ETRI-Knowledge Sharing Plaform

KOREAN
논문 검색
Type SCI
Year ~ Keyword

Detail

Conference Paper Fault Type Classification of Rotating Machinery based on Machine Learning with Fourier Transform Feature Extraction
Cited 0 time in scopus Share share facebook twitter linkedin kakaostory
Authors
Minseok Lee, Dohun Kim, Wonjong Kim
Issue Date
2023-10
Citation
International Conference on Consumer Electronics (ICCE) 2023 : Asia, pp.508-511
Publisher
IEEE
Language
English
Type
Conference Paper
DOI
https://dx.doi.org/10.1109/ICCE-Asia59966.2023.10326396
Abstract
This study presents an approach to classify failure types in rotating machinery using Fast Fourier Transform (FFT) for feature extraction and machine learning. Failures in rotating machinery can lead to substantial economic loss, and accurately classifying these failures is a complex task. The proposed method can contribute to enhancing the maintenance efficiency and system stability of rotating machinery. It can be applied across all industrial sectors involving rotating machinery. In this research, an FFT-based feature extraction method was implemented, transforming signals from the time domain to the frequency domain using window sizes corresponding to integer multiples of the machinery's rotation period. This process yielded significantly higher accuracy, up to 99.15%, compared to conventional deep learning methods when using window sizes corresponding to the rotation period. This represents a 21.96% improvement over existing deep learning techniques. The results demonstrate that feature extraction in the frequency domain using FFT is more effective than conventional time domain analysis.
KSP Keywords
Economic loss, FFT-based, Fast fourier transform (fft), Fault Type, Industrial sectors, Learning methods, Maintenance Efficiency, Rotating machinery, System stability, Type classification, complex task