Modified convolutional neural network with global average pooling for intelligent fault diagnosis of industrial gearbox

• Autorzy: Li Yaxin , Wang Kesheng

Identyfikatory

DOI

10.17531/ein.2020.1.8

Warianty tytułu

PL

Diagnostyka błędów przekładni przemysłowych z wykorzystaniem zmodyfikowanej splotowej sieci neuronowej z globalnym uśrednieniem wartości dla poszczególnych kanałów

• Języki publikacji: EN

Abstrakty

EN

Gearboxes are key transmission components and widely used in various industrial applications. Due to the possible operational conditions, such as varying rotational speeds, long period of heavy loads, etc., gearboxes may easily be prone to failure. Condition Monitoring (CM) has been proved to be an effective methodology to improve the safety and reliability of gearboxes. Deep learning approaches, nowadays, further enable the CM with more powerful capability to exploit faulty information from massive data and make intelligently diagnostic decisions. However, for most of conventional deep learning models, such as Convolutional Neural Network (CNN), a large amount of labelled training data is a prerequisite, while to obtain the labelled data is usually a laborious and time-consuming job and sometimes even unattainable. In this paper, to handle the case of only a limited labelled data is available, a modified convolutional neural network (MCNN) is proposed by integrating global average pooling (GAP) to reduce the number of trainable parameters and simplify the architecture of deep learning model. The proposed MCNN improves the traditional CNN’s ability in fault diagnosis with limited labelled data. Two experimental gearbox datasets are utilized to demonstrate the effectiveness of the proposed MCNN method. Compared with traditional deep learning approaches, namely LSTM, CNN and its variant methods, the experimental results show that the proposed MCNN with higher discrimination and generalization ability in fault classification and diagnostics under the scenario of limited labelled training samples.

PL

Przekładnie stanowią kluczowe elementy układów napędowych i jako takie znajdują szerokie zastosowane w przemyśle. Ze względu na warunki eksploatacji, takie jak różne prędkości obrotowe czy długie okresy pracy pod dużym obciążeniem itp., przekładnie mogą łatwo ulegać uszkodzeniom. Udowodniono, że monitorowanie stanu skutecznie poprawia bezpieczeństwo i niezawodność przekładni. Podejścia oparte na uczeniu głębokim umożliwiają ponadto monitorowanie stanu z większym wykorzystaniem informacji o błędach pochodzących z dużych zbiorów danych i podejmowanie inteligentnych decyzji diagnostycznych. Jednak w przypadku większości konwencjonalnych modeli uczenia głębokiego, takich jak splotowe sieci neuronowe (convolutional neural networks, CNN), wymagana jest duża ilość etykietowanych danych uczących, których pozyskanie jest zwykle zadaniem praco- i czasochłonnym, a czasem wręcz niemożliwym do wykonania. W niniejszej pracy, przedstawiono zmodyfikowaną splotową sieć neuronową (modified convolutional neural network, MCNN), która rozwiązuje problem dostępności danych etykietowanych poprzez zastosowanie globalnego uśrednienia względem kanałów (global average pooling), co pozwala na zmniejszenie liczby możliwych do wyuczenia parametrów i uproszczenie architektury modelu głębokiego uczenia. W porównaniu do tradycyjnych sieci CNN, proponowana sieć MCNN zwiększa możliwości diagnozowania błędów przy ograniczonych danych etykietowanych. Skuteczność proponowanej metody wykazano na przykładzie dwóch zbiorów danych doświadczalnych dotyczących błędów przekładni. Wyniki eksperymentalne pokazują, że, w porównaniu z tradycyjnymi metodami uczenia głębokiego, takimi jak LSTM, CNN oraz warianty tej ostatniej, proponowane podejście MCNN daje większe możliwości rozróżniania i uogólniania podczas klasyfikacji i diagnostyki błędów w przypadku ograniczonej dostępności etykietowanych danych uczących.

Słowa kluczowe

EN

modified convolutional neural network; global average pooling; intelligent fault diagnosis; industrial Gearbox

PL

zmodyfikowana splotowa sieć neuronowa; globalne uśrednienie względem kanałów; inteligentna diagnostyka błędów; przekładnia przemysłowa

• Wydawca: Polskie Naukowo-Techniczne Towarzystwo Eksploatacyjne
• Czasopismo: Eksploatacja i Niezawodność
• Rocznik: 2020
• Tom: Vol. 22, no. 1
• Strony: 63--72
• Opis fizyczny: Bibliogr. 44 poz., rys., tab.

Twórcy

autor

Li Yaxin

• Department of Mechanical and Electrical Engineering University of Electronic Science and Technology of China Chengdu, 610059, China

autor

Wang Kesheng

• Department of Mechanical and Electrical Engineering University of Electronic Science and Technology of China Chengdu, 610059, China

Bibliografia

• 1. Cai B P, Huang L, Xie M. Bayesian networks in fault diagnosis. IEEE Transactions on industrial informatics 2017; 13(5): 2227 - 2240, https://doi.org/10.1109/TII.2017.2695583.
• 2. Cai B P, Liu Y H, Fan Q ,Zhang Y W, Liu Z K, Yu S L, Ji R J. Multi-source information fusion based fault diagnosis of ground-source heat pump using Bayesian network. Applied Energy 2014; 114: 1-9, https://doi.org/10.1016/j.apenergy.2013.09.043.
• 3. Cai B P, Liu H L, Xie M. A real-time fault diagnosis methodology of complex systems using object-oriented Bayesian networks. Mechanical Systems and Signal Processing 2016; 80: 31-44, https://doi.org/10.1016/j.ymssp.2016.04.019.
• 4. Cai B P, Shao X Y, Liu Z K, Kong X D. Remaining useful life estimation of structure systems under the influence of multiple causes: Subsea pipelines as a case study. IEEE Transactions on Industrial Electronics 2019; 99:1-1, https://doi.org/10.1109/TIE.2019.2931491.
• 5. Chen D, Yang S, Zhou F. Transfer learning based fault diagnosis with missing data due to multi-rate sampling. Sensors 2019; 19(8):1826, https://doi.org/10.3390/s19081826.
• 6. Chen Yuejian, Liang Xihui, Zuo Ming J. Sparse time series modeling of the baseline vibration from a gearbox under time-varying speed condition. Mechanical Systems and Signal Processing 2019; 134: 106342, https://doi.org/10.1016/j.ymssp.2019.106342.
• 7. Chen ZY, Gryllias K, Li WH. Mechanical fault diagnosis using Convolutional Neural Networks and Extreme Learning Machine. Mechanical Systems and Signal Processing 2019; 133: 106272, https://doi.org/10.1016/j.ymssp.2019.106272.
• 8. Duan R, Lin Y, Zeng Y. Fault diagnosis for complex systems based on reliability analysis and sensors data considering epistemic uncertainty. Eksploatacja i Niezawodnosc - Maintenance and Reliability 2018; 20(4): 558-566, https://doi.org/10.17531/ein.2018.4.7.
• 9. Feng K, Wang K, Ni Q. A phase angle based diagnostic scheme to planetary gear faults diagnostics under non-stationary operational conditions. Journal of Sound and Vibration 2017; 408:190-209, https://doi.org/10.1016/j.jsv.2017.07.030.
• 10. Feng Z P, Zhu W P, Dong Zhang. Time-Frequency demodulation analysis via Vold-Kalman filter for wind turbine planetary gearbox fault diagnosis under nonstationary speeds. Mechanical Systems and Signal Processing 2019; 128: 93-109, https://doi.org/10.1016/j.ymssp.2019.03.036.
• 11. Han T, Jiang D, Qi Z, Lei W, Kai Y. Comparison of random forest, artificial neural networks and support vector machine for intelligent diagnosis of rotating machinery. Transactions of the Institute of Measurement and Control 2018; 40: 2681-93, https://doi.org/10.1177/0142331217708242.
• 12. Han, T, Jiang, D, Sun, Y, Wang, N, Yang, Y. Intelligent fault diagnosis method for rotating machinery via dictionary learning and sparse representation-based classification. Measurement 2018; 118:181-193, https://doi.org/10.1016/j.measurement.2018.01.036.
• 13. Han T, Liu C, Wu Lj, Sarkar S, Jiang DX. An adaptive spatiotemporal feature learning approach for fault diagnosis in complex systems. Mechanical Systems and Signal Processing 2019; 117:170-187, https://doi.org/10.1016/j.ymssp.2018.07.048.
• 14. Han T, Liu C, Yang WG, Jiang DX. Deep transfer network with joint distribution adaptation: A new intelligent fault diagnosis framework for industry application. ISA Transactions 2019, https://doi.org/10.1016/j.isatra.2019.08.012.
• 15. Han T, Liu C, Yang WG, Jiang D. Learning transferable features in deep convolutional neural networks for diagnosing unseen machine conditions. ISA Transactions, https://doi.org/10.1016/j.isatra.2019.03.017.
• 16. Huang W Y, Cheng J H. An improved deep convolutional neural network with multi-scale information for bearing fault diagnosis. Neurocomputing 2019, https://doi.org/10.1016/j.neucom.2019.05.052.
• 17. Jiao J, Zhao M, Lin J. Deep Coupled Dense Convolutional Network with Complementary Data for Intelligent Fault Diagnosis. IEEE Transactions on Industrial Electronics 2019; 66(12): 9858 - 9867, https://doi.org/10.1109/TIE.2019.2902817.
• 18. Kaluer S, Fekete K, Jozsa L, Klai Z. Fault diagnosis and identification in the distribution network using the fuzzy expert system. Eksploatacja i Niezawodnosc - Maintenance and Reliability 2018; 20(4): 621-629, https://doi.org/10.17531/ein.2018.4.13.
• 19. Lei J, Liu C, Jiang D. Fault diagnosis of wind turbine based on Long Short-Term memory networks. Renewable Energy 2019; 133: 422-432, https://doi.org/10.1016/j.renene.2018.10.031.
• 20. Lin M, Chen Q, Yan S C, Network in Network, Neural and Evolutionary Computing. arXiv:1312.4400.
• 21. Liu H , Zhou J , Zheng Y. Fault diagnosis of rolling bearings with recurrent neural network-based autoencoders. ISA Transactions 2018; 77:167-178, https://doi.org/10.1016/j.isatra.2018.04.005.
• 22. Liu XL, Zhang XY, Wang LY. Fault Diagnosis Method of Wind Turbine Gearbox Based on Deep Belief Network and Vibration Signal. Society of Instrument and Control Engineers of Japan.
• 23. PHM, Phm data challenge 2009., https://www.phmsociety.org/competition/PHM/09, 2009.
• 24. Rezaei M, Yang H J, Meinel C. Deep Neural Network with l2-norm Unit for Brain Lesions Detection. arXiv:1708.05221.
• 25. Shao H D, Jiang H K, Zhao K. A novel tracking deep wavelet auto-encoder method for intelligent fault diagnosis of electric locomotive bearings. Mechanical Systems and Signal Processing 2018; 110: 193-209, https://doi.org/10.1016/j.ymssp.2018.03.011.
• 26. Shao SY, Wang P, Yan R Q. Generative adversarial networks for data augmentation in machine fault diagnosis. Computers in Industry 2019; 106: 85-93, https://doi.org/10.1016/j.compind.2019.01.001.
• 27. Sikora M, Szczyrba K, Wróbel, Michalak M. Monitoring and maintenance of a gantry based on a wireless system for measurement and analysis of the vibration level. Eksploatacja i Niezawodnosc - Maintenance and Reliability 2019; 21(2): 341-350, https://doi.org/10.17531/ein.2019.2.19.
• 28. Tang G J, Pang B. Gearbox Fault Diagnosis Based on Hierarchical Instantaneous Energy Density Dispersion Entropy and Dynamic Time Warping. Entropy 2019; 21(6): 593, https://doi.org/10.3390/e21060593.
• 29. Tyagi S, Panigrahi S K. A Hybrid Genetic Algorithm and Back-Propagation Classifier for Gearbox Fault Diagnosis. Applied Artificial Intelligence 2017; 1-20, https://doi.org/10.1080/08839514.2017.1315502.
• 30. Verstraete D, Ferrada A, Droguett E.L, Meruane V, Modarres M. Deep learning enabled fault diagnosis using time-frequency image analysis of rolling element bearings. Shock and Vibration 2017; 2017: 1-17, https://doi.org/10.1155/2017/5067651.
• 31. Wang J, Li S, An Z. Batch-normalized deep neural networks for achieving fast intelligent fault diagnosis of machines. Neurocomputing 2019; 329: 53-65, https://doi.org/10.1016/j.neucom.2018.10.049.
• 32. Wang K S, Heyns P S. Application of computed order tracking, Vold-Kalman filtering and EMD in rotating machine vibration. Mechanical Systems and Signal Processing 2011; 25(1): 416-430, https://doi.org/10.1016/j.ymssp.2010.09.003.
• 33. Wang K S, Heyns P S. The combined use of order tracking techniques for enhanced Fourier analysis of order components. Mechanical Systems and Signal Processing 2011; 25(3): 803-811, https://doi.org/10.1016/j.ymssp.2010.10.005.
• 34. Wen L, Gao L, Li X, Wen L, Gao L, Li X. A new deep transfer learning based on sparse auto-encoder for fault diagnosis. IEEE Transactions on systems, man, and cybernetics: systems 2017; 1-9.
• 35. Xu H and Chen G. An intelligent fault identification method of rolling bearings based on LSSVM optimized by improved PSO. Mechanical Systems and Signal Processing 2013; 35(1-2): 167-175, https://doi.org/10.1016/j.ymssp.2012.09.005.
• 36. Yang J, Guo Y Q, Zhao W L. Long short-term memory neural network based fault detection and isolation for electro-mechanical actuators. Neurocomputing 2019; 360: 85-96, https://doi.org/10.1016/j.neucom.2019.06.029.
• 37. Yu J, Xu YG, Liu K. Planetary gear fault diagnosis using stacked denoising autoencoder and gated recurrent unit neural network under noisy environment and time-varying rotational speed conditions. Measurement Science and Technology 2019; 30: 095003, https://doi.org/10.1088/1361-6501/ab1da0.
• 38. Zhang M, Wang K, Li Y. Motion Periods of Planet Gear Fault Meshing Behavior. Sensors 2018; 18(11), https://doi.org/10.3390/s18113802.
• 39. Zhang M, Wang K S, Wei D D. Amplitudes of characteristic frequencies for fault diagnosis of planetary gearbox. Journal of Sound and Vibration 2018; 432:119-132, https://doi.org/10.1016/j.jsv.2018.06.011.
• 40. Zhang W, Li X and Ding Q. Deep residual learning-based fault diagnosis method for rotating machinery. ISA Transactions 2018, https://doi.org/10.1016/j.isatra.2018.12.025.
• 41. Zhang W, Peng G, Li C, Chen Y, Zhang Z. A new deep learning model for fault diagnosis with good anti-noise and domain adaptation ability on raw vibration signals. Sensors 2017; 17(2): 425, https://doi.org/10.3390/s17020425.
• 42. Zhang X L, Wang B J, Chen X F. Intelligent fault diagnosis of roller bearings with multivariable ensemble-based incremental support vector machine. Knowledge-Based Systems 2015; 89: 56-85, https://doi.org/10.1016/j.knosys.2015.06.017.
• 43. Zhang Z Z, Li S M. General normalized sparse filtering: A novel unsupervised learning method for rotating machinery fault diagnosis. Mechanical Systems and Signal Processing 2019; 124: 596-612, https://doi.org/10.1016/j.ymssp.2019.02.006.
• 44. Zhao X L, Jia M P. A new Local-Global Deep Neural Network and its application in rotating machinery fault diagnosis. Neurocomputing, https://doi.org/10.1016/j.neucom.2019.08.010.

Uwagi

Opracowanie rekordu ze środków MNiSW, umowa Nr 461252 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2020).