The classification of partial discharge (PD) or partial breakdown (PB) is an important issue that helps to identify the cause of this electrical phenomenon in pin type insulators. In this work, a PD classification method based on neural networks (NN) is proposed. This sorting technique consists of three parts. First, the detection and measurement of partial discharge are achieved by using a digital finite impulse response (FIR) filter, whose main objective is to obtain electrical charges with significant characteristics of the PB. The second part of the process deals with the classification of partial discharge. A statistical analysis is implemented to obtain PD patterns or fingerprints which are classified by a neural network. Finally, the third part of the proposed method focuses on interpreting the information obtained from the NN and determining the PD current in the pin isolator. The results show that this proposed technique of analysis and detection of partial discharge in pin type isolators is successful in optimizing the time of analysis and classification of PB.
|Número de páginas||6|
|Estado||Publicada - 16 ene 2018|
|Evento||2017 IEEE International Autumn Meeting on Power, Electronics and Computing, ROPEC 2017 - Ixtapa, México|
Duración: 8 nov 2017 → 10 nov 2017
|Conferencia||2017 IEEE International Autumn Meeting on Power, Electronics and Computing, ROPEC 2017|
|Título abreviado||ROPEC 2017|
|Período||8/11/17 → 10/11/17|