Drilling control

Patent metadata

Field	Value
Publication number	US-12297732-B2
Application number	US-202117304151-A
Country	US
Kind code	B2
Filing date	Jun 15, 2021
Priority date	Jun 15, 2021
Publication date	May 13, 2025
Grant date	May 13, 2025

Abstract

A method can include receiving sensor data; determining a rate of penetration drilling parameter value using a trained neural network and at least a portion of the sensor data; and issuing a control instruction for drilling a borehole using the determined rate of penetration drilling parameter value.

First claim

What is claimed is:

1. A method comprising:
receiving sensor data;
performing training to generate a trained neural network, wherein the training comprises utilizing a trained digital avatar to train the neural network as an agent;
determining a rate of penetration drilling parameter value using the trained neural network and at least a portion of the sensor data, wherein the rate of penetration drilling parameter value comprises a set point value, and wherein the trained neural network is trained using a reward function that comprises a plurality of terms; and
issuing a control instruction for drilling a borehole using the determined rate of penetration drilling parameter value.

2. The method of claim 1 , wherein the set point value comprises a weight on bit value.

3. The method of claim 1 , wherein the set point value comprises a torque value.

4. The method of claim 1 , wherein the set point value comprises a pressure value.

5. The method of claim 1 , wherein the set point value is a set point for at least one of proportional control and integral control.

6. The method of claim 5 , wherein the set point value corresponds to a control loop for at least one of weight on bit, torque, and differential pressure.

7. The method of claim 1 , comprising training another neural network to generate the trained digital avatar, wherein the another neural network comprises a deep Kalman filter.

8. The method of claim 7 , wherein the training to generate the trained digital avatar comprises feeding output of multiple heads of the another neural network to one or more long short-term memory components.

9. The method of claim 1 , wherein the trained digital avatar comprises another trained neural network that comprises at least one convolution neural network and at least one long short-term memory component.

10. The method of claim 1 , wherein the trained digital avatar comprises at least two heads for input.

11. The method of claim 10 , wherein the at least two heads comprise a state head for input of state information and a control head for input of control information.

12. The method of claim 1 , wherein the issuing the control instruction for drilling a borehole using the determined rate of penetration drilling parameter value comprises issuing the control instruction to an autodriller controller that controls a drawworks based on the set point value.

13. A system comprising:
a processor;
memory accessible to the processor;
processor-executable instructions stored in the memory and executable by the processor to instruct the system to:
receive sensor data;
perform training to generate a trained neural network, wherein the training comprises utilizing a trained digital avatar to train the neural network as an agent;
determine a rate of penetration drilling parameter value using the trained neural network and at least a portion of the sensor data, wherein the rate of penetration drilling parameter value comprises a set point value, and wherein the trained neural network is trained using a reward function that comprises a plurality of terms; and
issue a control instruction for drilling a borehole using the determined rate of penetration drilling parameter value.

14. One or more non-transitory computer-readable storage media comprising computer-executable instructions executable to instruct a computing system to:
receive sensor data;
perform training to generate a trained neural network, wherein the training comprises utilizing a trained digital avatar to train the neural network as an agent;
determine a rate of penetration drilling parameter value using the trained neural network and at least a portion of the sensor data, wherein the rate of penetration drilling parameter value comprises a set point value, and wherein the trained neural network is trained using a reward function that comprises a plurality of terms; and
issue a control instruction for drilling a borehole using the determined rate of penetration drilling parameter value.

Assignees

• Schlumberger Technology Corp

Inventors

• Wicks Nathaniel
• Yu Yingwei
• Meehan Richard John
• Mansour Darine

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/092

  Reinforcement learning · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title