Paper Title
Classification of Corona Discharge Signals using Support Vector Machines

Abstract
The most common problem in the high voltage systems is the partial discharge (PD). If the location and amplitude of the partial discharges that occur in transformers, high voltage cables or other high voltage components can not be determined accurately, they may cause significant permanent damage over time. PD’s are defined as corona in overhead high voltage transmission lines. In this study, two different types of partial discharge electrodes are used to perform the corona phenomenon over the capacitor connected to the high voltage transformer. Voltage level is started from a low value and gradually raised, and partial discharge corona signals are recorded as digital data using digital oscilloscope and computer. Tables of statistical property values are obtained from recorded data of different electrodes and partial discharges are classified using support vector machines. The support vector machine is firstly trained with the majority values of the experiments and the remaining data were used to test the classification and successful results are obtained. Keywords - Corona, Classification, High Voltage, Partial Discharge, Support Vector Machines