You are on page 1of 9

SPE/IADC-189432-MS

Step Change in Wellbore Positioning Accuracy

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
Mahmoud ElGizawy, Darren Aklestad, and Ross Lowdon, Schlumberger

Copyright 2018, SPE/IADC Middle East Drilling Technology Conference and Exhibition

This paper was prepared for presentation at the SPE/IADC Middle East Drilling Technology Conference and Exhibition held in Abu Dhabi, UAE, 29-31 January 2018.

This paper was selected for presentation by an SPE/IADC program committee following review of information contained in an abstract submitted by the author(s).
Contents of the paper have not been reviewed by the Society of Petroleum Engineers or the International Association of Drilling Contractors and are subject to correction
by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers or the International Association of Drilling Contractors,
its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers or
the International Association of Drilling Contractors is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations
may not be copied. The abstract must contain conspicuous acknowledgment of SPE/IADC copyright.

Abstract
A survey program is designed for every well drilled to meet the well objective of penetrating the target
reservoir and to avoid colliding with other offset wells. The selection of the wellbore survey tools within the
survey program are limited to the current accuracy available to the industry. A newly developed wellbore
survey technique has proven to have superior accuracy compared to the current standard measurement-
while-drilling (MWD) surveys with in-field referencing and multi-station analysis (MSA).
In almost every drilling bottom hole assembly (BHA), there is an MWD survey tool to survey the wellbore
while drilling. Accuracy of the MWD surveys has been improved over the years by correcting potential error
sources such as misalignment of the survey package from the borehole, drillstring magnetic interference,
limited global geomagnetic reference, and gravity model accuracy. This new positioning technique takes the
accuracy of MWD surveys to the next level by combining surveys from two independent survey packages.
The second survey package is installed inside the rotary steerable system (RSS). Surveys from both packages
are retrieved while drilling.
Results have been obtained from multiple runs worldwide, enabling comparisons between the new
technique and standard MWD surveys from both an enhanced accuracy and true wellbore placement point
of view. A proposed error model is based on both the theoretical improvements in accuracy and the empirical
proof from the data analyzed. The improved accuracy while drilling assures higher confidence that the well
placement will maximize reservoir production and avoid collision with nearby offset wells.
In reservoir sections, the wellbore survey accuracy limits the lateral spacing, and this constrains the
reservoir production. In top and intermediate sections, wellbore survey accuracy limits the well plan, and
this affects how close the well can be drilled in proximity to other offset wells. This directly impacts the
complexity of the directional work and the cost per drilled foot. The new technique unlocks the potential
to significantly improve the wellbore positioning accuracy.

Introduction
MWD surveys have been the backbone of wellbore surveying for many years. MWD surveys are subject
to a number of errors which significantly reduce their accuracy. These errors are outlined and numerically
defined in Williamson (1999), and are nominally called the ISCWSA MWD error models. These errors
2 SPE/IADC-189432-MS

are broadly broken into several parts, reference, measurement and misalignment errors. With the use of
more accurate local magnetic fields to improve reference errors, multistation analysis to reduce the effect of
drillstring interference and the use of diurnal field correction when necessary MWD tools have been able to
improve lateral accuracy by over 60% when compared to the ISCWSA MWD error model. This has allowed
for improvements in well spacing, more efficient directional drilling from the increased spacing between
error ellipses and the accurate placement of ever longer ERD wells ElGizawy et al. (2014).

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
However, as the industry continues to drill in more complex and congested fields, drilling ever longer
wells and requiring tighter well spacing, an additional increase in accuracy is required. The use of gyroscopic
surveying systems to improve accuracy when using wireline are not a realtime answer and gyro while
drilling systems are often prohibitively expensive for the majority off wellbores.
Some rotary steerable systems (RSS) have 6 axis magnetic surveying systems which were developed
primarily for steering. They are however also of the same specification as the MWD systems used to provide
the definitive surveys for the wellbore. These rotary steerable systems have low levels of magnetization
in order that the steering systems work, and are therefore able to be used as a comparison to the MWD
surveys. This comparison is similar in process to the concept of dual inclination Berger, Sele (1998), where
two surveying systems are capable of being compared in such a way that a reduction in the TVD MWD
error ellipses is possible.
The application of this technique for azimuth as opposed to or in addition to TVD has not been tried
before. Most near bit survey sensors lack magnetometers as this saves tool construction material and other
methods such as bending moments can be used to derive pseudo azimuth and so control the RSS while
drilling. The 6 axis RSS surveying package takes both static (full 6 axis surveys normally taken before
surveying or after connection) and continuous surveys where the axial magnetometer is used to provide
continuous azimuth. It does this by taking the static surveys as a starting point for azimuth and then using
a single axis to measure deflection from this azimuth. It works very well when the axial component of
the magnetic field is low or near zero (drilling East/West) where minor changes azimuth are easily seen.
However, it is subject to considerable error when running along the Earth's magnetic field line which
generally goes north/South as small changes in direction are swamped by the magnetometer seeing a large
part of the earth's magnetic field.
The effects of shock and vibration on axial magnetic sensors has not been discussed in detail previously,
the effect of vibration an axial accelerometer was discussed with the resulting analysis suggesting that as
long as the bit stays on bottom when drilling axial shocks are limited. This has been shown in practice
to be true with tri axial shock sensors from MWD/LWD and drilling dynamics tools showing that cross
axial shock is far more prevalent than axial Reckmann et al. (2010). With this in mind the effect of axial
shock when drilling should be minimal with a clipping of the data as opposed to corruption occurring. If
you consider the difference between the RSS and MWD axial surveys, they are often subject to different
shock levels. The fact that there appears to be slight difference in the azimuth they show would support
this agreement.
There are several issues to consider when making a comparison between the RSS and MWD azimuthal
surveys. While the survey packages are of the same quality, the MWD tool will have a lower level of
drillstring interference and subsequently this needs to be addressed. The RSS survey package is somewhat
compromised by the bit which is less than 9 ft. from the survey package.
Equally in order to do a comparison of the surveys from the different sensor packages they need to be
processed similarly and be compared at the same measured depth. This is done in this instance by combining
the static and continuous surveys from each survey package, deriving and then solving for drillstring
interference and interpolating the surveys over a 10ft interval. This method to build high definition surveys
(HDS) from MWD surveys has been verified in some detail in Monterrosa et al. (2016) and ElGizawy et
al. (2017).
SPE/IADC-189432-MS 3

Methodology
Collect continuous data from RSS and MWD, combine with static surveys, (HDS) then do comparisons
over multiple runs. Run MSA to correct both data sets, get deltas check SD between runs, look at difference
in results apply new error model see if it fits new surveys and their variation between them.
To do a comparison between multiple the RSS and MWD surveys it is necessary to not only look at
the static survey data which whilst well understood may not be at sufficient survey intervals to accurately

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
describe the wellbore position and show the relatively small differences in azimuth, or indeed my overly
flatter the data showing an improved result without description of the well bore architecture and the physical
misalignments therein.
With this in mind, the survey data was processed incorporating continuous and static survey data into
high definition surveys (HDS). This allowed for an improved understanding of the wellbore architecture
which helped to provide a more realistic answer for the differences between RSS and MWD surveys. This
leads onto an improved understanding of the improvements that could be made to a RSS/MWD combined
error model and the quality control (QC) limits needing to be applied in these cases.
The continuous surveys are typically at a 1ft interval and the static at 90ft. There are some statistical
outliers in the continuous survey data which must be resolved, and all the static surveys must pass a field
acceptance criteria (FAC) prior to being used. The continuous data azimuth is derived from the static
surveys, as they measure deviation from a known point, and as such they are intrinsically linked.
From the static surveys, you can derive a compensation from drillsting interference using multi-station
analysis (MSA), this is derived for all 3 magnetometers and the axial component is applied to the continuous
data. Once the data is combined a QC is run where the computed azimuth trajectory from the continuous
data between surveys points agree with the static azimuth at the same points.
The HDS surveys are then interpolated at 10ft or 3m survey intervals, which sufficiently describes the
wellbore architecture Stockhausen, Lowdon, Lesso (2012), and these surveys are compared to each other.
The difference between the RSS and the MWD surveys is derived at the 10ft survey intervals. The mean
and standard deviation between the two data sets is derived for an entire run in order to see the magnitude of
the differences and also to assess the variation or consistency of the differences, which could be described
as the accuracy of the result.
The declination error term in the MWD error model is irrelevant in these results, as it will be the same for
both surveys. The correction of drillstring interference is much more significant as any residual magnetic
drillsting interference after processing through MSA (which is designed to reduce it) could have a significant
impact on the results, certainly in terms of the mean differences between the RSS and MWD surveys. For
this reason, long spaced magnetically clean BHA's have been chosen for analysis, minimizing the impact
of residual drillstring interference on the results.
Additionally, it was also important to verify any new error model with the impact of different trajectory
types, the main types being tangent and build sections, as the error model needs to be applicable for the entire
wellbore. In this paper four build and seven tangent sections were compared and analyzed for inclusion in
the analysis of a dual azimuth error model.
A combined error model was constructed following on from previous work to combine surveys Chia,
Phillips, Aklestad (2003) and Ledroz et al. (2016). An equal quality weighting for the MWD and RSS
surveys was used for this study, further work to explicitly take into account specific QC of individual stations
could be used in the future for an alternative weighting.
In this regard, the baseline error models used were from the Operators Wellbore Survey Group (OWSG),
and were modified to develop a +DUAL (dual sensor package) model. As previously demonstrated Bang
et el. (2016), common error model elements can be directly combined with the appropriate weighting to
develop a new error model, which will be equivalent to the more rigorous development of combined EOU's
from Chia et al. (2003).
4 SPE/IADC-189432-MS

Results
Data Sets
The data sets consist of two categories; build sections and lateral horizontal sections. MWD and RSS
measurements from seven wells are used for the lateral horizontal section analysis, while data from four
wells are used for the build section analysis. The horizontal sections directions include East and West

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
directions and build sections are usually 12 ¼" hole size, where the wells build from 30 deg and land at 90
deg at the reservoir section, the average length of the build sections is 6000 ft. The lateral horizontal section
is geosteered in the reservoir with RSS tool. A typical well profile is illustrated in Figure 1.

Figure 1—Typical Well Profile

The different well profiles are summarized in Table 1. The well is turning west 30 deg from south west
direction (220deg) on one of the build sections, while another is running 9 deg from south west direction
(226 deg). The other two build sections, the well direction is almost fixed at 200 deg Azimuth and 324
deg, respectively. The horizontal lateral sections average length is 15,000 ft. The well is turning towards
west direction in two of the lateral sections, while another two lateral sections the well is turning toward
east directions. Hence the selected well profiles represented the most challenging directions for magnetic
surveys.

Table 1—Well profiles Summary

Section Section Type Well Inclination (deg) Well Direction (deg)

1 Build from 45 to 90 from 220 to 250

2 Build from 47 to 89 at 200

3 Build from 60 to 88 from 226 to 235

4 Build from 30 to 86 at 324

5 Horizontal 90 from 253 to 275

6 Horizontal from 86 to 89 from 200 to 135

7 Horizontal from 85 to 91 at 315

8 Horizontal from 81 to 90 from 217 to 270

9 Horizontal from 90 to 85 from 65 to 90

10 Horizontal from 85 to 89 from 180 to 90

11 Horizontal from 87 to 91 at 315


SPE/IADC-189432-MS 5

Data Processing
The measurements from MWD and RSS are processed with multi-station analysis based on robust regression
and high definition surveys algorithms. Surveys every 10 ft are generated and corrected for borehole
misalignment. In-field references (IFR) are used for all wells, where total Earth's magnetic field reference,
magnetic declination angle and the dip angle are generated at every survey station from a 3D crustal model.
Azimuth and inclination from two different survey packages within the same drill string are compared. An

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
example of azimuth comparison and inclination comparison are plotted in Figure 2.

Figure 2—(Left) Azimuth from MWD and RSS. (Right) Inclination from MWD and RSS

The difference between the corrected azimuth and inclination from the two survey packages are further
analyzed and an example is illustrated in Figure 3 and Figure 4, respectively. The blue dots represent delta
azimuth in Figure 3, and represents delta inclination in Figure 4. Both should fall within a predefined
tolerance for the surveys to qualify for this technique introduced in this article.

Figure 3—delta Azimuth (MWD Azi - RSS Azi)


6 SPE/IADC-189432-MS

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
Figure 4—delta Inclination (MWD Inc - RSS Inc)

The difference between the two-corrected azimuths at the same depth are computed for the entire section.
For all data of the wells within this analysis, the average azimuth difference is 0.11 deg with a standard
deviation of 0.2 deg. The 95-percentile of delta azimuth between MWD and RSS is 0.4 deg, where 95-
percentile of the inclination difference are less than 0.18 deg. Results are summarized in Table 2.

Table 2—MWD and RSS delta Inclination and Azimuth Statistics

Well 95%
Section Well Direction delta Inc Avg delta delta % delta Azi Number
Section Inclination Percentile
Type (deg) < 0.18 deg Azi (deg) Azi STD < 0.4 deg of Surveys
(deg) delta Azi

1 Build from 45 to 90 from 220 to 250 95% 0.28 0.15 78% 0.52 752

2 Build from 47 to 89 at 200 89% 0.12 0.21 90% 0.47 721

3 Build from 60 to 88 from 226 to 235 97% 0.1 0.14 99% 0.32 802

4 Build from 30 to 86 at 324 95% −0.21 0.21 85% 0.5 376

5 Horizontal 90 from 253 to 275 98% −0.02 0.21 94% 0.42 1016

6 Horizontal from 86 to 89 from 200 to 135 98% 0.01 0.52 94% 0.43 748

7 Horizontal from 85 to 91 at 315 95% 0.12 0.16 96% 0.39 930

8 Horizontal from 81 to 90 from 217 to 270 100% −0.07 0.09 100% 0.2 999

9 Horizontal from 90 to 85 from 65 to 90 99% 0.06 0.27 83% 0.5 990

10 Horizontal from 85 to 89 from 180 to 90 97% 0.08 0.16 98% 0.35 1164

11 Horizontal from 87 to 91 at 315 100% −0.21 0.1 98% 0.37 1063

Well Position
As described in the methodology above, azimuth and inclination are combined to construct the weighted
average azimuth and inclination at common depths every 10 ft. An example plot containing two well
trajectories generated from corrected MWD (red) and RSS (green) surveys along with composite trajectory
generated using the dual survey packages (blue) are presented in Figure 5.
SPE/IADC-189432-MS 7

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
Figure 5—MWD Well Trajectory (Red), RSS Well Trajectory (Green), and Dual Packages Well Trajectory (Blue)

Ellipse of Uncertainty
The wellbore survey accuracy is defined by the EOU size that is calculated from an error model relevant
to the survey tool. The MWD and RSS surveys are assigned MWD+IFR+MS+SAG error model, while the
Dual packages surveys are assigned MWD+IFR+MS+SAG+Dual. The dual-package error model has been
updated to account for the improved accuracy of the accelerometers and magnetometers biases and scale
factors. The three EOUs are overlapping each other, and the three trajectories fall inside the three generated
EOUs as illustrated in Figure 6.

Figure 6—MWD Surveys EOU (Red), RSS Surveys EOU (Green), and Dual Packages Surveys EOU (Blue)

EOU is computed for the MWD surveys using MWD-STD error model, for corrected MWD and RSS
surveys using MWD+IFR+MS+SAG error model and for dual-package surveys using MWD+IFR+MS
+SAG+DUAL. It can be noted that the smallest semi-major error is observed for build section 4 where the
well direction is close to North, while the largest semi-major error is observed for horizontal section 5,9
and 11 where the well direction is in East-West direction. The error models EOU of the different surveys
are summarized in Table 3.
8 SPE/IADC-189432-MS

Table 3—Error Models EOU Summary of the Different Surveys

MWD+IFR MWD+IFR+MS MWD+IFR MWD+IFR+MS


MWD-STD MWD-STD
Section MD (ft) +MS+SAG +SAG+DUAL +MS+SAG +SAG+DUAL
EOU Major TVD Uncert.
EOU Major EOU Major TVD Uncert. TVD Uncert.

1 13910 214.4 99.7 95.1 73.8 46.9 40.8

2 13110 191.2 89.1 81 72.9 45.6 39.2

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
3 14450 262.6 111.6 106.5 108 52.5 45

4 9560 99 50.6 48.5 32.8 25 23

5 25810 450.9 200.7 192.1 150.2 89.7 84.8

6 26900 421.3 187.1 177.7 167.2 98.3 81.9

7 26070 428.1 197.3 186.6 173.9 102.9 87.2

8 20940 334.9 153.3 147.8 127.3 76.8 67.1

9 26290 444.8 204 194.8 176 103.8 88.6

10 26070 351.5 170.8 160.9 173.2 105.1 90.4

11 25580 462.5 209.1 197.3 178.1 104.4 86.8

Conclusions
– It is possible to compare two azimuths downhole even if they are from an RSS and MWD with
different shock and vibration regimes.
– The results conform to the error model proposed.
– More work could be done on the error model to improve its accuracy and development.
– More data needs to be analyzed to continue the development of this surveying process.
– It is an industry first, it promises considerable improvements in EOU in the future and it is deployable
in any wellbore with RSS tools.

Acknowledgements
The authors would like to thank Schlumberger for the permission to publish this article. The Schlumberger
field crews are acknowledged for collecting the data necessary to accomplish this work, and Natalia
Androsovish (Schlumbeger) is acknowledged for helping in data processing.

References
Williamson, H. S. 1999. Accuracy Prediction for Directional MWD. SPE-56702-MS presented at the 1999 SPE Annual
Technical Conference and Exhibition, Houston, Texas, USA 3-6 October 1999.
ElGizawy, M., Grini, M., Al Junaibi, H., Rachi, H., Chandran, K., Adewumi, F., Day, T., & Batu, A. 2014. Accurate
Geometric Well Placement of the First ERD Producer Well Drilled from the Artificial Island in UAE. SPE-171819-MS
presented at the Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, UAE 10-13 November
2014.
Berger, P. E., & Sele, R. 1998. Improving Wellbore Position Accuracy of Horizontal Wells by Using A Continuous
Inclination Measurement From A Near Bit Inclination MWD Sensor. SPE-50378-MS presented at the 1998 SPE
International Conference on Horizontal Well Technology, Calgary, Alberta, Canada 1-4 November 1998.
Reckmann, H., Jogi, P., Kpetehoto, F. T., Chandrasekaran, S., & Macpherson, J. D. 2010. MWD Failure Rates Due to
Drilling Dynamics. SPE-127413-MS presented at the 2010 SPE Drilling Conference and Exhibition, New Orleans,
Louisiana, USA 2-4 February 2010.
Monterrosa, L. C., Rego, M. F., Zegarra, E., & Lowdon, R. 2016. Statistical Analysis Between Different Surveying
Instruments to Understand the Reliability of MWD/RSS High Resolution Surveys and its Effect in Well Trajectory
Characterization. SPE-178830-MS presented at the 2016 IADC/SPE Drilling Conference and Exhibition, Fort Worth,
Texas, USA 1-3 March 2016.
SPE/IADC-189432-MS 9

ElGizawy, M., Androsovich, N., Rachi, H., Abbas, N., Al Katheeri, Y., Kolakkodan, S., Al Noukhatha, J., & Hargreaves,
B. 2017. Is While-Drilling "True Vertical Depth" True?. SPE-188921-MS presented at the Abu Dhabi International
Petroleum Exhibition & Conference, Abu Dhabi, UAE 13-16 November 2017.
Stockhausen, E. J., Lowdon, R., & Lesso, W. 2012. Directional Drilling Tests in Concrete Blocks Yield Precise
Measurements of Borehole Position and Quality. SPE-151248-MS presented at the 2012 IADC/SPE Drilling
Conference and Exhibition, San Diego, California, USA 6-8 March 2012.
Chia, C. R., Phillips, W. J., & Aklestad, D. L. 2003. A New Wellbore Position Calculation Method. SPE Drill & Compl,

Downloaded from http://onepetro.org/SPEMEDT/proceedings-pdf/18MEDT/2-18MEDT/D021S007R003/1235580/spe-189432-ms.pdf/1 by University of Trinidad & Tobago user on 07 May 2021
September 2003, 209–213 SPE-85111-PA
Ledroz, A., Bang, J., & Weston, J. 2016. New Instrument Performance Models for Combined Wellbore Surveys: A Move
Toward Optimal Use of Survey Information. SPE-178826-PA presented at the 2016 IADC/SPE Drilling Conference
and Exhibition, Fort Worth, Texas, USA 1-3 March 2016.

You might also like