System and method for isolating failures in a locomotive

• US 5,845,272 A
• Filed: 11/29/1996
• Issued: 12/01/1998
• Est. Priority Date: 11/29/1996
• Status: Expired due to Term

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Morjaria, Mahesh Amritlal, Smith, William David, Nash, James Weston, Bush, James Arthur, Azzaro, Steven Hector, Smith, Myron Lee
• Primary Examiner(s): Hafiz, Tariq R.
• Assistant Examiner(s): Shah, Sanjiv
• Application Number:
  US08/753,583
• Time in Patent Office: 732 Days
• Field of Search: 395/50-54, 395/60-61, 395/183.02, 364/550, 364/148, 706/50, 706/45
• US Class Current: 706/50
• CPC Class Codes: B60L 2200/26 Rail vehicles
• CPC Class Codes: B60L 3/00 Electric devices on electrically-propelled vehicles for safety purposes; Monitoring operating variables, e.g. speed, deceleration or energy consumption methods or circuit arrangements for monitoring or controlling batteries or fuel cells B60L58/00
• CPC Class Codes: B60L 3/12 Recording operating variables ; Monitoring of operating variables
• CPC Class Codes: B61C 17/00 Arrangement or disposition of parts; Details or accessories not otherwise provided for; Use of control gear and control systems
• CPC Class Codes: B61C 5/00 Locomotives or motor railcars with IC engines or gas turbines characterised by power transmission B61C9/00; engines F02
• CPC Class Codes: G06F 11/2257 using expert systems
• CPC Class Codes: Y02T 50/60 Efficient propulsion technologies, e.g. for aircraft

Abstract

The present invention discloses a system and method for isolating failures in a locomotive. In the present invention, the cause of failures occurring in the locomotive sub-systems are determined quickly and efficiently, while minimizing the need for human interaction. However, the present invention can incorporate information provided by an operator to produce a more refined diagnosis. In the present invention, incidents are recorded in an incident log and mapped to indicators, which are sent to a fault isolator for diagnosis. The fault isolator uses a diagnostic knowledge base having diagnostic information between failures occurring in locomotive and observable symptoms happening therein. A diagnostic engine then processes the mapped indicators with the diagnostic information in the diagnostic knowledge base and produces a list of the most likely causes for any failures. The present invention then provides a course of action to correct the failures. The present invention can also be used to automatically isolate and diagnose failures occurring in process having a plurality of equipment.

Claims

1. A system for isolating failures in a locomotive having a plurality of sub-systems, comprising:

means for supplying information on incidents occurring in each of the plurality of sub-systems during operation of the locomotive;

means for mapping some of the incidents to indicators, each indicator representative of an observable symptom detected in a sub-system;

a fault isolator coupled to the mapping means for determining causes for any failures associated with the incidents, the fault isolator comprising a diagnostic knowledge base having diagnostic information about failures occurring in each of the plurality of sub-systems and the indicators, wherein the diagnostic information comprises a plurality of causal networks, each causal network having a plurality of nodes for each of the plurality of sub-systems, each causal network having a cause and effect relationship between some of the plurality of nodes, wherein some of the nodes represent root causes associated with failures in each of the plurality of sub-systems and some of the nodes represent observable manifestations of the failures and a diagnostic engine for processing the mapped indicators with the diagnostic information in the diagnostic knowledge base, wherein the diagnostic engine comprises an indicator evaluator for invoking the plurality of causal networks according to the mapped indicators and evaluating the mapped indicators with the probabilities in the invoked causal networks and a network solver for recalculating the probabilities of the invoked causal networks according to the status of the mapped indicators; and

means for providing a course of action to be performed for correcting the failures.

2. The system according to claim 1, wherein each of the root cause nodes in the causal networks has a prior probability indicating the likelihood of a failure in the absence of any additional knowledge.

3. The system according to claim 2, wherein each of the nodes in the causal networks has conditional probability information representing the strength of the relationships of the node to its causes.

4. The system according to claim 3, wherein the diagnostic engine further comprises a diagnostic evaluator for evaluating the recalculated probabilities in the invoked causal networks and providing a list of the most likely causes for any failures.

5. The system according to claim 4, wherein the diagnostic evaluator prompts an operator to check a manually examined indicator that may have a further impact on the diagnosis of the locomotive.

6. The system according to claim 5, wherein the indicator evaluator invokes the plurality of causal networks according to the status of the manually examined indicator.

7. The system according to claim 6, wherein the network solver recalculates the probabilities in the invoked causal networks according to the status of the manually examined indicator.

8. The system according to claim 7, wherein the diagnostic evaluator evaluates the recalculated probabilities in the invoked causal networks and provides a list of the most likely causes for any failures.

9. A method for isolating failures in a locomotive having a plurality of sub-systems, comprising the steps of:

recording incidents occurring in each of the plurality of sub-systems during operation of the locomotive;

mapping some of the recorded incidents to indicators, each indicator representative of an observable symptom detected in each sub-system;

determining causes for any failures associated with the incidents with a diagnostic knowledge base having diagnostic information about failures occurring in each of the plurality of sub-systems and the indicators, wherein the diagnostic information comprises a plurality of causal networks, each causal network having a plurality of nodes for each of the plurality of subsystems, each causal network having a cause and effect relationship between some of the plurality of nodes, wherein some of the nodes represent root causes associated with failures in each of the plurality of sub-systems and some of the nodes represent observable manifestations of the failures and a diagnostic engine for processing the mapped indicators with the diagnostic information in the diagnostic knowledge base;

invoking the plurality of causal networks according to the mapped indicators and evaluating the mapped indicators with the probabilities in the invoked networks;

recalculating the probabilities of the invoked causal networks according to the status of the mapped indicators; and

providing a course of action to be performed for correcting the failures.

10. The method according to claim 9, wherein each of the root cause nodes in the causal networks has a prior probability indicating the likelihood of a failure in the absence of any additional knowledge.

11. The method according to claim 10, wherein each of the nodes in the causal networks has conditional probability information representing the strength of the relationships of the node to its causes.

12. The method according to claim 11, further comprising the steps of evaluating the recalculated probabilities in the invoked causal networks and providing a list of the most likely causes for any failures.

13. The method according to claim 12, further comprising the step of prompting an operator to check a manually examined indicator that may have a further impact on the diagnosis.

14. The method according to claim 13, further comprising the step of invoking the plurality of causal networks according to the status of the manually examined indicator.

15. The method according to claim 14, further comprising the step of recalculating the probabilities in the invoked causal networks according to the status of the manually examined indicator.

16. The method according to claim 15, further comprising the steps of evaluating the recalculated probabilities in the invoked causal networks and providing a list of the most likely causes for any failures.

17. A system for automatically isolating failures occurring in a process having a plurality of equipment, comprising:

a plurality of sensors located about the process for measuring physical phenomena associated with the plurality of equipment;

means for mapping some of the sensor measurements to indicators, each indicator representative of an observable symptom detected in a piece of equipment;

a fault isolator coupled to the mapping means for determining causes for any failures associated with the equipment, the fault isolator comprising a diagnostic knowledge base having diagnostic information about failures occurring in each of the plurality of equipment and the indicators, the diagnostic information comprising a plurality of causal networks each having a plurality of nodes for each of the plurality of equipment, each causal network having a cause and effect relationship between some of the plurality of nodes, wherein some of the nodes represent root causes associated with failures in each of the plurality of equipment and some of the nodes represent observable manifestations of the failures, and a diagnostic engine for processing the mapped indicators with the diagnostic information in the diagnostic knowledge base, wherein the diagnostic engine comprises an indicator evaluator for invoking the plurality of causal networks according to the mapped indicators and evaluating the mapped indicators with the probabilities in the invoked causal networks and a diagnostic evaluator for evaluating the recalculated probabilities in the invoked causal networks and providing a list of the most likely causes for any equipment failures; and

means for providing a course of action to be performed for correcting the failures.

18. The system according to claim 17, wherein each of the root causes in the causal networks has a prior probability indicating the likelihood of a failure in the absence of any additional knowledge.

19. The system according to claim 18, wherein each of the nodes in the causal networks has conditional probability representing the strength of the relationships of the node to its causes.

20. The system according to claim 19, wherein the diagnostic engine further comprises a diagnostic evaluator for evaluating the recalculated probabilities in the invoked causal networks and providing a list of the most likely causes for any equipment failures.

21. The system according to claim 20, wherein the diagnostic evaluator prompts an operator to check a manually examined indicator that may have a further impact on the diagnosis of the process.

22. The system according to claim 21, wherein the indicator evaluator invokes the plurality of causal networks according to the status of the manually examined indicator.

23. The system according to claim 22, wherein the network solver recalculates the probabilities in the invoked causal networks according to the status of the manually examined indicator.

24. The system according to claim 23, wherein the diagnostic evaluator evaluates the recalculated probabilities in the invoked causal networks and provides a list of the most likely causes for any equipment failures.

25. A method for isolating failures occurring in a process having a plurality of equipment, comprising the steps of:

providing a plurality of causal networks each having a plurality of nodes for each of the plurality of equipment, each causal network having a cause and effect relationship between some of the plurality of nodes, wherein some of the nodes represent root causes associated with failures in each of the plurality of equipment and some of the nodes represent observable manifestations of the failures;

transforming the causal networks into a knowledge base having diagnostic information about failures occurring in each of the plurality of equipment and observable symptoms detected in each of the plurality of equipment;

coupling a diagnostic engine to the knowledge base;

measuring physical phenomena associated with the equipment;

mapping some of the measurements to indicators, each indicator representative of an observable symptom detected in a piece of equipment;

using the diagnostic information and the diagnostic engine to determine causes for any failures associated with the mapped indicators;

invoking the plurality of causal networks according to the mapped indicators and evaluating the mapped indicators with the probabilities in the invoked causal networks;

recalculating the probabilities of the invoked causal networks according to the status of the mapped indicators; and

providing a course of action to be performed for correcting the failures.

26. The method according to claim 25, wherein each of the root cause nodes in the causal networks has a prior probability indicating the likelihood of a failure in the absence of any additional knowledge.

27. The method according to claim 26, wherein each of the nodes in the causal networks has conditional probability information representing the strength of the relationships of the node to its causes.

28. The method according to claim 27, further comprising the steps of evaluating the recalculated probabilities in the invoked causal networks and providing a list of the most likely causes for any equipment failures.

29. The method according to claim 28, further comprising the step of prompting an operator to check a manually examined indicator that may have a further impact on the diagnosis.

30. The method according to claim 29, further comprising the step of invoking the plurality of causal networks according to the status of the manually examined indicator.

31. The method according to claim 30, further comprising the step of recalculating the probabilities in the invoked causal networks according to the manually examined indicator.

32. The method according to claim 31, further comprising the steps of evaluating the recalculated probabilities in the invoked causal networks and providing a list of the most likely causes for any equipment failures.