Research On The Forecasting Method For The ESDD And NSDD Of The Suspension-type Porcelain Insulator

The changing of the leakage current (LC) can reflect the contaminating and moistening process. A lot of researches about the influence of the equivalent salt deposit density (ESDD) and the relative humidity (RH) on the LC of the insulator have been carried out, but few studies are done about the influence of the non-soluble deposit density (NSDD) on LC. The laboratory experiments are carried out in artificial fog cabinet. The effect of ESDD, NSDD and RH on the LC is analyzed. In order to predict the pollution degree of the insulator, the least squares support vector machine (LS-SVM) is employed to build the prediction model for ESDD and NSDD. The LS-SVM model is optimized by genetic algorithms (GA) and suitable parameters of the model are obtained. It shows that the optimized LS-SVM model has achieved good predicting performance.