Systems and methods for high impedance fault detection in electric distribution systems

Patent metadata

Field	Value
Publication number	US-11709194-B2
Application number	US-202117403650-A
Country	US
Kind code	B2
Filing date	Aug 16, 2021
Priority date	Aug 16, 2021
Publication date	Jul 25, 2023
Grant date	Jul 25, 2023

Abstract

Systems, methods, and computer-readable media are disclosed for high impedance detection in electric distribution systems. An example method may include calculating, by a processor, a relative randomness of a signal, wherein the relative randomness is a derivative of a first scale wavelet transform divided by an energy of the signal. The example method may also include calculating, by the processor, one or more scales of a wavelet transform of the signal. The example method may also include calculating, by the processor, one or more energy ratios between energy of the wavelet transform in the one or more scales. The example method may also include calculating, by the processor, a zero-crossing phase difference between a third harmonic and a fundamental component of the signal. The example method may also include determining, by the processor, that a high impedance fault occurs based on at least one of: the relative randomness, a comparison between the one or more scales of the wavelet transform, and the zero-crossing phase difference.

First claim

That which is claimed is:

1. A method comprising:
calculating, by a processor, a relative randomness of a signal, wherein the relative randomness is a derivative of a first scale wavelet transform divided by an energy of the signal;
calculating, by the processor, one or more scales of a wavelet transform of the signal;
calculating, by the processor, one or more energy ratios between energy of the wavelet transform in the one or more scales;
calculating, by the processor, a zero-crossing phase difference between a third harmonic and a fundamental component of the signal;
determining, by the processor, that a high impedance fault occurs based on at least one of: the relative randomness, a comparison between the one or more scales of the wavelet transform, and the zero-crossing phase difference;
receiving a first signal from an intelligent electronic device located at a feeder head of a distribution system; and
applying one or more filters to filter out noise and signals generated by normal operations from the first signal to generate a filtered signal;
wherein the signal comprises the filter signal.

2. The method of claim 1 , further comprising:
calculating, by the processor, a first randomness by taking a derivative of an energy of a phase current or residual current of the signal,
wherein determining that the high impedance fault occurs is further based on the first randomness.

3. The method of claim 1 , further comprising:
calculating, by the processor, a second randomness by applying a band-pass filter onto an energy of a phase current or residual current of the signal,

wherein determining that the high impedance fault occurs is further based on the second randomness.

4. The method of claim 1 , further comprising:
calculating a first ratio of a first scale of the one or more scales of the wavelet transform to a second scale of the wavelet transform; and
calculating a second ratio of the second scale of the one or more scales of the wavelet transform to a third scale of the wavelet transform,
wherein determining that the high impedance fault occurs is further based on the first ratio and the second ratio.

5. The method of claim 1 , wherein calculating the zero-crossing phase difference comprises:
calculating an odd and even harmonic ratios to the fundamental component.

6. The method of claim 1 , wherein determining that the high impedance fault occurs comprises:
determining respective weights for each of the relative randomness, the one or more scales of the wavelet transform, and the zero-crossing phase difference;
determining that each of the respective weights exceeds a threshold; and
determining that the high impedance fault occurs based on each of the respective weights exceeding the threshold.

7. The method of claim 6 , wherein determining the respective weights comprises:
training a machine learning model using online data and offline data; and
inputting the relative randomness, the one or more scales of the wavelet transform, and the zero-crossing phase difference into the machine learning model,
wherein determining that the high impedance fault occurs is based on the machine learning model.

8. The method of claim 7 , wherein the online data is obtained from a feedback of an occurrence of high impedance fault.

9. The method of claim 1 , further comprising:
sending an alarm signal indicative of an occurrence of the high impedance fault to one or more monitoring and computing devices; and
performing one or more corrective actions in response to the occurrence of the high impedance fault.

10. A system comprising:
a computer processor operable to execute a set of non-transitory computer-readable instructions; and
a memory operable to store the set of non-transitory computer-readable instructions operable to:
calculate, by a processor, a relative randomness of a signal, wherein the relative randomness is a derivative of a first scale wavelet transform divided by an energy of the signal;
calculate, by the processor, one or more scales of a wavelet transform of the signal;
calculate, by the processor, one or more energy ratios between energy of the wavelet transform in the one or more scales;
calculate, by the processor, a zero-crossing phase difference between a third harmonic and a fundamental component of the signal;
determine, by the processor, that a high impedance fault occurs based on at least one of: the relative randomness, a comparison between the one or more scales of the wavelet transform, and the zero-crossing phase difference;

receive a first signal from an intelligent electronic device located at a feeder head of a distribution system; and
apply one or more filters to filter out noise and signals generated by normal operations from the first signal to generate a filtered signal;
wherein the signal comprises the filter signal.

11. The system of claim 10 , wherein the non-transitory computer-readable instructions are further operable to:
calculate, by the processor, a first randomness by taking a derivative of an energy of a phase current or residual current of the signal,
wherein determining that the high impedance fault occurs is further based on the first randomness.

12. The system of claim 10 , wherein the non-transitory computer-readable instructions are further operable to:
calculate, by the processor, a second randomness by applying a band-pass filter onto an energy of a phase current or residual current of the signal,
wherein determining that the high impedance fault occurs is further based on the second randomness.

13. The system of claim 10 , wherein the non-transitory computer-readable instructions are further operable to:
calculate a first ratio of a first scale of the one or more scales of the wavelet transform to a second scale of the wavelet transform; and
calculate a second ratio of the second scale of the one or more scales of the wavelet transform to a third scale of the wavelet transform;
wherein determining that the high impedance fault occurs is further based on the first ratio and the second ratio.

14. The system of claim 10 , wherein calculating the zero-crossing phase difference comprises:
calculate an odd and even harmonic ratios to the fundamental component.

15. The system of claim 10 , wherein determining that the high impedance fault occurs comprises:
determine respective weights for each of the relative randomness, the one or more scales of the wavelet transform, and the zero-crossing phase difference;
determine that each of the respective weights exceeds a threshold; and
determine that the high impedance fault occurs based on each of the respective weights exceeding the threshold.

16. The system of claim 15 , wherein determining the respective weights comprises:
train a machine learning model using online data and offline data; and
input the relative randomness, the one or more scales of the wavelet transform, and the zero-crossing phase difference into the machine learning model,
wherein determining that the high impedance fault occurs is based on the machine learning model.

17. The system of claim 16 , wherein the online data is obtained from a feedback of an occurrence of a high impedance fault.

18. The system of claim 10 , wherein the non-transitory computer-readable instructions are further operable to:
send an alarm signal indicative of an occurrence of the high impedance fault to one or more monitoring and computing devices; and
perform one or more corrective actions in response to the occurrence of the high impedan

Assignees

• General Electric Technology Gmbh

Inventors

• Li Xiaoming
• Ha Hengxu
• Xiang Ling

Classifications

• G01R31/088 Primary

  Aspects of digital computing · CPC title

• G06N20/00

  Machine learning · CPC title

• H02H1/0007

  concerning the detecting means (in general G01R or other subclasses of G01; reed switches H01H71/2445) · CPC title

• H02H1/0092

  concerning the data processing means, e.g. expert systems, neural networks · CPC title

• H02H7/22

  for distribution gear, e.g. bus-bar systems; for switching devices {(detecting mechanical or electrical defects in gas-insulated switchgears H02B13/065)} · CPC title