基于改进深度学习的配电网故障辨识系统研究  

Research on Distribution Network Fault Identification System Based on Improved Deep Learning

作　　者：沈雪红 SHEN Xuehong(College of Electronic and Communication Engineering,Zhejiang Technical College of Posts and Telecom,Shaoxing,Zhejiang 312366)

机构地区：[1]浙江邮电职业技术学院电子与通信工程学院,浙江绍兴312366

出　　处：《绵阳师范学院学报》2024年第8期93-97,共5页Journal of Mianyang Teachers＇ College

基　　金：浙江省教育厅科研项目(Y202250907)。

摘　　要：针对配电网运行过程中设备故障辨识难度大、处理不及时等问题,通过将长期记忆引入到深度学习模型,提出了一种改进深度学习模型,设计了基于改进深度学习的配电网故障辨识系统.该系统包括数据源、接口层、数据库、数据服务、计算层和信息层,实现对配电网故障的准确辨识预警.实验结果表明:提出的改进深度学习模型与现有方法相比,故障预测精度能够达到94.19%,优于现有模型,具有良好的应用价值.In response to the problems of difficulty in identifying equipment faults and delayed processing dur-ing the operation of distribution networks,an improved deep learning model was proposed by introducing long-term memory into the deep learning model.A distribution network fault identification system based on improved deep learning was designed.The system included a data source,interface layer,database,data service,computing layer,and information layer to achieve accurate identification and warning of distribution network faults.The experimental results showed that the proposed improved deep learning model had a fault prediction accuracy of 94.19%com-pared to existing methods,which is superior to existing models.This study provides guidance for distribution net-work fault identification and warning,and shows its application value.

关 键 词：配电网 长期记忆 改进深度学习 故障辨识 

分 类 号：TM732[电气工程—电力系统及自动化]