Product prediction and optimization of shift reactor based on hybrid modeling

doi:10.16085/j.issn.1000-6613.2020-2564

Abstract

Abstract:

The intelligent construction of the petrochemical industry is an important part of “Made in China 2025”. Strengthening the construction of CPS is an important guarantee for the realization of intelligence. Based on the CPS construction needs of a refining, this paper focuses on the modeling of transformation devices, and uses a hybrid modeling method to develop the model. Through the Aspen plus simulation of the production process and sensitivity analysis, a series of data were generated. A hybrid model was developed by mining the training set data to estimate the kinetic equation parameters of the R2204 reactor, and the model was validated using the prediction set data. Results showed that the relative deviation of the model in predicting the composition of CO at the reactor outlet was relatively small (the average deviation was 2.78%; the maximum deviation of a group does not exceed 10%), which proved that the model was reliable. Finally, based on this model, the temperature of the reactor inlet stream S2220 and the H₂O/CO ratio were optimized. Results showed that when the feed temperature was 201℃ and the H₂O/CO ratio was 51.35, the CO mole fraction at the outlet of the reactor was 0.406%, which met the technical production indicators. The model established by the hybrid modeling method in this paper can not only accurately predict the CO composition of the product composition of the reactor, but also can be integrated into the factory CPS system to provide a basis for subsequent decision-making optimization.

Key words: cyber-physical systems (CPS), hybrid modeling, data mining, parameter estimation, optimization

摘要：

石化行业智能化建设是“中国制造2025”的重要内容，加强信息物理系统（CPS）建设是实现石化行业智能化的重要保障。本文基于某炼化企业CPS建设需要，以变换装置建模为重点，采用混合建模方法进行模型的研发。通过Aspen plus模拟生产流程并进行灵敏度分析，模拟产生系列运行数据；利用混合建模方法对训练组数据进行挖掘，估算了反应器R2204的动力学参数，建立了混合模型，并采用预测组数据进行验证。结果表明，通过该方法所得模型在预测反应器出口最关键组分CO组成时相对偏差较小（25组数据平均偏差为2.78%，最大一组数据偏差不超过10%），证明模型可靠；基于此模型，对反应器进口流股S2220的温度和H₂O/CO摩尔比值进行了模拟优化，结果表明进料温度为201℃、H₂O/CO摩尔比值为51.35时，反应器出口CO的干基物质的量分数最小为0.406%，满足技术生产指标。本文采用混合建模方法建立的模型不仅可以准确地预测反应器的产品CO组成，还可以集成到工厂CPS系统中，为后续的决策优化提供依据。

关键词: 信息物理系统（CPS）, 混合建模, 数据挖掘, 参数估计, 优化

CLC Number:

TQ015

GAO Shilei, PAN Yanqiu, LI Pengfei, ZHANG Chunchao, YU Lu, WANG Zhenxing. Product prediction and optimization of shift reactor based on hybrid modeling[J]. Chemical Industry and Engineering Progress, 2021, 40(12): 6604-6612.

高石磊, 潘艳秋, 李鹏飞, 张春超, 俞路, 王振兴. 基于混合建模的变换反应器产品预测与优化[J]. 化工进展, 2021, 40(12): 6604-6612.

Figures/Tables 14

References 26

1	中国信息物理系统发展论坛. 信息物理系统白皮书 (2017)[EB/OL]. [2020-12-11]. .
	China Cyber Physics System Development Forum. Cyber-physical system white paper (2017)[EB/OL]. [2020-12-11]. .
2	中国电子技术标准化研究院, 中国信息物理系统发展论坛. 信息物理系统建设指南 (2020)[EB/OL]. [2020-12-11]. .
	China Elctronics Standardization Institute, China Cyber Physics System Development Forum. Cyber-physical system construction guide (2020) [EB/OL]. [2020-12-11]. .
3	林融. 智能自动化技术推动石油化工企业数字化转型升级[J]. 自动化仪表, 2020, 41(6): 1-7.
	LIN Rong. Intelligent automation technology promote digitalized transforming and upgrading in petrochemical enterprises[J]. Process Automation Instrumentation, 2020, 41(6): 1-7.
4	冯应国. 智能制造新模式——现代煤化工企业信息化建设方向探析[J]. 煤化工, 2019, 47(2): 72-75.
	FENG Yingguo. New mode of intelligent manufacturing—Analysis of informatization construction direction of modern modern coal chemical enterprise[J]. Coal Chemical Industry, 2019, 47(2): 72-75.
5	JI X, HE G, XU J J, et al. Study on the mode of intelligent chemical industry based on cyber-physical system and its implementation[J]. Advances in Engineering Software, 2016, 99: 18-26.
6	钱锋，杜文莉，钟伟民，等. 石油和化工行业智能优化制造若干问题及挑战[J]. 自动化学报, 2017, 43(6): 893-901.
	QIAN Feng, DU Wenli, ZHONG Weimin, et al. Problems and challenges of smart optimization manufacturing in petrochemical industries[J]. Acta Automatica Sinica, 2017, 43(6): 893-901.
7	覃伟中. 石油化工智能制造[M]. 北京: 化学工业出版社, 2019: 80.
	QIN Weizhong. Petrochemical intelligent manufacturing[M]. Beijing: Chemical Industry Press, 2019: 80.
8	ZENDEHBOUDI S, REZAEI N, LOHI A. Applications of hybrid models in chemical, petroleum, and energy systems: a systematic review[J]. Applied Energy, 2018, 228: 2539-2566.
9	郭晶晶, 徐金金, 杜文莉, 等. 自适应迭代混合建模及在碳二加氢过程的应用[J]. 化工学报, 2018, 69(11): 4814-4822.
	GUO Jingjing, XU Jinjin, DU Wenli, et al. Self-adaptive iterative hybrid modeling and its application in acetylene hydrogenation process[J]. CIESC Journal, 2018, 69(11): 4814-4822.
10	ZHANG S N, WANG F L, HE D K, et al. Soft sensor for cobalt oxalate synthesis process in cobalt hydrometallurgy based on hybrid model[J]. Neural Computing and Applications, 2013, 23(5): 1465-1472.
11	RUZ M L, GARRIDO J, VÁZQUEZ F, et al. A hybrid modeling approach for steady-state optimal operation of vapor compression refrigeration cycles[J]. Applied Thermal Engineering, 2017, 120: 74-87.
12	NOVAES L D R, DE RESENDE N S, SALIM V M M, et al. Modeling, simulation and kinetic parameter estimation for diesel hydrotreating[J]. Fuel, 2017, 209: 184-193.
13	孙延吉. 加氢装置CPS平台中的过程管控关键问题研究[D]. 大连: 大连理工大学, 2019.
	SUN Yanji. Research on key issues of process management and control in hydrogenation unit level CPS platform[D]. Dalian: Dalian University of Technology, 2019.
14	朱炳辰. 化学反应工程[M]. 4版. 北京: 化学工业出版社, 2012: 65-68.
	ZHU Bingchen. Chemical reaction engineering[M]. 4th ed. Beijing: Chemical Industry Press, 2012: 65-68.
15	孙兰义. 化工过程模拟实训——Aspen Plus教程[M]. 2版. 北京: 化学工业出版社, 2017: 47.
	SUN Lanyi. Chemical process simulation training—Aspen Plus tutorial[M]. 2nd ed. Beijing: Chemical Industry Press, 2017: 47.
16	YI Q, FENG J, LI W Y. Optimization and efficiency analysis of polygeneration system with coke-oven gas and coal gasified gas by Aspen Plus[J]. Fuel, 2012, 96: 131-140.
17	张欣欣. 水煤气变换工段的模拟与仿真[D]. 西安: 西安科技大学, 2015.
	ZHANG Xinxin. Modeling and simulation of water gas shift process[D]. Xi’an: Xi’an University of Science and Technology, 2015.
18	JOHN Y M, MUSTAFA M A, PATEL R, et al. Parameter estimation of a six-lump kinetic model of an industrial fluid catalytic cracking unit[J]. Fuel, 2019, 235: 1436-1454.
19	WANG X F, CHEN J D, LIU C B, et al. Hybrid modeling of penicillin fermentation process based on least square support vector machine[J]. Chemical Engineering Research and Design, 2010, 88(4): 415-420.
20	冯云飞, 江莉龙, 曹彦宁, 等. 新型钴钼系耐硫变换催化剂本征动力学[J]. 工业催化, 2013, 21(2): 32-37.
	Yunfei FEN, JIANG Lilong, CAO Yanning, et al. Intrinsic kinetics of a new cobalt molybdenum sulfur-tolerant shift catalyst[J]. Industrial Catalysis, 2013, 21(2): 32-37.
21	郁磊. MATLAB智能算法30个案例分析[M]. 2版. 北京: 北京航空航天大学出版社, 2015: 154-156, 242.
	YU Lei. Analysis of 30 cases of matlab intelligent algorithm[M]. 2nd ed. Beijing: Beihang University Press, 2015: 154-156, 242.
22	刘全生, 金恒芳, 王国军, 等. BX型中温变换催化剂上CO变换反应本征动力学研究[J]. 工业催化, 1999, 7(4): 13-19.
	LIU Quansheng, JIN Hengfang, WANG Guojun, et al. Intrinsic kinetics of CO shift reaction over BX HTS catalyst[J]. Industrial Catalysis, 1999, 7(4): 13-19.
23	潘银珍, 邹世平, 陈富生, 等. SB-5型钴钼耐硫变换催化剂反应动力学的研究[J]. 化肥工业, 1993, 20(6): 8-13.
	PAN Yinzhen, ZOU Shiping, CHEN Fusheng, et al. Study on reaction kinetics of SB-5 cobalt-molybdenum sulfur-tolerant shift catalyst[J]. Journal of Chemical Fertilizer Industry, 1993, 20(6): 8-13.
24	魏广学, 陈五平, 荣桂安. 钴钼系催化剂一氧化碳加压变换反应动力学研究[J]. 氮肥设计, 1994, 32(3): 14-18.
	WEI Guangxue, CHEN Wuping, RONG Gui’an. Study on pressured CO shift reaction kinetics on cobalt-molybdenum series catalysis[J]. Nitrogen Fertilizer Design, 1994, 32(3): 14-18.
25	隋志军. 化工数值计算与MATLAB[M]. 上海: 华东理工大学出版社, 2015: 154-156.
	SUI Zhijun. Chemical numerical calculation and MATLAB[M]. Shanghai: East China University of Science and Technology Press, 2015: 154-156.
26	YAZDANI M, JOLAI F. Lion optimization algorithm (LOA): a nature-inspired metaheuristic algorithm[J]. Journal of Computational Design and Engineering, 2016, 3(1): 24-36.

对比指标	S2220（进口）				S2221（出口）
对比指标	设计值	模拟值	相对偏差%	设计值	模拟值	相对偏差%
组成（物质的量分数）/%
H₂	36.7590	36.7594	0.0011	37.1850	37.1853	0.0008
CO	0.7150	0.7152	0.0280	0.2890	0.2893	0.1038
CO₂	27.9890	27.9793	0.0347	28.4150	28.4052	0.0345
H₂O	31.9060	31.9154	0.0295	31.4800	31.4895	0.0302
其他	2.6310	2.6307	0.0114	2.6310	2.6307	0.0114
流量/kmol·h^-1	16905.6	16905.6	0.0000	16905.6	16905.6	0.0000
温度/℃	210.0000	208.6808	0.6282	214.2000	214.2000	0.0000
压力/MPa	3.4500	3.4500	0.0000	3.4100	3.4100	0.0000

对比指标	S2220（进口）				S2221（出口）
对比指标	设计值	模拟值	相对偏差%	设计值	模拟值	相对偏差%
组成（物质的量分数）/%
H₂	36.7590	36.7594	0.0011	37.1850	37.1853	0.0008
CO	0.7150	0.7152	0.0280	0.2890	0.2893	0.1038
CO₂	27.9890	27.9793	0.0347	28.4150	28.4052	0.0345
H₂O	31.9060	31.9154	0.0295	31.4800	31.4895	0.0302
其他	2.6310	2.6307	0.0114	2.6310	2.6307	0.0114
流量/kmol·h^-1	16905.6	16905.6	0.0000	16905.6	16905.6	0.0000
温度/℃	210.0000	208.6808	0.6282	214.2000	214.2000	0.0000
压力/MPa	3.4500	3.4500	0.0000	3.4100	3.4100	0.0000

序号	S2220 /kmol·h^-1	T2220 /℃	组成（物质的量分数）/%
序号	S2220 /kmol·h^-1	T2220 /℃	CO_in	H₂O_in	H_2，in	CO_2，in	CO_out
1	13956.80	201.55	0.7675	26.9378	39.4469	30.0248	0.3104
2	14362.24	203.33	0.7548	29.0003	38.3333	29.1772	0.3017
—	—	—	—	—	—	—	—
75	16319.02	207.55	0.7225	31.2196	37.1351	28.2652	0.2922
76	16724.47	208.91	0.7050	32.8871	36.2348	27.5800	0.2852
—	—	—	—	—	—	—	—
100	18557.80	213.55	0.6519	37.9358	33.5089	25.5052	0.2637

序号	S2220 /kmol·h^-1	T2220 /℃	组成（物质的量分数）/%
序号	S2220 /kmol·h^-1	T2220 /℃	CO_in	H₂O_in	H_2，in	CO_2，in	CO_out
1	13956.80	201.55	0.7675	26.9378	39.4469	30.0248	0.3104
2	14362.24	203.33	0.7548	29.0003	38.3333	29.1772	0.3017
—	—	—	—	—	—	—	—
75	16319.02	207.55	0.7225	31.2196	37.1351	28.2652	0.2922
76	16724.47	208.91	0.7050	32.8871	36.2348	27.5800	0.2852
—	—	—	—	—	—	—	—
100	18557.80	213.55	0.6519	37.9358	33.5089	25.5052	0.2637

优化算法	目标函数值	动力学参数
优化算法	目标函数值	k₀	E/kJ·kmol^-1	a	b	c	d
GA	0.05526	18.28179	747.97970	1.99999	0.00002	-0.00001	-0.00001
PSO	0.05457	15.18988	0	2.00000	0	0	0
SA	0.05792	15.00843	26.27020	1.99968	0.00017	-0.00001	-0.00054