# US 2023/0065532 A1 · Machine learning based techniques for predicting component corrosion likelihood

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0065532 A1
- Document ID: US20230065532A1
- Status: Published application
- Kind: A1
- Application number: 17/878,511
- Filed: 2022-08-01
- Published: 2023-03-02
- Earliest claimed: 2021-08-09
- Assignees: Baker Hughes Holdings LLC
- Applicants: Baker Hughes Holdings LLC
- Inventors: Godine Kok Yan Chan; Okja Kim; Tarik Borogovac; Shyam Boriah; Mohamad Bagheri Esfe
- Web page: https://patentgrep.com/patent/US20230065532A1

## CPC

- G06N 3/042: Knowledge-based neural networks; Logical representations of neural networks
- G06N 3/09: Supervised learning
- G01N 17/006: {of metals}
- G06N 3/08: Learning methods
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/20: Ensemble learning
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q)

## Abstract

A machine learning based method for determining a likelihood of corrosion of a component is provided. The method comprises receiving data associated with a portion of at least one component, the data describing one or more operating conditions of the portion of the at least one component, applying, to the data associated with the portion, a first machine learning model, determining, responsive to the applying of the first machine learning model, a likelihood of corrosion specific to the at least one component based at least in part on the one or more operating conditions of the portion, and outputting, automatically and without user intervention, the likelihood of corrosion specific to the at least one component on a display.

## Claims

### Claim 1

A method comprising:
receiving data associated with a portion of at least one component, the data describing one or more operating conditions of the portion of the at least one component;
applying, to the data associated with the portion, at least one machine learning model;
determining, responsive to the applying of the at least one machine learning model, a likelihood of corrosion specific to the at least one component based at least in part on the one or more operating conditions of the portion; and
outputting, automatically and without user intervention, the likelihood of corrosion specific to the at least one component on a display.

### Claim 2 (depends on 1)

The method of claim 1, wherein the at least one machine learning model includes a first machine learning model that is based on a sequential deep learning neural network.

### Claim 3 (depends on 1)

The method of claim 1, wherein the operating conditions are associated with at least one of an amount of moisture or an addition amount of corrosion byproduct on the portion of the at least one component.

### Claim 4 (depends on 1)

The method of claim 1, further comprising:
applying, to the data associated with the portion, the at least one machine learning model, wherein the at least one machine learning model includes a second machine learning model trained for predicting a depth measurement associated with an additional component; and
predicting, responsive to the applying of the second machine learning model, a depth value specific to the additional component.

### Claim 5 (depends on 4)

The method of claim 4, wherein the depth value specific to the additional component is associated with a portion of the additional component and is larger that at least one addition portion of the additional component.

### Claim 6 (depends on 4)

The method of claim 4, further comprising:
predicting, responsive to the applying of the second machine learning model, a thickness value of the additional component at the depth value that is associated with the portion of the additional component.

### Claim 7 (depends on 4)

The method of claim 4, wherein the second machine learning model is based on random forest regression.

### Claim 8 (depends on 1)

The method of claim 1, further comprising:
classifying the at least one component as including the portion that is corroding based on the likelihood of corrosion satisfying or exceeding a threshold value;
receiving additional data associated with at least an additional portion of at least one additional component, the additional data describing one or more operating conditions specific to the at least one additional component;
applying the at least one machine learning model to the additional data, wherein the at least one machine learning model includes a first machine learning model;
determining, responsive to the applying of the first machine learning model, an additional likelihood of corrosion specific to the at least one additional component; and
classifying the at least one additional component as independent of corrosion at a first time based on the additional likelihood of corrosion satisfying or exceeding an additional threshold value.

### Claim 9 (depends on 8)

The method of claim 8, further comprising:
applying, to the additional data associated with the at least one additional component, the at least one machine learning model, wherein the at least one machine learning model includes a third machine learning model; and
applying, based on the third machine learning model, a bounded margin to the at least one addition component.

### Claim 10 (depends on 9)

The method of claim 9, further comprising reclassifying the at least one additional component as including the corrosion at a second time.

### Claim 11

A system, comprising:
at least one processor, and
memory storing instructions that, when executed by the at least one processor, cause the at least one processor to perform operations comprising:
receiving data associated with a portion of at least one component, the data describing one or more operating conditions of the portion of the at least one component;
applying, to the data associated with the portion, at least one machine learning model;
determining, responsive to the applying of the at least one machine learning model, a likelihood of corrosion specific to the at least one component based at least in part on the one or more operating conditions of the portion; and
outputting, automatically and without user intervention, the likelihood of corrosion specific to the at least one component on a display.

### Claim 12 (depends on 11)

The system of claim 11, wherein the at least one machine learning model includes a first machine learning model that is based on a sequential deep learning neural network.

### Claim 13 (depends on 11)

The system of claim 11, wherein the operating conditions are associated with at least one of an amount of moisture or an addition amount of corrosion byproduct on the portion of the at least one component.

### Claim 14 (depends on 11)

The system of claim 11, wherein the operations further comprising:
applying, to the data associated with the portion, the at least one machine learning model, wherein the at least one machine learning model includes a second machine learning model trained for predicting a depth measurement associated with an additional component; and
predicting, responsive to the applying of the second machine learning model, a depth value specific to the additional component.

### Claim 15 (depends on 14)

The system of claim 14, wherein the depth value specific to the additional component is associated with a portion of the additional component and is larger that at least one addition portion of the additional component.

### Claim 16 (depends on 14)

The system of claim 14, wherein the operations further comprising predicting, responsive to the applying of the second machine learning model, a thickness value of the additional component at the depth value that is associated with the portion of the additional component.

### Claim 17 (depends on 14)

The system of claim 14, wherein the second machine learning model is based on random forest regression.

### Claim 18 (depends on 11)

The system of claim 11, wherein the operations further comprising:
classifying the at least one component as including the portion that is corroding based on the likelihood of corrosion satisfying or exceeding a threshold value;
receiving additional data associated with at least an additional portion of at least one additional component, the additional data describing one or more operating conditions specific to the at least one additional component;
applying the at least one machine learning model to the additional data, wherein the at least one machine learning model includes a first machine learning model;
determining, responsive to the applying of the first machine learning model, an additional likelihood of corrosion specific to the at least one additional component; and
classifying the at least one additional component as independent of corrosion at a first time based on the additional likelihood of corrosion satisfying or exceeding an additional threshold value.

### Claim 19 (depends on 18)

The system of claim 18, wherein the operations further comprising:
applying, to the additional data associated with the at least one additional component, the at least one machine learning model, wherein the at least one machine learning model includes a third machine learning model; and
applying, based on the third machine learning model, a bounded margin to the at least one addition component.

### Claim 20 (depends on 19)

The system of claim 19, wherein the operations further comprising reclassifying the at least one additional component as including the corrosion at a second time.
