摘要
通过输电线路接地网腐蚀试验获得44样本数据,对样本数据进行主成分分析,根据计算结果确定pH值、Cl^(-)含量和NO_(3)^(-)含水量为输电线路接地网腐蚀预测模型的输入量。针对狼群算法存在的缺陷,对探狼游走行为和猛狼运动步长进行改进,得到改进狼群算法。利用改进狼群算法对最小二乘支持向量机进行优化,建立基于IWPA-LSSVM的输电线路接地网腐蚀预测模型。采用IWPA-LSSVM预测模型对线路接地网腐蚀程度进行预测,结果表明,IWPA-LSSVM预测模型的平均相对误差、全局最大相对误差和均方根误差分别为3.47%、4.35%和0.0164,均优于其他预测模型,验证了提出的输电线路接地网腐蚀预测方法的正确性和实用性。
Data on 44 samples were obtained from the corrosion test of transmission line grounding grid,and principal component analysis was conducted on the data.According to the calculation results,the pH value,Cl^(-) content,NO_(3)^(-) content and water content were determined as the input of the corrosion model of transmission line grounding grid.In view of the defects of wolf swarm algorithm,the walking behavior of wolf detection and the step size of fierce wolf were improved to obtain the improved wolf swarm algorithm.An improved wolf swarm algorithm was used to optimize the least squares support vector machine(LSSVM),and a corrosion prediction model of transmission line grounding grid based on IWPA-LSSVM was established.The results show that the average relative error,global maximum relative error and root mean square error of IWPA-LSSVM prediction model are 3.47%,4.35%and 0.0164 respectively,which are better than other prediction models.The correctness and practicability of the proposed method are verified.
作者
林肖斐
夏圣峰
吴簪麟
江南
张润昊
LIN Xiaofei;XIA Shengfeng;WU Zanlin;JIANG Nan;ZHANG Runhao(State Grid Fuzhou Power Supply Company,Fuzhou 350000,China)
出处
《电瓷避雷器》
CAS
北大核心
2022年第1期1-7,共7页
Insulators and Surge Arresters
基金
国家自然科学基金项目(编号:61876097)。
关键词
改进狼群算法
最小二乘支持向量机
输电线路
接地网
腐蚀
improved wolf pack algorithm
least squares support vector machine
transmission line
grounding grid
corrosion
