METHOD AND APPARATUS FOR IMPROVING FAULT CLASSIFICATIONS

• US 20030195675A1
• Filed: 04/12/2002
• Published: 10/16/2003
• Est. Priority Date: 04/12/2002
• Status: Active Grant

• Current Assignee: Honeywell International Inc.
• Original Assignee: Honeywell International Inc.
• Inventors: Magnuson, Randy R., Felke, Timothy J., Kolbet, David M., van der Zweep, Jeff J.
• Granted Patent: US 6,662,089 B2
• Time in Patent Office: Days
• US Class Current: 701/29
• CPC Class Codes: G05B 23/0278 Qualitative, e.g. if-then rules; Fuzzy logic; Lookup tables; Symptomatic search; FMEA

Abstract

A method for improving fault classifications for failure modes in complex systems, and a software program embodying the method or a corresponding method together with a computer based aircraft diagnostic system using such software is described. The method includes: capturing an observation that is consistent with a discrepancy of the complex system; identifying a set of failure modes of the complex system that are consistent with the observation; ranking, when the set includes more than one failure mode, the failure modes according to a degree of support the observation provides for each failure mode, the degree of support including a historical component; determining additional possible observations that are consistent with any failure mode included in the set; and ranking the possible observations according to how rapidly they reduce a number of failure modes included in the set; then preferably selecting a further observation and repeating the above steps until only one failure mode remains.

Claims

1. A method for improving fault classifications for failure modes in complex systems, the method including the steps of:

capturing an observation that is consistent with a discrepancy of the complex system;

identifying a set of failure modes of the complex system that are consistent with said observation;

ranking, when said set includes more than one failure mode, said failure modes according to a degree of support said observation provides for each failure mode, said degree of support including a historical component;

determining additional possible observations that are consistent with any failure mode included in said set; and

ranking said possible observations according to how rapidly they reduce a number of failure modes included in said set.

2. The method of claim 1 further including a step of selecting a further observation from said possible observations and repeating, using said observation, said further observation and any additional possible observations that will be selected, said steps of identifying, ranking said failure modes, determining said additional possible observations, ranking said additional possible observations and selecting further possible observations until said set includes only one failure mode and then reporting said one failure mode.

3. The method of claim 1 wherein said step of capturing said observation includes converting an observed discrepancy into one or more standard observations selected from a standard observations list and wherein said possible observations are possible standard observations that are members of said standard observations list.

4. The method of claim 1 wherein said step of identifying said set of said failure modes includes identifying a standard failure mode selected from a standard failure modes list, said standard failure mode representing a set of observations consistent with said standard failure mode.

5. The method of claim 1 wherein said step of ranking said failure modes further uses an assessment of experience with said observation occurring with said failure mode to increase said degree of support.

6. The method of claim 1 wherein said ranking said possible observations uses an algorithm to rank said possible observations according to the degree to which each possible observation divides the set of failure modes in half.

7. The method of claim 1 wherein said possible observations can be reported as one of present, absent, and unknown.

8. The method of claim 1 used within a diagnostic tool and applied to a plurality of fault models for a plurality of aircraft.

9. A software program comprising software instructions arranged to run on a processor to improve fault classification for failure modes in a complex system, the software program when installed and operating on a processor resulting in the processor:

capturing an observation that is consistent with a discrepancy of the complex system;

identifying a set of failure modes of the complex system that are consistent with said observation;

ranking, when said set includes more than one failure mode, said failure modes according to a degree of support said observation provides for each failure mode, said degree of support including a historical component;

determining additional possible observations that are consistent with any failure mode included in said set; and

ranking said possible observations according to how rapidly they reduce a number of failure modes included in said set.

10. The software program of claim 9 further including selecting a further observation from said possible observations and repeating, using said observation, said further observation and any additional possible observations that will be selected, said identifying, said ranking said failure modes, said determining said additional possible observations, said ranking said additional possible observations and said selecting further possible observations until said set includes only one failure mode and then reporting said failure mode.

11. The software program of claim 9 wherein said capturing said observation includes converting an observed discrepancy into one or more standard observations selected from a standard observations list and wherein said possible observations are possible standard observations that are members of said standard observations list.

12. The software program of claim 9 wherein said identifying said set of said failure modes includes identifying a standard failure mode selected from a standard failure modes list, said standard failure mode representing a set of observations consistent with said standard failure mode.

13. The software program of claim 9 wherein said ranking said failure modes further uses an assessment of experience with said observation occurring with said failure mode to increase said degree of support.

14. The software program of claim 9 wherein said ranking said possible observations uses an algorithm to rank said possible observations according to the degree to which each possible observation divides the set of failure modes in half.

15. The software program of claim 9 wherein said possible observations can be reported as one of present, absent, and unknown.

16. The software program of claim 9 used within a diagnostic tool and applied to a plurality of fault models for a plurality of aircraft.

17. An aircraft maintenance and operations diagnostic system for improved assistance with fault classification for a fault condition within an aircraft based on a fault model for the aircraft, the system comprising in combination:

a user interface;

a computer, coupled to the user interface, having memory for storing software instructions and databases and a processor for;

executing said software instructions to process information to facilitate the fault classification for the fault condition according to a fault model, the software program resulting in the computer;

capturing a standard observation from an observation list that is part of the fault model and that is consistent with the fault condition;

identifying a set of fault codes for the aircraft that are consistent with said standard observation;

ranking, when said set includes more than one fault code, said fault codes according to a degree of support said standard observation provides for each fault code, said degree of support including a historical component;

determining additional possible standard observations that are consistent with any fault code included in said set; and

ranking said possible standard observations according to how rapidly they reduce a number of fault codes included in said set.

18. The system of claim 17 further including selecting a further standard observation from said possible standard observations and repeating, using said standard observation, said further standard observation and any additional possible standard observations that will be selected, said identifying, said ranking said fault codes, said determining said additional possible standard observations, said ranking said additional possible standard observations and selecting further possible standard observations until said set includes only one fault code and then reporting said fault code.

19. The system of claim 17 wherein said capturing said standard observation includes converting an observed discrepancy into one or more standard observations.

20. The system of claim 17 wherein said identifying said set of said fault codes includes identifying a fault code selected from a standard fault code list that is part of the fault model, said fault code representing a set of standard observations consistent with the fault condition.

21. The system of claim 17 wherein said ranking said fault codes further uses an assessment of a number of occurrences of a fault code and a number of co-occurrences of said standard observation and said fault code to increase said degree of support.

22. The system of claim 17 wherein said ranking said possible observations uses an algorithm to rank said possible observations according to the degree to which each possible observation divides said set of fault codes in half.

23. The system of claim 17 further including displaying said set of possible observations as ranked on said user interface to facilitate a user reviewing and selecting a further standard observation.

24. The system of claim 17 further including a portal to a wide area network to provide remote access to said system to provide diagnostic support for fault classification to remote users.