HYBRID FAULT REASONING AND GUIDED TROUBLESHOOTING SYSTEM

• US 20090276387A1
• Filed: 05/05/2008
• Published: 11/05/2009
• Est. Priority Date: 05/05/2008
• Status: Active Grant

• Current Assignee: Rtx Corporation
• Original Assignee: United Technologies Corporation (Rtx Corporation)
• Inventors: Rajamani, Ravi, Baliga, Baban V., Paquet, Michel D., Gagnon, Andrew Paul
• Granted Patent: US 7,788,209 B2
• Time in Patent Office: Days
• US Class Current: 706/52
• CPC Class Codes: G06N 5/045 Explanation of inference; Explainable artificial intelligence [XAI]; Interpretable artificial intelligence

Abstract

A fault reasoning system combines both a model-based reasoning approach based upon predicted root causes of a plurality of fault signals, and a case-based reasoner that looks at the actual root cause of actual cases.

Claims

1. A fault reasoning system comprising:

a model-based reasoner to take in a plurality of fault signals, and predict a likely root cause for the plurality of fault signals based upon predictions developed utilizing a model of a particular piece of equipment;

a case database storing a plurality of prior fault signals, and associating a particular combination of fault signals with a particular root cause that has been identified in actual equipment in the past; and

a case-based reasoner having access to said case database, said case-based reasoner taking in the plurality of fault signals, and searching the case database for a prior case having a similar set of fault signals, with said case-based reasoner predicting a likely root cause if there is a prior case having a number of similar fault signals, and said model-based reasoner being utilized to predict a root cause when said case-based reasoner does not have such a previously stored case.

2. The system as set forth in claim 1, wherein the plurality of fault signals is fault codes initially sent to said model-based reasoner, and said model-based reasoner inverting the fault codes to predict likely root causes as isolation solutions to be sent to said case-based reasoner.

3. The system as set forth in claim 2, wherein said fault codes are also sent to said case-based reasoner.

4. The system as set forth in claim 1, wherein said fault reasoning system is programmed to identify root causes of problems with a gas turbine engine.

5. The system as set forth in claim 1, wherein said case database is updated when a unique set of fault signals is identified with a new case, and to relate that unique set of fault signals with a finally identified root cause.

6. The system as set forth in claim 1, wherein an output of the fault reasoning system is directed to a guided troubleshooting tool to provide maintenance directives.

7. The system as set forth in claim 1, wherein said model-based reasoner and said case-based reasoner are included onto a computer.

8. The system as set forth in claim 1, wherein said case database is dedicated and mounted with said case-based reasoner.

9. The system as set forth in claim 1, wherein said case database is remotely accessible to said case-based reasoner.

10. A method of operating a fault reasoning system comprising:

taking in a plurality of fault signals and making predictions at a model-based reasoner developed utilizing a model of a particular piece of equipment;

storing a plurality of prior fault signals on a case database, and associating a particular combination of fault signals with a particular root cause that has been identified in actual equipment in the past; and

allowing a case-based reasoner access to said plurality of prior fault signals, said case-based reasoner taking in the plurality of fault signals, and searching the case database for a prior case having a similar set of fault signals, with said case-based reasoner predicting a likely root cause if there is a prior case having a number of similar fault signals, and said model-based reasoner being utilized to predict a root cause when said case-based reasoner does not have such a previously stored case.

11. The method as set forth in claim 10, wherein said model-based reasoner inverting the fault signals to predict likely root causes as isolation solutions to be sent to said case-based reasoner.

12. The method as set forth in claim 10, wherein said plurality of fault signals are also sent to said case-based reasoner.

13. The method as set forth in claim 10, wherein said fault reasoning system is programmed to identify root causes of problems with a gas turbine engine.

14. The method as set forth in claim 10, wherein said case database is updated when a unique set of fault signals is identified with a new case, and to relate that unique set of fault signals with a finally identified root cause.

15. The method as set forth in claim 10, wherein an output of the fault reasoning system is directed to a guided troubleshooting tool to provide maintenance directives.

16. The method as set forth in claim 10, wherein observations made by personnel are included with said plurality of fault signals and utilized to make said predictions at said model-based reasoner, and said case-based reasoner.