System apparatus and method for diagnosing a flow system

• US 20040133289A1
• Filed: 01/23/2004
• Published: 07/08/2004
• Est. Priority Date: 03/06/2001
• Status: Active Grant

• Current Assignee: GoalArt AB
• Original Assignee: GoalArt AB
• Inventors: Ohman, Bengt, Larsson, Jan Eric, Dahlstrand, Fredrik, Tuszynski, Jan
• Granted Patent: US 7,062,358 B2
• Time in Patent Office: Days
• US Class Current: 700/83
• CPC Class Codes: G05B 23/0248 Causal models, e.g. fault tree; digraphs; qualitative physics
• CPC Class Codes: G05B 23/0278 Qualitative, e.g. if-then rules; Fuzzy logic; Lookup tables; Symptomatic search; FMEA
• CPC Class Codes: G05B 9/02 electric

Abstract

The present invention deals with diagnosing of a flow system, using status signals carrying information about said flow system, whereby new status information is generated, which relates to a plurality of aspects of the flow system based on said status signals and dependent on a functional model and on a set of predetermined rules. The invention comprises a diagnostic apparatus (20) arranged to analyze the status signals for example by means of an alarm analyzer (23a) arranged to separate alarms into primary and secondary alarms, a sensor fault detector (23b) arranged to detect sensor values that probably are wrong or a fault diagnozer (23c) arranged to find the root causes for unfulfilled goals. Further the invention can comprise a failure mode analyzer (23d) arranged to detect functions that can stop working in the future or a process measurement analyzer (23e) arranged to analyze the status signals of a specific process.

Claims

1. An apparatus for diagnosing a flow system of components, the flow system having at least a status signal indicative of the status of a component, comprising:

an interface arranged to receive said status signal from said flow system;

a data storage device arranged to store a functional model describing the functionality of said flow system, a set of predetermined rules describing the relationship between components of said flow system, and information about said status signal;

a data processing unit provided with a set of service functions arranged to generate status information relating to said flow system based on said status signal, and dependent on said functional model and on said set of predetermined rules; and

a control unit arranged to control said set of service functions dependent on selectable or predetermined information and arranged to activate or deactivate at least one service function.

2. The apparatus according to claim 1, wherein said service function is arranged to analyze a status signal such as an alarm, an alarm signal, a measurement, or an event recorded in said apparatus database.

3. The apparatus according to claim 1, wherein said service function comprises an alarm analyzer arranged to separate status signals corresponding to alarm into primary and secondary alarms.

4. The apparatus according to claim 1, wherein said service function comprises a sensor fault detector arranged to detect a sensor signal corresponding to a sensor value and/or a sensor measurement that probably is wrong.

5. The apparatus according to claim 1, wherein said service function comprises a fault diagnozer arranged to find a root cause for an unfulfilled goal by reducing the functional model to be traversed during the fault diagnosis.

6. The apparatus according to claim 1, wherein said service function comprises a failure mode analyzer arranged to detect a function that may stop working in the future and to estimate the time when the function may stop working.

7. The apparatus according to claim 1, wherein said service function comprises a process measurement analyzer arranged to analyze the received status signal and to generate information about the quality of an alarm signal, an alarm and a measurement.

8. The apparatus according to claim 1, wherein said control unit is arranged to monitor the operation of said service function.

9. The apparatus according to claim 1, further comprising an engineering station communicatively connectable to said control unit by means of a communication link, wherein said engineering station comprises means for configuration, means for building said functional model of said flow system and means for storing said functional model and configuration data.

10. The apparatus according to claim 1, wherein said service function is arranged to transmit said generated status information to a target system arranged to present said generated status information on a display unit.

11. The apparatus according to claim 1, wherein said set of predetermined rules comprises a set of consequence propagation rules.

12. The apparatus according to claim 1, wherein said functional model is a multilevel flow model (MFM) representing the flow system.

13. A method for diagnosing a flow system of components, the flow system having at least a status signal indicative of the status of a component, comprising the steps of:

receiving said status signal in an instrumentation and control database;

comparing said status signal with limit values;

setting a status signal by generating;

an alarm signal;

an alarm, if the status signal is outside the range of the limit values;

or

an event;

outputting said status signal from said instrumentation and control database;

receiving said status signal in a diagnostic apparatus;

storing said status signal in a data storage device;

detecting at least one change in said status signal;

controlling by means of a control unit said service function dependent on selectable or predetermined information; and

starting by means of said control unit at least one service function generating status information based on said status signal and dependent on a functional model of said flow system and on a set of predetermined rules.

14. The method according to claim 13, wherein the step of generating status information comprises the steps of:

searching for a cause to a changed alarm among the other functions in a network of said functional model, associated with alarm signals;

indicating the alarm signal as secondary, if the alarm signal for a function is caused by another function in the network, otherwise indicating the alarm signal as primary;

propagating causal effects to other functions in the network, if the alarm signal is indicated as primary;

propagating alarm signals to connected functions of conditional relations, if the goal connected to the network has conditional relations; and

repeating the step of searching for a cause to a changed alarm for each function where the alarm signals changes due to the propagation rules.

15. The method according to claim 13, wherein the step of generating status information comprises the steps of:

determining all alarm signals for the functions in a network of said functional model;

finding all possible root causes;

collecting a combination of possible root causes such that it explains all the observed alarm signals for the functions in the network;

marking all alarm signals not caused by any other alarm signal in the network as primary, and all other alarm signals in the network as secondary;

repeating the step of marking all alarm signals not caused by any other alarm signal in the network as primary, and all other alarm signals in the network as secondary for each combination of possible root causes collected;

propagating causal effects to other functions in the network if the alarm signal is marked as primary;

propagating alarm signals to the connected functions of the conditional relations if the goal connected to the network has conditional relations; and

repeating the step of determining all alarm signals for the functions in the network of said functional model for each function where the alarm signal changes due to the propagation along conditional relations.

16. The method according to claim 13, wherein the step of generating status information comprises the steps of:

determining if an alarm signal of a function in a network of said functional model, connected via a condition to a non-achieved goal, is not in the expected failure state; and

indicating that alarm signal as inconsistent.

17. The method according to claim 13, wherein the step generating status information comprises the steps of:

checking the alarm signals for each pair of connected functions in a network of said functional model; and

indicating the functions as inconsistent, if their alarm signals do not agree with the consequence propagation rules.

18. The method according to claim 13, wherein the step of generating status information comprises the steps of:

finding a network of said functional model with at least two primary alarms;

changing the alarm state of a function to another state for each function in the network;

storing the achieved situation and indicating the investigated alarm signal as inconsistent if the number of primary alarms in the network decreases; and

repeating the step of finding a network of said functional model with at least two primary alarms for each additional network with at least two primary alarms.

19. The method according to claim 13, wherein the step of generating status information comprises the steps of:

choosing a goal comprised in said functional model for diagnosis;

propagating the search downwards from the selected goal, via achieved relations, into a connected network of flow functions;

asking a diagnostic question associated to a flow function in order to find out whether the corresponding component is currently realizing the flow function, or finding information about the working order of the flow function by means of a sensor connected to the component;

setting the appropriate alarm state of the flow function; and

recursively investigating a connected sub-goal if a flow function connected to the sub-goal by means of a condition relation is found to be faulty or has no means of being checked, or skipping the part of the sub-tree that comprises the flow function if the flow function is working.

20. The method according to claim 13, wherein the step of generating status information comprises the steps of:

setting the failure time of a flow function comprised in said functional model to 0;

propagating the alarms as far as possible according to the consequence propagation rules and setting the failure time of the currently examined flow function to the current failure time, if the flow function has not been reached before or if the failure time stored for the flow function is larger than the current failure time;

updating the current failure time when passing condition relations and storage functions; and

stopping the propagation if the failure time of the currently examined flow function is less than the current failure time.

21. The method according to claim 13, wherein the step of generating status information comprises the steps of:

storing statistical information about said status signal; and

calculating properties of said status signal, e.g., mean or variance of a measurement signal, or number of times an alarm or event signal is triggered per time unit.

22. The method according to claim 13, further comprising the step of monitoring the operation of said service function.

23. The method according to claim 13, further comprising the steps of:

storing said generated status information in said data storage device;

detecting a change in said generated status information; and

outputting said generated status information.

24. The method according to claim 23, further comprising the steps of:

receiving said generated status information in said instrumentation and control database;

detecting said generated status information in said instrumentation and control database; and

presenting said generated status information on a display unit.

25. The method according to claim 13, wherein said set of predetermined rules comprises a set of consequence propagation rules.

26. The method according to claim 13, wherein said functional model is a multilevel flow model (MFM) representing said flow system.

41. A computer program product for use in apparatus for diagnosing a flow system, comprising software code portions for directing a data processing system to perform the method steps of claim 13.

27. A method for diagnosing a flow system of components, the flow system having at least a status signal indicative of the status of a component, comprising the steps of:

receiving said status signal in a diagnostic apparatus;

storing said status signal in a data storage device;

detecting at least one change in said status signal;

controlling by means of a control unit said service function dependent on selectable or predetermined information; and

starting by means of said control unit at least one service function generating status information based on said status signal and dependent on a functional model of said flow system and on a set of predetermined rules.

28. The method according to claim 27, wherein the step of generating status information comprises the steps of:

searching for a cause to a changed alarm among the other functions in a network of said functional model, associated with alarm signals;

indicating the alarm signal as secondary, if the alarm signal for a function is caused by another function in the network, otherwise indicating the alarm signal as primary;

propagating causal effects to other functions in the network, if the alarm signal is indicated as primary;

propagating alarm signals to connected functions of conditional relations, if the goal connected to the network has conditional relations; and

repeating the step of searching for a cause to a changed alarm for each function where the alarm signals changes due to the propagation rules.

29. The method according to claim 27 wherein the step of generating status information comprises the steps of:

determining all alarm signals for the functions in a network of said functional model;

finding all possible root causes;

collecting a combination of possible root causes such that it explains all the observed alarm signals for the functions in the network;

marking all alarm signals not caused by any other alarm signal in the network as primary, and all other alarm signals in the network as secondary;

repeating the step of marking all alarm signals not caused by any other alarm signal in the network as primary and all other alarm signals in the network as secondary for each combination of possible root causes collected;

propagating causal effects to other functions in the network if the alarm signal is marked as primary;

propagating alarm signals to the connected functions of the conditional relations if the goal connected to the network has conditional relations; and

repeating the step of determining all alarm signals for the functions in the network of said functional model for each function where the alarm signal changes due to the propagation along conditional relations.

30. The method according to claim 27, wherein the step of generating status information comprises the steps of:

determining if an alarm signal of a function in a network of said functional model, connected via a condition to a non-achieved goal, is not in the expected failure state; and

indicating that alarm signal as inconsistent.

31. The method according to claim 27, wherein the step generating status information comprises the steps of:

checking the alarm signals for each pair of connected functions in a network of said functional model; and

indicating the functions as inconsistent, if their alarm signals do not agree with the consequence propagation rules.

32. The method according to claim 27, wherein the step of generating status information comprises the steps of:

finding a network of said functional model with at least two primary alarms;

changing the alarm state of a function to another state for each function in the network;

storing the achieved situation and indicating the investigated alarm signal as inconsistent if the number of primary alarms in the network decrease; and

repeating the step of finding a network of said functional model with at least two primary alarms for each additional network with at least two primary alarms.

33. The method according to claim 27, wherein the step of generating status information comprises the steps of:

choosing a goal comprised in said functional model for diagnosis;

propagating the search downwards from the selected goal, via achieved relations, into a connected network of flow functions;

asking a diagnostic question associated to a flow function in order to find out whether the corresponding component is currently realizing the flow function, or finding information about the working order of the flow function by means of a sensor connected to the component;

setting the appropriate alarm state of the flow function; and

recursively investigating a connected sub-goal if a flow function connected to the sub-goal by means of a condition relation is found to be faulty or has no means of being checked, or skipping the part of the sub-tree that comprises the flow function if the flow function is working.

34. The method according to claim 27, wherein the step of generating status information comprises the steps of:

setting the failure time of a flow function comprised in said functional model to 0;

propagating the alarms as far as possible according to the consequence propagation rules and setting the failure time of the currently examined flow function to the current failure time, if the flow function has not been reached before or if the failure time stored for the flow function is larger than the current failure time;

updating the current failure time when passing condition relations and storage functions; and

stopping the propagation if the failure time of the currently examined flow function is less than the current failure time.

35. The method according to claim 27, further comprising the step of monitoring the operation of said service function.

36. The method according to claim 27, wherein the step of generating status information comprises the steps of:

storing statistical information about said status signal; and

calculating properties of said status signal, e.g. mean or variance of a measurement signal, or number of times an alarm or event signal is triggered per time unit.

37. The method according to claim 27, further comprising the steps of:

storing said generated status information in said data storage device;

detecting a change in said generated status information; and

outputting said generated status information.

38. The method according to claim 27, further comprising the step of presenting said generated status information on a display unit.

39. The method according to claim 27, wherein said set of predetermined rules comprises a set of consequence propagation rules.

40. The method according to claim 27, wherein said functional model is a multilevel flow model (MFM) representing said flow system.