摘要
覆冰脱落后导线上下振动会减小导线之间的相间距离,严重时可能造成相间距离小于绝缘间隙,引起输电线路发生相间闪络、跳闸等电气事故,严重影响输电线路安全运行。为准确得到脱冰后的最大跳跃高度,采用鲸鱼算法(WOA)混合遗传算法(GA)对广义回归神经网络(GRNN)参数进行优化,以分裂数、档距、覆冰厚度、脱冰率和导线型号作为参数输入,脱冰后最大跳跃高度作为输出,构建脱冰跳跃高度预测模型。通过有限元分析生成模型训练数据集和测试集,并采用评价指标法评估其准确性。与设计规程公式和工程简化公式的计算结果相比,该预测模型的平均相对误差最小,拟合效果更优,脱冰跳跃高度的计算准确度更高。该模型可以较准确、快速地得出最大脱冰跳跃高度,为输电线路防灾减灾设计提供支撑。
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.
作者
蔡德成
王岭
柏晓路
汪峰
王艳君
胡守松
CAI Decheng;WANG Ling;BAI Xiaolu;WANG Feng;WANG Yanjun;HU Shousong(HVDC Technology Department of CEEC Engineering Institute,Wuhan 430071,China;Central Southern China Electric Power Design Institute Co.,Ltd.of China Power Engineering Consulting Group,Wuhan 430071,China;Wuhan Power Supply Company of State Grid Hubei Electric Power Co.,Ltd.,Wuhan 430013,China;China Three Gorges University,Yichang 443002,China)
出处
《电力勘测设计》
2023年第12期37-43,共7页
Electric Power Survey & Design
基金
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)。
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
line conductor
ice-shedding jump height
whale algorithm
genetic algorithm
generalized regression neural network
