基于Double-RNN模型的电力设备故障定位与检测  

Fault location and detection of power equipment based on Double-RNN model

作　　者：李光华 赵小明[1] 何亚东 LI Guang-hua;ZHAO Xiao-ming;HE Ya-dong(Collega of Electrical Engineering Sichuan Univercity,Chengdu 610000,China;Electrical and New Energy Faculty of China Three Gorges Univercity,Yichang 443000,Hubei Province,China)

机构地区：[1]四川大学电气工程学院,成都610000 [2]三峡大学电气与新能源学院,湖北宜昌443000

出　　处：《信息技术》2025年第3期144-150,共7页Information Technology

摘　　要：研究设计了一个四个步骤的热故障诊断模型:数据预处理、特征提取、设备热故障识别和设备热故障定位。Double-RNN网络模型罗列出数据处理过程并对电力设备的热故障识别进行了深入研究,该模型主要由两层RNN网络构成,分别用于数据特征抽取和类别预测。结果显示,在平均精度上,Double-RNN为90.06%,优于其他算法,并在检测速度上表现出了优异的实时性。由此可见,Double-RNN网络控制模型在电力设备故障定位与检测方面显示出了优越的性能表现,可以在电力设备的故障定位与检测中被广泛采用。A four step thermal fault diagnosis model is designed,including data preprocessing,feature extraction,equipment thermal fault recognition,and equipment thermal fault localization.The Double RNN network model lists the data processing process and conducts in-depth research on thermal fault identification of power equipment.The model is mainly composed of two layers of RNN networks,which are used for data feature extraction and category prediction.The results show that in terms of average accuracy,Double RNN is 90.06%better than other algorithms,and exhibits excellent real-time performance in detection speed.Therefore,it can be seen that the Double RNN network control model has shown superior performance in fault location and detection of power equipment,and is worthy of widespread adoption in fault location and detection of power equipment.

关 键 词：Double-RNN 电力设备 热故障 故障检测 

分 类 号：TM63[电气工程—电力系统及自动化]