You are on page 1of 16

SPE/IADC-194099-MS

Unplanned Tortuosity Index: Separating Directional Drilling Performance


from Planned Well Geometry

John D'Angelo, Pradeepkumar Ashok, and Eric van Oort, The University of Texas at Austin; Mojtaba Shahri, Taylor
Thetford, Brian Nelson, Michael Behounek, and Matthew White, Apache Corp.

Copyright 2019, SPE/IADC International Drilling Conference and Exhibition

This paper was prepared for presentation at the SPE/IADC Drilling International Conference and Exhibition held in The Hague, The Netherlands, 5-7 March 2019.

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
Wellbore tortuosity is an important metric of wellbore quality; however, it is not always an appropriate
reflection of directional drilling performance. Drilling planned tortuous features will increase wellbore
tortuosity, but this in itself says nothing about directional drilling performance. Not only is there a need for
a metric of wellbore tortuosity, it isalso necessary to have a metric of "unplanned" wellbore tortuosity. The
former provides information about wellbore quality, whereas the latteris reflective of directional drilling
performance.
The directional drilling literature contains metrics for both wellbore and unplanned tortuosity; however,
they are largely unique and difficult to relate to one another. It is desirable for the planned and unplanned
tortuosity metrics to be relatable. This would allow operators to not only quantify overall wellbore tortuosity
in real time, but also to understand how much of that tortuosity is unplanned and possibly avoidable. This,
then, opens up avenues for directional drilling performance improvement.
In this paper, a new metric, the "Unplanned" Tortuosity Index is developed on the basis of an existing
metric of wellbore tortuosity. This is done by systematically removing the effects of intended tortuous
features from the wellbore tortuosity analysis, retaining only those tortuous features in the wellbore
trajectory that are unplanned. The unplanned tortuosity index is then tested on two distinct sets of survey
data from actual wells drilled. The results are compared between the sets and with the wellbore tortuosity
metric from which the new index was derived. It is shown that thenewly developed unplanned tortuosity
index canhelp operators and directional drilling companies discern their directional drilling performance,
especially forwell paths with multiple, planned tortuous features.

Introduction
Wellbore quality is one of the many metrics that defines how successfully a borehole is drilled from surface
to target depth. While there are many approaches to quantifying the quality of a wellbore, the most intuitive
of these approaches is geometric. Tortuosity is a quantitative, geometric concept that defines the deviation of
a path from being straight or smooth. This allows us to directly quantifythe geometric quality of a well path.
2 SPE/IADC-194099-MS

When used to describe a wellbore, tortuosity can define either the deviation of a wellbore from a smooth
or straight hole (Bang, 2016; Zhou, 2016) or the deviation of a well path from a planned trajectory (Gaynor,
2002). For consistency, this paper will refer to the former as wellbore tortuosity and the latter as unplanned
tortuosity.Planned tortuosity will be used to describe the tortuosity of planned or intentionally-drilled
features on the well path. Wellbore tortuosity is the combination of both planned and unplanned tortuosities
(Gaynor, 2002). This concept is illustrated in Fig. 1 below. Note how the planned trajectory (dotted)
contains a large, curved build-up section. This is a planned tortuous feature. The as-drilled trajectory (solid
line) consists of various unplanned deviations around the planned trajectory. These are unplanned tortuous
features. The combination of these two features, regardless of intention, makes up wellbore tortuosity.

Figure 1—Planned, unplanned, and wellbore tortuosity.

Tortuosity can also be thought of in terms of scale. Macro- (at or above 90 [ft])) and micro- (below macro
level) tortuosities are believed to arise from drilling technique, bit gauge length, formation characteristics,
and the impact ofdrill string vibrations. This concept is illustrated in Fig. 2 below.
SPE/IADC-194099-MS 3

Figure 2—Macro vs. micro tortuosity.

Two factors that significantly influence the frequency and magnitude of tortuous features in a wellbore
are directional drilling methodand directional driller experience. The two primary directional drilling
methodologies involve the utilization of either a steerable downhole mud motor or a rotary steerable system
(RSS). The former is a cost-effective and popular tool used in directional drilling. Steerable mud motors
allow for changes in inclination and azimuth while drilling by sliding and rotating a bent drill string.
However, the directional driller must carefully alternate rotating and sliding the drill string to change the
trajectory to reach or stay within target zones. Immediate changes or corrections of the trajectory are not
easily achieved. An unexperienced driller could easily produce severe tortuous bends in the well by over-/
under-estimating the required tool settings. RSS steering mechanisms allow for continuous, directional
drilling control by pushing or pointing the bit through some mechanism, but in most areas these systems are
still steered and guided by humans rather than automated systems. This means that over-/under-shooting
corrections to the drill string trajectory are still a possibility. Wellbore spiraling, a common form of micro-
tortuosity (Gaynor, 2001), is also possible with either steering method.
Tortuous features, both planned and unplanned, have an impact on wellbore quality (Aadnøy, 2003;
Gamperl, 2017; Menand, 2013; Naganawa, 2017; Romero, 1995), but despite the potential negative effects
of planned curvature, it is typically safe to assume that planned tortuous features have been accounted
for and are desired. Modern directional drilling techniques have given rise to drill plans that intentionally
exhibit more and more tortuosity. Curving and bending the well path can be important when trying to e.g.
avoid wellbore collisions, to intersectmultiple targets in one well path, or to efficiently reach a reservoir
and maximize its contact area. Given the obvious intent by which certain wells are planned with a certain
tortuosity, it would not make sense to "penalize" the well path for such planned tortuous features.
The point of this discussion is that both wellbore quality and directional drilling performance are of
interest. It is necessary to have a means of distinguishing or separating unplanned tortuosity from wellbore
tortuosity, such that both quantities are preserved. This means that there is a need for a means of quantifying
both wellbore and unplanned tortuosity.
The directional drilling literature contains various attempts at developing a metric of tortuosity that can
be easily calculated and interpreted in real-time (Gaynor, 2002). However, these attempts have shown that
one metric cannot meet both needs. Two existing methods for quantifying unplanned tortuosity are the Dog
Leg Severity Index (DLSI) (Romero, 2016) and Friction Factor Back Calculation (FFBC) (Gaynor, 2002).
Dog Leg Severity (DLS) is the rate of angle change in the unplanned trajectory typically expressed in
degrees per survey interval (degree/100 [ft]). One method for calculating tortuosity in the wellbore is based
on the DLS. The DLSI quantifies to what degree the as-drilled well deviates from the planned trajectory, by
4 SPE/IADC-194099-MS

taking the difference between the sum of the planned and as-drilled DLS values at different depths along
the well, and then normalizing by the total measured depth of the well:

(1)

FFBC on the other handquantifiesunplanned tortuosity based on friction factors calculated by torque and
drag software (Gaynor, 2002). Tortuosity was defined as having three components:
(2)
where TA is the tortuosity of the as-drilled well, TP is the planned tortuosity of the well, TL is the largescale
tortuosity of the as-drilled well, and TM is the micro-tortuosity of the as-drilled well. For the FFBC study,
"micro-tortuosity" considerations were limited to wellbore spiraling. The purpose of the study was to show
that the removal of micro-tortuosity meaningfully reduces the frictional effects during tripping drill pipe
and on the overall tortuosity of the well path.
Friction factors are often used in torque and drag models as a "fudge factor" to account for parameters not
directly and explicitly accounted forin the torque and drag model (Gaynor, 2002). The FFBC methodused the
friction factors calculated for corresponding planned and as-drilled well paths to approximate the unplanned
tortuosity of the well path. The authors, however, did not provide an exact equation for their index. The
FFBC was, in fact,a scale factor used to explain the difference between the planned and as-drilled friction
factors. This appears to be areasonable assumption, based on the earlier assumption that the unplanned
tortuosity, TL + TM, was the only factor that was not considered in the planned well calculation.
Both methods described above capture unplanned tortuosity, but there seems to be no direct means of
determining the degree of wellbore tortuosity and relating it to the unplanned tortuosity that was developed.
In addition to unplanned tortuosity metrics, there are also metrics that quantify wellbore tortuosity without
distinguishing between planned and unplanned tortuosity. A more complete discussion of these metrics
is available in D’Angelo et al. 2018, but in this paper, we will focus on oneparticular method that was
developed in Zhou et al. 2016 and then modified in D’Angelo et al. 2018.
The Wellbore Tortuosity Index calculates tortuosity by analyzing curve turns along the well path. New
curve turns are scored by ratio of their arc length (Lcsi) to chord length (Lxsi), i.e. the ratio . The chord
and arc lengths are derived from the survey data, which is approximated using the Minimum Curvature
Method (MCM).
(3)

(4)

These scores are summed, weighted by a frequency modifier , where n is the number of curve turns,
and normalized by a constant, Lc. The result is the wellbore tortuosity index shown in Equations 5 and 6.

(5)

(6)

Where TIIncl is the inclination component of the wellbore tortuosity index at curve turn n in the inclination
data, and TIAzm is the azimuth component of the wellbore tortuosity index at curve turn m in the azimuth data.
SPE/IADC-194099-MS 5

An important note about this methodis that curve turn detection is key to good results. Detecting distinct
curve turns in a 3D path is not a simple task. This method utilizes the inclination and azimuth data
independently to approximate the location of curve turns. The inflection points are detected in both the
inclination and azimuth data. This is done by determining the sign change of the second derivative of the
inclination with respect to the measured depth. The same process is done for the azimuth data. This is done
independently, so we are left with a set of inflection points for the azimuth data and for the inclination
data, respectively. Looking at one set at a time, every other inflection point denotes a curve turn due to that
deviation angle. This results in a set of azimuth curve turns and a set of inclination curve turns which are
separately rated using either Equation 5 or 6. These independent tortuosity estimates are then combined
into a 3D tortuosity value.

(7)

This can be visualized as taking a set of two-dimensional views of the wellbore tortuosity and combining
them into a three-dimensional view as shown in Fig. 3.

Figure 3—Conceptual illustration of combining TIIncl and TIAzm into TI3D.

A detailed flow chart of this process is provided in the Appendix, Fig. A1.
Note that the tortuosity index on its own is a low number typically between 10-5 and 10-7 in magnitude.
This is not desirable for interpretation. To account for this, the results in this paper have been scaled by a
factor of 107. This provides a tortuosity index in the range of 0 to 100.

Method for Unplanned Tortuosity Index


The Tortuosity Indexmethoddescribed above captures overall wellbore tortuosity. This means that it does
not discriminate between planned and unplanned tortuous features. This section describes a new method
that attempts to remove the effects of planned or intentional tortuous features from the existing method.
Equation 2 defines the overall wellbore tortuosity as the combination of planned and unplanned tortuosity.
Since the TI method captures overall wellbore tortuosity, Equation 2 can be rewritten as follows:
(8)
Where TIis the result of applying the TImethod to the as-drilled well and capture the overall wellbore
tortuosity of the as-drilled well. PTI is the result of applying the TI method to the planned well and captures
6 SPE/IADC-194099-MS

the planned tortuosity. Please note that in the original publications this TI is referred to as TI3D, but for
consistency, we will simply refer to it as TI. Lastly, UPTI is the unplanned tortuosity index which can be
viewed as the difference between the TI and PTI as shown in Equation 9 below. This effectively removes
the impact of the planned tortuosity from the TI result.
(9)
Based on this explanation, the direct subtraction of the PTI from the TIseems like an obvious approach.
However, this method fails in practice because the PTI is not directly relatable to the tortuosity index of
the as-drilled well, TI.
The approach proposedby this paper is meant to overcome this difficulty by using the planned well survey
to approximate the location of intended tortuous features on the as-drilled well path. First, the TI is calculated
for the as-drilled well path. The planned survey is then interpolated using MCM to match the survey interval
of the as-drilled survey (Hassan, 2012). Curve turn analysis is performed on the interpolated, planned survey
to determine where tortuous features are planned to occur. Lastly, these planned locations are mapped as
functions of measured depth onto the as-drilled well path to provide a set of expected curve turn locations.
For example, say we detected the following planned locations on a planned well that is 10,000 ft long:

For this example, let us assume we want to map locations A and B onto the as-drilled well that is 10,100
ft long. Location A is at 25% of the total planned well length and location B is at 27.5% of the total planned
well length. These would map to the following locations on the as-drilled well:

If an exact match is not found for the mapped point, the closest match on the as-drilled survey is selected.
These mapped points comprise the end-points of the plannedtortuous features on the as-drilled well path.
The tortuosity of these features is then calculated using the TI. This yields a step function of the planned
wellbore tortuosity of the as-drilled well. These mapped tortuosity values are linearly interpolated to match
the interval of the TI analysis of the as-drilled well. This set of interpolated, mapped tortuosity index data
points will be referred to as MTI. This MTI is then subtracted from the TI to yield the UPTI, a quantitative
measure of the tortuosity due to unplanned tortuous features as a function of measured depth along the
wellbore.
(10)
This method, summarized in Fig. 4, provides us not only with the TI, a measure of the overall wellbore
tortuosity of the as-drilled well, but also the UPTI, a measure of the unplanned tortuosity of the well path.
SPE/IADC-194099-MS 7

Figure 4—Flow chart of unplanned tortuosity index assessment.

Conceptually, the unplanned tortuosity index is a reasonable approach towards calculating the degree
of unplanned tortuosity by leveraging the TI. However, in some cases, the rate of increase of the mapped
tortuosity of the well is higher than the rate of increase of the as-drilled tortuosity. As a result, it is possible for
the calculated value of the unplanned tortuosity index to decrease. This would not make sense for an overall
tortuosity index, since tortuosity can only be added and not removed from a well (unless the geometry were
changed, such as when a well is plugged back and cemented and subsequently re-drilled along a different
path). To prevent this decrease, the unplanned tortuosity increase is simply assumed to remain constant if
the difference between the TI and MTI values were to decrease. The consequences of this are discussed
later in the paper.
The proportional mapping described above is useful for handling minor discrepancies between planned
and as-drilled surveys, but proportionality may not be the best relationship for mapping points between
surveys. Especially in cases of severe deviation between the planned and as-drilled survey where the scope
of the project changesand discrepancies between the planned and as-drilled survey could exceed hundreds
of feet. At such a point, proportional mapping would be insufficient to capture such deviation and the
unplanned tortuosity index should be ignored until the planned survey can be updated.
8 SPE/IADC-194099-MS

Application of Unplanned Tortuosity Index


The unplanned tortuosity index was tested on horizontal wells drilled in North America. These wells were
divided into two sets, one set (Set 1) in which the wells had only one planned build section, and a second
set (Set 2) in which wells had two planned build sections. The analyses for Sets 1 and 2 are available in the
Appendix. In this section, we will focus exclusively on one representative well from each set.
The expectation for this analysis was that wells from Set 1 would inherently have lower tortuosity values
than Set 2 due to the lack of the second build section. Fig. 5 shows the wellbore tortuosity analysis of a
well from Set 1. This well was representative of the other wells in the set. The vertical section was drilled
straight, followed by a build section, followed subsequently by a tortuous lateral section. The tortuosity
index value for most of the wells in Set 1 was between 20 and 30. Looking at the planned well, it has a
relatively low degree of tortuosity caused almost entirely by the influence of the build section.

Figure 5—Wellbore tortuosity analysis of as-drilled and planned well surveys for a well from Set 1.

Fig. 6 shows the unplanned tortuosity index at work. After the TI(blue curve) is calculated, MTI is
calculated and subtracted from the TI, resulting in the UPTI(black curve). This analysis suggests that the
curve section had some degree of planned tortuosity, but that most of the tortuous features in the well were
actually unplanned. Also note that MTI (green curve) is not the same as PTI(orange curve). MTI captures
the impact of intentional tortuous features that, in practice, may deviate slightly from the original plan.
This allows for some flexibility in the event that real-time decisions are not immediately corrected in the
planned survey. However, for the most accurate comparison, major changes to the planned trajectory should
be reflectedin the planned survey as quickly as possible.
SPE/IADC-194099-MS 9

Figure 6—Unplanned tortuosity analysis for a well from Set 1.

The wellshown in Fig. 7 is representative of the Set 2 wells. Notice the two significant curve sections
in the well path. The TI analysis of the as-drilled well shows a higher value than most of the set 1 wells.
Looking at the PTI analysis, these increases in TI are closely tracked by the PTI.

Figure 7—Wellbore tortuosityanalysis of as-drilled and planned well surveys for a well from Set 2.

The calculated UPTI(Fig. 8 below)suggests that there was significant tortuosity in the region between the
two curve sections, but the rest of the well was drilled with little significant unplanned tortuosity. Another
wayto interpret the flat section from 6,000 ft to target depth is that the plan for the well incorporated a
large tortuous feature. From a planning perspective, this means that the well was planned, designed, and
deemed to be capable of handling a certain degree of tortuosity. Since these planned parameters were not
exceeded by the rate increase of as-drilled wellbore tortuosity, the tortuosityintroduced by the directional
drilling process is deemed to be rather insignificant.
10 SPE/IADC-194099-MS

Figure 8—Unplanned tortuosity analysis for a well from Set 2.

Discussion
The unplanned tortuosity index was evaluated on horizontal shale wells drilled in North America. One set
of wells had only one intentional build section (Set 1), while the other set had two intentional build sections
(Set 2). Set 2 wells consistently scored higher on the wellbore tortuosity index than Set 1 wells as shown
in Fig. 9.

Figure 9—Wellbore tortuosity index comparison of Sets 1 and 2.

The unplanned tortuosity index showed that Set 2 wells owed a great deal of their tortuosity to the multiple
build sections planned into them. The unplanned tortuosities calculated by removing the tortuosity of these
intended features showed that the performance of the Set 2 wells was in fact comparable to the performance
of many of the wells in Set 1, as shown in Fig. 10 below.
SPE/IADC-194099-MS 11

Figure 10—Unplanned tortuosity index comparison of Sets 1 and 2.

This validated the idea that the wellbore tortuosity index must not be tied indiscriminatelyto directional
drilling performance. The wellbore tortuosity index captures the overall tortuosity of a well path, and in more
complex wells, this will include intentional tortuous features. The unplanned tortuosity index, on the other
hand, comparesthe wellbore tortuosity index for the as-drilled well with the intended wellbore tortuosity
index of the well. It thereby provides a much more appropriate indicator of directional drilling performance.

Conclusions
There is a need for a metric or set of metrics that can quantify both actual and unplanned tortuosity in a real-
time, easy-to-interpret fashion. This paper has taken the following steps towards filling this need:

• The unplanned tortuosity index was developed to quantify unplanned tortuosity by leveragingan
existing metric of wellbore tortuosity.
• The unplanned tortuosity index was used to successfully discriminate contributions of unplanned
tortuous features from those of planned tortuous features in two sets of test cases obtained for wells
drilled in the field.
• The unplanned tortuosity indexwas used as a metric of tortuosity that could be easily compared
between the two disparate sets of wells.
The wellbore and unplanned tortuosity indices have the potential to be important components of the
directional drilling, decision-making process, but not if they are just numbers. It is necessary to provide
the driller with concise, easily-digestible visualizations and recommendations based on these indices. Fig.
A6 in the Appendix is an early-stage concept of such a recommendation system. The system shows the
driller the results of the tortuosity analyses in the context of a vertical and a plan view of the developing
well path. A dialogue box provides updated recommendations based on factors such as the magnitude of
wellbore or unplanned tortuosity.

Acknowledgements
The authors are grateful to the Rig Automation and Performance Improvement in Drilling (RAPID) Program
at the University of Texas at Austin and its sponsors for providing the data and financial support for the
development of this work. The authors would also like to thank Apache Corporation for their continued
assistance and permission to publish this work.
12 SPE/IADC-194099-MS

Nomenclature
DLSI = Unplanned tortuosity comparison of dogleg severity of planned and as-drilled surveys
DLSdrilled = Dogleg severity of the as-drilled well path
DLSplanned = Dogleg severity of the planned well path
EW = Measurement in positive easting/westing direction
Lc = Total length of planned well
Lcsi = Along hole distance from one survey point to another
Lxsi = Straight line distance from one survey point to another
MD = Measured depth
MD0 = Initial measured depth
MDTD = Measured depth at final target depth
MTI = Tortuosity of the as-drilled well path calculated from points mapped from the plan
m= Current number of curve turns detected in azimuth data
NS = Measurement in positive northing/southing direction
n= Current number of curve turns detected in inclination data
PTI = Tortuosity calculated for planned well path
TA = Tortuosity of the as-drilled well
TL = Large-scale tortuosity of as-drilled well path
TM = Micro-scale tortuosity of as-drilled well path
TP = Tortuosity of planned well
TD = Target depth
TI = Tortuosity index of 3D well path
TIInc/Azm = Tortuosity index calculated in the inclination or azimuth projection of 3D well path
TVD = True vertical depth
UPTI = Unplanned tortuosity index

References
Bang, J., Jegbefume, O., Ledroz, A., Weston, J., & Thompson, J. (2017, May 1). Analysis and Quantification of Wellbore
Tortuosity. Society of Petroleum Engineers. doi:10.2118/173103-PA
D'Angelo, J., Ashok, P., van Oort, E., Shahri, M., Nelson, B., Thetford, T., & Behounek, M. (2018, August 9).
Monitoring Wellbore Quality in Real-Time Using a Geometrically Derived Tortuosity Metric. Unconventional
Resources Technology Conference. doi:10.15530/URTEC-2018-2901598
Gamperl, F. (2017). Analysis of the Impact of Wellbore Quality. Montanuniversität Leoben.
Gaynor, T. M., Chen, D. C.-K., Stuart, D., & Comeaux, B. (2001, January 1). Tortuosity versus Micro-Tortuosity - Why
Little Things Mean a Lot. Society of Petroleum Engineers. doi:10.2118/67818-MS
Gaynor, T., Hamer, D., Chen, D. C.-K., & Stuart, D. (2002, January 1). Quantifying Tortuosities by Friction Factors in
Torque and Drag Model. Society of Petroleum Engineers. doi:10.2118/77617-MS
Hassan, I. (2012). Survey interpolation: A software for calculating correct wellpath between survey stations (Master's
thesis, University of Stavanger, Norway).
Menand, S. (2013, February). Borehole Tortuosity Effect on Maximum Horizontal Drilling Length Based on Advanced
Buckling Modeling. In Oral presentation of paper AADE-13-FTCE-21 at the AADE National Technical Conference
and Exhibition, Oklahoma City, Oklahoma (pp. 26–27).
Naganawa, S., Kudo, H., & Matsubuchi, H. (2016, November 7). Simulation Study on Influences of Wellbore Tortuosity
on Hole Cleaning in Extended-Reach Drilling. Society of Petroleum Engineers. doi:10.2118/183409-MS
Romero, J., Mack, M. G., & Elbel, J. L. (1995, January 1). Theoretical Model and Numerical Investigation of Near-
Wellbore Effects in Hydraulic Fracturing. Society of Petroleum Engineers. doi:10.2118/30506-MS
Zhou, Y., Zheng, D., Ashok, P., & van Oort, E. (2016, March 1). Improved Wellbore Quality Using a Novel Real-Time
Tortuosity Index. Society of Petroleum Engineers. doi:10.2118/178869-MS
SPE/IADC-194099-MS 13

Appendix

Figure A1—Flow chart of wellbore tortuosity index calculation (after Zhou et al., 2016).
14 SPE/IADC-194099-MS

Figure A2—Set 1 well surveys.

Figure A3—Unplannedtortuosity analysis of Set 1 wells.


SPE/IADC-194099-MS 15

Figure A4—Set 2 well surveys.

Figure A5—Unplanned tortuosity analysis of Set 2 wells.


16 SPE/IADC-194099-MS

Figure A6—Real-time tortuosity index analysis, 2D View.

You might also like