基于电场分布法和支持向量机的复合绝缘子劣化检测方法研究
Research on degradation detection method of composite insulator based on electric field distribution method and SVM
邵九 1傅洪全 1王官健2
作者信息
- 1. 国网江苏省电力有限公司技能培训中心,江苏苏州 215004
- 2. 武汉大学电气与自动化学院,湖北武汉 430072
- 折叠
摘要
针对劣化绝缘子检测中电场分布法缺乏精确数值化算法判据的问题,提出结合支持向量机的复合绝缘子劣化检测方法.首先通过开展复合绝缘子轴向电场强度测量试验得到训练样本和测试样本,再利用交叉验证法完成惩罚参数与核函数的参数寻优,同时建立识别劣化绝缘子的相关支持向量机分类器模型,最终利用训练模型完成复合绝缘子的劣化识别.结果显示:基于电场分布法与支持向量机的劣化绝缘子检测方法能够有效识别复合绝缘子串中的劣化绝缘子,且识别准确率较高.
Abstract
Aiming at the problem of the lack of accurate numerical algorithm criteria of electric field distribution method in the detection of degraded insulators,a composite insulator degradation detection method combined with SVM(support vector machine)is proposed.Firstly,the axial electric field intensity measurement test of composite insulators is carried out to obtain training samples and test samples.Secondly,the parameters of penalty and the kernel function are optimized by K-CV(K-cross validation)method.At the same time,a correlation SVM classifier model for identifying degraded insulators is established.Finally,the training model is used to identify the degradation of composite insulators.The results show that the degraded insulator detection method based on electric field distribution method and SVM can effectively identify the degraded insulators in the composite insulator string,and the identification accuracy is high.
关键词
电场分布法/支持向量机/劣化/复合绝缘子Key words
electric field distribution method/SVM/degradation/composite insulator引用本文复制引用
出版年
2024
NETL
NSTL
万方数据