Partial discharge detection model training method and device, and detection information determination method and device

Abstract

The embodiment of the invention discloses a partial discharge detection model training method and device and a detection information determination method and device. The partial discharge detection model training method comprises the steps of: obtaining reference data which has corresponding first label information; determining a selected data set corresponding to the reference data; according to the first label information corresponding to the reference data, determining second label information corresponding to each piece of selected data in the selected data set; determining target data based on the reference data, the first label information, each piece of selected data and the second label information; training the partial discharge detection model based on the target data to obtain a trained partial discharge detection model. Based on the embodiment of the invention, the quantity of training data can be increased, the workload of manual labeling can be reduced, detection data obtained by detecting a power distribution network through multiple sensors can be effectively fused, and the influence of missing historical detection data on training of the partial discharge detection model is reduced.

Description

technical field [0001] The invention relates to the technical field of power distribution network detection, in particular to a training method for a partial discharge detection model, a method and a device for determining detection information. Background technique [0002] In the power distribution network, the generation of partial discharge is inevitable. If there is partial discharge in the equipment for a long time, it will accelerate the aging of the equipment, which will lead to the failure of the power distribution network. The traditional method of detecting partial discharge is mainly through manual detection by inspectors, but it is difficult to detect slight partial discharge by this manual detection method. With the continuous development of science and technology, a single method for detecting partial discharge, such as ultrasonic waves, ground waves, and infrared rays, has been proposed. However, the detection accuracy of this single method for detecting part...

AI Technical Summary

Problems solved by technology

[0004] (1) The deep neural network model requires a large amount of labeled detection data as training data, and the workload of labeling detection data is huge;

[0005] (2) When training the deep neural network model, it is necessary to effectively fuse the detection data collected by different sensors, and the fusion effect is difficult to guarantee;

[0006] (3) For the historical detection data collected by multiple sensors, there is often no detection data collected at the same time, which brings certain difficulties to the training of the deep neural network model

Images