Neural network fault diagnostics systems and related method

• US 5,919,267 A
• Filed: 04/09/1997
• Issued: 07/06/1999
• Est. Priority Date: 04/09/1997
• Status: Expired due to Term

• Current Assignee: Mcdonnell Douglas Corporation (The Boeing Co.)
• Original Assignee: Mcdonnell Douglas Corporation (The Boeing Co.)
• Inventors: Bond, William E., Urnes, James M.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Wright, Norman Michael
• Application Number:
  US08/833,770
• Time in Patent Office: 818 Days
• Field of Search: 395/183.02, 395/183.14, 395/183.13, 06/10, 06/23, 06/25
• US Class Current: 714/26
• CPC Class Codes: G06F 11/0736 in functional embedded systems, i.e. in a data processing system designed as a combination of hardware and software dedicated to performing a certain function testing or monitoring of automated control systems G05B23/02
• CPC Class Codes: G06F 11/0739 in a data processing system embedded in automotive or aircraft systems
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06N 3/045 Combinations of networks

Abstract

A fault diagnostics system for monitoring the operating condition of a host system, e.g., an aircraft, which includes a plurality of subsystems. The fault diagnostics system is preferably implemented in software running on a high-speed neural network processor. The fault diagnostics system constructs a neural network model of the performance of each subsystem in a normal operating mode and each of a plurality of different possible failure modes. The system preferably dynamically predicts the performance of each subsystem based upon the response of each of the neural network models to dynamically changing operating conditions, compares the actual performance of each subsystem with the dynamically predicted performance thereof in each of the normal and possible failure modes, and determines the operating condition of the host system on the basis of these comparisons. In a preferred embodiment, the determining step is carried out by performing a statistical analysis of the comparisons made in the comparing step, e.g., by emloying a comparison voting technique. A related method is also disclosed.

Claims

1. A fault diagnostics system for monitoring the operating condition of a host system which includes a plurality of subsystems, comprising:

neural network means for modeling the performance of each subsystem in a normal operating mode and a plurality of different possible failure modes;

comparator means for comparing the actual performance of each subsystem with the modeled performance thereof in each of the normal and possible failure modes; and

,

processing means for making a determination of the operating condition of the host system on the basis of the comparisons made by the comparator means.

2. The fault diagnostics system as set forth in claim 1, wherein the neural network means dynamically models the performance of each subsystem in response to dynamically changing operating conditions.

3. The fault diagnostics system as set forth in claim 2, wherein the neural network means, the comparator means, and the processing means are implemented in software running on a digital processor.

4. The fault diagnostics system as set forth in claim 3, wherein the digital processor comprises a neural network processor.

5. The fault diagnostics system as set forth in claim 2, wherein the processing means is programmed to make a determination of the operating condition of the host system on the basis of a statistical analysis of the comparisons made by the comparator means.

6. The fault diagnostics system as set forth in claim 2, further comprising sensor means coupled to components of each of the subsystems for generating feedback signals indicative of the dynamically changing operating conditions.

7. The fault diagnostics system as set forth in claim 1, wherein the processing means is programmed to make a determination of the operating condition of the host system on the basis of a statistical analysis of the comparisons made by the comparator means.

8. The fault diagnostics system as set forth in claim 1, wherein the host system comprises an aircraft.

9. The fault diagnostics system as set forth in claim 1, wherein the processing means is programmed to employ a comparison voting technique to make a determination of the operating condition of the host system.

10. The fault diagnostics system as set forth in claim 1, wherein:

the comparator means is programmed to make a plurality of comparisons between the actual performance of each subsystem and the modeled performance of the normal operating mode thereof, during a prescribed first time interval;

the processing means is programmed to determine, for each subsystem, whether or not there is a prescribed level of agreement between a prescribed percentage of the comparisons made by the comparator means relative thereto during the prescribed first time interval, and to produce a first output signal indicative of a normal operating mode thereof if the determination is affirmative;

the comparator means is further programmed to make, for each subsystem determined to have a fault, a plurality of comparisons between the actual performance of each subsystem and the modeled performance of each failure mode thereof, during successive prescribed time intervals; and

,

the processing means is programmed to determine, for each subsystem, whether or not there is a prescribed level of agreement between a prescribed percentage of the comparisons made by the comparator means relative thereto during each of the successive second time intervals, and to produce a second output signal indicative of the particular failure mode, if any, which is determined to have occurred.

11. A method for monitoring the operating condition of a host system which includes a plurality of subsystems, comprising the steps of:

modeling the performance of each subsystem in a normal operating mode and each of a plurality of different possible failure modes;

comparing the actual performance of each subsystem with the modeled performance thereof in each of the normal and possible failure modes; and

,

determining the operating condition of the host system on the basis of the comparisons made in the comparing step.

12. The method as set forth in claim 11, wherein the modeling step is performed by constructing a neural network model of the normal operating mode and each of the plurality of different possible failure modes of each subsystem.

13. The method as set forth in claim 12, further comprising the step of dynamically predicting the performance of each subsystem based upon the response of each of the neural network models to dynamically changing operating conditions.

14. The method as set forth in claim 13, wherein the modeling step is performed by a neural network processor.

15. The method as set forth in claim 12, wherein the determining step is carried out by performing a statistical analysis of the comparisons made in the comparing step.

16. The method as set forth in claim 11, wherein the determining step is carried out by performing a statistical analysis of the comparisons made in the comparing step.

17. The method as set forth in claim 11, wherein the host system comprises an aircraft.

18. The method as set forth in claim 11, wherein the determining step is carried out by performing a comparison voting technique.

19. The method as set forth in claim 11, wherein the comparing and determining steps are carried out by:

making a plurality of comparisons between the actual performance of each subsystem and the modeled performance of the normal operating mode thereof, during a prescribed first time interval;

determining, for each subsystem, whether or not there is a prescribed level of agreement go between a prescribed percentage of the comparisons made during the prescribed first time interval, and indicating that the operating condition of the respective subsystem is normal if there is the prescribed level of agreement;

making a plurality of comparisons between the actual performance of each subsystem determined to have a fault and the modeled performance of each failure mode thereof, during successive prescribed time intervals; and

,

determining, for each subsystem, whether or not there is a prescribed level of agreement between a prescribed percentage of the comparisons made during each of the successive second time intervals, and indicating the particular failure mode, if any, which is determined to have occurred.

20. A fault diagnostics system for monitoring the operating condition of a host system which includes a plurality of subsystems, comprising:

neural network facilities for modeling the performance of each subsystem in a normal operating mode and a plurality of different possible failure modes;

facilities for comparing the actual performance of each subsystem with the modeled performance thereof in each of the normal and possible failure modes; and

,

facilities for making a determination of the operating condition of the host system on the basis of the comparisons made by the comparator facilities.