Gabriel Eugenio CHAO, Dispersive Surface Acoustic Waves in Poroelastic Media

Dispersive surface acoustic waves
in poroelastic media
Dispersive surface acoustic waves
in poroelastic media
P ROEFSCHRIFT
ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,
op gezag van de Rector Magnificus, prof.dr.ir J.T. Fokkema,
voorzitter van het College voor Promoties,
in het openbaar te verdedigen op dinsdag 10 mei 2005 om 15:30 uur
door
Gabriel Eugenio CHAO
Licenciado en Ciencias Fisicas, Universidad de Buenos Aires

geboren te Buenos Aires, Argentinië
Dit proefschrift is goedgekeurd door de promotoren:
Prof.dr.ir. J.T.Fokkema
Prof.dr.ir. M.E.H. van Dongen
Samenstelling promotiecommissie:
Rector Magnificus, voorzitter

Prof.dr.ir. J.T. Fokkema, Technische Universiteit Delft, promotor
Prof.dr.ir. M.E.H. van Dongen, Technische Universiteit Eindhoven, promotor
Dr.ir. D.M.J. Smeulders, Technische Universiteit Delft, toegevoegd promotor
Prof.dr.ir. C.P.A. Wapenaar , Technische Universiteit Delft
Prof.dr.ir. G. Ooms, Technische Universiteit Delft
Prof.ir. M. Peeters, Colorado School of Mines, USA
Prof.dr. R.J. Arts , Technische Universiteit Delft
This research was financially supported by the ISES program.

c
Copyright 2005 by G.E. Chao
ISBN 90-9019411-8
Printed by Universiteitsdrukkerij TU Eindhoven, Eindhoven, The Netherlands
Cover design by Paul Verspaget
To my parents
vi
CONTENTS
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
3. Surface Waves along Partially Saturated Porous Media . . . . . . . . 35

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
viii Contents
3.2 Acoustic Properties of a Partially Saturated Porous Medium: The

Oscillating Gas Bubble Model . . . . . . . . . . . . . . . . . . . . 36
3.3 Saturation Effects on the Velocities and Attenuation of the Surface
Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.4 Conclusions and Discussion . . . . . . . . . . . . . . . . . . . . . 50
3.5 Appendix 3A: Gas Pocket Model Revisited . . . . . . . . . . . . . 51
3.6 Appendix 3B: Comparison Between the Two Models . . . . . . . . 55
3.7 Appendix 3C: Matrix elements for the gas pocket model . . . . . . 57
4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experi-

ments . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.2 Theory Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . 63
4.3 Permeability Effects on the Pseudo-Stoneley Wave . . . . . . . . . 66
4.4 Experimental Setup and Results . . . . . . . . . . . . . . . . . . . 72
4.5 Conclusions and Discussion . . . . . . . . . . . . . . . . . . . . . 76
4.6 Appendix 4A: Potential Formulation of Biot’s Theory in Cylindrical
Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.7 Appendix 4B: Matrix coefficients for the confined reservoir . . . . . 79
4.8 Appendix 4C: Matrix coefficients for the radially infinite porous for-
mation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5. Surface and Guided Waves along a Cylindrical Interface between a Liquid

and a Liquid-Saturated Porous Medium . . . . . . . . . . . . . . 83
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.2 Borehole modes in Poroelastic Formations . . . . . . . . . . . . . . 84
5.2.1 Pseudo-Stoneley and Pseudo-Rayleigh Waves . . . . . . . . 84
5.2.2 Radius Effects . . . . . . . . . . . . . . . . . . . . . . . . 87
5.2.3 Dispersion Relation in the Bounded Reservoir . . . . . . . . 89
5.3 Modification of the Shock Tube . . . . . . . . . . . . . . . . . . . 92
5.4 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
6. Seismic Signatures of Partial Saturation on Acoustic Borehole Modes . . 103

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.2 Borehole Waves in Partially Saturated Porous Media . . . . . . . . 104
6.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
7. Conclusions and Recommendations . . . . . . . . . . . . . . . 113
Samenvatting . . . . . . . . . . . . . . . . . . . . . . . . . 117
Contents ix
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . 120
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 129
Curriculum Vitae . . . . . . . . . . . . . . . . . . . . . . . . 131

x Contents
SUMMARY
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
1.1 Introductory Remarks
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)).
of the reservoir. Theoretical and experimental investigations in boreholes surrounded

by poroelastic formations provide promising evidence about the dependence of the
permeability and porosity of the reservoir on the frequency-dependent velocities and
attenuations of the surface modes. However, there is a lack of a full understanding
of the underlying physical mechanisms which govern the propagation of the surface
waves along poroelastic formations. Moreover, further experimental studies under
well-defined laboratory conditions are required to validate the theoretical models.
The particle motion driven by a surface wave follows an elliptical trajectory. This
rather complicated elliptical polarization can be expressed as a combination of the
compressional and shear bulk modes in the medium. This is a fundamental differ-
ence between the surface waves and the bulk or body waves. While the bulk waves
propagate independently in an unbounded medium, the presence of the boundaries
induces a coupled motion of the compressional and shear waves which characterizes
the surface waves. Therefore an accurate description of the bulk waves dynamics in
a porous material is an essential step in order to gain a better understanding of the
surface waves propagating along a poroelastic layer.
1.2. Literature Review 3
1.2 Literature Review
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.
1.2.1 Wave Propagation in Porous Media

From a purely historical point of view, the first investigations on acoustic wave propa-
gation in porous media can be traced back to the works of Zwikker and Kosten (1941)
and Frenkel (1944). Despite their significance and correct qualitative description of
the wave phenomena, these models neglected significant compressibility (Zwikker
and Kosten) and inertial (Frenkel) effects. Maurice A. Biot developed a theory which
includes the viscous dissipation due to the relative fluid/solid motion, and describes
the mechanical interaction between the fluid and solid phases (Biot, 1956a,b; Biot
and Willis, 1961). An important aspect of this theory is the accurate description
of a second compressional wave which propagates through a fully saturated porous
medium. This so-called slow compressional wave has since then been detected exper-
imentally and shown to correctly predict wave speeds (Plona, 1980; van der Grinten
et al., 1985; Kelder and Smeulders, 1997). However, Biot’s theory fails to account
for the level of attenuation of the fast compressional wave observed in laboratory and
field experiments. This suggests the existence of loss mechanisms which are not con-
sidered in the original theory. A critical comparison of the attenuation predicted by
the Biot theory with existing experimental data for natural rocks was given by Gist
(1994a).
1.2.2 Attenuation Mechanisms

The unability of the Biot theory to properly describe the attenuation values observed
in liquid-saturated rocks and sediments encouraged a series of investigations on the
influence of the wave induced fluid-flow loss mechanisms. At present it is assumed
that the presence of cracks, gas inclusions and lithological inhomogeneities are re-
sponsible for the most relevant dissipative mechanisms ignored in the Biot theory.
The influence of cracks at the grain scale on the wave propagation through a
porous material has been discussed by Mavko and Nur (1979), Dvorkin and Nur
(1993) and Dvorkin et al. (1995). At this microscopic scale the dilatation of the
rock due to the passage of an elastic wave causes the fluid to flow from the crack
to the pore space. This dissipative mechanism is denominated “squirt flow” and
4 1. Introduction
can be described in terms of a characteristic squirt-flow length. It has been shown

that the incorporation of this local dissipative mechanism significantly increases the
attenuation values, but still is unable to fully explain the observed attenuation in
seismic surveys.
Gas inclusions in the pore space of a liquid-saturated porous rock are also re-
sponsible for losses since wave-induced oscillations of the trapped gas bubbles cause
local dissipation. In this case, two immiscible fluid phases coexist in the void space.
White (1975) assumed that the gas phase distribution can be represented by spherical
gas pockets regularly arranged in a cubic lattice and formulated the so-called “gas
pocket” model. Numerical studies based on Biot’s theory for this system have been
published by Dutta and Ode (1979a,b). Their results show an important increase in
the attenuation though they are restricted to frequencies where the wavelengths are
considerably larger than the size of the inhomogeneities. A similar model has been
proposed by Smeulders and van Dongen (1997). The main difference with respect to
the work of Dutta and coauthors lies in the dissipative mechanisms induced by the
bubble oscillations. In this model the thermal interaction between the bubble and the
surrounding porous medium is also included. Their theoretical results were compared
with experimental data measured in shock wave experiments.
One of the basic assumptions of the Biot model is related to the isotropy of the
material that constitutes the rock. Heterogenous composites have wave-induced flow
losses due to compressibility changes which are not considered in the original Biot
formulation. Recently, Pride and Berryman (2003a), (2003b) have developed from
first principles a set of governing equations for double-porosity, dual-permeability
materials. These heterogeneous materials are constituted by two distinct porous ma-
terials with different lithological and poromechanical properties.
1.2.3 Surface Waves in Porous Media
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).
1.2.4 Acoustic Borehole Modes in Porous Reservoirs

The propagation of surface and head waves along cylindrical interfaces between a
liquid and a liquid-saturated porous reservoir has drawn attention due to the practi-
cal implications in the oil industry related to reservoir characterization. Rosenbaum
(1974) applied Biot’s theory to solve the high-frequency modes of a fluid-filled bore-
hole surrounded by a poroelastic formation in order to correlate the permeability
of the formation with the properties of the second compressional wave. The low-
frequency limit, where the (pseudo) Stoneley wave is referred as the tube wave, was
derived by Chang et al. (1988). Norris (1989) included the compressibility of the
solid matrix. Schmitt et al. (1988) studied the wave response to a point source in
both the time and frequency domain using Biot’s theory modified according to a ho-
mogenization theory. They presented results for the (pseudo) Stoneley mode and the
first pseudo-Rayleigh mode. Cheng et al. (1987) applied Biot’s theory to interpret the
permeability dependence of the properties of the (pseudo) Stoneley wave observed
in field data. Liu and Johnson (1997) simulated the effect of a mudcake layer that
6 1. Introduction
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).
1.2.5 Experimental Studies on Bulk and Surface Waves in Poroelastic Media
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
1.3 Objectives of this Thesis
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.
1.4 Thesis Outline
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
following a Sommerfeld formulation. At the end of this chapter approximate models

for surface wave propagation along boreholes in porous formations are described and
developed. Chapter 3 is focused on the influence of the gas saturation on the speed
and attenuation of the surface modes propagating along a liquid/poroelastic plane
interface.
The shock-induced experiments performed in a Berea sandstone are reported and
compared with numerical calculations based on Biot’s theory in Chapter 4. An anal-
ysis of the permeability dependence of the properties of the pseudo-Stoneley wave is
also given in Chapter 4, followed by a quantitative explanation based on the relative
contribution of the bulk waves to the radial displacement induced by this mode. A
series of experiments in a modified shock tube configuration, aimed to improve the
enhancement of the surface waves, is analyzed in Chapter 5. Also in this chapter
acoustic borehole modes in poroelastic formations are studied in detail. The purpose
of Chapter 6 is to investigate the saturation effects on the frequency-dependent phase
velocity and attenuation of the pseudo-Stoneley wave that propagates along a liq-
uid/poroelastic cylindrical interface. The main results of this thesis are summarized
in Chapter 7.
2. WAVE PROPAGATION IN FULLY SATURATED POROUS MEDIA:
BULK MODES AND INTERFACE WAVES
2.1 Dynamics of a Continuum Porous Medium: The Biot theory
2.1.1 Stress-Strain Relations

The formulation of a compact and consistent theory for the mechanical interaction
between the fluid and solid phase in a porous material can be attributed to M.A. Biot,
who developed a theory which also includes the viscous dissipation due to the relative
fluid/solid motion (Biot, 1956a,b; Biot and Willis, 1961). In this theory the stress in
the solid and the pressure in the fluid are coupled and the stress-strain relations are
written as:
τij = (P − 2N ) ukk δij + 2N uij + QUkk δij (2.1)
and
−1
p= (Qukk + RUkk ) , (2.2)
φ
where summation over repeated indexes is assumed, τ ij is the stress in the solid, p
the pore pressure, u denotes the displacement of the solid matrix, U denotes the fluid
displacement, uij = 21 (∂ui /∂xj + ∂uj /∂xi ), and δij is the Kronecker delta. The
porosity of the material φ is the ratio between the volume of the void space and the
total volume in the representative element of volume. The representative element
of volume REV, is a volume for which the average of all physical quantities over
that volume is independent of the volume size. Therefore the REV is a volume cell
containing all the information regarding the geometry and the physical properties of
the material. N is the shear modulus of the composite material and P , R and Q
are the so-called generalized elastic coefficients. They are related to the porosity,
the solid frame bulk modulus Kb , 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, (2.3)
Kb Ks 3
1−φ− +φ
Ks Kf
10 2. Wave Propagation in Fully Saturated Porous Media

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)).
2.1.2 The Lagrangian Approach to the Biot Equations

In this subsection the Biot equations are derived following a Lagrangian formulation
(see e.g. Johnson (1986)). For a rigorous treatment of the Lagrangian formulation
for continuous systems the reader is referred to Goldstein (1980). Consider the two-
phase system consisting of the solid matrix and the fluid saturating the pore space.
The kinetic energy T can be written as follows:
1
T = ρ11 |u̇|2 + 2ρ12 u̇ · U̇ + ρ22 |U̇|2 , (2.6)
2
where ρ11 , ρ12 and ρ22 are the density terms that are related to the density of the fluid
phase ρf and the solid phase ρs by:
ρ11 = (1 − φ) ρs + (α∞ − 1) φρf , (2.7)
ρ12 = (1 − α∞ ) φρf , (2.8)

ρ22 = α∞ φρf . (2.9)
The parameter α∞ is the tortuosity. It takes into account the microgeometry
of the channels in the pore space. Its value is always larger than 1. The potential
energy V associated with the deformation of the porous material can be expressed as
(Johnson, 1986):
1h i
V = P (∇ · u)2 + 2Q (∇ · u) (∇ · U) + R (∇ · U)2 + N |∇ × u|2 . (2.10)
2
Equations 2.6 and 2.10 define the volumetric Lagrangian density L = T − V for this
continuum system. In order to describe the viscous dissipation mechanism due to the
relative fluid-solid motion, the dissipation pseudo-potential D is introduced:
1 ηφ2 2
D= u̇ − U̇ , (2.11)
2 κ0
2.1. Dynamics of a Continuum Porous Medium: The Biot theory 11
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
Due to the intrinsic frequency-dependent nature of the viscous interaction, it is useful

to consider the Biot equations in the frequency domain. Assuming a temporal depen-
dence e−iωt it is possible to transform the Biot equations to the frequency domain by
the Fourier Transform to obtain:

−ω 2 ρe11 eu + ρe12 U
e = P ∇ (∇ · e u) + Q∇(∇ · U)e − N∇ × ∇ × e u, (2.15)

u + ρe22 U
−ω 2 ρe12 e e = R∇ ∇ · U u) ,
e + Q∇ (∇ · e (2.16)
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)
Here we have introduced a frequency-dependent tortuosity:
iηφF (ω)
α
e(ω) = α∞ + , (2.20)
k0 ωρf
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
2.1.3 Displacement Potential Formulation. Body Waves

In this subsection the main properties of the body waves’s solutions of the Biot equa-
tions will be discussed. The most remarkable result is the prediction of a highly
dispersive compressional wave, usually called slow P wave. In agreement with the
classical results of linear elasticity, this two-phase theory predicts the presence of P
and S waves. However, the existence of viscous interaction due to the relative motion
of the fluid with respect to the solid matrix causes the classical compressional and
shear waves to become dispersive and attenuated. It is first shown how the Biot equa-
tions can be solved in the frequency domain by means of a displacement potential
formulation. Then the expressions for the velocity of the different bulk modes are de-
rived. In order to simplify the notation, in the remaining of the chapter the tilde above
the functions and quantities in the frequency domain is omitted. The tilde above the
density terms and the tortuosity is used to denote the frequency-dependent nature of
these parameters.
For compressional P waves the displacements of the solid and fluid phase can be
described by the following scalar potentials:
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)
where ρ is the density matrix given by:

ρe11 ρe12
ρ= , (2.26)
ρe12 ρe22
Φ is a vector defined as [Φs Φf ]T and M is a matrix containing the generalized elastic

coefficients:

P Q
M= . (2.27)
Q R
Assuming that the solutions of equations 2.25 can be expressed as

Φc1 = Ac1 ekc1 ·r and Φ2 = Ac2 ek2 ·r , the linear system is reduced to an eigenvalue
problem. The eigenvalues are the squared complex-valued wavenumbers of the com-
pressional waves that propagate in the porous material. The wavenumbers k c1 and
kc2 are given by:
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 velocities of the compressional waves follow from:

ω
cj (ω) = j = 1, 2. (2.31)
Re[kcj (ω)]
The attenuation of the compressional modes can be expressed in terms of the

damping coefficient, Dj (ω) = |Im[kcj (ω)]| or the inverse quality factor,

2Im[kcj (ω)]
−1
Qj (ω) = j = 1, 2. (2.32)
Re[kcj (ω)]
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:
Φs = Φc1 + Φc2 (2.33)

and
Φf = Gc1 Φc1 + Gc2 Φc2 , (2.34)
where Gc1 and Gc2 are given by:
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)
U = Gsh ∇×Ψs , (2.38)

where
α
e−1
Gsh = . (2.39)
α
e
The velocity of the shear wave becomes:
1
N 2
csh = . (2.40)
ρe11 − ρe212 /e
ρ22
2.1.4 The Slow Compressional Wave

The accurate prediction of the frequency-dependent velocity and attenuation of the
slow compressional wave is the most important result of the Biot theory. This wave
was detected experimentally by Plona (1980) at ultrasonic frequencies (500 kHz-
2.25 MHz). The experiments were performed in water-saturated porous materials
made out of glass beads. Some years later, van der Grinten et al. observed the slow
compressional wave in shock wave experiments on porous columns consisting of
agglutinated sand particles (van der Grinten et al., 1985, 1987). Nagy et al. (1990)
observed the slow wave in air-saturated natural rocks. Kelder and Smeulders (1997)
provided experimental evidence for the slow compressional wave in water-saturated
rocks. They observed the effects of the slow wave in transmission experiments in a
Nivelsteiner sandstone. A detailed review of the different attempts to measure the
slow wave using different experimental techniques is given by Smeulders (2005).
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.
Berea sandstone Synthetic glass

Solid density ρs (kg/m3 ) 2644 2539
Porosity φ 0.20 0.465
Permeability k0 (D) 0.36 10
Tortuosity α∞ 2.4 1.93
Frame bulk modulus Kb (GPa) 10.37 2.53
Shear modulus N (GPa) 7.02 2.9
Grain bulk modulus Ks (GPa) 36.5 36.5
Fluid density ρf (kg/m3 ) 1000 1000
Fluid viscosity η (mPa s) 1.0 1.0
Fluid bulk modulus Kf (GPa) 2.25 2.25
Table 2.1: Physical properties of the porous formations.
The slow wave is highly dispersive. It is diffusive in the low-frequency limit

where the viscous dissipation dominates and propagative at high frequencies where
the inertial effects prevail. The transition between these two different regimes is
marked by the critical frequency:
ηφ
fc = . (2.41)
2πκ0 α∞ ρf
The slow compressional wave is characterized by an out-of-phase displacement of
the solid matrix with respect to the pore fluid. Therefore, a large influence of the per-
meability and the viscosity on the properties of this wave is to be expected. This is
analyzed in Figure 2.1 where the permeability and frequency effects on the velocity
and attenuation of the slow compressional wave are depicted for a Berea sandstone
saturated with water. The material properties of the rock and the saturating fluid
are given in Table 2.1. The velocity increases with increasing permeability and fre-
quency. In the zero-frequency limit the velocity tends towards zero. The specific
attenuation increases with decreasing permeability due to increasing friction. The
diffusive character of the slow wave at low frequencies results in a flattening of the
attenuation coefficient. At higher permeability values, the relative flow of the fluid
with respect to the solid matrix facilitates and therefore the attenuation decreases.
2.2 Poroelastic Surface Waves along Plane Interfaces
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.
2.2.1 Dispersive Surface Modes

In this subsection a displacement potential formulation is developed in order to de-
scribe the surface waves that propagate along a plane interface between a fluid half-
space and a liquid-saturated porous half-space. The configuration studied is displayed
in Figure 2.2. The surface modes propagate parallel to the interface, depend expo-
nentially on the distance z from the interface and can be expressed in terms of the
bulk mode solutions. In the liquid (z < 0), the compressional waves are described
by the following potential:
Φf = Af eγf z ei(kx x−ωt) . (2.42)
The potentials associated to each of the bulk modes which propagate in the porous
half-space are:
Φc1 = Ac1 e−γc1 z ei(kx x−ωt) , (2.43)
Φc2 = Ac2 e−γc2 z ei(kx x−ωt) , (2.44)

and
Ψsh = Be−γsh z ei(kx x−ωt) êy , (2.45)
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
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
where cj is the velocity of the corresponding bulk mode.

The surface modes can be written as a frequency-dependent linear combination
of the potentials stated above. The different contributions of the bulk modes are
determined by the boundary conditions, namely: continuity of averaged normal dis-
placement, total stress and pressure. The displacements of the solid phase and the
fluid phase in the porous medium can be expressed as follows:
u = ∇ (Φc1 + Φc2 ) + ∇ × Ψsh , (2.47)
and
U = Gc1 ∇Φc1 + Gc2 ∇Φc2 + Gsh ∇ × Ψsh , (2.48)
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)
can be expressed as:

∂Φc1 ∂Φc2
(1 − φ + φGc1 ) + (1 − φ + φGc2 ) +
∂z ∂z
∂Ψsh ∂Φf
+ (1 − φ + φGsh ) = . (2.50)
∂x ∂z
The continuity of the normal component of the total stress implies:
τzz − φp = −pf . (2.51)
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:
det [N (kx , ω)] = 0. (2.56)
At a fixed frequency ω, equation 2.56 is numerically solved for complex k x using

a Newton-Raphson algorithm. In this way, frequency-dependent phase velocities,
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
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.
2.2.2 Synthetic Seismograms: The Influence of the Leaky Modes
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
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.
(1980). The incident pressure wave field Pin ,

A(ω) i(k0 R−ωt)
Pin = e , (2.57)
R
can then be expressed as (see e.g. Brekhovskikh (1980), Chapter 4):
Z π −i∞
2
−iωt
Pin = A(ω)e ik0 J0 (k0 r sin θ)eik0 cos θ|z−z0 | sin θ dθ, (2.58)
0
where the integration is performed over the complex angle θ. The path of integration
is chosen in the form of the contour Γ0 , shown in Figure 2.5. In equation 2.58, A(ω)
denotes the amplitude of the incident wave, R is the distance between the source and
the observation point O, k0 is the wavenumber of the acoustic waves in the liquid,
z0 is the vertical position of the source with respect to the interface, r and z are the
horizontal and vertical coordinates of the observation point respectively, and J 0 is the
Bessel function of zero order and first kind.
Analogously, the reflected pressure wave field Pr can be expressed in terms of
cylindrical wave fronts as follows:
Z π −i∞
2
−iωt
Pr = A(ω)e ik0 R(θ)J0 (k0 r sin θ)eik0 cos θ|z+z0 | sin θ dθ. (2.59)
0
In the above equation R(θ) is the reflection coefficient for a plane wave which im-
pinges upon a plane interface between a liquid (in this case water) and a fully satu-
rated porous material. Reflection and transmission coefficients for this configuration
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
Pressure (a.u.) 1
−1
−2
−3
−4
PSfrag replacements −5
0 5 10 15 20
Time (µs)
Figure 2.6: Incident Blackman-Harris pressure pulse.
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.)
−0.5 Fast P P-Stoneley

Fluid
−1
PSfrag replacements −1.5
−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.
2.3 Surface Modes in Borehole Configurations: Approximate Models
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.
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.
2.3.1 Rigid Borehole Wall

In this subsection a review is presented for the acoustic guided waves that can prop-
agate along a fluid-filled cylindrical cavity in a rigid solid. This configuration has
been studied in the past and has an analytical solution for these modes White (1983).
Therefore, it constitutes a test for the accuracy of our numerical search routine.
We consider a water-filled hole with a radius R of 38.5 mm. The wave equation
in cylindrical coordinates for guided acoustic waves travelling in the fluid can be
solved. It follows that the potential function is given by:
ϕ = Acw J0 (kr r)ei(ωt−kz) (2.62)
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
8000
7000 n=1 n=2 n=3 n=4

Phase Velocity (m/s)
6000
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
where xn is the nth zero of the J1 (x) Bessel function.

A graphical representation of these modes is given in Figure 2.9, where both the
analytical (solid lines) and numerical results (square dots) are given. The numerical
calculations are performed with the zero-search Newton-Raphson algorithm in the
complex plane that was previously used to study the surface modes in the flat in-
terface. There is one wave which is non-dispersive and propagates at the speed of
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.
2.3.2 Quasi-1D Model for a Permeable Borehole Wall
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
applied to the control volume reads:

Z
∂ρ ∂vz 1 ∂vr
+ ρ0 + dV = 0. (2.67)
CV ∂t ∂z r ∂r
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:
vr (R) = φvr,I (R). (2.70)
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
and the constitutive equation:
1 ∂ρI 1 ∂pI
= , (2.72)
ρI ∂t Kf ∂t
where Kf is the bulk modulus of the fluid, to obtain:
1 ∂pI
+ ∇ · vI = 0. (2.73)
Kf ∂t
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
In order to solve equation 2.75 a solution of the type pI (r, z, t) = A(r)ei(ωt−kz z) is

assumed. Substitution of the solution in equation 2.75 leads to:

d2 A(r) 1 dA(r)
iωA(r) = D + , (2.77)
dr2 r dr
which can be written as:
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:
A(r) = bK0 (λr), (2.79)
where K0 is the zeroth-order Kelvin function. It is assumed that there is no flow

restriction at the boundary between the pore fluid and the bore fluid which implies
the continuity of pressure. Therefore the solution at the borehole wall, r = R, should
match the solution for the pressure induced by the surface mode inside the borehole
p = p0 ei(ωt−kz z) . With these considerations it follows that the pore pressure in the
reservoir can be written as:
K0 (λr) i(ωt−kz z)
pI (r, z, t) = p0 e . (2.80)
K0 (λR)
From Darcy’s law (equation 2.74) we finally obtain for the radial velocity in the
porous medium:
κ(ω) ∂pI (r, z, t)

vr,I = − . (2.81)
ηφ ∂r
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
where C(ω) is given by:

iω 2ρ0 κ(ω) K1 (λR)
C(ω) = iρ0 ω + λ . (2.85)
c20 Rη K0 (λR)
The quadratic equation for kz can be solved directly:

 s 
2
1 ig g 4C(ω) 
kz = −  2 − − 2 − . (2.86)
2 c0 c0 ρ0
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
1500
1250
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).
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.
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
n32 = −[(P − 2N ) + Q + Gc1 (Q + R)]( cω1 )2 + 2N γc1

2 ,
n33 = −[(P − 2N ) + Q + Gc2 (Q + R)]( cω2 )2 + 2N γc2

2 ,
n34 = −2N ikx γsh ,

n41 = ω 2 ρf ,
n42 = −( cω1 )2 1 (Q + RGc1 ),
φ
n43 ω
= −( c2 ) 1 (Q + RGc2 ),
2
φ
n44 = 0.
3. SURFACE WAVES ALONG PARTIALLY SATURATED POROUS
MEDIA
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
phenomena in a porous layer adjacent to a water table.

Despite all the efforts and attention to study saturation effects on seismic and
acoustic waves, there is, to our best knowledge, no study concerning the influence
of the liquid saturation on surface waves. The purpose of this work is to investigate
the effects of the gas fraction on the propagation of surface waves along a plane in-
terface between a liquid and a partially saturated porous medium. The bulk acoustic
properties of the partially saturated porous medium are described according to the
model of Smeulders and van Dongen (1997). The high-frequency properties of the
surface waves for the fully saturated case were studied in detail by Feng and John-
son (1983a,b). There are three surface modes that can propagate depending on the
relation between the mechanical properties of the porous material and the liquid and
the characteristics of the interface regarding the possibility for the liquid to flow be-
tween the two half-spaces (surface permeability). The three modes are the Stoneley
wave, the pseudo-Stoneley wave and the pseudo-Rayleigh wave. The Stoneley wave
is a true surface wave which propagates almost undamped along the interface with
an exponential decay in the normal direction away from the interface. The pseudo
modes are significantly damped in the direction of propagation and radiate energy
into the slow compressional wave in the porous medium (pseudo-Stoneley wave) and
both into the slow compressional wave and the acoustic wave in the liquid half-space
(pseudo-Rayleigh wave). We will study the high-frequency limit for the partially
saturated case and also the frequency-dependent dispersion of the pseudo modes.
The chapter is organized as follows. In section 3.2 we review the theoretical
model for acoustic wave propagation for the case that a liquid-gas mixture saturates
the porous material. In section 3.3 the results for the velocity and attenuation of the
surface modes propagating along a liquid-poroelastic plane interface are presented
and discussed. First the high-frequency limit is examined and the existence of the
different waves is discussed, followed by the analysis of the frequency-dependent
results. The chapter is summarized and the conclusions are given in section 3.4. We
review the gas pocket model in detail in Appendix 3A. The influence of the different
interaction and dissipative mechanisms on the velocity and specific attenuation of the
bulk waves is analyzed in Appendix 3B.
3.2 Acoustic Properties of a Partially Saturated Porous Medium: The

Oscillating Gas Bubble Model
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
or heterogeneous mixtures, scattering effects cannot be neglected and it is no longer

possible to define a spatially homogeneous external driving pressure.
Let us consider a spherical air bubble immersed in a fully water saturated porous
medium in the presence of an external oscillating pressure field. The dynamics of the
bubble is determined by the solution of the Biot equations at the spherical interface
between the air-saturated and water-saturated porous media. Mathematically it is
possible to solve the Biot equations in spherical coordinates in the two domains,
inside and outside the bubble. We will first focus on the external domain (fully water-
saturated porous medium outside the bubble). In the remaining of the chapter and in
order to simplify the notation, the tilde above the functions and quantities in the
frequency domain is omitted. The tilde above the density terms and the tortuosity is
used to denote the frequency-dependent nature of these functions.
We introduce the displacement potentials Φc1 and Φc2 associated with the fast
and slow compressional waves as follows:
u = ∇Φc1 + ∇Φc2 , (3.4)
and
U = Gc1 ∇Φc1 + Gc2 ∇Φc2 , (3.5)
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)
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:
∆Vg = 4πa2 [(1 − φ)ur + φUr ]. (3.13)
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
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
cj = (1 + ikj a)(1 − φ + φGcj ). (3.18)
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
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.
Solid density ρs (kg/m3 ) 2644

Porosity φ 0.20
Permeability k0 (mD) 360
Tortuosity α∞ 2.4
Frame bulk modulus Kb (GPa) 10.37
Shear modulus N (GPa) 7.02
Grain bulk modulus Ks (GPa) 36.5
Liquid bulk modulus Kl (GPa) 2.25
Gas pressure (bulk modulus) pg (GPa) 0.01
Liquid density ρl (Kg/m3 ) 1000
Gas density ρg (Kg/m3 ) 100
Liquid viscosity ηl (mPa s) 1
Gas viscosity ηg (mPa s) 1.5 10−2
Gas thermal diffusivity ag 1.87 10−7
Table 3.1: Physical properties of the Berea sandstone and the saturating fluids: water and air.
highly incompressible (Kg −→ ∞) and Kf = Kl /s. In this limit Kf increases with

concentration of air in the water. The transition between the two limits shows a min-
imum in the compressibility of the mixture, which corresponds to the anti-resonance
frequency of the bubble. At this frequency the bubbles oscillates out-of-phase with
the external pressure field, which results in a highly incompressible medium. The
relevant parameters in this model are the pressure of the gas, its saturation in the pore
space and the radius of the gas bubbles. The outcome of the velocities and attenuation
of the compressional waves that propagate in this partially saturated porous media are
shown in Figure 3.2 for a liquid saturation s of 0.95 and a bubble radius of 1 mm. The
results for the wave velocities can be explained by the arguments about the changes
in the compressibility of the mixture discussed above. The presence of air lowers the
bulk modulus of the mixture at low frequencies which results in compressional waves
propagating slower in the partially saturated case. This behavior is reversed at high
frequencies where the compressional waves propagate faster when the air saturation
is increased. More interesting are the modifications induced by the air phase in the
attenuation coefficients. A significant increase of the attenuation is found for the fast
compressional wave. The slow compressional wave presents a maximum in the at-
tenuation for the partially saturated case. This maximum is not observed for the fully
saturated case. The model presented here assumes that the shear wave is influenced
by the presence of the gas phase only due to changes in density.
In a way this model has some similarities with the previous work of White (1975)
3500
fast P
Phase Velocity (m/s) 3000
2500
2000
shear
1500
1000
slow P
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.
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
Figure 3.3: Liquid/partially saturated porous medium plane interface.
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.
3.3 Saturation Effects on the Velocities and Attenuation of the Surface

Waves
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
pore boundary case.

We first examine the high-frequency limit, for which the velocities of the bulk
modes become real-valued and the slow wave is propagative. It also holds that the
bulk modulus of the mixture saturating the pore space becomes real-valued (K l /s)
and therefore only compressibility effects influence the speed of the bulk modes.
The dissipative mechanisms discussed in section 3.2 are not present in this limit as
can be observed in Figure 3.1. The dependence of the surface wave velocities and
attenuation on the water saturation is shown in Figure 3.4. For reference, the bulk
wave velocities are also displayed.
For the fully water saturated case s = 1, two surface modes are found: the
pseudo-Stoneley wave and the pseudo-Rayleigh wave. The pseudo-Stoneley wave
has a velocity which is faster than the velocity of the slow wave and slower than the
speed of the rest of the bulk modes. This implies that it radiates energy into the slow
wave and therefore it is called a pseudo or leaky mode. The pseudo-Rayleigh wave
leaks energy into the fluid half-space and into the slow wave, its velocity is faster
than that of the slow wave and the fluid wave but slower than that of the shear and the
fast wave. The velocity of the slow wave decreases with increasing water saturation
as can be observed in Figure 3.4, while the shear mode speed is slightly affected due
to density effects only. The behavior of the slow wave as a function of saturation and
its relation with the other bulk modes plays an important role in the properties of the
surface waves. For water saturations higher than 0.46 the velocity of the slow wave is
lower than that of the shear wave and the fluid wave. In this range of saturations both
the pseudo-Stoneley and the pseudo-Rayleigh waves propagate. The velocity of the
pseudo-Stoneley wave decreases with increasing water saturation. Furthermore, for
s values below 0.46 the pseudo-Stoneley wave becomes a true Stoneley wave due to
the fact that the slow wave becomes faster than the speed of sound in the water. This
transition is neatly illustrated in the attenuation coefficient (Figure 3.4b). The pseudo-
Rayleigh wave ceases to radiate into the slow wave for water saturations lower than
0.43. In this range of saturations the velocity of the pseudo-Rayleigh wave decreases
sharply with increasing water saturation. Figure 3.4b shows the attenuation in terms
of the inverse quality factor Q−1 . The attenuation of the pseudo-Stoneley wave has
a sharp minimum for a saturation of 0.64 and over the entire range of saturations it
is significantly less damped than the pseudo-Rayleigh mode. The attenuation charac-
teristics of the pseudo-Rayleigh wave as a function of saturation are quite remarkable
as well. A sharp increase in Q−1 is observed for water saturation below 0.43 when
the pseudo-Rayleigh wave stops radiating into the porous medium. It is worthwhile
to note that very low values of liquid saturation s may be unrealistic from a physical
point of view, since the model is based on the assumption that there is no interaction
among the bubbles. In any case, the study of the high-frequency limit provides a first
insight on the compressibility effects on the surface modes due to the presence of the
2000
1900
1800 shear
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.
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.
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).
1880
s = 0.8
1860
s = 0.9
1840
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).
3.4 Conclusions and Discussion
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
3.5 Appendix 3A: Gas Pocket Model Revisited
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.
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)
In the spherical shell the expressions for the potentials are:

(2) (2) (2)
Φc1 (r, ω) = Bc1 (ω)j0 (kc1 r) + Cc1 (ω)n0 (kc1 r), (3.24)
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
is coherent with the hypothesis of non-interacting shells. In terms of the potentials

these boundary conditions are expressed as follows: at the boundary between the gas
pocket and the liquid saturated porous solid (r = a) the continuity of the solid and
fluid displacement leads to,
(1) (1) (2) (2)
∂Φc1 ∂Φc2 ∂Φc1 ∂Φc2
[ + = + ]r=a (3.26)
∂r ∂r ∂r ∂r
and
(1) (1) (2) (2)
(1) ∂Φc1 (1) ∂Φc2 (2) ∂Φc1 (2) ∂Φc2
[Gc1 + Gc2 = Gc1 + Gc2 ]r=a , (3.27)
∂r ∂r ∂r ∂r
(j)
where the Gci parameters are defined according to equations 3.6 and 3.7 in each
domain. The continuity of the stress and pressure results in:
(1) (1) (1) (1) (1) (1)
2 ∂Φc1 ∂ 2 Φc1 2 ∂Φc2 ∂ 2 Φc2 ∂ 2 Φc1 ∂ 2 Φc2
[β11 ( + 2
) + β21 ( + 2
) + 2N ( 2
+ )=
r ∂r ∂r r ∂r ∂r ∂r ∂r2
(2) (2) (2) (2)
2 ∂Φc1 ∂ 2 Φc1 2 2 ∂Φc2 ∂ 2 Φc2
β12 ( + ) + β 2 ( + )+
r ∂r ∂r2 r ∂r ∂r2
(2) (2)
∂ 2 Φc1 ∂ 2 Φc2
+2N ( + )]r=a (3.28)
∂r2 ∂r2
and
(1) (1) (1) (1) (2) (2)
2 ∂Φc1 ∂ 2 Φc1 1 2 ∂Φc2 ∂ 2 Φc2 2 2 ∂Φc1 ∂ 2 Φc1
[ξ11 ( + ) + ξ 2 ( + ) = ξ 1 ( + )+
r ∂r ∂r2 r ∂r ∂r2 r ∂r ∂r2
(2) (2)
2 ∂Φc2 ∂ 2 Φc2
+ξ22 ( + )]r=a (3.29)
r ∂r ∂r2
where we assume that the porosity is the same in both domains. At r = b, the radial
stress equals the external stress and therefore:
(2) (2) (2) (2)
2 ∂Φc1 ∂ 2 Φc1 2 ∂Φc2 ∂ 2 Φc2
−τ0 = [ β12 ( + 2
) + β22 ( + )+
r ∂r ∂r r ∂r ∂r2
(2) (2)
∂ 2 Φc1 ∂ 2 Φc2
+ 2N ( + )]r=b (3.30)
∂r2 ∂r2
In equations (3.28-3.30) the coefficients βij and ξij read:
(j)
βij = P (j) − 2N + Q(j) Gci , (3.31)
and
(j)
ξij = Q(j) + R(j) Gci . (3.32)
The jacketed condition, u2 = U2 implies:

(2) (2) (2) (2)
∂Φc1 ∂Φc2 (2) ∂Φc1 (2) ∂Φc2
[ + = Gc1 + Gc2 ]r=b . (3.33)
∂r ∂r ∂r ∂r
Solution of the system of equations 3.26-3.33 for the potential amplitudes allows
a complete solution for the problem. The linear system of equations can be written
as:
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
where ρef f is the effective density of the system given by:
ρef f = (1 − φ)ρs + φ(sρl + (1 − s)ρg ). (3.37)
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
3.6 Appendix 3B: Comparison Between the Two Models
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-
3300
3200

3100
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 ,
(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).
4. SHOCK-INDUCED BOREHOLE WAVES IN POROUS FORMATIONS:
THEORY AND EXPERIMENTS 1
ABSTRACT
The characteristics of the pseudo-Stoneley wave along boreholes in porous forma-

tions are studied in a broad band of frequencies (100 Hz-200 kHz). Experiments
are performed using a shock tube technique to excite the pseudo-Stoneley wave in
a water saturated confined reservoir. The formation is a natural Berea sandstone.
Frequency-dependent phase velocities and damping coefficients are measured using
this technique. Quantitative agreement between the experimental results and the the-
oretical predictions is found for the phase velocity in the frequency range from 10
to 50 kHz. Theoretically, the influence of the permeability on the phase velocity, at-
tenuation, radial displacement, and pore pressure is studied on the basis of the Biot
theory and the contribution of the different bulk modes to the average radial displace-
ment is analyzed in the frequency domain. The numerical results indicate that the
permeability dependence at low frequencies is caused by the Biot slow wave.
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
low frequencies and a propagative behavior at high frequencies.

The influence of the slow Biot wave in the phase velocity and attenuation of the
surface waves that propagate along the interface between a fluid and a fully saturated
porous medium has been subject of study in the past. Feng and Johnson (1983a,b)
applied Biot’s theory to the study of surface acoustic waves in flat interfaces. They
considered the high-frequency limit of the Biot theory for which the bulk waves be-
come nonattenuated and a lossless medium is obtained. They found that there are
three types of surface waves that can propagate in these configurations: The Stoneley
wave, the pseudo-Stoneley wave and the pseudo-Rayleigh wave. For lower frequen-
cies the phase velocities and the attenuation coefficients of the surface waves are
predominantly influenced by the permeability of the formation, since the fluid inflow
and outflow in the interface region is the main dissipative mechanism. Experimental
data regarding flat interfaces have been reported by Mayes et al. (1986), Nagy (1992)
and recently by Allard et al. (2002, 2003) who studied surface waves in air-saturated
porous media.
Also in the case of a cylindrical porous medium-fluid interface, the behavior of
the surface waves has been discussed. Rosenbaum (1974) applied Biot’s theory to
solve the high-frequency modes of a fluid-filled borehole surrounded by a porous
formation. The low-frequency limit, where the predominant surface wave is referred
as the tube wave, was derived by Chang et al. (1988). Norris (1989) extended White’s
formulation (White, 1983) for tube waves to take into account the compressibility of
the solid matrix. Schmitt et al. (1988) studied the wave response to a point source in
both the time and frequency domain using Biot’s theory modified according to ho-
mogenization theory. They presented results for the Stoneley mode and the pseudo-
Rayleigh mode. Cheng et al. (1987) applied Biot’s theory to interpret the permeabil-
ity dependence of the properties of the Stoneley wave observed in field data. Liu and
Johnson (1997) simulated the effect of a mudcake in the wall of the borehole with
an elastic membrane. The influence of a thin impairment layer in the wall of the
borehole on the dispersion relation of the surface waves was analyzed by Tichelaar
et al. (1999). Liu (1988) studied borehole modes in a cylindrical water-saturated
permeable medium.
Laboratory data for surface waves along boreholes in porous formations are
scarce. Winkler et al. (1989) measured phase velocities and damping coefficients
of 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 elas-
tic formation. They performed measurements using stainless steel and polyethylene
as formations.
In this work we study the frequency-dependent behavior of the pseudo-Stoneley
wave along boreholes in a fully saturated porous formation. We investigate the use
of shock waves to generate surface waves in a borehole. The Biot theory is applied
4.2. Theory Formulation 63
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.
4.2 Theory Formulation
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
Φ0 (kz , ω, r) = A0 (kz , ω)J0 (kf r), (4.1)
where kz refers to the wavenumber in the z−direction, kf to the radial wavenumber,

ω is the angular frequency, and r the radial coordinate. In the formation both fast
and slow compressional waves, and shear waves can propagate. We introduce for the
wave potentials Φc1 , Φc2 and Ψsh the following expressions (Ψsh is the component
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
Figure 4.1: Scaled borehole configuration in a confined reservoir.
of Ψsh in the θ-direction):
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
Φ2 (kz , ω, r) = A2 (kz , ω)J0 (kf r) + B2 (kz , ω)H0 (kf r). (4.6)

4.2. Theory Formulation 65
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:
U0 = (1 − φ)u1 + φU1 at r = R0 (4.7)
and
(1 − φ)u1 + φU1 = U2 at r = R1 . (4.8)
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 = τrr1 − φp1 at r = R0 , (4.9)
τrr1 − φp1 = −p2 at r = R1 , (4.10)

τrz1 = 0, at r = R0 , R1 . (4.11)
It also holds 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
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)
The boundary conditions are expressed in terms of displacements, pressures and

stresses. These quantities are obtained using the wave potentials according to Biot’s
theory (see appendix 4A). Substitution of the wave potentials 4.1-4.4 and 4.6 into the
boundary conditions leads to a linear system of 9 equations and 9 unknowns,
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
det[M(kz , ω)] = 0. (4.17)
At a fixed frequency ω, the equation is numerically solved for complex k z using

a Newton-Raphson algorithm. In this way, frequency-dependent phase velocities,
V (ω) = ωRe−1 (kz ), and damping coefficients, D(ω) = Im(kz ), are obtained.
There are an infinite number of wavemodes satisfying the boundary conditions. One
of these is the so-called pseudo-Stoneley wave which is generated in our experimen-
tal set-up (as we will show). Following Feng and Johnson (1983a), we define the
pseudo-Stoneley wave as the surface wave that propagates at a speed that is higher
than the slow wave and lower than that of the rest of the bulk modes. The main
difference with the true surface Stoneley wave is that the pseudo-Stoneley wave is
strongly attenuated in the direction of propagation. We will compare the results for
the pseudo-Stoneley wave in our configuration, when the reservoir has a finite lateral
dimension R1 (confined reservoir) with a more realistic situation where R1 → ∞
(laterally infinite reservoir). In this case, Ac1 = Ac2 = Ash = A2 = B2 = 0, and
the dimension of the matrix reduces to 4 × 4. The matrix elements for this case are
given in Appendix 4C.
4.3 Permeability Effects on the Pseudo-Stoneley Wave
In this section we investigate the permeability dependence of the properties of the

pseudo-Stoneley wave for laterally infinite reservoirs. We also examine how good
4.3. Permeability Effects on the Pseudo-Stoneley Wave 67
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)
1000
Vps (36 mD)
800
PSfrag replacements
600 Vps (360 mD)
400
confined (360 mD)

200
Vc2 (360 mD)
2 3 4 5
10 10 10 10
Frequency (Hz)
Figure 4.2: Frequency-dependent phase velocity for the p-Stoneley wave for different perme-
ability values. The solid lines correspond to the infinite reservoir configuration.
The results for the confined reservoir configuration for k0 = 360 mD are shown
in dotted line. The slow wave and shear wave velocities are also displayed for
this case (k0 = 360 mD) in dashed lines.
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-
2
10

1
10 Dc2 (360 mD)
confined (360 mD)

PSfrag replacements
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-
Inner radius R0 (mm) 6.3

External radius R1 (mm) 37.1
Density of the solid ρs (kg/m3 ) 2644
Porosity φ 0.20
Permeability k0 (mD) 360
Bulk modulus of the solid Kb (GPa) 10.37
Bulk modulus of the solid grains Ks (GPa) 36.5
Density of the fluid ρf (kg/m3 ) 1000
Bulk modulus of the fluid Kf (GPa) 2.25
Viscosity of the fluid η (mPa s) 1.0
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-
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
0.9
0.8
0.7
|p/p(R0)|
0.6
0.5
0.4 360 mD confined (360 mD)

0.3
PSfrag replacements
0.2
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.
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 pseudo-Stoneley wave on the permeability can now be explained by considering

(pS) (pS)
the contribution γc2 θc2 of the slow wave. By nature, the slow wave is highly
permeability-dependent. We notice that it plays an important role at low frequen-
cies and thus also the phase velocity and damping of the pseudo-Stoneley wave are
strongly dependent on permeability at those low frequencies (see Figures 4.2 and
4.3). At higher frequencies the contributions of the fast and shear waves become
more important and consequently the dependence on permeability of the properties
of the pseudo-Stoneley wave diminishes.
4.4 Experimental Setup and Results
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-
0
10
Normalized pressure amplitude

−1
10
−2
10
−3
10
−4
10
10
0 50 100 150 200 250 300 350 400 450 500
Frequency (kHz)
Figure 4.10: Normalized absolute pressure values of the incident wave.
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
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
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.
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.
4.5 Conclusions and Discussion
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
4.6 Appendix 4A: Potential Formulation of Biot’s Theory in Cylindrical

Coordinates
Biot’s equation of poroelasticity can be written in terms of the displacements as fol-

lows:
∂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:
ρ11 = (α∞ − 1)φρf + (1 − φ)ρs , (4.25)

ρ12 = (1 − α∞ )φρf , (4.26)

ρ22 = α∞ φρf , (4.27)
where ρs is the density of the constitutive material (quartz in our case), ρ f is the
density of the saturating pore fluid and α∞ is the tortuosity. The frequency-dependent
formation factor F (ω) is given below according to the formulation of Johnson et al.
(1987),
iωρf α∞ κ0 1
F (ω) = (1 − )2 , (4.28)
2ηφ
where η is the viscosity of the pore fluid and κ0 is the steady-state permeability.
Introducing the displacements wave potentials associated with each of the three
different body waves that can propagate in the porous medium, Φ c1 , Φc2 , Ψsh , the
solid and fluid displacement can be expressed as follows (e θ is the unit vector in the
θ-direction):
u = ∇Φc1 + ∇Φc2 + ∇ × (Ψsh eθ ), (4.29)
and,
U = Gc1 ∇Φc1 + Gc2 ∇Φc2 + Gsh ∇ × (Ψsh eθ ), (4.30)
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:
τij = (P − 2N )ukk δij + 2N uij + QUkk δij (4.38)
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)
4.7 Appendix 4B: Matrix coefficients for the confined reservoir
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:
γc1 = (1 − φ + φGc1 )kc1 , (4.42)
γc2 = (1 − φ + φGc2 )kc2 , (4.43)

and
γsh = (1 − φ + φGsh )ikz . (4.44)
m11 = −kf J1 (kf R0 ),

m12 = γc1 J1 (kc1 R0 ),
m13 = γc1 H1 (kc1 R0 ),
m14 = γc2 J1 (kc2 R0 ),
m15 = γc2 H1 (kc2 R0 ),
m16 = γsh J1 (ksh R0 ),

m17 = γsh H1 (ksh R0 ),
m21 = −ρ0 ω 2 J0 (kf R0 ),
2
m22 = ω2 J0 (kc1 R0 )Cc1 + 2N kc1 2 [J (k R ) − J1 (kc1 R0 ) ],
0 c1 0 kc1 R0
vc1
2 H 1 (kc1 R0 )
m23 = ω2 H0 (kc1 R0 )Cc1 + 2N kc1 2 [H (k R ) −
0 c1 0 ],
vc1 kc1 R0
2
0 c2 0 kc2 R0
vc2
2 H 1 (kc2 R0 )
m25 = ω2 H0 (kc2 R0 )Cc2 + 2N kc2 2 [H (k R ) −
0 c2 0 ],
vc2 kc2 R0
J (k R )
m26 = 2N ikz ksh [J0 (ksh R0 ) − 1 sh 0 ],
ksh R0
H1 (ksh R0 )
m27 = 2N ikz ksh [H0 (ksh R0 ) − ],
ksh R0
m32 = −2ikz kc1 N J1 (kc1 R0 ),
m33 = −2ikz kc1 N H1 (kc1 R0 ),
m34 = −2ikz kc2 N J1 (kc2 R0 ),
m35 = −2ikz kc2 N H1 (kc2 R0 ),
2
m36 = [2kz2 − ω2 ]N J1 (ksh R0 ),
vsh
2
m37 = [2kz2 − ω2 ]N H1 (ksh R0 ),
vsh
m41 = ρ0 ω 2 J0 (kf R0 ),
2 J (k R )
m42 = ω2 0 c1 0 [−RGc1 − Q],
vc1 φ
2 H (k R )
m43 = ω2 0 c1 0 [−RGc1 − Q],
vc1 φ
2 J (k R )
m44 = ω2 0 c2 0 [−RGc2 − Q],
vc2 φ
2 H (k R )
m45 = ω2 0 c2 0 [−RGc2 − Q],
vc2 φ
m52 = γc1 J1 (kc1 R1 ),
m53 = γc1 H1 (kc1 R1 ),
m54 = γc2 J1 (kc2 R1 ),
m55 = γc2 H1 (kc2 R1 ),
m56 = γsh J1 (ksh R1 ),
m57 = γsh H1 (ksh R1 ),
m58 = −kf J1 (kf R1 ),
m59 = −kf H1 (kf R1 ),
2
0 c1 1 kc1 R1
vc1
4.7. Appendix 4B: Matrix coefficients for the confined reservoir 81
2
m63 = ω2 H0 (kc1 R1 )Cc1 + 2N kc1 2 [H (k R ) − H1 (kc1 R1 ) ],
0 c1 1 kc1 R1
vc1
2
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.
4.8 Appendix 4C: Matrix coefficients for the radially infinite porous
formation
m11 = −kf J1 (kf R0 ),

m12 = γc1 H1 (kc1 R0 ),
m13 = γc2 H1 (kc2 R0 ),
m14 = γsh H1 (ksh R0 ),
m21 = −ρ0 ω 2 J0 (kf R0 ),
2
0 c1 0 kc1 R0
vc1
2
0 c2 0 kc2 R0
vc2
H1 (ksh R0 )
m24 = 2N ikz ksh [H0 (ksh R0 ) − ],
ksh R0
m32 = −2ikz kc1 N H1 (kc1 R0 ),
m33 = −2ikz kc2 N H1 (kc2 R0 ),
2
m34 = [2kz2 − ω2 ]N H1 (ksh R0 ),
vsh
m41 = ρ0 ω 2 J0 (kf R0 ),
2 H (k R )
m42 = ω2 0 c1 0 [−RGc1 − Q],
vc1 φ
2 H (k R )
m43 = ω2 0 c2 0 [−RGc2 − Q],
vc2 φ
m31 =m44 = 0.
5. SURFACE AND GUIDED WAVES ALONG A CYLINDRICAL
INTERFACE BETWEEN A LIQUID AND A LIQUID-SATURATED
POROUS MEDIUM
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
averaged fluid-solid radial displacement induced by the pseudo-Stoneley wave in the

poroelastic formation demonstrated that the loading of the fluid-bore column plays
a fundamental role in the generation of the surface modes. This result inspired a
modification of the shock tube setup in order to enhance the loading of the water
column and thus the generation of the surface waves. We especially aim to enhance
the excitation of the pseudo-Stoneley mode not only in the previously mentioned
Berea formation but also in higher permeable samples. These are synthetic samples
constituted of sintered crushed glass. The samples are characterized in Table 5.1.
They represent slow formations, where the shear wave velocity is lower than the
compressional wave in the borehole liquid.
The chapter is organized as follows. In section 5.2 the numerical calculations for
the dispersion relation of the guided modes travelling along boreholes are presented
and analyzed for the different configurations and formations studied in this chapter.
The modification to the experimental setup is presented in section 5.3. The measured
speed and specific attenuation coefficients and the comparison with the numerical
results are discussed in section 5.4. Finally the conclusions are drawn in section 5.5.
5.2 Borehole modes in Poroelastic Formations
5.2.1 Pseudo-Stoneley and Pseudo-Rayleigh Waves

There is an infinite number of wave modes that can propagate along a cylindrical
interface between a liquid and an elastic or poroelastic medium. An extensive study
of the different modes for an elastic formation is given by Sinha and Asvadurov
(2004). The modes can preliminary be classified according to their axial symmetry.
Axisymmetric modes propagating along an elastic formation include the Stoneley
wave and an infinite number of pseudo-Rayleigh waves. Higher nonaxisymmetric
set of borehole modes comprise the flexural modes and the quadrupole modes. We
only consider the axisymmetric case where there is no dependence of the displace-
ments and stresses on the azimuthal angle. The mathematical procedure to obtain the
frequency-dependent wave velocities and attenuation coefficients has been described
in detail in section 4.2. First a borehole radius of 6.3 cm is considered. The liquid is
water. Figure 5.1 shows the results for the formations characterized in Table 5.1. In
all cases the pseudo-Stoneley wave is observed. It is important to recall, in order to
avoid confusion, that in the liquid/poroelastic case the classical Stoneley wave that
exists in a liquid/elastic interface becomes a pseudo or leaky wave due to radiation
into the slow compressional wave. The pseudo-Stoneley wave velocity goes to zero
for the zero-frequency limit. This is not the case for a plane interface, where a finite
velocity is obtained in the zero-frequency limit (see section 2.2). At sufficiently low
frequencies the radial wavelength of the pseudo-Stoneley wave is considerably larger
5.2. Borehole modes in Poroelastic Formations 85
1
1600 pR−3 10
water pR−1 pR−2
1400
shear
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 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
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
1600 pR−1 pR−2 pR−3 pR−4

water
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).
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
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.
5.2.2 Radius Effects

The purpose of this subsection is to analyze the influence of the borehole radius on
the speed and attenuation of the surface waves. A small radius of 6.3 mm is con-
sidered here, which corresponds to the inner radius of the porous samples that were
used in the experiments. The computational results for the dispersive phase velocity
and attenuation are compared with the numerical calculations presented earlier for
a radius of 6.3 cm. Intuitively it is expected that the solutions for the two different
radii will approach the same value at high frequencies and that the dispersion results
for the smaller radius (6.3 mm) will roughly be shifted to higher frequencies with
respect to the larger radius (6.3 cm). Figures 5.2 and 5.3 show the calculated disper-
sion results in terms of the phase velocity and specific attenuation coefficients Q −1
of the pseudo-Stoneley and the first pseudo-Rayleigh waves for the Berea sandstone
and synthetic formation N1. The comparison of the numerical results for different
radii does not constitute a rigorous scaling, in this respect it should be remarked that
the numerous amount of independent parameters present in the equations makes the
application of scaling techniques unfeasible.
Figure 5.2 indicates that there is a clear similarity between the results for the wave
velocity of the pseudo-Stoneley wave for both radii, taking into account a frequency
scaling. At low frequencies the pseudo-Stoneley wave propagates faster for the larger
borehole radius. This was also noticed by Schmitt et al. (1988). They compared the
results for two different borehole radii of 7 cm and 12 cm. Their investigations are
restricted to relatively low frequencies (0.1 - 25 kHz) and are based on a modified
Biot’s formulation (Auriault, 1980; Auriault et al., 1985). Our numerical results show
that for high frequencies, the pseudo-Stoneley wave propagates faster in the slim
borehole. Obviously at sufficiently high frequencies both solutions merge to the high
frequency limit for the plane interface.
The attenuation of the pseudo-Stoneley mode is considerably higher in the slim
borehole (Figure 5.2b). It is clear that the features of the frequency-dependent atten-
1400
0
1300 10
1200
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
1100 0
10
1000
Q−1
900
800
−1
10
700
600
500
PSfrag replacements
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
2100 0.12
2000 0.1
1900 0.08
Q−1
1800 0.06
1700 0.04
1600 0.02
PSfrag replacements PSfrag replacements

1500 0
0 20 40 60 80 100 120 140 160 0 20 40 60 80 100 120 140 160
Frequency (kHz) Frequency (kHz)
(a) (b)
Figure 5.3: Borehole radius dependence of the phase velocity (a) and attenuation coefficient
(b) of the first pseudo-Rayleigh mode in the N1 formation. The results correspond
to a radius of 6.3 mm (solid lines) and 6.3 cm (dashed lines).
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.
5.2.3 Dispersion Relation in the Bounded Reservoir

In acoustic logging surveys, the formation surrounding the borehole can be assumed
to be radially infinite. In the previous chapter it was shown that the finite dimen-
sions of the shock tube samples and the presence of the shock tube wall augment
the complexity of the problem. Nevertheless, it was demonstrated that the properties
of the pseudo-Stoneley wave in a low-permeable sample were nearly unaffected by
these factors. Therefore the confined reservoir configuration reproduced in the shock
wave experiments, accurately represents the situation in a radially infinite reservoir
for the pseudo-Stoneley wave. In this subsection we complete the analysis by dis-
cussing the full set of axisymmetric modes. The physical description of the borehole
waves which leads to the dispersion relation was outlined in the previous chapter (see
section 4.2 and appendixes 4A and 4B).
Figure 5.4 shows the guided modes for the Berea sandstone confined reservoir.
Apart from the pseudo-Stoneley wave, which was extensively studied in the previous
chapter, several other modes are identified. The set of L-modes is also encountered
in liquid-filled pipes (Lafleur and Shields, 1995; Sinha et al., 1992) and in water-
saturated porous cylinders bounded by a cylindrical shell (Wisse, 1999). In our con-
5000
4500
4000
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.
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
0.2 30 kHz 0.2

90 kHz
PSfrag replacements 60 kHz PSfrag replacements

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R1 r/R1
(a) (b)
Figure 5.5: Radial dependence of the weighted radial displacement (a) and pore pressure (b)
associated with the S1 mode in the confined reservoir at frequencies of 30, 60
and 90 kHz. R1 = 38.5 mm
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
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.
5.3 Modification of the Shock Tube
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-
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 .
5.4 Experimental Results
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
1500
1000
500
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.
4500
4000
3500
2500
L3
2000
1500
1000
p-Stoneley
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.
underestimation of the theory indicates the existence of other loss mechanisms.

Figure 5.10 shows the frequency-dependent phase velocities of the guided waves
that are excited in the shock wave experiments performed on the N1 sample. Two
modes are distinctly observed, the pseudo-Stoneley wave and the L3 mode (see sub-
section 5.2.3). The speed of the pseudo-Stoneley wave is correctly predicted by the
numerical computations although fairly higher values are measured at high frequen-
cies. We report experimental data for the pseudo-Stoneley wave at frequencies as
low as 3 kHz. The L3 mode is related to the curvature of the geometry and cannot
be directly associated with any of the surface modes propagating along the borehole
wall. We have found an excellent agreement for the phase velocity between theory
and experiments for this mode even close to the cut-off frequency. The specific at-
tenuation coefficients for the pseudo-Stoneley wave and the L3 mode are shown in
Figure 5.11. The experimental Q−1 values reported in Figure 5.11 are determined
using the FFT-Prony method only, since it was not possible to isolate the waveforms
corresponding to each mode in the time domain and therefore it was not feasible to
apply the spectral ratio method. The attenuation data show an intense scatter and only
a qualitative comparison is feasible. Strikingly, the measured attenuation data for the
pseudo-Stoneley wave is well below the attenuation predicted by the Biot theory, this
may be a consequence of the limitations of the FFT-Prony method to determine the
attenuation coefficient.
1.5
1
Q−1
0.5
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
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).
3500
3000
2500
2000
1500

0 20 40 60 80 100 120 140 160
Frequency (kHz)
(a)
0
10
−1
10
Q−1
−2
10
−3
10
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.
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.
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.
6.2 Borehole Waves in Partially Saturated Porous Media
We consider a borehole penetrating a partially saturated porous formation. The con-

figuration is schematized in Figure 6.1. The material physical properties of the poroe-
lastic formation are given in Table 6.1 and have been taken from literature (Carcione
et al., 2003). They correspond to a solid matrix constituted by a mixture of 90%
quartz and 10% clay. The saturating liquid is brine and the properties of the gas
phase are chosen to simulate actual reservoir conditions. The bulk acoustic proper-
ties of the reservoir saturated by the mixture of brine and gas are described according
to the formulation outlined in Section 3.2. The resulting compressibility plots are
given in Figure 6.2. The dispersive borehole modes are obtained by solving a bound-
ary value problem in the frequency domain (see section 4.2). The different bulk
modes in the poroelastic formation and in the borehole fluid are coupled through the
boundary conditions. We have seen in the previous chapter that there are several sur-
face modes that can propagate in borehole geometries. We restrict ourselves to the
pseudo-Stoneley wave in this chapter.
6.2. Borehole Waves in Partially Saturated Porous Media 105
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-
Solid densityρs (kg/m3 ) 2585

Porosity φ 0.3
Permeability k0 (D) 0.55
Frame bulk modulus Kb (GPa) 8.67
Grain bulk modulus Ks (GPa) 34.3
Brine bulk modulus Kl (GPa) 2.4
Brine density ρl (kg/m3 ) 1040
Brine viscosity ηl (mPa s) 1.8
Gas bulk modulus Kg (GPa) 0.01
Gas density ρg (kg/m3 ) 100
Gas viscosity ηg (mPa s) 1.8
Gas thermal diffusivity ag 1.87 10−7
Gas polyatomic constant γ 1.4
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,
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
0.1 s=0.99 s=0.95

0 1 2 3 4 5
10 10 10 10 10
Frequency (Hz)
(b)
Figure 6.2: Frequency-dependent bulk modulus of the brine-air mixture. The real part ac-
counts for the compressibility of the system and the imaginary part represents the
loss mechanisms. The bubble radius is 1 mm. Computations for a bubble radius
of 1 cm and saturation s = 0.95 are shown in dotted line.
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
1500
1400
1300
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
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.
s f = 7.5 kHz f = 10 kHz f = 20 kHz

1 0.6539 0.8750 2.1805
0.99 1.2441 1.5629 3.0505
0.95 2.365 2.8794 4.9507
0.9 3.1298 3.7579 6.1225
0.8 3.9576 4.6871 7.2737
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
1500
1400 r = 1 cm
1300
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
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
PSfrag replacements 0.2 s=0.95

0.1
s=0.9
0
1 1.1 1.2 1.3 1.4 1.5
r/R
Figure 6.5: Pore-pressure radial distribution induced by the pseudo-Stoneley wave for differ-
ent brine saturations at a frequency of 10 kHz.
of the sensitivity we compute the following quantities:

s ∂v
ςv = (6.1)
v ∂s
and
s ∂D
ςD = , (6.2)
D ∂s
where v and D are the phase velocity and the damping coefficient of the pseudo-
Stoneley wave. ςv and ςD are the sensitivity parameters, also called normalized partial
derivatives (Cheng et al., 1982). We conduct this analysis at a frequency of 10 kHz.
Figure 6.6 shows the results. The sensitivity of the damping coefficient increases
monotonically with liquid saturation and it is very sensitive to this parameter, espe-
cially at very low concentrations of gas. The phase velocity is less sensitive though
a sharp increase for low gas saturations is found. These calculations indicate that
an accurate determination of the damping coefficient can provide useful information
regarding the saturation of the reservoir. On the other hand, the phase velocity may
be useful to detect small gas fractions.
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
2
10
1
10
Sensitivity
0
10
ς
D
−1 ς
10 v
−2
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.
saturated poroelastic formation.

Our numerical results indicate that there is a strong dependence of the phase ve-
locity and the damping coefficient on the liquid saturation in the porous reservoir.
The damping coefficient shows a clear increase with increasing gas fraction through-
out the band of frequencies studied (10 Hz - 50 kHz). The dependence of the phase
velocity is restricted to low frequencies. The frequency-dependent properties of the
phase velocities and damping coefficients and their dependence on the gas fraction
can be explained in terms of compressibility variations in the mixture due to the pres-
ence of the gas.
The outcome of these investigations suggests that information regarding the sat-
uration of a porous reservoir can be inferred through the analysis of the damping
coefficient of the pseudo-Stoneley wave. Remarkably, in the band of frequencies
related to borehole acoustics techniques the variation and the sensitivity of the damp-
ing coefficient to changes in the saturation is quite appreciable. However, it should
be pointed that our results depend considerably on the size of the gas bubbles, which
is a priori unknown. Also, other factors including permeability and the existence of
a mudcake influence the attenuation of the pseudo-Stoneley wave and complicate the
interpretation.
7. CONCLUSIONS AND RECOMMENDATIONS
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
pendence on the gas fraction can be primarily explained in terms of compressibility

changes in the mixture due to the gas concentration. The dissipative mechanisms in-
duced by the bubble oscillations appear to have little influence on the characteristics
of the pseudo-Stoneley wave.
116 7. Conclusions and Recommendations
SAMENVATTING
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
zandsteen verbeterde in hoge mate, en ze bleken bij lage frequenties nu uitstekend

overeen te komen met de theoretische voorspellingen. Een vergelijking tussen exper-
imentele waarden en numerieke berekeningen toont aan dat de oscillerende vloeistof-
stroming aan het boorgatoppervlak het belangrijkste verliesmechanisme voor de
pseudo-Stoneleygolf vertegenwoordigt, en dat dit goed wordt beschreven door de the-
orie van Biot bij frequenties lager dan 40 kHz. Bij hogere frequenties is er sprake van
systematische onderschattingen door de numerieke voorspellingen, wat veroorzaakt
wordt door verliesmechanismen die in het model van Biot veronachtzaamd worden.
In de experimenten werden ook geleide golven van hogere orde opgewekt, die
samenhangen met de compressiegolf in het poreuze gesteente en met de cilinder-
vorm van de schokbuis, en er werd zo gedetailleerde informatie verkregen over de
frequentieafhankelijke fasesnelheid en demping in bovengemiddeld poreuze en per-
meabele materialen. De gemeten demping van de geleide golf die samenhangt met
de compressiegolf, vertoont regelmatige oscillerende patronen die samenhangen met
radiale resonanties. Dit oscillerende gedrag wordt ook numeriek voorspeld, maar de
gemeten dempingswaarden zijn een orde groter dan de theoretische. Over het alge-
meen worden de fasesnelheden van deze golven van hogere orde correct voorspeld
door de theorie.
Om het effect van een kleine hoeveelheid gas in de poriën op de oppervlaktegol-
ven te onderzoeken, is het model van Biot uitgebreid om de extra interacties tussen de
verschillende fasen te kunnen beschrijven: akoestische straling, warmte-uitwisseling
en viskeuze dilatatie. Een numerieke studie op basis van deze theorie toont aan dat de
dispersieve fasesnelheden en dempingscoëfficiënten van de oppervlaktegolven langs
een recht oppervlak tussen een vloeistof en een gedeeltelijk verzadigd poro-elastisch
materiaal sterk afhankelijk zijn van de hoeveelheid gas in de poriën.Vergroting van
de hoeveelheid gas zorgt voor een toename van de demping over het gehele frequen-
tiedomein. Maxima in de dempingcoëfficiënt van de pseudo-Stoneleygolf worden
waargenomen in het interval van 10 tot 20 kHz, wat belangrijk is voor boorgatmetin-
gen. De grootte van het dempingsmaximum en de bijbehorende karakteristieke fre-
quentie hangen sterk af van de hoeveelheid gas in de poriën. In de limiet voor hoge
frequenties, waar alleen compressie-effecten het gedrag van de oppervlaktegolven
bepalen, wordt een overgang gevonden van de pseudo-Stoneleygolf naar de echte
Stoneleygolf. Deze overgang vindt plaats bij een karakteristieke hoeveelheid gas
waar de langzame compressiegolf en de akoestische golf in de vloeistof even snel
zijn.
De verzadigingseffecten zijn ook bestudeerd in een boorgatconfiguratie. In dit
geval wordt eveneens een met water verzadigd poro-elastisch gesteente beschouwd.
Ook hier wordt een duidelijk verband tussen de demping van de pseudo-Stoneleygolf
en de hoeveelheid gas gevonden. Over het gehele bestudeerde frequentiegebied (10
Hz-50 kHz) neemt de demping toe met toenemende hoeveelheid gas. De resul-
SAMENVATTING 119
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.
Aki, Keiiti and Richards, Paul G. (1980). Quantitative Seismology. W. H. Freeman

and Company, San Fransisco.
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.
Allard, J. F. (1993). Propagation of sound in porous media. Elsevier Science.
Archie, G.E. (1942). The electrical resistivity log as an aid in determining some
reservoir characteristics. Trans AIME, 146, 54–62.
Auriault, J.L. (1980). Dynamic behaviour of a porous medium saturated by a newto-

nian fluid. Int. J. Eng. Sci., 18, 775–785.
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. (1956a). Theory of propagation of elastic waves in a fluid-saturated

porous solid I. Low-frequency range. J. Acoust. Soc. Am., 28, 168–178.
122 BIBLIOGRAPHY
Biot, M. A. (1956b). Theory of propagation of elastic waves in a fluid-saturated

porous solid II. Higher frequency range. J. Acoust. Soc. Am., 28, 179–191.
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.
Brekhovskikh, L.M. (1980). Waves in Layered Media. Academic Press.
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.
Champoux, Y. and Allard, J. F. (1991). Dynamic tourtuosity and bulk modulus in

air-saturated porous media. J. Appl. Phys., 70, 1975–1797.
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
Feng, S. and Johnson, D.L. (1983a). High-frequency acoustic properties of a

fluid/porous solid interface. I. New surface mode. J. Acoust. Soc. Am., 74, 906–
914.
Feng, S. and Johnson, D.L. (1983b). High-frequency acoustic properties of a

fluid/porous solid interface. II. The 2D reflection Green’s function. J. Acoust. Soc.
Am., 74, 915–914.
Frenkel, J. (1944). On the theory of seismic and seismoelectrivc phenomena in a

moist soil. Journal of Physics (USSR) VIII, 4, 230–241.
Gist, Grant A (1994a). Fluid effects on velocity and attenuation in sandstones. J.

Acoust. Soc. Am., 96, 1158–1173.
Gist, Grant A (1994b). Interpreting laboratory velocity measurements in partially

gas-saturated rocks. Geophysics, 59, 1100–1109.
Goldstein, Herbert (1980). Classical mechanics. Addison-Wesley.
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. (1986). Recent developments in the acoustic properties of porous

media. In Frontiers of physical acoustics. North Holland Publishing Company,
Amsterdam.
Johnson, D.L. (2001). Theory of frequency dependent acoustics in patchy-saturated

porous media. J. Acoust. Soc. Am., 110(2), 682–694.
BIBLIOGRAPHY 125
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.
Jorgensen, C.S. and Kundu, T. (2002). Measurements of material elastic constants

of trabecular bone: a micromechanical analytical study using a 1 GHz acoustic
microscope. J. Orthopaed. Res., 20(1), 151–158.
Kelder, O. (1998). Frequency-dependent wave propagation in water-saturated

porous media. Ph.D. thesis, Delft University of Technology.
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.
Lafleur, L. D. and Shields, F.D. (1995). Low-frequency propagation modes in a

liquid-filled elastic tube waveguide. J. Acoust. Soc. Am., 97, 1435–1445.
Liu, H.L. (1988). Borehole modes in a cylindrical fluid-saturated permeable medium.

J. Acoust. Soc. Am., 84, 424–431.
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.
Marple, S.L. (1987). Digital spectral analysis: with applications. Prentice-Hall.
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
Nagy, Peter B. (1992). Observation of a new surface mode on a fluid-saturated per-

meable solid. Appl. Phys. Lett., 60(22), 2735–2737.
Norris, Andrew N. (1989). Stoneley-wave attenuation in permeable formations. Geo-

physics, 54(3), 330–341.
Paillet, F.L. and Cheng, C.H. (1991). Acoustic waves in boreholes- the theory and
application of acoustic full-waveform logs. CRC Press.
Plona, T.J. (1980). Observation of a second bulk compressional wave in a porous

medium at ultrasonic frequencies. Appl. Phys. Lett., 36, 259–261.
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.
Rosenbaum, J.H. (1974). Synthetic microseismograms: Logging in porous forma-

tions. Geophysics, 39, 14–32.
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. (1992). On wave propagation in saturated and partially saturated

porous media. Ph.D. thesis, Eindhoven University of Technology.
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. (1983). Underground sound-Application of seismic waves. Elsevier Sci-

ence.
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. (1999). On frequency dependence of acoustic waves in porous cylinders.

Ph.D. thesis, Delft University of Technology.
128 BIBLIOGRAPHY
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.
Wood, A.B. (1955). A Textbook of Sound. Bell.
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.
Zwikker, C. and Kosten, C. (1941). Extended theory of the absortion of sound by

compressible wall-coverings. Physica VIII, 9, 968–978.
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
1975 Born in Buenos Aires, Argentina.
2000 Diploma (MSc) in Physics, University of Buenos Aires, Argentina.
1998-2001 Teaching Assistant

Department of Physics, University of Buenos Aires, Argentina.
2001-2005 Research Assistant, Applied Geophysics and Petrophysics Group,

Delft University of Technology, The Netherlands.
2001-2005 Research Assistant, Fluid Dynamics Group,

Eindhoven University of Technology, The Netherlands.

Gabriel Eugenio CHAO, Dispersive Surface Acoustic Waves in Poroelastic Media

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Gabriel Eugenio CHAO, Dispersive Surface Acoustic Waves in Poroelastic Media

Caricato da

Copyright:

Formati disponibili

Dispersive surface acoustic waves

ter verkrijging van de graad van doctor

Gabriel Eugenio CHAO

Licenciado en Ciencias Fisicas, Universidad de Buenos Aires

Rector Magnificus, voorzitter

This research was financially supported by the ISES program.

3. Surface Waves along Partially Saturated Porous Media . . . . . . . . 35

3.2 Acoustic Properties of a Partially Saturated Porous Medium: The

4. Shock-Induced Borehole Waves in Porous Formations: Theory and Experi-

5. Surface and Guided Waves along a Cylindrical Interface between a Liquid

6. Seismic Signatures of Partial Saturation on Acoustic Borehole Modes . . 103

7. Conclusions and Recommendations . . . . . . . . . . . . . . . 113

Curriculum Vitae . . . . . . . . . . . . . . . . . . . . . . . . 131

1.1 Introductory Remarks

of the reservoir. Theoretical and experimental investigations in boreholes surrounded

1.2 Literature Review

1.2.1 Wave Propagation in Porous Media

1.2.2 Attenuation Mechanisms

can be described in terms of a characteristic squirt-flow length. It has been shown

1.2.3 Surface Waves in Porous Media

1.2.4 Acoustic Borehole Modes in Porous Reservoirs

1.2.5 Experimental Studies on Bulk and Surface Waves in Poroelastic Media

1.3 Objectives of this Thesis

1.4 Thesis Outline

following a Sommerfeld formulation. At the end of this chapter approximate models

2.1 Dynamics of a Continuum Porous Medium: The Biot theory

2.1.1 Stress-Strain Relations

2.1.2 The Lagrangian Approach to the Biot Equations

ρ11 = (1 − φ) ρs + (α∞ − 1) φρf , (2.7)

ρ12 = (1 − α∞ ) φρf , (2.8)

Due to the intrinsic frequency-dependent nature of the viscous interaction, it is useful

Here we have introduced a frequency-dependent tortuosity:

2.1.3 Displacement Potential Formulation. Body Waves

where ρ is the density matrix given by:

Φ is a vector defined as [Φs Φf ]T and M is a matrix containing the generalized elastic

Assuming that the solutions of equations 2.25 can be expressed as

The velocities of the compressional waves follow from:

The attenuation of the compressional modes can be expressed in terms of the

Φs = Φc1 + Φc2 (2.33)

Φf = Gc1 Φc1 + Gc2 Φc2 , (2.34)

where Gc1 and Gc2 are given by:

U = Gsh ∇×Ψs , (2.38)

2.1.4 The Slow Compressional Wave

Berea sandstone Synthetic glass

Table 2.1: Physical properties of the porous formations.

The slow wave is highly dispersive. It is diffusive in the low-frequency limit

2.2 Poroelastic Surface Waves along Plane Interfaces

2.2.1 Dispersive Surface Modes

Φf = Af eγf z ei(kx x−ωt) . (2.42)

Φc1 = Ac1 e−γc1 z ei(kx x−ωt) , (2.43)

Φc2 = Ac2 e−γc2 z ei(kx x−ωt) , (2.44)

Ψsh = Be−γsh z ei(kx x−ωt) êy , (2.45)

where cj is the velocity of the corresponding bulk mode.

u = ∇ (Φc1 + Φc2 ) + ∇ × Ψsh , (2.47)

U = Gc1 ∇Φc1 + Gc2 ∇Φc2 + Gsh ∇ × Ψsh , (2.48)

can be expressed as:

τzz − φp = −pf . (2.51)

det [N (kx , ω)] = 0. (2.56)

At a fixed frequency ω, equation 2.56 is numerically solved for complex k x using

2.2.2 Synthetic Seismograms: The Influence of the Leaky Modes

(1980). The incident pressure wave field Pin ,

Figure 2.6: Incident Blackman-Harris pressure pulse.

−0.5 Fast P P-Stoneley