Partial discharge fusion feature extraction method considering electrical feature quantity and graphic features

Abstract

The invention provides a partial discharge fusion feature extraction method considering an electrical feature quantity and a graphic feature. The method comprises the steps of combining the electricaldevice partial discharge electrical feature quantity and an image feature quantity to obtain a new complete feature vector; and performing weight evaluation on the complete feature vector, performingdimension reduction processing on the obtained complete feature vector after weight evaluation, and extracting the optimal feature quantity of partial discharge. Information loss caused by single type of characteristic quantity and single mode dimension reduction mode for characteristic quantity extraction is avoided, and the situation of misjudgment of electrical equipment partial discharge modeidentification due to insufficient characteristic quantity information is avoided.

Description

technical field [0001] The disclosure belongs to the technical field of partial discharge signal detection of electrical equipment, and in particular relates to a partial discharge fusion feature extraction method that takes electrical feature quantities and graphic features into consideration. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] During the production, manufacture and use of high-voltage electrical equipment, it is inevitable to cause burrs or insulation aging and deterioration. Partial discharge is a partial non-penetrating discharge phenomenon that occurs in the insulating medium of electrical equipment under the action of voltage. Long-term partial discharge will accelerate the insulation degradation of electrical equipment and lead to breakdown. But at the same time, partial discharge is also an important phenomenon refle...

AI Technical Summary

Problems solved by technology

[0005] For partial discharge, its electrical characteristic quantities mostly reflect its discharge intensity, discharge frequency and other information, but the description of its distribution on the spectrum diagram is not specific enough, and the information of such characteristic quantities is highly concentrated, and the details are relatively rough. Contributes to the accuracy of partial discharge pattern recognition

[0006] For the image information feature extraction method of partial discharge PRPD spectrum, that is, partial discharge phase distribution spectrum, which is generally obtained by experimental methods such as pulse current method, pattern recognition from the perspective of image processing can better describe the details of its discharge spectrum , but the characteristics of partial discharge itself are not considered enough, and there are certain limitations

Moreover, the dimensionality reduction of characteristic parameters by principal component analysis is not conducive to the investigation of the amount of information contained in each characteristic quantity. The current extraction method of partial discharge characteristic quantity of electrical equipment cannot objectively and comprehensively reflect the characteristics of partial discharge, which is easy to cause partial discharge Misjudgments in pattern recognition

Images