You are on page 1of 16

SPE 56702

Accuracy Prediction for Directional MWD
Hugh S. Williamson, SPE, BP Amoco

Copyright 1999, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the 1999 SPE Annual Technical Conference and Exhibition
held in Houston, Texas, 3–6 October 1999.
This paper was selected for presentation by an SPE Program Committee following review of information
contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The
material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers,
its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial
Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of
any part of this paper for commercial purposes without the written consent of the Society of Petroleum
Engineers 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 where
and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 750833836, U.S.A., fax 01-972-952-9435.

Abstract
This paper describes a new method for predicting wellbore position
uncertainty which responds to the current needs of the Industry. An
error model applicable to a basic directional MWD service is presented
and used for illustration. As far as possible within the limitations of
space, the paper is a self-contained reference work, including all the
necessary information to develop and test a software implementation of
the method.
The paper is the product of a collaboration between the many
companies and individuals cited in the text.
Introduction
As the Industry continues to drill in mature oil provinces, the dual
challenges of small geological targets and severe well congestion increase
the importance of quantifying typical wellbore positional errors. The
pioneering work of the 1970’s culminated in the paper by Wolff and de
Wardt1. Their approach, albeit extensively modified and added to, has
remained the de facto Industry standard to this day. At the same time,
various shortcomings of the method have been identified2,3,4, but are not
discussed further here.
In recent years, a number of factors have created the opportunity for
the Industry to develop an alternative method:
• risk-based approaches to collision avoidance and target hitting
require position uncertainties with associated confidence levels,
something which Wolff and de Wardt specifically avoided.
• changing relationships brought about by integrated service contracts
have forced directional drilling and survey companies to share
information on tool performance.

• the development of several new directional software products and
their integration with sub-surface applications has provided the
necessity and the opportunity to develop new means of communicating
and visualising positional uncertainty.
This paper provides a three-part response to this need:
1. Error Model for Basic MWD, based on the current state of
knowledge of a group of Industry experts. There are several reasons
why directional MWD is the most suitable survey service to illustrate a
new method of error modelling:
• the error budget is dominated by environmental effects, so that
accuracy differences attributable to tools alone are minimal.
• it is the survey tool of choice for most directional wells, where
position uncertainty is of greatest concern.
• the physical principles of its operation, including the navigation
equations, are in the public domain.
2. Mathematical Basis, being a rigorous description of the
propagation of errors in stationary tools. Fit-for-purpose error models
using the same basis are in development for inertial and continuous
gyroscopic tools, although some simplification and compromise is
inevitable. A rigorous treatment of continuous survey tools would
probably have too restricted a cognoscenti to be practical.
3. Standard Examples and Results. Despite the apparent simplicity
of the Wolff and de Wardt method, different software implementations
generally give subtly different results. While an effort has been made in
this paper to provide a comprehensive description of the new method,
there will surely remain some areas of ambiguity or confusion. In such
cases, reproduction of the numerical results at the end of the paper will
act as a powerful criterion for “validation”. The results were generated
by the author using experimental software and have since been
duplicated by commercially available software.
Genesis of the Work. The content of this paper is the fruit of two
collaborative groups.
ISCWSA. The Industry Steering Committee on Wellbore Survey
Accuracy is an informally constituted group of companies and

2

HUGH S. WILLIAMSON

individuals established following the SPWLA Topical Conference on
MWD held in Kerrville, Texas in late 1995. The Group’s broad
objective is “to produce and maintain standards for the Industry relating
to wellbore survey accuracy”. Much of the content of this paper, and
specifically the details of the basic MWD error model, had its genesis in
the Group’s meetings, which were distinguished by their open and cooperative discussions.
Four Company Working Group. The ISCWSA being too large a
forum to undertake the detailed mathematical development of an error
propagation model, this was completed by a small working group from
Sysdrill Ltd., Statoil, Baker Hughes INTEQ and BP Exploration. The
mathematical model created by the group and described below has been
made freely available for use by the Industry.
Assumptions and Definitions
The following assumptions are implicit in the error models and
mathematics presented in this paper:
1. Errors in calculated well position are caused exclusively by the
presence of measurement errors at wellbore survey stations.
2. Wellbore survey stations are, or can be modelled as, three-element
measurement vectors, the elements being along-hole depth, D,
inclination, I, and azimuth A. The propagation mathematics also requires
a toolface angle, τ, at each station.
3. Errors from different error sources are statistically independent.
4. There is a linear relationship between the size of each measurement
error and the corresponding change in calculated well position.
5. The combined effect on calculated well position of any number of
measurement errors at any number of survey stations is equal to the
vector sum of their individual effects.
No restrictive assumptions are made about the statistical distribution
of measurement errors.
Error sources, terms and models. An error source is a physical
phenomenon which contributes to the error in a survey tool
measurement. An error term describes the effect of an error source on a
particular survey tool measurement. It is uniquely specified by the
following data:
• a name
• a weighting function, which describes the effect of the error ε on
the survey tool measurement vector p. Each function is referred to by a
mnemonic of up to four letters.
• a mean value, µ.
• a magnitude, σ, always quoted as a 1 standard deviation value.
• a correlation coefficient ρ1 between error values at survey stations
in the same survey leg. (In a survey listing made up of several
concatenated surveys, a survey leg is a set of contiguous survey
stations acquired with a single tool or, if appropriate, a single tool type).
• a correlation coefficient ρ2 between error values at survey stations
in different survey legs in the same well.
• a correlation coefficient ρ3 between error values at survey stations

SPE 56702

in different wells in the same field.
To ensure that the correlation coefficients are well defined, only four
combinations are allowed.
Propagation Mode
Random (R)
Systematic (S)
Per-well (W)
Global (G)

ρ1
0
1
1
1

ρ2
0
0
1
1

ρ3
0
0
0
1

ρ1, ρ2 and ρ3 are to be considered properties of the error source, and
should be the same for all survey legs.
An error model is a set of error terms chosen with the aim of
properly accounting for all the significant error sources which affect a
survey tool or service.
An Error Model for “Basic” MWD
For the survey specialist in search of a “best estimate” of position
uncertainty it is tempting to differentiate minutely between tools types
and models, running configurations, BHA design, geographical location
and several other variables. While justifiable on technical grounds, such
an approach is impractical for the daily work of the well planner. The
time needed to find out this data for historical wells, and for many
planned wells, is simply not available.
The error model presented in this section is intended to be
representative of MWD surveys run according to fairly standard quality
procedures. Such procedures would include
• rigorous and regular tool calibration
• survey interval no greater than 100 ft
• non-magnetic spacing according to standard charts (where no axial
interference correction is applied)
• not surveying in close proximity to existing casing strings or other
steel bodies
• passing standard field checks on G-total, B-total and dip.
The requirement to differentiate between different services may be met
by defining a small suite of alternative error models. Examples covered
in this paper are:
• application or not of an axial interference correction
• application or not of a BHA sag correction
Alternative models would also be justified for:
• In-field referenced surveys
• In-hole (gyro) referenced surveys
• Depth-corrected surveys
The model presented here is based on the current state of knowledge and
experience of a number of experts. It is a starting point for further
research and debate, not an end-point.
Sensor Errors. MWD sensors will typically show small shifts in
performance between calibrations. We may make the assumption that

SPE 56702

ACCURACY PREDICTION FOR DIRECTIONAL MWD

the shifts between successive calibrations are representative of the shifts
between calibration and field performance. On this basis, two major
MWD suppliers compared the results of successive scheduled
calibrations of their tools. Paul Rodney examined 288 pairs of
calibrations, and noted the change in bias (ie. offset error), scale factor
and misalignment for each sensor. Wayne Phillips did the same for 10
pairs of calibrations, except that sensor misalignments were not
recorded.
Andy Brooks has demonstrated that if a sensor is subject to a scale
error and two orthogonal misalignments, all independent and of similar
magnitude, the combination of the three error terms is equivalent to a
single bias term. This term need not appear explicitly in the error model,
but may be added to the existing bias term to create a “lumped” error.
This eliminates the need for 20 extra weighting functions corresponding
to sensor misalignments.
The data from the MWD suppliers suggest that in-service sensor
misalignments are typically smaller than scale errors. As a result, only a
part of the observed scale error was “lumped” with the misalignments
into the bias term, leaving a residual scale error which is modeled
separately. In this way, four physical errors for each sensor were
transformed into two modeled terms. The results were as follows:
Error Source
Accelerometer biases
Accelerometer scale factors
Magnetometer biases
Magnetometer scale factors

weighting
function
ABX,Y,Z
ASX,Y,Z
MBX,Y,Z
MSX,Y,Z

magnitude
0.0004 g
0.0005
70 nT
0.0016

prop.
mode
S
S
S
S

These figures include errors which are correlated between sensors, and
which therefore have no effect on calculated inclination and azimuth (the
exception being the effect of correlated magnetometer errors on
interference corrected azimuths). It could be argued that the
magnetometer scale factor errors in particular (which may be influenced
by crustal anomalies at the calibration sites) should be reduced to
account for this.
BHA magnetic interference. Magnetic interference due to steel in
the BHA may be split into components acting parallel (axial) and
perpendicular (cross-axial) to the borehole axis.
Axial Interference. Several independent sets of surface
measurements of magnetic pole strengths have now been made.
Observed root-mean-square values are:

Item
RMS

Pin
Box
pole strength(sample size)

Source

Drill collar

Stabiliser

Motor

505 µWb (8)
605 µWb (11)

3

435 µWb (11)
511 µWb (4)

Grindrod, Wolff6
Lotsberg5
McElhinney 7

177 µWb (6)
396 µWb (10)
369 µWb (5)

189 µWb (10)
408 µWb (10)

Grindrod, Wolff
Lotsberg
McElhinney

340 µWb (12)

419 µWb (10)

Lotsberg

Oddvar Lotsberg also computed pole strengths for 41 BHAs from the
results of an azimuth correction algorithm. The RMS pole strength was
369 µWb (micro-Webers).
These results suggest that 400 µWb is a reasonable estimate for the 1
s.d. pole strength of a steel drill string component where further
information is lacking. This is useful information for BHA design, but
cannot be used for uncertainty prediction without a value for nonmagnetic spacing distance. Unfortunately, there is no “typical” spacing
used in the Industry, and we must find another way to estimate the
magnitude of this error source.
A well-established Industry practice is to require non-magnetic
spacing sufficient to keep the azimuth error below a fixed tolerance
(typically 0.5° at 1 s.d.) for assumed pole strengths and a given hole
direction. This tolerance may need to be compromised in the least
favourable hole directions. For a fixed axial interference field, and
neglecting induced magnetism, azimuth error is strongly dependent on
hole direction, being proportional to sinIsinAm. Thus to model the
azimuth error in uncorrected surveys, we require a combination of error
terms which
• predicts zero error if the well is vertical or magnetic north/south
• predicts errors somewhat greater than the usual tolerance if the well
is near horizontal and magnetic east/west
• predicts errors near the usual tolerance for other hole directions.
These requirements could be met by constructing some artificial
weighting function, but this would violate our restriction to physically
meaningful error terms. A constant error of 0.25° and a directiondependent error of 0.6°sinIsinAm is perhaps the best we can achieve by
way of a compromise. It is legitimate to consider these values
representative of 1 standard deviation, since the pole strength values
which underlie the non-magnetic spacing calculations are themselves
quoted at 1 s.d.
Both error terms may be propagated as systematic, although there is
theoretical and observational evidence4 that this error is asymmetric,
acting in the majority of cases to swing magnetic surveys to the north in
the northern hemisphere. Giving the direction-dependent term a mean
value of 0.33° and a magnitude of 0.5° reproduces this asymmetry (with
about 75% of surveys being deflected to the north), while leaving the
root-mean-square error unchanged.
Axial interference errors are not modelled for surveys which have been
corrected for magnetic interference.
Cross-Axial Interference Cross-axial interference from the BHA is
indistinguishable from magnetometer bias, and propagates in the same

4

HUGH S. WILLIAMSON

way. Anne Holmes 8 analysed the magnetometer biases for 78 MWD
surveys determined as a by-product of a multi-station correction
algorithm. Once a few outliers - probably due to magnetic “hot-spots”
and hence classified as gross errors - had been eliminated, the remaining
observations gave an RMS value of 57nT. This figure is somewhat
smaller than the 70nT attributable to magnetometer bias alone. The
conclusion must be that cross-axial interference does not, in the average,
make a significant contribution to the overall MWD error budget, and
may be safely left out of the model.
Tool Misalignment. Misalignment is the error caused by the alonghole axis of the directional sensor assembly being out-of-parallel with
the centre line of the borehole. The error may be modeled as a
combination of two independent phenomena:
BHA sag is due to the distortion of the MWD drill collar under
gravity. It is modelled as confined to the vertical plane, and proportional
to the component of gravity acting perpendicular to the wellbore (ie.
sinI). The magnitude of the error depends on BHA type and geometry,
sensor spacing, hole size and several other factors. Two-dimensional
BHA models typically calculate inclination corrections of 0.2° or 0.3°
for poorly stabilised BHAs in horizontal hole5. For well stabilised
assemblies the value is usually less than 0.15°. In the absence of better
information, 0.2° (at 1 s.d.) may be considered a realistic input into the
basic error model.
Sag corrections, if they are applied, are calculated on the often
unjustified assumptions of both the hole and stabilisers being in gauge.
Data comparisons by the author suggest a typical efficiency of 60% for
these corrections, leaving a post-correction residual sag error of 0.08°.
Assuming similar BHAs throughout a hole section, all BHA sag errors
may be classified as systematic.
Radially symmetric misalignment is modelled as equally likely to
be oriented at any toolface angle. John Turvill made an estimate of its
magnitude based on the tolerances on several concentric cylinders:
• Sensor package in housing. Tolerances on three components are
(i) clearance, 0.023°, (ii) concentricity, 0.003°, (iii) straightness of
sensor package, 0.031°.
• Sensor housing in drill collar. For a probe mounted in a
centralised, retrievable case, 0.063°.
• Collar bore in collar body. Typical MWD vendors’ tolerance is
0.05°.
• Collar body in borehole. The API tolerance on collar
straightness equates to 0.03°. MWD vendors’ specifications are
typically somewhat more stringent.
The root-sum-square of these figures is 0.094°. Being based on
maximum tolerances, it is probably an over-estimate for stabilised rotary
assemblies.
An analysis by the author of the variation in measured inclination over
46 rotation shots produced a root-mean-square misalignment of 0.046°.
Simulations show that within this figure, about 0.007° is attributable to
the effect of sensor errors.

SPE 56702

An additional source of misalignment - collar distortion outside the
vertical plane due to bending forces - may be estimated using 3dimensional BHA models. 0.04° seems to be a typical value. This error
differs from those above by not rotating with the tool. It should
therefore strictly have its own weighting function. Being so small, it
seems justifiable on practical (if not theoretical) grounds, to include it
with the other sources of radially symmetric misalignment. This leaves
us with an estimate for the error magnitude of 0.06°. This figure may be
a significant underestimate where there is an aggressive bend in the BHA
or a probe-type MWD tool is in use. This error term may be considered
systematic.
Magnetic field uncertainty. For basic MWD surveys, only the value
assumed for magnetic declination affects the computed azimuth.
However, conventional corrections for axial interference require
estimates of the magnetic dip and field strength. Any error in these
estimates will cause an error in the computed azimuth.
A study commissioned from the British Geological Survey by Baker
Hughes INTEQ 9 investigated the likely error in using a global
geomagnetic model to estimate the instantaneous ambient magnetic field
downhole. Five sources of error were identified:
• Modelled main field vs. actual main field at base epoch
• Modelled secular variation vs. actual secular variation
• Regular (diurnal) variation due to electrical currents in the
ionosphere
• Irregular temporal variation due to electrical currents in the
magnetosphere
• Crustal anomalies
By making a number of gross assumptions, and by considering typical
drilling rates, the current author has distilled the results of the study into
a single table:
Error Source

Main field model
Secular variation
Daily variation
Irregular variation
Crustal anomaly

error magnitude
declination
dip
field
0.012°*
0.005°
0.017°*
0.013°
0.045°†
0.011°†
0.110°†
0.043°†
0.476°
0.195°

total

prop.
mode

3 nT
10 nT
11 nT†
45 nT†
120 nT

G
G
R/S‡
R/S‡
G

* below 60° latitude N or S
† at 60° latitude N or S
‡ daily and irregular variation are partially randomised between surveys.
Correlations between consecutive stations are approximately 0.95 and
0.5 for the two error sources.

The dominant error source is crustal anomalies, caused by varying
magnetisation of rocks in the Earth’s crust. The figures shown are
representative of the North Sea. Some areas, particularly those at higher
latitudes and where volcanic rocks are closer to the surface, will show
greater variation. Other areas, where sedimentary rocks dominate, will
show less.

SPE 56702

ACCURACY PREDICTION FOR DIRECTIONAL MWD

In the absence of any other information, the uncertainty in an estimate
of the magnetic field at a given time and place provided by a global
geomagnetic model may be obtained by summing the above terms
statistically. There is one complication - some account must be taken of
the increasing difficulty of determining declination as the horizontal
component of the magnetic field decreases. This can be achieved by
splitting this error into two components: one constant and one inversely
proportional to the horizontal projection of the field, BH. For the
purposes of the model, the split has been defined somewhat arbitrarily,
while ensuring that the total declination uncertainty at Lerwick, Shetland
(BH = 15000nT) is as predicted by the BGS study (0.49°). Being
dominated by the crustal anomaly component, all magnetic field errors
may be considered globally systematic and summarised thus:
Error Source
Declination (constant)
Declination (BH-dependent)
Dip angle
Total field

weighting
function
AZ
DBH
MFD
MFI

magnitude
0.36°
5000°nT
0.20°
130nT

prop.
mode
G
G
G
G

Along-Hole Depth Errors. Roger Ekseth10 identified 14 physical
sources of drill-pipe depth measurement error, wrote down expressions
to predict their magnitude, and by substituting typical parameter values
into the expressions predicted the total error for a number of different
well shapes. He then proposed a simplified model of just four terms,
and chose the magnitudes of each to match the predictions of the full
model as closely as possible. The results were as follows.
Error Source

Random ref.
Systematic ref.
Scale
Stretch-type

error
error magnitude (1 s.d.)
proportional land rig
floating rig
to

1
1
D
D.V

0.35 m
0m
2.4×10-4
2.2×10-7 m-1

2.2 m
1m
2.1×10-4
1.5×10-7 m-1

prop.
mode

R
S
S
G

For the purposes of the basic model, the values for the land rig (or,
equivalently a jack-up or platform rig) may be chosen. The stretch-type
error, which dominates the other terms in deep wells, models two
physical effects - stretch and thermal expansion of the drill pipe. Both
these effects generally cause the drill string to elongate, so it may be
appropriate to apply this term as a bias (see below). If this is done, a
mean value of 4.4×10-7 m-1 should be used, since Ekseth effectively
treated his estimates of these errors as 2 s.d. values.
Errors omitted from the Basic MWD Model. Some errors known to
affect MWD surveys have nonetheless not been included in the basic
error model.
Tool electronics and resolution. The overall effect on accuracy

5

caused by the limitations of the tool electronics and the resolution of the
tool-to-surface telemetry system is not considered significant. Such
errors will tend to be randomised over long survey intervals.
External magnetic interference. Ekseth10 discusses the influence
of remanent magnetism in casing strings on magnetic surveys, and gives
expressions for azimuth error when drilling (a) out of a casing shoe and
(b) parallel to an existing string. Although certainly not negligible, both
error sources are difficult to quantify, and equally difficult to
incorporate within error modelling software. It seems preferable to
manage these errors by applying quality procedures designed to limit
their effect.
Effect of survey interval and calculation method. The method
presented in this paper relies on the assumption that error-free
measurement vectors p will lead to an error-free wellbore position vector
r. If minimum curvature formulae are used for survey calculation, this
assumption will only be true when the well-path between stations is an
exact circular arc. The resulting error may be significant for sparse data,
but may probably be neglected so long as the station interval does not
exceed 100 ft.
Gravity field uncertainty. Differences between nominal and actual
gravity field strengths will typically have no effect on MWD accuracy
since only the ratio of accelerometer measurements are used in the
calculation of inclination and azimuth.
Gross errors. Any attempt at a comprehensive discussion of MWD
error sources must at least acknowledge the possibility of gross errors sometimes called human errors. These errors lack the predictability and
uniformity of the physical terms discussed above. They are therefore
excluded from the error model, with the assumption that they are
adequately managed through process and procedure.
Propagation Mathematics
The mathematical algorithm by which wellbore positional uncertainty is
generated from survey error model inputs is based on the approach
outlined by Brooks and Wilson3. The development of this work
described here was carried out by the Working Group referred to in the
Introduction.
A physical error occurring at a survey station will result in an error,
in the form of a vector, in the calculated well position. From ref. 3:
ei = σi

d r ∂p
dp ∂εi

…..(1)

where ei is a vector-valued random variable, (a vector error), σi is the
magnitude of the ith error source, ∂p/∂εi is its “weighting function” and
dr/dp describes how changes in the measurement vector affect the
calculated well position. It is sufficient to assume that the calculated
displacement between consecutive survey stations depends only on the
survey measurement vectors at these two stations. Writing ∆rk for the
displacement between survey stations k-1 and k, we may thus express
the (1 s.d.) error due to the presence of the ith error source at the kth

6

HUGH S. WILLIAMSON

survey station in the lth survey leg as the sum of the effects on the
preceding and following calculated displacements:
 d∆rk d ∆rk +1  ∂p k
ei ,l ,k = σi,l 
+

 dp k
d p k  ∂εi

…..(2)

where σi,l is the magnitude of the ith error source over the lth survey leg,
and pk is the instrument measurement vector at the kth survey station.
The total position error at a particular survey station K in survey leg
L, will be the sum of the vector errors ei,l,k taken over all error sources i
and all survey stations up to and including K. The uncertainty in this
position error is expressed in the form of a covariance matrix:
CK =

∑ ∑ ∑

errors k1≤ K k2 ≤ K
i

(

)

ρ εi, l1 ,k 1 , εi, l2 , k2 e i,l1 ,k1 . eiT,l2 ,k 2

…..(3)

Weighting functions. The weighting function for a particular error
source is a 3×1 vector, the elements of which describe the effect of a
unit error on the measured along-hole depth, inclination and azimuth.
For example, the weighting functions for constant and BH-dependent
magnetic declination errors are:


0
∂p


=
0

∂ε DBH 
1 / ( B cosΘ ) 

…(4,5)

∂p
∂ε SAG

∂p
∂ε AMID



0


= 
0

sin I sin Am 

…(6,7)

and for reference, scale and stretch-type depth errors they are:
∂p
∂ε DREF

 1
 
=  0
 0

D
∂p
 
=  0
∂ε DSF
 0 

 DV 
∂p


= 0 
∂ε DST
 0 

(

...(11)

)

 G B −G B
G2x + G 2y + Gz2
x y
y x

Am = tan −1 
 B z Gx2 + G2y − Gz G x B x + G y B y

(

)





...(12)

G y = − G sin I cos τ

G z = G cos I

(

)

and making use of the inverse relations:
G x = − G sin I sin τ

…(13,14,15)
Bx = B cosΘcos I cos Am sinτ − B sinΘsin I sinτ + B cosΘsin Am cosτ

Bz = B cos Θ sin I cos Am + B sin Θ cos I

…(16,17,18)

Taking the X-accelerometer bias (ABX) as an example,
∂I
− 1 ∂ cos I
− 1  GxGz 
cos I sinτ
=
=
−
=−
3
∂G x sin I ∂Gx
sin I  G 
G

…(19)

and similarly,
∂Am ( cosI sin Am sin τ − cos Am cosτ) tanΘ + cot I cosτ
=
…(20)
∂G x
G
The appropriate weighting function is therefore:
0

 …(21)
1

=
− cos I sin τ

∂ε ABX G 
( cos I sin Am sin τ − cos Am cos τ) tan Θ + cot I cos τ
∂p

For BHA sag and direction-dependent axial magnetic interference they
are:
 0 


= sin I 
 0 



Gz


I = cos−1 
2
2
2
 Gx + Gy + Gz 

By = BcosΘcos I cos Am cosτ − B sinΘsinI cosτ − B cosΘsin Am sinτ

where ρ(εi,l1,k1,εi,l2,k2) is the correlation coefficient between the value of
the ith error source at the k 1th station (in the l1th leg) and the k 2th
station (in the l2th leg). In practice, it is more convenient to sum
separately the contributions of errors with different propagation
characteristics. Details are in Appendix A.

0
∂p
 
= 0
∂ε AZ  
 1 

SPE 56702

…(8,9,10)

Weighting functions for sensor errors. Tool axes and toolface
angle, τ, are defined in Fig 1. There are 12 basic sensor error sources (a
bias and scale factor for each of 3 accelerometers and 3 magnetometers)
and each requires its own weighting function. These are obtained by
differentiating the standard navigation equations for inclination and
azimuth:

Effect of axial interference correction. When a simple axial
magnetic interference correction is applied, (12) is no longer used, and
different weighting functions are required for sensor errors. The
following analysis is due to Andy Brooks.
The details of the interference corrections differ from method to
method, but since all such methods suffer from similar limitations, it is
reasonable to characterise them all with a single example. Methods
which ignore the BZ measurement and find the solution which minimises
the vector distance between the computed and expected values of the
magnetic field vector will satisfy (16), (17) and:

( B cosΘ − B$ cos Θ$ ) + ( B sin Θ − B$ sin Θ$ )
2

2

= minimum

…(22)

$ are the estimated values of total field strength and dip
where B$ and Θ
angle respectively. Solving these three equations for azimuth leads to
Psin Am + Q cos Am + Rsin Am cos Am = 0
where

…(23)

SPE 56702

ACCURACY PREDICTION FOR DIRECTIONAL MWD

(

)
Q = −( Bx cosτ − B y sinτ )

$ sin I cos I
P = Bx sin τ + B y cos τ cos I + B$ sin Θ

…(24)
…(25)

$ sin2 I
R = B$ cos Θ

…(26)

The sensitivities of the computed azimuth to errors in the sensor
measurements are found by differentiating (23).
Magnetic Field Uncertainty. The weighting function for magnetic
declination error is given above. Those for magnetic field strength and
dip angle, which are required when an axial magnetic interference
correction is in use, are derived by differentiating (23) with respect to
$.
B$ and Θ
Misalignment Errors. Brooks and Wilson3 model tool axial
misalignment as two uncorrelated errors corresponding to the x and y
axes of the tool. Their expressions for the associated inclination and
azimuth errors lead directly to the following weighting functions


0
∂p


=
sin τ

∂ε MX
 − cosτ / sin I 



0
∂p


=
cosτ 
∂εMY 
sin τ / sin I 

7

…(27,28)

Table 2 contains expressions for all the weighting functions not cited in
this section which are required to implement the error models described
in this paper.
Calculation Options
The method of position uncertainty calculation described here admits a
number of variations. It can still claim to be a standard, in that selection
of the same set of conventions should always yield the same results.
Along-Hole Depth Uncertainty. The propagation model described
above is appropriate for determining the position uncertainty of the
points in space at which the survey tool came (or will come) to rest.
These may be called uncertainties “at survey stations”.
Thorogood2 argues that it is more meaningful to compute the position
uncertainties of the points in the wellbore at the along-hole depths
assigned to the survey stations. These may be called the uncertainties
“at assigned depths”. This approach allows computation of the position
uncertainty of points (such as picks from a wireline log) whose depths
have been determined independently of the survey. Thorogood made
this calculation by defining a weighting function incorporating the local
build and turn rates of the well. The approach described in Appendix A
achieves the same result without the need for a new weighting function.
The results of the two approaches differ only in the along-hole
component of uncertainty. The along-hole uncertainty at a survey
station includes the uncertainty in the station’s measured depth, while
the uncertainty at an assigned depth does not.
The correct choice of approach depends on the engineering problem
being tackled - in many cases it is immaterial. The user of well-designed
directional software need not be aware of the issue.

Survey Bias. Not to be confused with sensor biases (which might
better be termed offset errors), survey bias is the tendency for the most
likely position of a well to differ from its surveyed position. The only
bias term defined by Wolff and deWardt was for magnetic interference in
“poor magnetic” surveys. The claims for stretch and thermal expansion
of drill-pipe to be treated as bias errors are at least as strong.
Some vendors of directional software have neglected to model survey
bias on the grounds that (a) such errors should be corrected for and (b)
engineers don’t like/understand them. The first objection can be
countered by the observation “yes, but they aren’t!”, the second by
careful software design.
The sign convention for position bias is from survey to most likely
position (ie. opposite to the direction of the error). Since drill pipe
generally elongates downhole, most likely depths are greater than survey
depths and bias values are positive. For axial drillstring interference,
most likely azimuths are greater than survey azimuths when the
weighting function, sinIsinAm is positive, so bias values are again
positive (at least in the northern hemisphere). The additional
mathematics required to model survey bias is included in Appendix A.
Calculation conventions. The calculation of position uncertainty
requires a wellbore survey consisting of discrete stations, each of which
has an associated along-hole depth, inclination, azimuth and toolface
angle. Clearly, these data will not be available in many cases, and certain
conventions are required whereby assumed values may be calculated.
The following are suggested.
Along-hole depth. For drilled wells, actual survey stations should
be used. For planned wells, the intended survey interval should be
determined, and stations should be interpolated at all whole multiples of
this depth within the survey interval. Typically, an interval of 100 feet
or 30 metres should be used.
Inclination and azimuth. For drilled wells, measured values should
be used. For planned wells, the profile should be interpolated at the
planned survey station depths using minimum curvature.
Toolface. If actual toolface angles are available, they should be used.
If not, several means of generating them are possible:
• Random number generation. Possibly close to reality, but results are
not repeatable and will tend to be optimistic.
• Worst-case. Several variations on this idea are possible, but each will
require some additional calculation. The principle is questionable, and
the computational overhead is probably not justified.
• Borehole toolface (ie. the up-down-left-right change in borehole
direction). This angle bears little relation to survey tool orientation, but
is at least well-defined, and may be computed directly from inclination
and azimuth data. This approach will tend to limit the randomisation of
toolface dependent errors, giving a conservative uncertainty prediction.
This is the convention used in the examples at the end of the paper.
Formulae for borehole toolface are given in Appendix B.

8

HUGH S. WILLIAMSON

Standard Profiles
At the 8th meeting of the ISCWSA participants were set the task of
designing a number of well profiles suitable for:
• testing software implementations of the error models and
propagation mathematics
• studying and highlighting the behaviour of different error models
(magnetic and gyroscopic) and individual error sources
• demonstrating to a non-specialist audience the uncertainties to be
expected from typical survey programs.
The ideas generated at the meeting were used to devise a set of three
profiles:
ISCWSA #1: an extended reach well in the North Sea
ISCWSA #2: a “fish-hook” well in the Gulf of Mexico, with a long
turn at low inclination
ISCWSA #3: a “designer” well in the Bass Strait, incorporating a
number of difficult hole directions and geometries.
Figs. 2, 3 illustrate the test profiles in plan and section. Their full
definition, given in Table 4, includes location, magnetic field, survey
stations, toolface angles and depth units.
Example Results
The error models for basic and interference-corrected MWD have been
applied to the standard well profiles to generate position uncertainties in
each well. The results of several combinations are tabulated in Table 5.
Examples 1 and 2 compare the basic and interference-corrected models
in well ISCWSA #1. Being a high inclination well running approximately
east-north-east, the interference correction actually degrades the
accuracy. The results are plotted in Fig 4. Examples 3 to 6 all represent
the basic MWD error model applied to well ISCWSA #2. They differ in
that each uses a different permutation of the survey station/assigned
depth and symmetric error/survey bias calculation options. The
variation of lateral uncertainty and ellipsoid semi-major axis,
characteristic of a “fish-hook” well, is shown in Fig 5. Finally, example
7 breaks well ISCWSA #3 into 3 depth intervals, with the basic and
interference-corrected models being applied alternately. This example is
included as a test of error term propagation.
Taken together, the examples form a demanding test set for
implementations of the method and models described in this paper.
Conclusions and Recommendations
This paper, and the collaborative work which it describes, establishes a
common starting point for wellbore position uncertainty modelling. The
standardised elements are:
• a nomenclature (see below)
• a definition of what constitutes an error model
• mathematics of position uncertainty calculation
• an error model for a basic directional MWD service
• a set of well profiles for investigating error models
• a set of results for testing software implementations
The future work which these standards were designed to facilitate

SPE 56702

includes:
• establishment of agreed error models for other survey services,
including in-field referencing and gyroscopic tools.
• interchangeability of calculated position uncertainties between
survey vendor, directional drilling company and operator.
Useful though this work is, it is only a piece in a larger jigsaw. Taking
a wider view, the collaborative efforts of the extended survey
community should now be directed towards:
• standardisation of quality assurance measures
• strengthening the link between quality assurance specifications
and error model parameters
• better integration of wellbore position uncertainty with the other
aspects of oilfield navigation..
Acknowledgments
The author thanks all participants in the ISCWSA for their enthusiasm
and support over several years and in the review of this paper.
Particular contributions to the MWD error model were made by John
Turvill and Graham McElhinney, both now with PathFinder Energy
Services, formerly Halliburton Drilling Systems; Wayne Phillips,
Schlumberger Anadrill; Paul Rodney and Anne Holmes, Sperry-Sun
Drilling Services; and Oddvar Lottsberg, formerly of Baker Hughes
INTEQ .
Participants in the Working Group on error propagation were David
Roper, Sysdrill Ltd; Andy Brooks and Harry Wilson, Baker Hughes
INTEQ ; and Roger Ekseth, formerly of Statoil.
The results in Table 5 were checked by Jerry Codling, Landmark.
The author also wishes to thank BP Amoco for their permission to
publish this paper.
Nomenclature
ISCWSA Nomenclature*
D
along-hole depth
I
wellbore inclination
A
wellbore azimuth
Am
wellbore magnetic azimuth
τ
toolface angle
N
north co-ordinate
E
east co-ordinate
V
true vertical depth
δ
magnetic declination
Θ
magnetic dip angle
B
magnetic field strength
G
gravity field strength
X, x, Y, y, Z, z
tool reference directions - see fig. 1.
* adopted by ISCWSA participants as a standard for all technical
correspondence.

SPE 56702

ACCURACY PREDICTION FOR DIRECTIONAL MWD

Special Nomenclature
b
component of wellbore position bias vector

B$

estimated magnetic field strength
wellbore position uncertainty covariance matrix
e
1 s.d. vector error at an intermediate station
e
1 s.d. vector error at the station of interest
E
sum of vector errors from slot to station of interest
ε
particular value of a survey error
H,L
used in calculation of toolface
m
bias vector error at an intermediate station
m
bias vector error at the station of interest
M
wellbore position bias vector
µ
mean of error value
σ
standard deviation of error value, component of wellbore
position uncertainty
p
survey measurement vector (D,I,A)
P,Q,R intermediate calculated quantities
r
wellbore position vector
∆rk
increment in wellbore position between stations k-1 and k
ρ
correlation coefficient
$
estimated magnetic dip angle
Θ
v
along-hole unit vector
w
factor relating error magnitude to uncertainty in measurement

C

subscripts and counters
hla
borehole referenced frame
i
a survey error term
k
a survey station
K
survey station of interest
Kl
number of stations in lth survey leg
l
a survey leg
L
survey leg containing the station of interest
nev
earth-referenced frame
superscripts
dep
at the along-hole depth assigned to the survey station
rand
random propagation mode
svy
at the point where the survey measurements were taken
syst
systematic propagation mode
well
per-well or global propagation mode
References
1. Wolff, C.J.M. and de Wardt, J.P., Borehole Position Uncertainty Analysis of Measuring Methods and Derivation of Systematic Error
Model, JPT pp.2339-2350, Dec. 1981
2. Thorogood, J.L., Instrument Performance Models and their
Application to Directional Survey Operations, SPEDE pp.294-298,
Dec. 1990.

9

3. Brooks, A.G. and Wilson, H., An Improved Method for Computing
Wellbore Position Uncertainty and its Application to Collision and
Target Intersection Probability Analysis, SPE 36863, EUROPEC,
Milan, 22-24 Oct 1996.
4. Dubrule, O., and Nelson, P.H., Evaluation of Directional Survey
Errors at Prudhoe Bay, SPE 15462, 1986 ATCE, New Orleans, Oct
5-8.
5. Minutes of the 7th Meeting of the ISCWSA, Houston, 9 Oct
1997.
6. Grindrod, S.J. and Wolff, J.M., Calculation of NMDC Length
Required for Various Latitudes Developed From Field
Measurements of Drill String Magnetisation, IADC/SPE 11382,
1983 Drilling Conference, Houston.
7. Minutes of the 6th Meeting of the ISCWSA, Vienna, 24-25 Jun
1997.
8. Minutes of the 8th Meeting of the ISCWSA, Trondheim, 19 Feb
1998.
9. Macmillan, S., Firth, M.D., Clarke, E., Clark, T.D.G. and
Barraclough, D.R., “Error estimates for geomagnetic field values
computed from the BGGM”, British Geological Survey Technical
report WM/93/28C, 1993.
10. Ekseth, R, Uncertainties in Connection with the
Determination of Wellbore Positions, ISBN 82-471-0218-8,
ISSN 0802-3271, PhD Thesis no. 1998:24, IPT report 1998:2, The
Norwegian University of Science and Technology, Trondheim,
Norway.
Appendix A
 Mathematical Description of Propagation
Model
The total position uncertainty at a survey station of interest, K (in
survey leg L) is the sum of the contribution from all the active error
sources. It is convenient computationally to group the error sources by
their propagation type and to sum them separately.
Vector errors at the station of interest. Recall that the vector error
due to the presence of error source i at station k is the sum of the effect
of the error on the preceding and following survey displacements:
 d ∆rk d ∆rk +1  ∂p k
ei , l , k = σi , l 
+

 dp k
d p k  ∂εi

…(A-1)

Evaluating this expression using the minimum curvature well
trajectory model is cumbersome. There is no significant loss of accuracy
in using the simpler balanced tangential model:
∆rj =

 sinI j−1 cos Aj−1 + sinI j cos Aj 
Dj − Dj−1 

 sinI j−1 sin Aj−1 + sinI j sin A j 
2


cosI j−1 + cos I j

…(A-2)

The two differentials in the parentheses in (A-1) may then be
expressed as:

10
d ∆r j
dp k

HUGH S. WILLIAMSON
 d ∆r j
=
 dD k
d ∆r j

and

dDk

d ∆r j
dI k

d ∆r j
dA k

d ∆r j

d∆ r j 

dI k 

dI k

where j = k , k+1

− sin I cos A − sin I cos A 
j −1
j −1
j
j

1
=  − sin I j− 1 sin A j −1 − sin I j sin Aj 
2

− cos I j −1 − cos I j

(
(

)
)

D −D

j
j− 1 cos I k cos Ak 
1 
=
D − D j −1 cos I k sin Ak 
2 j

 − D j − D j −1 sin I k 

(

(

)

)

…(A-5)

…(A-6)

For the purposes of computation, the error summation terminates at
the survey station of interest. Vector errors at this station are therefore
given by:
ei , L, K = σi , L

d∆ rK ∂p K
dpK ∂εi

Undefined weighting functions. For some combinations of weighting
function and hole direction, a component of the measurement vector
(usually azimuth) is highly sensitive to changes in hole direction and the
vector ∂p/∂εi is apparently undefined. There are two cases:
Vertical hole. In this case, dr/dp is zero and the vectors ei,l,k and
ei, L, K are still finite and well-defined. They may be computed by
forming the product (A-1) algebraically and evaluating it as a whole.
Take as an example the weighting function for an x-axis radially
symmetric misalignment. Substituting the expression for ∂p/∂εMX (27)
and the well trajectory model equations (A-3 to A-6) into (A-1) and (A7), and setting I equal to zero gives
…(A-8)

 sin ( A + τ) 
σi ,L ( DK − DK −1 ) 

 − cos( A + τ) 
2


0

Cirand
=
,l

Kl

∑ ( ei,l,k ) .( ei ,l ,k )

…(A-10)

T

k =1

L −1

K −1

l=1

k =1

rand
Crand
i , K = ∑ Ci, l + ∑ ( ei , L ,k ) .( ei , L ,k ) + ( ei, L , K ) . ( ei, L, K )
T

…(A-9)

There are similar expressions for Y-axis axial misalignment and X- and
Y-axis accelerometer biases. These are given in Table 3. Equivalent

T

…(A-11)
Systematic errors. The contribution to survey station uncertainty
from a systematically propagating error source i over survey leg l (not
containing the point-of-interest) is:
 Kl
  Kl


Cisyst
=
e
,l
 ∑ i ,l ,k  . ∑ ei ,l ,k 
 k =1
  k =1

T

…(A-12)

and the total contribution over all survey legs is:
T

 K −1
  K −1

=∑
+  ∑ ei , L , k + ei , L , K  . ∑ ei , L , k + ei , L, K 



 k =1
  k=1

l =1
…(A-13)
Per-Well and Global errors. Each of these error types is systematic
between all stations in a well. The individual vector errors can therefore
be summed to give a total vector error from slot to station
Cisyst
,K

L −1

Csyst
i ,l

L −1 K l
 K −1
Ei , K = ∑  ∑ ei ,l , k  + ∑ ei , L , k + ei , L , K
 k =1
l =1  k =1

and
ei, L ,K =

Summation of errors. Vector errors are summed into position
uncertainty matrices as follows.
Random errors. The contribution to survey station uncertainty from
a randomly propagating error source i over survey leg l (not containing
the station of interest) is:

and the total contribution over all survey legs is:

affects only the preceding survey displacement. In what follows we
reserve the notation ei,l,k for vector errors at intermediate stations, which
affect both the preceding and following displacements.

 sin( A + τ) 
σi ,l ( Dk +1 − Dk −1 ) 

− cos( A + τ )
2


0

Other hole directions. Some error sources really are unbounded in
certain hole directions. The examples in this paper are sensor errors after
axial interference correction in a horizontal and magnetic east/west
wellbore - a so-called “90/90” well. In such cases, the assumptions of
linearity break down, and computed position uncertainties are
meaningless. Software implementations should include an error-catching
mechanism for this case.

…(A-7)

The notation ei , L , K indicates that a measurement error at this station

e i,l ,k =

expressions may be used for evaluating bias vectors in vertical hole, with
m i , l ,k , m i , L , K , and µi,l substituted for e i , l ,k , ei , L , K , and σi,l
respectively.

…(A-4)

)

− D − D

j
j −1 sin I k sin Ak 
1 
=
D − Dj −1 sin I k cos Ak 
2 j

0



(

…(A-3)

SPE 56702

…(A-14)

The total contribution to the uncertainty at survey station K is
T
Ciwell
, K = E i, K . E i, K

…(A-15)

Total position covariance. The total position covariance at survey

SPE 56702

ACCURACY PREDICTION FOR DIRECTIONAL MWD

station K is the sum of the contributions from all the types of error
source:
=∑

C svy
K

i∈R

Crand
i, K

+∑
i∈S

Csyst
i, K

+
Cwell
i, K
i∈{ W ,G}

…(A-16)

where the superscript svy indicates the uncertainty is defined at a
survey station.
Survey bias. Error vectors due to bias errors are given by expressions
entirely analogous with (A-1) and (A-7):
 d ∆rk d ∆r k +1  ∂p k
m i ,l , k = µi ,l 
+

 dp k
d p k  ∂εi
d ∆r K ∂p K
m i , L ,K = µi , L
d p K ∂εi

…(A-17)
…(A-18)

The total survey position bias at survey station K, M svy
K , is the sum
of individual bias vectors taken over all error sources i, legs l and
stations k:
…(A-19)

Position uncertainty and bias at an assigned depth
Defining the superscript dep to indicate uncertainty at an assigned
depth, it may be shown that:
dep

svy

ei , L , K = ei, L , K −σ i , L wi , L , K v K
dep

…(A-20)

svy

ei ,l , k = ei ,l , k

mi , L, K = mi , L ,K − µi , L wi , L, K v K

dep

svy

…(A-22)

dep
mi ,l ,k

svy
mi ,l ,k

…(A-23)

Relative uncertainty between wells. When calculating the
uncertainty in the relative position between two survey stations
(KA,KB) in wells (A,B), we must take proper account of the correlation
between globally systematic errors. The uncertainty is given by:

[

Csvy rK A − rK B

[

M svy rK A − rKB

]

(

)(

svy
svy

= C K A + C KB − ∑  Ei , K A . Ei , KB
i ∈G 

) + ( E ).(E )
T




…(A-24)

T

i , KB

]

svy

svy

= M K A − M KB

…(A-25)

Substitution of equations (A-20) to (A-23) into these expressions
gives the equivalent results at the along-hole depths assigned to the
stations.
Transformation into Borehole Reference Frame. The results
derived above are in an Earth-referenced frame (North, East, Vertical subscript nev). The transformation of the covariance matrices and bias
vectors into the more intuitive borehole referenced frame (Highside,
Lateral, Along-hole - subscript hla) is straightforward:
C hla = T T C nev T

…(A-26)

b H 
 
T
 b L  = M hla = T M nev
 b A 

cos I K cos AK

T =  cos I K sin AK
 − sin I K

…(A-27)

i, K A

− sin AK
cos AK
0

sin I K cos AK 

sin I K sin AK 

cos I K

…(A-28)

is a transformation matrix. Uncertainties and correlations in the principal
borehole directions are obtained from:

[ ]

etc.

…(A-29)

[ ]

etc.

…(A-30)

σ H = Chla 1 , 1

…(A-21)

where wi,L,K is the factor relating error magnitude to measurement
uncertainty and vK is the along-hole unit vector at station K. Figs. 6, 7
illustrate these results. Substituting these expressions into (A-12 to A16) yields the position uncertainty at the along-hole depth assigned to
each survey station.
Survey bias at an assigned depth is calculated by substituting the
following error vectors into (A-19):

=

The relative survey bias is simply:

where

 L −1 Kl

 K −1
= ∑  ∑  ∑ m i ,l ,k  + ∑ m i , L ,k + m i , L , K 
 k =1
i  l=1  k =1

svy
MK

11

ρHA =

C hla 1 , 2
σHσL

Appendix B
 Calculation of Toolface Angle
The following formulae may be used to calculate a synthetic toolface
angle from successive surveys:
H K = sin I K cos I K −1 cos( A K − AK −1 ) − sin I K −1 cos I K

…(B-1)

LK = sin I K sin( A K − AK −1 )

…(B-2)

If HK > 0,

τK = tan-1(LK/HK)

…(B-3)

If HK < 0,

τK = tan (LK/HK) + 180°

…(B-4)

-1

If HK = 0, τ K = 270°, 0° or 90° as LK < 0, LK = 0 or LK > 0...(B-5)

12

HUGH S. WILLIAMSON

SPE 56702

Table 1—Summary of Basic MWD Error Models
Weighting
Function

Sensors
ABX
ABY
ABZ
ASX
ASY
ASZ
MBX
MBY
MBZ
MSX
MSY
MSZ
ABIX
ABIY
ABIZ
ASIX
ASIY
ASIZ
MBIX
MBIY
MSIX
MSIY

Basic
model

with axial
correction

Prop.
Mode

0.0004 g
0.0004 g
0.0004 g
0.0005
0.0005
0.0005
70 nT
70 nT
0.0016
0.0016

S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S

0.0004 g
0.0004 g
0.0004 g
0.0005
0.0005
0.0005
70 nT
70 nT
70 nT
0.0016
0.0016
0.0016

Weight.
Func.

Basic model

with axial
correction

Prop.
Mode

0.2°
0.06°
0.06°

S
S
S

Misalignment
SAG
0.2°
MX
0.06°
MY
0.06°
Axial magnetic interference
AZ
0.25°
AMID
0.6°
Declination
AZ
0.36°
DBH
5000°nT

S
S or B*

0.36°
5000°nT

G
G

Total magnetic field and dip angle
MDI
0.20°
MFI
130 nT
Along-hole depth
DREF
0.35 m
DSF
2.4 × 10-4
DST
2.2 × 10-7 m-1

0.35 m
2.4 × 10-4
2.2 × 10-7 m-1

G
G

R
S
G or B†

* when modelled as bias: µ = 0.33°, σ = 0.5°
† when modelled as bias: µ = 4.4 × 10-7 m -1, σ = 0

Table 2—Error Source Weighting Functions not Given in the Text
Sensor Errors (without axial interference correction)
ABX



0
1


cos
I
sin
τ

G
 cos I sin A sin τ − cos A cos τ tanΘ + cot I cosτ 
m
m

(

ABY



0
1

− cos I cos τ

G
 cos I sin A cos τ + cos A sin τ tanΘ − cot I sin τ
m
m

MBY

AB
Z

MSX



0

 MSY
0


− (cos A sin τ + cos I sin A cos τ ) / ( B cos Θ )
m
m




0
1


sin
I

G
 tanΘ sin I sin Am 

ASY

ASZ



0



sin
I
cos
I


tan Θ sin I cos I sin Am 

(

)

)



0


2
sin I cos I cos τ


− tan Θ sin I cos I sin A cosτ + cos A sin τ − cos I sin τ cosτ 
m
m

(

)



0


0


(cos A cos τ − cos I sin A sin τ ) / ( B cos Θ) 
m
m



0


2
sin
I
cos
I
sin
τ


− tanΘ sin I cos I sin A sin τ − cos A cosτ + cos I cosτ sin τ
m
m

(

)

(

MBX

ASX

(

)

)


0

0

 cos I cos A sin τ − tan Θ sin I sin τ + sin A cos τ cos A cos τ − cos I sin A sin τ
m
m
m
m

(

)(





)


0

0

− cos I cos A cosτ − tan Θ sin I cos τ − sin A sin τ cos A sin τ + cos I sin A cosτ
m
m
m
m

(

MBZ

)(



0


0


− sin I sin Am / ( B cos Θ) 

MSZ





)



0


0


 − (sin I cos Am + tan Θ cos I ) sin I sin Am 

SPE 56702

ACCURACY PREDICTION FOR DIRECTIONAL MWD

13

Table 2—Cont.
Sensor Errors (with axial interference correction)

0
1 

cos
I sin τ
G 
2
 cos I sin Am sin τ tan Θ cos I + sin I cos Am − cos τ tan Θ cos Am − cot I

0
1 
− cos I cos τ

G
2
 cos I sin Am cosτ tan Θ cos I + sin I cos Am + sin τ tan Θ cos Am − cot I

ABIX

ABIY




2
2
/ 1 − sin I sin Am 



2
2
/ 1 − sin I sin Am 

(

(

)

(

)) (

)

(

(

)

(

)) (

)



0


2
sin I cos I sin τ


− sin τ sin I cos2 I sin A sin τ tanΘ cos I + sin I cos A − cosτ tan Θ sin I cos A − cos I / 1 − sin 2 I sin 2 A 
m
m
m
m




0


2
sin
I
cos
I
cos
τ


− cos τ sin I cos 2 I sin A cos τ( tanΘ cos I + sin I cos A ) + sin τ( tanΘ sin I cos A − cos I ) / (1 − sin 2 I sin 2 A )
m
m
m
m 



0


0


− cos I cos A sin τ − tanΘ sin I sin τ + sin A cos τ cos I sin A sin τ − cos A cos τ / 1 − sin 2 I sin 2 A 
m
m
m
m
m 

ASIX

(

ASIY

(

)

)) (

(

(

MSIX

)

(

MSIY

MBIX

MBIY

)

)(

) (

)



0


0


2
2
− (cos I cos Am cos τ − tanΘ sin I cosτ − sin Am sin τ)(cos I sin Am cos τ + cos Am sin τ) / (1 − sin I sin Am ) 






0
0



1
ABIZ
0

sin
I




G
− ( cos I sin A sin τ − cos A cos τ) / B cos Θ(1 − sin 2 I sin 2 A ) 
2
2

m
m
m 

 sin I cos I sin Am ( tanΘ cos I + sin I cos Am ) / (1 − sin I sin Am ) 




0
0




ASIZ
0


− sin I cos I


− ( cos I sin A cos τ + cos A sin τ) / B cos Θ(1 − sin 2 I sin 2 A ) 
 sin I cos2 I sin A tan Θ cos I + sin I cos A / 1 − sin2 I sin2 A 
m
m
m 

m
m
m

(

)

(

(

)

(

)

)) (

(

Magnetic Field Errors (with axial interference correction)
MFI


0

0

− sin I sin A ( tan Θ cos I + sin I cos A ) / B(1 − sin 2 I sin 2 A )
m
m
m

(





MDI

)



0


0


− sin I sin Am( cosI − tan Θsin I cos Am ) / (1− sin2 I sin2 Am) 

Table 3—Error Vectors in Vertical Hole where Weighting Function is Singular
Sensor Errors (with or without axial interference correction)
ABX
or
ABIX

ei ,l, k =

(

)


(
)
σi ,l Dk+1 − Dk −1 − sin A + τ 
(
)
cos
A
+
τ


2G


0

ABY
or
ABIY

ei ,l , k =

(

)



σi ,l D k +1 − D k −1 − cos( A + τ ) 

sin
(
A
+
τ
)


2G


0

Misalignment Errors
MX

e i,l ,k =

 sin( A + τ) 
σi ,l ( Dk +1 − Dk −1 ) 

− cos( A + τ )
2


0

MY

ei ,l ,k =

cos( A + τ)
σi ,l ( Dk +1 − Dk − 1) 

sin ( A + τ) 
2


0

)

14

HUGH S. WILLIAMSON

SPE 56702

Table 4—Standard Well Profiles
ISCWSA #1 - North Sea Extended Reach Well
Lat. 60°N, Long. 2°E, Total Field, 50,000 nT, Dip 72°, Declination
4°W, Station interval 30m, VS Azimuth 75°

ISCWSA #2 - Gulf of Mexico Fish Hook Well
Lat. 28°N, Long. 90°W, Total Field 48,000nT, Dip 58°, Declination
2°E, Station interval 100ft, VS Azimuth 21°

MD
m
0.00
1200.00
2100.00
5100.00
5400.00
8000.00

MD
ft

Inc
deg
0.000
0.000
60.000
60.000
90.000
90.000

Azi
North
deg
m
0.000
0.00
0.000
0.00
75.000 111.22
75.000 783.65
75.000 857.80
75.000 1530.73

East
m
0.00
0.00
415.08
2924.62
3201.34
5712.75

TVD
m
0.00
1200.00
1944.29
3444.29
3521.06
3521.06

VS
m
0.00
0.00
429.72
3027.79
3314.27
5914.27

DLS
°/30m
0.00
0.00
2.00
0.00
3.00
0.00

Inc
deg
0.00 0.000
2000.00 0.000
3600.00 32.000
5000.00 32.000
5525.54 32.000
6051.08 32.000
6576.62 32.000
7102.16 32.000
9398.50 60.000
12500.00 60.000

Azi
North
East
TVD
VS
DLS
deg
ft
ft
ft
ft
°/100ft
0.000
0.00
0.00
0.00
0.00 0.00
0.000
0.00
0.00 2000.00
0.00 0.00
2.000 435.04
15.19 3518.11
411.59 2.00
2.000 1176.48
41.08 4705.37 1113.06 0.00
32.000 1435.37 120.23 5253.89 1383.12 3.00
62.000 1619.99 318.22 5602.41 1626.43 3.00
92.000 1680.89 582.00 6050.92 1777.82 3.00
122.000 1601.74 840.88 6499.44 1796.70 3.00
220.000 364.88 700.36 8265.27
591.63 3.00
220.000-1692.70 -1026.15 9816.02 -1948.01 0.00

ISCWSA #3 - Bass Strait Designer Well
Lat. 40°S, Long. 147°E, Total Field 61,000nT, Dip -70°, Declination 13°E, Station interval 30m, VS Azimuth 310°
MD
m

Inc
Azi
North
East
TVD
VS
DLS
deg
deg
m
m
m
m
°/30m
0.00 0.000 0.000
0.00
0.00
0.00
0.00 0.00
500.00 0.000 0.000
0.00
0.00 500.00
0.00 0.00
1100.00 50.000 0.000 245.60
0.00 1026.69
198.70 2.50
1700.00 50.000 0.000 705.23
0.00 1412.37
570.54 0.00
2450.00 0.000 0.000 1012.23
0.00 2070.73
818.91 2.00

MD
Inc
Azi
North
m
deg
deg
m
2850.00 0.000 0.000 1012.23
3030.00 90.000 283.0001038.01
3430.00 90.000283.000 1127.99
3730.00 110.000193.000 996.08
4030.00 110.000193.000 721.40

East
m
0.00
-111.65
-501.40
-727.87
-791.28

TVD
VS
DLS
m
m
°/30m
2470.73
818.91 0.00
2585.32
905.39 15.00
2585.32 1207.28 0.00
2520.00 1197.85 9.00
2417.40 1069.86 0.00

Table 5—Calculated Position Uncertainties (at 1 standard deviation)
• uncertainty at tie-line (MD=0) is zero • stations interpolated at whole multiples of station interval using minimum curvature
• instrument toolface = borehole toolface

uncertainties along
borehole axes
No.

Well

1
2
3
4
5
6

#1
#1
#2
#2
#2
#2

7

#3

Depth interval(s)

Model Option

0 m - 8000 m
0 m - 8000 m
0 ft - 12500 ft
0 ft - 12500 ft
0 ft - 12500 ft
0 ft - 12500 ft
(1)
0 m - 1380 m
(2) 1410 m - 3000 m
(3) 3030 m - 4030 m

basic
ax-int
basic
basic
basic
basic
basic
ax-int
basic

Key to error models:
Key to calculation options:

basic
ax-int
S, sym
S, bias
D, sym
D, bias

S, sym
S, sym
S, sym
D, sym
S, bias
D, bias
S, sym
S, sym
S, sym

σH

σL

σA

20.11 m
84.33 m 8.62 m
20.11 m 196.41 m 8.62 m
16.17 ft
29.66 ft 10.12 ft
16.17 ft
29.66 ft 9.16 ft
15.69 ft
27.41 ft 8.61 ft
15.69 ft
27.41 ft 8.50 ft

5.64 m

5.76 m

9.59 m

correlations between
borehole axes

ρHL

ρHA

ρLA

-0.015
-0.006
+0.032
+0.032
+0.052
+0.052

+0.676
+0.676
-0.609
-0.426
-0.602
-0.569

-0.003
+0.004
+0.060
+0.084
+0.157
+0.160

-0.186

-0.588

+0.297

survey bias along
borehole axes

bH

bL

-6.79 ft -12.41ft
-6.79 ft -12.41ft

bA

+11.70 ft
-4.76 ft

Basic MWD
Basic MWD with axial interference correction
Uncertainty at survey station, all errors symmetric (ie. no bias)
Uncertainty at survey station, selected errors modelled as biases (see table 1)
Uncertainty at assigned depth, all errors symmetric (ie. no bias)
Uncertainty at assigned depth, selected errors modelled as biases (see table 1)

ACCURACY PREDICTION FOR DIRECTIONAL MWD

15

Vertical Section

X-axis

-1000m

HighSide

τ

Y-axis

2000m

True Vertical Depth

SPE 56702

4000m

6000m

2000m

ISCWSA#3
VS Azi = 310o

ISCWSA#2

ISCWSA#1
VS Azi = 75 o

VS Azi = 21 o

4000m

τ

Fig.3  Vertical section plot of standard well profiles. Note
different section azimuths.

Z-axis

= toolface angle

(down hole)
Example 1 : Basic MWD
Example 2 : MWD with axial interference
correction

Fig.1  Definition of tool sensor axes and toolface angle
200
1500m
North

ISCWSA#3

6
1500m

160

1000m

“corrected” model
deteriorates rapidly
near “90/90”

“corrected” model is
marginally more
accurate at low
inclination

6000m

ISCWSA#1
1000m

-1000m

East

ISCWSA#2

1000m

1 s.d. Lateral Uncertainty (m)

4
5500m

120

80

2

1200

1500

1800

2100

-1000m

Fig.2  Plan view of standard well profiles

40

inset

0
2000

4000
6000
Measured Depth (m)

8000

Fig.4  Comparison of basic and interference corrected MWD
error models in well ISCWSA#1

16

HUGH S. WILLIAMSON

SPE 56702

Example 4 : Ellipsoid semi-major axis
Example 4 : Lateral uncertainty
Recorded (and calculated)
survey station position

40

True position where
tool came to rest

1 s.d. Uncertainty (ft)

True well position at depth
assigned tosurvey station

ellipsoid semi-major axis
reduces as well returns
below surface location

30

Calculated well path
True well path

lateral uncertainty and
ellipsoid semi-major
axis are equal while
azimuth is constant

20

10
at mid-turn, lateral
direction co-incides
with ellipsoid minor axis

0
0

4000

8000

12000

Measured Depth (ft)

Fig.5  Variation of lateral uncertainty and ellipsoid semimajor axis in a fish-hook well - ISCWSA#2

Recorded (and calculated)
survey station position
True position where
tool came to rest
True well position at depth
assigned to survey station
Calculated well path
True well path

depth error at
earlier station
=

(2) vector errors for last survey
station and last assigned depth
due to depth error at earlier station
must therefore be the same:
e dep
= esvy
k
k

σ i wi , k

(1) with no depth error at last station,
true positions at survey station and
at its assigned depth coincide

Fig.7  Vector errors at the last station (point of
interest) due to an along-hole depth error at a
previous station.

vector error at last
assigned depth due to
depth error at last station
dep
svy
= eK = eK − σi wi, K v K

vector error at last
station due to depth
error at last station
=

e Ksvy

depth error at
last station
=

σi wi, K

Fig.6  Vector errors at the last station (point of interest) due
to an along-hole depth error at the last station.