A dry-type transformer life prediction method and system based on a deep residual network

CN122286129APending Publication Date: 2026-06-26FUZHOU INNOVATION ELECTRONICS SCIE & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
FUZHOU INNOVATION ELECTRONICS SCIE & TECH
Filing Date
2026-02-27
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing methods for predicting the lifespan of dry-type transformers lack the ability to fully integrate multi-dimensional state information and efficiently extract deep degradation features, resulting in insufficient accuracy and generalization ability of the prediction results, which cannot meet the reliability requirements of power systems.

Method used

A deep residual network-based approach is adopted to construct a multi-dimensional time-series feature matrix by acquiring multi-source heterogeneous state data, and then train it using a one-dimensional convolutional deep residual network model to predict the loss of dry-type transformers.

Benefits of technology

It enables accurate loss prediction of dry-type transformers, improves prediction accuracy and process efficiency, ensures the continuity and reliability of prediction results, and facilitates engineering applications.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122286129A_ABST
    Figure CN122286129A_ABST
Patent Text Reader

Abstract

This invention relates to the field of transformer life prediction technology, and particularly to a method and system for predicting the life of dry-type transformers based on deep residual networks. First, multi-source heterogeneous state data of the dry-type transformer to be predicted during operation is acquired. This multi-source heterogeneous state data includes electrical parameter data, vibration information data, temperature information data, ambient temperature data, core grounding current data, and partial discharge data. Next, a multi-dimensional time-series feature matrix is ​​obtained based on the multi-source heterogeneous state data. Then, a deep residual network model based on one-dimensional convolution is constructed, and the multi-dimensional time-series feature matrix is ​​used as training data to train the deep residual network model, resulting in a trained life prediction model. Finally, the multi-dimensional time-series feature matrix is ​​input into the life prediction model to obtain the loss status of the dry-type transformer to be predicted, thereby achieving accurate prediction of the dry-type transformer's loss status.
Need to check novelty before this filing date? Find Prior Art