Method of training neural network models and using same for drilling wellbores

• US 8,417,495 B2
• Filed: 11/06/2008
• Issued: 04/09/2013
• Est. Priority Date: 11/07/2007
• Status: Active Grant

• Current Assignee: Baker Hughes Incorporated (Baker Hughes Company)
• Original Assignee: Baker Hughes Incorporated (Baker Hughes Company)
• Inventors: Dashevskiy, Dmitriy
• Primary Examiner(s): Rivas, Omar Fernandez
• Assistant Examiner(s): Saxena, Akash
• Application Number:
  US12/265,879
• Publication Number:
  US 20090114445A1
• Time in Patent Office: 1,615 Days
• Field of Search: 703/10, 703/7, 702/9
• US Class Current: 703/7
• CPC Class Codes: E21B 2200/22 Fuzzy logic, artificial intelligence, neural networks or the like
• CPC Class Codes: E21B 44/00 Automatic control systems specially adapted for drilling operations, i.e. self-operating systems which function to carry out or modify a drilling operation without intervention of a human operator, e.g. computer-controlled drilling systems; Systems specially adapted for monitoring a plurality of drilling variables or conditions
• CPC Class Codes: G01V 11/002 Details, e.g. power supply systems for logging instruments, transmitting or recording data, specially adapted for well logging, also if the prospecting method is irrelevant means for transmitting well survey signals E21B47/12; signal transmission systems in general G08C; transmission in general H04B

Abstract

A method of creating and using a neural network model for wellbore operations is disclosed. The method, in one aspect, may include defining a plurality of a wellbore parameter; calculating a plurality of output values of a tool operating parameter using the plurality of values of the wellbore parameter as input to a preexisting model; and obtaining a neural network model by using the plurality of values of the wellbore parameter and the calculated plurality of output values of the tool operating parameter. The neural network may be utilized for any suitable wellbore operation, including in conjunction with a drilling assembly for drilling a wellbore.

Claims

1. A method of operating a tool in a wellbore, comprising:

defining at least one value of at least one tool parameter,

defining a plurality of values of at least one wellbore parameter,

using the at least one value of the at least one tool parameter and the plurality of values of the at least one wellbore parameter as inputs to a selected model to estimate a bending moment of the tool;

obtaining an expected value of the bending moment by inputting into the neural network model the defined plurality of values of the at least one wellbore parameter and the estimated bending moment of the tool; and

comparing the expected value of bending moment of the tool obtained from the neural network model to the estimated bending moment of the tool obtained from the selected model to train the neutral network;

using the obtained neural network model to operate the tool in the wellbore.

2. The method of claim 1, wherein the at least one tool parameter is selected from a group consisting of:

(i) a length;

(ii) a weight;

(iii) a weight distribution;

(iv) a diameter;

(v) an outside dimension profile of the tool; and

(vi) a material.

3. The method of claim 1, wherein the at least one wellbore parameter is selected from a group consisting of:

(i) inclination;

(ii) wellbore geometry;

(iii) azimuth;

(iv) wellbore inside dimension; and

(v) curvature.

4. The method of claim 1, further comprising determining as an output value at least one tool operating parameter selected from a group consisting of:

(i) stress;

(ii) vibration;

(iii) whirl;

(iv) stick-slip;

(v) oscillation;

(vi) deflection;

(vii) tilt;

(viii) natural frequency;

(ix) tool fatigue; and

(x) tool health.

5. The method of claim 4 further comprising:

estimating the output value of the at least one tool operating parameter when the tool is in use in the wellbore (“

in-situ value”

) using as input a value of the wellbore parameter into the neural network; and

altering a tool operation in response to the in-situ value when the tool is in the wellbore.

6. The method of claim 1, wherein obtaining the neural network model further comprises using as input a value of at least one dynamic drilling parameter.

7. The method of claim 6, wherein the at least one dynamic drilling parameter is selected from a group consisting of:

(i) a side force applied by the tool to the wellbore;

(ii) direction of a side force applied by the tool to the wellbore;

(iii) stiffness of the tool;

(iv) weight-on-bit;

(v) rotational speed of a drill bit; and

(vi) rotational speed of the tool.

8. The method of claim 1 further comprising:

obtaining a value of the bending moment using the neural network model during drilling of the wellbore; and

altering a tool operation during drilling of the wellbore when the obtained value of the bending moment using the neural network is outside a selected criterion.

9. The method of claim 8, wherein altering the tool operation comprises altering at least one:

(i) drilling direction;

(ii) force applied on an inner wall of the wellbore by a force application device;

(iii) weight-on-bit;

(iv) rotational speed of a drilling bit;

(v) rotational speed of a drill string;

(vi) drilling motor rotational speed that drives a drill bit;

(vii) inclination;

(viii) azimuth; and

(ix) orientation.

10. The method of claim 1 further comprising:

obtaining a value of the bending moment when the tool is in use in the wellbore (“

in-situ value”

) using as input a value of a wellbore parameter into the neural network;

obtaining an actual value of the bending moment using a sensor downhole; and

updating the neural network model using the obtained actual value of the bending moment.

11. The method of claim 1, wherein the at least one value of the at least one tool parameter is fixed and the plurality of values of the at least one wellbore parameter falls within a selected range of values.

12. The method of claim 1 further comprising using at least one value of at least one dynamic drilling parameter as an input to the selected model to obtain the bending moment.

13. An apparatus for use in a well bore, comprising:

a drilling assembly configured to drill the wellbore;

a data storage medium;

a neural network model stored in the data storage medium, which neural network model is created by;

obtaining a bending moment of the drilling assembly by using as input to a selected model at least one value of at least one tool parameter and a plurality of values of at least one well bore parameter, and

obtaining an expected value of the bending moment of the drilling assembly by inputting into the neural network mode! the plurality of values of the at least one well bore parameter and the estimated bending moment of the drilling assembly, and

comparing the expected value of the bending moment of the drilling assembly obtained from the neural network model to bending moment of the drilling assembly obtained from the selected model to train the neural network; and

a processor configured to use the trained neural network to operate the drilling assembly in the wellbore.

14. The apparatus of claim 13 wherein the processor is further configured to determine the value of the bending moment of the drilling assembly using the neural network model during drilling of the wellbore.

15. The apparatus of claim 14 further comprising a force application device that is configured to apply force onto a wall of the wellbore during drilling of the wellbore to change drilling direction.

16. The apparatus of claim 15, wherein the processor is further configured to alter the force applied by the force application device to cause a change in drilling direction in response to the value of the bending moment of the drilling assembly determined by using the neural network model.

17. The apparatus of claim 13, wherein the processor is further configured to determine an output value of an operating property of the tool selected from a group consisting of:

(i) stress;

(ii) vibration;

(iii) whirl;

(iv) stick-slip;

(v) oscillation;

(vi) deflection; and

(vii) tilt.

18. The apparatus of claim 13 further comprising a drilling motor that rotates a drill bit for drilling the wellbore.

19. The apparatus of claim 13 further comprising a measurement-while-drilling device for estimating a property of a formation surrounding the wellbore.

20. The apparatus of claim 19, wherein the measurement-while-drilling device is selected from a group consisting of:

(i) a resistivity device;

(ii) an acoustic device;

(iii) a nuclear device;

(iv) a nuclear magnetic resonance device; and

(v) a formation testing device.

21. The apparatus of claim 13, wherein the processor is located at one of:

(i) in the drilling assembly; and

(ii) at the surface.

22. The apparatus of claim 13, wherein the processor is configured to cause altering an operation of the tool based on the value of the bending moment of the drilling assembly determined using the neural network model, which tool operation is selected from a group consisting of:

(i) drilling direction;

(ii) force applied on a wellbore inner surface by a force application device to drill the wellbore along a desired path;

(iii) weight on bit;

(iv) speed of a drilling bit;

(v) rotational speed of a drill string; and

(vi) mud motor speed;

(vii) inclination;

(viii) azimuth; and

(ix) orientation.