摘要
现有的输电线路故障分类方法使用的特征参数非常庞大,计算成本较高且会增加错误率,对此,提出基于联合神经网络的输电线路故障分类方法。使用卷积神经网络(CNN)模型,将故障特征以时序矩阵的形式输入卷积层和池化层,实现有效的特征提取和计算,从而简化分析过程并提高预测故障的能力。使用RNN模型对提取的特征进行进一步的学习与分析,以推断相间故障的接地类型。结果表明,所提方法准确度达到99.6%,时间为18.3s,具有较强的实用性。
The existing fault classification methods for transmission lines use very large feature parameters,high computational costs,and will increase error rates.Therefore,a joint neural network-based fault classification method for transmission lines is proposed.Using a Convolutional Neural Network(CNN)model,fault features are input into the convolutional and pooling layers in the form of a temporal matrix to achieve effective feature extraction and calculation,thereby simplifying the analysis process and improving the ability to predict faults.Further learning and analysis of the extracted features using RNN models to infer the grounding type of phase faults.The results show that the proposed method has an accuracy of 99.6%and a time of only 18.3 seconds,indicating practicality.
作者
马子劼
臧志斌
赵建伟
王佩光
马胜
张永欣
MA Zi-jie;ZANG Zhi-bin;ZHAO Jian-wei;WANG Pei-guang;MA Sheng;ZHANG Yong-xin(State Grid Siji Location Service Co.,Ltd.,Beijing 102200,China)
出处
《信息技术》
2025年第12期187-192,共6页
Information Technology
关键词
联合神经网络
特征提取
故障识别
输电线路
joint neural network
feature extraction
fault identification
transmission line
