Automatic analysis of drill string dynamics

Patent metadata

Field	Value
Publication number	US-11506044-B2
Application number	US-202016936878-A
Country	US
Kind code	B2
Filing date	Jul 23, 2020
Priority date	Jul 23, 2020
Publication date	Nov 22, 2022
Grant date	Nov 22, 2022

Abstract

Disclosed are methods, systems, and computer-readable medium to perform operations including: capturing, using an image sensor directed at a drilling system, an image feed of movement of a component of the drilling system with respect to a vertical reference line; converting the image feed into a digital representation of the movement of the component, the digital representation including a number of offset pixels from the component to the vertical reference line in the image feed; converting the digital representation into a machine learning (ML), the ML representation including a plurality of vectors each including the number of offset pixels from the component to the vertical reference line at a respective time; training a ML model using the ML representation to characterize the movement of the component as normal or abnormal; and using the trained ML model to characterize the movement of the component in real-time as normal or abnormal.

First claim

What is claimed is:

1. A method comprising:
capturing, using an image sensor directed at a drilling system, an image feed of movement of a component of the drilling system with respect to a vertical reference line;
converting the image feed into a digital representation of the movement of the component, the digital representation comprising a number of offset pixels from the component to the vertical reference line in the image feed;
converting the digital representation into a machine learning (ML) representation, the ML representation including a plurality of vectors each comprising the number of offset pixels from the component to the vertical reference line at a respective time, wherein the number of offset pixels represents a distance of misalignment from the component and the vertical reference line;
training a ML model using the ML representation to characterize the movement of the component as normal or abnormal; and
using the trained ML model to characterize the movement of the component in real-time as normal or abnormal,
wherein the image sensor comprises a network of image sensors arranged about the drilling system at respective angles, and wherein the image feed comprises a plurality of image feeds capturing the movement of the component from a plurality of angles.

2. The method of claim 1 , wherein each vector of the ML representation comprises an indication of offset pixels from the component to the vertical reference line in the plurality of image feeds at the respective time.

3. The method of claim 1 , wherein the component is a drill string, and wherein capturing, using the image sensor, the image feed of the movement of the component comprises:
capturing the movement of the drill string with respect to the vertical reference line at a plurality of predetermined survey points along the vertical reference line.

4. The method of claim 3 , wherein characterizing the movement of the component in real-time as abnormal comprises:
characterizing the movement as stick-slip, axial vibration, tangential vibration, or lateral vibration.

5. The method of claim 1 , wherein the component is a drilling rig, and wherein characterizing the movement of the component in real-time as abnormal comprises:
characterizing the movement as a rig misalignment or a derrick misalignment.

6. The method of claim 1 , further comprising:
training the ML model using the image feed to characterize drilling operations, lithology changes, and kick off points in real-time.

7. The method of claim 1 , wherein the ML representation is further based on drilling surface parameters of the drilling system.

8. The method of claim 1 , further comprising:
in response to characterizing movement of the component in real-time as abnormal, performing a remedial action.

9. The method of claim 8 , wherein the remedial action comprises displaying a representation of the abnormal movement on a display device.

10. The method of claim 8 , wherein the remedial action comprises triggering an alarm.

11. The method of claim 8 , wherein the remedial action comprises controlling the drilling system to adjust a drilling parameter.

12. A non-transitory computer-readable medium storing one or more instructions executable by a computer system to perform operations comprising:
capturing, using an image sensor directed at a drilling system, an image feed of movement of a component of the drilling system with respect to a vertical reference line;
converting the image feed into a digital representation of the movement of the component, the digital representation comprising a number of offset pixels from the component to the vertical reference line in the image feed;
converting the digital representation into a machine learning (ML) representation, the ML representation including a plurality of vectors each comprising the number of offset pixels from the component to the vertical reference line at a respective time, wherein the number of offset pixels represents a distance of misalignment from the component and the vertical reference line;
training a ML model using the ML representation to characterize the movement of the component as normal or abnormal; and
using the trained ML model to characterize the movement of the component in real-time as normal or abnormal,
wherein the image sensor comprises a network of image sensors arranged about the drilling system at respective angles, and wherein the image feed comprises a plurality of image feeds capturing the movement of the component from a plurality of angles.

13. The non-transitory computer-readable medium of claim 12 , wherein each vector of the ML representation comprises an indication of offset pixels from the component to the vertical reference line in the plurality of image feeds at the respective time.

14. The non-transitory computer-readable medium of claim 12 , wherein the component is a drill string, and wherein capturing, using the image sensor, the image feed of the movement of the component comprises:
capturing the movement of the drill string with respect to the vertical reference line at a plurality of predetermined survey points along the vertical reference line.

15. The non-transitory computer-readable medium of claim 14 , wherein characterizing the movement of the component in real-time as abnormal comprises:
characterizing the movement as stick-slip, axial vibration, tangential vibration, or lateral vibration.

16. The non-transitory computer-readable medium of claim 12 , wherein the component is a drilling rig, and wherein characterizing the movement of the component in real-time as abnormal comprises:
characterizing the movement as a rig misalignment or a derrick misalignment.

17. A system, comprising:
one or more processors; and
a non-transitory computer-readable storage medium coupled to the one or more processors and storing programming instructions for execution by the one or more processors, the programming instructions instructing the one or more processors to perform operations comprising:
capturing, using an image sensor directed at a drilling system, an image feed of movement of a component of the drilling system with respect to a vertical reference line;
converting the image feed into a digital representation of the movement of the component, the digital representation comprising a number of offset pixels from the component to the vertical reference line in the image feed;
converting the digital representation into a machine learning (ML) representation, the ML representation including a plurality of vectors each comprising the number of offset pixels from the component to the vertical reference line at a respective time, wherein the number of offset pixels represents a distance of misalignment from the component and the vertical reference line;
training a ML model using the ML representation to characterize the movement of the component as normal or abnormal; and
using the trained ML model to characterize the movement of the component in real-time as normal or abnormal,
wherein the image sensor comprises a network of image sensors arranged about the drilling system at respective angles, and wherein the image feed comprises a plurality of image feeds capturing the movement of the component from a plurality of angles.

Assignees

• Saudi Arabian Oil Co

Inventors

• Mora Arturo Magana
• Affleck Michael
• Gooneratne Chinthaka Pasan
• Contreras Otalvora William
• Machocki Krzysztof Karol

Classifications

• G06N20/00 Primary

  Machine learning · CPC title

• E21B44/08

  in response to the amplitude of the movement of the percussion tool, e.g. jump or recoil · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• E21B2200/22

  Fuzzy logic, artificial intelligence, neural networks or the like · CPC title

• E21B47/002 Primary

  by visual inspection · CPC title