The complexity of multi-mode industrial processes poses challenges for traditional fault diagnosis methods in achieving satisfactory results. This challenge arises from the non-stationarity caused by changes in modes, which often overlaps with temporal variations in fault generation processes. To address this issue, we propose a fault diagnosis method based on multi-mode categorization and dynamic transfer entropy graph analysis. This method can uncover the causal information changes resulting from both mode switches and fault occurrence. Firstly, modes are categorized using the criterion of maximum intra-segment similarity and minimum inter-segment similarity, and a static causal graph is constructed for each mode. Secondly, a framework for dynamic transfer entropy graph is established to extract short-term causal information transfer using transfer entropy and compare it with the static causal graph of normal modes. Time points with significant changes in causal relationships are selected as fault time points for further analysis. Finally, an anomaly score is designed to identify critical fault nodes by selecting nodes with significant causal changes before the fault time point, thereby accurately locating the fault root nodes. To validate the performance of the proposed method, a case study using the Tennessee Eastman process is provided. Simulation results demonstrate the effectiveness and feasibility of the proposed method.
GoelPDattaAMannanMS.Industrial alarm systems: challenges and opportunities. J Loss Prev Process Ind2017; 50: 23–36.
2.
LindnerBAuretLBauerM, et al. Comparative analysis of granger causality and transfer entropy to present a decision flow for the application of oscillation diagnosis. J Process Control2019; 79: 72–84.
3.
LiuYChenHSWuH, et al. Simplified granger causality map for data-driven root cause diagnosis of process disturbances. J Process Control2020; 95: 45–54.
4.
HuWWangJChenT, et al. Cause-effect analysis of industrial alarm variables using transfer entropies. Control Eng Pract2017; 64: 205–214.
5.
TianCZhaoCFanH, et al. Causal network construction based on convergent cross mapping (ccm) for alarm system root cause tracing of nonlinear industrial process. IFAC-PapersOnLine2020; 53(2): 13619–13624.
6.
JiangQWangZYanS, et al. Data-driven soft sensing for batch processes using neural network-based deep quality-relevant representation learning. IEEE Trans Artif Intell2023; 4(4): 602–611.
7.
ZhouYCaoZLuJ, et al. Objectives, challenges, and prospects of batch processes: arising from injection molding applications. Korean J Chem Eng2022; 39(12): 3179–3189.
8.
YaoLGeZ.Industrial big data modeling and monitoring framework for plant-wide processes. IEEE Trans Ind Inform2021; 17(9): 6399–6408.
9.
HuangKWeiKZhouL, et al. Multimode process monitoring and mode identification based on multiple dictionary learning. IEEE Trans Instrum Meas2021; 70: 1–12.
10.
WangMZhouDChenM.Recursive hybrid variable monitoring for fault detection in nonstationary industrial processes. IEEE Trans Ind Inform2022; 18(10): 7296–7304.
11.
ZhuJGeZSongZ.Non-gaussian industrial process monitoring with probabilistic independent component analysis. IEEE Trans Autom Sci Eng2017; 14(2): 1309–1319.
12.
LuoLBaoSTongC.Sparse robust principal component analysis with applications to fault detection and diagnosis. Ind Eng Chem Res2019; 58(3): 1300–1309.
13.
ShaoWHanWLiY, et al. Enhancing the reliability and accuracy of data-driven dynamic soft sensor based on selective dynamic partial least squares models. Control Eng Pract2022; 127: 105292.
14.
FentonWGMcGinnityTMMaguireLP.Fault diagnosis of electronic systems using intelligent techniques: a review. IEEE Trans Syst Man Cybern C2001; 31(3): 269–281.
15.
SiYWangYZhouD.Key-performance-indicator-related process monitoring based on improved kernel partial least squares. IEEE Trans Ind Electron2021; 68(3): 2626–2636.
16.
DengZChenXXieS, et al. Semi-supervised discriminative projective dictionary pair learning and its application to industrial process. IEEE Trans Ind Inform2023; 19(3): 3119–3132.
17.
PengGHuangKWangH.Dynamic multimode process monitoring using recursive gmm and kpca in a hot rolling mill process. Syst Sci Control Eng2021; 9(1): 592–601.
18.
HuangKWuYYangC, et al. Structure dictionary learning-based multimode process monitoring and its application to aluminum electrolysis process. IEEE Trans Autom Sci Eng2020; 17(4): 1989–2003.
Quiñones-GrueiroMPrieto-MorenoAVerdeC, et al. Data-driven monitoring of multimode continuous processes: a review. Chemometr Intell Lab Syst2019; 189: 56–71.
21.
HwangDHHanC.Real-time monitoring for a process with multiple operating modes. Control Eng Pract1999; 7(7): 891–902.
22.
JiCMaFWangJ, et al. Statistical method based on dissimilarity of variable correlations for multimode chemical process monitoring with transitions. Process Saf Environ Prot2022; 162: 649–662.
23.
ZhuZSongZPalazogluA.Process pattern construction and multi-mode monitoring. J Process Control2012; 22(1): 247–262.
24.
ZhouLZhengJGeZ, et al. Multimode process monitoring based on switching autoregressive dynamic latent variable model. IEEE Trans Ind Electron2018; 65(10): 8184–8194.
25.
TongCEl-FarraNHPalazogluA, et al. Fault detection and isolation in hybrid process systems using a combined data-driven and observer-design methodology. AIChE J2014; 60(8): 2805–2814.
26.
PengXTangYDuW, et al. Multimode process monitoring and fault detection: a sparse modeling and dictionary learning method. IEEE Trans Ind Electron2017; 64(6): 4866–4875.
27.
ZhaoCChenJJingH.Condition-driven data analytics and monitoring for wide-range nonstationary and transient continuous processes. IEEE Trans Autom Sci Eng2021; 18(4): 1563–1574.
28.
LiuYZengJBaoJ, et al. A unified probabilistic monitoring framework for multimode processes based on probabilistic linear discriminant analysis. IEEE Trans Ind Inform2020; 16(10): 6291–6300.
29.
WuDZhouDZhangJ, et al. Multimode process monitoring based on fault dependent variable selection and moving window-negative log likelihood probability. Comput Chem Eng2020; 136: 106787.
30.
YangFXiaoD.Progress in root cause and fault propagation analysis of large-scale industrial processes. J Contr Sci Eng2012; 2012: 1–10.
31.
YangFDuanPShahSL, et al. Capturing connectivity and causality in complex industrial processes. Springer Science & Business Media, 2014.
32.
LiuFFZuoHFSuY, et al. Review of research progresses for graph-based models in fault diagnosis method. China Mech Eng2013; 24(5): 696.
33.
HararyF.On the notion of balance of a signed graph. Mich Math J1953; 2(2): 143–146.
34.
DahlstrandF.Consequence analysis theory for alarm analysis. Knowl Syst2002; 15(1–2): 27–36.
35.
YimSYAnanthakumarHGBenabbasL, et al. Using process topology in plant-wide control loop performance assessment. Comput Chem Eng2006; 31(2): 86–99.
36.
HeYLTianYXuY, et al. Novel soft sensor development using echo state network integrated with singular value decomposition: application to complex chemical processes. Chemometr Intell Lab Syst2020; 200: 103981.
37.
GrangerCWJ. Investigating causal relations by econometric models and cross-spectral methods. Econometrica1969; 37: 424–438.
38.
SchreiberT.Measuring information transfer. Phys Rev Lett2000; 85(2): 461–464.
39.
BauerMCoxJWCavenessMH, et al. Finding the direction of disturbance propagation in a chemical process using transfer entropy. IEEE Trans Control Syst Technol2007; 15(1): 12–21.
40.
BauerMThornhillNF.A practical method for identifying the propagation path of plant-wide disturbances. J Process Control2008; 18(7–8): 707–719.
41.
BauerMCoxJWCavenessMH, et al. Nearest neighbors methods for root cause analysis of plantwide disturbances. Ind Eng Chem Res2007; 46(18): 5977–5984.
42.
CaiBHuangLXieM.Bayesian networks in fault diagnosis. IEEE Trans Ind Inform2017; 13(5): 2227–2240.
43.
LiGQinSJYuanT.Data-driven root cause diagnosis of faults in process industries. Chemometr Intell Lab Syst2016; 159: 1–11.
DuanPYangFChenT, et al. Direct causality detection via the transfer entropy approach. IEEE Trans Control Syst Technol2013; 21(6): 2052–2066.
46.
LobierMSiebenhühnerFPalvaS, et al. Phase transfer entropy: a novel phase-based measure for directed connectivity in networks coupled by oscillatory interactions. Neuroimage2014; 85(Pt 2): 853–872.
47.
ZhaoXSunYLiX, et al. Multiscale transfer entropy: measuring information transfer on multiple time scales. Commun Nonlinear Sci Numer Simul2018; 62: 202–212.
48.
ZhangNZhaoX.Quantile transfer entropy: measuring the heterogeneous information transfer of nonlinear time series. Commun Nonlinear Sci Numer Simul2022; 111: 106505.
49.
StaniekMLehnertzK.Symbolic transfer entropy. Phys Rev Lett2008; 100(15): 158101.
50.
DuanSZhaoCWuM.Multiscale partial symbolic transfer entropy for time-delay root cause diagnosis in nonstationary industrial processes. IEEE Trans Ind Electron2023; 70(2): 2015–2025.
51.
ZhangXHuWYangF, et al. A new transfer entropy approach based on information granulation and clustering for root cause analysis. Control Eng Pract2023; 140: 105669.
52.
LiuDChenXPengD.Some cosine similarity measures and distance measures between q-rung orthopair fuzzy sets. Int J Intell Syst2019; 34(7): 1572–1587.
53.
QiuLYangH.Transfer entropy calculation for short time sequences with application to stock markets. Physica A2020; 559: 125121.
54.
DownsJJVogelEF.A plant-wide industrial process control problem. Comput Chem Eng1993; 17(3): 245–255.
55.
YanhongXChengaoSYuanL.Fault monitoring of batch process based on moving window SVDD. Inf Control2015; 44(5): 531–537.
56.
ChenHSYanZYaoY, et al. Systematic procedure for granger-causality-based root cause diagnosis of chemical process faults. Ind Eng Chem Res2018; 57(29): 9500–9512.
57.
SongPZhaoCHuangB, et al. Explicit representation and customized fault isolation framework for learning temporal and spatial dependencies in industrial processes. IEEE Trans Neural Netw Learn Syst2024; 35(3): 2997–3011.
ParkDHoshiYKempCC.A multimodal anomaly detector for robot-assisted feeding using an LSTM-based variational autoencoder. IEEE Robot Autom Lett2018; 3(3): 1544–1551.
60.
ArnoldALiuYAbeN. Temporal causal modeling with graphical granger methods. In: Proceedings of the 13th ACM SIGKDD international conference on Knowledge discovery and data mining, 2007, pp. 66–75.
61.
SongPZhaoCHuangB.Mpge and rootrank: a sufficient root cause characterization and quantification framework for industrial process faults. Neural Netw2023; 161: 397–417.
62.
SongPZhaoCHuangB, et al. Sparse and time-varying predictive relation extraction for root cause quantification of nonstationary process faults. IEEE Trans Instrum Meas2022; 71: 1–13.
63.
ZhangJZhouDChenM, et al. Data-driven anomaly monitoring methods for multimode processes: overview and perspectives. Sci Sin Informations2023; 53: 2087–2106.
64.
ElshenawyLMYinSNaikAS, et al. Efficient recursive principal component analysis algorithms for process monitoring. Ind Eng Chem Res2010; 49(1): 252–259.
