You are on page 1of 13

Computers & Geosciences 37 (2011) 1437–1449

Contents lists available at ScienceDirect

Computers & Geosciences


journal homepage: www.elsevier.com/locate/cageo

Analytical solution of coupled stress-flow-transport processes in a single


rock fracture
Zhihong Zhao a,n, Lanru Jing a, Ivars Neretnieks b, Luis Moreno b
a
Department of Land and Water Resources Engineering, Royal Institute of Technology, Sweden
b
Department of Chemical Engineering and Technology, Royal Institute of Technology, Sweden

a r t i c l e i n f o abstract

Article history: A closed-form solution is presented for modeling the coupled stress-flow-transport processes along a
Received 9 July 2010 single fracture embedded in a porous rock matrix. Necessary assumptions were made to simplify the
Received in revised form subject into a two-dimensional (2D) problem, considering the changes of fracture aperture and matrix
20 December 2010
porosity under various stress conditions. The cubic law was assumed to be valid for the fluid flow in the
Accepted 22 February 2011
fracture, with an impermeable rock matrix. For transport mechanisms, advective transport along
Available online 11 March 2011
the fracture, longitudinal hydrodynamic dispersion in the flow direction, and the matrix diffusion were
Keywords: considered in three different transport models under constant concentration or constant flux (Danck-
Single rock fractures werts’) inlet boundary conditions. This analytical solution can be used as a constitutive model, or as an
Analytical solution
example for validation of similar constitutive models, for modeling the coupled hydro-mechanical-
Stress
chemical (HMC) processes in fracture networks of crystalline rocks. The influences of stress/deformation
Fluid flow
Transport processes on different transport mechanisms in a single fracture under different inlet boundary conditions
were studied for the first time. The results show that changes of fracture, as controlled by a combination of
normal closure and shear dilatancy, have a significant influence on the solute concentration distribution
both along the fracture and in the rock matrix, as well as on the solute residence/breakthrough time,
especially when shear-induced dilatancy occurs. Under compressions, the decreasing matrix porosity
slightly increases the solute concentration along the fracture and in the rock matrix.
& 2011 Elsevier Ltd. All rights reserved.

1. Introduction closed-form solution for the coupled stress-flow-transport processes


in a single rock fracture, for the first time, to the authors’ knowledge.
The fracture networks in crystalline rocks, for example, gran- Mass transport through single rock fractures is a complex
ites, provide the major pathways for groundwater flow and natural phenomenon including different mechanisms, such as
contaminant migration. The fractures’ dominating influence on advection and hydrodynamic dispersion, matrix diffusion, and
flow and transport processes is an important issue in many sorption reactions (Bodin et al., 2003a, 2003b). Since the 1970s, a
scientific or technical fields, such as geological radioactive waste large number of analytical, semianalytical, or numerical models
repositories, naturally fractured petroleum, and geothermal reser- have been proposed to simulate contaminant transport processes
voirs in crystalline rocks (Jing, 2003; Jing and Stephansson, 2007). in single rock fractures and fracture networks, using random walk
Due to the in situ stresses, the pattern and flow rate of ground- particle tracking techniques (Neretnieks, 1980; Tang et al., 1981;
water in crystalline rocks are influenced by fracture deformation, Moreno and Rasuson, 1986; Moreno et al. 1988). However, the
which, in turn, have an impact on mass transport processes. In the effects of the coupled stress-flow processes on these complicated
past, the mechanical, hydraulic, and mass transport processes in rock transport mechanisms have not been studied, compared with its
fractures were most often studied separately in different branches of potential importance in practice for subsurface engineering and
geosciences, and the combined effects of stress/deformation on fluid environmental protections.
flow and mass transport have not been investigated using analytical Mathematical models of coupled stress/deformation, ground-
solutions, despite the fact that the individual processes of stress, water flow, and solute transport processes in fractured rocks
flow, or transport have been extensively investigated by using could be developed by using either equivalent porous continuum
analytical or numerical modeling. In this paper, we present a or discrete element method (DEM) models. For the equivalent
continuum models, the governing equations, derived based on the
principles of continuum mechanics, are solved numerically to
n
Corresponding author. describe the macroscopic behavior of the fractured rocks, while
E-mail address: zhzhao@kth.se (Z. Zhao). overlooking the details of the fracture network geometry. It is the

0098-3004/$ - see front matter & 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cageo.2011.02.015
1438 Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449

most commonly applied modeling approach used for large-scale Besides the parallel plate model as noted above, in order to
problems of either sparsely or heavily fractured rocks. The DEM derive an analytical solution, the following basic assumptions
models, on the other hand, explicitly represent the geometry and were also adopted:
contact mechanisms of blocks (rock matrix) and fracture systems,
based on the solutions of the equations of motion of the blocks, (1) The model was defined in a 2D space, with the fracture
and the constitutive models of rock matrix (block materials) and aperture much smaller than its length.
fractures, respectively. A closed-form solution for the coupled (2) The hydraulic aperture of the fracture was assumed to be
stress/deformation, fluid flow, and mass transport processes in a equal to its mechanical aperture, and a residual aperture
single fracture, therefore, can serve as an effective tool contribut- always existed even under high compression stress. This
ing to establish the constitutive models of rock fractures for DEM could be regarded as an indirect consideration of the effects
modeling approaches. of rock fractures surface roughness in reality.
The main objective of this paper is to develop an analytical (3) The effect of fluid pressure and its variation on fracture
solution for the coupled stress-flow-transport problems in a deformation in the flow direction was assumed to be negli-
single fracture bounded by porous rock matrix, considering the gible, which indicates that the fracture surfaces always
change of fracture aperture and rock matrix porosity resulting remain parallel during the deformation/displacement pro-
from stress/deformation, steady fluid flow, and different disper- cess. This assumption is reasonable because usually ground-
sive and diffusive processes of solute transport. The model water flows under small hydraulic pressure gradients
represented by this solution can be especially useful for safety/ through a fracture in subsurface rocks in practice.
performance assessments of underground radioactive waste repo- (4) The fracture aperture changes due to normal closure or shear
sitories in fractured crystalline rocks, since the simulation of dilatancy were described by commonly adopted mechanical
radioactive nuclide transport process depends largely on the constitutive models of rock fractures in the field of rock
validity and reliability of the models of the stress-flow-transport mechanics, in order to simplify the derivation of the final
in individual rock fractures. analytical solution. This assumption is reasonable, since the
scope of the study is the effect of the stress on the flow and
transport processes, not the complicated stress/deformation
processes in details.
2. Conceptual model (5) Transverse dispersion within the fracture was neglected,
compared to longitudinal dispersion. This means that the
The simplest mathematical model of a single rock fracture is solute concentration was assumed to be equal at the fixed
the smooth parallel plate model, which assumes that the two fracture section.
walls of a fracture can be represented by two nominally smooth (6) The permeability of the porous matrix was sufficiently low to
and parallel surfaces (Snow, 1965), separated by an initial hinder water flow, but contaminant particles could enter the
aperture of 2b [L] (Fig. 1). Considering a single fracture embedded rock matrix by molecular diffusion.
in a saturated porous rock matrix, with a contaminant source
located at the origin of the fracture (x¼ 0), the solute particles are
With the above conceptualization and assumptions, we can
then transported by the groundwater flow along the fracture (the
explicitly describe the individual mechanical, hydraulic, and
advection), and may also diffuse in and out of the porous matrix.
transport processes in single fractures, separately, and then
Although the parallel plate assumption differs from the natural
combine them together to generate a closed-form solution for
fractures with rough surfaces, it remains, however, the basis of
coupled stress-flow-transport processes in a single rock fracture.
many complicated modeling approaches for flow and transport
processes at the local scale (Bodin et al., 2003b). It can provide the
basic insights for fluid flow and solute transport through an
individual rock fracture under different stress conditions. More 3. Coupled stress-flow-transport processes
importantly, the parallel plate model is the necessary condition
for deriving a closed-form solution for such a complex problem of 3.1. Mechanical process
coupled stress-flow-transport processes in a single fracture,
because such an analytical solution does not exist for natural When stress is applied to rocks, both the fractures and the rock
fractures of rough surfaces. This, however, does not prevent matrix will move and/or deform, which results in changes in
considerations of some mechanical effects of roughness on the fracture aperture and matrix porosity. These changes influence
mechanical component of the closed-form solution indirectly, as groundwater flow and contaminant transport in fractures, as well
described in Section 3. as the diffusion of solutes into and out of the rock matrices.

z
T: tensile strength
kn: nonlinear
normal stiffness
Ss S ks T D: dilation
2b 0 D
x
kn S: slider
Ss: shear strength
ks: nonlinear shear

Fig. 1. Schematic illustration of parallel plate model for fractures (after Bodin et al., 2003b).
Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449 1439

3.1.1. Normal closure behavior (1985) suggested that in situ fractures probably behave in a
It has been demonstrated that the normal closure of a rough manner similar to the third or fourth loading cycles (Fig. 2a).
rock fracture is typically nonlinear under compressive normal
stress; i.e., the normal stiffness of a fracture increases with 3.1.2. Shear behavior
increasing normal stress (Goodman, 1976; Swan, 1983; Cook, Fracture displacement during shear is inherently a coupled
1992; Rutqvist and Stephansson, 2003), and finally levels off to process, in which both normal and shear displacements occur
some asymptotic value at high values of the compressive normal (Jaeger et al., 2007). Fig. 3a shows the typical, although much
stress (Jaeger et al., 2007). The most commonly applied fracture idealized, shear stress–shear displacement behavior of a fresh
normal closure model is Bandis’ hyperbolic function (Fig. 2b; rock fracture under constant normal stress conditions, during
Bandis et al. (1983)): direct shear tests under laboratory test conditions. During this
  process, the most important mechanism is the so-called shear
un
sn ¼ kn0 , ð1Þ dilatancy, i.e., the increase of fracture aperture during shear
1un =unmax displacement due to roughness of the fracture surfaces. Various
empirical models have been suggested to simulate the dilatancy
where sn [M/LT2] is the normal stress (compression positive), kn0
phenomenon in the past, with varying degrees of success. The
[M/L2T2] is the initial normal stiffness at a prescribed stress state,
simplest approach is to use a dilation angle relating the normal
un [L] is the normal displacement of the fracture, and unmax [L] is
and tangential displacements of a fracture:
the maximum normal displacement, approached asymptotically
with increasing normal stress. The basic parameters kn0 and unmax ud ¼ us tan fd , ð3Þ
are determined by experiments (Barton et al., 1985). The frac- where ud [L] is the shear dilatancy, us [L] is the shear displacement
ture’s normal closure can be expressed by rearranging Eq. (1) as and fd is the dilation angle. Note that we did not consider a
sn dilation angle that varies with normal stress and shear displace-
un ¼ : ð2Þ ment in this study, but assumed that the dilation angle was a
kn0 þ sn =unmax
constant for deriving a closed-form solution.
It is noted here that those parameters in the above equation A simplified model, as shown in Fig. 3b, was adopted to
depend on the stress paths of normal compressive loading– approximate the behavior of rock fractures during shear
unloading cycles (Jing and Stephansson, 2007). Barton et al. (Fig. 3a), so that the final analytical model can capture the main

σn
σn

un un

Fig. 2. Fracture closure behavior under normal compression (after Bandis et al., 1983). (a) Compressive loading-unloading cycles and (b) normal stress — normal
displacement curve.

 

p p

r r

us us
(I) (II) (III)
ud ud

udmax udmax

φd
us us
up ucs

Fig. 3. Mechanical behavior of fractures during direct shear testing under constant normal stress (after Jing and Stephansson, 2007). (a) Conceptual model and (b)
simplified model.
1440 Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449

mechanical features without losing necessary generality. The Therefore, the stress-dependent porosity of rock matrix can be
peak shear stress, tp [M/LT2], is determined by the Mohr– written as
Coulomb criterion, characterized by a cohesion (C) and a frictional a3 sn
ea3 sni Þ
angle (f):
y ¼ 1ð1yi Þea1 ðsn sni Þa2 =a3 ðe , ð12Þ

9t9r C þ sn tan f ¼ tp : ð4Þ where yi is the initial matrix porosity at zero stress. This relation
links stress with rock porosity that, in turn, is related to matrix
Before reaching the peak shear stress, the shear stiffness, ks, is diffusion process.
assumed to be constant. In the case of
jtj Z tp , ð5Þ 3.2. Fluid flow process
one has
The conceptual model shown in Fig. 1 a fracture model for
t ¼ signðus Þtp , ð6Þ
fluid flow in rock fractures with a possible analytical solution
where sign( ) is the sign function, which extracts the sign of the (Zimmerman and Bodvarsson, 1996), represented by the well-
shear displacement, i.e., the shear direction. known cubic law. With a horizontal fracture of constant cross-
Dilatancy starts to occur at the onset of slip of the fracture, section in the x direction, one can argue that the fluid velocity is
governed by the dilation angle. The accumulated dilation is negligible in the y and z directions. With these simplifications and
limited by a critical value of shear displacement, ucs [L], corre- the ‘‘no-slip’’ boundary condition, the relation between the flow
sponding to the observation that crushing of asperities at large rate and the hydraulic pressure gradient can be represented by
shearing would eventually prevent continued increase of dila- the cubic law (Bear, 1972):
tancy. In this way, three functions are employed to describe the
2b3 dp
dilatancy during three different stages (I, II, and III in Fig. 3b) of Qx ¼ w , ð13Þ
3m dx
the shear displacement:
8 where Qx [L3/T] is the volumetric flow rate, w [L] is the fracture
> 0 us r up with up ¼ tp =ks ðIÞ
< width, m (Pa s) is the dynamic viscosity, x [L] is the coordinate
ud ¼ us tan fd up rus r ucs ðIIÞ : ð7Þ along the fracture axis, and dp/dx [MT2/L2] is the hydraulic
>
: u tan f
cs d u s Z ucs ðIIIÞ pressure gradient.
The average flow velocity of a steady flow state flow can be
obtained by dividing the flux, Qx, by the cross-sectional area, 2wb [L2]:
3.1.3. Hydraulic aperture
Experimental results typically show that an apparent residual Qx b2 dp
v¼ ¼ : ð14Þ
aperture, bres [L], exists in rock fractures even under high 2wb 3m dx
compressive stresses, when fracture appears to be nominally This equation links the change of flow velocity in fracture with
closed by compression (Witherspoon et al., 1979; Raven and the stress-dependent aperture. Furthermore, it transfers the
Gale, 1985; Pyrack-Nolte et al., 1987; Cook, 1992; Renshaw, effects of stress to the advective transport in fracture, as well as
1995). Additionally, the residual aperture may vary with sizes of other transport mechanisms.
the tested fracture specimen, which was not considered in this
paper. Therefore, the resultant hydraulic aperture can be
expressed as 3.3. Transport process

2b ¼ bres þ ðunmax un Þ þ ud , ð8Þ Generally speaking, contaminant transport and spreading in
where un and ud are determined by Eqs. (2) and (7), respectively. the fractured porous media are influenced by many causes, such
as advective transport along the connected fractures, longitudinal
hydrodynamic dispersion within the fractures in the flow direc-
3.1.4. Rock matrix porosity
tion, matrix diffusion (solute particles going into and out of the
Rock matrix porosity, y, changes under the in situ stresses and
micropores in rock matrices), adsorption onto the fracture walls,
can be defined as a function of volume change of micropores in
adsorption within the matrix, or decay for radioactive nuclides. It
the rock matrix (Han and Dusseault, 2003). In this study, Eq. (9)
is expected that some of them are affected significantly by the
relates the increment in matrix porosity to the increment in
in situ stresses. Many more details on stress effects on particle
compressive stress and was adopted because of the explicit
transport in single fracture are presented in Section 4.
inclusion of stress (Zimmerman, 1991):
In the past three decades, the analytical solutions for represent-
dy ¼ ½Cbc ð1yÞCm dsn , ð9Þ ing solute transport in single fractures for different initial and
2 2 2 2
where Cbc [L T /M] and Cm [L T /M] are effective bulk compressi- boundary conditions have been extensively studied and developed.
bility and rock matrix compressibility, respectively. Note that sn Among them, Tang et al. (1981) derived a general analytical solution
is the compressive stress acting on the rock matrix. In this study for contaminant transport along a discrete fracture seated in a
we assumed that it was equal to the stress applied on the fracture porous rock matrix, considering all the transport mechanisms noted
surfaces. Because Cm is usually small enough to be negligible, above. Additionally, various solutions in semianalytical forms for
Eq. (9) can be simplified into different boundary conditions at the inlet were also proposed
(Neretnieks, 1980; Rasmuson and Neretnieks, 1981; Moreno and
dy ¼ Cbc ð1yÞdsn : ð10Þ Rasuson, 1986; Berkowitz and Zhou, 1996; Sun and Buschec, 2003;
In reality, Cbc is stress dependent and can be measured by Shih, 2007). Although these models considered many processes,
laboratory tests. Based on experimental data, Zimmerman (1991) mechanisms, and factors, they did not consider the effect of the
derived an empirical relation for stress-dependent Cbc as stress/displacement process of the fractures and the effects of matrix
pore volume changes due to matrix deformation. It is also difficult to
Cbc ¼ a1 þa2 ea3 sn , ð11Þ
identify or explore the contributions of different individual transport
where a1, a2 and a3 are constants determined from curve fitting of mechanisms in a fundamental study. Therefore, in the following
the test data; the unit of sn is MPa here. section, three transport models and two inlet boundary conditions
Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449 1441

are considered in this research, in order to examine the stress effects the constrictivity and tortuosity are difficult to measure and to be
on different transport mechanisms. separated experimentally, an empirical relation of Archie’s law
To represent the contaminant transport processes in the (Archie, 1942) is often used to relate the effective diffusion
conceptual fracture model as shown in Fig. 1, two coupled, one- coefficient to the porosity:
dimensional (1D) governing equations are usually needed: one for m
De ¼ D y : ð20Þ
the fracture and the other for the porous matrix. The coupling
conditions are the continuity of fluxes and concentrations along The exponent m can take the values between 1.3 and 4 for
the fracture surface. In this study, we first assume that the different rocks but is often taken to be 1.6 for granites (Ruffet
continuous source concentration is constant c0 at the inlet, the 1:6
et al., 1995). Therefore, one has De ¼ y D and Dm ¼ y D .
0:6

same as defined by Tang et al. (1981). The boundary conditions 2


Da [L /T] is the apparent diffusion coefficient, which is defined as
for the concentration along the fracture are
De
cf ð0,tÞ ¼ c0 , ð15aÞ Da ¼ , ð21Þ
ðy þ rm Km Þ

cf ð1,tÞ ¼ 0, ð15bÞ where rm [M/L3] is the bulk density of the matrix, and Km is the
distribution coefficient that was defined by Freeze and Cherry
cf ðx,0Þ ¼ 0: ð15cÞ (1979) as the mass of solute adsorbed per unit volume of matrix
The boundary conditions for the concentration of matrix are divided by the concentration of solute in solution. Neretnieks (1980)
clarified the distinction between effective and apparent diffusivities.
cm ðx,b,tÞ ¼ cf ðx,tÞ, ð16aÞ
Da is usually used for describing instationary diffusion (Eq. (18)),
whereas De is used in the literature for flux calculation (Eq. (17)).
cm ð1,z,tÞ ¼ 0, ð16bÞ
The term (y þ rmKm) is called rock capacity, a. The volume sorption
cm ðx,z,0Þ ¼ 0, ð16cÞ coefficient rmKm is equal to zero if no sorption or retardation occurs.
The value of rock capacity a is then equal to that of porosity y; i.e.,
where cf [M/L3] is the volumetric concentration of solute in the Da ¼Dm for the conservative contaminant transport.
fracture, written as a function of location and time cf(x,t); cm [M/L3] Without detailed derivation, we present only the final solu-
is the volume concentration of solute in the porous matrix, written tions of Eqs. (17) and (18) below, under the boundary conditions
as cm(x,z,t); N represents the infinite. The boundary condition of (Eqs. (15) and (16)) (Carslaw and Jaeger, 1959; Neretnieks, 1980):
constant inlet concentration is the most used one in the modeling,
which may corresponds to a contaminant which is dissolved in the tw De !
cf b Da0:5
fluid to a concentration determined by its solubility in practice ¼ erfc ðfor fractureÞ, ð22Þ
c0 2ðttw Þ0:5
(Moreno and Rasuson, 1986). However, the contaminant is trans-
ported by advection and hydrodynamic dispersion in the fractures, !
and the concentration at the locations close to the inlet can be cm ðtw =bÞðDe =D0:5 0:5
a Þ þ zb=Da
significantly influenced by the boundary conditions when the ¼ erfc ðfor matrixÞ, ð23Þ
c0 2ðttw Þ0:5
value of the Péclet number is small. In other words, much more
contaminant mass may be introduced into the fractures due to the where tw is water residence time, and is equal to tw ¼ x=v.
large dispersive flux at the inlet boundary when dispersion is
strong. In order to meet the mass conservation condition, the
constant flux (Danckwerts’) inlet boundary conditions were pro-
posed (Danckwerts, 1953), and the corresponding solutions were 3.3.2. Transport model II
derived by Rasmuson (1986) and Moreno and Rasuson (1986). Based on the transport model I presented above, we added the
They are presented as transport model III in later section. hydrodynamic dispersion term in the fracture. The governing
equation for the solute transport process of a fracture then becomes
3.3.1. Transport model I 
The simplest particle transport model in a single fracture is the @cf @cf @2 cf De @cm 
þv Df  ¼ 0, ð24Þ
one only considering the advective transport in the flow direction @t @x @x2 b @z z ¼ b
and matrix diffusion, without dispersion or adsorption. Consider-
ing the mass balance of solute in the fracture and in the pores of where Df [L2/T] is the hydrodynamic dispersion coefficient, which
the matrix, respectively, one can obtain the following differential describes the spreading of a tracer pulse by local variations in
equations of solute transport (Neretnieks, 1982): velocity and molecular diffusion. It can be expressed as Df ¼
 aL vþ D , where aL [L] is the dispersivity in the direction of the
@cf @cf De @cm  fracture axis (Bear, 1972). Detwiler et al. (2000) proposed a complex
þv  ¼ 0 ðfor fractureÞ, ð17Þ
@t @x b @z z ¼ b theoretical model of Df dependent on the Péclet number (Pe), but
we use the first one for simplicity in this study. Together with
@cm @2 cm 2
Eq. (18), one obtains the following solutions for t 4 x2 =4x DL :
¼ Da ðfor rock matrixÞ, ð18Þ
@t @z2
where v [L/T] is the mean fluid velocity in the fracture (Eq. (14)); !
 Z 1
De [L2/T] is the effective diffusion coefficient, which is related to cf 2 v 2 x2 v2
the pore diffusion coefficient Dm [L2/T] and the rock matrix ¼ pffiffiffiffi exp x exp x  2
c0 p 2Df l 16x Df 2
porosity in the following form: !
De ¼ yDm : ð19Þ x2 De 1
erfc 2D
pffiffiffiffiffiffi 2
dx ðfor fractureÞ, ð25Þ
D 2 0:5
4bx f a 2ðtðx =4x Df ÞÞ
Neretnieks (1980) defined the pore diffusion coefficient as
Dm ¼ D ðd=o2 Þ, where D* [L2/T] is the molecular diffusion coeffi-  Z !
cm 2 v 1
2 x2 v2
cient in the fluid, d is the pore constrictivity ( o1), and o is the ¼ pffiffiffiffi exp x exp x  2
tortuosity, which is a function of rock matrix porosity. Because
c0 p 2Df l 16x Df 2
1442 Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449

0 1
2 D
determine the actual integration range. The equations can also
B4bxx2 D pe ffiffiffiffi þ Dzb0:5 C
be solved directly by Mathmatica or other commercial computa-
B f Da a C
erfcB  0:5 C dx ðfor matrixÞ: ð26Þ
@ A tion software.
x2
2 t 2
4x D f

4. Model behavior verification


3.3.3. Transport model III
As noted before, constant inlet concentration boundary con- Because there are no available experimental results published
dition is an approximation only, if the hydrodynamic dispersion is in the existing literatures and no laboratory experiments con-
present (Rasmuson, 1986). The correct one ensuring the mass sidering coupled stress-flow and transport processes with matrix
conservation is the constant flux (Danckwerts’) boundary at inlet, diffusion and dispersion were ever performed, we apply the above
which is expressed as (Danckwerts, 1953) solutions to an example to demonstrate the effect of stress on the
@ contaminant transport of a rock fracture in a porous rock matrix.
vc0 ¼ vcf ð0,tÞDf c ð0,tÞ: ð27Þ
@x f We test the sensitivity of the different transport mechanisms,
It is corresponding to a constant flux vc0 at x ¼0 and there is no such as advection, matrix diffusion, and longitudinal hydrody-
hydrodynamic dispersion for xo0 (Moreno and Rasuson, 1986). namic dispersion, to the normal closure, shear dilatancy, and
The solutions for the governing equations of Eqs. (24) and (18) stress-dependent matrix porosity, respectively. The material
under the constant inlet flux boundary condition (Eq. (27)) properties are given in Table 1.
were derived by Rasmuson (1986) and Moreno and Rasuson The three transport models with different longitudinal disper-
(1986): sivities under the initial condition of zero stress applied after 10
 Z 1   years are compared in Fig. 4. Without normal stress applied, the
cf v v v
¼ exp x exp  x initial fracture aperture was 30 mm, and the groundwater velocity
c0 D Df Df in the fracture was 7.5  10  7 m/s for a hydraulic pressure gradient
" f x
!
 Z 1 of 10 Pa/m (or water head gradient of 0.001 m/m). If only consider-
2 v 2 x2 v2
pffiffiffiffi exp x exp x  2
ing the pure advection in the fracture, the contaminant would be
p 2Df l 16x Df 2 expected to travel far enough after 10 years, to make contaminant
0 1 3
x2 De 1 concentration equal to c0 in the fracture. However, the relative
B pffiffiffiffiffiffi 0:5 C 7
 erfc@4bx2 Df Da  2
Adx5dy ðfor fractureÞ, concentration in the fracture decreased to a very small value after
2 t x2 x¼8 m, if the matrix diffusion is considered, indicating that diffu-
4x Df
sion of the contaminant (for example, radioactive nuclides) into the
ð28Þ rock matrix probably is one of the more significant transport
 Z 1   mechanisms, especially for final safety assessment of a potential
cm v v v
¼ exp x exp  x repository of radioactive waste (Neretnieks, 1980).
c0 Df Df Df
" x
!
For larger longitudinal dispersivity of aL ¼1 m, the longitudinal
 Z hydrodynamic dispersion increases the contaminant concentra-
2 v 1
2 x2 v2
 pffiffiffiffi exp x exp x  2 tion in the fracture or in the rock matrix at the initial state under
p 2Df l 16x Df 2
0 1 3 the constant concentration inlet boundary (Fig. 4a and c), which is
2
De ffiffiffiffi
x
2
p þ zb
0:5
in line with the results shown by Tang et al. (1981). However,
B4bx Df Da Da C 7 under constant flux boundary conditions, the relative solute
erfcB
@
 0:5 Cdx7dy
A 5 ðfor matrixÞ: ð29Þ
2 t x2
2
concentration estimated by transport model III is smaller than
4x Df
that from transport model I in the locations closer to the inlet
(x o3 m), due to the influence of longitudinal hydrodynamic
3.4. Solutions for coupled stress-flow-transport process dispersion. After that part, longitudinal hydrodynamic dispersion

Combining Eqs. (8) and (14), we obtain the average velocity of


groundwater for a steady-state flow in a single fracture as Table 1
Material properties of groundwater, rock matrix, and fracture.
1  2 dp
v¼ bres þðunmax un Þ þ ud : ð30Þ
12m dx Material parameters Value
Then substituting the updated half aperture b of the fracture
Groundwater
(Eq. (8)), the mean fluid velocity v obtained from Eq. (30), and the Hydraulic pressure gradient, dp/dx (Pa/m) 10
effective and apparent diffusion coefficients (Eqs. (20) and (21)) Dynamic viscosity, m (Pa s) 1.0  10  3
due to the changed matrix porosity represented by Eq. (12) into
Intact rock
the three transport models presented above (Eqs. (22), (23), (25), Initial matrix porosity, y 0.01
(26), (28), and (29)), one can have the solutions for the coupled Empirical constants, a1, a2, a3 (  10  4,  10  4, 1) 0.82, 5.35, 0.12
stress-flow-transport process in the fracture–matrix system as Cementation factor, n 1.6
shown in Fig. 1, and examine the effect of stress on solute Fractures
transport in the single fracture. We did not present the complete Normal stiffness at zero stress state, kn0 (GPa/m) 434
expression here to save space. Shear stiffness, ks (GPa/m) 434
Dispersivity along fracture, aL (m) 0.5
To evaluate the integrals in the transport models II and III
Diffusivity in water, Dn (m2/s) 1.6  10  9
(Eqs. (25), (26), (28), and (29)), we used the Gauss–Legendre Frictional angle, | (deg) 24.9
quadrature technique with 100 Gauss points. It was found that Cohesion, c (MPa) 0
the numerically significant portion generally extends over a Critical shear displacement for dilation, Ucs (mm) 3
smaller range, even though the integrand theoretically extends Shear dilation angle, fd (deg) 5.0
Residual aperture, bres (mm) 5
from l to infinity (Tang et al., 1981). In this way, before the Maximum normal displacement, Dunmax (mm) 25
integration, a prior scanning procedure was carried out to
Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449 1443

αL=1m αL=0.1m
1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III
R e la tive c o n c e n tra tio n c /c0

R e la tive c o n c e n tra tio n c /c0


For fracture

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0 2 4 6 8 0 2 4 6 8
Distance along fracture (m) Distance along fracture (m)

1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III
R e la tive c o n c e n tra tio n c /c0

R e la tive c o n c e n tra tio n c /c0


For matrix

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Distance into matrix (m) Distance into matrix (m)

Fig. 4. Comparison among the different transport models under the initial conditions without stress applied. Two longitudinal dispersion coefficients (aL ¼1 and 0.1 m)
were assumed to study the sensitivity of the transport model to the inlet boundary conditions.

increases the contaminant concentration in the fracture (Fig. 4a). process of increasing normal stress, and it increases the contami-
The same phenomenon exists in the matrix for the constant flux nant concentration along the fracture and especially in the rock
inlet boundary also (Fig. 4c). Three transport models give similar matrix. In Fig. 5c and d, the relative concentration from the
concentration profiles for smaller longitudinal dispersivity of transport model III is higher than that from transport model I,
aL ¼0.1 m (Fig. 4b and d), which indicates that the longitudinal indicating indirectly an important role of dispersion in the
dispersion plays a negligible role in this case. The concentration fractures. Without considering the hydrodynamic dispersion, it is
discontinuity at the inlet for transport model III is also illustrated almost impossible for particles to diffuse into rock matrix under
in Fig. 4a. The effects of constant concentration or flux inlet compressive stress of 10 MPa at this short time scale (t¼10 years).
boundary conditions are demonstrated here, and the same trend Fig. 6 presents the same information as in Fig. 5, except that it
can be found in the following results with stress applied. is at a long time scale (t ¼50 years). With the longer time, the
(1) Normal closure effect by pure normal compression: In this contaminant concentrations both along the fracture and in the
case, we mainly study the effect of the fracture aperture closure rock matrix further increase (compared with the results at the
under an increasing normal stress on the mass transport along the time scale of 10 years), and the effects of stress become more
fracture and into the rock matrix, without considering the shear obvious. However, the general trend remains the same. Fig. 6c
dilatancy and matrix porosity changes. The issues of the shear shows that the relative concentration estimated by transport
dilatancy and matrix porosity change are treated in the later model I without dispersion becomes larger than that of transport
sections. model III after a longer time, but the relative concentration of
Fig. 5 shows the concentration profiles along the fracture and transport model I is still smaller than that of transport model III
in the rock matrix (x¼2 m) at a short time scale (t¼10 years), in Fig. 6d. Considering the relative concentration in the fracture at
under two normal stresses of 5 and 10 MPa, respectively. The x¼2 m, transport model I gives a larger value in Fig. 6a than that
results illustrate that the contaminant concentrations both along of transport model III, and inverse in Fig. 6b. It is concluded that
the fracture and in the rock matrix decrease with the increasing larger concentrations in the fracture induce larger concentrations
normal stress value, regardless of what transport models are used. in the rock matrix, and after a long enough time the concentration
For transport model III, the concentration at the inlet decreases of transport model I will be larger than that of transport model III,
with increasing stress too. The fracture’s normal closure reduces both in fracture and in matrix.
the transport activity in the fractured rocks. However, the effect of Figs. 7 and 8 show the influence of the normal stress on
hydrodynamic dispersion becomes more significant during the breakthrough curves for the fracture (x¼2 m) and the matrix
1444 Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449

5 MPa 10MPa
1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III
R e la tive c o n c e n tra tio n c /c 0

R e la tive c o n c e n tra tio n c /c 0


For fracture

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0 2 4 6 8 0 2 4 6 8
Distance along fracture (m) Distance along fracture (m)

1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III
R e la tive c o n c e n tra tio n c /c 0
R e la tive c o n c e n tra tio n c /c 0
For matrix

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Distance into matrix (m) Distance into matrix (m)

Fig. 5. Concentration profiles along the fracture and into the matrix (x ¼2 m) at the short time scale (t¼ 10 years), under normal stress of 5 and 10 MPa.

(x¼2 m, z¼0.1 m), respectively. Under the initial condition of no becomes more significant, when the fluid velocity (or fracture
stress, the breakthrough curves for models I and II overlap very aperture) becomes smaller.
much after longer times, indicating that the influence of hydro- (2) Shear dilatancy effect: To examine the effect of shear
dynamic dispersion is negligible then. However, at the beginning displacement of the fracture, we fix the normal stress at 5 MPa,
period of transport, relative concentration increases more quickly and increase the shear displacement gradually to produce shear
with longitudinal dispersion under constant concentration inlet dilatancy. From Eqs. (4) and (7), the fracture aperture would start
conditions. Due to less contaminant mass introduced under to increase by shear dilatancy, when shear displacement us
constant flux inlet conditions, the relative concentrations of becomes larger than 5.3 mm. Fig. 9 shows the concentration
transport model III are lower than the others after a long time. profiles along the fracture and in the rock matrix (x ¼2 m) at
The normal stress decreases the fracture aperture and flow rate, the short time scale (t ¼10 years), with the shear displacement of
and in turn the rate of contaminant concentration in the fracture 0.1 and 0.2 mm, respectively. The influence of increasing shear
and the rock matrix. When a normal stress of 5 MPa is applied, displacement is just the inverse of that of increasing normal
the hydrodynamic dispersion in transport model II increases the stress. The shear dilatancy increases fracture aperture, then flow
rate of contaminant concentration in the fracture and the rock rate, and finally the rate of contaminant concentration in the
matrix compared with transport model I. The same phenomena fracture and the rock matrix. The influence of inlet boundary
happen for transport model III after the transport start immedi- conditions becomes less significant when large shear dilatancy
ately, but its relative concentration is smaller after a long time. happens, because the advection becomes more significantly
From the above results, it can be seen that the fracture induced from increasing fluid velocity.
aperture closure induced by normal stress plays an important (3) Effect of matrix porosity change: Under normal stress
role in the flow and transport in fractured rocks. It is expected conditions, the influences of matrix porosity change on transport
that the extent of fracture closure would become smaller with mechanisms are investigated by comparison between the results
further increase of the normal stress, because the normal stiffness with and without considering matrix porosity changes, under the
increases with the increasing normal stress. In this way, the compressive normal stress of 5 and 10 MPa, respectively (Fig. 10).
impact of normal stress would decrease under high normal For the concentration profiles along the fracture, if we consider
stresses, especially when fracture aperture is close to its residual the matrix porosity change under normal stresses, the concentra-
aperture value. The influence of hydrodynamic dispersion tion along the fracture increases significantly, compared with the
Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449 1445

5 MPa 10 MPa
1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III

R e la tive c o n c e n tra tio n c /c 0


R e la tive c o n c e n tra tio n c /c 0
For fracture

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0 2 4 6 8 0 2 4 6 8

Distance along fracture (m) Distance along fracture (m)

1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III
R e la tive c o n c e n tra tio n c /c 0

R e la tive c o n c e n tra tio n c /c 0


0.6 0.6
For matrix

0.4 0.4

0.2 0.2

0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Distance into matrix (m) Distance into matrix (m)

Fig. 6. Concentration profiles along the fracture and into the matrix (x¼ 2 m) at the long time scale (t ¼50 years), under normal stress of 5 and 10 MPa.

1.0

0.8
R e la tive c o n c e n tra tio n c /c0

0.6

0.4

Model I
0.2 Model II
Model III

0.0
0 20 40 60 80 100 120 140 160
Time (years)
Fig. 7. Comparison of breakthrough curves for the fracture at x ¼ 2 m, between the initial case with zero normal stress (solid lines) and the case under the normal stress of
5 MPa (dash lines).
1446 Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449

1.0

0.8

R e la tive c o n c e n tra tio n c /c0 0.6

0.4

Model I
0.2 Model II
Model III

0.0
0 20 40 60 80 100 120 140 160
Time (years)

Fig. 8. Comparison of breakthrough curves for the matrix (x¼ 2 m, z¼ 0.1 m), between the initial case with zero normal stress (solid lines) and the case under the normal
stress of 5 MPa (dash lines).

0.1mm 0.2mm
1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III
R e la tive c o n c e n tra tio n c /c0

R e la tive c o n c e n tra tio n c /c0


For fracture

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0 2 4 6 8 0 2 4 6 8
Distance along fracture (m) Distance along fracture (m)

1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III
R e la tive c o n c e n tra tio n c /c0

R e la tive c o n c e n tra tio n c /c0

0.6
For matrix

0.6

0.4 0.4

0.2 0.2

0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Distance into matrix (m) Distance into matrix (m)

Fig. 9. Concentration profiles along the fracture and into the matrix (x ¼2 m) at the short time scale (t¼ 10 years), under shear displacements of 0.1 and 0.2 mm.
Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449 1447

5 MPa 10 MPa
1.0 1.0

Model I Model I
Model II Model II
0.8 0.8 Model III
Model III

R e la tive c o n c e n tra tio n c /c 0


R e la tive c o n c e n tra tio n c /c 0
For fracture

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0 2 4 6 8 0 2 4 6 8
Distance along fracture (m) Distance along fracture (m)

1.0 1.0

Model I Model I
Model II Model II
0.8 Model III 0.8 Model III
R e la tive c o n c e n tra tio n c /c 0

R e la tive c o n c e n tra tio n c /c 0


0.6
For matrix

0.6

0.4 0.4

0.2 0.2

0.0 0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5
Distance into matrix (m) Distance into matrix (m)

Fig. 10. Comparison of concentration profiles for the fracture and matrix (x¼ 2 m) between the cases with (dash lines) and without (solid lines) considering the matrix
porosity change under normal stress conditions, at the short time scale (t ¼10 years).

case of constant matrix porosity. Under higher normal stress, this based on the assumptions listed in Section 2. Some important
difference becomes more obvious (Fig. 10a and b). The decreasing outstanding issues are addressed below.
matrix porosity caused by increasing the normal stress, inducing The 2D geometrical model was used mainly for its simplicity,
decreasing diffusion coefficients, makes the particles more diffi- with homogeneous material properties for the rock matrix, in
cult to diffuse into the rock matrix. The concentration in the rock order that a closed-form solution is possible. The stress/deforma-
matrix also increases with decreasing matrix porosity, due to the tion relation of the matrix, the groundwater flow velocity, and
increasing normal stress, compared with the case with constant contaminant transport along fracture or into rock matrix are
matrix porosity (Fig. 10c and d). The possible reason is that larger assumed to be 1D. Their validities depend on the configuration
contaminant concentration gradient at the fracture wall tends to of the particular rocks and fractures. For example, if the rock
push more particles into the matrix pores. matrix is heterogeneous with large porosity, the contaminant
distribution in the rock matrix may become a 2D problem, and
the groundwater flow through porous matrix should not be
5. Discussion and conclusions neglected. Therefore, it is emphasized here that the present model
is only appropriate for rocks with small matrix porosity. If there
An analytical solution for coupled stress-flow-transport processes are significant groundwater exchanges between the micropores in
in a single rock fracture was developed to consider the influence of matrix and fractures, the model presented in this paper should
stress/displacement on the contaminant transport through individual not be directly used.
rock fractures. This model explicitly integrates the mechanical, flow, Across the fracture section, the groundwater velocity and
and transport processes in a single rock fracture into a compact contaminant concentration are assumed to be uniform, following
mathematical form, by linking the stress and transport processes the parallel plate model of a constant aperture. For fractures with
through changes of aperture and matrix porosity, based on the cubic variable apertures, the above assumption is not valid. However,
law and Archie’s law, respectively. Therefore, it can be directly used for fractures with aperture of micrometer scale, under relatively
as a simple constitutive model for modeling the coupled hydro- lower hydraulic pressure gradient and at the time scale of
mechanical-transport processes in fracture networks in crystalline performance/safety assessments for underground radioactive
rocks. However, it is noted here that the validity of this model is waste repositories, the above assumption may still be applicable.
1448 Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449

The model in this paper was derived based on some empirical but the dilatancy increases the corresponding contaminant
mechanical models, so its validity largely depends on the range of concentrations.
validity of those empirical models and associated parameters. (2) The importance of the hydrodynamic dispersion depends on
Actually, the main contribution of this paper is to highlight the the fluid velocity, inlet boundary conditions, and time scales.
importance of the coupled stress-flow-transport processes in (3) With decreasing matrix porosity, the concentrations both
fractured crystalline rocks. Any other mechanical models can along the fracture and in the rock matrix increase.
replace those models used here, as long as they can consider (4) The transport models should be chosen according to practical
the changes of fracture aperture and matrix porosity under stress, conditions, after the nonnegligible transport mechanisms are
explicitly in compact mathematical function forms, to meet the identified first.
requirements of special rocks or sites, and the resultant equations
can be integrated analytically.
The above outstanding issues and limitations are identified for
There are a large number of material parameters and empirical
future research. Among them, the validity range of the parallel
constants in the present model. How to obtain the reasonable
plate model of the fractures, the effect of fracture roughness, and
values for them from laboratory or in situ experiments remains to
the influence of water pressure in fracture and pore pressure in
be a difficult problem and no attempt has been made to test
micropores of matrix will be considered in future studies.
coupled stress-flow-transport processes in a single rock fractures
with quantitative representations of transport mechanisms, vari-
ables, and parameters, so far, due to experimental difficulty. The
uncertainties in measuring the material properties have an Acknowledgment
important influence on the final assessments of radioactive waste
repositories. Moreover, these properties and parameters always We acknowledge the financial support from the Swedish
vary among different sites. Nuclear Fuel and Waste Management Co. (SKB) through the
Constant concentration inlet boundary results in large dispersive international DECOVALEX-2011 project.
flux at the inlet boundary for large hydrodynamic dispersion coeffi-
cient at the beginning of transport, and more solute mass is injected
in the fracture and rock matrix after a long time. Constant flux References
boundary conditions, however, can eliminate the extra dispersive flux
at the inlet, which is more proper physically, but induces an artificial Archie, G.E., 1942. The electrical resistivity log as an aid in determining some
reservoir characteristics. Petroleum Technology 146, 54–62.
discontinuity at the inlet boundary. If the longitudinal hydrodynamic
Bandis, S., Lumsden, A.C., Barton, N.R., 1983. Fundamentals of rock joint informa-
dispersion is neglected, the constant flux boundary condition tion. International Journal of Rock Mechanics and Mining Sciences & Geome-
becomes the same as the constant concentration boundary condition. chanics Abstracts 20, 249–268.
The different results between these two boundary conditions indicate Barton, N.R., Bandis, S., Bakhtar, K., 1985. Strength, deformation and conductivity
coupling of rock joints. International Journal of Rock Mechanics and Mining
that it is a crucial factor for modeling. The concentrations noted in Sciences & Geomechanics Abstracts 22, 121–140.
this paper are resident concentrations, and more information on the Bear, J., 1972. Dynamics of Fluids in Porous Media. Elsevier, New York, 764 pp.
transformation and to distinguish between flux concentrations and Berkowitz, B., Zhou, Z., 1996. Reactive solute transport in a single fracture. Water
Resource Research 32 (4), 901–913.
resident concentrations can be found in Kreft and Zuber, (1978), Van Bodin, J., Delay, F., de Marsily, G., 2003a. Solute transport in a single fracture with
Genuchten and Parker (1984), Rasmuson (1986), and Moreno and negligible matrix permeability: 1. Fundamental mechanisms. Hydrogeology
Rasuson (1986). Journal 11, 418–433.
Bodin, J., Delay, F., de Marsily, G., 2003b. Solute transport in a single fracture with
In the transport models presented in this paper, adsorption on negligible matrix permeability: 2. Mathematical formalism. Hydrogeology
the walls of fracture and within the matrix pores, and the Journal 11, 434–454.
radioactive decay were not included, but there are many available Carslaw, H.S., Jaeger, J.C., 1959. Conduction of Heat in Solids, New York, 2nd ed.
Oxford University Press, 510 pp.
transport models considering these issues in the literature (Tang Cook, N.G.W., 1992. Natural joints in rock: mechanical, hydraulic and seismic
et al., 1981; Sudicky and Frind, 1984; Cormenzana, 2000; Sun and behaivor and properties under normal stress. International Journal of Rock
Buschec, 2003). The basic mechanical and flow models presented Mechanics and Mining Sciences & Geomechanics Abstracts 29 (3), 198–223.
Cormenzana, I., 2000. Transport of a two-member decay chain in a single fracture:
here can be directly coupled with other transport models con-
simplified analytical solution for two radionuclides with the same transport
sidering adsorption and decay without major difficulties. The properties. Water Resource Research 36 (5), 1339–1346.
present model can also be further extended to a system with Danckwerts, P.V., 1953. Continuous flow systems: distribution of residence times.
multiple parallel fractures or other initial boundary conditions, Chemical Engineering Science 2, 1–13.
Detwiler, R.L., Rajaram, H., Glass, R.J., 2000. Solute transport in variable-aperture
with the known analytical transport solutions (Sudicky and Frind, fractures: an investigation of the relative importance of Taylor dispersion and
1982; West et al., 2004). macrodispersion. Water Resource Research 36 (7), 1611–1625.
From the present study, stresses do have a significant influence Freeze, R.A., Cherry, J.A., 1979. Groundwater. Prentice-Hall, Englewood Cliffs, NJ
604 pp.
on the groundwater flow and contaminant transport in rock Goodman, R.E., 1976. Methods of Geological Engineering in Discontinuous Rocks.
fractures, through changing the fracture aperture and matrix West Publishing Company, San Francisco, 472 pp.
porosity. The fracture closure under normal stress and the Han, G., Dusseault, B.M., 2003. Description of fluid flow around a wellbore with
stress-dependent porosity and permeability. Journal of Petroleum Science and
dilatancy during shear are the two major mechanisms controlling Engineering 40, 1–16.
the fluid flow and transport in the fractures, and consequently Jaeger, J.C., Cook, N.G.W., Zimmerman, R.W., 2007. Fundamentals of Rock
influence the groundwater velocity in fractures, contaminant Mechanics, 4th ed. Blackwell Publishing, 488 pp.
Jing, L., 2003. A review of techniques, advances and outstanding issues in
concentration distribution, and breakthrough time. The results numerical modelling for rock mechanics and rock engineering. International
show that matrix porosity changes corresponding to actual stress Journal of Rock Mechanics and Mining Sciences 40, 283–353.
conditions should not be neglected in practice, even though we Jing, L., Stephansson, O., 2007. Fundamentals of Discrete Element Methods for
Rock Engineering—Theory and Application. Elsevier Science Publishers,
assumed that the porosity was really small in this study and only
Rotterdam 562 pp.
matrix diffusion was considered. Kreft, A., Zuber, A., 1978. On the physical meaning of the dispersion equation and
Some basic concluding remarks are presented below. its solutions for different initial and boundary conditions. Chemical Engineer-
ing Science 33, 1471–1480.
Moreno, L., Rasuson, A., 1986. Contaminant transport through a fissured porous
(1) Fracture’s normal closure decreases the contaminant con- rock: impact of the inlet boundary condition on the concentration profile in
centrations both along the fracture and into the rock matrix, the frock matrix. Water Resource Research 22 (12), 1728–1730.
Z. Zhao et al. / Computers & Geosciences 37 (2011) 1437–1449 1449

Moreno, L., Tsang, Y.W., Tsang, C.F., Hale, F.V., Neretnieks, I., 1988. Flow and tracer Snow, D.T., 1965. A parallel model of fractured permeable media. Ph.D. disserta-
transport in a single fracture: a stochastic model and its relation to some field tion. University of California, 331 pp.
observations. Water Resource Research 24 (12), 2033–2048. Sudicky, E.Q., Frind, E.O., 1982. Contaminant transport in fractured porous media:
Neretnieks, I., 1980. Diffusion in the rock matrix: an important factor in radio- analytical solutions for a system of parallel fractures. Water Resource Research
nuclide retardation? Journal of Geophysical Research 85 (B8), 4379–4397. 18 (6), 1634–1642.
Neretnieks, I., 1982. Tracer movement in a single fissure in granitic rock: some Sudicky, E.A., Frind, E.O., 1984. Contaminant transport in fractured porous media:
experimental results and their interpretation. Water Resource Research 18 (4), analytical solutions for a two-member decay chain in a single fracture. Water
849–858. Resource Research 20 (7), 1021–1029.
Pyrack-Nolte, L.J., Myer, L.R., Cook, N.G.W., Witherspoon, P.A., 1987. Hydraulic and Sun, Y., Buschec, T.A., 2003. Analytical solutions for reactive transport of N-mem-
mechanical properties of natural fractures in low-permeability rock. In: ber radionuclide chains in a single fracture. Journal of Contaminant Hydrology
Herget, G., Vongpaisal, S. (Eds.), Proceedings 6th International Congress on
62–63, 695–712.
Rock Mechanics. Montreal, Canada, pp. 225–232.
Swan, G., 1983. Determination of stiffness and other joint properties from rough-
Rasmuson, A., 1986. Exact solution of some models for the dynamics of fixed beds
ness measurements. Rock Mechanics and Rock Engineering 16, 19–38.
using Danckwerts’ inlet condition. Chemical Engineering Science 41 (3),
Tang, D.H., Frind, E.O., Sudicky, E.A., 1981. Contaminant transport in fractured
599–600.
porous media: analytical solution for a single fracture. Water Resource
Rasmuson, A., Neretnieks, I., 1981. Migration of radionuclides in fissured rock: the
influence of micropore diffusion and longitudinal dispersion. Journal of Research 17 (3), 555–564.
Geophysical Research 86 (B5), 3749–3758. Van Genuchten, M.Th., Parker, J.C., 1984. Boundary conditions for displacement
Raven, K.G., Gale, J.E., 1985. Water flow in a natural rock fracture as a function of experiments through short laboratory soil columns. Soil Science Society of
stress and sample. International Journal of Rock Mechanics and Mining America Journal 48, 703–708.
Sciences & Geomechanics Abstracts 22 (4), 251–261. West, M.R., Kueper, B.H., Novakowski, K.S., 2004. Semi-analytical solutions for
Renshaw, C.E., 1995. On the relationship between mechanical and hydraulic solute transport in fractured porous media using a strip source of finite width.
apertures in rough-walled fractures. Journal of Geophysical Research 100 Advances in Water Resources 27, 1045–1059.
(B12), 24629–24636. Witherspoon, P.A., Amick, C.H., Gale, J.E., Iwai, K., 1979. Observations of a potential
Ruffet, C., Dardot, M., Gueguen, Y., 1995. Surface conductivity in rocks: a review. size effect in experimental determination of the hydraulic properties of
Surveys in Geophysics 16 (1), 83–105. fractures. Water Resource Research 15, 1142–1146.
Rutqvist, J., Stephansson, O., 2003. The role of hydromechanical coupling in Zimmerman, R.W., 1991. Compressibility of Sandstones, Developments in Petro-
fractured rock engineering. Hydrogeology Journal 11, 7–40. leum Science, vol. 29. Elsevier, New York, 181 pp.
Shih, D.C.F., 2007. Contaminant transport in one-dimensional single fractured Zimmerman, R.W., Bodvarsson, G.S., 1996. Hydraulic conductivity of rock frac-
media: semi-analytical solution for three-member decay chain with pulse and tures. Transport in Porous Media 23, 1–30.
Heaviside input sources. Hydrological Processes 21, 2135–2143.

You might also like