Optimization method for adaptive sensor reading scheduling and delayed alarm evaluation in real-time diagnostic systems

• US 5,237,518 A
• Filed: 10/27/1990
• Issued: 08/17/1993
• Est. Priority Date: 10/27/1990
• Status: Expired due to Term

• Current Assignee: Osaka Gas Company Limited, Vanderbilt University
• Original Assignee: Osaka Gas Company Limited, Vanderbilt University
• Inventors: Okuda, Koji, Miyasaka, Nobuji, Padalkar, Samir, Sztipanovits, Janos, Karsai, Gabor, Biegl, Csaba
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): AUCHTERLONIE, THOMAS
• Application Number:
  US07/602,945
• Time in Patent Office: 1,025 Days
• Field of Search: 340/825.07, 340/825.08, 340/825.10, 340/825.11, 340/825.12, 340/825.13, 340/825.16, 340/825.54, 340/505, 340/518, 340/517, 364/550, 364/551.01, 364/474.19, 364/579, 364/580, 364/185, 395/911-918, 371/15.1, 371/29.1
• US Class Current: 702/185
• CPC Class Codes: G05B 2219/24063 Select signals as function of priority, importance for diagnostic
• CPC Class Codes: G05B 2219/24064 Sample rate variable as function of importance of alarm signals
• CPC Class Codes: G05B 2219/24065 Real time diagnostics
• CPC Class Codes: G05B 2219/24085 Analyze, trace fault signals according to tree, table
• CPC Class Codes: G05B 23/0248 Causal models, e.g. fault tree; digraphs; qualitative physics
• CPC Class Codes: G05B 23/0297 Reconfiguration of monitoring system, e.g. use of virtual sensors; change monitoring method as a response to monitoring results
• CPC Class Codes: G08B 29/14 checking the detection circuits
• CPC Class Codes: Y10S 706/911 Nonmedical diagnostics

Abstract

An optimization method is used in a real-time diagnostic system that monitors alarm sources and detects possible sources of failure in an industrial process. When the diagnostic system detects one or more alarms, it determines which components are likely sources of failure. The system determines the level of criticality of certain other, currently off-line, alarms sources and schedules the off-line alarms for evaluation according to their levels of criticality. The system refines its analysis of likely sources of failure by evaluating the status of the previously off-line alarm sources Those previously off-line alarm sources are added to the grouping of possible sources of failure prior to repeating the evaluation process. The evaluation process is repeated, with the analysis of likely causes of failure being continually refined.

Claims

1. For use in a diagnostic system for analyzing operations of a monitored process, the system including

alarms at each of a selected number of monitoring locations for providing alarm signals indicative of a function being monitored,

a real-time monitoring method comprising the steps of:

selecting a first group of alarms to be monitored,

processing signals from said first group of alarms for determining the occurrence of process fault conditions and identifying potential fault sources of said conditions,

determining, upon the occurrence of at least one process fault condition, at least one other alarm to be selected from a group of alarms which are off-line,

dynamically processing signals from said first group of alarms and said at least one other alarm for determining, with expected greater accuracy, said potential fault sources, and,

wherein the alarms of said first group are on-line alarms and said other alarms are off-line alarms when said selecting and processing step is performed.

2. The method of claim 1 further comprising the step of

modeling said process as a graph model having

a plurality of on-line nodes representing functions being monitored by alarms in said first group of alarms,

a plurality of off-line nodes representing functions being monitored by alarms not in said first group of alarms,

a plurality of nodes representing functions of the monitored process not having alarms associated therewith, and

a plurality of propagation paths selectively interconnecting pairs of nodes having a functional relationship.

3. The method of claim 2 wherein said step of determining said at least one other alarm comprises the steps of

determining each originating fault source from which said alarms providing process fault condition could have originated;

tracing forward from each said determined originating source, using said graph model, along said propagation paths to said off-line and on-line nodes,

increasing a criticality parameter value of each off-line node each time it is encountered during said tracing step, and

selecting alarms corresponding to off-line nodes having a higher criticality parameter value as said other alarms to be processed.

4. The method of claim 2 wherein each propagation path is characterized by minimum and maximum propagation times corresponding to the time for said signals to propagate along said propagation path.

5. The method of claim 2 wherein each propagation path is characterized by a probability value corresponding to the likelihood that said signals will propagate along said propagation path.

6. The method of claim 2 wherein each said propagation path is characterized by a minimum time for said signals to propagate along said propagation path, and where said other alarm determining step comprises the steps of:

determining each originating fault source node from which said alarms providing fault source conditions could have originated,

determining times at which said signals originated at each one of said possible originating fault source nodes,

comparing the time within which said signals, originated at a node, reach a later destination node, with the minimum propagation time along each of said propagation paths,

eliminating the paths where said minimum propagation time along the propagation path is greater than the time in which said signals reached a destination node,

tracing forward from each said originating node along said propagation paths to a said monitoring node; and

increasing a criticality parameter value of each off-line node encountered on said propagation paths from each said originating node to each said monitoring node.

7. The method of claim 2 wherein each said propagation path is characterized by a maximum time that said signals require to propagate along said propagation path, and wherein

said determining step includes the steps of determining each originating fault source node from which said alarm providing fault source conditions could have originated;

determining the times at which said signals originated at each one of said possible originating fault source nodes;

comparing the time within which said signals, originated at a node, reached a later destination node, with the maximum propagation time along each of said propagation paths;

eliminating the paths where said maximum propagation time along the propagation path is less than the time in which said signals reached a said destination node,

tracing forward from each said originating node along said propagation paths to a said monitoring node;

increasing a criticality parameter value of each off-line node encountered on said propagation paths from each said originating node to said monitoring node.

8. The method of claim 1 wherein said other alarm processing step comprises the steps of

requesting alarm signals from at least one selected other alarm, and

determining, in response to said signals from said first group of alarms and said requested signals from said selected other alarms, the potential fault sources of said conditions.

9. The method of claim 8 further comprising the step of iteratively repeating said requesting and fault source determining steps for more accurately identifying the fault sources of said conditions.

10. For use in a diagnostic system for analyzing operations of a monitored process, the system including alarms at each of a selected number of monitoring locations for providing alarm signals indicative of a function being monitored,

a real-time monitoring apparatus comprising

means for selecting a first group of alarms to be monitored,

means for processing signals from said first group of alarms for determining the occurrence of process fault conditions and for identifying potential fault sources of said conditions,

means for determining, upon the occurrence of at least one process fault condition, at least one other alarm to be selected from a group of alarms which are off-line,

means for dynamically processing signals from said first group of alarms and said at least one other alarm for determining, with expected greater accuracy, said potential fault sources, and

wherein the alarms of said first group are on-line alarms and said other alarms are initially off-line alarms.

11. The apparatus of claim 10 further comprising

means for modeling said process as a graph model having

a plurality of on-line nodes representing functions being monitored by alarms in said first group of alarms,

a plurality of off-line nodes representing functions being monitored by alarms not in said first group of alarms,

a plurality of nodes representing functions of the monitored process not having alarms associated therewith, and

a plurality of propagation paths selectively interconnecting pairs of nodes having a functional relationship.

12. The apparatus of claim 11 wherein said means for determining said at least one other alarm comprises

means for determining each originating fault source from which said alarms providing process fault condition could have originated;

means for tracing forward from each said determined originating source, using said graph model, along said propagation paths to said off-line and on-line nodes,

means for increasing a criticality parameter value of each off-line node each time it is encountered by said tracing means, and

means for selecting alarms corresponding to off-line nodes having a higher criticality parameter value as said other alarms to be processed.

13. The apparatus of claim 11 further comprising

means for characterizing each propagation path by minimum and maximum propagation times corresponding to the time for said signals to propagate along said propagation path.

14. The apparatus of claim 11 further comprising

means for characterizing each propagation by a path probability value corresponding to the likelihood that said signals will propagate along said propagation path.

15. The apparatus of claim 11 wherein each said propagation path is characterized by a minimum time for said signals to propagate along said propagation path, and where said other alarm determining means comprises:

means for determining each originating fault source node from which said alarms providing fault source conditions could have originated,

means for determining times at which said signals originated at each one of said possible originating fault source nodes,

means for comparing the time within which said signals, originated at a node, reach a later destination node, with the minimum propagation time along each of said propagation paths,

means for eliminating the paths where said minimum propagation time along the propagation path is greater than the time in which said signals reached a destination node,

means for tracing forward from each said originating node along said propagation paths to a said monitoring node; and

means for increasing a criticality parameter value of each off-line node encountered on said propagation paths from each said originating node to each said monitoring node.

16. The apparatus of claim 11 wherein each said propagation path is characterized by a maximum time that said signals require to propagate along said propagation path, and wherein said determining means comprises

means for determining each originating fault source node from which said alarm providing fault source conditions could have originated;

means for determining the times at which said signals originated at each one of said possible originating fault source nodes;

means for comparing the time within which said signals, originated at a node, reach a later destination node, with the maximum propagation time along each of said propagation paths;

means for eliminating the paths where said maximum propagation time along the propagation path is less than the time in which said signals reached a said destination node,

means for tracing forward from each said originating node along said propagation paths to a said monitoring node;

means for increasing a criticality parameter value of each off-line node encountered on said propagation paths from each said originating node to said monitoring node.

17. The apparatus of claim 10 wherein said other alarm processing means comprises

means for requesting alarm signals from at least one selected other alarm, and

means for determining, in response to said signals from said first group of alarms and said requested signals from said selected other alarms, the potential fault sources of said conditions.

18. The apparatus of claim 17 further comprising the means for iteratively repeating the operation of said requesting and fault source determining means for more accurately identifying the fault sources of said conditions.