Calibration technique for rules used with asset monitoring in industrial process control and automation systems

Patent metadata

Field	Value
Publication number	US-10657199-B2
Application number	US-201615053942-A
Country	US
Kind code	B2
Filing date	Feb 25, 2016
Priority date	Feb 25, 2016
Publication date	May 19, 2020
Grant date	May 19, 2020

Abstract

A method includes identifying a statistical performance of a monitoring rule associated with an asset monitoring system. The monitoring rule includes logic configured to identify one or more faults with at least one asset, and the statistical performance includes an effectiveness of the monitoring rule in identifying the one or more faults. The method also includes identifying an economic performance of the monitoring rule, where the economic performance is based on costs associated with different outcomes of the monitoring rule. The method further includes updating or replacing the monitoring rule based on the economic performance.

First claim

What is claimed is:

1. A method for monitoring assets in an industrial process control and automation system, the method comprising:
obtaining from an asset monitoring system, historical operational data associated with at least one asset of the asset monitoring and obtaining historical failure data associated with faults of the at least one asset;
generating alerts or notifications by executing at least one monitoring rule using the obtained historical operational data;
identifying known faults by comparing the generated alerts or notifications with the faults of the at least one asset;
identifying a statistical performance of the at least one monitoring rule associated with the asset monitoring system, the statistical performance comprising an effectiveness of the monitoring rule in identifying the known faults by identifying rates associated with different outcomes of the monitoring rule, the different outcomes including true positive outcomes, true negative outcomes, false positive outcomes, and false negative outcomes;
identifying an economic performance of the monitoring rule, by generating at least one economic cost function associated with the at least one monitoring rule, wherein the processing device executes a calibration analytics querying of a user or a database to obtain values for the costs associated with the true positive, true negative, false positive, and false negative outcomes of the monitoring rule;
identifying, by the calibration analytics, one or more tuning parameter values of the monitoring rule based on the at least one economic cost function; and
updating or replacing the monitoring rule based on the economic performance to optimize asset monitoring by varying the at least one tuning parameter value of the monitoring rule to identify which parameter value provides improved or optimized cost.

2. The method of claim 1 , wherein identifying the economic performance of the monitoring rule comprises generating a cost function based on the rates associated with the different outcomes of the monitoring rule and the costs associated with the different outcomes of the monitoring rule.

3. The method of claim 2 , wherein:
the true positive outcomes denote instances where the monitoring rule correctly identifies that the one or more faults with the at least one asset exist;
the true negative outcomes denote instances where the monitoring rule correctly identifies that the one or more faults with the at least one asset do not exist;
the false positive outcomes denote instances where the monitoring rule incorrectly identifies that the one or more faults with the at least one asset exist; and
the false negative outcomes denote instances where the monitoring rule fails to identify that the one or more faults with the at least one asset exist.

4. The method of claim 3 , wherein the cost function is expressed as:
J=C TP ×R TP +C TN ×R TN +C FP ×R FP +C FN ×R FN
wherein:
J denotes a cost of the monitoring rule;
C TP denotes a cost associated with the true positive outcomes;
R TP denotes a rate associated with the true positive outcomes;
C TN denotes a cost associated with the true negative outcomes;
R TN denotes a rate associated with the true negative outcomes;
C FP denotes a cost associated with the false positive outcomes;
R FP denotes a rate associated with the false positive outcomes;
C FN denotes a cost associated with the false negative outcomes; and
R FN denotes a rate associated with the false negative outcomes.

5. The method of claim 1 , wherein identifying the statistical performance of the monitoring rule comprises:
obtaining historical operational data;
executing the monitoring rule with the historical operational data to generate notifications; and
comparing the notifications with occurrences of known faults.

6. The method of claim 1 , wherein updating or replacing the monitoring rule comprises:
obtaining one or more tuning parameter values for the monitoring rule; and
identifying a statistical performance and an economic performance of the monitoring rule with the one or more tuning parameter values and identifying an economic benefit of using the monitoring rule with the one or more tuning parameter values.

7. The method of claim 6 , wherein updating or replacing the monitoring rule further comprises:
repeatedly obtaining the one or more tuning parameter values for the monitoring rule during multiple iterations;
for each iteration, identifying the statistical performance and the economic performance of the monitoring rule with the one or more tuning parameter values for that iteration and identifying the economic benefit of using the monitoring rule with the one or more tuning parameter values for that iteration; and
updating or replacing the monitoring rule using the one or more tuning parameter values associated with a selected one of the iterations.

8. The method of claim 7 , wherein updating or replacing the monitoring rule further comprises:
identifying the selected iteration as the iteration in which the one or more tuning parameter values provide a maximum economic benefit.

9. The method of claim 1 , wherein updating or replacing the monitoring rule comprises providing the updated or new monitoring rule to the asset monitoring system for use in monitoring the at least one asset.

10. An apparatus for monitoring assets in an industrial process control and automation system comprising:
at least one processing device configured to:
obtain from an asset monitoring system, historical operational data associated with at least one asset and obtaining historical failure data associated with faults of the at least one asset;
generate alerts or notifications by executing at least one monitoring rule using the obtained historical operational data;
identify known faults by comparing the generated alerts or notifications with the faults of the at least one asset;
identify a statistical performance of the at least one monitoring rule associated with an asset monitoring system, the statistical performance comprising an effectiveness of the monitoring rule in identifying the known faults by identifying rates associated with different outcomes of the monitoring rule, the different outcomes including true positive outcomes, true negative outcomes, false positive outcomes, and false negative outcomes;
identify an economic performance of the monitoring rule, by generating at least one economic cost function associated with the at least one monitoring rule, wherein the processing device executes a calibration analytics querying of a user or a database to obtain values for the costs associated with the true positive, true negative, false positive, and false negative outcomes of the monitoring rule;
identify one or more tuning parameter values of the monitoring rule based on the at least one economic cost function; and
update or replace the monitoring rule based on the economic performance to optimize asset monitoring by varying at least one parameter value of the monitoring rule to identify which parameter value provides improved or optimized cost.

11. The apparatus of claim 10 , wherein, to identify the economic performance of the monitoring rule, the at least one processing device is configured to generate a cost function based on the rates associated with the different outcomes of the monitoring rule and the costs associated with the different outcomes of the monitoring rule.

12. The apparatus of claim 11 , wherein:
the true positive outcomes denote instances where the monitoring rule correctly identifies that the one or more faults with the at least one asset exist

Assignees

• Honeywell Int Inc

Inventors

• Stewart Gregory E
• Germann David J
• Light Marc
• Venkatesha Sharath

Classifications

• G06Q40/12

  Accounting · CPC title

• G06Q50/04

  Manufacturing · CPC title

• G06F17/18 Primary

  for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)} · CPC title

• G05B23/0297

  Reconfiguration of monitoring system, e.g. use of virtual sensors; change monitoring method as a response to monitoring results · CPC title

• G06Q10/0639

  Performance analysis of employees; Performance analysis of enterprise or organisation operations · CPC title