Method and apparatus for optimizing operation of a power generating plant using artificial intelligence techniques

• US 20060178762A1
• Filed: 02/08/2005
• Published: 08/10/2006
• Est. Priority Date: 02/08/2005
• Status: Active Grant

• Current Assignee: GE Digital Holdings LLC (GE Aerospace)
• Original Assignee: Pegasus Technologies Incorporated (Evergreen Energy, Inc.)
• Inventors: Parikh, Neel J., Katrompas, Alexander M., Wroblewski, David
• Granted Patent: US 7,584,024 B2
• Time in Patent Office: Days
• US Class Current: 700/30
• CPC Class Codes: G05B 13/027 using neural networks only

Abstract

A method and apparatus for optimizing the operation of a power generating plant using artificial intelligence techniques. One or more decisions D are determined for at least one consecutive time increment, where at least one of the decisions D is associated with a discrete variable for the operation of a power plant device in the power generating plant. In an illustrated embodiment, the power plant device is a soot cleaning device associated with a boiler.

Claims

1. A method for optimizing the operation of a power generating plant having a plurality of power plant devices, the method comprising:

generating a model that models operation of the power generating plant, said model receiving a plurality of input parameters associated with operation of the power generating plant, and generating one or more output parameters in response to said plurality of input parameters, wherein each said output parameter is associated with a goal for the power generating plant;

determining for one or more consecutive time increments, one or more decisions D that realize a desired goal for the power generating plant, wherein at least one of said decisions D is associated with at least one discrete variable for operation of a power plant device based upon said model; and

operating the power generating plant in accordance with at least one of the decisions D.

2. A method according to claim 1, wherein said at least one discrete variable has a value of ON or OFF with respect to activation of a power plant device.

3. A method according to claim 1, wherein said power plant device is a soot cleaning device.

4. A method according to claim 1, wherein said method further comprises:

determining whether each decision D will violate any rules or constraints associated with operation of said power generating plant.

5. A method according to claim 4, wherein said step of determining whether each decision D will violate any rules or constraints associated with operation of said power generating plant includes:

predicting power plant parameter values for a subsequent time increment using a model.

6. A method according to claim 1, wherein said decisions D form a decision tree for one or more consecutive time increments.

7. A method according to claim 1, wherein said method further comprises:

ranking each of said decisions D, wherein said rank is associated with achievement of a desired goal for the power generating plant.

8. A method for operating a power generating plant including a plurality of power plant devices, the method comprising:

providing a model of the operation of the power generating plant;

determining at least one decision D that realizes a desired goal for the power generating plant for each of one or more time increments, wherein at least one of said decisions D is associated with at least one discrete variable for operation of a power plant device based upon said model; and

selecting at least one of said decisions D to optimize operation of the power generating plant.

9. A method according to claim 8, wherein said discrete variable has a value of ON or OFF with respect to activation of a power plant device.

10. A method according to claim 8, wherein said power plant device is a soot cleaning device.

11. A method according to claim 8, wherein said method further comprises:

determining whether each decision D will violate any rules or constraints associated with operation of said power generating plant.

12. A method according to claim 11, wherein said step of determining whether each decision D will violate any rules associated with operation of said power generating plant includes:

predicting power plant parameter values for a subsequent time increment using a model.

13. A method according to claim 8, wherein said decisions D form a decision tree for a one or more consecutive time increments.

14. A method according to claim 8, wherein said method further comprises:

ranking each of said decisions D, wherein said rank is associated with achievement of a desired goal for the power generating plant.

15. A system for optimizing the operation of a power generating plant comprised of a plurality of power plant devices, the system comprising:

means for generating a model that models operation of the power generating plant, said model receiving a plurality of input parameters associated with operation of the power generating plant, and generating one or more output parameters in response to said plurality of input parameters, wherein each said output parameter is associated with a goal for the power generating plant;

means for determining for one or more consecutive time increments, one or more decisions D that realize a desired goal for the power generating plant, wherein at least one of said decisions D is associated with at least one discrete variable for operation of a power plant device based upon said model; and

means for operating the power generating plant in accordance with at least one of the decisions D.

16. A system according to claim 15, wherein said discrete variable has a value of ON or OFF with respect to activation of a power plant device.

17. A system according to claim 15, wherein said power plant device is a soot cleaning device.

18. A system according to claim 15, wherein said system further comprises:

means for determining whether each decision D will violate any rules or constraints associated with operation of said power generating plant.

19. A system according to claim 18, wherein said system further comprises:

means for predicting power plant parameter values for a subsequent time increment using a model.

20. A system according to claim 15, wherein said decisions D form a decision tree for one or more consecutive time increments.

21. A system according to claim 15, wherein said system further comprises:

means for ranking each of said decisions D, wherein said rank is associated with achievement of a desired goal for the power generating plant.

22. A system for operating a power generating plant including a plurality of power plant devices, the system comprising:

a model of the operation of the power generating plant;

means for determining at least one decision D that realize a desired goal for the power generating plant for each of one or more time increments, wherein at least one of said decisions D is associated with at least one discrete variable for operation of a power plant device based upon said model; and

means for selecting at least one of said decisions D to optimize operation of the power generating plant.

23. A system according to claim 22, wherein said discrete variable has a value of ON or OFF with respect to activation of a power plant device.

24. A system according to claim 22, wherein said power plant device is a soot cleaning device.

25. A system according to claim 22, wherein said system further comprises:

means for determining whether each decision D will violate any rules or constraints associated with operation of said power generating plant.

26. A system according to claim 25, wherein said system further comprises:

means for predicting power plant parameter values for a subsequent time increment using a model.

27. A system according to claim 22, wherein said decisions D form a decision tree for one or more consecutive time increments.

28. A system according to claim 22, wherein said system further comprises:

means for ranking each of said decisions D, wherein said rank is associated with achievement of a desired goal for the power generating plant.

29. A method for optimizing the operation of at least one soot cleaning device group in a power generating plant, wherein each said soot cleaning device group is comprised of one or more soot cleaning devices, the method comprising:

generating a model that models operation of the power generating plant, said model receiving a plurality of input parameters associated with operation of the power generating plant, and generating one or more output parameters in response to said plurality of input parameters, wherein each said output parameter is associated with a goal for the power generating plant;

repeatedly adjusting at least one of said plurality of input parameters received by the model, to obtain output parameters that achieve a desired goal for the power generating plant;

determining one or more decisions D for activation of a soot cleaning device group that is associated with achievement of said desired goal for the power generating plant;

ranking the decisions D in accordance with a predetermined criteria; and

selecting at least one soot cleaning device group for activation at the present time in accordance with said ranked decisions D.

30. A method according to claim 29, wherein said input parameters include at least one time value parameter associated with operation of each of the at least one soot cleaning device group, and a plurality of plant parameters.

31. A method according to claim 30, wherein said at least one time value parameter is selected from the group consisting of:

(1) a DutyCycle parameter representative of the total duration of time for which soot cleaning devices in a soot cleaning device group were active within a defined time period;

(2) an OffTime parameter representative of the total duration of time for which soot cleaning devices in a soot cleaning device group were inactive;

(3) an OnTime parameter representative of the total duration of time for which soot cleaning devices in a soot cleaning device group were active;

(4) a MaxOffTime parameter representative of the maximum total duration of time for which soot cleaning devices in a soot cleaning device group may be left inactive subsequent to a soot cleaning operation; and

(5) a MinOffTime parameter representative of the minimum total duration of time for which soot cleaning devices in a soot cleaning device group should be kept inactive subsequent to a soot cleaning operation.

32. A method according to claim 30, wherein said at least one plant parameter is selected from the group consisting of:

combustion parameters, fuel parameters, post-combustion parameters, emissions parameters, performance parameters, given conditions and state indicators.

33. A method according to claim 29, wherein said output parameters of the model are selected from the group consisting of:

NOx emission, heat rate, and boiler efficiency.

34. A method according to claim 29, wherein said method further comprises the step of:

dynamically adjusting the desired goals associated with said one or more output parameters.

35. A method according to claim 29, wherein said step of ranking the decisions D in accordance with a predetermined criteria includes the steps of:

determining for each soot cleaning device group the associated DutyCycle parameter (Opt Result) associated with said adjusted input parameters;

determining for each soot cleaning device group a current DutyCycle parameter (CurrentValue); and

determining for each soot cleaning device group an optimization cycle time (OptCycleTime).

36. A method according to claim 35, wherein said step of ranking the decisions D in accordance with a predetermined criteria further comprises:

determining a Target Factor by the following expression;

Target Factor=(OptResult−

CurrentValue)/OptCycleTime.

37. A system for optimizing the operation of at least one soot cleaning device group in a power generating plant, wherein each said soot cleaning device group is comprised of one or more soot cleaning devices, the system comprising:

means for generating a model that models operation of the power generating plant, said model receiving a plurality of input parameters associated with operation of the power generating plant, and generating one or more output parameters in response to said plurality of input parameters, wherein each said output parameter is associated with a goal for the power generating plant;

an optimizer for repeatedly adjusting at least one of said plurality of input parameters received by the model, to obtain output parameters that achieve a desired goal for the power generating plant;

means for determining one or more decisions D for activation of a soot cleaning device group that is associated with achievement of said desired goal for the power generating plant;

means for ranking the decisions D in accordance with a predetermined criteria; and

means for selecting at least one soot cleaning device group for activation at the present time in accordance with said ranked decisions D.

38. A system according to claim 37, wherein said input parameters include at least one time value parameter associated with operation of each of the at least one soot cleaning device group, and a plurality of plant parameters.

39. A system according to claim 38, wherein said at least one time value parameter is selected from the group consisting of:

(1) a DutyCycle parameter representative of the total duration of time for which soot cleaning devices in a soot cleaning device group were active within a defined time period;

(2) an OffTime parameter representative of the total duration of time for which soot cleaning devices in a soot cleaning device group were inactive;

(3) an OnTime parameter representative of the total duration of time for which soot cleaning devices in a soot cleaning device group were active;

(4) a MaxOffTime parameter representative of the maximum total duration of time for which soot cleaning devices in a soot cleaning device group may be left inactive subsequent to a soot cleaning operation; and

(5) a MinOffTime parameter representative of the minimum total duration of time for which soot cleaning devices in a soot cleaning device group should be kept inactive subsequent to a soot cleaning operation.

40. A system according to claim 38, wherein said at least one plant parameter is selected from the group consisting of:

combustion parameters, fuel parameters, post-combustion parameters, emissions parameters, performance parameters, given conditions and state indicators.

41. A system according to claim 37, wherein said output parameters of the model are selected from the group consisting of:

NOx emission, heat rate, and boiler efficiency.

42. A system according to claim 37, wherein said system further comprises:

means for dynamically adjusting the desired goals associated with said one or more output parameters.

43. A system according to claim 37, wherein means for ranking the decisions D in accordance with a predetermined criteria includes:

means for determining for each soot cleaning device group the associated DutyCycle parameter (Opt Result) associated with said adjusted input parameters;

means for determining for each soot cleaning device group a current DutyCycle parameter (CurrentValue); and

means for determining for each soot cleaning device group an optimization cycle time (OptCycleTime).

44. A system according to claim 43, wherein said means for ranking the decisions D in accordance with a predetermined criteria further comprises:

means for determining a Target Factor by the following expression;

Target Factor=(OptResult−

CurrentValue)/OptCycleTime.

45. A method for optimizing the operation of at least one soot cleaning device group in a power generating plant, wherein each said soot cleaning device group is comprised of one or more soot cleaning devices, the method comprising:

generating a model that models operation of the power generating plant, said model receiving a plurality of input parameters associated with operation of the power generating plant, and generating one or more output parameters in response to said plurality of input parameters, wherein each said output parameter is associated with a goal for the power generating plant;

repeatedly adjusting at least one of said plurality of input parameters received by the model, to obtain output parameters that achieve a desired goal for the power generating plant;

determining respectively for one or more consecutive time increments, one or more decisions D for activation of a soot cleaning device group that is associated with achievement of said desired goal for the power generating plant; and

selecting at least one soot cleaning device group for activation at the present time in accordance with said decisions D.

46. A method according to claim 45, wherein said method further comprises:

determining whether the power generating plant will respond to each decision D in a manner that does not violate any rules associated with operation of said power generating plant.

47. A method according to claim 46, wherein said step of determining whether the power generating plant will respond to each decision D in a manner acceptable to operation of said power generating plant includes:

predicting power plant parameter values for a subsequent time increment using a model.

48. A method according to claim 46, wherein said decisions D form a decision tree for one or more consecutive time increments.

49. A method according to claim 48, wherein said method further comprises:

evaluating said decisions D in the decision tree to select at least one group of soot cleaning devices for activation at the present time.

50. A method according to claim 45, wherein said method further comprises:

ranking each of said decisions D, wherein said rank is associated with achievement of said desired goal for the power generating plant.

51. A system for optimizing the operation of at least one soot cleaning device group in a power generating plant, wherein each said soot cleaning device group is comprised of one or more soot cleaning devices, the system comprising:

means for generating a model that models operation of the power generating plant, said model receiving a plurality of input parameters associated with operation of the power generating plant, and generating one or more output parameters in response to said plurality of input parameters, wherein each said output parameter is associated with a goal for the power generating plant;

an optimizer for repeatedly adjusting at least one of said plurality of input parameters received by the model, to obtain output parameters that achieve a desired goal for the power generating plant;

means for respectively determining for one or more consecutive time increments, one or more decisions D for activation of a soot cleaning device group that is associated with achievement of said desired goal for the power generating plant; and

means for selecting at least one soot cleaning device group for activation at the present time in accordance with said decisions D.

52. A system according to claim 51, wherein said system further comprises:

means for determining whether the power generating plant will respond to each decision D in a manner that does not violate any rules associated with operation of said power generating plant.

53. A system according to claim 52, wherein said means determining whether the power generating plant will respond to each decision D in a manner acceptable to operation of said power generating plant includes:

a model for predicting power plant parameter values for a subsequent time increment.

54. A system according to claim 52, wherein said decisions D form a decision tree for one or more consecutive time increments.

55. A system according to claim 54, wherein said system further comprises:

means for evaluating said decisions D in the decision tree to select at least one group of soot cleaning devices for activation at the present time.

56. A system according to claim 51, wherein said system further comprises:

means for ranking each of said decisions D, wherein said rank is associated with achievement of said desired goal for the power generating plant.