System and method for detecting an abnormal situation associated with a process gain of a control loop

• US 7,660,701 B2
• Filed: 06/09/2005
• Issued: 02/09/2010
• Est. Priority Date: 06/12/2004
• Status: Active Grant

• Current Assignee: Fisher-Rosemount Systems Incorporated (Emerson Electric Company)
• Original Assignee: Fisher-Rosemount Systems Incorporated (Emerson Electric Company)
• Inventors: Sharpe, Joseph H. Jr.
• Primary Examiner(s): Feliciano; Eliseo Ramos
• Assistant Examiner(s): Suglo; Janet L
• Application Number:
  US11/149,001
• Publication Number:
  US 20060020423A1
• Time in Patent Office: 1,706 Days
• Field of Search: 702/183
• US Class Current: 702/183
• CPC Class Codes: C10G 11/187 Controlling or regulating
• CPC Class Codes: G05B 13/024 in which a parameter or coefficient is automatically adjusted to optimise the performance
• CPC Class Codes: G05B 23/0254 based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks

Abstract

In a method for monitoring a control loop in a process plant, process gain data associated with a control loop may be collected. The collected process gain data may be used to determine an expected process gain behavior. For example, expected values of a process variable for given values of a load variable may be determined. As another example, expected changes in a process variable for given changes in a load variable may be determined. Then, during operation of the control loop, the process gain may be monitored. If the monitored process gain substantially deviates from the expected behavior, this may indicate an abnormal situation associated with the control loop.

Claims

1. A method for monitoring operation of a control loop in a process plant, comprising:

collecting, at one or more of a plurality of computing devices, process gain data associated with a first operating region of a control loop in a process plant, the control loop associated with a unit operation in the process plant, wherein a load variable varies in the first operating region;

determining, at one of the plurality of computing devices, an expected process gain behavior in the first operating region based on the collected process gain data, wherein an expected value of the process gain varies as the load variable varies;

wherein determining the expected process gain behavior in the first operating region comprises fitting a curve to the collected process gain data, and wherein the curve corresponds to the expected process gain behavior in the first operating region;

monitoring, at one of the plurality of computing devices, the process gain during operation of the control loop in the first operating region;

determining, at one of the plurality of computing devices, when the monitored process gain substantially deviates from the curve corresponding to the expected process gain behavior in the first operating region;

determining, at one of the plurality of computing devices, an abnormal situation associated with at least one of the control loop or the unit operation based at least on a substantial deviation from the curve corresponding to the expected process gain behavior in the first operating region; and

generating, at one of the plurality of computing devices, an abnormal situation indicator when an abnormal situation is determined.

2. A method according to claim 1, wherein collecting process gain data associated with the first operating region of the control loop comprises at least one of:

collecting data regarding a process variable versus the load variable;

collecting data regarding the load variable versus the process variable;

collecting data regarding a process gain versus the load variable;

collecting data regarding the load variable versus a process gain;

or

collecting data regarding the process variable versus one or more other process variables.

3. A method according to claim 2, wherein the load variable comprises at least one of a control output or another process variable.

4. A method according to claim 1, wherein determining the expected process gain behavior in the first operating region comprises at least one of:

determining expected values of a variable;

or

determining expected rates of change of the variable.

5. A method according to claim 1, wherein determining the abnormal situation comprises determining an abnormal situation associated with the unit operation based at least on a substantial deviation from the curve corresponding to the expected process gain behavior in the first operating region of the control loop and a substantial deviation from an expected process gain behavior of a different control loop associated with the unit operation.

6. A method according to claim 1, further comprising:

collecting, at one of the plurality of computing devices, process gain data associated with at least a second operating region of the control loop in the process plant;

determining, at one of the plurality of computing devices, an expected process gain behavior in the at least the second operating region based on the collected process gain data;

wherein determining the expected process gain behavior in the at least the second operating region comprises fitting another curve to the collected process gain data associated with the at least a second operating region, and wherein the other curve corresponds to the expected process gain behavior in the at least the second operating region;

monitoring, at one of the plurality of computing devices, the process gain during operation of the control loop in the at least the second operating region;

determining, at one of the plurality of computing devices, when the monitored process gain substantially deviates from the curve corresponding to the expected process gain behavior in the at least the second operating region; and

wherein determining the abnormal situation comprises determining the abnormal situation based at least on a substantial deviation from the curve corresponding to the expected process gain behavior in the first operating region or a substantial deviation from the curve corresponding to the expected process gain behavior in the second operating region.

7. A method according to claim 1, wherein collecting process gain data associated with the first operating region of the control loop comprises collecting process gain data regarding a plurality of process gains associated with the control loop, the plurality of process gains including at least a first process gain and a second process gain;

wherein determining the expected process gain behavior in the first operating region comprises determining an expected behavior of the first process gain with respect to at least the second process gain;

wherein the curve corresponding to the expected process gain behavior in the first operating region corresponds to the expected behavior of the first process gain with respect to at least the second process gain;

wherein monitoring the process gain during operation of the control loop in the first operating region comprises monitoring the first process gain and monitoring the second process gain;

wherein determining when the monitored process gain substantially deviates from the curve corresponding to the expected process gain behavior in the first operating region comprises determining when the monitored first process gain substantially deviates from the expected behavior of the first process gain with respect to at least the second process gain.

8. A method according to claim 1, wherein determining when the monitored process gain substantially deviates from the curve corresponding to the expected process gain behavior in the first operating region comprises at least one of:

determining when the monitored process gain is below the curve corresponding to the expected process gain behavior in the first operating region for a specified period of time;

or

determining when the monitored process gain is above the curve corresponding to the expected process gain behavior in the first operating region for the specified period of time.

9. A method according to claim 1, wherein determining the expected process gain behavior in the first operating region comprises determining a confidence interval for the first operating region;

wherein determining when the monitored process gain substantially deviates from the expected process gain behavior in the first operating region comprises determining at least when the monitored process gain is outside of the confidence interval in the first operating region.

10. A method according to claim 9, wherein determining when the monitored process gain substantially deviates from the expected process gain behavior in the first operating region comprises determining when the monitored process gain is outside of the confidence interval in the first operating region for a specified period of time.

11. A method according to claim 1, if an abnormal situation is determined, further comprising at least one of:

adjusting, at one of the plurality of computing devices, a control parameter associated with the control loop;

initiating, at one of the plurality of computing devices, a diagnostic procedure;

or

shutting down, via one or more computing devices, equipment associated with the control loop.

12. A method according to claim 1, wherein generating an abnormal situation indicator comprises generating an alert.

13. A method according to claim 1, wherein determining the abnormal situation comprises determining at least one of:

if the abnormal situation has occurred, or

if the abnormal situation will likely occur in the future.

14. A method according to claim 1, further comprising:

determining, at one of the plurality of computing devices, when the control loop is operating in a second operating region for which process gain data has not yet been collected; and

collecting, at one of the plurality of computing devices, process gain data associated with the second operating region of the control loop in the process plant after determining that the control loop is operating in the second operating region of the control loop;

monitoring, at one of the plurality of computing devices, the process gain during operation of the control loop in the second operating region; and

determining, at one of the plurality of computing devices, when the monitored process gain substantially deviates from the expected process gain behavior in the second operating region.

15. A method according to claim 14, further comprising;

prompting, at one of the plurality of computing devices, an operator whether to collect process gain data associated with the second operating region of the control loop in the process plant after determining that the control loop is operating the second operating region of the control loop;

wherein collecting process gain data associated with the second operating region of the control loop comprises process gain data associated with the second operating region if the operator indicates that process gain data associated with the second operating region of the control loop should be collected.

16. A method according to claim 14, wherein a unit of the process plant comprises the control loop;

wherein determining when the control loop is operating in the second operating region comprises determining when the unit of the process plant is operating in an operating region for which process gain data associated with the unit has not yet been collected.

17. A tangible medium having stored thereon machine executable instructions, the machine executable instructions capable of causing one or more machines to:

collect process gain data associated with a first operating region of a control loop in a process plant, wherein a load variable varies in the first operating region;

determine an expected process gain behavior in the first operating region based on the collected process gain data, wherein an expected value of the process gain varies as the load variable varies;

wherein determining the expected process gain behavior in the first operating region comprises fitting a curve to the collected process gain data, and wherein the curve corresponds to the expected process gain behavior in the first operating region;

monitor the process gain during operation of the control loop in the first operating region;

determine when the monitored process gain substantially deviates from the curve corresponding to the expected process gain behavior in the first operating region;

determine an abnormal situation associated with at least one of the control loop or the unit operation based at least on a substantial deviation from the curve corresponding to the expected process gain behavior in the first operating region; and

generate an abnormal situation indicator if the abnormal situation is determined.

18. A method for monitoring operation of a control loop in a process plant, comprising:

collecting process gain data associated with a first operating region of a control loop in a process plant, wherein a load variable varies in the first operating region;

determining an expected process gain behavior in the first operating region based on the collected process gain data associated with the first operating region, wherein an expected value of the process gain varies as the load variable varies;

wherein determining the expected process gain behavior in the first operating region comprises fitting a curve to the collected process gain data, and wherein the curve corresponds to the expected process gain behavior in the first operating region;

providing data indicative of the curve corresponding to the expected process gain behavior in the first operating region to an expert engine;

providing process gain data associated with the control loop during operation of the control loop to the expert engine;

utilizing the expert engine to detect an abnormal situation associated with the control loop based on the data indicative of the curve corresponding to the expected process gain behavior in the first operating region and the process gain data associated with the control loop during operation of the control loop; and

generating an abnormal situation indicator if the abnormal situation is determined.

19. A method according to claim 18, wherein utilizing the expert engine to detect the abnormal situation associated with the control loop comprises determining whether the process gain in the first operating region substantially deviates from the curve corresponding to the expected process gain behavior in the first operating region.

20. A method according to claim 18, further comprising:

providing process variable statistical data associated with the control loop during operation of the control loop to the expert engine; and

wherein utilizing the expert engine to detect the abnormal situation associated with the control loop comprises utilizing the expert engine to detect the abnormal situation associated with the control loop further based on the process variable statistical data associated with the control loop.

21. A method according to claim 20, wherein the process variable statistical data associated with the control loop comprises statistical data generated by field devices associated with the control loop.

22. A system for monitoring operation of a control loop in a process plant, the system comprising:

a process gain signature generator configured to generate a signature of expected process gain behavior associated with a control loop in a process plant, wherein the signature is indicative of a gain of a process variable of the control loop versus a load variable of the control loop, and wherein the gain of the process variable versus the load variable is expected to vary as the load variable varies;

wherein the process gain signature generator is configured to fit a curve to collected process gain data, and wherein the signature of expected process gain behavior comprises the curve;

a process gain evaluator configured to determine if an actual process gain substantially deviates from the curve; and

an abnormal situation detector configured to detect an abnormal situation associated with a process unit associated with the control loop based at least in part on whether the actual process gain substantially deviates from the curve and to generate an abnormal situation indicator if the abnormal situation is detected.

23. A system according to claim 22, further comprising:

an interval generator configured to generate an interval associated with the signature of expected process gain behavior;

wherein the process gain evaluator is configured to determine if the actual process gain substantially deviates from the curve based on the interval associated with the signature of expected process gain behavior.

24. A system according to claim 22, further comprising an expert system, wherein the expert system comprises at least one of the process gain evaluator or the abnormal situation detector.

25. A system according to claim 24, wherein the expert system is configured to detect an abnormal situation associated with a process unit associated with the control loop based at least in part on whether the actual process gain substantially deviates from the curve.

26. A system according to claim 22, further comprising a process gain data collector configured to collect data to be used by the process gain signature generator to generate the signature of expected process gain behavior associated with the control loop.

27. A system according to claim 22, wherein the abnormal situation detector is configured to detect at least one of:

if the abnormal situation has occurred, or

if the abnormal situation will likely occur in the future.

28. A method for facilitating monitoring operation of at least a portion of a process plant, comprising:

collecting, at one of a plurality of computing devices, process gain data indicative of respective process gains associated with respective unit operations in a process plant, wherein the respective process gains are associated with an operating region, wherein respective load variables vary in the operating region;

determining, at one of the plurality of computing devices, expected process gains associated with respective unit operations based on the collected process gain data, wherein expected values of the respective process gains vary as the respective load variables vary;

wherein determining the expected process gains associated with respective unit operations comprises fitting respective curves to the collected process gain data, and wherein the respective curves correspond to respective unit operations;

providing, at one of the plurality of computing devices, a common set of criteria for determining for each unit operation an abnormal situation associated with the unit operation based at least on whether the process gain associated with the unit operation substantially deviates from the curve associated with the unit operation;

permitting, at one of the plurality of computing devices, a user to modify the common set of criteria for a particular unit operation to generate a modified set of criteria;

utilizing, at one of the plurality of computing devices, the modified set of criteria to determine for the particular unit operation an abnormal situation associated with the particular unit operation;

utilizing, at one of the plurality of computing devices, the common set of criteria to determine for each of at least one other unit operation an abnormal situation associated with the other unit operation; and

generating, at one of the plurality of computing devices, an abnormal situation indicator when an abnormal situation for a unit operation is determined.

29. A method according to claim 28, wherein the common set of criteria comprises expert rules to be applied by an expert engine.

30. A method for monitoring operation of a control loop in a process plant, comprising:

collecting, at one of the plurality of computing devices, process gain data associated with an operating region of a control loop in a process plant, the control loop associated with a unit operation in the process plant, wherein the process gain data is indicative of a gain of a process variable of the control loop versus a load variable of the control loop, and wherein the gain of the process variable versus the load variable varies as the load variable varies;

determining, at one of the plurality of computing devices, an expected process gain behavior in the operating region based on the collected process gain data, wherein the process gain is expected to vary as the load variable varies;

wherein determining the expected process gain behavior in the operating region comprises fitting a curve to the collected process gain data, and wherein the curve corresponds to the expected process gain behavior in the operating region;

monitoring, at one of the plurality of computing devices, the process gain during operation of the control loop in the operating region;

determining, at one of the plurality of computing devices, when the monitored process gain substantially deviates from the curve;

determining, at one of the plurality of computing devices, an abnormal situation associated with at least one of the control loop or the unit operation based at least on a substantial deviation from the curve; and

generating, at one of the plurality of computing devices, an abnormal situation indicator when an abnormal situation is determined.

31. A method for facilitating monitoring operation of at least a portion of a process plant, comprising:

collecting, at one of a plurality of computing devices, process gain data indicative of process gains associated with respective unit operations in a process plant, wherein each of the process gains is a gain of a process variable of the respective unit versus a load variable of the respective unit, and wherein each of the process gains varies as the respective load variable varies;

determining, at one of the plurality of computing devices, expected process gains associated with respective unit operations based on the collected process gain data, wherein each of the expected process gains varies as the respective load variable varies; and

wherein determining the expected process gains associated with respective unit operations comprises fitting respective curves to the collected process gain data, and wherein the respective curves correspond to respective unit operations;

providing, at one of the plurality of computing devices, a common set of criteria for determining for each unit operation an abnormal situation associated with the unit operation based at least on whether the process gain associated with the unit operation substantially deviates from the curve associated with the unit operation.