Transient detector and fault classifier for a power distribution system

• US 8,336,352 B2
• Filed: 01/25/2010
• Issued: 12/25/2012
• Est. Priority Date: 01/25/2010
• Status: Active Grant

• Current Assignee: Aclara Technologies LLC (Hubbell, Inc.)
• Original Assignee: Aclara Power-Line Systems Incorporated (Hubbell, Inc.)
• Inventors: Abiprojo, Priotomo, Gregg, Gordon, Rieken, David
• Primary Examiner(s): COSIMANO, EDWARD R
• Application Number:
  US12/692,919
• Publication Number:
  US 20110184671A1
• Time in Patent Office: 1,065 Days
• Field of Search: 734/321, 738/658, 738/663, 324/72, 324/76.11, 324/76.12, 324/76.38, 324/102, 340/500, 340/540, 340/635, 340/657, 340/659, 340/660, 340/664, 340/870.01, 340/870.07, 702/1, 702/57, 702/58, 702/64, 702/65, 702/66, 702/69, 702/127, 702/182, 702/183, 702/187, 702/189, 708/100, 708/105, 708/131, 708/160, 708/200
• US Class Current: 72/69
• CPC Class Codes: G01R 31/086 in power transmission or distribution networks, i.e. with interconnected conductors
• CPC Class Codes: G01R 31/50 Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections testing of sparking plugs H01T13/58
• CPC Class Codes: H02H 1/04 Arrangements for preventing response to transient abnormal conditions, e.g. to lightning or to short duration over voltage or oscillations; Damping the influence of dc component by short circuits in ac networks
• CPC Class Codes: H02H 3/50 responsive to the appearance of abnormal wave forms, e.g. ac in dc installations
• CPC Class Codes: Y04S 10/52 Outage or fault management, e.g. fault detection or location

Abstract

Detection of fault occurrences within a power distribution system. A transient detector and fault classification system (10) implements an algorithm that detects transients which may result from the occurrence of a fault in the power distribution system. The system includes a detection module (18) which processes a set samples obtained from electrical waveforms propagated of through the power distribution system and appear to be statistically anomalous compared to other sample data. This is done using an adaptive detection algorithm that is applied when large changes occur in a waveform over a relatively short period of time. The identified samples are then provided to a signal classifier module (20) which processes sets of samples to classify a transient they represent as a likely fault occurrence or some other type of anomaly which is likely not a fault occurrence. If a transient is classified as representing a likely fault occurrence, a polling module (30) polls users of the distribution system to determine if a fault has occurred within the distribution system, and, if so, where.

Claims

1. In an electrical power distribution system, a method for detecting the occurrence of a fault comprising:

monitoring waveforms on the distribution system;

sampling the current of the waveforms and processing the samples obtained to detect any anomalies which may indicate a fault has occurred within the distribution system, the processing including detecting any transients within a set of waveform samples by identifying anomalous samples contained within the set relative to a statistical pattern of recent waveform samples; and

,

classifying samples meeting pre-established criteria indicative of a fault to then classify the result of a sampling as representing a likely fault occurrence or some other type of anomaly which is likely not a fault occurrence, the classification including comparing transients against a predetermined set of criteria to determine which transients result from faults rather than from other events within the distribution system, and features used for classifying a transient as a fault including identifying a peak current magnitude of those samples within a transient and measuring how much the magnitude of the current changes between the peak magnitude and the resulting magnitude after the transient has settled.

2. The method of claim 1 in which the sampling includes sampling the current on each feeder phase of the distribution system at a distribution substation of the system.

3. The method of claim 1 in which set of samples includes a current sample, at least one previously obtained sample, and at least one incoming sample.

4. The method of claim 1 further including classifying an anomaly as a feeder-switch event rather than a fault.

5. The method of claim 1 in which if a sample is classified as representing a likely fault occurrence, polling users of the distribution system to determine if a fault has occurred within the distribution system, and, if so, where.

6. The method of claim 1 in which the features used for classifying a transient as a fault further include the duration of the transient.

7. The method of claim 6 in which the method further includes establishing a classification for whether or not a transient represents an actual fault based upon selected values for each of the classifying features.

8. A transient detector and classification system for analyzing transients occurring in waveforms propagated through a power distribution system to determine if a transient is indicative of a fault somewhere in the system that may result in a power outage comprising:

a detection means monitoring the waveforms and obtaining continuous samples of the waveforms;

processor means processing the samples to determine if a transient occurred that may represent the signature of a fault occurrence, the processing means detecting any transients within the waveform samples by identifying anomalous samples relative to a statistical pattern of recent waveform samples; and

,

a classifier evaluating a set of samples representing the transient in accordance with predetermined criteria to classify the transient as representing a likely fault occurrence or some other type of anomaly which is likely not a fault occurrence, the classifier comparing each transient against a predetermined set of criteria to determine which transients result from faults rather than transients resulting from other events within the distribution system, the classifier classifying a transient as a fault based upon identifying a peak current magnitude of those samples within the transient and by a measure of how much the magnitude of the current changes between the peak magnitude and the resulting magnitude after the transient has settled.

9. The system of claim 8 in which the detection means samples the current of waveforms on each feeder phase of the distribution system at the distribution substation of the system.

10. The classifier of claim 8 in which the classifier further classifies a transient as a fault further based upon the duration of the transient.

11. The system of claim 8 in which the classifier further classifies an anomaly as a feeder-switch event rather than a fault.

12. A transient detector and classification system for analyzing transients occurring in waveforms propagated through a power distribution system comprising:

a detector means monitoring waveforms for each feeder phase of the power distribution system, at a distribution substation of the system, and obtaining samples of transients appearing within the waveforms to produce a set of samples of a transient;

a processor means processing the samples to determine if a transient has occurred that may represent the signature of a fault occurrence, the processing means detecting any transients within the waveform samples by identifying anomalous samples relative to a statistical pattern of recent waveform samples; and

,

a signal classifier evaluating each set of samples to classify a transient as one resulting from normal occurrences within the distribution system or one resulting from an anomaly occurring within the distribution system, the classifier performing its classification based upon preselected features exhibited by the samples contained within the set of samples including comparing each transient against a predetermined set of criteria to determine which transients result from faults rather than transients resulting from other events within the distribution system, and the utility performing an action based upon the result of the classification if the classification indicates the transient resulted from an anomalous occurrence.

13. The system of claim 12 in which if the classifier classifies the transient as indicative of a fault, at least a portion of the power distribution system is polled to determine if the fault resulted in an outage;

and, if so, the extent thereof.

14. The system of claim 12 in which if the classifier classifies the transient as indicative of a feeder-switch event, a message is given to the utility that a certain portion of their power distribution network may have been reconfigured.

15. A transient detector and classification system for analyzing transients occurring in waveforms propagated through a power distribution system comprising:

a detector means monitoring waveforms for each feeder phase of the power distribution system, at a distribution substation of the system, and obtaining samples of transients appearing within the waveforms to produce a set of samples of a transient;

a signal classifier evaluating each set of samples to classify a transient as one resulting from normal occurrences within the distribution system or one resulting from an anomaly occurring within the distribution system, the classifier performing its classification based upon preselected features exhibited by the samples contained within the set of samples, and the utility performing an action based upon the result of the classification if the classification indicates the transient resulted from an anomalous occurrence; and

,

if the classifier classifies the transient as indicative of a fault, at least a portion of the power distribution system is polled to determine if the fault resulted in an outage; and

, if so, the extent thereof.