You are on page 1of 5

Review: GSI and HoekBrown Procedure

The presence of geological structures within a rock mass (joints, shears, etc.), requires that
consideration be given to the combined influence of intact rock blocks and discontinuities when
calculating the rock mass' response to tunnelling. This is in contrast to most testing carried out during
geotechnical investigations, which is usually restricted to laboratory testing of intact rock samples. In
situ tests are often prohibitively expensive and associated with their own issues of reliability,
repeatability and scale. This has led to the development of a number of systems that link rock mass
propertiestoobservationsoftherockmasscharacteristics.Amongthese,theGeologicalStrengthIndex
(GSI) coupled with the HoekBrown failure criterion has become one of the industry standards for
estimatingrockmasspropertiesoninternationaltunnellingprojects.
The HoekBrown failure criterion is an empirical formulation for estimating the confinement
strengthrelationshipofarockmass.It'snonlinearformdistinguishesitfromthelinearMohrCoulomb
failurecriterion(Fig.1a).Thecriterionwasoriginallyconceivedbasedonexperienceswithbrittlefailure
in hard rock and developed to assume that rock mass failure was controlled by jointing but with no
preferred failure directions (Hoek & Brown 1980); i.e. the rock mass responds as an equivalent
continuum. Later revisions saw the HoekBrown failure criterion coupled with Bieniawski's Rock Mass
Rating(RMR)systemasameanstoscalelaboratoryintactrockpropertiestothoseforthejointedrock
mass(Hoek&Brown1988),andimprovementstobetteraccountforpoorerqualityrockmasses(Hoek
et al. 1992). Further experience with the latter found that it was difficult to apply RMR to very poor
quality rock masses. This led to the introduction of the Geological Strength Index (GSI) as a
characterization system based more heavily on fundamental geological observations and less on
'numbers' (Hoek et al. 1995). The most uptodate version, Hoek et al. (2002), represents a major re
examination of the entire HoekBrown criterion and includes new derivations for the relationships
betweenthedifferentinputparametersandGSI.
ThegeneralizedformofthenonlinearHoekBrownfailurecriterionis:
o
1
i
= o
3
i
+ o
ucs
[m
b
c
3
|
c
ci
+s
u
(1)
where o
1
i
and o
3
i
are the major and minor effective principal stresses at failure, o
c
is the uniaxial
compressivestrengthoftheintactrock,andm
b
,sandaarematerialconstantsfortherockmass.These
constantsaredeterminedfortherockmassusingGSIasperHoeketal.(2002):
m
b
= m

cxp [
uSI-100
28-14
(2)
s = cxp [
uSI-100
9-3
(3)
o =
1
2
+
1
6
(c
-uSI15
- c
-203
) (4)
From above, m

is a curve fitting parameter derived from triaxial testing of intact rock. The parameter
m
b
is therefore a reduced value of the intact rock value m

, which accounts for the strength reducing


effectsoftherockmassconditionsdefinedbyGSI.Strengthreductionfortheparameterssandafollow
accordingly (Fig. 1b). D is a disturbance factor that can account for blast damage and stress relaxation,
withvaluesrangingfrom0forundisturbedconditionsto1forverydisturbedrockmasses.

Figure1.a)ComparisonoflinearMohrCoulombandnonlinearHoekBrownfailureenvelopesplottedagainsttriaxialtestdata
forintactrock;b)ScalingofHoekBrownfailureenvelopeforintactrocktothatforrockmassstrength.

GSIisestimatedinthefieldfromthechartofMarinosetal.(2005);seeFigure2.Intheabsenceof
GSI values or asa secondary check, GSI can be converted from RMR
89
values (Bieniawski 1989), as was
thepracticeinearlyapplicationsoftheGSIsystemusingtherelationship:
0SI = RHR
89
i
- S(forRHR
89
i
>23) (5)
where RHR
89
i
is a modified version of RMR
89
in which the groundwater rating is set to 15 and the
adjustment for joint orientation is set to zero. These adjustments avoid double counting the effects of
groundwater(aneffectivestressparameterinthenumericalanalysis)andjointorientation(treatedasa
specific input for structural analysis) when deriving rock mass properties to be used in numerical
analyses.Forverypoorqualityrockmasses(RHR
89
i
<23),theconversionwasfoundtobeunreliableand
analternativerelationshipusingBartonetal.'s(1974)Qparameterwasrecommendedinstead:
0SI = 9Iog
c

i
+44(forRHR
89
i
<23) (6)
Similar to the adjustment for RHR
89
i
, Q' represents the modified Q value where the J
w
/SRF quotient is
dropped from the calculation (the influences of groundwater and insitu stresses being explicitly
accountedforintheeffectivestressnumericalanalysesbeingemployed).
Since Eqns. (5) and (6) were first published, Hoek (2007) has found that these correlations have
proventobeunreliableandhenowrecommendsthatGSIshouldbeestimateddirectlybymeansofthe
charts published on its use (Fig. 2). Experience has shown that most geologists and engineering
geologistsarecomfortablewiththedescriptiveandlargelyqualitativenatureoftheGSItablesandhave
littledifficultyinarrivingatanestimatedvalue.
ItshouldalsobeemphasizedthattheGSIsystemisnotareplacementfortheRMRorQsystemsas
it has no rock mass reinforcement or support design capability its only function is the estimation of
rockmassproperties(Marinosetal.2005).Furthermore,Hoeketal.(2005)emphasizethattheGSIwas
not developed for the purpose of specifying anticipated or changing tunnelling conditions and strongly
oppose its use for this purpose. Thus, as applied in the Preliminary Design report, it is recommended
thatRMR
89
beusedasaframeworkforclassifyingandcommunicating changingground conditionsand
changingsupportrequirementsduringtunnelconstruction.

Figure2.GeologicalStrengthIndex(GSI)lookupchartforjointedrockmasses(afterMarinos&Hoek2000).

A further consideration is that most geotechnical design calculations are written for the Mohr
Coulomb failure criterion, it is often necessary to calculate equivalent rock mass friction angles and
cohesive strengths from the HoekBrown parameters. Moreover, most practitioners have more
experience,andthereforeanintuitivefeelingforthephysicalmeaningsofcohesionandfriction,which
isnotthe caseform
b
,sand a.Intermsofequivalencies,the parameter m
b
isrelated tothefrictional
strengthoftherockmass,ands,whichisameasureofhowfracturedtherockmassis,isrelatedtothe
rock mass cohesion. These are only descriptive relationships, however. Where MohrCoulomb
parametersarerequired,thefittingofthelinearMohrCoulombenvelopetothenonlinearHoekBrown
enveloperesultsinthefollowingequationsforfriction
i
andcohesivestrengthc
i
:

i
= sin
-1
_
6um
b
(s+m
b
c
3n
|
)
c-1
2(1+u)(2+u)+6um
b
(s+m
b
c
3n
|
)
c-1
_ (7)
c
i
=
c
ci
|(1+2u)s+(1-u)m
b
c
3n
|
](s+m
b
c
3n
|
)
c-1
(1+u)(2+u)_1+_
6cm
b
(s+m
b
o
3n
|
)
c-1
(1+c)(2+c)
_
(8)
whereo
3n
i
= o
3mux
i
o
c
.Notethatthevalueofo
3mux
i
representstheupperlimitofconfiningstressover
which the relationship between the HoekBrown and MohrCoulomb failure envelopes is considered.
Fordeeptunnels,o
3mux
i
canbecalculatedfromtheempiricalrelationship(Hoeketal.2002):
c
3mcx
|
c
cm
|
= u.47 [
c
cm
|
yH

-0.94
(9)
wherey istheunitweightoftherockmass,Histhedepthofthetunnelbelowsurface,ando
cm
i
isthe
"global"rockmassstrengthforthestressrangeo
t
< o
3
i
< o
c
givenby:
o
cm
i
= o
c
_
|m
b
+4s-u(m
b
-8s)][
m
b
4+s

c-1
2(1+u)(2+u)
_ (10)
Incaseswherethehorizontalstressishigherthantheverticalstress,thehorizontalstressvalueshould
beusedinplaceofyEinEqn.(9).
These procedures form the basis for the Rocscience software package ROCLAB, which can be
downloadedforfreefrom:http://www.rocscience.com/products/RocLab.asp.
Incarryingoutthesecalculations,itmustbeemphasizedthatthequantitativeconversionofHoek
Brown to MohrCoulomb parameters is done by fitting an average linear relationship to the nonlinear
HoekBrownenvelopeforarangeofminorprincipalstressvaluesdefinedby
t
<
3
<o
3mux
i
(Hoeket
al.2002).Notethatthevalueofo
3mux
i
,theupperlimitofconfiningstressoverwhichtherelationship
between the HoekBrown and MohrCoulomb criteria is considered, has to be determined for each
individualcase(Fig.3).
WheregeotechnicaldesignsoftwareacceptsHoekBrowninputdirectly,itispreferabletousethis
inputratherthanestimatesoftheMohrCoulombparameterscandextrapolatedfromthenonlinear
HoekBrown failure envelope. This eliminates the uncertainty associated with the conversion that is
demonstratedinFigure3.

Figure3.FittingoflinearMohrCoulombfailureenvelopesalongtwodifferentstressrangesofanonlinearHoekBrownfailure
envelope.Notethechangeincohesionandfrictionanglevaluesforthetwodifferentstressrangesspecified.
References

Barton,N.,Lien,R.&Lunde,J.(1974).Engineeringclassificationofrockmassesforthedesignoftunnel
support.RockMechanics6(4):189236.
Bieniawski, Z.T. (1989). Engineering Rock Mass Classifications: A Complete Manual for Engineers and
GeologistsinMining,Civil,andPetroleumEngineering.NewYork:Wiley,272pp.
Hoek,E.(2007).PracticalRockEngineering.Toronto:Rocscience,ebook.
Hoek, E. & Brown, E.T. (1980). Underground Excavations in Rock. London: Institution of Mining and
Metallurgy,527pp.
Hoek, E. & Brown, E.T. (1988). The HoekBrown failure criterion a 1988 update. In J.H. Curran (ed.),
Proc.15thCanadianRockMech.Symp.,Toronto.Toronto:UniversityofToronto,pp.3138.
Hoek, E., Kaiser, P.K. & Bawden, W.F. (1995). Support of Underground Excavations in Hard Rock.
Rotterdam:Balkema,215pp.
Hoek, E., Wood, D. & Shah, S. (1992). A modified HoekBrown criterion for jointed rock masses. In J.
Hudson (ed.), Rock Characterization: ISRM Symp, Eurock 92, Chester, UK. London: Thomas
Telford,pp.209213.
Marinos, P. & Hoek, E. (2000). GSI: A geologically friendly tool for rock mass strength estimation. In
GeoEng2000,Melbourne.Lancaster:TechnomicPublishingCompany,CDROM.
Marinos, V, Marinos, P. & Hoek, E. (2005). The geological strength index: applications and limitations.
BulletinofEngineeringGeologyandtheEnvironment64(1):5565.
Rocscience(2007).RocLab,version1.031.Toronto:Rocscience,Inc.

You might also like