PhD Thesis FAN2014

Shock-induced borehole waves in fractured
formations
Shock-induced borehole waves in fractured
formations
PROEFSCHRIFT
ter verkrijging van de graad van doctor

aan de Technische Universiteit Delft,
op gezag van de Rector Magnificus prof.ir. K.C.A.M. Luyben,
voorzitter van het College voor Promoties,
in het openbaar te verdedigen op woensdag 11 juni 2014 om 12.30 uur
door Huajun FAN
Master of Geological Engineering

China University of Geosciences (Beijing), China
geboren te Suichang, Zhejiang, China
Dit proefschrift is goedgekeurd door de promotor:
Prof.dr.ir. D.M.J. Smeulders
Samenstelling promotiecommissie:
Rector Magnificus, Technische Universiteit Delft, voorzitter
Prof.dr.ir. D.M.J. Smeulders, Technische Universiteit Eindhoven, promotor
Prof.dr. N. Li, PetroChina Research Institute, Beijing
Prof.dr.-ing. H. Steeb, Ruhr-Universität Bochum
Prof.dr. G. Bertotti, Technische Universiteit Delft
Prof.dr. A.V. Metrikine, Technische Universiteit Delft
Prof.dr. R.J. Schotting, Universiteit Utrecht
Prof.dr.ir. C.P.A. Wapenaar, Technische Universiteit Delft
This work is financially supported by China Scholarship Council (CSC) and

Delft University of Technology.
ISBN 978-90-8891-910-7
c 2014 by H. FAN.
All rights reserved. No part of the material protected by this copyright notice
may be reproduced or utilized in any form or by any means, electronic or
mechanical, including photocopying, recording or by any information storage
and retrieval system, without permission from the author.
Published by: Uitgeverij BOX Press, Oisterwijk, The Netherlands

Printed by: Proefschriftmaken.nl
To my family
Contents
Summary xi
Samenvatting xiii
1 Introduction 1
1.1 Ray tracing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Stoneley wave . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Literature survey . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3.1 Bulk waves in porous media . . . . . . . . . . . . . . . 5
1.3.2 Waves along flat interfaces . . . . . . . . . . . . . . . . 5
1.3.3 Stoneley waves in a borehole . . . . . . . . . . . . . . 6
1.3.4 Stoneley waves in a borehole intersected by fractures . 7
1.4 Thesis outline . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 Constitutive relations and momentum equations 9

2.1 Constitutive equation . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Momentum equations . . . . . . . . . . . . . . . . . . . . . . 11
2.3 One-dimensional field equations . . . . . . . . . . . . . . . . . 12
3 Tube wave propagation in a fluid-filled borehole intersected

by a single horizontal layer 17
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Borehole waves in flexible and fractured formations . . . . . . 19
3.2.1 Low-frequency approximation . . . . . . . . . . . . . . 19
3.2.2 Dynamic parameters . . . . . . . . . . . . . . . . . . . 25
3.3 Exact solutions for fracture and borehole waves . . . . . . . . 29
3.3.1 Rigid fracture surfaces . . . . . . . . . . . . . . . . . . 29
3.3.2 Fracture elasticity effects . . . . . . . . . . . . . . . . 32
3.3.3 Tube wave propagation in the borehole adjacent to the
fracture . . . . . . . . . . . . . . . . . . . . . . . . . . 34
vii
viii Contents
3.4 Reflection and transmission . . . . . . . . . . . . . . . . . . . 37

3.4.1 Fracture bounded by rigid solid formations . . . . . . 37
3.4.2 Fracture bounded by rigid permeable formations . . . 43
3.5 Fast Fourier transform . . . . . . . . . . . . . . . . . . . . . . 43
4 Experimental setup 47
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 Vacuum procedure . . . . . . . . . . . . . . . . . . . . . . . . 50
4.3 Sample overview and preparation . . . . . . . . . . . . . . . . 52
5 Shock-induced borehole waves and fracture effects 55

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.2 Theoretical formulation . . . . . . . . . . . . . . . . . . . . . 57
5.3 Shock-tube experiment . . . . . . . . . . . . . . . . . . . . . . 60
5.3.1 Experiment with closed fracture . . . . . . . . . . . . 60
5.3.2 Experiment with open fracture . . . . . . . . . . . . . 62
5.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
6 Shock-induced wave propagation over porous and fractured

borehole zones: theory and experiments 65
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.2 Theoretical formulation . . . . . . . . . . . . . . . . . . . . . 66
6.3 Borehole impedance . . . . . . . . . . . . . . . . . . . . . . . 68
6.3.1 Formation impedance . . . . . . . . . . . . . . . . . . 68
6.3.2 Fracture impedance . . . . . . . . . . . . . . . . . . . 71
6.4 Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . 73
6.5 Experiment results . . . . . . . . . . . . . . . . . . . . . . . . 74
6.5.1 Borehole fractures . . . . . . . . . . . . . . . . . . . . 74
6.5.2 Porous sample . . . . . . . . . . . . . . . . . . . . . . 84
6.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.7 Fractured porous sample . . . . . . . . . . . . . . . . . . . . . 87
7 Fracture effects in mandrel sample 89
8 Conclusions 97
Appendix A Derivation of the plane fracture wave equation 99
Appendix B Coherence method 101

Contents ix
Appendix C Porosity and permeability determination 103

C.1 Porosity determination . . . . . . . . . . . . . . . . . . . . . . 103
C.2 Permeability determination . . . . . . . . . . . . . . . . . . . 103
Bibliography 107
Acknowledgements 117
Curriculum Vitae 119

x Contents
Summary
Natural or hydraulic fractures are of major importance for the productivity

of hydrocarbon reservoirs. Besides fracture detection, also the aperture and
extension of the fractures are essential for a correct reservoir productivity
estimate. There are many ways to detect and measure fractures, such as
borehole televiewers and electrical borehole scans. A practical approach to
investigate fracture properties is by means of acoustic logging. In this thesis,
borehole waves along fractured media are investigated theoretically and ex-
perimentally. Theoretically, the effect of a fracture intersecting a vertical
borehole can be described by the introduction of a frequency-dependent (dy-
namic) borehole fluid compressibility which is measured in the laboratory.
The dynamic fluid bulk modulus comprises the intrinsic fluid stiffness, the
borehole wall distensibility, and the radial fluid seepage into the adjacent
(horizontal) permeable zones. The latter two effects tend to diminish the in-
trinsic fluid’s stiffness, giving rise to a lower effective bulk modulus amplitude
and thus to a lower wave speed in the borehole. The radial oscillatory fluid
seepage causes viscous friction in the adjacent zones and results in a phase
lag between the pressure increase and the compression of the borehole fluid,
leading to attenuation of the borehole waves. This seepage effect is expressed
in terms of a so-called borehole dynamic wall impedance specifying the ra-
dial fluid velocity at the borehole wall as a function of the borehole pressure
variations. If a borehole wave travels down from an undamaged zone into a
fracture zone, it will encounter an impedance contrast causing the wave to
partially reflect and partially transmit, thus revealing the presence of per-
meable fracture zone adjacent to the borehole.
Stoneley wave propagation in porous and fractured formations is studied
experimentally by means of a vertical shock tube facility. In this set-up, shock
waves in air are generated that travel downwards into a water-saturated cyl-
indrical rock sample that has a borehole drilled along the center axis. In this
way, high-energy borehole waves can be generated with excellent repeatab-
ility. A logging probe is installed in the borehole to measure the pressure
xi
xii Summary
profiles.
Reflection from the water-sample interface and from the free water inter-
face can be recorded by means of a fixed pressure transducer mounted in the
wall of the shock tube above the sample in the water layer. The fractures
in the formation are represented by small horizontal slits in composite cyl-
inders whose upper and lower parts are separated by small spacer poles. In
this way, a variable horizontal fracture (slit) aperture can be obtained. Obvi-
ously these fractures form an open connection between the borehole fluid and
the fluid outside the cylinder. Also mandrel samples are used for horizontal
slits that are not open to the fluid outside the cylinder, thus representing
fractures with finite radial extension. Wave experiments show that varying
fracture widths significantly alter the recorded Stoneley wave pressure signal
at fixed depth. The reflection and transmission of borehole tube waves over
1 and 5 mm fractures are correctly predicted by theory. Other wave experi-
ments show that attenuation in boreholes adjacent to porous zones without
fractures can be predicted by theory. This technique even allows a direct
measurement of the permeability, although the acoustically measured per-
meability and the permeability measured by falling-head technique still show
a significant discrepancy. This technique is directly applicable to fractured
porous reservoir core samples.
Samenvatting
Natuurlijke en hydraulische scheurvorming is van groot belang voor de pro-

ductiviteit van olie- en gasreservoirs. Voor een correcte schatting van deze
productiviteit zijn, naast de detectie van de scheuren, ook de bepaling van
de wijdte en de lengte van de scheuren van belang. Er zijn veel manieren om
deze scheuren te detecteren en te meten, zoals met boorgatcamera’s en door
middel van elektrische boorgatmetingen. Scheureigenschappen kunnen echter
ook op praktische wijze onderzocht worden door middel van akoestische boor-
gatmetingen. In dit proefschrift wordt de voortplanting van boorgatgolven
theoretisch en experimenteel onderzocht in het geval dat het boorgat een
breukzone doorsnijdt.
Theoretisch kan het effect van een breukvlak op golfvoortplanting in een
verticaal boorgat dat dat breukvlak doorsnijdt, beschreven worden door de
frequentie-afhankelijke (dynamische) samendrukbaarheid (bulkmodulus) van
de vloeistof in het boorgat. Dit kan worden gemeten in het laboratorium.
De dynamische bulkmodulus van de vloeistof beschrijft de intrinsieke stijfheid
van die vloeistof, de flexibiliteit van de wand van het boorgat en de zijdelingse
vloeistofpenetratie in de aangrenzende (horizontale) permeabele gesteentela-
gen of breukzones. Deze laatste twee verschijnselen verminderen de intrins-
ieke vloeistofstijfheid, waardoor de bulkmodulus kleiner wordt. Hierdoor
neemt de golfvoortplantingssnelheid in het boorgat af. De zijdelingse osciller-
ende vloeistof ondervindt wrijving in de aangrenzende poreuze gesteentelagen
of breukzones. Dit resulteert in een faseverschil tussen de druktoename in
het boorgat en de samendrukking van de vloeistof in het boorgat. Het gevolg
is dat de golven in het boorgat demping ondervinden. De zijdelingse vloeis-
tofbeweging wordt beschreven met behulp van de zogenaamde dynamische
wandweerstand van het boorgat, waarbij de zijdelingse vloeistofsnelheid ter
plaatse van de wand wordt geformuleerd als functie van de drukvariaties in
het boorgat. Als een drukgolf in een boorgat op zijn weg naar beneden ter
plaatse van een breukzone aankomt, zal hij beı̈nvloed worden door de ver-
anderende bulkmodulus van de boorgatvloeistof. Dit zorgt ervoor dat de golf
xiii
xiv Samenvatting
deels zal reflecteren en deels zal verdergaan, waardoor de aanwezigheid van

een permeabele breukzone kan worden aangetoond daar waar hij het boorgat
doorsnijdt.
De voortplanting van Stoneleygolven in poreuze formaties met breukzones
is experimenteel onderzocht met behulp van een verticale schokbuisopstelling.
In deze opstelling worden schokgolven opgewekt in lucht. Deze schokgolven
reizen omlaag naar een met water verzadigd gesteentemonster met daarin een
gecentraliseerd boorgat. Op deze manier worden hoog-energetische boorgat-
golven gemaakt met een uitstekende reproduceerbaarheid. In het boorgat
wordt een verplaatsbare detector gemonteerd om het drukprofiel te regis-
treren.
Reflecties vanaf het grensvlak tussen het water en het gesteentemonster,
en van het vrije wateroppervlak kunnen vastgelegd worden door een vaste
drukopnemer die in de wand van de schokbuis is gemonteerd, in de waterlaag
boven het gesteentemonster. De scheuren in de formatie worden gesimuleerd
door nauwe horizontale spleetvlakken tussen kunststof cilinders, die met ver-
schillende kleine afstandshoudertjes van elkaar worden gehouden. Op deze
manier wordt een variabele horizontale spleetopening verkregen. Deze spleten
vormen dus een open verbinding tussen de vloeistof in het boorgat en de
vloeistof buiten de cilinder. Ook is gebruik gemaakt van horizontale spleten
die niet in contact staan met de vloeistof buiten de cilinder. Deze doornvor-
mige monsters simuleren dus breuken met een eindige lengte in de radiale
richting. Uit onze golfexperimenten blijkt dat een variërende spleetopening
een aanzienlijke verandering teweeg brengt in de gemeten druksignalen van
de Stoneleygolf op constante diepte. De reflectie en transmissie van boorgat-
golven worden door de theorie correct voorspeld voor spleetopeningen van 1
en 5 mm. Andere golfexperimenten laten zien dat de theorie ook geschikt is
om bij afwezigheid van breukzones de verzwakking van golven in boorgaten
door poreuze gesteentelagen te kunnen voorspellen. Deze techniek is zelfs
geschikt voor een directe metingen van de permeabiliteit, hoewel de akoes-
tisch gemeten permeabiliteit en de stationaire permeabiliteitsmetingen nog
steeds een aanzienlijk verschil vertonen. De hier besproken methodieken zijn
direct toepasbaar op reservoirgesteenten die zowel poreus zijn als breukzones
bevatten.
Chapter 1
Introduction
Borehole logging is the technique to obtain quantitative information on the

formation adjacent to the wellbore by means of measuring tools inside the
borehole. These measuring tools are typically combined in a tool string which
is lowered into the wellbore by means of a long cable. The cable thus carries
Borehole Formation Borehole Formation
Formaon
Head wave wave front
Wave front
Formaon
wave front
Source Source
Figure 1.1: Sketches of a borehole and the adjacent formation. A source generates a fluid
wave in the borehole. At the interface between the formation and the borehole, the fluid
wave partially reflects back into the borehole and partially transmits into the formation
to be converted into formation body waves (left panel). A head wave in the borehole is
generated (right panel).
1
2 1. Introduction
the weight of the tool string and also transmits measurement data to the
surface. Acoustic logging is one of the borehole logging techniques. In this
technique, sound chirps in the kHz frequency range are generated that travel
through the formation and thus carry information on its consistency. The
sound waves are recorded by the tool and interpreted for reservoir properties.
Here we introduce the basic concepts of monopole acoustic logging.
In Fig. 1.1, a borehole is sketched as well as the adjacent formation. In the
left panel of Fig. 1.1, a source generates a fluid wave which propagates as a
spherical wave. When the fluid wave reaches the interface between the form-
ation and the borehole, the wave partially reflects back into the borehole and
partially transmits into the formation to be converted into formation body
waves. The formation body waves comprise compressional waves (also called
P-waves) and shear waves (also called S-waves). The propagation speeds Vp
and Vs of the compressional and shear waves depend on the formation prop-
erties. In a so-called fast formation, Vs > Vf , in a slow formation, Vs < Vf ,
where Vf is the borehole fluid wave speed.
According to Huygens’ principle, each point at the fluid-solid interface
becomes a secondary wave source. The secondary source generates two kinds
of waves: the borehole head wave (velocity Vp or Vs ) and the secondary
compressional and shear waves in the formation. In most situations, we ignore
the secondary body waves, but focus on the head waves in the borehole. By
using this method, one can measure the Vp and Vs of the formation in the
borehole. In a slow formation, there is no shear head wave present.
1.1 Ray tracing

Another method to visualize wave propagation is by ray tracing. A ray indic-
ates the direction in which the wave propagates. It is always perpendicular
to the wavefront. A ray path is the fastest path between two points. At
the interface between two media which have different acoustic impedances,
the ray path will change direction according to Snell’s law. In the left panel
of Fig. 1.2, the fluid wave (with the speed Vf ) in the borehole propagates
to the interface between the formation and the borehole with incident angle
θ1 . This wave partially reflects back into the borehole with the same angle
θ1 , and partially transmits into the formation with angle θ2 . We distinguish
compressional body waves and shear body waves. According to Snell’s law,
it holds that:
sin θ1 sin θp sin θs
= = . (1.1)
Vf Vp Vs
1.1. Ray tracing 3
Strictly speaking, the ray tracing method is only valid when the fluid wavelen-
gth in the borehole is much smaller than the borehole diameter or when the
fluid wavefront can be considered plane. Although these conditions are not
always fully met, ray tracing is still a very powerful technique for problem
visualization. The ray tracing method helps to better understand the travel
paths of the waves and provides valuable information for logging tool design.
Borehole Formation 2 1 1 2
Reflected
P-wave Transmied Receiver array
P(S)-wave
2
1
Incident
P-wave
Transmitter
Source
Figure 1.2: A fluid wave in the borehole propagates to the interface between the formation
and the borehole with incident angle θ1 . This wave partially reflects back into the borehole
with the same angle θ1 , and partially transmits into the formation with angle θ2 . The
compressional body waves and shear body waves can be distinguished by using the Snell’s
law (left panel of this figure). In the right panel of this figure, different ray paths can be
identified. By using a sophisticated acoustic tool with multiple receivers (receiver array),
detailed information on formation 1 and 2 can be obtained, as well as the position of the
interface between them.
Acoustic logging can also be used for more complicated formations, where,
for example, the formation properties are altered with radial penetration
depth. In the right panel of Fig. 1.2, different ray paths can be identified. By
using a sophisticated acoustic tool with multiple receivers (receiver array),
detailed information on formation 1 and 2 can be obtained, as well as the
position of the interface between them.
4 1. Introduction
1.2 Stoneley wave

After the arrival of head waves and the fluid wave, there are also arrivals
associated with the surface waves propagating along the borehole wall. The
most prominent one is the Stoneley wave. The speed of the Stoneley wave
VSt is lower than the shear head wave Vs and the borehole fluid wave Vf ,
and it is slightly dispersive. The attenuation of the Stoneley wave is also
frequency-dependent. The attenuation of the Stoneley wave is caused by
viscous effects where borehole fluid interacts with porous formation or frac-
tures. This attenuation is explained in more detail in the forthcoming. In
the high-frequency range, the amplitude of Stoneley wave decreases very fast
with the distance from the borehole wall. In the low-frequency range, the de-
crease of the amplitude of Stoneley wave is less. In the limiting case of long
wavelengths, the Stoneley wave is called a tube wave with constant amp-
litude across the borehole diameter. Stoneley waves are typically carrying
information on surface impedance, such as fractures and permeable zones.
The Stoneley wave propagation through permeable formations and fractures
is illustrated in Fig. 1.3. We identify transmitted Stoneley waves, reflected
Stoneley waves, and fracture waves. When a Stoneley wave propagates over
Receiver
Fracture Fracture
Permeable Zone Permeable Zone
Transmitter
Figure 1.3: A Stoneley wave propagating along the borehole wall is affected by permeable
zones and fractures intersecting the borehole.
1.3. Literature survey 5
permeable zones, the wave partially reflects and partially transmits as in-
dicated in Fig. 1.3. The presence of a permeable zone causes the Stoneley
wave to become dispersive, so that velocity and attenuation are induced as
a function of frequency. When the Stoneley wave crosses a fracture, similar
effects occur. One can obtain information on fractures and permeable zones
by analyzing the reflected and transmitted Stoneley waves.
1.3 Literature survey

1.3.1 Bulk waves in porous media
A complete theoretical description of wave propagation in fluid-saturated
infinite porous media was developed by Biot (1956a,b) and De Josselin de
Jong (1956), although earlier results were published by Zwikker and Kosten
(1941, 1949) and Frenkel (1944). Two compressional waves and one shear
wave were predicted in Biot’s theory. The first compressional wave (also
called the fast P-wave) corresponds to an in-phase motion of the solid and
the fluid, while the second compressional wave (also called the slow P-wave)
corresponds to an out-of-phase motion of the solid and the fluid. The latter
results in high attenuation of the slow P-wave. A further contribution to the
Biot theory was made by Johnson et al. (1987), who introduced the concept
of “dynamic permeability”. The dynamic permeability of porous media was
measured by Smeulders et al. (1992c), Johnson et al. (1994), and Kelder
(1998). Wave propagation in heterogeneous porous media was investigated by
Schoenberg (1984), Schoenberg and Sen (1986), Berryman and Wang (1995,
2000), and Pride and Berryman (2003a,b). In this thesis, heterogeneity is
limited to radially and circumferentially isotropic layered media.
An elegant technique to study waves in porous media is by means of a
shock tube facility. Van der Grinten et al. (1985, 1987), Sniekers et al. (1989),
Smeulders et al. (1992b), Smeulders and van Dongen (1997), and Brown et al.
(2000) performed wave experiments on water-saturated, partially saturated,
and dry samples using a shock-tube facility.
1.3.2 Waves along flat interfaces

On interfaces between different media, waves may propagate. The wave
which propagates at the interface between vacuum and an elastic half space
is called a Rayleigh wave (Rayleigh, 1885). Stoneley (1924) investigated
wave propagation at the interface between two solids and Scholte (1948)
studied wave propagation at the interface between solid and fluid. Apart
6 1. Introduction
from the Stoneley wave, also pseudo-Rayleigh wave exists. Viktorov (1967),
Überall (1973), and Brekhovskikh (1980) investigated acoustic surface waves
in general. The first studies of surface wave propagation at the interface
between fluid and fluid-saturated porous media can be attributed to Der-
esiewicz (1960, 1961, 1962) and Deresiewicz and Skalak (1963). By invest-
igating the high-frequency range of Biot’s theory where all body waves in
the porous medium are non-attenuated, Feng and Johnson (1983a,b) found
that there are three types of surface waves existing at the interface between
fluid and fluid-saturated porous media. These are pseudo-Rayleigh waves,
pseudo-Stoneley waves and true Stoneley waves. The pseudo-Stoneley wave
is attenuated as it leaks energy into the slow compressional wave. The true
Stoneley wave is non-attenuated and it has a velocity lower than the slow
compressional wave (Chao, 2005).
Other studies were performed by Edelman and Wilmanski (2002), Albers
(2006), Markov (2009) and van Dalen et al. (2010, 2011a). For the experi-
mental investigations, we refer to the work of Mayes et al. (1986), Adler and
Nagy (1994), Allard et al. (2004) and van Dalen (2011b).
1.3.3 Stoneley waves in a borehole

Rosenbaum (1974) was the first to apply Biot’s theory in fluid-filled boreholes
surrounded by poro-elastic formations to investigate the permeability effects
on Stoneley wave propagation in the high-frequency range. Paillet and White
(1982) used the plane-geometry model to illustrate the effects of the borehole
fluid on surface and body waves propagating along the borehole in an elastic
solid. White (1983) developed a model for tube wave propagation in rigid
permeable formations in the low-frequency range. Kimball and Marzetta
(1984) developed a method to detect arrivals of the borehole waves in the
time domain by computing the scalar semblance for the arrival times and
their associated slownesses of different wave modes. Hsui and Toksöz (1986)
developed a model for determining formation permeability by using tube
wave attenuation. Lang et al. (1987) used Prony method for processing full-
waveform logging data and found that the velocity dispersion of the borehole
modes have a good agreement with the theoretical computations. Schmitt et
al. (1988) studied the wave response generated by a point source in the fluid-
filled borehole using Biot’s theory in both time and frequency domains. They
presented results for both the Stoneley wave and the first pseudo-Rayleigh
wave. Chang et al. (1988) and Norris (1989) used low-frequency asymptotics
of the Biot-Rosenbaum theory for the Stoneley wave propagation. The results
are compared with White’s model. They confirm that the White model is a
1.3. Literature survey 7
representative model for Stoneley wave propagation in a rigid formation in the

low-frequency range. Winkler et al. (1989) performed wave experiments for
Stoneley wave propagation and found a good agreement between experiment
and theory. Tang et al. (1991a) developed a simplified model for Stoneley
wave propagation in a permeable borehole and compared it with the full Biot-
Rosenbaum model. He found that the simple model and the Biot-Rosenbaum
model yield practically the same result in the hard-formation case. The
simplified Biot-Rosenbaum model was used for the inversion of formation
permeability from Stoneley wave logs by Tang and Cheng (1996).
Borehole surface waves are commonly used for fracture evaluation around
the borehole. Some field measurements showed the potential for the mode
converted sonic waves at a fracture to investigate the fracture extension
around the borehole (Gelinski and Cheng, 1998; Yamamoto et al., 1998). Also
VSP (Vertical Seismic Profiling) data are widely used for the study of the
fractures. Fracture waves induced by incident seismic waves at the fracture
surfaces are identified in VSP data for fracture characterization (Beydoun
et al., 1985; Cicerone and Toksöz, 1990; Toksöz et al., 1992). Similarly, the
diffraction behavior of elastic waves at the fracture tip and fracture surfaces
can be used for this (Van der Hijden and Neerhoff, 1984; Coutant, 1989; Liu
et al., 1997). Henry (2005) used the direct scattering problem for investigat-
ing the wave response of a single horizontal fracture intersecting a fluid-filled
borehole.
1.3.4 Stoneley waves in a borehole intersected by fractures

Stoneley waves for borehole fracture evaluation were investigated by Paillet
and White (1982), Cheng et al. (1987) and Hsu et al. (1987). Hornby et al.
(1989) used the reflected Stoneley wave to estimate the effective fracture aper-
tures. Tang and Cheng (1993) considered a zone containing large amounts
of microfractures to be a highly permeable zone for Stoneley wave propaga-
tion. They found that the effective permeability of the fracture zone is more
important than the sum of all fracture apertures. Kostek et al. (1998b) gave
complete numerical models of Stoneley wave propagation including borehole
washout effects and fracture effects. Formation elasticity effects were studied
by Tang (1990) and Kostek et al. (1998b). Stoneley wave propagation in a
borehole with vertical fractures was investigated by Holzhausen and Gooch
(1985), Tang and Cheng (1989), Tang et al. (1991b), Paige et al. (1992, 1995)
and Groenenboom (1998). Saito et al. (2004) used frequency-domain Stone-
ley wave field data to detect permeable fractures. Field data for fracture
detection by means of Stoneley wave propagation was discussed by Cheng
8 1. Introduction
and Toksöz (1981), Hardin et al. (1987), Hornby et al. (1987), Hsu et al.
(1987), Hsu and Esmersoy (1992), and Brie et al. (2000).
Hardin et al. (1987) reported an approach for borehole Stoneley wave at-
tenuation based on the steady fluid flow into the fracture. Tang and Cheng
(1989) developed a wave propagation theory in a single fracture bounded by
two rigid half spaces based on the Navier-Stokes equations. The viscous ef-
fects of the fluid were taken into account. Ferrazzini and Aki (1987) presented
a theory of wave propagation in a fracture bounded in two elastic half spaces
where viscous effects were neglected. Kostek et al. (1998b) and Korneev
(2008, 2010) discussed wave propagation in a fracture which includes both
the fluid viscous effects and the fracture surfaces elasticity effects.
1.4 Thesis outline

Natural or hydraulic fractures are of paramount importance for the pro-
ductivity of hydrocarbon reservoirs. Besides fracture detection, also the
aperture and extension are essential to estimate the reservoir productivity. A
practical approach to investigate fracture properties is by means of acoustic
monitoring. In this thesis experimental shock-induced acoustic wave data
are compared with theoretical predictions. Chapter 2 gives a comprehensive
introduction to wave propagation in porous media (Biot theory). In chapter
3, an overview of existing techniques to describe Stoneley wave attenuation
along porous and fractured formation is given. Secondly, a full frequency-
dependent theory for Stoneley waves along fractured formations is derived,
and compared with existing theory. The experimental facility is described
in chapter 4, and original measurements in fractured (mandrel) samples are
discussed in chapter 5. New types of samples and wave experiments therein
are detailed in chapter 6, where also an acoustic funnel is introduced for wave
enhancement. Finally, additional mandrel sample experiments using the fun-
nel for wave enhancement, are treated in chapter 7. Conclusions are given in
chapter 8.
Chapter 2
Constitutive relations and

momentum equations
2.1 Constitutive equation

We consider a fluid-filled elastic frame which has a statistical distribution of
interconnected pores. We have the following assumptions for this system:
• The wavelength is larger than the pore size;
• The displacements of both the fluid and solid phases are assumed to be
sufficiently small, so that the equations can be linearized;
• The fluid does not react to a shear force in the solid. The fluid cannot
sustain shear forces;
• The matrix is elastic and isotropic and there is no dissipation effect in

the matrix. Only the dissipation due to the interaction between the
solid and the fluid is taken into account;
• There are no thermo-elastic or chemical reaction effects in the system.
We consider a cube of unit size of bulk material. The total normal tension
force per unit bulk area applied to the fluid faces of the unit cube τ can be
described by:
pAf
τ =− = −φp, (2.1)
Ab
where φ is porosity of the system, p is the pressure of the fluid in the pores,
Af and Ab are total surfaces of the interconnected pores and the bulk area
9
10 2. Constitutive relations and momentum equations
respectively. The total normal tension force per unit bulk area applied to
that portion of the cube faces occupied by the solid τij can be described by:
τij = −σij − (1 − φ)pδij , (2.2)
where the intergranular stress σij refers to the intergranular forces percent
bulk area. The Kronecker symbol δij is introduced because the pore fluid
cannot exert nor sustain any shear forces on the macroscopic scale.
The stress-strain relations can be written as follows (Biot, 1955):
τij = 2μeij + Aekk δij + Qkk δij , (2.3)
τ = Qekk + Rkk , (2.4)

1−φ 1−φ
σij = −2μeij − A − Q ekk δij − Q − R kk δij , (2.5)
φ φ
where eij = 1/2(∂usi /∂xj + ∂usj /∂xi ), ij = 1/2(∂uf i /∂xj + ∂uf j /∂xi ), and
summation over repeated indices is assumed. A, Q and R are generalized
elastic parameters which can be related via Gedanken experiments to poros-
ity, bulk modulus of the solid Ks , bulk modulus of the fluid Kf , bulk modulus
of the porous drained matrix Kb , and shear modulus μ of both the drained
matrix and of the composite (Biot and Willis, 1957):
A = Kb − 2μ/3 + Kf (1 − φ − Kb /Ks )2 /φeff , (2.6)
Q = φKf (1 − φ − Kb /Ks )/φeff , (2.7)
R = φ2 Kf /φeff , (2.8)
where the effective porosity φeff = φ + Kf /Ks (1 − φ − Kb /Ks ). If we assume
that the frame and pore fluid are much more compressible than the solids
themselves (Kb /Ks 1 and Kf /Ks 1), we can write
(1 − φ)2 2
A= Kf + Kb − μ, (2.9)
φ 3
Q = (1 − φ)Kf , (2.10)
R = φKf . (2.11)
For this case, the stress-strain relations are simplified. Substituting (2.9),
(2.10) and (2.11) into (2.4) and (2.5), we obtain:
2
σij = −(Kb − μ)ekk δij − 2μeij , (2.12)
3
and
1−φ
p=− Kf ekk − Kf kk . (2.13)
φ
2.2. Momentum equations 11
2.2 Momentum equations

The equations of motion for a fluid-saturated porous medium were formu-
lated by Biot (1956a,b). Zwikker and Kosten (1941, 1949) and Frenkel (1944)
published earlier results, but some aspects were ignored in their papers. It
is assumed that for long-wavelength disturbances, we can define average val-
ues of the local displacements us (r, t) in the solid and uf (r, t) in the fluid.
Denoting the solid and fluid densities ρs and ρf , respectively, the linearized
momentum equations are written as:
∂2 ∂σji ∂p
(1 − φ)ρs 2
usi = − − (1 − φ) + fi , (2.14)
∂t ∂xj ∂xi
∂2 ∂p
φρf 2
uf i = −φ − fi . (2.15)
∂t ∂xi
The interaction term is specified in its linear form:

∂ ∂2
fi = b0 + φρf (α∞ − 1) 2 (uf i − usi ). (2.16)
∂t ∂t
The viscous damping factor b0 = ηφ2 /k0 , with η the dynamic fluid viscosity
and k0 the permeability. The tortuosity is denoted α∞ . The equations of
motion resulting from momentum conservation and the stress-strain relations
can now be written as:
∂2
μ∇2 us + (A + μ)∇∇ · us + Q∇∇ · uf = (ρ11 us + ρ12 uf )
∂t2
∂
+b0 (us − uf ), (2.17)
∂t
∂2 ∂
Q∇∇ · us + R∇∇ · uf = 2
(ρ12 us + ρ22 uf ) − b0 (us − uf ), (2.18)
∂t ∂t
where the density terms are given by
ρ12 = −(α∞ − 1)φρf , (2.19)
ρ11 = (1 − φ)ρs − ρ12 , (2.20)
ρ22 = φρf − ρ12 = α∞ φρf . (2.21)

2.3 One-dimensional field equations

Essential features of wave propagation in porous media are brought out by
considering the one-dimensional situation. Introducing P = A + 2μ, (2.17)
and (2.18) can be rewritten as:

∂2 ∂ ∂2 ∂2 ∂ ∂2
ρ11 2 + b0 − P 2 usx = Q 2 + b0 − ρ12 2 uf x , (2.22)
∂t ∂t ∂x ∂x ∂t ∂t

∂2 ∂ ∂2 ∂2 ∂ ∂2
ρ22 2 + b0 − R 2 uf x = Q 2 + b0 − ρ12 2 usx . (2.23)
∂t ∂t ∂x ∂x ∂t ∂t
Combination of both equations yields that

∂2 ∂ ∂2 ∂2 ∂ ∂2
ρ22 2 + b0 − R 2 ρ11 2 + b0 − P 2 usx
∂t ∂t ∂x ∂t ∂t ∂x
2 2
∂ ∂ ∂2
= Q 2 + b0 − ρ12 2 usx . (2.24)
∂x ∂t ∂t
Developing this expression, we find that

∂4 ∂4 ∂4
(ρ11 ρ22 − ρ212 ) 4 − Γ 2 2 + (P R − Q2 ) 4
∂t ∂t ∂x ∂x
2 2

∂ ∂ H ∂
+b0 ρ 2
− usx = 0, (2.25)
∂t ∂t ρ ∂x2
where Γ = P ρ22 + Rρ11 − 2Qρ12 , ρ = ρ11 + ρ22 + 2ρ12 , and H = P + R + 2Q.
Dividing by ρ11 ρ22 − ρ212 , we find that
4
∂ 2 2 ∂4 2 2 ∂
4 ∂
4
− (V + + V − ) 2 2
+ V V
+ − 4
+ θ −1
∂t ∂t ∂x ∂x ∂t
2 2
∂ ∂
2
− V02 2 usx = 0, (2.26)
∂t ∂x
with

2 Γ± Γ2 − 4(ρ11 ρ22 − ρ212 )(P R − Q2 )
V+,− = , (2.27)
2(ρ11 ρ22 − ρ212 )
H
V02 = . (2.28)
ρ
We notice that V+,− are real-valued high-frequency phase velocities, and V0
is a low-frequency phase velocity. Moreover,
b0 ρ
θ −1 = (2.29)
ρ11 ρ22 − ρ212
2.3. One-dimensional field equations 13
is the frequency which characterizes the transition from low-frequency vis-

cosity dominated flow to high-frequency inertia dominated flow. Equation
(2.26) can be rewritten as
2 2
2 2
2 2

∂ 2 ∂ ∂ 2 ∂ −1 ∂ ∂ 2 ∂
− V+ 2 − V− 2 + θ − V0 usx
∂t2 ∂x ∂t2 ∂x ∂t ∂t2 ∂x2
= 0.(2.30)
We now consider harmonic waves of the form exp i(ωt − κx), where κ is the
wavenumber. Substitution into (2.30) and introduction of the phase speed
c = ω/κ, yields that

4 iθ −1 2 2 2 iθ −1 2
c 1− − c V+ + V− − V + V+2 V−2 = 0. (2.31)
ω ω 0
From (2.31), the squared phase speed is expressed as
√
2 V+2 + V−2 − i(ωθ)−1 V02 ± D
c = , (2.32)
2[1 − i(ωθ)−1 ]
with
D = (V+2 − V−2 )2 + 2i(ωθ)−1 [2V+2 V−2 − V02 (V+2 + V−2 )] − (ωθ)−2 V04 . (2.33)
The result of the dispersion relation (2.32) is that there are two distinct
longitudinal modes which we call mode 1 (first wave) and mode 2 (second
wave). The solution (2.32) is somewhat modified by introducing a frequency-
dependent friction factor. It corrects for the fact that above the characteristic
frequency θ −1 , deviations from low-frequency Stokes’ flow become important
owing to inertia
effects. In the limit of high frequencies, the viscous skin
depth δ = 2ν/ω (ν denotes the kinematic viscosity η/ρf ) eventually be-
comes much smaller than a characteristic viscous length scale Λ (Smeulders,
1992a). The steady-state friction factor b0 in (2.29) is then replaced by a
more realistic frequency-dependent friction factor b(ω) = b0 F (ω), where the
viscous correction factor F (ω) is given by:

b(ω) 1 ω
F (ω) = = 1 + iM , (2.34)
b0 2 ωc
where
ηφ
ωc = (2.35)
α∞ k0 ρf
3500
3000 Fast
Phase Velocity (m/s)
2500
2000
1500 Shear
1000
500 Slow
0 −2 −1 0 1 2
10 10 10 10 10
ω/ωc
2
10
Slow
0 Shear
10
Attenuation (m 1)
−
Fast
−2
10
−4
10
−6
10 −2 −1 0 1 2
10 10 10 10 10
ω/ω
c
Figure 2.1: Phase velocity and attenuation factors of the fast, slow and shear wave for a
water-saturated porous medium. The parameters of the porous medium are listed in Table
6.1.
is the frequency for which inertia and viscous forces are of equal importance
in rigid porous solids, and M is the viscous shape factor defined as
8α∞ k0
M= . (2.36)
φΛ2
2.3. One-dimensional field equations 15
It was shown that for many porous materials the shape factor M = 1 in
good approximation, so that the characteristic viscous
length scale
is directly
related to other material properties by Λ = 8k0 α∞ /φ = 8ν/ωc . We
remark that there is direct relation between the characteristic frequencies ωc
and θ −1 :
(1 − φ)ρs + (1 − α−1 )φρf
ωc = θ −1 . (2.37)
(1 − φ)ρs + φρf
In the limiting case of large tortuosity, ωc = θ −1 .

From (2.17) and (2.18) we have for the shear wave
∂2 ∂2 ∂
μ usy = (ρ11 usy + ρ12 uf y ) + b0 (usy − uf y ), (2.38)
∂x2 ∂t2 ∂t
∂2 ∂
0= 2
(ρ12 usy + ρ22 uf y ) − b0 (usy − uf y ). (2.39)
∂t ∂t
The above equations can also be written as

∂2 ∂ ∂2 ∂ ∂2
ρ11 2 + b0 − μ 2 usy = b0 − ρ12 2 uf y , (2.40)
∂t ∂t ∂x ∂t ∂t

∂ ∂2 ∂2 ∂
b0 − ρ12 2 usy = ρ22 2 + b0 uf y . (2.41)
∂t ∂t ∂t ∂t
Combination of both equations yields that

∂2 ∂ ∂2 ∂2 ∂
ρ11 2 + b0 − μ 2 ρ22 2 + b0 usy
∂t ∂t ∂x ∂t ∂t
2
∂ ∂2
= b0 − ρ12 2 usy . (2.42)
∂t ∂t
Developing this expression, we find that
2
2 ∂4 ∂4 ∂ ∂ μ ∂2
(ρ11 ρ22 − ρ12 ) 4 − μρ22 2 2 + b0 ρ − usy
∂t ∂t ∂x ∂t ∂t2 ρ ∂x2
= 0, (2.43)
where ρ = ρ11 + ρ22 + 2ρ12 . Dividing by ρ11 ρ22 − ρ212 , we find that
4 2
∂ 2 ∂4 −1 ∂ ∂ 2 ∂
2
− Vs 2 2 + θ − V0 usy = 0, (2.44)
∂t4 ∂t ∂x ∂t ∂t2 ∂x2
with
μρ22
Vs2 = , (2.45)
ρ11 ρ22 − ρ212
μ
V02 = . (2.46)
ρ
We note that Vs is real-valued high-frequency phase velocity of the shear

wave, and V0 is a low-frequency phase velocity of the shear wave. As already
described in (2.29), θ −1 is the transition frequency. Repeating the same
procedure as for the compressional waves, we obtain
Vs2 − i(ωθ)−1 V02
c2 = . (2.47)
1 − i(ωθ)−1
As an example, we plot the phase velocities and the attenuation factors of
the fast, slow and shear waves in the water-saturated case (see Fig. 2.1).
The parameters of the porous medium are listed in Table 6.1. The value of
the tortuosity is estimated from porosity by Berryman (1980) and can be
expressed as
1+φ
α∞ = . (2.48)
2φ
It is clear that the phase velocity of the slow wave is frequency-dependent
while the velocities of the other two waves are almost frequency-independent.
The velocity of the fast wave is around 3000 m/s, the velocity of the shear
wave is around 1300 m/s, the velocity of the slow wave is below 1000 m/s
for all frequencies. The attenuation factor of the fast wave has the smallest
value, the attenuation factor of the slow wave has the largest value because
of the out-of-phase character of the slow wave which is described before, the
attenuation factor of the shear wave has the value between the fast wave and
the slow wave.
Chapter 3
Tube wave propagation in a

fluid-filled borehole intersected by
a single horizontal layer
3.1 Introduction
We consider a single horizontal layer, having a height h, intersecting the
borehole (see Fig. 3.1). The layer can be a fracture with aperture h, or a
permeable zone. The layer is of infinite extent in the radial (r) direction.
The borehole has radius b. The plane z = 0 is in the center of the layer.
The z-coordinate is pointing downward. We assume that the borehole fluid
pressure is uniform across the borehole. The 1-D borehole wave equation is
d2 ψ
+ κ2 ψ = 0, (3.1)
dz 2
where ψ is the displacement potential, and κ is the wavenumber. The fluid
pressure p and the axial fluid displacement U in the borehole are given by
p = ρf ω 2 ψ, (3.2)
dψ
U= , (3.3)
dz
where ρf is the fluid density, and ω is the angular frequency. Borehole wave
propagation in the z-direction is described by
ψ = A+ e−iκ1 z + A− eiκ1 z for z < −h/2, (3.4)
ψ = B + e−iκ2 z + B − eiκ2 z for − h/2 < z < h/2, (3.5)
17
18 3. Tube wave propagation in a fluid-filled borehole
2b
Δu z
Borehole
ub
Formation Formation
r
Δz h Viscous fluid
Formation Formation
z
Figure 3.1: Borehole intersected by a single horizontal fracture. The distortion of an
elementary volume of borehole fluid by fluid compression and wall expansion is indicated
by the dashed box. The axial compression of the volume element is denoted by Δuz . The
radial wall displacement is denoted ub .
ψ = C + e−iκ1 z for z > h/2. (3.6)
In the region z < −h/2, A+ e−iκ1 z and A− eiκ1 z represent the incident wave
propagating in the positive z direction and the reflected wave propagating in
the negative z direction, respectively. Note that κ1 is the fluid wavenumber in
the undisturbed borehole, and κ2 is the fluid wavenumber where the borehole
is intersected by the layer. A+ is the incident amplitude, and A− is the
reflected amplitude (see Fig. 3.2). In the region −h/2 < z < h/2, B + and
B − are the amplitudes of the waves propagating in the positive and negative
z directions, respectively. C + is the amplitude of the transmitted wave in
the region z > h/2. The fluid displacement and the pressure should be
continuous at z = h/2 and z = −h/2. The coefficients A− , B + , B − and C +
can now be calculated as a function of the incident amplitude coefficient term
A+ :
A− /A+ = −2i(κ22 − κ21 ) sin(κ2 h)/D, (3.7)
B + /A+ = 2κ1 (κ1 + κ2 )eiκ2 h /D, (3.8)

3.2. Borehole waves in flexible and fractured formations 19
2b
A+ A-
B+ B- r
h
C+
z
Figure 3.2: Stoneley wave propagation in the borehole intersected by a single horizontal
fracture. A+ and A− represent incident and reflected Stoneley wave amplitude in z <
−h/2 respectively; B + and B − represent incident and reflected Stoneley wave amplitude
in −h/2 < z < h/2 respectively; C + represents transmitted Stoneley wave amplitude in
z > h/2.
B − /A+ = 2κ1 (κ2 − κ1 )e−iκ2 h /D, (3.9)
C + /A+ = 4κ1 κ2 eiκ1 h /D, (3.10)
where D is given by
D = (κ1 + κ2 )2 eiκ2 h − (κ1 − κ2 )2 e−iκ2 h . (3.11)
The above equations were also found by Tang and Cheng (1993). We will
now discuss tube wave propagation over different types of layers.
3.2 Borehole waves in flexible and fractured form-

ations
3.2.1 Low-frequency approximation
The description of wave propagation in the fluid in a borehole is greatly
simplified if the wavelengths involved are long compared with the borehole
diameter (low frequencies). In that case we can assume piston-like motion

in the axial (z) direction. Due to a pressure increase in the water, (i) the
water is compressed, (ii) the walls of the borehole are displaced outward over
a distance ub , and (iii) water is squeezed into the surrounding permeable
formation. For the first contribution (i), we have that the volume change of
the water ΔV1 = πb2 (∂uz /∂z)Δz (see Fig. 3.1), so that the relative volume
change of the water in this case ΔV1 /V0 = ∂uz /∂z. For the second contribu-
tion (ii), the volume change of the water ΔV2 = 2πbΔzub , so that the relative
volume change becomes ΔV2 /V0 = 2ub /b. The third contribution (iii) will be
discussed later. We know that the relation between excess pressure change

p (p = pe + p , where pe is the equilibrium pressure) and the volume change
is given by

p ΔV ΔV1 + ΔV2
=− =− , (3.12)
Kbf V0 V0
where Kbf is the incompressibility of the bore fluid. Substitution of both
contributions yields that

p ∂uz 2ub
− = + . (3.13)
Kbf ∂z b
This equation must be further developed into an expression involving only
pressure and axial displacement. The relation between radial displacement
and pressure change can be expressed in terms of a distensibility D:

bp
D= . (3.14)
2 ub
Substitution of this relations yields that the mass balance equation becomes

∂uz 1 1
= −p + . (3.15)
∂z Kbf D
The momentum balance in the z direction yields that

∂ 2 uz ∂p
ρbf 2
=− , (3.16)
∂t ∂z
where ρbf is the density of the borehole fluid. Combination (3.16) with (3.15)
yields that

∂ 2 uz ∂ 2 uz 1 1 −1
ρbf = + . (3.17)
∂t2 ∂z 2 Kbf D
This equation has the well-known solution: uz = u+ (z − cT t) + u− (z + cT t),

where u+ and u− are the downward and upward waves respectively, and cT
is the tube wave velocity. It shows that a fluid column in a flexible tube is
capable of supporting pulses of any waveform, travelling in either direction,
without dispersion, where the speed of the tube waves can be expressed as

1 1 1
= ρbf + . (3.18)
c2T Kbf D
For the third contribution (iii), we assume an outward fluid velocity wb at
the wall. Assuming harmonic variations wb = ŵb exp(iωt), the outward
fluid displacement is simply ŵb /(iω). The third contribution thus becomes
ΔV3 /V0 = 2ŵb /(iωb), and the wave velocity is subsequently modified to be-
come

1 1 1 2
= ρbf + + , (3.19)
c2T Kbf D iωbZb
where the relation between radial velocity and pressure change is expressed
in terms of a wall impedance Zb :

p
Zb = . (3.20)
wb
For further evaluation, we need to specify D and Zb . For a thick-walled
tube of inner radius b, and outer radius a, Lamb (1928) derived the classical
relation
1 E(a2 − b2 )
D= , (3.21)
2 (1 + ν)(a2 + b2 ) − 2νb2
where E is Young’s modulus and ν is Poisson’s ratio. For a borehole in an
infinite solid, a is very large compared with b, and D approaches E/2(1 + ν),
which equals the shear rigidity G of the solid.
The wall impedance can be evaluated from reservoir considerations. We
simply apply Darcy’s law as

k0 ∂p
φwr = − , (3.22)
η ∂r
where η is the viscosity of the fluid inside the pore, and wr is the fluid velocity
in the pore. Mass balance for a cylindrical coordinate system is

∂p ∂wr wr
+ ρf + = 0, (3.23)
∂t ∂r r
where ρf is the density of the fluid inside the pore. Combining (3.22) and
(3.23), we find that
2
∂p ∂ p 1 ∂p
= Dh + , (3.24)
∂t ∂r 2 r ∂r
where Dh = k0 Kf /(ηφ) is the hydraulic diffusivity, with Kf the bulk modulus

of the pore fluid. For harmonic variations p = p̂ exp(iωt), (3.24) becomes
2
∂ 1 ∂ 2
+ − κ p̂ = 0, (3.25)
∂r 2 r ∂r
with
ω2 iω
κ2 = =− , (3.26)
c2f Dh
where cf is the speed of sound in the fluid. The solution to (3.25), which is
finite at large distance is
p̂ = AK0 (κr), (3.27)
where K0 (κr) is a modified Bessel function of zeroth order and complex

argument, and A is an arbitrary constant. Substitution of the boundary
condition p̂(b) = p̂0 into (3.27) yields that
K0 (κr)
p̂ = p̂0 . (3.28)
K0 (κb)
Following Darcy’s law (3.22) , we find that the fluid velocity in the pores can
be expressed as
k0 K1 (κr)
φŵr = p̂0 κ , (3.29)
η K0 (κb)
with K1 (κr) a modified Bessel function of first order and complex argument.
At the borehole wall, we have that
k0 K1 (κb)
φ(ŵr )b = p̂0 κb , (3.30)
ηb K0 (κb)
so that the wall impedance can be expressed as
1 ŵb φ(ŵr )b k0 K1 (κb)
= = = κb . (3.31)
Zb p̂0 p̂0 ηb K0 (κb)
2
10
1
10
ηb/(k0Zb)
0
10 Re
−1 Im
10
−2
10 −2 −1 0 1 2
10 10 10 10 10
b(ω/Dh)1/2
Figure 3.3: Real and imaginary parts of the reduced wall impedance.
A plot of the reduced wall impedance is given in Fig. 3.3. For high frequencies,
√
both the real and imaginary parts tend to infinity as ω. From (3.19) and
(3.31), the tube wave number can be obtained

ω ω 2k0 Kbf κ K1 (κb) Kbf
κT = = 1+ + . (3.32)
cT cf iωηb K0 (κb) G
It is now possible to compute the tube wave velocity as a function of per-

meability and porosity. For a borehole radius of 10 cm, a fluid bulk modulus
Kbf = 2.0 GPa, and a fluid viscosity η = 1.0 mPa·s, we first compute some
values for the hydraulic diffusivity (see Table 3.1). Assuming that the bore-
Table 3.1: Parameters values.
φ (%) k0 (mD) Dh (m2 /s) ωc (MHz)

16 10.0 0.125 0.107
20 100.0 1 0.667
26 1000.0 7.692 4.41
hole is filled with the same fluid as the reservoir and that ρf = ρbf = 1000
kgm−3 , we now compute the wave speed and attenuation of the tube wave as
a function of frequency using (3.32). For the shear modulus we take G = 3.0
1200
1000
a
800
b
600
c
400
200
0 −2 0 2 4 6
10 10 10 10 10
f (Hz)
1
10
0
10
Attenuation (m−1)
−1
10 c
b
−2
10 a
−3
10
−4
10 −2 0 2 4 6
10 10 10 10 10
f (Hz)
Figure 3.4: Tube wave phase velocity and attenuation along a pore formation from (3.32).
a: k0 =10 mD, φ=16%, b: k0 =100 mD, φ=20%, c: k0 =1000 mD, φ=26%.
GPa. Results are plotted in Fig. 3.4. The dependence on porosity and per-
meability is clearly visible. The higher the permeability, the more attenuation
is observed due to the increased infiltration into the formation.
When the permeable layer is replaced by a horizontal fracture (see Fig. 3.1),
the wavenumber κ2 in the fracture zone can also be obtained from (3.19), but
with different Zb . The difference with respect to the fluid flow in the per-
meable formation is that the relation between fracture velocity and fracture
pressure is now given by Poiseuille’s law:
h2 ∂ p̂
ŵr = − . (3.33)
12η ∂r
The wall impedance Zb can then be written as
1 h2 K1 (κb)
= κb , (3.34)
Zb 12ηb K0 (κb)
and the effective fluid bulk modulus is given by
1 1 1 h2 κ K1 (κb)
= + + . (3.35)
Keff Kbf D 6iωηb K0 (κb)
We thus easily obtain the wavenumber κ2 in the borehole adjacent to the
fracture:

ω ω h2 Kbf κ K1 (κb) Kbf
κ2 = = 1+ + . (3.36)
cT cf 6iωηb K0 (κb) D
Obviously it is assumed here that 1-D approach is valid which is realistic for
λ/2b 1, for b the borehole radius. The vertical motion of the layer’s wall
is discussed in section 3.3.2.
3.2.2 Dynamic parameters

For higher frequencies, Stokes’ flow can no longer describe the flow behavior
as inertia effects become important. This can be described by a modified per-
meability and wave number in the porous media to be substituted in (3.32).
A model for dynamic permeability k(ω) of porous media was developed by
Johnson et al. (1987):
k0 ω
= i + F (ω), (3.37)
k(ω) ωc
where the viscous correction factor F (ω) is defined in (2.34), and the rollover
frequency ωc is defined in (2.35).
If the porous medium is replaced by a fracture, we can also calculate the
dynamic permeability of the fracture. For a fracture with aperture h, we
know that k0 /φ = h2 /12, so the viscous shape factor M from (2.36) has a
value of 2/3 (Λ = h in this case). The critical frequency ωc is given by
12η
ωc = . (3.38)
h2 ρf
1200
1000
a
800
b
600
c
400
200
0 −2 0 2 4 6
10 10 10 10 10
f (Hz)
1
10
0
10
Attenuation (m )
−1
−1
10 c
b
−2
10 a
−3
10
−4
10 −2 0 2 4 6
10 10 10 10 10
f (Hz)
Figure 3.5: Tube wave phase velocity and attenuation along a porous formation from
(3.32). The solid lines are for constant permeability and the dashed lines are for dy-
namic permeability (3.37) and wavenumber (3.40). a: k0 =10 mD, φ=16%, b: k0 =100 mD,
φ=20%, c: k0 =1000 mD, φ=26%.
The dynamic permeability of fracture can now be obtained from (3.37):

2
iωh ρf iωh2 ρf
k0
= + 1+ . (3.39)
k(ω) 12η 36η
1200
1000
800
a
600
b
400
200
c
0 −2 0 2 4 6
10 10 10 10 10
f (Hz)
4
10
2
10
Attenuation (m )
−1
0 c
10
b
a
−2
10
−4
10 −2 0 2 4 6
10 10 10 10 10
f (Hz)
Figure 3.6: Tube wave phase velocity and attenuation in the borehole adjacent to the
fracture from (3.36) (solid lines) compared with where the dynamic parameters are included
(dashed lines). Fracture apertures are a: 1 μm, b: 10 μm, c: 100 μm.
The introduction of inertia effects will also affect the wave number κ,
which was so far diffusive in nature (3.26). The modified wave number in the
porous medium can be written as (Johnson et al., 1987)
ω
κ̃ = α(ω). (3.40)
cf
4
10
2
10
0
10
−2
10 −2 0 2 4
10 10 10 10
f (Hz)
4
10
3
10
Attenuation (m−1)
2
10
1
10
0
10 −2 0 2 4
10 10 10 10
f (Hz)
fracture from (3.36) (solid lines) compared with where the dynamic parameters are included
(dashed lines). Fracture aperture is 1 cm.
We noted that the dynamic interaction between solid and fluid in porous
media can be described by a dynamic permeability which extends the low-
frequency viscous effects toward high-frequency inertia effects. Likewise, all
these effects are also comprised in a dynamic tortuosity, which extends the
3.3. Exact solutions for fracture and borehole waves 29
high-frequency inertia effects toward low-frequency viscous effects:

iF (ω)b0
α(ω) = α∞ − , (3.41)
φωρf
where b0 = ηφ2 /k0 is the viscous damping factor of the porous media.
For a fracture, (3.41) becomes

12iη 12iη iωh2 ρf
α(ω) = 1 − 2 F (ω) = 1 − 2 1+ . (3.42)
h ωρf h ωρf 36η
This means that eqs. (3.31), (3.32), (3.34) and (3.36) still hold, but we simply
replace k0 by k(ω) and κ by κ̃.
The phase velocity and attenuation of the tube wave in porous forma-
tions from (3.32) are compared with and without dynamic parameters (see
Fig. 3.5). For the phase velocity the two situations agree very well in both
low- and high- frequency ranges. For the attenuation, they agree very well
in low frequencies but have some discrepancy in the high-frequency range.
Above the critical frequencies ωc (see Table 3.1), inertia effects become dom-
inant and the attenuation departures from the viscosity-dominated region
(see Fig. 3.5).
Tube wave phase velocity and attenuation in the borehole adjacent to the
fracture from (3.36) are compared with where the dynamic parameters are
included (see Fig. 3.6). For 1 μm and 10 μm fractures, the two situations
agree well in all frequencies with regard to both the phase velocity and atten-
uation. When the fracture aperture increases to 100 μm, they begin to show
different behavior around 100 Hz. When the fracture aperture increases to 1
cm (see Fig. 3.7), discrepancies become very large as inertia effects become
dominant.
3.3 Exact solutions for fracture and borehole waves

3.3.1 Rigid fracture surfaces
For plane fractures, it is also possible to solve the full Navier-Stokes equation.
We now consider the wave propagation inside the fracture. Tang and Cheng
(1989) derived a theory of wave propagation in a fractured borehole, where
also the wave effects in the fracture itself were included. The linearized
Navier-Stokes equation can be written as:
∇p̂ ν
iω ŵ = − + ν∇2 ŵ + ∇ (∇ · ŵ) , (3.43)
ρf 3
where ν is the kinematic viscosity. Using vector decomposition (Miklowitz,

1978),we obtain
ŵ = ∇Φ̂ + ∇ × Ψ̂. (3.44)
We substitute (3.44) into (3.43) and we obtain that
∇p̂
iω ∇Φ̂ + ∇ × Ψ̂ + = ν ∇2 ∇Φ̂ + ∇ × Ψ̂ +
ρf
ν
∇ ∇ · ∇Φ̂ + ∇ × Ψ̂ . (3.45)
3
After multiplication of this equation by iω, we find that

2 iω 2
−ω ∇Φ̂ + ∇ × Ψ̂ = ν + cf ∇ ∇ · ∇Φ̂ + ∇ × Ψ̂ +
3
iων∇2 ∇Φ̂ + ∇ × Ψ̂ , (3.46)
where we have used that

∇p̂ 1
= − c2f ∇(∇ · ŵ). (3.47)
ρf iω
From (3.46), we obtain

2 2 4iω
∇ × −ω Ψ̂ − iων∇ Ψ̂ = ∇ ν + c2f 2 2
∇ Φ̂ + ω Φ̂ . (3.48)
3
The above equation is satisfied if
ω2
∇2 Φ̂ + Φ̂ = 0, (3.49)
c2f + 43 iων
ω
∇2 Ψ̂ + Ψ̂ = 0. (3.50)
iν
We now consider cylindrical coordinates (r, φ, z). We obtain
∂ 2 Φ̂ 1 ∂ Φ̂ ∂ 2 Φ̂ ω2
+ + + Φ̂ = 0, (3.51)
∂r 2 r ∂r ∂z 2 c2f + 43 iων
∂ 2 ψ̂ 1 ∂ ψ̂ ∂ 2 ψ̂ ψ̂ ω
2
+ + 2
− 2 + ψ̂ = 0. (3.52)
∂r r ∂r ∂z r iν
The solution to (3.51) and (3.52) is given by
Φ̂ = H0 (κr)[A cos(f z) + B sin(f z)], (3.53)
ψ̂ = H1 (κr)[C cos(f z) + D sin(f z)], (3.54)

where the ‘effective’ wavenumbers f and f¯ are given by
ω2
f2 = − κ2 , (3.55)
c2f + 43 iων
ω
f¯2 = − κ2 . (3.56)
iν
Using Helmholtz decomposition (3.44), we know that
∂ Φ̂ ∂ ψ̂θ
wr = − , (3.57)
∂r ∂z
∂ Φ̂ ∂ ψ̂ ψ̂
wz = + + . (3.58)
∂z ∂r r
Substitution of (3.53) and (3.54) in (3.57) and (3.58) yields that
wr = −H1 (κr)[Aκ cos(f z) + Bκ sin(f z)
−C f¯ sin(f¯z) + D f¯ cos(f¯z)], (3.59)
wz = H0 (κr)[−f A sin(f z) + f B cos(f z)

+Cκ cos(f¯z) + Dκ sin(f¯z)]. (3.60)
From the boundary conditions, z = ±h/2, wr = wz = 0, we obtain the
system of equations Ax = 0, where the elements of A are specified in the
Appendix A. Solving this system, we obtain the characteristic equation

2 h¯ ¯ h
κ tan f + f f tan f = 0. (3.61)
2 2
The parameters are given by
B = 0, (3.62)
C = 0, (3.63)
κ cos(f h2 )
D=− A. (3.64)
f¯ cos(f¯h )
2
Thus the fracture wavenumber κ can be obtained by solving (3.61). Res-
ults are plotted in Fig. 3.8. Phase velocity of the fracture wave increases
with increasing the fracture apertures and the frequency, attenuation of the
fracture wave increases with the increasing frequency but it decreases when
enlarging the fracture apertures. The results are also compared with that
obtained from (3.40), where α(ω) is the frequency-dependent tortuosity of
the fracture from (3.42). We notice that they are identical, for all practical
purposes.
1500
c
1000
b
500
0 1 2 3 4 5
10 10 10 10 10
f (Hz)
4
10
3
10
Attenuation (m 1)
−
2
10 a
1 b
10
0
10 c
−1
10 1 2 3 4 5
10 10 10 10 10
f (Hz)
Figure 3.8: Phase velocity and attenuation of the wave in the fracture from the exact
solution (3.61) (solid lines) compared with the generalized solution (3.40) (dashed lines).
Fracture apertures are a: 1 μm, b: 10 μm, c: 100 μm.
3.3.2 Fracture elasticity effects

The dispersion relation for guided waves in a fluid-filled infinite fracture was
developed by Ferrazzini and Aki (1987). In their approach the viscous effects
of the fluid in the fracture is neglected. From continuity of shear and normal
1500
1 mm
1000 100 μm
10 μm
1 mm
500
100 μm
10 μm
0 1 2 3 4 5
10 10 10 10 10
f (Hz)
Figure 3.9: Phase velocity of the wave in the fracture bounded by two elastic half spaces
from (3.65) (dashed lines) compared with two rigid half spaces from (3.61) (solid lines).
Three fracture apertures 10 μm, 100 μm and 1 mm are considered.
forces at the fluid-solid interface, and continuity of normal displacement, it

can be derived that:
⎡ ⎤
1 − 2f
κh ρ (2 − 2 )2
1 − 2f + ×⎣ − 4 1 − 2s ⎦ = 0,(3.65)
s s
coth
2 ρf 4s 1− 2
p
where f = ω/(κcf ), p = ω/(κcp ), and s = ω/(κcs ). The parameter ρs is

the density of the elastic formation, cp and cs are the compressional wave
and shear wave velocity in the elastic formation respectively. The above
equation is also given by Kostek et al. (1998b). We assume ρs = 2495 kg/m3 ,
cp = 1687 m/s and cs = 1097 m/s (see Table 6.1). Results are plotted in
Fig. 3.9. In Fig. 3.9, three fracture apertures 10 μm, 100 μm and 1 mm
are considered. The phase velocities of the fracture waves increase when
increasing the frequency and the fracture apertures. As a reference, we also
plot the results from (3.61) where a rigid formation is considered. Apparently,
the introduction of fracture elasticity lowers the propagation speed as we have
seen in borehole waves as well for elastic formations. However, in order to take
all effects into account, full Biot solution should be considered for fracture
elasticity on porous formation which is beyond the scope of the present thesis.
Other theories for the fracture waves have been given by Korneev (2008,
2010), who takes both the viscosity effect of the fracture fluid, and the form-
ation elasticity into account.
3.3.3 Tube wave propagation in the borehole adjacent to the

fracture
We now describe the tube wave propagation in the borehole adjacent to a
fracture. The wall impedance is defined by averaged parameters over the
fracture aperture h:
< p̂b >
Zb = , (3.66)
< ω̂b >
where p̂b is the fluid pressure at the fracture opening and ω̂b is the radial
velocity of the fluid along the fracture opening. The notation < .. > denotes
averaging over the domain [−h/2 < z < h/2], at r = b. By using (3.44),
(3.47), (3.53) and (3.64), the fluid pressure in the fracture is
iωρf A
p̂ = H0 (κr) · cos(f z). (3.67)
−1 − 4iων
3c2
f
From (3.59), we find that

ˆ h

2 h tan(f h2 ) tan(f¯h2 )
wr dz = −H1 (κr)2Aκ cos(f ) − . (3.68)
−h
2
2 f f¯
By using (3.55) and (3.61), the above equation can be written as

ˆ h
2 h 1 ω 2 /κ2
wr dz = −2Aκ sin(f )H1 (κr) . (3.69)
−h 2 f c2f + 43 iων
2
At the fracture opening r = b, (3.67) is averaged over the fracture aperture

h to match the borehole fluid pressure. So
⎧ ⎫
⎨ˆ h2 iωρ A ⎬
f
p̂0 = H 0 (κb) cos(f z)dz /h. (3.70)
⎩ −h −1 − 4iων ⎭
2 3c2 f
From the above equation, we know that

4iων
p̂0 −1 − 3c2 f2h
f
A= . (3.71)
iωρf H0 (κb) sin(f h2 )
1200
a
1000
800
b
600
400
c
200
0 1 2 3 4 5
10 10 10 10 10
f (Hz)
1
10
c
Attenuation (m 1)
0
−
10
b
−1
10 a
−2
10 1 2 3 4 5
10 10 10 10 10
f (Hz)
fracture from (3.75) (solid lines) compared with (3.36) in which the dynamic parameters
are included (dashed lines). Fracture apertures are a: 1 μm, b: 10 μm, c: 100 μm.
Now we can obtain the average fluid velocity at the fracture opening using
(3.69) and (3.71):

iω H1 (κb)
< ω̂b >= 2 2 − < p̂b > κ , (3.72)
κ cf ρf H0 (κb)
1200
1000
800
600
400
c b a
200
0 1 2 3 4
10 10 10 10
f (Hz)
0.9
10
c
0.7 b
Attenuation (m−1)
10
0.5
10
a
0.3
10
0.1
10
1 2 3 4
10 10 10 10
f (Hz)
fracture from (3.75) (solid lines) compared with (3.36) in which the dynamic parameters
are included (dashed lines). Fracture apertures are a: 0.1 mm, b: 1 mm, c: 1 cm.
so that the wall impedance becomes

iκc2f ρf H0 (κb)
Zb = . (3.73)
ωH1 (κb)
3.4. Reflection and transmission 37
Similar as in (3.35), the effective fluid bulk modulus is now

1 1 2 H1 (κb) Kbf
= 1− + . (3.74)
Keff Kbf κb H0 (κb) G
We obtain the wavenumber in the borehole adjacent to the fracture:

ω 2 H1 (κb) Kbf
κ2 = 1− + , (3.75)
cf κb H0 (κb) G
where κ can be chosen according to (3.61) or (3.40) depending on whether

it is the exact solution or the generalized solution. Tube wave velocity and
attenuation in the borehole adjacent to the fracture as a function of frequency
is plotted in Fig. 3.10 for three small fracture apertures and in Fig. 3.11 for
three larger fracture apertures. The results are compared with (3.36) in which
the dynamic parameters are included. For both the small fracture apertures
and larger fracture apertures, both approaches obtain the same result. In
Fig. 3.10, attenuation of the tube wave increases when increasing the fracture
aperture, and the velocity of the tube wave increases when decreasing the
fracture aperture. The velocity approaches approximately 1100 m/s for the
high frequency values. In Fig. 3.11, the attenuation of the tube wave increases
when increasing the fracture aperture, and the velocity of the tube wave also
increases when the fracture aperture increases. After the fracture aperture
has reached a few centimeters, a further increase does no longer affect the
phase velocity and attenuation of the borehole wave.
3.4 Reflection and transmission

3.4.1 Fracture bounded by rigid solid formations
Obviously relations (3.7) and (3.10) define reflection and transmission over a
horizontal fracture. Another possibility to compute the reflection and trans-
mission coefficients was suggested by Hornby et al. (1989). Here it is assumed
that reflection and transmission take place over the fracture center (z = 0,
see Fig. 3.12). Using relations (3.2), (3.4) and (3.6), we define
p− = ρf ω 2 (A+ e−iκz + A− eiκz ), (3.76)
p+ = ρf ω 2 (C + e−iκz ). (3.77)
Continuity of pressure at z = 0 yields that

A+ + A− = C + , (3.78)
2b
A+ A-
r
h
C+
z
Figure 3.12: Stoneley wave propagation in the borehole intersected by a single horizontal
fracture of Hornby’s approach. A+ and A− represent incident and reflected Stoneley wave
amplitude in z < 0 respectively; C + represents transmitted Stoneley wave amplitude in
z > 0.
so that the transmission coefficient T = C + /A+ and the reflection coefficient

R = A− /A+ are related as
T = 1 + R. (3.79)
Continuity of volume flux at z = 0 implies that

p0
(wz+ − wz− )πb2 + 2πbh = 0, (3.80)
Zb
where wz is fluid velocity in the z direction, Zb is the wall impedance, and
p0 is the pressure at z = 0. Using (3.3), (3.4) and (3.6), we now find that
(Zf = ρf cf ):
1 C+
[C + − (A+ − A− )]πb2 + 2πbh = 0, (3.81)
Zf Zb
which can be rewritten as

2h Zf
1−R=T 1+ . (3.82)
b Zb
0.8
0.6
|R|
5 cm
0.4
3 cm
0.2
1 cm
0
0 500 1000 1500 2000
f (Hz)
1
1 cm
0.8
3 cm
0.6
5 cm
|T|
0.4
0.2
0
0 500 1000 1500 2000
f (Hz)
Figure 3.13: Amplitude of reflection and transmission coefficients over a horizontal frac-
ture bounded by rigid solid formations. The fracture apertures are 1 cm, 3 cm and 5 cm
respectively. Solid lines are from (3.7) and (3.10) where borehole flexibility is taken into
account, dashed lines are from (3.83) where borehole flexibility is neglected.
Combination of (3.79) and (3.82) allows us to find expressions for the reflec-
tion and transmission coefficients:

h Zf h Zf h Zf
R=− / 1+ ; T = 1/ 1 + . (3.83)
b Zb b Zb b Zb
180
170
Phase of R (degree)
160
150 1 cm
3 cm
140
5 cm
130
0 500 1000 1500 2000
f (Hz)
80
70
Phase of T (degree)
60
50
40
30
20 5 cm
10 3 cm
1 cm
0
0 500 1000 1500 2000
f (Hz)
Figure 3.14: Phase of reflection and transmission coefficients over a horizontal fracture
bounded by rigid solid formations. The fracture apertures are 1 cm, 3 cm and 5 cm
respectively. Solid lines are from (3.7) and (3.10) where borehole flexibility is taken into
account, dashed lines are from (3.83) where borehole flexibility is neglected.
For the wall impedance, we adopt here the expression given by Hornby:
(1)
Zf iH1 (κ̃b)
= (1)
. (3.84)
Zb α(ω)H0 (κ̃b)
0.8
0.6
|R|
5 cm
0.4 3 cm
0.2 1 cm
0
0 500 1000 1500 2000
f (Hz)
1
1 cm
0.8
3 cm
0.6
|T|
5 cm
0.4
0.2
0
0 500 1000 1500 2000
f (Hz)
ture bounded by rigid solid formations. The fracture apertures are 1 cm, 3 cm and 5 cm
respectively. Solid lines are from (3.7) and (3.10), dashed lines are from (3.83). Borehole
flexibility is neglected in both cases.
We now compare expressions (3.83) with (3.7) and (3.10) for b = 10 cm. In
(3.7) and (3.10) we have used that

ω Kbf
κ1 = 1+ , (3.85)
cf G
0.8
0.6
|R|
5 cm
0.4
0.2 3 cm
1 cm
0
0 500 1000 1500 2000
f (Hz)
1
1 cm
0.8 3 cm
0.6 5 cm
|T|
0.4
0.2
0
0 500 1000 1500 2000
f (Hz)
ture bounded by rigid permeable formations. Solid lines: permeability k0 =0.01 D, porosity
φ=16%. Dashed lines: permeability k0 =10 D, porosity φ=26%. The fracture apertures are
1 cm, 3 cm and 5 cm respectively.
and κ2 is according to (3.75). Kbf = 2.0 GPa and G = 3.0 GPa, so that
cf = 1414 m/s. Results are given in Figs. 3.13 and 3.14. We note that
over the frequency domain up to 2000 Hz the absolute values are reasonably
in agreement although the borehole flexibility gives rise to lower reflectivity
3.5. Fast Fourier transform 43
Table 3.2: Properties of the two permeable formations.
ρbf η Kf G φ k0 Dh
(kg · m−3 ) (mPa · s) (GPa) (GPa) (%) (D) (m2 · s−1 )
1000 1.0 2.0 3.0 16 0.01 0.125
1000 1.0 2.0 3.0 26 10 76.92
and higher transmittivity. For the sake of completeness, in Fig. 3.14 also the
phase values of the reflection and transmission coefficients are given. Note
that the reflected waves are largely out-of-phase, whereas the transmitted
waves tend to be fully in-phase, for higher frequencies.
The same procedure is now repeated for 1/G = 0. From the results in
Fig. 3.15, it can be seen that the Hornby approach is in agreement with our
layered approach (3.7) and (3.10), for all practical purposes.
3.4.2 Fracture bounded by rigid permeable formations

We now consider the upper and lower formations to be rigid, but permeable,
which obviously affects the value κ1 , which becomes equal to κT , expressed
by (3.32). The wavenumber κ2 is computed by (3.36). In both cases, full
dynamic convention is applied by means of the dynamic permeability k0 →
k(ω), and κ → κ̃ = κf α(ω). In Figs. 3.16, the results are plotted for two
different formations, whose properties are given in Table 3.2. Note that the
upper and lower formations are always identical to each other. Unmistakably,
the trends closely resemble our earlier results, but it can be seen that a strong
contrast in permeability between the formation and the fracture leads to a
higher reflectivity and, thus, to a lower transmittivity.
3.5 Fast Fourier transform

In this section we describe the numerical procedure of tube wave propagation
in a fluid-saturated borehole (see Fig. 3.1). We assume that the borehole
radius b is 6.25 mm. The saturating fluid of borehole and formation is water
with a density of 1000 kgm−3 , a bulk modulus of 2.0 GPa, and a viscosity
of 1.0 mPa · s. As input signal in the time domain at the top surface of the
borehole, a Heaviside step function H(t − t0 ), where t0 = 0, 1 ms, is used.
Using the FFT routine in Matlab, the input signal in the frequency domain
is obtained.
We take two kinds of formations for the modeling. Formation 1 is imper-
meable and has a horizontal fracture intersecting the borehole. The fracture
Table 3.3: Properties of the two formations.
G (GPa) φ (%) k0 (mD) ρb (kg/m3 ) h (mm)

Formation 1 1.7 – – 1427 1
Formation 2 3.0 16 100 1310 –
aperture h is 1 mm and the center of the fracture is at z = 0 which is 200

mm below the borehole top surface. Formation 2 is permeable, and is not
intersected by fractures. The properties of the two formations are listed in
Table 3.3.
For formation 1, the pressure in the borehole in the zone above the frac-
ture (see Fig. 3.1) is given by

p̂ = ρf ω 2 A+ e−iκ1 z + (A− /A+ )eiκ1 z , (3.86)
where the ratio A− /A+ is defined by (3.7). Using the boundary condition
that p̂ is p̂0 at z = −d, where d is the distance between the borehole top and
the fracture center, yields that
−iκ z
e 1 + (A− /A+ )eiκ1 z
p̂ = pˆ0 iκ1 d . (3.87)
[e + (A− /A+ )e−iκ1 d ]
In the zone below the fracture, we have that
C+ e−iκ1 z
p̂ = pˆ0 , (3.88)
A+ [eiκ1 d + (A− /A+ )e−iκ1 d ]
where the ratio C + /A+ is defined by (3.10). The wavenumber κ1 is obtained
from (3.85) and κ2 is obtained from (3.75).
Fig. 3.17 shows 3 snapshots of the borehole pressure profile at location
above and below the fracture. For the signal at z = −52.5 mm, there is
no attenuation of the step wave because the formation is impermeable. The
reflected wave from the fracture arrives around t = 0.26 ms causing a pressure
decrease in time. For the signal at z = 37.5 mm, the wave is dispersed because
it has passed the fracture.
The modeling of formation 2 can be considered as a special case of (3.88),
where C + /A+ = 1, A− /A+ = 0 and d = 0:
p̂ = pˆ0 e−iκ1 z , (3.89)
where κ1 is obtained from (3.32). Note that the dynamic parameters are
included. Fig. 3.18 shows 3 snapshots of the borehole pressure profile at
location z = 0, 30, and 50 mm from the borehole top surface. It can clearly
3.5. Fast Fourier transform 45
Pressure (bar) 0.8
0.6
0.4
0.2 z=−200 mm
z=−52.5 mm
0 z=37.5 mm
−0.1 0 0.1 0.2 0.3 0.4 0.5

time (ms)
Figure 3.17: Step wave propagation in a borehole intersected by a fracture. The fracture
aperture h = 1 mm, the center of the fracture is 200 mm below the borehole top. The
formation is defined in Table 3.3 (Formation 1).
0.8
Pressure (bar)
0.6
0.4
0.2 z=0 mm
z=30 mm
0 z=50 mm
−0.1 0 0.1 0.2 0.3 0.4 0.5

time (ms)
Figure 3.18: Step wave propagation in a borehole adjacent to formation 2 (see Table 3.3).
Dispersion of the incident step wave is clearly visible.
be seen that the attenuation causes the amplitude of the borehole wave to
decrease, but that the original pressure profile is largely maintained. The
attenuation of the tube wave is due to the radial flow from the borehole into
the formation where wave energy is dissipated by friction.
Chapter 4
Experimental setup
4.1 Introduction
Wave experiments are carried out in a vertical shock-tube (see Figs. 4.1 and
4.2). This setup was first used for porous medium research by Van der
Grinten et al. (1985, 1987). Next, research on wave propagation in partially
saturated porous media was carried out in this setup (Smeulders and van
Dongen, 1997). Also, Wisse (1999) and Chao (2005) performed shock-tube
Figure 4.1: Picture of the shock tube Figure 4.2: Shock tube within support
support structure. structure.
47
48 4. Experimental setup
2.02 m
High pressure
section
77 mm diaphragm
4.32 m
test section
Shock tube wall

water fracture
P2
water
probe
P1
formation formation
1.1 m
fracture
0.4 m
Sample
P2
Figure 4.3: Schematic of the shock tube set up. The fractured sample is enlarged in the
inset.
experiments on surface waves in porous medium. In Fig. 4.1 the high-rising

scaffold is visible (blue) which gives access to the steel tube via two platforms
at 3.3 m and 5.9 m above the floor level. A crane (green element in Fig. 4.1)
is installed above the upper platform to manoeuvre the steel tube. The tube
itself is visible in Fig. 4.2 and is located in the heart of the scaffold structure.
Openings in the platforms are provided where the tube runs through. The
schematic of the shock tube is shown in Fig. 4.3. The tube is approximately
7.4 meters long. The low-pressure section (test section) is filled with water.
The water level height can be checked by means of optical access windows
which are installed 1.1 meters above the shock tube end plate (see Fig. 4.4).
The porous test sample is installed in the low-pressure section, and also
saturated with water. It is a cylindrical sample with a diameter of 76 mm in
which a centralized borehole is drilled. The detailed steps of the saturation
procedure will be discussed later. The test sample has a typical length of 0.4
m. The sample can be provided with a horizontal fracture as shown in the
inset of Fig. 4.3. Pressure transducers P1 and P2 record the pressure profiles
in the wave experiments. Pressure transducer P2 (Druck PDCR81 sensor,
without cover plate) is installed on top of a long probe so that it can run along
the axial direction of the borehole to measure the pressure at the different
4.1. Introduction 49
Figure 4.4: Test section of the shock tube.
positions inside the borehole (see Fig. 4.5). The Kistler transducer P1 (603B,
see Fig. 4.3) is mounted in the shock tube wall, which is used for recording
the pressure in the shock tube. We use a four-channel LeCroy 6810 data
acquisition system. Kistler 5001 amplifier is used to increase signal voltage.
The inner diameter of the tube is 77 mm. The high-pressure section is about
2 meters long. The low-pressure section is about 5.4 meters long. The wall
thickness of the tube section is 2.5 cm. The test section and the high-pressure
section are separated by a diaphragm. The membrane (diaphragm) section
is shown in Fig. 4.6. The membrane is made of polyethylene terephthalate
(PET), with thickness 0.07 mm. The diaphragm has a red tape stuck to it to
hold the debris after rupture. The black arrow is drawn on the membrane to
provide reproducible cutting from the PET sheet. Below, in the low-pressure
section, the test sample is installed (see Fig. 4.4). All elements specified
above can be found back in this picture.
A wave experiment proceeds as follows: The pressure in the high pressure
Figure 4.5: The end plate of the shock Figure 4.6: Top view of the membrane
tube, PDCR 81 transducer is installed in section. To allow unobstructed view, the
the center of the end plate, and can be high-pressure section is removed. The
moved up and down. diaphragm has a red tape stuck to it to
hold the debris after rupture. The arrow
is drawn on the membrane to provide re-
producible cutting from the PET sheet.
section is increased to 1-5 bars. Rupture of the diaphragm is caused by

means of an electric current pulse. A shock wave in air is generated which
travels downward. It is transmitted into the water layer on top of the sample.
The wave is partially reflected and partially transmitted into the borehole.
The pressures at different positions inside the borehole are measured by the
mobile transducer P2 (see Fig. 4.3). Transducer P1 is used to trigger the
data acquisition system. By performing a series of wave experiments, we can
measure the full pressure profile in the borehole.
4.2 Vacuum procedure

In order to carry out a careful saturation procedure, the shock tube and the
sample should be evacuated prior to filling them with degassed tap water.
A rotary QDP80 Drystar Pump (4 kW) is used. This pump is connected to
the shock tube, and to an evacuation vessel (see Fig. 4.7) at the same time.
The evacuation vessel is filled with tap water. After approximately 1 hour
of continuous pumping, the water in the evacuation vessel is fully degassed.
Next, this water is carefully infiltrated (gravity driven flow) into the sample
and the test section of the shock tube. Degassed water slowly enters the test
section via water inlet ‘a’ (see Fig. 4.8), which can be closed by means of a
valve. This water is distributed over the porous column, the borehole, and
the annulus between the outer sample wall and the inner shock tube wall.
4.2. Vacuum procedure 51
Figure 4.7: Evacuation vessel.
Figure 4.8: Schematic of the end plate of the shock tube.

Figure 4.9: Schematic of the mandrel sample (left) and picture of the mandrel sample
(right).
Figure 4.10: Schematic of the composed sample (left) and picture of the composed sample
(right).
4.3 Sample overview and preparation

For our experiments, we used the following samples: a) three Polyvinyl chlor-
ide (PVC) samples; b) two porous samples (clean sandstone). The samples
4.3. Sample overview and preparation 53
120o
360o
240o
Figure 4.11: Schematic of sample #5 (left) and picture of sample #5 (right), composed
of two parts separated by a small horizontal slit.
which are used for the experiments are given in Table 4.1. A schematic and
photograph of sample #1 is given in Fig. 4.9. Sample #1 consists of a man-
drel and a host cylinder (parts A and B), in which a 10 mm diameter borehole
is drilled. The fracture aperture is determined by shim stock holding parts
A and B distance h apart. The fracture extension length L is also given
in the Table. Sample #2 is exactly the same as sample #1 except for the
borehole diameter which is 12.5 mm here. We also used PVC sample #3,
whose schematic and photograph are given in Fig. 4.10. This sample consists
of two cylindrical parts. The lower part is 195 mm and the upper part is
396 mm. The two parts are connected by three small steel poles holding
the parts distance h apart and thus creating a horizontal slit over the entire
cylinder cross-section. The three steel poles are arranged at 120o azimuthal
separation. The slit aperture can be changed by using different lengths of the
steel poles. Sample #4 is a sandstone sample and is also specified in Table
4.1. The length of the sample is 348 mm, the borehole diameter is 12.5 mm
Table 4.1: Samples that are used for the shock tube experiments.
Sample #1 #2 #3 #4 #5
type mandrel mandrel composed unfractured composed
clean clean
material PVC PVC PVC
sandstone sandstone
k0 (D) − − − 15±2 15±2
φ (%) − − − 47.6 47.6
r (mm) 5 6.25 6.1 6.25 6.25
h (mm) 0.1−5 0.1−5 0.1−5 − 0.1−5
L (mm) 25 23.75 32.15 − 32.15
and the outside sample diameter is 76.8 mm. The porosity and permeability
determination of this sample is discussed in detail in the Appendix C. Sample
#5 was made from sample #4 by cutting it into 2 pieces. One piece is 30 mm
and the other is 302 mm (see Fig. 4.11). A similar procedure for attaching
both cylinder parts to each other is used as for the non-porous sample #3.
Here as well, the slit aperture is depending on the length of the steel poles.
Chapter 5
Shock-induced borehole waves and

fracture effectsi
We perform wave experiments using a vertical shock tube setup. Shock waves
are generated by the rupture of a thin membrane. In the test section, the in-
cident pressure waves generate borehole-guided waves along water-saturated
samples. The tube is equipped with side wall gages and a mobile pressure
probe, so that the attenuation and reflection of the wave can be measured.
The computation for a single horizontal fracture intersecting a vertical bore-
hole gives a quantitative prediction of reflection and transmission of borehole-
guided waves. Three different fracture apertures are used for the calculation.
Fracture aperture significantly affects both reflection and transmission coef-
ficients. Large fractures increase reflectivity and decrease transmissivity. In
the experiment, we found that both pressures above and below the fracture
are influenced by the fracture aperture indeed, thus indicating the potential
for fracture detection by borehole waves.
5.1 Introduction
Fractured reservoir is ubiquitous in hydrocarbon reservoir engineering (Chilin-
garian et al., 1992; Reza Naimi-Tajdar et al., 2007; Ramirez et al., 2009). The
Stoneley wave (St) has been used for the detection and characterization of
fracture zones (Hornby et al., 1989; Kostek et al., 1998a; Qobi et al., 2001;
Saito et al., 2004). The borehole fluid flow into the fracture and scattering at
the fracture can lead to attenuation of the St (Hornby et al., 1989; Tang and
Cheng, 1989). The relationship between St propagation and permeability was
i
This chapter has been published as a journal paper in Transport in Porous Media. 93,
263-270 (Fan and Smeulders, 2012) and is slightly modified here for consistency in notation.
55
56 5. Shock-induced borehole waves and fracture effects
2.02 m
High pressure
section
77 mm diaphragm
4.32 m
test section
Shock tube wall

water fracture
P2
water
probe
P1
formation formation
1.1 m
fracture
0.4 m
Sample
P2
Figure 5.1: Schematic of the shock tube set up. The fractured sample is enlarged in the
inset.
studied by Rosenbaum (1974). Theoretical models of borehole St propaga-

tion in permeable materials were evaluated by laboratory experiments by
Winkler et al. (1989).
In this article, we will introduce a new experimental technique for fracture
detection by means of borehole wave propagation. We use a vertical shock
tube to generate a borehole wave in a core sample in the test section of the
tube. The shock tube is shown in Fig. 5.1. The high-pressure and the low-
pressure parts are separated by a membrane. The dimensions of each section
are indicated in the figure. A cylinder which has a centralized borehole is
mounted in the test section of the shock tube. A probe inside the borehole
can record the pressures at different positions. Similar experiments were
performed by Winkler et al. (1989), but they used a conventional piezoelectric
source. The advantage of using shock-induced guided waves is that we cover a
wide band of frequencies in one experimental run, and that shock amplitude is
easily varied over a wide range of amplitude levels. The frequency band width
was investigated in a previous article (Chao et al., 2004) and found to be
between 0.5 and 80 kHz, approximately. Van der Grinten et al. (1985, 1987),
Sniekers et al. (1989), and Smeulders and van Dongen (1997) also performed
shock-tube experiments on water-saturated and partially saturated samples,
but here we focus on borehole wave along water-saturated samples.
5.2. Theoretical formulation 57
The tube is filled with water. The membrane is ruptured by an electrical

current, and a stepwise shock wave in air is generated. The shock wave then
travels downward, partially reflecting, and partially transmitting into the
water layer on top of the sample. The transmitted wave reaches the sample
where it partially propagates as a borehole wave. The pressures inside the
borehole are recorded by the mobile probe P2. The transducer P1 is mounted
in the shock-tube wall and is used to trigger the data recording system (see
Fig. 5.1).
5.2 Theoretical formulation

We consider a single horizontal fracture, having a width L, intersecting the
borehole (see Fig. 5.2). The fracture is of infinite extent in the radial (r)
direction. The borehole has radius R. The plane z = 0 is in the middle
of the fracture. The z-coordinate is pointing downward. The maximum
L r
fracture O
borehole
Figure 5.2: The cylindrical coordinates of the borehole fracture.
frequency attained in our experiments is some 80 kHz, which means that the
wavelength λ is always larger than 2.5 cm, so that for all practical purposes,
λ/d >> 1. We thus assume that the borehole fluid pressure is uniform across
the borehole (Tang and Cheng, 1993). The borehole wave equation is
d2 ψ
+ κ2 ψ = 0, (5.1)
dz 2
1
Reflection coefficient amplitude
0.8
5 mm
0.6
0.4
1 mm
0.2
0.5 mm
0
0 2000 4000 6000 8000 10000
f (Hz)
Figure 5.3: Amplitude of reflection coefficient above the fracture at three different fracture
apertures.
1
0.5 mm
Transmission coefficient
0.8
1 mm
0.6
0.4 5 mm
0.2
0
0 2000 4000 6000 8000 10000
f (Hz)
Figure 5.4: Transmission coefficient over the fracture at three different fracture apertures.
where ψ is the displacement potential, and κ is the wavenumber. The fluid

pressure P and the axial fluid displacement U in the borehole are given by
dψ
P = ρf ω 2 ψ, U= , (5.2)
dz
where ρf is the fluid density, and ω is the angular frequency. Borehole wave
propagation is described by
ψ = A+ eiκ1 z + A− e−iκ1 z for z < −L/2, (5.3)
+ iκ2 z − −iκ2 z
ψ=B e +B e for − L/2 < z < L/2, (5.4)
+ iκ1 z
ψ=C e for z > L/2. (5.5)
In the region z < −L/2, A+ eiκ1 z and A− e−iκ1 z represent the incident wave
propagating in the positive z direction and the reflected wave propagating in
the negative z direction, respectively. Note that κ1 is the fluid wavenumber in
the undisturbed borehole, and κ2 is the fluid wavenumber where the borehole
is intersected by the fracture. A+ is the incident amplitude, and A− is the
reflected amplitude. In the region −L/2 < z < L/2, B + and B − are the
amplitude of the waves propagating in the positive and negative z directions,
respectively. C + is the amplitude of the transmitted wave in the region
z > L/2. The fluid displacement and the pressure should be continuous at
z = L/2 and z = −L/2. The coefficients A− , B + , B − and C + can now be
calculated as a function of the incident amplitude coefficient term A+ :
A− /A+ = 2i(κ22 − κ21 )sin(κ2 L)/D, (5.6)
+ + −iκ2 L
B /A = 2κ1 (κ1 + κ2 )e /D, (5.7)
− +
B /A = 2κ1 (κ2 − κ1 )e iκ2 L
/D, (5.8)
+ + −iκ1 L
C /A = 4κ1 κ2 e /D, (5.9)
where D is given by
D = (κ1 + κ2 )2 e−iκ2 L − (κ1 − κ2 )2 eiκ2 L . (5.10)
The above equations were also found by Tang and Cheng (1993).
Assuming a rigid formation, the wavenumber in the region z < −L/2
and z > L/2 is simply κ1 = ω/cf = ω ρf /Kf , where cf is acoustic velocity
in the fluid, and Kf is the fluid bulk modulus. The wavenumber κ in the
fracture is given by the fracture dispersion equation (Tang and Cheng, 1989)
L L
κ2 tan( f¯) + f ftan(
¯ f ) = 0. (5.11)
2 2
The parameters f and f¯ can be expressed as follows:
ω2
f2 = 2 4 − κ2 , (5.12)
cf − 3 iων
ω
f¯2 = − κ2 . (5.13)
−iν
The kinematic viscosity ν = μ/ρf , with μ the dynamic viscosity. The effective
wavenumber κ2 at region −L/2 < z < L/2 in the borehole can now be
expressed as follows:

ω 2 H1 (κR)
κ2 = 1− (5.14)
cf κR H0 (κR)
where H0 and H1 are Hankel functions of zeroth and first orders, respectively.
This relation is derived from the assumption that an oscillatory Poiseuille flow
exists in the fracture satisfying the no-slip boundary conditions at z = ∓L/2.
We assume the borehole radius R to be 10 mm and the fluid velocity cf to be
1.5 km/s. Using equations (5.6) and (5.9), we plot the reflection coefficient
A− /A+ and the transmission coefficient C + /A+ as a function of frequency
for different fracture apertures (see Figs. 5.3, 5.4). The amplitude of the
reflection coefficient increases with fracture aperture, but decreases with fre-
quency. The trends of the transmission coefficient are obviously opposite.
5.3 Shock-tube experiment

5.3.1 Experiment with closed fracture
We performed several experiments with sample #1 (Cp =2.3 km/s, Cs =1.1
km/s) for calibration. The length of the sample is 400 mm. The diameter of
the borehole, the probe, and the sample are 10, 6.3 and 76.5 mm, respectively.
A typical recording of the trigger channel P1 is given in Fig. 5.5. Scaling
is with respect to the incident pressure. The system is triggered at time
t = 0 by the arrival of the pressure step wave in water. At time 0.535
ms, the reflected wave from the sample top arrives at the transducer P1.
Therefore, the wave propagation time from the transducer P1 to the top of the
sample is 0.268 ms. At time 0.899 ms, the wave reflected from the free water
surface arrives at P1. Also in the borehole, experiments were performed.
The two test positions of the mobile pressure probe are 225 and 245 mm
from the sample top, respectively. From the 225-mm TVD (true vertical
depth) measurement, at 0.367 ms, the first arrival of the P-wave is recorded,
followed by the arrival of Stoneley wave (St) at 0.506 ms (see Fig. 5.6). From
the 245-mm TVD recording, we note that it has similar behavior as the 225
TVD measurement, but that the P and St events arrive a little later. Now
we can easily deduce the P-wave velocity recorded by the probe inside the
borehole to be 2268.5 m/s, which is in agreement with independent Polyvinyl
chloride velocity determination by means of ultrasound. The St velocity is
983.9 m/s. Obviously, frequency domain analysis would add more insight
5.3. Shock-tube experiment 61
1.5
1
a.u.
0.5
−0.5
−0.2 0 0.2 0.4 0.6 0.8 1 1.2
time (ms)
Figure 5.5: Recorded pressure signal P1, which is used to trigger the data acquisition
system.
2.5
2
Pressure (bar)
1.5
0.5
St
0 225mm from top

245mm from top
−0.5
0.3 0.4 0.5 0.6 0.7
P time (ms)
Figure 5.6: Recorded pressure signals from the mobile probe P2 at 225- and 245-mm
TVD.
to the frequency-dependent velocity and attenuation of the individual wave

modes (one of which is the St), but that would require sophisticated FFT-
Prony-Spectral Ratio analysis, as performed by Chao et al. (2004), to discern
the individual wave modes. Here, the focus is on the St mode only.
5.3.2 Experiment with open fracture

Now we describe an experiment which has a single horizontal fracture inter-
secting the sample. The fracture is located at 200-mm TVD. The position
of the probe is below the fracture, at 230-mm TVD. The fracture aperture
is 5 mm. In reality, 1-2 mm fractures might be anticipated in the real earth.
0.6
0.5 230mm from top
0.4
Pressure (bar)
0.3
0.2
0.1
−0.1
P St
−0.2
0.3 0.4 0.5 0.6 0.7
time (ms)
Figure 5.7: Recorded pressure signal from the mobile probe P2 at 230-mm TVD. The
fracture aperture is 5 mm. The fracture is at 200-mm TVD.
In that respect, our 5-mm fracture is much larger (taking borehole scaling
into consideration). For that reason also a much smaller fracture aperture
(0.1 mm) is studied in our setup (see Fig. 5.8). We also note that Tang and
Cheng (1993) modeled centimeter scale fractures that we want to compare
against our experimental results. The recorded pressure profile with a 5-mm
fracture is given in Fig. 5.7. In this figure, the P wave arrival is at 0.369 ms
and the St arrival is at 0.526 ms. The pressure between 0.423 and 0.526 ms
is dramatically decreasing because of the presence of the fracture. We also
note that in the previous Fig. (5.6) the influence of the (closed) fracture was
already visible as a small pressure rise at t ≈ 0.43 ms, indicating that a frac-
ture once created cannot easily be undone. The theoretical predictions have
transmission coefficients always below 1, which indicates that there should be
a pressure decrease over the fracture. In the experiment, the pressure drop
is almost complete, which is much larger than predicted by theory.
5.4. Conclusion 63
Pressure (bar) 1.5
0.5
0 0.1mm fracture
St 5mm fracture
−0.5
0.3 0.4 0.5 0.6 0.7
P time (ms)
Figure 5.8: Recorded pressure signals from the mobile probe P2 at 90-mm TVD. The
fracture apertures are 0.1 and 5 mm. The fracture is at 200-mm TVD.
We also performed two new experiments with different fracture apertures:

0.1 mm and 5 mm. The probe is 110 mm above the fracture which is at 90-
mm TVD. In Fig. 5.8, we notice the arrival of the P-wave at 0.308 ms and
of the St at 0.368 ms. Obviously, fracture aperture does not affect primary
arrivals above the fracture, so that both measurement curves are indistin-
guishable. From 0.4 ms onward, both arrivals start to become different. The
pressure of the 5-mm fracture experiment decreases significantly after 0.4 ms
compared with the 0.1-mm fracture aperture experiment. This is probably
due to reflection from the fracture that affects the recorded pressure profiles.
5.4 Conclusion
The computation for a single horizontal fracture intersecting a vertical bore-
hole gives a quantitative prediction of reflection and transmission of borehole-
guided waves. Three different fracture apertures are used for the calculation.
Fracture aperture significantly affects both reflection and transmission coef-
ficients, as was also predicted in previous literature.
The shock tube can generate borehole-guided waves in a broad frequency
band. Reflection from the water-sample interface and from the free water
interface can be recorded by means of a fixed pressure transducer mounted
in the wall of the shock tube above the sample in the water layer. We also
performed borehole pressure measurements on different fracture apertures

at different borehole positions using a specially designed mobile probe. The
presence of fractures significantly affects borehole-guided waves. Large frac-
tures increase reflectivity and decrease transmissivity. We found that both
pressure above and below the fracture are altered by its presence, thus in-
dicating the potential for fracture detection. Agreement with theory is still
only qualitative.
Chapter 6
Shock-induced wave propagation

over porous and fractured borehole
zones: theory and experimentsi
Borehole waves are strongly affected by adjacent porous zones or by frac-

tures intersecting the borehole. A theoretical description for both porous
and fracture zones is possible based on the introduction of an effective bore-
hole fluid bulk modulus, characterizing the wave attenuation via borehole
wall impedance. This impedance can be calculated for both porous and
fracture zones adjacent to the borehole, thus predicting borehole wave at-
tenuation, transmission, and reflection over such zones. A shock tube setup
generates borehole tube waves that are used for porous and fracture zone
characterization. Sample #3 is used to introduce and vary fractures in a cyl-
indrical sample. Shock wave experiments show that attenuation in boreholes
adjacent to porous zones can be predicted by theory. The transmittivities of
borehole tube wave over 1 and 5 mm fractures are correctly predicted, thus
showing the potential of borehole wave experiments for fracture detection
and characterization.
6.1 Introduction
The Stoneley wave has been used to detect and characterize fracture zones
(Hornby et al., 1989; Kostek et al., 1998a,b; Qobi et al., 2001; Saito et al.,
2004). In the horizontal fracture case, Stoneley waves are attenuated because
i
This chapter (except Section 6.7) has been published as a journal paper in Journal of
the Acoustical Society of America. 134 (6), 4792-4800 (Fan and Smeulders, 2013). With
respect to the publication, additional figures were added here.
65
66 6. Wave propagation over porous and fractured borehole zones
of flow into the fracture and scattering by the fracture (Tang and Cheng,
1989). The fracture zone is traditionally modeled as a fluid-filled narrow
parallel-plate channel in which fracture waves propagate. These fracture
waves carry part of the energy of the borehole wave radially outward away
from the borehole and thus attenuate the borehole wave itself. If the viscous
skin depth of the fracture wave is on the order of the fracture aperture,
these fracture waves are also attenuated by viscous effects, and thus some
of the energy of the borehole wave is dissipated. Moreover, because the
fracture zone is experienced by the borehole Stoneley wave as a zone with
a contrast in borehole impedance, Stoneley wave reflection and transmission
will be induced. The reflected Stoneley waves carry information about the
size, shape, and orientation of the fracture and are thus of potential interest
for fractured reservoir characterization.
In this paper, first theory and application of Stoneley wave propagation in
a single horizontal fracture plane are discussed. Next we describe the shock
tube facility. This facility was used previously to detect borehole surface wave
modes (Chao et al., 2004, 2006; Smeulders and van Dongen, 1997; Sniekers
et al., 1989), and is comparable with the setup used by Winkler et al. (1989).
Measurements of the wave experiments in the shock tube are presented and
discussed. Finally, conclusions are drawn.
6.2 Theoretical formulation

A schematic of borehole configuration is given in Fig. 6.1. The borehole
radius is R, the fracture width is h. We consider cylindrical coordinates
(r, φ, z), where r is the radial distance from the borehole axis with infinite
extent, φ is the polar angle and z is the vertical coordinate. The center of
the fracture opening is at z = 0. We assume that the experiments are carried
out at frequencies lower than the cutoff frequency of any mode other than
of the fundamental tube wave, thus the borehole fluid pressure is considered
uniform across the borehole. The wave equation is given by:
d2 ψ
+ κ2 ψ = 0, (6.1)
dz 2
where ψ is the wave displacement potential and κ is the axial wavenumber.
In the region z > h/2 and z < −h/2, the wavenumber is expressed as κ1 ,
and in the layer where −h/2 < z < h/2, the wavenumber is κ2 .
The fluid pressure p and the axial displacement U are given as follows:
p = ρf ω 2 ψ, (6.2)
2R
Borehole
Formation Formation
r
h Viscous fluid Viscous fluid
Formation Formation
z
Figure 6.1: Borehole intersected by a single horizontal gap representing the fracture.
dψ
U= , (6.3)
dz
where ρf is the fluid density and ω is the angular frequency. We now con-
sider wave propagation and reflection in the borehole. Wave propagation is
described by
1
ψ = A+ e−iκ1 z + A− eiκ1 z for z < − h, (6.4)
2
1 1
ψ = B + e−iκ2 z + B − eiκ2 z for − h < z < h, (6.5)
2 2
1
ψ = C + e−iκ1 z for z > h. (6.6)
2
In the region z < −h/2, where A+ e−iκ1 z represents the incident wave propagat-
ing in the positive z direction and A+ is the incident amplitude coefficient,
A− eiκ1 z is the reflected wave propagating in the negative z direction and A−
is the reflected amplitude coefficient. In the region −h/2 < z < h/2, B + and
B − are the amplitude coefficients for waves propagating in the positive z and
negative z directions, respectively. In the region z > h/2, C + is the amp-
litude coefficient of the transmitted waves. As for the boundary conditions,
at z = h/2 and z = −h/2, the fluid displacement and the pressure should
be continuous. We then obtain the coefficients A− , B + , B − , and C + as a
function of the incident amplitude coefficient term A+ :
A− /A+ = −2i(κ22 − κ21 ) sin(κ2 h)/D, (6.7)
B + /A+ = 2κ1 (κ1 + κ2 )eiκ2 h /D, (6.8)
B − /A+ = 2κ1 (κ2 − κ1 )e−iκ2 h /D, (6.9)
C + /A+ = 4κ1 κ2 eiκ1 h /D, (6.10)
where D is given by
D = (κ1 + κ2 )2 eiκ2 h − (κ1 − κ2 )2 e−iκ2 h . (6.11)
The above equations were also found by Tang and Cheng (1993). In the
upper zone, i.e., the zone above the fracture, the pressure in the borehole is
given by

p̂ = ρf ω 2 A+ e−iκ1 z + (A− /A+ )eiκ1 z , (6.12)
where the ratio A− /A+ is defined by (6.7). Using the boundary condition
that p̂ is p̂0 at z = −d, where d is the distance between the sample top and
the fracture center, yields
−iκ z
e 1 + (A− /A+ )eiκ1 z
p̂ = pˆ0 iκ1 d , (6.13)
[e + (A− /A+ )e−iκ1 d ]
in the upper zone. In the lower zone (below the fracture), we have that
C+ e−iκ1 z
p̂ = pˆ0 , (6.14)
A+ [eiκ1 d + (A− /A+ )e−iκ1 d ]
where the ratio C + /A+ is defined by (6.10).
6.3 Borehole impedance

6.3.1 Formation impedance
In the borehole, the wave propagation is defined by an effective fluid bulk
modulus (Chang et al., 1988):
1 1 1 2
= + + , (6.15)
Keff Kf G iωRZR
6.3. Borehole impedance 69
where Kf is the fluid bulk modulus, G is the shear modulus of the formation,
and ZR is the wall impedance, describing the pressure-velocity ratio at the
borehole wall. It can be expressed as (Chang et al., 1988)
1 k0 K1 (κr R)
= κr R , (6.16)
ZR ηR K0 (κr R)
where k0 is the permeability of the formation, η is the viscosity of the fluid,
and K1 and K0 are modified Bessel functions of first and zeroth order, re-
spectively. In the above equation, the radial wavenumber κr is given by
κ2r = −iω/Dh , (6.17)
for incompressible dynamic (Darcy) fluid motion in a rigid formation. Dh =
k0 Kf /(ηφ) is the hydraulic diffusivity, with φ the porosity of the formation.
As indicated by Kostek et al. (1998b),
for wave propagation in the fracture,
κr can simply be computed from α(ω)ω/cf , where cf is the fluid wave
speed and α(ω) is the dynamic tortuosity as defined by Johnson et al. (1987).
Therefore, in the limiting case for low frequencies, we have that

α(ω)ω iω
lim =− . (6.18)
ω→0 cf Dh
Here we have assumed that the attenuation of the borehole wave is governed
by viscous effects due to the oscillatory radial “breathing” fluid motion, where
the elasticity of the formation is of minor importance. The elasticity, how-
ever, cannot be ignored in the expression for the effective borehole fluid bulk
modulus (6.15). We have to note that this approach [Eqs. (6.15), (6.16), and
(6.17)] is only valid for low frequencies, where the Stoneley wave becomes a
tube wave. The wavenumber and velocity of the tube wave are now easily
given by

ω ρf
κ= =ω . (6.19)
cT Keff
For a borehole radius of 10 cm, a fluid bulk modulus Kf = 2.0 GPa, and
a fluid density is ρf = 1000 kgm−3 , we compute the phase speed and the
attenuation of the tube wave as a function of frequency, for three different
porosity-permeability combinations. For the shear modulus we use G = 3.0
GPa. Results are plotted in Fig. 6.2. Curves “a” refer to a porosity of
16% and a permeability of 10 mD, curves “b” to a porosity of 20% and a
permeability of 100 mD, and curves “c” to the highest values for porosity and
permeability of 26% and 1000 mD, respectively. The dependence on porosity
and permeability is clearly visible. The higher the permeability, the more
1200
1000
a
800
b
600
c
400
200
0 −2 0 2 4 6
10 10 10 10 10
f (Hz)
1
10
0
10
Attenuation (m−1)
−1
10 c
b
−2
10 a
−3
10
−4
10 −2 0 2 4 6
10 10 10 10 10
f (Hz)
Figure 6.2: Phase velocity and attenuation of the tube wave along a porous formation.
a: k0 =10 mD, φ=16%, b: k0 =100 mD, φ=20%, c: k0 =1000 mD, φ=26%.
attenuation is observed due to the increased oscillatory infiltration into the

formation (“breathing”). In the high-frequency limit, the impedance term
tends to zero, so that the effective bulk modulus simply becomes 1/Keff =
1/Kf + 1/G. If also the inverse shear modulus would vanish, we retrieve the
relation Keff = Kf , so that the tube wave velocity equals the fluid wave speed
cf .
6.3. Borehole impedance 71
6.3.2 Fracture impedance

In this case, the wall impedance is defined by averaged parameters over the
fracture aperture h:
< p(R, z) >
ZR = , (6.20)
< v(R, z) >
where p(R, z) is the fluid pressure at the fracture opening and v(R, z) is the
radial velocity of the fluid along the fracture opening. The notation < .. >
denotes averaging over the domain [−h/2 < z < h/2], at r = R. Thus,
the velocity and pressure distribution in the fracture need be known. It was
found by Tang and Cheng (1989) that
vr (r, z) = −H1 (κr r)[Aκr cos(f z) + D f¯ cos(f¯z)], (6.21)
vz (r, z) = H0 (κr r)[−f A sin(f z) + Dκr sin(f¯z)], (6.22)
−iωρf A
p(r, z) = H0 (κr r) cos(f z), (6.23)
1 + 43 iων/c2f
where H0 and H1 are Hankel functions of zeroth order and first order, respect-
ively, and ν = η/ρf is the kinematic viscosity. A and D are dimension-full
arbitrary constants. The modified wavenumbers f and f¯ are expressed as
follows:
ω2
f2 = − κ2r , (6.24)
c2f + 43 iων
ω
f¯2 = − κ2r . (6.25)
iν
The wavenumber κr in the fracture is given by the dispersion equation (Tang
and Cheng, 1989)
h h
κ2r tan( f¯) + f f¯ tan( f ) = 0. (6.26)
2 2
This equation is solved numerically, and the resulting pressure and velocity
field in the fracture is obtained from (6.21) and (6.23). Next, averaging over
the fracture opening is performed and the fracture impedance is calculated
from (6.20):
iκr c2f ρf H0 (κr R)
ZR = . (6.27)
ωH1 (κr R)
1200
a
1000
800
b
600
400
c
200
0 1 2 3 4 5
10 10 10 10 10
f (Hz)
1
10
c
Attenuation (m 1)
0
−
10
b
−1 a
10
−2
10 1 2 3 4 5
10 10 10 10 10
f (Hz)
Figure 6.3: Phase velocity and attenuation of the tube wave along a fracture. Fracture
apertures are a: 1 μm, b: 10 μm, c: 100 μm.
The effective bulk modulus and hence the borehole wavenumber are now
known. For the same borehole configuration as before, the results for the
phase velocity and attenuation are plotted in Fig. 6.3, for identical borehole
parameters as before, and fracture apertures 1 μm, 10 μm and 100 μm, re-
spectively. From the figures, it becomes clear that similar trends are observed
6.4. Experimental setup 73
as for the porous reservoir. The larger the aperture, the larger the attenu-
ation becomes, which corresponds with larger permeability effects. However,
the absolute attenuation values are significantly larger for the fracture case,
at corresponding frequencies. This isbecause the fracture apertures are on
the order of the viscous skin depth 2η/(ρf ω) in these examples. Other,
perhaps more complete theories have been derived by Korneev (2008, 2010),
who relaxes the assumption of wall rigidity in the derivation of the dispersion
equation for fracture waves.
6.4 Experimental setup

The vertical shock tube is shown in Fig. 6.4. It is 7.44 m long and consists
of a high-pressure section and a low-pressure section, separated by a dia-
phragm. The wall thickness of the tube is 2.5 cm. The shock tube was also
used for other borehole measurements in the past (Chao et al., 2004, 2006).
The dimensions of the sections are indicated in Fig. 6.4. A cylinder with a
centralized borehole is mounted in the test section of the shock tube and sat-
urated with water. The length of the cylinder is L. It has a single horizontal
fracture intersecting the borehole. A miniature pressure transducer is moun-
ted in a probe (P2), so that it can be positioned along the axial direction of
61 mm
2.02 m
High pressure
77 mm
section 12.6 mm
77 mm diaphragm
4.32 m
test section 120

wall
mm
water
tube
P1
1.1 m
water fracture
funnel
P2
L
Shock
fracture
probe
L
formation formation
Sample
P2
Figure 6.4: Schematic of the shock tube setup. An acoustic funnel is installed on top of
the sample for borehole wave enhancement (see inset for details).
the borehole. In order to enhance the excitation of borehole waves and sup-
press body wave generation in the sample itself, an acoustic funnel (see inset
of Fig. 6.4) is installed approximately 1 mm above the top of the sample,
consisting of a thick-walled cylinder with decreasing internal cross-sectional
open area in the downward direction.
A wave experiment proceeds as follows: The pressure in the high pressure
section is increased to 1 to 5 bars. Rupture of the diaphragm is caused by
means of an electric current pulse. A shock wave in air is generated which
travels downward and is transmitted into the water layer (see Fig. 6.4). The
wave is partially reflected and partially transmitted into the funnel and the
borehole. The pressure at different positions inside the borehole is meas-
ured by P2. The shock tube wall is equipped with pressure transducer P1,
which is used to trigger the data acquisition system. By repeating the wave
experiments, we can measure the full pressure profiles in the borehole.
6.5 Experiment results

A typical pressure recording of the trigger channel P1 is shown in Fig. 6.5.
At t = 0, the incident pressure wave arrives at P1. Between t = 0.37 and
0.54 ms, there is a gradual pressure increase caused by the combined wave
reflections from within the funnel. The initial pressure decrease at t = 0 is
caused by a precursor mode in the shock tube wall (Van der Grinten et al.,
1987).
In order to determine the effect of the funnel on the input signal, we
measured the pressure development in the funnel by means of sensor P2.
The results are shown in Fig. 6.6. The measured pressure p0 just below the
funnel is used as input signal for the borehole wave computations described
in Section 6.2. Conventional fast Fourier transformation (FFT) is used to
convert the temporal signal p0 to the frequency signal p̂0 .
6.5.1 Borehole fractures

A PVC cylinder in which a borehole was drilled, is cut into two pieces. By
positioning one piece on top of the other but slightly apart, a horizontal slit
between the two pieces is generated which can be varied in aperture by means
of separator poles. The three separator poles are arranged at 120o azimuths
(see Fig. 6.7). The slit aperture can be changed by using different lengths
of the poles. The borehole, fracture, and shock tube are carefully saturated
with water. The properties of sample #3 are in Table 6.1.
6.5. Experiment results 75
2.5
1.5
Pressure (bar)
0.5
−0.5
−1
−0.2 0 0.2 0.4 0.6 0.8
time (ms)
Figure 6.5: Recorded pressure signal P1, which is used to trigger the data acquisition
system. At t = 0, the incident step wave arrives. Between t = 0.37 and 0.54 ms, the
combined reflections from within the funnel cause a gradual pressure increase.
2.5
1.5 a
Pressure (bar)
1 b
0.5
c
0
−0.5
−1
0.2 0.3 0.4 0.5 0.6 0.7 0.8
time (ms)
Figure 6.6: Development of the measured pressure profile in the funnel: 5 mm above the
funnel (a), in the middle of the funnel (b), and 1 mm below the funnel (c).
We start with experiments where h = 1 mm. Probe P2 was consecutively

displaced over 5 mm distances between 329 and 434 mm from the sample top.
In this interval 23 shock wave experiments were carried out. The resulting
76.5
12.2
120o
A 396
Separator
pole
360o
h
B 195
240o
Figure 6.7: Schematic of sample #3. The three separator poles are arranged at 120o
azimuthal separation. The fracture aperture is h. Dimensions are in mm.
microseismogram is shown in Fig. 6.8. The tube wave is clearly visible (St).
The slope of the line St connecting all first arrivals of the tube waves in
Fig. 6.8 corresponds with a tube wave speed of 960±40 m/s. In Fig. 6.8, also
fluid wave mode E1 is indicated which propagates with a speed of 1500 m/s.
The identification of the different wave modes was performed by using the
so-called semblance cross correlation method (Kimball and Marzetta, 1984).
This method picks wave arrivals by computing the scalar semblance in a
time window for a large number of possible arrival times and slownesses. The
maximum values of semblance are interpreted as arrivals and their associated
slownesses are plotted in a slowness-time coherence graph in Fig. 6.9. In
Fig. 6.9, the different colors stand for the different values of the coherence.
The value is from 0 (black) to 100 (white), which means that in the black
area there is no coherence and in the white area the coherence is maximal.
Table 6.1: Properties of the samples.
L φ k0 Kb G ρb ρs
Sample
(mm) (%) (D) (GPa) (GPa) (kg/m3 ) (kg/m3 )
#3 591+h 0 0 7.8 1.7 1427 1427
#4 348 47.6 15 ± 2 7.1 3.0 1310 2495
Probe position from fracture centre (mm)

Probe position from sample top (mm)
E1 St
329 −67.5
339 −57.5
349 −47.5
359 −37.5
369 −27.5
379 −17.5
389 −7.5
396.5 0
404 7.5
414 17.5
424 27.5
434 37.5
0.2 0.4 0.6 0.8 1 1.2

time (ms)
Figure 6.8: Microseismogram comprising 23 shock wave experiments in sample #3. The
fracture center is at 396.5 mm from the sample top. The fracture aperture is 1 mm. The
line St connects first arrivals of the Stoneley wave. E1 is the fluid wave arriving earlier
than the Stoneley wave and having a velocity of 1500 m/s.
In Fig. 6.9, there is also a coherence plot maximum E2 that cannot clearly
be distinguished in the microseismogram (Fig. 6.8).
Two selected pressure recordings are compared with theory in Figs. 6.10.
In Fig. 6.10(a), the position of the transducer is 67.5 mm above the fracture
center. In Fig. 6.10(b), the position of the transducer is 27.5 mm below the
fracture center. In Fig. 6.10(a), the agreement between experiment and the-
ory is very good. The amplitude of the first peak perfectly matches theory.
Note that the theoretical result stems from (6.13) and (6.14), followed by a
standard inverse FFT routine to convert the signals back to the temporal
domain. The input parameters for the theory were obtained from independ-
ent laboratory experiments and no data fitting procedure was applied. The
pressure trough A is not measured. It is associated with precursor tube mode
(Van der Grinten et al., 1987) in the input signal p0 (Fig. 6.6). In Fig. 6.10(b),
where we compare theory and experiment at some distance below the frac-
ture, the agreement between experiment and theory is even better. We note
that in Figs. 6.10(a) and 6.10(b) event E1 does not appear in the theoretical
prediction. This is because E1 is associated with a fluid bulk mode that is
not part of the theoretical description given in Section 6.2.
Next, all amplitudes of the tube waves are compared with theory in
Figs. 6.11. These amplitudes are the maximum pressure values in all 23
5 90
4.5 80
4 70
Slowness (s/km)
3.5
60
3
50
2.5
40
2
St 30
1.5 E1 E2 20
1
0.5 10
0.35 0.4 0.45 0.5 0.55 0.6

time (ms)
Figure 6.9: Coherence plot for wave identification within sample #3. Event E1 is the
fluid wave arriving earlier than the Stoneley wave (St) and having a velocity of 1500 m/s.
E2 is a second fluid wave event that cannot clearly be identified in the microseismogram.
snapshots. In both Figs. 6.11(a) and 6.11(b), the vertical lines represent
the fracture position. We note that there is a strong decrease in amplitude
caused by the presence of the fracture. Surprisingly, the decrease in amplitude
starts somewhat earlier than where the fracture is located. In Fig. 6.11(b),
the modeled amplitude also predicts this sharp decrease over the fracture
very well. The pressure level before pressure decay (1.8 bar) is in agreement
with theory. The pressure level after pressure decay is around 1.2 bar for the
experiment and around 1.4 bar for the modeling. This means that the tube
wave loses somewhat more energy over the fracture in the experiment than
predicted by theory.
Next, new separator poles were used to obtain a fracture aperture of 5
mm. The center of the fracture is now at 398.5 mm from the sample top.
Again, 23 shock wave experiments were carried out. The resulting micro-
seismogram is shown in Fig. 6.12. Apart from the arriving tube wave, also
a reflection from the fracture can now be distinguished. The line St again
connects all first arrivals of the tube wave. The velocity is determined to
be 960±40 m/s, which is in agreement with the previous velocity measure-
ments. The dotted horizontal line indicates the position of the fracture. The
slowness-time coherence is plotted in Figs. 6.13. In Fig. 6.13(a), besides the
arrival of the Stoneley wave, again event E1 can be identified. The reflected
2
Experiment
1.5 Modeling
Pressure (bar)
1
0.5 E1
−0.5 A
−1
0 0.2 0.4 0.6 0.8
time (ms)
(a) z=-67.5 mm
2
Experiment
Modeling
1.5
Pressure (bar)
E1
0.5
−0.5 A
−1
0 0.2 0.4 0.6 0.8
time (ms)
(b) z=27.5 mm
Figure 6.10: Experimental and modeled pressure signals in sample #3 with 1 mm fracture
aperture at 67.5 mm above the fracture center (a), and 27.5 mm below the fracture center
(b). E1 is fluid wave arriving earlier than the Stoneley wave (St) and having a velocity of
1500 m/s. The precursor mode A is also visible.
Stoneley wave is identified in the slowness-time coherence Fig. 6.13(b).

Again, two snapshots are compared with theory in Figs. 6.14. In Fig.
1.8
Pressure (bar)
1.6
1.4
1.2
1
320 340 360 380 400 420 440
Depth from the sample top (mm)
(a) Experiment
1.9
1.8
Pressure (bar)
1.7
1.6
1.5
1.4
1.3
320 340 360 380 400 420 440
(b) Theory
Figure 6.11: Amplitude of the Stoneley wave at different positions in the borehole of
sample #3 with 1 mm fracture. The dashed vertical line is the center of the fracture and
the two solid lines are the borders of the fracture.
6.14(a), the position of the transducer is 72.5 mm above the fracture center.
In Fig. 6.14(b), the position of the transducer is 27.5 mm below the fracture

E1 St
RSt
331 −67.5
341 −57.5
351 −47.5
361 −37.5
371 −27.5
381 −17.5
391 −7.5
398.5 0
406 7.5
416 17.5
426 27.5
0.2 0.4 0.6 0.8 1 1.2

time (ms)
line St connects first arrivals of the Stoneley wave; the line RSt connects the inflection
points representing the reflected Stoneley wave. E1 is the fluid wave arriving earlier than
the Stoneley wave and having a velocity of 1500 m/s.
5 90 5 90
4.5 80 4.5 80
4 70 4 70
Slowness (s/km)
Slowness (s/km)
3.5 3.5
60 60
3 3
50 50
2.5 2.5
2
40
2
RSt 40
E1 St 30 30
1.5 1.5
1 20 20
1
0.5 10 0.5 10
0.35 0.4 0.45 0.5 0.55 0.6 0.4 0.45 0.5 0.55 0.6 0.65
time (ms) time (ms)
(a) (b)
Figure 6.13: Coherence plots for wave transmission (a) and wave reflection (b) within
sample #3. The fracture aperture is 5 mm. Events E1, St, and Rst can clearly be identified.
center. In Fig. 6.14(a), the amplitude of the first peak is perfectly predicted
by theory, and also the gradual oscillatory pressure decrease, albeit some
time lag between predicted and recorded peaks and troughs. Note that also
reflectivity from the fracture is included in the theory (see Section 6.2). In
Fig. 6.14(b), where we compare theory and experiment at some distance
2
Experiment
Modeling
1.5
Pressure (bar)
0.5 E1
−0.5 A
−1
0 0.2 0.4 0.6 0.8
time (ms)
(a) z=-72.5 mm
2
Experiment
1.5 Modeling
Pressure (bar)
1
E1
0.5
−0.5
A
−1
0 0.2 0.4 0.6 0.8
time (ms)
(b) z=27.5 mm
(b). E1 is the fluid wave, and A is the precursor pressure trough.
below the fracture, the agreement between experiment and theory is also
good. Again, event E1 does not appear in the theoretical prediction because
1.8
Pressure (bar) 1.6
1.4
1.2
0.8
0.6
0.4
320 340 360 380 400 420 440
(a) Experiment
2.2
1.8
Pressure (bar)
1.6
1.4
1.2
0.8
320 340 360 380 400 420 440

(b) Theory
E1 is associated with a bulk water mode that is not part of the theoretical
description given in Section 6.2. It can be seen in Fig. 6.14(b) that the first
Probe position from sample top (mm) 18

E1
St
28
38
48
58
68
78
88
98
108
118
0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6

time (ms)
line St connects the first arrivals of the Stoneley wave. E1 is the fluid wave traveling with
a speed of 1500 m/s.
peak of the Stoneley wave is slightly overpredicted by theory. By comparing

Fig. 6.14(b) with Fig. 6.10(b), we find that the Stoneley wave amplitude
decreases much more over the 5 mm fracture than over the 1 mm fracture.
The amplitudes of the tube waves are compared with theory in Figs.
6.15. For 23 traces, the maximum amplitude was determined and plotted
in Figs. 6.15. We note that there is a strong decrease in amplitude caused
by the presence of the 5 mm fracture. This decrease is larger than for the
1 mm fracture case. The pressure level before pressure decay is perfectly
predicted by theory again. The pressure level after pressure decay is around
0.5 bar for the experiment and around 0.7 bar for the modeling. This again
shows that the tube wave loses somewhat more energy over the fracture in
the experiment than predicted by theory.
6.5.2 Porous sample

Next, we use a porous sample #4 for tube wave attenuation measurements.
No fracture is present here. The length of the sample is 348 mm, the bore-
hole diameter is 12.5 mm, and the outer sample diameter is 76.8 mm. The
properties of sample #4 are given in Table 6.1. Probe P2 was consecutively
displaced over 5 mm distances from 18 to 123 mm from the sample top. In
this interval 22 shock wave experiments were carried out. The resulting mi-
4 80
3.5 70
Slowness (s/km) 3 60
2.5 50
2 40
E1 St
1.5 30
1 20
0.5 10
0.25 0.3 0.35

time (ms)
Figure 6.17: Coherence plot for wave identification in sample #4. Both Events E1 and
St are clearly identified.
0.8
0.6
a
Pressure (bar)
0.4
b
0.2
c
−0.2
−0.2 0 0.2 0.4 0.6 0.8

time (ms)
Figure 6.18: Experimental and modeled pressure signals in sample #4 at 48 mm from

the sample top. Three different permeabilities 0.5 D (a), 5 D (b) and 50 D (c) are used for
the modeling.
croseismogram is shown in Fig. 6.16. The slope of the line St connecting all
first arrivals of the tube waves in Fig. 6.16 corresponds with a wave speed of
905±40 m/s. The slowness-time coherence is given in Fig. 6.17. In Fig. 6.17,
besides the arrival of the Stoneley wave, again fluid wave event E1 can be
distinguished.
In Fig. 6.18, one pressure snapshot at 48 mm from the sample top is
plotted. In the plot, the results are compared with theory using three different
permeabilities: 0.5 Darcy (D), 5 D, and 50 D. It can be seen from Fig. 6.18
that permeability around 5 D would accurately predict the measured pressure
curve. The permeability was also determined in an independent falling head
laboratory experiment, from which the permeability was actually found to
be 15 ± 2 D. The discrepancy between the effective permeability of 5 D and
the actual permeability of 15 D can be attributed to the fact that in the
theory so far only low-frequency viscous effects are incorporated, whereas
also high-frequency tortuosity effects need be taken into account. Moreover,
due to long residence times of the sample in the water-filled shock tube, we
measured that fouling of the borehole wall decreased permeability over time.
We thus argue that the permeability from the separate falling head test was
probably too high.
6.6 Conclusions
Tube waves are strongly affected by fractures intersecting the borehole. A
theoretical description for both porous samples and fracture zones is given
based on the introduction of an effective borehole fluid bulk modulus. This
effective fluid bulk modulus contributes to the wave attenuation through the
borehole wall impedance. This impedance can be calculated for both porous
and fracture zones adjacent to the borehole, thus predicting borehole wave
attenuation, transmission and reflection over such zones. Our shock tube
setup generates borehole tube waves that are used for porous and fracture
zone characterization. We use sample #3 to introduce and vary fractures
in a cylindrical sample. Shock wave experiments show that attenuation in
boreholes adjacent to porous zones can be predicted by theory, although
the permeability fit still has a significant discrepancy. The reflection and
transmission of borehole tube wave over 1 and 5 mm fractures are correctly
predicted by theory, thus showing the potential of borehole wave experiments
for fracture detection and characterization.
6.7. Fractured porous sample 87

5.3 −27.5
15.3 −17.5
25.3 −7.5
32.8 0
40.3 7.5
50.3 17.5
60.3 27.5
70.3 37.5
80.3 47.5
90.3 57.5
100.3 67.5
110.3 77.5
0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65
time(ms)
fracture center is at 32.8 mm from the sample top. The fracture aperture is 5.3 mm. For
comparison see Fig. 6.16.
5 90 5
4.5 80 4.5 80
4 70 4 70
Slowness (s/km)
Slowness (s/km)
3.5 3.5
60 60
3 3
50 50
2.5 2.5
40 40
2 2
30 30
1.5 1.5
20 20
1 1
0.5 10 0.5 10
0.3 0.35 0.4 0.45 0.3 0.35 0.4 0.45

time (ms) time (ms)
(a) (b)
sample #5. The fracture aperture is 5.3 mm. Events E1, St, and Rst can not be identified
(for comparison see Fig. 6.17).
6.7 Fractured porous sample

Next, sample #4 was cut into two pieces that were kept 5.3 mm apart by
means of three steel poles to form sample #5. The center of the fracture is
now at 32.8 mm from the sample top. Again, 23 wave experiments were per-
formed. The resulting microseismogram is shown in Fig. 6.19. The coherence
plots are given in Figs. 6.20 for the incident and reflected waves, respectively.
Note that the original coherence method can only identify downgoing waves.
For reflected upgoing waves, the interpretation method was modified so that
also these wave types could be captured. In the coherence plots, no clear
identification of the Stoneley wave could be made, although in Fig. 6.19
wave-type behavior is clearly visible, albeit only in the downgoing direction.
We also notice that the downgoing disturbance is strongly attenuated, and
that there is a strong agreement with the recordings in Fig. 6.16. The fact
that the Stoneley waves have disappeared from the coherence plots is prob-
ably caused by the relatively long residence time in the shock tube which
has caused sample degradation and fouling of the pores. Unfortunately, also
in Fig. 6.19, no upgoing reflected waves can be traced. The presence of the
fracture obviously causes attenuation of the transmitted borehole waves, as
can be seen from the fact that not even the water wave is recognized any
more in the coherence plots.
Chapter 7
Fracture effects in mandrel sample
We use a mandrel sample #2 to introduce fractures in a cylindrical sample,

and to vary fracture widths. It consists of a mandrel and a host cylinder,
in which a 12.5 mm diameter borehole is drilled. The fracture width is
controlled by shim stock holding the two parts apart. As compared with
the experiments in Chapter 5, an acoustic funnel was installed on top of the
sample (see also Chapter 6), so that also the diameter of the borehole was
somewhat larger. The new configuration is given in Fig. 7.1. The borehole,
fracture, and shock tube are carefully saturated with water as indicated in
76.5
60
12.5
Fracture
width
400
200
Shim
stock
h
10
Figure 7.1: Schematic of sample #2.
89
90 7. Fracture effects in mandrel sample
Table 7.1: Physical properties of sample #2.
φ (%) k0 (mD) Kb (GPa) G (GPa) ρ (kg/m3 )

0 0 7.9 1.7 1426
Fig. 6.4. The physical properties of sample #2 are shown in Table 7.1.
Probe P2 is used to measure the borehole tube waves. It was translated
axially from 160 to 270 mm from the sample top. In this interval 23 shock
wave experiments were carried out. We started with experiments where all
shim stock was removed so that the fracture was fully closed. The resulting
microseismogram is shown in Fig. 7.2. The slope of the line St connecting
all first arrivals of the tube waves in Fig. 7.2 corresponds with a tube wave
speed of 900±40 m/s. The identification of the different wave modes was
performed by using the semblance cross correlation method (Kimball and
Marzetta, 1984). This method picks wave arrivals by computing the scalar
semblance in a time window for a large number of possible arrival times and
slownesses. The maximum values of semblance are interpreted as arrivals and
their associated slownesses are plotted in a slowness-time coherence graph in
Fig. 7.3. In Fig. 7.3, the different colors stand for the different values of the
coherence. The value is from 0 (black) to 100 (white), which means that in
the black area there is no coherence and in the white area the coherence is
St
160
170
180
190
200
210
220
230
240
250
260
270
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
time (ms)
fracture is closed. The line St connects first arrivals of the Stoneley wave.
91
5 90
4.5 80
4
Slowness (s/km) 70
3.5
3 60
2.5 50
2
St 40
1.5
30
1
20
0.5
0.4 0.45 0.5 0.55 0.6
time (ms)
Figure 7.3: Coherence plot in sample #2. The fracture is closed.

St RSt
135 −67.5
145 −57.5
155 −47.5
165 −37.5
175 −27.5
185 −17.5
195 −7.5
202.5 0
210 7.5
220 17.5
230 27.5
E1
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1

time(ms)
line St connects first arrivals of the Stoneley wave; the line RSt represents the reflected
Stoneley wave. E1 is the fluid wave arriving earlier than the Stoneley wave and having a
velocity of 1500 m/s.
maximum.
Next, shim stock was added to obtain a fracture aperture of 5 mm. The
5 90 5 90
4.5 80 4.5 80
4 70 4 70
Slowness (s/km)
Slowness (s/km)
3.5 3.5
60 60
3 3
50 50
2.5 2.5
40 40
2 2 RSt
30 30
1.5 St 1.5
E1 20 20
1 1
10 0.5 10
0.5
0.35 0.4 0.45 0.5 0.55 0.45 0.5 0.55 0.6 0.65
time (ms) time (ms)
(a) (b)
sample #2. The fracture aperture is 5 mm. Events E1, St, and RSt can clearly be identified.
center of the fracture is now at 202.5 mm from the sample top (see Fig. 7.1).
Again, 23 shock wave experiments were carried out with probe P2 each time
at a different axial position along the borehole. The resulting microseismo-
gram is shown in Fig. 7.4. Apart from the arriving tube wave, now also
a reflection from the fracture can be distinguished. The velocity is now
determined to be 921±40 m/s, which is in agreement with the previous ve-
locity measurements, within experimental accuracy. The dotted horizontal
line indicates the position of the fracture. The maximum values of semblance
are interpreted as arrivals and their associated slownesses are plotted in the
slowness-time coherence graph in Fig. 7.5. The incident Stoneley wave in
Fig. 7.5(a) and the reflected Stoneley wave from the fracture in Fig. 7.5(b)
have the same speed which is identical to the Stoneley wave speed in Fig. 7.4,
within experimental accuracy.
Two selected pressure recordings are compared with theory in Figs. 7.6.
In Fig. 7.6(a), the position of the transducer is 62.5 mm above the fracture
center. In Fig. 7.6(b), the position of the transducer is 27.5 mm below the
fracture center. In Fig. 7.6(a), the agreement between experiment and the-
ory is very good. The amplitude of the first peak perfectly matches theory.
Note that the theoretical result stems from (6.13) and (6.14), followed by a
standard inverse FFT routine to convert the signals back to the temporal
domain. Also reflectivity from the fracture is included in the theory (see
Section 6.2). The pressure trough A is not measured. It is associated with
precursor tube mode (Van der Grinten et al., 1987) in the input signal p0
(Fig 6.6). In Fig. 7.6(b), where we compare theory and experiment at some
distance below the fracture, the agreement between experiment and theory
93
2
Experiment
Modeling
1.5
Pressure (bar)
1
0.5
−0.5 A
−1
0 0.2 0.4 0.6 0.8
time (ms)
(a) z=-62.5 mm
2
Experiment
Modeling
1.5
Pressure (bar)
1
E1
0.5
−0.5
A
−1
0 0.2 0.4 0.6 0.8
time (ms)
(b) z=27.5 mm
(b). E1 is the fluid wave, and A is the precursor pressure trough.
is less good. Event E1 does not appear in the theoretical prediction because
E1 is associated with a bulk water mode that is not part of the theoretical
1.8
1.6
Pressure (bar)
1.4
1.2
0.8
0.6
0.4
120 140 160 180 200 220 240
(a) Experiment
2.2
1.8
Pressure (bar)
1.6
1.4
1.2
0.8
120 140 160 180 200 220 240

(b) Theory
description given in Section 6.2. It can be seen in Fig. 7.6(b) that the first
peak of the Stoneley wave is slightly overpredicted by theory, and the exper-
95
imental signal remains at a relatively high plateau level as compared with

theory.
The amplitudes of the tube waves are compared with theory in Figs. 7.7.
These amplitudes are obtained from pressure-time traces. For 23 traces, the
maximum amplitude was determined and plotted in Figs. 7.7. In both panels,
the vertical lines represent the fracture position. We note that there is a
strong step-wise decrease in amplitude caused by induced wave reflection and
reduced transmittivity. The decrease in amplitude starts somewhat earlier
than where the fracture is located. In the lower panel of Figs. 7.7, the modeled
amplitude also predicts this sharp decrease over the fracture very well. The
pressure levels before pressure decay (around 1.8 bar), are perfectly predicted
by theory. The pressure levels after pressure decay, however, are somewhat
overestimated by theory.
Chapter 8
Conclusions
Stoneley waves are strongly affected by fractures intersecting the borehole. A

theoretical description for both permeable and fracture zones is given based
on the introduction of a complex-valued frequency-dependent effective bore-
hole fluid bulk modulus specifying the wellbore fluid’s reaction to pressure
variations induced by sonic logging tools. The effective fluid bulk modulus
comprises the intrinsic fluid stiffness, the borehole wall distensibility, and the
radial fluid seepage into the adjacent (horizontal) permeable zones. The lat-
ter two effects tend to diminish the intrinsic fluid’s stiffness, giving rise to
a lower effective bulk modulus amplitude and thus to a lower wave speed in
the borehole. The radial oscillatory fluid seepage causes viscous friction in
the adjacent zones and results in a phase lag between the pressure increase
and the compression of the borehole fluid, leading to attenuation of the bore-
hole waves. This seepage effect is expressed in terms of a so-called borehole
dynamic wall impedance specifying the radial fluid velocity at the borehole
wall as a function of the borehole pressure variations. These combined effects
of wall distensibility and fluid seepage lead to a vertical stacking of borehole
fluids, where each layer in the borehole has its effective compressibility de-
termined by the permeable zone in the adjacent formation. If a borehole wave
travels down from one of these fluid layers into the next, it will encounter
an impedance contrast causing it to partially reflect and partially transmit,
thus revealing the presence of permeable and/or fracture zones adjacent to
the borehole.
The dynamic wall impedance is investigated at low frequencies, when
only steady-state Stokes’ flow in the permeable formation zones is important.
For higher wave frequencies, Stokes’ flow can no longer describe the flow
behaviour, as also inertia effect become important. In our evaluations, the
concepts of dynamic permeability and dynamic tortuosity are used to model
97
98 8. Conclusions
the dynamic wall impedance over the full frequency domain. Exact solutions
for slit flow confirm that fractures have a frequency-dependent permeability
indeed.
Straightforward computation on a single horizontal fracture gives a quant-
itative prediction of velocity, attenuation and reflection of borehole Stoneley
waves. Fracture aperture significantly affects both reflection and transmission
coefficients. Stoneley wave propagation in porous and fractured formations
is studied experimentally by means of a vertical shock tube facility. In this
set-up, shock wave in air are generated that travel downwards into a water-
saturated cylindrical rock sample that has a borehole drilled along the center
axis. In this way, high-energy borehole waves can be generated with excel-
lent repeatability. A logging probe is installed in the borehole to measure the
pressure profiles. We show that our shock tube facility can generate borehole
Stoneley waves in a broad frequency band under conditioned circumstances.
The wave can be analyzed for velocity and attenuation determination. Re-
flection from the water-sample interface and from the free water interface
can be recorded by means of a fixed pressure transducer mounted in the wall
of the shock tube above the sample in the water layer. The fractures in the
formation are manufactured by means of composite cylinders whose upper
and lower parts are separated by small spacer poles so that a variable ho-
rizontal fracture (slit) aperture can be obtained. Obviously these fractures
form an open connection between the borehole fluid and the fluid outside the
cylinder. Also mandrel samples are used for horizontal slits that are not open
to the fluid outside the cylinder, thus representing fractures with finite radial
extension. Wave experiments on PVC samples show that varying fracture
widths significantly alter the recorded Stoneley wave pressure signal at fixed
depth. The reflection and transmission of borehole tube waves over 1 and
5 mm fractures are correctly predicted by theory, thus showing the poten-
tial of borehole wave experiments for fracture detection and characterization.
Other wave experiments show that attenuation in boreholes adjacent to por-
ous zones can be predicted by theory, although the permeability fit still has
a significant discrepancy. The technique is easily extensible for detecting the
presence of multiple fractures in rock samples, and has a potential for per-
meability prediction of fractured porous samples. This research has already
successfully been applied for fracture evaluation of carbonate samples.
Appendix A
Derivation of the plane fracture

wave equation
From a combination of boundary conditions for a rigid plane fracture, the

secular equation Ax = 0 is obtained, where A is given by
⎡ ⎤
kcos( h2 f ) ksin( h2 f ) −f¯sin( h2 f¯) f¯cos( h2 f¯)
⎢ −f sin( h f ) f cos( h f ) kcos( h2 f¯) ksin( h2 f¯) ⎥
A=⎢ 2 2
⎣ kcos( f ) −ksin( f ) f¯sin( f¯)
⎥,
h
2
h
2
h
2 f¯cos( h2 f¯) ⎦
f sin( h2 f ) f cos( h2 f ) kcos( h2 f¯) −ksin( h2 f¯)
so that
kcos( h2 f ) ksin( h2 f ) −f¯sin( h2 f¯) f¯cos( h2 f¯)
−f sin( h2 f ) f cos( h2 f ) kcos( h2 f¯) ksin( h2 f¯)
|A| = ,
0 −2ksin( 2 f ) 2f¯sin( h2 f¯)
h
0
0 2f cos( h2 f ) 2kcos( h2 f¯) 0
which can be elaborated as follows:
kcos( h2 f ) −f¯sin( h2 f¯) f¯cos( h2 f¯)
h
|A| = 2ksin( f ) −f sin( h2 f ) kcos( h2 f¯) ksin( h2 f¯)
2
0 2kcos( h2 f¯) 0
kcos( h2 f ) ksin( h2 f ) f¯cos( h2 f¯)

¯ h¯
+2f sin( f ) −f sin( 2 f ) f cos( h2 f ) ksin( h2 f¯)
h
2
0 2f cos( h2 f ) 0
! "
h h¯ kcos( h2 f ) f¯cos( h2 f¯)
= 2ksin( f ) −2kcos( f )
2 2 −f sin( h2 f ) ksin( h2 f¯)
99
100 A. Derivation of the plane fracture wave equation
! "
¯ h¯ h kcos( h2 f ) f¯cos( h2 f¯)
+2f sin( f ) −2f cos( f )
2 2 −f sin( h2 f ) ksin( h2 f¯)
kcos( h2 f ) f¯cos( h2 f¯) h h

= h ¯ [−4k2 sin( f )cos( f¯)−
−f sin( 2 f ) ksin( 2 f )
h
2 2
h h
4f f¯sin( f¯)cos( f )] = 0.
2 2
From the above equations, we find that
! 2
k sin( h2 f )cos( h2 f¯) + f fsin(
¯ h ¯ h
2 f )cos( 2 f ) = 0
k 2 cos( h2 f )sin( h2 f¯) + f f¯sin( h2 f )cos( h2 f¯) = 0,
so that
!
k 2 tan( h2 f ) + f f¯tan( h2 f¯) = 0
k 2 tan( h2 f¯) + f f¯tan( h2 f ) = 0.
As the first equation does not have zeros in the complex plane, the second
characteristic relation remains to be solved.
Appendix B
Coherence method
An artificial signal consisting of two wave types was generated to test the
coherence method. The frequency of the first wave is 50 kHz, that of the
second wave is 20 kHz. We loaded 10 traces into the program. The distance
between two traces is 5 cm (see Fig. B.1). The velocity of the first wave is
2500 m/s, that of the second one is 500 m/s. The identification of the different
wave modes was performed by using the so-called semblance cross correlation
method (Kimball and Marzetta, 1984). This method picks wave arrivals by
computing the scalar semblance in a time window for a large number of
Probe position from sample top (cm)
0
5
10
15
20
25
30
35
40
45 First Wave Second Wave
0 0.5 1 1.5 2
time (ms)
Figure B.1: Microseismogram comprising 10 traces of artificial signal. The velocity of

the first wave is 2500 m/s, that of the second wave is 500 m/s.
101
102 B. Coherence method
90
80
2
70
Second Wave
Slowness (s/km)
60
1.5
50
40
1 30
20
First Wave
0.5 10
0 0.05 0.1 0.15 0.2 0.25 0.3

time (ms)
Figure B.2: Coherence plot of the artificial signal. The two waves are clearly identified
by the program.
possible arrival times and slownesses. The maximum values of semblance

are interpreted as arrivals and their associated slownesses are plotted in a
slowness-time coherence graph in Fig. B.2. In Fig. B.2, the different colors
stand for the different values of the coherence. The value is from 0 (black)
to 100 (white), which means that in the black area there is no coherence
and in the white area the coherence is maximal. Two constant velocities are
displayed in the plot corresponding with the two waves.
Appendix C
Porosity and permeability

determination
C.1 Porosity determination

In the wave experiments, porous samples #4 and #5 were used. They stem
from one long cylindrical core sample in which a borehole was drilled. We use
a segment of this sample for the determination of porosity and permeability.
The radius of the borehole is 6.25 mm, the outer radius of this sample is 38.4
Table C.1: Experimental parameters for permeability determination.
L At Asample η ρ g
(mm) (m2 ) (m2 ) (Pa · s) (kg/m3 ) (m/s2 )
302 3.22 · 10−3 4.51 · 10−3 10−3 998 9.8
mm. The length of the sample is 302 mm. We first dry the sample in an
oven. The dry weight of the sample was found to be 1779 g. Next, the sample
was carefully saturated with water and the buoyancy weight was measured
to be 1067.5 g. The porosity of the sample can now be obtained from
Mdry − Mbuoy (1779 − 1067.5) · 10−3
φ = 1− = 1− = 0.47±0.01 (C.1)
V · ρwater 1361.3 · 10−6 · 998
C.2 Permeability determination

The schematic of the setup for the permeability test is shown in Fig. C.1.
The permeability is determined by means of a so-called falling head test,
where the saturated porous cylinder is loaded with a water column on top.
103
104 C. Porosity and permeability determination
Air tube
Water tube
h0 h
Pressure gauge
Flexible rubber hose
L
Porous sample
Water tank
Figure C.1: Schematic of the permeability measurement setup.
The water head is measured as a function of time as the water flows through
the sample by means of gravity forces.
Obviously water must not escape from the side walls of the sample. There-
fore, the inner and outer walls are sealed with rubber hoses. The inner hose
(in the borehole) is pushed against the side wall by means of a small over-
pressure through an air tube (see Fig. C.1). The original water height in the
water tube h0 = 140 cm. After time t the water height reaches h. A pressure
gauge (Endress+Hauser Deltabar) is connected to the top of the sample to
determine the water height as a function of time. As the cross-sectional area
of the air tube is very small compared with that of the water tube, the flow
rate can be obtained from
At · dh
Q= , (C.2)
dt
where At is the cross-sectional area of the water tube. We have that
k0 (h + L)ρg
Q=− · · Asample , (C.3)
η L
where Asample is the cross-sectional area of the sample, k0 is the permeability
of the sample, ρ and η are the density and viscosity of the water, respectively.
Their values can be found in Table C.1. Combination of (C.2) and (C.3) gives:
1 −k0 ρgAsample
· dh = · dt. (C.4)
h+L LηAt
C.2. Permeability determination 105
−8
x 10
3.5
2.5
H (m2⋅ s)
1.5
0.5
0
0 500 1000 1500 2000 2500
time (s)
Figure C.2: Reduced water height recorded as a function of time. The slope is a measure
for permeability.
Integration gives that

1 h+L
− · ln = k0 · t, (C.5)
α h0 + L
where
ρgAsample
α= . (C.6)
LηAt
We recorded a series of water heights h and the corresponding time t. We
plot H = −α−1 ln hh+L
0 +L
as a function of time (see Fig. C.2). From Fig. (C.2),
we can easily obtain the permeability of the sample to be 15 ± 2 D.
106 C. Porosity and permeability determination
Bibliography
Adler, L., and P.B. Nagy (1994). Measurements of acoustic surface waves on
fluid–filled porous rocks. J. Geophys. Res., 99, 17863–17869.
Albers, B. (2006). Monochromatic surface waves at the interface between

poroelastic and fluid halfspaces. Proc. R. Soc. A, 462, 701-723.
Allard, J.F., M. Henry, C. Glorieux, W. Lauriks, and S. Petillon (2004). Laser

induced surface modes at water-elastic and poroelastic solid interfaces. J.
Appl. Phys., 95, 528–535.
Berryman, J.G. (1980). Confirmation of Biots theory. Appl. Phys. Lett., 37,
382-384.
Berryman, J.G., and H.F. Wang (1995). The elastic coefficients of double-
porosity models for fluid transport in jointed rock. J. Geophys. Res., 100,
24611-24627.
Berryman, J.G., and H.F.Wang (2000). Elastic wave propagation and atten-
uation in a double-porosity dual-permeability medium. Int. J. Rock Mech.
Min. Sci., 37, 63-78.
Beydoun, W.B., C.H. Cheng, and M.N. Toksöz (1985). Detection of open
fractures with vertical seismic profiling. J. Geophys. Res., 90, 4557–4566.
Biot, M.A. (1955). Theory of elasticity and consolidation for a porous aniso-
tropic solid. J. Appl. Phys., 26, 182–185.
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.
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.
107
108 BIBLIOGRAPHY
Biot, M.A., and D.G. Willis (1957). The elastic coefficients of the theory of
consolidation. J. Appl. Mech., 24, 594–601.
Brekhovskikh, L.M. (1980). Waves in layered media. Academic Press, New

York.
Brie, A., T. Endo, D.L. Johnson, and F. Pampuri (2000). Quantitative Form-
ation Permeability Evaluation from Stoneley Waves. SPE Reservoir Eval.
& Eng., 3, 109–117.
Brown, P., M. Batzle, M. Peeters, S. Dey-Sakar, and G. Steensma (2000).

Shock tube experiments and the observation of the Biot slow wave in nat-
ural rocks. SEG Expanded Abstracts, 1846–1849.
Chang, S.K., H.L. Liu, and D.L. Johnson (1988). Low-frequency tube waves
in permeable rocks. Geophysics, 53, 519–527.
Chao, G., D.M.J. Smeulders, and M.E.H. van Dongen (2004). Shock-induced
borehole waves in porous formations: Theory and experiments. J. Acoust.
Soc. Am., 116, 693-702.
Chao, G. (2005). Dispersive surface acoustic waves in poroelastic media.

Ph.D. thesis, Delft University of Technology.
Chao, G., D.M.J. Smeulders, and M.E.H. van Dongen (2006). Measurements
of shock-induced guided and surface acoustic waves along boreholes in
poroelastic materials. J. Appl. Phys., 99, 094904.
Cheng, C.H., and M.N. Toksöz (1981). Elastic wave propagation in a fluid-
filled borehole and synthetic acoustic logs. Geophysics, 46, 1042–1053.
Cheng, C.H., J. Zhang, and D.R. Burns (1987). Effects of in-situ permeability
on the propagation of Stoneley (tube) waves in a borehole. Geophysics, 52,
1279–1289.
Chilingarian, G.V., S.J. Mazzullo, and H.H. Rieke (1992). Carbonate Reser-
voir Characterization: A Geologic-Engineering Analysis, Part I. Elsevier,
Amsterdam, 1–35.
Cicerone, R.D., and M.N. Toksöz (1990). A dynamic model of tube wave
generation at fracture in hydrophone VSP data. Reservoir delineation –
Vertical Seismic Profiling Consortium, Annual Report.
Coutant, O. (1989). Numerical study of the diffraction of elastic waves by

fluid-filled cracks. J. Geophys. Res., 94, 17805–17818.
BIBLIOGRAPHY 109
De Josselin de Jong, C. (1956). What happens in soil during pile driving?

De Ingenieur, 68, B77–B88.
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. Seismol. 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. Seismol.
Soc. Am., 51, 51–59.
Deresiewicz, H. (1962). The effect of boundaries on wave propagation in a

liquid-filled porous solid: IV. Surface waves in a half-space. Bull. Seismol.
Soc. Am., 52, 627-638.
Deresiewicz, H., and R. Skalak (1963). On uniqueness in dynamic poroelasti-

city. Bull. Seismol. Soc. Am., 53, 783–788.
Edelman, I., and K. Wilmanski (2002). Asymptotic analysis of surface waves

at vacuum/porous medium and liquid/porous medium interfaces. Con-
tinuum Mech. Thermodyn., 14, 25–44.
Fan, H., and D.M.J. Smeulders (2012). Shock-induced borehole waves and
fracture effects. Transp Porous Med, 93, 263–270.
Fan, H., and D.M.J. Smeulders (2013). Shock-induced wave propagation over
porous and fractured borehole zones: Theory and experiments. J. Acoust.
Soc. Am., 134, 4792–4800.
Feng, S., and D.L. Johnson (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 D.L. Johnson (1983b). High-frequency acoustic properties of

a fluid/porous solid interface. II. The 2D reflection Green’s function. J.
Acoust. Soc. Am., 74, 915–924.
Ferrazzini, V., and K. Aki (1987). Slow waves trapped in a fluid-filled infinite
crack: implication for volcanic tremor. J. Geophys. Res., 92, 9215–9223.
Frenkel, J. (1944). On the theory of seismic and seismoelectric phenomena

in a moist soil. J. Phys., USSR, 8, 230–241.
Gelinski, S., and C.H. Cheng (1998). Integrated borehole acoustic fracture
characterization. Proceedings, 4th SEGJ Tokyo, 111–116.
110 BIBLIOGRAPHY
Groenenboom, J. (1998). Acoustic monitoring of hydraulic fracture growth.

Ph.D. thesis, Delft University of Technology.
Hardin, E.L., C.H. Cheng, F.L. Paillet, and J.D. Mendelson (1987). Fracture
characterization by means of attenuation and generation of tube waves in
fractured crystalline rock at Mirror Lake, New Hampshire. J. Geophys.
Res., 92, 7989–8006.
Henry, F. (2005). Characterization of borehole fractures by the body and in-

terface waves. Ph.D. thesis, Delft University of Technology.
Holzhausen, G.R., and R.P. Gooch (1985). Impedance of hydraulic fractures:

its measurement and use for estimating fracture closure pressure and di-
mensions. SPE13892, 411–422.
Hornby, B.E., D.L. Johnson, K.W. Winkler, and R.A. Plumb (1987). Fracture
evaluation from the borehole Stoneley wave. SEG Expanded Abstracts, 688–
691.
Hornby, B.E., D.L. Johnson, K.W. Winkler, and R.A. Plumb (1989). Fracture
evaluation using reflected Stoneley–wave arrivals. Geophysics, 54, 1274–
1288.
Hsu, K., A. Brie, and R.A. Plumb (1987). A new method for fracture iden-
tification using array sonic tools. J. Petrol. Tech., 39, 677–683.
Hsu, K., and C. Esmersoy (1992). Parametric estimation of phase and group
slownesses from sonic logging waveforms. Geophysics, 57, 978–985.
Hsui, A.T., and M.N. Toksöz (1986). Application of an acoustic model to

determine in situ permeability of a borehole. J. Acoust. Soc. Am., 79,
2055–2059.
Johnson, D.L., J. Koplik, and R. Dashen (1987). Theory of dynamic per-

meability and tortuosity in fluid-saturated porous media. J. Fluid Mech.,
176, 379–402.
Johnson, D.L., D.L. Hemmick, and H. Kojima (1994). Probing porous media
with first and second sound. I. Dynamic permeability. J. Appl. Phys., 76,
104–114.
Kelder, O. (1998). Frequency-dependent wave propagation in water-saturated

porous media. Ph.D. thesis, Delft University of Technology.
BIBLIOGRAPHY 111
Kimball, C.V., and T.L. Marzetta (1984). Semblance processing of borehole

acoustic array data. Geophysics, 49, 274–281.
Korneev, V. (2008). Slow waves in fractures filled with viscous fluid. Geo-
physics, 73, N1–N7.
Korneev, V. (2010). Low-frequency fluid waves in fractures and pipes. Geo-

physics, 75, N97–N107.
Kostek, S., D.L. Johnson, and C.J. Randall (1998a). The interaction of tube
waves with borehole fractures, Part I: Numerical models. Geophysics, 63,
800-808.
Kostek, S., D.L. Johnson, K.W. Winkler, and B.E. Hornby (1998b). The
interaction of the tube waves with borehole fractures, Part II: Analytical
models. Geophysics, 63, 809–815.
Lamb, H. (1928). Statics. Cambridge University Press.
Lang, S.W., A.L. Kurkjian, J.H. McClellan, C.F. Morris, and T.W. Parks
(1987). Estimating slowness dispersion from arrays of sonic logging wave-
forms. Geophysics, 52, 530–544.
Liu, E., S. Crampin, and J.A. Hudson (1997). Diffraction of seismic waves by
cracks with application to hydraulic fracturing. Geophysics, 62, 253–265.
Markov, M.G. (2009). Low-frequency Stoneley wave propagation at the in-

terface of two porous half-spaces. Geophys. J. Int., 177, 603-608.
Mayes, M.J., P.B. Nagy, L. Adler, B.P. Bonner, and R. Streit (1986). Excit-
ation of surface waves of different modes at fluid-porous solid interface. J.
Acoust. Soc. Am., 79, 249–252.
Miklowitz, J. (1978). The theory of elastic waves and waveguides. North–

Holland Publishing Company, Amsterdam.
Naimi-Tajdar, R., C. Han, K. Sepehrnoori, T.J. Arbogast, and M.A. Miller

(2007). A fully implicit, compositional, parallel simulator for IOR processes
in fractured reservoirs. SPE Journal, 12, 367–381.
Norris, A.N. (1989). Stoneley-wave attenuation and dispersion in permeable

formations. Geophysics, 54, 330–341.
Paige, R.W., J.D.M. Roberts, L.R. Murray, and D.W. Mellor (1992). Fracture
measurement using hydraulic impedance testing. SPE 24824, 605–614.
112 BIBLIOGRAPHY
Paige, R.W., L.R. Murray, and J.D.M. Roberts (1995). Field application of
hydraulic impedance testing for fracture measurement. SPE Production &
Facilities, 7–12.
Paillet, F.L., and J.E. White (1982). Acoustic modes of propagation in the
borehole and their relationship to rock properties. Geophysics, 47, 1215–
1228.
Pride, S.R., and J.G. Berryman (2003a). Linear dynamics of double-porosity
dual-permeability materials. I. Governing equations and acoustic attenu-
ation. Physical Review E, 68, 036603.01-036603.10.
Pride, S.R., and J.G. Berryman (2003b). Linear dynamics of double-porosity
dual-permeability materials. II. Fluid transport equations. Physical Review
E, 68, 036604.01-036604.10.
Qobi, L., A. Kuijper, X.M. Tang, and J. Strauss (2001). Permeability determ-
ination from Stoneley waves in the Ara group carbonates, Oman. GeoAr-
abia, 6, 649-666.
Ramirez, B., H. Kazemi, M. Al-Kobaisi, E. Ozkan, and S. Atan (2009). A
critical review for proper use of water/oil/gas transfer functions in dual-
porosity naturally fractured reservoirs: Part I. SPE Reservoir Evaluation
& Engineering, 12, 200–210.
Rayleigh, L. (1885). On waves propagated along the plane surface of an elastic
solid. Proc. London Math. Soc., s1–17, 4–11.
Rosenbaum, J.H. (1974). Synthetic microseismograms: logging in porous
formation. Geophysics, 39, 14–32.
Saito, H., K. Hayashi, and Y. Iikura (2004). Detection of formation bound-
aries and permeable fractures based on frequency-domain Stoneley wave
logs. Exploration Geophysics, 35, 45-50.
Schmitt, D.P., M. Bouchon, and G. Bonnet (1988). Full-wave synthetic acous-
tic logs in radially semiinfinite saturated porous media. Geophysics, 53,
807–823.
Schoenberg, M. (1984). Wave propagation in alternating solid and fluid layers.
Wave Motion, 6, 303–320.
Schoenberg, M., and P.N. Sen (1986). Wave propagation in alternating solid
and viscous fluid layers: Size effects in attenuation and dispersion of fast
and slow waves. Appl. Phys. Lett., 48, 1249–1251.
BIBLIOGRAPHY 113
Scholte, J.G. (1948). On the large displacements commonly regarded as

caused by Love waves and similar dispersive surface waves. Proc. Konink.
Ned. Akad. Wetensch, 51, 533–543.
Smeulders, D.M.J. (1992a). On wave propagation in saturated and partially

saturated porous meida. Ph.D. thesis, Eindhoven University of Technology.
Smeulders, D.M.J., J.P.M. De la Rosette, and M.E.H. van Dongen (1992b).

Waves in partially saturated porous media. Transp Porous Med, 9, 25-37.
Smeulders, D.M.J., R.L.G.M. Eggels, and M.E.H. van Dongen. (1992c). Dy-
namic permeability: reformulation of theory and new experimental and
numerical data. J. Fluid Mech., 245, 211–227.
Smeulders, D.M.J., and M.E.H. van Dongen (1997). Wave propagation in

porous media containing a dilute gas-liquid mixture: theory and experi-
ments. J. Fluid Mech., 343, 351–373.
Sniekers, R.W.J.M., D.M.J. Smeulders, M.E.H. van Dongen, and H. van der
Kogel (1989). Pressure wave propagation in a partially water–saturated
porous medium. J. Appl. Phys., 66, 4522-4524.
Stoneley, R. (1924). Elastic waves at the surface of separation of two solids.

Proc. R. Soc. Lond. A, 106, 416-428.
Tang, X.M., and C.H. Cheng (1989). A dynamic model for fluid flow in open
borehole fractures. J. Geophys. Res., 94, 7567–7576.
Tang, X.M. (1990). Acoustic logging in fractured and porous formations.

Ph.D. thesis, MIT.
Tang, X.M., C.H. Cheng, and M.N. Toksöz (1991a). Dynamic permeabil-
ity and borehole Stoneley waves: A simplified Biot–Rosenbaum model. J.
Acoust. Soc. Am., 90, 1632–1646.
Tang, X.M., C.H. Cheng, and M.N. Toksöz (1991b). Stoneley–wave propaga-
tion in a fluid–filled borehole with a vertical fracture. Geophysics, 56, 447–
460.
Tang, X.M., and C.H. Cheng (1993). Borehole Stoneley wave propagation
across permeable structures. Geophysical Prospecting, 41, 165–187.
Tang, X.M., and C.H. Cheng (1996). Fast inversion of formation permeab-
ility from Stoneley wave logs using a simplified Biot–Rosenbaum model.
Geophysics, 61, 639–645.
114 BIBLIOGRAPHY
Toksöz, M.N., C.H. Cheng, and R.D. Cicerone (1992). Fracture detection and
characterization from hydrophone vertical seismic profiling data. Interna-
tional Geophysics, 51, 389–414.
Überall, H. (1973). Surface waves in acoustics. Physical Acoustics, Principles

and Methods, 10, Academic Press, New York, NY.
Van Dalen, K.N., G.G. Drijkoningen, and D.M.J. Smeulders (2010). On wave-
modes at the interface of a fluid and a fluid-saturated poroelastic solid. J.
Acoust. Soc. Am., 127, 2240-2251.
Van Dalen, K.N., G.G. Drijkoningen, and D.M.J. Smeulders (2011a). Pseudo
interface waves observed at the fluid/porous–medium interface. A compar-
ison of two methods. J. Acoust. Soc. Am., 129, 2912–2922.
Van Dalen, K.N. (2011b). Multi–component acoustic characterization of por-

ous media. Ph.D. thesis, Delft University of Technology.
Van der Grinten, J.G., M.E.H. van Dongen, and H. van der Kogel (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., M.E.H. van Dongen, and H. van der Kogel (1987).
Strain and pore pressure propagation in a water–saturated porous medium.
J. Appl. Phys., 62, 4682–4687.
Van der Hijden, J.H.M.T., and F.L. Neerhoff (1984). Scattering of elastic
waves by a plane crack of finite width. J. Appl. Mech., 51, 646–651.
Viktorov, I.A. (1967). Rayleigh and Lamb waves: physical theory and applic-
ations. Plenum Press, New York.
White, J.E. (1983). Underground sound: application of seismic waves. El-

sevier, New York.
Winkler, K.W., H.L. Liu, and D.L. Johnson (1989). Permeability and bore-
hole Stoneley waves: Comparison between experiment and theory. Geo-
physics, 54, 66–75.
Wisse, C.J. (1999). On frequency dependence of acoustic waves in porous

cylinders. Ph.D. thesis, Delft University of Technology.
Yamamoto, H., S. Watanabe, H. Mikada, T. Endo, and A. Brie (1998). Frac-

ture imaging using borehole acoustic reflection survey. Proceedings, 4th
SEGJ International Symposium, 375–382.
BIBLIOGRAPHY 115
Zwikker, C., and C.W. Kosten (1941). Extended theory of the absorption of
sound by compressible wall–coverings. Physica VIII, 9, 968–978.
Zwikker, C., and C.W. Kosten (1949). Sound Absorbing Materials. Elsevier,
New York.
116 BIBLIOGRAPHY
Acknowledgements
First of all, I would express my gratitude to my supervisor and my promotor,

David Smeulders. David is strict and rigorous in scientific research. He is
also a detail-oriented person. I have learned it from his request to repeat
experiments whenever it comes to any doubts, and to revise the thesis and
papers over and over again. Moreover, David is a very nice person. At the be-
ginning of my PhD study, I encountered some problems and felt very difficult
because of the shifting topic (from my MSc study) as well as from failing to
adapt to the culture difference. His patience, tolerance and encouragement,
helped me through. He is very responsible. Our weekly (biweekly) discussion
has never been interrupted even after his transfer to TU Eindhoven. David,
thank you! Even though I know a simple “Thank you” is far from enough.
Many thanks are also addressed to the following professors for particip-
ation in my thesis committee: Andrei Metrikine, Giovanni Bertotti, Holger
Steeb, Kees Wapenaar, Li Ning, and Ruud Schotting. Thank you for your
time to read the manuscript and give me valuable comments.
I would like to thank the China Scholarship Council (CSC) for the 4
year PhD fellowship support. I would thank the Research Institute of Petro-
leum Exploration & Development (RIPED) of PetroChina to fund a research
project at TU Delft after the end of the 4th year of my PhD study. The
experimental work has been crucial for this PhD project. I appreciate the
great help from the technicians Jan Etienne and Karel Heller. Without your
assistance, it would not have been possible for me to accomplish this large
amount of experiments with such a large set-up. I would also like to thank
the other staff in our laboratory. It is very nice to work with you and I really
enjoyed the moments that we spent during the coffee break. I would like to
express my appreciation to Karel van Dalen: discussions with you are always
very inspiring. Thanks for your help in Matlab programming, as well as for
the good advices for my PhD study. I would like to thank those who also gave
me advice on literature study and Matlab/LaTex using: Evert Slob, Gabriel
Chao, Menne Schakel, Christiaan Schoemaker, Bouko Vogelaar, Jürg Hun-
117
118 Acknowledgements
ziker and Joost van der Neut. My thanks also go to my Chinese colleagues
Dong Yufei, He Yuanyuan and Guo Hua. I appreciate your help during my
PhD study and my stay in Delft. You helped me quickly getting used to
working in the laboratory and living in Delft. Alimzhan Zhubayev and Alex
Kirichek, the interesting discussions with you makes working in the laborat-
ory very lively. Ralph Feld, it is a pleasant to work in the same office with
you. I appreciate your help to translate Dutch letters for me. Especially, I
would thank you for translating the Summary part of my thesis into Dutch.
I also find it is interesting to teach you Chinese language.
I enjoyed the activities organized by DOGS (Delft Organization of Geo-
physics Students), e.g. the Thursday talks, the DOGS drinks, and the Com-
pany visits. I would also like to thank my alumni from China University of
Geosciences (Beijing) in the Netherlands: Yu Yanqing, Zhang Mengmeng,
Tan Shuhong, Li Tianqi, Li Jiaguang. I enjoy all the gatherings with you.
I would like to give my thanks to Wang Meng, Cui Haiyang, Xu Min, Yang
Xiaogang and other friends in Delft. It is nice to meet you in Delft and thank
you for your help to me and my family.
Finally, I address my appreciation to my families. I appreciate the great
help and encouragement from my parents and parents-in-law during my PhD
study. Thank you for coming to the Netherlands when my son was born. I
would like to thank my wife, Liao Fang. Dear, thank you for your continuous
support and encouragement since we met in the first year of our university
life in Beijing. I also appreciate it so much that you gave up the job oppor-
tunity in the USA after you obtained a MSc degree there and joined me in
the Netherlands. I am grateful for your company during the good time and
also the tough time in the Netherlands. My little son Fan Yuxiu, you bring
me so much joy. Your crying, laughing and your struggling to talk are a gift
for me. I enjoy every moment with you. I love you so much!
Huajun Fan
Delft, May, 2014
Curriculum Vitae
Personal information
Huajun Fan was born on 7 September 1983 in Suichang, Zhejiang Province
of China. He went to China University of Geosciences (Beijing) in 2001 to
study as a Bachelor student and obtain his BSc in Exploration and Engin-
eering in 2005. In the same year, he started as a Master student at the same
university. From March 2006 to October 2006, he interned as a geophysicist
in the Research Institute of Petroleum Exploration & Development (RIPED)
of PetroChina in Beijing. He obtained his MSc in Geological Engineering in
2008, and afterward became a PhD student at the Faculty of Civil Engin-
eering and Geosciences of Delft University of Technology. Since 2013, he has
been working in a project cooperating with RIPED as a research scientist at
Delft University of Technology.
Journal publications
• Fan, H., and D.M.J. Smeulders (2013). Shock–induced wave propaga-
tion over porous and fractured borehole zones: Theory and experi-
ments. Journal of the Acoustical Society of America, 134, 4792–4800.
• Fan, H., and D.M.J. Smeulders (2012). Shock–induced borehole waves
and fracture effects. Transport in Porous Media, 93, 263–270.
Conference publications
• Fan, H., and D.M.J. Smeulders (2013). Interaction between fracture
zones and shock–induced borehole waves. Poromechanics V, 267–275.
• Fan, H., and D.M.J. Smeulders (2012). Fracture evaluation using
shock–induced borehole waves. 28th International Symposium on Shock
Waves, 805–810.
119
120 Curriculum Vitae
• Fan, H., and D.M.J. Smeulders (2011). Open borehole shock–induced

Stoneley waves in fractured formations and mandrel samples. SEG
Expanded Abstracts, 30, 459–463.
• Fan, H., and D.M.J. Smeulders (2010). Shock–induced Stoneley waves

in fractured and permeable formations. SEG Expanded Abstracts, 29,
558–562.

PhD Thesis FAN2014

Uploaded by

Copyright:

Available Formats

PhD Thesis FAN2014

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PhD Thesis FAN2014

Uploaded by

Copyright:

Available Formats

Shock-induced borehole waves in fractured

ter verkrijging van de graad van doctor

door Huajun FAN

Master of Geological Engineering

This work is ﬁnancially supported by China Scholarship Council (CSC) and

Published by: Uitgeverij BOX Press, Oisterwijk, The Netherlands

2 Constitutive relations and momentum equations 9

3 Tube wave propagation in a ﬂuid-ﬁlled borehole intersected

3.4 Reﬂection and transmission . . . . . . . . . . . . . . . . . . . 37

5 Shock-induced borehole waves and fracture eﬀects 55

6 Shock-induced wave propagation over porous and fractured

7 Fracture eﬀects in mandrel sample 89

Appendix A Derivation of the plane fracture wave equation 99

Appendix B Coherence method 101

Appendix C Porosity and permeability determination 103

Curriculum Vitae 119

Natural or hydraulic fractures are of major importance for the productivity

Natuurlijke en hydraulische scheurvorming is van groot belang voor de pro-

deels zal reﬂecteren en deels zal verdergaan, waardoor de aanwezigheid van

Borehole logging is the technique to obtain quantitative information on the

Borehole Formation Borehole Formation

1.1 Ray tracing

1.2 Stoneley wave

Permeable Zone Permeable Zone

1.3 Literature survey

1.3.2 Waves along ﬂat interfaces

1.3.3 Stoneley waves in a borehole

representative model for Stoneley wave propagation in a rigid formation in the

1.3.4 Stoneley waves in a borehole intersected by fractures

1.4 Thesis outline

Constitutive relations and

2.1 Constitutive equation

• The wavelength is larger than the pore size;

• The matrix is elastic and isotropic and there is no dissipation eﬀect in

• There are no thermo-elastic or chemical reaction eﬀects in the system.

2.2 Momentum equations

The interaction term is speciﬁed in its linear form:

ρ12 = −(α∞ − 1)φρf , (2.19)

ρ11 = (1 − φ)ρs − ρ12 , (2.20)

ρ22 = φρf − ρ12 = α∞ φρf . (2.21)

2.3 One-dimensional ﬁeld equations

is the frequency which characterizes the transition from low-frequency vis-

In the limiting case of large tortuosity, ωc = θ −1 .

Combination of both equations yields that

We note that Vs is real-valued high-frequency phase velocity of the shear

Tube wave propagation in a

ψ = A+ e−iκ1 z + A− eiκ1 z for z < −h/2, (3.4)

ψ = B + e−iκ2 z + B − eiκ2 z for − h/2 < z < h/2, (3.5)

ψ = C + e−iκ1 z for z > h/2. (3.6)

A− /A+ = −2i(κ22 − κ21 ) sin(κ2 h)/D, (3.7)

B + /A+ = 2κ1 (κ1 + κ2 )eiκ2 h /D, (3.8)

B − /A+ = 2κ1 (κ2 − κ1 )e−iκ2 h /D, (3.9)

C + /A+ = 4κ1 κ2 eiκ1 h /D, (3.10)

3.2 Borehole waves in ﬂexible and fractured form-

diameter (low frequencies). In that case we can assume piston-like motion

This equation has the well-known solution: uz = u+ (z − cT t) + u− (z + cT t),

p̂ = AK0 (κr), (3.27)

where K0 (κr) is a modiﬁed Bessel function of zeroth order and complex

It is now possible to compute the tube wave velocity as a function of per-

Table 3.1: Parameters values.

where f = ω/(κcf ), p = ω/(κcp ), and s = ω/(κcs ). The parameter ρs is