Method and apparatus for in-situ detection and isolation of aircraft engine faults

• US 7,280,941 B2
• Filed: 12/29/2004
• Issued: 10/09/2007
• Est. Priority Date: 12/29/2004
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Brunell, Brent Jerome, Bonanni, Pierino Gianni
• Primary Examiner(s): Barlow; John
• Assistant Examiner(s): Le; John H.
• Application Number:
  US11/025,145
• Publication Number:
  US 20060142976A1
• Time in Patent Office: 1,014 Days
• Field of Search: 702/34, 702/181, 702/183, 702/189, 702/185, 700/30, 700/83, 701/13, 701/29, 701/100, 324/555, 714/736, 731/16, 703/2, 703/7
• US Class Current: 702/189
• CPC Class Codes: G05B 23/0254 based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks
• CPC Class Codes: G05B 23/0281 Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis

Abstract

A method for performing a fault estimation based on residuals of detected signals includes determining an operating regime based on a plurality of parameters, extracting predetermined noise standard deviations of the residuals corresponding to the operating regime and scaling the residuals, calculating a magnitude of a measurement vector of the scaled residuals and comparing the magnitude to a decision threshold value, extracting an average, or mean direction and a fault level mapping for each of a plurality of fault types, based on the operating regime, calculating a projection of the measurement vector onto the average direction of each of the plurality of fault types, determining a fault type based on which projection is maximum, and mapping the projection to a continuous-valued fault level using a lookup table.

Claims

1. A method for performing a fault estimation based on a plurality of residuals of a plurality of detected signals from a machine, comprising:

determining an operating regime based on a plurality of parameters;

extracting predetermined noise standard deviations of said residuals corresponding to said operating regime and scaling said residuals;

calculating a magnitude of a measurement vector of said scaled residuals and comparing said magnitude to a decision threshold value;

extracting a mean direction and a fault level mapping for each of a plurality of fault types, based on said operating regime;

calculating a protection of said measurement vector onto said mean direction of each of said plurality of fault types;

determining a fault type based on a maximum projection of said measurement vector onto said mean direction of each of said plurality of fault types; and

mapping said projection to a continuous-valued fault level using a lookup table, wherein said calculating the magnitude of the measurement vector and comparing the same to the decision threshold value, determining the fault type based on the maximum projection of the measurement vector, and mapping said projection to the continuous-valued fault level are provided for subsequent use in performing said fault estimation of the machine.

2. The method of claim 1, wherein, if said magnitude is below said decision threshold value, it is determined that there is no fault.

3. The method of claim 1, wherein said parameters comprise flight envelope parameters.

4. The method of claim 1, wherein said scaling comprises normalizing said residuals by dividing said residuals by said noise standard deviations.

5. The method of claim 1, further comprising correlating said measurement vector with said mean direction of each of said plurality of fault types.

6. The method of claim 1, wherein said residuals comprise extended Kalman filter residuals.

7. The method of claim 1, wherein said detected signals comprise at least one of actuator signals, sensor signals, and engine signals.

8. The method of claim 1, wherein said detected signals comprise actuator signals, sensor signals, and engine signals.

9. The method of claim 1, wherein said fault types comprise at least one of a sensor fault, an actuator fault, a first machine fault and a second machine fault.

10. The method of claim 1, wherein said fault types comprise a sensor fault, an actuator fault, a first machine fault and a second machine fault.

11. The method of claim 1, wherein said mean direction for each fault type comprises a set of p-dimensional vectors, where p represents a number of sensors.

12. A computer program product for enabling a computer to implement operations for performing a fault estimation based on a plurality of residuals of a plurality of detected signals from a machine, the computer program product comprising a computer readable medium and instructions on the computer readable medium, the operations comprising:

determining an operating regime based on a plurality of parameters;

extracting predetermined noise standard deviations of said residuals corresponding to said operating regime and scaling said residuals;

calculating a magnitude of a measurement vector of said scaled residuals and comparing said magnitude to a decision threshold value;

extracting a mean direction and a fault level mapping for each of a plurality of fault types, based on said operating regime;

calculating a projection of said measurement vector onto said mean direction of each of said plurality of fault types;

determining a fault type based on a maximum projection of said measurement vector onto said mean direction of each of said plurality of fault types; and

mapping said projection to a continuous-valued fault level using a lookup table, wherein said calculating the magnitude of the measurement vector and comparing the same to the decision threshold value, determining the fault type based on the maximum projection of the measurement vector, and mapping said projection to the continuous-valued fault level are provided for subsequent use in performing said fault estimation of the machine.

13. The computer program product of claim 12, wherein, if said magnitude is below said decision threshold value, it is determined that there is no fault.

14. The computer program product of claim 12, wherein said parameters comprise flight envelope parameters, wherein said detected signals comprise at least one of actuator signals, sensor signals, and engine signals, and said system comprises an engine, wherein said fault types comprise at least one of a sensor fault, an actuator fault, a first machine fault and a second machine fault.

15. The computer program product of claim 12, wherein said scaling comprises normalizing said residuals by dividing said residuals by said noise standard deviations, and wherein said residuals comprise extended Kalman filter residuals.

16. The computer program product of claim 12, wherein the operations further comprise:

detecting a plurality of signals; and

determining a residual of each of said plurality of signals, and

wherein the fault estimation is performed for a system that includes an aircraft engine.

17. A method for detecting and isolating faults in a system, comprising:

detecting a plurality of signals;

determining a residual of each of said plurality of signals;

determining an operating regime based on a plurality of parameters;

extracting predetermined noise standard deviations of said residuals corresponding to said operating regime and scaling said residuals;

calculating a magnitude of a measurement vector of said scaled residuals and comparing said magnitude to a decision threshold value;

extracting a mean direction and a fault level mapping for each of a plurality of fault types, based on said operating regime;

calculating a projection of said measurement vector onto said mean direction of each of said plurality of fault types,

determining a fault type based on a maximum projection of said measurement vector onto said mean direction of each of said plurality of fault types; and

mapping said projection to a continuous-valued fault level using a lookup table wherein said calculating the magnitude of the measurement vector and comparing the same to the decision threshold value, determining the fault type based on the maximum projection of the measurement vector, and mapping said projection to the continuous-valued fault level are provided for subsequent use in performing said fault estimation of the machine.

18. The method of claim 17, further comprising correlating said measurement vector with said mean direction of each of said plurality of fault types;

wherein said measurement vector is determined by dividing each residual by a noise standard deviation;

wherein said residuals comprise extended Kalman filter residuals; and

wherein, if said magnitude is below said decision threshold value, it is determined that there is no fault.

19. The method of claim 17, wherein said parameters comprise flight envelope parameters, wherein said detected signals comprise at least one of actuator signals, sensor signals, and engine signals, and said system comprises an engine, wherein said fault types comprise at least one of a sensor fault, an actuator fault, a first machine fault and a second machine fault, and wherein said system comprises an aircraft engine.

20. An apparatus for detecting and isolating faults in a system based on a plurality of residuals of a plurality of detected signals from the system, said apparatus comprising:

a processor configured to determine an operating regime based on a plurality of parameters, extract predetermined noise standard deviations of said residuals corresponding to said operating regime and scale said residuals, calculate a magnitude of a measurement vector of said scaled residuals and compare said magnitude to a decision threshold value, extract a mean direction and a fault level mapping for each of a plurality of fault types, based on said operating regime, calculate a projector of said measurement vector onto said mean direction of each of said plurality of fault types, determine a fault type based on a maximum projection of said measurement vector onto said mean direction of each of said plurality of fault types and map said projection to a continuous-valued fault level using a lookup table, wherein the processor is configured to calculate the magnitude of the measurement vector and compare the same to the decision threshold value, determine the fault type based on the maximum projection of the measurement vector, and map said projection to the continuous-valued fault level for subsequent use in detecting and isolating the faults in the system.

21. The apparatus of claim 20, wherein said processor is configured to operate in real time.

22. The apparatus of claim 20, wherein said apparatus is disposed in an aircraft and said system comprises an aircraft engine.

23. The apparatus of claim 20, wherein said processor comprises an aircraft engine controller.

24. The apparatus of claim 20, wherein said apparatus is configured to detect and isolate faults in a plurality of actuators, a plurality of sensors, and an engine.

25. A system for detecting and isolating faults based on a plurality of residuals of a plurality of detected signals, said system comprising:

a detector which detects said detected signals;

an extended Kalman filter which compares said detected signals with estimates of said detected signals and outputs a plurality of residuals; and

a processor which performs hypothesis testing on said residuals to determine a fault type and a fault level, wherein said processor is configured to determine an operating regime based on a plurality of parameters, extract predetermined noise standard deviations of said residuals corresponding to said operating regime and scale said residuals, calculate a magnitude of a measurement vector of said scaled residuals and compare said magnitude to a decision threshold value, extract a mean direction and a fault level mapping for each of a plurality of fault types, based on said operating regime, calculate a projection of said measurement vetcor onto said mean direction of each of said plurality of fault types, determine a fault type based on a maximum projection of said measurement vector onto said mean direction of each of said plurality of fault types, and map said projection to a continuous-valued fault level using a lookup table.

26. The system of claim 25, wherein said detector comprises a plurality of sensors.

27. The system of claim 25, wherein said hypothesis testing comprises Bayesian hypothesis testing.

28. The system of claim 25, wherein said processor is configured to operate in real time.

29. The system of claim 25, wherein said processor comprises an aircraft engine controller and is disposed in an aircraft and said system comprises an aircraft engine, and wherein said system is configured to detect and isolate faults in a plurality of actuators, a plurality of sensors, and an engine.

30. A method for performing fault estimation based on a plurality of residuals of a plurality of detected signals from a machine, said method comprising:

comparing said detected signals with estimates of said detected signals, based on an extended Kalman filter, and outputting said residuals; and

determining a fault type and a fault level by performing hypothesis testing on said residuals, said determining further comprises determining an operating regime based on a plurality of parameters, extracting predetermined noise standard deviations of said residuals corresponding to said operating regime and scale said residuals, calculating a magnitude of a measurement vector of said scaled residuals and compare said magnitude to a decision threshold value, extracting a mean direction and a fault level mapping for each of a plurality of fault types, based on said operating regime, calculating a projection of said measurement vetcor onto said mean direction of each of said plurality of fault types, determining a fault type based on a maximum projection of said measurement vector onto said mean direction of each of said plurality of fault types, and mapping said projection to a continuous-valued fault level using a lookup table, wherein said calculating the magnitude of the measurement vector and comparing the same to the decision threshold value, determining the fault type based on the maximum projection of the measurement vector, and mapping said projection to the continuous-valued fault level are provided for subsequent use in performing said fault estimation of the machine.

31. The method of claim 30, wherein said hypothesis testing comprises Bayesian hypothesis testing.