Title:	Determination of the Optimal Reference Signal Frequency in the Task of Power Spectrum Pattern Classification
Authors:	Olkhovskiy, Mikhail Müllerová, Eva Martínek, Petr
Citation:	OLKHOVSKIY, M. MÜLLEROVÁ, E. MARTÍNEK, P. Determination of the Optimal Reference Signal Frequency in the Task of Power Spectrum Pattern Classification. In Proceedings of the 2022 22nd International Scientific Conference on Electric Power Engineering (EPE 2022). Ostrava: VSB - Technical University of Ostrava, 2022. s. 1-4. ISBN: 978-1-66541-056-4
Issue Date:	2022
Publisher:	IEEE
Document type:	konferenční příspěvek ConferenceObject
URI:	2-s2.0-85135107653 http://hdl.handle.net/11025/53859
ISBN:	978-1-66541-056-4
Keywords in different language:	convolutional neural networks;reference signal processing;signal spectrum;signal analysis;cable insulation;classification accuracy
Abstract in different language:	This article describes the results of an experiment in which a reference signal with its subsequent processing is used to obtain information on the degree of cable insulation degradation. A one-dimensional convolutional neural network is used for signal classification. The reference signal was tested with two methods of insulation degradation, thermal and electrical. The signal was preprocessed to obtain its power spectrum density. After that the obtained result was sent to the input of the neural network. Based on the learning and validation curves and the confusion classification matrices, the optimal frequencies of the reference signal were found.
Rights:	Plný text je přístupný v rámci univerzity přihlášeným uživatelům © IEEE
Appears in Collections:	Konferenční příspěvky / Conference papers (RICE) Konferenční příspěvky / Conference Papers (KEE) OBD

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.