为了提高氧化锌避雷器的故障检测精度,文章利用收敛因子非线性变化和莱维飞行策略对灰狼(Grey Wolf Optimization,GWO)算法进行改进,得到收敛性能更好的改进灰狼(Im-proved Grey Wolf Optimization,IGWO)算法,再采用IGWO算法对支持向量机(Support Vector Ma-chine,SVM)的惩罚系数和核带宽进行优化,建立基于IGWO-SVM的避雷器故障检测模型.利用氧化锌避雷器监测数据进行故障检测实例分析,将IGWO-SVM模型的故障检测结果与现有避雷器故障检测模型的检测结果对比,结果表明,IGWO-SVM模型的检测精度更高,验证了该模型在氧化锌避雷器故障检测方面的优越性.
Fault Detection of Zinc Oxide Arrester Based on IGWO-SVM
In order to improve the fault detection accuracy of zinc oxide arrester,this paper improves the grey wolf optimization(GWO)algorithm using nonlinear changes in convergence factors and Levi's flight strategy.The improved grey wolf optimization(IGWO)algorithm with better convergence performance is obtained.The IGWO algorithm is used to optimize the penalty coefficient and kernel bandwidth of support vector machine(SVM),and a lightning arrester fault detection model based on IGWO-SVM is estab-lished.The monitoring data from zinc oxide arrester are applied for fault detection case analysis,and the fault detection results of the IGWO-SVM model are compared with the results of existing arrester fault de-tection models.The results show that the detection accuracy of IGWO-SVM model is higher,and the supe-riority of the model in zinc oxide arrester fault detection is verified.
NETL
NSTL
万方数据
