Proposal of a new method for learning of diesel generator sounds and detecting abnormal sounds using an unsupervised deep learning algorithm
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This study is to find a method to learn engine sound after the start-up of a diesel generator installed in nuclear power plant with an unsupervised deep learning algorithm (CNN autoencoder) and a new method to predict the failure of a diesel generator using it. In order to learn the sound of a diesel generator with a deep learning algorithm, sound data recorded before and after the start-up of two diesel generators was used. The sound data of 20 min and 2 h were cut into 7 s, and the split sound was converted into a spectrogram image. 1200 and 7200 spectrogram images were created from sound data of 20 min and 2 h, respectively. Using two different deep learning algorithms (CNN autoencoder and binary classification), it was investigated whether the diesel generator post-start sounds were learned as normal. It was possible to accurately determine the post-start sounds as normal and the pre-start sounds as abnormal. It was also confirmed that the deep learning algorithm could detect the virtual abnormal sounds created by mixing the unusual sounds with the post-start sounds. This study showed that the unsupervised anomaly detection algorithm has a good accuracy increased about 3% with comparing to the binary classification algorithm.
Abstract
This study is to find a method to learn engine sound after the start-up of a diesel generator installed in nuclear power plant with an unsupervised deep learning algorithm (CNN autoencoder) and a new method to predict the failure of a diesel generator using it. In order to learn the sound of a diesel generator with a deep learning algorithm, sound data recorded before and after the start-up of two diesel generators was used. The sound data of 20 min and 2 h were cut into 7 s, and the split sound was converted into a spectrogram image. 1200 and 7200 spectrogram images were created from sound data of 20 min and 2 h, respectively. Using two different deep learning algorithms (CNN autoencoder and binary classification), it was investigated whether the diesel generator post-start sounds were learned as normal. It was possible to accurately determine the post-start sounds as normal and the pre-start sounds as abnormal. It was also confirmed that the deep learning algorithm could detect the virtual abnormal sounds created by mixing the unusual sounds with the post-start sounds. This study showed that the unsupervised anomaly detection algorithm has a good accuracy increased about 3% with comparing to the binary classification algorithm.
참고문헌 (18)
- 최광희, "디젤발전기 엔진 상태 진단 기술의 발전소 적용 연구", 동력시스템공학회지, 17(4), pp.17~22 (2013), 한국동력기계공학회
- S.G. Lee, "Development status of emergency diesel generator status diagnosis technology for nuclear power plants", J. Electr. World, pp.82~88 (2014)
- 이상국, "발전용 비상디젤발전기 엔진 상태진단 프로그램 개발 연구", 동력시스템공학회지, 19(5), pp.67~72 (2015), 한국동력기계공학회[DOI]
- 강경원, "스펙트로그램 이미지를 이용한 CNN 기반 자동화 기계 고장 진단 기법", 융합신호처리학회 논문지, 21(3), pp.121~126 (2020), 한국융합신호처리학회[DOI]
- S. H. Lee, "CNN based actuator fault cause classification system using noise", Proceedings of the Winter Conference of the Korean Society of Computer and Information Engineers, 26, pp.7~8 (2018)
- Thanh Tran, "Drill Fault Diagnosis Based on the Scalogram and Mel Spectrogram of Sound Signals Using Artificial Intelligence", IEEE Access, 8, pp.203655~203666 (2020), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- Hongmei Liu, "Rolling Bearing Fault Diagnosis Based on STFT-Deep Learning and Sound Signals", pp.6127479~ (2016), Hindawi Publishing Corporation Shock and Vibration
- K.M. Lee, "A Machine Fault Diagnosis System Based on Convolutional Neural Network Using Acoustic Power Spectral Density", pp.2019
- Hisahide Nakamura, "A Diagnosis Method of Bearing and Stator Fault in Motor Using Rotating Sound Based on Deep Learning", Energies, 14(5), pp.1319~ (2021), MDPI AG[DOI]
- C. Vununu, "A Sound Diagnosis System with Deep Learning for Machine Fault Detection", pp.2017
- Dong Li, "A fault diagnosis equipment of motor bearing based on sound signal and CNN", Journal of Physics: Conference Series, 2010(1), pp.012159~ (2021), IOP Publishing[DOI]
- S. H. Lee, "Deep-learning based PHM embedded system using noise · vibration", Proceedings of the Summer Conference of the Korean Society of Computer and Information Engineers, 25, pp.9~10 (2017)
- Shutao Zhao, "Fault diagnosis method for energy storage mechanism of high voltage circuit breaker based on CNN characteristic matrix constructed by sound-vibration signal", Journal of Vibroengineering, 21(6), pp.1665~1678 (2019), JVE International Ltd.[DOI]
- Yong Yao, "End-To-End Convolutional Neural Network Model for Gear Fault Diagnosis Based on Sound Signals", Applied Sciences, 8(9), pp.1584~ (2018), MDPI AG[DOI]
- "From Autoencoder to Beta-VAE" (2018)[URL]
- S. Jin, "Introduction to the spectrum spectrogram", J. Kor. Soc. Laryngol. Phoniatr. Logop, 19, pp.101~106 (2008)
- S.A. Broughton, "Discrete Fourier Analysis and Wavelets: Applications to Signal and Image Processing" (2018), Wiley
- F. Chollet, "Deep Learning with Python" (2017), Manning