电晕Determination of voltage level from audible DC corona noise by using neutral network.pdfVIP

Determination of Voltage Level from Audible DC Corona Noise by Using Neural Network Suna Bolat Sert1, Özcan Kalenderli2 1, 2 Electrical Engineering Department, Istanbul Technical University, 34469 Maslak-Istanbul, Turkey bolats@elk.itu.edu.tr, ozcan@elk.itu.edu.tr Abstract terms of installation maintenance than the traditional measuring methods [6]. In this study, a signal recognition application is presented In this study, by means of probabilistic neural networks for classification of positive DC corona voltage level using (PNN) a new approach has is presented that the sound recorded sound data of corona (electrical discharge) and recordings of positive DC corona (electrical discharge) are used utilizing probabilistic neural network (NN). The recorded to determine the voltage level, which causes the electrical positive DC corona sound data are acquired experimentally discharge [7-8]. The data that are used to classify the voltage from a test set-up intentionally producing corona sound by from audible DC corona noise using the neural networks have applying different levels of positive DC high-voltage. been collected from a high-voltage line model installed in a Recordings have been used in training and test sets of the high-voltage laboratory. In order to apply the sound recordings neural network. The main object