Method for implementing indirect controller

• US 20040249480A1
• Filed: 06/05/2003
• Published: 12/09/2004
• Est. Priority Date: 06/05/2003
• Status: Active Grant

• Current Assignee: General Electric Company
• Original Assignee: NeuCo Incorporated (General Electric Company)
• Inventors: Lefebvre, W. Curt, Kohn, Daniel W.
• Granted Patent: US 7,194,320 B2
• Time in Patent Office: Days
• US Class Current: 700/31
• CPC Class Codes: G05B 17/02 electric
• CPC Class Codes: G05B 7/02 electric
• CPC Class Codes: G06F 18/24133 Distances to prototypes
• CPC Class Codes: G06V 10/82 using neural networks
• CPC Class Codes: Y10S 706/903 Control
• CPC Class Codes: Y10S 706/906 Process plant
• CPC Class Codes: Y10S 706/914 Process plant

Abstract

A system and method for implementing an indirect controller for a plant. A plant can be provided with both a direct controller and an indirect controller with a system model or a committee of system models. When the system model has sufficient integrity to satisfy the plant requirements, i.e., when the system model has been sufficiently trained, the indirect controller with the system model is automatically enabled to replace the direct controller. When the performance falls, the direct controller can automatically assume operation of the plant, preferably maintaining operation in a control region suitable for generating additional training data for the system model. Alternatively, the system model incorporates a committee of models. Various types of sources for errors in the committee of models can be detected and used to implement strategies to improve the quality of the committee.

Claims

1. A computer-implemented method for implementing an indirect controller with a system model in a plant operable for performing a process, the plant having a plurality of inputs, the inputs being manipulable by a controller to operate the plant, a particular combination of inputs received by the plant defining an operating region, comprising:

using the indirect controller to operate the plant;

detecting that the performance of the system model is low;

using a direct controller to operate the plant by maintaining operation around a particular operating region corresponding to detecting that the performance of the system model is low;

adapting the system model;

detecting that the performance of the system model is acceptable; and

selecting the indirect controller to operate the plant.

2. The method of claim 1, wherein fuzzy logic is used for detecting that the performance of the system model is low or detecting that the performance of the system model is acceptable.

3. The method of claim 1, wherein a threshold is used for detecting that the performance of the system model is low or detecting that the performance of the system model is acceptable.

4. The method of claim 1, wherein an overall error is used for detecting that the performance of the system model is low or detecting that the performance of the system model is acceptable.

5. The method of claim 1, wherein a local error is used for detecting that the performance of the system model is low or detecting that the performance of the system model is acceptable.

6. The method of claim 1, wherein maintaining operation around a particular operating region is coordinated with operating constraints, including satisfying a plant performance objective.

7. The method of claim 1, wherein the system model is implemented as a neural network.

8. The method of claim 1, wherein the system model is implemented using genetic programming.

9. A computer-implemented method for implementing an indirect controller with a committee of system models in a plant operable for performing a process, the plant having a plurality of inputs, the inputs being adjustable according to control settings received from a controller, comprising:

selecting the indirect controller to operate the plant;

obtaining a set of possible control settings from each system model in the committee that is operative;

evaluating confidence in the committee of models using the sets of possible control settings; and

if the confidence in the committee is acceptable, generating a set of desired control settings from the sets of possible control settings.

10. The method of claim 9, wherein evaluating confidence in the sets of possible control settings includes determining a confidence interval for the sets of possible control settings.

11. The method of claim 10, wherein evaluating confidence in the sets of possible control settings includes determining a confidence ratio for the sets of possible control settings and comparing the confidence ratio to a threshold.

12. The method of claim 9, if the confidence in the committee is not acceptable, further comprising selecting an alternate controller to operate the plant.

13. The method of claim 12, further comprising adapting the system models in the committee and shifting control to the direct controller when the performance of the committee is not acceptable.

14. The method of claim 13, wherein adapting the system models in the committee includes generating a new system model and including it in the committee.

15. The method of claim 12, wherein the alternate controller is a direct controller.

16. The method of claim 15, further including the direct controller operating the plant around an operating region corresponding to where the confidence in the committee is not acceptable.

17. The method of claim 16, further comprising adapting the system models in the committee and shifting control to the indirect controller when the performance of the committee is acceptable.

18. The method of claim 9, wherein generating a set of desired control settings includes combining the sets of possible control settings.

19. The method of claim 18, wherein combining the sets of possible control settings includes averaging the sets of possible control settings.

20. The method of claim 9, wherein fuzzy logic is used to determine whether the confidence is acceptable or not acceptable.

21. The method of claim 9, wherein at least one of the system models in the committee is a neural network.

22. The method of claim 9, wherein at least one of the system models in the committee is a genetically programmed model.

23. The method of claim 9, further comprising, if the confidence in the committee is not acceptable, detecting whether the system models in the committee have systematic errors or random errors and implementing a strategy for correction dependent on whether systematic errors or random errors are detected.

24. The method of claim 9, further comprising selecting a subset of the system models in the committee to be operative, the subset comprising the system models in the committee having the best performance.

25. A computer program product, residing on a computer readable medium, for use in implementing an indirect controller with a system model in a plant operable for performing a process, the plant having a plurality of inputs, the inputs being manipulable by a controller to operate the plant, a particular combination of inputs received by the plant defining an operating region, the computer program product comprising instructions for causing a computer to:

select the indirect controller to operate the plant;

detect that the performance of the system model is low;

select a direct controller to operate the plant by maintaining around a particular operating region corresponding to detecting that the performance of the system model is low;

adapt the system model;

detect that the performance of the system model is acceptable; and

select the indirect controller to operate the plant.

26. The computer program product of claim 25, wherein the instruction cause the computer to detect that the performance of the system model is low using fuzzy logic.

27. The computer program product of claim 25, wherein the instruction cause the computer to detect that the performance of the system model is low using a threshold of acceptable performance.

28. The computer program product of claim 25, wherein the instructions cause the computer to detect that the performance of the system model is low using an overall error.

29. The computer program product of claim 25, wherein the instructions cause the computer to detect that the performance of the system model is low using a local error.

30. The computer program product of claim 25, wherein the instructions cause the computer to instruct the controller to maintain operation around a particular operating region coordinated with operating constraints, including satisfying a plant performance objective.

37. The computer program product of claim 29, further comprising instructions for causing the computer to combine the sets of possible control settings to generate the set of desired control settings.

31. A computer program product, residing on a computer readable medium, for use in implementing an indirect controller with a committee of system models in a plant operable for performing a process, the plant having a plurality of inputs, the inputs being adjustable according to control settings received from a controller, the computer program product comprising instructions for causing a computer to:

select the indirect controller to operate the plant;

obtain a set of possible control settings from each system model in the committee that is operative;

evaluate confidence in the committee of models using the set of possible control settings; and

if the confidence in the committee is acceptable, generate a set of desired control settings from the set of possible control settings.

32. The computer program product of claim 31, wherein the instructions cause the computer to determine a confidence interval for the sets of possible control settings.

33. The computer program product of claim 32, wherein the instructions cause the computer to determine a confidence ratio for the sets of possible control settings.

34. The computer program product of claim 31, further comprising instructions for causing the computer to, if the confidence in the committee is not acceptable, select an alternate controller to operate the plant.

35. The computer program product of claim 34, further comprising instructions for causing the computer to instruct the alternate controller to operate the plant around an operating region corresponding to where the confidence in the committee is not acceptable.

36. The computer program product of claim 32, further comprising instructions for causing the computer to adapt the system models in the committee and shift control to the indirect controller when the performance of the committee is acceptable.

38. The computer program product of claim 35, further comprising instructions for causing the computer to average the sets of possible control settings.

39. The computer program product of claim 35, further comprising instructions for causing the computer to detect whether the system models in the committee have systematic errors or random errors and implement a strategy for correction dependent on whether systematic errors or random errors are detected.

40. A system for implementing indirect control in a plant operable for performing a process, the plant having a plurality of inputs subject to control to operate the plant, a particular combination of inputs received by the plant defining an operating region, comprising:

means for modeling the operation of the plant;

means for providing indirect control of the plant;

means for providing direct control of the plant, including means for maintaining operation around a particular operating region;

means for evaluating the performance of the means for modeling the operation of the plant;

means for selecting indirect control or direct control to operate the plant; and

means for adapting the means for modeling the operation of the plant.

41. A system for implementing indirect control in a plant operable for performing a process, the plant having a plurality of inputs, the inputs being adjustable according to control settings received from a means for control, comprising:

means for modeling the operation of the plant to generate a plurality of sets of possible control settings;

means for evaluating confidence in the plurality of sets of possible control settings;

means for providing indirect control to operate the plant using the sets of possible control settings;

means for providing direct control to operate the plant; and

means for selecting indirect control or direct control to operate the plant.