Applying S-transform and SVM to evaluate insulator's pollution condition based on leakage current

The flashover of polluted insulators is a serious threat to the safe operation of overhead power lines. The leakage current on surface of insulators is closely related to the developing process of pollution flashover, affected by many factors, such as wetting degree, system voltage and structures of insulators. At the same time, the leakage current is easy to be monitored on-line continuously. Hence, as one of on-line monitoring methods for polluted insulators. Based on artificial pollution tests under different relative humidity (RH) and salt deposit density (SDD), a method incorporates S-transform and support vector machine (SVM) for evaluating safety condition of polluted insulators is proposed. Firstly, the approach utilizes S-transform to extract the amplitude Ih, phase θ and total harmonic distortion (THD) of leakage current. Then the I h , θ and THD are as characteristic parameters to put into SVM, the output parameters of SVM are fuzzifled into four fuzzy subsets for evaluating safety condition of polluted insulators. Lastly, the performance of SVM is validated by 8 pairs of testing samples, and the influences of SDD and RH on the performance of insulators are discussed. The results show that, the leakage current becomes inductive when strong local arcs occur, decreases with the increase of RH, and Ih has no relation with SDD when the RH is low but increases with the increase of SDD when RH is high. THD of leakage current increases slowly with the increase of RH when the pollution is light, but decreases firstly and then increases with the increase of RH when the pollution is middle or heavy, THD take the minimum at the RH of about 80%.