Classification of Radiometric Partial Discharge Signals Using Signal Conditioning System

Due to multiple partial discharge sources in a substation, it becomes necessary to classify partial discharge signals to identify the sources of partial discharge. Therefore, this study aims to classify radiometric PD signals using a signal conditioning system developed in a previous study. The conditioning system used can reduce the need for sampling, resulting in reduced costs with the acquisition system and enabling the monitoring of partial discharges in substations. This paper will evaluate the signal conditioning system for classifying partial discharge and interference signals measured in a 230 kV substation, utilizing machine learning algorithms such as Support vector machine and random forest. The objective is to evaluate the effectiveness of the system under substation conditions.