Design of self-optimizing control structure for continuous catalytic reforming reaction process based on surrogate model

doi:10.11949/0438-1157.20250102

Abstract

Abstract:

A strategy for constructing a surrogate model for the global self-optimizing control (gSOC) problem was proposed. This strategy, tailored to the specific characteristics of the gSOC problem, integrates partitioned space design and a hybrid adaptive sampling method for sub-models, thereby improving construction efficiency. Based on this, the gSOC algorithm process was optimized by replacing inefficient process simulation software and accelerated the solution of the optimal combination matrix, thus extending the $a l g o r i t h m' s$ applicability to complex and large-scale processes. The improved algorithm was applied to the control structure design of a continuous catalytic reforming (CCR) unit. Using a reaction kinetic model comprising 27 lumped components, an Aspen dynamic model of the CCR process was developed. By integrating the improved gSOC algorithm with a simulation and heuristic integrated framework, the self-optimizing control structure of the CCR unit was systematically analyzed and designed, addressing key issues related to parameter uncertainty disturbances and faults, such as feedstock properties, recycle hydrogen flow rate, and reactor inlet temperatures. The results of dynamic simulation experiments show that the designed SOC structure exhibits significant real-time optimization performance. This research provides theoretical guidance for the control design of industrial CCR units.

Key words: optimization, process control, self-optimizing control, control structure, continuous catalytic reforming, surrogate model, dynamic simulation

摘要：

首先，提出了一种全局自优化控制（gSOC）问题的代理模型构建策略。该策略针对gSOC问题的特殊性，融合了划分空间设计和子模型混合自适应采样方法，提升了构建效率。在此基础上，对gSOC算法流程进行了优化，加速了最优组合矩阵的求解，扩展了算法对复杂和大规模过程的适用性。其次，改进后的算法应用于连续重整（CCR）反应过程的控制结构设计。通过结合改进的gSOC算法与模拟和启发式集成框架，系统地分析和设计了CCR反应过程的自优化控制结构，缓解了进料性质、循环氢流量及反应器入口温度的参数不确定性扰动与故障带来的芳烃损失。最后，动态模拟实验结果表明，所设计的SOC结构表现出显著的实时优化性能。该研究为工业CCR装置的控制提供了理论指导。

关键词: 优化, 过程控制, 自优化控制, 控制结构, 连续重整, 代理模型, 动态仿真

CLC Number:

TE 624

Yilei ZHOU, Zhi LI, Xin PENG. Design of self-optimizing control structure for continuous catalytic reforming reaction process based on surrogate model[J]. CIESC Journal, 2025, 76(9): 4499-4511.

周轶磊, 李智, 彭鑫. 基于代理模型的连续重整反应过程自优化控制结构设计[J]. 化工学报, 2025, 76(9): 4499-4511.

Figures/Tables 15

Fig. 1 Comparison of the two hierarchical control structures

Fig. 2 Flow diagram of the CCR reaction process

Fig. 3 27 lumped reaction network of catalytic reforming[29]

Table 1 The workflow of the gSOC algorithm

步骤	具体操作
1	使用Monte Carlo模拟方法对扰动空间进行采样，生成N个扰动场景 d_i，i=1,2,…,N。
2	选择一个参考点，估计该处的增益矩阵 $G r e f y$ 和Hessian矩阵 $J u u, r e f$ 。
3	对每个扰动场景求解操作优化问题，以求得最优测量值 $y o p t, i T$ ，根据式（10）构造过渡矩阵 Y 和 $Y ˜$ 。
4	根据式（12）求解最优组合矩阵 H。

Table 1 The workflow of the gSOC algorithm

步骤	具体操作
1	使用Monte Carlo模拟方法对扰动空间进行采样，生成N个扰动场景 d_i，i=1,2,…,N。
2	选择一个参考点，估计该处的增益矩阵 $G r e f y$ 和Hessian矩阵 $J u u, r e f$ 。
3	对每个扰动场景求解操作优化问题，以求得最优测量值 $y o p t, i T$ ，根据式（10）构造过渡矩阵 Y 和 $Y ˜$ 。
4	根据式（12）求解最优组合矩阵 H。

Table 2 The workflow of the improved gSOC algorithm based on a surrogate model

步骤	具体操作
1	使用Monte Carlo模拟方法对扰动空间进行采样，生成N个扰动场景 d_i，i=1,2,…,N。
2	选择一个参考点，估计该处的增益矩阵 $G r e f y$ 和Hessian矩阵 $J u u, r e f$ 。
3	采用2.2节所述方法构建经济目标函数J( u, d )的Kriging代理模型。
4	基于代理模型，对每个扰动场景求解操作优化问题，得最优操作 $u o p t d i$ 。
5	将 $u o p t d i$ 输入流程模拟器，求得最优测量值 $y o p t, i T$ ，并根据式（10）构造过渡矩阵 Y 和 $Y ˜$ 。
6	根据式（12）求解最优组合矩阵 H。

Table 2 The workflow of the improved gSOC algorithm based on a surrogate model

步骤	具体操作
1	使用Monte Carlo模拟方法对扰动空间进行采样，生成N个扰动场景 d_i，i=1,2,…,N。
2	选择一个参考点，估计该处的增益矩阵 $G r e f y$ 和Hessian矩阵 $J u u, r e f$ 。
3	采用2.2节所述方法构建经济目标函数J( u, d )的Kriging代理模型。
4	基于代理模型，对每个扰动场景求解操作优化问题，得最优操作 $u o p t d i$ 。
5	将 $u o p t d i$ 输入流程模拟器，求得最优测量值 $y o p t, i T$ ，并根据式（10）构造过渡矩阵 Y 和 $Y ˜$ 。
6	根据式（12）求解最优组合矩阵 H。

Table 3 Disturbances considered in this study

编号	描述	范围
D1	原料芳烃潜含量	±20.00%
D2	循环氢气流量	±20.00%
D3	第4反应器入口温度	±2.00%
D4	第1反应器入口温度	±2.00%
D5	第2反应器入口温度	±2.00%
D6	第3反应器入口温度	±2.00%

Fig. 4 The distribution of disturbance scenarios in Monte Carlo sampling

Fig. 5 Self-optimizing control structure of the CCR reaction process

Table 4 Description of dynamic simulation experiments

实验	扰动	阶跃幅度	时刻
1	D1	+20.00% / -20.00%	2 h
2	D2	+20.00% / -20.00%	2 h
3	D3	+2.00% / -2.00%	2 h
4	D4	+2.00% / -2.00%	3 h
	D5	+2.00% / -2.00%	5 h
	D6	+2.00% / -2.00%	7 h
5	D1	-18.89% / 11.73%	2 h
	D2	-16.07% / 15.63%	2 h
	D3	0.88% / -1.82%	2 h

Table 5 Steady-state aromatics loss under different disturbance scenarios for two control structures

扰动场景	AL-PICS/(kg/d)	AL-SOCS/(kg/d)
+20% D1	129.00	7.68
-20% D1	206.62	4.65
+20% D2	389.39	8.74
-20% D2	379.93	1.08
+2% D3	1871.87	3.94
-2% D3	2013.55	0.28
+2%D4,5,6	6570.67	0.99
-2% D4,5,6	7557.04	1.38
d_r1	2142.68	7.96
d_r2	3407.21	1.41

Fig. 6 The dynamic response in experiment 1

Fig. 7 The dynamic response in experiment 2

Fig. 8 The dynamic response in experiment 3

Fig. 9 The dynamic response in experiment 4

Fig. 10 Comparison of the dynamic changes in OD values in experiment 5

References 40

[1]	孔德金, 杨为民. 芳烃生产技术进展[J]. 化工进展, 2011, 30(1): 16-25.
	Kong D J, Yang W M. Advance in technology for production of aromatic hydrocarbons[J]. Chemical Industry and Engineering Progress, 2011, 30(1): 16-25.
[2]	徐承恩. 催化重整工艺与工程[M]. 2版. 北京: 中国石化出版社, 2014: 1-87.
	Xu C E. Catalytic Reforming Process and Engineering[M]. 2nd ed. Beijing: China Petrochemical Press, 2014: 1-87.
[3]	孟凡辉, 纪传佳, 杨纪. 惠州石化有限公司连续重整装置工艺流程模拟与优化[J]. 化工进展, 2017, 36(7): 2724-2729.
	Meng F H, Ji C J, Yang J. Process simulation and optimization for CNOOC Huizhou company's continuous reforming unit[J]. Chemical Industry and Engineering Progress, 2017, 36(7): 2724-2729.
[4]	宋倩倩, 慕彦君, 侯雨璇, 等. 中美两国石油化工产业实力对比分析[J]. 化工进展, 2020, 39(5): 1607-1619.
	Song Q Q, Mu Y J, Hou Y X, et al. Comparative analysis of the strength of petrochemical industry between China and USA[J]. Chemical Industry and Engineering Progress, 2020, 39(5): 1607-1619.
[5]	王钧炎, 黄德先. 基于差分进化算法和HYSYS机理模型的催化重整过程优化[J]. 化工学报, 2008, 59(7): 1755-1760.
	Wang J Y, Huang D X. Process optimization of catalytic reforming based on differential evolution and HYSYS mechanism model[J]. Journal of Chemical Industry and Engineering (China), 2008, 59(7): 1755-1760.
[6]	李鸿亮, 陆金桂, 侯卫锋, 等. 基于混合遗传算法的催化重整过程多目标优化[J]. 化工学报, 2010, 61(2): 432-438.
	Li H L, Lu J G, Hou W F, et al. Multi-objective optimization based on hybrid genetic algorithm for naphtha catalytic reforming process[J]. CIESC Journal, 2010, 61(2): 432-438.
[7]	Wei M, Yang M L, Qian F, et al. Integrated dual-production mode modeling and multiobjective optimization of an industrial continuous catalytic naphtha reforming process[J]. Industrial & Engineering Chemistry Research, 2016, 55(19): 5714-5725.
[8]	Ma X C, He C, Chen Q L, et al. Modeling and optimization for the continuous catalytic reforming process based on the hybrid surrogate optimization model[J]. Computers & Chemical Engineering, 2024, 191: 108841.
[9]	Chen C Y, Joseph B. On-line optimization using a two-phase approach: an application study[J]. Industrial & Engineering Chemistry Research, 1987, 26(9): 1924-1930.
[10]	Wu J G, Xu H L, Zhang C, et al. A sequential Bayesian partitioning approach for online steady-state detection of multivariate systems[J]. IEEE Transactions on Automation Science and Engineering, 2019, 16(4): 1882-1895.
[11]	Xie S, Yang C H, Wang X L, et al. A data-driven adaptive multivariate steady state detection strategy for the evaporation process of the sodium aluminate solution[J]. Journal of Process Control, 2018, 68: 145-159.
[12]	Xie S, Yang C H, Yuan X F, et al. A novel robust data reconciliation method for industrial processes[J]. Control Engineering Practice, 2019, 83: 203-212.
[13]	Matias J, Oliveira J P C, Le Roux G A C, et al. Steady-state real-time optimization using transient measurements on an experimental rig[J]. Journal of Process Control, 2022, 115: 181-196.
[14]	Engell S. Feedback control for optimal process operation[J]. Journal of Process Control, 2007, 17(3): 203-219.
[15]	Skogestad S. Plantwide control: the search for the self-optimizing control structure[J]. Journal of Process Control, 2000, 10(5): 487-507.
[16]	Jäschke J, Cao Y, Kariwala V. Self-optimizing control—a survey[J]. Annual Reviews in Control, 2017, 43: 199-223.
[17]	叶凌箭. 化工过程的自优化控制：原理、发展与应用展望[J]. 化工学报, 2023, 74(11): 4445-4465.
	Ye L J. Self-optimizing control for chemical processes: principle, developments and outlooks[J]. CIESC Journal, 2023, 74(11): 4445-4465.
[18]	叶凌箭, 马修水, 宋执环. 不确定性间歇过程的一种实时优化控制方法[J]. 化工学报, 2014, 65(9): 3535-3543.
	Ye L J, Ma X S, Song Z H. A real-time optimization approach for uncertain batch processes[J]. CIESC Journal, 2014, 65(9): 3535-3543.
[19]	王平, 赵辉, 杨朝合. 基于多目标优化的两段提升管重油催化裂解自优化控制[J]. 化工学报, 2016, 67(8): 3491-3498.
	Wang P, Zhao H, Yang C H. Self-optimizing control based on multi-objective optimization for heavy oil catalytic pyrolysis in two-stage riser[J]. CIESC Journal, 2016, 67(8): 3491-3498.
[20]	Jin B, Yao W X, Liu K L, et al. Self-optimizing control and safety assessment to achieve economic and safe operation for oxy-fuel combustion boiler island systems[J]. Applied Energy, 2022, 323: 119397.
[21]	李啸晨, 苏宏业, 谢磊, 等. 全局自优化控制策略及其测量变量子集选择[J]. 化工学报, 2021, 72(3): 1585-1594.
	Li X C, Su H Y, Xie L, et al. Research on the measurement subset selection for global self-optimizing control strategy[J]. CIESC Journal, 2021, 72(3): 1585-1594.
[22]	Halvorsen I J, Skogestad S, Morud J C, et al. Optimal selection of controlled variables[J]. Industrial & Engineering Chemistry Research, 2003, 42(14): 3273-3284.
[23]	Alstad V, Skogestad S. Null space method for selecting optimal measurement combinations as controlled variables[J]. Industrial & Engineering Chemistry Research, 2007, 46(3): 846-853.
[24]	Alstad V, Skogestad S, Hori E S. Optimal measurement combinations as controlled variables[J]. Journal of Process Control, 2009, 19(1): 138-148.
[25]	Ye L J, Cao Y, Yuan X F. Global approximation of self-optimizing controlled variables with average loss minimization[J]. Industrial & Engineering Chemistry Research, 2015, 54(48): 12040-12053.
[26]	梁超, 张泉灵. 催化重整装置反应器的建模与仿真[J]. 化工学报, 2012, 63(11): 3591-3596.
	Liang C, Zhang Q L. Modeling and simulation of continuous catalytic reforming reactor[J]. CIESC Journal, 2012, 63(11): 3591-3596.
[27]	Wei M, Yang M L, Qian F, et al. Dynamic modeling and economic model predictive control with production mode switching for an industrial catalytic naphtha reforming process[J]. Industrial & Engineering Chemistry Research, 2017, 56(31): 8961-8971.
[28]	Yusuf A Z, John Y M, Aderemi B O, et al. Modelling, simulation and sensitivity analysis of naphtha catalytic reforming reactions[J]. Computers & Chemical Engineering, 2019, 130: 106531.
[29]	周红军, 石铭亮, 翁惠新, 等. 芳烃型催化重整集总反应动力学模型[J]. 石油学报(石油加工), 2009, 25(4): 545-550.
	Zhou H J, Shi M L, Weng H X, et al. Lumped kinetic model of aromatic type catalytic naphtha reforming[J]. Acta Petrolei Sinica (Petroleum Processing Section), 2009, 25(4): 545-550.
[30]	LimaF S, Alves V M C, Araujo A C B. Metacontrol: a Python based application for self-optimizing control using metamodels[J]. Computers & Chemical Engineering, 2020, 140: 106979.
[31]	Khezri V, Panahi M, Yasari E, et al. Application of surrogate models as an alternative to process simulation for implementation of the self-optimizing control procedure on large-scale process plants: a natural gas-to-liquids (GTL) case study[J]. Industrial & Engineering Chemistry Research, 2021, 60(13): 4919-4929.
[32]	Sacks J, Welch W J, Mitchell T J, et al. Designs and analysis of computer experiments[J]. Statistical Science, 1989, 4(4): 409-435.
[33]	Johnson M E, Moore L M, Ylvisaker D. Minimax and maximin distance designs[J]. Journal of Statistical Planning and Inference, 1990, 26(2): 131-148.
[34]	McKay M D, Beckman R J, Conover W J. A comparison of three methods for selecting values of input variables in the analysis of output from a computer code[J]. Technometrics, 2000, 42(1): 55-61.
[35]	Murthy Konda N V S N, Rangaiah G P, Krishnaswamy P R. Plantwide control of industrial processes: an integrated framework of simulation and heuristics[J]. Industrial & Engineering Chemistry Research, 2005, 44(22): 8300-8313.
[36]	Skogestad S. Control structure design for complete chemical plants[J]. Computers & Chemical Engineering, 2004, 28(1/2): 219-234.
[37]	李艳强. 炼油厂催化重整装置的常见故障综述[J]. 化工设计通讯, 2017, 43(10): 83.
	Li Y Q. Summarization of common faults in catalytic reforming unit of refinery[J]. Chemical Engineering Design Communications, 2017, 43(10): 83.
[38]	王洪峰. 重整装置电加热器故障分析及处理[J]. 广东化工, 2020, 47(10): 126, 144.
	Wang H F. Failure analysis and treatment of electric heater in catalytic reforming unit[J]. Guangdong Chemical Industry, 2020, 47(10): 126, 144.
[39]	Li Z, Zhong W M, Wang X Q, et al. Control structure design of an industrial crude terephthalic acid hydropurification process with catalyst deactivation[J]. Computers & Chemical Engineering, 2016, 88: 1-12.
[40]	Vasudevan S, Rangaiah G P. Criteria for performance assessment of plantwide control systems[J]. Industrial & Engineering Chemistry Research, 2010, 49(19): 9209-9221.