Improving Wellbore Position with High-Resolution Data

45th General Meeting

March 17th, 2017

The Hague, The NetherlandsWellbore Positioning Technical Section

Improving Wellbore Position

with High-Resolution DataMarc Willerth, Shawn DeVerse,

Stefan Maus, & Andrew Brooks

1


March 17th, 2017


• Marc Willerth

• VP of Survey Technologies

• March 17, 2017

• Magnetic Variation Services LLC

• Specializes in • Wellbore Positioning• Survey Corrections• Uncertainty Models

Speaker Information

2

• MagVAR / Surcon

• Services:• Magnetic Modelling

• High accuracy local magnetic models (IFR1)• Real-time local magnetic observatories (IFR2)• High Resolution Global Magnetic Model (MVHD)

• Survey Management• Real-time survey corrections• Survey quality monitoring• Fit-for-purpose uncertainty modelling

Company / Affiliation Information


March 17th, 2017

The Hague, The Netherlands Wellbore Positioning Technical Section

High Level View: Errors of Process

• Relate to the manner by which the survey is conducted

• Independent of sensor accuracy or performance

• Associated with known compromises in operations

• Comprise the largest “acceptable” error sources

4


March 17th, 2017


Examples of Compromises

• BHA Design• Drillstring Interference, Sag

• Magnetic Reference • Crustal Anomalies, Solar Weather

• Survey Frequency• “Aliasing” in minimum curvature

5


March 17th, 2017


6

Wellbore trajectories assume circular arcs between

stations

Data outside of survey set can weaken this assumption

Simple solution: More frequent surveys (<30ft)

More often a compromise is made (95ft+)

Positional Error From Survey Frequency


March 17th, 2017


This Problem is Not New

• High-frequency gyro surveying since the late 70s

• Continuous MWD surveying since late 80s

• Formal problem statement: Stockhausen & Lesso 2003

7


March 17th, 2017


Visualizing the Problem

8

Survey 1 Survey 2

RotateSlide

Positional Error

True Position

Reported Position


March 17th, 2017


Balanced Slides

9

Survey 1 Survey 2

Positional Error

Set slide location to minimize error

Stockhausen & Lesso, 2003


March 17th, 2017


Simulated Survey Points

10

Survey 1 Survey 2

Positional Error

New “survey” based on Slide Sheet

Stockhausen & Lesso, 2003


March 17th, 2017


Additional Surveys – Near Bit Inc Tool

11

Survey 1 Survey 2

Positional Error

Survey 3 – From a second sensor

Lawrence, Mojsin, & Strachan (2010)


March 17th, 2017


Additional Surveys – Continuous Inclination

12

Survey 1 Survey 2

Positional Error

Many additional measurements taken while drilling

Berger & Sele (1998),

Monterrosa, Rego, Zegarra, & Lowdon

(2016),

Countless others…


March 17th, 2017


What’s the Hold-up?• Changing drilling practices is often a non-starter

• Altering slide placement

• Reduce motor yield / increase slide ratio

• Surveying more frequently

• Reluctance to add “extra” surveys• Particularly if they are artificial points

• Concerns about measurement accuracy & handling large data sets

• These issues are amplified for a “Factory Drilling” environment!

13


March 17th, 2017


What Has Worked Before?

• Other process errors are being successfully managed

• Drillstring interference Multi-station analysis

• BHA Sag Sag Corrections

• Adjustment to existing surveys is an acceptable compromise

14


March 17th, 2017


Can This Really Be Corrected For?

• Current corrections improve the measurement accuracy

• This is not a measurement issue, orientation is accurate

• But orientation is not the primary product of a wellbore survey

• What if we correct thinking about position instead?

15


March 17th, 2017


Revisiting Our Problem

16

Survey 1 Survey 2

RotateSlide Corrected Survey

TVD Error


March 17th, 2017


Revisiting Our Problem

17

Survey 1 Survey 2

RotateSlide Corrected Survey

Optimizing Position, not Orientation!

True Orientation

True Position


March 17th, 2017


Implementation

• Utilize desired data sources to determine optimum position

• Could be continuous survey, slide sheet, or other data

• Optimize a minimum curvature survey to get there

• Apply the appropriate correction to the measured survey

18


March 17th, 2017


85

86

87

88

89

90

91

92

93

94

95

11950 12150 12350 12550 12750 12950 13150 13350 13550 13750 13950

Incl

inat

ion

(deg

)

Measured Depth (ft)

Continuous Inclination

Stationary Surveys

Example Correction with Continuous Data

19

Accumulated area between

curves correlates with TVD

error


March 17th, 2017


85

86

87

88

89

90

91

92

93

94

95

11950 12150 12350 12550 12750 12950 13150 13350 13550 13750 13950

Incl

inat

ion

(deg

)

Measured Depth (ft)


Stationary Surveys

Corrected Surveys


20

Apply an appropriate offset at each

station, Similar to a sag correction


March 17th, 2017


85

86

87

88

89

90

91

92

93

94

95

11950 12150 12350 12550 12750 12950 13150 13350 13550 13750 13950

Incl

inat

ion

(deg

)

Measured Depth (ft)


Corrected Surveys


21

Differences are now

balanced

above and below the

surveys.

Error will not accumulate


March 17th, 2017


Challenge: Optimizing for Real-Time

22

Real-time corrections can create

dependencies for future solutions

RotateSlide

Corrected Survey


March 17th, 2017



23

Real-time corrections can create

dependencies for future solutions

RotateSlide

Corrected Survey


March 17th, 2017



24

RotateSlide

Corrected Surveys

Naïve application of corrections can induce

positional error and false tortuousity


March 17th, 2017



25

RotateSlide

Corrected Surveys

Naïve application of corrections can induce

positional error and false tortuousity


March 17th, 2017


Optimizing for Real-Time

• Viable real-time solutions have added challenges

• Minimal “future” data -- Cannot strictly solve for position

• Ideally avoid introducing future errors into position

• Solution: Balance positional accuracy & correction stability

26


March 17th, 2017


Verifying the Correction Quality

• Derive expression for TVD error from Continuous Survey:

𝛥𝑇𝑉𝐷 =1

2

𝑖=2

𝑛−1

(𝑀𝐷𝑖−1−𝑀𝐷𝑖+1) sin 𝐼𝑖 (𝐼𝑖 − 𝐼𝑖∗)

• Similar to Appendix A in SPE 67616

• Compare accumulated error between survey sets

27


March 17th, 2017


-30

-25

-20

-15

-10

-5

0

5

8000 9000 10000 11000 12000 13000 14000 15000 16000 17000 18000

Acc

um

ula

ted

TV

D E

rro

r (f

t)

Measured Depth (ft)

Corrected Surveys

Raw Surveys

Error Accumulation Across a 10,000ft Lateral

28

Raw surveys

show errors of

~2.5 ft / 1000ft

Corrected surveys

are nearly

equivalent to a

continuous inc.

derived survey


March 17th, 2017


Correction Caveats

• Corrected surveys are less accurate for orientation

• Should not be used for motor yield, projection to bit, etc

• Data that feeds the correction is subject to its own QC

• Corrected surveys are only as accurate as the source data

29


March 17th, 2017


30

Conclusion• Drilling & surveying priorities conflict on survey frequency

• Prior solutions have idiosyncrasies preventing their adoption

• Equivalent accuracy achievable via corrections to existing stations

• Remove the error in a method acceptable to current workflows


March 17th, 2017


Questions?

31

Improving Wellbore Position with High-Resolution Data

Documents

Transcript of Improving Wellbore Position with High-Resolution Data