Corrosion analysis using magnetic flux leakage measurements

Patent metadata

Field	Value
Publication number	US-12031900-B2
Application number	US-202217970873-A
Country	US
Kind code	B2
Filing date	Oct 21, 2022
Priority date	Oct 21, 2022
Publication date	Jul 9, 2024
Grant date	Jul 9, 2024

Abstract

Aspects of the subject technology relate to systems, methods, and computer-readable media for corrosion analysis using magnetic flux leakage measurements. The present technology can magnetic flux leakage data obtained by a magnetic flux leakage tool placed in a pipe within a wellbore and convert the magnetic flux leakage data into image data. Further, the present technology can provide the image data to a machine learning model. The machine learning model is configured to identify one or more physical parameters associated with corrosion present on the pipe.

First claim

What is claimed is:

1. A method comprising:
disposing a magnetic flux leakage tool in a pipe within a wellbore;
operating the magnetic flux leakage tool to generate magnetic flux leakage data for the pipe within the wellbore;
converting the magnetic flux leakage data into image data;
providing the image data to a machine learning model;
identifying, by applying the machine learning model to the image data, one or more physical parameters associated with corrosion present on the pipe.

2. The method of claim 1 , further comprising:
outputting, via the machine learning model with the image data as input, the one or more physical parameters associated with the corrosion.

3. The method of claim 1 , wherein the one or more physical parameters associated with the corrosion include a defect circumferential width, a defect axial length, a detect penetration, a volume of metal loss, and a combination thereof.

4. The method of claim 3 , further comprising:
providing the one or more physical parameters outputted by the machine learning model into a second machine learning model to identify a second physical parameter associated with the corrosion.

5. The method of claim 1 , wherein providing the image data to the machine learning model includes:
identifying at least one anomaly in the image data;
determining one or more features associated with the at least one anomaly; and
providing the one or more features associated with the at least one anomaly to the machine learning model.

6. The method of claim 5 , wherein the one or more features associated with the at least one anomaly include at least one of an anomaly circumferential width, an anomaly axial length, a peak deflection from a baseline, mean deflection from a baseline, width and length in a percentile, and an anomaly contour length.

7. The method of claim 1 , further comprising:
providing measurement parameters to the machine learning model, wherein the measurement parameters include at least one of an inner indicator, an outer indicator, a logging direction, an outside diameter (OD) of the pipe, an inside diameter (ID) of the pipe, a logging speed, and an eccentricity between the pipe and another outer pipe.

8. The method of claim 1 , wherein the machine learning model is a regression model.

9. The method of claim 1 , wherein the magnetic flux leakage tool comprises a magnet configured to magnetize the pipe and one or more magnetic field sensors disposed along a circumference of the pipe and configured to measure flux leakage.

10. The method of claim 1 , wherein the magnetic flux leakage tool comprises a magnet configured to at least partially saturate the pipe and one or more eddy current sensors disposed along a circumference of the pipe and configured to measure secondary magnetic fields generated by eddy currents.

11. The method of claim 1 , wherein the image data is a 2-dimensional measurement of a circumference of the pipe with respect to a measured depth of the pipe.

12. The method of claim 1 , wherein the machine learning model comprises a plurality of machine learning models, wherein each of the plurality of machine learning models is configured to identify one of the one or more physical parameters associated with the corrosion.

13. The method of claim 1 , wherein the machine learning model is trained with training data including at least one of data obtained through logging a test fixture, simulation data from a test fixture, and data obtained through logging a field log.

14. A system comprising:
one or more sensors of a magnetic flux leakage (MFL) tool configured to receive magnetic flux leakage data when the MFL tool is disposed and operating downhole within a wellbore
a memory; and
one or more processors coupled to the memory, the one or more processors being configured to:
receive the magnetic flux leakage data obtained by the MFL tool placed in a pipe within the wellbore;
convert the magnetic flux leakage data into image data;
provide the image data to a machine learning model;
identify, by applying the machine learning model to the image data, one or more physical parameters associated with corrosion present on the pipe.

15. The system of claim 14 , wherein the one or more processors are configured to:
output, via the machine learning model with the image data as input, the one or more physical parameters associated with the corrosion.

16. The system of claim 14 , wherein the one or more physical parameters associated with the corrosion present in the pipe include a defect circumferential width, a defect axial length, a detect penetration, a volume of metal loss, and a combination thereof.

17. The system of claim 14 , wherein providing the image data to the machine learning model includes:
identifying at least one anomaly in the image data;
determining one or more features associated with the at least one anomaly; and providing the one or more features associated with the at least one anomaly to the machine learning model.

18. The system of claim 17 , wherein the one or more features associated with the at least one anomaly include at least one of an anomaly circumferential width, an anomaly axial length, a peak deflection from a baseline, mean deflection from a baseline, width and length in a percentile, and an anomaly contour length.

19. The system of claim 14 , wherein the one or more processors are configured to:
provide measurement parameters to the machine learning model, wherein the measurement parameters include at least one of an inner indicator, an outer indicator, a logging direction, an outside diameter (OD) of the pipe, an inside diameter (ID) of the pipe, a logging speed, and an eccentricity between the pipe and another outer pipe.

20. The system of claim 14 , wherein the machine learning model is a regression model.

Assignees

• Halliburton Energy Services Inc

Inventors

• Sayed Sadeed
• Fouda Ahmed E
• Hill Freeman Lee

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/084

  Backpropagation, e.g. using gradient descent · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06N20/00

  Machine learning · CPC title