Surveying Uncertainty: Drilling Operations

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 30

Surveying Uncertainty

Drilling & Measurements | Training Centers | Drilling Operations


Revisions
Revision Date Description Changed by

Created by Elena McMorris


1.3 Oct 2012 Initial release Reviewed by Bertrand Cozon

1.4 Oct 2013 Minor corrections A. de Verteuil

Schlumberger Private
1.5 Nov 2013 Slides added and notes corrected for grammar A. de Verteuil
Added slide on error definitions and ellipse variation in fish-hook A. de Verteuil & Farooq Qureshi
1.6 Jan 2015 well

Rev 1.6 Jan 2015

2
Objectives
At the end of this presentation you should be able to:

 Name the sources of errors that affect wellbore positioning


 Name the error model used as the standard in the oil industry today

Schlumberger Private
 Identify the sources of errors that are not accounted for by the error model
 Explain the importance of the company standards in regards to uncertainty
estimation
 Explain what Ellipsoid of Uncertainty (EOU) is and its relationship to
probability analysis
 Be able to calculate the EOU dimensions using Drilling Office application

Rev 1.6 Jan 2015

3
Wellbore Positional Uncertainty

Uncertainty is a quantification of the accuracy

Uncertainty estimation is based on mathematic formulae within


the error model

Schlumberger Private
A number of error models have existed through the history of
wellbore surveying. To set the standard a group of companies
established “The Industry Steering Committee on Wellbore
Survey Accuracy” or ISCWSA

ISCWSA objective is “to produce and maintain standards for


the industry relating to wellbore survey accuracy”

Rev 1.6 Jan 2015

4
Survey Tool Errors and Propagation
• Random errors - change from station to station & propagate as the square root
of MD

example - magnetic disturbance small contribution to error ellipse

• Systematic errors - same at every station & errors propagate proportionally to

Schlumberger Private
MD

example - BHA sag large contribution to error ellipse

• Bias errors* - predictable and, in principle, correctible

example - drill-pipe stretch shifts error ellipse away from survey position

*Note SAG is not a Bias but stretch is


Rev 1.6 Jan 2015

5
ISCWSA Error Model – Modeled Errors

Standard accuracy is calculated taking the


following main types of errors into account:

• Sensor errors

Schlumberger Private
Sensor misalignment
• SAG
• Drillstring Magnetic Interference
• Reference errors
• Depth Errors

All errors are independent of each other


and when combined are assumed to
estimate the wellbore position uncertainty

Rev 1.6 Jan 2015

6
ISCWSA Error Model – Not Estimated Errors

Assumptions within ISCWSA error model are:

• Survey does not contain gross errors

Schlumberger Private
Errors due to typo mistakes, incorrect inputs,
incorrect calibration

• The depth tracking is correct


• Surveys are taken not more than 100ft apart
• There is no external sources of interference
• Tool measurements are within specifications
• Drillstring interference error is within 0.5ᵒ
• Magnetic reference used for surveys is
within tolerances
Rev 1.6 Jan 2015

7
ISCWSA Error Model – Not Estimated Errors

Assumptions within ISCWSA error model are:

• Survey does not contain gross errors

Schlumberger Private
• The depth tracking is correct

Inaccurate length measurements or calibration,


forgotten stands

• Surveys are taken not more than 100ft apart


• There is no external sources of interference
• Tool measurements are within specifications
• Drillstring interference error is within 0.5ᵒ
• Magnetic reference used for surveys is
within tolerances
Rev 1.6 Jan 2015

8
ISCWSA Error Model – Not Estimated Errors
70 Assumptions within ISCWSA error model are:

• Survey does not contain gross errors


• The depth tracking is correct
60 • Surveys are taken no more than 100ft apart

Schlumberger Private
Inclination

Minimum curvature assumes the well is a


perfect arc. It is not always the case.
50
• There is no external sources of interference
• Tool measurements are within specifications
• Drillstring interference error is within 0.5ᵒ
• Magnetic reference used for surveys is
40 within tolerances
Measured depth
Static survey
High Definition Surveys

Rev 1.6 Jan 2015

9
ISCWSA Error Model – Not Estimated Errors
Assumptions within ISCWSA error model are:

• Survey does not contain gross errors


• The depth tracking is correct
• Surveys are taken not more than 100ft apart

Schlumberger Private
• There is no external source of magnetic
interference

Casing interference, fish left in offset wells,


solar storms, magnetized formation

• Tool measurements are within specifications


• Drillstring interference error is within 0.5ᵒ
• Magnetic reference used for surveys is
within tolerances

Rev 1.6 Jan 2015

10
ISCWSA Error Model – Not Estimated Errors
Assumptions within ISCWSA error model are:

• Survey does not contain gross errors


• The depth tracking is correct
• Surveys are taken not more than 100ft apart
• There is no external sources of interference

Schlumberger Private
• Tool measurements are within specifications

Calibration, rotational shots, roll tests,


checkshots

• Drillstring interference error is within 0.5ᵒ


• Magnetic reference used for surveys is
within tolerances

(above) SlimPulse ORM, D&I specification


Rev 1.6 Jan 2015

11
ISCWSA Error Model – Not Estimated Errors
Assumptions within ISCWSA error model are:

• Survey does not contain gross errors


• The depth tracking is correct
• Surveys are taken not more than 100ft apart

Schlumberger Private
• There is no external sources of interference
• Tool measurements are within specifications
• Drillstring interference error is within 0.5ᵒ and
Interfering Field criteria is acceptable

Azimuth error and FAC are controlled by


standards and procedures (EDI)

• Magnetic reference used for surveys is


within tolerances

Rev 1.6 Jan 2015

12
ISCWSA Error Model – Not Estimated Errors
Total Field Intensity Anomaly, nT

Assumptions within ISCWSA error model are:

• Survey does not contain gross errors


• The depth tracking is correct

Schlumberger Private
• Surveys are taken not more than 100ft apart
• There is no external sources of interference
Northing

• Tool measurements are within specifications


• Drillstring interference error is within 0.5ᵒ
• Magnetic reference used for surveys is
within tolerances
The biggest contributor in azimuth error (apart
from drillstring interference) is magnetic
declination. Error model accounts only for
some errors
Easting
Rev 1.6 Jan 2015

13
ISCWSA Error Model

Not modeled error/not included


Modeled errors • Gross errors
• Sensor errors • Depth tracking errors

Schlumberger Private
• Sensor misalignment • Survey frequency
• SAG • Consideration for poor procedures
• Drillstring Magnetic Interference • Excessive error due to BHA mag
• Reference errors interference
• Depth Errors • Excessive error due to local
magnetic crustal anomalies

Rev 1.6 Jan 2015

14
Error Distribution in One Dimension

Assumptions:

Errors are independent from each other

Schlumberger Private
Errors are distributed normally

Normal distribution is characterized by


Mean  and Standard Deviation 
1std.dev Mean 1std.dev  ( x   )2
1 2 2
Confidence Interval f ( x)  e
 2
(above)The probability density function f(x)
Rev 1.6 Jan 2015

15
Example of 1 Dimensional Error Distribution

Short Average Tall


Confidence level and confidence
intervals for 1 measurement

Schlumberger Private
68.3% - 1 sigma
95.4% - 2 sigma
99.7% - 3 sigma

Average American is 5ft 10in tall


0.14% 2.1% 13.6% 68% 13.6% 2.1% 0.14%
Approximately 68% of them are within 3”
Heigh

5’10”

5’11”
5’3”

5’4”

5’5”
5’6”

5’7”
5’8”

5’9”

6’1”

6’2”

6’3”

6’4”
6’
t

1 3 4 6 7 12 17 17 12 7 6 4 3 1
from average: 5ft 10in +/- 3”
unt
Co

3” in this case is a standard deviation


(above)The probability density function/Height distribution of American men
Rev 1.6 Jan 2015

16
Example of 2 Dimensional Error Distribution
Height Distribution curve
5’10”

Schlumberger Private
0.14% 2.1% 13.6% 68% 13.6% 2.1% 0.14%

39.9% (2D) 1 sigma

Weight Distribution curve


89.9% (2D) 2 sigma

13.6%
69kg

68%

85kg
13.6%
101kg

Rev 1.6 Jan 2015

17
For wellbore positioning 3 Dimensional
Uncertainty is required
 The size of the error ellipsoid is dependent upon the confidence level
and number of dimensions considered.
 Our standard is 3D, 95% confidence or 2.79 sigma

Schlumberger Private
Rev 1.6 Jan 2015

18
Error Distribution in 1, 2 or 3 Dimensions
Dimensionality Confidence Standard
Level Deviations
1D 68% 1 sigma
1D 95% 1.96 sigma 1D or 2D may be used for Driller’s
target sizing but 3D must be used for

Schlumberger Private
1D 99% 2.58 sigma
definitive surveys and AC analysis.
2D 68% 1.51 sigma
2D 95% 2.44 sigma
2D 99% 3.03 sigma
3D 68% 1.88 sigma
3D 95% 2.79 sigma
3D 99% 3.37 sigma

Rev 1.6 Jan 2015

19
Errors at aofSurvey
Ellipsoid Station
Uncertainty (EOU)

A xis
jo r
Ma

Schlumberger Private
Vertical Survey Vector

EOU – 3 axis Ellipsoid of


Uncertainty
Rev 1.6 Jan 2015

20
EOU’s –Major, Minor and Vertical axes

Major

Vertical

Schlumberger Private
Minor

 Inclination Uncertainty
 Azimuth Uncertainty
 Vertical Uncertainty

Rev 1.6 Jan 2015

21
Survey Tool Errors Example

Schlumberger Private
GMAG MWD-STD
123.19ft 532.68ft
DMAG

244.05ft

HOW BIG IS THE UNCERTAINTY? 123ft? 244 ft? 533 ft?


Rev 1.6 Jan 2015

22
Ellipsoid of Uncertainty (3-D)

Semi-Major
(driven by azimuth errors)

The uncertainty at each survey

Schlumberger Private
station can be represented as an
ellipsoid with three axes defined
as Semi-Major, Semi-Minor and
Vertical . This reference frame is
also called NEV (North, East,
Vertical)
Vertical uncertainty
Semi-Minor (driven by TVD errors) There is another representation
(driven by inclination errors) called HLA (High side, Lateral,
Along hole)

Rev 1.6 Jan 2015

23
EOU Propagation

Uncertainty at the current survey station consist EOU = Errorssurvey1


of error related to the current survey and all

Schlumberger Private
errors collected before. This can include
uncertainty of the well and uncertainty of the rig
and slot location.
EOU = Errorssurvey1 + Errorssurvey2

EOU = Errorssurvey1 +Errorssurvey2 + Errorssurvey3


Wellbore
Rev 1.6 Jan 2015

24
Surveyed Position Uncertainty
Tie-on point or WRP WRP (Well Reference Point)
defined as last known point of
departure (EOU =zero)
normally taken at ground entry
or seabed.
Survey Points

Schlumberger Private
Ellipsoids of Uncertainty
Based on 3D Dimensionality, 95%
confidence, 2.79 sigma

Cone of Uncertainty

Rev 1.6 Jan 2015

25
Error Ellipse Variation in a Fish-Hook Well
At constant azimuth,
lateral error increases When well turns,
most rapidly ellipse rotation
lags behind

In vertical hole,
ellipses
are circular

Schlumberger Private
As well passes back
under itself, ellipse shrinks
(systematic errors cancel out)

As step-out increases again,


ellipse expands rapidly PLAN VIEW
Rev 1.6 Jan 2015

26
Uncertainty Application

Uncertainty affects many decisions:

- Where do we place the well?

Schlumberger Private
- Which survey instruments do we use?
- Do we require a correction technique?
- How frequently do we need to survey?
- When do we stop?

Rev 1.6 Jan 2015

27
Are you be able to …?

 Name the sources of errors that affect wellbore position estimations


 Name the error model used as a standard in the oil industry today
 Identify the sources of errors that are not accounted for by the error model

Schlumberger Private
 Explain the importance of the company standards in regards to uncertainty
estimation
 Explain what Ellipsoid of Uncertainty (EOU) is and its relationship to
probability analysis
 Be able to calculate the EOU dimensions using Drilling Office application

Rev 1.6 Jan 2015

28
Questions?

Schlumberger Private
Rev 1.6 Jan 2015

29
Schlumberger Private
Rev 1.6 Jan 2015

30

You might also like