Early identification of process deviation based on convolutional neural network

doi:10.1016/j.cjche.2022.07.034

Abstract

Abstract: A novel process monitoring method based on convolutional neural network (CNN) is proposed and applied to detect faults in industrial process. By utilizing the CNN algorithm, cross-correlation and autocorrelation among variables are captured to establish a prediction model for each process variable to approximate the first-principle of physical/chemical relationships among different variables under normal operating conditions. When the process is operated under pre-set operating conditions, prediction residuals can be assumed as noise if a proper model is employed. Once process faults occur, the residuals will increase due to the changes of correlation among variables. A principal component analysis (PCA) model based on the residuals is established to realize process monitoring. By monitoring the changes in main feature of prediction residuals, the faults can be promptly detected. Case studies on a numerical nonlinear example and data from two industrial processes are presented to validate the performance of process monitoring based on CNN.

Key words: Process monitoring, Residual, Principal component analysis, Process systems, Systems engineering

摘要： A novel process monitoring method based on convolutional neural network (CNN) is proposed and applied to detect faults in industrial process. By utilizing the CNN algorithm, cross-correlation and autocorrelation among variables are captured to establish a prediction model for each process variable to approximate the first-principle of physical/chemical relationships among different variables under normal operating conditions. When the process is operated under pre-set operating conditions, prediction residuals can be assumed as noise if a proper model is employed. Once process faults occur, the residuals will increase due to the changes of correlation among variables. A principal component analysis (PCA) model based on the residuals is established to realize process monitoring. By monitoring the changes in main feature of prediction residuals, the faults can be promptly detected. Case studies on a numerical nonlinear example and data from two industrial processes are presented to validate the performance of process monitoring based on CNN.

关键词: Process monitoring, Residual, Principal component analysis, Process systems, Systems engineering

Fangyuan Ma, Cheng Ji, Jingde Wang, Wei Sun. Early identification of process deviation based on convolutional neural network[J]. Chinese Journal of Chemical Engineering, 2023, 56(4): 104-118.

Fangyuan Ma, Cheng Ji, Jingde Wang, Wei Sun. Early identification of process deviation based on convolutional neural network[J]. 中国化学工程学报, 2023, 56(4): 104-118.

References

[1] V. Venkatasubramanian, R. Rengaswamy, K. Yin, S.N. Kavuri, A review of process fault detection and diagnosis: Part I: Quantitative model-based methods, Comput. Chem. Eng. 27 (3) (2003) 293-311.
[2] V. Venkatasubramanian, R. Rengaswamy, S.N. Kavuri, A review of process fault detection and diagnosis: Part II: Qualitative models and search strategies, Comput. Chem. Eng. 27 (3) (2003) 313–326.
[3] V. Venkatasubramanian, R. Rengaswamy, K. Yin, S.N. Kavuri, A review of process fault detection and diagnosis: Part I: Quantitative model-based methods, Comput. Chem. Eng 27 (3) (2003) 293–311.
[4] L.H. Chiang, E.L. Russell, R.D. Braatz, Fault Detection and Diagnosis in Industrial Systems, Springer, London, 2001.
[5] Z.Q. Ge, Z.H. Song, F.R. Gao, Review of recent research on data-based process monitoring, Ind. Eng. Chem. Res. 52 (10) (2013) 3543–3562.
[6] S. Yin, S.X. Ding, X.C. Xie, H. Luo, A review on basic data-driven approaches for industrial process monitoring, IEEE Trans. Ind. Electron. 61 (11) (2014) 6418–6428.
[7] S.J. Qin, Survey on data-driven industrial process monitoring and diagnosis, Annu. Rev. Control 36 (2) (2012) 220–234.
[8] X.G. Deng, X.M. Tian, S. Chen, C.J. Harris, Fault discriminant enhanced kernel principal component analysis incorporating prior fault information for monitoring nonlinear processes, Chemom. Intell. Lab. Syst. 162 (2017) 21–34.
[9] X. Zhang, W.W. Yan, H.H. Shao, Nonlinear multivariate quality estimation and prediction based on kernel partial least squares, Ind. Eng. Chem. Res. 47 (4) (2008) 1120–1131.
[10] 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.
[11] Y.N. Dong, S.J. Qin, Regression on dynamic PLS structures for supervised learning of dynamic data, J. Process. Control 68 (2018) 64–72.
[12] I. T. Jolliffe, Principal component analysis. Springer, New York, 2022.
[13] C.D. Tong, T. Lan, H.Z. Yu, X. Peng, Distributed partial least squares based residual generation for statistical process monitoring, J. Process. Control 75 (2019) 77–85.
[14] F. Ma, D. Lin, M. Xu, J. Wang, W. Sun, Early identification of small shift in process unit based on multivariate statistical method, Chem. Ind. & Eng. Pro. 39 (4) (2020) 1267-1272.
[15] T. Lan, C.D. Tong, H.Z. Yu, X.H. Shi, L.J. Luo, Nonlinear process monitoring based on decentralized generalized regression neural networks, Expert Syst. Appl. 150 (2020) 113273.
[16] H. Wu, J.S. Zhao, Self-adaptive deep learning for multimode process monitoring, Comput. Chem. Eng. 141 (2020) 107024.
[17] Y. LeCun, Y. Bengio, G. Hinton, Deep learning, Nature 521 (7553) (2015) 436–444.
[18] M.T. McCann, K.H. Jin, M. Unser, Convolutional neural networks for inverse problems in imaging: A review, IEEE Signal Process. Mag. 34 (6) (2017) 85–95.
[19] Y. Zhang, B. Wallace, A sensitivity analysis of (and practitioners' guide to) convolutional neural networks for sentence classification, arXiv preprint (2015) arXiv:1510.03820.
[20] U. Pak, J. Ma, U. Ryu, K. Ryom, U. Juhyok, K. Pak, C. Pak, Deep learning-based PM_2.5 prediction considering the spatiotemporal correlations: A case study of Beijing, China, Sci. Total. Environ. 699 (2020) 133561.
[21] D. Cannizzaro, A. Aliberti, L. Bottaccioli, E. Macii, A. Acquaviva, E. Patti, Solar radiation forecasting based on convolutional neural network and ensemble learning, Expert Syst. Appl. 181 (2021) 115167.
[22] W. Zhu, Y. Ma, Y. Zhou, J. Romagnoli, Deep learning based soft sensor and its application on a pyrolysis reactor for compositions predictions of gas phase components, In: The 13th International Symposium on Process Systems Engineering (PSE 2018), San Diego, USA, 2018.
[23] A. Krizhevsy, I. Sutskever, G.E. Hinton, ImageNet classification with deep convolutional neural networks, in: International Conference on Neural Information Processing Systems (NIPS), Doha, Qatar, 2012.
[24] M. Rastegari, V. Ordonez, J. Redmon, A. Farhadi, XNOR-net: ImageNet classification using binary convolutional neural networks, Comput. Vis. – ECCV 2016 (2016): 525–542.
[25] A.A.M. Al-Saffar, H. Tao, M.A. Talab, Review of deep convolution neural network in image classification, 2017 International Conference on Radar, Antenna, Microwave, Electronics, and Telecommunications (ICRAMET). Jakarta, Indonesia. IEEE, 26–31.
[26] T. He, R. Kong, A.J. Holmes, M. Nguyen, M.R. Sabuncu, S.B. Eickhoff, D. Bzdok, J. Feng, B. Yeo, Deep neural networks and kernel regression achieve comparable accuracies for functional connectivity prediction of behavior and demographics, NeuroImage 206 (2020) 116276.
[27] F.Y. Ma, D.X. Lin, J.T. Zhong, X.Y. Han, J.D. Wang, W. Sun, CNN based process monitoring of spatially distributed system, 2019 CAA Symposium on Fault Detection, Supervision and Safety for Technical Processes (SAFEPROCESS). Xiamen, China. IEEE, 695–700.
[28] T. Kourti, J.F. MacGregor, Multivariate SPC methods for process and product monitoring, J. Qual. Technol. 28 (4) (1996) 409–428.
[29] X.Y. Han, S.W. Tian, J.A. Romagnoli, H. Li, W. Sun, PCA-SDG based process monitoring and fault diagnosis: Application to an industrial pyrolysis furnace, IFAC-PapersOnLine 51 (18) (2018) 482–487.
[30] D.Y. Tzeng, R.S. Berns, A review of principal component analysis and its applications to color technology, Color Res. Appl. 30 (2) (2005) 84–98.
[31] J.W. Yan, X.X. Tian, Q. Zhou, Y.C. Yang, Improvement of scanlan's nonlinear model based on residual analysis, KSCE J. Civ. Eng. 23 (1) (2019) 280–286.
[32] J. Shang, M.Y. Chen, H.W. Zhang, Fault detection based on augmented kernel Mahalanobis distance for nonlinear dynamic processes, Comput. Chem. Eng. 109 (2018) 311–321.
[33] J.J. Downs, E.F. Vogel, A plant-wide industrial process control problem, Comput. Chem. Eng. 17 (3) (1993) 245–255.
[34] A. Bathelt, N.L. Ricker, M. Jelali, Revision of the Tennessee Eastman process model, IFAC-PapersOnLine 48 (8) (2015) 309–314.