Improving the accuracy of mechanistic models for dynamic batch distillation enabled by neural network: An industrial plant case

doi:10.1016/j.cjche.2024.04.018

中国化学工程学报 ›› 2024, Vol. 73 ›› Issue (9): 290-300.DOI: 10.1016/j.cjche.2024.04.018

Improving the accuracy of mechanistic models for dynamic batch distillation enabled by neural network: An industrial plant case

Xiaoyu Zhou¹, Xiangyi Gao², Mingmei Wang¹, Erwei Song¹, Erqiang Wang¹

1. School of Chemical Engineering, University of Chinese Academy of Sciences, Beijing 100049, China;
2. School of Information Science and Engineering, Yanshan University, Qinhuangdao 066004, China

收稿日期:2023-12-24 修回日期:2024-04-07 接受日期:2024-04-07 出版日期:2024-11-21 发布日期:2024-05-24
通讯作者: Erqiang Wang,E-mail:wangerqiang@ucas.ac.cn
基金资助:
This work is supported by Beijing Natural Science Foundation (2222037) and by the Fundamental Research Funds for the Central Universities.

Improving the accuracy of mechanistic models for dynamic batch distillation enabled by neural network: An industrial plant case

Xiaoyu Zhou¹, Xiangyi Gao², Mingmei Wang¹, Erwei Song¹, Erqiang Wang¹

1. School of Chemical Engineering, University of Chinese Academy of Sciences, Beijing 100049, China;
2. School of Information Science and Engineering, Yanshan University, Qinhuangdao 066004, China

Received:2023-12-24 Revised:2024-04-07 Accepted:2024-04-07 Online:2024-11-21 Published:2024-05-24
Contact: Erqiang Wang,E-mail:wangerqiang@ucas.ac.cn
Supported by:
This work is supported by Beijing Natural Science Foundation (2222037) and by the Fundamental Research Funds for the Central Universities.

摘要/Abstract

摘要： Neural networks are often viewed as pure ‘black box’ models, lacking interpretability and extrapolation capabilities of pure mechanistic models. This work proposes a new approach that, with the help of neural networks, improves the conformity of the first-principal model to the actual plant. The final result is still a first-principal model rather than a hybrid model, which maintains the advantage of the high interpretability of first-principal model. This work better simulates industrial batch distillation which separates four components: water, ethylene glycol, diethylene glycol, and triethylene glycol. GRU (gated recurrent neural network) and LSTM (long short-term memory) were used to obtain empirical parameters of mechanistic model that are difficult to measure directly. These were used to improve the empirical processes in mechanistic model, thus correcting unreasonable model assumptions and achieving better predictability for batch distillation. The proposed method was verified using a case study from one industrial plant case, and the results show its advancement in improving model predictions and the potential to extend to other similar systems.

关键词: Batch distillation, Mechanistic models, Neural network, GRU, LSTM

Abstract: Neural networks are often viewed as pure ‘black box’ models, lacking interpretability and extrapolation capabilities of pure mechanistic models. This work proposes a new approach that, with the help of neural networks, improves the conformity of the first-principal model to the actual plant. The final result is still a first-principal model rather than a hybrid model, which maintains the advantage of the high interpretability of first-principal model. This work better simulates industrial batch distillation which separates four components: water, ethylene glycol, diethylene glycol, and triethylene glycol. GRU (gated recurrent neural network) and LSTM (long short-term memory) were used to obtain empirical parameters of mechanistic model that are difficult to measure directly. These were used to improve the empirical processes in mechanistic model, thus correcting unreasonable model assumptions and achieving better predictability for batch distillation. The proposed method was verified using a case study from one industrial plant case, and the results show its advancement in improving model predictions and the potential to extend to other similar systems.

Key words: Batch distillation, Mechanistic models, Neural network, GRU, LSTM

Xiaoyu Zhou, Xiangyi Gao, Mingmei Wang, Erwei Song, Erqiang Wang. Improving the accuracy of mechanistic models for dynamic batch distillation enabled by neural network: An industrial plant case[J]. 中国化学工程学报, 2024, 73(9): 290-300.

参考文献

[1] D.W.T. Rippin, Simulation of single- and multiproduct batch chemical plants for optimal design and operation, Comput. Chem. Eng. 7 (3) (1983) 137-156.
[2] Mujtaba I M. Batch distillation: Design and Operation, World Scientific Publishing Company, Singapore, 2004.
[3] Hangos K, Cameron I. Process Modelling and Model Analysis, Acad Press, San Diego, 2001.
[4] M.L. Thompson, M.A. Kramer, Modeling chemical processes using prior knowledge and neural networks, AlChE. J. 40 (8) (1994) 1328-1340.
[5] H.I. Furlonge, C.C. Pantelides, E. Soerensen, Optimal operation of multivessel batch distillation columns, AlChE. J. 45 (4) (1999) 781-801.
[6] U.M. Diwekar, K.P. Madhavan, Multicomponent batch distillation column design, Ind. Eng. Chem. Res. 30 (4) (1991) 713-721.
[7] U.M. Diwekar, How simple can it be? -a look at the models for batch distillation, Comput. Chem. Eng. 18 (1994) S451-S457.
[8] Meadows E L. Multicomponent batch-distillation calculations on a digital computer,Chem. Eng. Symp. Ser, 59(1963) 46-48.
[9] G.P. Distefano, Mathematical modeling and numerical integration of multicomponent batch distillation equations, AlChE. J. 14 (1) (1968) 190-199.
[10] Boston J F, Britt H I, Jirapongphan S, & Shah, V. B. An advanced system for the simulation of batch distillation operations[J]. Foundations of Computer-Aided Chemical Process Design, (2)(1981)203-237.
[11] U. Diwekar, K.P. Madhavan, R.E. Swaney, Optimization of multicomponent batch distillation columns, Ind. Eng. Chem. Res. 28 (7) (1989) 1011-1017.
[12] G.Q. Wu, W.T.G. Yion, K. Le Nguyen Quang Dang, Z. Wu, Physics-informed machine learning for MPC: Application to a batch crystallization process, Chem. Eng. Res. Des. 192 (2023) 556-569.
[13] Y.Z. Zheng, Z. Wu, Physics-informed online machine learning and predictive control of nonlinear processes with parameter uncertainty, Ind. Eng. Chem. Res. 62 (6) (2023) 2804-2818.
[14] Y.Z. Zheng, C. Hu, X.N. Wang, Z. Wu, Physics-informed recurrent neural network modeling for predictive control of nonlinear processes, J. Process. Contr. 128 (2023) 103005.
[15] N. Sharma, Y.A. Liu, A hybrid science-guided machine learning approach for modeling chemical processes: A review, AlChE. J. 68 (5) (2022) e17609.
[16] M. Agarwal, Combining neural and conventional paradigms for modelling, prediction and control, Int. J. Syst. Sci. 28 (1) (1997) 65-81.
[17] H. Kwon, K.C. Oh, Y. Choi, Y.G. Chung, J. Kim, Development and application of machine learning-based prediction model for distillation column, Int. J. Intell. Syst. 36 (5) (2021) 1970-1997.
[18] M.S.F. Bangi, J.S.I. Kwon, Deep hybrid modeling of chemical process: Application to hydraulic fracturing, Comput. Chem. Eng. 134 (2020) 106696.
[19] V. Mahalec, Y. Sanchez, Inferential monitoring and optimization of crude separation units via hybrid models, Comput. Chem. Eng. 45 (2012) 15-26.
[20] I.M. Mujtaba, M.A. Hussain, Optimal operation of dynamic processes under process-model mismatches: Application to batch distillation, Comput. Chem. Eng. 22 (1998) S621-S624.
[21] Mujtaba I M, Macchietto S. Holdup issues in batch distillation-binary mixtures[J]. Chem. Eng. Sci. 1998, 53(14): 2519-2530.
[22] Nomikos P, MacGregor J F. Monitoring batch processes using multiway principal component analysis[J]. AIChE J, 40(8) (1994) 1361-1375.
[23] J. Chung, C. Gulcehre, K. Cho, Y. Bengio, Empirical evaluation of gated recurrent neural networks on sequence modeling, (2014): 1412.3555.
[24] D.P. Kingma, J. Ba, Adam: a method for stochastic optimization, (2014): 1412.6980.
[25] W.H. Wang, N. Yang, F.R. Wei, B.B. Chang, M. Zhou, Gated Self-Matching Networks for Reading Comprehension and Question AnsweringProceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Vancouver, Canada. Stroudsburg, PA, USA: Association for Computational Linguistics, (2017): 189-198.

Improving the accuracy of mechanistic models for dynamic batch distillation enabled by neural network: An industrial plant case

Improving the accuracy of mechanistic models for dynamic batch distillation enabled by neural network: An industrial plant case

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	Zongyu Yao, Qingchao Jiang, Xingsheng Gu. Distributed process monitoring based on Kantorovich distance-multiblock variational autoencoder and Bayesian inference[J]. 中国化学工程学报, 2024, 73(9): 311-323.
[2]	Zhiwei Zhu, Minglei Yang, Wangli He, Renchu He, Yunmeng Zhao, Feng Qian. A deep reinforcement learning approach to gasoline blending real-time optimization under uncertainty[J]. 中国化学工程学报, 2024, 71(7): 183-192.
[3]	Cheng Yang, Chao Jiang, Guo Yu, Jun Li, Cuimei Bo. Transferable adversarial slow feature extraction network for few-shot quality prediction in coal-to-ethylene glycol process[J]. 中国化学工程学报, 2024, 71(7): 258-271.
[4]	Bo Wang, Han Zhou, Shan Jing, Qiang Zheng, Wenjie Lan, Shaowei Li. A data-driven model of drop size prediction based on artificial neural networks using small-scale data sets[J]. 中国化学工程学报, 2024, 66(2): 71-83.
[5]	Jian Long, Kai Deng, Renchu He. Closed-loop scheduling optimization strategy based on particle swarm optimization with niche technology and soft sensor method of attributes-applied to gasoline blending process[J]. 中国化学工程学报, 2023, 61(9): 43-57.
[6]	Xi Luo, Xiayuan Feng, Xu Ji, Yagu Dang, Li Zhou, Kexin Bi, Yiyang Dai. Extraction and analysis of risk factors from Chinese chemical accident reports[J]. 中国化学工程学报, 2023, 61(9): 68-81.
[7]	Shanwei Xiong, Li Zhou, Yiyang Dai, Xu Ji. Attention-based long short-term memory fully convolutional network for chemical process fault diagnosis[J]. 中国化学工程学报, 2023, 56(4): 1-14.
[8]	Haoshan Duan, Xi Meng, Jian Tang, Junfei Qiao. Prediction of NO_x concentration using modular long short-term memory neural network for municipal solid waste incineration[J]. 中国化学工程学报, 2023, 56(4): 46-57.
[9]	Kun Ren, Zheng Jiao, Xiaolong Wu, Honggui Han. Multivariable identification of membrane fouling based on compacted cascade neural network[J]. 中国化学工程学报, 2023, 53(1): 37-45.
[10]	Denglong Ma, Ruitao Wu, Zekang Li, Kang Cen, Jianmin Gao, Zaoxiao Zhang. A new method to forecast multi-time scale load of natural gas based on augmentation data-machine learning model[J]. 中国化学工程学报, 2022, 48(8): 166-175.
[11]	Ruoshi Qin, Jinsong Zhao. Adaptive multiscale convolutional neural network model for chemical process fault diagnosis[J]. 中国化学工程学报, 2022, 50(10): 398-411.
[12]	Qiaoyi Xu, Wenli Du, Jinjin Xu, Jikai Dong. Neural network-based source tracking of chemical leaks with obstacles[J]. 中国化学工程学报, 2021, 33(5): 211-220.
[13]	Huichao Lv, Dayong Tian. Designing and optimizing a parallel neural network model for predicting the solubility of diosgenin in n-alkanols[J]. 中国化学工程学报, 2021, 29(1): 288-294.
[14]	Mirollah Hosseini, Hamid Hassanzadeh Afrouzi, Sina Yarmohammadi, Hossein Arasteh, Davood Toghraie, A. Jafarian Amiri, Arash Karimipour. Optimization of FX-70 refrigerant evaporative heat transfer and fluid flow characteristics inside the corrugated tubes using multi-objective genetic algorithm[J]. 中国化学工程学报, 2020, 28(8): 2142-2151.
[15]	Beverly S. Chittoo, Clint Sutherland. Column breakthrough studies for the removal and recovery of phosphate by lime-iron sludge: Modeling and optimization using artificial neural network and adaptive neuro-fuzzy inference system[J]. 中国化学工程学报, 2020, 28(7): 1847-1859.