基于最小二乘支持向量机的绝缘子盐密光纤在线检测  

Insulator ESDD online detection based on least square support vector machine using fiber technology

作　　者：焦尚彬[1] 刘丁[1] 

机构地区：[1]西安理工大学信息与控制工程研究中心,西安710048

出　　处：《仪器仪表学报》2008年第11期2335-2340,共6页Chinese Journal of Scientific Instrument

基　　金：陕西省教育厅专项科研基金(06jk218);西安理工大学博士启动基金资助项目

摘　　要：本文介绍了利用光纤技术在线检测绝缘子盐密的基本原理,并在对受污石英玻璃棒和绝缘子进行试验研究的基础上,针对光通量衰减与盐密、环境湿度和尘埃比率之间复杂的非线性关系,利用最小二乘支持向量机,建立了以光通量衰减、环境湿度、尘埃比率等3个变量作为输入参数,等值附盐密度作为输出参数的智能预测模型,并通过部分实验数据验证了模型的有效性。最后应用预测模型开发了绝缘子盐密在线检测系统。In this paper, the basic principle of insulator salt density online detection using fiber technology is introduced. Based on laboratory simulation experiments of the contaminated silex sensor and insulator, aiming at the complicated nonlinear relationship between the luminous flux attenuation and various factors, the luminous flux attenuation, environmental humidity and ash density are chosen as three input variables, the ESDD is chosen as output variable, and an intelligent prediction model using least squares support vector machine （LS-SVM） is built. The feasibility of the method is proved by the tests in laboratory. Then the insulator ESDD online detection system is developed based on fiber technology.

关 键 词：等值附盐密度 最小二乘支持向量机 光纤技术 在线检测 

分 类 号：TM852[电气工程—高电压与绝缘技术]