摘要
故障检测对于确保电力系统正常运行具有重要意义。近年来随着深度学习在目标识别领域的重大进展,基于深度学习的电力输电线故障目标检测逐渐成为电力系统故障目标检测领域的研究热点。目前深度学习领域的主流算法包括两阶段目标检测算法和单阶段目标检测算法。本文对此两类算法进行简要介绍,分析其中具有代表性的几种卷积神经网络算法的优缺点,并总结电力系统故障目标检测存在的问题,以及未来的发展方向。
Fault detection is of great significance to ensure the normal operation of the power system.In recent years,with the great development of deep learning in the filed of target recognition,fault target detection of power transmission lines based on deep learning has gradually become a research hot spot in the field of fault target detection in power system.At present,the mainstream algorithms in the field of deep learning include two-stage and single-stage target detection algorithms.This paper briefly introduces the two types of algorithms,analyzes the advantages and disadvantages of several representative convolutional neural network algorithms,and summarizes the problems as well as the direction of future development of fault target detection in power system.
作者
杨天宇
孙新娟
YANG Tianyu;SUN Xinjuan(School of Electronic Engineering of North China University of Water Resources and Electric Power,Zhengzhou Henan 450046,P.R.China)
出处
《重庆电力高等专科学校学报》
2023年第1期1-4,23,共5页
Journal of Chongqing Electric Power College
关键词
深度学习
目标检测
故障检测
电力系统
deep learning
target detection
fault detection
power system
