak rocks and it should be checked by specIaI cahbra-
tion studies wherever such a correlation is important in practice.
The point load strength test is quick and simple, and it can be in the
field at the site of drilling. The cores are broken but not destroye?, the
fractures produced tend to be cIean, single breaks that can be dlstmgUlshed
from preexisting fractures sampled by the drilling operation .. load
results can be shown on the drilllog, along with other geotechmcal mformatlOn,
and repetition of tests after the core has dried out the of
natural water conditions on strength. Values of the pomt load mdex are glven
for a number of typical rocks in Table 2.5.
2.7 Slaking and Durability
Durability of rocks is fundamentally important for all in
the properties of rocks are produced by exfoliation, hydratlOn, decrepItatlOn
(slaking), solution, oxidation, abrasion, and other processes. In sorne
and sorne volcanic rocks, radical deterioration in rock quality occurs r.apldly
after a new surface is uncovered. Fortunately, such changes usualIy act lmper-
ceptibly through the body of the rock and only the immediate surface. i.s de-
graded in tens of years. At any rate, sorne index to the degree of alterablhty of
rock is required. Since the paths to rock destruction devised by nature are
many and varied, no test can reproduce expectable service conditions for more
38 Classification and n d e ~ Properties of Rocks
than a few special situations. Thus an index to alteration is useful mainly in
offering a relative ranking of rock durability.
One good index test is the slake durability test proposed by Franklin and
Chandra (1972). The apparatus consists of a drum 140 mm in diameter and 100
mm long with sieve mesh forming the cylindrieal walls (2 mm opening); about
500 g of rock is broken into 10 lumps and loaded inside the drum, which is
tumed at 20 revolutions per minute in a water bath. Mter 10 min of this slow
rotation, the percentage ofrock retained inside the drum, on a dry weight basis,
is reported as the slake durability index (Id). Gamble (1971) proposed using a
second lO-min cyc1e after drying. Values ofthe slake durability index for repre-
sentative shales and c1aystones tested by Gamble varied over the whole range
from O to 100%. There was no discernible connection between durability and
geologieal age but durability increased linearly with density and inversely with
natural water content. Based upon his results, Gamble proposed a c1assification
of slake durability (Table 2.6).
Morgenstern and Eigenbrod (1974) expressed the durability of shales and
claystones in terms of the rate and amount of strength reduction resulting from
soaking. They showed that noncemented claystone or shale immersed in water
tends to absorb water and soften until it reaches its liquid limito The latter can
be determined by a standard procedure described in ASTM designation D423-
54T after disaggregating the rock by shaving it with a knife and mixing the
shavings with water in a food blender. Materials with high liquid limits are more
severely disrupted by slaking than those with low liquid limits. Classes of,
amounts of slaking were therefore defined in terms of the value of the liquid
limit as presented in Table 2.7. The rate at whieh slaking occurs is independent
of the liquid limit but can be indexed by the rate of water content change
following soaking. The rate of slaking was c1assified in terms of the change in
liquidity index (M
L
) following immersion in water for 2 h; M
L
is defined as
AT _ aw
I.UL-
WL - Wp
Table 2.6 Gamble's Slake Durability Classification
Group Name
Very high durability
High durability
Medium high durability
Medium durability
Low durability
Very low durability
% Retaioed after Ooe
lO-mio Cycle
(Dry Weight Basis)
>99
98-99
95-98
85-95
60-85
<60
(2.13)
% Retaioed after Two
lO-mio Cycles
(Dry Weight Basis)
>98
95-98
85-95
60-85
30-60
<30
2.S Sonie Velocity as an l n d e ~ lo Degree of Fissuring 39
Table 2.7 Description of Rate and Amount of Slaking
a
Amount of Slaking
Very low
Low
Medium
High
Very high
Rate of Slaking
Slow
Fast
Very fast
Liquid Limit (%)
<20
20-50
50-90
90-140
>140
Change in Liquidity Index after Soaking 2 h
<0.75
0.75-1.25
>1.25
a After Morgenstem and Eigenbrod (1974).
where a W is the change in water content of the rock or soil after soaking for 2 h
on filter paper in a funnel
wp is the water content at the plastic limit
WL is the water content at the liquid limit
AH the water contents are expressed as a percentage of the dry weight. These
indexes and procedures for determining them are described in most textbooks
on soil mechanics (e.g., Sowers and Sowers, cited in Chapter 9).
2.8 Sanie Velocity as an l n d e ~
to Degree of Fissuring
Measurement of the velocity of sound waves in a core specimen is relatively
simple and apparatus is available for this purpose. The most popular method
pulses one end of the rock with a piezoelectric crystal and receives the vibra-
tions with a second crystal at the other end. The travel time is determined by
measuring the phase difference with an oscilloscope equipped with a variable
del ay lineo It is also possible to resonate the rock with a vibrator and then
ca1culate its sonie velocity from the resonant frequency, known dimensions,
and density. Both longitudinal and transverse shear wave velocities ~ a n . be
determined. However, the index test described here requires the determlOatlOn
of only the longitudinal velocity VI, which proves the easier to measure. ASTM
Designation D2845-69 (1976) describes laboratory determination ofpulse veloc-
ities and ultrasonic elastic constants of rock.
Theoretically, the velocity with which stress waves are transmitted through
rock depends exc1usively upon their elastic properties and their density (as
explored in Chapter 6). In practice, a network of fissures in the specimen
40 Classifieation and n d e ~ Properties of Hoeks
superimposes an overriding effect. This being the case, the sonic velocity can
serve to index the degree of fissuring within rock specimens.
Fourmaintraux (1976) proposed the following procedure. First calculate the
longitudinal wave velocity (V*) that the speciment would have if it lacked
pores or fissures. If the mineral composition is known, V* can be calculated
from
_1 = C (2.14)
V* i V/,i
where VI, is the longitudinal wave velocity in mineral constituent i, which has
volume proportion C in the rock. Average velocities of longitudinal waves in
rock-forming mineral s are given in Table 2,8. Table 2,9 lists typical values of
V* for a few rock types.
Now measure the actual velocity of longitudinal waves in the rock speci-
men and form the ratio V/V*, As a quality index define
VI
IQ% = V* x 100% (2.15)
I
Experiments by Fourmaintraux established that IQ is affected by pores (spheri-
cal holes) according to
IQ% = 100 - 1.6n
p
% (2.16)
where np% is the porosity of nonfissured rock expressed as a percentage.
However, ifthere is even a small fraction offtat cracks (fissures), Equation 2.16
breaks down.
Table 2.8 Longitudinal
Velocities of Minerals
Mineral VI (mis)
Quartz 6050
Olivine 8400
Augite 7200
Amphibole 7200
Muscovite 5800
Orthoclase 5800
Plagioclase 6250
Calcite 6600
Dolomite 7500
Magnetite 7400
Gypsum 5200
Epidote 7450
Pyrite 8000
From Fourmaintraux (1976).
2.8 Sonie Velocity as an I n d e ~ to Degree of Fissuring 41
Table 2.9 Typical Values
ofV" for Rocks
Rock
Gabbro
Basalt
Limestone
Dolomite
Sandstone and quartzite
Granitic rocks
a From Fourmaintraux (1976).
vt (mis)
7000
6500-7000
6000-6500
6500-7000
6000
5500-6000
}> '/ ' f
: I
For example, a sandstone with np equals 10% had IQ equal to 84%. After
heating the rock to a high temperature that produced an additional increment of
ftat cracks amounting to 2% pore space (np = 10%, n = 12%), IQ fell to 52%.
(Heating opens grain boundary cracks in minerals with different coefficients of
thermal expansion in different directions, in this case quartz.)
Because of this extreme sensitivity of IQ to fissuring and based upon labo-
ratory measurements and microscopic observations offissures, Fourmaintraux
proposed plotting IQ versus porosity (Figure 2.3) as a basis for describing the
degree of fissuring of a rock specimen. Entering the figure with known porosity
* o
11%
Figure 2.3 Classification scheme for fissuring in rock specimens.
(After Fourmaintraux 1976.)
42 Classification and lnde" Properties of Rocks
and calculated IQ defines a point in one of the five fields: (1) nonfissured to
slightly fissured, (I!) slightly to moderately fissured, (II!) moderately to
strongly fissured, (IV) strongly to very strongly fissured, and (V) extremely
fissured. Although it would be better to determine the length, distribution, and
extent of fissures by direct microscopic techniques, this necessitates tools and
procedures that are not generally available. On the other hand, using Figure
2.3, the degree of fissuring can be appreciated and named readily and inexpen-
sively in almost any rock mechanics laboratory.
2.9 Other Physical Praperties
Many other physical properties are important to specific engineering tasks in
rock. The hardness ofrock affects drillability. Elasticity and stress-strain coeffi-
cients are basic to engineering for dams and pressure tunnels. The thermal
properties-heat conductivity and heat capacity and the coefficient of linear
expansion-affect storage of hot and cold ftuids in cavems and geothermal
energy recovery. The following chapters consider sorne of these rock specimen
attributes further. As noted previously, an overriding inftuence on rock behav-
ior in many instances stems from the characteristics of the discontinuities,
includingjoints, bedding, foliation, and fractures. This is addressed by a mean-
ingful system of rock classification that attempts to overlay index properties of
rocks and of discontinuities.
2.10 Classificatian af Rack Masses
far Engineering Purpases
It is not always convenient to make a definitive test in support of engineering
decision involving rock, and sometimes it is not even possible. Frequently,
experience and judgment are strained in trying to find answers to design deci-
sions involving rock qualities. Where there are particular and recurrent needs
for quantitative values from rock, useful index tests are used routinely as in
evaluating the need for continued grouting below a dam, deepening a pier shaft
before filling it with concrete, or establishing the thickness of shotcrete lining in
a newly excavated stretch of a rock tunnel. Thus it is not surprising that
numerous schemes have been devised to guide judgment through standardized
procedures and descriptions. Three especially well-received classification sys-
tems, originally advanced for tunneling, are those developed by Barton, Lien,
and Lunde (1974), Bieniawski (1974, 1984), and Wickham, Tiedemann, and
Skinner (1974).
2.10 Classification of Rock Masses for Engineering Purposes 43
Bieniawski's Geomechanics Classification system provides a general rock
mass rating (RMR) increasing with rock quality from O to 100. It is based upon
five universal parameters: strength of the rock, drill core quality, groundwater
conditions,joint and fracture spacing, andjoint characteristics. A sixth param-
eter, orientation of joints, is entered differently for specific application in tun-
neling, mining, and foundations. Increments ofrock mass rating corresponding
to each parameter are summed to determine RMR.
The strength of the rock can be evaluated using a laboratory compression
test on prepared core, as discussed in the next chapter. But for rock classifica-
tion purposes, it is satisfactory to determine compressive strength approxi-
mately using the point load test described previously on intact pieces of drill
coreo To simplify class boundaries, Bieniawski revised Equation 2.12 to qu =
25I
s
. The rock mass rating increment corresponding to compressive strength
values are listed in Table 2.10.
Drill core quality is rated according to the rock quality designation (RQD)
introduced by Deere (1963). Although the RQD is widely used as asole parame-
ter for classification of rock quality, it is preferable to combine it with other
parameters accounting for rock strength, joint character, and environmental
factors as done here, since the RQD alone ignores these features. The RQD of a
rock is evaluated by determining the percentage recovery of core in lengths
greater than twice its diameter. The index was first applied solely to NX core,
usually 2.125 in. in diameter, the percentage core recovery being modified to
reject from the "recovered" category any fragments less than 4 in. in length.
The rock mass rating increments corresponding to five bands of RQD values
are given in Table 2.11.
The spacing of joints is also evaluated from drill core, if available. It is
assumed that the rock mass contains three sets of joints in general and the
spacing entered in Table 2.12 to determine the rating increment should reftect
that joint set considered to be most critical for the particular application. If the
Table 2.10 Rack Mass Rating lncrements
far Campressive Strength of the Rack
Unconfined
Point Load Compressive
Index (MPa) OR Strength (MPa) Rating
>10 >250 15
4-10 100-250 12
2-4 50-100 7
1-2 25-50 4
Don't use 10-25 2
Don't use 3-10 1
Don't use <3 O
44 Classiflcation and lndex. Properties of Rocks
Table 2.11 Rock
Mass Rating
Increments for
Drill Core Quality
RQD (%) Rating
90-100 20
75-90 17
50-75 13
25-50 8
<25 3
rock mass has fewer sets ofjoints, the rating may be established more favorably
than indicated in this tableo The condition of joints is also examined with
respect to the joint sets most likely to inftuence the work. In general, the
descriptions ofjoint surface roughness and coating material should be weighted
toward the smoothest and weakestjoint set. Joint condition ratings are given in
Table 2.13. Further discussion of the inftuence of joint roughness and spacing
on the properties of rocks is presented in Chapter 5.
Groundwater can strongly inftuence rock mass behavior so the geome-
chanics elassification attempts to inelude a groundwater rating term as given in
Table 2.14. If an exploratory adit or pilot tunnel is available, measurements of
water inftows or joint water pressures may be used to determine the rating
increment directly. The drill core and drilling log can be used in lieu of such
Table 2.12
Increments of
Rock Mass Rating
for Spacing of
Joints of Most
Injluential Set
Joint
Spacing
(m) Rating
>2.0 20
0.6-2.0 15
0.2-.6 10
0.06-
0.2 8
<0.06 5
2.10 Classiflcation of Rock Masses for Engineenng Purposes 45
Table 2.13 Rock Mass Rating
Increments for Joint Condition
Description
Very rough surfaces of limited
extent; hard wall rock
Slightly rough surfaces; aperture
less than 1 mm; hard wall rock
Slightly rough surfaces; aperture
less than 1 mm; soft wall rock
Smooth surfaces, OR gouge filling
1-5 mm thick, OR aperture of
1-5 mm; joints extend more than
several meters
Open joints filled with more than
5 mm of gouge, OR open more
than 5 mm; joints extend more
than several meters
Rating
30
25
20
10
o
information to assign the rock to one of four categories from which the rating
increment is assigned-completely dry, moist, water under moderate pressure,
or severe water problems.
Since the orientation ofthejoints relative to the work canhave an inftuence
on the behavior of the rock, Bieniawski recommended adjusting the sum of the
first five rating numbers to account for favorable or unfavorable orientations,
according to Table 2.15. No points are subtracted for very favorable orienta-
tions of joints, up to 12 points are deducted for unfavorable orientations of
joints in tunnels, and up to 25 for unfavorable orientations in foundations. It is
difficult to apply these corrections by universal charts because a given orienta-
Table 2.14 Increments of Rock Mass Rating
Due to Groundwater Condition
Joint Water
Inftow per 10 m Pressure Divided
Tunnel Length OR by, Major Principal OR
(Llmin) Stress
None O
<10 0.0-0.1
10-25 0.1-0.2
25-125 0.2-0.5
>125 >0.5
General
Condition
Completely dry
Damp
Wet
Dripping
Flowing
Rating
15
10
7
4
O
46 Classification and Properties of Rocks
Table 2.15 Adjustment in RMR for Joint Orientations
Assessment of
Inftuence of Orientation
on the Work
Rating Increment
for Tunnels
Rating Increment
for Foundations
Very favorable
Favorable
Fair
Unfavorable
Very unfavorable
o
-2
-5
-10
-12
o
-2
-7
-15
-25
tion may be favorable or unfavorable depending upon the groundwater and
joint conditions. Thus, applying Table 2.14 requires advice from an engineering
geologist familiar with the particular rock formations and the works in ques-
tion. The orientation of joint sets cannot be found from normal, routine drilling
of rock mas ses but can be determined from drill core with special tools or
procedures, as reviewed by Goodman (1976) (work cited in Chapter 1). Logging
of the borehole using a television or camera downhole will reveal orientations
ofjoints, and absolute orientations will also be obtained from logging shafts and
adits.
Por applications in mining, involving assessments of caveability, drillabil-
ity, blasting, and supports, Laubscher and Taylor (1976) modified Tables 2.10
to 2.15 and introduced factors to adjust for blasting practice, rock stress, and
weathering. They also presented atable to find joint spacing ratings given the
separate spacings of alljoint sets. The overall RMR rating of a rock mass places
the rock in one ofthe five categories defined in Table 2.16. Specific applications
of the rock mass rating are presented in later chapters.
Table 2.16 Geomechanics Classlflcation
of Rack Masses
Description
Class of Rock Mass
1 Very good rock
11 Good rock
III Fair rock
IV Poor rock
V Very poor rock
RMR
Sum of Rating Increments
from Tables 2.9-2.14
81-100
61-80
41-60
21-40
0-20
2.10 Classification of Rock Masses for Engineering Purposes 47
The Q system by Barton, Lien, and Lunde (1974) (also called the NGI
system) combines six parameters in a multiplicative function:
Q = (RQD/J
n
) x (J,/Ja) x (JjSRF) (2.16)
where RQD is the Rock Quality Designation
J
n
relates to the number of joint sets
J relates to the roughness of the most important joints
a relates to the wall rock condition and/or filling material
J
w
relates to the water flow characteristics of the rock
SRF relates to looseness and stress conditions.
The first term of Equation 2.16 is a measure of the sizes of joint blocks, the
second factor expresses the shear strength of the block surfaces, and the last
factor evaluates the important environmental conditions influencing the behav-
ior ofthe rock mass. Numerical values are assigned to each parameter ofthe Q
system according to detailed descriptions to be found in Barton et
al., which are abbreviated in Table 2.17. Table 2.18 asslgns quahtattve elasses
to the rock according to the overall value of Q.
The Q system and the RMR system inelude
and therefore cannot be strictIy correlated. EquatlOn 2.17 IS an approxlmate
connecting relationship proposed by Bianiawski, based upon a study of a large
number of case histories (standard deviation = 9.4).
RMR = 9 log Q + 44
Table 2.17 Values of the Parameters
in the Q System
Number 01 SeIs 01 Disconlinuities In
Massive 0.5
One set 2.0
Two sets 4.0
Three sets 9.0
Four or more sets 15.0
Crushed rock 20.0
Roughness 01 Discontinuities 1/
Noncontinuous joints 4.0
Rough, wavy 3.0
Smooth, wavy 2.0
Rough, planar 1.5
Smooth, planar 1.0
Slick, planar 0.5
"Filled" discontinuities 1.0
* Add 1.0 if mean joint spacing
exceeds 3 m
(2.17}
4S Classification and Properties of Rocks
FiIling and Wall Rock Alteration
Essentially unfilled
Healed
Staining only; no alteration
Silty or sandy coatings
Clay coatings
Filled
Sand or crushed rock filling
Stiff clay filling <5 mm thick
Soft clay filling <5 mm thick
Swelling clay filling <5 mm thick
Stiff clay filling >5 mm thick
Soft clay filling >5 mm thick
Swelling clay filling >5 mm thick
Water Conditions
Dry
Medium water inflow
Large inflow with unfilled joints
Large inflow with filled joints that
wash out
High transient inflow
High continuous inflow
0.75
1.0
3.0
4.0
4.0
6.0
8.0
12.0
10.0
15.0
20.0
Jw
1.0
0.66
0.5
0.33
0.2-0.1
0.1-0.05
Stress Reduction Class SRF*
Loose rock with clay-filled
discontinuities 10.0
Loose rock with open discontinuities 5.0
Rock at shallow depth 50 m) with
clay-filled discontinuities 2.5
Rock with tight, unfilled discontinuities
under medium stress 1.0
* Barton et al. also define SRP values corresponding
to degrees of bursting, squeezing, and swelling rock
conditions.
The use of engineering classification systems for rock is still somewhat contro-
versial. Proponents point to the opportunities they offer for empiricism in
design of tunnels, mines, and other works in rock. an attempt to
fill out the tables of values required by these schemes dlSClphnes the observer
and produces a careful, thorough scrutiny of mass. ?n the other hand,
these classifications tend to promote generallzatlons tbat m sorne cases are
Table 2.18 After Barton, Lien,
and Lunde (1974)
References 49
Q Rock Mass Quality for Tunneling
<0.01
0.01- 0.1
0.1 - 1.0
1.0 - 4.0
4.0 - 10.0
10.0 - 40.0
40.0 -100.0
100.0 -400.0
>400.0
Exceptionally poor
Extremely poor
Very poor
Poor
Pair
Good
Very good
Extremely good
Exceptionally good
inadequate to describe the full range of specifics of real rocks. Whichever
argument prevails in a particular case, there can be no doubt that classification
systems are proving valuable to many in various aspects of applied rock me-
chanics.
References
Aastrup, A. and Sallstrom, S. (1964) Purther Treatment of Problematic Rock Pounda-
tions at Bergeforsen Dam. Proc. Eighth Congo on Large Dams, Edinburgh, p. 627.
Barton, N. (1976) Recent experiences with the Q-system of tunnel support design,
Proceedings of Symposium on Explorationfor Rock Engineering (Balkema, Rotter-
dam), Vol. 1, pp. 107-118.
Barton, N., Lien, R., and Lunde, J. (1974) Engineering classification ofrock masses for
the design of tunnel support, Rock Mech. 6: 189-236.
Bernaix, J. (1969) New Laboratory methods of studying the mechanical properties of
rock, lnt. J. Rock Mech. Min. Sci. 6: 43-90.
Bieniawski, Z. T. (1974) Geomechanics classification ofrock masses and its application
in tunneling, Proc. 3rd Congo ISRM (Denver), Vol. 2A, p. 27.
Bieniawski, Z. T. (1976) Rock mas s classifications in rock engineering, Proceedings of
Symposium on Exploration for Rock Engineering (Balkema, Rotterdam), Vol. 1,
pp. 97-106.
Bieniawski, Z. T. (1984) Rock Mechanics Design in Mining and Tunneling, Balkema,
Rotterdam.
Brace, W. P. and Riley, D. K. (1972) Static uniaxial deformation of 15 rocks to 30 kb,
lnt. J. Rock Mech. Mining Sci. 9: 271-288.
Brace, W. P., Walsh, J. B., and Prangos, W. T. (1968) Permeability of granite under
high pressure, J. Geoph. Res. 73: 2225-2236.
50 Classification and n d e ~ Properties of Rocks
Broch, E. and Franklin, J. A. (1972) The point load strength test, Int. J. Rock Mech.
Mining Sci. 9: 669-697.
Clark, S. P. (Ed.) (1966) Handbook of Physical Constants, Geological Society of Amer-
ica, Memoir 97.
Daly, R. A., Manger, G.I., and Clark, S. P., Jr. (1966) Density ofrocks. In S. P. Clark,
Ed., Handbook of Physical Constants, rey. ed., Geological Society of America,
Memoir 97, pp. 19-26.
Davis, S. N. and DeWiest, R. J. M. (1966) Hydrogeology, Wiley, New York.
Deere, D. U. (1963) Technical description of rock cores for engineering purposes, Rock
Mech. Eng. Geol. 1: 18.
Dobereiner, L. and de Freitas, M. H. (1986) Geotechnical properties of weak sand-
stones, Geotechnique 36: 79-94.
Fourmaintraux, D. (1976) Characterization of rocks; laboratory tests, Chapter IV in La
Mcanique des roches applique aux ouvrages du gnie civil by Marc Panet et al.
Ecole Nationale des Ponts et Chausses, Paris.
Franklin, J. A. and Chandra, R. (1972) The slake durability index, Int. J. Rock Mech.
Min. Sci. 9: 325-342.
Franklin, J. A., Vogler, U. W., Szlavin, J., Edmond, J. M., and Bieniawski, Z. T. (1979)
Suggested methods for determining water content, porosity, density, absorption
and related properties and swelling and slake durability index properties for ISRM
Commission on Standardization of Laboratory and Field Tests, Int. J. Rock Mech.
Min. Sci. 16: 141-156.
Gamble, J. C. (1971) Durability-plasticity classification of shales and other argillaceous
rocks, Ph. D. thesis, University of Illinois.
Hamrol, A. (1961) A quantitative classification of the weathering and weatherability of
rocks, Proceedings, 5th International Conference on Soil Mechanics and Founda-
tion Engineering (Paris), Vol. 2, p. 771.
Kulhawy, F. (1975) Stress deformation properties ofrock and rock discontinuities, Eng.
Geol. 9: 327-350.
Laubscher, D. H. and Taylor, H. W. (1976) The importance of geomechanics classifica-
tion of jointed rock masses in mining operations, Proceedings of Symposium on
Explorationfor Rock Engineering (Johannesburg), Vol. 1, pp. 119-135.
Morgenstern, N. R. and Eigenbrod, K. D. (1974) Classification ofargillaceous soils and
rocks, J. Geotech. Eng. Div. (ASCE) 100 (GT 10): 1137-1158.
Mller-Salzburg, L. (1963, 1978) Der Felsbau, Vols. 1 and 3, (In German), Ferdinand-
Enke, Stuttgart.
Nakano, R. (1979) Geotechnical properties of mudstone of Neogene Tertiary in Japan,
Proceedings of International Symposium on Soil Mechanics in Perspective (Oax-
aca, Mexico), March, Session 2 (International Society of Soil Mechanics and Foun-
dation Engineering).
Rocha, M. and Franciss, F. (1977) Determination of permeability in anisotropic rock
mas ses from integral samples, Rock Mech. 9: 67-94.
Rummel, F. and Van Heerden, W. L. (1978) Suggested methods for determining sound
velocity, for ISRM Commission on Standardization of Laboratory and Field Tests,
Int. J. Rock Mech. Min. Sci. 15: 53-58. .
Rzhevsky, V. and Novik, G. (1971) The Physics of Rocks, Mlr, Moscow.
Problems 51
Snow, D. T. (1965) A parallel plate model offractured permeable media, Ph.D. thesis,
University of California, Berkeley.
Snow, D. T. (1968) Rock fracture spacings, openings, and porosities, J. Soil Mech.
Foundations Div. (ASCE) 94 (SM 1): 73-92.
Techter, D. and Olsen, E. (1970) Stereogram Books of Rocks, Minerals & Gems,
Hubbard, Scientific. Northbrook, IL.
Underwood, L. B. (1967) Classification and identification of shales, J. Soil Mech. Foun-
dations Div. (ASCE) 93 (SM 6): 97-116.
Wickham, G. E., Tiedemann, H. R., and Skinner, E. H. (1974) Ground support predic-
tion model-RSR concept, Proc. 2nd RETC Con! (AIME), pp. 691-707.
Winchell, A. N. (1942) Elements of Mineralogy, Prentice-Hall, Englewood Cliffs, NJ.
Problems
1. A shale of Cretaceous age is composed of 60% illite, 20% chlorite, and 20%
pyrite. The porosity values at different depths are as follows: n equals
33.5% at 600 ft; 25.4% at 2500 ft; 21.1% at 3500 ft, and 9.6% at 6100 ft.
Estimate the vertical stress at 6000 ft depth in this shale (assuming a contin-
uous thickness of shale from the surface to depth 6000 ft and saturation with
water).
2. Three samples of rock were subjected to diametral point load tests. The
pressure gage readings at rupture were 250, 700, and 1800 psi. If the ram
area was 2.07 in.
2
, and the diameter of the cores tested was 54 mm, ca1cu-
late an estimate for the unconfined compressive strength of each rock.
(Ignore a size correction.)
3. A sandstone core composed of quartz and feldspar grains with ca1cite ce-
ment is 82 mm in diameter and 169 mm long. On saturation in water, its wet
weight is 21.42 N; after oven drying its weight is 20.31 N. Ca1culate its wet
unit weight, its dry unit weight, and its porosity.
4. Another core specimen from the same formation as the rock of Problem 3,
displays large voids. Its wet unit weight is 128 Ib/ft
3
Assuming its specific
gravityis the same as for the rock in Problem 3, estimate its porosity.
5. A granitic rock is composed of a mixture of 30% quartz, 40% plagioclase,
and 30% augite. Its porosity is 3.0% and its longitudinal wave velocity
measured in the laboratory is 3200 mis. Describe its state of fissuring.
6. A sandstone with porosity of 15% is composed of a mixture of 70% quartz
grains and 30% pyrite grains. Determine its dry density in pounds per cubic
foot and meganewtons per cubic meter.
52 Classification and n d e ~ Properties of Rocks
7. Determine the water content of the aboye rock when it is saturated with
water.
8. A rock is injected with mercury by subjecting it to a high pressure. Derive a
formula expressing its porosity in terms of the measured mercury content,
the specific gravity of mercury, and the specific gravity of the component
minerals.
9. If a rock has a permeability of 1 millidarcy, how much water will flow
through it per unit of time and area under a gradient of unity? (The water
temperature is 20C.)
10. What will be the vertical stress in the ground at a depth of 5000 ft in the
Pennsylvanian age shale whose porosity is given in Table 2.1 and whose
density is given in Table 2.3 (Oklahoma shale). (Integrate the varying den-
sity depth relation.) Express your answer in psi and MPa.
11. A rock mass has field conductivity of 10-
5
cm/s. Assuming the rock itself is
impervious and three orthogonal sets of smooth fractures recur with spac-
ing 1 m, calculate the aperture (e) of the fractures.
12. Derive a formula expressing the conductivity k (cm/s) of a rock mass with
orthogonal fractures characterized by identical spacing S and aperture e if
the fractures are filled with soil having permeability k (cm/s).
13. A moist rock mass is characterized by the following parameters: joint water
pressure is nil; the point load index = 3 MPa; thejoint spacing = 0.5 m; and
RQD = 55%. Prepare atable of rock mass rating versus joint condition
using the terminology of Table 2.16 for the former and Table 2.13 for the
latter.
14. An orthogonally jointed rock mass has a field permeability of 55.0 darcies.
The mean joint spacing is 0.50 m. Calculate the corresponding average
aperture of the fractures.
15. A frequently used estimator of rock mass hydraulic conductivity is the
waterloss coefficient (C) determined with "pump-in" tests. A section of an
exploratory borehole is isolated by packers, and the pressure is brought to
an elevated level (ll.p) aboye the initial water pressure in the middle of the
test section, while the flow rate (q) into the hole is monitored. For steady
state flow, arate of water loss of 1 "lugeon" corresponds to q = 1 L/min
per meter length of the test section at a pressure difference (ll.p) of 10
atmospheres (= 1 MPa) applied at the test section. How many lugeons of
water loss corresponds to a flow of 4.0 gal/min in a lO-foot-long test section
under a differential pressure (ll.p) of 55 psi?
16. A rock mass has initial unit weight equal to 'Y and, after loosening, it
Problems 53
assumes unit weight 'YI. A coefficient ofloosening (n) was defined by Mller
(1978) as
'Y - 'YI
n=---
'Y
(a) A jointed sedimentary rock mass assumes a value of n = 0.35 after
loosening and 0.08 after recompaction. Calculate the corresponding val-
ues of 'YI. ('Y = 27 kN/m
3
)
(b) Crystalline igneous rocks like granite, gneiss, and diabase have a range
of values of n = 0.35 to 0.50 after loosening and 0.08 to 0.25 after
recompaction. Compute the corresponding values of the unit weight
('Y).
Rock Strength and
Failure Criteria
Chapter 3
Whenever we place an engineered structure against rock, we ask the following
two questions: Will the stresses in the rock reach the maximum levels that are
tolerable, with consequent local or gross rock failure? Will the displacements
of the rock under the loads to be applied produce such large strains in the
structure that they cause its damage or destruction? This chapter discusses the
first question. Assuming that we can estimate the initial stresses in the rock
mass and that we can predict how these stresses will be modified by the con-
struction and operation of the engineering work, how may we discover if the
rock will flow, yield, crush, crack, buckle, or otherwise give way in service?
For this we utilize "criteria of failure" -equations that link the limiting combi- -
nations of stress components separating acceptable from inadmissible condi-
tions. Before we can propose meaningful criteria, however, we should examine
how rocks usually fail, that is, whether in bending, shearing, crushing, or
otherwise.
3.1 Modes of Rock FaHure
The varieties of load configurations in practice are such that no single mode of
rock failure predominates. In fact, flexure, shear, tension, and compression can
each prove most critical in particular instances. Flexure refers to failure by
with develo)'ment and propagation oftensile cracks. This may tend to
Occur lO the layersAibove a mine roof (Figure 3.ta). As the "immediate roof'
detaches from the rock aboye, under gravity, a gap forms and a beam of rock
downward under its own weight. As the beam begins to crack, its neutral
axiS advances upward; eventually, the cracks extend right through the beam,
after which sections of rocks may come loose and fallo Flexural failure can also
56 Rock Strength and Failure Criteria
(a)
----- - --- ----------
-------- ------------------------
(d)
(b)
Figure 3.1 of failure involving breakage of rock. (a) Flexure.
(b) (e) Crushmg and tensde cracking, followed by shear. (d and e) Di-
rect tenSlOn.
occur in rock slopes with steeply dipping layers as the layers overturn toward
the free space ("toppling failure").
Shear failure refers to formation of a surface of rupture where the shear
stresses have critical, folIowed by release of the shear stress as the
suffers a dlsplacement along the rupture surface. This is common in slopes
cut In weak, soil-like rocks such as weathered clay shales and crushed rack f
fault.zones. It may occur i.n a mine with stiff ore and a softer, weaker roof
floor, the shear stresses In the roof or pillar base can allow the pillar to
3.1 Modes of Rock Failure 57
"punch" relatively upward into the roof (Figure 3.1h) or downward into the
floor. Rock cutters employing "drag bits" or "picks" owe their cutting action
partIy to shear along fractures caused by compression under the edge of the bit
(Figure 3.1e). The vibration of such cutters as they advance reflects the peri-
odic formation and removal of rock chips.
Direet tension is occasionally set up in rock layers resting on convex up-
ward slope surfaces (e.g., in sheeted granites (Figure 3.1d)) and in sedimentary
rocks on the flank of an anticIine. The base of the slope has layers incIined more
steeply than friction will allow and the balance of support for the weight of the
layers is the tensile pull from the stable part of the slope aboye. Direct tension
also is the mechanism of failure in rock slopes with nonconnected, short joint
planes; the formation of tension cracks severs the rock bridge s and allows a
complete block of rock to translate downward en masse (Figure 3.1e). When
rock breaks in tension, the surface of rupture is rather rough and free from
crushed rock particles and fragments. With shear failure, on the contrary, the
surface of failure is slick and there is much powder from crushing and com-
munition of rock. Direct tension failure also occurs when the circumference of
a borehole or a tunnel is stretched owing to internal water or gas pressure. The
former situation arises when a pressure tunnel is operated at excessive pressure
and when a drill hole is "hydraulically fractured" by pumping water to a high
pressure in a sectionisolated by "packers." Detonation of an explosive agent
in a borehole will raise gas pressure against the wall to millions of pounds per
square inch; tensile failure then creates a series of radial cracks beyond the
immediate periphery of the borehole, which may be crushed or in extreme
cases actually melted. Sorne extension joints in bedrock are believed to have
arisen from circumferential strain accompanying large amounts of uplift over
broad geographic belts ("epeirogeny").
Crushing or eompression failure occurs in intensely shortened volumes or
rock penetrated by a stiff puncho Examination of processes of crushing shows it
to be a highly complex mode, incIuding formation of tensile cracks and their
growth and interaction through flexure and shear. When the particIes and sliv-
ers formed by cracking are not free to move away from the zone of compres-
sion, they become finely comminuted. This happens under sorne drill bits and
under disk cutters of boring machines. In a mine pillar, overextraction of ore
can lead to pillar failure by splitting and shear, although the destruction of the
load-carrying capacity of the pillar through growth and coalescence of cracks is
sometimes spoken of as "compression failure."
It may be appreciated that the actual destruction of a load-carrying rock
mass is rather complex and involves one or more of the modes mentioned. It is
no wonder then that no single method of testing rock has been advanced to the
exclusion of others. In fact, the theory of failure makes use of a variety of
laboratory and field testing techniques adapted to the special nature of the
problem at hand.
58 Rock Strength and Failure Critera
3.2 ~ o m m o n Laboratory Strength Tests
To characterize the strength of rock specimens, unconfined and confined com-
pression tests, shear tests, and direct and indirect tension tests are used widely.
Other test configurations are preferred for special applications and a great
variety of procedures has been investigated. We review here the important
features of the most widely used tests-unconfined compression, triaxial com-
pression, splitting tension ("Brazilian tests"), beam bending, and ring shear.
Figure 3.2 shows rock preparation equipment required to prepare specimens
for such tests.
Figure 3 . .2 Equipment for preparing rock specimens for
laboratory tests. (a) A drill press modified for feed
under constant pressure and equipped with a vise to
retain arbitrary block s during drilling. (The drill press
was devised by Quentin Gorton.)
Figure 3 .2 Equipment for preparing rock speci-
mens for laboratory tests. (b) A diamond saw. (e)
A surface grinder adapted from a milling machine
by adding a diamond wheel and water bath.
60 Rack Strength and Failure Criteria
Unconfined compression (Figure 3.3a) is the most frequently used strength
test for rocks, yet it is not simple to perform properly and results can vary by a
factor of more than two as procedures are varied. The test specimen should be
a rock cylinder of length-to-width ratio in the range 2 to 2.5 with flat, smooth,
and parallel ends cut perpendicularly to the cylinder axis. Procedures are rec-
ommended in ASTM designation D2938-71a and by Bieniawski and Bernede
(1979). Capping of the ends with sulfur or plaster to specified smoothness is
thought to introduce artificial end restraints that overly strengthen the rock.
However, introduction of Teflon pads to reduce friction between the ends and
the loading surfaces can cause outward extrusion force s producing a premature
splitting failure, especially in the harder rocks. When mine pillars are studied, it
is sometimes preferable to machine the compression specimen from a large
cylinder to achieve loading through rock of the upper and lower regions into the
more slender central region. In the standard laboratory compression test, how-
ever, cores obtained during site exploration are usually trimmed and com-
pressed between the crosshead and platen of a testing machine. The compres-
--
(a)
(b)
(e)
(e) (dI
Figure 3.3 Common laboratory tests for characterizing rock strength criteria. (a)
Unconfined compression. (b) Triaxial compression. (e) Splitting tension (Brazilian).
(d) Four-point flexure. (e) Ring shear.
3.2 Common Laboratory Strength Tests 61
sive strength qu is expressed as the ratio of peak load P to initial cross-sectional
area A:
P
qu = A
Representative values of qu are listed in Table 3.l.
(3.1)
Triaxial compression (Figure 3.3b) refers to a test with simultaneous com-
pression of a rock cylinder and application ofaxisymmetric confining pressure.
Recommended procedures are described in ASTM designation D2664-67 (1974)
and in an ISRM Committee report by Vogler and Kvari (1978).
Table 3.1 Unconfined Compressive Strength (qu) and Ratio ofCompressive
to lndirect Tensile Strength (quiTo) for Specimens of Representative Rocks
qu
Description
a
MPa psi qjTob Reference
c
Berea sandstone 73.8 10,700 63.0 5
Navajo sandstone 214.0 31,030 26.3 5
Tensleep sandstone 72.4 10,500 1
Hackensack siltstone 122.7 17,800 41.5 5
Monticello Dam s.s. (greywacke) 79.3 11,500 4
Solenhofen limestone 245.0 35,500 61.3 5
Bedford limestone 51.0 7,400 32.3 5
Tavemalle limes tone 97.9 14,200 25.0 5
Oneota dolomite 86.9 12,600 19.7 5
Lockport dolomite 90.3 13,100 29.8 5
Flaming Gorge shale 35.2 5,100 167.6 3
Micaceous shale 75.2 10,900 36.3 2
Dworshak Dam gneiss
45 to foliation 162.0 23,500 23.5 5
Quartz mica schist J.. schistocity 55.2 8,000 100.4 5
Baraboo quartzite 320.0 46,400 29.1 5
Taconic marble 62.0 8,990 53.0 5
Cherokee marble 66.9 9,700 37.4 5
Nevada Test Site granite 141.1 20,500 12.1 7
Pikes Peak granite 226.0 32,800 19.0 5
Cedar City tonalite 101.5 14,700 15.9 6
Palisades diabase 241.0 34,950 21.1 5
Nevada Test Site basalt 148.0 21,500 11.3 7
John Day basalt 355.0 51,500 24.5 5
Nevada Test Site tuff 11.3 1,650 10.0 7
a Description of rocks listed in Table 3.1:
Berea sandstone, from Amherst, Ohio; fine grained, slightly porous; cemented. Navajo sand-
stone, from Olen Canyon Dam site, Arizona; friable, fine to medium grained. (Both sandstones are
Table Footnote (continued)
62 Rack Strength and Failure Criteria
predominately composed of quartz grains.) Tens/eep sandstone, Pennsylvanian-age sandstone
from Alcova Powerhouse, Wyoming, (near Casper); calcite cemented; medium grained. Hacken-
sack siltstone, New Jersey; from Triassic Newark Series; cemented with hematite; argillaceous.
Monticello Dam greywacke, Cretaceous sandstone from the Monticello dam foundation, Califor-
nia; medium to coarse grained, cemented feldspar, quartz, and other components; sorne feldspars
altered to mica. So/enhofen limestone, from Bavaria; very fine, interlocked crystalline texture.
Bedford limestone, Indiana; slightly porous, oolitic, bioclastic limestone. Taverna/le limestone,
from Carthage, Missouri; fine grained, cemented and interlocked crystalline limestone with fossils.
Oneota d%mite, Kasota, Minnesota; fine-grained interlocking granular texture with mottled ap-
pearance due to disseminated calcite veins. Lockport d%mite, Niagara Falls, New York; very
fine-grained cemented granular texture grading to interlocking crystalline texture; sorne anhydrite
grains. F/aming Gorge sha/e, from Flaming Gorge damsite, Utah, Wyoming border. Micaceous
sha/e, from the Jonathan mine, Ohio; the clay mineral is kaolinite. Dworshak dam gneiss, from
Orofino, Idaho; fine to medium-grained granodiorite gneis s with prominent foliation. Quartz mica
schist with crenulated schistocity; origin unknown. Baraboo quartzite, from Wisconsin; fine-
grained, brittle, massive Pre-Cambrian quartzite with tightly interlocking crystalline texture. Ta-
conic white marb/e, Rutland, Vermont; uniform, fine-grained massive marble, with sugary texture.
Cherokee marb/e, from Tate, Georgia; medium- to coarse-grained massive marble with tightly
interlocking crystalline texture. Nevada Test Site "granite," granodiorite from Piledriver Experi-
ment; coarse-grained. Pikes Peak granite, Colorado Springs, Colorado; fine- to medium-grained
dense; interlocked crystalline texture. Cedar City tonalite, somewhat weathered quartz monzonite,
with porosity of 4.9%, from Cedar City, Utah. Pa/asades diabase, from West Nyack, New York;
medium-grained. Nevada Test Site basa/t, from Buckboard Mesa; fine, olivine basalto John Day
basa/t, from John Day dam site, Arlington, Oregon. Nevada Test Site tuff, from "Red Hot"
experiment; welded volcanic ash; porosity 19.8%.
b Tensile strengths were determined by point load tests for all entries corresponding to reference 5;
determined by Brazilian test for entries corresponding to references 6 and 7. The point load tensile
strength To in megapascals was calculated from the load at failure (F), in meganewtons for point
loading across the rock core diameter (d), in meters; T
o
= 6.62 10-
3
F/d
2
(Reichmuth, 1963).
e References for Table 3.1:
Genera/
Kulhawy, F. (1975) cited in references at the end of this chapter.
Lama, R. D. and Vutukuri, V. S., cited in references in Chapter 1.
Specific
1. Balmer, G. G. (1953) Physical properties of sorne typical foundation rocks, U. S. Bureau of
Reclamation Concrete Lab Report SP-39.
2. Blair, B. E. (1956) Physical properties of mine rock, Part IV, U. S. Bureau of Mines Rep. Inv.
5244.
3. Brandon, T. R. (1974) Rock mechanic properties of typical foundation rocks, U. S. Bureau of
Reclamation Rep. REC-ERC 74-10.
4. Judd, W. R. (1969) Statistical methods to compile and correlate rock properties, Purdue Univer-
sity, Department of Civil Engineering.
5. Miller, R. P. (1965) Engineering classification and index properties for intact rock, Ph.D. Thesis,
University of Illinois.
6. Saucier, K. L. (1969) Properties of Cedar City tonalite, U. S. Army Corps of Engineers, WES
Mise. Paper C-69-9.
7. Stowe, R. L. (1969) Strength and deformation properties of granite, basalt, Jimestone, and tuff,
U. S. Army Corps of Engineers, WES Mise. Paper C-69-1.
3.2 Camman Labaratary Strength Tests 63
At the peak load, the stress conditions are (TI = P/A and (T3 = p, where P is
the highest load supportable parallel to the cylinder axis, and p is the pressure
in the confining medium. The confinement effect, that is, the strengthening of
the rock by the application of confining pressure p, is realized only if the rock is
enclosed in an impervious jacket. The confining fluid is normally hydraulic oil
and the jacket is oil-resistant rubber (e.g., polyurethane); for tests of short
duration, bicycle inner tube is suitable. Most rocks show a considerable
strengthening effect due to confining pressure and it has become routine to
conduct triaxial compression tests on rocks.
Many varieties of triaxial cells are in use in rock mechanics laboratories
and several types are available from commercial suppliers. Figure 3.4a shows
two cells used at the University ofCalifornia, Berkeley. The one on the left was
designed by Owen Olsen for the U. S. Bureau of Reclamation. It provides extra
room for inserting instruments and gages and is easily adapted for pore pres-
sure and other special measurements; however, the diameter of the piston is
considerably larger than the diameter of the specimen, with the result that a
large uplift force from the confining pressure must be reacted by the axial
loading machine. The chamber on the right, based on a design by Fritz
mel, avoids this problem. The rock specimen, with strain gages attached, wlll
be jacketed before insertion in the triaxial 3.4b shows a high-
pressure, high-temperature triaxial test facility at the TerraTek
Salt Lake City, Utah. This computer-controlled apparatus can supply confinmg ,
Figure 3.4 Equipment for triaxial compression tests. (a) Two types of cells used
at Berkeley.
64 Rack Strength and Failure Criteria
Figure 3.4 Equipment for triaxial compression tests. (h) A
high-pressure, high-temperature facility at TerraTek Salt Lake
City, Utah. '
pressures to 200 MPa to specimens as large as 10 cm in diameter at tempera-
tures as hIgh as 200C (5-cm-diameter specimens can be heated up to 5350C).
The usual for conducting a triaxial compression test is first to
apply the pressure aH round the cylinder (i.e., (TI = (T3 = p) and then
to apply load (TI - P as the lateral pressure is held constant. In this
c.ase, the experiment can be interpreted as the superposi-
tlOn of a umaxIaI compresslOn test on an initial state of all-round compression.
However, the actual path of loading in service may be quite different; since
s?me rocks demonstrate strong path effects it may then be desirable to follow
procedures. For example, the stresses in the rock at the front of a
travehng plan e wave are applied simultaneously in all directions. With com-
3.2 Common Laboratory Strength Tests 65
puter or manual feedback control, it is possible tp foHow almost any prescribed
path of loading, a1though, as will be shown later, not aH paths can result in
fracture under load. For the best results and a clear interpretation of the effects
of load, both the axial shortening, and the lateral expansion of the specimen
should be monitored during loading as discussed later.
The Brazilian test, described for cylindrical concrete specimens in ASTM
designation C496-71, I is convenient for gaining an estimate of the tensile
strength ofrock. It has been found that a rock core about as long as its diameter
will split along the diameter and paraHel to the cylinder axis when loaded on its
side in a compression machine (Figure 3.2c). The reason for this can be demon-
strated by examining the stress inside a disk loaded at opposite sides of a
diametral planeo In such a configuration the horizontal stresses perpendicular
to the loaded diameter are uniform and tensile with magnitude
2P
(Tt,B = 'TT'dt
(3.2)
where P is the compression load, d is the cylinder diameter, and t is the
thickness of the disk (the length of the cylinder). It is much easier to perform
this type of test than to arrange the precise alignment and end preparation
required for a direct tensile test.
The "Brazilian tensile strength" is estimated from the test result by report-
ing the value of (Tt.B corresponding to the peak compression load. It should be
understood, however, that the actual cause offailure may also reflect the action
of the vertical stress along the vertical diameter in concertwith the horizontal
tension; the vertical stress is nonuniform increasing from a compressive stress
of three times (Tt B at the center of the disk to progressively higher values as the
ends are According to the Griffith theory of failure, the critical
point ought to be the center where the ratio of compression to tension is 3. With
a principal stress ratio of 3, failure ought to result from the application of the
tensile stress alone, without any complication from the simultaneous compres-
sion paraHel to the eventual rupture planeo In fact, the Brazilian test has be en
found to give a tensile strength higher than that of the direct tension test,
probably owing to the effect of fissures. Short fissures weaken a direct tension
specimen more severely than they weaken a splitting tension specimen. The
ratio of Brazilian to direct tensile strength has been found to vary from unity to
more than ten as the length of preexisting fissures grows larger (Tourenq and
Denis, 1970).
Aflexural test causes failure of a rock beam by bending. Like the Brazilian
test, flexural tests also can be run on rock cores lacking machined ends. Four-
point flexuralloading (Figure 3.3d), with the bottom of the core supported on
I Standard Method of Test for Splitting Tensile Strength of Cylindrical Concrete Specimens,
ASTM Committee C-9 on Concrete and Concrete Aggregate.
66 Rack Strength and Failure Criteria
points near the ends and the top of the core loaded from aboye at the third
points, produces uniform moment in the central third ofthe specimen and gives
better reproducibility ofresults than three-point loading in which the upper load
is central. The flexural strength or "modulus of rupture" is the maximum
tensile stress on the bottom of the rock corresponding to peak load. It is
calculatell from simple beam theory assuming elastic conditions throughout.
The flexural strength is found to be two to three times as great as the direct
tensile strength. For four-point bending of cylindrical rock specimens, with
loads applied at L/3 from each end and reactions at the ends, the modulus of
rupture (MR) is:
T - 16PrnaxL
MR - 31Td3
(3.3)
where P rnax is the maximum load, L is the length between load reactions on the
lower surface, and d is the diameter of the coreo
The ring shear test (Figure 3.3e) provides a relatively simple method to test
intact rock strength as a function of confining pressure (Lundborg, 1966). In
contrast to compression tests, core specimens for the ring shear test do not
require perfectly square and smooth ends. As with the triaxial test, the results
permit an appreciation of the rate of increase of strength with confining pres-
sure. The latter is provided by the load parallel to the axis ofthe coreo Two sets
of complex fracture surfaces form along the two planes of imposed shear as the
load is applied to the plunger.
If Pis the peak load on the plunger, the peak shear stress ('r
p
) is called the
"shear strength" and is calculated by
P
'rp = 2A
where A is the area across the core sample.
3.3 Slress-Slrain Behavior in Compression
STRESS AND STRAIN
(3.4)
In discussing the deformations of rock undergoing compression from various
directions, it proves useful to divide the stresses into two parts. Nondeviatoric
stresses (<T
rnean
) are compressions equally applied in all directions, that is, a
hydrostatic state of stress. Deviatoric stresses (<Tdev) are the normal and shear
stresses that remain after subtracting a hydrostatic stress, equal to the mean
normal stress, from each normal stress component. In the triaxial compression
experiment, for example, the principal stresses are <TI = P/A and <T2 = <T3 = p.
3.3 Stress-Strain Behaviar in Campressian 67
The nondeviatoric stress is given by !(<TI + 2p) all around while the deviatoric
stress is then what remains: <T,dev = - p) and <T3,dev = <T2,dev = -!C<TI - p).
There is strong motivation for doing this: deviatoric stress produces distortion
and destruction of rocks while nondeviatoric stresses gene rally do not (as
discussed in the next section). In the triaxial test, the initial pressuring is
nondeviatoric; subsequently, both deviatoric and nondeviatoric stresses are
raised simultaneously.
Normal strains in a triaxial compression specimen can be measured with
surface-bonded electric resistance strain gages. A gage parallel to the specimen
axis record s the longitudinal strain eaxial = tllll, while a strain gage affixed to the
rock surface in the circumferential direction yields the lateral strain elateral = tld/
d, where d is the diameter of the rock and 1 is its length (see Figure 3.4a).
Assuming that the strain gage readings are zeroed after the confining pressure
has been applied, we can write
elateral = - veaxial
(3.5)
in which the constant of proportionality v is called Poisson's ratio. In fact,
proportionality is maintained only in the restricted range of loading during
which there is no initiation and growth of racks. For linearly elastic and
isotropic rocks, v must lie in the range O to 0.5 and is often assumed equal to
0.25. Because a rock expands late rally as its shortens axially (Figure 3.5), a
negative sign is introduced to define ratio as a positive quantity. For
strains of less than several %, the volume change per unit of volume, tl VIv., is
T
6/
\E d---
Figure 3.5 Deformations in compression tests.
6S Rock Strength and Failure Criteria
closely approximated by the algebraic sum of the three normal strains. In the
triaxial compression experiment then,
IlV
V = Baxial + 2Blateral
or
IlV
V = Baxial(1 - 2v) (3.6)
Volumetric strain produced either by deviatoric or nondeviatoric stresses can
be measured indirectly using surface strain gages and applying Equation 3.5 or
directly by monitoring the ftow of oil into or out of the confining vessel as the
confining pressure is held constant by a servomechanism.
HYDROSTATIC COMPRESSION
Applying a nondeviatoric stress to a rock produces a volume decrease and
eventually changes the rock fabric permanently, as pores are crushed. How-
ever, it cannot produce a peak load response; that is, the rock can always
accept an added increment of load, apparently for as high a pressure as one can
generate. Tests have been conducted into the megabar region (millions of psi)
producing phase changes in the solido The pressure, volumetric strain curve is
generally concave upward as shown in Figure 3.6 with four distinct regions. In
the first, which may be the principal region for many good rocks in civil engi-
neering service, preexisting fissures are closed and the mineral s are slightly
compressed. When the load is removed, most of the fissures remain closed and
there is a net deformation or "per-def." The fissure porosity is related to the
per-def.
Mter most of the fissures have closed, further compression produces bulk
rock compression, consisting of pore deformation and grain compression at an
approximately linear rateo The slope of the pressure-volumetric strain curve in
this region is called the bulk modulus,
2
K. In porous rocks like sandstone,
chalk, and clastic limestone, the pores begin to collapse due to stress concen-
trations around them; in well-cemented rocks, this may not occur until reaching
a pressure of the order of 1 kbar (100 MPa or 14,500 psi), but in poorIy or
weakly cemented rocks, pore crushing can occur at much lower pressures.
Finally, when all the pores have been closed, the only compressible elements
remaining are the grains themselves and the bulk modulus beco mes progre s-
sively higher. Nonporous rocks do not demonstrate pore "crush up" but show
uniformly concave-upward deformation curves to 300 kbar or higher. Pore
2 The compressibility e is l/K.
111
e
E
11
t;)
3.3 Stress-Strap Behavior in Compression 69
Deviatoric stress
constant Locking
E lastic compression
F issure closing
6 V/V
(hydrostatic compression)
Figure 3.6 Volumetric compressjot( under increasing mean
stress, with constant deviatoric stress.
crushing is destructive in very porous rocks like chalk and pumice, which are ,
converted to a cohesionless sediment on removal from the test chamber.-
DEVIATORIC COMPRESSION
Applying deviatoric stress produces strikingly different results as shown in
Figure 3.7. With initial application of the deviatoric stress, fissures and sorne
pores begin to close, producing an ineIastic, concave-upward stress-strain sec-
tion. In most rocks, this is followed by linear relationships between axial stress
and axial strain and between axial stress and lateral strain. At point B (Figure
3.7a), the rate of lateral strain begins to increase reIative to the rate o ~ i a l
strain (Poisson's ratio increases) as new cracks begin to form inside the most
critically stressed portions of the specimen-usually near the sides of the mid-
section. A microphone attached to the rock will begin to pick up "rock noise"
as new cracks form and old ones extend parallel to the direction of (T\. In the
regio n between stress B and stress e, cracks are considered to be "stable"
rheaning that with each increment of stress they grow to a finite length, and
then stop growing. Mter point e, cracks that form propagate to the edges ofthe
specimen and a system of intersecting, coalescing cracks is deveIoped, which
eventually form a semicontinuous rupture surface termed a "fault." Figures
3.7c and d, from Wawersik and Brace (1971), show this deveIopment.
Bieniawski (1967a, b) suggested that point e corresponds to the yield point in
70 Rack Strength and Failun' entena
the axial stress-axial strain curve. The peak load, point D, is the usual object of
failure criteria. However, the rock may not fail when the load reaches this
point, as is discussed later. In a stiffloading system, it is possible to continue to
shorten the specimen, as long as stress is reduced simultaneously. If the volu-
metric strain.is plotted against the deviatoric stress as in Figure 3.7b, it is seen
that the attainment of the crack initiation stress (B) is marked by a beginning of
an increase in volume associated with sliding and buckling of rock slivers
between cracks and opening of new cracks. At a stress level corresponding to
stress point e, the specimen may have a bulk volume larger than at the start of
the test. This increase in volume associated with cracking is termed dilatancy.
al, de\liatoric
IV
V
Lateral
strain ---
Extension
D
Axial
strain
V
Macrocracking
by joining
of microcracks
Contraction Normal strain
!(a) /
Mean stress
Axial normal strain
(b)
Figure 3.7 Deformation under increasing deviatoric stress, with constant
mean stress (hypothetical curves). (a) Axial and lateral normal strain with
increasing deviatoric axial stress. (b) Volumetric strain with increasing axial
normal strain (dilatancy).
72 Rock Strength and Failure Criteria
EFFECT OF CONFINING PRESSURE
Most rocks are signifis;antly strengthened by confinement. This is especially
striking in a highly fissured rock, can be imagined as a mosaic of perfectly
matching pieces. Sliding along the fissures is possible if the rock is free to
displace normal to the average surface of rupture, as shown in Figure 3.8. But
under confinement, the normal displacement required to move along such a
jagged rupture path requires additional energy input. Thus it is not uncommon
for a fissured rock to achieve an increase in strength by 10 times the amount of
Original fissured rock
(grain boundary cracks)
Locus of faulting
Figure 3.S Dilatancy caused by roughness of the rupture sur-
face.
3.3 Stress-Sira in 8ehavior in Compression 73
a small increment in mean stress. This is one reason why rock bolts are so
effective in strengthening tunnels in weathered rocks.
As mean pressure is increased, the rapid decline in load carrying capacity
after the peak load (point D in Figure 3.7) becomes gradually less striking until,
at a value of the mean pressure known as the brittle-to-ductile transition pres-
sure the rock behaves fully plastically (Figure 3.9). That is, after point D
continued deformation of the rock is possible without any decrease in stress.
(a)
a
6V/V
(b)
Nonfracture
region
Figure 3.9 Behavior in triaxial compression. (a) Transition from brittle-to-
ductile behavior. (b) Volumetric compression and dilatancv.
74 Rock Strength and Fa1ure Criteria
("Stress-hardening" behavior is sometimes observed at even higher pressures,
meaning that there is actually a strengthening of the rock as it deforms without
any "peak stress.") Figure 3.10 shows triaxial test data for a crystalline rock
(norite) and a clastic rock (sandstone), both of which demonstrate loss of
brittleness with confining pressure.
The brittle-to-ductile transition occurs at pressures far beyond the regio n of
interest in most civil engineering applications. However, in evaporite rocks and
soft clay shales, plastic behavior can be exhibited at engineering service loads.
Table 3.2 lists sorne values of the transition pressure. Without confining pres-
sure, most rocks tested past point D of Figure 3.7 will form one or more
fractures parallel to the axis of loading (Figure 3.9a). When the ends are not
smooth, the rock will sometimes split neatly in two, parallel to the axis, like a
Brazilian specimen. As the confining pressure is raised, the failed specimen
demonstrates faulting, with an inclined surface of rupture traversing the entire
specimen. In soft rocks, this may occur even with unconfined specimens. Ifthe
specimen is too short, continued deformation past the faulting regio n will drive
the edges of the fault block s into the testing machine platens, producing com-
plex fracturing in these regions and possibly apparent strain-hardening behav-
ior. At pressures aboye the brittle-to-ductile transition, there is no failure per
se, but the deformed specimen is found to contain parallel inclined lines that are
the loci of intersection of inclined rupture surfaces and the surfaces of the
specimen. Examination of the deformed rock will show intracrystalline twin
gliding, intercrystal slip, and rupture.
The effect of confining pressure is also expressed in changing volumetric
strain response as shown for a series of triaxial compression tests in Figure
3.9b. At successively higher confining pressures, the volumetric strain curves
shift smoothly upward and to the right. These curves are the algebraic sum of
hydrostatic compression under increasing mean stress (e.g., distance ae) and
dilatancy under increasing deviatoric stress (eb). The response shown in Figure
3.9b applies when the ratio of (T3 to (TI is sufficiently small. When this ratio is
Table 3.2 Brittle-to-Ductile Transition
Pressures for Rocks (At Room TemperatureJ
Gage Pressure
Rock Type (MPa) (psi)
Rock salt O O
Chalk <10 <1500
Compaction shale 0-20 0-3000
Limestone 20-100 3000-15,000
Sandstone >100
>15,000
Granite ; 100
; 15,000
a, -p
50
40
"
30
o-
'b
~
~
20
(/)
10
a, - p
100
80
"
60 o-
M
o
~
~
40
ii
20
8
3.3 Stress-Strain 8ehavior in Compression 75
Axial strain - 10-
3
in./in.
(a)
Uniaxial
12 16 20
Axial strain - 10-
3
in./in.
(b)
,/e
,/
Lateral pressur
I
24
28 32
Figure 3.10 Stress difference versus axial strain curves as a function of con-
fining pressure in triaxial compression experiments on sandstone (a) and
norite (h). [From Bieniawski (1972).]
76 Rock Strength andFailure entena
larger than a value of the order of 0.2, fracture does not occur and dilatancy is
suppressed (as discussed in Section 3.8). In the usual triaxial test procedure,
the principal stress ratio is decreased progressively during application of the
deviator stress, until fracture occurs. But in practice, loading may occur such
tha! the principal stress ratio remains fixed or increases.
3.4 The Meaning af ((Rack Strength"
The word "failure" connotes an almost totalloss of integrity in a sample of
rock; in an engineering context, it usually implies los s of ability to perform the
intended function. Obviously phenomena that constitute failure will depend on
the function-varying from loss of a commodity in storage to structural col-
lapse, property damage, and death. Even in a single specimen of rock, how-
ever, the concept of "failure" is unclear, for a totalloss of cohesion occurs or
does not occur in a single given rock according to the way it is loaded. The
reason for such apparentIy fickle response is that the destruction of a rock by
load is partIy dependent on the loading system and is not a true rock property.
Fof purposes of engineering design, it is useful to map peak stress values (i.e.,
the stresses corresponding to point D of Figure 3.7), and the criteria of failure
discussed later will relate to such points. However, the compression test does
not have to end in rupture at that point, but may proceed all the way to point E
orbeyond ifthe loading system is very stiff. The rock will exhibit what has been
called "a complete stress-strain curve" if tested in a stiff system because the
sys!em responds to gradual deterioration in load-carrying capacity through
automatic reduction in the applied load.
A testing machine is a reaction frame in which a screw or a hydraulic
cylinder is operated to load a specimen. A screw-powered machine is charac-
terized in Figure 3.11a. The rock specimen is fitted between a test table and an
upper platen that are connected by stiff screws parallel to the axis of the
specimen. A motor below the table turns a gear that causes the screws to turn
thus bringing the upper platen up or down. If the screws are turned so that the
rock specimen feels a load and then the motor is switched off, any subsequent
movement of the upper platen relative to the table must alter the load at arate
given by the stiffness of the testing machine km' The family of lines marked A
through J in Figure 3.11b describes the machine stiffness at different platen
positions. Operation of the testing machine to build up load on a specimen
corresponds to moving across the family of curves as shown in the figure. At
the point of peak load of the rock, assuming we turn the motor off, the speci-
men will fail or not depending upon the relative values of kr and km where k
r
is
the slope of the post peak portio n of the complete stress-strain curve. For
example, rock 1, Figure 3.11b, will continue to deform without sudden rupture
as the testing machine is continuously shortened, whereas even with the ma-
G
3.4 The Meaning of IIRock Strength" 77
V J - o ~ - - Screw
". " .....
. . '
.. ::
o',
........ " ..
", ': .... ~
I . I - - - - - - - - ~ I
II+---..r'::t+...r I
A_J \
Motor Gear
(a)
Upper platen
~ ~ - - (1) Nonbrittle rock
Downward
displacement
of upper platen
(bl
Figure 3.11 Influence of testing machine stiffness on failure. (a)
A testing machine and its representation by a freebody. (h) Sta-
ble and unstable samples.
chine turned off, rock 2 will fail because the machine as it "unstretches"
cannot reduce the load sufficientIy to satisfy the unloading requirements of the
rock. However, if the motor were quickly switched to run backward and the
upper platen were moved upward, it would be possible to follow the complete
stress-strain curve of rock 2 as the system returned to curves G, F, etc. In a
78 Rock Strength and Falure Critera
servocontrolled testing machine, this can be done automatically, the motor
responding to the commands of an electrical circuit processing the signal of a
deformation transducer attached to the specimen. Even without such a ma-
chine, it is possible to determine the complete stress-strain curve of a brittle
rock by quick manual response, switching back and forth from loading to
unloading modes. Figure 3.12 shows an actual record from such an operation
with a specimen of coarse-grained marble (Chino marble, California). When the
rock was loaded in the usual way in a 160,OOO-lb-capacity screw testing ma-
chine, the peak load was followed by a violent rupture that reduced the speci-
men to rock powder. By quickly unloading whenever the x-y plotter record
revealed the onset of yielding, a series of hysteresis loops was created, the
envelope to which estimates the right side of the complete stress-strain curve.
At the end of the test, the specimen displayed a continuous fault but was still
integral. Since the stiffness of the rock is proportional to its cross-sectional
area, such a test is relatively easier to perform using small specimens. A con-
vincing demonstration of the influence of the loading system stiffness on the
mode of failure can be achieved by running two tests with varying machine
stiffness while the rock stiffness is held constant. This can be achieved by
adding a spring in series with the rock for one of the two tests.
4000
3200
3200 lb
2400
cE
-ti
ro
o
-'
1600
~
800
0_015 0_2
Cross-head motion, in.
Figure 3.12 A complete stress-strain curve obtained by
load cycling on a moderately stiff testing machine. The
specimen was a coarse marble cylinder, 0.8 in. in diam-
eter and 1.45 in. long.
3.5 Applications of the Complete Stress-Stran Cwve 79
3.5 Applications of the
Complete Stress-Strain Curve
Normally when stresses become high enough to cause fractures in the wall of a
tunnel or mine, rock simply spalls off, producing a destressed zone that drives
the flow of stress away from the opening.
In a well-designed mine, the roof load will find somewhere else to go when
a pillar collapses. But if a room and pillar mine were made with very wide
rooms, the loss of one pillar might be insufferable.
These varying behavior modes in practice are understandable in terms of
the complete stress-strain curve concept. In the mine with very wide rooms,
the deflection of the roof due to the removal of one pillar can be calculated by
assuming the roof span to be two rooms and one pillar wide. U sing beam
formulas or numerical model methods, the ratio of peak pillar load to the
increment of roof deflection caused by removal of the pillar defines the system
stiffness. If this stiffness is greater in magnitude than the slope of the postpeak
part ofthe complete force-displacement curve, the mine can survive the failure
of a pillar.
The complete stress-strain curve can also be used to preduct failure of rock
as a result of creep. As shown in Figure 3.13, the locus of a creep test in the
stress-strain graph is a horizontalline. If the initial stress in the rock is c10se to
the peak load, creep will terminate in rupture when accumulated strain is such
as to intersect the falling part of the complete stress-strain curve. A creep test
started at A will terminate in rupture at point B after a relatively short time. A
a, - p
G -------
of long-term
creep tests
~ = = - - - - - - - - - - - - - - - - - - - - - - - - f 10ng
Figure 3.13 Creep in relation to the complete stress-strain curve.
so Rock Strength and Failure Criteria
al -p
Figure 3.14 Response to dynamic loading, in relation to the complete stress-
strain curve.
creep test begun at e will terminate in rupture at D after a much longer time.
And a creep test initiated at E below critical stress level G will approach point F
without rupture after a long time (compare with Figure 6.16).
A similar concept applies to cyclic loading beneath the peak load level, as
shown in Figure 3.14. Cycles of loading and unloading produce "hysteresis
loops" as energy is consumed in sliding on cracks and fissures inside the rock
volume. Multiple load cycles begun at point A such that the peak load is never
surpassed will cause a migration of the envelope of hysteresis loops which
terminate in rupture at point B.
3
3.6 The Mohr-Coulomb Failure Criterion
We have noted that the peak stress of rock undergoing deviatoric loading will
increase if the rock is confined. The variation of peak stress 0"1 with confining
pressure 0"3 is known as a criterion of failure. The simplest and best-known
criterio n of failure for rocks is the Mohr-Coulomb criterion; as shown in Figure
3.15, this consists of a linear envelope touching all Mohr's circles representing
J This suggests a method of relieving stored energy along faults near the point by
cyclic loading. Professor B. Haimson of the University of Wisconsin proposed be done by
cyclic pumping of water from wells in the fault zone. The effect of water pressure IS considered in
Section 3.7.
I
I
: I
'51
:u I
gl
I
3.6 The Mohr-Coulomb Failure Criterion 81
T
/
__
a, a, al
Figure 3.15 The Mohr-Coulomb failure criterio n with a tension
cutoff.
critical combinations of principal stresses. Stated in terms of normal and shear
stresses on the plane represented by the point of tangency of a Mohr circle with
the envelope
Tp = Si + O" tan <p (3.7)
<p is called the angle of internal friction, for like a friction angle for sliding
between srfaces, it describes the rate of increase of peak strength with normal
stress. T
p
is the peak shear stress, or shear strength.
The Mohr-Coulomb criterion is also used to represent the "residual"
strength, that is, the minimum strength reached by a material subjected to
deformation beyond the peak, as shown in Figure 3.10. In this case, the sub-
script r may be used with each of the terms of Equation 3.7 to identify them as
parameters of residual strength. The residual shear strength (Si,r may approach
zero while the residual friction angle <Pr will usually lie between zero and the
peak friction angle. However, i compaction shales rich in montmorillonite,
like the Cretaceous Bearpaw and Pierre shales, values as low as 4_6
0
are
reported, even in "drained" test arrangements that prevent pore water pres-
sure accumulation during deformation (Townsend and Gilbert, 1974).
Equation 3.7 has the following physical interpretation. "Failure" occurs
when the applied shear stress les s the frictional resistance associated with the
normal stress on the failure plane becomes equal to a constant of the rock, Si.
Since it would not be reasonable to admit a frictional resistance in the presence
of a tensile normal stress, this equation then loses its physical validity when the
value of O" crosses into the tensile region; O" represents the normal stress on the
plane of failure. The mnimum principal stress 0"3 may be tensile as long as O"
remains compressive. Other theories of failure (e.g., the Griffith theory) are
82 Rock Strength and Failure Criteria
more exact in the tensile region. However, the Mohr-Coulomb theory has the
merit of simplicity and will be retained here by extrapolating the Mohr-Cou-
10mb line into the tensile regio n up to the point where 0'3 becomes equal to the
uniaxial tensile strength - T
o
. The minor principal stress can never be less than
-To
Respecting the last as a constraint on the criterion of failure is, in effect,
recognizing a "tension cutoff' superimposed on the Mohr-Coulomb criterion
offailure as shown in Figure 3.15. The actual envelope of critical Mohr's circles
with one principal stress negative willlie beneath the Mohr-Coulomb criterion
with the superimposed tension cutoff as indicated in Figure 3.16, so it is neces-
sary to reduce the tensile strength T
o
and the shear strength intercept Si when
applying this simplified failure criterion in any practical situation.
In terms of the principal stresses at peak load conditions, the Mohr-Cou-
10mb criterion can be written
(3.8)
where O'I,p is the major principal stress corresponding to the peak of the stress-
strain curve, and qu is the unconfined compressive strength. The change of
variables leads to the following relationship between shear strength intercept Si
and unconfined compressive strength qu
~ ~ L - ~ __ ~ __________ ____ ______________ ~ a
Figure 3.16 Comparison of empirical envelope and Mohr-Coulomb
criterion in the tensile region. Inside the ruled region, the Mohr-
Coulomb criterio n with tension cutoff overestimates the strength.
(3.9)
3.6 The Mohr-Coulomb Faz1ure Criterion S3
Table 3.3 Representative Values for Shear Strength lntercept
(Si) and Angle of lnternal Friction ( ~ ) for Selected Rocks
a
Descripti n
Berea sandstone
Bartlesville sandstone
Pottsville sandstone
Repetto siltstone
Muddy shale
Stockton shale
Edmonton bentonitic shale
(water content 30%)
Siou)i quartzite
Texas slate; loaded
30 to c1eavage
90 to c1eavage
Georgia marble
Wolf Camp limestone
Indiana limestone
Hasmark dolomite
Chalk
Blaine anhydrite
Inada biotite granite
Stone Mountain granite
Nevada Test Site basalt
Schistose gneiss
90 to schistocity
30 to schistocity
Porosity
(%)
18.2
14.0
5.6
4.7
44.0
0.3
19.4
3.5
40.0
0.4
0.2
4.6
0.5
1.9
a Data frorn Kulhawy (1975) (Ref. 5).
Si
(MPa)
27.2
8.0
14.9
34.7
38.4
0.34
0.3
70.6
26.2
70.3
21.2
23.6
6.72
22.8
O
43.4
55.2
55.1
66.2
46.9
14.8
Range of
Confining
Pressure
cp (MPa)
27.8 0-200
37.2 0-203
45.2 0-68.9
32.1 0-200
14.4 0-200
22.0 0.8-4.1
7.5 0.1-3.1
48.0 0-203
21.0 34.5-276
26.9 34.5-276
25.3 5.6-68.9
34.8 0-203
42.0 0-9.6
35.5 0.8-5.9
31.5 10-90
29.4 0-203
47.7 0.1-98
51.0 0-68.9
31.0 3.4-34.5
28.0 0-69
27.6 0-69
Reference
b
4
3
8
4
4
2
9
3
6
6
8
3
8
4
1
3
7
8
10
2
2
b 1. Dayre, M., Dessene, J. L., and Wack,'B. (1970) Proc. 2nd Congress ofISRM, Belgrade, Vol.
1, pp. 373-381.
2. DeKlotz, E., Heck, W. J., and Neff, T. L. (1964) First Interirn Report, MRD Lab Report
64/493, U. S. Arrny Corps of Engineers, Missouri River Division.
3. Handin, J. and Hager, R. V. (1957) Bull. A.A.P.G. 41: 1-50.
4. Handin, J., Hager, R. V., Friedrnan, M., and Feather, J. N. (1963) Bull. A.A.P.G. 47: 717-
755.
5. Kulhawy, F. (1975) Eng. Geol. 9: 327-350.
6. McLamore, R. T. (1966) Strength-deforrnation characteristics of anisotropic sedirnentary
rocks, Ph.D. Thesis, University of Texas, Austin.
7. Mogi, K. (1964) Bull. Earthquake Res. Inst., Tokyo, Vol. 42, Part 3, pp. 491-514.
8. Schwartz, A. E. (1964) Proc. 6th Symp. on Rock Mech., Rolla, Missouri, pp. 109-151.
9. Sinclair, S. R. and Brooker, E. W. (1967) Proc. Geotech. Conf on Shear Strength Properties
of Natural Soils and Rocks, Oslo, Vol. 1, pp. 295-299.
10. Stowe, R. L. (1969) U. S. Arrny Corps of Engineers Waterways Experirnent Station. Vicks-
burg, Mise. Paper C-69-1.
84 Rack Strength and.Failure Criteria
The maximum tension criterion must be superimposed on Equation 3.8, that is,
failure is presumed to occur because of tensile stress whenever <T3 becomes
equal to - T
o
, regardless of the value of <TI
Typical values of the peak shear strength intercept Si and the peak angle of
internal friction <p for a representative set of rock specimens are listed in Table
3.3. The ratio ofunconfined compressive to tensile strength aulTo for a sampling
of rock types is given in Table 3.1.
3.7 The Effect of Water
Sorne rocks are weakened by the addition of water, the effect being a chemical
deterioration of the cement or clay binder. A friable sandstone may, typically,
lose 15% of its strength by mere saturation. In extreme cases, such as mont-
morillonitic clay shales, saturation is totally destructive. In most cases, how-
ever, it is the effect of pore and fissure water pressure that exerts the greatest
inftuence on rock strength. If drainage is impeded during loading, the pores or
fissures will compress the contained water, raising its pressure.
Development of pore pressure and consequent loss in strength of a Penn-
sylvanian shale tested in triaxial compression is shown in Figure 3.17. Two
separate test results are presented in this diagram: the circles represent triaxial
compression of a saturated specimen under conditions such that excess pore
pressures could drain away rather than accumulate ("drained conditions"); the
triangles represent a saturated shale specimen tested without drainage, so that
excess pore pressures that develop must accumulate ("undrained conditions").
The curve of differential axial stress versus axial strain for the drained test
displays a peak and then a descending tail as depicted in Figure 3.7a. Since the
mean stress increases simultaneously with the axial stress in a triaxial test, the
curve of volumetric strain shown in Figure 3.17 is the sum of hydrostatic
compression (Figure 3.6) and dilatancy behavior (Figure 3.7b). Initially, the
volume decreases by hydrostatic compression until the specimen begins to
dilate, whereupon the rate of volume decrease slows, eventually becoming
negative, meaning the volume increases on subsequent load increments. In the
undrained test, the tendency for volume change cannot be fully realized be-
cause the water filling the voids undergoes compression rather than drainage.
As a result, the water pressure Pw inside the pores begins to increase. This
dramatically lowers the peak stress and ftattens the postpeak curve.
Many investigators have confirmed the validity of Terzaghi's effective
stress law for rocks, which states that a pressure of Pw in the pore water of a
rock will cause the same reduction in peak normal stress as caused by a reduc-
tion of the confining pressure by an amount equal to pw We can make use of
this result by introducing the term efJective stress <T' defined by
<T' = <T - Pw (3.10)
3.7 The Effect of Water S5
200
o Drained
1:> Undrained
160
120
'[
U
80
01 - 03
40
__ -L ______ L-____
?fi 1.0
c:
'" 2.0
E
:J
o
> 3.0
4.0 '---__ -'-___ .1..-___ ..1.-__ ---1
O 2.0 4.0 6.0
Axial strain. %
Figure 3.17 Drained and undrained triaxial com-
pression test results for a shale of Pennsylvanian
age; w is the initial water content; Pw is the pore
water pressure. (From Mesri and Gibala, 1972.)
8.0
Differential stress (<TI - <T3) is unaffected by water pressure since <T; - =
(<TI - Pw) - (<T3 - Pw) = <TI - <T3.
T?e effect of water pressure can be input in the failure criterion simply by
restatmg the conditions foi" failure in terms of effective stresses. In a test on a
dry rock, there is no difference between normal stresses and effective normal
For a saturated rock, rewrite Equation 3.8 in terms of effective stress
y mtroducing , on the normal stress terms:
= qu + 0"] tan
2
(45 +
(3.11)
S6 Rock Strength and Failure Criteria
or
(3.12)
Since the differential stress is unaffected by pore pressure, Equation 3.12 may
also be written
(T1,p - (T3 = qu + (T3 - Pw) [tan
2
(45 + 1) -1]
Solving for Pw, we can calculate the water pressure in the pores or fissures of a
rock required to initiate failure from an initial state of stress defined by (TI and
(T3:
(TI - (T3) - qu
Pw = (T3 - tan2(45 + 1>/2) - 1
(3.13)
Figure 3.18 portrays this condition graphicaHy. The buildup ofwater pressure
in the rock near a reservoir or in an aquifer can cause rock failure and earth-
quakes, if the rock is initiaHy stressed near the limit. However, earthquakes
induced by reservoir construction and by pumping water into deep aquifers are
believed to originate from rupture along preexisting faults in determined orien-
tations. The mechanism is similar but the equations contain the inftuence of the
relative directions of initial stress as discussed in Chapter 5 (compare with
Equation 5.9).
- - L - - L ~ ________ ~ __________________ -L ________ t - __ a
Figure 3.18 Water pressure required to initiate failure of an intact rock from a
given initial state of stress.
3.8 The lJifluence of the
Principal Stress Ratio on FaHure
3.9 Empirical Criteria of Failure 87
In the usual triaxial compression experiment, the rock is seated with a hydro-
static stress, that is, with a principal stress ratio K = (T3/(T1 equal to unity.
Thereafter, as (TI is raised, the value of K is reduced until cracking occurs, and
eventuaHy peak strength is reached. This may not be a realistic loading path for
aH situations, and it may be desirable to consider rock behavior when the
principal stress ratio is fixed at sorne value. In examining the conditions result-
ing from excavating an underground chamber in a rock mass, for example, the
directions and relative magnitudes of principal stresses can be found through-
out the region of inftuence of the opening. Changing the assumption about the
magnitudes of the initial stresses will increase or decrease these stresses but
will not alter any value of K as long as the rock behaves elasticaHy. Therefore,
there is sorne merit in expressing the criterion offailure in terms ofthe principal
stress ratio, as discussed by Hoek (1968). When this is done, it is easily seen
that there is a value of K aboye which failure cannot occur, and this can be
verified by tests. In terms of the Mohr-Coulomb criterion of failure, dividing
both sides of Equation 3.8 by (TI,P and introducing K = (T3/(T1 yields
(T1,p = 1 - K tan
2
(45 + 1>/2)
(3.14)
from which we can see that the peak major principal stress becomes large when
K approaches cot
2
(45 + 1>/2). For example, for 1> = 45, failure cannot occur
aboye a principal stress ratio K = 0.17.
3.9 Empirical Criteria of FaHure
While the Mohr-Coulomb criterion is easy to work with and affords a useful
formula for manipulation in practical situations, a more precise criterion of
failure can be determined for any rock by fitting an envelope to Mohr' s circles
representing values of the principal stresses at peak conditions in laboratory
tests. As shown in Figure 3.19, this envelope will frequently curve downward.
Jaeger and Cook (1976)4 and Hoek (1968) demonstrated that the failure enve-
lopes for most rocks lie between a straight line and a parabola: The Griffith
theory of failure predicts a parabola in the tensile stress region. This theory is
premised on the presence of randomly oriented fissures in the rock that act to
create local stress concentrations, facilitating new crack initiation. However,
the Griffith theory has no physical basis in the region where both principal
4 See references, Chapter 1.
ss Rock Strength and Falure Criteria
3.9 Emprical Criteria of Failure 89
T T
Figure 3.19 An empirical criterio n of failure defined by the envelope to a series of
Mohr's circ1es: A, direct tension; B, Brazilian; e, unconfined compression; D, tri-
axial compression.
stresses are compressive. In practice, empirical curve fitting is the best proce-
dure for producing a criterio n offailure tailored to any given rock type. See, for
example, Herget and Unrug (1976). A satisfactory formula for many purposes
will be afforded by the union of a tension cutoff, a3 = - T
o
, and a power law
(Bieniawski, 1974):
al,p = 1 + N (a
3
)M (3.15)
qu qu
The constants N and M will be determined by fitting a curve to the family of
points
(a
3
, al,p _ 1)
qu qu
Another approach fits data from the ring shear test (Figure 3.3e) to find an
empirical equation for a Mohr envelope of intact rock (Lundborg, 1966). The
peak shear strength (7
p
) (Equation 3.4) is plotted against a to define the strength
envelope, Figure 3.20a. Lundborg found that such data define a curved enve-
lope with intercept Si and asymptote Sf, fit by
.t'a
7
p
= Si + ----!...--:---
'a
1 + --,-.t_-=-
Sr- Si
(3.16)
----------------
Si
_1_
Tp -Si
(a)
________________________________
(b)
Figure 3.20 An empirical criterio n of failure derived
from the ring shear test. (After Lundborg, 1966.) (a) A
plot of equation 3.16; (b) graphical determination of L'.
a
(The .t' has been used in place of Lundborg's .t to distinguish this
from the Mohr-Coulomb coefficient of internal friction, .t = tan <p.)
EquatlOn 3.16 can also be written
1 1 1
7
p
- Si = .t'a + Sf- Si (3.17)
so .t' is determined as the inverse of the slope of the line obtained by
(7p - S)-I as ordinate against (a)-I as abscissa (Figure 3.20b). Table
3.4 hsts typical values of Lundborg's parameters. Strengths determined by ring
90 Rock Strength and Failure Critera
Table 3.4 Sorne Values of
Constants for Lundborg's
Strength Equation
Si S
IL' (MPa) (MPa)
Granite 2.0 60 970
Pegmatite 2.5 50 1170
Quartzite 2.0 60 610
Slate 1.8 30 570
Limestone 1.2 30 870
shear tend to be slightly higher than corresponding strengths determined by
triaxial tests.
3.10 The Effect of Size on Strength
Rocks are composed of crystals and grains in a fabric inc1udes
fissures; understandably, rather large samples are reqmred to obtam statlstl-
cally complete collections of all the components that influence strength. ":'hen
the size of a specimen is so small that relatively few cracks are present, fatlure
is forced to involve new crack growth, whereas a rock mas s loaded through a
larger volume in the field may present preexisting criticallocations.
Thus rock strength is size dependent. Coal, altered gramt1c rocks, and
other rocks with networks of fissures exhibit the greatest degree of Slze depen-
dency, the ratio offield to laboratory strengths sometimes attaining values of 10
or more. .
A few definitive studies have been made of size effect in compresslve
strength over a broad spectrum of specimen sizes. Bieniawski (1968) reported
tests on prismatic in situ coal specimens up to 1.6 x 1.6 x. 1 m, prepared by
cutting coal from a pillar; the specimens were then capped wlth concrete
and loaded by hydraulic jacks. Jahns (1966) reported results of sImIlar tests on
cubical specimens of calcareous iron ore; the specimens were prepared
means of slot cutting with overlapping drill holes. Jahns a specl-
men size such that 10 discontinuities intersect any edge. Larger speClmens are
more expensive without bringing additional size reduction, while smaller speci-
mens yield unnaturally high strengths. Available data are too sparse to
Jahn's recommendation for aH rock types but it does appear that there .1S
generally a size such that larger specimens suffer no further decrease m
3.10
The E.ffect of Size on Strength
150
100
70
50
Calcareous iron ore, Jahns (1966)
0."
:2
30
J
-5
01
e
1;;
Q)
>
Cedar City quartz diorite (altered)
Q)
a.
Pratt, Black, Brown, and Brace (1972)
E
O
U
O
O
4
X Coal, Bieniawski (1968)
3
2
O
Specimen edge length, m
Figure 3.21 Effect of specimen size on unconfined compressive strength. (After
Bieniawski and Van Heerden, 1975.)
91
3.0
strength. Figure 3.21 demonstrates this pattern ofbehavior in a summary ofthe
tests on coal and iron ore, as well as tests on an altered and fissured quartz
diorite by Pratt et al. (1972). This clever series of tests included specimens of
equilateral triangular cross section 6 ft (1.83 m) on edge, and 9 ft (2.74 m) long,
loaded via stainless steel flat jacks in a vertical slot at one ende Figure 3.22a
shows a specimen being freed by drilling a slot inc1ined at 60 and Figure 3.22b
shows the surface of the specimen, with completed slots, jacks in place on one
end, and extensometers positioned for strain measurements on the surface. The
quartz diorite tested displayed a large size effect because it contains highly
fractured plagioclase and amphibole phenocrysts in a finer-grained ground
rnass with disseminated clay; ihe porosity of this rock is 8-10%.
The influence of size on shear and tension tests is less weH documented but
undoubtedly as severe for rocks that contain discontinuities. The subject of
scale effect will be considered further in Chapter 7 in the context of under-
ground openings.
92 Rock Strength and Failure Criteria
J(b)
Figure 3.22 Large uniaxial compression tests conducted in-situ by TerraTek
on Cedar City Quartz Diorite. (a) Drilling a line of 1-1I2-inch diameter holes
plunging 60 to create an inclined slot forming one side of the triangular prism
"specimen." (b) A view of the test site showing ftat jacks at one end and ex-
tensometers for relative displacement measurement during loading. (Courtesy
of H. Pratt.)
3.11 Anisotropic Rocks 93
3.11 Anisotropic Rocks
Variation of compressive strength according to the direction of the principal
stresses is termed "strength anisotropy." Strong anisotropy is characteristic of
rocks composed of parallel arrangements of ftat minerals like mica, chlorite,
and clay, or long mineral s like hornblende. Thus the metamorphic rocks, espe-
cially schist and slate, are often markedly directional in their behavior. For
example, Donath (1964) found the ratio of minimum to maximum unconfined
compressive strength of Martinsburg slate to be equal to 0.17. Anisotropy also
occurs in regularly interlayered mixtures of different components, as in banded
gneisses, sandstone/shale alternations, or chert/shale alternations. In all such
rocks, strength varies continuously with direction and demonstrates pro-
nounced minima when the planes of symmetry of the rock structure are oblique
to the major principal stress.
Rock mas ses cut by sets of joints also display strength anisotropy, except
where the joint planes lie within about 30 of being normal to the major principal
stress direction. The theory of strength for jointed rocks is discussed in Chap-
ter 5.
Strength anisotropy can be evaluated best by systematic laboratory testing
of specimens drilled in different directions from an oriented block sample.
Triaxial compression tests at a set of confining pressures for each given orienta-
tion then determine the parameters Si and ~ as functions of orientation. Ex-
panding on a theory introduced by Jaeger (1960), McLamore (1966) proposed
that both Si and ~ could be described as continuous functions of direction
according to
Si = SI - S2[COS 2(t/J - t/Jmin,sW (3.18)
and
tan ~ = TI - T
2
[cos 2(t/J - t/Jmin,ci]m (3.19)
where SI. S2, TI. T
2
, m,' and n are constants
t/J is the angle between the direction of the cleavage (or schistocity,
bedding or symmetry plane) and the direction (TI
t/Jmin,s and t/Jmin,ci> are the values of t/J corresponding to minima in Si and ~ ,
respectively
For a slate, McLamore determined that friction and shear strength inter-
cept minima occur at different values of t/J, respectively 50 and 30. The
strength parameters for the slate are
Si = 65.0 - 38.6[cos 2(t/J - 30)]3 (MPa)
(3.18a)
and
tan ~ = 0.600 - 0.280 cos 2(t/J - 50) (3.19a)
94 Rock Strength and Failure Criteria
50
0.." 30
::;:
20
10
O
r
I
I
I
I
11/1 . ~ 30
I mm
I
I
I
I
Slate. a
3
= 69 MP.
(McLamore.1966)
(a)
(b)
O 1/1 m;n = 45 - rp/2
For jointed
rack
Figure 3.23 Strength anisotropy in triaxial compression.
theory for:
Continuously
variable strength
Discontinuous
rock (Chapter 5)
References 95
In general, the entire range of I/J from O to 90 cannot be well fit with one set of
constants since the theory (Equations 3.18 and 3.19) would then predict
strength at I/J = 0 to be less than the strength at I/J = 90; in fact, the strength
when loading is parallel to slaty cleavage, schistosity, or bedding is usually
higher than the strength when the loading is perpendicular to the planes of
weakness within the rock. (Compare Figures 3.23a and b.) For oH shale, a
repetitive layering of marlstone and kerogen, McLamore used one set of con-
stants for the region 0 ~ I/J < 30 and a second set of constants for 30 ~ I/J ~
90.
The variation of the friction angle with direction proves generally less
severe than the variation of the shear strength intercept. As a simplification,
as sume n = 1, I/Jmin,s = 30, and <P independent of direction (m = O). Then the
strength anisotropy can be evaluated from compressive tests run at I/J = 30 and
I/J = 75 (see Problem 12).
References
Bieniawski, Z. T. (1967a) Stability concept of brittle fracture propagation in rock, Eng.
Geol. 2: 149-162.
Bieniawski, Z. T. (1967b) Mechanism ofbrittle fracture ofrock, Int. J. Rock Mech. Min.
Sci. 4: 395-430.
Bieniawski, Z. T. (1968) The effect of specimen size on compressive strength of coal,
Int. J. Rock Mech. Min. Sci. 5: 325-335.
Bieniawski, Z. T. (1972) Propagation of brittle fracture in rock, Proceedings, 10th
Symposium on Rock Mechanics (AIME), pp. 409-427.
Bieniawski, Z. T. (1974) Estimating the strength of rock materials, J. South African
Inst. Min. Metal/. 74: 312-320.
Bieniawski, Z. T. and Bernede, M. J. (1979) Suggested methods for determining the
uniaxial compressive,strength and deformability ofrock materials, for ISRM Com-
mission on Standardization of Laboratory and Field Tests, Int. J. Rock Mech. Min.
Sci. 16 (2).
Bieniawski, Z. T. and Hawkes, 1. (1978) Suggested methods for determining tensile
strength of rock materials, for ISRM Commission on Standardization of Lab and
Field Tests, Int. J. Rock Mech. Min. Sci. 15: 99-104.
Bieniawski, Z. T. and Van Heerden, W. L. (1975) The significance of in-situ tests on
large rock specimens, Int. J. Rock Mech. Min. Sci. 12: 101-113.
Broch, E. (1974) The influence uf water on some rock properties, Proc. 3rd Congo ISRM
(Denver), Vol. 11 A, pp. 33-38.
Brown, E. T., Richards, L. W., and Barr, M. V. (1977) Shear strength characteristics of
Delabole slates, Proceedings, Conference on Rock Engineering (British Geotechni-
cal Society, Vol. 1, pp. 33-51.
Cook, N. G. W. and Hodgson, K. (1965) Some detailed stress-strain curves for Rock, J.
Geophys. Res. 70: 2883-2888.
96 Rock Strenglh and Failure Critera
Donath, F. A. (1964) Strength variation and deformational behavior in anisotropic
rocks. In W. Judd (Ed.), State of Stress in the Earth's Crust, Elsevier, New York,
pp. 281-300. . . . .
Fairhurst, C. (1964) On the validity of the BrazIltan test for bnttle matenals, Int. J. Rock
Mech. Min. Sci. 1: 535-546.
Haimson, B. C. (1974) Mechanical behavior of rock under cyclic loading, Proc. 3rd
Congo ISRM (Denver), Vol. 11 A, pp. 373-378.
Hallbauer, D. K., Wagner, H., and Cook, N. G. W. (1973) Sorne observations concem-
ing the microscopic and mechanical behavior of quartzite specimens in stiff triaxial
compression tests, Int. J. Rock Mech. Min. Sci. 10: 713-726.
Heard, H. C. (1967) The influence of environment on the brittle failure of rocks, Pro-
ceedings, 8th Symposium on Rock Mechanics (AIME), pp. 82-93.
Herget, G. and Unrug, K. (1976) In situ rock strength from triaxial testing, Int. J. Rock
Mech. Min. Sci. 13: 299-302.
Heuze, F. E. (1980) Scale effects in the determination of rock mass strength and de-
formability, Rock Mech. U (3).
Hoek, E. (1968) Brittle failure of rock. In K. Stagg and O. Zienkiewicz (Eds.), Rock
Mechanics in Engineering Practice, Wiley, New York.
Hoek, E. and Brown, E. T. (1980). Empirical strength criterion for rock masses, J.
Geotech. Eng. ASCE 106: 1013-1035.
Hoek, E. and Franklin, J. A. (1968) Sample triaxial cell for field or laboratory testing of
rock, Trans. Section A Inst. Min. Metal. 77: A22-A26.
Hudson, J. A., Crouch, S. L., and Fairhurst, C. (1972) Soft, stiff and servo-controlled
testing machines: A review with reference to rock failure, Engl. Geol. 6: 155-189.
Hustrulid, W. and Robinson, F. (1972) A simple stiff machine for testing rock in com-
pression, Proceedings, 14th Symposium on Rock Mechanics (ASCE), pp. 61-84.
Hustrulid, W. A. (1976) A review of coal pillar strength formulas, Rock Mech. 8: 115-
145.
Jaeger, J. C. (1960) Shear failure of anisotropic rocks, Geol. Mag. 97: 65-72.
Jahns, H. (1966) Measuring the strength ofrock in-situ at an increasing scale, Proc. 1st
Congo ISRM (Lisbon), Vol. 1, pp. 477-482 (in German).
Kulhawy, F. H. (1975) Stress-deformation properties ofrock and rock discontinuities,
Eng. Geol. 9: 327-350.
Lundborg, N. (1966) Triaxial shear strength of sorne Swedish rocks and ores, Proc. 1st
Congo ISRM (Lisbon), Vol. 1, pp. 251-255.
Maurer, W. C. (1965) Shear failure of rock under compression, Soco Petrol. Eng. 5:
167-176.
McLamore, R. T. (1966) Strength-deformation characteristics of anisotropic sedimen-
tary rocks, Ph.D. thesis, University of Texas, Austin.
Mesri, G. and Gibala, R. (1972) Engineering properties of a Pennsylvanian shale, Pro-
ceedings, 13th Symposium on Rock Mechanics (ASCE), pp. 57-75.
Pratt, H. R., Black, A. D., Brown, W. D., and Brace, W. F. (1972) The effect of
specimen size on the mechanical properties of unjointed diorite, Int. J. Rock Mech.
Min. Sci. 9: 513-530.
Reichmuth, D. R. (1963) Correlation of force-displacement data with physical proper-
ties of rock for percussive drilling systems, Proceedings, 5th Symposium on Rock
Mechanics, p. 33.
Problems 97
Robinson, L. H. Jr. (1959) Effects of pore pressures on failure characteristics of sedi-
mentary rocks, Trans. AIME 216: 26-32.
Toure?q, C. an.d Denis, A. (1970) The tensile strength ofrocks, Lab de Ponts el Chaus-
sees-Pans, Research Report 4 (in French).
Townsend, .F. C. and Gilbert, P. A. (1974) Engineering properties of c1ay shales, Co s
of EnglOeers, WES Tech, Report S-7i-6. rp
Vogler, U. W. .K. (1978) methods for determining the strength of
rock matenals m compresslon, for ISRM Commission on Standardization
of and PIe.ld Tests, Int. J. Rock Mech. Min. Sci. 15: 47-52.
W. R. (1972) Tu?e dependent rock behavior in uniaxial compression, Pro-
14th Sympos/Um on Rock Mechanics (ASCE), pp. 85-106.
W. R. and Brace, W. F. (1971) Post failure behavior of a granite and a
dtabase, Rock Mech. 3: 61-85.
Wawersik, W'. R. and F.airhurst, C. (1970) A study of brittle rock fracture in laboratory
expenments, Int. J. Rock Mech. Min. Sci. 7: 561-575.
Yudhblr, Lemanza, W., and Prinzl, F. (1983) Proc. 5th Congo ISRM (Melboume) p
BI-B8. ' p.
Problems
l. In a series of triaxial compression tests on a sandstone, the following repre-
sent the stresses at peak load conditions:
Test
U"3(MPa) U"(MPa)
1 1.0
9.2
2 5.0
28.0
3 9.5
48.7
4 15.0
74.0
Determine values of Si and cp that best fit the data.
2. !he initial state of stress at a point in the ground, in sandstone of Problem 1,
lS
<T3 = 1300 psi
<TI = 5000 psi
pore water (Pw) will be raised by the construction of a
reserVOIr. What value of Pw will cause fracture of the sandstone in situ
(assume fracture occurl) at peak stress).
3. the rock of l .. 2, what is the value of the principal stress
rabo (<T31<T1) aboye WhlCh faIlure" cannot occur?
9S Rack Strength and Failure Critera
4. In aplane wave front, the pressures in the plane of the wave front are
v/(1 - v) times the pressure normal to the wave front. What is the least
value of Poisson's ratio (v) such that compressive or shear failure does not
occur as the wave front sweeps through the rock? (See Problem 3.)
5. Triaxial compression tests of porous rock yield Si equal to 1. O MPa and cf>
equals 35. Calculate the unconfined compressive strength and estimate the
tensile strength for this rock.
6. In an area underlain by rock of Problem 5, a porous limestone, the in situ
stresses at a given point near a reservoir site are (TI = 12 MPa, and (T3 = 4
MPa. How deep a reservoir must be built before the pore pressure increase
from a corresponding rise in the groundwater levels could fracture the
rock? (Express your answer in MPa and psi units for the required increase
in water pressure and meters or feet for the reservoir depth.)
7. If a bilinear failure criterio n is used with a tension cutoff superimposed on
the Mohr-Coulomb criterion, at what value of (T (expressed in terms of T
o
,
Si, and cf are shear failure and tensile failure criteria satisfied simulta-
neously?
8. (a) For the sandstone of Figure 3.10, determine the value of cf>p and Sip that
best fits the peak strengths given. (The subscript p denotes "peak.")
(b) Find best fit values of cf>r and Sir describing the residual strength of the
sandstone.
(c) Find values of M and N to express the peak strength of the sandstone in
Figure 3.10 according to Equation 3.15.
(d) Find cf>p and Sip for the norite of Figure 3.10.
(e) Find cf>r and Sir for the norite.
(1) Find M and N for the norite.
9. Derive an expression for the modulus of ruptur T MR for a test with three-
point bending of a core sample (circular cross section).
10. Show that the volume change per unit of volume (a V/V) of a rock core
undergoing compression is approximately equal to the sum of the three
normal strains.
11. (a) Plot Mohr envelopes of strength for (1) 1/1 = 0, (2) 1/1 = 30, (3) 1/1 = 60,
and (4) 1/1 = 90 for the slate whose strength anisotropy is described by
Equation 3.18a and 3.19a.
(b) Plot the peak compressive strength of the slate as a function of 1/1 for
(1) (T3 = O and (2) (T3 = 30 MPa.
12. A set ofunconfined compressive strength tests on slate specimens oriented
with 1/1 = 30 and 1/1 = 75 yields strength values qu,30 and qu,75 , respectively.
Show that the directional strength of the rock can be approximated by
Problems 99
(TI,P = (T3 tan
2
(45 + *) + qu,75 - (Qu,75 - Qu,30)COS 2(1/1 - 30)
13. It has been observed that the degree of anisotropy, as expressed by the
ratio of maximum to minimum strength for aH directions through a rock,
decreases as the confining pressure increases. What explanation can you
offer?
14. The four-point loading configuration for the modulus-of-rupture test is de-
sirable because it generates uniform moment with zero shear between the
inner load points. Can you find another beam testing configuration such that
the central portio n of the rock beam receives zero moment with constant
shear?
15. (a) Derive an empirical expression similar to Equation 3.15 appropriate for
residual strength.
(b) Find values for the constants M and N fitting the residual strength data
for the sandstone in Figure 3.10.
(c) Do the same for the residual strength of the norite.
16. E. Hoek and E. T. Brown (1980) introduced the empirical criterion of
failure for rocks:
(TI,p (T3 ((T3 )112
--=-+ m-+s
qu qu qu
where m and s are constants;
(
qu rock mass )2
s = qu rock substance
(a) Compare this with Equation 3.15 for the case m = O and s = 1.
(b) Hoek and Brown, in studying many sets of data, found the foHowing
approximate values: m = 7 for carbonate rocks, m = 10 for argillaceous
rocks, m = 15 for sandstone and quartzite, m = 17 for volcanic rocks,
and m = 25 for granitic and other intrusive rocks. Compare the peak
compressive strengths as a function of confining pressure for a marble, a
rhyolite, and a granite, aH having qu = 100 MPa. What is the physical
meaning of parameter m?
17. Equation 3.15 was generalized by Yudhbir et al. (1983) by replacing the
unity on the right side by the constant A. Permitting A to vary between 1
and O offers a continuous variation of rock mass quality in the criterion of
failure. They propose linking Ato Barton's Q (see Chapter 2) according to
the relation A = 0.0176QM.
(a) Find a corresponding relationship between A and RMR.
(b) Based on the answer to part (a), use Equation 3.15 to express the peak
major principal stress as a function of confining pressure for a weak
sandstone with M = 0.65, N = 5, Qu = 2.0 MPa, and RMR = 50.
Initial Stresses in
Rocks and Their
Measurernent
Chap
ter
4
Any undisturbed mass of rock in situ contains nonzero stress components due
to weight of overlying materials, confinement, and past stress history. Near the
surface in mountainous regions the in situ stress may approach zero at sorne
points or le c10se to the tock strength at others. In the former case, rocks may
fall from surface and underground excavations because joints are open and
weak; in the latter case, disturbance of the stress field by tunneling or perhaps
even surface excavation may trigger violent release of stored energy. This
chapter concerns determination of the magnitude and direction of the initial
stresses at the site of a work.
4.1 lnfluence o/ the lnitia' Stresses
It is often possible to estimate the order of magnitude of stresses and their
directions, but one can never be certain of the margin of error without backup
measurements. Application of such measurements is fairly common in mining
practice, but since stress measurements tend to be expensive they are not
routine for civil engineering applications. There are several civil engineering
situations, however, when knowledge of the state of stress can be helpful or
lack of knowledge might prove so costly that a significant stress measurement
program is warranted. For example, when choosing the orientation for a cav-
ern, one hopes to avoid aligning the long dimension perpendicular to the great-
est principal stress. Ifthe initial stresses are very high, the shape will have to be
selected largely to minimize stress concentrations. Knowledge ofrock stresses
also aids in layout of c,omplex underground works. An underground power-
JIU
(J
102 lnitial Stresses in Hocles and Their Measurement
house for example, consists of a three-dimensional array of openings including
a machine hall, a transformer gallery, low-voltage lead shafts, pressure tunnels,
surge shafts, rock traps, access tunnels, ventilation tunnels, muck hauling tun-
neIs, penstocks, draft tubes, and other openings. Cracks that initiate at one
opening must not run into another (Figure 4.1a). Since cracks tend to extend in
the plane perpendicular to 0'3 knowledge of the direction of the stresses permits
choosing a layout to reduce this risk. Pressure tunnels and penstocks can be
constructed and operated in rock without any lining if virgin stress is greater
than the internal water pressure, so for such applications stress measurement
might permit large cost savings. When displacement instruments are installed
in an underground or surface excavation, to monitor the rock performance
during construction and service, stress measurements beforehand pro vide a
framework for analysis of the data and enhance their value. When making large
surface excavations with presplitting techniques, economies will be realized if
the excavation is oriented perpendicular to 0'3 (Figures 4.1b,c). With under-
ground storage of fluids in reservoir rocks, knowledge of the initial state of
stress will help evaluate the potential hazard of triggering an earthquake.
These are a few examples of situations in which a knowledge of the state of
stress can be integrated in engineering designo In a more general sense, how-
ever, the state of stress can be considered a basic rock aUribute whose magni-
tudes and directions affect the overall rock strength, permeability, deformabil-
ity, and other important rock mass characteristics. Thus it is rarely irrelevant to
know the initial stress state when dealing with rock in situ.
Sometimes initial stresses are so high that engineering activities can trigger
rock failure. Whenever the major stress in the region of an excavation is more
than about 25% of the unconfined compressive strength, new cracking can be
expected as a result of construction no matter how carefully it is performed.
This derives from two observations: (1) the maximum stress concentration
around an underground opening cannot be less than 2; and (2) cracking occurs
in an unconfined compression specimen when the stress reaches about half of
the unconfined compressive strength. Close to steep valley sides, where the
angle from the excavation to the mountain top is greater than 25, data show
that rock stress problems tend to occur in Norwegian fjord country whenever
the weight of rock cover is greater than about 0.15qu (Brekke and Selmer-
Olsen, 1966; Brekke, 1970). Such stress problems can vary from slabbing and
overbreak of rock on the tunnel wall nearest the valley side, to isolated violent
detachment of rocks from the walls or even destructive bursts. Conditions for
rock bursts are found underground in deep mines, as in the Canadian Kirkland
Lake District, the South African gold mines, and the Idaho Coeur d'Alene
district, where mining is pursued at depths of as much as 11,000 ft. In civil
engineering work, in addition to the valleyside stress problem noted, railroad
and road tunnels under high mountains, such as, the Mont Blanc Tunnel in the
Alps, have encountered severe rock stress problems. In shales and other rocks
M
"
r---
I
I
I ~
I
~
iii
I (;)
I
I
L
---1
I
I
I
I
I
I
I
___ J
M
"
1::
o
o:;
"
Q)
IJ)
Q)
u
. .:
u
~
~
c..
e
o
e
.9
-
u
. ~
"O
'"
'"
<1)
.b
'"
"O
<1)
u
e
<1)
::l
~
<1)
..c
- '+-<
o
'"
i
~
<1)
<1)
e
o
S
CIl
""1
.,,;
i
104 lnitial Stresses in Rocks and Their Measurement
with low values of qu, conditions for rock failure due to concentration of initial
stress may lead to slow compression ("squeeze") and destruction of tunnel
supports rather than violent collapse, but difficulties can still be significant. The
"stand-up time" of a tunnel, that is, the maximum duration for erecting sup-
ports, is closely related to the ratio of maximum initial stress to qu.
4.2 Estimating the lnitial Stresses
VERTICAL STRESS
It is generally safe to assume that the vertical normal stress is equal to the
weight of the overlying rock, 0.027 MPa/m or 1.2 psi/ft on the average. Near
horizontal ground, the principal stress directions are vertical and horizontal. It
is often assumed that they are also vertical and horizontal at depth (Figure
4.2a); however, this isjust an assumption to reduce the number ofunknowns,
an assumption that finds reinforcement in Anderson's observations that normal
and reverse faults often dip at 60 and 30, respectively (see Jaeger and Cook,
1976). The simplifying assumption that the principal stresses are vertical and
horizontal has been widely adopted in practice. Of course, this breaks down at
shallow depths beneath hilly terrain, because the ground surface, lacking nor-
mal and shear stresses, always forms a trajectory of principal stress (Figure
4.2). Beneath a valley side, one principal stress is normal to the slope and
equals zero, while the other two principal stresses lie in the plane of the slope
1
(b)
(a)
Figure 4.2 The inftuence of topography on initial stresses.
4.2 Estimating the lnitial Stresses 105
(Figure 4.2b). These stresses likewise approach zero where the rock slope is
convex upward but grow larger where the slope is concave upward. Beneath
the sharp notch of a V -shaped valley, the in situ stresses may be close to or at
the strength of the rock.
Over any significant horizontal surface within the ground, the average
vertical stress must equilibrate the downward force of the weight of overlying
rock, hence the rule stated previously:
(Tu = yZ (4.1)
where (Tu is the average total vertical stress at depth Z in rock with unit weight
y. This rule has been supported by numerous measurements (Figure 4.7a) and
is one of the reliable formulas of stress in situ. However, it can be violated over
limited horizontal distances owing to effects of geological structure. Figure 4.3,
for example, shows how the vertical stress might vary along horizontal planes
cutting through a succession of rigid and compliant beds folded into synclines
and anticlines. Along line AA' the stress varies from perhaps 60% greater than
yZ under the syncline to zero just beneath the anticline, the more rigid layer
serving as a protective canopy and directing the ftow of force down the limbs of
the fold. A tunnel driven along line BB' could expect to pass from relatively
understressed rock in the compliant shales to highly stressed rock as it crossed
z
N
"
Figure 4.3 The 'inftuence of folds in heterogeneous, layered rock
on vertical stresses.
106 lnitial Stresses in Rocks and Their Measurement
into more rigid sandstone in passing under the trough ofthe syncline. Ifthere is
a low-strength sheared zone along the contact, produced by slip between the
layers during folding, the vertical stress could be expected to jump in crossing
the contact. Since geological structure can alter the vertical stresses and the
direction of principal stresses, it is wise to investigate geological effects
through analysis in important applications wherever geological heterogeneities
can be expected to deflect the lines of force away from the vertical. Figure 4.4
shows the result of one such analysis, performed using the finite element
method, in a region with heterogeneous geology superimposed on a sharply
notched topography.
HORIZONTAL STRESS
In regard to the magnitude ofthe horizontal stresses, it is convenient to discuss
the ratio of horizontal to vertical stresses. Let
K = ITh
lT
u
(4.2)
In a region ofrecent sedimentation, such as the Mississippi Delta, the theory of
elasticity can be invoked to predict that K will be equal to v/(l - v). This
expression derives from the symmetry of one-dimensionalloading of an elastic
material over a continuous plane surface, which infers a condition of no hori-
zontal strain; such a formula has no validity in a rock mass that has experienced
cycles of loading and unloading. Consider an element of rock at depth Zo with
initial value of K = K
o
, which is then subjected to unloading by remo val of !l.Z
thickness of overburden (Figure 4.5). Due to unloading of y!l.Z vertical stress,
the horizontal stress is reduced by y!l.Zv/(l - v). Therefore, after erosion of a
thickness of rock equal to !l.Z, the horizontal stress at depth Z = Zo - !l.Z will
become equal to KoYZo - y!l.Zv/(l - v), and
K(Z) = Ko + [( Ko - 1 J !l.zJ
Z
(4.3)
Thus, erosion of overlying rock will tend to increase the value of K, the hori-
zontal stress becoming greater than the vertical stress at depths less than a
certain value.
1
The hyperbolic relationship for K(Z) predicted by Equation 4.3
can be generated by other arguments. While the vertical stress is known to
equal yZ, the horizontal stress could lie anywhere in the range of values be-
tween the two extremes KalT u and KplT u shown in Figure 4.6. Ka corresponds to
conditions for normal faulting, Figure 4.6b, in which the vertical stress is the
With the restriction K :s; Kp given by (4.5). Thermal effects have been ignored.
4.2 Estimating the lnitial Stresses 107
900
r--------------400-------
-------------500----
______ _
=
1000
1100---__
L.LJ
Scale
(al
(b)
Figure 4.4 Comparison of maximum shear stresses beneath valleys in
(a) and heterogeneous (b) formations. Units of shear stress are hundreds of pouo s
per square foot.
lOS lnitial Stress es in Rocks and Their Measurement
Figure 4.5 The effect of erosion on stresses at depth.
major principal stress and failure is by horizontal extension. Assuming Cou-
lomb's law
(4.4)
Kp corresponds to conditions for reverse faulting (Figure 4.6c), in which the
vertical stress is the minor principal stress and failure is by horizontal compres-
sion, giving
K = tan
2
(45 + + qu . .l
p 2 Y Z
(4.5)
Values of these extreme horizontal stresses are tabulated for an assumed set of
rock properties in Table 4.1. If there is no existing fault, we observe that the
range of possible values of K such that Ka ::; K ::; K
p
is quite vast. However,
near a preexisting fault, qu can be assumed equal to zero and the range of K is
considerably reduced. Although tension is possible, it has rarely been mea-
sured and is to be considered an unusual situation.
Brown and Hoek (1978) examined a number of published values of in situ
4.2 Estimating the lnitial Stresses 109
T
/
/
\ /
\/
7'
/
/
(a)
(b)
(e)
---..................
"" \
\
\
\
>-----r----<.Kpov
Figure 4.6 Stresses required to initiate normal and reverse faults.
stress (Figure 4.7b) and independently discerned a hyperbolic relation for the
limits of K(Z), as
100 - 1500
0.3 + Z < K < 0.5 + Z
(4.6)
where Z is the depth in meters and K is the average to
vertical stress. The range in extreme values of K given by thls empmcal cnte-
Table 4.1 El'treme Values for Possible Horizontal Stresses Corresponding
to Conditions for Normal and Reverse Faulting 'Y = 25.9 kN/m
3
Depth
(m)
10
20
40
60
100
150
200
400
750
1000
2000
""""' ...
3 "1
a 'f'
...
=""'\:1:1
>
... .,::;-
.....
o (JC o
...,
ION",
-..lO",
('p _
q=
.,
'"
'" = ...
,-..",
::,
""'
=""'
Q.<
o'
a
Before Faulting Occurs;
No Preexisting Fault
Horizontal Stress U'h
qu l3.8 MPa
cJ>
40
Vertical
Stress Normal Reverse
U'v Faulting Faulting
(MPa) (MPa) (MPa)
0.26 -2.94 14.99
0.52 -2.88 16.18
1.04 -2.77 18.56
1.55 -2.66 20.95
2.59 -2.43 25.72
3.89 -2.15 31.68
5.18 -1.87 37.64
10.36 -0.74 61.49
19.43 1.23 103.2
25.90 2.64 l33.0
51.80 8.28 252.4
Depth Z. m
N I'\l - -
U' "'" o (,]1
8 o o o
o o o 10
o
U' ....
=. .. 4",
"'. == O - 41..\ lo
4 _. tH , \
-, @ ...,..., \
<O ero
4
= ...,
t--r"t..- tH 4 1.. .....
1
_
--.......... 4\;;:! . : .. -.. !1, U'
.... N 4.
"'\. ..., J
ti) "',. ,
1
- - '- .,.. ii 11 N
N 8 , re j1 o
. - + '" \ ..
O \ o o ,
. 1- \. o ..JJ N
tH
+
9
o N \: ,
N o I
<.JI
\ '"
\ TI o
\
\
4 ,
44
1
qu
cJ>
Normal
Faulting
(MPa)
-0.85
-0.73
-0.47
-0.22
0.29
0.92
1.56
4.10
8.54
11.72
24.42
'"
,
<
2 MPa
20
Reverse
Faulting
(MPa)
2.53
3.06
4.11
5.17
7.28
9.92
12.57
23.l3
41.62
54.83
107.6
After Faulting Has Occurred
and a Fault Exists
Horizontal Stress U' h
qu O qu O
cJ>
40
cJ>
20
Normal Reverse Normal Reverse
Faulting Faulting Faulting Faulting
(MPa) (MPa) (MPa) (MPa)
0.06 1.19 0.l3 0.53
0.11 2.38 0.25 1.06
0.23 4.76 0.51 2.11
0.34 7.15 0.76 3.17
0.56 11.91 1.27 5.28
0.84 17.87 1.90 7.92
1.13 23.82 2.54 10.57
2.25 47.64 5.08 21.13
4.22 89.33 9.52 39.62
5.63 119.1 12.70 52.83
11.26 238.2 25.40 105.6
Depth Z. m
8
o
o
9g>g>
o e
.. ::> -+
Q.
:;
.. O ::>
'!? <
g O !!
..
4
O
N
<1'
o
o
4 r
e (') l> i
(/) g ::J
l>[ Q.
o
@
4
...,
'"
8
,
.. %.- r
O ;;:
'4iI ::.
I g
'" -
O !!'.
ii
g:
q
.. <
O -
O
O
'"
...,
N
O
" O
t..:I
S-
5r
S'
!
1
:
...
...
...
112
4.:l Estimating the Initia' Stresses 113
rion is considerably less than the range Ka to Kp given by (4.4) and (4.5) when qu
is not equal to zero, due in part to the fact that average horizontal stress is being
considered, whereas the previous criteria refer to maximum and minimum
values of horizontal stress. In any event, aH the equations for K(Z) presented
and the actuaHy measured data are consistently found to be inverse with Z.
Thus, even without measurements one can estimate, within broad limits, the
variation of horizontal stress with depth. While the magnitude of the horizontal
stress might be estimated only approximately, it is often possible to offer good
estimates for the directions of the horizontal stresses.
HORIZONTAL STRESS DIRECTION
If the present state of stress is a remnant of that which caused visible geological
structure, it will be possible to infer the directions of stresses from geological
observations. Figure 4.8 shows the relationship between principal stress direc-
tions anddifferent types of structures. The state of stress that causes a normal
fault has (TI vertical, and (T3 horizontal pointed perpendicularly to the fault
trace as seen in the plan. In the case of reverse faulting, the stresses that caused
the rupture have (T3 vertical, while (TI is horizontal and directed perpendicular
to the fault trace. Axial planes of folds also define the plane of greatest principal
stress. Strike-slip faults are created by a state of stress in which (TI is horizontal
and inclined about 30 with the fault trace, clockwise or counterclockwise as
dictated by the sense of motion on the fault. These directions of horizontal
stresses are not those of crustal block s caught and squeezed between pairs of
paraHel f a u l ~ such block s , the primary stress state ofthe crust that is linked
directly to the primary rupture surfaces will have superimposed on it the effects
of the strain from accumulated fault motions, as discussed by Moody and Hill
(1956).
Another line of observations comes from dikes and flank volcanoes formed
around larger craters. Sorne dikes represent hydraulic fractures, in which case
they He perpendicular to (T3. The perpendicular to a radius from a master crater
to a flank volcano similarly identifies the direction of least horizontal stress.
2
Seismologists are able to indicate the directions of primary stresses from first
motion analysis of earthquakes. If the directions of the vectors from the focus
to different seismic stations are plotted on a stereographic projection of a unit
reference hemisphere, it will be seen that two regions contain vectors to sta-
tions that received compressive first motion, while the other two regions con-
tain vectors that received extensile first motion (Figure 4.81). Two great circles
are drawn to divide these fields and their point of intersection defines the
direction of (T2. The direction of (TI is 90 from the direction of (T2 approxi-
2 K. Nakamura (1977) VoJcanoes as possible indicators of tectonic stress orientation-PrinC-
pIe and proposal. J. Volcanol. Geothermal Res. 2: 1-16.
"
---
M
"
~
t
e
~
Cl
e
~
Cl
c.
::J
/ "
1
M
"
I
1
e
O
. ~
E
e
O
.;;
e
l!l
x
w
t
e
O
. ~
'"
ti
E
O
u
4.3 Techniques for Measurement of ln-Situ Stresses 115
mately along the great circle bisecting the angle between the dividing great
circles in the extension first motion field. The direction of (T3 is the perpendicu-
lar to the plane of (TI and (T2' (Stereographic projection principIes are presented
in Appendix 5.)
Another approach to determining stress directions comes from the Occur-
rence of rock breakage on the walls of wells and boreholes, which tends to
create diametrically opposed zones of enlargement, termed "breakouts."
These features can be seen in caliper logs, photographs, and televiewer logs of
boreholes and have been found to be aligned from hole to hole in a region.
Haimson and Herrick (1985) reported experimental results confirming that
breakouts occur along the ends of a borehole diameter aligned with the least
horizontal stress as depicted in Figure 4.8g.
Directions of horizontal stresses in the continental United States, inferred
from a variety oftechniques, are shown in Figure 4.7c, prepared by Zoback and
Zoback (1988). This map also indicates the styles of deformation, that is, exten-
sion witld:e least principal stress horizontal or contraction with the greatest
principal stress horizontal.
4.3 Techniques fur Measurement
uf In-Situ Stresses
Stresses in situ can be measured in boreholes, on outcrops, and in the walls of
underground galleries as well as back calculated from displacements measured
underground. The available techniques summarized in Table 4.2 involve a vari-
ety of experimental approaches, with an even greater variety of measuring
tools. Three ofthe best known and most used techniques are hydraulicfractur-
ing, theflatjack method, and overcoring. As will be seen, they are complemen-
tary to each other, each offering different advantages and disadvantages. All
stress measurement techniques perturb the rock to create a response that can
then be measured and analyzed, making use of a theoretical model, to estimate
part of the in situ stress tensor. In the hydraulic fracturing technique, the rock
is cracked by pumping water into a borehole; the known tensile strength of the
rock and the inferred concentration of stress at the well bore are processed to
yield the initial stresses in the plane perpendicular to the borehole. In the flat
jack test, the rock is partIy unloaded by cutting a slot, and then reloaded; the in
situ stress normal to the slot is related to the pressure required to n u ~ l the
displacement that occurs as a result of slot cutting. In the overcoring test, the
rock is completely unloaded by drilling out a large core sample, while radial
displacements or surface strains of the rock are monitored in a central, parallel
borehole. Analysis using an unloaded thick-walled cylinder model yields stress
in the plane perpendicular to the borehole. In each case stress is inferred, but
Table 4.2 Methods for Measuring the Absolute State of Stress in Rocks
PrincipIe
Complete strain
relief
Partial strain
relief
Procedure
Overcore a radial defor-
mation gage in a central
borehole (D. S. Bureau of
Mines method)
Overcore a soft inclusion
containing strain gages
(LNEC and CSIRO
methods)
Overcore a borehole with
strain gages on its walls
(Leeman method)
Drill around a rosette
gage placed on a rock
face
Overcore a rosette gage
placed on the bottom of a
drill hole (doorstopper
method)
Overcore a soft photoelas-
tic inclusion
Measure time dependent
strains on faces of a rock
after its removal from the
ground
Null displacements caused
by cutting a tabular slot in
a rock wall (flat jack
method)
Overcore a stiff photo-
elastic inclusion with
down-hole polariscope
(glass stress meter)
Overcore a stiff inclusion
to freeze stresses into it;
measure frozen streses in
the laboratory (cast inclu-
sion method)
Overcore a stiff instru-
mented inclusion (stiff
inclusion method)
Drill in the center of a
rosette array on the sur-
face of a rock face (under-
coring method)
Monitor radial displace-
ments on deepening a
borehole (borehole deep-
ening method)
Reference
Merrill and Peterson (1961)
Rocha et al. (1974),
Worotnicki and Walton (1976)
Leeman (1971),
Hiltscher et al. (1979)
Olsen (1957)
Leeman (1971)
Riley, Goodman, and
Nolting 1977)
Emery (1962)
Voight (1968)
Bernede (1974)
Rocha et al. (1966)
Roberts et al. (1964, 1965)
Riley, Goodman, and
Nolting (1977)
Hast (1958)
Nichols et al. (1968)
Duvall, in Hooker et al.
(1974)
De la Cruz and Goodman
(1970)
4.3 Techniques for Measurement of ln-Situ Stresses 117
Rock flow or
fracture
Correlation
between rock
properties
and stress;
other
techniques
Measure strain to fracture
a borehole with a bore-
hole jack (Jack fracturing
technique)
Measure water pressures
to create and extend a
vertical fracture in a bore-
hole (Hydraulic fractur-
ing)
Measure strains that accu-
mulate in an elastic inclu-
sion placed tightly in a
viscoelastic rock
Core disking-observe
whether or not it has
occurred
Resistivity
Rock noise (Kaiser effect)
Wave velocity
X-ray lattice spacing mea-
surements in quartz
Dislocation densities in
crystals
De la Cruz (1978)
Fairhurst (1965)
Haimson (1978)
Obert and Stephenson (1965)
Kanagawa, Hayashi, and
Nakasa (1976)
Friedman (1972)
displacements are actuaHy measured. Precisions are seldom great and the
results are usually eonsidered satisfactory if they are intemaHy consistent and
yield values believed to be correet to within about 50 psi (0.3 MPa). The main
problem of aH stress measurement teehniques is that the measurement must be
conducted in a regon that has been disturbed in the process of gaining aeeess
for the measurement; this paradox is handled by accounting for the effect of the
disturbance in the analytical technique, as shown below.
HIDRAULIC FRACTURING
The hydraulic fracturing method makes it possible to estimate the stresses in
the rock at considerable depth using boreholes. Water is pumped into a section
of the borehole isolated by packers. As the water pressure increases, the initial
compressive stresses on the walls of the borehole are reduced and at sorne
points become tensile. When the stress reaches -T
o
, a crack is formed; the
down-hole water pressure at this point is Pct (Figure 4.9a). Ifpumping is contin-
118 lnitial Stresses in Rocks and Their Measurement
Figure 4.9 Hydraulic fracturing. (a) Pressure versus time data as water is
pumped into the packed-off section. (b) Experiment in progress. (Photo by
Tom Doe.)
ued, the crack will extend, and eventually the pressure down the hole will faH
to a steady value p., sometimes called "the shut-in pressure."
To interpret the data from the hydraulic frac:turing experiment in terms of
initial stresses, we need to determine the orientation of the hydraulically in-
duced fracture ("hydrofac"). The greatest amount of information coincides
with the case of a vertical fracture, and this is the usual result when conducting
4.3 Techniquesfor Measurement of ln-Situ Stresses 119
tests below about 800 m. The orientation of a fracture could be observed by
using down-hole photography or television; however, a crack that closes upon
depressuring the hole to admit the camera would be difficult to see in the
photograph. It is better to use an impression packer, such as one available from
Lynes Company, which force s a soft rubber lining against the waH while inter-
nal pressure is maintained, recording the fracture as an impression on the
rubber surface.
The analysis of the pressure test is simplified if it is assumed that penetra-
tion of the water into the pores of the rock has little or no effect on the stresses
around the holeo Making such an assumption, it is possible to use the results of
the known distribution of stress around a circular hole in a homogeneous,
elastic, isotropic rock (the "Kirsch solution") to compute the initial stresses at
the point of The tangential stress on the wall of the hole reaches the
least magnitude at A and A I (Figure 4.10) where it is
U" 8 = 3U" h.mn - U" h.max (4.7)
When the water pressure in the borehole is p, a tensile stress is added at aH
points around the hole equal (algebraically) to -p. _The conditions for a new,
vertical tensile crack are that the tensile stress at point A should become equal
to the tensile strength - TIJ. Applying this to the hydraulic fracturing experi-
ment yields as a condi!ion for creation of a hydraulic fracture
Plan
view
A'
3U"h,mn - U"h,max - Pel = -To
I I
..
..
----...... Oh.max
..
Figure 4.10 Location of critical points around the borehole used for
hydraulic fracture.
(4.8)
120 Initia' Stresses in Rocks and Their Measurement
Once formed, the crack will continue to propagate as long as the pressure is
greater than the stress normal to the plane of the fracture. If the pressure of
water in the crack were less than or greater than the normal stress on this
crack, it would close or open accordingly. In rocks, cracks propagate in the
plane perpendicular to <T3 In the context of hydraulic fracturing with a vertical
fracture, this means that the stress normal to the plane of the fracture is equal
to the shut-in pressure Ps:
<Th,min = Ps
(4.9)
Equations 4.8 and 4.9 allow the major and minor normal stresses in the plane
perpendicular to the borehole to be determined if the tensile strength of the
rock is known. If the borehole pressure is dropped and once again raised aboye
the value Ps, the hydraulic fracture will close and then reopen. Let the new
peak pressure, smaller than PcI , be called Pe2. Replacing T
o
and PcI of Equation
4.8 with the values O and Pe2, respectively, and subtracting Equation 4.8 from
the resulting equation yields a formula for the tensile strength of the rock
around the borehole applicable to the conditions of the experiment:
To = PcI - Pe2 (4.10)
Assuming that the vertical stress equals yZ, and is a principal stress, the state
of stress is now completely known, for the experiment yields the values and
directions of the major and minor normal stress in the plane perpendicular to
the borehole.
If the rock is pervious, water will enter cracks and pores creating an inter-
nal pressure gradient whereas the theory aboye presumed a sudden pressure
drop across the borehole wall. The effect is to lower the value of PcI and round
the peak of Figure 4.9. Haimson (1978) shows how to modify the analysis to
solve for the principal stresses in this case.
The hydraulic fracturing experiment does not yield the aboye results if the
fracture is horizontal. Conditions for propagation of a horizontal fracture are
met if the internal pressure beco mes equal to the vertical stress plus the tensile
strength. Assuming that the tensile strengths for propagation of horizontal and
vertical fractures are the same, the vertical fracture could form only at depths
below which the vertical stress obeys
<Tv ~ (3N - 1)<Th,max (4.11)
where N = <Th,minl<Th,max. To permit an estimate of the minimum depth for
vertical fracturing, it is useful to express Equation 4.11 in terms of K, the ratio
of mean horizontal stress to vertical stress. In these terms, a vertical fracture
will form at a depth such that K is less than (1 + N)/(6N - 2) where N is <Th,minl
<Th max (with N restricted to values greater than k). The minimum depths for a
vertical fracture, corresponding to the upper and lower limits of K(Z) given in
Equatitm 4.6, are presented for various Values of N in Table 4.3. When the
4.3 Techniques for Measurement of In-Situ Stresses 121
Table 4.3 Minimum Depths for a Vertical Hydraulic Fracture
er h,minl er h,max
(N)
:::;0.33
0.40
0.50
0.60
0.667
0.70
0.80
0.90
1.00
Transition Value
a
of K = hlerv
(K
T
)
00
3.5
1.5
1.0
0.833
0.773
0.643
0.559
0.500
Minimum Depth (meters) for a
Vertical Hydrofrac Assuming
z - ( 100 )
K - 0.3
o
31
83
143
188
211
292
386
500
z = ( 1500 )
K - 0.5
o
500
1500
3000
4505
5495
10,490
25,424
00
value of N is small, or when the mean horizontal stress tends toward the lower
values in the range of experience, vertical fractures can occur at shallow
depths. This has in fact been experienced by the oil industry, which has pro-
duced more than a million hydrofracs for artificial stimulation of oil and gas
wells.
THE FLAT JACK METHOD
Hydraulic fracturing can be performed only in a borehole. If one has access to a
rock face, for example, the wall of an underground gallery, stress can be
measured using a simple and dependable technique introduced by Ti!1celin in
France in 1952. The method involves the use offlat hydraulicjacks,consisting
of two plates of steel welded around their edges an,d a nipple for introducing oil
into the intervening space. Through careful welding and the use of preshaping
bends, or internal fillets, it is possible to achieve a pressure of 5000 psi or higher
in such a jack without rupture: The first step is to install one or more sets of
measuring points on the face of the rock. The separation of the points is typi-
cally 6 in., but must conform to the gage length of available extensometers.
Then a deep slot is installed perpendicular to the rock face between the refer-
ence points (Figure 4.l1b); this may be accomplished by drilling overlapping
jackhammer holes, by using a template to guide the drill, or by diamond sawing
(Rocha et al., 1966). As a result of cutting the slot, the pin separation will
decrease from do to a smaller value if the rock was under an initial compression
normal to the plane of the slot (Figure 4.11c). The initial normal stresses could
122 lnitial Stresses in Rocks and Their Measurement
e
O
';:;
'" c.
el
,<:
o..
<
(a)
Time
Jack
ttt
(e)
(b)
Jack pressure
os'" Pe
Figure 4.11 The flat jack test.
Pe
be calculated from the measured pin displacement if the elastic constants of the
rock were known. However, a self-compensating method of stress determina-
tion is preferred making it unnecessary to determine the rock properties explic-
itly. The flat jack is inserted into the slot, cemented in place, and
When the pins have been returned to do, their initial separation, the pressure In
the jack (Pe) approximates the initial stress normal to the jack. In theory, the
initial stress parallel to the slot and the geometric differences between the
inside of the jack and the inside of the slot require a correction to this result
(Alexander, 1960). However, the correction is often within. band of uncer-
tainty anyway, and if a diamond sawed slot is used, it is neghglbly .small; thus Pe
(the "canceIlation pressure of the jack") is an acceptable ,estlmate for the
average stress normal to the jack.
4.3 Techniques for Measurement of ln-Situ Stresses 123
In the flat jack test we have a large, rugged, and inexpensive method for
determining one stress component of the stress tensor. The equipment can be
fabricated on site and is virtually indestructible, an important consideration in
any instrumentation or measurement program underground. A serious limita-
tion of the method is that the measured stress les in the region of disturbance of
the gallery introduced for the purpose of taking the measurement. If the gallery
is carefully executed, this disturbance might be calculated by conducting an
independent stress concentration investigation, using numerical methods (e.g.,
the finite element method). In general, if the stresses normal to the plane of the
jack are determined at three points around the section of the gallery, yielding
values CT!IA, CT8B, CT8C for the tangential stresses (stresses parallel to the surface
of the opening) near the surface at these points, the initial stresses in the plane
perpendicular to the gallery can be calculated by inverting the relationship:
{
CT
8,A} _ (al! a12 a13){CT x}
CT8,B - a21 a22 a23 CTy
CT8,C a31 a32 a33 Txy
(4.12)
where the coefficients aij are determined by the numerical study. For example,
suppose flat jacks were placed at R and W, in the roof and side wall, respec-
tively, of a perfectly circular underground opening; if the initial stresses were
known to be horizontal and vertical, and if the tunnel radius were large com-
pared to the width of the jacks, then Equation 4.12 would simplify to
{
CT8,W} = (-1 _ 3) {CThOriZ}
CT8,R 3 1 CTver!
(4.13)
whereupon
and (4.14)
The stresses around an underground gaIlery vary inversely with the radius
squared (see Equations 7.1). Therefore, if stresses are measured in a borehole
at least one gaIlery diameter in depth, the results should correspond to the
initial state of stress before driving the measurement gallery. This can be ac-
complished using the overcoring test.
OVERCORING
First one drills a small-diameter borehole and sets into it an instrument to
respond to changes in diameter. One such instrument is the U. S. Bureau of
Mines six-arm deformation gage (Figures 4.12a and 4. 13a), a relatively rugged
124 lnitial Stresses in Rocks and Their Measurement
6.d
Delormation
gage
Large dril!
------- = = = t - - - - ~
Plane 01 buttons
..---
/ ""
( ~ \
\ ~ J
\ /
'-/
Distance drilled
Figure 4.12 The overcoring method, using the Bureau
of Mines gage.
tool that uses the bending of a cantilever equipped with strain gages to give
output voltage proportional to displacement. There are three opposed pairs of
carbon-carbide-tipped buttons, each pressing against a cantilever arm fixed to a
base plate, tightened against the wall ofthe borehole by a spring. By choosing a
button of appropriate size in each of the six positions, each of the cantilevers
can be pre-bent to yield an initial output in the center of the linear region and
the borehole diameter changes can be monitored along three diameters simulta-
neously, whether the borehole becomes smaller or larger. After the gage is
inserted, the output wires are threaded through a hollow drill and out through
the water swivel and a larger hole is cored concentrically over the first (Figure
4. 13b). This produces a thick-walled cylinder of rock, detached from the rock
mass and therefore free of stress. If the rock had been under an initial compres-
sion, the deformation gage will record an enlargement along two or all of the
monitored directions in response to the "overcoring" (Figure 4. 12b)-all radii
4.3 Techniques for Measurement of ln-Situ Stresses
125
Figure 4.l3 In-situ stress measurements by overcoring from a rock outcrop. (a)
Six component borehole deformation gage and the overcored measuring hole: (b)
experiment in progress. (Photos by Rick Nolting; Courtesy of TerraTek.)
126 lnitia' Stresses in Rocks and Their Measurement
expanding if the ratio of minor to major normal stress in the plane perpendicu-
lar to the borehole is greater than one-third. As a result of the experiment, the
change in borehole diameter will be known along three diameters, 60 aparto
Select an x axis convenientIy in the plane perpendicular to the hole and let () be
the angle counterelockwise from Ox to a pair of buttons that yields results
tld()). Let the plane perpendicular to the borehole be the xz plane, with the
borehole parallel to y. Then, the deformations measured are related to the
initial stresses in the xyz coordinate system according to
tld( ()) = (F JI + (F yl2 + (F J3 + T xzi4 (4.15)
1 - v
2
dv
2
where II = d(1 + 2 cos 2()) -E- + E
12=_dv
E
1 - v
2
dv
2
13 = d(1 - 2 cos 2()) -E- + E
1 - v
2
14 = d( 4 sin 2()) -E-
In the aboye, E is Young's modulus, vis Poisson's ratio, and d is the diameter
of the borehole in which the measurement is conducted. Equation 4.15 ex-
eludes the two shear stress components T xy and T zy parallel to the borehole
because these have no inftuence on the diameter of the borehole. Gray and
Toews (1968) showed that only three linearly independent equations are obtain-
able from repeated diametral measurements in different orientations, so the
general state of stress cannot be computed from diameter changes recorded in
one borehole. However, a solution can be found if one ofthe stress components
is known or can be assumed. If the measurement is conducted in a borehole
perpendicular to a rock face and at shallow depth, (F y might be taken as zero. If
the value of (Fy were known, or assumed, on the other hand, the terml2(Fy could
be taken to the left side of the equal sign in each of three equations representing
measurements along different directions and the remaining three stress compo-
nents could be determined. In this way, the state of stress in the plane perpen-
dicular to the borehole could be computed as a function of (Fy alone. An alterna-
tive approach, discussed later, is to combine measurements from three or more
nonperpendicular boreholes, adopting a single, universal coordinate system
into which the unknown stresses from each borehole are transformed. The
resulting set of equations will be redundant, and, furthermore, since it is impos-
sible to occupy the same volume of rock in more than one measurement, the
results will be scattered.
In the usual situation where measurements are conducted in one borehole
parallel to y, and a value of (F y is assumed for purposes of computation, diame-
4.3 Techniques for Measurement of ln-Situ Stresses 127
ter change measurements are conducted in directions ()I , ()I + 60, and ()I + 120,
yielding three equations in three unknowns:
{
tld( ()I) - Ji (F y } (111 113 f4){(F x}
tld()1 + 60) - 12(Fy = 121 123 124 (Fz
tld()1 + 120) - Ji(Fy 131 133 134 T
xz
(4.16) .
Inversion of Equations 4.16, after assuming a value for (Fy, yields the stress
components in the plane perpendicular to the borehole.
The overcoring test thus can be used to measure the stresses at sorne
distance from a rock face. There is a practicallimit to how far one borehole can
be drilled concentrically over another. With a template to collar the drillhole
and homogeneous, nonfractured rock, it might be possible to proceed for as
much as 30 m from a face; but normally the test has to be discontinued beyond
about 5 m.
The Swedish State Power Board has perfected the mechanical aspects of
overcoring and has succeeded in conducting Leeman-type triaxial measure-
ments at depths of more than 500 m. These tests are performed by cementing
strain gage rosettes to the walls of a 36-mm hole drilled exactly in the center of
the bottom of a 76-mm-diameter borehole. Extending the larger borehole over-
cores the former and strains the rosettes (Hiltscher, Martna, and Strindell,
1979; Martna, Hiltscher, and Ingevald 1983).
The principal disadvantage ofthe U. S. Bureau ofMines overcoring test is
the linear dependence ofthe stresses upon the elastic constants. The Bureau of
Mines determines E and v directIy on the overcore by compressing it in a
speciallarge-diameter triaxial compression chamber, while the borehole defor-
mation gage responds inside. Another approach, applicable in horizontal holes,
is to as sume a value for v and use the value of E that makes the vertical
component of stress, at sorne distance behind the wall, agree with the value of
the unit weight of rock times depth below ground. Another approach altogether
is to replace the deformation gage with a stiffer gage (e.g., glass or steel)
forming a "stiff elastic inelusion." In such a case, the stresses inside the
inclusion on overcoring are almost independent of the elastic modulus of the
rock. However, the precision of measurement is reduced making the experi-
ment more difficult. Another difficulty with the overcoring method is the re-
quirement to use large drill cores (e.g., 6-in.-diameter). There is no theoretical
demand that the outer diameter be any specific value, and, in fact, the stresses
deduced from the experiment will be unaffected by choice of outer diameter. In
practice, however, difficulty is experienced with rock breakage if the- outer
diameter is less than at least twice the inner diameter.
In the doorstopper method (Figure 4.14) strain gages are fixed to the center
of the stub of rock at the bottom of the hole which is then isolated from the
surrounding rock by continuing the original hole (Leeman, 1971). This permits
128 Initial Stresses in Rocks and Their Measurement
To bridge circuit
Insertion tool
Potting material (liquid rubber)
Strain rosette gage
Metal foil
z
Figure 4.14 In situ stress measurement by the "doorstop-
per" technique.
4.3 Techniques for Measurement of In-Situ Stresses 129
the interpretation of stress at greater depth but the interpretation of the data is
more precarious. The test is performed as follows. First, drill a borehole to the
site of measurement. Then install a flat, noncoring drill bit to grind the bottom
to a smooth flat surface. Clean the bottom surface and then cement onto it a
piece of metal foil bearing a strain gage rosette on its upper surface. When the
cement has hardened, thread the bridge wires through the drill and deepen the
holeo This releases the stresses in the bottom, yielding strains ex, e
z
, 'Yxz (with
the y axis parallel to the borehole and the x, z axes along two perpendicular
lines in the bottom, selected at will). Appendix 2 shows howto convert strain
rosette readings to strain components ex, ez, 'Yxz
The changes in stress at the bottom of the hole (da-x.B, da-y,B, dT xy,B) can be
calculated from the strain components by the stress-strain relationship for lin-
ear, elastic isotropic bodies:
[
1
da-xB
da-' = _E __ v
Ld 1-.' O
(4.17)
The initial stresses in x, y, z coordinates are related to the stress changes on
the bottom of the hole by
{ } ( )
{
a-
x
}
da-x,B a c b O
a-
da- = - b c a O y
dTx:': O O O d ~ :
(4.18)
Constants a, b, c, and d have been evaluated by several independent workers.
De la Cruz and Raleigh (1972) give the following values, based upon a finite
element analysis:
a = 1.30
b = (0.085 + 0.15v - v
2
)
c = (0.473 + 0.91v)
d = (1.423 - 0.027v)
(4.19)
As in the overcoring test, a-y must be assumed or evaluated independently.
Then
{
a-
x
} _ (a b 0)-1 {da-X'B .+ ca-
y
}
a-
z
- - b a O da-z,B + ca-y
T
xz
O O d dTxz,B
. (4.20)
The "doorstopper" method can be pursued at the bottom of a shaft as well
as in a drill holeo
130 Initial Stresses in Rocks and Their Measurement
Measuremeos Made Directly 00 the Rock Surface If a machine-bored
shaft or tunnel is available for rock mechanics work, stress measurements may
be made directly on the wall if the rock is not highly fractured. There are at
least two methods for doing this: overdrilling a strain gage rosette applied
directly to the rock surface, and drilling a central hole amid a set of measuring
. points (undercoring).
Strain gage rosettes applied to the rock surface have been used in boreholes
by Leeman (1971) with an ingenious device to transport, glue, and hold the
rosettes at several points simultaneously. Upon overcoring the hoie, these
rosettes then report strain changes that can be transformed to yield the com-
plete state of stress (o-)xyz. In the present context, we can overcore strain gage
rosettes cemented to points directly on the rock surface. Appendix 2 presents
formulas for calculating the state of strain (ex, e
z
, Yxz) from the readings of the
component gages of the rosette when the rock to which they are attached is
overcored. These strains can then be converted to stresses using (4.17).
Undercoring is a name applied by Duvall (in Hooker et al., 1974) to a
procedure for measuring stresses on an exposed surface by monitoring radial
displacements of points around a central borehole (Figure 4.15). Expressions
for the radial and tangential displacements of a point located at polar coordi-
nates r, () from the central hole of a radius a are given in Equations 7.2 for plane
strain; these expressions are changed to plane stress by substituting v/(1 + v) in
place of v as discussed in the derivation of Equations 7.1 and 7.2 (Appendix 4).
Equations 7.2 are developed for the condition where the major and minor
principal stress directions in the measuring plane are known. For the stress
measurement problem, these directions will not be known apriori so an arbi-
o
o o
o
Figure 4.15 Undercoring.
4.3 Techniques for Measurement of In-Situ Stresses 131
p(r, O)
l l r ~
Vo
~ __ __ ________ ~ X
Figure 4.16 Coordinate system for the
displacement equations.
trary choice ofaxes (x, z) is made (Figure 4.16). The stresses {a}xz ~ a ~ then be
determined from radial displacement measurements (ur ) at three poslhons (r, ())
using the following equation for each point in turn
3
:
1 a
2
where JI = 2E -; [(1 + v) + H cos 2())
1 a
2
J2 = 2E -; [(1 + v) - H cos 2())
1 a
2
13 = E -; (H sin 2())
a
2
H = 4 - (1 + v) r
2
(4.21)
With radial displacement, Ur,1 measured at rl, ()I, Ur,2 at r2, ()2, and U
r
,3 at r3, ()3,
Equation 4.21 yields
(4.22)
3 We assume that the tangential displacement v. does not influence the measured radial dis-
placement.
132 Initial Stresses in Rocks and Their Measurement
Figure 4.17 Coordinate systems for stress measurements
on the walls of a tunnel.
whic.h.can be inverted to determine the stresses. This method cannot yield good
perClSlon unless the measuring points are close to the surface of the central
hole, or the rock is deformable; otherwise, the values of U
r
will be quite small.
Duvall placed the measuring pins on a 10-in.-diameter circle and created a 6-in.
central hole by out an EX pilot holeo Vojtec Mencl used undercoring to
stresses m the toe of a landslide in soft rock,4 where measurable
dlsplacements were experienced despite relatively small stresses (0.6 MPa)
the value .of E was quite low. A variant of undercoring using a central
expanSlOn cell (dilato meter) to null the initial radial displacements of
the central hole was used by Dean, Beatty, and Hogan at Broken
HIll Mme, Australia.
5
Virgin stresses (the initial stresses at the test site befo re it was excavated)
can be calculated from the stress components measured on the rock walls if the
stress concentrations at the measuring points are known. The problem at hand
resembles that discussed previously in connection with the flat jack test. Since
the shape of the bored gallery is circular with smooth walls, the required stress
can be obtained from the classical Kirsch solution (whose deri-
vabon can be .followed in Jaeger and Cook (1976. For our purposes, the
adopted coordmate system is shown in Figure 4.17; unprimed coordinates
4 O. Zaruba and V. Mencl (1969) Landslides and Their Control, EIsevier, New York.
5 Rock stress measurements using cylindrical jacks and fiat jacks at North Broken Hill Ud.
Hill Mine Monography No. 3 (1968), Australian Inst. Min. Metal. Melbourne, Austra-
lia (399 Llttle Collins St.).
4.3 Techniques for Measurement of In-Situ Stresses 133
XY Z , X2Y2Z2, etc., refer to local coordinate directions at each measuring site
1, 2, etc., with YY2 . always in the direction of the normal to the surface
(radius ofthe tunnel or shaft) and XX2, etc., parallel to the axis ofthe measuring
tunnel or shaft. The x' , y', z' are global coordinates with y' parallel to the axis
ofthe shaft or tunnel, and x' and z' any convenient orthogonal axes in its cross
section.
The surface stress concentrations can then be obtained from the general
Kirsch formulas given by Leeman (1971) (see Appendix 4), substituting r = ato
identify points on the wall at the site of the measurement: with the aboye
coordinates, at each site
O'r = O'y = O
0'0 = O'z
O'long = O'x
TrO = Tyz = O
Tlong,O = T xz
Tlong,r = Txy = O
O e O O
1 h O O
O O n p
where d = 1 - 2 cos 2e
e = 1 + 2 cos 2e
f = -4 sin 2e
g = -2v cos 2e
h = 2v cos 2e
i = -4v sin 2e
n = -2 sin e
p = 2 cos e
O'x'
i)
O'y'
O'z'
Tx'y'
(4.23)
Ty'z'
Tz'x'
Two or more sites for surface stress measurement (e.g., (1) the roof e =
90, and (2) the wall e
2
= O (Figure 4.17) yield six equations whose solution.
determines the complete state of stress. Depending on the choice of sites, the
coefficient matrix might be singular, necessitating a third location (with redun-
dant data) to obtain a complete stress solution.
Principal Stresses If the stresses are determined with reference to two arbi-
trarily 'chosen directions x and z in the plane of measurement, the values of
normal stress will depend on the choice ofaxes. It is better to convert the
results to the form of principal stresses and directions. (If the xz plane is not a
principal plane, it is still possible to find, within it, directions in which the shear
134 Initial Stresses in Rocks and Their Measurement
stress is zero; these are then called "secondary principal stresses. ") Given {Ix,
{I z' and T
xz
, the principal stresses are found from
{Irnajor = !({Ix + (Iz) + [T;z + !({Ix - (Iz)2]l/2
and (4.24)
_ 1( + ) [2 + l( )2] 1/2
{Irninor - 2 {Ix {Iz - T
xz
4 {Ix - {Iz
The major principal stress, {II acts in a direction O, measured counterclockwise
from Ox, given by
tan 20 = (4.25)
{Ix - {Iz
Since the are tan is multivalued, we must observe the following rules.
6
Let ex =
tan-
I
[2T
xz
/({Ix - (Iz)] with -'TT/2 :5 ex :5 'TT/2 then
20 = ex
20 = ex + 'TT
20 = ex - 'TT
if {Ix> {Iz
if {Ix < {Iz and T
xz
> O
if {Ix < {Iz and T
xz
< O
Measurement o' Stresses in Three Dimensions Civil engineering and
mining work rarely require that all stress components be known. If such knowl-
edge is desired, methods exist to yield the complete state of stress from a single
experiment (e.g., Leeman, 1971; Rocha et al., 1974). AIso, data from tech-
niques enumerated aboye can be combined to permit computation of the com-
plete stress matrix. A procedure for doing this was already discussed for the
case of strain measurements on the surface of a drilled shaft or tunnel. Data can
also be combined for overcoring, doorstopper, and other approaches. In each
case, the strategy is to transform the measured stress components to a global
coordinate system to combine data from nonparallel directions at different
measuring sites.
For example, consider overcoring measurements in several nonparallel
boreholes. In hole A, adopt coordinate axes XA, YA, ZA with YA parallel to the
axis of the borehole; diametral displacements are measured in directions OAI,
OA2, and OA,3' Application of 4.15 to each direction yields
(4.26)
6 These rules were suggested to the writer by Professor Steven Crouch, University of Minne-
sota.
Rtiferences 135
where the coefficientsfij are defined for each O for jj of (4.15). Now transform
the stresses in XAYAZA coordinates to sorne convenient set ofaxes x', y', z'
(referred to henceforth as the "global axes"). This can be written
{
{Ix A}
{Iy,A = (Ter) {{I}x'y'Z'
{I (4 x 6) (6 x 1)
z,A
Txz,A
(4.27)
in which (Ter) is a 4 x 6 matrix corresponding to rows 1, 2, 3, and 5 of the
coefficient matrix defined in Equation 23 of Appendix 1, and {{I}x'y'Z' is the
column of the six stress components of the same equation. Let (fA) denote the
3 x 4 coefficient matrix in (4.26). Equations 4.26 and 4.27 can then be combined
into
3 XI 3x44x6 6 xl
(4.28)
Similarly for borehole B, nonparallel with A,
{tld}B = (fB)(Ter)o{{I}x'y'Z'
(4.29)
3 xl 3x44x6 6 xl
Combining the six rows of (4.28) and (4.29) gives six equations with {I x'y'Z'
as the right-hand vector. Gray and Toews (1968), however, showed that the
coefficient matrix thus derived is singular. Thus three nonparallel boreholes
will be required to yield sufficient information to solve for {{I}x'y'z" One can
reject surplus rows to achieve a solvable set of six equations. Even better, one
can use a least-squares solution scheme. Panek (1966) and Gray and Toews
(1975) showed how to handle the redundancy and minimize error associated
with variation in the state of stress from one measuring site to another.
A similar procedure can be followed to combine the results from "door-
stopper tests" in three nonparallel holes to determine the complete state of
stress.
References
Alexander, L. G. (1960) Field and lab. test in rock mechanics, Proceedings, Third
Australia-New Zealand Conference on Soil Mechanics, pp. 161-168.
Bemede, J. (1974) New Developments in the flatjack test (in French), Proc. 3rd Congo
ISRM (Denver), Vol. 2A, pp. 433-438.
Booker,.E. W. and Ireland, H. O. (1965) Earth pressures at rest related to stress history,
Can. Geot. J. 2: 1-15.
Brekke, T. L. (1970) A survey of large permanent underground openings in Norway,
Proceedings ofConference on Large Permanent Underground Openings, pp. 15-
28 (Universitets Forlaget, Oslo).
136 lnitial Stresses in Rocks and Their Measurement
Brekke, T. L. and Selmer-Olsen, R. (1966) A survey ofthe main factors influencing the
stability ofunderground construction in Norway, Proc. 1st Congo ISRM (Lisbon),
Vol. 11, pp. 257-260.
Brown, E. T. and Hoek, E. (1978) Trends in relationships between measured in situ
stresses and depth, Int. J. Rock Mech. Min. Sci. 15: 211-215.
De la Cruz, R. V. (1978) Modified boreholejack method for elastic property determina-
tion in rocks, Rock Mech. 10: 221-239.
De la Cruz, R. V. and Goodman, R. E. (1970) Theoretical basis ofthe borehole deepen-
ing method of absolute stress measurement, Proceedings, 11 th Symposium on Rock
Mechanics (AIME), pp. 353-376.
De la Cruz, R. V. and Raleigh, C. B. (1972) Absolute stress measurements at the
Rangely Anticline, Northwestem Colorado, Int. J. Rock Mech. Min. Sci. 9: 625-
634.
Emery, C. L. (1962) The measurement of strains in mine rock, Proceedings, Interna-
tional Symposium on Mining Research (Pergamon), Vol. 2, pp. 541-554.
Fairhurst, C. (1965) Measurement of in-situ stresses with particular reference to hydrau-
lic fracturing, Rock Mech. Eng. Geo!. 2: 129-147.
Friedman, M. (1972) X-ray analysis of residual elastic strain in quartzose rocks, Pro-
ceedings, 10th Symposium on Rock Mechanics (AIME), pp. 573-596.
Gray, W. M. and Toews, N. A. (1968) Analysis ofaccuracy in the determination ofthe
ground stress tensor by means of borehole devices, Proceedings, 9th Symposium
on Rock Mechanics (AIME), pp. 45-78.
Gray, W. M. and Toews, N. A. (1975) Analysis ofvariance applied to data obtained by
means of a six element borehole deformation gauge for stress determination, Pro-
ceedings, 15th Sympasium on Rock Mechanics (ASCE), pp. 323-356.
Haimson, B. C. (1976) Pre-excavation deep hole stress measurements for design of
underground chambers-case histories, Proceedings, 1976 Rapid Excavation and
Tunneling Conlerence (AIME), pp. 699-714.
Haimson, B. C. (1978) The hydrofracturing stress measurement technique-method and
recent field results., Int. J. Rock Mech. Sci. 15: 167-178.
Haimson, B. C. and Fairhurst, C. (1967) Initiation and extension of hydraulic fractures
in rock, Soco Petr. Eng. J. 7: 310-318.
Haimson, B. C. and Herrick, C. G. (1985) In situ stress evaluation from borehole
breakouts-Experimental studies, Proceedings, 26th U.S. Symposium on Rock
Mechanics (Balkema), pp. 1207-1218.
Hast, N. (1958) The measurement ofrock pressure in mines, Sveriges Geo!. Unders6kn-
ing Arsbok 52 (3).
Hiltscher, R., Martna, F. L., and Strindell, L. (1979) The measurement of triaxial
stresses in deep boreholes and the use of rock stress measurements in the design
and construction of rock openings, Proceedings 01 the Fourth International Con-
gress on Rack Mechanics, Montreux, (ISRM) Vol. 2, 227-234.
Hooker, V. E., Aggson, J. R. Bickel, D. L., and Duvall, W. (1974) Improvement in the
three component borehole deforrnation gage and overcoring technique, U.S.B.M.
Rep. Inv. 7894; with Appendix by Duvall on the undercoring method.
Jaeger, J. C. and Cook, N. G. W. (1976) Fundamentals al Rock Mechanics, 2d ed.,
Chapman & Hall, London.
References 137
Kanagawa, T., Hayashi, M., and Nakasa, H. (1976) Estimation of spatial geostress in
rock samples using the Kaiser effect of acoustic emission, Proceedings, 3rd Acous-
tic Emission Symposium (Tokyo; separately available from Central Research Inst.
of Elec. Power Ind., Japan).
Leeman, E. R. (1971) The CSIR "Doorstopper" and triaxial rock stress measuring
instruments, Rock Mech. 3: 25-50.
Lindner, E. N. and Halpem, J. A. (1978) In situ stress in North America-a compila-
tion, Int. J. Rock Mech. Sci. 15: 183-203.
Martna, J., Hiltscher, R., and Ingevald, K. (1983), Geology and rock stresses in deep
boreholes at Forsmark in Sweden, Proceedings olthe Fifth International Congress
on Rock Mechanics, Melboume, (ISRM), Section F, pp. 111-116.
Merrill, R. H. and Peterson, J. R. (1961) Deformation of a borehole in rock, U.S.B.
Mines R.I. 5881.
Moody, J. D. and Hill, M. J. (1956) Wrench fault tectonics, Bull. Geol. Soco Am. 67:
1207-1246.
Nichols, Jr., T. C., Abel, Jr., J. F., and Lee, F. T. (1968) A solid inclusion borehole
probe to determine three dimensional stress changes at a point of a rock mass, U. S.
G. S. Bulletin 1258C, pp. CI-C28.
Obert, L. and Duvall, W. (1967) Rock Mechanics and the Design 01 Structures in Rock,
Wiley, New York.
Obert, L. and Stephenson, D. E. (1965) Stress conditions under which core discing
occurs. Trans. Soco Min. Eng. 232: 227-234.
Olsen, O. J. (1957) Measurements of residual stress by the strain relief method, Qtly.
Colorado School 01 Mines, Vol. 52, July, pp. 183-204.
Panek, L. A. (1966) Calculation ofthe average ground stress components from measure-
ments of the diametral deformation of a drill hole, in ASTM Spec. Tech. Pub!. 402
(American Society of Testing and Materials). pp. 106-132.
Riley, P. B., Goodman, R. E., and Nolting, R. E. (1977) Stress by over-
coring cast photoelastic inclusions, Proceedings, 18th Symposium on Rock Me-
chanics, paper 4C4.
Roberts, A. et al. (1964, 1965) The photoelastic stress meter, Int. J. Rack Mech. Sci. 1:
441-454; 2: 93-103.
Rocha, M. (1971) A new method of integral sampling of rock masses, Rock Mech. 3: 1.
Rocha, M., Baptista Lopes, J., and DaSilva, J. (1966) A new technique for applying the
method ofthe flatjack in the determination of stresses inside rock masses, Proc. 1st
Congo ISRM (Lisbon), Vol. 2, pp. 57-65.
Rocha, M., Silverio, A., Pedro, J., and Delgado, J. (1974) A new development of the
LNEC stress tensor gauge, Proc. 3rd Congo ISRM (Denver), Vol. 2A, pp. 464-467.
Stauder, W. (1962) The focal mechanism of earthquakes, Adv. Geophys. 9: 1-75.
Terzaghi, K. and Richart, R. E. (1952) Stresses in rock about cavities. Geotechnique, 3:
57-90.
Tincelin, M. E. (1952) Measurement of pressure in the iron mines in the East:
(in.French), Supplement to Annales ITBTP, October.
Voight, B. (1968) Determination of the virgin state of stress in the vicinity of a borehole
from measurements of a partial anelastic strain tensor in drill holes, Rock Mech.
Eng. Geo!. 6: 201-215.
138 lnitial Stresses in Rocks and Their Measurement
Voight, B. and Sto Pierre, B. H. P. (1974) Stress history and rock stress, Proc. 3rd
Congo ISRM (Denver), Vol. 2A, pp. 580-582.
Worotnicki, G. and Denham, D. (1976) The state ofstress in the upper part ofthe earth's
crust in Australia according to measurements in mines and tunnels and from seis-
mic observations, Proceedings, Symposium on Investigation of Stress in Rock
(ISRM) (Sydney, Australia), pp. 71-82.
Worotnicki, G. and Walton, R. J. (1976) Triaxial "hollow inc1usion" gauges for determi-
nation ofrock stresses in situ, Proceedings, Symposium on Investigation of Stress
in Rock (Sydney), Supplement 1-8 (ISRM and Inst. of Engineers of Australia).
Zoback, M. D. and Healy, J. H. (1984) Friction, faulting, and "in situ" stress, Ann.
Geophys. 2: 689-698.
Zoback, M. D., Moos, D., and Mastin, L. (1985) Well bore breakouts and in situ stress,
J. Geophys. Res. 90: 5523-5530.
Zoback, M. L. and Zoback M. D. (1980) State of stress in the conterminous United
States, J. Geophys. Res. 85: 6113-6156.
Zoback, M. L. and Zoback, M. D. (1988), Tectonic stress field of the continental U.S.,
in GSA Memoir (in press), Geophysical Framework of the Continental United
States, edited by L. Pakiser and W. Mooney.
Problems
1. Estimate the vertical and horizontal stresses at a depth of 500 m in a zone of
normal faulting in Paleozoic sedimentary rocks. Use Figure 4.7b and as-
sume the normal faulting is recent.
2. A vertical hydraulic fracture was initiated in a borehole at a depth of 3000
ft. Assume the ground is saturated continuously from the surface and that
the pressure in the ground water is hydrostatic. The water pressure was
first raised 710 psi aboye the original groundwater pressure and then it was
not possible to raise it further. When pumping stopped, the water pressure
fell to a value 110 psi aboye the original groundwater pressure. After a day,
the pressure was raised again, but it could not be pumped to a value higher
than 100 psi aboye the previous pressure (the "shut-in" pressure). Estimate
the horizontal stresses at the site of measurement, the tensile strength of
the rock, and the vertical pressure at the site.
3. A borehole is drilled and instrumented and then overcored, yielding the
following values of the stress components in the plane perpendicular to the
borehole:
U'x = 250 psi
U'y = 400 psi
Txy = -100 psi
Problems 139
The x axis was horizontal and to the right. Find the magnitudes and direc-
tions of the major and minor principal stresses in the plane perpendicular to
the borehole.
4. A natural slope rises at 45 for 1000 m and then levels off. The rock has an
unconfined compressive strength of 50 MPa. A tunnel is to be driven for an
underground pressure pipe beginning with a portal at the base of the slope
and continuing directIy into the mountain. Based upon Norwegian experi-
ence, at what distance from the portal would you first expect to encounter
rock pressure problems?
5. In a zone of active thrust faulting (low-angle reverse faulting), in rock with
cp = 30, qu = 1000 psi, and unit weight of 150 Ib/ft3, estimate the major and
minor principal stresses at a depth of 3500 ft, assuming conditions for
faulting. Compare your estimate with that of Figure 4.7.
6. A rock mass at a depth of 5000 m had a value of K (= ratio of horizontal to
vertical stresses) equal to 0.8. If Poisson's ratio is 0.25, what should K
become after erosion of 2000 m of rock?
7. What form would the data of the flat jack test assume if the initial stress
normal to the plane ofthejack were tensile? How could the data be worked
to estimate the magnitude of the tensile stress?
8. Two flatjacks, 12 in. square, are placed in the wall and roof of an approxi-
mately circular test gallery 8 ft in diameter. Flat jack 1 is horizontal, and
placed in the side wall. Flat jack 2 is vertical, with its edge parallel to the
axis of the gallery. The cancellation pressures measured were 2500 psi with
FJ 1 and 900 psi with FJ 2. Estimate the initial stresses (vertical and hori-
zontal). List your assumptions.
9. The U. S. Bureau of Mines overcoring method is used to measure stresses
in a borehole drilled perpendicularly to a tunnel wall. The site of the mea-
surement (the plane of the measuring pins) is 5 ft deep in a test gallery 10 ft
in diameter. The measuring borehole has a diameter of 1.25 in. The first pair
ofbuttons is horizontal, pair 2 is oriented 60 counterclockwise from button
pair 1; button pair 3 is 120 counterclockwise from pair 1. Deformations .
were measured as a result of overcoring as follows: pair 1 moved outward
3 x 10-
3
in.; pair 2 moved outward 2 x 10-
3
in.; and pair three moved
outward 1 x 10-
3
in. If E = 2 X 10
6
psi and v = 0.20, determine the stress
components in the plane perpendicular to the borehole, and the major and
minor normal stresses in this plane and their directions. (Assume tb,e initial
stress parallel to the borehole is insignificant.)
10. Stress measurements in a horizontal rock outcrop using a series of vertical
flat jacks all give a cancellation pressure of about 80 MPa. The rock is
140 Initial Stresses in Rocks and Their Measurement
granite with E = 5 X 10
4
MPa and v = 0.25. If the rock started its life at a
depth of 10 km with CTh = CTy and was brought to the surface by erosion,
what should be the value of the horizontal stress? (y = 0.027 MN/m
3
.) If
there is a discrepancy, explain why.
11. Making use of the effective stress principIe (Chapter 3), derive a formula
corresponding to Equation 4.8 expressing the effective pressure (Pc\ - Pw)
for crack initiation in hydraulic fracturing when the rock has a pore pres-
sure Pw.
12. Bearpaw shale was loaded from O to 2000 psi vertically over a broad area so
horizontal strain could be assumed equal to zero. Poisson's ratio was 0.40
during loading. Subsequently, the vertical load was reduced to 1000 psi.
Poisson's ratio was 0.31 during unloading. (a) Estimate the horizontal pres-
sure corresponding to the maximum and final verticalloads. (b) What natu-
ral events could bring about a similar stress history?
13. In a rock with v = 0.3 and E = 3.0 X 10
4
MPa, "doorstopper" measure-
ments yield the following strains in the arms of a 60 rosette gauge on the
bottom of a vertical borehole 10 m deep: eA = -20 X 10-
4
in the gage
parallel to OX (east-west direction); eB = -3.8 X 10-
4
in the gage oriented
60 counterc1ockwise from OX; and ee = -5.0 x 10-
4
in the gage aligned
120 counterc1ockwise from OX. The hole is parallel to the y axis. Assuming
CTy is due to rock weight alone and y = 0.027 MN/m3, compute the greatest
and least normal stresses in the plane ofthe hole bottom (xz plane) and their
directions.
Planes ofWeakness
in Rocks
5.1 lnlroduction
Chap
ter
5
Those who excavate into rock know it to be a material quite apart from what a
mathematician might choose for tractable analysis. The engineer's rock is het-
erogeneous and quite often discontinuous. The latter is especially true ifblast-
ing has be en excessively violent and uncontrolled, but even if the engineer were
as careful as a sculptor, he or she would be confronted with naturally occurring
planes of weakness traversing the rock mass and separating it into perfectly
fitted blocks. Furthermore, the process of excavation will induce new fractures
in the surrounding rock by virtue of stress readjustments (Figure 5.la, b).
Small cracks that we have been referring to asfissures may be seen in many
hand specimens. Joints (Figure 5.2) are usually present in rock optcrops. They
appear as approximately parallel planar cracks separated by several centi-
meters up to as much as 10 m. One set of joints commonly forms parallel to
bedding planes and there are usually at least two other sets in other directions.
Igneous and metamorphic rocks may have regular jointing systems with three
or more sets. Rocks that have been deformed by folding often contain roughly
parallel seams of sheared and crushed rock produced by interlayer slip or minor
fault development. These shears are usually spaced more widely than joints
and are marked by several millimeters to as much as a meter thickness of soft
or friable rock or soil. Shears parallel to bedding planes occur in unfolded strata
near valley sides, due to interlayer slip as the rock mass relaxes horizontally.
Faults that offset all other crossing structures may also occur in the rock of an
engineering site. Thus there is a full range of planar weaknesses in rock mas ses
with a statistical distribution of spacings and orientations at all scales. Figures
5.3a and b show histograms of discontinuities observed at two dam sites stud-
ied by B. Schneider (1967). The fractures were studied using a combination of
aerial photographic interpretation, field observation, and microscopic study of
Figure 5.1 Fractures induced by mining a longwall face at great depth in quartzite,
South Africa. In both photos, the new fractures terminate in the roof against a preex-
isting, ftat-Iying joint developed along a shale parting. (a) The slabs formed by the
mining-induced fractures can be seen toppling towards the excavated space. (b) The
new fractures terminate downward in a preexisting shear zone inclined towards the
excavation. (Courtesy South African Chamber of Mines.)
Figure 5.2 Discontinuous rocks. (a) Rock-bolted sheet
joints in Navajo Sandstone, Glen Canyon, Arizona. (b)
Cross-bedded and sheet-jointed Navajo sandstone, Glen
Canvon.
144 Planes ofWeakness in Rocks
Figure 5.2 Discontinuous rocks. (e) Bedded sandstone, shale, and coaly
sediments in the foundation of Bennett Dam, Peace River, B. C. (d) Grout-
ing open joints between beds of quartzite, Akosombo Dam, Ghana.
5.1 Inlroduction 145
Fiw.re 5.2 Discontinuous rocks. (e) Stream controlled by primatic jointing.
(j) Formation of an arch by deterioration of wall rock liberated by a persist-
entjoint. [(e), (j) are in Devonian siltstone and shale, Enfield Glen, near
Ithaca, N.Y.].
Figure 5.2 Discontinuous rocks. (g) A long fracture surface formed by
linking of nonpersistent individual joints. (h) Short cross joints and one long
discontinuity [as in (g)], Navajo sandstone, Zion Canyon, Utah.
-
" 3
-
g
3
...
LI ...
1-
- o
3
-
3
-
o
n
3
-
n
3
Spacing
(a)
Spacing
(b)
3
3 3
3
3
3
o
g
3
3
o
g
3
3
o
3
3
o
3
3
5.1 lnlroduction 147
Faults
Shears
Joints
Fissures
Microfissures
Faults
Shears
Joints
F issu res
Microfissures
Figure 5.3 Relative distributions of discontinuities at
(a) Ta Chien Dam Site, in quartzite and (b) Malpasset
Dam Site, in schistose gneiss. [Mter B. Schneider
(1967).]
stained thin sections of rock. At Ta-Chien dam site, Formosa, planes of weak-
ness traversed the rock with spacings from 50 m down to 10 cm. At the Malpas- .
set Dam site, the rock showed cracks of various types with spacings from more
than 100 m to less than 5 mm. The Malpasset Dam failed because of the
discontinuous character of the rock in the abutments.
. . that are spaced more than about 20 m apart can be shown
mdlvldually in site sections and plans and considered individually in analysis.
On the other hand, planes of weakness that are more closely spaced occur in
large numbers and the only feasible way to appreciate their impact is often to
modify the properties of the rock mass, for example, by reducing
ItS modulus of elasticity. Figure 5.4 shows examples of single features (S) and
148 Planes ofWeakness in Hoeks
s
(al
(bl
Figure 5.4 Siogly occurring discontinuties (S) and
multiple features (M) in the region of iofluence of exca-
vatioos.
(al
(6)
Figure 5.5 Examples of the influence of joints aod other dscoo-
tinuities 00 fouodatoos and excavatioos.
150 Planes ofWeakness in Rocks
multiple features (M) in the rock around a tunnel and a surface excavation.
Notice how the locations of joints have affected the shape of the tunnel in
Figure 5.4a. N ormally the precise locations of the multiple features will not be
shown on the drawings, a pattem like that of Figure 5.4a being diagrammatic
only. Single features, however, can and should be plotted, for details of their
occurrence can radically affect the quality and cost of the work.
The importance of planar weaknesses stems from the special properties
that such features superimpose on rock. Basically, the rock mass becomes
weaker, more deformable, and highly anisotropic because there is reduced
shear strength and higher permeability parallel to discontinuities and increased
compressibility as well as reduced tensile strength (essentially zero) perpendic-
ular to them. These factors combine to create a variety of potential problems.
Foundations onjointed rocks (Figure 5.5a) may settle significantly as thejoints
close under load even if the rock itself is very stiff. Dams underlain by discon-
tinuous rock (Figure 5.5b) may initiate slip of rock blocks along one or more
weak surfaces; more than one dam failure has been attributed to this mecha-
nism (including Malpasset mentioned above-see Figure 9.15). Rock slopes
may fail as rock blocks move on single or multiple weakness planes. For
example, the rock slope in Figure 5.5e will most probably fail when the excava-
tion is deepened to the position of the dashed lines. Figure 5.4a showed how
block s might fall from the roof of a tunnel due to intersectngjoint planes. At a
larger scale, whole chambers can collapse owing to unfortunate intersections of
planar weaknesses. Another behavior pattem of jointed rock, in addition to
rigid translation of block s on planar surfaces, is bending of blocks under stress.
Flexural cracking and rock falls may follow such bending in a roof in thinly
bedded rock. Similarly, the rock cut of Figure 5.4b is subject to fiexure and
cracking of the inclined "cantilever beams" created by the steeply dipping
joints and contacts.
Since joint planes introduce such strongly directional weaknesses, the most
important joint attribute can be considered to be orientation. Fortunately, this
can be determined relatively reliably.
5.2 Joint Orientations
It is rare to encounter a rock mass with truly random fracturing. In every
instance where attitudes were measured for naturally occurring weakness
planes, the author's experience has been that planar weaknesses cluster around
one or more "preferred orientations." This is appreciated readily if the direc-
tions normal to the measured planes are plotted on a stereographic or equal
area projection. (The principies of stereographic projection ~ r e presented in
Appendix 5.) Either the upper hemisphere or the lower hemlsphere normal s
5.2 Joint Orientations 151
may be plotted. The upper hemisphere normal will be preferred here because it
has the same direction as that ofthe dip vector; that is, ifthe strike is north and
the dip 30 east, the upward normal will rise 60 to the east. Figure 5.6a shows a
series of normals that cluster about three preferred orientations roughly orthog-
onal to each other. In Figure 5.6b, there are two clusters but the scatter of
values is very great in one case (set 1) while a second cluster (set 2) has been
rotated and spread by folding so that its points are distributed about a segment
of a great circle of the sphere. If the normals to planes are distributed evenly
around a center, it is possible to pick a good representation for the preferred
orientation of the distribution by guessing the location of the point of greatest
concentration of normals. There are also methods for contouring the points
(see, for example, Hoek and Bray (1977), cited in Chapter 8). Another ap-
proach to describing the preferred orientation is to sum the normals vectorially.
Eachjoint normal can be considered a unit vector and the orientation ofthe
resultant of all the individual s of a cluster represents the preferred orientations
(the "mean") ofthe set. The summation can be accomplished by accumulating
the direction cosines (see Appendix 1). Let x be directed horizontally and
north, y horizontally and west, and z vertically upward. If a normal to a joint
plane rises at angle 8 aboye horizontal in a direction {3 measured counterclock-
wise from north, the direction cosines of the normal to the joints are
1 = cos 8 cos {3
m = cos 8 sin {3 (5.1)
and
n = sin 8
If many joints are mapped in one set, the preferred, or mean orientation of the
joint set is parallel with the line defined by direction cosines equal to the sums
of all individual l' s, m' s and n' s; dividing by the magnitude of .this resultant
vector gives the direction cosines (IR, mR, nR) of the mean joint orientation:
(5.2)
where
IRI = [(II)2 + (Im)2 + (In)2]112
The angle of rise 8
R
and the direction of rise (3R of the normal to the mean
orientation are obtained with Equations 5.1 together with rules for the correct
~ ~ ~ ~ O O ~ : .
8
R
= sin-1(nR) O 8
R
~ 90
{3R = + cos-
1
( ~ ) if mR ~ O
cos 0R
152 Planes of Weakness in Rocks
Upper hemisphere
(a)
1
Upper hemisphere
(b)
Figure 5.6 Distributions of normals to dis-
continuities, plotted on stereographic projec-
tions. (a) Two-well-defined sets and a third
more disperse set. (b) One very disperse set
and a second set distributed in a great circle
girdle.
and
~ R = - cos-
I
( ~ )
cos 8R
5.2 JoinlOrienlations 153
(5.3)
(In the aboye, the number represented by the cos-
I
term is presumed to He
between O and 180.)
One can also estimate the scatter of normals about the mean orientation by
comparing the length ofthe resultant with the number N ofjoints considered. If
the joints were aH parallel, the resultant would equal N, whereas if the joints
were widely varying in orientation, the resultant would be considerably less
than N. This is represented by a parameter K
F
:
(5.4)
KF becomes very large as the dispersion of joint orientations becomes small.
According to the hemispherical normal distribution (Fisher, 1953), the
probability P that a normal will make an angle of t/J degrees or less with the
mean orientation is described implicitly by
1
cos t/J = 1 + - ln(1 - P)
KF
(5.5)
Thus one can express the spread of values about the mean-the "disper-
sion" -corresponding to any degree of certainty. The standard deviation of the
hemispherical normal distribution (;;) can also be used to express the disper-
sion of normals about the mean:
- 1
t/J = v'K;
(5.6)
_ When calculating or estimating the orientation parameters 8R , ~ R , KF , and
t/J for each joint set, one must insure unbiased selection of individuals for
analysis. Unfortunately, as pointed out by Terzaghi (1965), outcrops and drill
holes introduce bias. Figure 5.7 a shows that the joints that paraHel an outcrop
surface (i.e., whose' normals are paraHel to the normal to an outcrop) cannot be
seen. If ao is the angle between a normal to a joint and the normal to the
outcrop, bias can be overcome by weighting the calculations such that the
singlejoint is replaced by a number ofjoints equal to 1/sin ao. Similarly, drill
holes will not reveal joints whose normals are perpendicular to the axis of the
hole (Figure 5.7b). Thus eachjoint individual oriented in a drill core should be
weighted by treating it in the analysis of orientations as if it were 1/ cos aH joints
where aH is the angle between the normal to the joint and the axis of the holeo
154 Planes of Weakness in Rocks
Generally, orientations of joints cannot be determined from drill hole data
because the core rotates an unknown amount as it is returned to the surface.
Methods for orienting core were reviewed by Goodman (1976).
The joint orientation parameters discussed here are fundamental properties
of the rock mass. In general, each joint set will also have a characteristic
physical description and a corresponding set of physical properties including,
most importantly, the parameters necessary to represent joint strength .
.
noutcrop njoint
0<0
---.....
nOint 1\
noutcrop
Lower hemisphere
Figure 5.7 Bias in occurrence of joints in (a) outcrops.
5.2 Joint Orientations 155
Lower hem isphere
Figure 5.7 Bias in occurrence of joints in (b) drill holes.
156 Planes of Weakness in Rocks
5.3 Joint Testing
When a rock mass is excavated, somejoints will close up while others will open
and sorne blocks will slide against others along joint surfaces. The properties
governing joint deformability and strength enable the magnitudes and direc-
tions of these movements to be calculated in practical problems. To obtain the
required descriptive properties, the engineer has two choices: (1) to use experi-
ence and judgment to select reasonable values for the joint properties based on
careful descriptions of joint characteristics as observed by geologists or
geotechnical engineers in outcrops and in core samples, or (2) to attempt to
measure the properties directly in field or laboratory tests. The latter is prefera-
ble but it is not often possible to obtain good samples for this purpose.
(a)
Clay dam
Liquid rubber
(b)
Fault gouge
Remolded fau It
gouge
Figure 5.8 Methods for gaining joint samples. (a) Oriented drill-
ing. (b) Molding and casting.
5.3 Joint Testing 157
Figure 5.8 Methods for gaining joint samples. (e) Models of joints ob-
tained as in (b): the left rectangle is a mold of a ripple-marked bedding
surface supplied by H. Schneider; the middle rectangle is a cast of the
same surface in plaster prepared for direct shear testing; the right squares
are the top and bottom of a portio n of the same surface after shear testing
in a smaller direct shear machine (note the wear along the slopes of the
ripples).
Samples for"laboratory testing can be acquired by drilling a large-diameter
core parallel to ajoint plane that outcrops, as depicted in Figure 5.8a. Alterna-
tively, a section of the joint surface could be molded with liquid rubber in a field
and facsimiles of the upper and lower blocks later cast in plaster, cement, or
sulfur in the laboratory (Figures 5.8b,c). The modeljoints will correctly repre-
sent the roughness of the surface and clay or mineral filling material collected in
the field can be spread on the model surfaces to simulate the actual field condi-
tion. Good results can be obtained this way ifthe tests are scaled by the ratio of
normal stress to compressive strength. For example, to study the shear of a
joint under 500 psi normal pressure in limestone having a compressive strength
of 16,000 psi, a model joint of sulfur (concrete cylinder capping compound
"Cylcap") having a compressive strength of 8000 psi should be tested at a
normal pressure of 250 psi.
Both triaxial and direct shear test methods may be adapted for testing
specimens withjoints. In the direct shear test (Figure 5.9a), the joint surface is
oriented parallel to the direction of applied shear load and the two halves of the
sample are fixed inside a shear box, using Cylcap, concrete, plaster, or epoxy.
158 Planes uf Weakness in Rucks
To avoid a moment and rotation of one block relative to the other, the shear
load may be inclined slightly, as in Figure 5.9b; but this prevents shear testing
at very low normalloads. Rotation during shear can occur in the shear appa-
ratus shown in Figure 5.9c, where the reactions to the normal and shear force s
are supplied by cables. A test in which rotations can occur during shearing
tends to underestimate the shear strength compared to one in which rotations
are prevented. Both conditions of loading exist in nature.
The stress conditions inside the shear box are represented by the Mohr
circle in Figure 5.9d. The normal stress (Iy and shear stress Txy on the failure
plane define point A'. The normal stress (Ix parallel to the joint is unknown and
N
N
(al
t
y
N
(bl
Leveling shims
"Potting" materia'l
4 T
Figure 5.9 Direct shear testing. (a) The arrangement of the specimen in
a shear box. (b) A system for testing with inclined shear force to avoid
m o m p . n t ~
"Failure" criterion
A'
__ ----r
__ -- xy
--
--
--
--
--
a
(d)
Figure 5.9 Direct shear testing. (e) A portable shear-test machine and load-
maintaining system based upon the ISRM-suggested method which accommo-
dates samples up to 115 by 115 mm. Courtesy of Roctest Inc., Plattsburg, N.Y.
(d) Approximation to the state of stress in the shear box; x and y are parallel
and perpendicular to the joint surface.
160 Planes ofWeakness in Rocks
a,
J
(a)
- - - - - - ~ ~ ~ - - - - - - - - - - - - - - - - - - - - - - ~ ~ ~ - - ~ ~ a
Direction
of a on
joint
(b)
Figure 5.10 Triaxial testing with jointed specimens. (a) Arrange-
ment of joint. (b) Stress conditions.
5.3 Joint Testing 161
Figure 5.10 Triaxial testing with jointed specimens. (e) A natural joint in
graphite schist oriented for a multistage triaxial test, viewed after the test. The
joint was initially closed and was opened by the first load stage.
can range from zero to a large proportion of CT
y
depending on the system used
for fixing the specimen inside the shear box. Since the shear stress perpendicu-
lar to the joint plane must be equal to T xy' the stress is represented by the Mohr
circle with diameter AA I (Figure 5.9d). This can lie partIy in the tensile stress
field if CTx is small.Accordingly, direct shear is not a good way to test intact
specimens of rock or soil. However, for sliding of joints, the direct shear test is
advantageous because normal and shear displacements during shearing can
both be measured easily and because shearing can proceed for long distances
such that wear is developed and the strength falls to its minimum value; the
latter is termed the residual strength. The direct shear test can also be con-
ducted in the field by excavating ajoint block in place. A series of direct shear
tests conducted at different confining pressures generate a series of points Al,
A2 , etc., through which ajoint shear strength curve or line will be drawn.
Another approach for joint testing is the triaxial compression test discussed
in Chapter 3. If the joint is oriented at an angle .p equal to 25 to 40 with the
direction ofaxialload, as shown in Figure 5.lOa, the joint will slip before the
rock has had a chance to fracture and the joint strength criterion can be estab-
lished. The strength criterion is no! an envelope to a series ofMohr circles as in
the case of intact rock, because failure is not free to develop in the most critical
orientation but is constrained to occur along the joint at angle .p. Figure 5.lOb
8
o
en
8
o
"
o
o
o
<D
';
c.
e-
e
M
t t t t
[U;J
_ N _
- -
o
!Sd "
'"
a)
e
(Ij
</)
'-'
....
</}
:.a
()
</}
a)
.<;::
..c:
g.
...
bIl
.S
;>;
.<;::
</}
o
....
</}
:.a
()
</}
o
....
-
a
c..
....
1::
'0
......
"C
a)
ca
a)
..c:
j
.S
1::
(Ij
1::
O
....
</)
a)
....
ca
.::
....
a)
....
....
(Ij-;
",",V'l
O a)
!l 5
-bIl
;:l .-
.S
'""
</}
'""
(Ij
115
1::
a)
.5
5.3 Joint Testing 163
shows a simple construction to locate the relevant point on the Mohr circle.
Using Bray's procedure (Appendix 1) for constructing the Mohr circle for the
triaxial test with (Ix = P and (Iy = (Iaxia' the focus is located at the right side of
the circle at point F (Figure 5.10). The normal and shear stress on the joint
plane corresponding to the stresses at peak load are then found by drawing a
line from ,F making an angle of t/J with the horizontal and piercing the Mohr
circle at point A (or alternatively making an angle of t/J with vertical from F' at
the left side of the circle). A series of triaxial tests conducted at different
normal pressures will yield a series of points Al, A2 , etc., through which ajoint
failure criterion can be drawn.
l'
T=-
A
N
o = A constant
Area =A
________ ______ ______________ ____ .. 6u
u,
6V(T)
______ ____ ____________ ______ 6u
/1,
Figure 5.12 Tangential and normal displacements during direct shear
of a rough joint.
164 Planes ofWeakness in Rocks
If the. test i.s started at a low eonfining pressure and the eonfining
pressure IS qUlekly ratsed after the peak axial stress is reaehed it will be
possible to generate a series of Mohr eirc1es for slip from a single est. Sueh a
proeedure is ealled "a multistage test. " Figure 5.11 shows a Mohr eircle family
generated from a multistage test for the joint in graphite sehist shown in Figure
5.lOc. (A multistage test ean also be programmed in direet shear.) If real roek
joints rather than models are tested, identieal specimens cannot be supplied for
a series ofpreviously untested roughjoints; but a multistage test could be used
to investigate the strength eriterion. Since wear will develop at higher normal
pressures, a multistage test will not give the same results as a series of identical
virgin specimens tested at different normal pressures; however, sorne compro-
mise is required if actual rock is to be tested.
Figure 5.12 shows the type of data obtained from the direct shear test. The
shear displacement across the joint, Ilu, is the difference in displacements of
the upper and lower block s measured parallel to the joint planeo If the joint is
rough, a meanjoint plane is recognized, passing through hilIs and over valleys
of the joint surface. Because of these undulations, the joint will tend to thicken
or "dilate" during shearing. The dilatancy Ilv is the difference between the
normal displacements of the upper and lower block as a result of shear dis-
plaeement. Opening (thickening) is reckoned as positive dilatancy. As the shear
stress builds, a period of adjustment with slight dilatancy is followed by rapid
increase in the rate of dilatancy; the dilatancy rate is greatest as the peak shear
stress (the "shear strength") is attained. Thereafter, the shear stress falls con-
tinuously and the joint dilates continuously until attaining the residual displace-
ment, sorne millimeters or even centimeters beyond the peak. In the field, with
very roughjoint surfaces, residual displacement may be achieved only after as
much as a meter of movement. The dilatancy and, to a considerable degree, the
joint strength are controlled by the joint roughness.
5.4 Joint Roughness
Imagine a joint surface with identical asperities rising at an angle i from the
meanjoint plane (Figure 5.13). Let the friction angle of a smoothjoint be 4J",. At
the moment of peak shear stress, the resultant force on the joint, R, is then
oriented at an angle 4J", with the normal to the surface on which motion is about
to oceur; since this surface is inclined i degrees with the joint plane, the joint
friction angle is 4J", + i when referred to the direction of the mean joint planeo
The aceuracy and utility of this simple eoncept was demonstrated by Patton
(1966).
Values of 4J", have been reported by many authors. A reasonable value
appears to be 30 with most values in the range 21 to 40. Byerlee (1978) found
5.4 Joint Roughness 165
t/l
eff
. = (t/l" + i)
Figure 5.13 The basis for Patton's law for joint shear strength.
4J", = 40 fit a variety of rocks with saw cut or ground surface up to (I equals 2
kbar (with 'rp = 0.5 + 0.6 (I kbar applicable at higher pressure-up to 16 kbar).
Frequently, 4J", can be much lower when mica, tale, chlorite, or other sheet
silicate minerals form the sliding surface or when clay gouge is present. There
is little opportunity for drainage of water from the pores of saturated c1ay .
loeked in between joint walls and values of 4J", as low as 6 have been reported
for montmorillonite c1ay-filled shears. The roughness angle i can be any value
from O to 40 or more at low pressures as diseussed later.
If the normal pressure is relatively large, it will be easier to shear the joint
through the teeth along its surface than to lift over them. Mobilizing sorne roek
strength by failure through the teeth generates a shear strength intereept Si and
a new friction angle 4Jr related to sliding on surfaees broken through the roek
and thus approximated by the residual friction angle for intact roek specimens.
(The residual frietion angle is the slope of the linear envelope to a series of
166 Planes ofWeakness in Rocks
Mohr circles through residual stress values for a series of triaxial compression
tests with intact rock specimens.) Figure 5.14 shows the bilinear failure crite-
rion for joints representing the merging of Patton' s law and the condition for
shearing through asperites:
T
p
= (J' tan(<p/L + i) for (J' smaH
and (5.7)
for (J' large
For many purposes, it is sufficient to replace <Pr in the second equation by <P/L
since these values are close. Actual data show a transition from the initial slope
at <P/L + i to the final slope at <Pr. Theories of joint strength effecting this
transition smoothly were presented by Ladanyi and Archambault (1970), Jaeger
(1971), and Barton (1973) and were reviewed by Goodman (1976).
Roughness control s not only the peak shear strength at low normal pres-
sures but the shape of the shear stress versus shear deformation curve and the
rate of dilatancy. This is shown in Figures 5.15 and 5.16 based on work by
Rengers (1970) and Schneider (1976). Suppose thejoint surface were accurately
profiled as in Figure 5 .15a. Any two points a distance nS apart along the profile
will define a line inclined i degrees with the mean plane through the joint
surface. If the measuring points are moved aH along the surface and the mea-
suring base length is varied, a series of angles will be measured; these angles
have been plotted against base length in Figure 5.15b, and envelopes to aH the
points have been drawn. It is seen that the maximum angles presented by a
rough surface approach zero as the measuring distance becomes appreciably
S
J
T
____ __________________________________
Figure 5.14 Bilinear shear strength criterion.
(a) Profile:
O 234567 B
r-sl
Envelope for shearing
top to right
30
20
10
..
..
a.
O
..
"O
..
-10
-20
-30
( )
T posltlve
\
"'"
/
"-
K:
..........
:--
--
s 2s 3s 4s
6s-
---
V
./
r-
'"
/ ""
i E nvelope for shearing top to left
(T negative)
(b)
Large plate Compass Small plate
(e)
7s
5.4 Joint Roughness 167
1c1'-0
9 10 11
Ss
Base distance
(d)
Mean plane
12
Largest plate
Smallest
plate
Figure 5.15 Rengers' analysis of roughness. (a) A rough surface. (b) Envelopes
to roughness angles as a function of base length. (e) and (d) Approximation to
Rengers' roughness angles by Fecker and Rengers' method.
168 Planes of Weakness in Rocks
1.0
\
" 25
i
tan
=
I
,
" 20
,
"-
Calculated
T(611)
Envelope to
roughness
angles
' .... 10
.... -
--
5
--
---
la) Base distance
2 Calculated dilatancy
---_L_
6ul T) ; base distance
lb)
rJ>, = 35
T(IlU) = (J(tan[c!>, + i
tan
(llu)])
6u ; base distance
(e)
Figure 5.16 Construction of the dilatancy curve and the shear stress-shear
deformation curve from the Rengers' envelope according to the method of
H. Schneider (1976).
5.5 lnterrelationships among Displacements and Strengths 169
larger than the wavelengths of the roughness anclwaviness of the surface. The
envelopes drawn in Figure 5.15b present a series of i values corresponding to
shear displacements numericaHy equal to each base distance nS. The upper
envelope corresponds to shearing of the upper block to the right while the
lower corresponds to shearing of the upper block to the left. Consider
shearing to the right. In Figure 5.16a, the appropriate envelope has been replot-
ted. Then, directly below it (Figure 5.16b), the dilatancy curve has been con-
structed by drawing a series of secants described by the appropriate i value for
each base distance. Next we construct the shear stress versus shear displace-
ment curve (Figure 5 .16c) from the dilatancy curve assuming that each shear
stress value r(au) can be calculated from the normal stress acting on a surface
whose effective friction angle is the sum of the current i value and the residual
friction angle (Schneider, 1976); that is,
r(au) = (I' tan[cPr + (5.8)
where itan(au) is the inclination of the tangent to the dilatancy curve at any
value of au.
We have seen that the roughness profile of a weakness "plane" is funda-
mentaHy valuable. It will be useful then to have more than one way of determin-
ing it. Figure 5.15c shows another method demonstrated by Fecker and Ren-
gers (1971). 'If the orientations of ftat plates placed down on a single rough
surface are compared, it will be seen that they are scattered about a mean
value. For a plate of a given size, the maximum angle from the mean of the
series of measurements, in other words the i value, can be obtained by plotting
the normals on a stereographic projection, drawing an envelope to aH the
points, and measuring the angle between the envelope and the mean orientation
(Figure 5.15d). Alternatively, this can be done mathematicaHy, emulating the
procedure discussedin Section 5.2 where a series of nonparaHel planes were
averaged. The difference here is that one plane only is being measured, the
different points representing the different orientations measured at different
place s on its surface.
5.5 lnterrelationships among
Displacements and Strengths
When a block containing a joint plane is subjected to a shear stress paraHel to
the joint, it can undergo both shear au and normal displacement av. If. com-
pressed normal to the joint, it will tend to shorten by joint closure and if puHed
apart normal to the joint, the block will eventuaHy separate into two block s as
the joint opens. AH these phenomena are coupled together as shown in Figure
5.17. The upper figure (5. 17a) shows the compression behavior ofthejoint; it is
170 Planes of Weakness in Rocks
6V
-It
o
O
(a)
Vme
6V
It
(b)
O
a
T
(e)
A B C D
.. a
a=O
a=B
__________ __________
a= C
a= D
No dilatancy permitted (path O, 3, 6)
-------
a=C
No dilatancy permitted (path O, 1, 2)
a=A
a=O
Figure 5.17 Coupling of the normal deformation, shear deformation,
and dilatancy laws for rough joints and analysis of path dependency.
5.6 Effect ofWater Pressure 171
highly nonlnear and becomes asymptotic to a maximum closure (Vme) related
to the initial thickness or aperture of the joint.
Suppose a virgin specimen were sheared without normal pressure. Dila-
tancy. would occur as shown in the upper curve of Figure 5.17 b while the shear
stress would never rise aboye zero since there is no frictional resistance in this
case (lowermost curve of Figure 5.17 c). If the specimen had been compres sed
initially to a value A, B, e, or D, the dilatancy and shear stress versus shear
displacement relationships would have been as depicted by the families of
curves in 5.17b and c. As the normal pressure grows, the dilatancy is gradually
reduced because a greater proportion of the asperities becomes damaged during
shearing. All the curves of dilatancy and shear assume the normal stress was
maintained constant at the indicated initial value throughout the shearing pro-
cess. This would not have been true if the normal displacement had been
restricted, as, for example, when a block moves into a tunnel between parallel
joint surfaces. However, the shear stress versus shear displacement function
for such a condition can be determined from the data presented. Suppose, for
example, the joint normal stress were zero to start but no dilatancy was allowed
during shearing. Then, the joint would acquire normal stress A when it had
sheared to point 1, with shear resistance appropriate to the point on the shear
stress/displacement curve corresponding to normal stress A. Thus, as shear
progressed, the shear stress would rise with displacement along the dashed
locus 0, 1, 2. In similar fashion, one could construct the shear stress/shear
displacement curve for a joint initially compres sed to normal stress A and then
sheared without dilatancy (path 0, 3, 6). In both cases, note that considerable
additional shear slrength was acquired by the restriction of normal displace-
ment and the behavior became plastic rather than brittle; that is, there was lttle
or no drop in strength after a peak stress had been reached. This helps explain
why rock bolt reinforcement has be en successful in stabilizing rock slopes and
excavations.
Mathematical relationships have already be en presented for the variation
ofpeak stress versus normal pressure (Equation 5.7). Additional formulas have
been demonstrated linking the decline of peak diliitancy with normal stress
(Ladanyi and Archambault, 1970) and for the compression ofjoints (Goodman,
1976).
5.6 Effect of Water Pressure
Joints obey the effective stress principIe discussed in Chapter 3. The water
pressure in a joint directly counteracts the strengthening effect of the normal
stress applied to the joint. To calculate the water pressure required to cause a
fault or joint to slip, one needs to determine the amount by which the Mohr
172 Planes of Weakness in Rocks
circle corresponding to the current state of stress has to travel to the left to
bring the normal and shear stress on the fault or joint plane to the limiting
condition represented by the criterion offailure (Figure 5.18). This calculation
is slightly more complicated than for the case of unjointed rock, because now,
in addition to initial stresses and strength parameters, the orientation of the
joint plane (t/J with the direction of (TI) needs to be considered. If the initial
stresses are (T3 and (TI, the water pressure that will produce fault slip is
S ( . sin t/J cos t/J)
Pw = -t J,J.. + (T3 + (TI - (T3) sm
2
t/J - t ,J..
an 'Pj an 'Pj
(5.9)
Pw is the minimum of values calculated from (5.9) using (a) Sj = O and ~ j = ~ + i
and (b) Sj * O and ~ j = ~ r '
This simple application of the effective stress principIe has been shown to
explain satisfactorily the occurrence of earthquakes due to water injection in a
deep waste disposal well near Denver, Colorado (Healy et al., 1968) and in the
Rangely oil field, western Colorado (Raleigh et al., 1971). It can be used to
consider the likelihood of triggering earthquakes by construction of reservoirs
a
Figure 5.18 Water pressure to cause slip 00 a joiot.
REiferences 173
near active faults. However, the initial stresses in the crust as well as the
frictional properties of the fault will have to be known.
References
Bartoo N. (1973) Review of a oew shear streogth criterio n for rockjoints, Eng. Geol. 7:
287-332.
Barton, N. (1976) The shear strength ofrock and rockjoints, Int. J. Rock Mech. Min.
Sci. 13: 255-279.
Barton, N. et al. (1978) Suggested methods for the quantitative description of discon-
tinuities in rock masses, for ISRM Commission on Standardization of Lab and
Field Tests, Int. J. Rock Mech. Min. Sci. 15: 319-368.
Barton, N. R. and Choubey, V. (1977) The shear strength ofrock joints in theory and
practice, Rock Mech. 10: 1-54.
Bray, J. W. (1967) A study ofjointed and fractured rock, 1. Fracture pattems and their
failure characteristics, Rock Mech. Eng. Geol. 5: 117-136.
Byerlee, J. (1978) Friction ofrocks, Pure and Applied Geophysics, American Geophysi-
cal Union.
Fecker, E. and Rengers, N. (1971) Measurement oflarge scale roughness ofrock planes
by mean s ofprofilograph and geological compass, Proceedings, International Sym-
posiuln on Rock Fracture, Nancy (ISRM), paper 1-18.
Fisher, R. A. (1953) Dispersion on a sphere, Proc. R. Soco London, Ser. A 217: 295.
Goodman, R. E. (1970) The deformability of joints, ASTM Spec. Tech. Publ. 477, pp.
174-196.
Goodman, R. E. (1976) reference given in Chapter 1.
Healy, J. H., Rubey, W. W., Griggs, D. T., and Raleigh, C. B. (1968) The Denver
earthquak.e, Science 161: 1301-1310.
Jaeger, J. C. (1971) Friction ofrocks and the stability ofrock slopes-Rankine Lecture,
Geotechnique 21: 97-134.
Ladanyi, B. and Archambault, G. (1970) Simulation of the shear behavior of a jointed
rock mass, Proceedings, 11th Symposium on Rock Mechanics (AIME), pp. 105-
125.
Patton, F. D. (1966) Multiple modes of shear failurein rock, Proc. 1st Congo ISRM
(Lisbon), Vol. 1, pp. 509-513.
Priest, S. D. and Hudson, J. A. (1976) Discontinuity spacings in rock, Int. J Rock Mech.
Min. Sci. 13: 135-148.
Raleigh, C. B., Healy, J. H., Bredehoeft, J. D., and Bohn, J. P. (1971) Earthquake
control at Rangely, Colorado, Trans AGU 52: 344.
Rengers, N. (1970) Influence of surface roughness on the friction properties of rpck
planes, Proc. 2nd Congo ISRM (Belgrade), Vol. 1, pp. 229-234.
Schneider, B. (1967) Reference given in Chapter 6.
Schneider, H. J. (1976) Commeot in Proceedings olInternational Symposium on Nu-
merical Methods in Soil Mechanics and Rock mechanics, G. Borm and H.
Meissner, Eds., pp. 220-223 (lost. fr Bodeomechanik und Felsmechanik of Karls-
ruhe Uoiversity, D-7500, Karlsruhe 1, Germaoy).
Terzaghi, R. (1965) Sources of error io joiot surveys, Geotechnique 15: 287.
174 Planes ofWeakness in Rocks
Problems
1. Determine by mathematical ca1culation the mean orientation and the Fisher
distribution parameter kfor each ofthejoint sets represented by the follow-
ing data collected in the field:
Joint or Strike Dip Joint or Strike Dip
Other Plane
(0)
e) Other Plane
(0) (0)
1 S40 E 35 NE 16 S38 W 62NW
2 S42 E 35 NE 17 S36W 63NW
3 S40 E 39NE 18 S38 E 41 NE
4 S30W 60NW 19 S25 E 38NE
5 S35 W 61 NW 20 S30W 58NW
6 S41 E 34NE 21 N30E 30 SE
7 S32 W 59NW 22 N35 E 32 SE
8 S35 W 62NW 23 N22 E 28 SE
9 S38 E 37 NE 24 N45 E 60NW
10 S40 E 37 NE 25 N55 E 58NW
11 S33 W 61 NW 26 N50E 59NW
12 S33 W 64NW 27 N30W 90
13 S40 E 37 NE 28 N40W 88 NE
14 S41 E 36NE 29 N40W 1 NE
15 S40W 62NW 30 N30E 24 SE
2. Plot the normal s to the joint planes of Problem 1 on an upper hemisphere
stereographic projection and compare the calculated preferred orientations
with what seem to be the points where the greatest density of normals
occur.
3. A multistage triaxial test with a sawed joint oriented 45 with the axis of the
core yielded the following data. Determine <PI".
Confining Pressure (P)
(MPa)
0.10
0.30
0.50
1.00
Maximum Axial Stress
(MPa)
0.54
1.63
2.72
5.45
Problems 175
4. A reverse fault in the rock of Problem 3 has a dilatancy angle of 5 and is
inclined 20 with the horizontal. What is the maximum horizontal stress that
could be sustained at a depth of 2000 m in this rock?
5. Trace the roughness profile of Figure 5.15a on a sheet of paper; then cut
along it carefully with scissors to produce a model of a direct shear speci-
meno Slide the top to the right past the bottom, without rotation and without
"crushing," and draw the path of any point on the top block. Compare this
path to the constructed dilatancy curve of Figure 5 .16b. Mark the locations
of potential crushing at different shear displacements.
6. A normal fault that is partly cemented with calcite mineralization dips 650
from horizontal. The fault slipped when the water pressure reached 10 MPa
at a depth of 600 m. If Sj = 1 MPa and <Pj = 35, what was the horizontal
stress before the fault slipped?
7. Sj = O and <pj = 28.2 for a sawed joint oriented 50 from vertical in a
saturated triaxial compression specimen. The confining pressure is 1.5 MPa
and the axial stress 0"1 = 4.5 MPa with zero joint water pressure. What
water ,pressure will cause the joint to slip if 0"1 and 0"3 are held constant?
8. The following data were taken in a direct shear test conducted in the field
along a rockjoint, with area 0.50 m
2
The weight of the block aboye the joint
is 10 kN.
,
T, Shear Force (kN) O 1.0 2.0 3.0 5.0 6.5 6.0 5.5 5.4 5.3
u, Shear Displacement (mm)O 0.5 1.0 1.5 3.0 5.2 7.5 9.5 11.5 2=12
Assuming that joint cohesion is zero, and that <PI" = <Presid, determine the
peak and residual friction angles, the shear stiffness (MPa/m), and the
dilatancy angle at peak and post peak displacements.
9. A jointed shear test specimen is drilled at angle a with the normal to the
shear plane and a model rock bolt is installed and tensioned to force F
B
(see
figure). Then a pair of shear force s Tare applied until the joint slips.
(a) What is the bolt tension F
B
just sufficient to prevent slip under shear
force T.
(b) What is the value of a that minimizes the value of F B required to prevent
joint slip?
(c) How are the answers to be changed if the joint tends to dilate during
shear, with dilatancy angle i, and the bolt has stiffness kb?
176 Planes ofWeakness in Rocks
Joint
T
T
10. John Bray (1967) derived the following expression for the limiting effeetive
stresses for joint slip:
_ <T} _ tan 11/1 1
Kf = <T - tan(11/I1 + CPj)
where 1/1 is the angle between the direction of <TI and the joint planeo (The
derivation for this useful formula is given in Appendix 4 in the derivations
to equations 7.11 to 7.16.) Draw a polar plot of the ratio <T}/ <T for limiting
eonditions as a funetion of I/I( -7T/2::5 1/1 ::5 7T/2) for values of (a) CPj = 20 and
(b) cpj = 30. Label the regions on these diagrams eorresponding to slip and
saje principal stress ratios.
11. Use the expression given in Problem 10 to re-solve Problem 5-7. (Hint:
Substitute <TI - Pw and <T3 - Pw in place of <T, <TD
12. (a) Sedimentation is inereasing the thiekness of overburden (z) and the
vertical stress (<T
v
) in a roek mass. Assume the roek strength is given by
Si = 1 MPa and cP = 30. With v = 0.2, and'Y = 0.025 MPa/m, draw the
limiting Mohr circle that causes shear failure in the rock and determine
the corresponding values of z, <Tv , and <Th
(b) Now as sume that shear fractures have formed, in the orientation deter-
Problems 177
mined by the shear failure in (a). The new shear joints have CPj = 20.
Draw the new Mohr circle after the failure and determine the new value
of <Th (no ehange in <T v has oeeurred).
(e) Assume additional sedimentation inereases the value of <T
v
to 1.5 times
its value in (b). What are the eorresponding values of z and <Th. Draw the
eorresponding Mohr circle.
(d) Now erosion begins, redueing <T
v
Assuming the eorresponding redue-
tion in <Th is given by d<Th = [v/(1 - v)] d<T
v
, draw a series of Mohr
eircles and determine the value of z when <Th = <Tv
(e) With further erosion, the shear joints formed in (a) are no longer rele-
vant to the stress circles sine e the major stress is now horizontal. New
joints form when the Mohr circle eontaets the roek strength envelope.
Draw this eircle and determine the corresponding values of z, <Th,
and <T
v
(f) Assume the Mohr circle is now limited by the new joints. Find the
appropriate new value for <Th (no ehange in <T
v
).
(g) Draw graphs showing the variations of <T
v
and <Th with z inereasing to the
max found in (e) and then deereasing to zero.
13. The average fracture frequeney (A) aeross a roek eore is the total number of
natural'fraetures divided by the totallength drilled.
(a) Suppose there is only one set of joints and that A is the fracture fre-
queney measured in a direetion normal to them. Derive an expression
giving A in a direetion 8 with respeet to the normal.
(b) There are two orthogonal sets offraetures with A values, respeetively Al
and A2. Derive an expression for A measured in a direetion 8 from Al.
(e) GivenAI = 5.0 and A2 = 2.0 fractures per meter. Find the values of8 and
A sueh that the fracture frequeney is a maximum. What is the average
fracture spacing in this direetion?
14. Barton (1973) proposed an empirieal eriterion of peak shear strength for
joints:
T = <T
n
tan[JCR log (JCS/<T
n
) + Pb]
where JCS is the eompressive strength ofthe wall roek, and JCR is thejoint
roughness eoefficient. (In this expression, the argument of tan is under-
stood to be expressed in degrees.) Compare this equation with Equa-
tion 5.8.
Deforrnability of
Rocks
6.1 lntroduction
Chapter 6
Deformability means the capacity of rock to strain under applied loads or in
response to unloads on excavation. The strains in rock concern engineering
even when there is Httle risk of rock failure because locally large rock displace-
ments can raise stresses within structures. For example (Figure 6.1), a dam
seated astride varying rock types having dissimilar deformability properties
will develop shear and diagonal tension stresses by virtue of the unequal deflec-
tions of the foundation. The dam can be structured to handle these deflection
tendencies if the properties of the rock are known and if the variation of
properties within the found3;tion are determined. Furthermore, in mass con-
crete structures like gravity dams, the deformability of the rock also relates to
thermal stresses in the concrete, which are calculated by the product of a
thermal expansion coefficient, a temperature change, and a deformability termo
There are many situations in which rock displacements should be calcu-
lated. To design pressure tunnels, one should know the expansion of the lining
under operating pressure, as well as the amount of recovery when the pressure
is lowered. The same is true of arch dams pressing against their abutments. Tall
bUildings on rock may transmit sufficient load to their foundations that rock
settlement becomes significant for designo For long-span, prestressed roof
structures and bridges anchorages in rock, structures pushing against rock, or
gravity blocks seated on rock, knowledge of the rock displacements and rota-
tions is basic to design details. And for any excavation that is monitored, the
expectable displacements should be calculated to pro vide a framework with
which to interpret the measurements.
ELASTIC AND NONELASTIC BEHAVIOR
It is not sufficient to characterize rock deformability by elastic constants alone:
for many rocks are nonelastic. Elasticity refers to the property of reversibility
179
180 Deformability of Rocks
Figure 6.1 Shear stress developed in a concrete dam due to variable
deformability in the foundation rock.
of deformation in response to load. Many fresh, hard rocks are elastic when
considered as laboratory specimens. But on the field scale, where the rock can
be expected to contain fissures, fractures, bedding planes, contacts, and zones
of aItered rock and clays with plastic properties, most rocks do not exhibit
perfect elasticity. The extent of irrecoverability of strain in response to load
cycles may be as important for the design as the slope of the load/deformation
curve. Figure 6.2 offers an example. As the reservoir behind the arch dam
rises, the rock under the arch responds along curve 1. The concave upward
curvature of this load/ deflection path is typical for fractured rocks on first
("virgin") loading because the fractures close and stiffen at low loads. When
the reservoir is lowered for whatever reason, the rock unloads along path 2;
with a permanent deflection. The dam will try to follow the loading, but since it
Elevation
of reservoir
surface
Displacement of abutment!;x
Figure 6.2 Permanent foundation deformation caused by cycles of reservoir filling
and emptying.
6.2 Elastic Constants 181
is often more elastic than the rock, it will move 1:\way!rom the rock on unload-
ing. This could openjoints in the rock or concrete or slmply lower compres-
ive stress flowing through the structure. Repeated cycles of loadmg and un-
in response to cyclic operation of the reservoir would produce the
eries of loop s ("hysteresis") depicted in Figure 6.2. Sorne sites have been
unacceptable for concrete dams because of large hysteresis even
though the modulus of elasticity ofthe reasonable: A
criterion in this regard is presented later m conJunctlon wlth the plate-beanng
test.
6.2 Elastic Constants
The deformations of linearIy elastic isotropic solids can be calculated for
known increments of stress if only two material constants are specified .. In
Chapter 4, these were taken as Young's modulus. E (the "modu!us of.
ity") and Poisson's ratio v. Hooke's law, generallzed to three dlmensIOns, IS
1 v v
O O O
E E E
v v
O O O
ex E' E E
G"x
e y v v 1
O O
G"y
O
(6.1)
e
z
E E E
G"z
'Yxy
2(1 + v)
O O
Txy
O O O
'Yyz
E Tyz
'Yzx
O O
2(1 + v)
O
Tzx
O O
E
O O
2(1 + v)
O O O
E
The quantities E and v can be determined directIy tests known .
stress is applied and strains are measured. Where strams ,are apphed and stress
changes are measured, it is more natural to use Lame s constant A and the
shear modulus G as the two elastic constants; these are defined by
G"x . + 2G A A O O O ex
G"y A . + 2G A O O O ey
G"z
A . A + 2G O O O e
z (6.2)
Txy
O O O G O O 'Yxy
T
yZ
O O O O G O 'YyZ
Tzx
O O O O O G
'Yzx
182 Dtiformabilily of Hoeks
The relationships between the two sets of constants are
E
G = 2(1 + v)
(6.3)
and
A = Ev
(1 + v)(1 - 2v)
(6.4)
One other constant is quite useful, the bulk modulus K, which expresses the
relationship between hydrostatic pressure p and volumetric strain AV/V. Let
AV
(6.5) p=K-
V
then
K= E
3(1 - 2v)
(6.6)
(The compressibility is the reciprocal of K.)
Many rock mas ses are anisotropic, that is, directional in their behavior,
due to regular bedding or jointing or oriented fabric or microstructure that
makes the rock itself anisotropic. It is rarely fe asible to deal with extreme
anisotropy but orthotropic symmetry can be entered into computations without
mathematical burden. In the latter case, there are three mutually perpendicular
directions of symmetry, referred to as principal symmetry directions. If the
rock has three perpendicular sets of joints, for example, it should behave
orthotropically .. If x, y, and z are chosen parallel to the orthotropic symmetry
directions, Hooke's law now requires constants:
_ v
yx _ Vu;
O O O
Ex Ey Ez
_ vyx 1 _ Vzy
O O O
ex Ey Ey Ez o"x
ey Vu; _ Vzy 1
O O O
O"y
e
z Ez Ez Ez u"z (6.7)
Yxy
1
O O
Txy
O O O
YyZ
Gxy TyZ
yzx
1
O
Tzx
O O O O
GyZ
O O O O O
Gzx
6.3 Measurement of Dtiformabilily Properties by Statie Tests 183
The Poisson's ratio Vij determines the normal strain in symmetry directionj
when stress is added in symmetry direction i. In the orthotropic rock mass
Vij = Vji
Ei Ej
(6.8)
The nine independent constants are reduced to five if the rock is isotropic
within aplane (transversely isotropic). This arises when two types of rock are
regularly interlayered. It also comes about when flat minerals like mica, tale,
chlorite, graphite, or serpentine are arrayed in parallel orientation, or when
long minerals (e.g., amphiboles) are oriented with their long axes randomly
pointed within parallel planes. The number of elastic constants is reduced to
four when transversely isotropic symmetry arises from one set ofregular joints,
for example, parallel to bedding, in otherwise isotropic rock (see Problem 6.11).
To describe transversely isotropic elasticity, let s and t be any two perpen-
dicular directions in the plane perpendicular to the axis of symmetry (e.g., in
the plane of bedding) and let n be the direction parallel to the axis of symmetry
(e.g., in the direction normal to the bedding planes). Then Es = El and VIS = VSI
and Equation 6.7 becomes
es
el
Yns
Ynl
YSI
V
sn
Es
V
sn
Es
O
O
O
V
sn
Es
1
Es
VSI
Es
O
O
O
VSI
Es
1
Es
O
O
O
O
O
O
1
G
ns
O
O
O
O
O
O
1
O
O
O
(6.9)
O
U"I
O
O
TsI
In routine engineering work, it is usual to as sume the rock to be isotropic,
but with schists this may be inappropriate. Measurements can be made on
samples cored in different directions to determine additional deformability con-
stants. When anisotropy derives from regular structures, the orthotropic con-
stants can be caleulated, as shown latero
6.3 Measurement of Deformabl1ity
Properties by Static Tests
Stress-strain relationships can be observed in static and dynamic tests con-
ducted in the laboratory or in the field. Deformability properties can then be
184 Deformabnity of Rocks
obtained from the data, assuming that sorne idealized model describes the rock
behavior in the test configuration. Deformability properties can also be back
calculated from instrumental data on the movements of a structure or excava-
tion, if the initial and final states of stress are known, using methods of Chapter
4 inversely.
The most widely used testing procedures for deformability measurements
are laboratory compression and bending tests, wave velocity measurements in
the lab or field, field loading tests using flatjacks or plate bearing apparatus, and
borehole expansion tests.
LABORATORY COMPRESSION TESTS
An unconfined compression test on a core of rock, with carefully smoothed
ends and length to diameter ratio of 2, yields a stress-strain curve like that of
Figure 6.3a. The axial strain can be measured with strain gages mounted on the
specimen or with an extensometer attached parallel to the length of the speci-
men; the lateral strain can be measured using strain gages around the circum-
ference, or an extensometer across the diameter. The ratio of lateral to axial
strain magnitudes determines Poisson's ratio v. With hard rocks, it is usually
not acceptable to determine axial strain from measured shortening of the test-
ing space ("crosshead motion") because relatively large displacement occurs
at the ends where the rock contacts the platens of the testing machine.
Figure 6.3b shows the difficulty of defining exactly what is meant by E. It is
not simply the slope ofthe virgin loading curve, for this embraces nonrecovera-
ble as well as elastic deformation. The unloading curve, or reloading curve after
a cycle ofload and unload, are better measures of E. This definition allows E to
be determined even after the peak load when the rock has become fractured
(Figure 6.3b).
Deere (1968) presented a classification graph for intact rock specimens
based upon the ratio of elastic modulus to unconfined compressive strength,
together with the absolute value of the latter. For most rocks, the ratio El qu lies
in the range from 200 to 500 but extreme values range as widely as 100 to about
1200. In general, the "modulus ratio" Elqu is higher for crystalline rocks than
for clastic rocks, with sandstones higher than shales. Table 6.1 gives the mea-
sured ratio of "modulus of deformation" to unconfined compressive strength,
and corresponding values of Poisson's ratio, for the set of rocks previously
considered in Table 3.1. The substitution of modulus of deformation in place of
modulus of elasticity indicates that the deformability property embraces both
recoverable and nonrecoverable deformation. In general, whenever the modu-
lus value is calculated directly from the slope of the rising portio n of a virgin
loading curve, the determined property should be reported as a modulus of
deformation rather than as a modulus of elasticity. U nfortunately, this is not
universal practice at present.
6.3 Measurement of Deformability Properties by Sta tic Tests 185
Lateral E.
strain
(b)
(e)
E
-f.
V =-
f,
0.5
a,
Ax ial strain
-- Slope 01 ........ , Slope 01
virgin........ reloading
curve -:::"..... curves
................
.......... ----
Calculated Irom
Calculated Irom reloading curve
L----
--
--
--
virgin curve
E,
Figure 6.3 Variation of modulus of elasticity (E) and Poisson's ratio (v)
with axial strain in a compression test.
186 Dtiformability of Rocks
Table 6.1 ModuIus Ratio E/qu and
Poisson's Ratio v for the Rock
Specimens of TabIe 3.1
a
Description E/qu v
Berea sandstone 261 0.38
Navajo sandstone 183 0.46
Tensleep sandstone 264 0.11
Hackensack siltstone 214 0.22
Monticello Dam greywacke 253 0.08
Solenhofen limes tone 260 0.29
Bedford limestone 559 0.29
Tavernalle limestone 570 0.30
Oneota dolomite 505 0.34
Lockport dolomite 565 0.34
Flaming Gorge shale 157 0.25
Micaceous shale 148 0.29
Dworshak Dam gneis s 331 0.34
Quartz mica schist 375 0.31
Baraboo quartzite 276 0.11
Taconic marble 773 0.40
Cherokee marble 834 0.25
Nevada Test Site granite 523 0.22
Pikes Peak granite 312 0.18
Ce dar City tonalite 189 0.17
Palisades diabase 339 0.28
Nevada Test Site basalt 236 0.32
John Day Basalt 236 0.29
Nevada Test Site tuff 323 0.29
a E reported here includes both recoverable and nonre-
coverable deformation, mixed in unknown proportions.
The negative slope of the tail of the complete stress-strain curve is not a
stress-strain curve in the conventional sense but is a yield function; in particu-
lar, it is the envelope of yield points from aH reloading curves. Figure 6.3c
shows the value of v calculated from lateral deformation of a compression
specimen on its virgin loading curve. The ratio of lateral to axial strain begins at
a value near 0.2 and increases graduaHy until near the peak load when it begins
to accelerate, even surpassing the theoretical maximum value of v for isotropic
materials-0.5. (Equation 6.6 shows that K approaches infinity as v tends
toward 0.5.) The rock cannot be described as elastic as it moves on the yield
surface after the peak since it is cracked and large lateral deformations occur
6.3 Measurement of Deformability Properties by Sta tic Tests 187
with movement of rock wedges. However, on unloading and reloading, lateral
strains occur with v < 0.5. Again, we can conelude that the elastic constants
should be defined with respect to the reloading curve.
The full description of rock deformability should inelude not only the elas-
tic coefficients E and v, but the permanent deformation associated with any
applied stress level. Figure 6.4 shows how to determine the modulus of perma-
nent deformation M, defined as the ratio of a stress to the permanent deforma-
tion observed on releasing that stress to zero. M is determined by running a
series ofload cyeles during a compression test. We may compute similarly the
Poisson's ratio V
p
corresponding to permanent lateral deformation increments.
e
B
/
--------1--
/1
/1
/ I
JI
/1
/
-----f---
/1
/1
/ I
/ I
/
/ I
I
I
I
p,e
Figure 6.4 Determination of the modulus of permanent deforma-
tion M from load cycling data in a compression test.
188 Deformability of Rocks
PLATE-BEARING TEST
The deformability of rock may be measured in the field by loading a rock
surface and monitoring the resulting deformation. This is easily arranged in an
underground gallery as shown in Figure 6.5a. The site must be selected care-
fully to exc1ude loose, highly fractured rock that might be unrepresentative of
the average rock condition. A relatively flat rock surface is sculptured and
leveled with mortar to receive circular bearing plates 50 cm to 1 m in diameter.
Dial gage
Bearing plate
Mortar
(a)
(b)
Tensioning
head
Cable
Anchorage
Figure 6.5 Plate-bearing test setup. (a) In a gallery. (b)At the surface.
plate
6.3 Measurement of Diformability Properties by Sta tic Tests 189
The depth of the rock volume affected is proportional to the diameter of the
loaded area, SO it is desirable to choose a large bearing plate; but it proves
difficult to apply loads greater than about 200 ton s so it may be necessary to
reduce the plate size to achieve desired contact pressure levels. The load can
be applied by hydraulic cylinders or screw jacks reacting against the opposite
wall of the gallery. Flat jacks, in series to allow sufficient "travel," have also
been adapted for this function. Displacement must be measured at several
points on the bearing plate to correct for rotations and plate bending. Displace-
ments are usually monitored by mounting dial gages on a rigid inertial reference
bar passing over the plateo It is also possible to use a benchmark set at depth in
a borehole centered on the plate (Figure 6.5a). Plate-bearing tests can be run at
the surface by jacking against a cable anchored at depth in a borehole drilled
through the center of the bearing plate (Zienkiewicz and Stagg, 1966).
The data obtained from the plate-bearing test are the radius of the plate a,
the plate pressure p (contact force per unit plate area), and the mean displace-
ment ;; corrected for rotation. The following equation can be derived from
Timoshenko and Goodier (1951) if we assume that the rock is a homogeneous
infinite half space of elastic isotropic material:
_ Cp(1 - v
2
)a
w= E
(6.10)
Assuming a value for v permits calculation of E.
C is a constant depending on the boundary conditions. If the plate is per-
fectly rigid, C = 'TT'12, whereas ifthe plate is flexible, C = 1.70; we see that there
is little difference in the calculated E corresponding to extremes in boundary
conditions as long as mean displacement of the plate is measured. However,
determining the mean displacement of a flexible, bending bearing plate corre-
sponding to a constant pressure boundary condition would require more dial
gages than there is generally room to accommodate. Unless the rock is very
hard, it will be simpler to attempt to achieve constant plate displacement condi-
tions by using thick steel plates and a stiff arrangement.
The plate-bearing test is conducted in a gallery more often than on a semi-
infinite medium, but Equation 6.10 is still used to compute E. A more important
influence on the results is departure from nonideal rock conditions. This can be
appreciated by using deep bench marks for displacement measurement. It
should be noted that almost any departure from conditions assumed will tend to
increase the measured displacements, so the plate-bearing test tends to under-
estimate the modulus of elasticity. Tests conducted vertically in galleries will
usually give even lower values of E because joints in the roof rock tend too open
under gravity.
Permanent and elastic deformations in plate bearing results can be sepa-
rated if the load is cyc1ed during the test. The elastic modulus should be calcu-
lated from the slope of the reloading portion of a load cycle:
190 Deformability of Rocks
E = Ca(1 - V
2
) jl
Jelas
(6.11)
where Welas is the mean plate displacement on reapplying plate pressure from
nearly zero to p. The modulus ofpermanent deformation M can be caleulated
from
M = Ca(1 - v
2
) --,-p--
J - Jel
as
(6.12)
Using plates of 14- to 50-cm radius, with plate pressure up to 200 bars,
Schneider (1967) tested foundations of a number of dam sites in various rock
types. He observed that permanent deformation greater than 0.01 mm/bar char-
acterized sites that had been considered to be unacceptable for foundations of
concrete dams. This translates to a value of M = 7700 MPa (1.1 x 10
6
psi) for
a = 50 cm, v = 0.3.
BOREHOLE AND GALLERY TESTS
Rock deformability may also be measured statically in boreholes. The dilatom-
eter test (Figure 6.6a) is a borehole expansion experiment conducted with a
rubber sleeve. The expansion ofthe borehole is measured by the oil or gas ftow
into the sleeve as the pressure is raised, or by potentiometers or linear variable
differential transformers built inside the sleeve. The gallery test is a similar
experiment conducted inside a bulkheaded section of a tunnel. The cost of
gallery tests has tended to minimize their application in recent years. The
borehole jack is similar to the dilatometer except that the loads are applied
unidirectionally across a diameter. Interpretation is similar but the formula
requires attention to the more difficult boundary conditions (Goodman et al.,
1972; Heuze and Salem, 1979). For the dilato meter or gallery test, in which the
pressure p is applied uniformly over the borehole or gallery surface of radius a,
the modulus of elasticity can be caleulated from the measured radial deforma-
tion il.u by
a
E = (1 + v)il.p il.u (6.13)
One problem with borehole deformability tests is that they affect a rela-
tively small volume of rock and therefore contain an incomplete sample of the
fracture system. Sorne would argue that the system is indeterminate and there-
fore that the test is useless. However, the borehole tests have the unique
advantage of giving an indication of the range of properties of the rock remote
from the surface at an early stage of investigation. Based upon the results of
such a program of tests, it is possible to appreciate potential site difficulties,
and it should be possible to subdivide the volume of rock in a foundation into
6.3 Measurement of Deformability Properties by Static Tests 191
\ I
/ /\
1;
"-, -j-
I -:::-/ ~ / ~ el /
- 1/_ /
-- I --
(al
Pressure !rom
curved
"flat jacks"
\ \ --
/'
(bl
Steel ring
sets and
longitudinal
members
Figure 6.6 Schemes for pressuring the interior of a cylindrical space. (a) Dilatometer
or gallery test, with fluid pressure inside the test region. (b) Radial jacking, or
TlW AG test, with pressure supplied by jacks reacting against interior sets.
approximately homogeneous subregions. Further testing can be carried out to
characterize each of these subregions. Field tests of a larger scale require
galleries and are more expensive. They too present difficulties in interpretation
since usually no field test will be as large as the rock volume affected by an
actual structure.
RADIAL JACKING TESTS
Among the largest in situ tests used to measure deformability of rock are radial
jacking tests (Figure 6.6b), an adaption of the "TIWAG" test used in Europe.
Loads are applied to the circumference of a tunnel by a series of jacks reacting'
against circular steel ring members. The test allows the direction of load to be
varied according to the plan for pressuring the jacks. Such tests were con-
ducted by the Bureau of Rec1amation at the site of the Aubum Dam-a "ite
where seams of tale schist raised questions about deformability and stability of
the abutments. The tests were expensive but were defensible in terms of the
total cost of this enormous project. However, lab tests, borehole tests, and
plate-bearing tests were also ron at the Aubum Dam site and through these a
good understanding of the variation and distribution of deformabiiity values
throughout the dam site was gained.
192 Deformability of Rocks
FLAT JACK TESTS
The ftat jack test, previously discussed in connection with stress measure-
ments, yields deformability properties as a by-product. A large volume of rock
can be loaded to pressure up to 70 MPa or higher using stainless steel ftat jacks
with special welding details.
The area of typical ftat jacks is of the order of 600 cm
2
and much larger
jacks have been used; thus very large loads are applied to the rock. Recall that
the pressuring stage of the ftat jack test provides data on the variation of pin
separation 2Ay with applied jack pressure p (Figure 6.7). If load cyc1es are
programmed, the reloading relationships will permit calculation of E using a
relationship derived by Jaeger and Cook (976):
E = p(2e) [o -v) ( ~ 1 + y2 - ~ ) + 1 + v l
2Ay e
2
e R2
1 +-
e
2
(6.14)
do = 2y
~ - - - - - - - - - - - - - - - - - - - 2 c - - - - - - - - - - - - - - - - - - ~
Figure 6.7 The flat jack test, with a slot obtained by drilling overlapping
holes.
6.4 Dynamic Measurements 193
where y is the distance from the jack center to each of a pair of measuring pins
and 2e is the length of the jack. (The modulus of permanent deformation M c a ~
also be obtained as discussed for the plate bearing test.)
6.4 Dynamic Measurements
The velocity of stress waves may be measured in laboratory rock specimens
and in the field. The laboratory pulse velocity test is run using sections of
cylindrical core with smooth, parallel ends to which piezoelectric crystals are
cemented (Figure 6.8a). A high-frequency electrical pulse transmitted to one
crystal creates a stress wave that is received by the second crystal and re con-
verted to an electrical signal. A delay line allows the received wave form to be
(a)
)
/
(b)
Transverse
geophone
Longitudinal
geophone
Figure 6.8 Scheme for dynamic modulus determination. (a) Pulse
velocity method in the laboratory. (b) Sound velocity measurements in
the field.
194 Deformability of Hoc1es
aligned to the sending wave form on an oscilloscope and the required delay to
achieve this measures the travel time for the puise through the specimen. The
use of longitudinal and shear cut crystals permits both longitudinal waves and
transverse waves to be observed so that both longitudinal wave velocity VI and
transverse wave velocity VI can be determined. If the rock were an ideal
elastic, isotropic solid of small diameter compared to the length, then E and G
could be calculated from
E = Vrp
and
G = Vp
where p is the mass density of the rock. Recalling that G
(Equation 6.3)
(6.15)
(6.16)
E/[2(1 + v)]
(6.17)
In the field, wave velocity can be measured by swinging a sledgehammer
against an outcrop and observing the travel time (milliseconds, typically) to a
geophone standing on the rock at a distance of up to about 50 m. Portable
seismographs available from several commercial sources are suited to such
measurements. Another method is to record the time for a shock to travel
between points in drill holes spaced 50 to 100 m aparto Both downhole hammers
and explosive sources are used for such measurements. If the "signature" of
the wave arriving at the geophone is displayed, both compressional wave ve-
locities V
p
and shear wave velocities V
s
can be determined. Then, assuming the
rock to be homogeneous, isotropic, and elastic,
_ ~ A + 2G
Vp - p
(6.18)
and
(6.19)
Recalling Equations 6.3 and 6.4
(6.20)
and
E = 2(1 + v)pV; (6.21)
6.5 Fractured Hoc1es 195
or
E = (1 - 2v)(1 + v) V2
(1 - v) p p
(6.22)
The stress loadings sent through the rock by these methods are small and
transient. Most rock masses and even rock specimens depart significantly from
the ideal material s hypothesized with respect to Equations 6.15 to 6.22. Conse-
quently, elastic properties ca1culated from these equations are often considera-
bly larger than elastic properties calculated from static loading tests like plate
bearing. This is particularly true in the case of fractured rocks. To distinguish
elastic properties measured by static methods from those obtained dynami-
calIy, the subscripts s and d will be introduced (Es, V
s
and E
d
, Vd for static and
dynamic constants, respectively).
6.5 Fractured Rocks
Plate-bearing tests in fractured rocks reported by Schneider (1967) typically
yielded a load deformation curve of the form shown in Figure 6.9, with a yield
p
Figure 6.9 Typical data from plate-bearing tests in fractured rock reported by
Schneider (1967); pis the average pressure on the plate and w is the average
displacement of the plateo
196 Deformability of Bocks
point effect. The slope of the envelope ofload cycles, the "yield function," is
termed r. Schneider found that in highly fractured rock with open cracks, the
ratio El r was as high as 45. He proposed the following classification of results
(Table 6.2):
Table 6.2
Class E/r
Compact rock <2
Moderately open 2-10
Very open > 10
If the rock is regularly crossed by a single set of joints, it is possible to
ca1culate elastic constants for an "equivalent" continuous material representa-
tive ofthe rock mass. We as sume the rock itselfis isotropic and linearly elastic,
with constants E and v (Figure 6.10). The joints are assumed to be regularly
spaced a distance S. Let ks-the shear stiffness-be the slope of the shear
stress-shear displacement curve until slip (Figure 5.12). We adopt axes n, t
normal and parallel to the joints and therefore in the principal symmetry direc-
tions of the rock mass. When shear stress Tnt is applied, each rock block un-
dergoes a displacement equal to (TntlG)S and eachjoint slips a distance (Tntlks)
(Figure 6.lOb). The shear deformation of a continuous material will be equiva-
lent to that of the jointed rock mass if it has shear modulus G
nt
such that
(TntIGnt)S is the sum of rock and joint displacements given aboye. Therefore,
111
-=-+-
Gnt G ksS
(6.23)
Similarly, we assign the joint a "normal stiffness" k
n
equal to the slope of the
joint compression curve U' versus Llv (Figure 5.17 a). Since the compression
curve is highly nonlinear, k
n
depends on the normal stress. The equivalent
continuous material has modulus of elasticity En such that (U'nIEn)S is the sum
ofrock deformation (U'IE)S andjoint deformation (U'lk
n
) (Figure 6.lOa). There-
fore,
111
-=-+-
En E KnS
(6.24)
The Poisson's ratio giving strain in the n direction caused by a normal
stress in the t direction is simply v:
Vtn = V (6.25)
The modulus of elasticity in the t direction is simply E:
Et =E (6.26)
6.5 Fractured Rocks 197
T
k.
_ T_
I
,
,
I
I
I
G
I
Rack
I
I
I
:
i
Jaint
...,
I
I
I
I
I
I
I
j
H
1.s+ 1.
G k.
(a)
(b)
~ T _
,
1
I
I
I
I I
I I
I
G
nt
I
I
I S
I
I
I
I
I
Equivalent
I
r
I
cantinuaus
I
I
rack
i 1 ~ O
Figure 6.10 Representation of a regularly joirited rock by an "equiva-
lent" transversely isotropic material.
Finally, symmetry of the stress-strain relationship requires Vtn l Et
giving
(6.27)
Equations 6.23 to 6.27 permit calculation of all five constants of the equivalent
transversely isotropic medium representing a regularly jointed rock mass.
198 Deformability of Rocks
If the rock is highly fractured in several directions, Equation 6.24 can be
used to find a "reduced modulus" representing the rock mass. The procedure
is as follows. For each test specimen or test site, determine a characteristic
average spacing between the joints of each set. From the measured value of the
modulus of elasticity and assigning a value E to intact rock, calculate a value
for k
n
using Equation 6.24. Then, input this value of k
n
in calculations with any
specified fracture spacing. The rock mass modulus can in this way be related to
degree of fracturing (Raphael and Goodman, 1979) or to RQD (discussed in
Chapter 2) (Kulhawy 1978).
Bieniawski (1978) showed that a rock mass modulus could be assigned
approximately if the rock were rated by the geomechanics c1assification system
discussed in Chapter 2. Figure 6.11 shows his values of in situ modulus of
deformation, determined by various large-scale field tests at a number of sites,
plotted against the rock mass rating (RMR). For rocks rating higher than 55, the
data points are fit approximately by
E = 2 RMR - 100
For softer rocks (10 < RMR < 50), Serafim and Pereira (1983) gave the
following correlation between rock mass modulus of elasticity and RMR:
E = lO(RMR -10)/40
0.." 80
M
o
",'
2 RMR - 100
c'
o
60 .:
ro
E
.E
Q)
'O
....
o
'" 40
::J
'O
o
E
.= o
.;;
.::
20
40
Rock mass rating
Figure 6.11 Relationship between rock mass rating and rock mas s
deformability. (After Bieniawski, 1978.)
100
6.5 Fractured Rocks 199
The term modulus of deformaton signifies that the value of E is calculated
from the data of the loading portion of the load/ deformation curve using both
elastic and permanent deformation. The units of E in the aboye are GPA (= 10
3
MPa). The data points embrace mudstone, sandstone, diabase, slate, phyllite,
and quartzite.
Dynamic moduli Ed in fractured rocks tend to be considerably higher than
rock mass moduli measured by static load tests Es or computed as aboye.
Schneider (1967) determined values ofthe ratio Ed/E
s
up to 13 in fractured hard
rocks. He observed that high frequencies are selectively attenuated in fractured
rock. This was also shown by King et al. (1975) (Figure 6.12). One would
Q)
'O
::J
O
::::
c.
E
<{
01------'--
6 fractures/m
ROO = 0.81
10 fractures/m
ROO = 0.60
16 fractures/m
ROO = 0.49
OL-____ ____ ____ -L ____
Time, lAS
Figure 6.12 Acousti.c logging signals received in
pegmatite fractured m vanous amounts. (After King et al., 1975.)
200 Deformability of Hoeles
(a)
(b)
E
~
M
~
:l
t>
~
' ~
'O ,s
o
z
(e)
[=2-
T
Figure 6.13 Relationship between transmitted vibration
frequency and degree of fracturing observed by Sch-
neider (1967). (a) Typical wave group traveling through
the rock as a result of a hammer blow. (h) Measurement
of period using a first return seismograph by separation
of time-distance graphs caused by changing the polarity
of the instrument. (e) Inverse relationship between fre-
quency and degree of fracturing.
6.5 Fraelured Hoeks 201
expect then that measurements of frequency Qr wave length received at a
standard distance from a constant type of seismic source would tend to corre-
late with Edl Es. Figure 6.13 confirms such a relationship for dam sites studied
by Schneider using a hammer blow source with an engineering seismograph
(MDl). The instrument yielded only the time for the arrival of the first wave
having a signal aboye an adjustable threshold. Switching the polarity on the
seismograph and repeating the experiment will cause a change in arrival time of
approximately half the period, as shown in Figure 6.13a, b. Thus the fre-
quency, velocity, and wavelength can be determined using a hammer source
and a simple seismograph. The dynamic elastic constants can then be related to
the static elastic constants through site calibration studies.
Alternatively, site studies can establish a direct relationship between in situ
static modulus of elasticity and shear wave frequency. For example, Figure
6.14 shows the relationship between in situ static modulus of deformation and
shear wave frequency using a hammer seismograph with standardized tech-
50
,'
e:
o
':;
'"
E
~
., 30
'O
'O
'"
.2
:l
'O
o
E
:
20
';;
C:
10
o
Data from Bieniawski (1978)
o Data from Schneider (1967) O,054[ - 9,2/
{
o ~ ~ /
/'
/0
o /
o/
/0 o
/ o
./
.7' .
Shear wave frequency, [, cps,
using a hammer seismograph
//
/
///
1000
Figure 6.14 Relationship between transmitted vibration fre-
quency and rock mas s deformability. (After Bieniawski,
1978.)
202 Deformability of Rocks
nique (Bieniawski, 1978). Both Schneider's results and Bieniawski's results are
fit by the same straight line:
E = 0.054f - 9.2
where E is measured in gigapascals (GPa),1 fis the shear wave frequency from
the hammer blow received at distances up to 30 m on a rock surface, andfis the
frequency in hertz (cycles per second).
6.6 The lnfluence of Time
on Rock Deformation
Thus far we have omitted all reference to time as a parameter of rock deforma-
tions. Since no effect can be truly instantaneous, time must be implicit in aH the
equations connecting stress and strain. In many cases, rock deformations can
be ca1culated satisfactorily ignoring the influence of time, but sometimes they
cannot.
Stress or displacements can change with time when the loads or pressures
on the rock change, as, for example, due to flow of water; the geometry of the
loaded or excavated region changes, as, for example, by further excavation;
the deformability properties ofthe rock change, as, for example, by weathering
or hydration; or the rock responds slowly to changes in stress or strain. AH but
the last factor can be accommodated by appropriate superposition of stress
increments in a series of elastic analyses. However, the last reason for time
dependency which we might term viscous behavior, requires further discus-
sion.
VISCOUS BEHAVIOR AND CREEP
We can view solids as bodies that retain their shape indefinitely, while liquids
as sume the shape of their containers. An apparentIy solid material that distorts
slowly and continuously in response to shearing stresses is then at least partly a
viscous liquido Dynamic viscosity 'Y/, depicted by a "dashpot" (Figure 6.15b),
expresses the proportionality between shear stress T and shear strain rate -y:
T = 'Y/'Y (6.28)
Since strain is dimensionless, the dimensions of 'Y/ are FL -2T, for example,
psi/min or MPa/s.
2
Most rocks exhibit both "instantaneous" and delayed de-
formation when loaded and are therefore spoken of as viscoelastic. As with
1 One gigapascal (1 GPa) = 1000 MPa.
2 The cornrnon units "poise" (P) will not be used here; 1 P = 0.1 Pa/s = 1.450 x 10-
5
psi/s.
6.6 TIte Influence of Time on Rock Deformation 203
~ ~ O O ~ , - __ _ . ~ T _
(a)
(b)
Figure 6.15 Elements of linear visco-
elastic models. (a) Linear spring. (b)
Linear dashpot shock absorber).
elasticity, real deformation data can display various non-linearities while the
majority of theory concerns linear viscoelasticity.
Laboratory data are mostly in the form of strain-time curves from creep
tests. In such tests, an increment of stress is applied quickly and then held
constant while the gradualIy increasing strain is recorded. An alternative exper-
iment, termed a relaxation test, monitors the decline in stress when strain is
held constant. Figure 6.16 shows the general form of the creep curve for rock.
Immediate "strain" is followed by primary creep in which strain occurs at a
decreasing rate with time. In sorne rocks, the primary creep curve approaches a
steady rate of creep, termed secondary creep. In specimens stressed near peak
strength, secondary creep may turn upward in tertiary creep, in which strain
rate increases with time, resulting in failure (creep rupture).
We can caH on two types of mechanisms to explain creep in rocks-mass
flow and cracking. Sorne rocks (e.g., rock salt, tar sands, and compaction
shales) will creep at relatively low deviator stress, even with unfissured, intact
specimens. In the case of salt and potash the process of creep involves move-
ment of dislocations and intracrystalline gliding, while creep in uncemented
clay rocks involves migration of water and movements of clay platelets ("con-
solidation"). Bituminous rocks like tar sand are inherently viscous, especially
at higher temperatures. Hard rocks like granite and limes tone can also exhibit
creep at deviatoric stresses sufficient to cause new crack growth (e.g., when SI
exceeds about one half qu in an unconfined compression specimen). An incre-
ment of applied stress will provoke a change in the network of cracks thi"ough
lengthening of old cracks and initiation of new ones. Such a process is nonlin-
ear because the rock is changed by each new load increment; to ca1culate
stresses and deformations in nonlinear viscoelastic materials, the properties
will have to be determined and used as functions of stress. There are probably
204 Deformability of Rocks
Rupture
rrm
o
Primary Secondary
Tertiary
Figure 6.16 Regions of behavior in creep.
no ideal1y linear viscoelastic rocks, not even salt. However, a theory of linear
viscoelasticity can still be used incrementally to approach time-dependent
problems in much the same way that the theory of linear elasticity is used to
ca1culate stress and strain for time-independent problems.
LINEAR VISCOELASTIC MODELS
It is possible to fit creep curves empirically using exponential or power func-
tions. If creep data are forced to fit models composed of springs and dashpots,
however, the results are more readily usable; therefore, this will be our pro-
cedure.
The theory of linear elasticity of an isotropic body is based on two con-
stants, as we have seen previously. One of these (K) may be associated with
purely volumetric deformation under hydrostatic loads. Then a second con-
stant alone (G) must account for all distortion. The question we now face is,
how many additional constants will be required to represent time-dependent
deformation?
Figure 6.17 shows five possible models with one, two, or three additional
constants. The series arrangement in Figure 6.17a is termed a Maxwell body or
G
(a)
G
I
:J
T
...
(b)
G
T
(e)
G
1
a
'Y
(d)
'Y
(e)
Figure 6.17 Simple linear viscoelastic models and their response to the .
creep test. (a) Two-constant liquid (Maxwell body). (b) Two-constant solid
(Kelvin body). (e) Three-constant liquid (generalized Maxwell body). (d)
Three-constant solido (e) Four-constant liquid (Burgers body).
205
206 Deformability of Rocks
a two-constant liquido It flows continuously at a constant rate when a shear
stress is applied suddenly and held constant. Figure 6.17 b is a Kelvin or Voight
body or a two-constant solido A suddenly applied constant shear stress causes
shear strain at an exponentially decreasing rate, approaching zero as t grows
without bound.
Figures 6.17 e and d show a three-constant liquid and a three-constant solid,
respectively. The former, termed a generalized Maxwell body, initially has an
exponential rate of shear strain, decaying to become asymptotic to a constant
rate of shear strain. The latter, termed a generalized Kelvin body, shows an
initial "instantaneous" strain followed by shear strain at an exponentially de-
creasing rate, eventually tapering off completely.
Finally, Figure 6.17 e shows a four-constant liquid termed the Burgers
body, composed of a Maxwell and a Kelvin body in series. Its response to a
suddenly applied and sustained shear stress is a combination of all the elements
we have seen in the previous models-initial "instantaneous" shear strain
followed by shear strain at an exponentially decreasing rate, becoming asymp-
totic to a line representing a constant rate of shear strain. In view of the form of
the general creep curve, Figure 6.16, this is the simplest model that can be used
to trace strain up to the onset of tertiary creep. More complicated models can
be invoked by adding additional springs and dashpots, but the Burgers body
will suffice and is preferable for many practical purposes. An informative com-
parison of various spring-dashpot models and empirical formulations for creep
data was reported by Afrouz and Harvey (1974) for sedimentary rocks. Of the
spring-dashpot models, the Burgers representation was consistently the best.
DETERMINING VISCOELASTIC CONSTANTS FROM
LABORATORY TESTS
The simplest procedure for evaluating viscoelastic constants is through uncon-
fined compression of cylindrical rock specimens over prolonged periods. This
requires constant stress and constant temperature and humidity over the whole
test duration, which may be hours, weeks, or longer. Load may be applied by
dead weights acting through levers bearing directly on the specimen or through
an oil pressure. Servocontrolled hydraulic pressure systems and compressed
springs are also used. Careful testing procedure insures correction of load for
changes in the cross sectional area of the specimen (see, for example, Rutter,
1972) and measurement of strain without long-term drift.
The axial strain with time e,(t) in a Burgers body subjected to constant
axial stress (T" is
e,(t) = 2(T, + ~ + (T, _ ~ e-(G,lhll) + ~ t
9K 3G
2
3G, 3G, 31/,
(6.29)
6.6 The Injluence of Time on Rock Diformation 207
a,
;0 (2+--.!... +--1-)
B '9K 3G
2
3G,
1 1
fa; a, (9K + 3G2)
Figure 6.18 Creep in uniaxial compression of a rock that behaves
as a Burgers body under deviatoric stress but as an elastic body
under hydrostatic compression.
where K = EI[3(1 - 2v)] is the bulk modulus, assumed to be independent of
time
3
and 1/" 1/2, G" and G
2
are properties of the rock to be evaluated as
follows.
Figure 6.18 is a graph of e, versus t corresponding to Equation 6.29. At t =
0, there is an intercept eo = (T,(2/9K + 1I3G
2
) while strain at large t falls along
the line with intercept eB = (T,(2/9K + 1/3G
2
+ 1I3G,) and slope (T,/31/2. Load
cannot be applied instantaneously and it may be preferable in practice t ~ find
3 The assumption that volumetric strain is non-time-dependent is not satisfied generally, but it
simplifies problems without usually introducing large errors. It is feasible to define additional
constants to represent time dependency of volumetric strain, but this will not be done here.
20S Dliformability of Hocks
the intercept eo by regression. Let q equal the positive distance between the
creep curve and the line asymptotic to the secondary creep curve (Figure 6.15).
Then
(
<T) G
log q = log 3G - 2.3'1/ t
(6.30)
A semilog plot 10glO q versus t has intercept <T/3G and slope -G/2.3'1/ deter-
mining G and '1/ .
If lateral strains e3 are monitored as well as axial strains e, the volumetric
strain is determined by V/V = e + 2e3 (see Equation 3.6) while the mean
stress is <T/3. Therefore K can be calculated by
(6.31)
and G2 can be calculated from
<T ( 1 2 )
3G
2
= en - <T 3G + 9K
(6.32)
As an example, we will find the Burgers constants for Indiana limestone
using the data from creep tests by Hardy et al. (1970) presented in Figure 6.19.
The rock is homogeneous limestone with a mean grain size of 14 mm and
porosity of 17.2%. The unconfined compressive strength qu was 9000-11,000
p:i (dry). Cylinders of limestone, 1.12 in. in diameter and 2.25 in. long, were
loaded by dead weights on levers, in increments so that creep curves were
obtained at different axial stresses. Time dependency was absent in tests with
axial stress less than 40% of qu, and secondary creep was unimportant when <T
was less than about 60% of qu. Table 6.3 summarizes the pertinent data for one
specimen. Straight lines were drawn asymptotic to each creep curve in Figure
6.19, giving slopes <T/3'1/2 and intercepts en. Regression analysis on a pocket
calculator determined the constants <T/3G and G/'1/. (Semilog plotting
could have been used instead.) The determined values of K, G, G
2
, '1/, and '1/2
are given in Table 6.4. Note that G and the viscosity terms are large for the first
two increments when there was no time dependency and beco me progressively
smaller as the axial stress augments. G
2
and K are almost independent of stress.
This is nonlinear viscoelasticity of the type derived from the initiation and
growth of cracks. These deformability constants have real physical meaning:
G2 is the elastic shear modulus; G control s the amount of delayed elasticity; '1/
determines the rate of delayed elasticity; and '1/2 describes the rate of viscous
flow.
.....
e
<1)
e
~
-
.,.
C'l 00 0'\ r-
"" C'l C'l C'l
0000
......
210 Deformability of Rocks
280
260
240
220
7
200 6
5
180
:..:::;
e
''
Increment 4
e
ro 120
'x
<t:
100
80
60
40
20
Time, min
Figure 6.19 Creep of Indiana limestone in
unconfined compression. (Data from Hardy et
al.,1970.)
32
Table 6.4 Burgers Body Constants Fit to Data of Table 6.3
Percent qu K G
1
G
2 1/1
Increment after Loading (10
6
psi) (10
6
psi) (10
6
psi) (10
6
psi/min)
1 and 2 39 3.7 Large 2.7
00
3 60 4.5 28.9 2.1 84
4 67 3.8 23.0 2.5 71.8
5 78 3.6 16.1 2.3 57.5
6 83 4.0 13.6 2.2 46.1
7 91 4.5 7.8 2.0 28.9
1/2
(10
6
psi/min)
00
00
2120
1630
640
620
6.6 The lnfluence of Time on Rock Deformation 211
DETERMINING VISCOELASTIC CONSTANTS
FROM FIELD TESTS
Any field test in which the load can be sustained for days or weeks can be used
to evaluate viscoelastic constants of rock masses. Corrections might be neces-
sary for changes in environmental conditions for tests conducted on the surface
(e.g., the plate-bearing test using cables). But in boreholes and in underground
galleries, temperature and humidity are often unvarying.
The dilatometer test is convenient for creep tests. Unlike the laboratory
compression test on rock cylinders, there is no change in the mean stress when
a dilatometer is pressured and expanded against the borehole. Therefore, there
is no term in K influencing the time history of radial displacement. The outward
radial displacement at the wall of the borehole (r = a) in a Burgers material is
pa pa pa _ pa
u(t) = - + - - -e (G/ITJ) +-t
r 2G
2
2G
1
2G, 21/2
(6.33)
in which p is the internal pressure in the dilatometer.
The displacement with time follows a curve, Figure 6.20, like that for the
compression test and the analysis of the data is as previously discussed, except
that the intercepts and slopes have different values. At t = 0, the radial dis-
placement is
pa
uo = 2G
2
The asymptote to the displacement-time curve has intercept
(6.33a)
(6.33b)
and slope pa/21/2. Again letting q equal the positive vertical distance between
the asymptote and the displacement-time curveat any time,
pa G,
log q = log - - --t
2G, 2.31/,
Thus a series of sustained pressure increments in dilato meter tests will
the constants G, G
2
, 1/ , and 1/2 to be determined.
The plate-bearing test can also be conducted to yield the viscoelastic con-
stants. Now, however, there are terms with the bulk modulus K because the
mean pressure as well as the deviatoric stress changes when the is
applied. For a constant pressure P applied suddenly to a flexible bearing plate of
212 Diformability of Rocks
1 1
u
B
= pa(2G1 + 2G2)
1
Uo = pa(
2G
2)
Figure 6.20 Creep response to a dilatometer or gallery test if the
rock behaves as a Burgers body under deviatoric stress.
circular shape and radius a, the mean displacement w varies with time accord-
ing to
As a simplification, it may be acceptable to as sume that the rack is incompress-
ible (K = 00, v = l). In this instance, (6.34) reduces to
w = 1.70pa (J... + .!.... + J... (1 - e-<G1fhll ))
4 G2 'Y/2 G1
(6.34a)
6.6 The lJifluence of Time on Rack Diformation 213
The displacement is then Wo = (1.70paI4)(l/G2) and after the delayed
elastIclty has occurred, the settlement of the plate tends to the line
WB = 1.70pa (J... + J... + .!....)
4 G2 G, 'Y/2
(6.34b)
An analysis of the field test results imitating that performed for the creep
test will therefore yield values for G" G2, 'Y/" and 'Y/2. (If vis less than 0.5, all
but 'Y/2 will be in error.)
For a test with a rigid bearing plate, it is tempting to replace the factor 1.70
by 7T12, in analogy to the elastic case. However, this would not be strictly
correct in the viscoelastic case.
Data fram long-term plate-bearing tests in schists and sandstones are pre-
sented by Kubetsky and Eristov (1970).
TERTIARY CREEP
Secondary creep at stresses approaching peak values will end with tertiary
creep and rupture, possibly violent (Figure 6.16). Recalling Figure 3.13, the key
parameter identifying the time when tertiary creep begins is the accumulated
strain, for when the stress and cumulative strain define a point on the right side
of the complete stress-strain curve, rupture will occur. Data by John (1974) for
norite, a basic igneous intrusive rack, demonstrate this principIe. Creep tests
were run fram different starting stresses in unconfined compression. Figure
6.21a presents plots ofaxial stress versus log time. The horizontallines, show-
ing the paths of creep tests, terminate in points along a negatively inclined
locus. Strain varies with time (although nonuniformly over the family of creep
tests), so Figure 6.21 can be mapped into stress-strain space.
EFFECT OF STRESS BATE
Figure 6.21a also shows that the strength of norite falls when the stress rate is
slowed. A specimen loaded to peak stress at 2.1 MPa/s (over about 100 hr)
developed about two-thirds of the strength of a specimen loaded to peak stress
at 1.8 x 10
4
MPa/s (over about a twentieth of a second). Figure 6.21b, also from
John, shows stress-strain curves for unconfined compression ofnorite at differ-
ent stress rates, demonstrating reduction in stiffness when the stress rate is
lowered. Note that the elastic modulus (the slope ofthe axial stress versus axial
stmin curve) is unaffected by stress rate until a certain stress has be en reached
and thereafter there is pronounced curvature and apparent yielding before
rupture (rupture points are not shown). These observations can be by
viscoelasticity theory.
214 Diformability of Hocles
400
:. 300
f
200
10-2
10-' 10
Time, s
Axial and lateral strain
(/)
(h)
- Continuously applied stress
------ Creep
Strength locus in continuous
loading tests
Strength locus for creep tests
o Rupture
:h:--t=-
-----== ----
--------
0.1 1.0 10 100
hr
Volumetric strain
Figure 6.21 Results of dynamic tests on norite by M. John (1974). (a) Stress history
on various loading paths. (b) Deformation versus axial stress at varying stress rates.
Stress rates are keyed as follows (MPa/s): (1) = 8.4 x 10-4, (2) = 4.1 x 10-
3
, (3) =
6.4 x 10-2 , (4) = 2.1, (5) = 2.5 x 1Ql, (6) = 2.2 x 1Q2, (7) = 3.9 x 10\ (8) = 1.8 x
1()4 and (9) = 2.8 x toS.
6.6 The lrifluence of Time on Hock Deformation 215
When a constantIy increasing stress (TI at a stress rate "I is applied to a
rock behaving as a Burgers body in distortion and an elastic material in hydro-
static compression, the axial strain is
(
1 1 2 ) "17]1 (T2
el = (TI 3G
2
+ 3G
I
+ 9K - 3G (1 - e-(G(Th/') + (6.35)
The stress-strain relation E is then dependent on "I. For example, consider
Indiana limestone whose Burgers body constants and bulk modulus were listed
n Table 6.4. Applying Equation 6.35 to each increment of stress in turn, with
the constants 7]1, 7]2, GI , G2, and K approximately selected from Table 6.4,
yields the values of E presented in Table 6.5. In Figure 6.22, these have been
plotted incrementally defining stress-strain paths OA, OB, OC, and OD for the
four values of stress rate arbitrarily selected for the example. Let us assume
that the complete stress-strain curve has a uniquely defined right side defined
a,
10,000 psi A
-1--'\0---------------- 8737
O
->.---------------- 7955
7174
----------- 6392
D
'\
5723
'\
'\
'\
'\
\.
\.
'\
'\
,
'\
5 X 10-
3
P ,
Figure 6.22 Stress versus strain calculated for Indiana limestone in unconfined
compression at various stress rates.
-o
o
11 ,-....
cC
f
.6' <18
'-'
..,.. 00
V) IQ 00 0\..,.. N
..,..00<,,>0\..,....,..
IQ..,..IQ..,..OIQ
....;- IQ" ., .,
--
Rtiferences 217
by line AP (for zero confining pressure). The stress-strain curves must reach
peak values where they meet this line, at points A, B, e, and D. In this idealized
example, we can appreciate how the rate ofloading can alter both deformability
and strength of a rock .
References
Mrouz, A. and Harvey, J. M. (1974) Rheology of rocks within the soft to medium
strength range, Int. J. Rock Mech. Min. Sci. 2: 281-290.
Benson, R. P., Murphy, D. K., and McCreath, D. R. (1969) Modulus testing ofrock at
the Churchill Falls Underground Powerhouse, Labrador, ASTM Spec. Tech. Rept.
477, 89-116.
Bieniawski, Z. T. (1978) Determining rock mas s deformabilitY-Experience from case
histories, Int. J. Rock Mech. Min. Sci. 15: 237-248.
Deere, D. V. (1968) Geological considerations, in Stagg, K. G. and Zienkiewkz, O. C.
(Eds.), Rock Mechanics in Engineering Practice, Wiley, New York.
Flgge, W. (1975) Viscoelasticity, 2d ed., Springer, Berlin.
Goodman, R. E. and Duncan, J. M. (1971) The role of structure and solid mechanics in
the design of surface and underground excavation in rock, Proceedings, Confer-
en ce on Structure, Solid Mechanics and Engineering Design, Part 2, paper 105, p.
1379, Wiley, New York.
Goodman, R. E., Van, T. K., and Heuze, F. E. (1972) The measurement of rock
deformability in boreholes, Proceedings, 10th Symposium on Rock Mechanics
(AIME). pp. 523-555.
Hardy, H. R., Jr., Kim, R. Y., Stefanko, R., and Wang, Y. J. (1970) Creep and micro-
seismic activity in geologic materials, Proceedings, 11th Symposium on Rock Me-
chanics (AIME), pp. 377-414.
Heuze, F. E. and Salem, A. (1979), Rock deformability measured ln-situ-Problems
and solutions, Proceedings, International Symposium on Fielfl Measurements in
Rock Mechanics (Balkema, Rotterdam), Vol. 1, pp.
Jaeger, J. C. and Cook, N. G. W. (1976) Fundamenials ofRock Mechanics, 2d ed.,
Chapman & Hall, London.
John, M. (1974) Time dependence of fracture processes of rock materials (in German),
Proc. 3rd Congo ISRM (Denver), Vol. 2A, pp. 330-335.
King, M. S., Pobran, V. S., and McConnell, B. V. (1975) Acoustic borehole logging
system, Proceedings, 9th Canadian Rock Mechanics Symposium (Montreal).
KUbetsky, V. L. and Eristov, V. S. (1970) In-situ investigations of creep in rock for the
design of pressure tunnellinings, Proceedings, Conference on In-situ Investiga-
tions in Soils and Rocks (British Geot. Soc.), pp. 83-91. .
Kulhawy, F. H. (1975) Stress-deformation properties of rock and rock discontinuities,
Eng. Geol. 9: 327-350.
Kulhawy, F. H. (1978) Geomechanical model for rock foundation settlement, J.
Geotech. Eng. Div. (ASCE) 104 (GT2): 211-228.
218 Deformabilily of Rocks
Lane, R. G. T. and Knill, J. L. (1974) Engineering properties ofyielding rock, Proc. 3rd
Congo ISRM (Denver), Vol. 2A, pp. 336-34l.
Raphael, J. M. and Goodman, R. E. (1979) Strength and deformability of highly frac-
tured rock, J. Geotech. Eng. Div. (ASCE) 105 (GT11): 1285-1300.
Rutter, E. H. (1972) On the creep testing of rocks at constant stress and constant force,
Int. J. Rock Mech. Min. Sci. 9:191-195.
Schneider, B. (1967) Moyens nouveaux de reconaissance des massifs rocheux, Supp. to
Annales de L'Inst. Tech. de Batiment et des Travaux Publics, Vol. 20, No. 235-
236, pp. 1055-1093. . .
Serafim J. L. and Pereira, J. P. (1983) Considerations of the geomechamcs classlfica-
of Bieniawski, Proceedings, International Symposium on Engineering Geol-
ogy and Underground Construction (L.N.E.C., Lisbon, Portugal) Vol. 1, Section
11, pp. 33-42. .
Timoshenko, S. and Goodier, J. N. (1951) Theory of Elasticity, 2d ed., McGraw-Hlll,
New York.
Van Heerden, W. L. (1976) In-situ rock mass property tests, Proceedings of Sympo-
sium on Explorationfor Rock Engineering, Johannesburg, Vol. 1, pp. 147-158.
Wawersik, W. R. (1974) Time dependent behavior of rock in compression, Proc. 3rd
Congo ISRM (Denver), Vol. 2A, pp. 357-363. .
Zienkiewicz, O. C. and Stagg, K. G. (1967) The cable method ofin-situ rock testmg, Int.
J. Rock Mech. Min. Sci. 4: 273-300.
Problems
1. Show that the stress-strain relationship connecting deviatoric strain eij and
deviatoric stress 'Tij consists of six uncoupled identical statements:
'Tij = 2Geij
i,j=1,3
("Deviatoric strain" is discussed in Appendix 2.)
2. Suppose a triaxial compression test is conducted by simultaneous change in
(TI and p; derive expressions for E and /J in terms of the axial and lateral
strains and the stresses (TI and p.
3. Describe a procedure for triaxial testing that raises the deviatoric stress
while the nondeviatoric stress remains constant.
4. The following force s and displacements were measured in. an .unconfined
compression test of a cylindrical claystone specimen 5.0 cm In dmmeter and
10.0 cm long.
Axial
Force
(N)
O
600
1000
1500
2000
2500
O
2500
3000
4000
5000
Axial
Shortening
(mm)
O
0.030
0.050
0.070
0.090
0.110
0.040
0.110
0.130
0.170
0.220
Lateral
Extension
(mm)
O
0.018
0.009
0.037
Axial
Force
(N)
O
2,500
5,000
6,000
7,000
7,500
O
7,500
9,000
10,000
O
Problems 219
Axial
Shortening
Lateral
Extension
(mm)
(mm)
0.080
0.140
0.220
0.260
0.300
0.330
0.120
0.330
0.400
0.440
0.160
0.016
0.056
0.025
0.075
0.035
Compute E and /J corresponding to elastic deformation and their
parts M and /Jp for permanent deformation from the aboye data.
5. A triaxial compression test is performed as follows: (a) An all-around pres-
sure is first applied to the jacketed rock specimen. Nondiviatoric stress ii is
plotted against nondeviatoric strain e and the slope DI = aiil ae is deter-
mined. (b) Then deviatoric stress is increased while nondeviatoric stress is
held constant and the axial deviatoric stress (T1,dev is plotted against the
axial deviatoric strain el,dev. The slope D
2
= a(Tl,devlael,dev is determined
from the graph. Derive formulas expressing E and /J in terms of DI and D
2
6. (a) Derive a relationship between E, the modulus of elasticity computed
from the reloading curve of stress and strain; M, the modulus ofpermanent
deformation; and Etota, the modulus of deformation computed from the
slope of the loading curve of stress and strain. (b) Show how M varies with
axial strain throughout the complete stress-strain c,urve.
7. In a full seismic wave experiment, the compressional and shear wave veloc-
ities were measured as Vp = 4500 mis, V
s
= 2500 mis. Assuming the density
of the rock is 0.027 MN/m3, calculate E and /J.
8. What physical phenomena could explain a plate-bearing pressure versus
displacement curve like that of Figure 6.9?
9. A rock mass is cut by one set of joints with spacing S = 0.40 m. (apf the
joint normal and shear deformations are assumed to be equal to that of the
rock itself, express ks and kn in terms of E and /J. (b) Assuming E = 10
4
MPa
and /J = 0.33, calculate all the terms of the strain-stress relationship for an
equivalent transversely isotropic medium, (corresponding to Equation 6.9).
220 DeformabUity of Rocks
10. Modify Equations 6.23 and 6.24 accordingly for a rock mass with three
mutually perpendicular sets of joints.
11 Show that for rock cut by one set of joints with spacing S, the normal
. strains and normal stresses referred to n, s, t coordinates are related by
(
::) = =: ] (::)
e
t
-v -v 1 (J't
where p = 1 + E/knS and where E and v.are Young's and .Pois-
son's ratio of the rock, k
n
is the normal stlffness of the jomts, and n lS the
direction perpendicular to the joint planes.
12. A jacketed rock cube, with edge length 50 cm, is subject to an ano
pressure p. The pressure versus volumetric strain curve lS
in the figure. Assume the rock contains three mutually perpend1cular jomt
sets all spaced 5 cm aparto Calculate the normal stiffness of the k
n
at
each of the normal pressures corresponding to the start of unloadmg paths
(2.4, 4.8, and 10.3 MPa).
10.0
8.0
'"
"-
6.0
Q.,
4.0
2.0
Av/v
0.0027 0.0043 0.0058
13. Let v
p
, v, and v be respectively the Poisson's plastic, total, and
elastic strain; that is, for strain applied in the x dlrectlOn, v = -ey/e
x
, etc.
Derive a formula expressing Vt as a function of E, M, v, and V
p
Applications of
Rock Mechanics in
Engineering for
Underground
Openings
7.1 Introduction
Chapter 7
Engineering underground space has many facets, sorne of which are unrelated
to rock conditions. However, rock mechanics has direct bearing on many ofthe
critical aspects of engineering work, for example, planning the location, dimen-
sions, shapes, and orientations of chambers, selecting supports, arranging for
construction access, engineering blasting, and designing instrumentation. Rock
mechanics provides information of utmost relevance: measurement of initial
stress; monitoring stresses developed in the peripheries of openings; measure-
ment of material properties; analysis of stresses, deformations, temperatures,
and water ftow in support of design; and interpretation of instrumentation
readings, especially displacements.
The uses of underground space are many, varying from simple openings in
the dry, to large complexes of openings in dimensions, fille.d with hot or
cold ftuids of varying viscosity and pressure. Tunnels built for highways and
railroads may be short sections beneath valle y sides, or very long structures
underneath major mountain ranges. Ventilation requirements for highway 1un-
neIs tend to make these very Iarge (e.g., 15 m wide). Water supply and sewage
tunnels are generally smaller, but they may be very long and frequently operate
under internal pressure. For hydroelectric power production, pressure' head-
222 Applications of Rock Mechanics in Engineering for Underground Openings
race tunnels lead water to smface or underground penstocks and thence into
surface or underground power stations. Large water pressures are supported
by rock alone in sorne cases. The main machine hall chambers are room s with
spans of the order of 25 m, while access tunnels and other openings may also be
quite large. These chambers are feasible only if the rock is essentially self-
supporting (Figure 7.1).
Pumped storage projects may require rock tunnels, underground power-
houses and other openings as well (Figure 7.2). Energy storage is now also
demanding underground space-for storage of oil (Figure 1.4), and eventually
for hot air or hot water used in peak demand energy conversion schemes of
various kinds. Liquefied natural gas (LNG) may be stored in rock caverns more
widely if the problems of thermal cracking and loss of product can be over-
come. Nuclear wastes are to be stored in specially mined repositories in rock
salt, chosen for its high heat conductivity and general continuity without frac-
tures. In the field of mining, there are two types of underground openings:
those that are intended to be stable while the ore is removed, and those that are
intentionally collapsed to produce broken rock that is drawn off as the ground
caves. For defense, deep cavities are required to protect installations from
shock. Finally, industry has need for underground space for product storage,
offices, and even public facilities like swimming pools.
With such a vast range of underground usage, many kinds of rock mechan-
ics considerations need to be addressed. However, there are certain features
common to all underground works. They are usually inaccessible until actual
construction. Occasionally, when an existing installation is being expanded,
the engineer will have access to the site at the initiation of the job; more
usually, however, he or she will have to begin deliberations from information
acquired in drill holes, shafts, and galleries. All underground workings are
constructed in rock that is initially stressed and all openings cause changes in
the initial stress when they are constructed. Most underground workings are
made below the water tableo And all openings are constructed in an environ-
ment of even temperature equal to the mean surface temperature plus the
product of geothermal gradient and depth. The gradient of temperature varies
from OSe/lOO m to as much as 5/100 m.
When working with rock mechanics below ground, there are certain condi-
tions that should be appreciated. The underground environment is very often
hostile for instruments due to water, blasting, and truck traffic. Working space
is often cramped, poorly lit, and wet. As a result, experiments and concepts for
instrumentation underground should be as simple as possible and the equip-
ment must be rugged. Overly sophisticated testing technique or data handling,
and overly precise measurements are to be avoided. However, almost any data
taken underground near the site of the job will be more useful than data ob-
tained remotely or from boreholes. As far as possible, then, major experiments
and measurements should be deferred until the opportunity to work in the
Figure 7.1 Photos of Churchill Falls underground powerhouse during construction;
Courtesy of D. R. McCreath, Acres Consulting Services, Niagara Falls, Canada. (a)
The machine hall, 297 m long, 25 m wide, and 47 m high; t was excavated at a
d.epth of about 300 m in gneiss. (b) The surge chamber, with draft tube entries on the
nght. The keyhole shape was determined by finite element analysis to -reduc<! the
extent of zones of tensile stress in the rock. The opening is approximately 275 m
long, 19.5 m wide at the maximum section, 12 m wide at the base, and 45 m high.
[See Benson et al. (1971).]
224 Applications of Rock Mechanics in Engineering for Underground Openings
INTERMEDlATE
RESERVOIR
SWITCHYARD
1
TRANSFORMER GALLFIIY-----l
DRAFT TUBE TUNNELS
LOWER PENSTOCK SHAFT
HEAVY HOIST SHAFT
TRANSFORMER GALLERY
VENT ANO CABLE
SHAFT
MAIN ACCESS SHAFT
VENT
Figure 7.2 A scheme for a two-stage underground pumped storage project.
(Reproduced from McCreath and Willett (1973) with permission.)
underground site itself. This may require that certain planning decisions'be
postponed until access is gained. For example, an
constructed at a depth of 800 m in Colombia was fixed 1ll locatlOn but not 1ll
orientation, pending the completion of the access tunnel, which was built in a
separate, early construction contract. Stress measurements, deformability
7.2 Openings in Competent Rock 225
measurements, and other tests were conducted in a gallery driven from the
access tunnel in time to permit their integration into the final design process.
Rock mechanics for underground engineering begins with proper apprecia-
tion of the character of the rock. Rock that is able to bridge across an opening
of 20 m or more without appreciable support could be considered competent. In
such rocks, design can be aided by con si deration s of stress concentrations
around the opening as deduced from the theory 01 elasticity.
When the rock is layered where bending and separation of strata are pos si-
ble, the theory of elastic beams and plates can be invoked. For rock that
presents time-dependent properties, such as rock salt, the theory 01 linear
viscoelasticity provides useful concepts. In weak rocks, stresses around open-
ings may reach the limit according to the criteria of failure, resulting in slow
convergence (squeeze); in such rocks, a solution for stresses and displacements
derived from the theory 01 plasticity provides a useful basis for engineering
work. Injointed rock, only individuallimit equilibrium analysis or studies with
numerical or physical models may be appropriate. The discussions of compe-
tent rock, layered rock, and plastic rock that follow are intended only to pro-
vide simple models for guidance in engineering practice. These models can
always be improved by using powerful numerical techniques, but the engineer
cannot resort to such techniques for every question-he or she has to have
sorne working tools to provide checks on computations, order of magnitud e
predictions, and sensitivity studies through parameter variation. This is the
spirit in which the following theory is presented.
7.2 Openings in Competent Rock
In rock stressed below its elastic limit, that is, below about one-half of the
compressive strength, and in which joints are widely spaced and tightly pre-
compres sed or healed, it is often acceptable to consider an opening as a long
hole of constant cross section in an infinite volume. This is the plane strain
l
equivalent of a hole in a plate, and we can use the solution to the problem of a
circular hole in a biaxially loaded plate of homogeneols, isotropic, continuous,
linearly elastic material-the Kirsch solution. A point located at polar coordi-
nate r, O near an opening with radius a (Figure 7.3) has stresses (Tn (T8, 7 r 8,
given by
(T = PI + P2 (1 _ a
2
) + PI - P2 (1 _ 4a
2
+ 3a
4
)cos 20
r 2 r
2
2 r
2
r
4
(7. la)
(T = PI + P2 (1 + a
2
) _ PI - P2 (1 + 3a.
4
) cos 20
8 2 r
2
2 r
4
. (7.lb)
PI - P2 ( 2a
2
3a
4
)
7 r 8 = - 2 1 + -;:'2 - 0 sin 20
(7.lc)
I The concept of plane strain is discussed in the derivation to Equation 7.1 in Appendix 4.
226 Applications of Rock Mechanics in Engineering for Underground Openings
-
_P,
-
(a)
aolp2
3 K = p,IP2
K = O I
-
~
I\. VK=I"O
~ ...
I
,f ........ K= 2.0
O
,---K = 3.0 2
3
rla
-1
~ O _ 1
2 3 4 5 6 7 8
()= 1.0 ~ L
;-
K=O
I
I I
V
r--2
I
/
I
r--3
rla
(b)
Figure 7.3 Stresses around a circular hole in an isotropic, lin-
early elastic, homogeneous continuum.
7.2 Openings in Competent Rock 227
where U'r is the stress in the direction of changing " and U'o is the stress in the
direction of changing 8.
Substituting the value , = a in Equation 7.1 gives the variation of stresses
on the walls of the opening. The radial and shear stresses are zero since this is a
free surface. The tangential stress U'o varies from a maximum of 3PI - P2 at 8 =
90 to a minimum of 3P2 - PI at 8 = 0 (results used in Chapter 4). Away from
the opening, the stress concentrations faH off quickly, as shown in Figure 7.3b
and Table 7.1.
The Kirsch solution allows calculations of the potential inftuence of joints
in the region of a tunnel. Assuming that a joint of given position and orientation
introduces no change in the stress field, we compare the shear and normal
stresses along its surface with the limiting values of shear stresses consistent
with the criteria of peak shear strength presented in Chapter 5. This exercise
defines a region of joint inftuence, which can be overlain on the actual or
assumed geological section to isolate potential problem areas in the roof or
walls.
Figure 7.4 shows regions of joint slip according to such an approach for
three directions of jointing for an example in which K = 2.33 (P2/PI = 0.43).
The joints were assumed to obey Coulomb's law with ~ j = 31. The contour
values give the number of times the lateral pressure of 1000 psi must be multi-
plied to produce slip on joints of the given orientation in the closed region
inside the contour. The contour values must be multiplied by the shear strength
intercept Sj of the joint and divided by 100. For example, ajoint striking parallel
to the tunnel and dipping at 30, with a friction angle of 31 and a shear strength
intercept of 50 psi, would slip throughout the region enclosed within the con-
tour marked 0.50 if the horizontal stress were equal to 250 psi and the vertical
stress were 108 psi.
An effective method for monitoring an underground opening is to measure
relative displacements of points on the walls, for example, witha precision tape
stretched between pairs ofpoints, orwith rods anchored at different depths in a
borehole (a "multiposition extensometer"). To interpret such data, it is helpful
to know the order of magnitude of displacements assodated with elastic behav-
ior. The displacements can be determined from the Kirsch solution, assuming
conditions of plane strain:
PI + P2 a
2
PI - P2 a
2
( a
2
)
Ur = - + - 4(1 - v) - - cos 28
4G, 4G, ,2
(7.2a)
and
PI - P2 a
2
( a
2
)
Vo = - 4G r 2(1 - 2v) + ,2 sin 28
(7.2b)
<:X:>
"t:l
t::
o:l
-o
11
<:X:>
"&
O
ex::
.5
b <'O
"- ci
b
0\
o
0\
o
200
20 40 60
03 (MPa)
Answers lo Problems 499
(See the Figure.) It can be appreeiated that parameter m is a generalization ofthe
Mohr-Coulomb frietion angle for the case of a downward-eurved failure enve-
lope.
17. (a) Equation 2.17 gives RMR = 9 log Q + 44
Solvng for Q yields
(
RMR - 44)
Q = e 9
Substituting this result for Q in the given expression gives
A = 0.0176e
M
[(RMR-44)!9J
(b) For M = 0.65, A = 0.0176e(0.072RMR-3.177) or A = e(0.0722RMR-7.217). Substituting in
Equation 3.15 withA in place of 1 and with N = 5, qu =.2, and RMR = 50 gives
A = e-
3
.
607
and
Finally
U",p = 0,054 + 6.37(U"3)0.65 (MPa)
CHAPTER4
l. U"u = 13.5 MPa = 1960 psi
(iH = 6,75 MPa = 979 psi
2. Original water pressure = 1300 psi
U"h,min = 9.72 MPa = 1410 psi
T
o
= 3.45 MPa = 500 psi
U"u = 24.7 MPa = 3580 psi
U"h,max = 18.76 MPa = 2720 psi if pore pressure is negleeted. If pore pressure is
eonsidered using the answer to Problem 4.11 with Pw = 13,000 psi: U"h,max = 1,420
psi = 9.79 MPa.
3. U"major = 3.10 MPa = 450 psi
U"minor = l. 38 MPa = 200 psi
O = -63.4
4.278 m
5. U" = U"h,max = 82.33 MPa = 11,938 psi
U"2 = U"h,min = ?
U"3 = U"u = 25.15 MPa = 3646 psi
Figure 4.7 gives K '= 1.91, or U"h,max = 48.0 MPa = 6957 psi.
6. K = 1.11
500 Answers to Problema
7.
c:
o
.;
~
6d
--
--
i
-----------,---
--
c:
'c.
.=
.,
'" c:
ca
.r.
U L-______________ .-
Time
I
I
I
I
I
o
.... ;.----u
t
_
Jack pressure
8. Assuming that the Kirsch solution applies, with a perfectly circular opening in a
homogeneous, isotropic, elastic, continuous mass, with the radius of the tunnel
much larger than the width of the jack; that the value of E for loading equals the
value for unloading; and that the slot and jack have the same dimension, then
CFhoriz = 4.48 MPa = 650 psi
CFvert = 7.24 MPa = 1050 psi
Note that if the stress concentrations are taken at the center of the jack, Equations
4.13 become
giving
Then
{
CFO.w} = (-0.635 2.47 ){CFhOriZ}
CFO,R 2.47 -0.635 CFvert
{
CFhOriZ} = (0.111 0.434){CFO.w}
CFvert 0.434 0.111 CFO.R
CFhoriz = 4.61 MPa = 669 psi
CFvert = 8.16 MPa = 1185 psi
9. In pounds and inches
giving
{
CFx} (182.5 -57.5 )-1{0.003}
CF
Z
= 10
8
2.5 122.5 207.85 0.002
T
xz
2.5 122.5 -207.85 0.001
{
CFx} (544.4 127.8
CF
z
= 10
3
-11.1 405.6
Txz 240.6
{
CFX} {2017.0}
CF
z
= 1183.0 psi
T
xz
241.0
CFmax = 2082 psi
CFmin = 1118 psi
a = 15.0
127.8 ){0.003}
405.6 0.002
-240.6 0.001
Answers to Problema 501
10. CFh = 180 MPa = 26,100 psi
The measurement gives CFh = 80 MPa = 11,600 psi.
The rock could not withstand such a high stress difference (CF
u
= O) and developed
fractures.
11. Pc1 - Pw ~ 3CFh.min - CFh.max - 2pw + To by substituting CFh.max - Pw, CFh.min - Pw, and
Pc1 - Pw In place of CFh.max, CFh.min, and PcI, respectively.
12. (a) CFh.max = 1333 psi
CFh.final = 1333 - 449 = 884 psi
(b) Glaciation, or sedimentation followed by uplift and erosion
13. CFx = 48.25, CFz = 10.15, Txy = -1.13, CFI = 48.28, CF2 = 10.12, a = -1.70
CHAPTER5
1. There are three sets of joints:
(a) Strike S 38.4 E (b) Strike S 34.3 W
dip 36.8 NE dip 62.2 NW
Kf = 557 Kf = 439
(c) Strike N 18S E
dip 63.2 SE
Kf = 238
2. See diagram.
North
West o
South
o o
o
o
o
o
Upward
Normals
to
joints
Upper hemlsphere
East
502 Answers to Problems
3. epp. = 34.5" (the values of G', Ton the sawedjoint during sliding are (0.32, 0.22), (0.97,
0.67), (1.61, 1.11), and (3.23, 2.23.
4.252 MPa
5. 6v
Calculated dilatancy
L ___ MeaSUred dilatancy
Roughness profile
6. G' v = 16.2, G'h = 10.6 MPa
,
7. Pw = 0.50 MPa
- 6.5 kN _ 3 kN/ 2 }
8. Tpeak - 0.5 m2 - 1 m
G' = = 20 kN/m
2
Tpeak = tan 33
5.3 kN k / 2
Tresid = 0.5 m2 = 10.6 N m
T resid = tan 27.9
Shear stiffness at peak:
ks = 1; = 2.50 MPa/m
. mm
The initial shear stiffness is 4.00 MPa/m. The peak dilataney angle is eppeak - epresid =
5.1 (assuming no wear on thejoint). Using Sehneider's equation (5.8), i varies as
U(mm)15.2 7.5 9.5 n.o 2:12
(O) 5.1 3.1 0.91 0.45 o
9. (a) At slip, with angle of friction epj on the joint
T - F B sin a + F B eos a tan epj
(1)
Then, to prevent slip
F = T
B eos a tan epj + sin a
(2)
(b) The force s aeross the joint are inclined at epj with the normal at the point of slip.
10.
T
Answers to Problems 503
The direetions of T and F
B
are given. Prom
the force triangle, F
B
is minimum when
(3)
So
I amin = 90 - epj I (4)
In this direetion, (2) gives
Fb = Teos ep (5)
(e) Pirst, dilation would ehange the strength so that in place of epj one should substi-
tute epj + i in. (2) and (4). Seeond, the dilatant displaeement eomponent in the
direetion of the bolt would eneounter resistanee due to the bolt stiffness k
B
eausing an inerement of bolt force:
!:lF
B
= k
B
(u tan i eos a + u sin a)
In place of (2), we obtain
FB = (ep T.) . + kBu(tan i eos a + sin a)
eos a tan :j + 1 + sm a
(6)
(7)
An additional resistanee would come from bolt shear stiffness when the steel hits
the sides of the borehole. As this induces roek erushing and sharp eurvature in
the steel, with eombined, nonhomogeneous stresses, the solution beeomes more
eomplex. Equation 5 can be simplified to
eos(epj + i) sin(a + 1) (8)
Fb =. A-') + kB U
sm(a + 'Pj + 1 eos 1
G'31G'
11/11 C)
epj = 20 epj = 30
0a
O, 180 0.000 0.000
5, 175 0.086 0.125
10, 170 0.305 0.210
15, 165 0.383 0.268
20, 160
0.434 0.305
25, 155
0.466 0.327
30, 150
0.484 0.333
504
Answers to Problems
0.490
0.484
0.466
0.434
0.383
0.305
0.086
0.000
-0.327
-1.000
-3.063
0.327
0.305
0.268
0.210
0.125
0.000
-0.188
-0.484
-1.000
-2.064
35, 145
40, 140
45, 135
50, 130
55, 125
60, 120
65, 115
70, 110
75, 105
80, 100
85,95
90
-00 -5.330
a ~ s ~ ~ ~ ~ ~ - - - - ~ - ~ o o
::;ee figures for .l.. = 20 'f'J and 30.
f'
l
Ana wers to Problems
11. cp. - 28 2 J - ., t/J = 50 ,Pw = 10 MP a, (j3 = 1.5 M
1.5 _ Pa, (jI = 4.5 MP
Pw _ tan 50 a
12. S - 10 I - ., cp = 30
"Y = 0.025 MPa/m
v = 02 . (see figure)
w tan 78.2 = 0.249
4.5 - P -
l.5-p -1 w - .121 - 0.249p
Pw = 0.379 w
0.751 = 0.505 MPa
505
'"
N
Answers lo Problems 507
qu
Equation 3.14 ~ er.p = 1 _ k tan
2
(45 + cp/2)
qu = 2S tan(45 + cp/2) = 2(1.0)tan 60 = 3.46 MPa
3.46
er - er - = 13.8 MPa
u - l,p - 1 - 0.25 tan2 60
Depth Z = 13.8 MPa/0.025 MPa/m = 554 m
erh = 0.25er
u
= 3.46 MPa
(b) Using Bray's equation, given in Problem 5.10, after the shear joints have
formed,
tan t/J tan 30
erh/
er
u = tan(t/J + cp) = tan 50 = 0.484
erh becomes 0.484 x 13.8 = 6.7 MPa
(e) eru = 1.5 x 13.8 = 20.7 MPa
(d)
erh = 0.484 x 20.8 = 10.0 MPa
Z = 831 m
Ilerh = 0.251ler u
10.0 - erh = 0.25(20.7 - er
u
)
erh = 4.825 + 0.25er
u
er
u erh
15 8.58
10 7.32
6.43 6.43
4 5.83
and erh = er
u
when
er u = 4.825 + 0.25er u
er u = 6.43 MPa
Z = 257 m
(e) The equation connecting erh and er
u
is
erh = 4.825 + 0.25er
u
at rupture, with erh = er and er u = er3
erh = qu + eru tan
2
(45 + cp/2)
giving
4.825 + 0.250"u = 3.46 + 3er
u
508 Answers to Problema
or
and
(Tu tan 30
(O - = -t 50 = 0.484
(Th an
(Tu = 1.365/2.75 = 0.50 MPa
(Th = 4.95 MPa
Z = 20m
Note that I/J = 30
0
is now to be measured from horizontal, whereas previously it
was measured from vertical:
(Tu 0.50
(Th = 0.484 = 0.484 = 1.03
(g) See figure.
Z(m)
13. See figures
(a) In direetion ~ fractures
A=!!
L
In direetion o from normal
Answers to Problema 509
r - - - - - - - - - ~ - - - - - - - - - - n t h fracture
Normal to fracture set
(b) 1..(0) = Adeos 011 + A21eos 0
2
1
= Adeos od + A21sin od
(e) Por max A,
dA
dO = -Al sin 01 + 1..2 eos 0
1
= O
or
tan 0= 1..
2
/1..
1
then
eos O = Al
VAT + A ~
and
. II 1..2
SIn u = , r-;;----o
v AT + A ~
giving
Amax = 'VAr + Ai = 5.39
Omax = 21.8
0
from normal to set 1
Average spaeing in this direction = (Amax)-I = 0.19.
510 Answers to Problems
14. By analogy with Equation 5.8, CPb is the residual friction angle ofthejoint and [JCR
log (JCS/ <Tn)] is the roughness angle of the joint surface. When <Tn is small compared
with JCS, the dilatancy angle is large and when JCS/<T is 10 the roughness angle
equals JCR (degrees). The roughness angle declines with increasing normal stress
until <Tn = JCS after which the roughness angle is zero. (H makes no sense to discuss
the shear strength, using Barton's equation, when the normal stress is higher than
JCS.)
CHAPTER 6
l. Denoting deviator strain components by e and mean strain by e, as in Appendix 2,
(1)
Similarly,
(2)
where <Tx,dev is deviatoric normal stress in the x direction and iT is the mean stress.
Substituting Equation 6.1 in (1) gives
Similar expressions arise by considering ey and ez' AIso, exy = Wxy = l(Txy/G) with
similar expressions for e
yZ
and ezx.
2. Let Slateral = R; then
Baxial
v = ---o;;:-;R;::--<T-"a":;xia'7 I_-----'-P__
p(2R - 1) - <Taxial
and
E = <Taxial - 2vp
Baxial
3. During a triaxial compression test, the mean stress is iT = (<Taxial + 2p)/3. To hold iT
constant as <Taxial is increased, it is sufficient to de crease the confining pressure from
its initial value p = iT to a value p = (3iT - <Taxial)/2, Manual feedback can achieve the
control for slow rates of load; a computer driving a servofeedback system is neces-
sary for fast load rates and for precise control near the peak. In terms of the change
in p required,
(i = l<T axial + 2p = constant; then ~ < T axial + 2 ~ p = O and ~ p = ! ~ < T axial
Answers to Problema 511
4. The data are plotted in the diagram. Using the load cycle from O to 5000 N
,
lateral
Axial
force kN
Axial shortening mm
E = 1821 MPa (elastic deformation)
M = 3188 MPa (permanent deformation)
V = 0.300 (elastic deformation)
v
p
= 0.400 (permanent deformation)
From the total deformation on loading,
Etotal = 1159 MPa and Vtolal = 0.336
E 2E
5. DI = 3K = 3 3(1 _ 2v) and D2 = 2G = 2(1 + v)
(se e Problem 1). Then
and v=
DI - D
2
2D + D
2
512 Answers to Problems
1 1 1
6. (a) -E = M + E
total
(b) See the diagram.
Permanent deformation
1iL:.L--l __ L_...L ___ _ . . . . . . i I ~ ~ __ taxial
7. V = 0.277
M
Secant modulus
of permanent
deformat ion
L-__________ ---:= ______ Eax,al
E = 43,900 MPa (=43.9 GPa) = 6.37 X I06 psi
8. Permanent deformation arises from irreversible c10sure of fissures, pores, and
joints. The second and fiatter slope (D could refiect sliding on joints.
9. (a) Joint normal deformation = ~
Rock normal deformation = i S = i 0.4 m
Equating 1 and 2 gives
E
kn = S = 2.5E
(1)
(2)
(3)
r
'1M.
Ji:
'1";
'\
Similarly,
(b)
Answers to Problems 513
G E 2.5E
ks = S 2(1 + v)S 2(1 + v)
(4)
1 1 1 1 1
En = E + knS = E + E = 2 X 10-
4
(MPa)-1
1 1 1 1 1 4(1 + v)
- = - + - = - + - = = 5.32 X 10-
4
(MPa)-1
~ G U G G E .
Es = E = 10
4
MPa
V
sn
= 0.33
En 0.5 X 10
4
Vns = E v = 104 (0.33) = 0.165
V
SI
= 0.33
Then the .terms of the strain-stress matrix corresponding to (6.9) are
2 -0.33 -0.33 O O O
-0.33 1 -0.33 O O O
10-
4
-0.33 -0.33 1 O O O
O O O 5.32 O O
O O O O 5.32 O
O
O O O O 2.66
10. Letjoint set 1 have spacing SI, etc. Define directions 1,2, and 3 normal to sets 1, 2,
and 3. Then,
etc., for 2 and 3 and
etc., for 23 and 31.
1 1 1
-=-+--
Enl E knlSI
c--
1 1 1 1
-=-+--+--
G
12
G kslSI kS2 S2
Enl
V12 = V13 = E v
E
n2
V21 = V23 = E v
En3
V31 = V32 = E v
514 Answers to Problems
11. We as sume the rock is isotropic and that the strain/stress matrix is symmetric. Since
s, and ( directions are both in the plane parallel to the joint, there is no difference in
deformability constants relating n to s and relating n (o (directions. Thus it suffices
to discuss: (1) Bn due to application of <Tn; (2) Bs due to application of <Ts ; and (3) Bs
due to application of <Tn
(1) Applying only <T
n
causes closure due to the joint, and due to the rock.
Then
_ _ <Tn <Tn _ <T (1 1 )
Bn - S + S - kS + E - E kS + E
<T(E
n
) 1 n
Bn = E knS + 1 = E (p)<Tn finally
(2) Applying only <T
s
causes strain in direction s due only to the rock:
1
Bs = Bs,r = <Ts = E (l)<Ts
(3) Applying only <T
n
causes lateral strain due only to the rock
-v 1
Bs = Bs.r = E <Tn = E (-v)<Tn
12. The slopes of the unloading, reloading ramps give as the bulk modulus kB
k
2.4
B = 0.0007 = 3430 MPa
4.8
kB = 0.0014 = 3430 MPa
k
10.3 34
B = 0.0030 = 30 MPa
Assuming the plastic deformation arises entirely due to unrecoverable (plastic) clos-
ing of the joints, then
= B + B + B . = 3 J!..-
V(plastic) 1
p
l."i< 2
p
l."i< 3
p
l","< knS
where S = the spacing of joints. Substituting S = 5 cm = 0.05 m, and solving yields
13
= vA! + vpE
. v, A! + E
P (MPa)
2.4
4.8
10.4
3 P
k
n
0.05
0.0034
0.0057
0.0088
0.0027
0.0043
0.0058
k
n
(MPa/m)
53,000
67,000
107,000
Answers to Problems
CHAPTER7
1.
Point 8 r <To
A 61 29.15 695
B 30 25.00 748
e -1 29.15 695
2. For 'Y = 1.1 psi/ft,
K O
1
3
<To/<Tv
2.0 1.67
<TO.max(psi) 2200 1837
Location Wall Wall
3. = 1.71 psi = 11.81 kPa
Umax = 0.136 in. = 3.45 mm
<Tmax,s.s = 363 psi = 2.50 MPa
<Tmax,shale = 72.7 psi = 0.50 MPa
<T
r
404
352
404
2.33
2563
Roof
4. (a) Q = 1.0396; R = 34.74a; b = 2.67 x 1Q10
( = 53,390
u = 556.5 in. a, meaning the tunnel fails
Elastic zone:
ex <T
n
-31 481
O 352
+31 481
-
-
-550
-
-
-
2
4.0 9.0
4400 9900
Roof Roof
<Tr = 4000 psi - 2.898 x 10
6
<To = 4000 psi + 2.898 x 106 _1_
(r/a)2
T
ns
-128.5
O
128.5
3
14.0
15,400
Roof
515
516 Answers to Problems
Plastic zone:
6000
5000
.;;
4000 c.
!:lo
Vi 3000
2000
1000
(
r)1.0396
er
r
= 40 -
a
ero = 2.0396er
r
37 41 45 49
---Elastic zone rla
(b) For Pi = 400 psi R = 3.79a, U
r
= -6.35 in. in elastic zone:
err = 4000 psi - 3.453 x 104 _1_
(r/ a)2
ero = 4000 psi + 3.452 x 10
4
in plastic zone: er
r
= 400
ero = 2.0396er
r
5. (a) 3608 lb
(b) 17.1 psi in limestone
113 psi in sandstone
6. PI = 4000 and P2 = 2000 psi
20 min 1 hr
Ur (in.)
1.15 1.15
1 week
1.73
12 hr
1.19
2 weeks
2.28
1 day 2 days 4 days
1.23 1.32 1.48
8 weeks 1 yr 10 yr
4.92 10.69
13.84
Answers to Problems '517
7. Anchor of bolt has coordinates 480", 30.
8.
Head of bolt has coordinates 300", 30.
E = 30 X 10
6
psi; area of bolt = 1.227 in.
2
u in.
!::'u in. !::'u since
Anchor Head Uhead - Uanchor t = 12 hr e
12 hr 0.83 1.23 0.40 O O
24 hr 0.86 1.27 0.41 0.01 5.56 x 10-
5
2 days 0.92 1.36 0.44 0.04 2.22 x 10-
4
4 days 1.04 1.55 0.51 0.11 6.11 x 10-
4
1 week 1.22 1.81 0.59 0.19 1.06 x 10-
3
2 weeks 1.63 2.41 0.78 0.38 2.11 x 10-
3
The bolts should beco me plastic after about 2 weeks.
---
er
psi
O
1,667
6,667
18,333
31,667
63,333
YIELD
Po = P, = P2
After 24 hr
Flb
O
2,045
8,182
22,500
38,360
77,710
For the tunnel in an axisymmetric stress field at r = a in Equation 7.18, we use A =
poa/2, B = e = O, giving
For the dilatometer, with PI internal pressure, Equation 6.33 is equal to
518 Answers to Problems
Thus the displacements are the same (but opposite in magnitude, of course). For a
depth of 1000 ft, with'Y = 150 P/ft3, Po = 1042.
Time
1 min
5 min
15 min
30 min
1 hr
3 hr
6 hr
12 hr
24 hr
36 hr
2 days
3 days
4 days
5 days
6 days
Time Since
Dilatometer
Applied
o
12 hr
24 hr
2 days
3 days
4 days
5 days
0.21
0.22
0.22
0.25
0.28
0.41
0.59
0.85
1.16
1.31
1.39
1.44
1.46
1.46
1.46
u(p)
o
-0.08
-0.11
-0.13
-0.14
-0.14
-0.14
Utotal
0.21
0.22
0.22
0.25
0.28
0.41
0.59
0.85
1.16
1.23
1.28
1.32
1.32
1.32
1.32
9. For RMR = 20, Figure 7.13 indicates that the following relationship exists:
Unsupported Span
(m)
0.8
1.2
1.8
2.3
Stand-up Time
(hr)
2
1
0.5
0.3
10. For a maximum unsupported span of 4 m, Figure 7.13 indicates the following:
Rock Mass Rating
34
39
53
67
Stand-up Time
5 hr
1 day
1 month
2 yr
11. (a) The center of gravity of block 1 lies aboye the edge of block 2 if Xl = SI2. The
center of gravity of blocks 1 and 2 considered as a unit overlies the edge of block
3 if X2 = S/4. Similarly X3 = S/6, X4 = S/8, etc. Thus, at depth n t below the
crown, the tunnel cannot maintain a width greater than w = L ~ I S/i. This is the
harmonic series, which diverges showing that the walls approach but never
become vertical.
12.
Answers to Problems 519
(b) Block 1 tends to fail by rotation about the upper comer of block 2, forming a
hinge at its own upper right comer (a) (see diagram). This causes point (b) to
swing up, sliding along the adjacent block. If there were a horizontal force on
that block, the resulting friction would exert a stabilizing moment on block 1.
Horizontal force on the lower rows of blocks would tend to shear them along the
layer boundaries, reducing the width of the span and increasing the stability of
the roof. This lateral displacement would reduce the horizontal force and the
tunnel wall would then become stabilized with the maximum horizontal stress
allowed by interlayer friction.
Bounds of slip
region for
q, = 20
520 Answers to Problems
13. (a) With x, y parallel to (TI, (T3 direetions and x' parallel to the normal to layers, a is
the angle fromx to x' and (Tx' and Tx'y' are given by Equations A1.2 (with Txy = O).
By definition of epj: ITx'y,1 (Tx, I :s tan epj. This lead to the limiting eondition
eot epi - eot a
(T- - (T
3 - tan a + eot epj 1
Bray's formula, Probo 10 of Chapo 5, eould also be used.
(b) (T3 = 0.31 MPa
(e) The aboye value of (T3 must be supplied by the aetion ofthe supports. This ean be
obtained by roek bolts, for example, ifthe foree in eaeh roek bolt divided by the
area tributary to one bolt = (T3, providing the bolts are closely spaeed.
14. For strata dipping 8 = 45 to the left as shown, interlayer slip oeeurs from () = 0 to
15, from () = 75 to 195 and from () = 225 to 360. a = 180 - () + 8.
Pb
()
a (MPa)
O 225 0.402
15 210 O
60 165 O
90 135 0.938
120 105 0.804
180 45 0.402
15. (a) The original volume, per unit thickness, is (see figure)
u = 7T(b
2
- a
2
) (1)
.c.v = kv
whieh bulks to a final volume
Uf = (1 + kB)u (2)
b The final volume ean also be expressed as
Uf = 7T(b
2
- (a - U
a
)2) (3)
where Ua is the inward displaeement of radius a. Assume U
b
is zero; then
inserting (1) and (2) and equating to (3) yie1ds
(4)
(This result was first published by Labasse, in Reuue Universelle des Mines,
Mareh 1949.)
(5)
Answers lo Problems 521
Sinee U
a
is always less than a, 2a - U
a
> O and kB is always positive (if U
b
= O).
k
- Ui2a - Ua) - Ub(2b - Ub ) (6)
(e) B - b2 _ a2
sinee 2a - U
a
= 2a and 2b - U
b
= 2b
kB =
2(aU
a
- bU
b
)
(7)
b
2
- a
2
16.
Limits of Ring t
(m) r :Ia (days) k
B Sense of Change
2.12-4.5 3.3 1.56 20 -0.0078 Compaeting
100 -0.0168 Compaeting
800 -0.0347 Compaeting
4.5-7 5.75 2.71 20 0.0007 Expanding
100 -0.0025 Compaeting
800 -0.0128 Compaeting
7-9.4 8.20 3.87 20 0.0021 Expanding
100 -0.0003 Compaeting
800 -0.0054 Compaeting
These results are plotted in the answer to Problem 17.
17. (a) The results of Problem 16 are plotted as k
B
versus r yielding the following:
t
(days)
20
100
800
'00
-'00
-200
-300
re
(m)
5.5
8.6
10.6
R = 2.7r
e
(m) Rla
14.8 7.0
23.2 11.0
28.6 13.5
re'" 5.5
re'" 8.6
522 Answers to Problems
p(l + ,,)
(b) Uelas = E '
0.4(1.2)
= 5000 '
Uelas = 9.6 x 10-
4
,
250
200
150
100
"
"
"
"
"
50
"
"
Elastic
2.12 4.5 7.0 9.4 11.7 17.2 25.7
r (m) (Iog scale)
This relationship has been plotted with the extenso meter data in U - log ,
coordinates. The extensometer data intersect the elastic displacement data de-
termining R as follows:
T R
(days) (m) R/a
20 11.7 5.5
100 17.2 8.1
800 25.7 12.1
18. (a) At limiting equilibrium, T = O" tan cf> on the vertical joints while the horizontal
joint opens, becoming free of stress. Summing force s in the vertical direction
gives
w = 'Ybh (per unit of thickness)
B - W + 2hO" tan cf> = O
B = 1 _ 20" tan cf>j
W "lb
Answers to Problems 523
(b) If B = O
b = 20" tan cf>j
max 'Y
19. Associated with shear displacement u, eachjoint tends to dilate by = u tan i. If
the wall rock were rigid, the normal strain increment of the block would be =
and the normal stress increment would be = Thus, in view ofthe
answer to 18(a),
B = 1
W
4E tan i tan cf> !!.
"lb b
20. The displacement path across Figure 5.17b would be inclined a upward from hori-
zontal. For initial normal stress equal to a, it would follow the dashed path shown
(neglecting initial shear stress on the joints) (see figure). The normal stress would
drop slightly; then it would increase almost to b, and then start to drop.
.:lV
--
--
_ _ Path for wedge
.:lU
a -------..... 6
21. Before gravity is "switched on" to the wedge, the stresses tend to flow tangentially
around the circular opening. Close to the periphery of the opening, the direction of
the tangent makes a larger angle (al) with the normal to the joints than it does nearer
the vertex ofthe wedge where the angle is al as shown in the figure. Thus the upper
portion of the block is effectively reinforced by shears along its slides. If the angle al
exceeds the friction angle for the joint, a portion of the block's weight will be
transferred farther up along the joint, setting up vertical tensile stresses in the
wedge. The block may therefore break into two parts, allowing the lower part to fall
while the upper part is restrained.
524 Answers to Problems
22. (a) By symmetry, there is no vertical force at the top center (O). Summing forces
vertically gives
v = ysb
Taking moments about O, when .:ly is very small
or with (1)
and
s
Hb + ysb 2. = Vs
s
H/V = 2b
(1)
(2)
(3)
The lines of action of the resultant of H, v, Wand the horizontal force through O
intersect at a single point; this establishes the location of the reaction.
(b) The block becomes unstable ifpoint O falls lower than H, which can happen for
block s in which b <1ii s. The critical case, is shown in the figure. The block has
rotated by an amount
O = tan-
1
(b/s) (4)
Answers to Problems 525
.:lx = s cos O + b sin O - s = s(cos O - 1) + b sin O (5)
Using (4), .:lx = vS'f+7)2 - s (This result can be seen directly in the figure).
1
23. (a) See figure. S = 1/2 ---? - = 2
S
(b) See figure.
G
o
1/4
-1--
1/2
_L __
526 Answers to Problems
1 = BD + DF + FH
BD = CB cos 45
DF = CE cos 8/2
AB cos 45
cos(45 + 8/2)
GH cos 45
FH = EH cos 45 = --:-'-..,...,..-::'-'-----:-::-:-
sin(45 + 8/2)
a sin 8/2 cos 45
cos(45 + 8/2)
a
= 2" cos 8/2
(a + l)sin 8/2 cos 45
sin(45 + 8/2)
(1)
(2)
(3)
(4)
. . . sin 8/2 + cos 8/2
SIn(45 + 8/2) = SIn 45 cos 8/2 + cos 45 SIn 8/2 = v2 (5)
. . cos 8/2 - sin 8/2
cos(45 + 8/2) = cos 45 cos 8/2 - SIn 45 SIn 812 = v2 (6)
Substituting (2), (3), and (4) with (5) and (6) in (1) and solying for 1 yields
1 ( cos 8/2 ) _ 2a sin 8/2 cos 8/2 a 8/2
sin 8/2 + cos 8/2 - cos
2
8/2 - sin
2
8/2 + 2" cos
Finally,
1 = a(1 + tan 8/2)(! cos 8/2 + tan 8)
Letting bolt spacing be determined by are length, then
s = a8
and
1 1
~ = (j (1 + tan 8/2)(! cos 8/2 + tan 8)
For 8 = 40, = 0.698 radians, as drawn, !:.. = 2.56a.
s
(7)
(8)
(9)
(10)
24. The angle bolts defend against diagonal tensile and shear failure aboye the
haunches. Figure 7.6c shows an opening diagonal tension crack aboye the left
haunch. Its growth releas es a complete block in the roof, which is falling in Figure
7.6d.
25. See figure for parts (a) and (b):
(a) 100
(b) 011
Dip and Dip Direction
30.0 70.0
50.0 140.0
60.0 270.0
90.0 0.0
See figure for part (c)
Horizontal
Answers to Problems 527
111
North wall
101
2
W---------E
Free face of a
block of JP 100
looking North at North wall.
528 Answers to Problems
26. See figure
--------- Horizontal
Projection of JP 101
CHAPTER8
1.
w
;'3
N
D,
o N
3
o
s
/'2
o
D
2
o
L.H.
N,
o
2.
E
Answers to Problems
529
Azimuth of Strike of Cut Steepest Governing
Radius to Cutting at Slope Dip Safe Failure
Point That Point Direction Slope Mode
O E S 90 None
15 S 75 E S 15 W 90 None
30 S 60 E S 30 W 90 None
45 S 45 E S 45 W 87 Sliding on 1
60 S 30 E S 60 W 53 Toppling on 2
75 S 15 E S 75 W 51 Toppling on 2
90 S W 50 Toppling on 2
105 S 15 W N 75W 51 Toppling on 2
120 S 30 W N600W 53 T oppling on 2
135 S 45 W N 4SOW 60 Wedge (1
12
)
150 S 60 W N 30
0
W 50 Wedge (Iu)
165 S 75 W N 15W 42 Wedge (1
12
)
180 W N 38 Wedge (1
12
)
195 N75W N 15 E 37 Wedge (1
12
)
210 N 60
0
W N 30 E 37 Wedge (1
12
)
225 N 45W N 45 E 39 Wedge (1
12
)
240 N 30
0
W N 60 E 44 Wedge (1
12
)
255 N 15W N 75 E 51 Wedge (1
12
)
270 N E 43 Toppling on 1
285 N 15 E S 75 E 41 Toppling on 1
300 N 30 E S 60 E 40 Toppling on 1
315 N 45 E S 45 E 40 Toppling on 1
330 N 60 E S 30 E 42 Toppling on 1
345 N 75 E S 15 E 83 Sliding on 2
The best orientation for a highway cut through a ridge in this rock mas s would be the
one that produces the steepest safe slope on both sides of the highway. A cut
striking east can have a slope of 90 on one side but only 38 on the other, thus it is
not optimum.
Maximum Slopes
from Kinematic
Strike of Cut Analysis
E 90 38
S 75 E 90 37
S 60 E 90 37
S 45 E 87 39
S 30 E 53 44
530 Answers tu Problems
S 15 E 51 51
S 50 43
S 15 W 51 41
S 30 W 53 40
S 45 W 60 40
S 60 W 50 42
S 75 W 42 90
The optimum is the one that minimizes the exeavation and ean be determined
graphically if the topographie profile is drawn.
3. (a) (See diagram.) The minimum bolt foree for a faetor of safety of 1.0 is the mini-
mum foree that when added to 400 ton s vertically will incline the resultant 20
from the vertical. The magnitude of this foree is 137 tons and it is applied in a
direetion rising 20 aboye the horizontal to the S 60 W. For a faetor of safety of
1.5, tan CPreq = tan cpj/1.5 giving CPreq = 21. Therefore the minimum bolt foree
rises 29 aboye horizontal to S 60 W with magnitude 194 tons.
N
w
L.H.
s
Answers tu Problems 531
(b) (See diagram.) From point e, representing the tip ofthe resultant with F = 1.5 a
foree c;D = 112 ton s direeted opposite to N will incline the new resultant 300
from N, and is therefore the foree to initiate slip. The pressure is 112 tonsl200 m2
= 0.56 tons/m
2
.
Water force =
CD = 112 tons
(e) The minimum foree direetion is not the direetion for shortest bolts. The latter is
parallel to N. The optimum direetion depends on the relative eosts of steel and
drill holes and lies somewhere between these two extremes.
4. (See diagram.) The frietion eircle for CPj = 33 lies partly in the upper hemisphere.
The diameter of the eircle is from e to A'. A' is found by first marking A" at the outer
eircle on line eN. A" is 30 from N. Lay off the additional3 along the outer eircle to
loeate point A. A' is at the interseetion. of lines OA and eA" where O is at north. (The
rationale for this eonstruetion is diseussed in Goodman (1976) Methods ofGeologi-
cal Engineering.)
(a) The minimum roek bolt foree Bmin must incline the resultant 42 from W. It is 200
sin 42 = 134 MN, 42 aboye horizontal to the east.
(b) With the bolts in direetion b, as shown in the diagram, the required rotation from
W is 46 (to point D). The angle between Wand b is 80 giving B = 255 MPa.
532 Answers to Problems
(e) The inertia force F = (Kg)m = (Kg)W/g = KW. The angle between the direetion
of F and point D is 80 and the required rotation for the orientation of the
resultant W + B is 20 to point E. The force triangle on the figure determines
F = 135 MN. K = F/W = 0.68. Thus the block slips when the aeeeleration
reaehes 0.68g.
o
w
s
Answers to Problems 533
N
w
534 Answers lo Problems
6. (a) Slides when 8 = cf>j.
(b) Overtums when 8 = tan-I(t/h).
The goveming condition is the one that yields the smaller value of 8. This then
depends on cf>j, and the dimensions of the block.
7. (a) If 8 > cf>j' both blocks slide.
(b) If tl/hl > tan 8 and t2/h2 > tan 8 and 8 < cf>j' the system is stable.
Answers lo Problems 535
(c) If tl/hl > tan 8 and t2/h2 < tan 8 and 8 < cf>j only the lower block rotates. In the
drawing, however, t2/h2 > tl/h
l
so this could not happen.
(d) If 8 < cf>j' t2/ h2 > tan 8 and tI / hl < tan 8, the upper block (1) tends to rotate. The
lower block (2) may tend to slide or to rotate and both conditions must be
checked. In either case, the force P transferred to the lower block is
P = WI(h l sin 8 - tI cos 8)
2h
2
If the lower block slides, the limiting condition is
WI(h
l
sin 8 - tI COS 8) _ W
2
(cos 8 tan cf>j - sin 8)
2h
2
- 1 - tan
2
cf>j
If the lower block rotates, the limiting condition is
W
I
(h
l
sin 8 - tI COS 8) _ W
2
(t2 COS 8 - h
2
sin 8)
2h
2
- 2(h
2
- t2 tan cf>j)
8. According to the theory discussed in most books on strength of materials, buckling
occurs when (TI, the stress parallel to the axis of a column, reaches Euler's critical
stress for buckling (TE:
(compare with discussion aboye Equation 7.5). The free-body diagram defines the
force polygon, which yields
0/ - "(/ sin(b -</Jj)
sin (90 + </Jj)
536 Answers lo Problems
The condition for initiation of failure is met if <T = <TE, giving
7T
2
Et
2
sin(90 + epj)
(a) I
max
= 3L2 Y sin(B - epj)
(b) I
max
= 59.9 m
9. The following values are found by trial, using Equation 8.12 (with (j = O):
Fb
epj (MN)
30 77.94
35 16.54
38 -19.31
36 4.53
37 -7.42
36.4
36.3 0.94
Since Fb = O, epavailable = 36.4.
(b) Repeating a similar procedure with W reduced to 6000 m
3
yields eprequired = 33.3.
Thus the factor of safety is now tan 36.4/tan 33.3 = 1.12.
epj Fh
30 33.53
32 13.28
34 -6.77
33.5 -1.77
33.3 0.23
(c) With W = 10,000 as originally given, and epj = 33.3, the required anchorage
force is 37.1 MN (8.3 x 10
6
lb = 4200 tons). This means approximately 42
anchors in the passive region for each meter of slide width.
10. If ep is fixed at a value corresponding to a defined factor of safety on plane 1,
Equation 8.12 can be solved for <T2 required at the limit of equilibrium and therefore
the factor of safety on plane 2. Thus there is an infinite combination of values ep , ep2
corresponding to limiting equilibrium. An appreciation of the sensitivity of the sys-
tem to changes in either ep or ep2 is gained by plotting the limiting values of tan ep
versus tan ep2.
11 d cot B(F sin B - cos B tan ep)
. cot ex = d(F sin B - cos B tan ep) - U tan ep - V(sin B tan ep + F cos B) + SjA
where d = hH2(1 - ZI H)2 cot B.
12. (a) Summing forces down the dip of P3, at limiting equilibrium gives
W sin B - B cos B - TA - T2A2 - = O (1)
Answers lo Problems 537
h
W = y!lh tan B
h
2
W
A = A2 = ! -- = -
tan B yl
At the limit of equilibrium
and
T3A3 = W cos B tan ep3
Inserting aH the aboye in (1) gives
B cos B = W sin B - 2<Tj ; tan epi - W cos B tan ep3
Dividing by W cos B yields
B <T. tan ep.
- = tan B - tan ep3 - 2 ---1 __ 1
W yl cos B
(b) BIW = O = tan 60 - tan 30 _ 2 <Tj tan 30
25 I cos 60
<T./l = tan 60 - tan 30)(25 cos 60)
J 2 tan 30
<Tj = 12.51 kN/m
I
<Tj Required
(m) (kPa) (psi)
1. 12.5 1.81
5. 62.5 9.06
10. 125. 18.1
20. 250. 36.3
Note that only a small side stress can stabilize a large block.
(2)
(3)
(4)
(5)
(6)
(7)
13. (a) If the abutment rocks are rigid, all dilatant displacement can be expressed as a
tendency toward normal strains, lle = lll/I in the block. Then
III = 2u tan i
and
ll<Tj = E MIL
Combining these equations gives
ll<Ti = 2E tan i uf[
B = tan B - tan CP3 _ (4E tan tan Cf'i) u
W yI2 cos B
538 Answers to Problems
(b) In Problem 12(b), aj = 12.51 with a expressed in kPa and 1 in meters. Here
tlaj = (2 x 10
7
(kPa(2 tan 10 u/l)
or
7.1 X 10
6
u/l = 12.51
u = 1.77 x 1O-
6
f2 (u and 1 in meters)
For equilibrium
1 (m) u (mm)
1 1.77 x 10-
3
2 7.1 X 10-
2
5 0.044
10 0.177
20 0.71
14. (a) W = 100 tons. Let B = the bolt force, determined from the triangle offorces (se e
figure) as follows:
giving
/
/
{
I
I
I
\
\
\
\
I
\
/
/
/
\
\
'\
"-
"-
......
......
,
-- .....
100 tons/sin 75 = B/sin 5
B = 9.02 ton s
r- __
J10 ....... ,
." ......
+b ......
;'
--
--- ----
./
" "-
/
/
/
'\
\
\
/
\2
\
\
\
\
\
.1
1
I
J Reference
J circle
/
/
/
/
Answers to Problems 539
(b) 5.6, or approximately 6, as shown on the stereographic projection in the figure.
This is an upper focal point ("lower hemisphere") projection.
w= loor
80
15. See the figure, constructed with a lower focal point ("upper hemisphere") projec-
tion.
o
I
/
/
Dip vector 3 - - - - - - -3--
\
2
Dip vector 2
540 Answers to Problems
16. See the figure (in two parts):
W = 50 ton s = 50,000 kg = 0.49 MN
N
s
The required water force is U determined from the triangle offorces as follows:
50/sin 6.3 = U/sin 23.7
giving
U = 183.2 tons = 1.795 MN
The water pressure (averaged over the face of plane 1) is
P
w
= U/7.5 m
2
= 0.239 MPa = 34.7 psi
17. (a) See the figure for the JP anaIysis. The only JP lacking any area inside the dashed
circle is 011. This is therefore the only JP defining removable blocks.
Answers to Problems 541
(b) See the figure for the stability analysis. For a factor of safety of2.0 on each plane
tan CPreq'd = tan 35/2.0
giving
CPreq'd = 19.3
~ r ~ m the stereographic projection, the required rotation of R from W is 13.30,
glvmg B = 90 tan 13.3 = 21.3 tons.
-
FS" 1.0
---
-.....
........
"-
FS " 2.0
CHAPTER 9
1. For the left circ1e in the diagram,
Ph = 2Sp tan (45 + %'-) = qu
Then Equation 3.8 gives
q = qu tan
2
( 45 + ir) + 2S
r
tan (45 + ir)
542 Answers to Problems
2. In addition to the forces considered in regard to Equation 8.2, we now have an
additional vertical force P sin f3 and an additional horizontal force P cos f3. Thus the
result for a force P bearing on the slide block follows from the following substitu-
tions in (8.2): in place of W, input W + P sin f3; in place of V, input V + P cos f3.
3. The block slides if the resultant of P and W is inclined cf>j with the normal to the
planeo
For sliding, from the law of sines applied to the force triangle we have
W P
sin[90 - (f3 + cf>j - a)] sin(cf>j - a)
(1)
or for limiting equilibrium in sliding
sin(cf> - a) W
Pslide = cos(f3 + cf>j - a)
(2)
Far case a: The block overturns when the resultant of P and W is oriented 8 = tan-
I
b/ h with the normal to the block. Therefore, for toppling, replace cf>j in (2) by 8 giving
P sin(8 - a) W (3)
topple = cos(f3 + 8 - a)
Far case b: The block topples when P = one-half of the value given by (3).
4. Following closely the derivation in the appendix, with an added force P gives in
place of Equation 9
W
1
sin(8
1
- cf>1)COS cf>3 P cos(f3 - 81 + cf>1)COS cf>3
N
3
= cos(8
1
- cf>l - cf>3) + cos(8
1
- cf>1 - cf>3)
(9a)
Equating this to (15) and solving for Fh gives as a final result
[W
1
sin(8
1
- cf>1) + P cos(f3 - 81 + cf>1)]cos(82 - cf>2 - cf>3) .
+ W
2
sin(8
2
- cf>2)cos(8
1
- cf>1 - cf>3)
Answers to Problems 543
5. For v, = Ve = 0.26 and embedment depths of a, 2a, 3a, and 4a, the results of
Osterberg and Gill for Ee/ E, = i are fitted closely by Equation 9.10 with JL = tan 59
as shown in the following table (Pend = O'y): '
y/a
1
2
3
4
Osterberg and Gill
0.44
0.16
0.08
0.03
Equation 9.10
0.44
0.19
0.09
0.04
However, 40 mm of settlement will most probably rupture the bond, reducing JL
to a value less than unity. Initially, the Osterberg and Gill results should apply.
Subsequently, the load transfer will shift to the distribution given by (9.10) with
JL<1.
6. There is no rational procedure, short offull-scale load tests, that can determine the
allowable bearing pressure. However, it can be estimated in several ways. First,
sin ce a 2-m diameter is about seven times the joint spacing, the qu given by the
small-Iab tests should be reduced by a scale factor of, say, 5. Moreover, a factor of
safety is warranted. Assume cf>p = 20, and (qu)field = 18 MPa/5 = 3.6 MPa. With a
factor of safety F equal to 3, the allowable bearing pressure for a surface footing is
(
tan
2
55) + 1)
qallow,footing = 3 3.6 = 3.65 MPa
For a pier at depth, this value may be multiplied by 2, giving
qallow,pier = 7.3 MPa
The bond strength is not governed by the same size-scaling factor as the end bearing
since the shear is confined to the contact. Assume Tbond = 0.05qu with qu = one-half
the lab value. Then Tbond = 0.45 MPa. With a factor of safety of 2, this gives
Tallow = 0.22 MPa
7. Using radius a = 1 m, lmax = 31.8 m. With the Osterberg and Gill results, any length
approaching this would transfer all load to the sides and we will require 1 = lmax'
Such a long pier (a "pile" in fact) is not the most economical solution. For a = 2 m,
lmax = 15.9 m; then with I1 = 6 as a trial, Pend/Ptotal = 0.07 giving Pend = T = 0.27 MPa,
which is much too small for the former and much too large for the latter.
As an alternative, if the bond is broken or prevented, the load transfer will
occur at a lesser rateo Using Equation 9. lOa with JL = tan 40, and ignoring shear
stress in the concrete since it is confined, the required lengths to satisfy the bearing
capacity of 2 MPa are shown in the following tableo The compressive stress in the
pier aboye the socket is also given.
544 Answers lo Problems
Socket
a 1 Volume
Pend
T
(I max,concrete
(m) (m) (m
3
)
Pendl Ptotal
(MPa) (MPa) (MPa)
1.0 8.97 28.2 0.31 2.00 0.24 6.37
0.9 9.54 24.3 0.255 2.00 0.28 7.86
0.8 9.94 20.0 0.201 2.00 0.32 9.95
0.7 10.15 15.6 0.154 2.00 0.38 12.99
0.6 10.13 11.5 0.113 2.00 0.46 17.68
0.5 9.85 7.74 0.079 2.00 0.60 25.46
0.4 9.27 4.66 0.050 2.00 0.82 39.79
0.3 8.63 2.95 0.024 2.00 1.20 70.74
If the concrete has compressive strength equal to 20 MPa, and it is desired to
achieve a factor of safety of 2, the solution with minimum volume socket is a pier
with radius 0.8 m and length 10 m. These results depend markedly on the choice for
/J- and Eel E, and, to a lesser extent, on v, and Ve'
Another solution is to use a pier seated without a socket on the surface of the
rock, or, if that surface is weathered or inc1ined, seated inside a socket of enlarged
diameter. The required pier radius is 1.78 m. The most economical choice between
the altematives depends on the volume of the pier passing through the soil.
8. Consider the sandstone roof as a continuous c1amped beam. The most critical condi-
tion is tensile stress at the ends on the upper surface. Using (7.5) with (Th = O, Y =
150 Ib/ft
3
, and T
o
= 2MPa gives L = 334 ft = 100 m. This is increased if (Th # O.
However, a beam 200 ft thick with L = 334 ft is too thick for thin-beam theory. (A
finite element analysis would be useful in a particular case.)
2h
9. H = B - 1
10. (a) Summing forces in the y (vertical) direction acting on the differential element
gives
S2 d(Tv + 4T S dy = O (1)
Substituting in (1)
T = (Th tan cf> (2)
and
(Th = k (Tu (3)
yields
d(Tu 4
-= --ktancf>dy
(Tu s
(4)
Solving gives
(Tu = A e-4k tan q,yls
(5)
when y = O, (Tu = q, giving A = q.
Answers lo Problems 545
The support pressure is the value of (Tv when y = t, which gives
I Pb = q e-4ktanq,lls
(b) k = cot
2
(45 + cf>/2) = 0.406
tls = 0.67, q = 21 kPa
Then (6) gives
Pb = 21 e-
O
.
50
? = 12.65 kPa (= 1.83 psi)
If s = l.5 m, the force per support is T = S2 (12.65) = 28.5 kN (=6400 lb).
11. (a) With self-weight, the free-body equilibrium gives
S2 d(Tv + 41's dy = ys2 dy
Substituting as in the answer to Problem lO(a) gives
d - ( 4k tan cf> )d
(Tv - Y - S (Tv Y
Let
4k tan cf>
z = y - S (Tv
then
dz = - 4k tan cf> d(T
s v
and (2) becomes
-s dz
-:-4k"--ta-n-cf>"" Z dy
whose solution is
z = Ae-4k tan q, yls
Finally, resubstituting (3) in (6) gives
Since (T u(y = O) = q
4k tan cf>
y - (T = Ae-4k tan q, yls
S u
A = y _ 4k tan cf> q
s
Finally, P
b
= (T u(Y = t), giving
P
b
= s (y _ (y _ 4k tan cf> q)e-4k tan q,Yls)
4ktancf> s
Simplifying yields
sy
'Pb = 4k tan cf> (1 - e-
4k
tan q, lis) + q e-4k tan q, lis
(6)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
546 Answers to Problems
b) P - (1.5)(27) _ -0.507 -0.507
( b - (4)(0.406)(0.47) (1 e ) + 21 (e )
= 33.75 kPa (=4.89 psi)
(e) If roek bolts were being installed, then an additional force would need to be
added to the equilibrium equation to aeeount for the aetion of the anchor end of
the bolt. (This is diseussed by Lang, Bisehoff, and Wagner (1979).)
APPENDIX 1
l. (a) l. U"x' = 27.7
2. U"x' = 20.0
3. U"x' = 30.0
4. U"x' = 50.0
(b) l. U"x' = 52.7
2. U"x' = 52.7
3. U"x' = 72.7
4. U"x' = 108.3
(a)
y
x
(1)
(b) y
t-------.x
Tx'y' = -18.7
Tx'y' = -10.0
Tx'y' = 20.0
Tx'y' = -20.0
Tx'y' = -7.3
Tx'y' = 7.3
Tx'y' = 27.3
Tx'y' = 0.0
y
(2)
y
y
+-
y
+-
x x x
(3) (4)
y
Io<-------x """----.............. x
(4)
Answers to Problems 547
2. (a) T
~ ~ - 7 ~ - - ~ - - - - - + - - - - ~ - - - - - L - - - - ~ + - - - ~ __ v
(b) T
~ ____________ ~ L - ____ - ~ ~ ~ ~ ~ _______________ v
3. For Problem la:
For Problem lb:
a = -31.7,58.3
U"I = 62.4, U"2 = 17.6
a = 22.5, 112S
U"I = 108.3
U"2 = 51.7
54S Answers lo Problems Answers lo Problems
549
- ,;n 20) {:;}
7.
4. u y' = (sin
2
a cos
2
a
i
T
xy
5. Ux' + u y' = (u
x
cos
2
a + u y sin
2
a)
~
Up Up
70
0
+ (Ux sin
2
a + u
y
cos
2
a)
30" 60"
':1
= Ux + u y
'T'
z 70
0
y'
6. (a) a = 67.50
Up
UI = 108.28 z 20
0
x
U2 = 51.72 (a) (b)
(b) a = -67.50
%
UI = 108.28
Up
U2 = 51.72
'
..-Ji
90
0
T
Z' x' z i Up90
(e) (d)
6(a)
a
Line Bearing {3 a m n
(a) x' East O 60 0.50 0.00 0.866
y' West 180 30 -0.866 0.00 0.50
z' South -90 O 0.00 -1.00 0.00
(b) x' S 20 W -110 20 -0.321 -0.883 0.342
y' N 20 E 70 70 0.117 0.321 0.940
z' N 70
0
W 160 O -0.940 0.342 0.00
T
(e) x' S 45 E -45 O 0.707 -0.707 0.00
y' N 45W 135 90 0.00 0.00 1.00
z' S 45 W -135 O -0.707 -0.707 0.00
6(b)
(d) x' 90 0.00 0.00 1.00
y' North 90 O 0.00 1.00 0.00
o a
z' West 180 O -1.00 0.00 0.00
8. Ux' Tx'y' Tx'z' ITx'maxl
(a) 593.30 234.81 -25.00 236.13
(b) 265.55 141.03 -10.28 141.40
(e) 100.00 35.36 50.00 61.24
a, (d) 700.00 O -50.00 50.00
10.
North
y;y
z' ----------jl------_X
--
POS, TU
-
Pos. T.,,'
t zx = 50 (given); acts to left
t' = - 50 (calculated); acts to left
11.1, = 1000F/L2
h = 225 X 10
3
(F/Ll)2
h = 11.75 X 10
6
(F/V)3
APPENDIX.2
1. Equations A1.2 together with the answer to Problem 4 in Appendix 1 for stress
transformation in two dimensions are
{
:x:} = ( cOs::; a : _:::
Tx :' -! sin 2a ! sin 2a cos 2a Txy
Replacing T by "1/2 and (T by B, with appropriate subscripts, yields:
{
::: } = : _! }
Yx'y' -sin 2a sin 2a cos 2a Yxy
552 Answers to Problems
2. For a gage with O!A = O, O!B = 60, O!c = 90, the coefficient matrix in (A2.3) becomes
The inyerse of the aboye is
3. (exey'rxy) =
(+ O ~ 5 oH
O
O
2.309
(a) (1.0 x 10-
3
, O, 5.774 x 10-
4
) ,
(b) (l.0 x 10-
2
,3.0 X 10-
2
, -1.155 x 10-
2
)
(e) (2.0 x 10-
4
, 5.33 X 10-
4
, -1.61 x 10-
4
)
4. (a) O! = 15.0, el = 1.077 x 10-3, e2 = -7.736 x 10-
5
(b) O! = -75.0, el = 3.155 x 10-
2
, e2 = 8.452 x 10-
3
(e) O! = -78.29, el = 5.344 x 10-4, e2 = 1.856 X 10-
4
APPENDIX5
1. x = tan(45 - (12).
2. The angle between (1) and (2) is 59. Their common plane strikes N 64 E and dips 78
N 26 W.
3. 1
12
plunges 37 to the S 84 E.
4. The answer is giyen in Figure A5.1. The line from O to -P is directed into the upper
hemisphere. It plots outside of the horizontal circle at position -p as shown.
5. Let the position of line OQ in a lower hemisphere projection be point q. Then the
position of the opposite to OQ when plotted in an upper hemisphere projection is
obtained by rotating the tracing 180. What was north must be re1abeled as south.
6. 15.
210 00
EQUATORIAL CON FORMAL STEREONET
Aastrup, A., 419
Adhesion, see Bond strength
Age of a rock:
effect on porosity, 29
terminology, 424
Alexander, L. G., 122
Alto Rabagao dam, 369
Alvarez, L. G., 122
Anchor piles, 382
Anhydrite:
specific gravity, 31
Mohr-Coulomb parameters, 83
Anisotropy, 13
bulbs of pressure, effect on, 355-361
in elasticity, 182, 183
in rock fabric or structure, 420-423
in strength, 93-95
in tunnel support needs, 274-280
Archambault, G., 166-171
Barr, M. V., 378
Barton, N., 42, 166, 177
Basalt,20
compressive and tensile strength, 61
conductivity, 35
density, 33
Hoek-Brown parameters, 99
modulus ratio and Poisson's ratio, 186
Mohr-Coulomb parameters, 83
point load strength, 37
porosity, 83
sound velocity, ideal, 41
Beams in mine roof, 233-237, 289
Bearing capacity:
analysis, 361-364
codes, 348-350
Beatty, R. A., 132
Bedding, 13, 144
Benson, R. P., 223
Bemaix, J., 34
Bemede, J., 60, 116
Bibliographies and indexes, 15
Bieniawski, Z. T., 75
geomechanics classification, 43-46
I n d e ~
modulus in-situ, 198
size effect, tests, 90, 91, 231, 233
standards for compression tests, 60
stand-up time in tunnels, 250
Bischoff, J. A., 546
Bishnoi, B. W., 364
Block sliding (plane sliding), 295, 296
kinematic analysis, 301-302
stability ana1ysis, 310-319
Blocks:
criteria for a key block, 258, 259
numbers of joint faces, 322
removability, 259, 260
types, 259
Block theory:
introduction, 257-262
for rock slopes, 320-325
for tunnels, 270-274
for underground chambers, 262-270
Bond strength, concrete to rock, 372, 374,
376
Borehole breakouts, 114, 115
Borehole deepening method, 116
Borehole jack test, 190, 191
Boyle, W., 258
Brace, W. F., 35, 69, 71
Bray, J. W.:
bulbs of pressure in anisotropic rock,
358-361
Mohr's circ1e construction, 163, 392
plastic zone around tunnels, 243-250,
454-464
slip on joints, 176, 455
toppling, 296
Brazilian test, 60, 65
Broch, l., 37
Brown, E. T., 99, 108, 111,378
Buckling:
of rock slopes, 337
of roof beams, 233
Building codes, 349, 350
Bulb of pressure, 355-361
Bulking of rock, 53, 287
Bulk modulus, 68, 182,441
555
556 I n d e ~
Burgers body, 205-214,445
Byerlee, Jo, 164
Canelles arch dam, 365-367
Chalk:
bearing capacity, 350
ductility, 83
Mohr-Coulomb parameters, 83
porosity, 83
Chandra, Ro, 38
Christian, Jo To, 354
Churchill Falls underground powerhouse, 223
Classification of rocks, see Rock classification
Cleavage, effect on strength, 83
Coal:
density, 33
point load strength, 37
size effect in strength, 90, 91
Coefficient of loosening, 53
Cohesion (shear strength intercept), 81
Compaction us consolidation, 28
Complete stress strain curve:
definition, 76
relation to creep, 79
relation to dynamic strength, 80
Compressibility, 182
Compression tests, standards, 60
Compressive strength, unconfined:
and point load index, 36
and shear strength, 82
and tensile strength, 61
test procedure, 60, 184
typical values for rocks, 61
Conductivity uSo permeability, 34
Conway, Ho Do, 358
Cook, No Go Wo, 104, 132,230,409
Cording, Eo Jo, 242
Core discing, 116
Coulomb's law, 166, 239
Creep, 79, 202-217
rupture (tertiary creep), 79, 213
in tunnels, 250-256
Criteria of failure:
Barton (for joints), 177
empirical, 87-90
Griffith, 87
Hoek-Brown, 99
Mohr-Coulomb, 80-84
Cross bedding, 143
Crushing failure, 57
Cuttability, 4-8, 57
Cyclic loading, 78
Cylcap, 157
Darcy, unit of permeability, 34
Darcy's law, 34
David, Eo Ho, 352, 370
Dean, Ao Ho, 132
Deep foundations, 370-378
Deere, Do 00,43, 184,242
Deformability, 179
of joints, 163, 164, 170, 220
reasons to know, 179
De la Cruz, 116, 117
Delayed elasticity, 208
Denis, Ao, 65
Density, 30
usefulness of knowing, 32
values, 33
Desai, Co So, 354
Deviatoric and nondeviatoric stresses, 66, 67
strains, 412
Diabase, 22
compressive and tensile strength, 61
fracture in compression, 70
modulus values and Poisson's ratio, 186
porosity, 29
Dilatancy:
and bearing capacity, 350-352
calculated for joints, 168
of joints, 163, 164
of rock, 70, 72
path effects on strength, 288, 338, 339
Dilatometer tests, 190, 191, 211
Direction cosines and strike and dip, 400-402
Direct shear test, 157-161
Direct tension test, 65, 88
Discontinuities:
classes, 13
problems caused by, 150
single and multiple, 147, 148, 150
spacing and aperture, 34, 147
Dolomite:
bearing capacity, 349, 350
compressive and tensile strength, 61
modulus ratio and Poisson's ratio, 186
Mohr-Coulomb parameters, 83
permeability, 32
point load strength, 37
porosity, 29, 83
sound velocity, 40, 41
specific gravity, 31
Donath, Fo Ao, 93
Doorstopper method, 127-129
Drillability, 4-7
Drilled shaft foundations, 346
load distribution in, 370-378
Dube, 287
Ductile behavior, 74
Durability, 37-39
Duvall, Wo, 116, 130,230,231,236
Dynamic measurements, 193-195
in fractured rocks, 199-202
Earthquake forces, slope stability, 316, 318,
319
Effective stress, 84
and joints, 171-173
Terzaghi's law of, 84
Eigenbrod, K. Do, 38, 39
Elastic behavior, 179, 180
Elastic constants, 181-183
Emery, Co L., 116
Empirical criteria of failure, 87-90, 177
Energy development, 6, 7, 12
Equilibrium equation, 456
Eristov, Vo So, 213
Erlikhman, So, 359, 361
Evaporite rocks, see Gypsum; Salt
Excavation Pyramid (EP), 260
for a rock slope, 323
for the roof of a chamber, 264
for the wall of a chamber, 266
Factor of safety, 312, 318, 338
Fairhurst, Co, 117
Faults, 141
gouge, swelling of, 380
relation to principal stresses, 106-115
Fecker, Eo, 167, 169
Finite element analysis, 352-354
Fisher, Ro Ao, 153
Fissures, 13, 14
aperture of, 36
effect on permeability, 35
effect on sonic velocity, 39-42
in thin sections, 24-26
Flatjacks:
for compression tests in situ, 91, 92
for deformability tests, 192, 193
for plate bearing tests, 189
for stress measurement, 115, 121-123
Flexural cracking, 55
Flexural test, 65, 66
Folds, infiuence on initial stresses, 105
Foose, Ro, 378
Footings, modes of failure, 351
Footwall failure, 337
Foundations on rock:
problems, 343
stresses in, 352-361
types, 343-348
Fourmaintraux, Do, 40
Fracture frequency, 177
Franklin, Jo Ao, 38
I n d e ~ 557
Frequency of stress waves, 199-200
Friction:
and bulb of pressure in layered rock, 357
internal,83
and load transfer in piers, 374
minerals with low values, 418
and mine roof stability, 237
Friction circle, 312-314
Friedman, Mo, 117
Gabbro, 20, 27o See also Norite
density, 33
sound velocity, ideal, 41
Gallery test, 190, 191
Gamble, Jo, 38, 379
Gardner, Wo So, 370, 377
Gaziev, E., 359, 361
Geological hazards, 4-7
Geological time scale, 424
Geomechanics classification, 43-47
and deformability, 198
and stand-up time in tunnels, 250
Gibala, Ro, 85
Gill, So Ao, 370-372
Gneiss,20
bearing capacity, 350
compressive and tensile strength, 61
modulus ratio and Poisson's ratio, 186
Mohr-Coulomb parameters, 83
porosity, 83
Goodier, Jo No, 189, 468
Goodman jack, see Borehole jack test
Gouge, swelling pressure, 380
Granite, strength, 61
in dam foundation, 369
Hoek-Brown constants, 99
Lundborg's strength parameters, 90
modulus ratio and Poisson's ratio, 186
Mohr-Coulomb parameters, 83
permeability, 33
porosity, 29, 83
size effect in weathered quartz diorite, 91
558 I n d e ~
Granite, strength (Continued)
sound velocity in, ideal, 41
weathering profile, 342
Gray, R. E., 379
Gray, W. M., 126, 135
Greer, D. M., 370, 377
Greywacke, see Sandstone
Griffith theory, 82, 87
Grout column foundations, 347, 348
Gypsum, 15, 20
density, 33
porosity, 31
sonic velocity, 40
Haimson, B. e., 80, 115, 117, 120
Hamrol, A., 30
Hardness, scratch, 416-418
Hardy, R. H., Jr., 209, 210, 216
Hast, N., 116
Hayashi, M., 117
Healy, J. H., 172
Hemispherical normal distribution, 153
Heuze, F. E., 190
Hinds, D. V., 378
Hocking, G., 378
Hoek, E., 99,108,111,296,475
Hoek-Brown criterion, 99
Hogan, T. K., 132
Hooker, V. E., 116, 130
Hustrulid, W. A., 233
Hydraulic fracturing, 3-7, 12, 117-121
depth required for a vertical fracture, 121
Hydrostatic compression, 68
Hysteresis, 80, 181
Identification of rocks, 419-424
Impression packer, 119, 378
Inclusions for stress measurement, 116,
187
Index properties, 27
Ingraffea, A., 361, 378
Initial stress:
directions of, 112-115
and faults, 106-115
and folds, 103
horizontal, 106-115
horizontal values, ratio of, 120
sedimentation values, ratio of, 101-104
reasons to know, 101-104
and topography, 104
vertical, 104
Invariants of stress, 403
Jaeger, J. e., 93, 104, 132, 166,230,409
Jahns, H., 90, 91
Jaspar, J. L., 380
Jes and JeR, 177
Jethwa, 286, 287
John, K. W., 318
John, M., 213
Joint orientation:
analysis, 150-154
bias in observations, 154, 155
and compressive strength, 93
distribution, 153
effect on works, 46
Joints, 141, 145, 146
Barton's equation, 177
Bray's equation, 176,455
condition, 45
deformability, normal and shear, 163, 164
dilatancy, 163
maximum closure, 170, 171
models, 157
problems caused by, 150
roughness, 164-169
strength tests, 156-164
trace map, 265, 267-269
JP (joint pyramid), 260, 263
Kaiser effect, 117
Kanagawa, T., 117
Kelly, B. l., 367
Kelvin solid, 204-206
creep in, 446-448
Keyblocks, 257
Kinematic analysis of models, 259, 331-333
Kinematic analysis of rock slopes, 301-310
King, M. S., 199
Kirsch solution:
circular hole in elastic medium, stresses
around, 119, 133, 225
elastic compared with plastic medium, 246
Kovari, K., 61
Kubetsky, V. L., 213
Kulhawy, F., 62, 81, 361, 378
Ladanyi, B., 166, 171,350,362,371,372,375
Lame's constant, 181, 182, 444
Lang, T. A., 546
Laubscher, D. H., 46
Leeman, E. R., 116, 127, 130, 133, 134,437
Lien, R., 42
Limestone:
bearing capacity, 350
compressive and tensile strength, 61
creep, 208-210, 216
in dam foundations, 367-369
density, 33
ductility, 74
karst, 341, 342
modulus ratio and Poisson's ratio, 186
Mohr-eoulomb parameters, 83
permeability, 35
point load strength, 37
porosity, 29, 83
Liquid limit, 38
Londe, P., 318
Lugeon test, 52
Lundborg, N., 66, 88-90
Lunde, J., 42
Lynes impression packer, 119
Mcereath, D. R., 223, 224
McLamore, R. T., 93-95
Mahar, J. W., 242
Malpasset dam, 13, 14, 34, 147
Manhattan schist, 8
Map of initial stress, 112
Marble:
complete stress-strain curve, 78
compressive and tensile strength, 61
modulus ratio and Poisson's ratio, 186
Mohr-eoulomb parameters, 83
porosity, 83
Markwell, S. D., 367
Maximum removable blocks of a tunnel,
271-274
Maxwell body, 204, 205
creep in, 447, 448
Meehan, R.L., 382
Menel, V., 132
Merrill, R.H., 116
Mesri, G., 85
Minerals:
identification, 416-418
sound velocity in, 40
specific gravity, 31
Mining:
open pits, design of, 306-310
room and pillar design, 231-237
stable and unstable methods, 9
subsidence over old mines, 379
Model of joints, 157
Modulus of elasticity, see Young's modulus
Modulus in situ, 198, 201
Modulus of rupture, 65
Modulus ratio, 186
Mohr-eoulomb criterion, 80-84
Mohr's circle, 392
I n d e ~ 559
for bearing capacity, 362, 363
for direct shear test, 159, 161
for strength criteria, 81, 82, 86
for triaxial test with joint, 160-163
for water pressure effects, 86, 172
Morgenstem, N. R., 38, 39
Mount Rushmore, 2
Mudstone, porosity, 29
point load strength, 37
Mller-Salzburg, L., 50
Multistage triaxial test, 162, 164
Muskhelishvili, N. l., 230
Mylonite, 20, 23
Nakamura, K., 113
Nichols, T. e., Jr., 116
Nakasa, H., 117
NGI rock classification, 47-49
Nolting, R. E., 116
Norite, see a/so Gabbro
creep and dynamic effects in, 214
in triaxial compression, 75
Normal stiffness, 196
Nuclear power, 6, 7
Obert, L., 117,230,231
Octahedral stresses, 404-406
Oil shale, density, 33
Olsen, O., 63
Open pit mines, kinematic analysis, 306-310
Orientation optimum for a tunnel, 274-280
Osterberg, J. O., 370, 372
Overcoring, 123-127
Panek, L. A., 135
Patoka dam, 367-369
Patton, F., 164
Patton's law, 164-166
Peak stress, meaning of, 76
Pegmatite, Lundborg's strength
parameters, 90
Penstocks, 9
Permanent deformation modulus, 187
Perrnanent deformation Poisson's Ratio, 220
Permeability, 32-36
Permeability us. conductivity, 34
Peterson, J. R., 116
Phenocrysts, 26, 91
Phillips, F. e., 475
560 I n d e ~
Photoelastic stress gage, 116
Piles and piers, 346
load distribution in, 370-374
load tests, 372, 373
settlement of, 376, 377
Pillars, 231
Plane sliding, see Block sliding
Plane stress and plane strain, 452
Plastic analysis of tunnels, 242-249, 454-464
Plate bearing test, 188-190
Point load index, 36
Poisson's ratio, 67, 181
measurement, 184
for permanent and total strain, 220
typical values, 186
variation with deviatoric compression,
70, 185
Pore "crush-up", 69
Pore pressure:
and fault slip, 171-173
and rock failure, 85, 86
and stress measurement, 140
Porosity:
effect of age on rocks, 29
measurement, 31
types of pore spaces, 13, 28, 29
values for selected rocks, 29, 83
and water content, 32
Poulos, H. G., 352, 370
Pratt, H. R., 91, 92
Principal strains, 412
Principal stresses, 133, 134, 393,402,403
Principal stress ratio:
effect on strength, 74-76, 87
initial values, 106-113
Principal symmetry directions, 182
Punching failure of footing, 351, 352
, Q system (classification), 47-49
Quartzite, 142, 144
compressive and tensile strength, 61
Hoek-Brown constants, 99
Lundborg's strength parameters, 90
modulus ratio and Poisson's ratio, 186
Mohr-Coulomb parameters, 83
Radial jacking test, 191
Radial permeability test, 34, 35
Raleigh, C. B., 172
Rangeley oil field, 172
Raphael, J., 363
Rate of load effects, 213-215
Reichmuth, D. R., 62
Relaxation, 203
Removable blocks, 258
Rengers, N., 166, 167, 169
Reservoir induced seismicity, 86, 171-172
Residual strength, 81, 161
Residual stresses, 116
Rhyolite, 20
density, 33
Riley, P. B., 116
Ring shear test, 66, 88-90
Roberts, A., 116
Rocha, M., 36, 116, 121, 134
Rock boIt design:
to control plastic zone of tunnels, 246
to prevent slip of joints in tunnels, 241
for rock slope stability, 279, 283, 317-319
to secure roof beams in layered rock, 236,
237, 281, 290, 334, 387
Rock bursts, 102
Rock classification:
comparison of Q and RMR, 47
fabrics, 20, 21
fissuring, 39-42
geological, 19, 419-425
NGI (Q) system, 47
RMR (Geomechanics) system, 43-46
for tunneling, 43-47
Rock loosening, 53, 247
Rock mas s rating, see Geomechanics
classification
Rock noise, 69
Rock reinforcement, 249. See also Rock bolt
design
Rock slopes:
buckling of, 337
types, 293, 294
Rock socket, 345, 371, 377
Rock test preparation equipment, 58, 59
Roof beams, buckling of, 233
Roof design, underground, 233-237
Rosettes, strain gages, 411, 412
RQD,44 '
Rummel, F., 63
Rutter, E. H., 206
Safety, see Factor of safety
Salem, A., 190
Sallstrom, S., 419
Salt, 15
creep properties, 252
density, 252
ductility, 74
Sandstone, 143, 144, 146
compressive and tensile strength, 61
(
foundations, 349, 350
Hoek-Brown constants, 99
joints in, 146
modulus ratio and Poisson's ratio, 186
Mohr-Coulomb parameters, 83
permeability, 35
point load strength, 37
porosity, 29, 83
in triaxial compression, 75
Schist, 8, 20, 21, 161
bearing capacity, 350
compressive and tensile strength, 61
density, 33
modulus ratio and Poisson's ratio, 186
permeability, 35
Schistosity, 13
and strength, 61, 83
Schneider, B., 147, 190, 199-201
Schneider, H. J., 157, 168, 169
Scratch hardness, 416-418
"Seamy rock" in building code, 349
Seismicity, reservoir induced, 87, 172
Serpentinite, 20, 21
Shale, 144-146
bearing capacity, 350
compressive and tensile strength, 37, 61
ductility, 74
Hoek-Brown constants, 99
modulus ratio and Poisson's ratio, 140, 186
Mohr-Coulomb parameters, 83
permeability, 35
pore water pressure, induced, 85
porosity, 29, 83
residual strength, 81
swelling pressure, 380
Shear failure, 56
Shear modulus, 182
of jointed rock, 196
Shear zones, 141
Shear stiffness, 163, 196
Shear strength intercept, 81
typical values, 83
variation with direction, 93
Shi, G. H., 257
Shi's theorem, 259, 260
Shield, R. T., 362
Shtenko, V. W., 382
Siltstone, 145, 146
compressive and tensile strength, 61
Mohr-Coulomb parameters, 83
porosity, 83
Size effect on strength, 90-93, 229
tests, 90, 91, 231, 233
Skinner, E. H., 42
Slake durability classification, 38
Slaking,37
Slate,20
anisotropy in strength, 93
bearing capacity, 350
I n d e ~ 561
Lundborg's strength parameters, 90
Mohr-Coulomb strength parameters, 83
Snow, D. T., 35
Sonic velocity, 39-42
Sowers, G., 39, 378
Specific gravity, 30-33
Spillways, 9
Splitting tension test, see Brazilian test
Squeezing ground, 104, 286
Stability analysis:
of block on aplane, 312-319
of general polyhedron, 325-331
of 2d active/passive system, 333-334
of 2d block, 287
of wedge, 319-320
Stagg, K. G., 189
Standards, 17
for compression tests, 60
Stephenson, D. E., 117
Stereographic projection, 475-493
equations for drawing great circles, 261
joint orientation analysis, 150-154
of joint system, 263
kinematic analysis of slopes, 301-310
limit equilibrium analysis for a given JP,
325-331
roughness angles of joints, 167
stability analysis of slopes, 312-320
Stereonet, 479, 480
Stiffness of joints, 196
of testing machines, 76
Strain in two dimensions, 409-413
Strength, 34, 36, 37, 76
Stress concentrations, 225, 228-231
Stresses, manipulation and operations,
389-408
Stress history related to sedimentation and
erosion, 176
Stresses in the crust map, 112
Stress measurement, 116, 117
complete state of stress, 132-135
doorstopper test, 127-129
flat jack method, 121-123
hydraulic fracturing, 117-121
inclusion methods, 116, 127
overcoring, 123-127
undercoring, 130, 131
562 I n d e ~
Stress-strain relations:
deviatoric values in terms of, 218
elastic, 181
for jointed rock, 195-197, 220
viscoelastic, 206
Strike and dip, defined, 400
Subsidence over old mines, 379
Swelling rock, 380-382
mineral s responsible for, 343, 419
Syenite, 33
Ta ehien dam site, 147
Taylor, H. W., 46
Temperature gradient in ground, 220
Tensile failure, 56, 57
Tensile strength, 82, 119
Terratek, Inc., 12,64,92
Terzaghi, K., 242
Terzaghi, R., 153
Three-dimensional analysis of stresses in situ,
134, 135
Tiedemann, H. R., 42
Timoshenko, S., 189, 468
Tincelin, M. E., 121
Toews, N. A., 126, 135
Topographic effects on stresses, 102,
104, 105, 107
Toppling failure, 298-300
kinematic analysis, 304-306
Tourenq, e., 65
Townsend, F. e., 81
Transversely isotropic rock, 183
Triaxial compression:
equipment, procedures, 61-65
for joints, 160-164
Tributary area, 231, 232
Trona, 20
Tuff,20
compressive and tensile strength, 61
modulus ratio and Poisson's ratio, 186
porosity, 29
Tunnel support spectrum, 274-280
Twin gliding, 74
Two block sliding analysis of slopes, 333, 334
Unconfined compressive strength, see
eompressive strength
Undercoring, 130, 131
Underground openings:
circular, elastic rock, 220, 225
circular, plastic rock, 242-249, 454-464
elliptical, oval, rectangular, 230
spherical, 230
time dependency, creep, 250-256
U.S. Bureau of Mines gage, 124
Van Heerden, W. L., 91
Velocity of sound in rocks, 39-42
longitudinal ve!ocity in minerals, 40
Virgin stresses, see Initial stress
Viscoelasticity:
spring and dashpot models, 204-206
and stress measurement, 117
Viscosity, 202
for water, 34
Vogler, U. W., 61
Voight, B., 116
Volcanic rocks, 20. See alsa individual
rack type
point load strength, 37
Volumetric strain, 68
Wagner, P. L., 546
Water, effect on rock strength, 84-87
Water loss coefficient, 52
Wave equation, 442
Wave velocity, 39-42
Wawersik, W. R., 69, 71
Weathering:
effect on foundations, 341, 343
effect on porosity, 24, 28, 30
Wedge slides, 295-297
kinematic analysis, 304
stability analysis, 319, 320
Wickham, G. E., 42
Willett, D. e., 224
Wilson, L. e., 385
Wittke, W., 318
Woodward, R. J., 370, 377
X-ray measurement of stresses, 117
Young's modulus, 181
fractured or jointed rocks, 195-202
measurement, 183-195
ratio to compressive strength, 186
variation with strain, 185
Yudhbir,99
Zaruba, Q., 132
Zienkiewicz, O. e., 189,352
Zoback, M., 112