Power grid failure detection system and method

• US 7,490,013 B2
• Filed: 12/08/2004
• Issued: 02/10/2009
• Est. Priority Date: 12/09/2003
• Status: Active Grant

• Current Assignee: OSISoft Incorporated (Schneider Electric SE)
• Original Assignee: Oslsoft Incorporated
• Inventors: Wells, Charles H.
• Primary Examiner(s): Barbee; Manuel L
• Application Number:
  US11/008,632
• Publication Number:
  US 20050141682A1
• Time in Patent Office: 1,525 Days
• Field of Search: 702/57, 702/58, 702/60, 702/66, 702/72, 702 75- 77, 702/78, 324/76, 324/19, 324 7621- 7624, 324/76.39, 324/76.77, 324/76.79, 700/22, 700/286, 700/292, 700/293, 700/294, 700/297
• US Class Current: 702/60
• CPC Class Codes: H02J 2203/20 Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
• CPC Class Codes: H02J 3/00 Circuit arrangements for ac mains or ac distribution networks
• CPC Class Codes: Y02E 60/00 Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
• CPC Class Codes: Y04S 40/20 Information technology specific aspects, e.g. CAD, simulation, modelling, system security

Abstract

Electric power grids are usually constantly monitored for AC frequency. The monitoring of the AC frequency, however, usually does not reveal information that would indicate impending failure in the power grid. This document describes several techniques for detecting impending failure in the power grid by examining line data in the frequency domain.

Claims

1. A power grid monitoring system comprising:

at least one power line sensor, each sensor measuring frequency of alternating current carried by the power grid; and

an analysis block, the analysis block processing data gathered by the at least one sensor, the analysis block determining instability in the power grid,

where the analysis block converts sensor data from a time based domain to a frequency domain,

where the analysis block identifies peaks in the frequency domain,

where identified peaks with dampening ratios less than 0.7 indicate grid instability.

2. The system of claim 1, further comprising a control block, the control block automatically adjusting control parameters of the power grid based on the grid instability.

3. The system of claim 1, where operators are alerted to the grid instability.

4. The system of claim 1, where the at least one sensor sends data to the analysis block via an internet connection.

5. The system of claim 1, where the at least one sensor measures data at least once per second.

6. The system of claim 1, where the at least one sensor measures data at least 20 times per second.

7. The system of claim 1, where the at least one sensor measures data at least 60 times per second.

8. The system of claim 1, where the at least one sensor measures frequency with a resolution of at least 0.05 Hz.

9. The system of claim 1, where the at least one sensor measure phase.

10. The system of claim 9, where the at least one sensor measures phase with a resolution of at least 0.5 degrees.

11. The system of claim 1, where the at least one sensor synchronizes time based on GPS signals.

12. The system of claim 1, where each control area of the power grid has at least one associated sensor.

13. The system of claim 1, where the analysis block uses fast Fourier transforms for the data conversion.

14. The system of claim 1, where identified peaks lasting more than a predetermined amount of time indicate grid instability.

15. The system of claim 1, where identified peaks comprise a first peak, where the analysis block calculates a first ratio of the first peak amplitude to a first normalized amplitude, where the first ratio being at least 2.25 indicates grid instability.

16. The system of claim 15, where identified peaks further comprise a second peak, where the analysis block calculates a second ratio of the second peak amplitude to a second normalized amplitude, where the second ratio being at least 1.5 indicates grid instability.

17. The system of claim 1, where identified peaks comprise a Hopf bifurcation.

18. The system of claim 1, where identified peaks comprise a cluster of peaks.

19. The system of claim 18, where said cluster comprises:

a central peak; and

at least one side peak.

20. The system of claim 19, where the analysis block calculates a ratio of a central peak amplitude to a corresponding normalized amplitude, where the ratio being at least 1.5 indicates grid instability.

21. The system of claim 19, where the analysis block calculates a ratio of at least one side peak amplitude to at least one corresponding normalized amplitude, where the ratio being at least 1.1 indicates grid instability.

22. The system of claim 1, where identified peaks comprise a first peak and a second peak, where a measured state consists of:

a) the first peak exceeding a first limit; and

/or

b) the second peak exceeding a second limit,

where the measured state existing for a predetermined amount of time indicates grid instability.

23. The system of claim 22, where the predetermined amount of time comprises 30 minutes.

24. The system of claim 22, where the first and second limits vary with the measured duration of the predetermined amount of time.

25. The system of claim 1, where the analysis block cross-correlates frequency data gathered by the at least one sensor.

26. The system of claim 25, where frequency data with low correlation indicates grid instability.

27. The system of claim 25, where correlation ratios below 0.9 indicate grid instability.

28. The system of claim 25, where frequency correlations are calculated at least once per second.

29. The system of claim 25, where the analysis block uses a moving window of data for correlation calculations.

30. The system of claim 1, where the analysis block calculates a moving window average of absolute phase difference between at least one pair of sensors, where the average indicates grid instability if the average exceeds a predetermined threshold.

31. The system of claim 30, where moving window comprises a predetermined amount of time.

32. The system of claim 30, where moving window comprises at least 5 minutes.

33. The system of claim 30, where moving window comprises at least 10 minutes.

34. The system of claim 1, where the analysis block calculates a moving window autoregressive model on signal data.

35. The system of claim 34, where a lack of whiteness of autoregressive model residuals indicates grid instability.

36. A method of detecting instability in a power grid comprising:

collecting data from at least one sensor, wherein the data represent measured frequency of alternating current carried by the power grid;

converting the data to a frequency domain; and

identifying characteristics in the data that indicate instability, wherein identifying characteristics in the data comprises identifying a first peak in the frequency domain with an amplitude at least 2.25 times a first normalized amplitude.

37. The method of claim 36, further comprising alerting power grid operators to grid instabilities.

38. The method of claim 36, further comprising automatically adjusting power grid controls to correct grid instabilities.

39. The method of claim 36, where converting the data to a frequency domain comprises using fast Fourier transforms to convert the data.

40. The method of claim 36, where identifying characteristics in the data comprises identifying peaks in the frequency domain.

41. The method of claim 36, where identifying characteristics in the data comprises identifying a second peak in the frequency domain with an amplitude at least 1.5 times a second normalized amplitude.

42. The method of claim 36, where identifying characteristics in the data comprises:

identifying a first peak in the frequency domain with an amplitude at least 1.5 times a first normalized amplitude; and

identifying a second peak in the frequency domain with an amplitude at least 1.1 times a second normalized amplitude.

43. The method of claim 36, where identifying characteristics in the data comprises identifying peaks in the frequency domain with dampening ratios less than 0.7.

44. The method of claim 36, where identifying characteristics in the data comprises identifying sensor data with low correlation.

45. The method of claim 36, where identifying characteristics in the data comprises identifying a phase difference between a pair of sensors that exceeds a predetermined threshold.

46. The method of claim 36, where identifying characteristics in the data comprises identifying a lack of whiteness of autoregressive model residuals.

47. A power grid monitoring system comprising:

at least one power line sensor, each sensor measuring frequency of alternating current carried by the power grid; and

an analysis block, the analysis block processing data gathered by the at least one sensor, the analysis block determining instability in the power grid,

where the analysis block calculates a moving window average of absolute phase difference between at least one pair of sensors, where the average indicates grid instability if the average exceeds a predetermined threshold,

where the moving window comprises a predetermined amount of time,

where the moving window comprises at least 5 minutes.