Contactless sensor monitoring of a drill string controlled by a drilling program

1. A monitoring system for a drill string controlled by a drilling program, the system comprising:

a contactless sensor comprising a camera and arranged to capture data at a first time, the data comprising a first image captured by the camera and being associated with a parameter at a position along a drill line associated with the drill string; and

a processor operatively connected to the contactless sensor and configured to:

access a machine learning (ML) model that is trained using the data captured at a second time, the second time prior to the first time, and an image analysis algorithm,

generate a first feature vector of the data captured at the first time, the first feature vector based on one or more distances between the drill line and a reference straight line as shown in the first image,

generate, using the trained ML model, a prediction that the data captured at the first time corresponds to a condition pertaining to a downhole event of the drill string based on analyzing the first feature vector,

identify, using the trained ML model, the condition pertaining to a downhole event of the drill string based on, at least in part, the prediction, and

cause an adjustment to the drilling program based on the identified condition.

2. The monitoring system of claim 1 , wherein:

the camera captures a second image of the drill line at the second time, and

the processor is further configured to train the ML model, the training of the ML model including:

determining, using a pixelation technique applied to the second image, a distance between the drill line, as shown in the image, and a reference straight line added to the second image in relation to the drill line, at a survey point of the reference straight line; and

generating a second feature vector of the second image based on one or more distances between the drill line and the reference straight line, as shown in the second image, the one or more distances being determined at one or more survey points of the reference straight line, as shown in the second image.

3. The monitoring system of claim 2 , wherein the training of the ML model further includes:

associating the second feature vector of the second image with either a first label corresponding to a normal state of the drill string, or a second label corresponding to the condition pertaining to the downhole event of the drill string, the associating being based on the one or more distances between the drill line and the reference straight line.

4. The monitoring system of claim 2 , wherein the camera is a first camera

and the contactless sensor further comprises a second camera that captures a third image of the drill line, the third image being captured at the second time, and

wherein the training of the ML model further includes:

generating a third feature vector of the third image based on one or more distances determined between the drill line, as shown in the third image, and the reference straight line added to the third image in relation to the drill line, the one or more distances being determined at the one or more survey points of the reference straight line, as shown in the third image; and

associating a union of the second first feature vector and the third feature vector with either a first label corresponding to the normal state of the drill string, or a second label corresponding to the condition pertaining to the downhole event of the drill string, the associating being based on the one or more distances between the drill line and the reference straight line.

5. The monitoring system of claim 4 , wherein the first camera captures a first angle of the drill line, and the second camera captures a second angle of the drill line.

6. The monitoring system of claim 2 , wherein the training of the ML model includes assigning a signature to a downhole condition based on a particular downhole event, and

wherein the identifying, using the ML model, of the condition pertaining to the downhole event of the drill string includes recognizing that the downhole condition occurred at the first time, the recognizing being based on matching the data captured at the first time and the signature.

7. The monitoring system of claim 1 , wherein a first portion of the data is captured by the contactless sensor at a first position along the drill line,

wherein a second portion of the data is captured by another sensor at a second position along the drill line, and

wherein the processor is further configured to aggregate and synchronize the first portion of the data and the second portion of the data.

8. The monitoring system of claim 1 , wherein a first portion of the data is captured by the contactless sensor at a first position along the drill line,

wherein a second portion of the data is captured by another sensor at a second position along the drill string, and

wherein the processor is further configured to aggregate and synchronize the first portion of the data and the second portion of the data.

9. The monitoring system of claim 1 , wherein the identifying, using the trained ML model, of the condition pertaining to the downhole event of the drill string includes:

classifying the downhole event as the condition based on analyzing, using the trained ML model, the data captured at the first time.

10. The monitoring system of claim 1 , wherein the identifying, using the trained ML model, of the condition pertaining to the downhole event of the drill string includes:

detecting an anomaly pertaining to the drill string based on detecting an outlier in the data captured at the first time.

11. The monitoring system of claim 1 , wherein the causing of the adjustment to the drilling program based on the identified condition includes performing a modification to an item of equipment associated with the identified condition.

12. The monitoring system of claim 1 , wherein the causing of the adjustment to the drilling program based on the identified condition includes:

generating an alert pertaining to the identified condition; and

causing display of the alert in a user interface of a client device.

13. A method for monitoring a drill string controlled by a drilling program, the method comprising:

using a contactless sensor comprising a camera and capturing data at a first time, the data comprising a first image captured by the camera and being associated with a parameter at a position along a drill line associated with the drill string;

accessing a machine learning (ML) model that is trained using the data captured at a second time, the second time prior to the first time, and an image analysis algorithm;

generating a first feature vector of the data captured at the first time, the first feature vector based on one or more distances between the drill line and a reference straight line as shown in the first image;

generating, using the trained ML model, a prediction that the data captured at the first time corresponds to a condition pertaining to a downhole event of the drill string based on analyzing the first feature vector;

identifying, using the trained ML model, the condition pertaining to a downhole event of the drill string based on, at least in part, the prediction, the identifying being performed by one or more hardware processors; and

causing an adjustment to the drilling program based on the identified condition.

14. The method of claim 13 , wherein the camera capturing a second image of the drill line at the second time, and

the method further comprises training the ML model, the training of the ML model including:

determining, using a pixelation technique applied to the second image, a distance between the drill line, as shown in the second image, and a reference straight line added to the second image in relation to the drill line, at a survey point of the reference straight line; and

generating a second feature vector of the second image based on one or more distances between the drill line and the reference straight line, as shown in the second image, the one or more distances being determined at one or more survey points of the reference straight line, as shown in the second image.

15. The method of claim 14 , wherein the training of the ML model further includes:

associating the second feature vector of the second image with either a first label corresponding to a normal state of the drill string, or a second label corresponding to the condition pertaining to the downhole event of the drill string, the associating being based on the one or more distances between the drill line and the reference straight line.

16. The method of claim 14 , wherein the camera is a first camera,

wherein the contactless sensor further comprises a second camera that captures a third image of the drill line, the third image being captured at the second time, and

wherein the training of the ML model further includes:

generating a third feature vector of the third image based on one or more distances determined between the drill line, as shown in the third image, and the reference straight line added to the third image in relation to the drill line, the one or more distances being determined at the one or more survey points of the reference straight line, as shown in the third image; and

associating a union of the second feature vector and the third feature vector with either a first label corresponding to the normal state of the drill string, or a second label corresponding to the condition pertaining to the downhole event of the drill string, the associating being based on the one or more distances between the drill line and the reference straight line.

17. The method of claim 14 , wherein the training of the ML model includes assigning a signature to a downhole condition based on a particular downhole event, and

wherein the identifying, using the ML model, of the condition pertaining to the downhole event of the drill string includes recognizing that the downhole condition occurred at the first time, the recognizing being based on matching the data captured at the first time and the signature.

18. The method of claim 13 , wherein the identifying, using the trained ML model, of the condition pertaining to the downhole event of the drill string includes:

classifying a downhole event as the condition based on analyzing, using the trained ML model, the data captured at the first time.

19. A non-transitory machine-readable storage medium comprising instructions that, when executed by one or more processors of a machine, cause the machine to perform operations comprising:

capturing data at a first time with a contactless sensor comprising a camera, the data comprising an image and being associated with a parameter at a position along a drill line associated with the drill string;

accessing a machine learning (ML) model that is trained using the data captured at a second time, the second time prior to the first time, and an image analysis algorithm;

generating a feature vector of the data captured at the first time, the feature vector based on one or more distances between the drill line and a reference straight line as shown in the image;

generating, using the trained ML model, a prediction that the data captured at the first time corresponds to a condition pertaining to a downhole event of the drill string based on analyzing the feature vector;

identifying, using the trained ML model, a condition pertaining to the downhole event of the drill string based on, at least in part, the prediction; and

causing an adjustment to the drilling program based on the identified condition.