You are on page 1of 8

SPE/IADC

SPE/IADC 13476

Reduction of Nonmagnetic Drill Collar Length Through Magnetic


Azimuth Correction Technique
by A.W. Russell, NL Sperry-Sun, and R.F. Roesler, NL MWD
SPE Members

Copyright 1985, SPEIIAOC 1985 Drilling Conference

This paper was presented at the SPE/IADC 1985 Drilling Conference held in New Orleans, Louisiana, March 6-8, 1985. The material is subject to correc-
tion by the author. Permission to copy is restricted to an abstract of not more than 300 words. Write SPE, P.O. Box 833836, Richardson, Texas
75083-3836.

ABSlRACT
a nonmagnetic drill collar (NMDC) of sufficient
Traditiona 1 compass-type surveying instruments length reduces this systematic error to a
with their ability to sense only the direction of negligible magnitude. This reducihle, systematic
the local magnetic field vector must be used in error in the azimuth measurement is qualitatively
conjunction with appropriate lengths of different from the (random) uncertainty inherent
nonmagnetic drill call ar to ensure that the 1ocal in the instrument which wi 11 be present even if
magnetic field vector is uncorrupted by drill- the tool is housed in a much longer NMDC.
string magnetization. It is now possible to
compensate for the effects of drill-string The generic solid-state survey/steering instrument
magnetization and, hence, utilize much shorter consists of transducers to measure the three
nonmagnetic drill collars by using solid-state orthoqona 1 components of both the 1oca 1 magnetic
surveying instruments which use gravity and field and the local gravitational acceleration.
magnetic transducers. Measurements in different MWD survey/steering instruments (both the wireline
bottom hole assemblies show that the required tool and mud pulse telemetry tool) and the
nonmagnetic collar lengths can be reduced in most electronic multishot utilize transducers. Prior to
situations using this compensation technique. the introduction of these solid-state instruments,
This ability to run with shorter collars has consideration of the effects of drill string and
significant impact on the use of MWD tools because 8HA magnetization consisted largely of empirical
it allows greater flexibility in the configuration
of the bottom hole assembly as well as allowing estimatesl. These were used to indicate the
the directional sensor and other MWD sensors to be magnitude of possible errors which might be
positioned closer to the bit. expected in any particular situation and to select
the 1ength of NMDC required. Using the wire 1i ne
steering tool to measure the components of the
IN lRODUC TION
magnetic field inside the NMDC resulted in a more
accurate quantization of NMDC length requirements 2
Surve.vi ng instruments which measure azimuth a1 for different BHAs. As indicated in reference 2,
orientation with respect to magnetic fields appear however, although the systematic error in the raw
in two varieties: a comoass-type which senses the azimuth measurement can be estimated for a given
direction of the local field and a solid-state RHA, the estimate cannot be used to correct the
type which measures the magnitudes of the three raw azimuth measured with a compass-type
orthogonal components of the local field. The instrument because the magnetic pole strengths
traditional compass-type surveying instruments needed for the computation are not known with
(magnetic single . shots and magnetic multi shots) sufficient accuracy.
measure the •raw• 'azimuth, which we define to be
any azimuth measurement made with respect to the Recause of their ability to individually measure
local magnetic field without correction. When the three orthogona 1 components of the 1oca 1
there is no magnetic interference from the magnetic field, the use of wireline steering tools
magnetic sections. of the bottom hole assembly as surveying instruments allowed the adjustment of
(RHA) above and below the instrument, the raw the raw azimuth (measured with respect to the
azimuth is the true azimuth measured with respect direction of the local magnetic field) to yield a
to magnetic north. Otherwise, an extraneous corrected azimuth (measured with respect to
magnetic field produces a systematic error in the magnetic north). This ability of the tool to
azimuth measurement and the raw azimuth differs correct for the systematic error in the raw
from the true azimuth. Housing the instrument in azimuth suggested the possibility of shortening
the length of the NMDC and making a larger
References and illustrations at end of paper. correction. No effort was made to reduce the
463
2 REDUCTION OF NON-MAGNETIC DRILL COLLAR LENGTH SPE 13476

length of the NMDC, though, because use of the field strength and direction to compensate for the
steering tool was generally 1imited to the corrupted OZ-axis measurement 3,4. An instrument
directi~nal phases of the drilling operation; used with the corrected azimuth technique requires
convent1onal compass~type instruments, requiring highly accurate calibration, because the absolute
relatively longer lengths of NMDC, were used for magnitudes of the field components are required.
the bulk of the surveying. The raw azimuth, however, requires only ratios of
How~ver,, because the mud pulse telemetry MWD the magnitudes of these components, thus reducing
str1ng. 1s used throughout .the entire drilling the calibration complexity and scale factor errors
operat1on, the use of the az1muth correction tech- for this measurement.
nique assumes greater importance, especially if it
can be used to reduce the amount of nonmagnetic Experiments were done in 1977 and 1978 in the
material needed. Without azimuth correction the Netherlands by Koninklijke/Shell Exploratie en
entire MWD string must be nonmagnetic; with s~ch a Producktie Laboratorium and NL Sperry-Sun/Russell
correction, only a relatively short length on Attitude Systems in order to confirm that the
either side of the transducers needs to be error field inside the NMDC can, in fact, be
nonmagnetic. In either event one or more properly modeled. Measurements of the transverse
additional NMDCs mav be reqJired below the (OXV) and longitudinal (OZ) components of the
string, especially when drilling with a motor. maqnetic field in different BHAs were made every
Reduction of the NMDC length requirement would n.s meters along the length of the nonmagnetic
result in the following advanta9es: sect ion. These NMDC sweeps were made downho 1e,
but not while drilling. Results from a t_ypical
• The cost of nonmagnetic collars and the sweep through two NMDCs in tandem (16.5 meter
corresponding lost-in-hole exposure is total length) are shown in Figures 2 and 3. The
ma~nitude and dip angle of the earth's magnetic
reduced.
field were 48~T and 67.8°, respectively. The
• The directional sensor can be positioned
closer to the bit because there are no borehole inclination was 41.5°. The relative
requirements for additional NMDCs below the variations in R0 z and Roxy along the length of the
tool.
nonmagnetic section are shown in Figure 2. The
• More flexibility can be achieved in longitudinal field strength, B0 z, changes rapidly
positioning other MWD sensors nearer the
bit, again because no additional NMDCs are as we near the magnetic sections of the BHA above
requi~ed ?elow the MWD string. Proximity to and below the NMDCs, but is relatively constant
the b1 t 1s very important when the natura 1 near the center; the transverse field strength,
gamma ray/resistivity combination is used, R0 xy' is relatively constant along the entire
for example, for real-time shale
overpressure prediction. 1ength of the nonmagnetic section. Knowing the
.borehole orientation and the earth's field
strength and direction, we can subtract from the
BASIC 11-IEORY measured B0 z that component which is due solely to
Basis of the Corrected Azimuth Technique the earth's magnetic field. The residual field is
the lonqitudinal error field produced by the
When using a solid-state instrument which measures magnetic sections of the BHA. The magnitude of
the three orthogonal components of the local this error field along the length of the
magnetic field individually, there are two basic nonmagnetic section is shown in Figure 3. These
approaches to computing an azimuth. The first results, along with the results from the other
assumes that there is negligible interference from downhole measurements, indicate that the
the maqnet i c sections of the RHA above and be 1ow transverse component of the magnetic field is
the instrument; then the components along the OX, unaffected by the BHA when the NMDC 1ength is
OV, and 0? axes (the coordinate svstem which will greater than about 10 feet. This result is used
be used throughout is shown in (i gure 1) may be in the patented azimuth correction technique.
combined to yield a raw azimuth3. It must be a
•raw• azimuth, because we are measuring the Theoretical Performance
components of the local magnetic field without
correcting for any magnetic interference which may In the sol id-s tate instrument ernp.loying a combi-
be present. Note that if there is magnetic inter- nation of gravity and magnetic transducers, the
ference from the BHA, then this computed raw instrument-related uncertainties in the derived
azimuth will include the same systematic error as survey parameters are directly related to the
would a compass-type instrument under the same uncertainties in the individual transducer
conditions. Thus, the survey instrument must be measurements. We define the •basel ine•
isol~ted. f.rom extraneous magnetic fields by uncertainty in the derived azimuth as the
hous1ng 1t 1n a NMDC of sufficient length. uncertainty in the raw azimuth when the tool is
housed in a very long NMDC (i.e,, no extraneous
The corrected azimuth is based upon a patented magnetic fields to corrupt the measurement). This
technique that uses the magnitudes of the magnetic uncertainty in the computed azimuth is always
--lo -l. ~
~resent and is not reducible, in contrast with the
field components, R0 x, B0 y, and Roz in conjunction
systematic error in the raw azimuth which is
with the known values of the earth's magnetic reducible by housing the instrument in a longer
464
SPE 13476 A.W. RUSSELL AND R.F. ROESLER 3
NMDC. Because we are correcting for the From the point of intersection, draw a hori zonta 1
systematic error in the raw azimuth, the corrected line to the left until it intersects the vertical
azimuth retains only a residual uncertainty, the figure-of-merit axis. The figure of merit ..is seen
magnitude of which is related to the individual to be 1.1. Thus, if the one standard deviation
transducer uncertainties and to the alqorithm used (1 a) baseline uncertainty for the Gulf of Mexico
to do the actual computation. It remains to show is 0. 25°, the 1 a uncertainty in the corrected
how this uncertainty in the corrected azimuth
compares to the baseline uncertainty in the raw azimuth would be (1.1 x 0.25° =) 0.28°.
azimuth.
Now assume that the well is 1ocated in the North
For a particular tool orientation, we can define a Sea, but that the orientation remains the same.
~
figure of merit for the azimuth correction The magnitude of BN in the Gulf of Mexico and the
technique as the expected uncertainty in the
correct azimuth relative to the haseline North <:iea is about 251JT and l71JT, respectively;.
uncertainty: therefore, because the uncertainty varies inverse-
-1...

figure of _ uncertaint~ in corrected azimuth ly with BN, the uncertainty will increase by a
merit - basel1ne uncerta1nty factor of (251JT/171JT-) 1.5. Thus, in the North
Sea, our 1 a uncertainty in the corrected azimuth
The north (hori zonta 1) component of the earth's
~ would be approximately (1.5 x 0.~8° = ) ~.42°.
magnetic field, RN, serves as the reference vector
for the azimuth measurement and both the baseline VERIFICATION
uncertainty and the uncertainty in the corrected
azimuth vary inversely with its magnitude.
~ - The Laboratory Measurements
magnitude of 8N decreases with increasing dip
angle, so the uncertainties generally increase The theoretical performance of the correction
with increasing 1ati tude. However, because the technique was verified at NL by subjecting a sur-
figure of merit is the ratio of two values which vey tool to an external magnetic field oriented
along the tool's axis; the field was generated by
both vary inversely with Rl\l, the figure of merit an electromagnetic collinear with the tool. The
is independent of location. The figure of merit tool and electromagnet were placed in a orient-
for the particular correction technique employed able, nonmagnetic test fixture, in which the
in NL's solid-state survey systems is shown in tool's inclination and azimuth could be determined
Figure 4 as a function of tool orientation. to 0.1. 0 • The entire experiment was conducted in
a nonmaqnetic enclosure (see Figure 6). The mea-
sured magnitude and dip angle of the local earth's
General Characteristics and Examples
magnetic field were 50.7 IJT and 59.8°,
Inspection of Figure 4 reveals the following respectively.
general characteristics of the correction
technique: Data were taken for 1ong itudi na 1 error fie 1ds of
5.01JT and 7.51JT at inclinations of 10°, 30°, 60°
1 The uncertainty in the corrected azimuth is
equa 1 to or · greater than the base 1i ne and 80° and azimuths of 180°, 225°, 315°, and
uncertainty. 0°. The transverse (OXY) error field was <0.11JT
throughout the experiment. In order to
1 When drilling within 45° of north-south, the demonstrate the ahi 1ity to accurately make 1arge
uncertainty is never more than 1.5 times the corrections a raw azimuth measurement, we show the
baseline uncertainty, regardless of results for our worst case experiment:
inclination.
measurements made with the too 1 oriented at 270°
1 The uncertainty is twice the baseline uncer- azimuth and subjected to a 7.5 IJT longitudinal
error field. The raw and corrected azimuths are
tainty at 60° inclination when drilling near shown in Fi·qure 7 as a function of inclination;
east-west. the 2 a uncertainties for each datum are also
shown. Included on the plot are the curves of the
1 Above 60° inclination, the uncertainty expected raw and corrected azimuths which were
increases rapidly when drilling near east- generated based upon the known orientation and
west. local magnetic field. We see excellent agreement
between the expected and measured values. This
In order to demonstrate the use of the fiqure of example demonstrates two important points. First,
merit chart, consider the case where a well is the correction technique can correct for large
being drilled in the Gulf of Mexico; the borehole systematic errors in the raw azimuth. Note that
inclination is 30° and the azimuth is 120° at the higher inclinations, the correction tech-
(S 60 E). Refer to Figure 5. Along the nique is accurately adjusting the raw azimuth by
horizontal axis find the endpoint of the arc cor- about 15°. Second, the sources of error in both
responding to a bearing of 60° from south. Trace the raw and corrected azimuths are we 11 under-
the arc counterc 1ockwi se until it intersects the stood; in all cases, the expected value of either
raw or corrected azimuth fell within the 2 a band
radial line corresponding to 30° inclination. around the measured value. These 2 a limits on
465
4 REDUCTION OF NON-MAGNETIC DRILL COLLAR LENGTH SPE 13476
the uncertainties correspond to the 1 worst case 1 these tools to be run in shorter NMDCs than is
condition in which all measurement errors sum in possible without this patented technique. The
the most adverse manner. short NMDC configuration for MWD has a positive
impact on the drilling operation: the BHA design
Downhole Measurements is flexible because there is no requirement for
long lengths of NMOC, and all the measurements,
Extensive downhole measurements, both stationary both survey/steering and formation evaluation, can
and while drilling, have been made in order to be made closer to the bit since there is no need
verifv that the results obtained under controlled to run additional NMDCs below the MWD string. In
laboratory conditions also hold in the inhospi- situations of high dip angle and high borehole
table downhole environment. The results for the inclination, however, the user must weigh these
NMDC sweep previously discussed (see Figures 2 and advantages against the accuracy limitations of the
3) are shown in Figure 8; one single-shot compass method.
azimuth and two gyro azimuths measured in the same
section of the borehole are also shown. The
results indicate that the corrected azimuth is ACKNOWLEDGMENTS
va1 i d over rough 1y 80% of the NMDC 1ength or, in
other words, to within roughly 5 feet of the mag- Many of the early experiments to measure magnetic
netic sections of the string. field components within NMDC were carried out in
conjunction with Koninklijke/Shell Exploratie en
The results of another field test in the North Sea Produckti e Laboratori urn and Nederl andse Aardo 1i e
(earth 1 s magnetic field strength and dip angle Maatschappi j. The authors are indebted to
personnel within these organizations for their
were 50 #LT and 71.5°, respectively) are shown in willingness over the years to assist in the
Figure 9. Two runs were made with NL Sperry-Sun 1 s development of newer and better surveying instru-
new electronic multishot: one run in a 120 foot ments and techniques.
length of NMDC (a sufficient length to reduce any
systematic error in the raw azimuth to a negli-
gible amount) and one in a 30 foot length of REFERENCES
NMDC. Shown in Figure 9 are the raw azimuth in
the long NMDC and both the raw and corrected l. Blythe, E.J., Jr. 11 Estimating Compass Spacing
azimuths in the short NMDC. Borehole inclination Inside Nonmagnetic Collars 11 , Drilling-DCW
varied between 35° and 55°. As expected, the raw EXPOSITION-IN-PRINT, 1~75.
azimuth in the 1ong NMDC agrees with the short
~. Grindrod, S.J. and Wolff, J.M., 11 Calculation
NMDC corrected azimuth to within 1°. Again, note of NMDC Length Required for Various Latitudes
that the correction technique is adjusting the Developed from Field Measurements of Drill
short NMDC raw azimuth by about 10°. String Magnetization .. , paper IADC/SPE 11382
presented at the 1983 IADC/SPE Dri 11 ing
The results of measurements made while drilling Conference, New Orleans, February 20-?.3.
with NL 1 s mud pulse telemetry tool in the Gulf of
Mexico (field strength and dip angle were 50~LT and 3. Russell, M.K. and Russell, .A.W., 11 Surveying of
liO.l 0 , respectively) are shown in Figure 10. This
Roreho 1es 11 , United States Patent 4 163 324,
MWO tool was run in a 16.5 foot NMDC with the August 7, J97~.
transducers positioned 6 feet from the bottom. 4. Roes 1er, R. F., 11 Surveying of Roreho 1es Using
The MWD NMDC was p 1aced between a 1ength of mag- Shortened Non-Magnetic Call ars 11 , United States
netic collar (below) and a magnetic stabilizer and Patent pending.
eros saver (above). In addition to the MWD
measurements, a magnetic single shot survey (taken
in a 37 foot NMDC positioned directly above the
aforementioned crossover) and a gyroscopic multi-
shot survey were run; these results are also shown
in Figure 10. The borehole inclination varied
between 18° and 21° over the 500 foot interval
shown. The three surveys shown exce 11 ent agree-
ment to within 1° over almost the entire
interval. The largest deviation (yet unexplained)
occurs near X500 feet, where the MWD corrected
azimuth and the gyro azimuth differ by about 4°.

CONCLUSIONS
The combination of rigorous theoretical analysis,
extensive laboratory and field verifications, and
widespread exoerience in measuring magnetic field
components inside NMDCs has confirmed the validity
and usefulness of the corrected azimuth technique
in the majority of drilling situations. Use of
the method with MWO survey/steering tools allows
466
,.._ NMDC ( TOP )

>I!C'"- ••' > • N

I
16.5m

'



10

20
• •

30

L
B0 z ERROR FIELD ( % EARTH'S TOTAL FIELD )

X • • •

Fig. 3-Magnitude of the residual longitudinal error field as a function of position along the NMDC.

Fig. 1-Earth-fixed (north-east-verticle) and tool-fixed (x-y-z) coordinate system referred to in the text.

90
3.0j ~~~c~ 1
2.0 60 "11-<1~
/0

1.5j
. . - NMDC ( TOP )
....
t: 1.4
a: 1.3
w
t - ------- -- ---- - 45
-- ---- - ------- - --
~Q
~~
~
I
~
?.__-o ·--· 0 0 u. 1.2
0
-- ------ --- --- ---- ---
o___..o--.- w
16.5m
~ 0~
o~ :
.
' ,;
\
0 Boz ( MEASURED )
x B
oxy
( MEASURED ) a: 1.1
::>
(!)
l. "II'• FLUX DENSITY ( B ) u::
o;
L
••
X

o..o )
o~o

,.._ NMDC (BOTTOM)


-o-- •. •.
o-----~~-------------o
x -------------0 ()
1.0 o~----~1~5----~3o-----4~5----6~0--7~590

~ BEARING FROM NORTH OR SOUTH (DEGREES)

--a Fig. 2-Relative change in the magnitudes of B 0 z and Boxy as a function of position along the NMDC. Fig. 4-Figure of merit for NL's azimuth correction technique as a function of borehole azimuth and
inclination.
\.N
~
:-J
0'
90
3.0 IJ\Ic
<:Ilk
2.0 - 60 '-<~)':
/0
1.5 ~0
I- -45 ~Q
a:w 1.4
1.3 ~~
::!
LL 1.2 ~'V
0
w
a: 1.1
:::J
(!)
u:

1.0 ~------~~------~------~------~~0
0 15 30 45 60 75 90
BEARING FROM NORTH OR SOUTH (DEGREES)

Fig. 5-Demonstration of the use of the figure of merit chart for a borehole azimuth of S 60 E (120°) and in-
clination of 30°.

ORIENTABLE TEST FIXTURE

'--_NONMAGNETIC ENCLOSURE
ELECTROMAGNET

Fig. &-Illustration of the laboratory facility and apparatus used to measure and confirm the characteristics of the
corrected azimuth technique.

SP.E 1 34 7.6
EXPECTED CORRECTED AZIMUTH

MEASURED CORRECTED AZIMUTH (±2u)


~

~MEASURED RAW AZIMUTH (±2u)

250~----~----~----~----~----~----~----~----~----~
0 10 20 30 40 50 60
INCLINATION (DEGREES)
Fig. 7-Measured and expected raw and corrected azimuths for the laboratory case of a 7.5~tT longitudinal error when
the true azimuth is 270°.

x RAW AZIMUTH
l I
o ADJUSTED AZIMUTH
_& GYRO AZIMUTH
[!] COMPASS SINGLE-SHOT AZIMUTH
0
I __.--x--x -----~IC---IC
I ,.....,..IC
I IIC/IC
16.5m 1° /al'i
10 ~~
I ~\IG
II ~~ICI,_ __
IC--IC
I 0 lo -----
1 I
I I
I I
I .& I'
~ ~i
311.7°+/-1.0°
\ .. _

310 312 314 316 318 320 322


AZIMUTH ( DEGREES )

Fig. 8-Raw and corrected azimuths for the NMDC sweep discussed in the text along with gyro and single·shot
confirmations.

SPf 1 34 7.6
AZIMUTH (DEGREES) AZIMUTH (DEGREES)
215 __ .., __ .., _..,.., --- - ' 145
X200 ·-- ·-- ........... . ......., .• .., •.
xooo
-CORRECTED AZIMUTH • MWD CORRECTED AZIMUTH
(SHORT NMDC)
•••••• RAW AZIMUTH
NL
SPERRY-SUN
• GYROSCOPIC MULTISHOT ....
' (LONG NMDC) ELECTRONIC A MAGNETIC SINGLE SHOT •
\ • - • RAW AZIMUTH
I (SHORT NMDC)
MULTISHOT
.....
..... •
I

'' I
I i='
w
' \ w
~

• •
\
~X500 \ I
1-
w \ l:l.
~ w
\ Cl

'',,,,
I
l:i:
~xooo •
...... .................. •
''· •
I
....

Fig. 10-A Gulf Coast MWD (mud pulse telemetry) survey with gyroscopic multishot and magnetic single-shot
confirmations.

Fig. 9-A North Sea survey with the electronic multishot in both long (120ft) and short (30ft) NMDC's.
~
......
~

\.)J

+-
.......J
C1'

You might also like