电力勘测设计Issue(12)：37-43,7.DOI:10.13500/j.dlkcsj.issn1671-9913.2023.12.007
基于WOA-GA-GRNN神经网络的输电导线脱冰跳跃高度预测
Prediction of Ice-Shedding Jump Height of Transmission Lines Based on WOA-GA-GRNN Neural Network
摘要
Abstract
The ice-shedding process causes the conductor to vibrate vertically,resulting in a reduction in the distance between the conductors.This reduction can cause the phase-to-phase distance to be less than the insulation distance,leading to serious electrical accidents such as phase-to-phase flashovers and line tripping,which significantly compromise the safe operation of the transmission line.To accurately determine the highest possible jump height after ice-shedding,we utilize a combination of the whale algorithm(WOA)and genetic algorithm(GA)to optimize key parameters of the generalized regression neural network(GRNN).By taking the number of splits,gear distance,ice cover thickness,ice-shedding rate,and conductor type as inputs,we develop a ice-shedding jump height prediction model with maximum jump height after ice-shedding as the output.The datasets for model training and testing have been produced via finite element analysis.The evaluation index method is employed to assess the accuracy.The prediction model displays the least average relative error and a superior fitting effect.It also ensures higher accuracy when calculating the ice-shedding jump height as compared to the results yielded by the design regulation formula and the engineering simplified formula.The model can derive the maximum ice-shedding jump height more accurately and efficiently.This development supports the design of measures for transmission line disaster prevention and mitigation.关键词
线路导线/脱冰跳跃高度/鲸鱼算法/遗传算法/广义回归神经网络Key words
line conductor/ice-shedding jump height/whale algorithm/genetic algorithm/generalized regression neural network分类
动力与电气工程引用本文复制引用
蔡德成,王岭,柏晓路,汪峰,王艳君,胡守松..基于WOA-GA-GRNN神经网络的输电导线脱冰跳跃高度预测[J].电力勘测设计,2023,(12):37-43,7.基金项目
中国电力工程顾问集团有限公司科技项目"输电线路不均匀脱冰与断线动力效应研究"(40-2B-KY201819-D101). （40-2B-KY201819-D101）
