基于深度卷积神经网络的变电一次设备故障检测方法研究

打开文本图片集

摘要：该文介绍了一种变电一次设备故障检测方法：通过不同光照环境收集一次设备的图像，创建设备数据集并进行预处理，通过深度卷积神经网络提取设备特征并加以检测。经检测，此方法能够明显降低变电一次设备故障的漏报和误报率。

关键词：卷积神经网络；变电站；一次设备；故障检测

doi：10.3969/J.ISSN.1672-7274.2024.10.010

中图分类号：TP 391.41                 文献标志码：A            文章编码：1672-7274（2024）10-00-03

Research on Fault Detection Method for Substation Primary Equipment Based on Deep Convolutional Neural Network

Abstract： This paper designs a substation primary equipment fault detection method： collecting images of the primary equipment in different lighting environments， creating equipment datasets and preprocessing them， and extracting equipment features for detection based on deep convolutional neural networks. After testing， this method can significantly reduce the missed and false alarm rates of primary equipment faults in substations.

Keywords： convolutional neural Network; substation; one device; fault detection

0   引言

变电一次设备指的是变电站系统中的电源接入、低压配电、变压配电设备等，其能够实现变电站功能。（剩余3768字）