DETECTION OF STATOR WINDING FAULTS IN ROTARY ELECTRIC MACHINES BY USING DEEP LEARNING

Noğay H. S.

5th INTERNATIONAL CONGRESS ON ENGINEERING SCIENCES AND MULTIDISCIPLINARY APPROACHES, İstanbul, Türkiye, 25 - 26 Şubat 2023, ss.288-293

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.288-293
  • Bursa Uludağ Üniversitesi Adresli: Evet

Özet

Abstract: The most basic failure condition in rotating electrical machines is stator winding failures. Stator winding faults can show itself in the form of stator windings burning due to reasons such as excessive current or overload, or stator windings breaking due to other reasons or wear of insulations. Stator winding faults can normally be noticed at first glance by any winders or electrical machine researchers. However, some stator faults are not immediately noticeable and cause loss of time in the troubleshooting process. In order to solve this problem to some extent, an image-based artificial intelligence application was carried out in this study. A deep learning model has been applied that can detect stator winding faults with 93.55 % rate of accuracy based on the image data. As a result of the study, it has been seen that image-based deep learning approaches can be used in automatic fault detection in rotary electric machines and stator winding failure can be distinguished with a high accuracy rate and this application can be developed by generalizing it. Keywords: Rotary Electric Machines, Stator Winding, Deep Learning, CNN, TL