Partial discharge defect type identification method and device

Abstract

The invention provides a partial discharge defect type identification method, which comprises the steps of A, acquiring partial discharge signals; B, extracting discharge statistical features of the partial discharge signals; C, calculating a sampling partition proportions; D, enabling the discharge statistical features and each sampling partition proportion to act as input of a first neural network classifier and a second neural network classifier, wherein the first neural network classifier and the second neural network classifier respectively output a discharge defect type and a confidence degree thereof; and E, comprehensively judging a final discharge defect type and a confidence degree thereof. The partial discharge defect type identification method not only considers statistical features of multiple partial discharge pulse signals, but also considers multi-peak and vibration conditions in the single partial discharge pulse signal, and comprehensively judges the final discharge defect type according to certain rules by means of the neural network classifiers, thereby solving adverse impacts imposed on identification by multiple peaks and waveform oscillation, and improving the accuracy and the reliability of the final identification result.

Description

technical field [0001] The invention relates to a method and device for identifying partial discharge defect types. Background technique [0002] Faults and defects may occur during the operation of the insulation structure of high-voltage power equipment, which will lead to insulation breakdown and failure of the entire power equipment after long-term operation, thereby affecting the reliability of the power system. Therefore, it is necessary to identify the insulation defects of high-voltage power equipment and assessment, while partial discharge detection has proven to be an effective means of revealing defects in high-voltage equipment and assessing their severity. [0003] The form of partial discharge signal is a single or continuous electrical pulse. The pulse waveform not only contains the information of the discharge mechanism of the defect, but also contains the information of the severity of the defect. Therefore, one of the main tasks of partial discharge detecti...

AI Technical Summary

Problems solved by technology

However, due to the complexity of the partial discharge mechanism, it has not been fully studied so far. In particular, under a certain voltage, defects will not only appear single-peak pulses, but also multi-peak pulses and oscillation pulses. The timing of appearance is random. Difficulty in identifying defect types

[0005] The statistical identification method based on phase resolution has been widely used in the identification technology of defect types. However, because this method is mainly based on the distribution of PD pulses in the power frequency phase, it is greatly affected by interference and often leads to wrong identification. As a result, it is impossible to accurately determine the severity of the defect

However, the identification method based on the characteristic parameters of the partial discharge pulse waveform (such as rise time, etc.), whether it is a time-domain method or a frequency-domain method, is more effective for single-peak pulses, but less effective for multi-peak or oscillating pulses.

Images