-
58 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES
for every site and will be organized into an overall database.
The data will provide the basis fora detailed quantitative
assessment of current design criteria of emergency spillways and
will assistin the development of any needed revisions.
I believe that Dr. Kirsten's idea of applying his erodibility
index (expressed as N) to a rockmass that forms an excavated
spillway channel should be explored in our (SCS) assessment.
Oncethe data base is established from the spillway performance
reports it would be productive todetermine an empirical
relationship between J, and the relative dip angle and block shape
forhydraulic erosion of spillway channel beds. The analysis should
be useful in our assessment ofdesign criteria for excavated rock
emergency spillways.
Nick Barton'
Rock Mass Classification and TunnelReinforcement Selection Using
theQ-System
REFERENCE: Barton, N., "Rock Mass Classification and Tunnel
Reinforcement SelectionUsing the Q-System," Rock Classification
Systems for Engineering Purposes, ASTM STP 984,Louis Kirkaldie,
Ed., American Society for Testing and Materials, Philadelphia,
1988, pp. 59-88.
ABSTRACT: This paper provides an overview of the Q-system and
documents the scope of caserecords used in its development. A
description of the rock mass classification method is given
usingthe following six parameters: core recovery (RQD), number of
joint sets, roughness and alterationof the least favorable
discontinuities, water inflow, and stress-strength relationships.
Examples offield mapping are given as an illustration of the
practical application of the method in the tunnelingenvironment,
where the rock may already be partly covered by a temporary layer
of shotcrete. Themethod is briefly compared with other
classification methods, and the advantages of the method
areemphasized.
KEY WORDS: rock mass, classification, tunnels, rock support,
shotcrete, rock bolts jointing
This paper provides an analysis of the Q-system of rockmass
characterization and tunnel supportselection. The 212 case records
utilized in developing the Q-system (Barton et al, 1974)
arereviewed in detail, so that application to new projects can be
related to the extensive range ofrock mass qualities, tunnel sizes,
and tunnel depths that constitute the Q-system data base.
Ultimately, a potential user of a classification method will be
persuaded of the value of aparticular system by the degree to which
he can identify his site in the case records used to developthe
given method. The most comprehensive data base of the seven or
eight classification systemsreviewed is utilized in the Q-system.
This body of engineering experience ensures that supportdesigns
will be realistic rather than theoretical, and more objective than
can be the case when fewprevious experiences are utilized to
develop a support recommendation.
Classification Systems Currently in UseTable I is an abbreviated
listing of most of the rock mass classification systems currently
in
use internationally in the field of tunneling. These are:
Terzhagi (1946) Rock Load ClassificationThis has been used
extensively in the United Statesfor some 40 years. It is used
primarily to select steel supports for rock tunnels. However, it
isunsuitable for modem tunnelling methods in which rock bolts and
shotcrete are used.
Lauffer (1958) Stand- Up Time ClassificationThis introduced the
concept of an unsupportedspan and its equivalent stand-up time,
which was a function of rock mass quality. It appearsexcessively
conservative when compared with present-day tunneling methods.
' Norwegian Geotechnical Institute, Oslo 8, Norway.
59
-
TABLE 1Major rock mass classification systems.
Name of
Originator CountryClassification and Date of Origin
Applications
Rock Loads Terzhagi (1946) USAStand-Up-Time Lauffer (1958)
AustriaRQD Deere et al (1967) and USA
Deere et al (1970)RSR Concept Wickham et al (1972)
USAGeomechanics Bieniawski (1973) S. Africa
(RMR System)Q-System Barton et al (1974) Norway
tunnels with steel supportstunnelingcore logging, tunnelling
tunnels with steel supportstunnels, mines, etc.
tunnels, large chambers
BARTON ON Q-SYSTEM 6160 ROCK CLASSIFICATION SYSTEMS FOR
ENGINEERING PURPOSES
The NATM relies on performance monitoring for prediction and
classification of groundconditions. It is adapted to each new
project based on previous experience. The classification isalso
adapted during a single project based on performance monitoring. A
particular classificationis therefore only applicable to the one
case for which it was developed and modified, so use byothers on
other projects may be difficult.
The NATM is essentially a design method in which the rock mass
is allowed to yield onlyenough to mobilize its optimum strength, by
utilizing light temporary support. With correct timingof final
support, this initial yielding is arrested in time to prevent loss
of strength. On occasion,the desire to allow deformation to occur
by installing canals of deformable material within theshotcrete,
and steel ribs with sliding joints, has resulted in loss of ground
control and severedamage to final concrete linings and bolt arrays
(Barton, 1982).
" Sec Bibliography for details.
Deere et al (1967) Rock Quality Designation (RQD)This is a
simple description of thecondition of recovered drill core. It has
been successfully adopted as part of subsequent
classificationsystems. On its own, it fails to account for the
condition of joint surfaces and filling materials,and may be overly
sensitive to orientation effects. Deere et al (1970) utilized RQD
to developsupport recommendations for 6 to 12 m span tunnels, but
pointed out that the details of jointing,weathering, and
groundwater should also be taken into account when selecting
support. Fourteencase records were utilized in developing these
recommendations.
Wickham, Tiedemann, and Skinner (1972) Rock Structure Rating
(RSR)This conceptintroduced numerical ratings and weightings to
relate rock mass quality, excavation dimensions,and steel support
requirements. The method was an immediate forerunner to the two
methods nowused most frequently on an international basis (the RMR
and Q systems).
Bieniawski (1973) RMR Geomechanics ClassificationThis evolved
from several earliersystems and has undergone several changes
(1974, 1975, 1976, and 1979) since its first introductionin 1973.
The method was eventually based on 49 case records, though details
of these cases withtheir relation to support recommendations have
not been published. Recent applications of theRMR system have been
made in mining, which has extended the data base considerably.
Barton, Lien, and Lunde (1974) Q-SystemThis classification
system was developedindependently of the Wickham et al (1972) and
Bieniawski (1973) methods, but it builds extensivelyon the RQD
method of Deere et al (1967), introducing five additional
parameters to modify theRQD value to account for the number of
joint sets, the joint roughness and alteration (filling), theamount
of water, and the various adverse features associated with
loosening, high stress, squeezing,and swelling. The classification
method and the associated support recommendations were basedon an
analysis of 212 case records. Full details of these cases are given
later in this paper.
NATM
One further tunnel support concept which should be mentioned
here is the New AustrianTunneling Method (NATM), which was
developed by Rabcewicz, Packer, and Muller (Rabcewicz,1963 and
Rabcewicz and Packer, 1975). As acknowledged by Milner (Salzburg),
almost everyoneusing this method has a different conception of it,
and numerous economic and technical failuresin past years
demonstrate the amount of confusion that prevails in this field. In
NATM's defense,it must be understood that the method is principally
utilized in squeezing ground conditions, whichcould present
technical and economic problems for any tunneling method.
Updating Case Records
Some of the support methods recommended by the above
classification methods are quite laborintensive and will need
updating as new support methods become more generally available.
Forexample, the development of high strength, but highly ductile,
steel fiber reinforced microsilicashotcrete is a revolutionary
advance in tunnel support. It can be applied by one robot operator
andone back-up person right at the tunnel face. The extra strength
of this product removes the needfor mesh in shotcrete, and it has
sufficient early strength to replace steel arches and cast
concreteunder a large range of tunnelling conditions.
Comparison of RMR and Q SystemsThe two classification systems
that appear to be in widest use in tunneling that do not rely
on
performance monitoring (though they can be used in conjunction
with monitoring) are the RMRand Q systems. These two systems are
therefore compared in some detail here.
Bieniawski (1976) rates the following six parameters in his RMR
system:
I. Uniaxial compressive strength of rock material.Drill core
quality RQD.Spacing of joints.Condition of joints.Groundwater
conditions.
6. Orientation of joints.
In contrast, the Q-system (Barton et al, 1974) rates the
following six parameters:
I. RQD.Number of joint sets.Joint roughness.Joint
alteration.Joint water.
6. Stress factor.
The common parameter used in both systems is Deere's RQD.
Bieniawski also includes jointspacing and orientation, while the
Q-system considers the number of joint sets. Orientation isincluded
implicitly in the Q-system by classifying the joint roughness and
alteration of only themost unfavorably oriented joint sets or
discontinuities.
Bieniawski (1975) appears to have favored the mean rating for
spacing and orientation of the
-
62 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES BARTON
ON 0-SYSTEM 63
different joint sets according to the example given in his
paper. He also indicates (1979) that whenonly two joint sets are
present the average spacing will prove conservative, since the
rating isusually based on the presence of three sets of joints.
The very detailed treatment of joint roughness and alteration,
perhaps the strongest feature ofthe Q-system, is not particularly
emphasized in the RMR Geomechanics Classification. In hisoriginal
version Bieniawski (1973) considered the condition of joints under
three descriptive terms:weathering (5 ratings), separation of
joints (5 ratings, 5 mm) and continuity ofjoints (5 ratings, not
continuous up to continuous with gouge). In his 1974 publication
Bieniawskicondensed these three terms to "condition of joints"
which again had five ratings; from very tight,separation 5 mm,
continuous gouge >5 mm. In his laterpublications (1976, 1979),
Bieniawski also includes joint roughness in his fourth
parameter"condition of joints."
In the RMR system, rock stress is not used specifically as a
parameter though it is apparentlywhen selecting support measures.
In his 1975 paper Bieniawski gives support recommendation fora
tunnel of 5 to 12 m span in which the vertical stress should be
less than 30 MPa. In his 1976version, the same support
recommendations are given specifically for 10 m span tunnels, with
thevertical stress limited to 25 MPa.
In the Q-system, the ratio (o.c/o,) (unconfined compression
strength/major principal stress) isevaluated when treating rock
stress problems. The onset of popping, slabbing, and rock
burstproblems can be quite accurately predicted in hard rocks. The
Q-system also accounts for looseningcaused by shear zones and
faults, and squeezing and swelling ground. However, very few
caserecords could be utilized in the squeezing category, so support
recommendations are tentative.
The Geomechanics Classification was based initially on Lauffer
(1958), which is nowacknowledged to be excessively conservative.
Bieniawski increased Lauffer's maximum unsupportedspan of 8 m (in
his 1973 version) to 20 m (1975 version) and finally to 30 m (1979
version).Despite these later modifications, Bieniawski's chart of
stand-up time versus unsupported span isstill seen to be very
conservative compared with the Q-system.
In a detailed comparison of various classification methods,
Einstein et al (1979) compared eachmethod's support estimates with
Cecil's (1970) unsupported cases. They showed that
Bieniawski's(1976) method predicted considerable support in all
cases, the Deere et al (1969) method evenlarger amounts of support.
The Q-system predicted no-support, since Cecil's cases formed
thebackbone of the method. Such discrepancies reflect the important
differences in support philosophiesbetween Scandinavia, South
Africa, and the United States. These differences appear to be
narrowinggradually.
The Einstein et al (1979) comparison of each method with Cecil's
supported cases indicatedthat the Q-system support recommendations
also agree well with the actual support. The Bieniawski(1976) and
Deere et al (1969) RQD method were more conservative.
One of the most recent reports of comparisons between rock mass
classification systems fortunneling was published by Einstein et al
(1983) from work performed at the Porter Square Station,a 168 by 14
by 21 m excavation located 20 to 30 m below the surface in
argillite, in Cambridge,Massachusetts. Five of the classification
methods listed in Table I were compared in detail duringseveral
stages of the project and by several observers.
The methods were compared while classifying drill core, mapping
an inspection shaft and apilot tunnel, and when the rock was
exposed in the final excavation. Only the Q-system was foundto be
applicable to the worst conditions encountered in the main
excavation. Category 32 supportpredicted by the Q-system,
consisting of systematic bolts at I m centers, and 40 to 60 cm
ofshotcrete, was "practically identical to the actual support"
used. Careful monitoring using MPBX(multiple position borehole
extensometers) and convergence measurements revealed
maximumdeformations of only 8 mm; in other words, stable conditions
were established.
Description of the Q-SystemRationale
The vast majority of the thousands of kilometers of tunnels
constructed world-wide every yeardo not have the benefit of
performance monitoring. Design decisions are nevertheless
requiredboth before and during construction. No matter how many
sophisticated rock mechanics testprograms and finite element
analyses are performed, design engineers will come back to the
basicquestion: "Is this bolt spacing, shotcrete thickness, or
unsupported span width reasonable in thegiven rock mass?"
At present we have to rely on engineering judgment, or on
classification methods, where thedesign is based on precedent, and
where a good classification method will allow us to extrapolatepast
designs to different rock masses and to different sizes and types
of excavation. Undergroundexcavations can be supported with some
confidence primarily because many others have beensupported before
them and have performed satisfactorily.
Method for Estimating Rock Mass Quality (Q)The six parameters
chosen to describe the rock mass quality (Q) are combined in the
following
way:
Q = (RQD/J,,) (J,1J) (J,/SRF) (1)
where
RQD = rock quality designation (Deere et al, 1967),J = joint set
number,J, = joint roughness number (of least favorable
discontinuity or joint set),J,, = joint alteration number (of least
favorable discontinuity or joint set),J. = joint water reduction
factor, and
SRF = stress reduction factor.
The three pairs of ratios (RQD/J, J,/.1, and J./SRF) represent
block size, minimum inter-blockshear strength, and active stress,
respectively. These are fundamental geotechnical parameters.
It is important to observe that the values of J, and J relate to
the joint set or discontinuity mostlikely to allow failure to
initiate. The important influence of orientation relative to the
tunnel axisis implicit.
Detailed descriptions of the six parameters and their numerical
ratings are given in Table 2.The range of possible Q values
(approximately 0.001 to 1000) encompasses the whole spectrumof rock
mass qualities from heavy squeezing ground up to sound unjointed
rock. The case recordsexamined included 13 igneous rock types, 26
metamorphic rock types, and 11 sedimentary rocktypes. More than 80
of the case records involved clay occurrences. However, most
commonly thejoints were unfilled and the joint walls were unaltered
or only slightly altered.
The Q-system is more detailed than any of the other methods as
regards the factors jointroughness (or degree of planarity), joint
alteration (filling), and relative orientation. The
classificationof "least favorable features" (for J, and .1,,)
represents one of the strongest features of the method.It also
seems to be a factor that is virtually ignored in the other
classification schemes. Forexample, in Bieniawski's RMR method,
although data for all joint set and discontinuities arecollected,
only the average data are incorporated in the numerical ratings.
Furthermore, in theRMR it is impossible to separately vary the
degree of joint roughness and the degree of infilling,as obviously
may occur in practice.
-
(J.) (Or)(approx.)(a) Rock wall contact
Tightly healed,hard,non-soften-ing,impermiable filling
i.e.quartz or epidote
0.75 ( - )Unaltered joint walls,surfacestaining only 1.0Slightly
altered joint walls.Non-softening mineral coatings,sandy particles,
clay-freedisintegrated rock etc.
2.0Silty-,or sandy-clay coatings,small clay fraction (non-soft.)
3.0
E. Softening or low friction claymineral coatings,
i.e.kaoliniteor mica. Also chlorite,talc,gypsum,graphite etc.,
andsmall quantities of swellingclays.
(b) Rock wall contact before10 .ms shear
BARTON ON Q-SYSTEM 6564 ROCK CLASSIFICATION SYSTEMS FOR
ENGINEERING PURPOSES
TABLE 2-Ratings for the six Q-system parameters(table continues
on pp. 65 and 66).
1. ROCK QUALITY DESIGNATION (RQD)
Very poor 0 - 25
Poor 25 - 50
Fair 50 - 75
Good 75 90E. Excellent
90 - 100
Note: (i) Where RQD is reported or measured as 10,
(including 0) a nominal value of 10 is usedto evaluate Q in
equation (1).
(ii) RQO intervals of 5, i.e. 100,95,90, etc. aresufficiently
accurate.
(J )2. JOINT SET NUMBER
Massive,no or few joints One joint set One joint set plus
random
Two joint sets Two joint sets plus random Three joint sets Three
joint sets plus random Four or more joint sets,random,heavily
jointed,"sugar cube" etc. .... 15
J. Crushed rock,earthlike 20
Note: (1) For intersections use (3.0 x Jn)Note: (ii) For portals
use 12.0 JD)
0.5 - 1.02
469
12
3. JOINT ROUGHNESS NUMBER
Rock wall contact andRock wall contact before (Jr)
10 ems shearDiscontinuous joints 4Rough or irregular,undulating
3
Smooth,undulating 2
Slickensided,undulating 1.5
Rough or irregular,planar 1.5Smooth,planar 1.0
Slickensided,planar 0.5
Note: (1) Descriptions refer to small scale featuresand
intermediate scale features,in thatorder.
(c) No rock wall contact when shearedZone containing clay
minerals thick enoughto prevent rock wall contact 1.0
J. Sandy,gravelly or crushed zone thick enoughto prevent rock
wall contact 1.0
Note: (ii) Add 1.0 if the mean spacing of the relevIntjoint set
is greeter than 3m.
(iii) J r w0.5 can be used for planar slickensidedjoints having
lineations,provided the line-ations are orientated for minimum
strength
TABLE 2-(continued).4. JOINT ALTERATION NUMBER
Sandy particles,clay-freedisintegrated rock etc.
4.0 (25-30')Strongly over-consolidatednon-softening clay
mineralfillings (continuous,but
-
66 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES BARTON
ON Q-SYSTEM 67
Method of Selecting Suitable SupportThe Q-system is essentially
a weighting process, in which the positive and negative aspects
of
a rock mass are assessed. A store of experience (case records),
which is itself based on earlierexperience, is searched to try to
find the most appropriate support measures for the given
excavationsand rock mass conditions. The whole procedure is
probably not dissimilar to the mental processoccurring when a very
experienced tunneling consultant is asked for his support
recommendations.While the assessment of most of the parameters is
admittedly subjective, the process of supportselection is organized
and reasonably objective. The trial-and-error adjustment and
readjustmentof parameter ratings necessary during the development
of the Q-system was an important factorin reducing the need for
subjective judgments on the part of the developers. The large
number ofcase records made it possible to generate the support
recommendations quite objectively.
Figure I shows that the tunnel or excavation span width and the
rock mass quality (Q) are thedecisive parameters for placing an
excavation in a given support category (Boxes 1 to 38).However,
there is an important user requirement for different degrees of
safety. The excavationsupport ratio (ESR), which reduces the
effective span in Fig. 1, reflects construction practice inthat the
degree of safety and support demanded by an excavation is
determined by the purpose,presence of machinery, personnel, etc.
The list of ESR values in Table 3 was developed throughexhaustive
trial and error, and seems to be the most workable solution to the
problem of variablesafety requirements.
Increased safety can be selected at will by reducing the ESR
value (e.g. by using ESR = 1.3for an important permanent mine
opening in place of 1.6). Similarly, support for oil storagecaverns
could be selected by using ESR = 1.6 instead of 1.3, assuming the
occasional fall ofsmall stones (from the walls) were acceptable.
The use of ESR = 1.0 for power station chambersand major road
tunnels ensures a high factor of safety, as obviously required.
Most of the 38 numbered "support boxes" shown in Fig. 1 contain
case records. The supportrecommendations for cases that plot in
these same boxes are listed in Table 4. Further details ofthe use
of these support recommendations are given by Barton, Lien, and
Lunde (1975). Roofsupport, wall support, and temporary support can
be selected as required.
Examples of Case RecordsA considerable data base for developing
the Q-system was provided by Cecil (1970), who
described numerous tunneling projects in Sweden and Norway,
including detailed evaluations ofthe rock, the jointing, the type
of support, and the apparent stability. Figure 2 shows three
examplesof Cecil's cases, and Table 5 gives the abbreviated
descriptions of the key rock mass parametersand describes the
support actually used. Note that the alphabetic descriptions given
in brackets inTable 5 can be checked directly against the same
letters given in Table 2. This convenientshorthand method can be
used during tunnel mapping, when writing conditions are
unfavorable.
Examples of Tunnel MappingExamples of some tunnel projects in
which the Q-system has been used extensively in day-to-
day follow-up mapping are shown in Figs. 3a and 3b. It will be
noticed that the treatment ofcrushed zones and major
discontinuities shown in Figs. 3a and 3b is often on an individual
basis.The quality of the rock mass between the zones is a decisive
factor in deciding between individual"stitching" and general
support.
One of the examples (Fig. 3b) is a tunnel excavated by a
full-face tunnel boring machine(TBM). The minimal disturbance
caused by TBM excavation makes it particularly important tomap as
close to the advancing face as possible (for optimal joint
definition), followed by repeated
6. STRESS REDUCTION FACTOR TABLE 2(continued).
(a) Weakness zones intersectingexcavation,which may
causeloosening of rock mass whentunnel is excavated. (SRF)
Multiple occurrences of weak-ness zones containing clay
orchemically disintegrated rock,very loose surrounding rock(any
depth)
10 Single weakness zones cont-aining clay or
chemicallydisintegrated rock(depth ofexcavation 50m) Single
weakness zones cont-aining clay or chemicallydisintegrated rock
(depth ofexcavation > 50m ) 2.5Multiple shear zones in
compet-ent rock (clay-free),loose surr-ounding rock (any depth)
7.5
E. Single shear zones in competentrock (clay-free) (depth
ofexcavation S 50m ) 5.0Single shear zones in competentrock
(clay-free) (depth of excav-ation 50m I 2.5Loose open
joints,heavily jointedor "sugar cube" etc. (any depth) 5.0
Note: (i) Reduce these values of SRF by25 - 50% if the relevant
shearzones only influence but do notintersect the excavation.
(b) Competent rock, rock stress problems
oc ia lof/al (SRF)
Low stress, near surface >200 >13 2.5Medium stress '00-10
13-0.66 1.0High stress, very tightstructure (usually fav-ourable to
stability,may be unfavourable forwall stability) 10-5 0.66-.33
0.5-2Mild rock burst(massive rock) 5-2.5 0.33-.16 5-10Heavy rock
burst(massive rock) 02.5 10,reduce oc and a t to 0.60cand 0.60 t ,
where o cunconfinedcompression strength, and a t = tensilestrength
(point load), and 0 1 and 0 3 arethe major and minor principal
str?,SeS.
(iii) Few case records available where depth ofcrown below
surface is less than spanwidth. Suggest SRF increase from 2.5 to
5for such cases (see H).
Squeezing rock:plastic flow of incompetentrock under the
influence of high rock pressure
(OAF)M. Mild squeezing rock pressure ,
5 100. Heavy squeezing rock pressure 10 20
swelling rock:chermecal swelling activitydepending on presence
of water
P. Mild swelling rock pressure 5 - 10R. Heavy swelling rock
pressure 10 15
-
R, R 0 e)
a
0
VS3 / NINS
8
8
8n
VICC
2 x
2 1
0.1
aatt"MI
a
9.11=111111119.1111111ONIMMIIIIn71
0.10=1=M11=1.nAlnnn111MIINIMti 1111111110111M=MMII
2281==MMIIMIIIIIIIIIMMEMIIInnnMEMIN
0 11
Ili
11111111111111111M1
1111111111011011=11111111111111111111111111.1.111111E111nn
11111111
08). 0O
al gamma...m.....rx====:::=E=se mowMOO :. 1.."1:=;;;I:mr61min
'=a0MMINIIIMMIIIIIMIMI .1111=NIMIIIIIIIIIMIIIMIMOE111111
MI
Ma IIIIIMEM
11 MOINE111111nn
,..._camOMISn1IMINUCIN=M11=1UM.IMEMINONINEMIEMEMIIUnnn
1111111essiuminn UMW
11111UOMMMMIN
11111111111n1 1n111111 nn1
Z0 0 0V
11
nnnnnnnnnN11181111111111MINMIN11611111111111111111=11116111111111111111
. 0)1-4 gsr 0
MI INIMIIIM=IMOMMOOMInnn
=11=11111n11,n11101111==n11n111n11MIIMMIIINn1111 161111111111/1/
sumwoommossoessammissBONIFIIIMIIMIIIn12nnn12.1
IIMMIINIMINMEMENIMOMMINI=M111INmill11111111111113111111111M11111111111111111111111MMII
1111n 1111111
11 111 In11111111MUM
68 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES BARTON
ON 0-SYSTEM 6900 0
el 0 0 in .1, PITABLE 3Excavation Support Ratio (ESR)pr a
variety of underground gveavations.
NumberESR of Cases
Temporary mine openings etc. ca. 3-5? 2Permanent mine openings,
water tunnels for hydro power (excluding high pressurepenstocks),
pilot tunnels, drifts and headings for large openings 1.6 83Storage
caverns, water treatment plants, minor road and railway tunnels,
surgechambers, access tunnels, etc. 1.3 25Power stations, major
road and railway tunnels, civil defense chambers,
portals,intersections 1.0 79
E. Underground nuclear power stations, railway stations, sports
and public facilities,factories ca. 0.8? 2
mapping before permanent support is chosen. There will then be
improved possibilities forobserving the character of narrow
clay-bearing discontinuities. The effective RQD of the zone ofrock
around a TBM excavated tunnel will generally be higher than that
around a blasted tunnelowing to the relatively slight disturbance
of incipient joints and tight structures.
Figure 4 illustrates ten parallel (100 m long) sewage treatment
caverns constructed near Oslo.At the feasibility and planning
stage, surface mapping and drill core analysis were interpreted
interms of the Q-system parameters. Support requirements were
predicted on this basis. Duringconstruction, support decisions were
also guided by the method outlined. The general improvementin rock
conditions as the parallel caverns advanced from the shale into the
nodular limestone wereclearly reflected in the six parameters and
support was reduced accordingly:
Shale: B 1.25 in c/c, L 3.5 m + S (mr) 12-15 cmNodular
limestone: B 1.5 m c/c, L= 3.5 m+ S 5 cm
(B = bolting, c/c = spacing, L = length, S = shotcrete, mr =
mesh reinforced)
The Q-system has also been used in the area of mine stability.
In a recent assessment of stabilityin two limestone mines with 13
to 15 m span rooms or drifts, respectively, the quality of
thelimestone varied as follows:
80 100 I 1.5 1Q = = 4 38 (fair -- good)4 9
1 2 1
In the great majority of the drifts the quality was "good" (Q =
18 38). By comparing thesequalities with the permanently
unsupported cases (Fig. 1, black circles) it was possible
todemonstrate satisfactory conditions for the great majority of the
excavations. However, in places,stability was apparently nearer the
"temporary mine openings" category (ESR = 3-5, Table 1).There were
in fact limited areas in the mines where fall-out occurred from the
pillars or walls. Alimited number of pillars in one of the mines
were instrumented as a precaution.
Rock Mass Requirements for Permanently Unsupported ExcavationsAn
important area of application for the Q-system is the recognition
of rock mass characteristics
required for safe operation of permanently unsupported openings.
The relationship between themaximum unsupported span and the
Q-value is clearly seen in Fig. I. Detailed analysis of
theavailable case records reveals the following requirements:
Type of Excavation
-
70 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES BARTON
ON Q-SYSTEM 71
TABLE 4-Support recommendations for the 38 categories shown in
Fig. I (see Barton et al, 1974, 1975for notes: table continues on
pp. 71 and 72). TABLE 4-(continued).
Cottilatonal factorsSupport Ro J. SPANrate-gory .1 1, ESR Tvpe
of .support NotesI" -
-
sblutg) -
-- -
sblutg) -- - - sblutg) -
4' - sblutg) -5 -- sblutg)
6 - - sblutg) -7' - - - sblutg) -g'
--
sblutg) -Note. The type of suppon to be used in categories I to
0 will depend on the
blasting technique. Smooth wall blasting and thorough
barring-downmay remove the need for suppon. Rough-wall blasting may
result in theneed for single applications of shotcrete, especially
where the excavationheight is >25 m Future cast records should
differentiate categories Ito 8.
9 020 - .sblutg))
-
cr)w
0
If)
O
-0
0.
O
O
72 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES BARTON
ON Q-SYSTEM 73
TABLE 4(continued).
Conditional factors5"PP.'" RQD 1, SPAN!UM-
80ry 1, .1, ESR Type of support Notes
-
TABLE 5-Comparison of support used and support recommended for
three case records described b y Cecil (1970).
CaseNo.
3.
DESCRIPTION OF ROCK MASSNature of instabilityPurpose of
excavation,location, reference
SPAN Height
(m ) ( m )
Depth
( m)Support used
RQD L._1.r.J
.3n
(Code :Tables
,w
SRF
1 to 6)
Q ESR SPAN Estimate ofpermanentroof support
ESR
24 60 m length, including a
1 m wide shear zone in Rock bolts, Category 22mylonite. Chrushed
wire mesh and 60 1.0 1.0 =13 1 mmylonite and non-softeningclay
seams and jointfillings. Intersectingjoint set. 2 joint setsplus
random, 5-30 cm
12.5 6.5 60 shotcrete 6 6
( 31 )
2.5
( SA.)
+S(mr)2.5-5 cm
spacing. Minor waterinflows (
-
ocO
.--
-
oo.
'....z
....,
I
.i l -....
.
", ,
o
O..4.4ir
(clay
60
40r.viv
4asample
eo 4,100A ,-,-,44
I.'ti,ItI
t
I, KS
830
825
815
810
805
,---
820
.".
t
,..
(C/C IL.4
+ Simi,10cmto in-vert
I. ;
..
8Slmr)15cm
.
. ,
7 '``.. '..,
..' ,
,.. .
,i,. .
. + ..,
,.. ..,...
?
.." ......
. V..
1,51C/C)-.
!oror
17,
Jy
E ..:
z , - 9
ROCK MASSDESCRIPTION
.
ice,1 .
,;
). ,-n .k-
-It TEMPORARY`I SUPPORT
t 'C
RECOMMENDEDSUPPORT
.4Z
=
'-'
z
,
;6--'
1
cC3 --nw
t..a ,;!.!lf,
ng
1. L3,Z
,5', -_-1- -.
C C"'
NOTES elo - 830 Heavily jointed , partly
crushed with a little c lay
Orginally large waterie akageROCK
Gneiss 810 - 817 Crushed zone with clay oar[25 3 SO
SiONet)
TUNNEL MAPPING - SUPPORT PROJEC TNO 71610LOCALITY HYLLEN , ULLA
- FORRE 5NEE T NO
ROCK MASSDiscontinuity,thickness 50cm
Crushed zone,without clay
Crushed zone,weathered withclay
80Strike/dip
Dikes and sills
0'.o- 1
o
...
PI lb. -
ritlipe0,1041.
11111.
111
4 (01.401,111.
10,., . .Are Like
-I-ntictine-r-
..,411,0,4 40
ri0 0
1090
1080-
1070-
1060
1050 -
!.!.
S I m 111111114
1441%
11111111111111
S 5 c m
10cm
S Im I10cm
.
..
...
;-
,-.
1
o -H.0
..2 m
. -
'c'l
I
a-
,....
-
-
1,,
.
_
a.,
e--,...o
.2: c-..1,,,,,,
-
o,, - -
. ,?, I - ' -
r-.. / t' '- ,'" \ , _r zt ' -_C-%.l,c,5 ROCK MASS ic TEMPORARY
RECOMMENDED
DESCRIPTION ": SUPPORT SUPPORTiLo
z7(.T.,
,_o
z
-L' a --,N,,
I
I
L'ezL---".,
t,x-, r
,.-_-,-sr."z
i,='n2, a -'
2a ET.
NOTES- 1060-1075 Fault 20 cm crushedwith 10 cm clay filling1080
Fault 3m crushed zone1090 2 m crushed zoneROCKNodular lime -
stone and shale 1100 Discontinuity with 5cm clay fitting A TI 80
s2,TUNNEL MAPPING - SUPPORT PROJECT NONO 716 28LOCALITY HOL ME N
(right I 50000 NO
SUPPORT
B L0 o fi Expansion bolts
Bolts with bands
Grouted bolts
Shotcrete
Mesh reinforcedshotcrete with grouted bolts
Cast concrete arch
S (mr) , ;
76 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES BARTON
ON Q - SYSTEM 77
FIG. 3aExample of tunnel mapping using the Q-system: 160 m 2
headrace tunnel. FIG. 3bExample of tunnel mapping using the
Q-system: 3.3 m diameter TBM driven sewage tunnel.
-
=CY T--)-5 --CM77:nF--116m
0
BO
1.0'At
60U.0
CO
02
20
0
BARTON ON 0-SYSTEM 7978 ROCK CLASSIFICATION SYSTEMS FOR
ENGINEERING PURPOSES
SPAN OR DIAMETER (m)
1
60m
40mNODULAR LIMESTONE AND SHALE
20m
==n==r-)==r-)==n 0FIG. 4Vertical section through the VEAS sewage
treatment plant, Oslofjord.
General requirements for permanently unsupported openings (i.e.,
it is preferable that):
J 9, J,.- 1.0, J-- 1.0, J,= 1.0, SRF 2.5 (see Table
2).Conditional requirements for permanently unsupported
openings:
If RQD 40, need J 2.If J = 9, need J, 1.5 and RQD
90.If f, = 1.0, need J, < 4.If SRF > 1, need J, % 1.5.If
SPAN > 10 m, need J, < 9.
7. If SPAN > 20 m, need 4 and SRF 1.
The shorthand in No. 3 gives the following recommendations:
3. If there are as many as three joint sets (J = 9), then one
needs joint roughness (J,) at leastequivalent to rough-planar or
smooth-undulating (J, % 1.5) and one needs RQD = 90
(i.e.,"excellent").
Existing natural and man-made openings indicate that very large
unsupported spans can besafely built and utilized, if the rock mass
is of sufficiently high quality. Our case records
describeunsupported man-made excavations having spans from 1.2 to
100 m.
Analysis of Q -System Case RecordsUltimately, a potential user
of a classification method will be persuaded of the value of a
particular system by the degree to which he can identify his
site in the case records used to developthe given method. In this
section, the characteristics of the 212 case records used to
develop theQ- system arc analyzed by means of histograms. The
potential user can evaluate to what extent hissite fits with the
data base, or whether he would be relying on few, extreme value
cases, whichwould inherently reduce the reliability of associated
support recommendations.
Histograms of the principal parameters used to codify the 212
case records used in the Q-systemhave been developed. The following
brief summary describes the extreme values and the mostcommon
values of the various factors affecting tunnel stability in the
case records considered.
Support Method
The large majority (180) of the 212 case records were supported
excavations. Only 32 caseswere permanently unsupported. Support
ranged from spot bolting (as little as 50 bolts over a roofarea of
6000 m 2) to very heavy rib-and-rock-bolt-reinforced concrete of 2
to 3 m thickness, poured
1-5 5-10 10-15 15-20 20-30 30-100
FIG. 5Histogram of tunnel spans.
in multiple arch and wall drifts. The predominant form of
support in the case records was rockbolts,or combinations of
rockbolts and shotcrete, often mesh reinforced. Occasionally,
extreme conditionscalled for extreme varieties of support (e.g.,
9.8 m long rockbolts on 0.9 m centers together with14.6 m long
bolts on overlapped 0.9 m centers) and mesh-reinforced gunite.
Dimensions
The cases studied ranged from unsupported 1.2 m wide pilot
tunnels to unsupported 100 mwide mine caverns (Fig. 5). The
predominant tunnel dimensions (span or diameter) were 5 to 10m (78
cases) and 10 to 15 m (59 cases). Excavation heights ranged from
extreme values of 1.8to 100 m. A significant body of the case
records came from hydroelectric projects; consequentlythere were
some 40 cases of large caverns with spans in the range of 15 to 30
m and wall heightsin the range of 30 to 60 m.
Depths
Excavation depths ranged from 5 to 2500 m, though most were
commonly in the range of 50to 250 m. The Scandinavian bias caused
predominantly by Cecil's (1970) case records is shownin Fig. 6.
Note that the other case records contribute most of the data on the
deeper-seatedexcavations, including numerous hydropower
caverns.
RQDThe "rock quality designation" (RQD) ranged in a quite
uniform manner from 0 up to 100%.
Forty (40) cases lay in the "very poor" category (0 to 25%) and
fifty-three (53) cases in the"excellent" category (90 to 100%).
Number of Joint Sets (.1,)The number of joint sets was most
commonly in the range of one set (plus random) to three
sets (plus random). Fifty-two (52) cases, the largest group, had
exactly three joint sets. Extremecases consisted of massive,
unjointed rock and completely crushed, disintegrated rock.
60 m
40m
20m
0
-
0-25 25-50 60-100 100-260 250-500 600-250
501-
0
30
20
10
0
16
10
O
80 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES BARTON
ON Q-SYSTEM 81
TUNNEL DEPTH (meters)
SCANDINAVIAN CASES (CECIL. 1970)
ME OTHER CASESFIG. 6Histogram of tunnel depths.
Joint Roughness (sr)The joint roughness numbers most commonly
found in the 212 case records were 1.0-1.5-2.0,
which represent smooth-planar, rough-planar, and
smooth-undulating surfaces, respectively.Extreme values consisted
of discontinuous joints in massive rock (16 cases) and plane
slickensidedsurfaces (17 cases) typically seen in faulted rock with
clay fillings.
Joint Alteration (.L)The joint alteration parameter most
commonly seen in the case records was represented by the
number 1.0 (unaltered or unweathered). One hundred and three
(103) cases were in this class.However, more than eighty (80) of
the case records involved clay mineral joint fillings of
variouskinds; these included twelve (12) swelling clay occurrences.
Thirteen (13) cases consisted ofhealed joints, which are obviously
very favorable for stability and for their inherently
lowpermeability.
Joint Water (.1w)The joint water reduction factor describing the
degree of water inflow was strongly biased in
the direction of "dry excavations or minor inflow" (
-
40
30
20
1
0
I BARTON ON 0-SYSTEM
TABLE 6Frequency of occurrence of rock types in examined case
records.
83
I. Igneous II. Metamorphic III. Sedimentary
Basalt 1 Amphibolite 8 Chalk 1Diabase 4 Anorthosite (meta-) 1
Limestone 3Diorite 2 Arkose I Marly LimestoneGranodiorite 1 Arkose
(meta-) 3 MudstoneQuartzdiorite I Claystone (meta-) 2 Calcareous
MudstoneDolerite 1 Dolomite 1 Sandstone 4Gabbro 2 Gneiss 14 Shale
2Granite 46 Biotite Gneiss 1 Clay Shale 2Aplitic Granite I Granitic
Gneiss 4 Siltstone 2Monzonitic Granite 1 Schistose Gneiss 2
MarlQuartz Monzonite 2 Graywacke I Opalinus ClayQuartz Porphyry 2
GrecnstoneTuff 2 Schistose meta Graywacke
Quartz HornblendeLeptite IMarbleMylonite 4Pegmatite 2Syenite
1Phyllite 1Quartzite 13Schist 17Biotite Schist 1Mica Schist
2Limestone Schist 1Sparag mite 2
82 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES
ROCK MASS QUALITY (0)
.001-.01 .01-0.1 0.1-1.0 1.0-4 4-10 10-40 i400 100-1000
-
-
-
-
-
-
EXCEPT.POOR
EXTREM.POOR
VERYPOOR POOR FAIR 0000
VERY0000
EXT.EXC.
GOOD
FIG. 8Histogram of Q-value for all 212 case records.
The predominant rockmass characteristic in the cases with
popping, slabbing, or rockburstingwas relatively massive rock, with
few joint sets and/or wide joint spacing. The mean RQD forthese
cases was 91% (range 70 to 100%), and the mean J, value was 5.0
(two joint setstwoplus random). Jointing ranged from three widely
spaced sets (J = 9) to massive intact rock (J= 1.0).
Squeezing or swelling problems (groups (c) and (d)) were
encountered in only nine of the caserecords, although a total of
twelve cases were listed as rock containing swelling clay such
asmontmorillonite.
Q-ValueThe whole spectrum of rock mass qualities exhibited by
the case records is shown in Fig. 8.
As expected, the majority of cases (76%) fall in the central
categories "very poor" (Q = 0.1 to1.0), "poor" (Q = I.0 to 4),
"fair" (Q = 4 to 10), and "good" (Q = 10 to 40). The wholespectrum
of case records utilized in the Q-system ranges from qualities of
0.001 (extreme squeezing)to 800 (essentially unjointed, massive
rock).
Rock TypesThe distribution of rock types represented in the case
records can be summarized as follows:
igneous rock (13 types), metamorphic rock (26 types), and
sedimentary rock (II types). Table 6provides a complete breakdown
on the rock types and the number of case record occurrences ofeach
type. The statistics are dominated by granite (48) and gneiss (21).
However, there aresignificant numbers of case records involving
schist (21), quartzite (13), leptite (11), and amphibolite(8).
Sedimentary rocks are relatively poorly represented with only 19
cases.
Conclusions
I. The large number of case records utilized to develop the
Q-system ensures that reliablesupport recommendations are provided
for a very wide range of tunnel sizes, types of excavation,depths,
and rock mass qualities.
Detailed analysis of the case records has revealed the overall
distribution of individual rockmass parameters such as joint
roughness, alteration, and stress-strength ratios, so that
extremevalue cases can be readily identified.
Squeezing ground is the only class of problems that is
inadequately represented in the originaldata base. Swelling,
slabbing, and rock bursting problems are represented in a
sufficient numberof case records for reliable determination of
support requirements. General tunneling conditionsare extremely
well represented, with 160 case records in the range of Q-values
from 0.1 (verypoor) to 40 (good).
Fifty individual rock types are represented in the case records.
Their characteristics arequantified in such a manner that the
individuality exhibited by many rock types is carried all theway
through to support selection. Application of the Q-system to other
rock types than thosedescribed in the case records can be performed
with confidence, provided that any specialcharacteristics of the
new rock type are adequately represented in the six parameters. A
case inpoint would be the susceptibility to alteration by exposure
to moisture. The environment expectedunder tunnel use must always
be carefully considered.
The Q-system has been used for several years in conjunction with
Norwegian tunnelssupported with fiber-reinforced microsilica
shotcrete, a revolutionary new material that is rapidlyreplacing
labor-intensive mesh-reinforced shotcrete. Very poor rock qualities
previously requiring
-
84 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES
DISCUSSION ON Q-SYSTEM 85
cast concrete arches are being successfully supported by
fiber-reinforced shotcrete and rock bolts.Updating of the Q-system
for use in countries with access to this new temporary and
permanentsupport method is underway at NGI.
BibliographyBarton, N., Lien, R., and Lunde, J., "Engineering
Classification of Rock Masses for the Design of Tunnel
Support," Rock Mechanics, Vol. 6, No. 4, 1974, pp.
189-236.Barton, N., Lien, R., and Lunde, J., "Estimation of Support
Requirements for Underground Excavations,"
in Proceedings, 16th Symposium on Design Methods in Rock
Mechanics, Minn., 1975, published byASCE, New York, 1977, pp.
163-177; discussion on pp. 234-241.
Barton, N., "Unsupported Underground Openings," Rock Mechanics
Discussion Meeting, Befo, SwedishRock Mechanics Research
Foundation, Stockholm, 1976, pp. 61-94.
Barton, N., Loset, F., Lien, R., and Lunde, J., "Application of
the Q-System in Design Decisions ConcerningDimensions and
Appropriate Support for Underground Installations," in Proceedings,
InternationalConference on Sub-Surface Space, Rockstore, Stockholm,
Sub-Surface Space, Vol. 2, 1980, pp. 553-561.
Barton, N., "Characterizing Rock Masses to Improve Excavation
Design," Panel Report, Theme II, 4th IAEGCongress, India, 1982.
Bieniawski, Z. T., "Engineering Classification of Jointed Rock
Masses," Transactions of the South AfricanInstitution of Civil
Engineering, Vol. 15, No. 12, 1973, pp. 335-344.
Bieniawski, Z. T., "Geomechanics Classification of Rock Masses
and Its Application in Tunnelling," inProceedings, 3rd
International Congress on Rock Mechanics, ISRM, Denver, Colo., Vol.
IIA, 1974, pp.27-32.
Bieniawski, Z. T., "Case Studies: Prediction of Rock Mass
Behavior by the Geomechanics Classification,"in Proceedings, 2nd
Aust.-N.Z. Conference on Geomechanics, Brisbane, Australia, 1975,
pp. 36-41.
Bieniawski, Z. T., "Rock Mass Classifications in Rock
Engineering," in Proceedings, Symposium Explorationfor Rock
Engineering, Johannesburg, A.A. Balkema, Vol. I, 1976, pp.
97-106.
Bieniawski, Z. T., "Rock Mass Classifications in Rock
Engineering Applications," in Proceedings, 4thInternational
Congress on Rock Mechanics, ISRM, Montreaux, Vol. 2, 1979, pp.
51-58.
Cecil, 0. S., 111, "Correlations of Rock Bolt-Shotcrete Support
and Rock Quality Parameters in ScandinavianTunnels," Ph.D. thesis,
University of Illinois, Urbana, 1970, 414 pp.
Deere, D. U., Hendron, A. J., Jr., Patton, F. D., and Cording,
E. J., "Design of Surface and Near-SurfaceConstruction in Rock," in
Failure and Breakage of Rock, C. Fairhurst, Ed., Society of Mining
Engineersof AIME, New York. 1967, pp. 237-302.
Deere, D. U., Peck, R. B., Parker, H. W., Monsees, J. E., and
Schmidt, B., "Design of Tunnel SupportSystems," Highway Research
Record, No. 339, 1970, pp. 26-33.
Einstein, H. H., Steiner, W., and Baecher, G. B., "Assessment of
Empirical Design Methods for Tunnels inRocks," in Proceedings, 4th
Rapid Excavation Tunneling Conference, AIME, New York, Vol. 1,
1979,pp. 683-706.
Einstein, H. H., Azzouz, A. S., McKown, A. F., and Thompson, D.
E., "Evaluation of Design andPerformancePorter Square Transit
Station Chamber Lining," in Proceedings, Rapid Excavation
andTunneling Conference, Chicago, Ch. 36, Vol. I, 1983, pp.
597-620.
Lauffer, H., "Gebirgsklassifizierung far den Stollenbau,"
Geologie und Bauwesen, Vol. 24, 1958, pp. 46-51.
Palmstrom, A., "The Volumetric Joint CountA Useful and Simple
Measure of the Degree of Rock MassJointing.'' in Proceedings,
Fourth Cong. Int. Assoc. of Engineering Geology. New Delhi, India,
Vol.V, Theme 2, 1982, pp. V.22I-V.228.
Rabcewicz, L., Bemessung von Hohlraurnbauten, Die "Neue
Osterreichische Bauweise" und ihr Einlluss aufGebirgsdruckwirkungen
und Dimensionierung, Felsmech. u. Ing. Geol., Vol. I, H. 3/4, 1963,
p. 224.
Rabcewicz, L. and Pacher, F., Die Elemente der "Neuen
Osterreichischen Tunnelbauweise" und ihregeschichtliche
Entwicklung," Osterr. Ing. Zschr., Vol. 18, Jg., 1975, p. 315.
Terzaghi, K., "Rock Defects and Loads on Tunnel Supports," in
Rock Tunneling with Steel Supports, R. V.Proctor and T. White,
Eds., Commercial Shearing Co., Youngstown, Ohio, 1946, pp.
15-99.
Wickham, G. E., Tiedernan, H. R., and Skinner, E. H., "Ground
Support Prediction Model (RSR Concept),"in Proceedings, 1st Rapid
Excavation Tunneling Conference, AIME, New York, 1972, pp.
43-64.
DISCUSSION
H. A. D. Kirsten' (written discussion)The table of Barton et al
[DI, Table 2] for the jointalteration number (.10 ) is quite
complex and relatively open to interpretation. The criteria may
besystematized and extended as shown in Table D-1 to overcome these
problems to a large extent.
The stress reduction factor (SRF) is not a geological property
that varies for different regimesof rock around an excavation. It
is a parameter that characterizes the loosening of the rock
aroundthe excavation as a whole.
Barton et al provide qualitative criteria for determining SRF
for incompetent, non-homogeneousrockmasses [DI, Table 3a1 and
competent, homogeneous rockmasses [Dl, Tables 3b, 3c, and34 The
non-homogeneities provided for involve single or multiple clay
filled weakness or clayfree shear zones. The determination of SRF
is open to interpretation, because of the qualitativenature of the
criteria and the difficulties that often arise in deciding whether
the rockmass ishomogeneous or not. Further difficulties arise in
the case of homogeneous rockmasses with regardto assessing the
degree of stressing, bursting, squeezing, and swelling.
These problems can be overcome by observing that, in the case of
non-homogeneous rock, SRFis related to the overall quality of the
rock, Q, and, in the case of homogeneous rock, to the fieldstress
state relative to the rockmass strength. The respective
relationships are given by theexpressions
TABLE D-1Joint alteration number (1,).
Description of Gouge
Joint Alteration Numberfor Joint Separation of
Less than1.0 mm" mmb
Larger than5.0 mm
Tightly healed, hard, non-softeningimpernieable tilling 0.75
Unaltered joint walls, surface staining only 1.0Slightly
altered, non-softening, non-cohesive rock mineral or
crushed rock tilling2.0 4.0 6.0
Non-softening, slightly clayey non-cohesive tilling 3.0 6.0'
10.0'Non-softening strongly over-consolidated clay mineral
tilling,
with or without crushed rock3.0' 6. 10.0
Softening or low friction clay mineral coatings and small
quantitiesof swelling clays
4.0 8.0' 13.0'
Softening moderately over-consolidated clay mineral tilling,
withor without crushed rock
4.0' 8.0 13.0
Shattered or micro-shattered (swelling) clay gouge, with or
without crushed rock5.0' 10.0 18.0
Joint walls effectively in contact.b Joint walls come into
contact before 100 mm shear.
Joint walls do not come into contact at all upon shear. Also
applies when crushed rock is present in clay gouge and there is no
rock wall contact.
Figures added to original data to complete sequences.
' Steffen, Robertson, and Kirsten, P. 0. Box 8856, Johannesburg
2000, South Africa.
-
/00
/0
5
/.0
0./
i`o
Z0
c r4
o R 2Cl...se 2E ..?:_.0 1' 0-) S ;
Contours,..,/-- for P h /
F4
.
J w z' 1,0
of SRFQ n .-1.0
Pc Pe.00e.0%
ei)11,7CIP0 / 6.4
/41.,,Z,o (ROD/JII)(JrAtIksq
/Jo)
,
,../0VO
P13Ltvo.3-.0
0,,,099,,, , .C15'
TN.,n5'*-
E
?iii
itiC>....-:
1$:60.
. 48.- 0C C elle 29rErta)0 .7,0 kE o
Contoursf n r n in
increasingloosening -.1,-4.--e-burstingpotential
r'ili)of SRF
---1 n \
increasing
potential-
1,0( H / UCS)
FIG. D-1Graphs of QhIQ, versus H/UCS for range of
(RQD1l)(.1,11).
SRF = 1.809 Q -0329SRFh = 0.244 K 0346 (H/UCS)' 722 +
0.176(UCS/H)' 413
whereQ = (RQD/J)(J,JJ)(J/SRF),K maximum-to-minimum principal
field stress ratio,H = head of rock corresponding to maximum
principal field stress, m, and
UCS = unconfined compressive strength of rock, MPa.
The corresponding rockmass indices are given by the
expressions
Qn = I ( RQD/J,,)(Jr/.1 )(4)1 .809)I"' - 329)Q, =
(RQD/J)(,/,/,/)1,//0.244 e346(H/UCS)1 322 + 0.176(UCS/H) 1 4 I
The ruling value for Q may be determined by calculation as the
minimum of these twoexpressions. The corresponding value for SRF
may also be determined. In the event of this valuecorresponding to
SRF h it may further be established which of the two terms in the
expression forSRFh is the larger. These two terms represent
respectively the stress and structurally controlledbehavior of
competent rock.
The proposed procedure for determining SRF replaces that given
by Barton et al IDI] and is
HEAVYROCK BURSTSQUEEZINGSWELLING
MILDROCK BURSTSOUSE ZINGSWELLING
MINIMALFRACTURINGSQUEEZINGSWELLING
NOFRACTURINGSQUEEZINGSWELLING
IIIIII11111.1n1111111111
III1111111illi IlaN=1 II11111I
:"." 4' Z
,...:........MMENEWMEM
roil1 IEEE".IIIIMIIIIMnleMMMINMENOIN=MO
11111111111=11011LBINIIMINIMINIIMICM III II IIIIMMOMMIl'
Illiek..III
I IIii
--4111alimmonNMI.
Mn10=113= e6,
__ mo IN mmiumnImlommmannsm211112INIMEIMIIELEINIMIoG END:g22/
TEFCAGHI(after Barton et al (1)) 111MMIIIII=11111+ BA RTON TABLE
3(al
...J.., to Kirsten)lCi CECIL(after Barton et d Mt
1lIllibrAlill 11111111111irEormdleiliiiikin_... fratOMIMMii.
imam,
I II Stress Reduction WollSRF .8090 0,329(applicable to
incompetent rock)
=linSELF SUPPORTED CASES(otter Barton et al (I))ii X R L m
1111POOR POOR
^
00 V RYPOOR GOODliMill XCGOOD GOOD4 40 /00 400 1000
ROCK MASS QUALITY 0
FIG. D-2Plot of stress reduction factor versus rock mass
quality.
illustrated graphically in Fig. D-1 for K = 1.0 and J,. = 1.0.
The significance of the proposedalternative procedure is
illustrated in Fig. D-2. The straight-line graph corresponding to
SRF =1.809 Q-"9
represents the squeezing behavior associated with shear or
weakness zones inincompetent ground. As such it represents the
lower bound values for SRF. The domain abovethe graph represents
the loosening behavior of competent rock, to which the expression
for SRFin terms of K, H, and UCS applies.
The various case histories shown plotted in Fig. D-2 confirm the
proposed alternative determinationof SRF. The correspondence
between Barton's and Terzaghi's implied relationships between
SRFand Q for incompetent ground is remarkable. The fact that
Terzaghi worked in tunnels which weregenerally of a squeezing
nature and not self-supporting is confirmed in Fig. D-2. It is also
evidentthat tunnels in ground in which the loosening behavior is
governed by shear or weakness zonesor by compressible gouge on the
joints are generally not self-supporting.
The behavioral characteristics of the ground surrounding a
tunnel are detailed in Table D-2.The proposed direct calculation of
SRF from a limited number of geological parameters is the
only quantification of the loosening potential or stress
relaxational behavior of tunnelled groundavailable to date. Barton
and his co-authors should be congratulated for this contribution,
althoughthey were not explicitly aware of the quantitative
implications of their findings.
0.00/ 0.006 05 4/00
/Op
Ca
0 /,
0/
0,0,0/
DISCUSSION ON Q-SYSTEM 8786 ROCK CLASSIFICATION SYSTEMS FOR
ENGINEERING PURPOSES
-
88 ROCK CLASSIFICATION SYSTEMS FOR ENGINEERING PURPOSES
TABLE D-2Behavioral characteristics of tunnelled ground in terms
of SRF (J,, = K = 1.0)Stress
ReductionControlling Factor
Criterion (SRF)
ProbableRock Mass
Quality(0
Excavation-InducedGround
Condition
Excavation-Nature ofInduced Ground-Ground Supporting
Displacement Action
Larger than 10 0.001-25/ Heavy rock-burst- Indefinite creep
Visco-plastic
Stress in re-spect ofcompetentrock/Shearor weakness
5-10
0.001-0.006
0.006-50/0.006-0.05
ing, squeezing,or swelling
Mild rock-bursting,squeezing, orswelling
arch
Terminating creep Partially plasticarch
zones in re-spect of in-competentground
2.5-5 0.05-100/0.05-0.4
Minimal fracturing,squeezing, orswelling
Large elastic de- Highly stressedformation elastic arch
0.4-2.5 0.4-640/ No facturing, Moderate elastic
Moderately0.4-100 squeezing, or
swellingdeformation stressed elastic
arch0.4I 6.0-640 No fracturing,
squeezing, orswelling
Limited elastic
Lowly stressed
deformation elastic arch
Geologicalstructure in
1-2.5 0.4-250 No fracturing,squeezing, orswelling
Random block Lowly stressed
displacement
elastic arch
respect ofcompetent
2.5-5 0.05-10() No fracturing,squeezing, or
Random block Keystone arching
displacement by mechanicalrock swelling interlocking of
blocksLarger than 5 0.001-50 No fracturing,
squeezing, orswelling
Block fall outs or
Non-arching in
runs of ground completelyloosenedground
The two probable ranges of Q correspond to the given controlling
criteria.
Discussion References[DI] Barton, N., Lien, R., and Lunde, J.,
"Engineering Classification of Rock Masses for the Design of
Tunnel Support," Rock Mechanics, Vol. 6, No. 4, 1974, pp.
189-236.
Page 1Page 2Page 3Page 4Page 5Page 6Page 7Page 8Page 9Page
10Page 11Page 12Page 13Page 14Page 15Page 16Page 17Page 18Page
19Page 20Page 21Page 22Page 23Page 24Page 25Page 26Page 27Page
28Page 29Page 30Page 31Page 32Page 33Page 34Page 35Page 36Page
37Page 38Page 39Page 40Page 41Page 42Page 43Page 44Page 45Page
46Page 47Page 48Page 49Page 50Page 51Page 52Page 53Page 54Page
55Page 56Page 57Page 58Page 59Page 60Page 61Page 62Page 63Page
64Page 65Page 66Page 67Page 68Page 69Class 1.PDFPage 1Page 2Page
3Page 4Page 5Page 6Page 7Page 8Page 9Page 10Page 11Page 12Page
13Page 14Page 15Page 16Page 17Page 18Page 19