Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
in poroelastic media
Dispersive surface acoustic waves
in poroelastic media
P ROEFSCHRIFT
door
Samenstelling promotiecommissie:
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Wave Propagation in Porous Media . . . . . . . . . . . . . 3
1.2.2 Attenuation Mechanisms . . . . . . . . . . . . . . . . . . . 3
1.2.3 Surface Waves in Porous Media . . . . . . . . . . . . . . . 4
1.2.4 Acoustic Borehole Modes in Porous Reservoirs . . . . . . . 5
1.2.5 Experimental Studies on Bulk and Surface Waves in Poroe-
lastic Media . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Objectives of this Thesis . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2. Wave Propagation in Fully Saturated Porous Media: Bulk Modes and Inter-
face Waves . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Dynamics of a Continuum Porous Medium: The Biot theory . . . . 9
2.1.1 Stress-Strain Relations . . . . . . . . . . . . . . . . . . . . 9
2.1.2 The Lagrangian Approach to the Biot Equations . . . . . . 10
2.1.3 Displacement Potential Formulation. Body Waves . . . . . 12
2.1.4 The Slow Compressional Wave . . . . . . . . . . . . . . . 14
2.2 Poroelastic Surface Waves along Plane Interfaces . . . . . . . . . . 16
2.2.1 Dispersive Surface Modes . . . . . . . . . . . . . . . . . . 17
2.2.2 Synthetic Seismograms: The Influence of the Leaky Modes 21
2.3 Surface Modes in Borehole Configurations: Approximate Models . 25
2.3.1 Rigid Borehole Wall . . . . . . . . . . . . . . . . . . . . . 26
2.3.2 Quasi-1D Model for a Permeable Borehole Wall . . . . . . 28
2.4 Appendix 2A: Matrix Coefficients for the Fluid/Porous Medium Plane
Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Samenvatting . . . . . . . . . . . . . . . . . . . . . . . . . 117
Contents ix
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . 120
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 129
In this thesis, the dispersive properties of the surface acoustic waves propagating
along poroelastic materials are theoretically and experimentally studied. The influ-
ence of the pore space saturation on the surface modes is numerically investigated in
the cases of plane and cylindrical interfaces. The theoretical formulation is based on
the two-phase Biot model which accounts for the frequency-dependent mechanical
interaction between the solid and the fluid constituents. Shock-induced wave exper-
iments on water-saturated porous samples were performed in order to validate the
theoretical and numerical findings.
The characteristics of the pseudo-Stoneley wave along boreholes in porous for-
mations are studied in a broad band of frequencies (100 Hz-200 kHz) using a dis-
placement potential formulation. The influence of the permeability on the phase
velocity, attenuation, radial displacement, and pore pressure is investigated and the
contribution of the different bulk modes to the average radial displacement is ana-
lyzed in the frequency domain. The numerical results indicate that the permeability
dependence at low frequencies is caused by the Biot slow wave, due to its relatively
high contribution to the displacements and pressures induced by the pseudo-Stoneley
wave.
Acoustic experiments were performed using a shock tube technique to excite the
pseudo-Stoneley wave in a water-filled borehole in a water-saturated Berea sand-
stone cylinder. Frequency-dependent phase velocities and damping coefficients were
measured using this technique. Quantitative agreement between the experimental re-
sults and the theoretical predictions is found for the phase velocity in the frequency
range from 10 to 50 kHz. The attenuation data are found to deviate from theory
in an oscillatory pattern, and only qualitative agreement is reached. In order to im-
prove the quality of the attenuation data, the shock tube was modified. A funnel-like
structure was mounted inside the tube just above the sample in order to channel the
acoustic energy into the bore and enhance the excitation of the surface modes. More-
over, a new FFT-Prony-Spectral Ratio method is implemented to transform the data
from the space-time domain to the frequency-wavenumber domain. For the Berea
sandstone, a significant improvement in the accuracy of the attenuation data and an
excellent agreement with theoretical predictions at low frequencies is found. The
comparison of the experimental results and the numerical calculations shows that
xii SUMMARY
the oscillating fluid flow at the borehole wall is the dominant loss mechanism for
the pseudo-Stoneley wave and that it is properly described by the Biot theory at fre-
quencies below 40 kHz. A higher frequencies, a systematic underestimation of the
numerical predictions is found, which can be attributed to the existence of other loss
mechanisms neglected in the Biot model.
Higher-order guided modes associated with the compressional wave in the porous
formation and the cylindrical geometry of the shock tube were excited in the exper-
iments, and detailed information was obtained on the frequency-dependent phase
velocity and attenuation in highly porous and permeable materials. The measured
attenuation of the guided wave associated with the compressional wave, reveals the
presence of regular oscillatory patterns that can be attributed to radial resonances.
This oscillatory pattern is also numerically predicted, although the measured attenu-
ation values are one order of magnitude higher than the theoretical ones. The phase
velocities of these higher-order modes are generally well predicted by theory.
In order to investigate the effect of partial gas saturation on the surface waves, the
Biot model is extended to include additional interactions between the distinct phases,
namely acoustic radiation, heat exchange and viscous dilatation. Within this theoret-
ical framework, a numerical study of the dispersive phase velocity and attenuation of
the surface waves along a liquid/partially saturated poroelastic plane interface reveals
an important dependence on the gas fraction. Increasing gas fraction causes increased
attenuation over the entire frequency range. Maximum values in the attenuation coef-
ficient of the pseudo-Stoneley wave are obtained in the 10-20 kHz frequency interval,
which is relevant for borehole logging purposes. The attenuation level and the char-
acteristic frequency of this maximum depend on gas fraction. In the high-frequency
limit, where only compressibility effects govern the dynamics of the surface modes, a
transition is found between a leaky pseudo-Stoneley wave and a true Stoneley mode.
This transition occurs at a typical gas fraction where the slow compressional bulk
wave and the acoustic wave in the liquid half-space have identical phase velocities.
The saturation effects are also studied for a borehole configuration. In this case
a poroelastic formation saturated by a brine-air mixture is considered. A clear de-
pendence of the damping of the pseudo-Stoneley wave on gas fraction is found. The
damping increases with gas fraction over the complete range of frequencies studied
(10 Hz-50 kHz). The interpretation of the results indicates that the compressibility
of the mixture governs the dynamics of the pseudo-Stoneley wave and accounts for
the saturation effects observed in the phase velocity, damping and pore pressure. A
quantitative analysis demonstrates the high sensitivity of the damping coefficient to
small amounts of gas in the pores. This effect not only holds for bulk modes, but for
surface modes as well.
1. INTRODUCTION
Surface elastic waves are quite common phenomena in nature. Perhaps the most
illustrative example follows from seismological observations. When an earthquake
occurs in the earth’s interior, part of the elastic energy radiates in the form of surface
seismic waves. Their energy is confined to a region close to the surface as the waves
spread out along the two-dimensional surface instead of the three-dimensional do-
main. These waves propagate almost unattenuated along the earth’s surface and are
responsible for most of the damage due to their large amplitude and associated ro-
tational movement (Aki and Richards, 1980). In this case, the wavelengths involved
are on the order of magnitude of meters to kilometers.
On the other hand, in ultrasound technological applications related to interdigital
transducers (IDTs) and more recently to material characterization by means of laser-
generated surface acoustic waves (SAW), the typical wavelengths are on the microm-
eters scale (Maznev et al., 2003). In fact, the study of surface elastic waves covers
a broad band of wavelengths and frequencies as illustrated in Figure 1.1. Allard
and coauthors (2003; 2004) reported studies on laser-irradiated surface waves along
porous materials. The analysis of the surface deformation caused by the surface mode
provides information on the mechanical properties of the porous solid. Among other
applications, the inversion of surface wave data for subsurface imaging is relevant
for seismological studies in general and in shallow geophysics for local soil charac-
terization. The wide spectrum of applications extends to bone biomechanics where
measurements of the distribution of material elastic constants of trabecular bones
and anisotropy in cortical bones have been lately carried out with promising results
(Jorgensen and Kundu, 2002; Raum et al., 2004).
Particularly in the frequency band relevant to seismic and borehole exploration
(10 Hz- 30 kHz) it is challenging to exploit at maximum the information provided
by the surface waves in order to obtain an accurate reservoir characterization. In
this respect there has been an increasing interest in understanding the fluid effects
on wave propagation over the last years. Most of the work is focused on the flow-
induced loss mechanisms in the compressional P wave. Scarce studies on surface
waves have shown the great sensitivity of these kind of waves to relevant parameters
2 1. Introduction
W a v e le n g th F re q u e n c y
k m In fra so u n d 1
S e is m ic W a v e s
1 0 2
m A u d ib le W e ll lo g g in g
so u n d 1 0 4
c m
m m 1 0 6
U ltra s o u n d ID T s
1 0 8
S A W p u ls e s
m m S A W g ra tin g s
1 0 1 0
[H z ]
Figure 1.1: Surface (elastic) acoustic waves are found over a broad frequency spectrum. Cur-
rent research extends from seismic waves in the infrasound region to interdigital
transducers (IDTs) and to laser-generated SAW (surface acoustic waves) pulses
and transient gratings in the ultrasound region (adapted from Hess (2002)).
In this section a literature review is presented for the research on propagation of bulk
and surface waves in poroelastic materials. Special attention is devoted to the exam-
ination of the existing knowledge on attenuation mechanisms and the propagation of
acoustic borehole modes in porous formations. The experimental findings and the
experimental techniques are reviewed in detail with emphasis on shock wave tech-
niques.
The first investigations on surface wave propagation in poroelastic materials are those
of Deresiewicz (1960), (1961) and Deresiewicz and Skalak (1963). In this series of
papers, the boundary conditions at the interface between a fluid and a fully saturated
porous material were derived. Obviously these conditions govern the dynamics and
the propagation of the surface modes. The most important difference with respect
to the elastic case is the existence of the second compressional wave, which adds an
additional degree of freedom and therefore an extra boundary condition is required to
completely describe and close the problem. Deresiewicz and Skalak linked this extra
boundary condition with the fluid pressure continuity at the boundary and introduced
the concept of surface permeability, which controls the pressure discontinuity at the
interface.
1.2. Literature Review 5
Feng and Johnson (1983a,b) applied these boundary conditions to study the ve-
locity and attenuation of surface waves propagating along air/porous and liquid/po-
rous plane interfaces. Their investigations are on the high-frequency limit of the Biot
theory where the second compressional wave, as well as the first compressional wave
and the shear wave, propagate without attenuation. They found that there are three
different surface waves that may propagate depending on the stiffness of the porous
medium and the surface permeability. Two of them, the pseudo-Rayleigh wave and
the pseudo-Stoneley wave are the generalization of the pseudo-Rayleigh wave and
the Stoneley wave which propagate along a liquid/elastic plane interface. In this case
the pseudo-Rayleigh wave radiates energy into the liquid. It is important to remark
that in the porous case the pseudo-Stoneley wave becomes attenuated because of ra-
diation into the second compressional wave. A new surface mode is predicted which
propagates slower than the second compressional wave and is denominated the true
Stoneley wave.
Recently, the influence of the Biot frequency-dependent viscous dissipative mech-
anisms on the propagation of surface waves for a very high-permeable porous medium
have been studied by Gubaidullin et al. (2004). Their numerical results for the
frequency-dependent phase velocity and attenuation of the surface waves indicate
that the true Stoneley wave becomes attenuated due to viscous friction. This loss
mechanism also governs the attenuation of the pseudo-Stoneley wave and the pseudo-
Rayleigh wave, particularly at low frequencies. Studies of surface wave propagation
along porous surfaces have also been carried out based on other models than the Biot
theory (Edelman and Wilmanski, 2002; Edelman, 2004).
may exist on the wall of the borehole. The influence of a thin impairment layer on
the wall of the borehole on the dispersion relation of the (pseudo) Stoneley wave was
analyzed by Tichelaar et al. (1999).
The first experiments which unambiguously linked a slower arrival with the second
compressional wave were reported by Plona (1980). The experiments were per-
formed in water-saturated sintered glass porous materials at ultrasonic frequencies
using a mode conversion technique. Using a shock tube technique, van der Grin-
ten et al. (1985), (1987) performed experiments in air- and water-saturated porous
columns consisting of agglutinated sand particles. They were able to detect the sec-
ond compressional wave. Using this method also the slow compressional wave in
water-saturated natural rocks was detected (Smeulders, 1992; Brown et al., 2000).
Nagy et al. (1990) successfully measured the second compressional wave in air-
saturated natural rocks. Using ultrasound, experimental evidence of the slow wave in
water-saturated rocks was also provided by Kelder and Smeulders (1997) and Kelder
(1998).
An interpretation of laboratory velocity measurements in a wide range of partially
gas-saturated rocks is given by Gist (1994b). Cadoret et al. (1995), (1998) present
experimental results using a resonant-bar technique to determine the velocity and
attenuation of acoustic waves in partially saturated limestones at a sonic frequency
of 1 kHz . Similar experiments were previously performed by Lucet et al. (1991).
Wave propagation experiments in porous media saturated by a mixture of water and
air were performed by Smeulders and van Dongen (1997).
Experimental data regarding surface waves along plane interfaces were reported
by Mayes et al. (1986) and Nagy (1992). The aim of their investigations was to detect
the new surface mode predicted in the earlier work of Feng and Johnson (1983a).
Recently, Allard et al. (2002), (2003), (2004) reported experimental data on surface
wave propagation along poroelastic materials using a laser-based technique. Wisse
et al. (2002) reported data for shock-induced guided waves along the outer surface of
water-saturated porous cylinders.
Laboratory data for surface waves along boreholes in porous formations are
scarce. Winkler et al. (1989) measured phase velocities and damping coefficients
of (pseudo) Stoneley waves in porous samples saturated with silicone oil. Hsu et al.
(1997) studied the influence of a cylindrical permeable mandrel on the tube wave
in an elastic formation. They performed measurements using stainless steel and
polyethylene as formations.
1.3. Objectives of this Thesis 7
From the literature review presented in the previous section it is obvious that there are
several open questions regarding the propagation of surface waves along poroelastic
surfaces. In the borehole case, the cause for the dependence of the pseudo-Stoneley
wave on the permeability of the formation is not fully understood. Moreover there is
an increasing demand for accurate independent experimental data to validate the the-
oretical models, since the only set of experimental data available is mostly restricted
to frequencies above 10 kHz (Winkler et al., 1989). In this thesis a shock wave tech-
nique is applied to study the propagation of borehole modes in poroelastic formations.
One of the advantages of this method is the broad band of frequencies involved in the
experiments (1 kHz- 150 kHz). We aim to obtain accurate frequency-dependent data
of the phase velocities and attenuation coefficients of the surface waves. These results
are compared with numerical calculations based on Biot’s theory in order to assert the
validity of the Biot loss mechanisms to describe the attenuation of the surface waves.
Also a theoretical analysis is performed to quantitatively explain the dependence of
the frequency-dependent pseudo-Stoneley wave on the permeability.
The study of the dissipative mechanisms induced by the presence of gas in the
pore space is relevant to many applications. It has been proven that even small frac-
tions of gas significantly affect the acoustic properties of a porous medium. The main
efforts so far have been conducted towards a description of the bulk acoustic proper-
ties of partially saturated porous materials. In this thesis the analysis is extended to
consider the influence of distributed bubbles on the properties of the surface waves.
Plane and cylindrical interfaces between a liquid and a partially saturated porous
medium are considered. Results on the flat interface problem are relevant to explo-
ration seismology in layered one-dimensional structures. Applications concerning
the cylindrical interface configuration are obviously related to borehole geophysics
and acoustic borehole logging techniques.
We also aim to study the surface waves in the time domain. An algorithm is
developed to calculate the reflected pressure field for a spherical wavefront imping-
ing upon a liquid/poroelastic flat interface. In this way we can compute synthetic
seismograms and analyze the influence of the surface waves in an actual waveform.
A review of the Biot theory is given in Chapter 2 where the main features of the slow
compressional wave are highlighted. Surface waves along liquid/porous plane inter-
faces are also discussed in this chapter in terms of the frequency-dependent phase
velocities and attenuations. The pressure response in the time domain for the re-
flected pressure field due to an incident spherical waveform is numerically solved
8 1. Introduction
Kb
1−φ−φKs
Ks
Q= , (2.4)
Kb Ks
1−φ− +φ
Ks Kf
φ2 K s
R= . (2.5)
Kb Ks
1−φ− +φ
Ks Kf
The above expressions for the generalized elastic coefficients are obtained through
‘Gedanken’ experiments (see e.g. Allard (1993)).
where η is the viscosity of the pore fluid and k0 is the steady-state permeability. The
coupled equations of motion are derived from the Hamilton principle which implies
that the Lagrangian should satisfy the following equation:
∂ ∂L ∂ ∂L ∂L ∂D
+ − + = 0. (2.12)
∂t ∂ q˙i ∂xj ∂ ∂qi ∂qi ∂ q˙i
∂xj
In this case the generalized coordinates qi are u and U. The equations obtained are:
∂2u ∂2U
ρ11 + ρ 12 = P ∇ (∇ · u) + Q∇ (∇ · U) − N ∇ × ∇ × u +
∂t2 ∂t2
ηφ2 ∂U ∂u
+ − (2.13)
κ0 ∂t ∂t
and
∂2u ∂2U ηφ2 ∂U ∂u
ρ12 2 + ρ22 2 = R∇ (∇ · U) + Q∇ (∇ · u) − − . (2.14)
∂t ∂t κ0 ∂t ∂t
where the density terms are now written in their generalized frequency-dependent
form:
ρe11 = (e
α − 1) φρf + (1 − φ) ρs , (2.17)
ρe12 = (1 − α
e) φρf , (2.18)
and
ρe22 = α
eφρf . (2.19)
iηφF (ω)
α
e(ω) = α∞ + , (2.20)
k0 ωρf
12 2. Wave Propagation in Fully Saturated Porous Media
where F (ω) is the viscodynamic operator describing the interaction between the pore
fluid and the solid matrix. Johnson et al. (1987) argued that the simplest possible
expression for F (ω) is:
s
iωρf α∞ κ0
F (ω) = 1− . (2.21)
2ηφ
For the sake of completeness we also give the expression for the frequency-
dependent permeability κ(ω), which reads,
iηφ
κ(ω) = . (2.22)
α(ω)ωρf
u = ∇Φs (2.23)
and
U = ∇Φf . (2.24)
The above potentials describe irrotational waves associated with the solid, Φ s ,
and the fluid, Φf . In terms of the potentials, the Biot equations in the frequency
domain can be written in a matrix form as follows:
−ω 2 ρΦ = M∇2 Φ, (2.25)
2.1. Dynamics of a Continuum Porous Medium: The Biot theory 13
ω2 h √ i
2
kc1 (ω) = P ρ
e 22 + Re
ρ 11 − 2Qe
ρ 12 − ∆ (2.28)
2 (P R − Q2 )
and
ω2 h √ i
2
kc2 (ω) = P ρ
e 22 + Re
ρ 11 − 2Qe
ρ 12 + ∆ , (2.29)
2 (P R − Q2 )
where ∆ is:
∆ = (P ρe22 + Re ρ12 )2 − 4 P R − Q2
ρ11 − 2Qe ρe11 ρe22 − ρe 212 . (2.30)
The potentials Φs and Φf are related to the eigenvectors Φc1 and Φc2 , associated
with the fast and slow P wave respectively, by the following relations:
and
P − vc1 2 ρ
e11
Gc1 = 2 (2.35)
vc1 ρe12 − Q
and
P − vc2 2 ρ
e11
Gc2 = 2 . (2.36)
vc2 ρe12 − Q
Shear waves are described by a vector potential Ψsh . The displacements induced
in the solid and fluid phase are respectively:
u = ∇ × Ψs , (2.37)
Figure 2.1: Velocity and specific attenuation of the slow compressional wave for a water-
saturated Berea sandstone. Biot’s theory predicts an important dependence of
the properties of this wave on the frequency and the permeability. The slow
wave is propagative at high frequencies and diffusive at low frequencies. Out-
of-phase movement between the fluid and the solid phase characterizes the slow
wave, therefore an increase in permeability results in lower values for the specific
attenuation and higher velocities.
16 2. Wave Propagation in Fully Saturated Porous Media
Surface waves are localized phenomena. They occur in the vicinity of an interface
between two different media. The energy of these waves is usually limited to the re-
gion close to the interface as these waves fade away exponentially when moving away
2.2. Poroelastic Surface Waves along Plane Interfaces 17
from the interface. The penetration in the material is on the order of the wavelength
involved. The properties of the surface modes depend on the properties of the two
media and the characteristics of the interface itself. In the case that one of the media
is a porous material the flow impedance of the surface plays a significant role. Feng
and Johnson (1983a,b) were the first to systematically study the velocity and attenua-
tion of the surface waves that propagate along a plane interface between a fluid and a
fully saturated porous medium. Their results are limited to the high-frequency limit
where the viscous effects are neglected and the slow compressional wave is unatten-
uated. In this limit there is also no influence of the permeability of the formation
on the features of the surface modes. Gubaidullin et al. (2004) included the viscous
effects and extended this study to consider the frequency-dependent properties of the
surface modes for the case of a plane interface between water and a water-saturated
highly permeable material (308 Darcy).
In this section the general procedure to compute the dispersion of the surface
modes is outlined. Then the surface waves for a water/Berea sandstone interface are
studied. Synthetic seismograms are obtained using a Sommerfeld integral formula-
tion for the incident and the reflected pressure fields.
The potentials associated to each of the bulk modes which propagate in the porous
half-space are:
where êy is the cartesian basis vector in the y− direction. The above potentials
describe waves that propagate parallel to the interface. The wavenumbers in the
18 2. Wave Propagation in Fully Saturated Porous Media
flu id
fu lly s a tu r a te d p o r o u s m e d iu m
Figure 2.2: Plane interface between a liquid and a fully saturated poroelastic medium. The
surface waves propagate along the interface in the x−direction and decay expo-
nentially in the z−direction.
z−direction are related to the horizontal wavenumber kx through the following rela-
tions:
s
ω2
γj = kx2 − 2 , j = 1, 2, sh, f, (2.46)
cj
and
where u refers to the displacement of the matrix and U to the displacement of the pore
fluid. In the liquid half-space, the displacement Uf is ∇Φf . Therefore the continuity
of average normal displacement at the interface,
(1 − φ) uz + φUz = Uf z , (2.49)
2.2. Poroelastic Surface Waves along Plane Interfaces 19
Using the Biot’s stress-strain relations (equations 2.1 and 2.2), the above equation
can be written in terms of the potentials as follows:
∂ 2 Φc1
[P − 2N + Q + Gc1 (Q + R)] ∇2 Φc1 + 2N + [P − 2N + Q +
∂z 2
∂ 2 Φc2 ∂ 2 Ψsh
+Gc2 (Q + R)]∇2 Φc2 + 2N + 2N = −ω 2 ρw Φf . (2.52)
∂z 2 ∂z∂x
The absence of tangential stress in the liquid requires τxz = 0 at the interface, and
this condition implies that:
2
∂ Φc1 ∂ 2 Φc2 ∂ 2 Ψsh ∂ 2 Ψsh
N 2 + + − = 0. (2.53)
∂z∂x ∂z∂x ∂x2 ∂z 2
Finally, the continuity of pressure leads to
1
− (Q + RGc1 ) ∇2 Φc1 + (Q + RGc2 ) ∇2 Φc2 = ρw ω 2 Φf . (2.54)
φ
Substituting equations (2.42-2.45) into equations (2.50,2.52-2.54) and after some
algebraic manipulations a linear system for the amplitudes of the potentials is found:
N(kx , ω) · a = 0, (2.55)
where the matrix N contains information about the mechanical properties of the fully
saturated porous medium and the water half-space and a is a vector containing the
amplitude of the wave potentials, aT = (A0 , Ac1 , Ac2 , B). The elements of the
matrix N are given in Appendix 2A. The surface modes satisfy the condition that the
determinant of N equals zero:
1800
1750 shear
p-Rayleigh
1700
Velocity (m/s)
1650
1600
1550
fluid
1500
1450
PSfrag replacements
1400
p-Stoneley
1350 3 4 5
10 10 10
Frequency (Hz)
(a)
1
10
slow
0
10
p-Rayleigh
−1
10
Q−1
−2
10
p-Stoneley
−3
PSfrag replacements 10
shear
−4 fast
10
3 4 5
10 10 10
Frequency (Hz)
(b)
Figure 2.3: Frequency-dependent velocities (a) and specific attenuations (b) of the surface
modes (solid lines). The velocities and specific attenuations of the bulk modes
are indicated in dashed lines. The material properties used in the numerical cal-
culations are given in Table 2.1
2.2. Poroelastic Surface Waves along Plane Interfaces 21
V (ω) = ωRe−1 (kx ), and specific attenuation coefficients are obtained. The spe-
cific attenuation coefficient, Q−1 (ω) = |2Im(kx )/Re(kx )|, represents the fraction of
energy dissipated into heat in an oscillation period.
There are three types of surface waves that may propagate on a flat interface be-
tween a fully saturated porous medium and a liquid: the Stoneley wave, the pseudo-
Stoneley wave and the pseudo-Rayleigh wave. The existence of these waves depends
on the mechanical properties of the porous material and the liquid and the perme-
ability of the interface. Feng and Johnson (1983a) have demonstrated that for the
completely open interface only the pseudo modes exist for hard rocks. The pseudo-
Stoneley wave leaks energy into the slow compressional wave while the pseudo-
Rayleigh wave radiates into both the slow compressional wave and the acoustic wave
in the fluid (z < 0). Due to radiation and viscous mechanisms, these waves are sig-
nificantly damped along the direction of propagation. Adler and Nagy (1994) mea-
sured the velocities and attenuation coefficients of the pseudo modes in diverse fluid-
saturated rocks at ultrasonic frequencies. Their results are consistent with the theo-
retical predictions of Feng and Johnson. In this subsection the frequency-dependent
phase velocities and specific attenuations of the pseudo modes are calculated for a
water/Berea sandstone plane interface. The open-pore boundary conditions are con-
sidered. The properties of the rock are listed in Table 2.1. The viscous effects are
considered in the complete range of frequencies studied (400 Hz-200 kHz).
Figure 2.3 shows the dispersion relation corresponding to the surface waves. The
two pseudo surface modes are found. The pseudo-Stoneley wave propagates slower
than the acoustic wave in the fluid at every frequency. The attenuation is larger than
that of the shear wave and has a maximum around 25 kHz. The velocity of the
pseudo-Rayleigh wave is very similar to the speed of propagation of the shear wave.
This pseudo wave has an attenuation which is almost frequency-independent. These
results are consistent with previous calculations carried out by Gubaidullin et al.
(2004). The results presented here correspond to a more realistic material in geo-
physical applications.
The purpose of this subsection is to study the reflected pressure field in the liquid
half-space due to the incidence of a spherical waveform, in order to investigate the
influence of the surface waves in the time domain. An acoustic monopole point
source M is located at a height z0 in the fluid as depicted in Figure 2.4. The spherical
wavefront is decomposed in cylindrical waves by means of the Sommerfeld integral.
For further details about this decomposition the reader is referred to Aki and Richards
22 2. Wave Propagation in Fully Saturated Porous Media
M
O
L iq u id - s a tu r a te d p o r o e la s tic m e d iu m
Figure 2.4: A spherical wave front impinges upon a liquid/poroelastic plane interface. The
reflected pressure response is calculated at the observation point O using a Som-
merfeld formulation for the reflected field.
Im (q )
p /2 R e (q )
G 0
Figure 2.5: The path of integration Γ0 in the complex plane used to evaluate the Sommerfeld
integral.
were studied by Wu et al. (1990), Santos et al. (1992) and Denneman et al. (2002).
The reflection coefficient follows from the boundary conditions at the water/porous
solid interface for oblique incidence and is obtained numerically at each frequency
of interest. Further algebraic manipulation of the integral in equation 2.59 enables us
to write the reflected pressure field as:
Z π
2
Pr = A(ω)e−iωt [ik0 R(θ)J0 (k0 r sin θ)eik0 cos θ|z+z0 | sin θ dθ +
0
Z ∞ p
+k0 R(ζ)J0 (k0 r ζ 2 + 1)ek0 |z+z0 |ζ dζ], (2.60)
0
where ζ = −i cos(θ).
The point source located at z0 is assumed to emit a Blackman-Harris acoustic
pulse of the form:
3
X 2πj 2 2πjt
A(t) = Bj cos , (2.61)
T0 T0
j=1
for t [0, T0 ] and 0 elsewhere. T0 is the duration of the pulse and B1 = −0.48829,
B2 = 0.14128, and B3 = −0.01168. We chose these coefficients in agreement with
the source used in the high-frequency investigations of Feng and Johnson (1983b).
The incident pulse is shown in Figure 2.6.
The complex amplitudes A(ω) in equation 2.60 are obtained through the Fast
Fourier Transform of the incident wave Pi . The modified Sommerfeld integrals of
24 2. Wave Propagation in Fully Saturated Porous Media
Pressure (a.u.) 1
−1
−2
−3
−4
PSfrag replacements −5
0 5 10 15 20
Time (µs)
equation 2.60 are then solved numerically for each frequency using a recursive adap-
tive Simpson quadrature algorithm. Finally the pressure traces are obtained via the
Inverse Fast Fourier Transform. This routine is implemented in a Matlab algorithm.
First the features of the reflected wave for the water-Berea sandstone interface are
investigated. The acoustic source is placed at z0 = -0.3 cm and the incident pulse has
a duration of 10 µs. The pressure response at z = -0.1 cm and r = 40 cm is depicted
in Figure 2.7. Both the source and the observation point are situated close to the
interface in order to improve the observation of the surface waves. The calculations
in the frequency domain discussed in the previous subsection showed that there are
two surface waves that can propagate along the interface in this case: the pseudo-
Stoneley wave and the pseudo-Rayleigh wave. In our computations, it is possible
to clearly recognize the pseudo-Stoneley wave. Also the head wave associated with
the fast compressional wave and the reflected fluid wave are observed. The arrows
indicating the arrivals in Figure 2.7 are calculated using the velocities values of the
different waves at the typical frequency of 150 kHz, which corresponds to the main
frequency of the incident wave. The pseudo-Rayleigh wave is not observed due to its
high attenuation. From the dispersion relation, the calculated value for the imaginary
part of the wavenumber predicts that the amplitude of the wave has decreased at least
to a factor of e−1 of its original amplitude for distances larger than 3 cm. The same
reasoning can be applied to the head wave associated to the slow compressional wave
2.3. Surface Modes in Borehole Configurations: Approximate Models 25
1.5
0.5
Pressure (a.u.)
−2
0.1 0.2 0.3 0.4 0.5
Time (ms)
Figure 2.7: Reflected pressure from a water/Berea sandstone plane boundary. The incident
pressure wave from the liquid is a Blackman-Harris spherical wavefront (Figure
2.6). The coordinates of the source are r0 = 0 and z0 = -0.3 cm. The observa-
tion point is placed at r = 40 cm and z = -0.1 cm. The properties of the Berea
sandstone are given in Table 2.1. Shear modulus N = 4.02 GPa.
which practically disappears when the travelling distances are beyond 6 cm.
We also investigated the behavior of the amplitude of the pseudo-Stoneley wave
as a function of the vertical position of the observation point z. The results for three
different values of z are shown in Figure 2.8. From the computed pressure signals it
can be concluded that the amplitude of the pseudo-Stoneley wave decreases when the
distance with respect to the interface is increased. We also remark that the fast wave
remains practically unaffected over the distances considered.
So far in this chapter only plane interfaces have been considered. In this section
the attention is focused on the study of the propagation of head and interface waves
along cylindrical interfaces. Practical applications are obviously related to hydro-
carbon prospecting. Acoustic borehole logging is a common technique where the
information carried by the guided modes, head and surface waves travelling along
the borehole wall is used to infer the properties of the porous formation and the satu-
rating fluid.
26 2. Wave Propagation in Fully Saturated Porous Media
z = -1 cm
4
Fast P P-Stoneley
3
Pressure (a.u.)
z = -0.5 cm
2
1 Fast P P-Stoneley
z = -0.1 cm
0
PSfrag replacements
−1 Fast P P-Stoneley
−2
−3
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Time (ms)
Figure 2.8: Pressure response at different values of z. The position of the source and the
properties of the porous medium are the same as in Figure 2.7, r = 40 cm.
where Acw is the amplitude of the wave, J0 the zero-order Bessel function, kr the
radial wavenumber, ω the angular frequency and k the wavenumber of the guided
waves in the z− direction. The term corresponding to the Hankel function is excluded
because it diverges at r = 0.
The relation between kr and k can be written as:
s
ω2
kr = − k2 , (2.63)
c2w
2.3. Surface Modes in Borehole Configurations: Approximate Models 27
8000
5000
4000
3000
PSfrag replacements
2000
1000
0
0
0 10 20 30 40 50 60 70 80 90 100
Frequency (kHz)
Figure 2.9: Analytical (solid lines) and numerical results (squares) for the phase velocities of
undamped waves in a fluid filled cylindrical cavity in a rigid solid. The fluid is
water, cw =1500 m/s. The radius of the cavity is 38.5 mm. Numerically only the
non-trivial modes are computed.
where cw is the speed of sound in water, 1500 m/s. The boundary condition at the
wall implies the radial displacement ur to vanish, which leads to:
s
ω2
ur (R) = −Acw −k 2 J1 (kr r)ei(ωt−kz) = 0, (2.64)
c2w
The solution k(ω) of equation 2.64 gives the dispersion relation. Phase velocities
cn (ω) can be obtained as the ratio between ω and k. Analytical solutions can easily
be found:
x c 2 −1/2
n w
cn (ω) = cw 1 − (2.65)
Rω
sound in water. The remaining waves are characterized by cut-off frequencies. The
cut-off frequency increases with the order of the mode. Figure 2.9 shows an excellent
agreement between numerical and analytical results.
In this subsection we illustrate the effects of the permeability on the borehole waves
using an approximate quasi-1D model. The cavity is now within a rigid perme-
able solid. Different authors have addressed this problem using slightly different
approaches. The model presented here resembles in some aspects the work done by
White (1983), Norris (1989) and Tang et al. (1991). Conservation laws are applied
to the bore-fluid and it is assumed that the exchange of volume fluid with the sur-
rounding formation occurs only in the radial direction and therefore is governed by
the radial component of the velocity. Figure 2.10 depicts the control volume CV for a
cylindrical fluid cavity. We assume that the borehole wall is permeable and the solid
d S
C V P o ro u s
d S r e s e r v o ir
liq u id
Figure 2.10: Borehole configuration: a liquid cylindrical cavity surrounded by a porous reser-
voir. The mass and momentum balance equation are applied in the control vol-
ume CV in order to obtain the dispersion relation of the surface wave. Radial
fluid volume exchange with the porous reservoir is considered.
matrix is rigid. The linearized mass balance equation for the fluid
∂ρ
+ ρ0 ∇ · v = 0, (2.66)
∂t
2.3. Surface Modes in Borehole Configurations: Approximate Models 29
Applying Gauss theorem in the second term and performing an integration over the
cross section yields:
∂ρ ∂vz 2
+ ρ0 + ρ0 vr (R) = 0. (2.68)
∂t ∂z R
The linearized momentum balance equation in the z−direction, after neglecting vis-
cous terms, is:
∂vz ∂p
ρ0 + ρ0 g + = 0. (2.69)
∂t ∂z
Equations 2.68 and 2.69 govern the dynamics of the surface wave. In order to find
the dispersion relation a travelling harmonic dependence is assumed for the velocity
vz = v0 ei(ωt−kz z) , and the pressure p = p0 ei(ωt−kz z) . The coupling to the properties
of the porous formation is expressed in this model by the radial component of the
velocity at the borehole wall vr (R) which satisfies the following boundary condition:
In the above equation vr,I is the radial velocity of the pore fluid in the rigid porous
formation. The subscript I is used to distinguish the dynamic quantities in the porous
medium from the ones in the borehole liquid.
In order to find an analytical expression for vr,I , we combine the mass balance
equation for the pore fluid
1 ∂ρI
+ ∇ · vI = 0, (2.71)
ρI ∂t
1 ∂ρI 1 ∂pI
= , (2.72)
ρI ∂t Kf ∂t
1 ∂pI
+ ∇ · vI = 0. (2.73)
Kf ∂t
30 2. Wave Propagation in Fully Saturated Porous Media
The relation between the pressure and the velocity of the fluid is given by the Darcy
law. A dynamic permeability is considered here in the way stated by Johnson et al.
(1987). The Darcy law then can be written:
κ(ω)
vI = − ∇pI . (2.74)
ηφ
Combining Darcy’s law and equation 2.73 we can write in cylindrical coordi-
nates:
2
∂pI ∂ pI 1 ∂pI
=D + , (2.75)
∂t ∂r2 r ∂r
where
κ(ω)Kf
D= . (2.76)
ηφ
d2 A(r) 1 dA(r)
+ − λ2 A(r) = 0, (2.78)
dr2 r dr
where λ2 = iω/D. This is the modified Bessel equation of zeroth order. Assuming
that the pressure is bounded when r increases the general solution is:
From Darcy’s law (equation 2.74) we finally obtain for the radial velocity in the
porous medium:
Substitution of the harmonic solutions for vz and pz together with vr,I into equa-
tions 2.68 and 2.69 yields the following system of equations:
iω 2ρ0 κ(ω)λ K1 (λR)
v0 + + p0 (−iρ0 kz ) = 0 (2.82)
c20 Rη K0 (λR)
g
v0 − ikz + p0 (iρ0 ω) = 0, (2.83)
c20
where c0 = (p0 /ρ0 )1/2 is the acoustic velocity in the liquid.
The surface mode is the nontrivial solution and its dispersion relation ω(k z ) is
found by imposing that the determinant of the associated matrix equals zero. The
following characteristic equation is obtained:
iρ0 g
ρ0 kz2 + kz + C(ω) = 0, (2.84)
c20
The frequency-dependent results for the phase velocity and the damping coeffi-
cient are shown in Figure 2.11. The material properties correspond to the synthetic
(glass) sample in Table 2.1. Two different values of the steady-state permeabil-
ity are compared and the effects of the frequency-dependent dynamic permeability
are analyzed with respect to the quasi-static model, which considers a frequency-
independent steady-state value for the permeability. This steady-state permeability is
the low-frequency limit of the dynamic permeability. Obviously at low frequencies
both models agree. Deviations are observed in the phase velocities for frequencies
higher than the critical frequency, 38.3 kHz for the 1 Darcy formation and 3.83 kHz
32 2. Wave Propagation in Fully Saturated Porous Media
1500
1250
Phase Velocity (m/s)
1000
750
500
κ0=1D
250
PSfrag replacements κ0=10D
0 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(a)
3
10
Damping coefficient (1/m)
2
10
1
κ0=10D
10
κ0=1D
0
10
PSfrag replacements
−1
10 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(b)
Figure 2.11: Phase velocity (a) and damping coefficient (b) for the surface wave predicted by
the quasi 1D model (solid lines). The results are compared with the output of
a steady state model where the permeability is taken equal to its low-frequency
value (dotted lines).
2.3. Surface Modes in Borehole Configurations: Approximate Models 33
for the 10 Darcy formation. The discrepancy is considerably larger when the per-
meability is increased. It is also interesting to note that while the quasi-static model
predicts significant differences in the entire range of frequencies (100 Hz-1 MHz), the
dependence on permeability in the full dynamic model is limited to relative low fre-
quencies and both curves converge for frequencies higher than 100 kHz. Interesting
features are observed in the damping coefficient. The quasi static model predicts a
monotonous increase of the damping coefficient with frequency. The dynamic model
predicts the existence of a maximum for the damping coefficient. The position of
the maximum in the frequency axis depends on the steady state permeability and
it decreases with increasing values of steady state permeability. The existence of a
maximum in the attenuation coefficient Q−1 has also been noticed by Norris (1989)
and Schmitt et al. (1988). It is generally accepted that the cause of this maximum
is a transition from a bore-dominated dispersion to a high frequency regime where
the surface wave behaves as in a plane interface since its wavelength is much smaller
than the bore radius. Note that the quasi-static model does not take into account
the inertial effects that exist at high frequencies. These effects are incorporated in
the dynamic formulation. The quasi static model predicts an increase on the damping
coefficient with increasing permeability due to pore fluid diffusion to the surrounding
medium, this diffusion eases at higher values of permeability.
34 2. Wave Propagation in Fully Saturated Porous Media
Appendixes to Chapter 2
2.4 Appendix 2A: Matrix Coefficients for the Fluid/Porous Medium Plane
Interface
In this appendix the elements nij of the matrix N of eq. 2.55 are explicitly given:
n11 = γf ,
n12 = γc1 (1 − φ + φGc1 ),
n13 = γc2 (1 − φ + φGc2 ),
n14 = −ikx (1 − φ + φGsh ),
n21 = 0,
n22 = 2N γc1 ikx ,
n23 = 2N γc2 ikx ,
n24 2 + k 2 )N ,
= (γsh x
n31 = ω ρf ,
2
3.1 Introduction
The presence of gas bubbles can dramatically influence the acoustic properties of a
liquid. The bulk modulus of the liquid becomes frequency-dependent and attenuation
effects arise due to oscillations of the bubbles (radiation) and heat transfer to the sur-
rounding liquid. It is particularly interesting to consider the problem of a gas-liquid
mixture filling the pore space of a porous medium. In this case, even more dissi-
pative mechanisms have to be taken into account, namely the interaction between
the gas and both the liquid and the solid elastic matrix. In the case that only liquid
saturates the pore space, the interaction between the liquid and the solid matrix can
be understood in terms of the Biot theory (Biot, 1956a,b). This theory was previ-
ously extended in order to include the effects of gas saturation on the bulk elastic
waves in partially saturated porous media by among others White (1975), Dutta and
Ode (1979a,b), Berryman et al. (1988), Smeulders and van Dongen (1997), Johnson
(2001) and Carcione et al. (2003).
A great deal of attention has been given to the influence of the gas saturation on
the velocities and attenuation of seismic waves since the pioneering work of White
(1975). The White model assumes the air fraction to be constituted as spherical
gas pockets distributed in a cubic array in the liquid-saturated porous medium. This
model will be referred to as the “gas pocket model”. Dutta and Ode (1979a,b) pro-
vided a more complete solution based on Biot’s theory for the bulk modulus of a
single bubble surrounded by a spherical fluid-saturated porous layer. Berryman et al.
(1988) formulated a model based on variational principles for the bulk acoustic prop-
erties of a porous medium saturated with a mixture of two fluids. Smeulders and van
Dongen (1997) carried out experimental and theoretical studies on compressional
wave propagation in porous columns saturated by a dilute air-water mixture. Their
theoretical model is based on the study of the response of a gas bubble in a fully satu-
rated porous medium to an external oscillating pressure field. Damping mechanisms
due to radiation into the two compressional waves, viscous dilatation at the bubble
surface and heat exchange with the solid matrix are considered. Degrande et al.
(1998) used this model to study the effects of saturation on the wave propagation
36 3. Surface Waves along Partially Saturated Porous Media
Acoustic wave propagation through a fully saturated porous media can be described
in terms of the Biot equations. In the frequency domain these equations are expressed
as (see also Chapter 2):
−ω 2 (e u + ρe12 U)
ρ11 e u + N ∇2 e
e = (P − N )∇∇ · e u + Q∇∇ · U
e (3.1)
3.2. Acoustic Properties of a Partially Saturated Porous Medium 37
and
ρ22 U
−ω 2 (e u) = R∇∇ · U
e + ρe12 e u,
e + Q∇∇ · e (3.2)
where u e is the solid displacement and U e is the fluid displacement. N is the shear
modulus of the composite material and P, Q and R are the so-called generalized
elastic coefficients. They are mathematically related to the porosity φ, the solid frame
bulk modulus Kb , the solid grain bulk modulus Ks , the pore-fluid modulus, Kf and
N through the so-called Gedanken experiments. The parameters ρe11 , ρe12 and ρe22 are
the complex-valued frequency-dependent densities. They are functions of the density
of the fluid ρf , the density of the solid ρs , the porosity φ and the frequency-dependent
tortuosity αe(ω). It is not the purpose of this section to review Biot’s theory and for
further details the reader is referred to Chapter 2 and classical books on the subject
(see e.g. Allard (1993), Bourbié et al. (1987)).
In our case, the pore space is saturated by a mixture of water and air. Therefore
new interaction mechanisms between the gas and the liquid and the gas and the solid
matrix have to be taken into account. The oscillations of the air bubbles will induce
radiation of the two compressional waves at the bubble surfaces. The liquid dilata-
tion at the bubble surface causes viscous attenuation. Finally, heat transport from the
bubble to the surrounding media is also considered. The dissipative phenomena men-
tioned above are described in terms of a complex-valued frequency-dependent bulk
modulus of the mixture of water and air by Smeulders and van Dongen (1997). In this
section we will review this theory and analyse its implications for the bulk modes. We
will include the compressibility of the solid grains which was neglected in the cited
paper. We will calculate the volume variation of a single bubble as a response to an
external oscillating pressure field. In this way the bulk modulus of the bubble can be
computed and, neglecting the interaction between the bubbles it will be considered as
the bulk modulus of the gas phase in the mixture, Kg (ω). The frequency-dependent
bulk modulus of the mixture, Kf (ω) is obtained through a modified Wood’s formula
(Wood, 1955):
1 s 1−s
= + , (3.3)
Kf (ω) Kl Kg (ω)
where Kl is the bulk modulus of the liquid phase and s is the liquid saturation. The
expression for Kf (ω) given in equation 3.3 differs from the original Wood’s formula
in which both the bulk modulus for the gas and liquid phase are constant.
Strictly speaking the original Wood’s formula is only valid for highly homoge-
neous mixtures and at frequencies sufficiently low so that the wavelengths are consid-
erably larger than the size of the heterogeneities. In this case it is possible to assume
that the external driving pressure is spatially homogeneous. This assumption is rea-
sonable for the range of bubble sizes considered in this work. At high frequencies
38 3. Surface Waves along Partially Saturated Porous Media
and
where,
P − vc1 2 ρ
e11
Gc1 = 2 , (3.6)
vc1 ρe12 − Q
and
P − vc2 2 ρ
e11
Gc2 = 2 . (3.7)
vc2 ρe12 − Q
In the above equations vc1 , vc2 and vsh refer to the frequency-dependent wave veloc-
ities of the fast wave, slow wave and shear wave respectively.
Assuming an eiωt temporal variation the linearized radial momentum equation
for the liquid phase can be written as follows:
∂Ur ∂pf ∂ur ∂Ur
iωφρf = −φ + (eα(ω) − 1) iωφρf − . (3.8)
∂t ∂r ∂t ∂t
The above equation is integrated from the bubble radius (r = a) to infinity in order
to find an equation of motion for the bubble, which reads:
φρf ω 2 (Gc1 Φc1 + Gc2 Φc2 ) = −φ(pf ∞ − pf a ) + ω 2 φρf (e
α(ω) − 1)
[Φc1a (1 − Gc1 ) + Φc2a (1 − Gc2 )]. (3.9)
3.2. Acoustic Properties of a Partially Saturated Porous Medium 39
The solution for the potentials outside the bubble can be expressed as:
Ac1 e−ik1 r
Φc1 = , (3.10)
r
and
Ac2 e−ik2 r
Φc2 = , (3.11)
r
where k1 and k2 are the radial wavenumbers associated with the fast and slow com-
pressional waves respectively. Then, substitution of the solutions given by Equations
3.10 and 3.11 into Equation 3.9 leads to a relation between the amplitudes A c1 and
Ac2 . The boundary conditions at the bubble surface provide the remaining relations
to close the problem. Inside the bubble we neglect the interaction between the air
and the solid matrix and the matrix is considered as acoustically compact. It can be
shown that this condition implies that the velocity of the solid phase linearly depends
on r. We assume continuity of the radial velocity of the solid phase and its radial
derivative across the bubble surface. This leads to the following relation:
∂ 2 ur + 1 ∂ur +
(a ) = (a ), (3.12)
∂r∂t r ∂t
which holds at the outside of the bubble (a+ ). The continuity of fluid volume provides
an equation for the change in the volume of the gas bubble ∆V g in terms of the fluid
and solid displacements at the bubble surface:
Finally we consider that the pressure difference across the bubble surface is balanced
by the radial viscous stress in the fluid at the bubble surface:
4 ∂ 2 Ur
pf (a+ ) − pg = η (a), (3.14)
3 ∂r∂t
where pf denotes the pressure outside the bubble evaluated at the bubble radius and
pg is the gas pressure inside the bubble.
Substitution of the expressions for Φc1 and Φc2 in the boundary conditions and
in the equation of motion followed by some algebraic manipulations lead to the fol-
lowing relation between the volume of the air bubble and the external pressure p f ∞ :
a1 b2 − a2 b1 4 iηφa1 a2 (Gc2 − Gc1 ) Vg
ω 2 ρf + = pf ∞ − pg , (3.15)
a1 c2 − a2 c1 3 ωρf (a1 c2 − a2 c1 ) 4πφa
40 3. Surface Waves along Partially Saturated Porous Media
where
1 + ikj a
aj = kj2 (1 − 3 ), (3.16)
kj2 a2
ρe12 + φρf α
e(ω)
bj = φGcj − (Gcj − 1) , (3.17)
ρf
and
These relations were also derived by Smeulders and van Dongen (1997).
The last dissipative mechanism considered in this model is the thermal damping.
It arises due to the heat exchange between the gas phase and the solid matrix induced
by the oscillations of the bubble. Its contribution to the bulk modulus of the gas phase
can be expressed as npg . Here we have introduced a complex-valued polytropic
coefficient, n:
−1
1/2
coth ψ (8α k
∞ 0 /φ) 1
n = γ 1 + 3 (γ − 1) 1/2
− 2 (3.19)
ψ (8α∞ k0 /φ) ψ (8α∞ k0 /φ) 1/2
where ψ = (1 + i)(ω/2ag ), ag being the thermal diffusivity of the gas and γ the
specific heat ratio of the gas (for air γ = 1.4). Champoux and Allard (1991) and
Henry et al. (1995) reported a different expression for the polytropic exponent n. It
can be shown that both expressions are similar and have the same features at the low-
and high-frequency limits.
Finally, the following expression is found for the frequency-dependent bulk mod-
ulus of the gas phase Kg = −Vg (∂Vg /∂pf ∞ )−1 :
1 2 2 3npg a1 b2 − a2 b1 4 iηφa1 a2 (Gc2 − Gc1 )
Kg (ω) = a ω ρf − − . (3.20)
3 a 2 ω 2 ρf a1 c2 − a2 c1 3 ωρf (a1 c2 − a2 c1 )
Substitution of Kg into equation 3.3 leads to the frequency-dependent bulk mod-
ulus of the mixture saturating the void space of the porous material. This is the ex-
pression that will be used for the calculations of the dispersion relation of the surface
waves. Figure 3.1 shows the absolute and phase values of the bulk modulus of the
mixture as a function of the frequency for different liquid saturations. The properties
of the porous material and the saturating fluids are given in Table 3.1. On one hand,
at low frequencies, the bulk modulus of the gas phase equals 0.01 GPa and therefore
a decrease in liquid saturation causes a decrease in the bulk modulus of the mix-
ture since Kl > Kg . On the other hand, at high frequencies the gas phase becomes
3.2. Acoustic Properties of a Partially Saturated Porous Medium 41
1.4
1.2
s=1
1
|Kf | /Kwater
s = 0.999
0.8
0.6
0.4
PSfrag replacements s = 0.99 s = 0.8
0.2 s = 0.90
s = 0.95
0 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(a)
120
100
Phase (degrees)
80
s = 0.8
60
s = 0.90
40 s = 0.95
PSfrag replacements
20
s = 0.99
s=1 s = 0.999
0 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(b)
Figure 3.1: Frequency-dependent bulk modulus for a mixture of water and air saturating a
Berea sandstone porous rock. The radius of the air bubbles is 1 mm and the gas
pressure is 0.01 GPa (100 bars). Different liquid saturation s are considered.
42 3. Surface Waves along Partially Saturated Porous Media
Table 3.1: Physical properties of the Berea sandstone and the saturating fluids: water and air.
3500
fast P
Phase Velocity (m/s) 3000
2500
2000
shear
1500
1000
slow P
PSfrag replacements 500
0 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(a)
1
10
0
10
slow P
−1
10
Q−1
−2
10
fast P (s=0.95)
shear
−3
10
PSfrag replacements fast P
−4
10
−5
10 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(b)
Figure 3.2: Phase velocities (a) and attenuation coefficients (b) of the body waves in a
water/air-saturated Berea sandstone. The effects of air saturation are shown for
the compressional waves in solid lines, the gas pressure is 0.01 GPa, the bubble
radius 1 mm and s = 0.95. The Biot predictions for the fully saturated case are
shown in dotted lines.
44 3. Surface Waves along Partially Saturated Porous Media
w a te r
p a r tia lly s a tu r a te d
B e re a s a n d s to n e
and Dutta and Ode (1979a). They considered a cubic array of air bubbles in a porous
medium and computed the bulk modulus of a single bubble surrounded by a liquid-
saturated spherical shell (see Figure 3.7). The volume of the individual cell consti-
tuted by the gas bubble and the shell is identical to the dimensions of the original
cubic cell. We will discuss this model in detail in the Appendix 3A. The main differ-
ence with the model described in this section are the newly introduced thermal and
viscous mechanisms. The influence of these mechanisms is discussed in Appendix
3B.
In this section we discuss the numerical results for the different surface wave modes.
The configuration is depicted in Figure 3.3. The mathematical procedure to obtain the
dispersion relation of the surface waves is outlined in section 2.2.1. The oscillating
gas bubble model is used to describe the bulk modulus of the fluid phase, which in this
case is constituted by a mixture of water and air. We adopt the surface wave terminol-
ogy given by Feng and Johnson (1983a). In order to avoid confusion, it is worthwhile
to mention that the pseudo-Stoneley wave propagating along a liquid/poroelastic in-
terface is the generalization of the classical Stoneley wave in a liquid/elastic interface.
In the poroelastic case it becomes a pseudo wave due to the radiation into the slow
P wave. It is important to note that in this work we will assume that the interface is
fully permeable so that continuity of pressure holds across the interface. The effect
of sealed or partially sealed pores at the interface has been modelled in the past using
the empirical concept of surface flow impedance. We restrict ourselves to the open
3.3. Saturation Effects on the Velocities and Attenuation of the Surface Waves 45
2000
1900
1800 shear
Phase Velocity (m/s)
1700 p-Rayleigh
1600
fluid
1500
Stoneley
1400 p-Stoneley
1300
PSfrag replacements
1200
1100
1000
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
s
(a)
0
10
p-Rayleigh
−1
10
−2
10
Q−1
−3 p-Stoneley
10
−4
10
−5
10
PSfrag replacements
−6
10
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
s
(b)
Figure 3.4: Saturation effects on the phase velocity (a) and attenuation (b) of the surface
waves that propagate in a flat interface between water and a porous Berea sand-
stone saturated by a mixture of air and water. The high-frequency limit is consid-
ered.
3.3. Saturation Effects on the Velocities and Attenuation of the Surface Waves 47
gas fraction.
We now extend the study to more realistic frequency-dependent surface waves
and we consider the dissipative mechanisms which were neglected previously. In
this case we calculate the dispersion relation of the leaky modes for different liquid
saturations. Figure 3.5 shows the results for the pseudo-Stoneley wave. The phase
velocity decreases with saturation at high frequencies. This is consistent with the
high-frequency results. At low frequencies an important decrease in the phase veloc-
ity is found for all the saturations with respect to the high-frequency values although
a clear trend with the saturation is not found. The results for the attenuation coef-
ficient show a monotonous increase with the air fraction occupying the pore space.
This influence of the gas fraction is most significant in the 1-100 kHz range. The
characteristic frequency for the maximum of Q−1 depends on saturation. It is found
that when s increases this characteristic frequency moves towards higher values.
The influence of the saturation on the properties of the pseudo-Rayleigh is de-
picted in Figure 3.6. At low and high frequencies the pseudo-Rayleigh wave prop-
agates slower when the liquid fraction in the pore space is decreased. Interesting
features occur at intermediate frequencies (1-150 kHz). In this range of frequencies
the speed of the pseudo-Rayleigh mode decreases with increasing saturation. Fur-
thermore, a peak in the phase velocity is predicted. For the values of liquid saturation
studied in this work, this maximum lies in frequencies between 10 and 20 kHz. The
position of this maximum on the frequency axis slightly depends on saturation; lower
characteristic frequencies are obtained for lower values of s. It is interesting to note
the presence of additional local maxima for the s = 0.95, s = 0.9 and s = 0.8 cases,
which become more pronounced for lower saturation values. The characteristic fre-
quency of this secondary maximum increases with decreasing saturation.
The higher attenuation values are obtained in the low-frequency range and a max-
imum is observed. This maximum is associated with the presence of air bubbles and
becomes sharper when the saturation decreases. For the lower liquid saturation cases
considered here, s = 0.9 and s = 0.8, a second local maximum is observed at higher
frequencies. The attenuation coefficient Q−1 diminishes at high frequencies and in
this limit a clear dependence on liquid saturation is found where the attenuation in-
creases with saturation.
48 3. Surface Waves along Partially Saturated Porous Media
1460
1440
Phase Velocity (m/s) s = 0.8
1420 s = 0.9
s = 0.95
s = 0.99
1400
1380
1360
PSfrag replacements
1340
1320 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(a)
0.16
0.14
0.12
0.1
s = 0.8
Q−1
0.08
s = 0.9
0.06
s = 0.95
0.04
PSfrag replacements
0.02
s = 0.99
0 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(b)
Figure 3.5: Frequency-dependent phase velocities (a) and attenuation coefficients (b) of the
pseudo-Stoneley wave along a water/partially saturated poroelastic interface. The
porous material is a Berea sandstone saturated with a water-air mixture. Different
liquid saturations s are considered (solid lines) and the results are compared with
the fully saturated case (dotted line).
3.3. Saturation Effects on the Velocities and Attenuation of the Surface Waves 49
1880
s = 0.8
1860
s = 0.9
1840
Phase Velocity (m/s)
1820 s = 0.95
1800
1780 s = 0.99
1760
1740
PSfrag replacements
1720
1700 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(a)
0.26
0.24
0.22
0.2
0.18
Q−1
0.16
0.14
s = 0.99
0.12 s = 0.95
PSfrag replacements s = 0.9
0.1 s = 0.8
0.08 2 3 4 5 6
10 10 10 10 10
Frequency (Hz)
(b)
Figure 3.6: Frequency-dependent phase velocities (a) and attenuation coefficients (b) of the
pseudo-Rayleigh wave in a water/partially saturated poroelastic interface. The
porous material is a Berea sandstone saturated with a water-air mixture. Different
liquid saturations s are considered (solid lines) and the results are compared with
the fully saturated case (dotted line).
50 3. Surface Waves along Partially Saturated Porous Media
In this work we have studied the saturation effects on the properties of the surface
waves that propagate along a plane interface between a liquid and a partially sat-
urated porous solid. The numerical results for the pseudo-Stoneley wave and the
pseudo-Rayleigh wave show interesting features when the pore space of the poroe-
lastic medium is filled with a mixture of water and air. In the high-frequency limit
where only compressibility effects are present, the full range of liquid saturations was
studied. A transition between the leaky pseudo-Stoneley and the true Stoneley wave
is found at a characteristic saturation for which the slow P wave propagates faster
than the acoustic wave in the water half-space. This transition is neatly illustrated in
the behavior of the attenuation coefficient Q−1 which drastically decreases for wa-
ter saturations lower than s = 0.46. This indicates that the pseudo-Stoneley wave
becomes a true unattenuated surface wave, the Stoneley wave.
When the frequency-dependent dissipative mechanisms are included interesting
features arise in the velocity and attenuation of the surface waves. The pseudo-
Stoneley wave shows a well-defined maximum in the attenuation. This maximum
is located in the range of frequencies which is relevant to borehole geophysical ap-
plications (5-30 kHz). The characteristic frequency of this maximum depends on the
liquid saturation. In acoustic borehole logging techniques the pseudo-Stoneley plays
an important role and therefore our numerical results indicate its attenuation is af-
fected by saturation effects. Similar conclusions can be drawn for the phase velocity
of the pseudo-Rayleigh wave though it should be noted that this wave is difficult to
detect in field or laboratory measurements.
3.5. Appendix 3A: Gas Pocket Model Revisited 51
Appendixes to Chapter 3
The gas pocket model stands perhaps as the most widely-used model to calculate
the frequency-dependent acoustic properties of a liquid-saturated porous solid which
contains gas pockets. The original idea was introduced by J.E. White in 1975 fol-
lowed by another publication by White and coauthors (White, 1975; White et al.,
1975). Dutta and Ode (1979a,b) and Dutta and Seriff (1979) removed several of
the White approximations through a more rigorous solution derived from Biot’s the-
ory and corrected some mistakes of the original paper. The purpose of this section
is to derive this model using a displacement potential formulation which includes a
frequency-dependent description for the viscous friction interaction between the fluid
phase (gas or liquid) and the solid matrix.
In the original work of White, the gas pockets are assumed to be spherical and
located in the center of the cubic cells of a cubical lattice. In order to calculate
the compressional bulk modulus and thereafter the velocity and attenuation of the
compressional waves propagating through this system, the variation in volume of
each cubic cell caused by an oscillating external stress is considered. The interaction
among the gas pockets is neglected and therefore the macroscopic bulk modulus can
be obtained by considering a single cell. The configuration is simplified by replacing
the cubic cell by a spherical shell of identical volume as depicted in Figure 3.7. Then
g a s -s a tu ra te d
liq u id - s a tu r a te d
Figure 3.7: Gas pocket model: The inner sphere is gas-saturated and is embedded in a liquid-
saturated porous spherical shell.
52 3. Surface Waves along Partially Saturated Porous Media
the bulk modulus of this representative volume, Kb = (V /∆V )(τ0 ), can be expressed
as function of the external pressure τ0 and the radial displacement at the outer radius
u(b) as follows:
τ0 b
Kb (ω) = − . (3.21)
3u(b, ω)
The main contribution of the work of Dutta and Ode was to compute u(b) by a rig-
orous solution of the Biot equations on both the gas bubble and the liquid-saturated
spherical shell. However, in their approach, the frequency-dependent viscous inter-
action between the fluid and the solid porous in both regions was neglected and this
assumption limited the validity of their results to very low frequencies.
Let us derive now the boundary value problem which leads to the calculation
of u(b). We will use a displacement potential formulation in line with the previous
derivations for the plane interface and the oscillating bubble (see sections 2.2.1 and
3.2). We will denote the gas pocket as domain 1 and the spherical shell as domain
2. The porosity, steady-state permeability, tortuosity and shear modulus are assumed
to be the same in both domains. Inside the gas pocket the displacement potentials
(1) (1)
associated with the fast and slow compressional waves, Φ c1 (r, ω) and Φc2 (r, ω) can
be written as follows:
(1) (1)
Φc1 (r, ω) = Ac1 (ω)j0 (kc1 r), (3.22)
and
(1) (1)
Φc2 (r, ω) = Ac2 (ω)j0 (kc2 r). (3.23)
and
(1) (2) (2)
Φc2 (r, ω) = Bc2 (ω)j0 (kc2 r) + Cc2 (ω)n0 (kc2 r). (3.25)
In the above expressions for the potentials j0 and n0 denote the spherical Bessel and
Neumann functions of zeroth-order and the superscripts are used to distinguish the
domains.
The boundary conditions considered here are the continuity of the solid and fluid
displacement and pressure and stress at radius a. At the outer radius b, the stress is
assumed to be equal to the external stress τ0 and the solid and liquid displacements are
considered to be identical. In this way the spherical shell is jacketed. This assumption
3.5. Appendix 3A: Gas Pocket Model Revisited 53
and
(j)
ξij = Q(j) + R(j) Gci . (3.32)
54 3. Surface Waves along Partially Saturated Porous Media
G(ω) · x = b, (3.34)
where the matrix G contains information about the poroelastic properties of the
porous solid and the fluids saturating the void space, x = [A c1 , Ac2 , Bc1 , Cc1 , Bc2 ,
Cc2 ] and b = [0, 0, 0, 0, −τ0 , 0]. The matrix elements of G are listed in Appendix
3C. Once the amplitudes are known, the displacement u(r = b) at the outer radius of
the spherical shell is:
(2) (2)
∂Φ ∂Φc2 (2) (2) (2)
e(r = b, ω) = [ c1 +
u ]r=b = −kc1 (Bc1 j1 (kc1 b) + Cc1 n1 (kc1 b)) −
∂r ∂r
(2) (2) (2)
−kc2 (Bc2 j1 (kc2 b) + Cc2 n1 (kc2 b)). (3.35)
It is assumed that the oscillating external stress is homogenous in space. In this way
the bulk modulus of the system is the bulk modulus of one single cell and which be
obtained from equation 3.21. The complex-valued velocity of the fast compressional
wave is given by:
s
Kb + 4N/3
vc1 = , (3.36)
ρef f
In equation 3.37 ρs , ρl andρg are the density of the solid, liquid and gas phase re-
spectively and s denotes the saturation of the liquid phase, s = 1 − a 3 /b3 . We note
that no change in the shear modulus N is expected due to the purely radial nature of
the flow and displacements.
3.6. Appendix 3B: Comparison Between the Two Models 55
The aim of this section is to compare the two models which describe the bulk acoustic
properties of a partially saturated porous medium (sections 3.2 and 3.5). The most
important point in common is the fact that both models assume that the wavelengths
involved are considerable larger than the diameter of the gas bubbles and larger than
the dimensions of the cells containing the gas pockets. This assumption implies a
limitation of these models to describe the acoustic properties at frequencies high
enough that the wavelengths becomes of the order of magnitude of the patches and
therefore scattering occurs. In this scenario, the hypothesis that the dynamics of the
patches are dictated by a spatially homogeneous oscillating external pressure (stress)
field, is not valid.
The oscillating gas bubble model is based on the computation of the frequency-
dependent bulk modulus of the gas phase. As we have seen in section 3.2 this bulk
modulus also depends on the properties of the liquid and the porous solid. Then the
bulk modulus of the mixture of fluids saturating the void space is obtained through a
straightforward average (equation 3.3). The resulting bulk modulus for the mixture is
then used in the framework of the Biot theory. In this way a description of the body
waves that propagate in the partially saturated porous medium is obtained.
The gas pocket model on the other hand is based on the calculation of the com-
pressional bulk modulus of the system illustrated in Figure 3.7. This model is based
on a rigorous description of the dynamics of the gas bubble which takes into account
the radial variation of the gas displacement and the solid displacement inside the
bubble and in the surrounding liquid-saturated porous spherical shell. Another inter-
esting feature of this model is that despite the assumption of non-interacting pockets,
the presence of the neighboring pockets is somehow considered by means of the pa-
rameter b. The main restraint of the gas pocket model is related to the fact that only
the fast compressional wave is predicted and no description of the properties of the
second compressional wave propagating in the partially saturated porous solid is pos-
sible. This is a consequence of the lack of an independent description for the bulk
modulus of the mixture filling the pore space.
Next we focus on a quantitative comparison for the prediction of the velocity
and attenuation of the fast compressional wave. Figure 3.8 shows the outcome of
the velocity as a function of frequency. The analysis is restricted to relatively low
frequencies in order to avoid the scattering effects mentioned above. At very low fre-
quencies both solutions catch up to the steady-state value of 2949 m/s predicted by
the Biot-Gassmann equation modified to account for the variation of compressibility
due to the gas fraction (see e.g. Mavko et al. (1998), page 177). The velocity pre-
dicted by the oscillating gas bubble model increases monotonically with frequency.
The gas pocket model also predicts a velocity increase. We observe a maximum fol-
56 3. Surface Waves along Partially Saturated Porous Media
3300
3200
3000
osc. b. model
gas p. model
2900
2800
PSfrag replacements
2700 2 3 4
10 10 10
Frequency (Hz)
Figure 3.8: Phase velocity comparison of the oscillating gas bubble model and the gas pocket
model. The rock properties are given in Table 3.1 and the water saturation is 0.9.
A bubble radius of 1 cm is considered.
lowed by an abrupt decrease. Dutta and Ode (1979b) have noticed that there is an
inherent cut-off frequency which corresponds to a wavelength ten times larger than
the distance between neighboring gas pockets. In our case this frequency is approxi-
mately 10 kHz.
The most substantial difference between the oscillating gas bubble model and the
gas pocket model lies on the dissipative mechanisms considered in each model. The
oscillating gas bubble model incorporates new loss processes based on thermal con-
duction from the gas phase to the surrounding solid grains and acoustic radiation into
the slow compressional wave which are absent in the gas pocket model. The study of
the compressional wave attenuation provides information about the relative influence
of the dissipative mechanisms. Figure 3.9 shows the calculated inverse quality factor
according to both models. For the oscillating gas bubble model we discriminate two
different cases depending on whether the heat exchange is incorporated or not. From
the computations it appears obvious that the higher levels of attenuation predicted by
the oscillating gas bubble model can be attributed to radiation into the second com-
pressional wave and it is more significant at very low frequencies. This radiation
mechanism is suppressed in the gas pocket model by the boundary condition at the
outer radius of the spherical shell which prevents the liquid and the porous matrix to
move out-of-phase.
3.7. Appendix 3C: Matrix elements for the gas pocket model 57
0.09
0.08
0.07
0.06
Q−1
0.05
0.04
0.03
0.02
osc. b. model
PSfrag replacements 0.01
gas p. model
0 2 3 4
10 10 10
Frequency (Hz)
Figure 3.9: Specific attenuation coefficient for the oscillating gas bubble model and the gas
pocket model. The rock properties are given in Table 3.1, the water saturation is
0.9 and a bubble radius of 1 cm is considered. In order to compare the influence
of the loss mechanisms the solution neglecting the thermal conduction is plotted
in dotted lines for the oscillating bubble model.
3.7 Appendix 3C: Matrix elements for the gas pocket model
In this appendix the elements of the matrix G (equation 3.34) are given:
(1) (1)
g11 = −kc1 j1(kc1 a),
(1) (1)
g12 = −kc2 j1(kc2 a),
(2) (2)
g13 = kc1 j1(kc1 a),
(2) (2)
g14 = −kc1 n1(kc1 a),
(2) (2)
g15 = kc2 j1(kc2 a),
(2) (2)
g16 = −kc2 n1(kc2 a),
(1)
g21 = Gc1 g11 ,
(1)
g22 = Gc2 g12 ,
(2)
g23 = Gc1 g13 ,
(2)
g24 = Gc1 g14 ,
(2)
g25 = Gc2 g15 ,
(2)
g26 = Gc2 g16 ,
58 3. Surface Waves along Partially Saturated Porous Media
(1) (1)
(1) 2kc1 j1 (kc1 a) (1)
g31 = (P (1) − 2N + Q(1) Gc1 )[− a ] + (P (1) + Q(1) Gc1 )
(1)2 (1) (1)
[−kc1 (j0 (kc1 a) − 23 j2 (kc1 a))],
(1) (1)
(1) 2kc2 j1 (kc2 a) (1)
g32 = (P (1) − 2N + Q(1) Gc2 )[− a ] + (P (1) + Q(1) Gc2 )
(1)2 (1) (1)
[−kc2 (j0 (kc2 a) − 23 j2 (kc2 a))],
(2) (2)
(2) 2kc1 j1 (kc1 a) (2)
g33 = −(P (2) − 2N + Q(2) Gc1 )[− a ] − (P (2) + Q(2) Gc1 )
(2)2 (2) (2)
[−kc1 (j0 (kc1 a) − 32 j2 (kc1 a))],
(2) (2)
(2) 2kc1 n1 (kc1 a) (2)
g34 = −(P (2) − 2N + Q(2) Gc1 )[− a ] − (P (2) + Q(2) Gc1 )
(2)2 (2) (2)
[−kc1 (n0 (kc1 a) − 32 n2 (kc1 a))],
(2) (2)
(2) 2kc2 j1 (kc2 a) (2)
g35 = −(P (2) − 2N + Q(2) Gc2 )[− a ] − (P (2) + Q(2) Gc2 )
(2)2 (2) (2)
[−kc2 (j0 (kc2 a) − 32 j2 (kc2 a))],
(2) (2)
(2) 2kc2 n1 (kc2 a) (2)
g36 = −(P (2) − 2N + Q(2) Gc2 )[− a ] − (P (2) + Q(2) Gc2 )
(2)2 (2) (2)
[−kc2 (n0 (kc2 a) − 32 n2 (kc2 a))],
(1) (1)
( 2kc1 j1 (kc1 a) (1)2 (1) (1)
g41 = φ1 (Q(1) + R(1) Gc1 1))[ a + kc1 (j0 (kc1 a) − 32 j2 (kc1 a))],
(1) (1)
g42 = 1
(Q (1) + R(1) G( 1))[ 2kc2 j1 (kc2 a) + k (1)2 (j (k (1) a) − 2 j (k (1) a))],
φ c2 a c2 0 c2 3 2 c2
(2) (2)
( 2k j (k a) (2)2 (2) (2)
g43 = − φ1 (Q(2) + R(2) Gc1 2))[ c1 1a c1 + kc1 (j0 (kc1 a) − 23 j2 (kc1 a))],
(2) (2)
( 2k n (k a) (2)2 (2) (2)
g44 = − φ1 (Q(2) + R(2) Gc1 2))[ c1 1a c1 + kc1 (n0 (kc1 a) − 32 n2 (kc1 a))],
(2) (2)
( 2k j (k a) (2)2 (2) (2)
g45 = − φ1 (Q(2) + R(2) Gc2 2))[ c2 1a c2 + kc2 (j0 (kc2 a) − 23 j2 (kc2 a))],
(2) (2)
( 2k n (k a) (2)2 (2) (2)
g46 = − φ1 (Q(2) + R(2) Gc2 2))[ c2 1a c2 + kc2 (n0 (kc2 a) − 32 n2 (kc2 a))],
g51 = 0,
g52 = 0,
(2) (2)
(2) 2kc1 j1 (kc1 b) (2) (2)2 (2)
g53 = (P (2) −2N +Q(2) Gc1 )[− b ]+(P (2) +Q(2) Gc1 )[−kc1 (j0 (kc1 b)−
2 (2)
3 j2 (kc1 b))],
(2) (2)
(2) 2k n (k b) (2) (2)2 (2)
g54 = (P (2) −2N +Q(2) Gc1 )[− c1 1b c1 ]+(P (2) +Q(2) Gc1 )[−kc1 (n0 (kc1 b)−
2 (2)
3 n2 (kc1 b))],
(2) (2)
(2) 2k j (k b) (2) (2)2 (2)
g55 = (P (2) −2N +Q(2) Gc2 )[− c2 1b c2 ]+(P (2) +Q(2) Gc2 )[−kc2 (j0 (kc2 b)−
2 (2)
3 j2 (kc2 b))],
(2) (2)
(2) 2k n (k b) (2) (2)2 (2)
g56 = (P (2) −2N +Q(2) Gc1 )[− c1 1b c1 ]+(P (2) +Q(2) Gc1 )[−kc1 (n0 (kc1 b)−
2 (2)
3 n2 (kc1 b))],
3.7. Appendix 3C: Matrix elements for the gas pocket model 59
g61 = 0,
g62 = 0,
(2) (2) (2)
g63 = (Gc1 − 1)kc1 j1 (kc1 b),
(2) (2) (2)
g64 = (Gc1 − 1)kc1 n1 (kc1 b),
(2) (2) (2)
g65 = (Gc2 − 1)kc2 j1 (kc2 b),
(2) (2) (2)
g66 = (Gc2 − 1)kc2 n1 (kc2 b).
60 3. Surface Waves along Partially Saturated Porous Media
4. SHOCK-INDUCED BOREHOLE WAVES IN POROUS FORMATIONS:
THEORY AND EXPERIMENTS 1
ABSTRACT
4.1 Introduction
The study of surface acoustic waves provides a useful tool in the characterization of
elastic materials in a wide range of applications. Several techniques in geophysics
for example, make use of surface acoustic waves in order to characterize water and
hydrocarbon reservoirs. In these situations, the effects of the presence of the fluid on
wave propagation cannot be neglected. Therefore, a realistic approach must consider
both the porous medium and the saturating fluid separately rather than a single elastic
material with averaged fluid/solid properties.
Acoustic wave propagation in a fully saturated porous medium can be described
in terms of Biot’s theory (Biot, 1956a,b). This theory assumes a frequency-dependent
coupling between the solid matrix and the saturating fluid. An extra longitudinal
wave, usually called slow P wave, is predicted in addition to the usual P and S
waves. This wave is strongly frequency-dependent, it shows a diffusive behavior at
1
THIS CHAPTER HAS BEEN PUBLISHED IN THE J. ACOUST. SOC. AM., 2004 116, 693-702
62 4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experiments
to calculate the phase velocity, attenuation coefficient and radial penetration in the
porous formation of the pseudo-Stoneley wave. The experiments are performed us-
ing a water-saturated natural rock sample placed in a vertical shock tube. This shock
tube has already been used to study diverse aspects of wave propagation in porous
media. van der Grinten et al. (1985, 1987) performed experiments in air and water-
saturated porous columns consisting of agglutinated sand particles. Wave propaga-
tion in partially saturated porous media was studied by Smeulders and van Dongen
(1997). Wisse et al. (2002) presented data for shock-induced guided waves along the
outer surface of porous cylinders.
The organization of this chapter is as follows. The theoretical considerations that
lead to the dispersion relation are summarized in section 4.2. The numerical results in
terms of the permeability effects on the phase velocity, damping, radial displacement
and pore pressure are discussed in section 4.3. The contribution of the slow P wave
to the radial displacement induced by the pseudo-Stoneley wave is analyzed too. The
experimental setup is described in section 4.4, followed by the experimental results
in terms of the frequency-dependent wave speeds and attenuation coefficients. The
conclusions are summarized in section 4.5.
Figure 4.1 shows the borehole configuration. It consists of a porous cylinder with a
radius R1 = 37.1 mm placed in the test section of a shock tube. In this cylindrically
confined reservoir, a borehole with a radius R0 = 6.3 mm was drilled. The borehole
and the pores are fully filled with water. The inner shock tube radius R 2 is 38.5
mm. Between the external radius of the cylinder and the wall of the shock tube
there is a small annulus filled with water. Our aim is to study the behavior of the
pseudo-Stoneley wave which can be observed in our experimental setup. All possible
wave solutions that may exist in our borehole configuration can be written as a linear
combination of the bulk modes in the borehole, the porous medium, and the fluid
annulus. The bulk modes are described as wave potentials. In the borehole fluid only
compressional waves can propagate and the wave potential for the fluid displacement
can be written as
R 2 = 3 8 .5 m m
R 1 = 3 7 .1 m m
R 0 = 6 .3 m m
r
s h o c k tu b e w a ll
p o ro u s
fo r m a tio n
w a te r
Φc1 (kz , ω, r) = Ac1 (kz , ω)J0 (kc1 r) + Bc1 (kz , ω)H0 (kc1 r), (4.2)
Φc2 (kz , ω, r) = Ac2 (kz , ω)J0 (kc2 r) + Bc2 (kz , ω)H0 (kc2 r), (4.3)
Ψsh (kz , ω, r) = Ash (kz , ω)J1 (ksh r) + Bsh (kz , ω)H1 (ksh r), (4.4)
where J0 and J1 are Bessel functions and H0 and H1 are Hankel functions of the first
kind. The subscript c1 refers to the fast compressional wave, c2 to the slow compres-
sional wave and sh to the shear wave. The above potentials describe axially symmet-
ric wave modes that propagate in the z−direction, where the exponential dependence,
ei(kz z−ωt) , is implicitly assumed. The relation between the radial wavenumbers k j
and kz is given by:
ω2
kj2 = − kz2 , j = 1, 2, sh, f, (4.5)
cj (ω)2
where cj denotes the free-field velocity of the corresponding body wave. In the water
annulus the following expression holds
The acoustic modes are determined by the boundary conditions. In this case there
are four boundary conditions that must be satisfied both at the borehole wall and at
the outer cylinder wall. The first boundary condition represents the continuity of
volume flux:
and
In equations 4.7 and 4.8 U refers to the radial component of the fluid displacement
and u to the radial component of the solid displacement. The porosity is denoted by
φ. The subscript 0 refers to the bore-fluid, 1 to the porous formation and 2 to the
water annulus. The continuity of the average stress in terms of the stress in the solid
τij , and the pressure in the fluid p yields that
p0 = p1 − iωZφ(U1 − u1 ). (4.12)
This last boundary condition defines the surface impedance Z, which can be used to
model permeability discontinuity effects such as the mudcake and impairment layers
(Liu and Johnson, 1997; Tichelaar et al., 1999; Deresiewicz and Skalak, 1963). In
this work we only consider the case Z=0, which corresponds to open-pore conditions
and leads to the continuity of pressure:
p0 = p 1 at r = R0 , (4.13)
and
p1 = p 2 at r = R1 . (4.14)
The shock tube sections are made out of steel and have a wall thickness of 24 mm
in order to minimize compliance effects. Therefore it is reasonable to consider the
66 4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experiments
shock tube wall as rigid. Then, at the inner shock tube wall, the radial displacement
must be equal to zero:
U2 = 0 at r = R2 . (4.15)
M(kz , ω) · a = b, (4.16)
where the matrix M contains information about the fluid and the porous formation, a
is a vector containing the amplitudes of the wave potentials, a T = (A0 , Ac1 , Bc1 , Ac2 ,
Bc2 , Ash , Bsh , A2 , B2 ) and b is a source term. In order to find the dispersion rela-
tion of the borehole modes, b = 0 is considered. The matrix elements are shown in
Appendix 4B.
The guided wave modes are obtained by solving the complex equation
this condition of lateral infinity is approached by our shock tube configuration which
has a confined reservoir at r = R1 .
In Figures 4.2 and 4.2 we compute the phase velocities and damping coefficients
as a function of frequency for different permeability values. We generally assumed
1400
Vsh (360 mD)
1200
Vps (3.6 mD)
Phase Velocity (m/s)
1000
Vps (36 mD)
800
PSfrag replacements
600 Vps (360 mD)
400
laterally unbounded media, except for the case k0 = 360 mD where also the bounded
reservoir is considered (dotted line). For comparison, the bulk slow and shear waves
are also shown for this case (dashed lines, k0 = 360 mD). The parameter values for
the porous reservoir and the saturating fluid are given in Table 4.1. It can be seen that
the pseudo-Stoneley wave is always faster than the slow wave and slower than the
rest of the bulk modes. Its attenuation is in between the damping coefficients of both
bulk modes. For frequencies over 30 kHz for the k0 = 360 mD case, the damping
decreases, but we found that a f 1 behavior is reached for f→ ∞ (not shown in the
Figure). It is also noticed that the shock tube configuration is a good approximation
for the radially infinite reservoir, especially at high frequencies. As this agreement is
expected to become better for lower permeability values (less influence of the shock
tube structure as will be explained below), all other permeability curves are computed
for the laterally infinite reservoir only.
A clear dependence on permeability was found for both the phase velocities and
damping coefficients. The increase of the attenuation with permeability can be ex-
68 4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experiments
2
10
0
10 Dps (360 mD)
Dsh (360 mD)
Dps (36 mD)
Dps (3.6 mD)
2 3 4 5
10 10 10 10
Frequency (Hz)
Figure 4.3: Frequency-dependent damping coefficient for the p-Stoneley wave for different
permeability values. The solid lines correspond to the infinite reservoir configu-
ration. The results for the confined reservoir configuration for k0 = 360 mD are
shown in dotted line. The slow wave and shear wave damping are also displayed
for this case (k0 = 360 mD) in dashed lines.
plained by the fact that a larger permeability implies a larger flow rate between the
pores and the borehole. This constitutes the main dissipative mechanism for the sur-
face waves and it is present in the complete range of studied frequencies (100 Hz-200
kHz). The dependence of the phase velocities on permeability is basically restricted
to the low-frequency range (100 Hz-10 kHz).
We also study the radial distribution of the pore pressure and the weighted radial
displacement ζ = φU1 +(1−φ)u1 , which are continuous across the fluid/porous solid
interfaces. This provides yet another way to compare radially bounded and infinite
reservoirs and helps to assess the validity of shock tube data for infinitely extended
reservoirs. Also the permeability dependence of the radial distributions of p and ζ is
studied (for the laterally infinite configuration).
Figures 4.4 and 4.5 show the radial distribution of ζ as a function of the per-
meability. The results are normalized with respect to the value at the borehole wall
(r = R0 ). In the borehole (r < R0 ) the weighted radial displacement linearly in-
creases with radius until the maximum value is reached at r = R 0 . In the porous
material, ζ decreases which is typical for non-radiating surface waves. The decrease
displays a 1/r dependence (incompressibility condition at low frequencies). For the
confined reservoir, the weighted radial displacement in the annulus linearly decreases
until it reaches zero at the rigid wall. Also here, a good agreement between the re-
4.3. Permeability Effects on the Pseudo-Stoneley Wave 69
Table 4.1: Physical properties of the Berea sandstone and the saturating fluid (water).
sults for confined and infinite reservoirs is obtained. This agreement is even better
for f = 5 kHz (Figure 4.5). Here the wavelengths involved are much smaller than
the radius of the borehole and the shock tube, so less influence of the shock tube wall
is to be expected.
0.9
0.8
0.7
|ζ/ζ(R0)|
0.6
0.5 360 mD
0.4
3.6 mD
0.3
confined (360 mD)
PSfrag replacements 0.2
0.1 36 mD
0
0 0.5 1 1.5
r/R2
Figure 4.4: Radial and permeability dependence of the average radial displacement at f = 500
Hz. The results for the confined reservoir are shown in dotted line.
Similar conclusions can be drawn for the radial distributions of the pore pressure
(Figures 4.6 and 4.7). The dependence on the permeability, however, is more promi-
70 4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experiments
0.9
0.8
0.7
|ζ/ζ(R0)|
0.6
0.5
0.4 3.6 mD
0.3 36 mD
PSfrag replacements
360 mD
0.2
0.1
confined (360 mD)
0
0 0.5 1 1.5
r/R2
Figure 4.5: Radial and permeability dependence of the average radial displacement at f = 5
kHz. The results for the confined reservoir are shown in dotted line.
nent in this case and it is more clear that the radial decay increases when the per-
meability is decreased. This implies that the use of lower permeability samples will
improve the agreement between the results corresponding to the radially bounded
and unbounded reservoirs. For f = 500 Hz (Figure 4.6) there is still a significant
discrepancy between the results for bounded and unbounded reservoirs but it has
disappeared at 5 kHz (Figure 4.7). These results correspond to the low-frequency
deviations between the bounded and unbounded configurations that are found in the
phase velocities and damping coefficients (Figures 4.2 and 4.3).
Next, we examine the relative contributions of the different bulk modes to the
weighted radial displacement caused by the pseudo-Stoneley wave (Figure 4.8). We
limit our analysis to the laterally infinite reservoir. Following the theoretical consid-
erations explained in Appendix 4A, the weighted radial displacement ζ (pS) induced
by the pseudo-Stoneley wave can be written as:
(pS) (pS) (pS) (pS)
ζ (pS) = γc1 Bc1 (kz , ω)H1 (kc1 r) + γc2 Bc2 (kz , ω)H1 (kc2 r) +
(pS) (pS)
+γsh Bsh (kz , ω)H1 (ksh r), (4.18)
(pS) (pS) (pS)
where the frequency-dependent coefficients γc1 , γc2 and γsh are given in Ap-
(pS) (pS) (pS)
pendix 4B. Furthermore, the potential amplitudes Bc1 , Bc2 and Bsh are at each
frequency linearly related to the potential amplitude of the fluid wave in the borehole,
(pS) (pS)
A0 , by frequency-dependent coefficients θj (ω) (this follows from the fact that
the pseudo-Stoneley wave mode satisfies eq. 4.17). Then, the contribution of each
4.3. Permeability Effects on the Pseudo-Stoneley Wave 71
0.9
0.8
0.7
|p/p(R0)|
0.6
0.5
0.1 36 mD
3.6 mD
0
0 0.5 1 1.5
r/R2
Figure 4.6: Radial and permeability dependence of the pore pressure at f = 500 Hz. The
results for the confined reservoir are shown in dotted line.
0.9
0.8
0.7
|p/p(R0)|
0.6
0.5
0.4
0.3
PSfrag replacements
0.2 confined (360 mD)
0.1
3.6 mD 36 mD
360 mD
0
0 0.5 1 1.5
r/R2
Figure 4.7: Radial and permeability dependence of the pore pressure at f = 5 kHz. The results
for the confined reservoir are shown in dotted line.
72 4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experiments
bulk mode to the average radial displacement of the pseudo-Stoneley wave is defined
as follows:
(pS) (pS) (pS) (pS)
ζj (ω) = γj (ω)θj (ω)A0 (ω), j = c1, c2, sh. (4.19)
(pS) (pS)
In Figure 4.8 the normalized mode contributions (γj θj ) are displayed rep-
resenting the relative effect of the different bulk modes c1, c2 and sh upon the radial
displacement ζ (pS) associated with the pseudo-Stoneley wave. The dependence of
2
10
0
10
(ps) (ps)
γ j θj
−2
10
slow wave
shear wave
PSfrag replacements
10
−4 fast wave
2 3 4 5
10 10 10 10
Frequency (Hz)
Figure 4.8: Bulk modes weight for the pseudo-Stoneley wave. The results are expressed in
terms of the frequency-dependent contribution of each bulk mode to the weighted
radial displacement.
The vertical shock tube is shown in Figure 4.9. It consists of a high-pressure section
and a low-pressure section, separated by a diaphragm which is ruptured by an electric
pulse. The dimensions of the sections are indicated in the figure.
4.4. Experimental Setup and Results 73
w a te r
s h o c k tu b e w a ll
P 2 p o ro u s
fo r m a tio n
p ro b e
Figure 4.9: Vertical shock tube. The measuring probe P2 is shown in detail.
A porous Berea cylinder with a concentric borehole is mounted in the test section
of the shock tube and carefully saturated with water. The water level is approximately
70 cm above the top of the sample. The properties of the sample used are given in
Table 4.1. A miniature pressure transducer is mounted in a mobile probe, P2, that can
be displaced along the axial direction of the borehole. In practice, the position of the
transducer is changed 0.5 cm after each shock tube run; then a new shock wave exper-
iment is carried out. Due to the excellent repeatability of the wave experiments, these
measurements can be combined in one single record. Combining 25 measurements,
a record of 25 equidistant pressure signals is obtained. The pressure transducer P1
mounted in the shock tube wall is used to trigger the data acquisition system. Figure
4.10 shows a typical FFT conversion of the signal registered by P1.
Figure 4.11 shows the combined record of 4 pressure measurements. Distances
are measured with respect to the water-porous medium interface. Knowing their ap-
proximate velocities, two arrivals can be distinguished. The first one corresponds to
the fast bulk wave followed by an arrival with a larger amplitude corresponding to
the pseudo-Stoneley wave. It can be observed how these two arrivals separate in time
when the depth increases. In order to find the dispersion relation and consequently
V (ω) and D(ω), the complete record of 25 measurements is transformed from the
time-space domain to the frequency-wave number domain applying a combined FFT-
Prony method (Wisse et al., 2002). For further details about Prony’s method we re-
74 4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experiments
0
10
−2
10
−3
10
−4
10
PSfrag replacements −5
10
0 50 100 150 200 250 300 350 400 450 500
Frequency (kHz)
4
Pressure (bar)
z=5.6 cm
3
z=4.6 cm
2
z=3.6 cm
1
PSfrag replacements
z=2.6 cm
0
pseudo-Stoneley wave
fast P wave
1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Time(ms)
Figure 4.11: Pressure recordings at different depths. For readability, traces have been shifted
upwards.
fer to Marple (1987) and Wisse (1999). The results are shown in Figures 4.12 and
4.13, together with the theoretical computations. Good agreement is obtained for the
phase velocity in a broad band of frequencies (10-50 kHz). At lower frequencies
there are deviations from the theoretical values. The resolution for low frequencies
is limited by the length of the sample and diminishes when the wavelength is compa-
4.4. Experimental Setup and Results 75
2000
1800
1600
Phase Velocity (m/s) 1400
1200
1000
800
600
400
200
PSfrag replacements
0
0 5 10 15 20 25 30 35 40 45 50
Frequency (kHz)
Figure 4.12: Comparison between experiments and theory for the phase velocity of the
pseudo-Stoneley wave. The formation is a Berea sandstone. Dots: experimental
data, solid line: theoretical computations.
1
Damping coefficient (1/m)
10
0
10
−1
10
−2
10
PSfrag replacements
0 5 10 15 20 25 30 35 40 45 50
Frequency (kHz)
Figure 4.13: Experimental (dots) and theoretical (solid lines) damping coefficients of the
pseudo-Stoneley wave. The formation is a Berea sandstone.
rable with that length. The experimentally determined damping coefficients present
large oscillations. Similar problems were reported by Hsu et al. (1997) and Wisse
et al. (2002). Ellefsen et al. (1993) have pointed out that the combined FFT-Prony
method has inherent limitations in the determination of the attenuation coefficients.
76 4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experiments
Moreover, Wisse et al. (2002) indicate that the accuracy of the determination of the
damping coefficient is very sensitive to the signal-to-noise ratio. We note that all
input parameters for the numerical computations (Table 4.1) were obtained from in-
dependent laboratory measurements without any adjustable parameters. Our results
are in agreement with those reported by Winkler et al. (1989), who performed similar
measurements using an acoustic source located in the borehole. In their technique,
however, considerably less oscillations in the damping were observed.
Using a shock tube technique, accurate measurements of the phase velocity of the
pseudo-Stoneley wave along boreholes in a water-saturated porous reservoir were
performed. The frequency band involved ranges from 1 to 50 kHz. Good agreement
with the computations based on Biot’s theory is observed. The quality of the attenu-
ation data is less good; intense scattering is observed and only qualitative agreement
with the theoretical calculations is possible.
Permeability effects were studied numerically. The damping of the pseudo-
Stoneley wave shows a strong dependence on permeability over the entire frequency
range. The phase velocity shows strong dependence on permeability for low frequen-
cies only (below 10 kHz). The contribution of the slow wave in comparison with the
other bulk modes has been calculated for the weighted radial displacement. The slow
wave contribution to the weighted radial displacement is one order of magnitude
larger than that of the fast compressional and shear waves in the porous reservoir.
The outcome of these calculations explains the strong dependence of the properties
of the pseudo-Stoneley wave on permeability for low frequencies.
Finally, we point out that the results obtain so far indicate that the shock tube
technique can be successfully implemented in order to study the propagation of sur-
face waves in porous infinite reservoirs. The agreement will improve when samples
with low permeability values are used, since the radial penetration of the weighted
radial displacement and pore pressure in the formation decreases with permeability.
4.6. Appendix 4A: Potential Formulation of Biot’s Theory in Cylindrical Coordinates 77
Appendixes to Chapter 4
∂2u ∂2U
ρ11 + ρ 12 = P ∇(∇ · u) + Q∇(∇ · U) − N ∇ × ∇ × u +
∂t2 ∂t2
ηφ2 ∂U ∂u
+ F (ω) − (4.20)
κ0 ∂t ∂t
and,
∂2u ∂2U ηφ2 ∂U ∂u
ρ12 2 + ρ22 2 = R∇(∇ · U) + Q∇(∇ · u) − F (ω) − , (4.21)
∂t ∂t κ0 ∂t ∂t
where u is the solid displacement and U is the fluid displacement. N is the shear
modulus of the composite material, P, Q and R are the so-called generalized elastic
coefficients. They are related to the porosity φ, the solid frame bulk modulus K b ,
the solid grain bulk modulus Ks , the pore fluid modulus, Kf and N according to the
following expressions:
Kb Ks
(1 − φ)(1 − φ − )Ks + φ Kb
Ks Kf 4
P = + N, (4.22)
Kb Ks 3
1−φ− +φ
Ks Kf
Kb
(1 − φ −
)φKs
Ks
Q= , (4.23)
Kb Ks
1−φ− +φ
Ks Kf
φ2 K s
R= . (4.24)
Kb Ks
1−φ− +φ
Ks Kf
The density terms are given by:
and,
where,
P − vc1 2 ρ
e11
Gc1 = 2 , (4.31)
vc1 ρe12 − Q
P − vc2 2 ρ
e11
Gc2 = 2 , (4.32)
vc2 ρe12 − Q
and
α
e−1
Gsh = . (4.33)
α
e
In the above equations vc1 , vc2 and vsh refer to the frequency-dependent wave ve-
locities of the fast, the slow and the shear wave respectively and α
e to the frequency-
dependent tortuosity given by (Johnson et al., 1987),
iηφ2 F (ω)
α
e = α∞ + . (4.34)
k0 φωρf
The density terms ρe11 , ρe12 and ρe22 in equations (31) and (32) are frequency-dependent
and are calculated using α e:
ρe11 = (e
α − 1)φρf + (1 − φ)ρs , (4.35)
4.7. Appendix 4B: Matrix coefficients for the confined reservoir 79
ρe12 = (1 − α
e)φρf , (4.36)
ρe22 = α
eφρf . (4.37)
The stress in the solid and the pore pressure in the porous reservoir can be obtained
using the Biot stress-strain relations given by:
and,
−1
p= (Qukk + RUkk ), (4.39)
φ
∂ui ∂u
where summation over repeated indexes is assumed, u ij = 21 ( ∂x j
+ ∂xji ) and δij is
the Kronecker delta. In the bore fluid the following expressions hold for the radial
displacement and pressure:
∂Φ0
U= , (4.40)
∂r
and,
p = ρ f ω 2 Φ0 . (4.41)
In this appendix we the matrix elements mij for the confined reservoir are given. In
order to simplify the notation, we introduce the following variables:
2
m63 = ω2 H0 (kc1 R1 )Cc1 + 2N kc1 2 [H (k R ) − H1 (kc1 R1 ) ],
0 c1 1 kc1 R1
vc1
2
m64 = ω2 J0 (kc2 R1 )Cc2 + 2N kc2 2 [J (k R ) − J1 (kc2 R1 ) ],
0 c2 1 kc2 R1
vc2
2 H 1 (kc2 R1 )
m65 = ω2 H0 (kc2 R1 )Cc2 + 2N kc2 2 [H (k R ) −
0 c2 1 ],
vc2 kc2 R1
J (k R )
m66 = 2N ikz ksh [J0 (ksh R1 ) − 1 sh 1 ],
ksh R1
H1 (ksh R1 )
m67 = 2N ikz ksh [H0 (ksh R1 ) − ],
ksh R1
2
m68 = −ρ0 ω J0 (kf R1 ),
m69 = −ρ0 ω 2 H0 (kf R1 ),
m72 = −2ikz kc1 N J1 (kc1 R1 ),
m73 = −2ikz kc1 N H1 (kc1 R1 ),
m74 = −2ikz kc2 N J1 (kc2 R1 ),
m75 = −2ikz kc2 N H1 (kc2 R1 ),
2
m76 = [2kz2 − ω2 ]N J1 (ksh R1 ),
vsh
2
m77 = [2kz2 − ω2 ]N H1 (ksh R1 ),
vsh
2 J (k R )
m82 = 2ω 0 c1 1
[−RGc1 − Q],
vc1 φ
2 H (k R )
m83 = ω2 0 c1 1 [−RGc1 − Q],
vc1 φ
2 J (k R )
m84 = ω2 0 c2 1 [−RGc2 − Q],
vc2 φ
2 H (k R )
m85 = ω2 0 c2 1 [−RGc2 − Q],
vc2 φ
m88 = ρ0 ω 2 J0 (kf R1 ),
m89 = ρ0 ω 2 H0 (kf R1 ),
m98 = −kf J1 (kf R2 ),
m99 = −kf H1 (kf R2 ),
and,
m18 = m19 = m28 = m29 = m31 = m38 = m39 = m46 = m47 =
m48 = m49 = m51 = m61 = m71 = m78 = m79 = m81 = m86 =
m87 = m91 = m92 = m93 = m94 = m95 = m96 = m97 = 0,
where,
Cc1 = (P − 2N ) + Gc1 Q + RGc1 + Q, and
Cc2 = (P − 2N ) + Gc2 Q + RGc2 + Q.
82 4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experiments
4.8 Appendix 4C: Matrix coefficients for the radially infinite porous
formation
5.1 Introduction
In this chapter we extend the study carried out in the previous chapter in order to in-
clude higher order surface modes and guided waves that propagate along a cylindri-
cal interface between a liquid and a fully saturated poroelastic reservoir. In acoustic
logging applications the diameter of the drilled boreholes commonly vary from 10
to 50 cm (5 to 20 inches)(Hearst et al., 2000). The effects of the borehole radius
on the phase velocity and attenuation of the surface waves are of great importance
since our experiments are carried out in boreholes with a radius which is one order
of magnitude smaller than the typical borehole dimensions encountered in acoustic
borehole logging. Another complicating factor in the interpretation of our experi-
mental findings is the presence of the boundaries imposed by the shock tube wall.
It is shown that when the reservoir is bounded, the complexity of the wave pattern
increases enormously. Despite the presence of numerous extra modes in the bounded
reservoir configuration, it is quite well possible to relate some of the waves to the
guided modes that exist in the radially infinite reservoir. This equivalence allows us
to study these modes in our confined reservoir configuration, the shock tube, under
well-defined laboratory conditions and to extrapolate our results to a more practical
situation where the reservoir can be considered as radially infinite. One of the waves,
for example, which can be excited in the bounded reservoir is a guided wave which is
strongly linked to the fast compressional wave. Therefore our experimental technique
can be applied to study the velocity of the fast P wave and its frequency-dependent
attenuation.
In previous experiments carried out in the shock tube we were able to measure
the frequency-dependent phase velocity and attenuation of the pseudo-Stoneley wave
propagating along the borehole wall of a Berea sandstone formation (Chao et al.,
2004). However, while the resolution of the phase velocity was reasonably good and
was found to be in quantitative agreement with numerical predictions based on Biot’s
theory, the attenuation data presented a strong scatter and only qualitative agreement
was obtained. The theoretical analysis of the different mode contributions to the
84 5. Surface and Guided Waves along Cylindrical Interfaces
1
1600 pR−3 10
water pR−1 pR−2
1400
shear
Phase Velocity (m/s)
0
10
1200 p−Stoneley
1000 −1
10
Q−1
800 slow P p−Stoneley
−2
10
600
pR−3
400 pR−2
−3
10
200 pR−1
PSfrag replacements −4
PSfrag replacements 0 10
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
(a) (b)
1
1800 10
pR−1 pR−2 pR−3 pR−4 pR−5 pR−6
1600
water
Phase Velocity (m/s)
0
10
1400
1200
−1
10
p−Stoneley
Q−1
1000 shear
p−Stoneley
800 slow P −2 pR−6
10 pR−5
600 pR−4
pR−3
400 −3 pR−2
10
200
pR−1
−4
PSfrag replacements 0 PSfrag replacements 10
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
(c) (d)
1
1800 10
0
1400 10
shear
1200 p−Stoneley
−1
10
Q−1
1000
slow P p−Stoneley
800 −2 pR−4
10
pR−3
600
pR−2
400 10
−3
pR−1
200
PSfrag replacements PSfrag replacements −4
0 10
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Frequency (kHz) Frequency (kHz)
(e) (f)
Figure 5.1: Frequency-dependent borehole modes in a porous reservoir. In solid lines the
phase velocities and attenuation coefficients of the pseudo-Stoneley wave and the
pseudo-Rayleigh waves (PR). The bulk velocities in the poroelastic formation are
indicated in dashed line. The results for the Berea sandstone are shown in (a) and
(b), for the N1 reservoir in (c) and (d) and for the N2 reservoir in (e) and (f).
86 5. Surface and Guided Waves along Cylindrical Interfaces
Berea sandstone N1 N2
Solid densityρs (kg/m3 ) 2644 2590 2539
Porosity φ 0.20 0.53 0.465
Permeability k0 (D) 0.36 2.9 10.8
Tortuosity α∞ 2.4 1.7 1.93
Frame bulk modulus Kb (GPa) 10.37 1.97 2.53
Shear modulus N (GPa) 4.02 1.3 2.9
Solid grain bulk modulus Ks (GPa) 36.5 36.5 36.5
Liquid density ρf (kg/m3 ) 1000 1000 1000
Liquid viscosity η (mPa s) 1.0 1.0 1.0
Liquid bulk modulus Kf (GPa) 2.25 2.25 2.25
Dry compressional velocity vp (m/s) 2727 1744 2170
Dry shear velocity vs (m/s) 1379 1033 1461
Table 5.1: Physical properties of the porous formations used in the experiments and the nu-
merical calculations. N1 and N2 are synthetic porous materials made out of sin-
tered glass, the Berea sandstone is a natural formation. The parameters reported
in this table were independently measured in laboratory tests.
than the borehole radius and the pressure oscillations in the borehole can be assumed
to be one-dimensional (see section 2.3.2). The borehole wall is fully permeable and
allows the liquid to diffuse in the formation, which leads to a zero velocity in the
zero-frequency limit. In the high-frequency regime the wavelengths are substantially
shorter than the radius of the borehole and therefore the interface waves do not “see”
the curvature of the borehole and behave like in a plane liquid/poroelastic interface
(see section 2.2). The attenuation of the pseudo-Stoneley wave is appreciable and
considered as a combination of flow-induced dissipation at the borehole wall and ra-
diation into the slow compressional wave in the formation. This constitutes the main
difference with the Stoneley wave in the elastic case which propagates almost unat-
tenuated. The Stoneley wave in the elastic case may radiate into the shear wave in the
case of slow elastic formations, especially at low frequencies (Sinha and Asvadurov,
2004). In the cases considered in this thesis there is no radiation into the shear wave
as can be observed in Figure 5.1 since the phase velocity of the pseudo-Stoneley wave
is always lower than the speed of the shear wave.
The pseudo-Rayleigh waves are characterized by cut-off frequencies. At the high-
frequency limit they catch up to the speed of sound in water (1500 m/s). The existence
of multiple pseudo-Rayleigh modes is a consequence of the cylindrical geometry and
was also observed in the case of a rigid borehole wall (see Figure 2.9). For the N1
formation the attenuation curves reveal a clear maximum in the first pseudo-Rayleigh
5.2. Borehole modes in Poroelastic Formations 87
mode for a frequency slightly higher than the cut-off frequency. Also a maximum in
the attenuation of the second pseudo-Rayleigh wave is observed in this case. It is
interesting to note that the attenuation of the pseudo-Rayleigh waves is lower than
the attenuation of the pseudo-Stoneley wave. This situation is reversed for the elastic
case where the pseudo-Rayleigh waves are stronger damped than the (true) Stoneley
wave (Schmitt, 1988; Sinha and Asvadurov, 2004). In the poroelastic scenario the
main dissipative mechanism which affects the propagation of the pseudo-Rayleigh
waves is the radiation into the shear and slow compressional waves. Apparently the
combination of viscous dissipation and radiation into the slow wave generates higher
attenuation than pure radiation into the slow wave and the shear wave.
1400
0
1300 10
1200
Phase Velocity (m/s)
1100
1000
900
Q−1
800
−1
10
700
600
500
400
PSfrag replacements 300 PSfrag replacements
0 2 4 2 3 4 5
10 10 10 10 10 10 10
Frequency (Hz) Frequency (Hz)
(a) (b)
1
1400 10
1300
1200
Phase Velocity (m/s)
1100 0
10
1000
Q−1
900
800
−1
10
700
600
500
PSfrag replacements
PSfrag replacements −2
400 1 2 3 4 5
10 1 2 3 4 5
10 10 10 10 10 10 10 10 10 10
Frequency (Hz) Frequency (Hz)
(c) (d)
Figure 5.2: Borehole radius dependence of the phase velocity and attenuation coefficient of
the pseudo-Stoneley mode in the water-saturated Berea sandstone ((a),(b)) and
the N1 synthetic formation ((c),(d)). The results correspond to a radius of 6.3
mm (solid lines) and 6.3 cm (dashed lines).
uation cannot simply be interpreted in terms of frequency scaling. For the synthetic
reservoir N1 (see Figure 5.2d) an intersection between the two solutions is observed;
at low frequencies the attenuation is higher for the slim borehole, but for higher fre-
quencies the results are almost identical.
We also computed the effects of a change in borehole radius on the phase velocity
and specific attenuation of the first pseudo-Rayleigh wave. This was only done for
the N1 formation because the cut-off frequencies of the pseudo-Rayleigh modes in-
crease drastically with decreasing radius and in the Berea formation not even the first
pseudo-Rayleigh mode is observed in the range of frequencies studied. The results
5.2. Borehole modes in Poroelastic Formations 89
2100 0.12
2000 0.1
Phase Velocity (m/s)
1900 0.08
Q−1
1800 0.06
1700 0.04
1600 0.02
are given in Figure 5.3. Undoubtedly the pseudo-Rayleigh waves are the waves that
are most affected by the variation of the radius of the borehole. The changes of the
properties of the pseudo-Rayleigh waves can be explained by taking into account that
this set of modes is strongly linked to the curvature of the borehole.
5000
4500
4000
Phase Velocity (m/s)
3500
3000
2500
2000 L2 L3 L4 L5
L1
PSfrag replacements
1500
1000 S1
500 D
p-Stoneley
0
0 20 40 60 80 100 120 140 160
Frequency (kHz)
(a)
1
10
L1
L3
0
10 D
Q−1
−1 p-Stoneley
10
PSfrag replacements
−2
10
S1
L2
0 20 40 60 80 100 120 140 160
L4 Frequency (kHz)
L5
(b)
Figure 5.4: Phase velocity and attenuation of the axysimmetric guided modes in the confined
reservoir configuration for the Berea sandstone material.
5.2. Borehole modes in Poroelastic Formations 91
1 1
0.8 0.8
|ζ/ζ(R1)|
|p/p(R1)|
30 kHz
0.6 0.6
60 kHz
0.4 0.4
90 kHz
figuration the first three L modes are propagative at all frequencies. The L2 mode
is related to the fast compressional wave at low frequencies as we will see below.
The higher order L-modes are characterized by cut-off frequencies quite similar to
the pseudo-Rayleigh waves in a borehole configuration or the undamped modes in
a liquid-filled cylindrical cavity (see Figure 2.9). There is one guided mode which
is associated with the slow compressional wave in the porous sample and this mode
is denoted D-mode (Wisse, 1999). An extra surface mode, S1, is also present; this
mode is a surface mode propagating at the interface between the outer surface of the
cylinder and the water gap. This can be inferred from the radial dependence of the
average radial displacement ζ = (1 − φ)ur + φUr , and the pore-pressure induced by
the S1 mode at representative frequencies of 30 kHz, 60 kHz and 90 kHz. Inspection
of Figure 5.5 confirms the assumption about the nature of the S1 mode. The radial
profiles of the average radial displacement and pore pressure depict the characteristic
decay in the porous reservoir correlated to a surface wave whose energy is confined
to the interface between the outer surface of the porous sample and the water annulus.
Next we discuss the correspondence between the modes that propagate in the con-
fined reservoir configuration and in the radially infinite reservoir configuration. In the
previous chapter the influence of the shock tube wall on the properties of the pseudo-
Stoneley wave for the Berea sandstone was investigated. The results indicated that
there is almost a one-to-one correspondence between the pseudo-Stoneley mode in
both configurations except at very low frequencies where the wavelengths are on the
order of the radius of the shock tube. This is the only surface wave that is observed in
a borehole of 6.3 mm radius surrounded by a radially infinite porous formation in the
92 5. Surface and Guided Waves along Cylindrical Interfaces
range of frequencies of interest for our experimental investigations (1-150 kHz), since
the cut-off frequencies of the pseudo-Rayleigh waves are well above the upper limit
of this frequency band. This situation changes for the N1 synthetic reservoir. In this
case there are two surface modes that propagate along the slim borehole in the range
of frequencies of interest: the pseudo-Stoneley and the first pseudo-Rayleigh wave.
Once more the properties of the pseudo-Stoneley wave are almost unaffected by the
presence of the tube wall. Unfortunately it is not possible to relate the first pseudo-
Rayleigh mode to any of the guided modes in the bounded reservoir and therefore it
cannot be studied in our experimental facility.
A picture of the test section of the vertical shock tube employed to perform the acous-
tics experiments is shown in Figure 5.6. The vertical shock tube is made out of steel.
o p tic a l a c c e s s
w in d o w
s h o c k tu b e
} a m p lifie rs
} p re ssu re g a u g e s
b o tto m p la te
Figure 5.6: Test section of the vertical shock tube. The picture covers an area of approxi-
mately 0.5 x 1.35 m.
The length of the tube is approximately 6.8 m and it has an inner diameter of 77 mm.
The shock tube consists of two sections: the high-pressure section and the test sec-
5.3. Modification of the Shock Tube 93
W a te r
F u n n e l
P o ro u s c o lu m n
B o re h o le
Figure 5.7: Schematic drawing of the funnel placed in the test section of the shock tube
tion which are separated by a plastic diaphragm (see Figure 4.9). The porous sample
is placed at the bottom of the test section and it is carefully saturated with water.
The water level is approximately 1 meter above the bottom plate of the test section.
The lengths of the samples varies from 25 cm to 40 cm. With respect to the setup
described in the previous chapter, the shock tube was slightly modified. In order to
enhance the excitation of the surface waves, a funnel was placed inside the tube. This
is schematically shown in Figure 5.7. This structure is designed to shield the top of
the sample from the direct impact of the incident wave, and to channel most of the
energy into the borehole, thereby strongly reducing the compressional body wave.
The funnel structure is attached to the shock tube wall and its base is separated from
the top of the sample by a distance of approximately 1 mm in order to avoid mechan-
ical contact. The variable cross section is intended to prevent resonances inside this
device.
A typical experiment proceeds as follows: The high pressure section is pressur-
ized up to 4 bars with respect to the test section which is at atmospheric pressure.
Then the plastic membrane is ruptured. In this way a weak shock wave is gener-
ated which travels downwards. It is partially transmitted and partially reflected at the
air-water interface. The transmitted wave becomes a step-like acoustic wave which
is focussed into the borehole by means of the funnel. The pressure inside the bore-
hole is measured by means of a miniature pressure transducer which is mounted in
a mobile probe that can be displaced along the borehole axis. In this way the pres-
94 5. Surface and Guided Waves along Cylindrical Interfaces
sure at different positions can be recorded. Due to the high reproducibility of the
experiments these recordings can be combined in a single array of measurements
consisting of 25 different pressure signals. The only source of discrepancy between
the different experiments is a slight variation of the amplitude of the incident wave in
the water. This amplitude variation is of the order of 1% at most. In order to correct
for this variation, the measured pressure in the borehole is normalized with respect
to the amplitude of the incident wave recorded at the transducer P1 (see Figure 4.9),
which is located above the upper aperture of the funnel. This transducer is also used
to trigger the data acquisition system.
We used the three samples mentioned earlier, a natural Berea sandstone sample
and two synthetic samples N1 and N2. The synthetic samples have a relatively high
permeability in comparison with ordinary rocks. They are made of sintered crushed
glass. In this way we cover a range of permeability values from 360 mD to 10.8 D.
The properties of the samples are given in Table 5.1.
In order to obtain the frequency-dependent phase velocities and specific atten-
uations, the pressure signals are transformed from the space-time domain to the
frequency-wavenumber domain. This is performed using a combined FFT-Prony-
Spectral Ratio technique. This newly developed method uses the output for the
frequency-dependent phase velocities from the conventional FFT-Prony method as
discussed in Chapter 4, as the input for the spectral ratio technique. This extra step
significantly improves the determination of the specific attenuation Q −1 .
In this section, the main experimental results are summarized. The first topic to be
addressed is the influence on the transmitted waves in the borehole of the enforced
energy canalization. In the previous experiments performed in the fully saturated
Berea sandstone, it was possible to distinguish two different arrivals. These arrivals
are associated with the fast P wave and the pseudo-Stoneley wave propagating along
the interface (see Figure 4.11). The aim of the funnel is to enhance the excitation
of the surface waves and in this particular case of the pseudo-Stoneley wave. Figure
5.8 shows the measured pressure as a function of time for representative experiments
performed with the funnel installed in the shock tube. With respect to Figure 4.11, the
P-wave arrival has clearly been suppressed. Also visible is a precursor mode which is
associated with compliance effects in the shock tube wall and should not be confused
with the neat P wave precursor detected in the previous series of experiments shown
in Figure 4.11.
The application of the FFT-Prony method in the complete series of measurements
performed in the Berea sandstone revealed the presence of only one wave mode, the
pseudo-Stoneley wave, as shown in Figure 5.9a. As in the previous chapter, there
5.4. Experimental Results 95
5
z=6.1 cm
Pressure (bar)
4
z=5.1 cm
3
2
z=4.1 cm
1
PSfrag replacements
z=3.1 cm
0
pseudo-Stoneley wave
−1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Time (ms)
Figure 5.8: Pressure recordings at different depths. The position of the pressure transducer
(z) is measured from the top of the sample. For readability, traces have been
shifted upwards.
is an excellent agreement between the experimental data and the theoretical predic-
tions for the frequency-dependent phase velocity. We remark that we were able to
measure frequencies as low as 750 Hz. Next the FFT-Prony-Spectral Ratio method
was implemented to process the attenuation data. The results in terms of the spe-
cific attenuation coefficient Q−1 are depicted in Figure 5.9b. Now there is also an
excellent agreement between theory and experiments for Q −1 , which is an enormous
improvement with respect to the previous chapter (see Figure 4.12). Apparently, both
the improvement of the shock tube setup and the improved data processing technique
result in an almost complete elimination of the scatter in the attenuation data that we
previously encountered. As an overall conclusion we state that regarding the pseudo-
Stoneley wave, the quality of the experimental data has improved considerably due
to the suppression of the fast compressional wave, which to a certain extent over-
lapped the surface wave in the previous series of experiments. We remark that the
numerical calculations for Q−1 slightly underestimate the attenuation measured, par-
ticularly at high frequencies (> 40 kHz). The shock wave experiments show that the
oscillating viscous flow at the borehole wall, which is the predominant dissipative
mechanism for the pseudo-Stoneley wave is properly described by the Biot theory
for frequencies lower than 40 kHz, in spite of the fact that it has been documented in
literature that the Biot theory underestimates the attenuation of seismic and acoustic
P and S waves observed in field surveys and laboratory experiments in natural rocks
(see e.g. Dvorkin and Nur (1993); Gist (1994a); Dvorkin et al. (1995); Pride et al.
(2004) and references therein). At frequencies higher than 40 kHz, the systematic
96 5. Surface and Guided Waves along Cylindrical Interfaces
1500
1000
500
PSfrag replacements 0
0 10 20 30 40 50 60 70 80
Frequency (kHz)
(a)
2
1.8
1.6
1.4
1.2
Q−1
0.8
0.6
0.4
0.2
PSfrag replacements
0
0 10 20 30 40 50 60 70 80
Frequency (kHz)
(b)
Figure 5.9: Frequency-dependent phase velocity (a) and specific attenuation coefficient (b)
for the pseudo-Stoneley wave in the sandstone confined reservoir. The experi-
mental results are shown in dots and the theoretical calculations based on Biot’s
theory in solid lines.
5.4. Experimental Results 97
4500
4000
3500
2500
L3
2000
1500
1000
p-Stoneley
PSfrag replacements 500
0
0 10 20 30 40 50
Frequency (kHz)
Figure 5.10: Frequency-dependent phase velocities and specific attenuation coefficient for
the pseudo-Stoneley wave and the L3 mode in the synthetic N1 confined reser-
voir. The numerical calculations based on Biot’s theory are displayed in solid
lines.
1.5
1
Q−1
0.5
PSfrag replacements 0
0 10 20 30 40 50
Frequency (kHz)
(a)
0.7
0.6
0.5
0.4
Q−1
0.3
0.2
0.1
PSfrag replacements 0
30 40 50 60 70 80 90 100
Frequency (kHz)
(b)
Figure 5.11: Specific attenuation coefficients Q−1 measured in the shock wave experiments
performed in the N1 formation. The experimental results for the pseudo-
Stoneley wave (a) and the L3 mode (b) are compared with the numerical calcu-
lations (solid lines).
5.4. Experimental Results 99
3500
3000
Phase Velocity (m/s)
2500
2000
1500
−1
10
Q−1
−2
10
−3
10
PSfrag replacements −4
10
0 20 40 60 80 100 120 140 160
Frequency (kHz)
(b)
Figure 5.12: Experimental obtained velocities and specific attenuation coefficients for the L2
mode (fast P wave) in the N2 reservoir. In solid lines the numerical results for
the L2 mode are shown. Biot’s predictions for the velocities and attenuation
coefficients of the fast compressional wave are displayed in dashed lines.
100 5. Surface and Guided Waves along Cylindrical Interfaces
Finally we discus the experimental data obtained in the N2 sample. In this case
only one mode was detected and therefore the full FFT-Prony-Spectral ratio method
could be used. This mode is associated with the fast compressional wave and ac-
cording to the notation introduced in the subsection 5.2.3 it is denoted L2. The
pseudo-Stoneley wave is not observed due to the high permeability of the porous
sample, which implies that the attenuation of this wave is such that its amplitude is
significantly damped and therefore not observed along the distances covered in the
measurements. Figure 5.12 depicts the outcome of the experiments and the compar-
ison with the numerical results for the L2 mode in the confined reservoir configura-
tion. In Figure 5.12 also the fast compressional wave is shown in dashed lines. For
this sample the correlation between the L2 mode and the fast P wave extends to the
complete range of frequencies of interest (1-160 kHz). The agreement between the
experimental data and the theoretical predictions for the wave velocity is fair with
an overestimation of the theory at low frequencies and a slight underestimation at
high frequencies. The attenuation of the L2 mode reveals the presence of regular
oscillatory patterns. These oscillations are associated with radial resonances in the
cylindrical cavity determined by the shock tube. Remarkably, the experimental at-
tenuation data show regularly spaced maxima and minima which correspond to the
oscillations predicted theoretically for the L2 mode. As an overall conclusion we re-
mark that the attenuation measured is one order of magnitude higher than the values
predicted numerically.
5.5 Conclusions
A novel shock wave experimental technique was developed to study the frequency-
dependent properties of the guided waves in a broad band of frequencies (1-160 kHz).
We were able to excite different kind of guided waves in several samples covering a
broad spectrum of material properties. The results for the phase velocity and specific
attenuation coefficient for a Berea sandstone material showed an excellent agreement
with numerical calculations based on Biot’s theory. We note that using this technique
we obtain experimental data at frequencies as low as 750 Hz which is a major im-
provement with respect to the experimental data available in literature (Winkler et al.,
1989; Chao et al., 2004).
Other guided modes are also successfully measured. The L2 mode is associated
with the fast compressional wave in the porous material. Measurements of this wave
performed in a highly permeable sample revealed an underestimation of the model
predictions to describe the level of attenuation observed in the experimental data.
The comparison between the numerical results based on Biot’s theory and the
outcome of our experiments allows us to draw some conclusions over the attenua-
tion mechanisms governing the different waves. It appears obvious that the fluid-
5.5. Conclusions 101
exchange interaction at the borehole wall which is the principal loss mechanism for
the pseudo-Stoneley wave is properly described by Biot theory at frequencies below
40 kHz. At higher frequencies, there is a systematic underestimation of the numer-
ical predictions which can be attributed to the existence of other loss mechanisms
neglected in the Biot model.
The outcome of this study may be of significance in a broad band of problems
related to exploration methods and reservoir characterization. The study of the propa-
gation of the borehole modes under well-defined laboratory conditions removes sev-
eral of the uncertainties encountered in field measurements. In this sense our ex-
perimental technique can be a valuable tool in the validation of theoretical models,
inversion algorithms and interpretation of seismic and borehole acoustic data.
102 5. Surface and Guided Waves along Cylindrical Interfaces
6. SEISMIC SIGNATURES OF PARTIAL SATURATION ON ACOUSTIC
BOREHOLE MODES
6.1 Introduction
One of the main objectives of logging techniques is the determination of the volume
fraction of the fluids saturating the pore space of a reservoir. Commonly oil, water
and gas coexist in prospective reservoirs and an accurate assessment of the oil content
is required in order to assert the economical value of the reservoir. It is common prac-
tice to estimate the oil saturation using resistivity logs by means of semi-quantitative
methods based on the empirical formula postulated by Archie (1942) (see e.g. Hearst
et al. (2000)). There are numerous limitations in the application of these techniques,
mainly related to the assumption that the reservoir rock is an electric insulator, which
is certainly not true for rocks with rich clay content. Therefore it is interesting to ex-
plore the possibility of the application of other logging techniques in order to extract
information about the saturation of a given reservoir.
Acoustic borehole methods have evolved rapidly since the early times when only
information about the porosity and the elastic velocities in the formation were ob-
tained through the identification of the distinct arrivals of the head waves. In recent
years, there has been an increasing interest in the interpretation of “full waveform”
acoustics logs in order to characterize the poroelastic properties of the reservoir and
the saturating fluid (see e.g. White (1983); Paillet and Cheng (1991); Tang and Cheng
(2004)). The emphasis of these investigations was directed to understand the relation
between the permeability and the dispersive velocity and attenuation of the pseudo-
Stoneley wave propagating along the borehole wall (Rosenbaum (1974); Schmitt
et al. (1988); Winkler et al. (1989); Tang and Cheng (1996); Chao et al. (2004)).
From these theoretical and experimental studies it has became obvious that a strong
dependence of the attenuation of the pseudo-Stoneley wave on the permeability can
be expected, though other dissipative mechanisms can influence the propagation of
this surface wave.
The purpose of this work is to study the influence of relatively small amounts
of gas in liquid saturated formations on the features of the pseudo-Stoneley wave.
We assume that the reservoir is saturated by a mixture of brine and air. The acous-
tic bulk properties of the partially saturated reservoir are described using a modi-
104 6. Seismic Signatures of Partial Saturation on Acoustic Borehole Modes
fied Biot’s theory of poromechanics according to the model described in section 3.2
(Biot (1956a,b); Smeulders and van Dongen (1997); Smeulders et al. (1992)). The
frequency-dependent properties of the pseudo-Stoneley wave are investigated in the
complete seismic and borehole band of frequencies (10 Hz- 50 kHz). We study the
sensitivity of the phase velocity and attenuation of the pseudo-Stoneley wave to vari-
ations on the gas saturation. We will also study the low-frequency limit where the
radial dependence of the pressure and the liquid displacement inside the borehole
can be neglected. In this limit it is possible to obtain an analytical solution for the
dispersive velocity and attenuation of the pseudo-Stoneley wave (White (1983); Nor-
ris (1989); Tang et al. (1991), section 2.3.2). These results are compared with the
rigorous solution based on a displacement potential formulation of the modified Biot
theory, which is also valid at higher frequencies, in order to determine the range of
validity of the approximate analytical solution.
The organization of this chapter is as follows: In section 6.2 we analyzed the nu-
merical results for the phase velocity and damping coefficient of the pseudo-Stoneley
wave for several saturation values. Then we proceed with the comparison between the
low-frequency analytical solution and the full numerical solution. This is followed
by the analysis of the influence of the gas bubbles radius in the frequency-dependent
speed and attenuation at a fixed liquid saturation. For a characteristic frequency we
study the pore-pressure radial distribution induced by the pseudo-Stoneley wave for
different liquid saturations. Finally we analyze quantitatively the sensitivity of both
the phase velocity and damping coefficient of the pseudo-Stoneley mode to the gas
fraction in the pore space. The conclusions are drawn in section 6.3.
R = 7 .5 c m
r
z
p a r tia lly
s a tu ra te d
p o ro u s
b r in e
fo r m a tio n
Figure 6.1: Borehole in a partially saturated reservoir. The properties of the formation are
given in Table 6.1
Figure 6.3 displays the phase velocities and damping coefficients of the pseudo-
Stoneley wave as a function of frequency for different brine saturations s. The phase
velocity shows a strong dependence on brine saturation at low frequencies. Due to
compressibility effects the phase velocity decreases with decreasing saturation in the
low-frequency range (10 Hz- 5 kHz). This relation is reversed at higher frequencies
where the loss mechanisms induced by the gas bubble oscillations become significant.
The damping coefficient in Figure 6.3b is expressed as the imaginary part of the
complex valued wavenumber of the pseudo-Stoneley wave. A notable dependence
on the brine saturation is found in our numerical calculations. Attenuation increases
with decreasing brine saturation over the complete range of frequencies (10 Hz-50
kHz). Frequencies involved in acoustic borehole logging range from 5 kHz to 30
kHz. In this band of frequencies the effect of saturation on the damping coefficient
is significant. The values of the damping coefficient for representative frequencies
of 7.5 kHz, 10 kHz and 20 kHz are listed in Table 6.2. We obtain discrepancies of
approximately a factor 2 when a gas fraction of only 1 % is introduced. These results
suggest that the attenuation of the pseudo-Stoneley wave is highly sensitive to the gas
fraction in the pore space.
In order to explain the frequency-dependent behavior of the phase velocity and
damping coefficient and their relation with the liquid saturation, it is illustrative to
examine the complex-valued frequency-dependent bulk modulus of the mixture sat-
urating the pore space of the porous medium. The complex-valued bulk modulus is
depicted in Figure 6.2. The real part accounts for the compressibility of the mix-
106 6. Seismic Signatures of Partial Saturation on Acoustic Borehole Modes
Table 6.1: Physical properties of the rock and the saturating fluids.
ture and the imaginary part reflects the loss mechanisms induced by the oscillations
of the gas bubbles. For low frequencies the real part is frequency-independent and
can be obtained by a simple average of the steady-state bulk modulus of the liquid
and the gas. At these frequencies is also obvious from Figure 6.2b that the dissipative
mechanisms induced by the oscillations of the gas bubbles are inactive, therefore only
compressibility effects are responsible for the change on the properties of the pseudo-
Stoneley wave with respect to the fully saturated case. As the frequency is increased
an important decrease of the compressibility is observed and also the damping mech-
anisms become important. We note that for frequencies below 30 kHz the bulk loss
mechanisms decrease with increasing air fraction. However, the damping coefficient
of the pseudo-Stoneley wave increases significantly with increasing gas volume. It is
evident that the propagation of the pseudo-Stoneley wave is mainly governed by the
frequency-dependent compressibility of the mixture. In spite of the fact that the loss
mechanisms may have an influence on the pseudo-Stoneley wave it is obvious they
are not responsible for the clear trend found in the damping coefficient.
It is attractive to examine the very low-frequency limit of the pseudo-Stoneley
wave because of the feasibility of an approximate analytical solution for the
frequency-dependent speed and damping coefficient. This analytical solution can
be used in inversion procedures to infer the saturation of the reservoir. Under the ba-
sic assumption that the wavelengths are sufficiently larger than the borehole radius,
6.2. Borehole Waves in Partially Saturated Porous Media 107
1.2
0.8 s=0.99
Re(Kf (ω))/Kl
s=0.95
0.6
0.4
0.2 s=0.9
PSfrag replacements
0
s=0.8
−0.2 1 2 3 4 5
10 10 10 10 10
Frequency (Hz)
(a)
0.9
0.8
0.7
Im(Kf (ω))/Kl
0.6 s=0.8
0.5
s=0.9
0.4
0.3
0.2
PSfrag replacements
the liquid motion inside the borehole can be assumed to be independent of the radial
coordinate. Furthermore, the fluid flow in the pore space is effectively described by
Darcy’s law. In this case a quasi-one dimensional model can be employed to describe
the pseudo-Stoneley wave propagation. This model has been derived in section 2.3.2
108 6. Seismic Signatures of Partial Saturation on Acoustic Borehole Modes
1500
1400
1300
Phase Velocity (m/s) 1200
1100
1000
s=1
900
800 s=0.99
PSfrag replacements
700
s=0.95
600 s=0.9
s=0.8
500 1 2 3 4
10 10 10 10
Frequency (Hz)
(a)
1
10
Damping coefficient (1/m)
0
10
s=0.8
s=0.9
s=0.95
−1
PSfrag replacements 10 s=0.99
s=1
1 2 3 4
10 10 10 10
Frequency (Hz)
(b)
Figure 6.3: Frequency-dependent phase velocity (a) and damping coefficient (b) of the
pseudo-Stoneley wave for different values of brine saturation s.
and here we modified it by considering the bulk modulus of the mixture saturating the
pore space to follow the frequency-dependent description given in section 3.2. Fig-
ure 6.4 shows the outcome of the phase velocity and damping coefficient. In order
to assess the validity of the low-frequency solution, also the full solution is plotted.
6.2. Borehole Waves in Partially Saturated Porous Media 109
Table 6.2: Damping coefficients of the pseudo-Stoneley wave for different saturations at rep-
resentative frequencies.
The quasi-1D solution reproduces the main features of the pseudo-Stoneley wave,
though its range of application seems to be confined to frequencies well below the
frequency band utilized in acoustic borehole measurements. Appreciable deviations
from the full solution are found for frequencies higher than 100 Hz for the phase
velocity and 800 Hz for the damping coefficient. Chang et al. (1988) have found
similar limitations for a low-frequency approximation of the full solution for the fully
liquid-saturated reservoir. They stated that the range of validity of their approximate
solution is restricted to frequencies below 300 Hz.
The bubble radius is a critical parameter which conditions the bubble dynamics
and therefore the acoustic properties of the partially saturated porous medium. We
analyze its influence on the pseudo-Stoneley wave. Figure 6.4 shows the calculations
for the phase velocity and the damping coefficient for two different bubble radii. For
comparison, also the dispersion curves for the fully brine-saturated case are shown in
dashed lines. We observe that by increasing the radius of the bubbles the dependence
on saturation decreases drastically and it is restricted to the low-frequency part of
the spectrum. With the aid of Figure 6.2a we can conclude that due to the increase
of compressibility at lower frequencies for the larger bubble radius, the range of
frequencies where the saturation influences the pseudo-Stoneley wave is restricted to
frequencies where the contrast of compressibility is significant.
Next we examine the radial penetration in the porous formation of the pseudo-
Stoneley wave. We compute the pore-pressure as a function of the radial coordinate
for different liquid saturations. We restrict our analysis to a characteristic frequency
of 10 kHz. The results are displayed in Figure 6.5. The radial penetration decreases
with decreasing brine saturation. This can be explained taking into account that
the radial penetration is inversely proportional to the compressibility of the mixture.
From Figure 6.2a it is clear that at 10 kHz the real part of the bulk modulus of the
mixture increases with brine saturation and therefore the compressibility decreases.
Finally the sensitivity of the phase velocity and the damping coefficient to the
variation of liquid saturation is analyzed. In order to obtain a quantitative estimation
110 6. Seismic Signatures of Partial Saturation on Acoustic Borehole Modes
1500
1400 r = 1 cm
1300
Phase Velocity (m/s) 1200
r = 1 mm
1100
1000
s=1
900
800
700
PSfrag replacements
600
500 1 2 3 4
10 10 10 10
Frequency (Hz)
(a)
1
10
r = 1 mm
Damping coefficient (1/m)
0
10
r = 1 cm
−1
10
s=1
PSfrag replacements
−2
10 1 2 3 4
10 10 10 10
Frequency (Hz)
(b)
Figure 6.4: Low-frequency quasi-one dimensional analytical solution for the phase velocity
(a) and specific attenuation coefficient (b) of the pseudo-Stoneley wave (dotted
lines). The radius of the bubbles is 1 mm and s = 0.95. The full solutions for
different bubble radius r for s = 0.95 are shown in solid lines. For comparison
also the full solution for s = 1 is displayed in dashed lines.
6.3. Conclusions 111
0.9
0.8
0.7
s=1
|p/p(R)|
0.6
0.5
s=0.99
0.4
0.3
6.3 Conclusions
This study reports numerical results on the influence of the gas saturation on the
propagation of the pseudo-Stoneley wave along a borehole surrounded by a partially
112 6. Seismic Signatures of Partial Saturation on Acoustic Borehole Modes
2
10
1
10
Sensitivity
0
10
ς
D
−1 ς
10 v
−2
PSfrag replacements 10
0.7 0.75 0.8 0.85 0.9 0.95 1
s
Figure 6.6: Sensitivity analysis of the saturation for the phase velocity and damping coeffi-
cient of the pseudo-Stoneley wave.
The aim of this thesis was to study the dispersive properties of the surface acoustic
waves that propagate along poroelastic materials. The main focus was towards the
understanding of the underlying physical mechanisms which govern the dynamics of
the surface waves and the sensitivity of the features of these modes to the different
properties of the medium. In this respect, the attenuation plays a fundamental role
due to its intrinsic relation to the flow-induced loss mechanisms. In this work two
fundamental dissipative mechanisms for the surface waves are considered, the wave-
induced exchange of fluid at the interface and the losses generated due to the presence
of a small amount of oscillating gas bubbles in the liquid-saturated pore space.
The surface waves that propagate along a borehole wall constitute an important
tool for the permeability assessment of a reservoir due to the high sensitivity of these
modes to the permeability. Among the different modes that propagate in this configu-
ration, the so-called tube wave is by far the most affected by the permeability and it is
the surface mode that is usually clearly observed in the acoustic recordings. It is im-
portant to remark that the tube wave is the low-frequency limit of the pseudo-Stoneley
wave. In Chapter 4, a detailed analysis of the permeability dependence of the pseudo-
Stoneley mode was given. A strong dependence was obtained for the attenuation in
the complete band of frequencies studied (100 Hz-200 kHz) while for the phase ve-
locity the sensitivity on the permeability is restricted to the low-frequency range (be-
low 10 kHz). It is shown in Chapter 4 that the significant permeability dependence
at low frequencies can be explained by considering the different bulk wave contribu-
tions to the average solid-liquid radial displacement induced by the pseudo-Stoneley
wave. The numerical calculations indicate that the amplitude associated with the
slow compressional wave is two orders of magnitude larger than that of the fast com-
pressional wave and the shear wave. Due to the intrinsic permeability-dependent
nature of the slow compressional wave, it can be asserted that these calculations pro-
vide a quantitative explanation for the low-frequency permeability sensitivity of the
pseudo-Stoneley wave.
A novel experimental technique has been developed to study borehole modes un-
der well-defined laboratory conditions. This technique is based on the use of shock
waves to excite the guided modes that propagate along a concentrically drilled bore-
hole in a cylindrical porous core sample. A crucial factor for the excitation of the
114 7. Conclusions and Recommendations
surface modes is the loading of the borehole liquid column. The most favorable re-
sults were obtained for a modified setup in which the energy of the incident wave was
canalized into the borehole liquid. Using this setup, accurate measurements of the
frequency-dependent phase velocity and intrinsic attenuation of the pseudo-Stoneley
wave along boreholes in water-saturated porous reservoirs were performed. The ex-
perimental results show an excellent agreement with numerical calculations based on
Biot’s theory. It appears obvious that the fluid exchange interaction at the borehole
wall, which is the main loss mechanism for the pseudo-Stoneley wave, is properly de-
scribed by the Biot theory for frequencies below 40 kHz. At higher frequencies, there
is a systematic underestimation of the numerical predictions which can be attributed
to the existence of other loss mechanisms neglected in the Biot model. Experimental
data for the velocity and attenuation of the surface modes at frequencies as low as 750
Hz are reported in this work, which constitutes a major improvement with respect to
the experimental data available in literature.
The effects of partial saturation on the surface waves along a liquid/poroelastic
plane interface were numerically investigated in Chapter 3. In this case the compress-
ibility variation and loss mechanisms induced by the oscillation of the gas bubbles
are described by means of a frequency-dependent bulk modulus of the mixture of
fluids (water and gas). The results indicate a high sensitivity of the intrinsic attenua-
tion of the pseudo-Stoneley wave to the liquid saturation. A maximum attenuation is
predicted at a characteristic frequency and attenuation level that considerably depend
on liquid saturation. In the high-frequency limit, where only compressibility affects
the surface modes, a neat transition between the leaky pseudo-Stoneley wave and the
true Stoneley wave is found at a characteristic saturation. For lower saturation values,
the slow compressional wave propagates faster than the acoustic wave in the liquid
and the true Stoneley wave propagates with no attenuation. For higher saturations, on
the other hand, this surface mode radiates energy into the slow compressional wave
and becomes a leaky pseudo-Stoneley wave.
A more realistic situation was studied in Chapter 6, where the interface is cylin-
drical (borehole configuration) and a porous reservoir saturated by a brine/air mixture
was considered. The main issue addressed in this thesis is whether information on the
saturation of a reservoir can be inferred using surface acoustic data. Consistent with
the results for the plane configuration, an important dependence of the attenuation on
the liquid saturation was found. The damping coefficient shows a clear increase with
increasing gas fraction throughout the band of frequencies studied (10 Hz-50 kHz).
Remarkably, in the band of frequencies related to borehole acoustic techniques, the
variation and the sensitivity of the damping coefficient to changes in saturation are
particularly significant. From the analysis of the frequency-dependent properties of
the phase velocity and damping coefficient of the pseudo-Stoneley wave in terms of
the complex-valued bulk modulus of the mixture, it can be concluded that their de-
115
In dit proefschrift wordt verslag gelegd van een theoretisch en experimenteel on-
derzoek naar de dispersieve eigenschappen van akoestische oppervlaktegolven in
poro-elastische materialen. Een numerieke studie is uitgevoerd naar de invloed van
de verzadigingsgraad van de poriën op de oppervlaktegolven voor rechte en cilin-
dervormige oppervlakken. De theoretische formulering is hierbij gebaseerd op het
tweefase model van Biot, dat de frequentieafhankelijke wrijvingsaspecten tussen
vloeistof en korrels in rekening brengt. Door een schok opgewekte golven zijn ge-
bruikt om bij waterverzadigde poreuze materialen de theoretische bevindingen ex-
perimenteel te onderbouwen.
In een breed frequentiespectrum (100 Hz - 200 kHz) zijn de eigenschappen van
de pseudo-Stoneleygolf in een boorgat in een poreus gesteente bestudeerd met be-
hulp van verplaatsingspotentialen. De invloed van de permeabiliteit op de fases-
nelheid, demping, radiale verplaatsing en poriedruk is onderzocht, en de bijdrage
van de verschillende volumegolven aan de gemiddelde radiale verplaatsing is in het
frequentiedomein geanalyseerd. De numerieke resultaten tonen aan dat de perme-
abiliteitsafhankelijkheid bij lage frequenties wordt veroorzaakt door de langzame vol-
umegolf van Biot, doordat deze een relatief grote bijdrage heeft in de verplaatsingen
en drukken veroorzaakt door de pseudo-Stoneleygolf.
Akoestische experimenten zijn uitgevoerd waarbij een schokbuis werd gebruikt
om de pseudo-Stoneleygolf op te wekken in een watergevuld boorgat in een wa-
terverzadigde cilinder van Berea zandsteen. Op deze manier zijn fasesnelheden en
dempingscoëfficiënten gemeten. Er bestaat een uitstekende kwantitatieve
overeenkomst tussen experimentele en theoretische resultaten voor wat betreft de fas-
esnelheid in het frequentiedomein van 10 tot 50 kHz. De waarden voor de demping
wijken in een oscillerend patroon van de theorie af, en vertonen alleen een kwali-
tatieve overeenkomst hiermee. Om de kwaliteit van de dempingsmetingen te ver-
beteren werd de schokbuis aangepast. Een trechtervormige constructie werd juist
boven het monster in de buis gemonteerd, met als doel de akoestische energie het
boorgat in te leiden en de opwekking van oppervlaktegolven te verbeteren. Boven-
dien werd er een nieuwe ”FFT-Prony-Spectral Ratio” methodiek geı̈mplementeerd
om de meetgegevens van het ruimte-tijddomein naar het golfgetal-frequentiedomein
te converteren. De nauwkeurigheid van de waarden voor de demping van de Berea
118 SAMENVATTING
taten tonen aan dat de samendrukbaarheid van het vloeistofmengsel het dynamische
gedrag van de pseudo-Stoneleygolf bepaalt en verantwoordelijk is voor het verzadi-
gingsafhankelijke gedrag in de fasesnelheid, demping en poriedruk. Een kwanti-
tatieve studie bewijst dat de dempingscoëfficiënt erg gevoelig is voor kleine hoeveel-
heden gas in de poriën. Dit gedrag geldt niet alleen voor volumegolven, maar ook
voor oppervlaktegolven.
120 SAMENVATTING
BIBLIOGRAPHY
Adler, L. and Nagy, P.B. (1994). Measurements of acoustic surface waves on fluid-
filled porous rocks. J. Geophys. Res. B, 99, 17863–17869.
Allard, J.F., Jansens, G., Vermeir, G., and Lauriks, W. (2002). Frame-borne surface
waves in air-saturated porous media. J. Acoust. Soc. Am., 111(2), 690–696.
Allard, J.F., Henry, M., Glorieux, C., Petillon, S., and Lauriks, W. (2003). Laser-
induced surface modes at an air-porous medium interface. J. Appl. Phys., 93(2),
1298–1304.
Allard, J.F., Henry, M., Glorieux, C., Lauriks, W., and Petillon, S. (2004). Laser-
induced surface modes at water-poroelastic solid interfaces. J. Appl. Phys., 95(2),
528–535.
Archie, G.E. (1942). The electrical resistivity log as an aid in determining some
reservoir characteristics. Trans AIME, 146, 54–62.
Auriault, J.L., Borne, L., and Chambon, R. (1985). Dynamics of porous saturated
media, checking of the generalized law of Darcy. J. Acoust. Soc. Am., 77, 1641–
1650.
Berryman, J. G., Thigpen, L., and Chin, R.C.Y. (1988). Bulk elastic wave propagation
in partially saturated porous solids. J. Acoust. Soc. Am., 84(1), 360–373.
Biot, M. A. and Willis, D. G. (1961). The elastic coefficients of the theory of consol-
idation. J. Appl. Mech., 24, 594–601.
Bourbié, Th., Coussy, O., and Zinszner, B. (1987). Acoustics of porous media. Gulf
publishing company.
Brown, P., Batzle, M., Peeters, M., Dey-Sakar, S., and Steensma, G. (2000). Shock
tube experiments and the observation of the Biot slow wave in natural rocks, pages
1846–1849. Soc. of Expl. Geophys.
Cadoret, T., Marion, D., and Zinszner, B. (1995). Influence of frequency and fluid
distribution on elastic wave velocities in partially saturated limestones. J. Geophys.
Res. B, 100, 9789–9803.
Cadoret, T., Mavko, G., and Zinszner, B. (1998). Fluid distribution effect on sonic
attenuation in partially saturated limestones. Geophysics, 63, 154–160.
Carcione, Jose M., Helle, Hans B., and Pham, Nam H. (2003). White’s model for
wave propagation in partially saturated rocks: Comparison with poroelastic nu-
merical experiments. Geophysics, 68, 1389–1398.
Chang, S.K., Liu, H.L., and Johnson, D.L. (1988). Low-frequency tube waves in
permeable rocks. Geophysics, 53, 519–527.
Chao, G., Smeulders, D.M.J., and van Dongen, M.E.H. (2004). Shock-induced bore-
hole waves in porous formations: theory and experiments. J. Acoust. Soc. Am.,
116(2), 693–702.
Cheng, C.H., Toksoz, M.N., and Willis, M.E. (1982). Determination of in situ atten-
uation from full waveform acoustic logs. J. Geophys. Res. B, 87, 5477–5484.
Cheng, C.H., Jinzhong, Z., and Burns, D.R. (1987). Effects of in-situ permeability
on the propagation of Stoneley (tube) waves in a borehole. Geophysics, 52, 1279–
1289.
BIBLIOGRAPHY 123
Degrande, G., Roeck, G. De, Broeck, P. Van Den, and Smeulders, D. (1998). Wave
propagation in layered dry, saturated and unsaturated poroelastic media. Int. J.
Solids Struct., 35, 4753–4778.
Denneman, A.I.M., Drijkoningen, G.G., Smeulders, D.M.J., and Wapenaar, K.
(2002). Reflection and transmission of waves at a fluid/porous-medium interface.
Geophysics, 67(1), 282–291.
Deresiewicz, H. (1960). The effect of boundaries on wave propagation in a liquid-
filled porous solid: I. Reflection of plane waves at a free plane boundary (non-
dissipative case). Bull. Seism. Soc. Am., 50, 599–607.
Deresiewicz, H. (1961). The effect of boundaries on wave propagation in a liquid-
filled porous solid: II. Love waves in a porous layer. Bull. Seism. Soc. Am., 51,
51–59.
Deresiewicz, H. and Skalak, R. (1963). On uniqueness in dynamic poroelasticity.
Bull. Seism. Soc. Am., 53, 783–788.
Dutta, N.C. and Ode, H. (1979a). Attenuation and dispersion of compressional waves
in fluid-filled porous rocks with partial gas saturation (White model)- Part I: Biot
theory. Geophysics, 44, 1777–1788.
Dutta, N.C. and Ode, H. (1979b). Attenuation and dispersion of compressional waves
in fluid-filled porous rocks with partial gas saturation (White model)- Part II: Re-
sults. Geophysics, 44, 1789–1805.
Dutta, N.C. and Seriff, A.J. (1979). On White’s model of attenuation in rocks with
partial gas saturation. Geophysics, 44, 1806–1812.
Dvorkin, Jack and Nur, Amos (1993). Dynamic poroelasticity: A unified model with
the squirt and the Biot mechanisms. Geophysics, 58, 524–533.
Dvorkin, Jack, Mavko, Gary, and Nur, Amos (1995). Squirt flow in fully saturated
rocks. Geophysics, 60, 97–107.
Edelman, I. (2004). Surface waves at vacuum/porous medium interface: low fre-
quency range. Wave Motion, 39, 111–127.
Edelman, I. and Wilmanski, K. (2002). Asymptotic analysis of surface waves at
vacuum/porous medium and liquid/porous medium interfaces. Continuum Mech.
Thermodyn., 14, 25–44.
Ellefsen, K.J., Burns, D.R., and Cheng, C.H. (1993). Homomorphic processing of
the tube wave generated during acoustic logging. Geophysics, 58, 1400–1407.
124 BIBLIOGRAPHY
Gubaidullin, A.A., Kuchugurina, O.Yu., Smeulders, D.M.J., and Wisse, C.J. (2004).
Frequency-dependent acoustic properties of a fluid/porous solid interfce. J. Acoust.
Soc. Am., 116(3), 1474–1980.
Hearst, Joseph R., Nelson, Philip H., and Paillett, Frederick L. (2000). Well Logging
for Physical Properties. A Handbook for Geophysicists, Geologists and Engineers.
John Wiley and Sons Ltd.
Henry, M., Lemarinier, P., Allard, J.F., Bonardet, J.L., and Gedeon, A. (1995). Eval-
uation of the characteristic dimensions for porous sound-absorbing materials. J.
Appl. Phys., 77, 17–20.
Hess, P. (2002). Surface acoustic waves in material sciences. Phys. Today, 55(3),
42–47.
Hsu, C-J, Kostek, S., and Johnson, D.L. (1997). Tube waves and mandrel modes:
Experiment and theory. J. Acoust. Soc. Am., 102, 3277–3289.
Johnson, D.L., Koplik, J., and Dashen, R. (1987). Theory of dynamic permeability
and tortuosity in fluid-saturated porous media. J. Fluid. Mech., 176, 379–402.
Kelder, O. and Smeulders, D.M.J. (1997). Observation of the Biot slow compres-
sional wave in water-saturated Nivelsteiner sandstone. Geophysics, 62, 1794–
1796.
Liu, H.L. and Johnson, D.L. (1997). Effects of an elastic membrane on tube waves
in permeable formations. J. Acoust. Soc. Am., 101, 3322–3329.
Lucet, N., Rasolofosaon, P. N., and Zinszner, B. (1991). Sonic properties of rocks
under confining pressure using the resonan bar technique. J. Acoust. Soc. Am., 89,
980–990.
Mavko, G. and Nur, A. (1979). Wave attenuation in partially saturated rocks. Geo-
physics, 44, 161–178.
Mavko, Gary, Mukerji, Tapan, and Dvorkin, Jack (1998). The Rock Physics Hand-
book. Tools for Seismic Analysis in Porous Media. Cambridge University Press.
Mayes, Michael J., Nagy, Peter B., and Adler, Laszlo (1986). Excitation of surface
waves of different modes at fluid-porous solid interface. J. Acoust. Soc. Am., 79(2),
249–252.
Maznev, A. A., Mazurenko, A., Zhuoyun, L., and Gostein, M. (2003). Laser-based
surface acoustic wave spectrometer for industrial applications. Rev. Sci. Instrum.,
74(1), 667–669.
Nagy, P.B., Adler, L., and Bonner, B.P. (1990). Slow wave propagation in air-filled
porous materials and natural rocks. Appl. Phys. Lett., 56, 2504–2506.
126 BIBLIOGRAPHY
Paillet, F.L. and Cheng, C.H. (1991). Acoustic waves in boreholes- the theory and
application of acoustic full-waveform logs. CRC Press.
Pride, S.R. and Berryman, J.G. (2003a). Linear dynamics of double-porosity dual-
permeability materials. I. Governing equations and acoustic attemuation. Phys.
Rev. B, 68, 036603.
Pride, S.R. and Berryman, J.G. (2003b). Linear dynamics of double-porosity dual-
permeability materials. II. Fluid transport equations. Phys. Rev. B, 68, 036604.
Pride, S.R., Berryman, J.G., and Harris, J.M. (2004). Seismic attenuation due to
wave-induced flow. J. Geophys. Res. B, 109, B01201.
Raum, Kay, hauer, Jan Reiß, and Brandt, Jörg (2004). Frequency and resolution
dependence of the anisotropic impedance estimation in cortical bone using time-
resolved scanning acoustic microscopy. J. Biomed. Mat. Res-A, 71A(3), 430–438.
Santos, J.E., Corbero, J.M., Ravazzoli, C.L., and Hensley, J.L. (1992). Reflection
and transmission coeffcients in fluid-saturated porous media. J. Acoust. Soc. Am.,
91(4), 1911–1923.
Schmitt, D.P. (1988). Shear wave logging in elastic formations. J. Acoust. Soc. Am.,
84, 2215–2229.
Schmitt, D.P., Bouchon, M., and Bonnet, G. (1988). Full-wave synthetic acoustic
logs in radially semiinfinite saturated porous media. Geophysics, 53, 807–823.
Sinha, B.K. and Asvadurov, S. (2004). Dispersion and radial depth of investigation
of borehole modes. Geophysical Prospecting, 52, 271–286.
Sinha, B.K., Plona, T.J., Kostek, S., and Chong, S.K. (1992). Axisymmetric wave
propagation in fluid-loaded cylindrical shells. I: Theory. J. Acoust. Soc. Am., 92,
1132–1143.
BIBLIOGRAPHY 127
Smeulders, D.M.J. (2005). Experimental evidence for the slow compressional wave.
J. Appl. Mech., to be published.
Smeulders, D.M.J. and van Dongen, M.E.H. (1997). Wave propagation in porous
media containing a dilute gas-liquid mixture: theory and experiments. J. Fluid
Mech., 343, 351–373.
Smeulders, D.M.J., de la Rosette, J.P.M., and van Dongen, M.E.H. (1992). Waves in
partially saturated porous media. Transp. Porous Media, 9, 25–37.
Tang, X.M. and Cheng, A. (2004). Quantitative Borehole Acoustic Methods. Elsevier.
Tang, X.M. and Cheng, C.H. (1996). Fast inversion of formation permeability from
Stoneley wave logs using a simplified Biot-Rosenbaum model. Geophysics, 61,
639–645.
Tang, X.M., Cheng, C.H., and Toksöz, M.N. (1991). Dynamic permeability and
borehole Stoneley waves: A simplified Biot-Rosenbaum model. J. Acoust. Soc.
Am., 90, 1632–1646.
Tichelaar, Bart W., Liu, Hsui-Lin, and Johnson, David L. (1999). Modeling of bore-
hole Stoneley waves in the presence of skin effects. J. Acoust. Soc. Am., 105(2),
601–609.
White, J.E. (1975). Computed seismic speeds and attenuation in rocks with partial
gas saturation. Geophysics, 40, 224–232.
White, J.E., Mykhaylova, N.G., and Lyakhovitsky, F.M. (1975). Low-frequency seis-
mic waves in fluid saturated layered rocks. Izvestijan Acad. Sci USSR, Phys Solid
Earth, 11, 654–659.
Winkler, K.W., Liu, H-S, and Johnson, D.L. (1989). Permeability and borehole
Stoneley waves: Comparison between experiment and theory. Geophysics, 54,
66–75.
Wisse, C.J., Smeulders, D.M.J., van Dongen, M.E.H., and Chao, G. (2002). Guided
wave modes in porous cylinders: Experimental results. J. Acoust. Soc. Am., 112(3),
890–895.
Wu, K., Xue, Q., and Adler, L. (1990). Reflection and transmission of elastic waves
from a fluid-saturated porous solid boundary. J. Acoust. Soc. Am., 87(6), 2349–
2358.
van der Grinten, J.G., van Dongen, M.E.H., and van der Kogel, H. (1985). A shock
tube technique for studying pore-pressure propagation in a dry and water-saturated
porous medium. J. Appl. Phys., 58, 2937–2942.
van der Grinten, J.G., van Dongen, M.E.H., and van der Kogel, H. (1987). Strain and
pore pressure propagation in a water-saturated porous medium. J. Appl. Phys., 62,
4682–4687.
ACKNOWLEDGEMENTS
First of all I express my gratitude to my advisors David Smeulders and Prof. Rini
van Dongen. I thank David for giving me autonomy and freedom to define the scope
of the thesis, his useful advice in different aspects of this work and for providing all
the means needed in order to work efficiently in this project. I am grateful to Rini
for his constant encouragement and particularly, for all the passionate discussions
about physics and other subjects, which despite the intensity were always friendly
and constructive.
My promotor at TU Delft, Prof. Jakob Fokkema was very helpful during the last
stage of my PhD. I really appreciate that he always found time to discuss about my
thesis and promotion despite his understandably very busy agenda. I am thankful
to the members of the committee: Prof. Max Peeters, Prof. Kees Wapenaar, Prof.
Gijs Ooms and Prof. Rob Arts for careful reading of the manuscript and for valuable
comments.
The experimental part of this thesis required the professional assistance and tech-
nical support of the technicians at the Fluid Dynamics Laboratory at TU/e and at the
Dietz Laboratory at TU Delft. I thank them for their contribution. In particular, I
would like to mention Ad Holten, Jan Willems and Herman Koolmees at TU/e, and
Karel Heller and Andre Hoving at TU Delft. I also thank Mico Hirschberg for useful
advice regarding the design of the experimental setup. I acknowledge that many peo-
ple from TU Delft have helped me (sometimes anonymously) dealing with technical
and administrative issues.
During my PhD, I had the enriching opportunity to work with students who con-
tributed to this work: Michaël Daenen, Rik Hansen and Robin van Gastel.
This work benefited from a cooperation project with Jonathan Woolley and Prof.
Max Peeters from the Colorado School of Mines, which was a fruitful and successful
experience.
The friendly working atmosphere at the Fluid Dynamics Group at TU/e was pro-
moted by all the staff, PhD students, technicians, secretaries and guests in Cascade.
I have fond memories of numerous coffee and tea breaks, group excursions, dinners
and borrels thanks to Dennis, Petra, Laurens, Marjan, Cyrille, Sveta, Brigitte, Ralph,
Honza, Geert, Annemie, Andrzej, Alejandro, Werner, Thijs, Marleen, Ruben, Vin-
cent, Gert and all (former) colleagues in the group. Diligent assistance in dealing
130 ACKNOWLEDGEMENTS
with all kind of paper work was kindly provided by secretaries Brigitte van Wijdeven
and Marjan Pepers.
I thank my friends: Vinayak, Amancay, David, Mathew, Mar, Pablo, Aloke,
Sylvie, Federico, Vishwa, Joseph, Dilna, Vinit, Merina, Irina and Ana Maria for all
the fun and good moments shared in Eindhoven, Buenos Aires, Barcelona, (the way
to) Paris, Roma, Ardennes and many other places; which included a friendly stay in
Vestide during the early times, very long long-distance communications, memorable
(table-) tennis and chess matches, climbing tries, sharing office and parties.
I sincerely thank my parents, Matilde and Luis for their support and confidence
throughout my career, especially to my mother, for her continuous interest in my
studies from the very beginning. I am also grateful to my brother, Lucas, and my
sister, Celeste, for their help and company.
Gabriel Chao,
Eindhoven, April 2005.
CURRICULUM VITAE