摘要
人工和无人机航拍读取避雷器动作次数不仅费时费力,且实施难度大,为了经济、快速高效的实现带间隙避雷器动作次数的在线计算。文中提出一种根据雷电定位系统、输电线路数据、防雷机理与GA-BP神经网络技术相结合的带间隙避雷器动作次数计算方法。首先,计算线路引雷区与落在引雷区内的最大雷电流;然后,确定GA-BP神经网络结构,利用训练好的网络判断雷击后避雷器是否动作。实例表明,该方法对反击雷致避雷器动作识别的准确率达到90%,对雷击后避雷器不动作样本判断的准确率为86.25%,对绕击雷致避雷器动作的准确率总体为80%,计算累加的避雷器动作次与实际动作次数基本相吻合,故该方法可较准确的实现避雷器动作次数的在线计算。
It is not only time-consuming and laborious to read the action times of arrester by artificial and unmanned aerial photographs,but also difficult to implement.In order to realize the online calculation of action times of arrester with clearance economically,quickly and efficiently.This paper presents a calculation method of action times of lightning arrester with clearance based on lightning location system,transmission line data,lightning protection mechanism and GA-BP neural network technology.Firstly,the maximum lightning current in the lightninginduced area and in the lightning-induced area is calculated.Then,the structure of GA-BP neural network is determined,and the trained network is used to judge whether the arrester acts after lightning strike.Examples show that the accuracy rate of action recognition of lightning-induced arresters by this method reaches 90%,the accuracy rate of non-action samples after lightning stroke is 86.25%,and the overall accuracy rate of action of lightning-induced arresters by shielding stroke is 80%.The calculated accumulative action times of lightning arresters are basically consistent with the actual action times,so this method can be compared.Accurate on-line calculation of the number of arrester actions is realized.
作者
邓红雷
唐崇旺
刘刚
张莉彬
DENG Honglei;TANG Chongwang;LIU Gang;ZHANG Libin(School of Electric Power,South China University of Technology,Guangzhou 510640,China)
出处
《高压电器》
CAS
CSCD
北大核心
2021年第6期189-195,202,共8页
High Voltage Apparatus
基金
国家自然科学基金项目(51777079)。
关键词
带间隙避雷器
雷电流参数
引雷区
GA-BP神经网络
动作次数
lightning arrester with gap
lightning current parameters
lightning guidance area
GA-BP neural network
operation number
