基于支持向量机高压静电场闪络信号的模糊识别  

Fuzzy Recognition for Flashover Signal in High Voltage Electrostatic Field Based on Support Vector Machine

作　　者：张瑾[1] 潘必超[1,2] 

机构地区：[1]广东机电职业技术学院信息工程学院,广东广州510550 [2]华南理工大学机械与汽车工程学院,广东广州510640

出　　处：《电网与清洁能源》2015年第8期1-5,共5页Power System and Clean Energy

基　　金：国家自然科学基金项目(5117518)~~

摘　　要：电场本体参数和静电闪络信号特征存在差异性,给控制器进行闪络信号识别带来了困难,传统识别方法误判率高,灵活性和鲁棒性较差。引入高斯分布噪声的虚拟训练集,结合实验电场实际采样信号构成复合训练集对闪络信号识别的支持向量机进行训练,解决训练集不均衡问题,并将得到的决策模型用于实际电场信号的实时识别。仿真与实际电场信号测试结果表明,所提方法具有运算速度快,准确率高,且具有很强的泛化能力。It is very difficult to identify a flashover signal correctly because of differences in the flashover signal and the existence of the interference signal. The conventional method is short of flexibility and robustness. This paper presents a method based on the support vector machine which is trained by a compound training data,and Gaussian noise is added to theoretical signal and the lab signal so that virtual samples can be constructed to solve the problem of uneven training data. The simulation and electrostatic field experimental results show that the proposed method can achieve a better performance on the accuracy rate,calculation speed,as well as the generalization ability.

关 键 词：支持向量机 虚拟样本 高压静电 火花检测 

分 类 号：TM83[电气工程—高电压与绝缘技术] TP181[自动化与计算机技术—控制理论与控制工程]