摘要
针对线路舞动形态特征识别定位误差大、舞动强度误差大和微风环境下监测位置误差大的问题,提出基于迁移学习AlexNet的关键输电线路舞动形态特征辨识技术。结合GPS定位和伪距差分法完成对关键输电线路舞动形态特征的精确辨识。经比较实验验证,所提方法振动曲线与实际振动位置相符,振动强度曲线峰值对应的频率与实验设置的振动频率相同;得到的舞动位置最接近实际振动位置。所提方法定位误差小、定位舞动强度误差小和微风环境下监测位置误差小,有利于关键输电线路舞动形态特征的辨识。
Aiming at the problems of large positioning error,large galloping intensity error and large monitoring position error in the breeze environment,a key transmission lines galloping shape features identification technology based on transfer learning AlexNet is proposed.The precise identification of galloping characteristics of key transmission lines is completed by combining GPS positioning and pseudo range difference method.The comparison test shows that the vibration curve of the proposed method is consistent with the actual vibration position.The frequency corresponding to the peak value of the vibration intensity curve is the same as the vibration frequency set in the experiment.The obtained vibration position is closest to the actual vibration position.The proposed method has small positioning error,small positioning galloping intensity error and small monitoring position error in breeze environment which is conducive to the identification of galloping morphological characteristics of key transmission lines.
作者
刘洋
LIU Yang(Jilin Power Supply Company of State Grid Jilin Electric Power Co.,Ltd.,Jilin 132001,China)
出处
《微型电脑应用》
2025年第3期247-250,共4页
Microcomputer Applications
关键词
迁移学习
AlexNet神经网络
输电线路舞动
形态特征
辨识技术
transfer learning
AlexNet neural network
transmission lines galloping
morphological feature
identification technology
