复杂非线性系统的子系统分解方法

doi:10.11949/0438-1157.20201746

摘要/Abstract

摘要：

利用社区发现算法研究了一种复杂非线性化工系统的子系统分解方法，并进行了分布式模型预测控制设计。引入信息图论的节点表示系统的状态、输入和输出变量，构建非线性过程系统的加权有向图，节点通过加权边连接，加权反映了节点间连接的强度，因而能够同时反映系统内部的连通性和连接强度。利用社区结构发现算法将所有变量分成子系统的群组，使得每个组内的关联比不同组间的相互作用强，从而得到复杂化工过程系统的子系统分解。针对连续搅拌反应釜过程，实施子系统分解，并设计分布式模型预测控制算法，结果表明，所提出的子系统分解方法更能考虑子系统之间的连接权重，得到更有利于分布式模型预测控制的子系统划分，提升系统控制的性能。

关键词: 过程系统, 过程控制, 社区发现, 子系统分解, 模型预测控制

Abstract:

The community detection algorithm is used to study a subsystem decomposition method of a complex nonlinear chemical system, and a distributed model predictive control design is carried out. The nodes of information graph were used to represent the state variables, input variables and output variables of nonlinear system. The weighted directed graph of nonlinear process system was then constructed. The nodes were connected by weighted edges and the weight reflects the strength of the connection between nodes. Thus, the weighted directed graph can better reflect the internal connectivity and connection strength of the system. The community structure detection algorithm was used to divide all variables into the combination of subsystems to get the subsystem decomposition of complex nonlinear system. The correlation within each group was much stronger than the interaction between different groups. For the process of continuous stirred tank reactor, the subsystem decomposition was implemented and the distributed model predictive control algorithm was designed. The results show that the proposed subsystem decomposition method can take the connection weight of subsystems into account, which is more conducive to improve the performance of chemical process system with distributed model predictive control.

Key words: process systems, process control, community detection, subsystem decomposition, model-predictive control

中图分类号:

谢苗苗, 张浪文, 谢巍. 复杂非线性系统的子系统分解方法[J]. 化工学报, 2021, 72(3): 1557-1566.

XIE Miaomiao, ZHANG Langwen, XIE Wei. Subsystem decomposition of complex nonlinear systems[J]. CIESC Journal, 2021, 72(3): 1557-1566.

图/表 8

图1 复杂非线性系统的子系统的划分流程

Fig.1 Flowchart of the proposed subsystem decomposition method of nonlinear systems.

图2 两个CSTR和一个分离釜的复杂反应分离器系统

Fig.2 Two CSTRs and a flash tank with recycle stream

表1 式（29）、式（30）中的参数值

Table 1 Parameter values in Eq.(29) and Eq.(30)

参数值	单位
T₁₀=300，T₂₀=300	K
F₁₀=5，F₂₀=5，F_r=1.9	m³/h
C_A10=4，C_A20=3	kmol/m³
V₁=1.0，V₂=0.5，V₃=1.0	m³
E₁=5×10⁴，E₂=5.5×10⁴	kJ/kmol
k₁=3×10⁶，k₂=3×10⁶	h^-1
ΔH₁=-5×10⁴，ΔH₂=-5.3×10⁴	kJ/kmol
H_vap=5	kJ/kmol
C_P=0.231	kJ/(kg·K)
R=8.314	kJ/(kmol·K)
$ρ = 1000$	kg/m³
$α A = 2, α B = 1, α C = 1.5, α D = 3$
MW_A=50，MW_B=50，MW_C=50	kg/kmol

表1 式（29）、式（30）中的参数值

Table 1 Parameter values in Eq.(29) and Eq.(30)

参数值	单位
T₁₀=300，T₂₀=300	K
F₁₀=5，F₂₀=5，F_r=1.9	m³/h
C_A10=4，C_A20=3	kmol/m³
V₁=1.0，V₂=0.5，V₃=1.0	m³
E₁=5×10⁴，E₂=5.5×10⁴	kJ/kmol
k₁=3×10⁶，k₂=3×10⁶	h^-1
ΔH₁=-5×10⁴，ΔH₂=-5.3×10⁴	kJ/kmol
H_vap=5	kJ/kmol
C_P=0.231	kJ/(kg·K)
R=8.314	kJ/(kmol·K)
$ρ = 1000$	kg/m³
$α A = 2, α B = 1, α C = 1.5, α D = 3$
MW_A=50，MW_B=50，MW_C=50	kg/kmol

表2 系统(29)的三种分解

Table 2 Three decompositions for system (29)

分解	子系统	状态	输入	输出
分解1	1	T₁，C_A1，C_A2，C_A3	Q₁	T₁
	2	T₂，C_B1，C_B2，C_B3	Q₂	T₂
	3	T₃，C_C1，C_C2，C_C3	Q₃	T₃
分解2	1	T₁，C_A1，C_B1，C_C1	Q₁	T₁
	2	T₂，C_A2，C_B2，C_C2	Q₂	T₂
	3	T₃，C_A3，C_B3，C_C3	Q₃	T₃
分解3	1	T₁，C_A1，C_A3	Q₁	T₁
	2	T₂，C_A2	Q₂	T₂
	3	C_B1，C_C1，C_B2，C_C2，T₃，C_B3，C_C3	Q₃	T₃

图3 分解1（蓝色点划线）、分解2（红色实线）和分解3（黑色虚线）的闭环系统的温度和输入轨迹

Fig.3 Temperature and input trajectories of the nominal closed-loop system under Decomposition 1 (blue dash dotted line), Decomposition 2 (red solid line), and Decomposition 3 (black dotted line)

表3 不同分解模型的迭代分布式MPC和集中式MPC的性能指标（J）和CPU时间

Table 3 Performance index (J) and CPU time of iterative distributed MPC with different decomposition models and centralized MPC

分解	J	CPU time/s
分解1	298.6897	26.622806
分解2	297.3096	25.406688
分解3	295.0141	25.270445
集中式	276.6084	24.151277

图4 分布式LMPC在迭代1次（红色实线），2次（黑色虚线），3次（蓝色点划线）时的温度和输入轨迹

Fig.4 The temperature and input trajectories of distributed LMPC when iteration times is 1(red solid line), 2 (black dotted line) and 3 (blue dash dotted line)

表5 不同迭代次数的分布式LMPC的性能指标和CPU时间

Table 5 Performance index and CPU time of distributed LMPC with different iteration times

迭代次数	J	CPU time/s
1次	338.5985	12.626477
2次	297.3096	25.406688
3次	295.3047	41.282411

参考文献 31

1	Daoutidis P, Zachar M, Jogwar S S. Sustainability and process control: a survey and perspective[J]. J. Process. Contr., 2016, 44: 184-206.
2	周红标, 张钰, 柏小颖, 等. 基于自适应模糊神经网络的非线性系统模型预测控制[J]. 化工学报, 2020, 71(7): 3201-3212.
	Zhou H B, Zhang Y, Bai X Y, et al. Model predictive control of nonlinear system based on adaptive fuzzy neural network[J]. CIESC Journal, 2020, 71(7): 3201-3212.
3	Qin S J, Badgwell T A. A survey of industrial model predictive control technology[J]. Control. Eng. Pract., 2003, 11(7): 733-764.
4	Rawlings J B, Stewart B T. Coordinating multiple optimization-based controllers: new opportunities and challenges[J]. J. Process. Contr., 2008, 18(9): 839-845.
5	Christofides P D, Scattolini R, de la Pena D M, et al. Distributed model predictive control: a tutorial review and future research directions[J]. Comput. Chem. Eng., 2013, 51: 21-41.
6	Raimondo D M, Magni L, Scattolini R. Decentralized MPC of nonlinear systems: an input-to-state stability approach[J]. Int. J. Robust. Nonlin. Contr., 2007, 17(17): 1651-1667.
7	Magni L, Scattolini R. Stabilizing decentralized model predictive control of nonlinear systems[J]. Automatica, 2006, 42(7): 1231-1236.
8	Stanković S S, Stipanović D M, Šiljak D D. Decentralized dynamic output feedback for robust stabilization of a class of nonlinear interconnected systems[J]. Automatica, 2007, 43(5): 861-867.
9	Hong C, Jacobsen E W. Performance limitations in decentralized control[J]. J. Process. Contr., 2002, 12(4): 485-494.
10	Zhang L W, Xie W, Liu J F. Robust control of saturating systems with Markovian packet dropouts under distributed MPC[J]. ISA Trans., 2019, 85: 49-59.
11	Liu S, Liu J F. Distributed Lyapunov-based model predictive control with neighbor-to-neighbor communication[J]. AIChE J., 2014, 60(12): 4124-4133.
12	Scattolini R. Architectures for distributed and hierarchical model predictive control: a review[J]. J. Process. Contr., 2009, 19(5): 723-731.
13	Farina M, Scattolini R. Distributed predictive control: a non-cooperative algorithm with neighbor-to-neighbor communication for linear systems[J]. Automatica, 2012, 48(6): 1088-1096.
14	Heidarinejad M, Liu J F, de la Peña D M, et al. Multirate Lyapunov-based distributed model predictive control of nonlinear uncertain systems[J]. J. Process. Contr., 2011, 21(9): 1231-1242.
15	Dunbar W B. Distributed receding horizon control of dynamically coupled nonlinear systems[J]. IEEE T. Automat. Contr., 2007, 52(7): 1249-1263.
16	Heo S, Marvin W A, Daoutidis P. Automated synthesis of control configurations for process networks based on structural coupling[J]. Chem. Eng. Sci., 2015, 136: 76-87.
17	Heo S, Daoutidis P. Control-relevant decomposition of process networks via optimization-based hierarchical clustering[J]. AIChE J., 2016, 62(9): 3177-3188.
18	Kang L X, Tang W T, Liu Y Z, et al. Control configuration synthesis using agglomerative hierarchical clustering: a graph-theoretic approach[J]. J. Process. Contr., 2016, 46: 43-54.
19	Yin X Y, Liu J F. Input-output pairing accounting for both structure and strength in coupling[J]. AIChE J., 2017, 63(4), 1226-1235.
20	Tang W T, Allman A, Pourkargar D B, et al. Optimal decomposition for distributed optimization in nonlinear model predictive control through community detection[J]. Comput. Chem. Eng., 2018, 111: 43-54.
21	Tang W T, Pourkargar D B, Daoutidis P. Relative time-averaged gain array (RTAGA) for distributed control-oriented network decomposition[J]. AIChE J., 2018, 64(5): 1682-1690.
22	Yin X Y, Decardi-Nelson B, Liu J F. Subsystem decomposition and distributed moving horizon estimation of wastewater treatment plants[J]. Chem. Eng. Res. Des., 2018, 134: 405-419.
23	Yin X Y, Arulmaran K, Liu J F, et al. Subsystem decomposition and configuration for distributed state estimation[J]. AIChE J., 2016, 62(6): 1995-2003.
24	Yin X Y, Liu J F. Subsystem decomposition of process networks for simultaneous distributed state estimation and control[J]. AIChE J., 2019, 65(3): 904-914.
25	Zhang L W, Yin X Y, Liu J F. Complex system decomposition for distributed state estimation based on weighted graph[J]. Chem. Eng. Res. Des., 2019, 151: 10-22.
26	杨晓峰, 谢巍, 张浪文. 加权有向图社区发现的子系统划分[J]. 控制理论与应用, 2020, 37(9): 1923-1930.
	Yang X F, Xie W, Zhang L W. Weighted directed graph based community detection for subsystem decomposition[J]. Control Theory & Applications, 2020, 37(9): 1923-1930.
27	Leicht E A, Newman M E J. Community structure in directed networks[J]. Phys. Rev. Let., 2007, 100(11): 118703.
28	Newman M E J. Modularity and community structure in networks[J]. PNAS, 2006, 103(23): 8577-8582.
29	Liu J F, Ohran B J, de la Pena D M, et al. Monitoring and handling of actuator faults in two-tier control systems for nonlinear processes[J]. Chem. Eng. Sci., 2010, 65(10): 3179-3190.
31	Pourkargar D B, Almansoori A, Daoutidis P. Impact of decomposition on distributed model predictive control: a process network case study[J]. Ind. Eng. Chem. Res., 2017, 56(34): 9606-9616.

[1]	陈哲文, 魏俊杰, 张玉明. 超临界水煤气化耦合SOFC发电系统集成及其能量转化机制[J]. 化工学报, 2023, 74(9): 3888-3902.
[2]	曹跃, 余冲, 李智, 杨明磊. 工业数据驱动的加氢裂化装置多工况切换过渡状态检测[J]. 化工学报, 2023, 74(9): 3841-3854.
[3]	康飞, 吕伟光, 巨锋, 孙峙. 废锂离子电池放电路径与评价研究[J]. 化工学报, 2023, 74(9): 3903-3911.
[4]	郑玉圆, 葛志伟, 韩翔宇, 王亮, 陈海生. 中高温钙基材料热化学储热的研究进展与展望[J]. 化工学报, 2023, 74(8): 3171-3192.
[5]	李贵贤, 曹阿波, 孟文亮, 王东亮, 杨勇, 周怀荣. 耦合固体氧化物电解槽的CO₂制甲醇过程设计与评价研究[J]. 化工学报, 2023, 74(7): 2999-3009.
[6]	邵远哲, 赵忠盖, 刘飞. 基于共同趋势模型的非平稳过程质量相关故障检测方法[J]. 化工学报, 2023, 74(6): 2522-2537.
[7]	苏晓丹, 朱干宇, 李会泉, 郑光明, 孟子衡, 李防, 杨云瑞, 习本军, 崔玉. 湿法磷酸半水工艺考察与石膏结晶过程研究[J]. 化工学报, 2023, 74(4): 1805-1817.
[8]	贠程, 王倩琳, 陈锋, 张鑫, 窦站, 颜廷俊. 基于社团结构的化工过程风险演化路径深度挖掘[J]. 化工学报, 2023, 74(4): 1639-1650.
[9]	王子宗, 索寒生, 赵学良. 数字孪生智能乙烯工厂研究与构建[J]. 化工学报, 2023, 74(3): 1175-1186.
[10]	张中秋, 李宏光, 石逸林. 基于人工预测调控策略的复杂化工过程多任务学习方法[J]. 化工学报, 2023, 74(3): 1195-1204.
[11]	张江淮, 赵众. 碳三加氢装置鲁棒最小协方差约束控制及应用[J]. 化工学报, 2023, 74(3): 1216-1227.
[12]	王雅琳, 潘雨晴, 刘晨亮. 基于GSA-LSTM动态结构特征提取的间歇过程监测方法[J]. 化工学报, 2022, 73(9): 3994-4002.
[13]	周乐, 沈程凯, 吴超, 侯北平, 宋执环. 深度融合特征提取网络及其在化工过程软测量中的应用[J]. 化工学报, 2022, 73(7): 3156-3165.
[14]	王琨, 侍洪波, 谭帅, 宋冰, 陶阳. 局部时差约束邻域保持嵌入算法在故障检测中的应用[J]. 化工学报, 2022, 73(7): 3109-3119.
[15]	杨岭, 崔国民, 周志强, 肖媛. 精细搜索策略应用于质量交换网络综合[J]. 化工学报, 2022, 73(7): 3145-3155.