Multiple input self-organizing-map ResNet model for optimization of petroleum refinery conversion units

doi:10.1007/s11705-022-2269-5

Front. Chem. Sci. Eng.

2023, Vol. 17

Issue (6) : 759-771 https://doi.org/10.1007/s11705-022-2269-5

RESEARCH ARTICLE

Multiple input self-organizing-map ResNet model for optimization of petroleum refinery conversion units

Jiannan Zhu¹, Vladimir Mahalec², Chen Fan¹, Minglei Yang^1,³(

), Feng Qian^1,³(

)

¹. Key Laboratory of Smart Manufacturing in Energy Chemical Process, Ministry of Education, East China University of Science and Technology, Shanghai 200237, China
². Department of Chemical Engineering, McMaster University, Hamilton, Ontario L8S 4L8, Canada
³. Engineering Research Center of Process System Engineering, Ministry of Education, East China University of Science and Technology, Shanghai 200237, China

Download: PDF(4284 KB) HTML
Export: BibTeX | EndNote | Reference Manager | ProCite | RefWorks

Abstract

This work introduces a deep-learning network, i.e., multi-input self-organizing-map ResNet (MISR), for modeling refining units comprised of two reactors and a separation train. The model is comprised of self-organizing-map and the neural network parts. The self-organizing-map part maps the input data into multiple two-dimensional planes and sends them to the neural network part. In the neural network part, residual blocks enhance the convergence and accuracy, ensuring that the structure will not be overfitted easily. Development of the MISR model of hydrocracking unit also benefits from the utilization of prior knowledge of the importance of the input variables for predicting properties of the products. The results show that the proposed MISR structure predicts more accurately the product yields and properties than the previously introduced self-organizing-map convolutional neural network model, thus leading to more accurate optimization of the hydrocracker operation. Moreover, the MISR model has smoother error convergence than the previous model. Optimal operating conditions have been determined via multi-round-particle-swarm and differential evolution algorithms. Numerical experiments show that the MISR model is suitable for modeling nonlinear conversion units which are often encountered in refining and petrochemical plants.

Keywords hydrocracking convolutional neural networks self-organizing map deep learning data-driven optimization

Corresponding Author(s): Minglei Yang,Feng Qian

Just Accepted Date: 17 January 2023 Online First Date: 06 March 2023 Issue Date: 17 May 2023

Cite this article:

Jiannan Zhu,Vladimir Mahalec,Chen Fan, et al. Multiple input self-organizing-map ResNet model for optimization of petroleum refinery conversion units[J]. Front. Chem. Sci. Eng., 2023, 17(6): 759-771.

URL:

https://academic.hep.com.cn/fcse/EN/10.1007/s11705-022-2269-5
https://academic.hep.com.cn/fcse/EN/Y2023/V17/I6/759

Fig.1 A simplified flow diagram of a two-stage-series hydrocracking process.

Tab.1 Inputs and outputs of the model

Fig.2 Distribution of raw data (yields of the eight products).

Tab.2 Statistics related to performances of different SOM sizes under 2000 iterations (prediction of 13 outputs)

Fig.3 Structure of MISR of (a) SOM part and (b) residual part (3 residual blocks).

Fig.4 Residual block (right) compared with the classical CNN (left).

Index	SOM-CNN without BN	SOM-CNN with BN
Iterations	2000	2000
Correlation coefficient $R 2$	0.9329	0.9468
MRE (10 samples)	2.543	2.316
MAE (10 samples)	0.4970	0.4464
Time cost/min	5.50	7.66

Tab.3 Comparison of SOM-CNN with and without BN

Fig.5 The loss of SOM-CNN with and without BN.

Index	SOM-CNN	Multi-input-SOM-CNN
Correlation coefficient $R 2$	0.9468	0.9533
MRE (test samples)	2.316	2.117
MAE (test samples)	0.4464	0.3870
Time cost/min	7.66	8.01

Tab.4 Comparison of SOM-CNN with and without multi-input

Fig.6 The loss of SOM-CNN with and without multi-input.

Tab.5 Performances of MISR with multiple residual blocks

Fig.7 The predicted yields vs. actual yields of the eight products on the testing data: (a)

H 2 S

, (b) GAS, (c) LPG, (d) LN, (e) HN, (f) kerosene, (g) diesel, and (h) bottom.

Tab.6 Performances of classical FNN models with different hidden layers

Tab.7 Performances of different networks

Fig.8 Loss curves of the four networks.

Fig.9 Interpolation test of different models to predict yields of the six products based on different feed ratios: (a) dry gas, (b) LPG, (c) LN, (d) HN, (e) kerosene, and (f) diesel.

Tab.8 Optimization effects of three methods

Tab.9 Profit prediction and real optimization results via SOM-CNN and MISR

Fig.10 Difference between true optimum profit and profit by SOM-CNN and MISR.

1	A Marafi, H Albazzaz, M S Rana. Hydroprocessing of heavy residual oil: opportunities and challenges. Catalysis Today, 2019, 329: 125–134 https://doi.org/10.1016/j.cattod.2018.10.067
2	E Iplik, I Aslanidou, K Kyprianidis. Hydrocracking: a perspective towards digitalization. Sustainability, 2020, 12(17): 7058 https://doi.org/10.3390/su12177058
3	J W Ward. Hydrocracking processes and catalysts. Fuel Processing Technology, 1993, 35(1): 55–85 https://doi.org/10.1016/0378-3820(93)90085-I
4	S Sánchez, M A Rodríguez, J Ancheyta. Kinetic model for moderate hydrocracking of heavy oils. Industrial & Engineering Chemistry Research, 2005, 44(25): 9409–9413 https://doi.org/10.1021/ie050202+
5	H Kumar, G F Froment. Mechanistic kinetic modeling of the hydrocracking of complex feedstocks, such as vacuum gas oils. Industrial & Engineering Chemistry Research, 2007, 46(18): 5881–5897 https://doi.org/10.1021/ie0704290
6	G Félix, J Ancheyta. Using separate kinetic models to predict liquid, gas, and coke yields in heavy oil hydrocracking. Industrial & Engineering Chemistry Research, 2019, 58(19): 7973–7979 https://doi.org/10.1021/acs.iecr.9b00904
7	J Singh, M Kumar, A K Saxena, S Kumar. Reaction pathways and product yields in mild thermal cracking of vacuum residues: a multi-lump kinetic model. Chemical Engineering Journal, 2005, 108(3): 239–248 https://doi.org/10.1016/j.cej.2005.02.018
8	S Qader, G Hill. Hydrocracking of gas oil. Industrial & Engineering Chemistry Process Design and Development, 1969, 8(1): 98–105 https://doi.org/10.1021/i260029a017
9	N Bhutani, A K Ray, G Rangaiah. Modeling, simulation, and multi-objective optimization of an industrial hydrocracking unit. Industrial & Engineering Chemistry Research, 2006, 45(4): 1354–1372 https://doi.org/10.1021/ie050423f
10	C S Laxminarasimhan, R P Verma, P A Ramachandran. Continuous lumping model for simulation of hydrocracking. AIChE Journal, 1996, 42(9): 2645–2653 https://doi.org/10.1002/aic.690420925
11	H M S Lababidi, F S AlHumaidan. Modeling the hydrocracking kinetics of atmospheric residue in hydrotreating processes by the continuous lumping approach. Energy & Fuels, 2011, 25(5): 1939–1949 https://doi.org/10.1021/ef200153p
12	R J Quann, S B Jaffe. Structure-oriented lumping: describing the chemistry of complex hydrocarbon mixtures. Industrial & Engineering Chemistry Research, 1992, 31(11): 2483–2497 https://doi.org/10.1021/ie00011a013
13	P J Becker, N Serrand, B Celse, D Guillaume, H Dulot. Comparing hydrocracking models: continuous lumping vs. single events. Fuel, 2016, 165: 306–315 https://doi.org/10.1016/j.fuel.2015.09.091
14	P J Becker, N Serrand, B Celse, D Guillaume, H Dulot. A single events microkinetic model for hydrocracking of vacuum gas oil. Computers & Chemical Engineering, 2017, 98: 70–79 https://doi.org/10.1016/j.compchemeng.2016.11.035
15	M RosliN Aziz. Review of neural network modelling of cracking process. In: Second International Conference on Chemical Engineering (ICCE). Bandung, Indonesia: IOP, 2016
16	N Bhutani, G P Rangaiah, A K Ray. First-principles, data-based, and hybrid modeling and optimization of an industrial hydrocracking unit. Industrial & Engineering Chemistry Research, 2006, 45(23): 7807–7816 https://doi.org/10.1021/ie060247q
17	H Fang, J Zhou, Z Wang, Z Qiu, Y Sun, Y Lin, K Chen, X Zhou, M Pan. Hybrid method integrating machine learning and particle swarm optimization for smart chemical process operations. Frontiers of Chemical Science and Engineering, 2022, 16(2): 274–287 https://doi.org/10.1007/s11705-021-2043-0
18	Y Ma, Z Gao, P Shi, M Chen, S Wu, C Yang, J Wang, J Cheng, J Gong. Machine learning-based solubility prediction and methodology evaluation of active pharmaceutical ingredients in industrial crystallization. Frontiers of Chemical Science and Engineering, 2022, 16(4): 523–535 https://doi.org/10.1007/s11705-021-2083-5
19	C McGreavy, M Lu, X Z Wang, E K T Kam. Characterisation of the behaviour and product distribution in fluid catalytic cracking using neural networks. Chemical Engineering Science, 1994, 49(24): 4717–4727 https://doi.org/10.1016/S0009-2509(05)80054-5
20	L M Ochoa-Estopier, M Jobson, R Smith. Operational optimization of crude oil distillation systems using artificial neural networks. Computers & Chemical Engineering, 2013, 59: 178–185 https://doi.org/10.1016/j.compchemeng.2013.05.030
21	F Yang, C Dai, J Tang, J Xuan, J Cao. A hybrid deep learning and mechanistic kinetics model for the prediction of fluid catalytic cracking performance. Chemical Engineering Research & Design, 2020, 155: 202–210 https://doi.org/10.1016/j.cherd.2020.01.013
22	W Song, V Mahalec, J Long, M Yang, F Qian. Modeling the hydrocracking process with deep neural networks. Industrial & Engineering Chemistry Research, 2020, 59(7): 3077–3090 https://doi.org/10.1021/acs.iecr.9b06295
23	Y Lecun, L Bottou, Y Bengio, P Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 1998, 86(11): 2278–2324 https://doi.org/10.1109/5.726791
24	S IoffeC Szegedy. Batch normalization: accelerating deep network training by reducing internal covariate shift. In: International Conference on Machine Learning. Lille, France: JMLR, 2015
25	A Krizhevsky, I Sutskever, G E Hinton. ImageNet classification with deep convolutional neural networks. Communications of the ACM, 2017, 60(6): 84–90 https://doi.org/10.1145/3065386
26	C SzegedyW LiuY Q JiaP SermanetS ReedD AnguelovD ErhanV VanhouckeA Rabinovich. Going deeper with convolutions. In: 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Boston, MA: IEEE, 2015
27	K HeX ZhangS RenJ Sun. Deep residual learning for image recognition. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV: IEEE, 2016
28	R Zhao, R Yan, Z Chen, K Mao, P Wang, R X Gao. Deep learning and its applications to machine health monitoring. Mechanical Systems and Signal Processing, 2019, 115: 213–237 https://doi.org/10.1016/j.ymssp.2018.05.050
29	G Serin, B Sener, A M Ozbayoglu, H O Unver. Review of tool condition monitoring in machining and opportunities for deep learning. International Journal of Advanced Manufacturing Technology, 2020, 109(3): 953–974 https://doi.org/10.1007/s00170-020-05449-w
30	R M Souza, E G Nascimento, U A Miranda, W J Silva, H A Lepikson. Deep learning for diagnosis and classification of faults in industrial rotating machinery. Computers & Industrial Engineering, 2021, 153: 107060 https://doi.org/10.1016/j.cie.2020.107060
31	J Yuan, Y Tian. A multiscale feature learning scheme based on deep learning for industrial process monitoring and fault diagnosis. IEEE Access: Practical Innovations, Open Solutions, 2019, 7: 151189–151202 https://doi.org/10.1109/ACCESS.2019.2947714
32	M Elhefnawy, A Ragab, M S Ouali. Fault classification in the process industry using polygon generation and deep learning. Journal of Intelligent Manufacturing, 2022, 33(5): 1531–1544 https://doi.org/10.1007/s10845-021-01742-x
33	A Glaeser, V Selvaraj, S Lee, Y Hwang, K Lee, N Lee, S Lee, S Min. Applications of deep learning for fault detection in industrial cold forging. International Journal of Production Research, 2021, 59(16): 4826–4835 https://doi.org/10.1080/00207543.2021.1891318
34	K SimonyanA Zisserman. Very deep convolutional networks for large-scale image recognition. In: International Conference on Learning Representations. San Diego, CA: OpenReview.net, 2015
35	S XieR GirshickP DollárZ TuK He. Aggregated residual transformations for deep neural networks. In: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI: IEEE, 2017
36	S ZagoruykoN Komodakis. Wide residual networks. In: Proceedings of the British Machine Vision Conference (BMVC). York, UK: BMVA, 2016

[1]

FCE-22094-OF-ZJ_suppl_1

Download

[1]	Chong Peng, Yanze Du, Xiang Feng, Yongkang Hu, Xiangchen Fang. Research and development of hydrocracking catalysts and technologies in China[J]. Front. Chem. Sci. Eng., 2018, 12(4): 867-877.

Viewed

Full text

Abstract

Cited

Shared

Discussed