Method and apparatus for detecting high-impedance faults in electrical power systems

• US 5,537,327 A
• Filed: 10/22/1993
• Issued: 07/16/1996
• Est. Priority Date: 10/22/1993
• Status: Expired due to Fees

• Current Assignee: New York State Electric & Gas Corporation (Iberdrola SA)
• Original Assignee: New York State Electric & Gas Corporation (Iberdrola SA)
• Inventors: Apostolov, Alexander P., Snow, Peter B., Bronfeld, Jefferson D.
• Primary Examiner(s): Voeltz, Emanuel T.
• Assistant Examiner(s): SHAH, KAMINI S
• Application Number:
  US08/141,308
• Time in Patent Office: 998 Days
• Field of Search: 364/483, 364/482, 364/481, 364/492, 361/93, 361/76, 361/30, 361/91-92, 324/508, 324/509, 395/907, 395/915, 395/21-23, 395/50, 395/54, 395/75, 395/77
• US Class Current: 700/293
• CPC Class Codes: G01R 31/52 Testing for short-circuits, leakage current or ground faults
• CPC Class Codes: G01R 31/58 Testing of lines, cables or conductors testing of electric windings G01R31/72
• CPC Class Codes: H02H 1/0015 Using arc detectors
• CPC Class Codes: H02H 1/0092 concerning the data processing means, e.g. expert systems, neural networks

Abstract

The present invention features a method and apparatus for detecting and enabling the clearance of high impedance faults (HIFs) in an electrical transmission or distribution system. Current in at least one phase in a distribution system is monitored in real time by sensors. Analog current signature information is then digitized for processing by a digital computer. Zero crossings are identified and current maxima and minima located. The first derivatives of the maxima and minima are computed and a modified Fast Fourier Transform (FFT) is then performed to convert time domain to frequency domain information. The transformed data is formatted and normalized and then applied to a trained neural network, which provides an output trigger signal when an HIF condition is probable. The trigger signal is made available to either a network administrator for manual intervention, or directly to switchgear to deactivate an affected portion of the network. The inventive method may be practiced using either conventional computer hardware and software or dedicated custom hardware such as a VLSI chip.

Claims

1. A method for detecting high impedance faults on a power line, the steps comprising:

a) sensing a plurality of cycles of currents attributable to a power line and generating a signal representative thereof;

b) computing a first derivative of cycle-to-cycle maxima or minima of said current;

c) transforming said derivatives into a frequency domain representation thereof, the power spectrum of which frequency domain representation being usable by means for processing information;

d) applying said power spectrum to means for processing information using artificial intelligence techniques to detect a high-impedance fault and to generate a fault signal indicative thereof; and

e) actuating a switch in response to said fault signal.

2. The method for detecting high impedance faults on a power line in accordance with claim 1, wherein said means for processing information comprises a neural network.

3. The method for detecting high impedance faults on a power line in accordance with claim 2, wherein said neural network is trained by applying signals representative of line currents indicative of a high impedance fault.

4. The method for detecting high impedance faults on a power line in accordance with claim 3, wherein said neural network training is accomplished by the use of simulated data.

5. The method for detecting high impedance faults on a power line in accordance with claim 4, wherein said simulation is based on superposition of actual data.

6. The method for detecting high impedance faults on a power line in accordance with claim 3, wherein said neural network training is accomplished by applying actual data.

7. The method for detecting high impedance faults on a power line in accordance with claim 1, wherein said transforming step (c) further comprises determining an envelope based on minimum and maximum signals and processing said envelope.

8. The method for detecting high impedance faults on a power line in accordance with claim 1, wherein said actuating step (e) results in operating a circuit clearing device.

9. The method for detecting high impedance faults on a power line in accordance with claim 8, wherein said clearing device is a clearing device disposed along said power line.

10. The method for detecting high impedance faults on a power line in accordance with claim 1, wherein said actuating step (e) results in operating an indicating device.

11. The method for detecting high impedance faults on a power line in accordance with claim 1, the steps further comprising:

f) prior to applying said power spectrum to said means for processing information, mathematically conditioning said power spectrum.

12. The method for detecting high impedance faults on a power line in accordance with claim 11, wherein step (f) comprises the steps of:

i) summing positive and negative power spectra; and

ii) averaging said sums.

13. A method for detecting high impedance faults on a power line, comprising:

a) providing a rules-driven expert system;

b) sensing voltages and/or currents attributable to a line of a power distribution system and generating a signal representative thereof;

c) computing a first derivative of cycle-to-cycle maxima or minima of said voltages and/or currents;

d) mathematically preconditioning said signal;

e) transforming said preconditioned signal representative of said line voltages and/or line currents into a frequency domain representation usable by said rules-driven expert system;

f) applying said frequency domain representation to said expert system to detect a high-impedance fault and to generate a fault signal indicative thereof; and

g) initiating an event in response to said fault signal.

14. The method for detecting high impedance faults on a line of a power distribution system in accordance with claim 13, wherein said rules-driven expert system comprises fuzzy logic.

15. The method for detecting high impedance faults on a line of a power distribution system in accordance with claim 13, wherein said transforming step (e) further comprises determining an envelope based on minimum and maximum signals and processing said envelope.

16. The method for detecting high impedance faults in a line of a power distribution system in accordance with claim 13, wherein said initiating step (g) results in operating a circuit clearing device.

17. The method for detecting high impedance faults in a line of a power distribution system in accordance with claim 13, wherein said initiating step (g) comprises the sub-step of recording the data.

18. The method for detecting high-impedance faults in a line of a power distribution system of claim 17, wherein said recording sub-step is performed on a digital fault recorder.

19. The method for detecting high impedance faults in a line of a power distribution system in accordance with claim 16, wherein said clearing device is a clearing device disposed along said line of said power distribution system.

20. The method for detecting high impedance faults on a power line in accordance with claim 13, wherein said initiating step (g) results in operating an indicating device.

21. The method for detecting high impedance faults on a power line in accordance with claim 13, the steps further comprising:

h) prior to applying said frequency domain representation to said expert system, mathematically conditioning said frequency domain representation.

22. The method for detecting high impedance faults on a power line in accordance with claim 21, wherein step (h) comprises the steps of:

i) summing positive and negative frequency domain representations; and

ii) averaging said sums.

23. A method for detecting high impedance faults on a power line, comprising:

a) sensing a plurality of cycles of current attributable to a power line and generating a signal representative thereof;

b) detecting changes in the cycle-to-cycle positive and negative peaks of the current waveform; and

c) initiating an event only when said changes are detected;

wherein the step (b) of detecting changes in the cycle-to-cycle positive and negative peaks of the current waveform comprises;

i) computing a first derivative of cycle-to-cycle maxima or minima of said current;

ii) transforming said derivatives into a frequency domain representation thereof, the power spectrum of which frequency domain representation being usable by means for processing information; and

iii) applying said power spectrum to means for processing information using artificial intelligence techniques to detect a high-impedance fault and to generate a fault signal indicative thereof.

24. The method for detecting high impedance faults on a power line in accordance with claim 23, wherein said transforming substep (ii) of detecting step (b) further comprises determining an envelope based on minimum and maximum signals and processing said envelope.

25. The method for detecting high impedance faults on a power line in accordance with claim 23, the steps further comprising:

d) prior to applying said power spectrum to said means for processing information, conditioning said power spectrum.

26. The method for detecting high impedance faults on a power line in accordance with claim 25, wherein step (d) comprises the steps of:

i) summing positive and negative power spectra; and

ii) averaging said sums.

27. A method for detecting high impedance faults on a power line, comprising:

a) sensing a plurality of cycles of voltage attributable to a power line and generating signal representative thereof;

b) detecting changes in the cycle-to-cycle positive and negative peaks of the voltage waveform; and

c) initiating an event only when said changes are detected;

wherein the step (b) of detecting changes in the cycle-to-cycle positive and negative peaks of the voltage waveform comprises;

i) computing a first derivative of cycle-to-cycle maxima or minima of said voltage;

ii) transforming said derivatives into a frequency domain representation thereof, the power spectrum of which frequency domain representation being usable by means for processing information; and

iii) applying said power spectrum to means for processing information using artificial intelligence techniques to detect a high-impedance fault and to generate a fault signal indicative thereof.

28. The method for detecting high impedance faults on a power line in accordance with claim 27, wherein said transforming substep (ii) of detecting step (b) further comprises determining an envelope based on minimum and maximum signals and processing said envelope.

29. The method for detecting high impedance faults on a power line in accordance with claim 27, wherein said initiating step (c) results in operating a circuit clearing device.

30. The method for detecting high impedance faults on a power line in accordance with claim 27, the steps further comprising:

d) prior to applying said power spectrum to said means for processing information, conditioning said power spectrum.

31. The method for detecting high impedance faults on a power line in accordance with claim 29, wherein said clearing device is a clearing device disposed along said power line.

32. The method for detecting high impedance faults on a power line in accordance with claim 30, wherein step (d) comprises the steps of:

i) summing positive and negative power spectra; and

ii) averaging said sums.