Transformer fault diagnosis device and method based on decision tree and random forest

Abstract

The invention discloses a transformer fault diagnosis device and method based on a decision tree and a random forest. The device comprises a serial data input interfaces, a decision tree continuous machine learning model implementation module, a random forest continuous machine learning model implementation module, a preferred data output selector and a prediction data output module. According tothe invention, a training result model packaged by an algorithm is packaged in a module in a hardware form, and a multi-core CPU is selected as hardware support, so that the system operation efficiency is improved; and the preferred data output selector is used for periodically retraining the two models according to the model output result, so that adaptive model adjustment of different transformers is realized, and continuous learning and self-improvement of the learning model are achieved.

Description

technical field [0001] The invention belongs to the technical field of transformer fault diagnosis, and in particular relates to a transformer fault diagnosis device and method based on decision trees and random forests. Background technique [0002] The power transformer is one of the most important equipment in the power system, and its operating condition is directly related to the safe and stable operation of the power system. Transformer state assessment and its sub-topic fault detection based on time series data have always been key research topics in the industry, and are of great significance to the stable operation of the power grid. [0003] With the development of the Energy Internet, traditional evaluation schemes and single machine learning transformer evaluation schemes have been initially explored. Although these methods have certain effects, there are problems such as low accuracy of fault judgment and failure to update in time for changes in the external env...

AI Technical Summary

Problems solved by technology

[0003] With the development of the Energy Internet, traditional evaluation schemes and single machine learning transformer evaluation schemes have been initially explored. Although these methods have certain effects, there are problems such as low accuracy of fault judgment and failure to update in time for changes in the external environment.

In recent years, scholars have proposed some higher-level machine learning algorithms to upgrade the diagnostic method, but there are still some shortcomings, such as: the fuzzy clustering method is not ideal for large-scale sample classification; SVM is also for large-scale Training samples are difficult to apply, etc.

Images