You are on page 1of 16

New Directions in Sonic Logging

Reaching the reservoir, and doing so safely, is made easier by recent


advances in shear-wave logs. Anisotropy measurements spotlight
more efficient ways to drill and stimulate formations and real-time
sonic-while-drilling alerts drillers to overpressured zones. These and
other developments are helping solve tough production problems.

Alain Brie
Takeshi Endo
David Hoyle
Fuchinobe, Japan
Daniel Codazzi
Cengiz Esmersoy
Kai Hsu
Sugar Land, Texas, USA

Michael C. Mueller
Amoco Exploration and Production
Houston, Texas
Tom Plona
Ram Shenoy
Bikash Sinha
Ridgefield, Connecticut, USA

Stan Denoo
Englewood, Colorado, USA

For help in preparation of this article, thanks to


Michael Kane and Christopher Kimball, SchlumbergerDoll Research, Ridgefield, Connecticut, USA; Philippe
Laurent and Julian Singer, Schlumberger Wireline &
Testing, Caracas, Venezuela; Frank Morris and Robert
Young, Schlumberger Wireline & Testing, Sugar Land,
Texas, USA.
CDR (Compensated Dual Resistivity), CMR (Combinable Magnetic Resonance), DSI (Dipole Shear
Sonic Imager), ELAN (Elemental Log Analysis), EPT
(Electromagnetic Propagation Tool), FMI (Fullbore
Formation MicroImager), GeoFrame, ISONIC (IDEAL
sonic-while-drilling), LSS (Long-Spaced Sonic Tool),
PowerPak, PowerPulse and UBI (Ultrasonic Borehole
Imager) are marks of Schlumberger.

40

1. Esmersoy C, Koster K, Williams M, Boyd A and Kane


M: Dipole Shear Anisotropy Logging, Expanded
Abstracts, 64th SEG Annual International Meeting and
Exposition, Los Angeles, California, USA, October
23-28, 1994, paper SL3.7.
Shear Wave Logging with Dipoles, Oilfield Review
2, no. 4 (October 1990): 9-12.
2. In the language of sonic logging, slownessthe
reciprocal of velocityis most commonly used. It is
identical to the interval transit time, which is a basic
measurement made by sonic logging tools.
3. Armstrong P, Ireson D, Chmela B, Dodds K,
Esmersoy C, Miller D, Hornby B, Sayers C,
Schoenberg M, Leaney S and Lynn H: "The Promise
of Elastic Anisotropy, Oilfield Review 6, no. 4
(October 1994): 36-47.
4. Aron J, Chang SK, Dworak R, Hsu K, Lau T, Masson
J-P, Mayes J, McDaniel G, Randall C, Kostek S and
Plona T: Sonic Compressional Measurements While
Drilling, Transactions of the SPWLA 35th Annual
Logging Symposium, Tulsa, Oklahoma, USA, June
19-23, 1994, paper SS.

One of the greatest challenges in optimizing


development of a reservoir is placing wells
correctly. Getting them in the right place,
whether vertical or horizontal, not only
decreases cost but improves recovery. Today,
service companies are focusing their effort
on developing new technology to help plan
field development, optimize well location
and improve drilling safety. Solutions to the
problem require integrating many types of
data. At the forefront are measurements that
take advantage of evolving technology in
sonic logging.
During the last decade important advances
have been made in sonic logging.1 Using
dipole sources that can excite flexural
wavesshear-like vibrations of the boreholethese tools are capable of measuring
sonic compressional and shear-wave slowness data in hard and soft formations.2 Shearwave anisotropy measurements are sensitive
to stress and fracture density and directions.3
This is vital information for those who want
to optimize production by drilling a borehole aimed at encountering as many fractures as possible.
Sonic logging is becoming a routine way to
plan well placement strategiesleading to
improvements in reservoir production.
Mechanical rock properties from sonic measurements can help predict formation
strength and potential sanding problems. In
both vertical and horizontal wells, this information allows prediction of the best direction to perforate for maximum production.
Stress magnitude derived from sonic logging
helps forecast maximum sand-free drawdown pressures.
Logging-while-drilling (LWD) sonic tools
have been improving, too. Since the first
LWD sonic tool designed to provide compressional-slowness measurements while
drilling was reported in 1994, many successful real-time and memory logs have been
recorded in hard and soft rocks.4 These realtime LWD sonic measurements bring fresh
information, obtained soon after the drill bit
penetrates the formation. This information is
vital to the driller, helping to avoid costly mistakes such as drilling into overpressured
zones without proper mud-weight adjustments. In addition, experience is showing
that the formations around the wellbore
change when exposed to drilling fluids. LWD
sonic logs from freshly drilled boreholes
compared with wireline measurements
usually taken days after the drilling has
exposed the formationshow remarkable
differences. Both bring important, but different, information about wellbore properties.

Oilfield Review

Fractures

CAUTION
Overpressure
ns

tio

ra

rfo

Pe

Maximum
stress

Minimum
stress

Spring 1998

With the introduction of dipole sonic logs,


the petrophysical community has the ability to
record high-quality shear and compressional
slownesses in a variety of formations, and for
the first time in slow formations. These measurements are helping to solve some of the
mysteries in formation interpretation.
In this article we look at how sonic shearwave anisotropy measurements are used to
find fractures and their orientation, understand stress directions in formations and
predict the best directions for perforating
or drilling stable vertical and horizontal
wellbores that yield optimum flow rates. We
discuss how real-time sonic LWD measurements are being used, first as a means to
avoid costly drilling mistakes, and then as an
effective way of determining unaltered formation properties and pay zones in hard and
soft formations. Both formation types present
special problems for LWD measurements.
Finally, we will see how petrophysicists are
interpreting sonic logs in gas-saturated shaly
sandsone of the most difficult environments for sonic logging.

41

Using Shear-Wave Anisotropy


to Improve Production

Drilling
fractures

Borehole radii

Smin

-1
Damage
-3

Breakouts

Smax
-5

-3

-1

Borehole radii

Mechanical state of rock and failure


mechanisms. The annulus around the
wellbore may be damaged by drilling
and tectonic stresses on the openhole.
Damage in the form of breakouts or
washouts usually first appears in the
direction of minimum stress (Smin), and
drilling-induced fractures occur with their
strike direction along the direction of
maximum stress (Smax).
5. Fletcher PA, Montgomery CT, Ramo GG, Miller ME
and Rich DA: Using Fracturing as a Technique for
Controlling Formation Failure, paper SPE 27899,
presented at the 64th SPE Annual Western Regional
Meeting, Long Beach, California, USA, March
23-25, 1994.
6. Upchurch ER, Montgomery CT, Berman BH and
Rael EL: A Systematic Approach to Developing
Engineering Data for Fracturing Poorly Consolidated
Formations, paper SPE 38588, presented at the
1997 SPE Annual Technical Conference and Exhibition,
San Antonio, Texas, USA, October 5-8, 1997.
7. Fletcher et al, reference 5.

To produce hydrocarbons efficiently, reservoir engineers need to know tectonic and


wellbore stresses and their direction. These
properties first affect wellbore stability
(breakouts), and then the ability to hydrofracture the well. Permeability impairment and
sanding problems can be influenced by
stress. The presence of a borehole will influence the state of stress of the rock and the
formation strength around the borehole (left).
Sonic measurements, such as those from
the UBI Ultrasonic Borehole Imager tool for
breakout measurements and the DSI Dipole
Shear Sonic Imager tool for shear-wave
anisotropy, provide a foundation for understanding formation geomechanics. Stress
directions can be determined by measuring
the locations of wellbore breakouts (below).
Shear-wave anisotropy is a more robust
method for determining stress directions
because the measurement is in the rock and
does not rely on formation wall failure
washoutwhich may not always be present
in the well (see Dipole Shear-Wave
Anisotropy Analysis, page 44).
Natural and induced fracture orientations
are also important considerations for reservoir management. Since hydraulic fractures
open in a plane perpendicular to the minimum stress, determining stress direction is
crucial. Consider the case of reservoir
drainage from a hydraulically fractured well

in a low-permeability formation. Since the


drainage pattern is elliptical, optimum
reservoir drainage depends on the correct
placement of multiple wells (next page, top
left).5 Stress direction is especially important
when fracturing from horizontal wells in
which control of the fracture orientation
with respect to the wellbore is important
(next page, top right). Since shear-wave
anisotropy measurements are sensitive to
fracture orientation, they provide useful
directions for drilling a horizontal section
aimed at encountering as many natural
fractures as possible.
Even in soft unconsolidated formations,
where sonic measurements are difficult,
shear-wave anisotropy measurements are recommended reservoir engineering practices
for planning fracture treatments.6 While
enhancing productivity through fracture stimulation is the primary goal, fracturing is also
used as a means of implementing effective
sand control in unconsolidated formations. It
has been shown that unprotected
(unpropped) perforations are a major cause
of sand production.7 Once stress direction is
determined, 180 phased oriented perforations can be used to optimize the fracture
treatment, as well as minimize the number of
unpropped perforations that cause sanding
problems (next page, bottom). In addition,
the use of 180 phased perforations aligned
normal to the minimum stress direction helps
minimize wellbore tortuosity after fracturing.
(continued on page 46)

X066

Images versus depth


4

Top

X067

2
X068

Depth

X66.7m

Hole deviation

37.7 degrees

Breakout

138.0 degrees N
111/2 degrees top
0.8 in.

Breakout

Breakout
-2

-4
-4

-2

Borehole radius, in.

Breakouts from the UBI tool. The UBI Ultrasonic Borehole Imager tool uses a pulse-echo
reflection measurement that provides high-resolution images (left) of borehole size and
shape. The radius plot (right) shows breakouts (red arrows). Breakouts, caused by the
borehole being in compression failure, have been observed worldwide to cause ovalization of the borehole with the ovals long axis parallel to the minimum stress.

42

Oilfield Review

Smin

Smin

Smin

Good Drainage

Incomplete Drainage

Smin

Smin

Injector
wells
Smin
Producer
wells
Good Areal Sweep

Poor Areal Sweep

Fracture azimuth and geometry determination. Incomplete reservoir


drainage can occur if fracture orientation is ignored when well spacing and placement are designed (top). An understanding of fracture
orientation can be useful in determining optimum well location for
waterflooding, and enhanced oil recovery (EOR) applications (bottom).
[Adapted from Lacey LL and Smith MB: Fracture Azimuth and Geometry Determination, in Recent Advances in Hydraulic Fracturing, SPE
Monograph No. 12. Henry L. Doherty Monograph Series. Richardson,
Texas, USA: Society of Petroleum Engineers (1989): 341-354.]

Effective
perforations

Ineffective
perforations

Fracturing horizontal wells. Wells drilled along the line of


minimum horizontal stress will fracture in planes perpendicular to the wellbore. Wells drilled in the direction normal to the minimum stress (Smin) will fracture along the
wellbore. Fracture placement can influence the drawdown along extended horizontal sections. Knowing stress
directions and magnitude from formation mechanical
properties helps orient perforations to contain fractures in
the desired directions.

Stable
perforations

Fracture

Smax
Smin

Unstable
perforations

Oriented perforations for sand control. In sand control, the 180 phased perforations ensure
that all the perforations connect to the fracture and are propped. This procedure eliminates the
unconnected perforations that produce sand during drawdown. The use of 180 phased perforations, oriented perpendicular to the minimum stress (Smin), is helpful in fracturing because
these perforations minimize breakout that causes borehole tortuosity.

Spring 1998

43

Dipole Shear-Wave Anisotropy Analysis

For sonic measurements, it is well recognized that


Fractures

sedimentary rocks generally exhibit some degree


of anisotropy.1 Anisotropy may arise from intrinsic
structural effects, such as aligned fractures and
layering of thin zones, or from unequal stresses
within the formation. These effects lead to differences in formation elastic properties, and if they
are on a smaller scale than the sonic wavelengths, then sonic wave propagation can be used
to detect and quantify the anisotropy.
Sonic waves travel fastest when the direction of
particle motionpolarizationis aligned with the

Y
X

Formation
fast axis

Formation
slow axis

materials stiffest direction. Shear-wave particle


motion is in a plane perpendicular to the wave

Fast shear
Slow shear

propagation direction. If the formation is


anisotropic in this plane, meaning that there is
one direction that is stiffer than another, then the

Shear-wave splitting.
Shear waves travel in an
anisotropic formation with
different speeds along the
directions of the formation
anisotropy. In this example,
anisotropy is caused by the
vertical fractures (or microcracks) with a strike direction
along the formation Y-axis,
and the fastest shear
wavewith the longer
wavelength component
will be polarized along the
fracture strike direction as it
propagates along the borehole (Z-axis). When shearwave splitting is the result of
stress anisotropy, the Y-axis
corresponds to the direction
of maximum stress, and the
X-axis corresponds to the
direction of minimum
stress.

shear-wave polarization aligned in the stiff direction will travel faster than one aligned in the
other, more compliant direction. As a result, the
shear wave splits into two components, one polarized along the formations stiff (or fast) direction,
and the other polarized along the formations compliant (or slow) direction.2
For example, in the case of vertically-aligned

dense microcracks or fractures, a shear wave that


is polarized parallel to the fracture strike will prop-

hole axis (next page). The measurement involves

This is done by minimizing the cross-receiver

agate faster than a shear wave polarized perpen-

recording the waveforms on receivers pointing in

energies, xy and yx. The rotated direction of the

dicular to it (above right). In general, a shear (or

directions parallel and normal to each transmitter

fastest shear wave becomes the fast-shear tool

flexural) wave, generated by a dipole source, will

along the tool x- and y-axes.3

azimuth; and the tool orientation, measured by a

Four sets of waveforms are recorded at each

magnetometer, is used to determine the fast shear

along the X- and Y-directions in the formation. As

depth and receiver level. These measurements

azimuth relative to true north. This rotation, called

they propagate along the borehole, the fast wave

are labeled xx, xy, yx and yy. The first direction

the Alford method, uses the fact that the

will be polarized along the direction parallel to the

refers to the transmitter and the second direction

anisotropy model expects the amplitude of the

fracture strike and a slow wave in the direction

to the receiver. The direction and speed of the fast

cross-receiver measurements to vanish when the

perpendicular to it.

and slow split shear waveforms traveling in the

measured axes x and y align with the anisotropy

formation can be easily determined by mathemati-

axes X and Y.4

split into two orthogonal components polarized

With two orthogonal dipole transmitters and

In addition to the fast and slow shear-wave

multiple receiver pairs aligned in orthogonal

cally rotating the measured waveforms through an

directions, the DSI Dipole Shear Sonic Imager tool

azimuthal angle so that they line up with the two

velocitiesdetermined by a slowness-time-

can measure the components of shear slowness in

orthogonal formation X- and Y-directions.

coherence (STC) processing on the rotated wave-

any direction in a plane perpendicular to the bore-

formsthree measurements of anisotropy are


computed.5 These are energy anisotropy, slowness anisotropy and time anisotropy.

44

Oilfield Review

Slowness anisotropy is the difference between


the fast and slow slownesses calculated by
STC on the rotated waveforms. It yields a quantitative measure of slowness anisotropy, and has
the best vertical resolution at about 3 ft [1 m]
the size of the receiver array. It can be compared
directly with seismic or core measurements of
slowness anisotropy.
Traveltime anisotropy is the arrival-time difference between the fast- and slow-shear waves at
the receivers. It is obtained from a cross-correlation between fast and slow shear-wave arrivals at
each receiver spacing. Time lags computed at
each receiver are referenced to the largest offset
receiver and averaged across the receiver array.
This is divided by the average of the fast and slow
arrival times to compute a percentage difference.
The traveltime anisotropy indicator is robust and
quantitative, and has the vertical resolution of the
average transmitter-receiver spacing, 13 ft [4 m].
Slowness and traveltime anisotropy indicators are
identical in formations with homogeneous beds
thicker than 13 ft.
Energy anisotropy is the energy in the crosscomponent waveforms as a percentage of energy

Flexural waves induced by


dipole transmitters. During logging, flexural waves are induced
by dipole transmitters fired
sequentially in two perpendicular
directions, first along the tool xand then the tool y-axes. In this
example, the fastest component
of the induced shear wave is
polarized along the formation Yaxis direction, which is aligned
along the fracture strike or maximum stress direction. The slowest component of the shear wave
is polarized along the formation
X-axis. Projections of these two
shear-wave components are
received by each of the dipole
sonic tool x- and y-receiver pairs.
The inline signals xx and yy are
the x-receiver and y-receiver
waveforms received when the xand y- transmitters are fired.
Cross-signal components xy and
yx are the y- and x-receiver
waveforms received as the x- and
y-transmitters are fired. The
Alford rotation angle, , is determined by minimizing the crosssignal components. This would
happen automatically if the tool
axes were rotated through an
angle, , and aligned with the two
orthogonal directions in the
anisotropic formation.

R8 y

Receiver-8 pair

R8x
R7 y

Receiver-7 pair

R7x
R6 y

R6x

Receiver-6 pair
R5 y

R5x

Receiver-5 pair

R4 y

R4x

Receiver-4 pair

R3 y

R3x
R2x

R2 y

R1x

Receiver
array

R1 y

Receiver-3 pair
Receiver-2 pair
Receiver-1 pair

Borehole
flexural wave
(exaggerated)

Undisturbed
borehole

Dipole
transmitter
pair
Ty
Tx

in all four components. In an isotropic formation,


energy anisotropy reads zero. In an anisotropic

formation, the reading depends on the degree of


X

anisotropy. Two curves are computed from the

Formation fast
shear wave axis
y

Tool axis

Tool orientation
relative to formation

waveforms: minimum and maximum cross-energy.


The minimum cross-energy is the energy in the

and time), energy anisotropy is a measure of both

cross-components when the tool measurement

slowness and amplitude differences of the fast-

axis lines up with the formation anisotropy axis.

and slow-shear waves. Large differences between

Minimum cross-energy reads zero in an ideal for-

the maximum and minimum values, especially

mation whether anisotropic or not. This curve is a

when the minimum energy is low, indicate zones of

good relative measure of whether the assumed

significant anisotropy. Energy anisotropy, though

model for anisotropy inversion fits the real forma-

qualitative, is little affected by processing, and is

tion. The maximum cross-energy is a measure of

the principal measure of anisotropy.

1. Armstrong et al, reference 3, main text.


2. For compressional waves, the particle motion is the same
direction as the wave propagation.
3. Esmersoy et al, reference 10, main text.
4. Alford RM: "Shear Data in the Presence of Azimuthal
Anisotropy: Dilley, Texas," Expanded Abstracts, 56th SEG
Annual International Meeting and Exposition, Houston,
Texas, USA, November 2-6, 1986, paper S9.6.
5. Kimball and Marzetta, reference 8, main text.

the amount or strength of anisotropy. Unlike the


two previous anisotropy measurements (slowness

Spring 1998

45

Aiming for Minimum Stress

Sonic velocity anisotropy was used by


Seneca Oil Company in the USA to determine the most productive position and direction for a horizontal well. A pilot hole for
horizontal drilling was drilled through a
moderate-porosity10 to 15 p.u.shaly
sand interval. The operators objective was to
determine natural fracture orientation and
understand tectonic stress direction (left and
below). DSI logs were recorded for Stoneley,
monopole P- and S-wave and both crossed
receivers (BCR) dipole modes. The BCR mode
was processed for anisotropy. Three different
computations of shear-wave anisotropy are
presented, dealing with the time and energy
of the dipole shear waveform.
First, shear-wave energy anisotropythe
minimum and maximum cross-component
energy differenceis the most obvious indication of anisotropy. Large energy differences, when the minimum stays low,
indicate significant shear-wave splitting and
signal zones of interest. The interval from
12,750 to 12,870 ft contains several zones
with significant anisotropy.

12,800

8. Kimball CV and Marzetta TL: Semblance


Processing of Borehole Acoustic Array Data,
Geophysics 40 (March 1984): 274-281.
9. Sinha B and Kostek S: Stress-Induced Azimuthal
Anisotropy in Borehole Flexural Waves, Geophysics
61 (November-December 1996): 1899-1907.
10. Esmersoy C, Kane M, Boyd A and Denoo S:
Fracture and Stress Evaluation Using Dipole-Shear
Anisotropy Logs, Transactions of the SPWLA
36th Annual Logging Symposium, Paris, France,
June 26-29, 1995, paper J.

12,900

Emax
0

100 5
%

Emin
0

100 0
%

Energy
anisotropy

deg

Fast shear wave azimuth

360 -90

0
deg

+90

Time anisotropy
50

Slowness anisotropy

Caliper
20
in.

Azimuth uncertainty

200
%

Gamma ray
150
API

0
%

Slow-wave slowness
250

50

250

50

s/ft
Fast-wave slowness
s/ft

Shear waveforms
1000

340
330
320

Processing window

Anisotropy evaluation. The difference between the minimum and maximum


cross-component shear energy, shown in the depth track, is an indicator of
anisotropy. The tool orientation (blue), track 2, is used to determine the absolute
fast-shear azimuthal direction (red), with its uncertainty (gray shading), track 3.
The interval between 12,750 and 12,870 ft [3886 and 3923 m] contains several
anisotropic zones. The average of the fast component of the shear-wave direction, shown in the azimuthal projection (inset, right), is between 20 and
30. Acoustic time anisotropy (black with shaded gray) is shown in track 4. This
measurement is more sensitive to acoustic properties deep within the formation
than surface effects such as drilling-induced fractures. Both fast (blue dashed)
and slow (red) components of the shear slowness are computed by STC processing and are shown in track 4. For visual quality control, the fast (red) and slow
(blue) waveforms from the largest spacing receiver are shown in track 5. The
light yellow band shows the shear-wave processing window, which should
include the first few cycles of the shear arrival. Both moveout and energy differences between the fast and slow shear waves are easily visualized and can be
verified on the display. Both waves would be identical in an isotropic formation.

46

5000

310
300
290
280
270

100
350 360 10 20
90
30
80
40
70
50
60
50
60
40
70
30
20
80
10
0
90

260

100

250

0 20 30 40 50 60 70 80 90 100
Frequency, %

Tool orientation
Depth, ft 0

110
120

240

130
140

230
220
210

200

150
190 180 170 160

Oilfield Review

Energy
anisotropy

Smax

Breakout
8

Shear waveforms

6
4
2

Fractures

in.

0
-2
-4

7600
-6
-8
-8

-6

-4

-2

in.

Four-arm caliper projections as qualitative


indicators of stress direction. The zone containing significant anisotropy seen by the
DSI tool also shows hole breakoutsshown
by enlarged hole size measured by the fourarm caliperin this vertical borehole. The
breakouts occur in the direction of minimum horizontal stress. Perpendicular to the
breakout direction is the principal tectonic
stress direction of 25. Bit size is shown by
the green circle, and the borehole walls are
compressed inward slightly along the maximum stress direction. The data inside the
bit size represent the calipers opening up
as they come off bottom.

Second, large shear-wave traveltime


anisotropy indicationsbased on arrival
time of the fast and slow wavescorrelate
well with the energy-anisotropy indications
throughout the interval.
The third anisotropy measurement, slowness anisotropy, is derived from the computed shear-wave, fast- and slow-component
slownessesusing slowness-time-coherence
(STC) processing to obtain slowness.8 This
deep-reading measurement suggests that the
anisotropic properties are from within the
reservoir and not from surface effects such as
shallow drilling-induced fractures.9
The fast shear-azimuth direction for this
interval is between 20 and 30 with respect
to north. This azimuth, aligned with the maximum horizontal stress direction, is the
direction of the current tectonic stress field
and is the orientation of any drilling-induced
fractures. This is also the direction that would
produce the least stable horizontal borehole,
and be least likely to intersect open fractures.
The oriented four-arm caliper summation
data over this zone give a qualitative overall
view of the borehole profile (above left).
Hole breakout is found in two opposing
quadrants, which straddle the axis of maximum horizontal stress. Perpendicular to the
breakout direction, the maximum horizontal
stress direction is found to be 25, which
agrees with the direction found by the fast
shear-wave azimuth.

Spring 1998

Fractures

7700

Fractures

Depth, ft
Offline 0
energy
0
1.0

Gamma ray
API

Fast shear-wave azmuth Fast shear-wave slowness


150 0

180 0

s/ft

Processing window

50 1000

6000

Shear-wave splitting in cased hole. The DSI tool differential-energy measurements,


shown in the depth track, identify the zones with high shear-wave anisotropy caused by
fractures. The difference in the two shear-wave amplitude components (track 5) increases
in the fractured zones.

The operator sidetracked the horizontal leg


of this well at right angles to the maximum
horizontal stress, along the direction intersecting the largest number of natural fracturesfor maximum production.
In another application, shear-wave anisotropy was used by Louisiana Land & Exploration Company in southwest Wyoming, USA
to find fractures in a cased well drilled in a
tight gas-bearing sandstone. Openhole logs
were not run because of poor wellbore conditions.10 However, shear-wave anisotropy

logging was used to find the fractured intervals behind casing (above). The fractured
zones are easily identified from the DSI differential-energy curves shown in the depth
track. Here the maximum in the energy difference between the fast and slow shear
waves quickly identifies three zones
in which large shear-wave velocity anisotropy existsbecause of the fractures. These
zones were perforated, and subsequent production logs show good gas entry from each
zone. This well subsequently produced
4.5 MMcf/D of gas from these perforations.

47

1.8

Measuring Stress with Anisotropy

Finding Faults with Anisotropy

In fracture systems, faults are major events


that impact not only fracture distribution, but
also the rock stresses. Significant rock deformation, fracturing and variations of the stress
fields are found near faults (below right).
An example of fault-induced fractures is
seen in an Egyptian oil-producing well
drilled in granite basement rock.13 Traditional
sonic techniques, such as Stoneley wave
attenuation and reflection were combined
with anisotropy using shear-wave splitting to
evaluate fractures near a fault in this hard
rock formation. These techniques react in different ways to the presence of fractures in the
formations, and their combined analysis
reveals the complex picture of this formation. The direction of the fast shear wave provides information on fracture orientation, but
the shear-wave azimuthal measurements differ somewhat from that of traditional
Stoneley fracture analysis; in particular
shear-wave anisotropy investigates a volume
of formation up to several borehole diameters farther from the borehole and can sense
fractures missed by other techniques.14
Combining measurements provides additional information about reservoir characteristics, and is especially helpful in locating
the fault.

Unstressed

Velocity, km/s

1.6
Fast shear wave
1.4
1.2

Slow shear wave

1
0

10

20

30
40
Frequency, kHz

50

60

1.8
Polarized parallel to stress
Stressed
5 MPa

1.6
Velocity, km/s

One major challenge is to distinguish


between stress-induced and other sources of
shear anisotropy, such as fractures. Currently,
this distinction is difficult to make from sonic
measurements alone, and remains a topic of
ongoing research. Recently, it has been
found that borehole stress concentrations
cause a crossover in the two flexural dispersion curves (right).11 This crossover is caused
by stress-induced radial gradients in the
acoustic-wave velocities that are different in
the two principal stress directions.
Other sources of intrinsic anisotropy
caused by finely layered dipping beds,
aligned fractures or microstructure found in
claysexhibit neither radial velocity gradients nor flexural dispersion crossover.
Consequently, a crossover in the flexural
dispersion curves can be used as an indicator of stress-induced anisotropy. In the presence of stress-induced anisotropy, the
fast-shear direction coincides with the maximum stress direction, and the magnitude of
the shear anisotropy is proportional to the
stress magnitude.12

1.4

Polarized
perpendicular
to stress

1.2
1
0

10

20

30
40
Frequency, kHz

50

60

Dipole dispersion crossover. Laboratory experimental (circles) and theoretical (solid


curves) flexural dispersion curves in Berea sandstone. The plots show the shear-wave
anisotropy velocities measured without stress (top), and with 5 MPa [725 psi] uniaxial
stress (bottom). The effects of stress on shear-wave anisotropy depend on signal frequency and stress magnitude. As stress increases, the shear-wave component polarized
parallel to the stress direction becomes the fastest component at low frequencies. The
reverse is true at high frequencies.

Compressional
stresses

ult

Fa



,,,,,



Prefaulting extensional
fractures

Extensional
stresses

Faults as a major disruption in stress and fracture orientation. A fault can cause a drag
zone in which rock deformation is large. Bending of the beds next to the fault causes
extensional stresses with fractures on one side of the bed, and compressional stresses
with conjugate shear fractures on the other side (inset). Measured stress directions will
change rapidly when a well crosses the fault.

48

Oilfield Review

Gamma ray
0
API 300
Energy
anisotropy

Fault signature
100

Caliper
in.

16

Processing window

Anisotropy
Slowness, %
Time, %

0
100

Shear slowness
Fast shear azimuth
Shear waveforms
Tool azimuth
s
7000
0
0 deg 360 270
+90 450
50 1000
s/ft
Depth, ft
X300
Caliper

Slow
shear
Slowness
anisotropy

Gamma
ray
Tool
azimuth

Fast
shear

Error bar

Maximum
energy

Fa
ul
t

X400

Time
anisotropy

Minimum
energy
X500

Fault found with anisotropy evaluation. Fast shear azimuth (track 3) shows major rotation of the fast shear azimuth from 315 to 20 across a relatively short depth interval centered at X400 ft, the fault location. Anisotropy indicators (track 4) are large below the
fault, indicating small cracks in the rocks, probably caused by high stress in this region.
Waveforms and the processing time window used to determine shear-slowness velocities
are shown in track 5.

The trick is to identify and locate the fracture system associated with the fault zone.
Many wells in the same field did not intercept the fault zone and as a result never
reached commercial production levels.
DSI logging, using monopole P-, S-wave,
Stoneley and BCR modes, was combined
with conventional openhole porosity logging
and FMI Fullbore Formation MicroImager
measurements to locate productive fractures
and their orientations. The FMI fracture orientation data (not shown) agree with the
results from the fast shear-wave azimuth in
the upper zones of the well and suggest a
major tectonic stress in the area oriented in a
NW (320) direction (above).

Spring 1998

The shear-wave anisotropy analysis shows


the fast shear azimuth changing gradually
from 320 to 340 below X325 ft. There is
substantial shear-wave splitting at X300 ft
indicated by the energy, slowness and time
differences. However, there are no indications of major fractures at this depth from the
traditional Stoneley analysis (not shown). The
FMI images confirm the anisotropy indications, showing primarily closed fractures
with their strike oriented at 315.

Major fracture events were recorded


between X380 and X460 ft with Stoneley
fracture analysis, and Stoneley permeability
indicators were strong down to X490 ft.
Porosity is 15 p.u. from X350 to X460 ft and
significantly less above and below this zone.
Shear-wave splitting shows an abrupt rotation of the fast shear azimuth from 315
above X400 ft to 20 below and then back to
0 in the lowest part of this interval. It is
worth observing that at X400 ft, where the
azimuth change is taking place, there is no
significant evidence of anisotropy. However,
shear-wave anisotropy is present above and
below this depth.
All evidence indicates that the well crossed
the fault at X400 ft. The events detected as
fractures in the Stoneley analysis are likely to
be fractures caused by the fault or the fault
itself. Faults are common in granite, and
though faults were identified in a higher section of this well using a vertical seismic profile, VSP measurements were not available in
this interval. However, in shear-wave
anisotropy logging, the fault signature is
clearthe fast shear azimuth starts changing
slowly 70 ft [21 m] above the fault, then
quickly changes by nearly 65 across the
fault and returns to an intermediate value
100 ft [30 m] below the fault until it finally
returns to the regional trend.
Faults typically have large effects on the
producibility and stability of a reservoir and
must be accounted for when completing a
well. The fault seen in this example exhibits
high permeability and is expected to have a
good production potential. The fast shearwave anisotropy indicates the presence of
fractures or stress, and the directional variation of the fast shear azimuth can be used to
detect faults and their associated highpermeability, fractured zones.
11. Winkler KW, Sinha BK and Plona TJ: Effects of
Borehole Stress Concentrations on Dipole
Anisotropy Measurements, Geophysics 63, no. 1
(January-February 1998): 11-17.
12. Sinha BK, Papanastasiou P and Plona TJ:
Influence of Triaxial Stresses on Borehole Stoneley
and Flexural Dispersions, Expanded Abstracts,
67th SEG Annual International Meeting and
Exposition, Dallas, Texas, USA, November 2-7,
1997, paper BH2.2.
13. Endo T, Ito H, Brie A, Badri M and El Sheikh M:
Fracture and Permeability Evaluation in a Fault Zone
from Sonic Waveform Data, Transactions of the
SPWLA 38th Annual Logging Symposium, Houston,
Texas, USA, June 15-18, 1997, paper R.
14. Sinha and Kostek, reference 9.

49

Real-time LWD sonic data in formations


where pressures are either unknown or
known to vary rapidly can be critical.
Knowledge of expected overpressured formations provides the ability to efficiently and
safely drill wells with correct mud weights.15
For example, after there were kicks attributed
to overpressured formations in two offset
wells nearby, real-time LWD overpressure
detection capability was tried in an exploratory well drilled offshore Angola, West Africa.
The 1338-in. casing was set down to X905 ft
and drilled out using a 1214-in. bit with the
sonic-while-drilling tool in the bottomhole
assembly (BHA).
A long single-bit run was conducted over a
7-day period that covered a depth from
X1000 to X8300 ft (below). The wellbore was
deviated 20 in this interval. The BHA consisted of a PowerPak mud motor, CDR
Compensated Dual Resistivity tool,
PowerPulse MWD telemetry systemfor realtime transmissionand an ISONIC sonic-

while-drilling tool. The ISONIC tool, placed


above the PowerPulse system, was approximately 104 ft [32 m] away from the bit. At the
average rate of drilling, the LWD measurements were made fewer than four hours after
the formation was first cut. Wireline sonic logging was run after the 7-day drilling run was
completed, and then only after circulating the
well for several hours.
The gamma ray log indicated that the entire
interval is primarily shale, with sand-shale
sequences dominating the bottom 2000 ft
[610 m] of the formation. A trend of decreasing slowness with increasing depth is
observed in the upper interval from X1000 to
X4800 ft. This trend is normal and caused by
the overburden stress, which compresses the
rock and decreases the porosity with depth.
As long as slowness readings follow this
trend, formation porosities are maintaining a
normal compaction with depth. If overpressured formations are encountered, the slowness data points will diverge from the
expected trend toward abnormally high values (above right).

Depth, ft
B
X2000

Compaction
trend

X3000

X4000
A
X5000

X6000

X7000

X8000
Gamma ray
0

API

Wireline slowness
Attenuation resistivity
Wireline slowness
ISONIC slowness
150
s/ft
50
0.2
ohm-m
20
ISONIC slowness
150 Phase shift resistivity 150
s/ft
50 150
s/ft
50
150
s/ft
50
0.2
ohm-m
20

Detecting overpressure while drilling. The log display shows a gamma ray, CDR phase
and attenuation resistivity and the LWD and wireline sonic slowness log comparisons in a
long 7000-ft [2134-m] bit run. The gamma ray log (track 1) indicates that the entire interval is primarily shale. The wireline and LWD slowness logs are shown in tracks 3 and 4,
respectively, and are shown overlaid in track 5. Real-time LWD sonic readings detected
an overpressured zone 5 hours ahead of expectations, Zone A. There is a consistent difference of up to 10 s/ft [33 s/m] between the ISONIC reading and wireline slownesses in
Zone B attributed to formation alteration near the borehole.

50

Increasing depth

Early Warning While Drilling

Normal
pressure
zone
Entering
overpressured
zone
Normal
compaction
trend

Increasing slowness

Real-time overpressure alert. Formation


compaction, resulting from increasing
overburden weight, causes porosity and
thus acoustic slowness to decrease with
depth. High deviation from the natural
trend is a sign of an overpressured zone.
Real-time LWD acoustic slowness measurement can warn a driller of entry into an
overpressured zone, early enough to allow
mud-weight adjustmentsavoiding costly
damage to the borehole.

This is what happens at about X4800 ft. The


real-time LWD slowness diverges from the
compaction trend, indicating the approach of
potential overpressured formations. Similarly,
the CDR resistivity log shows a departure
from its normal trend of increasing resistivity
with depth toward lower than normal resistivity values, because of increasing porosity
in the overpressured formations and resulting
higher saline-water content in shale. This
interval, showing logging values diverging
from their usual trends, extends down to
about X4850 ft. For the two wells previously
showing kicks, the tops of overpressured formations were found at X5000 ft200 ft
[60 m] below the first sign of overpressure
encountered in this well. The LWD real-time
sonic logs provided warning that the driller
was entering an overpressured formation
hours ahead of expectations.
Another example shows the use of realtime LWD sonic logging in soft formations
to locate a gas-sand pay zone. This vertical
well was drilled by an operator in the Gulf
of Mexico. The ISONIC tool was mounted
in a BHA 77 ft [23 m] behind the bit, and
at the average rate of penetration, the formation was logged less than an hour after
the bit cut through it. The slowness projections on the LWD memory semblance

Oilfield Review

Changing Formation

X100
A
Wireline
LWD

X300

X500
Gas
sand

ISONIC slowness
Density porosity
Deep induction
STC projection
s/ft
80 ISONIC slowness
p.u.
0 0.2
ohm-m
20 180
Depth, ft 60
LSS slowness
Neutron porosity
Medium induction
30
s/ft
210
s/ft
80
60
p.u.
0 0.2
ohm-m
20 180

Real-time gas detection in slow formations. The LWD sonic compressional slowness (red)
shows a large increase (track 4) in Zone B, a gas-sand pay zone. Comparisons with wireline sonic (blue) are generally good, except for a systematic difference in Zone A between
XX50 and X180 ft. This is attributed to stress relaxation in the shales caused by the drilling
process. In the gas-sand pay zoneconfirmed by the classic neutron-density gas crossover
(track 2)the wireline sonic log, track 4, is beginning to show signs of invasion (Zone B)
because its gas signature is not as strong as that of the real-time LWD sonic log.

analysisan indicator of data quality


verify the accuracy of the real-time compressional-slowness log (above). It is clear
that high semblance coherence for the
compressional arrival was obtained
throughout the entire interval. There are
good correlations between the sonic compressional and resistivity curves, especially
below X400 ft, where both are responding
to changes in porosity.
In the sand bed, Zone B, the neutron and
density curves show a large crossovera
classic gas signature. This is supported by the
high-resistivity readings. The compressional
slowness also reads extremely high in this
zonemuch higher than would be expected

Spring 1998

from a normal porosity variationindicating


gas. The gas sand appears to have a thin
shale stringer in the middle, which is also
visible on the porosity and resistivity logs. It
is interesting to note that the wireline sonic
log shows less response to gas in this zone,
particularly in the upper section of the gas
sand above the shale stringer. This is thought
to be due to borehole filtrate invasion gradually depletingover a few daysthe volume of gas in the annular zone surrounding
the borehole, thus decreasing the gas signature seen by the wireline tools. Details such
as these, seen in many LWD and wireline log
comparisons, help show the value of early
real-time answers.

In the previous offshore Angola example,


real-time sonic LWD data helped the operator safely drill the well. However, subsequent
comparison of the LWD logs with their wireline counterparts sparked the curiosity and
interest of log interpretation experts.
Although LWD and wireline logs agree in
many wells, experience has shown that
sometimes there are zones, especially in
shales, where the slowness measurements
can differ. In the upper interval, Zone B, of
the offshore Angola well, the LWD slowness
log reads consistently lower than the wireline log by 5 to 10 s/ft [16 to 33 s/m].
Everywhere else, the agreement between the
LWD and wireline sonic logs is excellent.
The agreement in most of the well indicates
that the differences are not caused by the different processing methodsSTC processing
was used for LWD; and for wireline, a firstmotion-detection scheme (FMD), which is
triggered by waveform amplitudes exceeding
a selected threshold.
Other explanations were sought. Sonic
measurements are known to be affected by
borehole washouts. These borehole irregularities can cause the measured slowness
values to oscillate around the true formation
slowness values. Although the LWD log was
not borehole-compensated and the wireline
was, the systematic difference between the
two observed measurements was probably
not caused by washouts.
First, the borehole should be in excellent
condition during drilling, thus making the
need for borehole compensation minimal.
This is particularly true since this borehole
was drilled with oil-base mud, which minimizes washout problems. Second, the consistently faster trend of the LWD slowness
values conflicts with the expected effect of
washouts, which make the uncompensated
slowness values fluctuate around the compensated wireline measurements. Because
the faster trend of the LWD sonic is persistent, the absence of borehole compensation
is probably not the cause of the discrepancy.
15. Hsu K, Minerbo G, Hashem M, Bean C and Plumb
R: Sonic-While-Drilling Tool Detects Overpressured
Formations, Oil & Gas Journal 95, no. 31 (August 4,
1997): 59-63, 66-67.

51

X1,000

Borehole
Altered shale

Sonic
logging
tool

Undamaged
formation

X1,200

X1,400

Bicompressional
arrival. Water takeup in certain shales
or stress relief near
the borehole can
change the elastic
moduli of the annular rock. The altered
zone traps wavefronts in the same
way as the borehole
does. The extraneous second arrival is
shown leading to a
second compressional wave
trapped in the
altered zone.

X1,600
Phantom
arrival
LWD
Wireline

Altered shale
compressional
arrival

Virgin formation
compressional
arrival

X1,800

X2,000
Wireline slowness
STC projection
Waveform VDL
180
s/ft
80
Depth, ft
ISONIC slowness
ISONIC slowness
500
s
3500
s/ft
200
180
s/ft
80 50

Time-lapse logging using sonic while drilling. Formation alteration is taking place in
the 6 to 7 days between the running of the LWD sonic and the wireline measurement.
Details in both slowness logs seen in track 2 correlate closely, but there is a systematic
shift to larger slowness in the wireline log. The reason for the shift can be seen in the
slowness-time-coherence (STC) projections from the ISONIC tool, track 3, where a second slower coherent arrival is emerging. This phantom arrival, too fast to be a shear
arrival in these slow formations, is likely to be caused by shale alteration.

It is clear that detailed features of the two


slowness logs in the upper shaly zone are
correlated (above left). However, the consistent difference between the two measurements in this interval and the STC projection
derived from the ISONIC recorded waveforms in track 3 provides clues to what was
happening. Specifically, the STC projection
shows that there is a second coherent arrival
becoming apparent about 10 to 20 s/ft [33
to 66 s/m] slower than the first arrival. This
second arrival is too fast to be the shear in
such a slow formation.
A slow-formation phenomenon known as
an altered zone provides a possible explanation for the second arrival. An altered zone is
created by the drilling process which may
damage the borehole wall and cause the
elastic modulus in an annular zone around
the borehole to change. This is particularly
true in soft formations. Water take-up in certain shales and shaly sands or stress relief
near the borehole can also change the elastic modulus of the annular rock, generally
reducing it. The altered zone traps wave-

52

fronts in the same way that a borehole does.


The extraneous second arrival in this exampleknown as a bicompressional wave
corresponds to a compressional wave
trapped in the altered zone (above right). In
North American test wells, many early wireline digital sonic field tests recorded bicompressional waveform arrivals.
The time delay between the LWD sonic
measurements and the wireline sonic measurements in the altered zone is about 6 to 7
days. The presence of a well-developed
altered zone can cause the wireline tool to
read slowness values somewhere between
the altered and unaltered formation-slowness values, depending on the type of tool
and spacing used. In this well LWD sonic
measured the formationonly hours after
rock was drilledbefore any significant
alteration had occurred.
With all the evidence collected, the operator concluded that the slowness difference
seen while drilling and at wireline time was
likely caused by changed formation properties over timea time-lapse effect. Confident
in the accuracy of the sonic-while-drilling
results and realizing the potential for real-

time overpressure detection, the operator


continued using the ISONIC tool in its highprofile wells in a neighboring block, later
making a major discovery there.
In cases like this when a systematic difference is observed between LWD and wireline
logs taken some time after drilling was completed, it usually happens in shaly zones.
Experience shows that the two logs generally
agree in hard rock and clean sands. The
slowness difference can be expected to
increase with time, due to shale swelling. In
these cases, the LWD slowness would
always be less than the wireline slowness.
In some wells the opposite occurs. A well
drilled in a plastic shale can lead to borehole
deformation due to compaction from high
vertical stress. In such cases, the high mean
stresses produce a decrease in porosity and a
stiffening of the bulk-frame modulus.16 The
overall effect of this consolidation phenomenon can decrease the shales slowness
with time, in an annular zone a foot or more
away from the borehole.17 The distance into
the formation will depend on formation permeability, magnitude of the pressure difference, rock properties and the time since the
well was drilled. In such cases, the LWD slowness can be larger than the wireline slowness.
This phenomenon may have occurred in the
previous Gulf of Mexico shaly-sand example.
In this well, the wireline LSS Long-Spaced
Sonic Tool slowness shows overall good
agreement with the real-time sonic slowness.
However, close examination shows that in the
upper shale interval, Zone A, the LWD slowness tends to read 3 to 5 s/ft systematically
larger than the wireline slownessagain indicating shale consolidation near the borehole,
caused by the drilling process.

Oilfield Review

Since the LWD measurement has the


potential of determining unaltered formation
slowness, it has significant implications for
two seismic applications, specifically, the
time-depth relationship and the synthetic
seismogram. Seismic waves, with their long
wavelength, see the bulk unaltered formation and are insensitive to small features such
as the borehole region. However, wireline
sonic logs are frequently affected by formation alteration, and under those conditions
they do not provide the most appropriate
velocities for comparison with seismic data.

Stoneley waves
Shear waves
Compressional waves
Drilling noise

XX900

X1000
Compressional
slowness

X1100
Shear
slowness

Hard Rock LWD

Although hard rock formations are considered routine for wireline sonic logging
because the formation signals are fast and
separate from the borehole mud signal
they present a major challenge for LWD
measurements. Experience indicates that the
level of drilling-induced noise is high in hard
formations, and under these conditions,
downhole waveform processing requires
special care.
In such wells, waveform stacking is indispensable. The stacked waveforms in the
ISONIC slowness tool are filtered with a narrow band-pass filter, tuned to frequencies at
which drilling noise is minimal and drill collar arrivalsnoiseare highly attenuated.
16. This situation would occur if the drilling fluid pressure was less than the pore pressure. Also, because
of the high ionic strength of the water phase in some
drilling mud (200 to 300 kppm calcium chloride),
the mud can pull water out of the formation, particularly the shales. This can lead to embrittlement of
the formation, and possible fracturing.
17. Hsu K, Hashem M, Bean CL, Plumb R and Minerbo
G: Interpretation and Analysis of Sonic While
Drilling Data in Overpressured Formations,
Transactions of the SPWLA 38th Annual Logging
Symposium, Houston, Texas, USA, June 15-18,
1997, paper FF.
18. Aron J, Chang SK, Codazzi D, Dworak R, Hsu K,
Lau T, Minerbo G and Yogeswaren E: Real-Time
Sonic While Drilling in Hard and Soft Rocks,
Transactions of the SPWLA 38th Annual Logging
Symposium, Houston, Texas, USA, June 15-18,
1997, paper HH.
19. The empirical time-averaged Wyllie model is a
sonic-porosity interpretation mixing law that works
in many low-porosity applications. It is a linear volumetric mix of the formation and fluid slownesses. It
does not account for the fact that real rocks are not
homogeneous, nor microscopically continuous, or
isotropic and not ideally elastic.
20. The first significant model of sonic properties
in fluid-filled porous rock was developed by
Gassmann. See Gassmann F: Elastic Waves
Through a Packing of Spheres, Geophysics 16,
no. 18 (1951): 673-685. Then Biot developed a
more complete model by allowing relative motion
between the fluid and rock matrix. See Biot M:
Theory of Propagation of Elastic Waves in FluidSaturated Porous Solids, Journal of the Acoustical
Society of America 28 (1956): 179-191. The Biot
equations converge to those of Gassmann at low
frequencies. For a general review of sonic properties
in fluid-filled porous media, see Ellis D: Well
Logging for Earth Scientists. New York, New York,
USA: Elsevier, 1987.

Spring 1998

X1200

X1300

Limestone bed

Depth,
m
0

Wireline comp.
slowness
ISONIC
200
s/ft
50
waveform VDL
Gamma ray
LWD comp. slowness
s
3400
s/ft
50 400
200
API
LWD shear slowness
150
200
s/ft
50

The stacked and filtered waveforms are then


archived in memory, while the downhole tool
microprocessor and digital signal processor
determine the slowness of the coherent
arrivals in the waveforms using STC processing and selecting the arrival that corresponds
to the compressional slownessknown as
labeling. Once the compressional slowness is
extracted downhole, it is transmitted to surface via the mud telemetry system. The
ISONIC slowness log is generated at surface
in real time as the tool passes the formation.
An example in a highly deviated North Sea
well shows the current capability of LWD
real-time sonic logging in hard rock
(above).18 In this well, the ISONIC logging
tool was attached 114 ft [40 m] behind the
drill bit, which put its measurements more
than 4 hours after drilling. The recorded
waveform variable density log (VDL) shows
the good compressional signal, followed by
a strong shear and even stronger Stoneley
arrivals. Some residual drilling noise is visible in the harder sections of the formation.
The compressional slowness from the LWD
matches the wireline results throughout this
well. Many fast carbonate stringers correlate
between both sonic logs and the gamma ray,
all dipping in value as the clean hard
stringers are passed by the logging tools. In
the lower limestone zone at X1300 m, a
Wyllie time-averaged porosity of about
7 p.u. is computed for this limestone bed.19
These results enable the operator to determine formation rock properties for planning
a hydraulic fracturing program.

LWD Sonic in fast


formations. The ISONIC
tool recorded waveforms
plotted in the VDL display
in track 3 show a compressional arrival near 1150
s, a strong shear near
1700 s and a stronger
Stoneley arrival near 2500
s. Compressional (red)
and shear (green) slowness
values are displayed in
track 2. The LWD compressional slowness matches
the wireline results (blue)
throughout this interval.
The formation slowness
and the amplitude of the
compressional arrival
correlate, especially just
below X850 m, where
the formation slowness
drops from 95 to 80 s/ft
[312 to 262 s/m] and the
A
amplitude of the compressional arrival increases
by about a factor of two.
In the fasterharder
portions of this interval,
Zone A, increases in
drilling noise are visible on
the early part of the VDL
waveforms, just before
the compressional arrival.

Petrophysics: Sonic Logs in Gas Sands

After the structure of the reservoir is


knowntraps are located, faults are mapped
and fractures and their orientations identifiedpay zones must be quantified (see
"Localized Maps of the Subsurface, page
56). Hydrocarbon identification, and determination of porosity and saturation are
important petrophysical applications for
sonic logging. The Biot-Gassmann theory is
widely used today to relate wet rock to dry
rock frame properties for sonic applications.20 With the introduction of dipole sonic
logs, and the ability to record high-quality
shear and compressional slownesses for the
first time in slow formations, trends can be
identified in sands and shales and matched
with semi-empirical correlations based on
Biot-Gassmann theory. These trends can be
used as a quality-control check on sonic logs
and for quicklook lithology interpretation.
However, the effects of partial saturation
pose special problems. Gas and liquid mixtures in soft formations make the interpretation more complicated, as these can affect
the sonic slowness significantly, in particular
the compressional slowness. The compressional-to-shear velocity ratio, Vp/Vs, has
been used in unconsolidated sands to qualitatively detect gas.
The fluid distribution in the pore system
at the microscopic level and the acoustic
frequency have a strong influence on
the strength of the gas effect on acoustic

53

Sonic
6

104

que

ncy 10 102
,H
101 100
z

80

60

Gas

20

40

,%
ation

3.00

satu

Relationship of compressional velocity


to frequency and gas saturation. Seismic
frequencies lie near the front of this plot,
which shows an abrupt change in velocity
with only a few percent increase in gas
saturation. Sonic logging measurements
are in the center of the graph, and show
a more gradual increase in velocity as gas
saturation increases.

slowness. Although the Biot-Gassmann


model together with Woods mixing law
have been successfully used at seismic frequencies for geophysical interpretation, they
give deceptive results in partially gas-saturated formations at sonic frequencies.21 The
Biot-Gassmann model itself is not the problem. The effect of gas on elastic-wave slowness has been observed by a number of
authors, and the problem lies in determining
the effective modulus of the pore-fluid mixture (above).22
To evaluate gas volume from sonic logs in
shaly sands, a new mixing law for computing the average sonic properties of gas-saturated fluids has been proposed recently.23
Kliquid

K f (Fluid bulk modulus)

K f = ( Kliquid - Kgas) S xo + Kgas

e=1
2
3
5
10

Kgas

100

Sxo (Invaded zone liquid saturation), %

Pore-fluid mixing laws. The observed


pore-fluid bulk modulus Kf changes from
Kliquid to Kgas in a nonlinear manner as
liquid is replaced by gas in the formation.
Field data suggest an average mixing-law
coefficient, e, of between 2 and 5. The
curves are scaled according to liquid saturation in the invaded zone, Sxo, because the
sonic properties of oil and water are similar,
and sonic measurements are mostly sensitive to fluid present in the invaded zone.
For high values of the mixing coefficient
(e 40), the mixing law behavior
approaches that of Woods mixing law.

54

Porosity,
p.u.

2.50
ale
s

Fre

2.0

ic

Seism

Sonic crossplot. Wet


formations, such as the
Wet Sand trend (blue
curve), show an increase
in their Vp/Vs velocity
ratio due to decreasing
shear velocity, caused
by fluid-rock coupling.
Shales (green curve)
show even greater effect.
The new mixing law for
computing the effective
fluid-bulk modulus of
gas-saturated formations
successfully explains
the observed trends in
logging data.

Sh

105

Unconsolidated
sediments

North Sea
shales
Malaysia
shaly sands
North Sea
water sand

Vp / Vs

10

Vp, km/s

3.50
2.5

30
90
2.00
20

10

Invaded
80 zone
70 fluid
saturation
60
50
Gas
40

Dolomite
Limestone
Wet sands
Shales
Dry or gas
sandstones
Anhydrite
Salt
Quartz

1.50
40

100

180

Compressional slowness, s/ft

The new mixing law differs from Woods


mixing law in that it relates the effective fluid
bulk modulus to the liquid and gas moduli
through a power law function of the liquid
saturation (left). The interpretation proceeds
straightforwardly from here. The apparent
pore-fluid bulk modulus can still be computed from the Biot-Gassmann model, which
relates it to the formation velocity ratio measured by the sonic logging tool; and the
porosity can be derived from neutron and
density logging. To derive the gas saturation,
the apparent fluid modulus is compared with
the one computed by the new mixing law
based on the liquid saturation in the invaded
zone. The crossplot of the velocity ratio versus compressional slowness illustrates how
gas-bearing formations are differentiated
from liquid-filled formations (above).
The Biot-Gassmann model along with
openhole density, porosity and lithological
interpretations are used to derive the
expected dry formation properties, which
enable computing fully liquid-saturated formation moduli and slowness values. These
results can serve as an effective quicklook gas

indicator on the log plots by comparing them


with the measured slowness values. The dry
compressional and shear slowness curves are
what the logging values would be if the formation was completely filled with dry gas.
Similarly the wet slowness curves are what
the logs would read if the formation was
completely filled with liquidoil or water.
The quicklook gas saturation is computed
from the difference between the derived wet
slowness and observed slowness.
These dry formation parameters also find
application as inputs to rock mechanics computations. These properties are independent
of the pore fluid in the rock, and therefore
provide a more reliable basis for rockstrength estimation than the moduli computed directly from the measured slowness.
Finally, the dry formation properties can be
used to estimate the acoustic response for any
pore fluid mixture as an input to amplitudeversus-offset (AVO) modelingknown as
fluid substitution.24 Woods mixing law,
which has a more dramatic effect from gas
saturation, should be used at seismic fre-

Oilfield Review

quencies because seismic compressional


velocities are more sensitive to gas saturation.
In practice, when gas is present, the dry
acoustic parameters provide a suitable
approximation to the seismic response.
An example in a well drilled through a shaly
sand formation with oil-base mud shows the
capability of sonic logging to quantitatively differentiate between liquid and gas (right). The
logged compressional and shear slownesses
fall between the expected dry-gas and wetliquid limits. Porosities are high, averaging
33 p.u., in the interval shown, and the compressional slowness shows a large gas separation, often as much as 35 s/ft [115 s/m],
which is characteristic of these soft formations. Hard formations show less separation.
The gas effect seen on the shear slowness is
small. This is expected because the shear
modulus is independent of pore fluid in the
Biot-Gassmann model, and the effect on
shear slowness is coming from density alone.
Over most of the interval shown, gas is
clearly seen in the sonic interpretation
except in the 5-m [16-ft] zone at X100 m,
where both compressional and shear slownesses agree with the expected wet values.
The traditional ELAN Elemental Log Analysis
openhole interpretation, using resistivity and
EPT Electromagnetic Propagation Tool data,
indicates hydrocarbons in this zone. This is
the only place where there is marked disagreement between the sonic logging- and
the resistivity-based interpretation. However,
this zone consists of fine-grained silt, and the
resistivity interpretation using Archies equation did not take into account the effects of
high irreducible water saturation associated
with silt in this zone, which leads to an incorrect saturation result.
Sonic compressional-slowness logs are
mostly sensitive to the fluid present in the
invaded zone. Compressional waves follow
refraction laws, and when the invaded zone
is faster than the deep formationas in a gas
zonethen only the invaded zone is seen.
The slower, gas-filled, virgin zone lying deep
in the formation is not measured because the
21. Woods law is frequently used in many geophysics
applications to compute the effective fluid complianceinverse bulk modulusfrom a linear volumetric mix of the gas and liquid compliances.
22. Murphy W, Reischer A and Hsu K: Modulus
Decomposition of Compressional and Shear
Velocities in Sand Bodies, Geophysics 58, no. 2
(February 1993): 227-239.
23. Brie A, Pampuri F, Marsala AF and Meazza O:
Shear Sonic Interpretation in Gas-Bearing Sands,
paper SPE 30595, presented at the 70th SPE Annual
Technical Conference and Exhibition, Dallas, Texas,
USA, October 22-25, 1995.
24. Chiburis E, Frank C, Leaney S, McHugo S and
Skidmore C: Hydrocarbon Detection with AVO,
Oilfield Review 5, no. 1 (January 1993): 42-50.

Spring 1998

Slowness, s/ft

350
Depth,
m

50

Dry Modulus, GPa


0
20

XX50

Shear
slowness

Wet

Sonic
Volume

ELAN
Volumes

50 p.u. 0 0
Compressional
slowness

100%
Clay

Dry

Gas

Log

Log

Quartz
Gas

Dry

Moduli
G
X100

Wet

Water

Liquid
Kdry
Feldspar

Total
porosity

Gas evaluation in shaly sands. The compressional and shear slowness logs are shown
in track 2, and compared with the dry and wet slowness predicted by the BiotGassmann model for fully gas- and water-saturated formations. The difference between
the expected wet and observed compressional and shear slownesses are used to determine gas volume, shown in track 3. Dry bulk and shear moduli of the rock shown in
track 2 can be used in rock mechanics applications. The ELAN interpretation, track 4,
shows fluid and formation volumes.

compressional wave in this slower formation


is refracted away from the wellbore. The
important conclusion is that fluid analysis
obtained from the sonic logs represents
invaded-zone saturation.
These sonic interpretation techniques are
implemented in the PetroSonic module of the
GeoFrame interpretation program. They can
be used to evaluate hydrocarbon saturation
from sonic slowness, remove pore-fluid
effects and analyze frame properties for rock
mechanics evaluation, and replace the pore
fluid present at logging time (invaded zone)
with another fluid combination, either to reproduce the original reservoir conditions, or to simulate other situations for AVO modeling.

Sonic measurements make important contributions to our knowledge of a field during every phase of reservoir lifedrilling,
completion, formation evaluation, stimulation, even fluid characterization and monitoring. Sonic waves dont solve every
problem, but when combined with other
measurementsneutron-density for gas
identification, FMI measurements for fractures or CMR Combinable Magnetic
Resonance Tool readings for permeability
they strengthen our understanding of the
subsurface and ultimately help find and
produce hydrocarbons more efficiently.
RCH

55

You might also like