Statistical pattern analysis methods of partial discharge measurements in high voltage insulation

• US 6,192,317 B1
• Filed: 03/06/2000
• Issued: 02/20/2001
• Est. Priority Date: 04/11/1997
• Status: Expired due to Fees

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Krahn, John Raymond, Yazici, Birsen
• Primary Examiner(s): Hoff, Marc S.
• Assistant Examiner(s): MILLER, CRAIG S
• Application Number:
  US09/517,088
• Time in Patent Office: 351 Days
• Field of Search: 702/57-59, 702/81-84, 702/179-185, 706/21, 706/47, 706/52, 324/512, 324/522, 324/536, 324/551, 324/772
• US Class Current: 702/58
• CPC Class Codes: G01R 31/1227 of components, parts or materials G01R31/1209, G01R31/1218, G01R31/18 take precedence; circuits therefor G01R31/14; testing vessels of electrodes G01R31/16
• CPC Class Codes: G01R 31/1272 of cable, line or wire insulation, e.g. using partial discharge measurements locating faults in cables G01R31/083

Abstract

Statistical methods of partial discharge analysis utilizing histogram similarity measures are provided by quality assessment and condition monitoring methods of evaluating high voltage electrical insulation. The quality assessment method is utilized to evaluate the quality of insulation within the electrical equipment. The condition monitoring method is utilized during normal operation of the equipment to identify degradation in the insulation and to predict catastrophic insulation failures.

Claims

1. A statistical method of condition monitoring the insulation in high voltage equipment utilizing partial discharge measurements, comprising the following steps:

calculating at least one analysis feature of undetermined insulation based on a plurality of partial discharge measurements;

estimating trend parameters of said at least one analysis feature based on a linear trend line generated based on a plurality of previously determined ones of said analysis feature;

predicting an estimated analysis feature statistical distance of said analysis feature with respect to said trend line;

comparing said predicted analysis feature statistical distance with an actual analysis feature statistical distance so as to determine whether the condition of said undetermined insulation has changed.

2. The statistical method of condition monitoring as recited in claim 1 further comprising the steps of:

conditioning said partial discharge measurements for computer manipulation;

generating a plurality of partial discharge measurements from said high voltage equipment with undetermined insulation so as to calculate at least one said analysis feature of said undetermined insulation.

3. The statistical method of condition monitoring as recited in claim 2 further comprising the steps of estimating said trend parameters of said analysis features based on a parabolic trend line generated in correspondence with previously determined ones of said analysis features.

4. The statistical method of condition monitoring as recited in claim 1 further comprising the steps of estimating said trend parameters of said analysis features based on a parabolic trend line generated in correspondence with previously determined ones of said analysis features.

5. The statistical method of condition monitoring as recited in claim 1 further comprising the steps of estimating said trend parameters of said analysis features based on a log linear-trend line generated in correspondence with previously determined ones of said analysis features.

6. The statistical method of condition monitoring as recited in claim 1 further comprising the step of generating a change in condition signal when the measured statistical distance is outside a normal operating limit.

7. The statistical method of condition monitoring as recited in claim 1 further comprising the steps of:

synchronizing the phase of said partial discharge measurements over a single cycle of a supply voltage;

normalizing the partial discharge magnitude of said partial discharge measurements; and

removing excitation noise from said partial discharge measurements.

8. The statistical method of condition monitoring as recited in claim 7 further comprising the step of taking the logarithm of said partial discharge magnitudes to minimize the low magnitude high partial discharge event count.

9. The statistical method of condition monitoring as recited in claim 8 further comprising the steps of comparing the latest calculated said trend parameters with the previously calculated ones of said trend parameters.

10. The statistical method of condition monitoring as recited in claim 9 further comprising the step of generating a pre-catastrophic insulation signal when the difference in a plurality of previous ones of said trend parameters and current said trend parameters are greater than a catastrophic confidence interval value.