Você está na página 1de 6

Analytical solution of a pressure transmission experiment

J.M. Huyghe
Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands

ABSTRACT: Analytical solutions are derived for a pressure transmission experiment of a saturated charged
compressible porous medium. The governing equations describe infinitesimal deformations of charged porous
media saturated with a mono-valent ionic solution. From the governing equations a coupled diffusion equation
in state space notation is derived for the electro-chemical potentials, which is decoupled introducing a set of
normal parameters, being a linear combination of the eigenvectors of the diffusivity matrix. The magnitude of the
eigenvalues of the diffusivity matrix correspond to the time scales for Darcy flow, diffusion of ionic constituents
and diffusion of electrical potential.

1 INTRODUCTION include an electrical potential to explicitly model the


ionisation of the material and the ensuing phenom-
Shales exhibit swelling and shrinkage behavior ena of Donnan-osmosis, electro-osmosis, streaming
induced by changes in external salt concentration. currents, diffusion currents, streaming potentials and
Swelling and shrinkage results from the interaction diffusion potentials. For this reason, (Molenaar et al.
between the fixed charges on the solid and ions present 1998) recasted the model of (Huyghe and Janssen
in the fluid. Donnan-osmosis, electroosmosis, stream- 1997) to include compressibility being an important
ing potentials and streaming currents, (Mitchell 1993) feature of shales. A thorough verification of FEMim-
are observed phenomena. Swelling of shales is of plementations of the models, require adequate analyt-
major concern for petroleum engineering during active ical solutions of transient problems. At present these
drilling, (Heidug and Wong 1996), (van Oort 1997). analytical solutions are available only for the incom-
(Sherwood 1993), (Ekbote and Abousleiman 2005) pressible case. In this paper the governing equations
developed a chemomechanical model of shale. This of the compressible quadriphasic model are presented
model extends Biots theory to include the diffusion- shortly for the case of infinitesimal deformation of
convection of salt within the deforming shale. (Heidug a mixture. These equations are decoupled following
and Wong 1996) use a thermodynamic framework (Biot 1956). Analytical solutions are presented for the
to build a chemomechanical constitutive model of pressure-transmission experiment under conditions of
hydration in water absorbing rocks. The aforemen- near electroneutrality.
tioned rock model show significant analogy to biome-
chanical models of cartilagenous tissues. (Lai et al.
1991) derived a small deformation triphasic formu- 2 THEORY
lation for soft charged hydrated tissues, considering
all constituents incompressible. (Huyghe and Janssen In this section the governing equations, as derived by
1997) extended the chemo-electro-mechanical model (Molenaar et al. 1998), are given in the special case of
to finite deformation behavior of a porous medium infinitesimal quadriphasic mechanics of compressible
consisting of an electrically charged solid, liquid, charged porous media. Specific constitutive material
cations and anions. This model has been implemented behavior is considered and corresponding relations
in a finite element formulation and is verified exper- for the osmotic pressure and electrical potential at
imentally on one dimensional transient behavior of equilibrium are given. Finally the governing equa-
annulus fibrous tissue by (Frijns et al. 1997). The tions are reduced for a 1-dimensional configuration
solution procedure has been verified using an analyti- and decoupled.
cal solution of confined swelling and compression by
(van Meerveld et al. 2003). While these biomechanical
2.1 Governing and balance equations
models include diffusion-convection of salts like the
rock models, they differ essentially from the models of We consider a porous solid saturated with a monova-
(Sherwood 1993; Heidug and Wong 1996) in that they lent ionic solution. The current volume fraction of the

57
Copyright 2005 Taylor & Francis Group plc, London, UK
solid is s , the current volume fraction of the ionic 2.2 Constitutive behavior
solution is f . The corresponding initial volume frac- The fluxes obey a coupled form of Darcys, Ficks and
f
tions are 0s , 0 . The solution is a molecular mixture Ohms law:
of water (w), a cation (+) and an anion (). The par-
tial densities of water, cations and anions are in the
current state , + and , and the in the initial state
0 , 0+ and 0 . The unstrained volume change of
constituent = , +, is: with K a positive definite symmetric permeabil-
ity matrix. In relationship (7) the electro-chemical
potentials, are defined as:


in which i is an intrinsic reference density such that
with and W the electrical potential and the energy
function respectively. The stress appearing in the
momentum balance (3) is the partial derivative of
the energy function W with respect to the infinites-
imal strain tensor  = 1/2(u + (u)T ), with u the
The momentum balance, neglecting inertia, reads: displacement vector of the solid:

with the Cauchy stress tensor. Mass balance of The energy function is the sum of the poroelastic strain
constituent requires: energy and a mixing energy:

with the unstrained volume fraction of constituent


and vs the unstrained volume flux of constituent
with respect to the solid. In this paper denotes all
constituents, i.e. electrically charged solid (s), water
(w), cations (+) and anions (); , , cover all con-
stituents except the charged solid matrix. The mixture
is assumed to be fully saturated giving the following
relation for the saturation condition:

The form (10) of W assumes linear isotropic elasticity


in which vs is the solid velocity, and M are the Biot and Donnan osmosis as the swelling mechanism of the
coefficients. A negligible departure from electroneu- shale. G is the shear modulus, is Poissons ratio, Kf

trality is considered for mathematical reasons. This is the volumetric modulus of the solution, 0 is the
results in the following expression for the electroneu- reference potential of constituent , R is gas constant,
trality condition: T the absolute temperature and  the osmotic coeffi-
cient. Given this expression for the energy function,
the electrochemical potentials (8) take the form:

with F, , z and V Faradays constant, capaci-


tance, the valence and molar volume of constituent
respectively.

58
Copyright 2005 Taylor & Francis Group plc, London, UK
throughout the sample and for all times after applica-
tion of the load. Substituting eq. (17) into eqs. (1113)
yields:

in which

or,

3 THE PRESSURE TRANSMISSION TEST

Figure 1 illustrates the experimental setup considered.


The shale sample is placed in a sleeve. Upstream, z = 0,
the sample is in contact with a solution of sodium chlo- in which C = (2 W / + )1 . We substitute
ride at a given pressure pz . Downstream, z = L, the eq. (20) in the electroneutrality condition (6):
shale is resting (uz = 0) on a closed reservoir filled
with sodium chloride solution. The outer surface of
the sleeve is subject to a pressure pr . The volume of
the closed reservoir is assumed infinitely small com-
pared to the size of the shale sample. The momentum
balance (3) is written in the axysymmetric case: Substitution of the frictional equation (7) into the mass
balance equation (4) yields:

Equation (21) is transformed into:

If we neglect the shear stresses, we find for the pressure


transmission test: with:

r
z pz

upstream reservoir
Substitution of equations (23) and (1113) into (22)
z=0 gives the diffusion equation:
pr

sample rubber sleeve

z=L with,
downstream reservoir

Figure 1. Experimental setup.

59
Copyright 2005 Taylor & Francis Group plc, London, UK
4 THE BOUNDARY AND INITIAL method of separation of variables a relation for
CONDITIONS the electro-chemical potentials is obtained, with

0 = (0, t):
A change in the external concentration is considered
along the upstream surface of the sample. The electro-
chemical potentials for the sodium chloride solution
in the upstream reservoir are equal to:

with Rn a diagonal matrix with on its diagonal ele-


ments:

with ce+ and ce the external concentration of cations


and anions respectively. The change in boundary con-
ditions at the upstream reservoir (z = 0) and at the
lower reservoir (z = L) corresponding with a stepwize and M the matrix transforming the normal variables
change in salt concentration in the upper reservoir is: i to the original variables ,+ , . The pressure in
the shale and in the reservoirs is given as a function of
the electrochemical potentials:

with

The ionic volumes are given by:

while the electrical potential is

The fixed charge density ccf , closely related to the


f
with ce+ , ce
the change in upstream concentra- cation exchange capacity CEC = 0 c fc /0s , is its
tion of cations and anions and t0 and t0+ time just local value in the shale, while it is zero in the reservoir.
before and after t0 when chemical loading is applied. The pressure, the ionic concentrations and the electri-
The change in the external concentration is considered cal potential exhibits a jump at the shale boundary. In
sufficiently small such that C and K are considered particular, the pressure in the downstream reservoir is
constant in time. given by
Since inhomogeneous boundary conditions on
are present at z = 0 together with homogeneous initial
conditions a new variable is introduced as,
The permeability matrix K is the inverse of a
frictional matrix:
giving a set of homogeneous boundary conditions
and inhomogeneous initial conditions. Utilizing the

60
Copyright 2005 Taylor & Francis Group plc, London, UK
11.8025 the sample. At the downstream end of the sample, the
axial displacement, the pore pressure, the downstream
11.802
pressure (Fig. 2), the chemical potential, the electro-
chemical potentials, the electrical potential (all show a
downstream pressure [MPa]

dual response indicating at least two time constants in


11.8015 the system: one in the order of 50100 hours, the other
in the order of 1000 hours. In addition many variables
11.801
change instantaneously after application of the load.

11.8005
6 DISCUSSION

11.8
0 50 100 150 200 250 300 350 400 450 500
The analytical solutions are presented for elec-
time [h] trochemomechanical analysis of the pressure trans-
mission experiment. The literature clearly relates
Figure 2. Downstream pressure at z = L. osmotic pressures in shales with ionisation of the clay
platelets. Unlike previous models of pressure trans-
which in turn depends on the hydraulic permeability mission, which associate the osmosis of the shale with
K, the ion diffusion coefficients D+ and D , and the a chemical phenomenon (Sherwood 1993; Detournay
formation factor f : et al. 2005), this model introduces the ionisation as
the source of the osmotic forces observed. Hence,
the experimental parameters needed are closer to the
physics of the system. The cation exchange capacity
of the shale stand therefore as a key input parameter
that governs the osmotic swelling and shrinking. Stan-
dard techniques are available to measure the cation
exchange capacity. At the output side of the model,
the electrical potential is computed that is a measur-
able quantity as well, and provides a sensitive check
for the correct description of the ionic diffusion-
convection in the shale. An important spin-off of the
inclusion of the ionisation of the shale into the system
of equations is that phenomena of electro-osmosis,
electrophoresis, streaming potentials and streaming
currents are included as well. These phenomena are
well-documented and have potential technological
applications in high specific surface soils (Yeung and
Mitchell 1993). The electrochemomechanical model
discussed here extends the model developed by (Yeung
and Mitchell 1993) to include deformation of the shale.
5 RESULTS The linearised equations are reduced to three diffu-
sion equations. To each of these equations one diffu-
For the material parameters considered the three eigen- sion coefficient is associated (eq. 40). Each of these
values equal: three eigenvalues is associated with a specific time
constant equal to the size of the sample squared divided
by the eigenvalue. Physically, one can identify the three
eigenvalues as the electric potential diffusion coeffi-
The upstream salt concentration is increased cient, the pressure diffusion coefficient and the ionic
with a small amount from ce+ = ce = 1.000 mol/l to diffusion coefficient (van Meerveld et al. 2003). The
ce+ = ce = 1.001 mol/l, with boundary condition given first coefficient is infinite if strict electro-neutrality is
in relation (28). At the upstream boundary, z = 0, handled. Because we included some capacitance  into
the computation typically shows no change of pres- the system, its value is large but finite. (van Meerveld
sure or electrochemical potentials because of the et al. 2003) showed that the capacitance does not affect
prescribed boundary conditions. The pressure, the the macroscopic behavior, provided its order of magni-
electrical potential and the concentrations, however, do tude is sufficiently low. All output values (Fig. 2) show
not take their values in the upstream reservoir. There a dual response indicating at least two time constants
is a jump of these quantities across the boundary of in the system: one in the order of 50100 hours, the

61
Copyright 2005 Taylor & Francis Group plc, London, UK
other in the order of 1000 hours. In fact they are asso- end of the spectrum and should therefore be very
ciated with the second eigenvalue (L2 /2 70 h) and sensitive to electrical potential gradients.
the third eigenvalue of eq. 40 (L2 /2 1300 h). The
initial decrease of the downstream pressure and its
subsequent recovery is consistent with experimental REFERENCES
observations (van Oort 1997).
The discontinuous jumps of the concentrations Biot, M. A. (1956). General solutions of the equations of elas-
across the interfaces between the sample and the ticity and consolidation. Journal of Applied Mechanics 78,
reservoir are at odds with earlier models of pres- 9196.
sure transmission (Sherwood 1993), (Detournay et al. Casagrande, L. (1952). Electro-osmotic stabilization of soils.
2005), but in agreement with elementary electrochem- Journal of the Boston Society of Civil Enginmeers 20,
istry. The sharp reduction of the shale pore pressure, 120.
Detournay, E., J. Sarout, and J. Caurel (2005). Chemoporoe-
following a contact with an environment of low chemi-
lastic parameter identification of a reactive shale. In
cal potential, has been observed repeatedly in the field J. Huyghe, P. Raats, and S. Cowin (Eds.), Physicochemi-
(Neuzil 2000). The importance of electrical potential cal and Electromechanical Ineractions in Porous Media,
gradients relative to concentration gradients and pres- pp. 12. Kluwer Academic Publishers, Dordrecht, The
sure gradients is related to the ratio between the fixed Netherlands.
charge density over the salt concentration with which Ekbote, S. and Y. Abousleiman (2005). Incorporating chem-
the shale is in contact. The upstream salt concentra- ical effectsin a porothermoelastic formulation and appli-
tions in pressure transmission experiments in water cation to inclined boreholes. In J. Huyghe, P. Raats, and
based muds is usually chosen according to water based S. Cowin (Eds.), Physicochemical and Electromechanical
Ineractions in Porous Media, pp. 12. Kluwer Academic
mud concentration in the field, typically in the range
Publishers, Dordrecht, The Netherlands.
of several moles per liter. The fixed charge density is Frijns, A., J. Huyghe, and J. Janssen (1997). A validation of
usually in the order of a few tenths of mole per liter. the quadriphasic mixture theory for intervertebral disc
This results in a fairly low contribution from electrical tissue. Int. J. Eng. Sci. 35, 14191429.
phenomena. In the field however, many shale forma- Heidug, W. and S. -W. Wong (1996). Hydration swelling of
tions have a relatively low salt concentrations, in other water-absorbing rocks: a constitutive model. Int. J. Num.
words they are equilibrium with a low salt concentra- Anal. Meth. Geomechanics 20, 403430.
tion in the order of magnitude of their own fixed charge Huyghe, J. M. and J. D. Janssen (1997). Quadriphasic
density, which is a order of magnitude lower than mechanics of swelling incompressible porous media. Int.
J. Eng. Sci. 35, 793802.
the typical mud concentration. An unusually complete
Lai, W. M., J. S. Hou, and V. C. Mow (1991, Aug). A tripha-
description of a shale formation is given in (Marine and sic theory for the swelling and deformation behaviors of
Fritz 1981). They describe a illite rich shale from a Tri- articular cartilage. J Biomech Eng 113, 245258.
assic sediment with a porosity of 0.035. They found Marine, I. and S. Fritz (1981). Osmotic model to explain
a osmotic pressure of 1.3 MPa for a total dissolved anomalous hydraulic heads.Water Resources Research 17,
solute concentration of 15 g/l against an external con- 7382.
centration of 1 to 2 g/l. These high osmotic pressures Mitchell, J. (1993). Fundamentals of soil behavior (2 ed.).
can be explained using a fixed charge density of the New York, United States: John Wiley & Sons, Inc.
order of 0.5 moleq/l, while the value of 15 g/l may at Molenaar, M., J. Huyghe, and P. Van den Bogert (1998). Con-
stitutive modeling of charged porous media. In SPE paper
most correspond to a corresponding equilibrium exter-
47332, SPE/ISRM EUROCK98,Trondheim, pp. 127132.
nal salt concentration of the order of a few tenths of Neuzil, C. E. (2000). Osmotic generation of anomalous
mole per liter. This shows a ratio of fixed charge den- fluid pressures in geological environments. Nature 403,
sity over salt concentration in the order of one in a 145150.
virgin formation. This indicates that electrical effects Sherwood, J. (1993). Biot poroelasticity of a chemically active
plays a mayor role in the virgin field, as well as away shale. Proc. Res. Soc. Lond. A 440, 365377.
from the borehole. Electro-osmotic permeability has van Meerveld, J., M. M. Molenaar, J. M. Huyghe, and
been shown to depend minimally on the pore size, F. T. P. Baaijens (2003). Analytical solution of compres-
while hydraulic permeability decreases many orders of sion, free swelling and electrical loading of saturated
charged porous media. Transport in porous media 50,
magnitude as pore size decreases, (Casagrande 1952).
111126.
The electro-osmotic permeability of fine sand is in van Oort, E. (1997). Physico-chemical stabilization of shales.
the order of 0.4108 m2 /Vs and for clays in the order In SPE International Symposium on Oilfield Chemistry,
of 0.5108 m2 /Vs. The corresponding hydraulic per- Houston, Texas, United States, pp. paper 37263.
meabilities are 109 m4 /Ns and 1015 m4 /Ns. Shale Yeung, A. T. and J. K. Mitchell (1993). Coupled fluid,
having a sub micron pore size is at the extreme lower electrical and chemical flows in soil. Geotechnique 1,
121134.

62
Copyright 2005 Taylor & Francis Group plc, London, UK

Você também pode gostar