Machine learning based online fault prognostics for nonstationary industrial process via degradation feature extraction and temporal smoothness analysis

Abstract

Fault degradation prognostic, which estimates the time before a failure occurs and process breakdowns, has been recognized as a key component in maintenance strategies nowadays. Fault degradation processes are, in general, slowly varying and can be modeled by autoregressive models. However, industrial processes always show typical nonstationary nature, which may bring two challenges: how to capture fault degradation information and how to model nonstationary processes. To address the critical issues, a novel fault degradation modeling and online fault prognostic strategy is developed in this paper. First, a fault degradation-oriented slow feature analysis (FDSFA) algorithm is proposed to extract fault degradation directions along which candidate fault degradation features are extracted. The trend ability assessment is then applied to select major fault degradation features. Second, a key fault degradation factor (KFDF) is calculated to characterize the fault degradation tendency by combining major fault degradation features and their stability weighting factors. After that, a time-varying regression model with temporal smoothness regularization is established considering nonstationary characteristics. On the basis of updating strategy, an online fault prognostic model is further developed by analyzing and modeling the prediction errors. The performance of the proposed method is illustrated with a real industrial process.

Keywords

fault prognostic / nonstationary / industrial process / fault degradation-oriented slow feature analysis(FDSFA) / temporal smoothness regularization

References

[1] RamezaniS, MoiniA, RiahiM, MarquezA C. A model to determining the remaining useful life of rotating equipment, based on a new approach to determining state of degradation[J]. Journal of Central South University, 2020, 27(8): 2291-2310

[2] ZhaoC, HuangB. A full-condition monitoring method for nonstationary dynamic chemical processes with cointegration and slow feature analysis [J]. AIChE Journal, 2018, 64(5): 1662-1681

[3] ChaiZ, ZhaoC. Enhanced random forest with concurrent analysis of static and dynamic nodes for industrial fault classification [J]. IEEE Transactions on Industrial Informatics, 2020, 16(1): 54-66

[4] DaiY, ChengS, GanQ, YuT, WuX, BiF. Life prediction of Ni-Cd battery based on linear Wiener process [J]. Journal of Central South University, 2021, 28(9): 2919-2930

[5] ZhaoC, GaoF. Online fault prognosis with relative deviation analysis and vector autoregressive modeling [J]. Chemical Engineering Science, 2015, 138: 531-543

[6] CosmeL B, CaminhasW M, D’angeloM F S V, PalharesR M. A novel fault-prognostic approach based on interacting multiple model filters and fuzzy systems [J]. IEEE Transactions on Industrial Electronics, 2019, 66(1): 519-528

[7] LiN, LeiY, LinJ, DingS X. An improved exponential model for predicting remaining useful life of rolling element bearings [J]. IEEE Transactions on Industrial Electronics, 2015, 62(12): 7762-7773

[8] WangY, PengY, ZiY, JinX, TsuiK L. A two-stage data-driven-based prognostic approach for bearing degradation problem [J]. IEEE Transactions on Industrial Informatics, 2016, 12(3): 924-932

[9] AdedigbaS A, KhanF, YangM. Dynamic failure analysis of process systems using principal component analysis and Bayesian network [J]. Industrial & Engineering Chemistry Research, 2017, 56(8): 2094-2106

[10] ZhangS, ZhaoC, HuangB. Simultaneous static and dynamic analysis for fine-scale identification of process operation statuses [J]. IEEE Transactions on Industrial Informatics, 2019, 15(9): 5320-5329

[11] DongS, LuoT. Bearing degradation process prediction based on the PCA and optimized LS-SVM model [J]. Measurement, 2013, 46(9): 3143-3152

[12] WANG Wen-yi. Toward dynamic model-based prognostics for transmission gears [C]//AeroSense 2002. Proc SPIE 4733, Component and Systems Diagnostics, Prognostics, and Health Management II, Orlando, FL, USA. 2002, 4733: 157–167. DOI: https://doi.org/10.1117/12.475505.

[13] LiG, QinS J, JiY, ZhouD. Reconstruction based fault prognosis for continuous processes [J]. Control Engineering Practice, 2010, 18(10): 1211-1219

[14] ZhaoC, SunY. Subspace decomposition approach of fault deviations and its application to fault reconstruction [J]. Control Engineering Practice, 2013, 21(10): 1396-1409

[15] ZhaoC, GaoF. Fault subspace selection approach combined with analysis of relative changes for reconstruction modeling and multifault diagnosis [J]. IEEE Transactions on Control Systems Technology, 2016, 24(3): 928-939

[16] ZhaoC, GaoF. Critical-to-fault-degradation variable analysis and direction extraction for online fault prognostic [J]. IEEE Transactions on Control Systems Technology, 2017, 25(3): 842-854

[17] JiaX, ZhaoM, DiY, YangQ, LeeJ. Assessment of data suitability for machine prognosis using maximum mean discrepancy [J]. IEEE Transactions on Industrial Electronics, 2018, 65(7): 5872-5881

[18] LiaoL, JinW, PavelR. Enhanced restricted boltzmann machine with prognosability regularization for prognostics and health assessment [J]. IEEE Transactions on Industrial Electronics, 2016, 63(11): 7076-7083

[19] MannH B. Nonparametric tests against trend [J]. Econometrica, 1945, 13(3): 245

[20] LuY, LiQ, PanZ, LiangS Y. Prognosis of bearing degradation using gradient variable forgetting factor RLS combined with time series model [J]. IEEE Access, 2018, 6: 10986-10995

[21] ByonE, ChoeY, YampikulsakulN. Adaptive learning in time-variant processes with application to wind power systems [J]. IEEE Transactions on Automation Science and Engineering, 2016, 13(2): 997-1007

[22] LouZ, WangY. Multimode continuous processes monitoring based on hidden semi-Markov model and principal component analysis [J]. Industrial & Engineering Chemistry Research, 2017, 56(46): 13800-13811

[23] LiX, ZhangW, DingQ. Deep learning-based remaining useful life estimation of bearings using multi-scale feature extraction [J]. Reliability Engineering & System Safety, 2019, 182208-218

[24] BringmannL F, HamakerE L, VigoD E, AubertA, BorsboomD, TuerlinckxF. Changing dynamics: Time-varying autoregressive models using generalized additive modeling [J]. Psychological Methods, 2017, 22(3): 409-425

[25] WU Wei, HU Jing-tao, ZHANG Ji-long. Prognostics of machine health condition using an improved ARIMA-based Prediction method [C]//2007 2nd IEEE Conference on Industrial Electronics and Applications. 2007, Harbin, China. IEEE, 2007: 1062–1067. DOI: https://doi.org/10.1109/ICIEA.2007.4318571.

[26] GaoX, YangF, HuangD. Model quality assessment and model mismatch detection: A temporal smoothness regularization approach [J]. IFAC-Papers OnLine, 2018, 51(18): 1-6

[27] HU Yun-yun, ZHAO Chun-hui. Online fault prognostics based on degradation-oriented slow feature analysis and temporal smoothness analysis [C]//2019 12th Asian Control Conference (ASCC). IEEE, 2019: 844–849.

[28] ShangC, HuangB, YangF, HuangD. Probabilistic slow feature analysis-based representation learning from massive process data for soft sensor modeling [J]. AIChE Journal, 2015, 61(12): 4126-4139

[29] HamedK H, RamachandraR A. A modified Mann-Kendall trend test for autocorrelated data [J]. Journal of Hydrology, 1998, 204(1–4): 182-196

[30] YueS, WangC Y. Regional streamflow trend detection with consideration of both temporal and spatial correlation [J]. International Journal of Climatology, 2002, 22(8): 933-946

[31] KendallM G. A new measure of rank correlation [J]. Biometrika, 1938, 30(12): 81-93

[32] ZouH, HastieT. Regularization and variable selection via the elastic net [J]. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 2005, 67(2): 301-320

[33] YangX, MaX, KangN, MaihemutiM. Probability interval prediction of wind power based on KDE method with rough sets and weighted Markov chain [J]. IEEE Access, 2018, 6: 51556-51565

[34] ChaiT, DraxlerR R. Root mean square error (RMSE) or mean absolute error (MAE)?-Arguments against avoiding RMSE in the literature [J]. Geoscientific Model Development, 2014, 7(3): 1247-1250