摘要
文章提出了一种基于深度学习的继电保护故障识别方法。利用自编码器(Autoencoder)对原始电气信号进行特征提取,压缩冗余信息并突出关键故障特征,构建BP神经网络分类模型,对提取后的特征向量进行多类故障状态识别。通过在仿真实验平台上构建电力系统故障数据集,开展训练与验证实验,结果显示文章方法表现出较高的识别准确率,最高可达98.6%,结合深度特征提取与神经网络分类器可显著提升继电保护系统的智能化水平,为实现电网故障智能诊断与快速响应提供了技术支撑。
This paper proposes a deep learning-based fault classification method for protective relaying.An autoencoder is employed to extract features from raw electrical signals,effectively compressing redundant information while emphasizing critical fault characteristics.Subsequently,a backpropagation(BP)neural network classifier is constructed to classify multiple fault types based on the extracted feature vectors.A power system fault dataset was generated on a simulation platform,and training/validation experiments were conducted.Results demonstrate that the proposed method achieves high classification accuracy,reaching up to 98.6%.The integration of deep feature extraction with a neural network classifier significantly enhances the intelligent capabilities of protective relaying systems,offering robust technical support for intelligent fault diagnosis and rapid response in power grids.
作者
蒋冰洁
Jiang Bingjie(Hongze Power Supply Branch,State Grid Jiangsu Electric Power Co.,Ltd.,Huaian 223001,China)
出处
《办公自动化》
2025年第23期64-66,共3页
Office Informatization
关键词
继电保护
故障识别
自编码器
BP神经网络
relay protection
fault identification
autoencoder
BP neural network
