Reconstruction-based stacked sparse auto-encoder for nonlinear industrial process fault diagnosis

• Autorzy: Weng Qihang , Ren Shaojun , Zhu Baoyu , Jin Yinfeng , Si Fengqi

Identyfikatory

DOI

10.17531/ein/175873

• Języki publikacji: EN

Abstrakty

EN

The reconstruction-based (RB) approach can effectively suppress the misdiagnosis problem due to the smearing effect in fault isolation. However, the current exploration of the RB approach for large-scale nonlinear systems is still limited. Therefore, this paper proposes a reliable and effective fault diagnosis method based on a reconstruction-based stacked sparse autoencoder (RBSSAE) for high-dimensional industrial systems. In RBSSAE, a reconstruction-based index achieved by the Steffensen iterative method is developed to check whether the given variable(s) are responsible for the faults efficiently. However, the number of possible faulty variable combinations grows exponentially with the system dimensionor actual abnormal variables, causing an unbearable computational burden for variable combination optimization. Hence, the proposed RBSSAE utilizes a sequential floating forward selection approach to rapidly isolate the most decisive combination of fault variables, meeting a requirement of online fault diagnosis. Finally, the effectiveness of the RBSSAE is verified on a numerical example and a real industrial case. Comparisons with other state-of-the-art methods are also presented.

Słowa kluczowe

EN

process monitoring; fault diagnosis; stacked sparse auto-encoder; reconstruction-based method; sequential floating forward selection

• Wydawca: Polskie Naukowo-Techniczne Towarzystwo Eksploatacyjne
• Czasopismo: Eksploatacja i Niezawodność
• Rocznik: 2024
• Tom: Vol. 26, no. 1
• Strony: art. no. 175873
• Opis fizyczny: Bibliogr. 32 poz., rys., tab., wykr.

Twórcy

autor

Weng Qihang

• School of Energy and Environment, Southeast University, China

autor

Ren Shaojun

• School of Energy and Environment, Southeast University, China

• https://orcid.org/0000-0002-0013-4522

autor

Zhu Baoyu

• School of Energy and Environment, Southeast University, China

autor

Jin Yinfeng

• Nanjing Nari-Relays Electric Co., Ltd, China

autor

Si Fengqi

• School of Energy and Environment, Southeast University, China

Bibliografia

• 1. Li Z, Bao S, Peng X, Luo L. Fault detection and diagnosis in multivariate systems using multiple correlation regression. Control Eng Pract [Internet]. 2021;116(January):104916. Available from: https://doi.org/10.1016/j.conengprac.2021.104916
• 2. Wen S, Zhang W, Sun Y, Li Z, Huang B, Bian S, et al. An enhanced principal component analysis method with Savitzky–Golay filter and clustering algorithm for sensor fault detection and diagnosis. Appl Energy [Internet]. 2023;337(July 2022):120862. Availablefrom: https://doi.org/10.1016/j.apenergy.2023.120862
• 3. Kong X, Jiang X, Zhang B, Yuan J, Ge Z. Latent Variable Models in the Era of Industrial Big Data: Extension and Beyond. Annu Rev Control [Internet]. 2022;(June). Available from: https://doi.org/10.1016/j.arcontrol.2022.09.005
• 4. Zhang J, Zhou D, Chen M. Self-learning sparse PCA for multimode process monitoring. IEEE Trans Ind Informatics. 2022;XX(XX). https://doi.org/10.1109/TII.2022.3178736
• 5. Wang Z, Zheng Y, Wong DSH. Trajectory-based operation monitoring of transition procedure in multimode process. J Process Control [Internet]. 2020;96:67–81. Available from: https://doi.org/10.1016/j.jprocont.2020.09.008
• 6. Kong XY, Cao ZH, Du BY, Luo JY. Quality-related multimodal fault detection technique based on partial least squares. Kongzhi yu Juece/Control Decis. 2019;34(12):2547–57.
• 7. Lin X, Sun R, Wang Y. Improved key performance indicator-partial least squares method for nonlinear process fault detection based on just-in-time learning. J Franklin Inst [Internet]. 2023;360(1):1–17. Available from: https://doi.org/10.1016/j.jfranklin.2022.11.029
• 8. Zheng Y, Zhou W, Yang W, Liu L, Liu Y, Zhang Y. Multivariate/minor fault diagnosis with severity level based on Bayesian decision theory and multidimensional RBC. J Process Control [Internet]. 2021;101:68–77. Available from: https://doi.org/10.1016/j.jprocont.2021.01.009
• 9. Van Den Kerkhof P, Vanlaer J, Gins G, Van Impe JFM. Analysis of smearing-out in contribution plot based fault isolation for Statistical Process Control. Chem Eng Sci. 2013;104:285–93. https://doi.org/10.1016/j.ces.2013.08.007
• 10. Yue HH, Qin SJ. Reconstruction-based fault identification using a combined index. Ind Eng Chem Res [Internet]. 2001;40(20):4403–14. Available from:https://doi.org/10.1021/ie000141+
• 11. Ren S, Si F, Zhou J, Qiao Z, Cheng Y. A new reconstruction-based auto-associative neural network for fault diagnosis in nonlinear systems. Chemom Intell Lab Syst. 2018;172(December):118–28. https://doi.org/10.1016/j.chemolab.2017.12.005
• 12. Mnassri B, El Adel EM, Ouladsine M. Reconstruction-based contribution approaches for improved fault diagnosis using principal component analysis. J Process Control [Internet]. 2015;33:60–76. Available from: http://dx.doi.org/10.1016/j.jprocont.2015.06.004
• 13. Alcala CF, Qin SJ. Reconstruction-based contribution for process monitoring. Automatica [Internet]. 2009;45(7):1593–600. Available from: http://dx.doi.org/10.1016/j.automatica.2009.02.027
• 14. Yan Z, Yao Y. Variable selection method for fault isolation using least absolute shrinkage and selection operator (LASSO). Chemom Intell Lab Syst [Internet]. 2015;146:136–46. Available from: http://dx.doi.org/10.1016/j.chemolab.2015.05.019
• 15. Hallgrímsson ÁD, Niemann HH, Lind M. Unsupervised isolation of abnormal process variables using sparse autoencoders. J Process Control [Internet]. 2021;99:107–19. Available from: https://doi.org/10.1016/j.jprocont.2021.01.005
• 16. Tang P, Peng K, Jiao R. A process monitoring and fault isolation framework based on variational autoencoders and branch and bound method. J Franklin Inst [Internet]. 2022;359(2):1667–91. Available from: https://doi.org/10.1016/j.jfranklin.2021.11.016
• 17. Rauber TW, Boldt FA, Munaro CJ. Feature selection for multivariate contribution analysis in fault detection and isolation. J Franklin Inst. 2020;357(10):6294–320. https://doi.org/10.1016/j.jfranklin.2020.03.005
• 18. He B, Yang X, Chen T, Zhang J. Reconstruction-based multivariate contribution analysis for fault isolation: A branch and bound approach. J Process Control. 2012;22(7):1228–36. https://doi.org/10.1016/j.jprocont.2012.05.010
• 19. He B, Zhang J, Chen T, Yang X. Penalized reconstruction-based multivariate contribution analysis for fault isolation. Ind Eng Chem Res. 2013;52(23):7784–94. https://doi.org/10.1021/ie303225a
• 20. Zheng Y, Hu C, Wang X, Wu Z. Physics-informed recurrent neural network modeling for predictive control of nonlinear processes. J Process Control [Internet]. 2023;128:103005. Available from: https://doi.org/10.1016/j.jprocont.2023.103005
• 21. Alcala CF, Qin SJ. Reconstruction-based contribution for process monitoring with kernel principal component analysis. Proc 2010 Am Control Conf ACC 2010. 2010;7022–7. https://doi.org/10.1109/ACC.2010.5531315
• 22. Yu W, Zhao C. Robust monitoring and fault isolation of nonlinear industrial processes using denoising autoencoder and elasticnet. IEEE Trans Control Syst Technol. 2020;28(3):1083–91. https://doi.org/10.1109/TCST.2019.2897946
• 23. Li S, Luo J, Hu Y. Nonlinear process modeling via unidimensional convolutional neural networks with self-attention on global and local inter-variable structures and its application to process monitoring. ISA Trans [Internet]. 2022;121:105–18. Available from: https://doi.org/10.1016/j.isatra.2021.04.014
• 24. Hines JW, Wrest DJ, Uhrig RE. ERATION MONITO. JW Hines, DJ Wrest, RE Uhrig, Plant wide Sens calibration Monit Proc IEEE Internatinal Symp Intell Control IEEE, 1996, pp 378–383 . 1996;0–5.
• 25. Pawlik P, Kania K, Przysucha B. Eksploatacja i Niezawodnosc –Maintenance and Reliability neural network not requiring training data from a faulty machine. 2023;25(3):0–3. https://doi.org/10.17531/ein/168109
• 26. Ren S, Si F, Cao Y. Development of Input Training Neural Networks for Multiple Sensor Fault Isolation. IEEE Sens J. 2022;22(15):14997–5009 https://doi.org/10.1109/JSEN.2022.3184078
• 27. Ren S, Jin Y, Zhao J, Cao Y, Si F. Nonlinear process monitoring based on generic reconstruction-based auto-associative neural network. J Franklin Inst [Internet]. 2023;360(7):5149–70. Available from: https://doi.org/10.1016/j.jfranklin.2023.03.041
• 28. Chen S, Yu J, Wang S. One-dimensional convolutional auto-encoder-based feature learning for fault diagnosis of multivariate processes. J Process Control [Internet]. 2020;87:54–67. Available from: https://doi.org/10.1016/j.jprocont.2020.01.004
• 29. Li X, Su K, He Q, Wang X, Xie Z. Research on Fault Diagnosis of Highway Bi-LSTM Based on Attention Mechanism Indexed by. Eksploat i Niezawodn. 2023;25(2):0–3. https://doi.org/10.17531/ein/162937
• 30. Özmen NG. Worm gear condition monitoring and fault detection from thermal images via deep learning method Monitorowanie stanui wykrywanie błędów przekładni ślimakowej na podstawie termogramów z wykorzystaniem metody głębokiego uczenia. 2020;22(3):544–56. https://doi.org/10.17531/ein.2020.3.18
• 31. Liu J, Wu J, Xie Y, Jie W, Xu P, Tang Z, et al. Toward robust process monitoring of complex process industries based on denoising sparse auto-encoder. J Ind Inf Integr [Internet]. 2022;30(November):100410. Available from: https://doi.org/10.1016/j.jii.2022.100410
• 32. Dunia R, Joe Qin S. A unified geometric approach to process and sensor fault identification and reconstruction: The unidimensional fault case. Comput Chem Eng. 1998;22(7–8):927–43. https://doi.org/10.1016/S0098-1354(97)00277-9