Systems and methods for classifying power line events

Patent metadata

Field	Value
Publication number	US-10422833-B2
Application number	US-201113155236-A
Country	US
Kind code	B2
Filing date	Jun 7, 2011
Priority date	Jun 7, 2010
Publication date	Sep 24, 2019
Grant date	Sep 24, 2019

Abstract

Systems and methods for classifying power line events are disclosed. Classifying power line events may include receiving measured data corresponding to a signal measured on a power line, such as proximate a substation bus or along the power line, determining from the measured data that the power line event has occurred, extracting at least one event feature from the measured data, and determining at least partially from the at least one event feature at least one probable classification for the power line event. The systems may include an Intelligent Electronic Device (IED) connected to the power line and a processor linked to the IED.

First claim

What is claimed is:

1. A computer implemented method for classifying a power line event on a power transmission or distribution system, the method comprising:
measuring, by an Intelligent Electronic Device (IED) connected to a power line of the power transmission or distribution system, first, second and third signals on the power line associated with first, second and third phases of a three phase system, respectively;
sending, by the IED, measured data corresponding to the first, second and third signals measured on the power line to at least one computer processor of a substation of the power transmission or distribution system;
receiving from the at least one computer processor at least one probable classification for the power line event, wherein the at least one computer processor is configured to:
determine from the measured data that the power line event has occurred and been cleared;
estimate from the measured data an estimated residual fault data;
estimate, using the estimated residual fault data, one or more characteristics of one or more fault clearing devices of the power line that responded to the power line event;
apply a transformation to the measured data associated with the first, second and third phases to generate a single complex signal;
isolate from the measured data transient data corresponding to the power line event based on the single complex signal generated by the transformation;
calculate at least one event feature from the isolated transient data;
project the at least one event feature onto a lower dimensional component space to obtain a feature vector for the power line event; and

determine from the feature vector the at least one probable classification for the power line event.

2. The computer implemented method of claim 1 , wherein the first, second and third signals are currents associated with first, second and third phases along the power line measured proximate a substation bus.

3. The computer implemented method of claim 1 , wherein the at least one computer processor being configured apply a transformation to the measured data associated with the first, second and third phases to generate a complex signal comprises the at least one computer processor being configured to apply Park's transformation to the measured data.

4. The computer implemented method of claim 1 , wherein the measured data is received from the IED, and the at least one computer processor is configured to:
define at least two monitoring zones for the IED; and
identify from the isolated transient data a probable one of the monitoring zones in which the power line event occurred.

5. The computer implemented method of claim 1 , wherein the power line event is a fault that occurred in at least one phase and was cleared, and the at least one computer processor is configured to:
subtract reference data from the measured data to estimate the estimated residual fault data, and wherein the one or more characteristics of the one or more fault clearing devices includes at least a range of device sizes for the one or more fault clearing devices.

6. The computer implemented method of claim 5 , wherein the at least one computer processor is configured to identify from the measured data a most probable monitoring zone in which the fault occurred.

7. The computer implemented method of claim 1 , wherein the at least one computer processor being configured to determine the at least one probable classification for the power line event comprises the at least one computer processor being configured to:
identify a plurality of event classification groups; and
determine for each one of the plurality of event classification groups a probability that the power line event belongs to the one of the plurality of event classification groups.

8. The computer implemented method of claim 1 , wherein the at least one computer processor is configured to:
identify a plurality of event classification groups;
calculate for the power line event a distance measure for each one of the plurality of event classification groups; and
determine a most probable classification for the power line event, wherein the most probable classification corresponds to an identified one of the plurality of event classification groups for which the distance measure is the smallest.

9. The computer implemented method of claim 8 , wherein the at least one computer processor is configured to:
acquire at least two classification threshold parameters based on the identified plurality of event classification groups;
indicate that the most probable classification has relatively high probability of being a correct classification for the power line event when the smallest distance measure is less than a first one of the at least two classification threshold parameters; and
indicate that the most probable classification is not the correct classification for the power line event when the smallest distance measure is greater than a second one of the at least two classification threshold parameters.

10. The computer implemented method of claim 1 , wherein the at least one computer processor being configured to determine the at least one probable classification for the power line event comprises the at least one computer processor being configured to:
identify a plurality of event classification groups;
calculate for the power line event a Mahalanobis distance for each one of the plurality of event classification groups;
identify a smallest one of the calculated Mahalanobis distances; and
determine that none of the plurality of event classification groups correspond to the probable classification for the power line event if the smallest one of the calculated Mahalanobis distances is larger than a predetermined threshold.

11. The computer implemented method of claim 1 , wherein the at least one computer processor is configured to:
retrieve tunable parameters for a plurality of power line events; and
wherein:
the at least one computer processor being configured to calculate the at least one event feature from the isolated transient data comprises the at least one computer processor being configured to calculate a spectrum for the isolated transient data; and
the at least one computer processor being configured to project the at least one event feature onto the relatively lower dimensional component space to obtain the feature vector for the power line event comprises the at least one computer processor being configured to project the spectrum onto the relatively lower dimensional component space to obtain the feature vector for the power line event.

12. The computer implemented method of claim 11 , wherein the at least one computer processor is configured to:
scale the isolated transient data, wherein the spectrum is calculated from the scaled transient data; and
subtract a mean spectrum from the spectrum calculated for the scaled transient data to obtain a mean-centered spectrum, wherein the at least one computer processor being configured to project the spectrum onto the relatively lower dimensional component space to obtain the feature vector for the power line event comprises the at least one computer processor being configured to project the mean-centered spectrum onto the relatively lower dimensional component space to obtain the feature vector for the power line event.

13. The computer implemented method of claim 11 , wherein the power line is a distribution feeder and the power line event is a distribution feeder event.

14. The computer implemented method of claim 1 , wherein the at least one computer processor is configured to estimate a probability th

Assignees

• Saarinen Kari
• Mousavi Mirrasoul
• Stoupis James
• Mcgowan John
• Abb Research Ltd

Inventors

• Saarinen Kari
• Mousavi Mirrasoul
• Stoupis James
• Mcgowan John

Classifications

• G01R31/58

  Testing of lines, cables or conductors (testing of electric windings G01R31/72) · CPC title

• G01R31/086

  in power transmission or distribution networks, i.e. with interconnected conductors · CPC title

• G01R11/25

  Arrangements for indicating or signalling faults · CPC title

• G01R31/3274 Primary

  Details related to measuring, e.g. sensing, displaying or computing; Measuring of variables related to the contact pieces, e.g. wear, position or resistance (measuring contact resistance G01R27/205) · CPC title

• Y02E60/724

  Cross-Sectional Technologies · mapped topic