11th Congress of the International Society for Rock Mechanics – Ribeiro e Sousa, Olalla & Grossmann (eds)

© 2007 Taylor & Francis Group, London, ISBN 978-0-415-45084-3

Shear behavior of rock joints under CNS conditions

Yongjun Kim, Sunki Kim & Juhwa Kim

Department of Civil Projects, Daelim Industrial Co. Ltd., Seoul, Korea

ABSTRACT: The shear characteristics of rock mass are subjected to not only to the shear force but also to the boundary
conditions given by neighboring rock mass. The boundary conditions of the rock mass can be classified into four categories
according to the stress state of the rock joint. The constant normal load (CNL) is the most widely used for shear test and
produces the lowest shear strength and different behavior. In this study, the shear behavior under constant normal stiffness
(CNS) condition is replicated by graphic method normalizing the test results under constant normal load (CNL) condition.

1 INTRODUCTION
The discontinuities in the rock mass reduce their resistance
capacity to shear loading and provoke the displacement of the
rock mass. In order to understand the behavior and stability of
rock structures, such conditions highlight the importance for
the engineer to apply analysis and test methods that comply
with circumstances of the field subjected to discontinuities.
Leichnitz (1985) studied and stressed the importance of the
boundary conditions in underground openings which is con-
strained by the surrounding rock masses using numerical
approach. Ohnishi (1990) attempted to simulate rock joints
conditions by installing a spring with stiffness at the top of
the rock joints, and Mouchaorab et al. (1994) developed an
experimental device which is able to simulate the rock joint
behavior using a hydraulic servo control system.
This study investigates a predictive method for the behavior
of rock joints under constant normal stiffness (CNS) using the
results of shear test that is performed under constant normal
load (CNL). The study processed by performing a series of
tests on saw-toothed specimens with constant angle, and then
predicting the shear behavior of CNS from the prior CNL test
results using partially modified graphic method of Saeb &
Amadei (1990).
2 BOUNDARY CONDITIONS
Apart from the structural features of the rock joints, the shear
characteristics of rock mass are subjected not only to the shear
force but also to the boundary conditions in the neighboring
the rock mass. For example, the boundary conditions of masses
of behaving freely in a slope differ from that in an under-
ground opening constrained by the surrounding rock masses.
Mouchaorab et al. (1994) classified the two boundary con-
ditions of rock masses into CNL and CNS conditions, as
illustrated in Figure 1.
As shown in the figure 1, the rock is not constrained under
CNL conditions and normal load acting on the rock joint
is constant, while CNS conditions stand for underground
openings, rock-socketed piles or grouted rock anchors where
normal load acting on the rock joint increases due to the
stiffness (K) of the surrounding rock mass when normal
displacement develops at the rock joint. The latter exhibits
larger shear strength than CNL conditions since normal load
increases due to the stiffness of the surrounding rock mass.
Numerous researchers made efforts to examine the behavior
of rock joints regarding the boundary conditions. According
to their results, the boundary conditions of rock masses can be
classified into 4 categories as follows with respect the stress
state developed in the rock joints.
• Constant Normal Load (K = 0): Normal load acts on the
joint surface is constant. In this case, the shear behavior
can be expressed in terms of the geometric conditions of
the rock joint and the initial normal load level.
• Constant Normal Stiffness (K = constant): Normal load
varies constantly during sliding. The variation of the nor-
mal load is governed by the size of the stiffness and the
characteristics of the normal displacement induced by the
roughness and strength of the rock joint.

Figure 1. Example of boundary conditions. (Mouchaorab & Benmokrane, 1994).

419

Figure 2. Ratio of required support force F to block weight W for
different boundary conditions. (a) Free body diagram. (b) F/W –
vertical displacement.
• Variable Normal Stiffness (K
= constant): Stiffness of the
rock mass is variable. As rock joint is closed, normal
load increase more than normal displacement in normal
compression test. In reality, stiffness of rock mass is
nonlinearity.
• Constant (Zero) Normal Displacement (K = ∞): In the
case of asperity strength of rock joint is low or normal load
level is very high, dilatancy is zero. Shear strength is devel-
oped by the asperity strength and surface of contact of the
rock joint instead of the dilation due to roughness. The shear
strength grows higher as the surface of contact and asperity
strength of the rock joint is larger.
Saeb & Amadei(1990) proposed a calculation method for
the support force according to the displacement of the block
in the underground opening. Figure 2 represents the varia-
tions of the support force with respect to boundary conditions.
Figure 2(a) illustrates a symmetric wedge block generated at
the top of an opening. Assuming that the displacement of the
wedge block which is generated by two discontinuities occurs
to the vertical direction, material properties of two discon-
tinuities is identical, and variations of the normal stress are
equal in both discontinuities. The sum of horizontal forces is
zero and the equilibrium of the whole block is defined by the
normal components as expressed in Equation 1.
420
Figure 3. Joint response curves for constant normal stress condi-
tions. (a) Joint closure curve. (b) Shear stress vs. shear displacement.
(c) Normal displacement vs. shear displacement.
Where F is the support force, W is the weight of the block, h
is the height of the wedge, γ is the specific weight of the rock
mass, and 2α is the angle between the two discontinuities.
Figure 2(b) plots the relationship of the support force
required to achieve equilibrium of the wedge block according
to the shear displacement for each type of boundary condi-
tions. These curves are similar to the ground reaction curves
used to calculate the support time and quantities of timbering
in tunnels. In Figure 2, it can be observed that the required
support force obtained under CNL conditions is excessive
compared to that under CNS conditions.
3 GRAPHIC METHOD
Saeb & Amadei(1990) suggested a graphic method that the
behavior of rock joints is accorded to the boundary conditions
and using the shear behavior which is obtained from normal
compression tests and CNL tests, of which results are plotted
in Figure 3. Figure 3(a) represents the relationship between the
normal stress and joint closure in the normal compression test,
where A is an arbitrary constant. Figure 3(b) represents shear
stress – shear displacement curves by means of a constant
peak shear displacement model. Figure 3(c) plots the normal
displacement of the joint by level of normal stress which is
accorded to the shear displacements of Figure 3(a).
The peak shear stress occurs at shear displacement U4
regardless of the normal stress levels, and the normal dis-
placement is assumed to be constant after U4.
Figure 4 represents the relationship between normal stress
and normal displacement according to the shear displacement,
using the results of Figure 3. The values of the normal stress
Figure 4. Normal stress vs. normal displacement curves at different
shear displacement levels. (Saeb & Amadei, 1990)
Figure 5. View of direct shear test apparatus for rock joint.
(σn ) and normal displacement (v) are the ones corresponding
to that of each of the shear displacements U0 , U1 , U2 , U3 ,
U4 , . . . represented in Figure 3(c). Observing Figure 4 gives
information on the variations of the normal stress and displace-
ment according to the shear displacement when the stiffness
(K) varies linearly from 0 to ∞ for an arbitrary normal stress
level or when the stiffness is not constant.
4 EXPERIMENTAL SETUP
In order to perform shear tests considering the boundary con-
ditions of the rock joint, a device has been developed to achieve
variations of the stiffness under the 4 categories of boundary
conditions (K = 0, K = constant, K
= constant, K = ∞). The
new direct shear test apparatus is constituted by an hydraulic
pump to produce compression, an actuator to apply normal
load and shear load, a load frame to support the loads devel-
oped by the actuator, a special shear box to remove undesirable
frictions occurring in the specimens, a controller to control
input and output signals from the measuring devices installed
in the apparatus, and a computer to control the apparatus and to
store data. Two LM guide bearings with very small coefficient
of friction below 0.0001 are disposed between the loading sys-
tem and the top of the shear box in order to control and prevent
the eccentric load and movement of the shear box during the
shear test.
Shear box is also fixed between the two roller bearings
installed at the end point of the shear boxes, and cyclic test is
possible. In addition, two load amplifiers are equipped in the
controller controlling the input and output of the apparatus in
order to ease control, and four servo-amplifiers are installed to
control displacement and loading in each direction. The setup
421
Figure 6. Test results normalized by graphic method. (a) Joint
closure curve. (b) Shear stress vs. shear displacement (c) Normal
displacement vs. shear displacement.
of the developed direct shear test apparatus is illustrated in
Figure 5.
5 PREDICTION OF THE SHEAR BEHAVIOR
OF ROCK JOINTS
Numerous researchers attempted to develop theoretical as
well as experimental approaches to identify systematically
the characteristics of the shear behavior under CNS condi-
tions. Among these researches, Dight et al. (1983), Skinas
et al. (1990) and Saeb & Amadei (1990) tried to predict the
shear behavior according to the stiffness (K) of the rock mass
subjected to CNS conditions using the results of the widely
performed CNL test. The common opinion of these researches
is that the variation of the normal stress and shear stress
obtained from the stress paths in the CNS test is equal to
the corresponding with the peak shear strength obtained under
CNL conditions or its slight lower values due to the roughness
reduction of the rock joint.
In addition, Ohnishi(1990) demonstrated experimentally
that the peak shear strength obtained under CNL conditions
appear in the stress path obtained from CNS test, regardless
of the initial normal stress (σni ) and variation of stiffness (K).
This study performs a series of shear test under CNL con-
ditions to predict the shear behavior under CNS conditions
using the graphic method proposed by Saeb & Amadei (1990).
The rock specimens used in the tests are rock-like specimens
of saw tooth shapes (i = 15◦ , amplitude 10 mm). Figure 6(a)
plots the normal compression test results. Figures 6(b) and
(c) present the results of CNL test performed at normal stress
(σn ) levels of 1, 3, 5, and 10 kg/cm2 . Figure 7 displays the nor-
mal stress(σn )-normal displacement(v) relationship regarding
the shear displacement(u) at intervals of 0.25 mm. Saeb &
Amadei (1990) normalized the normal stress-normal displace-
ment curves developed with respect to shear displacement as

Figure 7. Normal stress vs. normal disp. curves with respect to the
shear displacement Levels.
Figure 8. Shear behavior predicted by the modified graphic method.
(a) Normal displacement vs. shear displacement. (b) Shear stress vs.
shear displacement.
shown in Figure 4, assuming convergence to the constant nor-
mal compression curve for the specimen with matching joints.
The experimental results were normalized by means of their
proposed method and the predicted behaviors were compared
with results of CNS test. A little differences can be observed.
The differences may be attributed to the different normal com-
pression between matching joints and non-matching joints
during shear. Instead of applying convergence of the normal
stress-normal displacement curves developed during shear to
the normal compression curves, this study performed fitting of
each of the curves so as to normalize the normal stress-normal
displacement relationships regarding the shear displacement,
as shown in Figure 7. Accordingly, Figure 7 plots the normal
stress-normal displacement curves with respect to the shear
displacement for stiffness (K) of 20, 50, and 120 kg/cm3 under
an initial normal stress (σni ) of 1 kg/cm2 . The behavior under
422
CNS conditions was then predicted using the method pro-
posed by Saeb & Amadei (1990). The results were seemed to
be in good accuracy with experimental values as illustrated in
Figure 8.
6 CONCLUSIONS
It was shown that another approach can be applied to evalu-
ate shear behavior in rock structures with different boundary
conditions. In this study, the shear behavior under constant
normal stiffness condition was able to replicate by modified
graphic method normalized by the test results under constant
normal stress condition. The conclusions are as follows.
(1) The boundary conditions of the rock mass can be classified
into 4 categories according to the stress state of the rock
joint. Among them CNL condition is the most widely used
for shear test and produces the lowest shear strength.
(2) Using the results obtained from CNL test, the shear behav-
ior of CNS conditions has been predicted applying the
existing graphic method. But a little difference appears.
The differences maybe attributed to the different normal
displacement between matching joints and non-matching
joints during shear.
(3) Instead of applying convergence of curves obtained from
normal compression test of matching joint, this study per-
formed to fit the normal stress-normal displacement curves
of non-matching joint obtained from CNL test. The results
show a good agreement with behaviour predicted from
CNL test and behavior obtained from CNS test.
REFERENCES
Dight, P.M. & Chiu H.K. 1981. Prediction of shear behaviour of joints
using profiles, Int. J. Rock Mech. Min Sci. & Geom. Abstr., Vol.
18, 386–396.
Mouchaorab, K.S. & Benmokrane, B. 1994. A new combined servo-
controlled loading frame/direct-shear apparatus for the study of
concrete or rock joint behavior under different boundary and load-
ing conditions, Geotechnical Testing Journal, GTJODJ, Vol. 17,
No. 2, 233–242.
Leichnitz, W. 1985. Mechanical properties of rock joints, Int. Rock
Mech. Sci. & Geomech. Abstr. Vol. 22. No. 5, 313–321.
Ohnishi, Y. 1990. Shear behaviour of physical models of rock joints
under constant normal stiffness conditions, Proc. Int. Conf. on rock
joints, Leon, Norway, 267–273.
Saeb, S. & Amadei, B. 1990. Modelling joint response under constant
or variable normal stiffness boundary conditions, Int. Rock Mech.
Sci. & Geomech. Abstr., Vol. 27. No. 3, 213–217.
Skinas, C.A., Bandis, S.C. & Demiris, C.A. 1990. Experimental
investigations and modelling of rock joint behaviour under con-
stant stiffness, Proc. Int. Conf. on rock joints, Leon, Norway,
301–308.