Predictive engine maintenance apparatuses, methods, systems and techniques

Patent metadata

Field	Value
Publication number	US-12148508-B2
Application number	US-202217664700-A
Country	US
Kind code	B2
Filing date	May 24, 2022
Priority date	Nov 26, 2019
Publication date	Nov 19, 2024
Grant date	Nov 19, 2024

Abstract

A method includes inputting used oil analysis data to a pre-trained predictive model, the used oil analysis data including values quantifying a plurality of chemical components measured in a sample of used oil taken from an engine under analysis, determining a probability of at least one fail code with the pre-trained predictive model in response to the used oil analysis data, the at least one fail code corresponding to one of a plurality of predetermined engine failure types, providing the at least one fail code and the probability of the at least one fail code to an expert system, performing with the expert system a root cause analysis of the at least one fail code determine a root cause indicating a preventative maintenance action, and performing the predictive maintenance action on the engine under analysis.

First claim

The invention claimed is:

1. A method comprising:
providing a computing system configured with a combination of models including a classification model pre-trained by machine learning to output one or more probabilities of one or more fail codes in response to input including oil analysis data, a recommendation model configured to receive output of the pre-trained classification model and pre-trained by machine learning to output a related items dataset in response to the one or more probabilities of one or more fail codes and, and an expert system model configured to receive output of the recommendation model and pre-trained by machine learning to output a root cause indicating a preventative maintenance action in response the related items dataset;
inputting used oil analysis data to the classification model, the used oil analysis data including values quantifying a plurality of chemical components measured in a sample of used oil taken from an engine under analysis;
determining a probability of at least one fail code with the classification model in response to the used oil analysis data, the at least one fail code corresponding to one of a plurality of predetermined engine failure types;
determining with the recommendation model a related items dataset in response to the at least one fail code;
providing the related items dataset, the at least one fail code and the probability of the at least one fail code to the expert system model;
determining a root cause indicating a preventative maintenance action with the expert system model in response to the related items dataset, the at least one fail code and the probability of the at least one fail code; and
performing the predictive maintenance action on the engine under analysis.

2. The method of claim 1 wherein the related items dataset includes one or more other fail codes with a correlational relationship to the at least one fail code.

3. The method of claim 2 wherein wherein the related items dataset includes one or more other fail codes with a causal relationship to the at least one fail code.

4. The method of claim 1 wherein the models include classification model comprises a boosted decision tree model.

5. The method of claim 1 wherein the classification model is trained using an oil analysis training dataset including values quantifying a plurality of chemical components measured use engine oil samples taken during oil change maintenance events for a plurality of engines of a common type or model.

6. The method of claim 5 wherein the classification model is trained using an engine failure training dataset comprising a set of engine failure codes indicating one of a plurality of predetermined failure types for said plurality of engines and over a corresponding time period as oil analysis training dataset.

7. The method of claim 1 wherein the classification model is trained using an engine failure training dataset comprising a set of engine failure codes indicating one of a plurality of predetermined failure types for a plurality of engines.

8. The method of claim 1 wherein the expert system model is configured to identify root causes of the fail codes using an inference engine which operates on a knowledge base represented by a ruleset of if-then rules.

9. The method of claim 1 wherein the expert system model is configured to use a forward chaining operation.

10. The method of claim 1 wherein the expert system model is configured to use a backward chaining operation.

11. A system for predictive engine maintenance, the system comprising:
one or more computers configured to implement a combination of model components including a classification model component pre-trained by machine learning to output one or more probabilities of one or more fail codes in response to input including oil analysis data, a recommendation model component configured to receive output of the pre-trained classification model component and pre-trained by machine learning to output a related items dataset in response to the one or more probabilities of one or more fail codes and, and an expert system model component configured to receive output of the recommendation model component and pre-trained by machine learning to output a root cause indicating a preventative maintenance action in response the related items dataset, wherein
the classification model component is configured to receive an input including used oil analysis data and to determine a probability of at least one fail code in response to the used oil analysis data, the used oil analysis data including values quantifying a plurality of chemical components measured in a sample of used oil taken from an engine under analysis, the at least one fail code corresponding to one of a plurality of predetermined engine failure types;
the recommendation model component is configured to determine a related items dataset in response to the at least one fail code; and
the expert system model component is configured to receive an input including the related items dataset, the at least one fail code and the probability of the at least one fail code, perform a root cause analysis of the at least one fail code to determine a root cause in response to the related items dataset, the at least one fail code and the probability of the at least one fail code, and in response to the root cause indicate a preventative maintenance action for the engine under analysis.

12. The system of claim 11 wherein the related items dataset includes one or more other fail codes with a correlational relationship to the at least one fail code.

13. The system of claim 11 wherein the related items dataset includes one or more other fail codes with a causal relationship to the at least one fail code.

14. The system of claim 13 wherein the classification model component includes a boosted decision tree model.

15. The system of claim 11 wherein the classification model component is trained using an oil analysis training dataset including values quantifying a plurality of chemical components measured use engine oil samples taken during oil change maintenance events for a plurality of engines of a common type or model.

16. The system of claim 15 wherein the classification model component is trained using an engine failure training dataset comprising a set of engine failure codes indicating one of a plurality of predetermined failure types for said plurality of engines and over a corresponding time period as oil analysis training dataset.

17. The system of claim 11 wherein the classification model component is trained using an engine failure training dataset comprising a set of engine failure codes indicating one of a plurality of predetermined failure types for a plurality of engines.

18. The system of claim 11 wherein the expert system model component configured to identify root causes of the fail codes using an inference engine which operates on a knowledge base represented by a ruleset of if-then rules.

19. The system of claim 11 wherein the root cause analysis is performed using a forward chaining operation.

20. The system claim 11 wherein the root cause analysis is performed using a backward chaining operation.

Assignees

• Cummins Inc

Inventors

• Denton Ryan E
• Xue Xinjian
• Trobaugh Corey W
• Huth Anthony Joseph

Classifications

• G06N5/01

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

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G16B40/20 Primary

  Supervised data analysis · CPC title

• G06Q10/20 Primary

  Administration of product repair or maintenance · CPC title