A blast furnace fault monitoring algorithm with low false alarm rate: Ensemble of greedy dynamic principal component analysis-Gaussian mixture model

doi:10.1016/j.cjche.2022.09.012

Abstract

Abstract: The large blast furnace is essential equipment in the process of iron and steel manufacturing. Due to the complex operation process and frequent fluctuations of variables, conventional monitoring methods often bring false alarms. To address the above problem, an ensemble of greedy dynamic principal component analysis-Gaussian mixture model (EGDPCA-GMM) is proposed in this paper. First, PCA-GMM is introduced to deal with the collinearity and the non-Gaussian distribution of blast furnace data. Second, in order to explain the dynamics of data, the greedy algorithm is used to determine the extended variables and their corresponding time lags, so as to avoid introducing unnecessary noise. Then the bagging ensemble is adopted to cooperate with greedy extension to eliminate the randomness brought by the greedy algorithm and further reduce the false alarm rate (FAR) of monitoring results. Finally, the algorithm is applied to the blast furnace of a large iron and steel group in South China to verify performance. Compared with the basic algorithms, the proposed method achieves lowest FAR, while keeping missed alarm rate (MAR) remain stable.

Key words: Chemical processes, Principal component analysis, Gaussian mixture model, Process monitoring, Ensemble, Process control

摘要： The large blast furnace is essential equipment in the process of iron and steel manufacturing. Due to the complex operation process and frequent fluctuations of variables, conventional monitoring methods often bring false alarms. To address the above problem, an ensemble of greedy dynamic principal component analysis-Gaussian mixture model (EGDPCA-GMM) is proposed in this paper. First, PCA-GMM is introduced to deal with the collinearity and the non-Gaussian distribution of blast furnace data. Second, in order to explain the dynamics of data, the greedy algorithm is used to determine the extended variables and their corresponding time lags, so as to avoid introducing unnecessary noise. Then the bagging ensemble is adopted to cooperate with greedy extension to eliminate the randomness brought by the greedy algorithm and further reduce the false alarm rate (FAR) of monitoring results. Finally, the algorithm is applied to the blast furnace of a large iron and steel group in South China to verify performance. Compared with the basic algorithms, the proposed method achieves lowest FAR, while keeping missed alarm rate (MAR) remain stable.

关键词: Chemical processes, Principal component analysis, Gaussian mixture model, Process monitoring, Ensemble, Process control

Xiongzhuo Zhu, Dali Gao, Chong Yang, Chunjie Yang. A blast furnace fault monitoring algorithm with low false alarm rate: Ensemble of greedy dynamic principal component analysis-Gaussian mixture model[J]. Chinese Journal of Chemical Engineering, 2023, 57(5): 151-161.

References

[1] H.W. Zhang, M.Y. Chen, J. Shang, C.J. Yang, Y.X. Sun, Stochastic process-based degradation modeling and RUL prediction: from Brownian motion to fractional Brownian motion, Sci. China Inf. Sci. 64 (7) (2021) 1–20.
[2] H. Zhou, H.F. Zhang, C.J. Yang, Hybrid-model-based intelligent optimization of ironmaking process, IEEE Trans. Ind. Electron. 67 (3) (2020) 2469–2479.
[3] R. An, C. Yang, Y. Pan, Unsupervised change point detection using a weight graph method for process monitoring, Ind. Eng. Chem. Res. 58 (2019) 1624–1634.
[4] T.S. Zhang, W. Wang, H. Ye, D.X. Huang, H.F. Zhang, M.L. Li, Fault detection for ironmaking process based on stacked denoising autoencoders, 2016 American Control Conference (ACC). Boston, MA. IEEE, 3261–3267.
[5] L. Wang, C.J. Yang, Y.X. Sun, H.F. Zhang, M.L. Li, Effective variable selection and moving window HMM-based approach for iron-making process monitoring, J. Process. Control 68 (2018) 86–95.
[6] Y. Pan, C. Yang, R. An, Y. Sun, Robust principal component pursuit for fault detection in a blast furnace process, Ind. Eng. Chem. Res. 57 (2018) 283–291.
[7] B. Zhou, H. Ye, H.F. Zhang, M.L. Li, Process monitoring of iron-making process in a blast furnace with PCA-based methods, Control Eng. Pract. 47 (2016) 1–14.
[8] J. Shang, M. Chen, H. Ji, D. Zhou, H. Zhang, M. Li, Dominant trend based logistic regression for fault diagnosis in nonstationary processes, Control Eng. Pract. 66 (2017) 156–168.
[9] J. Shang, M. Chen, H. Zhang, H. Ji, D. Zhou, H. Zhang, M. Li, Increment-based recursive transformed component statistical analysis for monitoring blast furnace iron-making processes: An index-switching scheme, Control Eng. Pract. 77 (2018) 190–200.
[10] W.F. Ku, R.H. Storer, C. Georgakis, Disturbance detection and isolation by dynamic principal component analysis, Chemom. Intell. Lab. Syst. 30 (1) (1995) 179–196.
[11] S.W. Choi, I.B. Lee, Nonlinear dynamic process monitoring based on dynamic kernel PCA, Chem. Eng. Sci. 59 (24) (2004) 5897–5908.
[12] Y.N. Dong, S.J. Qin, A novel dynamic PCA algorithm for dynamic data modeling and process monitoring, J. Process. Control 67 (2018) 1–11.
[13] T.J. Rato, M.S. Reis, Defining the structure of DPCA models and its impact on process monitoring and prediction activities, Chemom. Intell. Lab. Syst. 125 (2013) 74–86.
[14] Z. Ge, Z. Song, F. Gao, Review of recent research on data-based process monitoring, Ind. Eng. Chem. Res. 52 (2013) 3543–3562.
[15] R.F. Li, X.Z. Wang, Dimension reduction of process dynamic trends using independent component analysis, Comput. Chem. Eng. 26 (3) (2002) 467–473.
[16] M. Kano, S. Tanaka, S. Hasebe, I. Hashimoto, H. Ohno, Combined multivariate statistical process control, IFAC Proc. Vol. 37 (1) (2004) 281–286.
[17] Z. Ge, Z. Song, Process monitoring based on independent component analysis- principal component analysis (ica-pca) and similarity factors, Ind. Eng. Chem. Res. 46 (2007) 2054–2063.
[18] T. Chen, J. Zhang, On-line multivariate statistical monitoring of batch processes using Gaussian mixture model, Comput. Chem. Eng. 34 (4) (2010) 500–507.
[19] X. Xie, H. Shi, Dynamic multimode process modeling and monitoring using adaptive Gaussian mixture models, Ind. Eng. Chem. Res. 51 (2012) 5497–5505.
[20] S.T. Chen, Q.C. Jiang, X.F. Yan, Multimodal process monitoring based on transition-constrained Gaussian mixture model, Chin. J. Chem. Eng. 28 (12) (2020) 3070–3078.
[21] S.W. Choi, J.H. Park, I.B. Lee, Process monitoring using a Gaussian mixture model via principal component analysis and discriminant analysis, Comput. Chem. Eng. 28 (8) (2004) 1377–1387.
[22] J. Yu, S.J. Qin, Multimode process monitoring with Bayesian inference-based finite Gaussian mixture models, AIChE J. 54 (7) (2008) 1811–1829.
[23] X.D. Jiang, H.T. Zhao, B. Jin, Multimode process monitoring based on sparse principal component selection and Bayesian inference-based probability, Math. Probl. Eng. 2015 (2015) 465372.
[24] K. Pearson, LIII. On lines and planes of closest fit to systems of points in space. The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science, 2(11) (1901) 559–572.
[25] H. Hotelling, Analysis of a complex of statistical variables into principal components, Journal of Educational Psychology, 24(6) (1933) 417-441.
[26] J. E. Jackson, Quality control methods for several related variables, Technometrics 1 (1959) 359–377.
[27] J. Zheng, Q. Wen, Z. Song, Recursive Gaussian mixture models for adaptive process monitoring, Ind. Eng. Chem. Res. 58 (2019) 6551–6561.
[28] J.M. Lee, C. Yoo, I.B. Lee, Statistical process monitoring with independent component analysis, J. Process. Control 14 (5) (2004) 467–485.
[29] J.M. Lee, C. Yoo, I.B. Lee, Statistical monitoring of dynamic processes based on dynamic independent component analysis, Chem. Eng. Sci. 59 (14) (2004) 2995–3006.