FAULT SPLITTING ALGORITHM

• US 20100198610A1
• Filed: 02/05/2009
• Published: 08/05/2010
• Est. Priority Date: 02/05/2009
• Status: Active Grant

• Current Assignee: Honeywell International Inc.
• Original Assignee: Honeywell International Inc.
• Inventors: Khalak, Asif, Dins, C. Arthur, Barton, Bradley John, Shao, Qingqiu Ginger, Kolbet, David Michael, Magnuson, Randy
• Granted Patent: US 8,175,846 B2
• Time in Patent Office: Days
• US Class Current: 705/2
• CPC Class Codes: G16H 40/40 for the management of medical equipment or devices, e.g. scheduling maintenance or upgrades
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

The present application relates to a method of splitting a fault condition including receiving evidence observations of a monitored system from monitors connected in a many-to-many relationship to the failure modes, generating a fault condition, computing a relative probability of failure for each failure mode in the fault condition. When there is more than one failure mode in the fault condition, the method includes computing a relative probability of each pair of failure modes in the fault condition, ranking the computed relative probabilities of the individual failure modes and the computed relative probabilities of the pairs of failure modes. If the highest ranked failure mode is a pair of failure modes, the fault condition is split based on the failure modes in the highest ranked pair of failure modes are split. If the highest ranked failure mode is an individual failure mode, a failure is isolated based on the ranking.

Claims

1. A method for splitting a fault condition based on a relative likelihood of the failure modes, the method comprising:

receiving evidence observations of a monitored system from monitors connected in a many-to-many relationship to the failure modes;

generating a fault condition comprising states of all failure modes for a reference model of the monitored system based on the evidence observations;

computing a relative probability of failure for each failure mode in the fault condition, the relative probability being based on a false alarm probability, a detection probability, and a ratio of prior probabilities of a candidate hypothesis to a null hypothesis of no active failure modes;

when there is more than one failure mode in the fault condition, computing a relative probability of each pair of failure modes in the fault condition, the relative probability being based on a false alarm probability, a detection probability, and a ratio of prior probabilities of a candidate hypothesis to a null hypothesis of no active failure modes;

ranking the computed relative probabilities of the individual failure modes and the computed relative probabilities of the pairs of failure modes;

if the highest ranked failure mode is a pair of failure modes, splitting the fault condition based on the failure modes in the highest ranked pair of failure modes; and

if the highest ranked failure mode is an individual failure mode, isolating a failure based on the ranking.

2. The method of claim 1, wherein splitting the fault condition comprises:

generating a first fault condition for states of failure modes for a reference model of a monitored system based on a first failure mode; and

generating a second fault condition for states of failure modes for the reference model of the monitored system based on a second failure mode.

3. The method of claim 2, wherein generating a first fault condition and generating a second fault condition comprises:

determining the proximity of each other-failure mode to the first failure mode;

determining the proximity of each other-failure mode to the second failure mode;

attaching to the first fault condition each of the other-failure modes that is closer to the first failure mode than to the second failure mode; and

attaching to the second fault condition each of the other-failure modes that is closer to the second failure mode than to the first failure mode.

4. The method of claim 2, further comprising:

computing a relative probability of failure for each failure mode in the first fault condition;

when there is more than one failure mode in the first fault condition, computing a relative probability of each pair of failure modes in the first fault condition;

ranking the computed relative probabilities of the individual failure modes in the first fault condition and the computed relative probabilities of the pairs of failure modes in the first fault condition;

if the highest ranked failure mode in the first fault condition is a pair of failure modes, splitting the first fault condition based on the failure modes in the highest ranked pair of failure modes in the first fault condition; and

if the highest ranked failure mode in the first fault condition is an individual failure mode, isolating a failure based on a ranking of the relative probability of failure for each failure mode in the first fault condition.

5. The method of claim 4, further comprising:

computing a relative probability of failure for each failure mode in the second fault condition;

when there is more than one failure mode in the second fault condition, computing a relative probability of each pair of failure modes in the second fault condition;

ranking the computed relative probabilities of the individual failure modes in the second fault condition and the computed relative probabilities of the pairs of failure modes in the second fault condition;

if the highest ranked failure mode in the second fault condition is a pair of failure modes, splitting the second fault condition based on the failure modes in the highest ranked pair of failure modes in the second fault condition; and

if the highest ranked failure mode in the second fault condition is an individual failure mode, isolating a failure based on a ranking of the relative probability of failure for each failure mode in the second fault condition.

6. The method of claim 1, further comprising:

receiving additional evidence observations of the monitored system from the monitors;

generating an updated fault condition for the states of failure modes for the reference model of the monitored system, the updated fault condition being based on the additional evidence observations received from the monitors;

computing a relative probability of failure for each failure mode in the updated fault condition;

when there is more than one failure mode in the updated fault condition, compute a relative probability of each pair of failure modes in the updated fault condition;

ranking the computed relative probabilities of the individual failure modes and the relative probabilities of the pairs of failure modes; and

splitting the fault condition based on the failure modes if the highest ranked failure mode is a pair of failure modes.

7. The method of claim 1, further comprising:

generating a ratio of the prior probabilities of the hypothesis to the null hypothesis;

generating a trigger factor associated with triggered monitors in the fault condition;

generating a quiescent factor associated with quiescent monitors in the fault condition; and

determining the relative probability from the ratio of the prior probabilities of the hypothesis to the null hypothesis, the trigger factor and the quiescent factor.

8. The method of claim 7, wherein generating the trigger factor associated with the triggered monitors in the fault condition comprises:

generating a probability for each monitor that the monitor reads TRUE due to a false alarm; and

generating a probability for each monitor to read TRUE due to a failure, the probability being generated for every failure mode in the fault condition.

9. The method of claim 1, wherein generating the quiescent factor associated with the quiescent monitors in the fault condition comprises:

generating a probability for each monitor to read TRUE due to a failure, the probability being generated for every failure mode in the fault condition.

10. The method of claim 1, further comprising:

generating the reference model of the monitored system; and

implementing Model-based software to generate the fault condition for the states-of-failure modes for the reference model of the monitored system.

11. A computer readable medium storing instructions to implement a method to:

generate a fault condition comprising states of all failure modes for a reference model of a monitored system, the fault condition being based on evidence observations received from monitors;

compute a relative probability of failure for each failure mode based on a false alarm probability, a detection probability, and a ratio of prior probabilities of a candidate hypothesis to a null hypothesis of no active failure modes;

when there is more than one failure mode in the fault condition, compute a relative probability of each pair of failure modes in the fault condition, the relative probability based on a false alarm probability, a detection probability, and a ratio of prior probabilities of a candidate hypothesis to a null hypothesis of no active failure modes;

rank the computed relative probabilities of the individual failure modes and the computed relative probabilities of the pairs of failure modes;

if the highest ranked failure mode is a pair of failure modes, split the fault condition based on the failure modes in the highest ranked pair of failure modes; and

if the highest ranked failure mode is the individual failure mode, isolate a failure based on the ranking

12. The computer readable medium of claim 11, wherein the instructions to split the fault condition if the highest ranked failure mode is a pair of failure modes comprise instructions to implement a method to:

generate a first fault condition for states of failure modes for the reference model of the monitored system based on a first failure mode; and

generate a second fault condition for states of failure modes for the reference model of the monitored system based on a second failure mode.

13. The computer readable medium of claim 12, wherein the instructions to generate a first fault condition and to generate a second fault condition comprise instructions to implement a method to:

determine the proximity of each other-failure mode, other than the first failure mode and the second failure mode, to the first failure mode;

determine the proximity of each other-failure mode to the second failure mode;

attach to the first fault condition each of the other-failure modes that is closer to the first failure mode than to the second failure mode; and

attach to the second fault condition each of the other-failure modes that is closer to the second failure mode than to the first failure mode.

14. The computer readable medium of claim 12, wherein the instructions further include instructions to implement a method to:

compute a relative probability of failure for each failure mode in the first fault condition;

when there is more than one failure mode in the first fault condition, compute a relative probability of each pair of failure modes in the first fault condition;

rank the computed relative probabilities of the individual failure modes in the first fault condition and the computed relative probabilities of the pairs of failure modes in the first fault condition;

if the highest ranked failure mode in the first fault condition is a pair of failure modes, split the first fault condition based on the failure modes in the highest ranked pair of failure modes in the first fault condition; and

if the highest ranked failure mode in the first fault condition is an individual failure mode, isolate a failure based on the ranking of the computed relative probability of failure for each failure mode in the first fault condition.

15. The computer readable medium of claim 14, wherein the instructions further include instructions to implement a method to:

compute a relative probability of failure for each failure mode in the second fault condition;

when there is more than one failure mode in the second fault condition, compute a relative probability of each pair of failure modes in the second fault condition;

rank the computed relative probabilities of the individual failure modes in the second fault condition and the computed relative probabilities of the pairs of failure modes in the second fault condition;

if the highest ranked failure mode in the second fault condition is a pair of failure modes, split the second fault condition based on the failure modes in the highest ranked pair of failure modes in the second fault condition; and

if the highest ranked failure mode in the second fault condition is an individual failure mode, isolate a failure based on the ranking of the computed relative probability of failure for each failure mode in the second fault condition.

16. The computer readable medium of claim 11, wherein the instructions further include instructions to implement a method to:

generate an updated fault condition for states of failure modes for the reference model of the monitored system, the updated fault condition being based on additional evidence observations received from monitors;

compute a relative probability of failure for each failure mode in the updated fault condition;

when there is more than one failure mode in the updated fault condition, compute a relative probability of each pair of failure modes in the updated fault condition;

rank the computed relative probabilities of the individual failure modes and the pairs of failure modes; and

split the fault condition based on the failure modes in the highest ranked pair of failure modes if the highest ranked failure mode is a pair of failure modes.

17. A health management system configured to split a fault condition based on a relative likelihood of a failure mode in a monitored system, the health management system comprising:

sensors to detect system states and to output evidence observations;

at least one processor configured to;

generate a fault condition including states of all failure modes that are connected to the monitors based on the evidence observations received from the sensors for a reference model of the monitored system;

execute algorithms configured to compute a relative probability of failure for each failure mode based on a false alarm probability, a detection probability, and a ratio of prior probabilities of a candidate hypothesis to a null hypothesis of no active failure modes; and

split the fault condition when additional evidence is received to indicate that a given fault condition includes two failure modes; and

a storage medium to store software executable by the at least one processor.

18. The health management system of claim 17, wherein the monitored system includes system components, wherein at least one system component is communicatively coupled to at least one other system component, and wherein the sensors are configured to monitor the system components and the linking between two system components.

19. The health management system of claim 17, further comprising a failure-mode-relative-likelihood algorithm stored in the storage medium.

20. The health management system of claim 17, further comprising a reference model of the monitored system stored in the storage medium.