Partial discharge source identification using PRPD data and deep learning techniques

Date

2024-1

Author

Altunyay, Özkan

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

360
views

3032
downloads

Cite This

Partial discharge is a phenomenon of localized electrical discharge that only partially bridges the insulation between conductors and which can or cannot occur adjacent to a conductor. It can occur within cables, transformers, and electrical equipment under high-voltage stress. As it is a dangerous event that progresses rapidly, it eventually causes the insulation materials to be damaged and the equipment to be out of service. In this thesis, the features of the Partial Discharge phenomenon are investigated. The studies on this field are revealed with the literature research. Test setups similar to real-life situations are constructed in the laboratory environment, and experiments are carried out related to corona, internal, and surface partial discharges. Grayscale images are created using PRPD data obtained from laboratory experiments, and they are used to train and test the deep-learning models for the classification of the partial discharge types. For this purpose, ResNet50, VGG16, VGG19, Inception, and Xception models are used via the transfer learning method. According to the results, the deep learning models have reached high accuracy levels.

Subject Keywords

Partial discharge measurement, PRPD, Partial discharge classification, Deep learning, Transfer learning

URI

https://hdl.handle.net/11511/108440

Collections

Graduate School of Natural and Applied Sciences, Thesis