Voltage transformer insulation fault identification method and system based on ensemble learning

Abstract

The invention relates to a voltage transformer insulation fault identification method and system based on integrated learning, and the method comprises the steps: obtaining the operation data of a voltage transformer, and constructing a first data set according to the operation data; performing cleaning and feature extraction on the first data set to obtain a second data set; screening out a plurality of transfer learning methods according to the cosine distance of the sample, and migrating the second data set by using the transfer learning methods to obtain a training set; a plurality of supervised learning methods are used for training by utilizing the training set to obtain a plurality of supervised learning models, transfer learning evaluation indexes are constructed, the weight of each supervised learning model is determined in combination with the correct rate, the weights are fused, and an integrated learning model is obtained; and identifying the fault of the to-be-tested voltage transformer by using the integrated learning model. According to the invention, a plurality of supervised learning models are fused through combination of transfer learning and ensemble learning, so that insulation faults of one or more voltage transformers of a transformer substation can be identified with high accuracy by using a small amount of data.

Description

technical field [0001] The invention belongs to the technical field of power equipment detection and deep learning, and in particular relates to a method and system for identifying an insulation fault of a voltage transformer based on integrated learning. Background technique [0002] Voltage transformer is one of the key equipment for power system information collection. As a high-voltage measurement device widely used in power system, its output voltage is the main source of power measurement, system status monitoring and relay protection in power system. It is one of the important equipments to speed up the promotion of the Energy Internet. [0003] In the long-term process, the voltage transformer is affected by environmental factors, and the components are gradually aging, and some weak points of insulation are prone to breakdown, causing the primary equipment to fault grounding, resulting in protection tripping, and affecting the stable operation of the primary equipme...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem of low accuracy of fault identification of multiple voltage transformers in the prior art, in the first aspect of the present invention, a method for identifying insulation faults of voltage transformers based on integrated learning is provided, which includes: obtaining the operating voltage transformer data, and construct a first data set based on it; the operating data includes at least secondary voltage data; the first data set is cleaned, and inter-group features and inter-phase features are extracted from the cleaned first data set, and It performs clustering to obtain a second data set containing multiple category features; uses multiple transfer learning methods to migrate the multiple second data sets to obtain multiple third data sets; uses each third data set Train a plurality of supervised learning models respectively, and determine the weight of each supervised learning model according to the accuracy rate, fuse the plurality of supervised learning models to obtain an integrated learning model; use the integrated learning model to identify the voltage to be measured Transformer failure

Images