Diagnosis of Bearing Faults Under Variable Speed Conditions Using Deep Learning

Öz

Bearings are fundamental and delicate elements directly influencing performance, efficiency, stability, and operational lifespan. However, harsh and fluctuating operating conditions not only jeopardize the safe working environment but also lead to abrupt and unforeseen component faults, resulting in economic losses. Diagnosing faults in bearings operating under variable speed conditions necessitates a shift from traditional methods towards more intricate signal processing techniques and artificial intelligence models with more challenging interpretations. Nevertheless, this research article aims to significantly reduce computational burden and complexity by employing simpler and more straightforward models both in the process of feature extraction and classification, utilizing deep learning methodologies. The research article encompasses the transformation of raw vibration data obtained from bearings operating under variable speed conditions into visual representations and their subsequent classification using the Long Short-Term Memory (LSTM), one of the deep learning models. The developed LSTM-based fault classification model, trained with very limited data, achieves 100% accuracy in classifying four different states of the bearing.

Anahtar Kelimeler

• Bearing
• fault diagnosis
• deep learning
• LSTM
• signal to image mapping

Destekleyen Kurum

This study is supported by TÜBİTAK - BİDEB 2211/C National PhD Scholarship Program in the Priority Fields in Science and Technology, 100/2000 Council of Higher Education (Yükseköğretim Kurulu - YÖK) Doctoral Scholarship Program and Fırat University Scientific Research Projects Unit (Fırat Üniversitesi Bilimsel Araştırma Projeleri - FÜBAP) with the project number ADEP.22.06.

Teşekkür

We would like to thank TÜBİTAK, YÖK and FÜBAP for their support.

Değişken Hız Koşullarında Rulman Arızalarının Derin Öğrenme Kullanılarak Teşhisi

Öz

Rulmanlar, performansı, verimliliği, stabiliteyi ve operasyonel ömrü doğrudan etkileyen temel ve hassas bileşenlerdir. Ancak, zorlu ve değişken çalışma koşulları, yalnızca güvenli çalışma ortamını tehlikeye atmakla kalmaz, aynı zamanda ani ve öngörülemeyen bileşen arızalarına yol açarak ekonomik kayıplara neden olmaktadır. Değişken hız koşulları altında çalışan rulmanlarda arıza teşhisi, geleneksel yöntemlerden daha karmaşık sinyal işleme tekniklerini ve yorumlanması daha zor yapay zekâ modellerini gerektirir. Buna rağmen, bu araştırma makalesi, hem özellik çıkarma hem de sınıflandırma sürecinde daha basit ve anlaşılır modeller kullanarak hesaplama yükünü ve model karmaşıklığını önemli ölçüde azaltmayı amaçlamaktadır. Araştırma makalesi, değişken hız koşulları altında çalışan rulmanlardan elde edilen ham titreşim verilerinin görsel temsillere dönüştürülmesini ve ardından derin öğrenme modellerinden biri olan LSTM ile sınıflandırılmasını kapsamaktadır. Geliştirilen LSTM tabanlı arıza sınıflandırma modeli, oldukça sınırlı verilerle eğitildiğinde, rulmanın dört farklı durumunu %100 doğrulukla sınıflandırmayı başarmaktadır.

Anahtar Kelimeler

• Rulman
• Arıza Teşhisi
• Derin Öğrenme
• LSTM
• Sinyal Görüntü Haritalama

Destekleyen Kurum

TÜBİTAK BİDEB 2211/C Yurtiçi Öncelikli Alanlar Doktora Burs Programı, YÖK 100/2000 Doktora Burs Programı, Fırat Üniversitesi Bilimsel Araştırmalar Programı - FÜBAP ADEP.22.06 projesi ile desteklenmektedir

Teşekkür

Desteklerinden dolayı TÜBİTAK, YÖK ve FÜBAP'a teşekkür ederiz

Kaynakça

1. Rao M., Zuo M.J., Tian Z., "A speed normalized autoencoder for rotating machinery fault detection under varying speed conditions", Mechanical Systems and Signal Processing, 189, 110109, 2023.
2. Chen J., Chen J., Chen Z., Liu S., He S., "Hybrid augmented network with balance domain window for few-shot fault diagnosis under sharp speed variation", Mechanical Systems and Signal Processing, 207, 110944, 2024.
3. Sun H., Gao S., Ma S., Lin S., "A fault mechanism-based model for bearing fault diagnosis under non-stationary conditions without target condition samples", Measurement, 199, 111499, 2022.
4. Aziz S., Khan M.U., Faraz M., Montes G.A., "Intelligent bearing faults diagnosis featuring Automated Relative Energy based Empirical Mode Decomposition and novel Cepstral Autoregressive features", Measurement, 216, 112871, 2023.
5. Lu R., Xu M., Zhou C., Zhang Z., He S., Yang Q., Mao M., Yang J., "A Novel Fault Diagnosis Method Based on NEEEMD-RUSLP Feature Selection and BTLSTSVM", IEEE Access, 11, 113965–113994, 2023.
6. Kumar A., Groza V., Raj K.K., Assaf M.H., Kumar S., Kumar R.R., "Comparative Analysis of Machine Learning Techniques for Bearing Fault Classification in Rotating Machinery", SACI 2023 - IEEE 17th International Symposium on Applied Computational Intelligence and Informatics, Proceedings, 575–580, 2023.
7. Zhou H., Huang X., Wen G., Dong S., Lei Z., Zhang P., Chen X., "Convolution enabled transformer via random contrastive regularization for rotating machinery diagnosis under time-varying working conditions", Mechanical Systems and Signal Processing, 173, 109050, 2022.
8. Zhao J., Yang S., Li Q., Liu Y., Gu X., Liu W., "A new bearing fault diagnosis method based on signal-to-image mapping and convolutional neural network", Measurement, 176, 109088, 2021.

9. ZHANG J., SUN Y., GUO L., GAO H., HONG X., SONG H., "A new bearing fault diagnosis method based on modified convolutional neural networks", Chinese Journal of Aeronautics, 33, 439–447, 2020.
10. Öcalan G., Türkoğlu İ., "Fault Diagnosis of Rotating Machines Using Raw Vibration Signals and Deep Learning", 2021 Innovations in Intelligent Systems and Applications Conference (ASYU), 1–7, 2021.
11. Neupane D., Seok J., "Bearing Fault Detection and Diagnosis Using Case Western Reserve University Dataset With Deep Learning Approaches: A Review", IEEE Access, 8, 93155–93178, 2020.
12. Dündar D.R., Sarıçiçek İ., Çınar E., Yazıcı A., "Machine Learning In Predictive Maintenance: Literature Research", Journal of Engineering and Architecture Faculty of Eskisehir Osmangazi University, 29, 256–276, 2021.
13. Huang H., Baddour N., "Bearing vibration data collected under time-varying rotational speed conditions", Data in Brief, 21, 1745–1749, 2018.
14. Huang H., Baddour N., "Bearing vibration data collected under time-varying rotational speed conditions", Mendeley Data, V2, [Online]. Available: https://data.mendeley.com/datasets/v43hmbwxpm/2, 2019.
15. Karakaya M., "LSTM: Understanding the Number of Parameters", Kaggle, [Online]. Available: , https://www.kaggle.com/code/kmkarakaya/lstm-understanding-the-number-of-parameters, 2024.
16. Karim R., "Animated RNN, LSTM and GRU", Medium, [Online]. Available: https://towardsdatascience.com/animated-rnn-lstm-and-gru-ef124d06cf45, 2024.
17. Sokolova M., Lapalme G., "A systematic analysis of performance measures for classification tasks", Information Processing & Management, 45, 427–437, 2009.