A dynamic detection method for distribution transformer faults based on aoelm

Abstract

The invention relates to an AOELM-based dynamic detection method for distribution transformer faults. This method collects the experimental monitoring data of the distribution transformer under normal state and fault state operation, and stores the offline training sample set after data normalization and labeling preprocessing, and then uses the extreme learning machine algorithm to establish The initial transformer fault detection model, and then, use the support vector data description method combined with the detection results of the initial transformer fault detection model to construct a new training sample set, so as to dynamically update the initial transformer fault detection model, and then solve the existing The transformer fault static detection method does not have the ability of active online dynamic update, and cannot adapt to the data distribution drift characteristics in the time-varying process. While ensuring the accuracy of fault online detection, it can further reduce the fault false alarm rate.

Description

technical field [0001] The invention relates to the field of fault detection of power transformers in distribution networks, in particular to a dynamic detection method for faults of low-voltage transformers in distribution networks based on an active online extreme learning machine (AOELM). Background technique [0002] As the core equipment used to realize the conversion and distribution of electric energy in the modern power grid, the health status of the power transformer directly affects the normal operation of the power system. Among them, the working environment of the distribution network transformer is complex and prone to failure. Therefore, it is very important for the safe and stable operation of the power system to timely detect transformer faults online and arrange staff to repair them as soon as possible. [0003] With the development of modern sensor technology and computer storage technology, a large amount of data containing transformer working condition i...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a dynamic detection method for distribution transformer faults based on AOELM to solve the problem that existing static detection methods for transformer faults do not have active online dynamic update capabilities and cannot adapt to the data distribution drift characteristics in the time-varying process problems, and further reduce the rate of fault false positives while ensuring the accuracy of fault online detection

Images