You are on page 1of 18

Estimating In Situ Deformation of Rock Masses Using

a Hardening Parameter and RQD


Ashutosh Trivedi1

Abstract: This study explores the utility of a pressure-dependent empirical hardening parameter in estimating deformation of rock masses. The
hardening parameter has drawn poor attention in classical rock mechanics. Review of previous literature and observation of test results show
that hardening is highly dependent on confining pressures and the plasticity already experienced, which modify joint parameters by mutual
interaction. This paper provides an insight into the published test data that reflects the dependencies of the in situ deformation modulus on in-
cremental joint parameters and the isotropic pressure ratio. There exists an empirical relation between modulus ratio and modified joint factor
derived from number of joints, roughness, joint inclination, and gouge parameters. The author has interpreted the modulus ratio of jointed rock
based on the modified joint factor. The author applied a hardening parameter obtained from the initial condition of stresses and that of the joint
on the modified joint factor to get a modulus ratio. Using a correlation of the modified joint factor and the rock quality designation (RQD), the
model successfully shows that operating hardening parameter on RQD envelops the published data of modulus ratio versus RQD. The main aim
of this paper is to show that the present model is valid for in situ deformation of the strain-hardening rock mass in structural foundation, mines,
and excavation design. DOI: 10.1061/(ASCE)GM.1943-5622.0000215. © 2013 American Society of Civil Engineers.
CE Database subject headings: Rock masses; Parameters; Joints; Deformation.
Author keywords: Modulus ratio; Rock mass; Hardening parameter; Modified joint factor; Rock quality designation (RQD).

Introduction The fracture stiffness is highly nonlinear even at low confine-


ment and deformation. Few hypoplastic models (Wu and Bauer
The interpretation of deformation properties of rock masses is highly 1993) indicate the potential to evaluate deformation of the rock
intricate because of the presence of discontinuities of variable mass without assuming a linear elastic or perfectly plastic behavior
compressibility along the joint surface itself and among the volume (Kolymbas 2007). Therefore, the already existing empirical rela-
of infill present in the joint. The compressibility depends on the tionship of modulus ratio and rock quality designation (RQD)
mean effective confining pressure, which is negligible at an open (Deere et al. 1967; Coon and Merritt 1970; Zhang and Einstein 2004)
joint to a peak pressure existing at the point of stress concentration. may improve on incorporating nonlinearity owing to confinement.
The peak pressure adjusts itself on localized deformations. This The initial mean confining pressure (p 5 I1 =3) sets in a progressive
process of adjustment sets a distinct path of initial pressure to shear stiffness with incremental storage of strain energy up to a limiting
failure. Even during the process of unconfined loading of the jointed value related to uniaxial compressive strength sr of the intact rock. A
rock mass, the joint experiences varied levels of confinement in portion of this stored strain energy dissipates at a rate depending on
relation to the peak uniaxial compression. It is recognized that mean initial mean effective confining pressure set to deform the rock mass.
initial confinement in relation to the joint condition and failure stress The uniaxial compressive strength conceals an equivalent initial
tend to rationalize a path-dependent deformation, as also proposed confining pressure, often characterized by the morphology of the
in the failure theories incorporating stress invariants (I1 , J2 , J3 ). rock masses normally considered in various rock-mass classification
Physically, I1 is mean confining pressure, J2 represent magnitude of systems.
shear stress, and J3 determines its direction (Yu 2006). While experiencing deformations, the rock masses undergo
Theoretically, during the deformation of rock masses, the con- isotropic and kinematic hardening. This invokes Ziegler’s (1959)
hardening rule, where the yield surface transforms in the stress space.
servative and dissipative components of the work input occur si-
It expands or contracts, changes its shape, and translates on loading.
multaneously. The total deformation may consist of the deformation
In their seminal work, Shield and Ziegler (1958) described the
of discontinuities, especially at low-stress conditions (Bandis et al.
behavior of a rigid work-hardening material. They envisioned an
1983). The equivalent continuum models (Kulhawy 1978; Gerrard
initial yield condition characterized by a state of stress first to set
1982) show the effect of the fracture stiffness on the rock-mass de-
a plastic flow. Next to come is a flow rule connecting the plastic
formation modulus. A ratio of deformation modulus of the jointed strain increment with stress and the stress increment. In addition,
rock mass Em to that of intact rocks Er is defined as the modulus ratio. there is a hardening rule that specifies the modification of the yield
1
condition in the course of plastic flow. In extending these concepts to
Professor, Dept. of Civil Engineering, Delhi Technological Univ. and rock masses, it is observed that with increasing initial isotropic
Univ. of Delhi, Delhi 110042, India. E-mail: prof.trivedi@yahoo.com
pressure ratio pi =sr , the joints get hardened at a nonlinear rate.
Note. This manuscript was submitted on March 15, 2011; approved on
March 7, 2012; published online on March 12, 2012. Discussion period Therefore, a nonlinear parameter Ch is recognized among the values
€I and bI such that the hardening parameter implies €I . Ch . bI. In
open until January 1, 2014; separate discussions must be submitted for
individual papers. This paper is part of the International Journal of Geo- other words, if Ch → €I, the material undergoes deformation with an
mechanics, Vol. 13, No. 4, August 1, 2013. ©ASCE, ISSN 1532-3641/ effect on confinement, shape, and size, and a very strong strain
2013/4-348–364/$25.00. gradient exist. Finally, if Ch → bI, the deformations are independent

348 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


of observed boundary conditions. As such, well within the limit of

Note: RMR 5 R1 1 R2 1 R3 1 R4 1 R5 1 R6; Emr 5 ½0:0028RMR2 1 0:9 exp ðRMR=22:82Þ=100 (Nicholson and Bieniawski 1990); Q 5 ½ðRQD=Jn ÞðJr =Ja ÞðJw =SRFÞ; Emr 5 10ðQsr =100Þ1=3 (Barton 1983);
deformations
Application

Tunnels,
inequalities, the experimental and realistic values of the hardening

(12)

-do-
-do-
parameter observed for Ch depends on effective isotropic pressure.
Researchers have defined plastic parameters, such as hardening
or softening, in a number of ways in engineering applications. There

Dilatancy
is no wide agreement among the researchers on a common possible

(11)



composition of the plastic parameter (Varas et al. 2005; Alejano and
Alonso 2005). However, there are two common components of this
parameter. One of them is a function of internal variables, and the

10–1
SRF
(10)


other is incremental (Vermeer and De Borst 1984). Interestingly,
pressure-dependent incremental hardening is one of the intrinsic

(R6 5 23020)
parameters of all granular materials, namely, powders, silts, ashes

Orientation
(Trivedi and Sud 2004), and even jointed rocks. In this study, the

0–90
(nb )
(9)



empirical hardening parameter for jointed rocks Ch is obtained
numerically for input of different initial mean effective confining

GSI 5 9 log ½ðRQD=Jn ÞðJr =Ja Þ 1 44 (from Q system); GSI 5 10 1 R1 1 R2 1 R3 1 R4 (from RMR); Emr 5 ðsr =100Þ1=2 10ðGSI210Þ=4 (Hoek and Brown 1997).
pressure ratios pi =sr set into the rock mass based on the criteria of
Johnston and Chiu (1984) and Johnston (1985). It tends to reach

Roughness (Jr )
a failure stress incorporated in the strength ratio smr obtained from

0.5–4
0.5–4
uniaxial and triaxial test data of jointed rocks determined by the

(8)

direction of shear stress J3 .

Scope of Study

for 10-m tunnel


1–1,000 L/min
Joints represent discontinuities at the boundary of the intact and

(R5 5 1620)
Inflow (Jk )
continuous material. The discontinuities may exist with fragments of

1–0.05
(7)


parent rock material often along with externally deposited gouges in
the joints, which may control the deformability of the rock mass.
During the process of compression, the rock mass closes the discon-
tinuities and becomes hardened. Therefore, a laboratory-controlled
hardening parameter is identified. Because laboratory test data on
small specimens are often insufficient to predict deformation of

(R4 5 23520) (R5 5 1620)


(Jw 2 u=s1 )
a rock mass, its relation to in situ testing is necessary. It is easier to Joint water
pressure

0–0.6
apply an empirical hardening parameter on the field test data of

(6)



RQD to find the in situ deformation modulus.
Table 1. Modulus Ratio from Various Rock-Mass Characterizations Systems (RMR, Q, GSI)

Hoek and Brown (1997) and Hoek and Diederichs (2006) con-
sidered the uniaxial compressive strength of intact rock and the
geological strength index (GSI) to propose a relation for rock-mass
(t=ta ) (Ja )
thickness

0–6 mm

0.75–20
0.75–20
Gouge

deformation, later incorporating a disturbance factor. The rock-mass


(5)

rating (RMR) of Bieniawski (1978), Serafim and Pereira (1983),


Nicholson and Bieniawski (1990), and Mitri et al. (1994) includes
the joint spacing and RQD to suggest a modulus ratio. Furthermore,
(number of joints/m)

RMR and GSI provide measures of qualitative assessment of rock


mass strength. The Q system (Barton et al. 1980; Barton 1983;
(R3 5 0225)
Joint spacing

0.01–10 m
0.01–10m

20–0.5

Barton 2002) considers six parameters inclusive of RQD, number


(4)

of joints, roughness, effect of joint filling and condition of joints,


groundwater condition, and stress-reduction factor to evaluate mod-
ulus ratio. Fossum (1985) and Kulhawy (1978) related modulus ratio
with mean discontinuity spacing. Taking a cue from the relationship
of discontinuity spacing and RQD proposed by Priest and Hudson
(R1 5 0216) (R2 5 0225)

(1976), Zhang (2010) presented an analytical solution incorporating


0–100

0–100
0–100
RQD
(3)

compiled field observations of deformation modulus. These systems


do not directly consider joint hardening.
The equivalent-continuum approach treats the jointed rock mass
as an equivalent continuum with deformability that reflects the de-
0–300 MPa
UCS (sr )

formation properties of both the intact rock and the discontinuities


(2)


(Zhang 2010). Arora (1987) considered a few significant rock joint


characteristics in a joint factor to propose a relationship for modulus
ratio. It should be noted that all methods need to use some of the
characterization

deformation properties of intact rock and discontinuities obtained


Q (0.001–400)
RMR (0–100)

GSI (0–100)

via laboratory or in situ tests. The comparison of modulus ratios


Rock-mass

evaluated by various investigators is presented in Tables 1 and 2.


(scale)

This study shows that the rock-mass deformation criteria should


(1)

be improved to have the ability to predict rock-mass deformation

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013 / 349


under different joint conditions and unifying the anisotropy, effect

— ap 5 l=C Laboratory and field

Note: Em 5 Er exp ðCh × Jf Þ; Jf 5 Jn =ðnb rÞ (Ramamurthy and Arora 1994); and Ch 5 20:0115 (Ramamurthy and Arora 1994); Em 5 Er exp ðCh × JfgÞ (present work); Ch 5 20:04ðpi =sr Þ21 for ap 5 20:005;
Ch 5 20:02ðpi =sr Þ21:25 for ap 5 20:009; Ch 5 20:001ðpi =sr Þ22 for ap 5 20:0125; Jfg 5 a exp ðb × RQDÞ (present work); Emr 5 exp ½a × Ch expðb × RQDÞ (present work); Emr 5 m ln ðRQDÞ 1 n (present
of orientation, discontinuities, presence of gouges, and especially

Application
hardening due to confinement and the shear stress direction from

(12)
Laboratory
initial conditions to failure.
The previous studies have left the scope for estimating the rock-

— ap 5 l=C In situ
mass hardening parameter as one of the unresolved issues. There-
fore, one of the aims of this study is to obtain a hardening parameter
Inflow (Jk ) Roughness (Jr ) Orientation (nb ) SRF Dilatancy
as a function of stress invariant. Second, the modulus ratio of jointed
(11)

rock has yet to relate a readily estimated hardening parameter to the
experimentally obtained quantities found by direct measurements,
such as modified joint factor and RQD. A modified joint factor
(10)

considers intrinsic rock joint characteristics, namely, discontinuities,


roughness, orientation, condition of joints and gouge thickness,
0.1–1, f ðnb Þ
0.1–1, f ðnb Þ

compaction, stress, and groundwater conditions in relation to the


f ðnb Þd

modulus ratio.
(9)

The author applied the strength criteria of Johnston and Chiu


(1984), using the joint-mapping Ramamurthy-Arora criteria
(RAC) of Ramamurthy and Arora (1994) and Trivedi (2010), and
invoked application of Ziegler’s (1959) hardening rule to obtain
r 5 f ðsm Þ
r 5 tan f

a hardening parameter. The range of the hardening parameter was


f ðsm Þd
(8)

examined and applied by the author to a large experimental (Emr


versus Jfg ) and in situ (Emr versus RQD) data set to show wide-
ranging application of the present proposal to evaluate rock-mass
deformation.
—e

—e
(7)

Definition of the Problem and Applications

A rock mass consists of discontinuities in the form of joints, fissures,


Joint water pressure

cracks, and fractures with infill that undergoes significant deformation


(Jw 2 u=s1 )

compared with intact rocks. In routine engineering applications, the


—e

—e

relation of rock-mass deformations to joint condition and hardening


(6)

receives very little attention, which is so essentially because of the fact


that many problems of rock mechanics are solved, avoiding evalua-
Table 2. Modulus Ratio from Various Rock-Mass Characterizations Systems (RAC, Jfg , RQD)

tion of path-dependent deformations, and second because of inherent


difficulties in estimating hardening. Early investigators, namely,
Bieniawski (1978) and Serafim and Pereira (1983), admited that
Joint spacing in m Gouge thickness

various types of in situ tests for deformation modulus can produce


(t=ta ) (Ja )

f ðt=ta Þd
f ðt=ta Þ

different results because of different loading and boundary conditions.


(5)

The loading and boundary conditions indirectly refer to the path-


dependent deformations as taken care of in the present model of
hardening. The aim of this study is to advance a consistent model for
jointed rocks that connects the final strength with a hardening pa-
UCS (sr ) RQD (number of joints/m)

rameter used for rock masses with changing discontinuities and joint
13–500/m

conditions. It has potential application in rock excavations, tunneling,


f ðJn Þd
f ðJn Þ
(4)

and foundations in rock masses. In view of the present trends in


modeling, the purpose of this study is not to obtain highly accurate
As per the depth of the water table relative to the joint.

values of deformation but rather to focus on the issue that considers the
effects of a hardening parameter on theoretical and practical engi-
neering problems. A large ensemble of studies on this topic reveals
(3)

that the hardening parameter and relative joint conditions are rarely
As per the relationship of Jfg and RQD.

taken in to consideration. Often, when they are considered, the ap-


proach is poorly drawn, consisting of a thumb rule (Ch 5 constant).
smr
smr
(2)

Nevertheless, this rule does not necessarily represent deformation


Ramamurthy and Arora (1994).

behavior of rock masses.


The results of model studies on rocks and modified masses il-
lustrate the potential of varied deformation modes owing to the
Modified joint factorb,c
RAC (Jf 5 132500)

highly intricate internal stress distribution within a jointed rock


mass. With progress in loading, there is a mutual adjustment of the
Trivedi (2010).
(Jfg 5 021,000)
characterization

micromechanical deformation parameters with the mean effective


RQDc (0–100)

Present work.

confining pressure. So far the deformation behavior of jointed rock


Rock mass

masses has been quantified as a function of joint orientation, joint


work).
(scale)

size, frequency, roughness, and waviness of the joints. There are


(1)

inherent difficulties in mapping these parameters for rock masses.


b

d
a

350 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


Fig. 1. Model deformation of jointed rocks with joint modification and isotropic pressure

Therefore, combined effect in terms of equivalent values adopted for 3p


a modified joint factor and initial confinement may capture rea- smr ¼ (1e)
sr
sonably well the deformation of a jointed rock mass right from a state
of low confinement to a heavily confined state using iterative inputs In an unconfined state, the strength ratio
of the resulting strength. A concept model for the deformation of
jointed rocks is shown in Fig. (1). ðsm Þ=3
smr ¼ (1f )
ðsr Þ=3
Preliminary Definitions sm
s¼ (1g)
sr
Strength and Modulus Ratio
In prefailure triaxial conditions, the modulus ratio is defined as
The strength and modulus of jointed rock is often evaluated in terms
of strength ratio smr and modulus ratio Emr , respectively. The pri- Em
Emr ¼ (2a)
mary aim of finding the strength and modulus ratio relationship with Er
a modified joint factor is to readily obtain the strength and modulus
of jointed rock by conducting a single uniaxial compression test on where Er and Em 5 tangent moduli of intact rock and rock mass,
the parent rock mass. The strength ratio smr is defined as the ratio of respectively, considered at 50% of sr and sm , respectively.
the strength of a rock mass sm with respect to uniaxial compressive Similarly, in prefailure unconfined conditions, the modulus ratio
strength of a same-sized intact rock sample sr of the same parent is defined as
material. If s1m , s2m , and s3m are triaxial principal stresses in the
Em ðat s2m ¼ s3m ¼ 0Þ
rock mass and sr is the uniaxial compressive strength of the intact Emr ¼ (2b)
rock sample, then in the triaxial state, at failure, the strength ratio is Er
defined by
Similarly, mean confining pressure ratio and shear stress ratio are
defined as
s1m þ s2m þ s3m
p¼ (1a)
3 ðs1m þ s2m þ s3m Þ=3
pmr ¼ (3a)
sr
q ¼ ½ðs1m 2 s3m Þ (1b)
s1m 2 s3m
qmr ¼ (3b)
ðs þ s2m þ s3m Þ=3 sr
smr ¼ 1m (1c)
ðsr Þ=3
Joint Number, Joint Factor and Modified Joint Factor
p
smr ¼ (1d) The number of joints per unit run of a scan line is joint number Jn .
ðsr Þ=3 The joint number is a directional parameter. Its orientation and its

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013 / 351


tendency to slip, spread, and obstruct expansion affects the Rock Quality Designation
equivalent number of joints per unit volume of rock mass. A joint
Rock masses essentially consist of strong and weak formations that
factor captures potential engineering possibilities within a joint,
appear to be a continuous deposit. The drilling process tends to re-
namely, number of joints, orientation, and friction. The author
cover a continuous core of strong material. The already existing
considered the condition of joints, namely, the presence of gouges
weaknesses reappear as discontinuities at irregular frequencies,
and water pressure in the rock mass, together in a modified joint
which are the average number of joints appearing per unit length.
factor. The joint number connects RQD, the joint factor, and the
The RQD (Deere et al. 1967) is normally considered as the
modified joint factor. According to Arora (1987) and Ramamurthy
percentage of a core recovery of spacing length of greater than or
and Arora (1994), the joint factor is a significant joint-mapping
equal to 100 mm. The RQD also captures the orientation of the
parameter in relation to the strength ratio of the rock mass. It
discontinuity relative to the scan line, and thus the orientation pa-
considers number of joints, joint orientation, and friction in the
rameter is the result. This information is not always recorded in the
expression as
RQD data sheet. Engineering applications usually consider RQD as
Jn the percentage of the borehole core in a drill run consisting of intact
Jf ¼ (4a) lengths of rock greater than or equal to 100 mm, which is represented
nb r
numerically as
where Jn 5 number of joints per unit length in the direction of P
n  
loading (joints per meter length of the sample); nb 5 orientation RQD ¼ 100 Li =Ln % (7)
i¼1
parameter; and r 5 reference roughness parameter (5 1). It is
presented in nondimensional form as
where Li 5 lengths of individual pieces of core in a drill run having
lengths more than or equal to 100 mm; and Ln 5 total length of the
Lna
Jf ¼ Jn (4b) drill run. Based on Eq. (7), Priest and Hudson (1976) found the
nb r
estimate of RQD related to discontinuity spacing measurements
made on core as
where Lna 5 reference length (5 1 m).
On the basis of the experimental data of Yaji (1984), Arora RQD ¼ 100ð1 þ 0:1Jn Þeð20:1Jn Þ (8)
(1987), Trivedi (1990), Arora and Trivedi (1992), and Trivedi and
Arora (2007), Trivedi (2010) modified the joint factor to incorporate For the number of joints ranging among 6 to 16 per meter, an ap-
varied engineering possibilities among rock masses as proximate linear relation is interpreted from the work of Priest and
Hudson (1976) as
Jfg ¼ cg Jf (5)
RQD ¼ 110:4 þ 3:68Jn (9a)
where cg 5 modification factor for the condition of the joint, water
Based on the observations of Grenon and Hadjigeorgiou (2003) that
pressure, and the gouge; that is
neglecting volumetric joint count would lead to an erroneous
quantification of the discontinuity nature of the rock mass, Palmstorm
Jt (2005) suggested use of number of joints per unit volume Jv instead
cg ¼ Jdj (6)
gd Jw of number of joints per unit length of scan line Jn

where Jdj 5 correction for depth of joint (joint stress parameter); Jt 5 RQD ¼ 115 2 3:3Jv (9b)
correction for thickness of gouge in joint (gouge thickness pa-
rameter); Jw 5 correction for groundwater condition; and gd 5 Eq. (9b) is applicable for RQD for engineering applications for
correction factor depending on the compactness or relative density of volumetric joint count Jv between 4.5 and 35. Palmstrom (2005)
the gouge in the joint (5 1 for fully compacted joint fill). For clean, provided a mutual relationship for RQD, joint spacing s, joint number
compact joints, cg 5 1. These discontinuities may consist of frag- Jn , and volumetric joint count Jv . Sen and Eissa (1991) showed that
ments of the parent material to varied extents of thickness, density, RQD lowered with increasing difference between the lengths of the
and orientation. An observation of a rock-core recovery captured gd block side, i.e., joint spacing. In fact, changing block size (Samadhiya
in terms of designated discontinuity condition or precisely by et al. 2008) modulates stress intensity on discontinuities and thus on
assigning a packing density of fragments in the discontinuity. The the volume of joint material. Such an observation calls for an ad-
packing in the discontinuity tends to compact, dilate, or crush during justment in RQD versus joint properties, namely, spacing, volume,
the process of loading. friction, gouge material, groundwater, and internal pressure. The
The progressive compression of the discontinuities influences volumetric joint count tends to have an exponential relation with RQD
significantly the deformation modulus of the rock mass. The gran- as block size increases (Sen and Eissa 1991).
ular material in the joint’s rupture zone undergoes shear deformation The work of Sen and Eissa (1991), Grenon and Hadjigeorgiou
depending on the relative density of the gouge. The relative density (2003), and Palmstrom (2005) provided a basis to propose an ex-
RD is considered conventionally as a ratio of the difference of ponential relationship for a number of horizontal joints with friction
the natural-state void ratio en and the minimum void ratio emin from
the maximum void ratio emax of the gouge material [RD 5 ðemax 2 en Þ= Jnr ¼ a9 expðb9 × RQDÞ (10)
ðemax 2 emin Þ]. The effect of pore pressure u is discounted as dissipated;
therefore, the mean effective confining pressure p9 is equal to mean where a9 and b9 take a value based on the rock block characteristics,
confining pressure p [p 5 ðs1 1 s2 1 s3 Þ=3, where s1 , s2 , and discontinuity, and friction in relation to RQD. The volumetric effects
s3 5 principal stresses]. However, under the field condition of on strength and deformation of jointed rocks, namely, spacing,
groundwater, Jw is proposed as a linear function (of u=s1m ) in the volume, friction, gouge material, groundwater, and internal pres-
same manner as the Q system. sure, are considered in the modified joint factor Jfg .

352 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


Experimental Input and Numerical Formulation strain is indicative of the potential for changes in significant joint
parameters ( j 5 dɛ ps =dq), which is designated as j. The subscripts ij
This work is based on the generalizations derived from the theo- for variables in Eqs. (15a)–(15l) indicate their dependence on the
retical framework of the hardening rule by Ziegler (1959) and ex- instantaneous values of i and j. The superscripted variables indicate
perimental outputs of Johnston and Chiu (1984) and Johnston (1985) iterative steps proceeding to the plastic flow among joints. The sub-
on strength criteria, Arora (1987) and Trivedi (1990) on the em- scripted strain invariants, namely, ɛ v and ɛ s , represent volumetric
pirical relation of strength and modulus ratio based on significant and shear components, whereas the superscripts on strain invariants,
joint characteristics, and Chun et al. (2009) and Zhang (2010) on namely, ɛ pv and ɛ ps , indicate plastic range of behavior. These steps are
relating modulus ratio and RQD. represented by
Johnston and Chiu (1984) and Johnston (1985) proposed a re-
lationship in normalized form for intact rocks [based on the ex- i1 ¼ di þ i, i2 ¼ di þ i1 , i3 ¼ di þ i2 ,
perimental observations of Brook (1979), Johnston (1985), Aldrich
(1969), and Swansson and Brown (1971)] as i4 ¼ di þ i3 , ..., in ¼ di þ in21 (15a)

s1N ¼ ½ðM=BÞs3N þ 1B (11) j1 ¼ dj þ j, j2 ¼ dj þ j1 , j3 ¼ dj þ j2 ,


j4 ¼ dj þ j3 , ..., jn ¼ dj þ jn21 (15b)
where s1N 5 s1 =sr ; s3N 5 s3 =sr ; s1 and s3 5 stresses; sr 5
uniaxial compressive strength of the rock sample; and M and B 5
empirical rock constants. On simplification, we get where in and jn are arrived at by incrementally stepping to an iterative
failure condition based on Eqs. (13) and (14) in a number of steps
(1 to n). Prior to the arrival of a failure, the rock mass has a maxi-
sr 2M mum conservative component of the input energy. As such, the final
¼ (12)
st B conditions of failure leave a scope for an evaluation of the initial
conditions captured in steps by trial.
where st 5 tensile strength of the rock sample. The twin factors incremental initial change in isotropic condition
Griffith theory (Griffith 1924) predicts that uniaxial compressive di and incremental changes in joint parameters dj influence the
strength at crack extension is eight times the uniaxial tensile strength. expansion-contraction, translation, and shear shift in the p-q space.
The frictional forces in the microcrack network tend to modify the Therefore, shear stress q controlled by variable incremental initial
ratio as per Eq. (12) beyond the limits of prediction of Griffith theory. isotropic pressure conditions di and with variable incremental
The modification of empirical rock constants consequent to the changes in joint parameters dj may be represented by
changes in initial isotropic condition and joint conditions is captured  
by the modified joint factor. f ðq, p, Ch Þ ¼ q 2 Cij p2 þ Cij9p þ Cij99 (15c)
Based on Johnston and Chiu (1984) and Johnston (1985), the
following relationships for rock mass are used:
For a variety of rocks with microfracture networks and equivalent
2 jointed rocks in a confined state, the author correlated mean effective
B ¼ 1 2 xb ðlog sm Þ (13) confining pressure through a simple numerical code to predict the
modulus ratio of jointed rocks with progress in joint condition and to
M ¼ x þ xm ðlog sm Þ2 (14) provide values of the empirical parameter for jointed rocks Ch . The
concentrations of stresses during loading tend to modify the joint
The values of the constants in Eqs. (13) and (14) are compiled in factor, which, in turn, modifies the stress pattern. The progress of
Table 3. Consequent to the loading, the joint factor is modified as loading modifies the isotropic pressure, shear stress, and joint
a result of the changes taking place in the cracks, microcracks, and conditions. The empirical fitting parameter for jointed rocks Ch is
fracture network. The joint factor as per Eqs. (4)–(6) gets modified, obtained by estimating the deformation modulus of the jointed rock
and the stress redistribution takes place and continues alternatively with varying values of effective confining pressure for different
as long as deformation progresses. The changes in the isotropic initial isotropic pressure ratios pi =sr .
pressure dp induces plastic volume change dɛ pv . It bring changes It is observed that incremental initial pressure tends to increase
in potential of the isotropic pressure (i 5 dp=dɛ pv ). The changes in the hardening parameter, whereas increasing the value for joint
plastic shear strain dɛ ps with changes in shear stresses dq are as- parameters tends to decrease it. The joint parameters capture the
sociated with changes in the isotropic pressure dp. The plastic shear possibilities among joints through the modified joint factor in this
study. It is difficult to precisely indicate at which level either of the
components of the modified joint factor and the initial pressure
Table 3. Trial Parameter for Rock Mass (Adopted from Johnston 1985) offsets each other. It is evident that the joint factor characterizes the
postfailure tendency, whereas the initial confining conditions pro-
Group Rock type xb x xm
duce a condition toward conservation of rock. The presence of
A Limestone 0.0172 2.065 0.170a strain-restrictive regimes, namely, plane strain or triaxial condition,
B Mud stone 0.231b provides a direction to the deformation. The roots of Eq. (15c) in-
C Sandstone 0.270c dicate possible negative and positive values, suggesting application
D Granite 0.659d of Eqs. (11)–(15) beyond the limits of prediction of Griffith theory.
All types 0.276b This implies an increasing intercept Cij99 on axis q with mean con-
a
Brook (1979). fining pressure p. An increasing intercept Cij99 on axis q with mean
b
Johnston (1985). confining pressure p conserves the energy so that the rock mass
c
Aldrich (1969). appears strong on yield. The strong-on-yield behavior is shown by
d
Swansson and Brown (1971). dense granular material (Schofield and Wroth 1968). The coefficient

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013 / 353


Cij9 tends to translate the p-q plot in the positive axis p as per the Table 4. End Conditions for Strength Ratio of Rock Mass from Trial
Ziegler (1959) hardening hypothesis, where the yield surface Parameters
translates into the stress space. The shape of the p-q plot in Eq. (15c) End conditions Aa Cb ap 5 l=C l A=C b pi =sr
is controlled by the coefficient Cij . The second derivative of the yield
function in Eq. (15c) is operated to find the parameter Cij as Triaxial (min) 3 22.75 20.009 0.0250 21.09 1
25 20.005 20.6 ,1
Triaxial (max) 22.75 20.0139 0.0383 21.09 1
d2 q 25 20.0076 20.6 ,1
¼ 2Cij (15d)
dp2
Plane strain (min) 5 22.75 20.0100 0.0275 21.81 1
The shear stresses are represented in a simplified form as 25 20.0055 21 ,1
Plane strain (max) 22.75 20.0151 0.0416 21.81 1
    25 20.0083 21 ,1
dq ¼ dq=dɛ ps dɛ ps =dɛ pv dɛ pv =dp dp (15e) a
Bolton (1986).
b
For rock mass with granular fill; Trivedi (2010).
   
0dq ¼ dq=dɛ ps ðFÞ dɛ pv =dp dp (15f )
Table 5. Calculation of Hardening Parameter in Eq. (18)
  
0dq=dp ¼ ðFÞ dq=dɛ ps dɛ pv =dp (15g) Coefficient (k) and power (h) for
End conditions isotropic pressure ratio
If dp=dɛ pv 5 L is constant and dɛ ps =dq 5 j, ap k h
20.005 20.040 21.00
0L jðdq=dpÞ ¼ F (15h) 20.008 20.022 21.23
20.009 20.020 21.25
The derivative of Eq. (15h) with respect to changes in pressure is 20.0100 20.011 21.42
20.0125 20.001 22.00
 
0L j d2 q=dp2 ¼ ðdF=dpÞ (15i)

  A solution to the Eq. (15l) is brought about by author based on the


0L j 2Cij dp ¼ dF (15j) experimental observations [Eqs. (13) and (14)] of Johnston and Chiu
(1984) and Eq. (15c). Based on this analysis, it was found that
The dissipation of plastic energy is proportional to plastic shear  
strain and means that confining pressure acting on it. To introduce Ch ¼ f ðpi =sr Þ, ap (15m)
compatibility to the change in flow rule dF with changes in pressure
dp along with modification of joint parameters and hardening due to Ch contains very sensitive effects of plastic yielding, hardening, and
plastic flow, the constants in Eq. (15j) are substituted as thermal characteristics of the joint material. Several models for soil
and rock deformation consider the shear and volumetric strain alone
to serve as a hardening parameter (Lade and Kim 1995). There are
2LCij ¼ Ciju (15k) serious difficulties associated with measurements of actual plastic
volumetric strain dɛ pv and plastic shear strain dɛ ps for rock masses.
0jCiju dp ¼ dF (15l) Therefore, it is necessary to advance a model to capture defor-
mability of jointed rocks that is applied without precise informa-
tion about actual plastic volume change based on Eq. (15m). The
Because the factor jCiju takes a direction according to ap to reach present expression as represented by Eq. (15m) has the ability to
a failure, therefore jCiju 5 f ðap Þ, where ap 5 l=C; l and C depend on provide a relevant solution for interpretation of the modulus ratio.
dilatancy and yielding among joints, respectively, as described by
Trivedi (2010). dF is defined as a change in the flow function with
Discussion of Results
a change in mean effective confining pressure for a rock mass to
reach yielding Y by frictional flow. Evaluation of the modulus ratio was based on preliminary obser-
To explain the dissipation of plastic energy, investigators in- vations of different investigators working on strength characteristics
troduced a time variable of inconsequential physical meaning to of rock masses. Griffith theory as a conceptual starting point for later
bring compatibility to the relationship of plastic volume change and strength criteria that take into account frictional forces on closed
incremental plastic potential (Alejano and Alonso 2005). Zhang and cracks and the influence of intermediate principal stress (Jaeger et al.
Zhang (2009) associated physical meaning to such a change of 2007) called for further investigations to bring satisfactory agree-
plastic volumetric strain in the process of rock-mass deformation ment with experimental results; therefore, workers in the field
having stress-induced and thermal components enabling thermo- resorted to empirical expressions (Parry 1995). Based on experi-
dynamic compatibility. The initial confining pressure ratio pi =sr and mental observations (Brown 1970; Brown and Trollope 1970;
joint parameters tend to transform empirical parameter Ciju as a result Einstein and Hirschfeld 1973; Yaji 1984; Arora 1987; Roy 1993;
of yielding, hardening, and thermal characteristics of the joint ma- Trivedi and Arora 2007; Jain 2011) of the strength ratio [Eqs. (1a)–
terial. Integrating the process described by Eqs. (15a)–(15l), the (1g)] corresponding to a joint factor [Eqs. (4a)–(4b)], a modified
author used a simple numerical analysis to propose a relationship for joint factor [Eqs. (5) and (6)], and numerical trials (Trivedi 2010), the
Ch that depends on the initial confining pressure ratio and ap author obtained ap from the relationship between strength and
(Tables 4 and 5). modified joint factor for rock masses

354 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


Fig. 2. (a) Back-calculated values of ap versus joint factor and modified joint factor; (b) values of Ch in confined and unconfined state versus joint factor
and modified joint factor; (c) hardening parameter Ch versus initial isotropic pressure ratio for varying ap ; (d) back-calculated values of Ch within upper
and lower bounds versus initial isotropic pressure ratio

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013 / 355


Fig. 2. (Continued.)

356 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


ap Jfg ¼ ln smr (16a) a common joint factor and modified joint factor captured in laboratory
conditions [Eq. (16d)]. Assuming a progressive failure similar to
The modified joint factor [Eq. (5)] is obtained from the joint factor foundations (Trivedi and Sud 2005), l and C are defined as dilatancy-
[Eqs. (4a) and (4b)] using cg 5 1 [Eq. (6)] for blank joints. The back- related parameters for jointed rocks that connect the dilatancy of soils
calculated values of ap are plotted in Fig. 2(a) corresponding to the (Bolton 1986) and rock masses in plane strain and triaxial conditions
joint factor and modified joint factor using selected experimental (Trivedi 2010; Trivedi and Kumar 2010). The upper and lower bounds
observations. The regression of data indicates a broad range of trend of ap are tabulated in Table 4 to show that rock masses having granular
lines for ap with probable dilational and yielding characteristics. fill among the joints can be modeled with the help of a path-dependent
Eq. (16a) is extended to obtain a modulus ratio in terms of parameter that joins the initial condition of pressure with the end
modified joint factor as conditions decided by strain restrictions. The strain restrictions may not
remain strictly plane strain or triaxial; rather, the conditions may allow
Ch Jfg ¼ ln Emr (16b) for in-between values of ap . The author obtained an empirically ob-
  served rock-mass modulus by a path-constrained hardening parameter
Emr ¼ exp Ch Jfg (16c) Ch and a modified joint factor Jfg .
The parameter Ch further varies according to the ratio of initial
where Jfg 5 joint factor, modified in the Eqs. (16a)–(16c); and Emr 5 mean effective confining pressure p9i and uniaxial compressive
modulus ratio for jointed rocks. Emr is obtained as per Eqs. (2a) and strength of intact rocks sr . Because no effect of pore pressure is
(2b). The factor Ch was defined as a fitting constant by many considered further in this analysis, p9i 5 pi and p9 5 p, respectively.
investigators [Ch 5 2 0:0115 by Arora (1987) and Ch 5 2 0:0018, The parameter Ch is sensitive to plastic yielding, hardening, and
2 0:00115, and 2 0:0007 for varied modes of joint failures by crushing characteristics of the joint material. The initial confining
Singh (1997)]. The back-calculated values of Ch corresponding to pressure ratio pi =sr and presence of a gouge material tend to
the joint factor and the modified joint factor using experimental transform empirical parameter Ch as a result of yielding, hardening,
observations (Brown 1970; Brown and Trollope 1970; Einstein and and crushing characteristics and end conditions of the joint. The
Hirschfeld 1973; Yaji 1984; Arora 1987; Roy 1993) of the modulus plastic potential lines as per Eq. (15c) for varying initial confining
ratio [Eqs. (2a) and (2b)] corresponding to a joint factor [Eqs. (4a) pressure ratios pi =sr support a hypothesis of a nonassociated flow
and (4b)] and a modified joint factor [Eqs. (5) and (6)] show a rule. The author used a simple numerical analysis to propose a re-
wide range between 20.1 and 20.001 for unconfined conditions and lationship for Ch that depends on the initial isotropic pressure ratio
20.02 and 20.001 for confined conditions in Fig. 2(b). Such a trend (Table 5). Based on this analysis, it was found that
is shown in Fig. 2(b) by solid and dotted lines for unconfined and
confined conditions, respectively, and supports pressure-dependent Ch ¼ kðpi =sr Þh (18)
hardening characteristics of rock masses.
The strength ratio in Eq. (16a) is expressed in terms of the where k and h 5 empirical constants depending on ap ; and pi =sr 5
modified joint factor as initial isotropic pressure ratio in percent.
  Fig. 2(c) shows the variation of the empirical hardening pa-
smr ¼ exp ap Jfg (16d) rameter with the percent initial effective isotropic pressure ratio
pi =sr for jointed rocks obtained through the numerical formulation
smr ¼ expðl=CÞJfg (16e) shown in Eqs. (15a)–(15m). A range of Ch values corresponding to
the numerical trials [Fig. 2(c)] indicates that the average experi-
mental values [Fig. 2(d)] are captured within the upper and lower
smr ¼ expðA=CÞIr (16f )
bound values interpreted by the numerical model.
The empirically observed rock-mass modulus related a pressure-
where Ir 5 relative dilatancy index, as defined for soils by Bolton dependent hardening parameter Ch with a modified joint factor and
(1986). According to Bolton (1986), A takes a value of 3 and 5 in the the RQD data. The rock masses consisting of continuous and dis-
axisymmetric and plane-strain cases, respectively. Trivedi (2010) continuous formations recover a continuous core. The already
selected values of C between 21 and 25 for high- and low-pressure existing weaknesses may appear as discontinuities that affect the
conditions, respectively, for jointed sandstones. Eq. (16f) controls deformations. The recorded discontinuities equal a number of joints
the behavior of the rock mass if gouges predominate the behavior. appearing per unit length Jn . The core recovery data also capture the
Owing to loading, a change in the joint factor is represented by dJfg . orientation of the discontinuity relative to the scan line and thus the
The joint conditions have both plastic and damage components (Zuo orientation parameter of discontinuities nb . The modified joint factor
et al. 2012) as a result of input energy. Under favorable circum- as per the Eqs. (4)–(6) uses supplementary data of core recovery
stances, the rock mass may conserve or dissipate energy to modify along with RQD.
Jfg . The incremental changes dJfgd and Jfgp precede the modified joint With the help of significant joint-mapping parameters such as
factor to damage the plastic state as well in the Eq. (17). Therefore, average number of discontinuities per unit length Jn , roughness
on incremental loading, a change in the modified joint factor is parameter r, joint inclination parameter nb , discontinuity condition
cg , and groundwater condition Jw , a modified joint factor Jfg was
p
dJfg ¼ dJfg
d
dJfg (17) evaluated corresponding to the RQD data points published by Chun
et al. (2009). The data of Jfg as obtained along RQD and Emr from the
The modified joint factor Jfg ð$1Þ is a significant joint-mapping published data of Chun et al. (2009) are shown in Figs. 3 and 4.
parameter in both stages of input energy that arrives in small steps
and goes eventually to plastic failure. Under damaged conditions
Relationship of Modulus Ratio with Jfg and Rock
alone, dJfg 5 dJfgd , whereas at plastic failure, dJfg 5 Jfgp . The earlier
Quality Designation
investigators proposed ap as a fitting constant [ap 5 2 0:005 to
2 0:025 according to Arora (1987), Trivedi (1990), Arora and Trivedi The author examined the result in light of Johnston’s generalization
(1992), Singh (1997), and Trivedi and Arora (2007)] in relation to evolving from triaxial testing. The author conjoined these findings to

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013 / 357


Fig. 3. Variation of modulus ratio with modified joint factor and constant hardening parameter Ch

Fig. 4. Variation of modified joint factor with RQD

358 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


propose a relationship for modulus ratio incorporating a pressure- Fig. (4) shows variation of the modified joint factor Jfg with RQD
dependant hardening parameter. along with the experimental data of Chun et al. (2009). It also shows
It was observed that at lower values ( , 50%), RQD may not trend lines corresponding to the linear and exponential relationships
directly capture all the engineering possibilities among the joints that of joint number versus RQD by Priest and Hudson (1976). It is
are often well expressed by the modified joint factor. Zhang and interesting to note that the significant joint parameters considered in
Einstein (2004) and Zhang (2010) report that modulus ratio in- the modified joint factor tend to stretch the relationship toward lower
terpretation may have significant variation because of the intrinsic and upper bounds.
insensitivity of RQD to directional effects, discontinuity spacing, The modulus ratio is plotted along with RQD data compiled by
and discontinuity conditions. External to the RQD, the methods for Zhang (2010) [based on the data sets of Chun et al. (2009), Isik et al.
determining rock-mass modulus, namely, the plate load test, various (2008), Yang (2006), Vukovik (1998), Labrie et al. (1997), Ebisu
jacking tests, and the borehole deformability test in vertical and et al. (1992), Bieniawski (1978), Coon and Merritt (1970)] in Fig.
horizontal directions, have significant influence on the dataset and on 5(a). The figure shows a significant scatter. Zhang (2010) captured
the proposed relationships. From laboratory-scale testing and direct this scatter by normal stiffness and joint numbers. It appears that
measurements, it is established that a joint factor consisting of three the data sets of many investigators have an apparent dissimilarity of
significant factors, namely, joint frequency, roughness, and orien- best-fit trends to propose a common relationship for the modulus
tation (Ramamurthy and Arora 1994) may be conveniently im- ratio. The trend lines obtained from the selected data set of Chun
proved by the author to capture discontinuity conditions to relate et al. (2009) and Bieniawski (1978) suggest their applicability to low
them to the deformation and strength of rock mass as per Eqs. (16c)– and high RQD ranges, respectively. Some of the data sets report
(16e). Moreover, over the years, it has been well realized that RQD a decrease in modulus ratio with increasing RQD (Yang 2006;
may be related to the in situ strength and deformation properties of Labrie et al. 1997) as opposed to the best fit arrived at by collective
rock masses (Deere et al. 1967; Coon and Merritt 1970) by in- consideration of the Emr versus RQD relation from various inves-
corporating horizontal and vertical stiffness (Zhang 2010). tigators [Fig. (5a)]. The modulus ratio based on rock-mass classi-
A numerical representation of RQD in relation to a modified joint fication systems shows that other variants over and above
factor captures varied engineering possibilities, namely, number and discontinuity spacing and normal and shear stiffness also have
condition of joint and friction [Eq. (10)], orientation, gouge, water significant influence. Apparently, there are several factors captured
pressure, block size, and redistribution of stresses. A modified joint by RMR, GSI, Q, and RAC, as considered together in a modified
factor varying in a scale (from 0–1,000) has an exponential re- joint factor (Tables 1 and 2), that can be helpful in interpretation of
lationship with RQD [Eq. (19a)]. Therefore, employing a relation- the modulus ratio in relation to RQD.
ship of joint-mapping characteristics with RQD, the modulus ratio is The author presents these results to show that the hardening
conveniently expressed as parameter operated on a modified joint factor to compute the mod-
ulus ratio of rock mass. The modulus ratio obtained by extending
Jfg ¼ a expðb × RQDÞ (19a) best fit from the data compiled by Zhang (2010) and those obtained
at overburden at various data points published by Chun et al. (2009)
where a and b depend on the relationship of modified joint factor is expressed empirically along with upper-bound values at appro-
with RQD. The factors a and b may be readily estimated from core priate hardening levels in Table 6 as
drilling data of the rock mass and geotechnical site investigation Emr ¼ m lnðRQDÞ þ n (20)
report.
The relationship of modulus ratio and RQD in normalized form is where the values of m and n correspond to a hardening in relation to
the statistical data set considered for the study [Fig. 5(b)]. The
Emr ¼ exp½aCh expðb × RQDÞ (19b) preliminary estimates of modulus ratio obtained from Eq. (20) must
be used with great care in the absence of the mutual relationship
where a 5 1,000; b 5 2 1:5 to 2 0:05 depending on the extent of between Jfg and RQD.
significant joint characteristics of the rock mass; and Ch 5 path- Eq. (19b) is used to find the modulus ratio if a relationship
dependent hardening parameter. The relationships among fre- between Jfg and RQD as per Eq. (19a) is available from the core
quently employed rock-mass classifications used for determining recovery and geotechnical investigation. Fig. 5(c) shows that using
the modulus ratio, namely, RMR, GSI, Q, RAC, Jfg , and RQD are Eq. (19a) judiciously envelops the data of Emr versus RQD compiled
shown in Tables 1 and 2 for the purpose of ready reference to the by Zhang (2010) and Chun et al. (2009). The trend-line descriptor
present work. shows that adopting a constant hardening parameter (Ch 5 2 0:01;
The relationship of modulus ratio and modified joint factor b 5 2 0:1; and b 5 2 0:05) may lead to an erroneous modulus ratio,
[Fig. (3)] has an advantage of mapping varied engineering particularly at lower values of RQD. The trend lines corresponding
parameters in addition to discontinuity spacing. The effect of initial to the hardening at overburden [for data of Chun et al. (2009)] and
mean effective confining pressure ratio pi =sr on hardening pa- best fit [for data compiled by Zhang (2010)] show a similar trend.
rameter has an advantage of placing the deformation criteria as a The present proposal to map the modulus ratio together with the
function of a stress invariant I1 . modified joint factor, RQD, and a hardening parameter provides
The modulus ratio, estimated corresponding to GSI and as per a fresh insight into largely discontinuous rock-mass behavior. The
Chun et al. (2009), is plotted corresponding to the modified joint variation in the modulus ratio with the hardening parameter and
factor [Fig. (3)]. The modulus ratio according to GSI corresponding modified joint factor corresponding to the RQD condenses our
to each data point falls within the range corresponding to a modified understanding of the deformation of rock masses.
joint factor for the hardening parameter corresponding to a mean
confining pressure. Among these limiting initial mean effective Comments on the Validity of the Present Model
confining pressure ratios, the constant hardening (20.005, 20.01,
20.1) parameter was numerically operated uniformly for the range The application of a hardening parameter envelops the available data
of data set for modeled rock mass. set of modulus ratio and RQD within the reasonable zone of

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013 / 359


Fig. 5. (a) Best fit for variation of modulus ratio with RQD; (b) modulus ratio versus RQD for selected range of pressure-dependent hardening
parameter; (c) modulus ratio versus RQD for a pressure-dependent and constant hardening parameter

360 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


Fig. 5. (Continued.)

Table 6. Calculation of Modulus Ratio from RQD in Eq. (20) • The deformations assumed for the axisymmetric and plane-
End conditions Coefficient (m) and intercept (n) for RQD strain cases are similar to those in soil mechanics, which im-
plicitly assumes that the rock mass is isotropic and that continuum
ap m n behavior prevails. In practice, contrary to the assumption,
20:005 0.166 20:340 rock masses are anisotropic, and the deformations may be
20:009 0.302 20:357 discontinuous.
• The values of peak hardening of the rock masses asymptotically
20:0125 0.163 0.420
increase with mean isotropic pressure and are affected by joint
conditions. Hence, the hardening parameter for rock masses is
prediction using the present relations. The data sets of several recognized as a value considered approaching the intact rock with
investigators are used in generalized form. However, the relations confining pressure. The rock masses are assumed to deform and
described earlier should be applied judiciously and with care for the cause an increase in the deformations below the peak modulus
following reasons: with an increasing number of joints. Therefore, the increasing
• The strength and deformation criteria cited in this study are values of Jfg for the rock masses are essentially associated with
essentially empirical in nature, and the constants associated with more deformations.
various relationships (Tables 1 and 2) provide only an approx- • While gouges within the joints undergo volume changes, the
imate interpretation of the modulus ratio. hardening may accompany the gouges and the joint closure and
• Johnston and Chiu (1984) and Hoek and Brown (1997) de- then may damage the rock mass. The modulus ratio considers all
veloped an empirical strength criterion for intact rocks and then these parts. To isolate these effects, there is a need to fine-tune the
extended it to rock masses. Ramamurthy and Arora (1994) used results.
the joint factor to propose a strength criterion for jointed rocks • This work does not intend to comment on the advantages and
based on laboratory studies. The author applied concepts of disadvantages of other elastoplastic and hardening models. The
progressive failure and modified joint factor with the progress of purpose of the present relations is to provide a framework to
loading, and it is extended to the modulus ratio based on a hardening handle hardening problems for one-dimensional deformation and
parameter. The process used for the development of these criteria to evaluate it with the help of a simple numerical tool.
was one of pure trial and error like that of early workers. Because of • The volume-change characteristics of rock mass are difficult to
the empirical nature of the strength and deformation criteria of rock estimate. The laboratory estimate often fall short of the field
masses, it is uncertain whether the criterion will adequately predict estimate of the consequent deformation; therefore, a hardening
the behavior in all possible conditions. model is proposed in which changes in deformation relative to
• The modulus ratio considers jointed and intact rock for the same intact rocks are captured by a hardening parameter and modified
size of sample, and therefore, size effects are assumed to be joint factor. The precise predictions of the modulus of jointed
discounted. However, in practice, the block size effects may be rocks depends on an appropriate choice of hardening parameter
captured differently. Ch progressing toward failure conditions specified by ap .

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013 / 361


These limitations should be considered to apply to the proposed Ir 5 relative dilatancy index;
relationships for the modulus ratio and RQD. €I 5 limiting value for variation of hardening parameter
such that €I . Ch . bI;
bI 5 limiting value for variation of hardening parameter
Conclusions
such that €I . Ch . bI;
This paper describes an approach to finding deformation in jointed i 5 potential for change in isotropic pressure with plastic
rocks in terms of empirically established joint characteristics cap- volumetric strain and change in plastic shear strain with
tured by a modified joint factor. The modulus ratio is conventionally shear stresses;
adopted in relation to a fixed set of joint-mapping parameters in Jf 5 joint factor;
various rock-mass classification systems. This work uses a modified Jfg 5 modified joint factor;
joint factor consisting of joint roughness, joint orientation, number Jn 5 number of joints in the direction of loading per unit
of joints, and gouge parameters. The joint-mapping parameter is length;
altered according to initial confining pressure pi and a direction Jt 5 gouge thickness parameter;
selected by stress conditions of plane strain and triaxial states. Small Jv 5 volumetric joint count;
changes in these considerations unfold multiplicity in interpretation j 5 potential for change in isotropic pressure with plastic
of empirical joint characteristics and hence modulus ratio Emr . volumetric strain and change in plastic shear strain with
The dependence of intrinsic joint characteristics on initial con- shear stresses;
fining pressure provides input to a numerical technique that incor- Lna 5 reference length 51 m;
porates these effects into the modulus ratio. Comparing the results M 5 empirical rock constant in Johnston and Chiu (1984)
of the proposed model with the test data indicated that multiplicity criteria;
appearing in the interpretation of the modulus ratio is essentially due m 5 fitting constant for experimental data of RQD;
to hardening and accompanying changes. The hardening parameter n 5 fitting constant for experimental data of RQD;
is presented as a function of stress invariant. This paper considers the nb 5 joint-orientation parameter depending on inclination
effect of hardening in varied confinement conditions in plane-strain of the joint plane b with respect to the direction of loading
and axisymmetric cases on the modulus by the coefficients Ch and ap
as per RAC;
of the readily estimated modified joint factor and in situ data of RQD
p 5 mean effective confining pressure;
for rock masses. These coefficients have potential application in rock
pa 5 reference pressure 51 kPa;
mechanic designs. The main advantage of this model is to provide an
estimate of the modulus ratio of jointed rocks in terms of the already pi 5 initial mean confining pressure;
established parameters, namely, Jfg and RQD. The relation between q 5 shear stress;
the modulus ratio and hardening and intrinsic joint characteristics RD 5 relative density of gouge;
provides a rationale to resolve the diversity in interpretation of the r 5 joint strength parameter;
deformation of jointed rocks with RQD. Scr 5 strength-reduction factor;
Y 5 yield function;
Acknowledgments ap 5 parameter depending on l and C as per Trivedi (2010);
dp 5 change in mean effective confining pressure;
The author is thankful to Prof. L. Zhang, Univ. of Arizona, Tucson, dq 5 change in shear stress;
Arizona, Prof. V. K. Arora, NIT, Kurukshetra, India, and Mr. M. ɛps 5 plastic shear strain;
Sagong, KRRI, Uiwang, South Korea, for granting permission to ɛ pv 5 plastic volumetric strain;
use their representative data. The author sincerely compliments h 5 empirical constant depending on ap ;
the reviewers for their savant inputs, comments, and questions, k 5 empirical constant depending on ap ;
which significantly improved the presentation of this paper. l 5 empirical coefficient for dilatancy of joints;
sm 5 uniaxial compressive strength of rock mass;
Notation smr 5 strength ratio;
sr 5 uniaxial compressive strength of intact rock;
The following symbols are used in this paper: s1 5 principal stress in intact rock;
A 5 empirical constant; takes a value 3 for axisymmetric s2 5 principal stress in intact rock;
and 5 for plane-strain cases as per Bolton (1986); s3 5 principal stress in intact rock;
a 5 fitting constant for experimental data of RQD; s1m 5 principal stress in a rock mass;
a9 5 fitting constant for experimental data of RQD; s2m 5 principal stress in a rock mass;
B 5 empirical rock constant in Johnston and Chiu (1984) s3m 5 principal stress in a rock mass;
criteria; x 5 empirical rock constant in Johnston and Chiu (1984)
b 5 fitting constant for experimental data of RQD; criteria;
b9 5 fitting constant for experimental data of RQD; x b 5 empirical rock constant in Johnston and Chiu (1984)
C 5 empirical coefficient for dilatancy of joints; criteria; and
Ch 5 hardening parameter; xm 5 empirical rock constant in Johnston and Chiu (1984)
cg 5 modification for joint factor; criteria.
Em 5 modulus for rock mass;
Emr 5 modulus ratio; References
Er 5 modulus of intact rock;
F 5 flow function controlled by a flow rule; Aldrich, M. J. (1969). “Pore pressure effects on Berea sandstone subjected
gd 5 modification for gouge density; to experimental deformation.” Geol. Soc. Am. Bull., 80(8), 1577–1580.

362 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


Alejano, L. R., and Alonso, E. (2005). “Considerations of the dilatancy angle Johnston, I. W. (1985). “Strength of intact geo-materials.” J. Geotech. Eng.,
in rocks and rock masses.” Int. J. Rock Mech. Min. Sci., 42(4), 481–507. 111(6), 730–749.
Arora, V. K. (1987). “Strength and deformational behaviour of jointed Johnston, I. W., and Chiu, H. K. (1984). “Strength of weathered Melbourne
rocks.” Ph.D. thesis, Indian Institute of Technology (IIT), Delhi, India. mudstone.” J. Geotech. Eng., 110(7), 875–898.
Arora, V. K., and Trivedi, A. (1992). “Effect of Kaolin gouge on strength of Kolymbas, D. (2007). “Incompatible deformations in rock mass.” Acta
jointed rocks.” Proc., Asian Regional Symp. on Rock Slope, Central Geotech., 2(1), 33–40.
Board of Irrigation and Power, New Delhi, India, 21–25. Kulhawy, F. H. (1978). “Geomechanical model for rock foundation set-
Bandis, S., Lumsden, A. C., and Barton, N. R. (1983). “Fundamentals of tlement.” J. Geotech. Engrg. Div., 104(2), 211–227.
rock joint deformation.” Int. J. Rock Mech. Min. Sci. Geomech. Abstr., Labrie, D., Plouffe, M., Haevey, A., and Major, C. (1997). “Distress blast
20(6), 249–268. testing at Sigma Mine: Experimentation and results.” Division Rep.
Barton, N. (1983). “Application of Q-system and index tests to estimate MMSL 97-143E, Mining and Mineral Sciences Laboratories, Ottawa.
shear strength and deformability of rock masses.” Proc., Int. Symp. on Lade, P. V., and Kim, M. K. (1995). “Single hardening constitutive model
Engineering Geology and Underground Construction, Vol. 1, Labo- for soil rock and concrete.” Int. J. Solids Struct., 32(14), 1963–1978.
ratório Nacional de Engenharia Civil (LNEC), Lisbon, Portugal, 51–70. Mitri, H. S., Edrissi, R., and Henning, J. (1994). “Finite-element modeling
Barton, N. (2002). “Some new Q value correlations to assist in site char- of cable-bolted stopes in hard rock ground mines.” Proc., SME Annual
acterization and tunnel design.” Int. J. Rock Mech. Min. Sci., 39(2), Meeting, Society for Mining, Metallurgy and Exploration, Albuquerque,
185–216. NM, 94–116.
Barton, N., Loset, F., Lien, R., and Lunde, J. (1980). “Application of the Nicholson, G. A., and Bieniawski, Z. T. (1990). “A nonlinear deformation
Q-system in design decisions.” Subsurface space, M. Bergman, ed., modulus based on rock mass classication.” Int. J. Min. Geol. Eng., 8(3),
Vol. 2, Pergamon, Oxford, U.K., 553–561. 181–202.
Bieniawski, Z. T. (1978). “Determining rock mass deformability: Experi- Palmstrom, A. (2005). “Measurements of and correlations between block
ence from case histories.” Int. J. Rock Mech. Min. Sci. Geomech. Abstr., size and rock quality designation (RQD).” Tunn. Undergr. Space
15(5), 237–247. Technol., 20(4), 362–377.
Bolton, M. D. (1986). “The strength and dilatancy of sands.” Geotechnique, Parry, R. H. G. (1995). Mohr circle stress path and geotechnics, E & FN
36(1), 65–78. Spoon, London.
Brook, N. (1979). “Estimating triaxial strength of rock.” Int. J. Rock Mech. Priest, S. D., and Hudson, J. (1976). “Discontinuity spacing in rock.” Int.
Min. Sci. Geomech. Abstr., 16(4), 261–264. J. Rock Mech. Min. Sci. Geomech. Abstr., 13(5), 135–148.
Brown, E. T. (1970). “Strength of models of rock with intermittent joints.” Ramamurthy, T., and Arora, V. K. (1994). “Strength prediction for jointed
J. Soil Mech. and Found. Div., 96(6), 1935–1949. rocks in confined and unconfined states.” Int. J. Rock Mech. Min. Sci.
Brown, E. T., and Trollope, D. H. (1970). “Strength of a model of jointed Geomech. Abstr., 31(1), 9–22.
rock.” J. Soil Mech. and Found. Div., 96(2), 685–704. Roy, N. (1993). “Engineering behavior of rock masses through study of
Chun, B. S., Ryu, W. R., Sagong, M., and Do, J. N. (2009). “Indirect es- jointed models.” Ph.D. thesis, Indian Institute of Technology (IIT),
timation of the rock deformation modulus based on polynomial and Delhi, India.
multiple regression analyses of the RMR system.” Int. J. Rock Mech. Samadhiya, N. K., Viladkar, M. N., and Al-Obaydi, M. A. (2008). “Nu-
Min. Sci. Geomech. Abstr., 46(3), 649–658. merical implementation of anisotropic continuum model for rock
Coon, R. F., and Merritt, A. H. (1970). “Predicting in situ modulus of masses.” Int. J. Geomech., 8(2), 157–161.
deformation using rock quality indices.” ASTM Spec. Tech. Publ. 477, Schofield, A. N., and Wroth, C. P. (1968). Critical state soil mechanics,
ASTM, West Conshohocken, PA, 154–173. McGraw Hill, London.
Deere, D. U., Hendron, A. J., Patton, F. D., and Cording, E. J. (1967). Sen, Z., and Eissa, E. A. (1991). “Volumetric rock quality designation.”
“Design of surface and near surface construction in rock.” Proc., 8th U.S. J. Geotech. Engrg., 117(9), 1331–1346.
Symp. on Rock Mechanics, C. Fairhurst, ed., American Institute of Serafim, J. L., and Pereira, J. P. (1983). “Consideration of the geomechanical
Mining Engineers, Minneapolis, 237–302. classification of Bieniawski.” Proc., Int. Symp. Engineering Geology
Ebisu, S., Aydan, O., Komura, S., and Kawamoto, T. (1992). “Comparative and Underground Construction, Vol. 1, Laboratório Nacional de
study on various rock mass characterization methods for surface Engenharia Civil (LNEC), Lisbon, Portugal, 33–44.
structures.” Proc., Int. Society for Rock Mechanics Symp.: Eurock ’92, Shield, R. T., and Ziegler, H. (1958). “On Prager’s hardening rule.”
Rock Characterization, J. A. Hudson, ed., British Geotechnical Society, Z. Angew. Math. Phys., 9(3), 260–276.
Chester, U.K., 203–208. Singh, M. (1997). “Engineering behaviour of jointed model materials.”
Einstein, H. H., and Hirschfeld, R. C. (1973). “Model studies on mechanics Ph.D. thesis, Indian Institute of Technology (IIT), Delhi, India.
of jointed rock.” J. Soil Mech. and Found. Div., 99(3), 229–248. Swansson, S. R., and Brown, W. S. (1971). “An observation of loading path
Fossum, A. F. (1985). “Effective elastic properties for a randomly jointed rock dependence of fracture in rock.” Int. J. Rock Mech. Min. Sci. Geomech.
mass.” Int. J. Rock Mech. Min. Sci. Geomech. Abstr., 22(6), 467–470. Abstr., 8(3), 277–281.
Gerrard, C. M. (1982). “Elastic models of rock masses having one, two and Trivedi, A. (1990). “Effect of gouge on the strength of jointed rocks.” M.E.
three sets of joints.” Int. J. Rock Mech. Min. Sci. Geomech. Abstr., 19(1), dissertation, Regional Engineering College (REC), Kurukshetra, India.
15–23. Trivedi, A. (2010). “Strength and dilatancy of jointed rocks with granular
Grenon, M., and Hadjigeorgiou, J. (2003). “Evaluating discontinuity net- fill.” Acta Geotech., 5(1), 15–31.
work characterization tools through mining case studies.” Proc., Soil Trivedi A., and Arora V. K., (2007). “Discussion of bearing capacity of
Rock America 2003, Vol. 1, ASCE-MIT & BSCE, Boston, 137–142. shallow foundations in anisotropic non-Hoek-Brown rock masses.”
Griffith, A. A. (1924). “Theory of rupture.” Proc. 1st Int. Congress of J. Geotech. Geoenviron. Eng., 133(2), 238–240.
Applied Mechanics, C. B. Biezeno and J. M. Burgers, eds., 55–63. Trivedi, A., and Kumar, N. (2010) “Strength of jointed rocks with granular
Hoek, E., and Brown, E. T. (1997). “Practical estimates of rock mass fill.” Proc., Int. Society for Rock Mechanics Symp.: 6th Asian Rock
strength.” Int. J. Rock Mech. Min. Sci., 34(8), 1165–1186. Mechanics Symp., Central Board of Irrigation and Power (CBIP), Delhi,
Hoek, E., and Diederichs, M. S. (2006). “Empirical estimation of rock mass India.
modulus.” Int. J. Rock Mech. Min. Sci., 43(2), 203–215. Trivedi, A., and Sud, V. K. (2004) “Collapse behavior of coal ash.” J.
Isik, N. S., Doyuran, V., and Ulusay, R. (2008). “Assessment of deformation Geotech. Geoenviron. Eng., 130(4), 403–415.
modulus of weak rock masses from pressuremeter tests and seismic Trivedi, A., and Sud, V. K. (2005). “Ultimate bearing capacity of footings
surveys.” Bull. Eng. Geol. Environ., 67(3), 293–304. on coal ash.” Granul. Matter, 7(4), 203–212.
Jaeger, J. C., Cook, N. G. W., and Zimmerman, R. W. (2007). Fundamentals Varas, F., Alonso, E., Alejano, L. R., and Fdez.-Manín, G. (2005). “Study of
of rock mechanics, 4th Ed., Wiley-Blackwell, Hoboken, NJ. bifurcation in the problem of unloading a circular excavation in a strain
Jain, N. K. (2011).“Effect of cemented fill on strength of jointed specimen” softening material.” Tunnelling and Underground Space Technology,
M.E. dissertation, Dept. of Civil Engineering, Delhi Univ., Delhi, India. 20(4), 311–322.

INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013 / 363


Vermeer, P. A., and De Borst, R. (1984). “Nonassociated plasticity for soils Zhang, L. (2010). “Method for estimating the deformability of heavily
concrete and rocks.” Heron, 29(3), 3–64. jointed rock.” J. Geotech. Geoenviron. Eng., 136(9), 1242–1250.
Vukovic, N. (1998). “Comparison of laboratory and field modulus of Zhang, L., and Einstein, H. H. (2004). “Using RQD to estimate the de-
elasticity of rocks.” M.S. thesis, McGill Univ., Montreal. formation modulus of rock masses.” Int. J. Rock Mech. Min. Sci.
Wu, W., and Bauer, E. (1993). “A hypoplastic model for barotropy and Geomech. Abstr., 41(2), 337–341.
pyknotropy of granular soils.” Modern approaches to plasticity, Zhang, S., and Zhang, F. (2009). “A thermo-elastic-viscoplastic model for
D. Kolymbas, ed., Elsevier, New York, 225–245. soft sedimentary rock.” Soils Found., 49(4), 583–595.
Yaji, R. K. (1984). “Shear strength and deformation of jointed rocks.” Ph.D. Ziegler, H. (1959). “A modification of Prager’s hardening rule.” Q. Appl.
thesis, Indian Institute of Technology (IIT), Delhi, India. Math., 17(1), 55.
Yang, K. (2006). “Analysis of laterally loaded drilled shafts in rock.” Ph.D. Zuo, Y., et al. (2012). “Numerical study of zonal disintegration within
thesis, Univ. of Akron, Akron, OH. rock mass around deep excavated tunnel.” Int. J. Geomech., 12(4),
Yu, H. S. (2006). Plasticity and geotechnics, Springer, New York. 471–483.

364 / INTERNATIONAL JOURNAL OF GEOMECHANICS © ASCE / JULY/AUGUST 2013


Copyright of International Journal of Geomechanics is the property of American Society of
Civil Engineers and its content may not be copied or emailed to multiple sites or posted to a
listserv without the copyright holder's express written permission. However, users may print,
download, or email articles for individual use.

You might also like