SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

FIELD TEST RESULTS OF AN EXPERIMENTAL FULLYTRIAXIAL INDUCTION TOOL

Richard Rosthal, Tom Barber, Steve Bonner, Kuo-Chang Chen, Sophia Davydycheva, Gary. Hazen, Dean
Homan, Charles Kibbe, Gerald Minerbo, Robert Schlein, Laurent Villegas, Hanming Wang, and Feng Zhou

Schlumberger

ABSTRACT
A new fully triaxial induction tool has completed a
series of field tests. This tool offers two significant
technical advances. First, it features triaxial
transmitter and receiver antenna modules. Each
module includes an antenna with an axially
oriented magnetic moment and a pair of orthogonal
transverse antennas. The ability to collocate all
orientations of the transmitter or receiver antennas
greatly simplifies the analysis of the data and
eliminates the problems of depth errors between
the measurements. All nine couplings between
transmitters and receivers are measured, giving the
most complete data set possible.
Second, the tool is designed to minimize
the borehole effect for the transverse couplings.
For an induction tool with transverse coils enclosed
in a nonconducting sleeve, the borehole effect in
waterbase mud can be orders of magnitude larger
than for a conventional induction array. In addition,
the borehole effect depends in a complicated way
upon the borehole geometry and formation
resistivity far from the location of the tool, making
it very difficult to correct. Understanding the
physics of this effect has allowed us to develop a
tool that provides a measurement with greatly
reduced and simplified borehole effect. The
remaining borehole effect is comparable in size to
that of a conventional tool with axially oriented
magnetic moments and it depends upon the
borehole and formation only near the tool location.
Experimental and modeling results show excellent
agreement and demonstrate the effectiveness of the
design.
This combination of collocated antennas
and handling of borehole effects allows the tensor
resistivity to be robustly computed at any dip angle
from all nine couplings of the triaxial arrays.

Modeling studies and comparison of field

logs with modeled results show the understanding
of the tool physics in complicated situations
involving dip, invasion and formation anisotropy.
Results from a very fast one-dimensional inversion
show the ability of the tool to determine formation
dip, bedding and tensor resistivity in an actual
formation.
WHY BUILD A TRIAXIAL TOOL?
For over fifty years, induction tools have been a
key measurement in well logging. In addition, the
idea of a triaxial induction tool is not new [Huston,
1970]. Conventional induction tools use coils that
lie in planes perpendicular to the tool axis. These
coils have magnetic moments with their axes along
the tool axis. Now companies are introducing
induction tools with coils oriented in other
directions [Kriegsh~iuser, 1999, 2000, Mollison
2001]. We may reasonably ask "What's wrong
with conventional tools?"
The answer, of course is nothing m so far as
they go. Conventional induction tools provide data
that lead to good interpretations of the formation
parameters. However they do not give as complete
an interpretation of the formation as is available
from a triaxial tool. In particular, the triaxial tool
provides data to improve interpretation of:

Formation anisotropy - - A conventional

induction tool run in a vertical well is sensitive
only to the horizontal resistivity of the
formation. A triaxial tool is sensitive to
anisotropy at any dip.

Formation dip and cross-bedding m While not

able to provide results as accurate as a
conventional
dipmeter,
a
fully-triaxial
induction tool can give good indication of
formation dip and dip azimuth. In formations
where the anisotropy directions do not lie
parallel to the bedding planes, we can obtain
an indication of the direction of the anisotropy.

t)Q

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

Fracture identification - - A fully-triaxial

induction tool can identify the presence and
orientation of fractures, providing important
information about formation stress directions.
This sensitivity is strong for fractures filled
with hydrocarbons.

Identification of formation geometry - - A

fully-triaxial tool can help to detect the
presence and direction of nearby beds. In
addition, it may give an indication of
pinchouts or directional invasion.
Note that all of these advantages are potentials with
a triaxial tool. The analysis required to actually
achieve them is in development.
One of the most important aspects of a
triaxial tool is the ability to determine formation
anisotropy in wells drilled at arbitrary angles. At
present, we are able to detect and quantify
formation anisotropy with LWD propagation tools
at high relative dip angles. This is due to the fact
that the currents due to such a tool at high dip
angles have a horizontal and a vertical component
and so the received voltage depends upon both
horizontal and vertical formation conductivity.
For point dipole transverse transmitters
and receivers in a homogeneous medium,

Vxx =

i colt
4~-r 3

ik h r

(- l + ikhr + [kh 2 + k 2 ] r 2)

Vzz

= ~ coil

2m.3

ik h r

(1 - i k hr)

k h - 4io)fltTh
k - a/icoltcr
Thus, the coupling Vxx between two transverse
dipoles is dependent upon both horizontal and
vertical conductivity, while the coupling Vzz
between two axial dipoles depends only upon the
horizontal resistivity. This is later statement is
obvious, since the currents that flow in a formation
from a vertical dipole lie entirely in the horizontal
plane.
T O O L DESCRIPTION
A simplified tool diagram is shown in Figure 1.
This tool is an experimental prototype and as such

was designed in order to test various concepts to be

applied to a full commercial tool. It was built using
as much of the commercial AIT-H* (Array
Induction Tool series H) electronics and sonde
mechanical construction as possible. The AIT-H
has a single transmitter and eight conventional
receivers. In order to save time, the experimental
prototype was limited to eight receiver channels.
We determined early on in the tool design
process that it would be advantageous to use
collocated triaxial coils for both transmitter and the
receivers. That is, each sensor consists of a
conventional z-coil, as well as two transverse coils
to form an orthogonal set, all located at a single
position. All nine couplings between the
transmitters
and
receivers
are
measured
simultaneously. This greatly simplifies the analysis
of data, since the change from one coordinate
system to another involves a simple coordinate
transformation. There are no problems caused by
having different couplings measured with different
spacings or with having them measured at different
depths.
While having all the sensors located at the
same point simplifies the interpretation of the data,
it significantly complicates the construction. In this
prototype, we have a triaxial transmitter and two
triaxial receiver arrays. The long array has a main
coil placed 39" from the transmitter and a bucking
coil placed 27" from the transmitter. The short
array has spacings of 27" and 21". Thus the sensor
array located at 27" consists of six collocated coils,
the main coils for the 27" array as well as bucking
coils for the 39" array. The construction of this
assembly and the ability to align and mutually
balance all the different sensors was a main
objective of this prototype. The tool also includes a
conventional short 9" array. The final measurement
is a current measurement that is associated with our
approach to eliminating much of the borehole
effect.
BOREHOLE EFFECT
The idea of building a triaxial induction tool is not
new. One of the major problems with this tool is
the large and complicated borehole effect.
Borehole effects have always been a problem with
induction tools, but the magnitude and character of
the borehole effect for a transverse coil induction
tool is much worse than anything that occurs for a
conventional tool. This effect arises particularly

Mark of Schlumberger

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

when the tool is eccentered in a conductive

borehole.
In order to kill the borehole effect, it is
important to understand the physical phenomena
involved. Consider a transmitter and a receiver
both with their axes in the x-direction,
perpendicular to the borehole. If this tool is
centered in a borehole in a homogeneous
formation, the currents will flow through the
formation in roughly circular paths. When these
currents get near the borehole, they will be
distorted and this will cause a change in the voltage
in the receiver. This case is the benign one and is
not much different in size or character than the
borehole effect for a conventional induction tool.
When the tool is eccentered in the ydirection (perpendicular to the direction of the
dipole moment), the situation is dramatically
different In this case; large currents are induced in
the borehole and flow up and down the borehole.
For conductive boreholes and resistive formations,
this current falls off very slowly as we move away
from the transmitter. An approximation to the
falloff can be given by the "telegrapher's equation"
which states that the current will falloff
exponentially with a characteristic length which is
proportional to the square root of the ratio of the
formation resistivity divided by the mud resistivity.
I ~: e -z/z

where

z0

,Of = formation resistivity

Pm=mud resistivity
When the formation is resistive and the mud is
conductive, this means that the falloff for the
current is very slow. In fact, we can even have the
unusual phenomena that the borehole effect for the
longer arrays can be greater than the borehole
effect for the shorter ones.
A second unusual characteristic of the
borehole effect is that it is non-local. That is, the
borehole effect does not just depend upon the
characteristics of the mud, borehole and formation
near the tool. Instead, it may depend upon the

formation at distances far removed from the tool.

This arises because the size of the currents in the
borehole depends upon the formation and borehole
character all along their paths. For instance,
suppose that the tool is in a thin resistive
hydrocarbon zone with conductive shale above and
an oil-water contact below. In this case, the current
in the borehole may travel up the borehole and leak
out through the shale and water zones. This current
could be much larger than it would be in a
homogeneous medium.
This current produces a borehole effect
which is more like that of an unfocussed galvanic
tool such as a normal or a lateral than it is like that
of a conventional induction tool. The various
designs of laterolog tools minimize their borehole
effects by canceling the current that would
otherwise flow up the borehole. We have
developed a tool design for a triaxial induction tool
that kills this current in much the same way. It has
significant conductive elements on the body of the
tool. Currents that would otherwise flow up the
borehole are controlled and are made to flow
through the tool. Figure 2 shows the current
patterns for an insulating body tool and for a
conductive body tool.
The remaining borehole effect for this
conductive tool is comparable in size and behavior
to that from a conventional induction tool. It is
much smaller and it declines for longer arrays.
Critical to the ability to correct for the remaining
borehole effect is that it is local, depending only
upon the formation and borehole characteristics in
the immediate vicinity of the tool.
To confirm our intuition and also to verify
that we could model this effect, we began by
building a transverse induction tool with a metal
body. Only the areas immediately about the
transmitter and receiver coils were insulated. This
simple experimental model could then be covered
with a thin insulating layer (actually a length of
shrink-wrap tubing) to test the difference between a
tool with a conventional (insulating body) design
and one with a conductive body. Figure 3 shows
the borehole effect for a conventional tool together
with that for a metal-body tool. The difference
between the borehole effect for a centered tool and
for one fully eccentered is reduced by a factor of
more than 100 by using a conductive tool body.
A method known as "frequency focusing"
has been proposed to cancel the borehole effect
from a transverse induction tool [Forgang, 1998].
This method utilizes the difference between the
tool responses at two frequencies (typically

QQ

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

differing by a factor of three) instead of a single

frequency. Among the problems with this method
is the great amplification in noise that comes from
this procedure. The difference in response of a tool
at two frequencies is of the same order of
magnitude as the X-signal. Thus frequency
focusing requires measuring differences between
two R-signals at two different frequencies, which is
of the same order of magnitude as the X-signals.
For high resistivity formation, it is very difficult to
do this accurately.
In addition, frequency focusing gives a
very complicated behavior in more complex
formations and leads to a borehole effect that is
non-local. That is, the borehole effect depends in a
complicated
manner
upon
the
formation
characteristics far from the tool and not just in the
immediate vicinity. This is shown in figure 4 where
a 27" tool is placed in the middle of a 12' thick, 50
f2-m (20 mS/m) bed with varying shoulder bed
resistivities. The tool is eccentered in a borehole
and we have plotted the difference between the
eccentered tool response and the centered response.
The insulating tool has a borehole effect of 500
mS/m, or 10 times the signal from the formation.
This shows the large difference between centered
and eccentered response that we observed in the
experiments. By comparison the conductive tool
response is small and flat, showing that the
conductive tool has much smaller sensitivity to
eccentering than does the insulating tool. The
frequency focused result is quite interesting. On the
left, when the shoulders have the same
conductivity as the bed (homogeneous medium),
frequency focusing does indeed reduce the effect of
eccentering considerably, although not quite as
much as the conductive tool does. However, when
the shoulders have a different resistivity than the
bed, frequency focusing doesn't work as well. For a
1000 mS/m shoulder, the additional borehole effect
is 51 mS/m, or more than twice the signal from the
formation.
While the actual tool is not built about a
conductive body, it does have carefully designed
conductive elements to control the borehole
currents and improve the response. The two triaxial
arrays and the uniaxial array take up seven of the
eight available measurement channels. The eighth
channel is used to measure the current flowing
through the tool between the conductive elements
placed above and below the transmitter.
M O D E L I N G RESULTS

Let's begin examining the tool with some modeling

results. Figure 5 shows all the couplings for a fullytriaxial tool in the Oklahoma 2 formation at 0 and
60 . This is a well-known formation that has been
used to benchmark resistivity tools for years. It
consists of a number of thin beds with high
contrast. For this example, there was no anisotropy
in the model. All the couplings are shown in a
linear conductivity presentation. This is necessary
because the transverse couplings (XX and YY) are
often negative. In fact, at each bed boundary, there
is an overshoot, with the apparent conductivity
reading low (often negative) on the high
conductivity side and high on the low conductivity
side. If nothing else, this provides good, robust data
to determine the location of bed boundaries unlike
the conventional ZZ coupling which merely has an
inflection point at the bed boundary. For a vertical
well, the symmetry of the formation insures that
there are no cross-coupling terms.
At a 60 relative dip there are cross-terms.
We have chosen the y-axis to be the one is along
the relative formation strike. That is, it is the tool's
y-axis is perpendicular to the axis of the tool and
lies in the bedding planes. In this case, the crosscouplings with Y are all zero.

V-

g~

0
This is the simplest form for the couplings that can
be written for this situation. In a real borehole,
where the rotation angle of the tool is arbitrary, the
coupling matrix will be full. However, since we
have a fully-triaxial tool, we can simplify it into
this form with a simple rotation. The rotation angle
determines the strike of the formation.
The curves show considerable activity and
behave in ways that are very different from what
we have become used to for conventional induction
tools. The ZZ couplings now have overshoots
("horns") at the boundaries, as we have come to
expect in high-contrast formations with dip. The
overshoots on the XX and YY coupling are now
different from each other. In addition, the
overshoots on them are smaller than they were at
zero dip. The behavior of the cross-terms is
fascinating. The XZ and ZX couplings generally
have opposite signs. When the tool is above a
conductive bed, the XZ coupling is positive and the
ZX coupling is negative. Conversely, when the tool
is above a resistive bed, the XZ coupling is
negative and the ZX coupling is positive. Thus this

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

coupling
provides
directional
information
concerning the location and conductivity of nearby
beds.
It is obvious from these examples that the
interpretation of the data from a triaxial tool will
not be simple. The addition of multiple arrays and
multiple frequencies will only exaggerate this
problem. In general, a meaningful interpretation of
the data will require some type of inversion
processing.
FIELD EXAMPLE
Our prototype tool has been field tested in a
Schlumberger test well and has been run in a
number of client wells. An example of the results
obtained is shown in Figure 6. This was a
sandstone formation with very little formation dip.
The well was drilled as a vertical 7 7/8" borehole
with fresh (= 2f2-m) mud. The pay in this well
consisted of several gas zones varying from 5' to
30' thick. Most of the well was moderately shaly
with resistivities from 5 to 10 f2-m. The pay zones
were cleaner and varied between 10 and 100 f2-m.
All the cross-terms are small, as we would expect
in a vertical well with flat bedding. Notice that the
XX and YY coupling overlay almost precisely.
These transverse dipole couplings are also almost
always significantly lower in conductivity than the
conventional ZZ coupling. This is typical of the
response of a triaxial tool to an anisotropic
medium. Notice also the increased activity on the
transverse couplings. This corresponds to the
overshoot response that we saw in with the
modeled data.
FORMATION PARAMETER INVERSION
In this well we were lucky enough to have an FMI*
(Formation Microlmager tool) as well as a standard
AIT-H tool. The FMI allows us to determine
visually which regions are anisotropic due to the
presence of thin beds and which do not show these
thin beds.
Previous work has been done on inversion
of triaxial induction tools [Anderson, 2001, 2002,
Mollison, 2001]. We have developed a fast 1-D
inversion algorithm to retrieve the formation
parameters in a TI anisotropic formation. Since our
problem is non-linear, the inversion algorithm is
designed to match the synthetic data with the
measured data using an iteratively updated estimate
of the formation. The inversion procedures will not
be terminated until the mismatch between synthetic

Mark of Schumberger

data and measured data is less than a desired

threshold. The inversion method is a weightedleast-square constrained minimization approach
based on a Gauss-Newton method. Anderson et al.
(2002) previously used this method successfully
for the inversion of tri-axial induction data. In
addition to the data misfit, some additional terms
have been added to the penalty function to improve
the stability and performance of the algorithm.
N

These

terms

include

~,;L/

N-1

and

i=1

where'/-

4Rye/Rhi

~ / L i / ~-1
i=1

is the square root of the

ratio of vertical to horizontal resistivity for the

i 'h layer and N is the number of layers.

The inversion includes the relative dip and

azimuth, the horizontal and vertical resistivity of
each layer, and the bed boundary positions. This
leads to a total of 3N+l parameters for a formation
with N layers.
From our study, we discover the
combinations of

O'xx/O'zz

and 2Crzz - Crxx have higher vertical

resolution and better signature of bed boundary
location than do any of the individual coupling at
any dip angle. We use them to gives us the initial
model of bed boundary location and later iterate on
the exact location of the boundaries.
In the framework of inversion procedure,
the input data is borehole effect corrected while the
invasion effect is ignored. The azimuth angle is
solved first by rotating the data matrix so that the
cross coupling O'xr, O'rx, O'yz and O'zr are zero.
This puts the new Y-axis along the relative strike
of the formation. That is, it lies in the intersection
of the bedding plane and the plane perpendicular to
the tool axis. The rest of the parameters: horizontal
and vertical resistivity, bed boundary locations, and
relative dip angle are inverted simultaneously
rather than separately. The ability to simply rotate
the data is one of the significant advantages of a
fully-triaxial tool.
In
the
inversion
procedure,
the
computation of sensitivity function or Jacobian
~O" i

matrix ( ~ p a r a m e t e r j )

generally costs 99% of

the CPU time. Normally, such sensitivity function

is computed by taking a finite-difference between
the logs at two closely spaced parameter values. In
order to improve the performance of the inversion

QQ

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

software to a new stage, we derived the analytical

expression of the sensitivity function in a 1-D TI
horizontally-layered medium. This sensitivity
function is computed exactly at each iteration with
little additional computational expense. This gives
both a more accurate sensitivity function and a
reduction in the time for each iteration. The final
code is an order of magnitude faster than one based
on regular finite difference method. The result is a
code that can be run in near real-time at the
wellsite.
Figure 7 shows the inversion of tri-axial
induction logging in a 15-layer chirp formation.
The horizontal and vertical resistivities of the
conductive beds are 1.0 f2-m, and 3.0 f2-m
respectively while the horizontal and vertical
resistivities of the resistive beds are 50.0 ~ - m and
5.0 f2-m respectively. The relative dip angle and
azimuth angle are 60 , and 30 . The synthetic data
is contaminated by 3% random noise. From this
figure we can see that even if the initial bed
boundary location is set 1 ft away from the true
location, the inversion result still can recover the
bed boundary position very well. The inverted dip
angle and azimuth angle are 59.99 and 30.04 .
The misfit between the synthetic data and the
reconstructed data is too small to be seen on this
plot. The example shows that our inversion
software can provide resistivity anisotropy as well
as tool relative dip and azimuth and formation bed
boundary locations. With the help of explicit
expression of sensitivity function, the inversion
processing is very fast. This example took less than
one minute on a 600 MHz laptop computer with
256K of memory.
The inversion code has been further tested
with the field data measured by the tool in a
vertical borehole. In this well we were lucky
enough to have an FMI (Formation Microlmager)
as well as a standard AIT-H. The FMI allows us to
determine visually which regions are anisotropic
due to the presence of thin beds and which do not
show these thin beds. From FMI image, light
anisotropy was observed in the area and the
formation dip is close to zero. Only XX and ZZ
coupling are needed for the inversion, since the
lack of cross-terms and the overlay of XX and YY
indicated flat bedding with zero dip. Figure 10
shows the tool response of the interval between
x000ft to x200 ft. The formation of the interval is
segmented into 100 layers. The number of
inversion parameters is 301. The sensitivity
function is computed within a 20ft window size. It
only took 53 seconds to compute the sensitivity

function for 400 logging points with respect to all

inversion parameters. It took less than 20 minutes
before the inversion procedure was terminated on
the same 600 MHz laptop. The figure shows the
inversion result of vertical, horizontal resistivity
and bed boundary position. Conventional AIT log
with 2 ft resolution is also shown on the same
figure. We can see the inverted horizontal
resistivity has very good agreement with
conventional AIT logs. There are a number of thin
beds and most of the formation is anisotropic. The
exceptions are the regions from X015 to X025 and
from X035 to X060. In these regions the FMI
shows much more isotropic beds and the logs show
that the XX and ZZ conductivities agree. The
disagreement
between
the
data
and
the
reconstruction is small.
Figure 11 shows the inversion result of
another interval from the same well. This region
has less bedding and moderate separation between
the XX and ZZ couplings. The anisotropy ratio in
the interval is consistently less than 1.5. Such light
anisotropy is confirmed by FMI image. The misfit
between the measured data and the reconstruction
is small.
Again, we can see good agreement
between horizontal resistivity and conventional
AIT logs has been obtained. From the FMI image
shown in a clean sand layer between y030 and
y050 ft is observed. The inversion result shows the
anisotropy ratio is close to 1.0 at that section.
CONCLUSIONS
A new experimental fully triaxial induction tool
has been built and field-tested. The tool
incorporates two features that distinguish it from
prior tools. First, it is fully triaxial with sensors
measuring all combinations of transmitter and
receiver orientations at the same point and at the
same time. This greatly simplifies the interpretation
of the data and allows the most robust and accurate
interpretation of formation parameters.
In addition, the tool design reduces the
borehole effect by at least two orders of magnitude
compared to a triaxial tool of normal design. The
remaining borehole effect is similar in size and
characteristics to that found with a conventional
induction tool.
ACKNOWLEDGEMENT
The authors would like to thank the oil companies
who consented to the use of their data.
REFERENCES

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

Anderson, B., Barber, T., Habashy, T., M., 2002,

The interpretation and Inversion of Fully Triaxial
Induction Data; A Sensitivity Study, Paper O, in
43r'iAnnual Logging Symposium Transactions:
Society of Professional Well Logging Analysts.
Anderson, B.I., 2001, Modeling and inversion
methods for the interpretation of resistivity logging
tool response, Ph.D. thesis, Delft University Press,
The Netherlands.
Forgang; S., Fanini, O., Tabarovsky, L., Method
and apparatus for transverse electromagnetic
induction well logging, US Patent 5,781,436, July
14, 1998.
Huston, O. H., Formation dip measuring methods
and apparatus using induction coils, US Patent
3,510,757, May 5, 1970
Kriegsh~iuser, B., Fanini, O., Forgang, S.,
Itskovich, G., Rabinovich, M., Tabarovsky, L., Yu,
L. Epov, M., and v.d. Horst, J., 1999, A new
multicomponent induction logging tool to resolve
anisotropic formations, Paper D, SPWLA 40 th
Annual Logging Symposium Transactions.
Kriegsh~iuser, B.F., Fanini, O.N., Forgang, S.,
Mollison, R.A., Yu, L., Gupta, P.K., Koelman,
J.M.F., and van Popta, J., 2000, Increased oil in
place in low resistivity reservoirs from
multicomponent induction log data, Paper A,
SPWLA 41st Annual Logging Symposium
Transactions.
Mollison, R.A., Fanini, O., Kriegsh~iuser, B., Yu,
L., and Ugueto, G., 2001, Impact of
multicomponent induction technology on a
deepwater turbidite sand hydrocarbon saturation
evaluation, Paper T, SPWLA 42 nd Annual Logging
Symposium Transactions.
AUTHORS
Richard Rosthal is a Principal Scientist at
Schlumberger's Sugarland Product Center. He has
been involved with the development of nearly
every resistivity tool that Schlumberger has
introduced in the past 20 years, including Array
Induction Tool, Compensated Dual Resistivity,
Resistivity At the Bit, and now 3D Induction. He
has a PhD in theoretical solid-state physics from
the University of Maryland.

Tom Barber is a Technical Advisor at

Schlumberger Sugar Land Technology Center, and
has worked on induction modeling, array design
and environmental corrections since coming to
Houston in 1978. Author of over 30 papers and
holder of thirteen patents, he was awarded the
SPWLA Distinguished Technical Achievement
Award in 1993 for significant contributions in
electromagnetic logging. He previously worked on
magnetic susceptibility logging measurements at
Schlumberger-Doll Research, Ridgefield, CT.
Before joining Schlumberger he worked at the
NASA Marshall Flight Center, Huntsville, AL and
at Brookhaven National Laboratory, Upton, NY.
(barber@ slb.com).
Steve Bonner holds a BSEE from the University
of Houston ('74) and an MSEE from MIT ('76).
He has worked for Schlumberger since 1976 on a
variety
of
engineering
and
management
assignments focusing on the design and
development of electromagnetic tools. For the past
3 years, he has been the manager of the Formation
Evaluation Wireline Department in their Sugar
Land Product Center.

Kuo-Chiang (KC) Chen is the manager of triaxial induction tool in the Schlumberger Sugar
Land Technology Center at Texas. He began his
career in the Schlumberger Reservoir Completion
Center in 1996 as a mechanical engineer. Before
the current position, he was the lead mechanical
engineer of the Wireline Oriented Perforating
Tools, Perforating Anchors, Wireline Perforating
Platform, and holders of several patents. KC holds
a BS degree in Mechanical Engineering from
National Taiwan University, and a MS degree and
PhD degree in Mechanical Engineering from
Massachusetts Institute of Technology.

Sofia Davydycheva graduated from the Physical

department of Moscow State University in 1984
and completed postgraduate studies at the same
university, Applied Mathematics department, in
1988. Her PhD thesis was on mathematical
problems in electrical oil prospecting. From 1988
to 2001 she was with the Central Geophysical
Expedition, Moscow. In 2002 Sofia joined
Schlumberger, Sugarland Product Center, where
she currently works as a senior research scientist.

QQ

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

Gary Hazen received his BSEE from the

University of South Alabama in 1973, and his
MSEE from Georgia Institute of Technology in
1974. He has worked for Schlumberger Oilfield
services since 1974 where he has designed various
downhole electronic systems for wireline formation
evaluation. He presently holds the position of
Engineering Advisor.
Dean M. Homan is a Senior Research Scientist at
Schlumberger Sugar Land Technology Center, and
has worked on induction modeling, array design
and environmental corrections since 1999. Author
of over 14 papers and holder of three patents.
Before joining Schlumberger he worked on the
degenerate Fermi gas experiment at Rice
University, Houston, TX. (homan@slb.com).
Gerald Minerbo is a Scientific Advisor at
Schlumberger Sugar Land Product Center.
Minerbo was one of the originators of the AIT and
HRLA tool concepts.
He is now providing
modeling support for electromagnetic logging. He
is investigating new concepts for array resistivity
measurements. Minerbo joined Schlumberger in
1981. Previously he was a staff physicist at Los
Alamos National Laboratory. He has a BS degree
in physics from Polytechnic University, New York,
and a PhD in theoretical physics from Cambridge
University, England.
Robert Schlein is a Senior Software Engineer at
Schlumberger. He received a BSEE from City
College of New York in 1972, and a Ph.D. in
Electrical Engineering from the University of
Illinois in 1978. He has worked on Surface
Acquisition, Calibration, Test, and Processing
software, and embedded systems software for
Wireline Logging tools for the past 20 years.
Laurent Villegas is an Electrical Engineer. He
received a MS degree from the National Institute of
Applied Sciences (INSA) in Lyon, France. He
worked on different downhole applications for
MWD, Nuclear tools and he is now developing the
hardware architecture for resistivity tools.
Hanming Wang received his BS degree from
Physics Department of Zhejiang Normal University
of China in 1985, MS degree from Petroleum
Exploration Department of University of Petroleum
of China in 1992, PhD degree from Electrical
Engineering Department of University of Houston
in 1999. He worked at Petroleum Exploration

Department of University of Petroleum as an

instructor form 1992-1996. He joined with
Schlumberger Sugar Land Product Center as a
Senior Engineer in 1999. He is a member of
SPWLA, SPE, SEG.
Feng Zhou obtained a Master degree at Cornell
University in Ithaca, New York, and a BS degree at
North Dakota State University in Fargo, North
Dakota, both in electrical engineering. After joined
Schlumberger in 1996, Feng has worked as a
design engineer for the development of several
nuclear and resistivity wireline tools. Now he is a
project manager for a new wireline nuclear tool.

SPWLA 44thAnnual Logging Symposium, June 22-25, 2003

Electronics Housing

Current Sensor
Triaxial Transmitter
Short Z Array
Conducting Sleeve
Short Triaxial Array

Long Triaxial Array

Mud Resistivity Sensor

Figure 1. Major design elements of the fully-triaxial tool.

QQ

2~ii

f
.....

~ , .

InsulatingTool

- ' 2
.

OmduetingTool

Figure 2. Borehole currents for insulating and conducting tools

eccentered in conducive boreholes.

S P W L A 44 th A n n u a l L o g g i n g S y m p o s i u m , J u n e 2 2 - 2 5 , 2 0 0 3

1.6
1.4

Insulated Tool
Metal Tool (Experiment)
Metal Tool (Theory)

~'1.2

.>_
"00.8
tO

o
"0.6
<0.4
0.2
-~2:

= ......
-1.5

,~,
-1

,'r
T
T
-0.5
0
0.5
Eccentering (in)

"r
1

~
1.5

-2

Figure 3. Borehole effect for conductive and insulating induction tool

with transverse oriented magnetic dipoles. Tool is eccentered in a
direction perpendicular to the orientation of the dipole. Rm=. 1 f2-m,
Rt=20 f2-m, Dh=7 7/8 inch.

o
o

o
o-

J I

Insulating Tool

+.,~0
(Do

4-o~o

Insulating Tool Freq. Focused

.,

..

,...

,,

0 ........................ i 0
0

TIO

A .
w

......................

Conducting Tool
I

100

1000

Sigma shoulder (mS/m)

Figure 4. Borehole effect for insulating and conductive tool centered in 12' bed.
O'bed -- 20 mS/m, (Ym - 5000 mS/m, Dh - 8".

10

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

0 -1000 -500
-120
'_
-110

0
~

500
.....:

1000 -1000 -500

I
/

400
'

0
I

500
I

-100 -50

100

50

100 150 200

-100
-90

R Signal (mS/m)

R Signal (mS/m)

R Signal (mS/m)

c ~ -

....

...............................

.....

-90
180

80

-70

-70

-60

-6o

-50

-50

-40

-40

-30

-30

-20

-20

-10

-10

50
,i ....

.............. .~

100

0
10

20

2O

3O

30 --:::::::i .~

(/

40
50

XX
y

..........................................................................

......................

XZ . . . . . . . .

yzZX

XX
xY

..................................

.............................

XZ

yx

zz

yx

yy

Sig T

YY

.......................................................

........................................
........................................

V7ZX
zy

....................................

....................................

zz

Figure 5. Response of 27" triaxial array in the Oklahoma 2

formation at 0 (left) and 60 (right).

Sig T
Figure 6. Field data from the
triaxial 39" array.

11

QQ

SPWLA 44 thAnnual Logging Symposium, June 22-25, 2003

*Gxx in p ut

Rh

*Gzzlnput

-- Reconstruted
Initial Guess

~ &Li

-100
0

mS/m

250

0.2

f2-m

200

!I

...........L.._..,

50

l.-m,:

100

............... | .....
]

!
|

i
1
]

I
i

150

~i

I .............................................................................

200
Figure 7. Reconstruction of an anisotropic chirp formation from a triaxial induction array.
12

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

*C~xxData
*(~zzData
: :

ah

"~,/!i ii i/i ~:~!)

i:

AT60
-100

mS/m

250 2

~-m

FMI Image

200

XO00~

j ......ll.........

_ NN

. . . . . . .

X050

~;-_.: :.

=~__-.

- .-~. ~~

~._ =-

X100

X150

: "-,'v.

:~.'-~

,~.~--',, !

.... ~~, . ~ i

X200
Figure 8. Inversion of field data from a triaxial induction tool.

13

~I

iim~

in

I / i n

ClCl

i~

SPWLA 44 th Annual Logging Symposium, June 22-25, 2003

*Oxx Data

*ozzData

ah

Rv
AT60
~-m

.... c~zz iVi e d e i

-100

mS/m

250

200

YOOOr

Y050

YIO0

Y150

Y200
Figure 9. Additional inversion of field data from a triaxial induction tool.

14

FMI Image