The prediction of post insulators leakage current from environmental data

• Main Authors: Ling Zhao, Jianwu Jiang, Shaohui Duan, Chunhua Fang, Jianguo Wang, Kang Wang, Pingmei Cao, Jian Zhou
• Format: Conference Proceeding
• Language: English
• Subjects: Environmental Data; Equations; Humidity; Insulators; Leakage current; Linear regression; Mathematical model; Pollution measurement; Post Insulators; Regression Analysis
• DOI: 10.1109/ICECENG.2011.6057235

Leakage current (LC) and environmental data online monitoring system are installed in 6 substations in Shenzhen, which can measure simultaneously post insulators LC and environment temperature, humidity, wind speed, rainfall amount and wind direction. This paper presents the predict results of LC from the environmental data using linear and nonlinear regression. The results show that although the model of linear regression is simple, but the fitting degree is not high; The nonlinear regression method combined with principal component analysis can not only analyze the main impact factors of the meteorological on LC accurately and reduce the number of independent variables, but also can establish high fitting degree non-linear regression equation about the meteorological factors and LC. LC can be used to predict accurately by this equation. Fitting monthly eight months of data in No.1 substation using nonlinear regression equation can find that the model of regression equation is the same, whereas the regression coefficients fluctuate to a certain extent with the insulators surface contamination degree changes at different times, the regression coefficients should be dynamically adjusted according to the measured data in actual applications.