基于迁移学习的有机硅单体分馏过程能耗建模

doi:10.11949/0438-1157.20200402

摘要/Abstract

摘要：

有机硅单体分馏过程需分离的物质多且分离技术要求高，是有机硅单体生产过程中能耗较高的单元之一。能耗模型的建立对分馏过程的生产优化有着重要的意义。虽然分馏过程中各提纯单元的塔器设备型号、操作条件存在差异，但由于各提纯单元存在相似性，如何充分利用这些相似性进行能耗建模，从而减少建模过程中对数据量、质的要求是要解决的问题。通过引入迁移学习算法，增大源单元和目标单元数据域的相似性，实现不同数据域的知识传递，充分利用现有数据中包含的知识，建立新的能耗模型。以某有机硅企业采集的现场数据进行算法验证，结果显示，当训练样本量不足时，迁移学习可以有效提升模型性能，从而验证了该算法的有效性与实际应用价值。

关键词: 迁移学习, 有机硅单体, 蒸馏, 预测, 集成, 能耗

Abstract:

Organosilicon monomer fractionation process needs to separate many substances and the separation requirements are high, leading to high energy consumption. The modeling of energy consumption is of great significance to the optimization of fractionation process. Although the type of equipment and operating conditions of each purification unit in the fractionation process are different, each purification unit has certain similarity according to the dynamic characteristics of the distillation process. Therefore, making full use of these similarities in modeling can reduce the cost of data acquisition in the modeling process. By introducing the transfer learning algorithm, the similarity between the data domains of the source unit and the target unit is increased, the knowledge transfer in different data domains is realized, the knowledge contained in the existing data is fully utilized, and a new energy consumption model is established. Finally, the simulation experiment was carried out with the field data of an organosilicon factory, and the results showed that when the sample size of target domain is little, the transfer learning can effectively improve the performance of the model, so the effectiveness and the practical application of the algorithm were verified.

Key words: transfer learning, organosilicon monomer, distillation, prediction, integration, energy consumption

中图分类号:

TP 181

平晓静,赵顺毅,栾小丽,刘飞. 基于迁移学习的有机硅单体分馏过程能耗建模[J]. 化工学报, 2020, 71(12): 5706-5714.

PING Xiaojing,ZHAO Shunyi,LUAN Xiaoli,LIU Fei. Modeling of energy consumption in organosilicon monomer fractionation process based on transfer learning[J]. CIESC Journal, 2020, 71(12): 5706-5714.

图/表 8

图1 有机硅单体分馏流程

Fig.1 Fractionation process of organosilicon monomer

图2 数据预处理前后能耗指标曲线

Fig.2 Energy consumption curves before and after data preprocessing

图3 数据预处理前后PCA得分图

Fig.3 Scores plots of PCA before and after data preprocessing

图4 迁移学习建模流程示意图

Fig.4 Modeling schematic diagram of transfer learning

表1 三种回归模型的RMSE和MAPE

Table 1 RMSE and MAPE of PLS, BP and RBF

回归算法	性能指标
回归算法	RMSE/%	MAPE/%
PLS	3.6399	0.4146
BP	4.0994	0.5689
RBF	6.2147	0.6185

图5 三种模型在训练集上的RMSE曲线

Fig.5 RMSE curves on training data set of three models

图6 不同目标域样本比例下的迁移学习模型性能提升

Fig.6 Performance improvement of transfer learning for different target domain data percentage

表2 源域和目标域的MMD距离

Table 2 MMD distance of source and target domains

三甲塔样本比例	MMD距离
三甲塔样本比例	迁移算法前	迁移算法后
0.05 0.1	0.9480 0.9855	0.3494 0.3817
0.2	0.9730	0.3512
0.3 0.4 0.5 0.6 0.7 0.8	1.0025 1.0071 0.9951 0.9935 0.9963 0.9874	0.4087 0.2561 0.2337 0.2248 0.2199 0.4398

参考文献 38

1	Eduok U, Faye O, Szpunar J. Recent developments and applications of protective silicone coatings: a review of PDMS functional materials[J]. Progress in Organic Coatings, 2017, 111: 124-163.
2	Zhang S. Application of mathematical modeling in optimization of synthesis process parameters of methylchlorosilane[J]. Journal of Photonic Materials and Technology, 2018, 4(2): 49-54.
3	Luo W, Wang G, Wang J, et al. Effect and role of Al addition in the direct synthesis reaction of methylchlorosilane monomer[J]. Journal of Chemical Engineering of Chinese Universities, 2005, 19(6): 803-807.
4	傅积赉. 对直接法合成甲基氯硅烷生产工艺复杂性的认识与建议[J]. 有机硅材料, 2004, 18(1): 1-4.
	Fu J L. Acquaintance & suggestions on complexity of production technology on direct synthesis of methylchlorosilane[J]. Silicone Material, 2004, 18(1): 1-4.
5	Hitsov I, de Sitter K, Dotremont C, et al. Economic modelling and model-based process optimization of membrane distillation[J]. Desalination, 2018, 436: 125-143.
6	Eleiwi F, Ghaffour N, Alsaadi A S, et al. Dynamic modeling and experimental validation for direct contact membrane distillation (DCMD) process[J]. Desalination, 2016, 384: 1-11.
7	Ruiz A, Blanco C, Ozkan L. Modeling of reactive batch distillation processes for control[J]. Computers and Chemical Engineering, 2019, 121: 86-98.
8	李凤莲, 任瑞平, 邹雪梅. 化工精馏高效节能技术开发及应用[J]. 化工管理, 2014, (18): 162.
	Li F L, Ren R P, Zou X M. Development and application of high efficiency and energy saving technology for chemical distillation[J]. Chemical Enterprise Management, 2014, (18): 162.
9	何小阳, 李健, 闵力, 等. 精馏塔的机理-神经网络混合建模[J]. 控制工程, 2009, 16(2): 94-96.
	He X Y, Li J, Min L, et al. Modeling of distillation column based on hybrid neural network and prior knowledge[J]. Control Engineering of China, 2009, 16(2): 94-96.
10	于丙芹, 张贝克, 孙军, 等. 精馏过程动态仿真建模[J]. 计算机与应用化学, 2011, 28(9): 1219-1223.
	Yu B Q, Zhang B K, Sun J, et al. Dynamic simulation and modeling of distillation process[J]. Computers and Applied Chemistry, 2011, 28(9): 1219-1223.
11	Pino-Mejías R, Pérez-Fargallo A, Rubio-Bellido C, et al. Comparison of linear regression and artificial neural networks models to predict heating and cooling energy demand, energy consumption and CO2 emissions[J]. Energy, 2017, 118: 24-36.
12	Wang Z, Shao C, Zhu L. Soft-sensing modeling and intelligent optimal control strategy for distillation yield rate of atmospheric distillation oil refining process[J]. Chinese Journal of Chemical Engineering, 2019, 27(5): 1113-1124.
13	Brahim A, Abderafi S, Bounahmidi T. Optimization of the distillation column of petroleum fractions using ANN method[C]//2018 IEEE 6th International Renewable and Sustainable Energy Conference (IRSEC), 2018: 1-6.
14	Ibrahim D, Jobson M, Li J, et al. Optimization-based design of crude oil distillation units using surrogate column models and a support vector machine[J]. Chemical Engineering Research and Design, 2018, 134: 212-225.
15	Osuolale F N, Zhang J. Energy efficiency optimisation for distillation column using artificial neural network models[J]. Energy, 2016, 106: 562-578.
16	Osuolale F, Zhang J. Exergetic Optimisation of Atmospheric and Vacuum Distillation System based on Bootstrap Aggregated Neural Network Models[M]. Switzerland: Springer, Cham, 2018: 1033-1046.
17	Chan L, Chen J. Economic model predictive control of distillation startup based on probabilistic approach[J]. Chemical Engineering Science, 2018, 186: 26-35.
18	Pan S, Yang Q. A survey on transfer learning[J]. IEEE Transactions on Knowledge and Data Engineering, 2009, 22(10): 1345-1359.
19	Wang X, Ren J, Liu S. Distribution adaptation and manifold alignment for complex processes fault diagnosis[J]. Knowledge-Based Systems, 2018, 156: 100-112.
20	Yang C, Chen B, Wang Z, et al. Transfer learning soft sensor for product quality prediction in multi-grade processes[C]//2019 IEEE 8th Data Driven Control and Learning Systems Conference (DDCLS). 2019: 1148-1153.
21	刘佳, 邵诚, 朱理. 基于迁移学习工况划分的裂解炉收率PSO-LS-SVM建模[J]. 化工学报, 2016, 67(5): 1982-1988.
	Liu J, Shao C, Zhu L. Modeling of cracking furnace yields with PSO-LS-SVM based on operating condition classification by transfer learning[J]. CIESC Journal, 2016, 67(5): 1982-1988.
22	周书恒, 杜文莉. 基于迁移学习的裂解炉产率建模[J]. 化工学报, 2014, 65(12): 4921-4928.
	Zhou S H, Du W L. Modeling of ethylene cracking furnace yields based on transfer learning[J]. CIESC Journal, 2014, 65(12): 4921-4928.
23	Han T, Liu C, Yang W, et al. Deep transfer network with joint distribution adaptation: a new intelligent fault diagnosis framework for industry application[J]. ISA Transactions, 2020, 97: 269-281.
24	Lu J, Yao Y, Gao F. Model migration for development of a new process model[J]. Industrial & Engineering Chemistry Research, 2009, 48(21): 9603-9610.
25	Lu J, Gao F. Process modeling based on process similarity[J]. Industrial & Engineering Chemistry Research, 2008, 47(6): 1967-1974.
26	孙建阳, 孔令启, 张传国, 等. 国内甲基氯硅烷分离技术现状及改进方法[J].有机硅材料, 2015, 29(1): 47-51.
	Sun J Y, Kong L Q, Zhang C G, et al. Status of separation technology of methyl chlorosilane and its improvements in China[J]. Silicone Material, 2015, 29(1): 47-51.
27	段继海, 王文建, 范军领, 等. 有机硅单体分离流程的模拟与优化[J]. 化学工程, 2012, 40(8): 69-73.
	Duan J H, Wang W J, Fan J L, et al. Simulation and optimization of organosilicon monomer separation processes[J]. Chemical Engineering(China), 2012, 40(8): 69-73.
28	高维平, 杨莹, 刘学线, 等. 化工精馏高效节能技术开发及应用[J]. 计算机与应用化学, 2008, 25(12): 1531-1536.
	Gao W P, Yang Y, Liu X X, et al. Development and application of high efficiency energy-saving distillation technology[J]. Computers and Applied Chemistry, 2008, 25(12): 1531-1536.
29	李英劼. 化工生产中降低精馏技术能耗的思路[J]. 石油和化工设备, 2011, 14(1): 58-60.
	Li Y J. Chemical Engineering produces inside lowers the way of thinking that tower consume[J]. Petro & Chemical Equipment, 2011, 14(1): 58-60.
30	黄克谨, 钱积新, 孙优贤, 等. 一种改进的多元精馏塔动态模型[J]. 化工学报, 1992, 43(4): 482-488.
	Huang K J, Qian J X, Sun Y X, et al. An improved dynamic model of multicomponent distillation column[J]. Journal of Chemical Industry and Engineering(China), 1992, 43(4): 482-488.
31	Hong J, Yin J, Huang Y, et al. TrSVM: a transfer learning algorithm using domain similarity[J]. Journal of Computer Research and Development, 2011, 48(10): 1823-1830.
32	刘强, 秦泗钊. 过程工业大数据建模研究展望[J]. 自动化学报, 2016, 42(2): 3-13.
	Liu Q, Qin S Z. Perspectives on big data modeling of process industries[J]. Acta Automatica Sinica, 2016, 42(2): 3-13.
33	Alma Ö. Comparison of robust regression methods in linear regression [J]. International Journal of Contemporary Mathematical Sciences, 2011, 6(9): 409-421.
34	Borgwardt K, Gretton A, Rasch M, et al. Integrating structured biological data by kernel maximum mean discrepancy[J]. Bioinformatics, 2006, 22(14): 49-57.
35	Pardoe D, Stone P. Boosting for regression transfer[C]//Proceedings of the 27th International Conference on Machine Learning. Haifa, Israel, 2010: 863-870.
36	Khan M, Heisterkamp D. Adapting instance weights for unsupervised domain adaptation using quadratic mutual information and subspace learning[C]//2016 IEEE 23rd International Conference on Pattern Recognition (ICPR). 2016: 1560-1565.
37	Dai W, Yang Q, Xue G, et al. Boosting for transfer learning[C]//Proceedings of the 24th International Conference on Machine Learning. Corvalis Oregon. USA, 2007: 193-200.
38	Sarstedt M, Ringle C, Hair J. Partial Least Squares Structural Equation Modeling[M]. Springer International Publishing, 2017.

[1]	杨欣, 王文, 徐凯, 马凡华. 高压氢气加注过程中温度特征仿真分析[J]. 化工学报, 2023, 74(S1): 280-286.
[2]	连梦雅, 谈莹莹, 王林, 陈枫, 曹艺飞. 地下水预热新风一体化热泵空调系统制热性能研究[J]. 化工学报, 2023, 74(S1): 311-319.
[3]	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.
[4]	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.
[5]	闫琳琦, 王振雷. 基于STA-BiLSTM-LightGBM组合模型的多步预测软测量建模[J]. 化工学报, 2023, 74(8): 3407-3418.
[6]	尹刚, 李伊惠, 何飞, 曹文琦, 王民, 颜非亚, 向禹, 卢剑, 罗斌, 卢润廷. 基于KPCA和SVM的铝电解槽漏槽事故预警方法[J]. 化工学报, 2023, 74(8): 3419-3428.
[7]	王光, 单发顺, 钱禹丞, 焦建芳. 基于集成学习传递熵的化工过程微小故障检测方法[J]. 化工学报, 2023, 74(7): 2967-2978.
[8]	余娅洁, 李静茹, 周树锋, 李清彪, 詹国武. 基于天然生物模板构建纳米材料及集成催化剂研究进展[J]. 化工学报, 2023, 74(7): 2735-2752.
[9]	郭雨莹, 敬加强, 黄婉妮, 张平, 孙杰, 朱宇, 冯君炫, 陆洪江. 稠油管道水润滑减阻及压降预测模型修正[J]. 化工学报, 2023, 74(7): 2898-2907.
[10]	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.
[11]	于源, 陈薇薇, 付俊杰, 刘家祥, 焦志伟. 几何相似涡流空气分级机环形区流场变化规律研究及预测[J]. 化工学报, 2023, 74(6): 2363-2373.
[12]	王光宇, 张锴, 张凯华, 张东柯. 微波加热干燥煤泥热质传递及其能耗特性分析[J]. 化工学报, 2023, 74(6): 2382-2390.
[13]	李艳辉, 丁邵明, 白周央, 张一楠, 于智红, 邢利梅, 高鹏飞, 王永贞. 非常规服役超临界锅炉的微纳尺度腐蚀动力学模型建立及应用[J]. 化工学报, 2023, 74(6): 2436-2446.
[14]	刘尚豪, 贾胜坤, 罗祎青, 袁希钢. 基于梯度提升决策树的三组元精馏流程结构最优化[J]. 化工学报, 2023, 74(5): 2075-2087.
[15]	杨灿, 孙雪琦, 尚明华, 张建, 张香平, 曾少娟. 相变离子液体体系吸收分离CO₂的研究现状及展望[J]. 化工学报, 2023, 74(4): 1419-1432.