Select Filter-Wrapper characteristic variables for yield prediction of fluid catalytic cracking unit

doi:10.11949/j.issn.0438-1157.20170841

Abstract

Abstract:

Fluid catalytic cracking is an important means to improve quality of heavy oil. Proper mathematical models are required to investigate influence of operating conditions and raw material properties on product distribution. Selection of suitable input variables largely affects model performances. The input variable selection is currently dependent on understanding mechanisms of FCC process. From the viewpoint of data-driven modeling, a method of selecting Filter-Wrapper characteristic subset was proposed by combination of Filter method using classical RReliefF algorithm and Wrapper method using GA-SVR algorithm. With no requirement of prior FCC knowledge, this method chose input variables by spontaneous data-driven selection of characteristic variables. A model with good prediction accuracy and proper number of input variables was established by taking advantage of a FCCU operating data and selecting input variables for prediction model of the unit dry gas and coke yield. Present work can not only provide a method to variable selection in FCCU modeling and process analysis, but also extend to other industrial process analysis.

Key words: model reduction, algorithm, genetic algorithm, feature selection, yield prediction

摘要：

催化裂化过程是重质油轻质化的重要手段，为了研究操作条件、原料性质等因素对产品分布的影响，通常需要对催化裂化过程建立准确可靠的数学模型。选择合适的输入变量对模型预测效果有着较大的影响，而在现有的催化裂化装置模型中，输入变量的选取主要依赖于对催化裂化机理的理解。本文从数据驱动建模的角度出发，提出一种Filter法与Wrapper法联合使用的特征子集选择方法。该方法在输入变量选取的过程中不依赖于催化裂化的先验知识，是一种数据驱动的自发的特征变量选择过程。以某炼油厂催化裂化装置为研究对象，利用该装置的生产数据分别选择用于干气和焦炭产率预测模型的输入变量，建立了预测精度高、输入变量数目适中的模型。此外，该方法为催化裂化装置建模的变量选取提供了新角度。

关键词: 模型简化, 算法, 遗传算法, 特征选择, 产率预测

CLC Number:

WANG Jie, CAO Daofan, LAN Xingying, GAO Jinsen. Select Filter-Wrapper characteristic variables for yield prediction of fluid catalytic cracking unit[J]. CIESC Journal, 2018, 69(1): 464-471.

王杰, 曹道帆, 蓝兴英, 高金森. 用于催化裂化装置产率预测的Filter-Wrapper特征变量选择方法[J]. 化工学报, 2018, 69(1): 464-471.

References 33

[1]	徐春明,杨朝合. 石油炼制工程[M]. 北京:中国石油工业出版社, 2009. XU C M, YANG C H. Petroleum Refinery Engineering[M]. Beijing:China Petroleum Press, 2009
[2]	张结喜, 齐艳华, 邱建章. 催化裂化关联模型的研究[J]. 计算机与应用化学, 2007, 24(11):81-84. ZHANG J X, LIU Y H, QIU J Z. Study on FCC correlation models[J]. Computers and Applied Chemistry, 2007, 24(11):81-84
[3]	BLANDING F H. Reaction rates in catalytic cracking of petroleum[J]. Industrial & Engineering Chemistry, 1953, 45(6):1186-1197.
[4]	翁惠新, 欧阳福生, 马军. 重油催化裂化反应集总动力学模型(Ⅰ):模型的建立[J]. 化工学报, 1995, 46(6):662-668. WENG H X, OUYANG F S, MA J. Lumping kinetic model for heavy oil catalytic cracking(Ⅰ):Model establishment[J]. Journal of Chemical Industry and Engineering(China), 1995, 46(6):662-668
[5]	翁惠新, 马军. 重油催化裂化反应集总动力学模型(Ⅱ):模型的工业验证[J]. 化工学报, 1995, 46(6):669-674. WENG H X, MA J. Lumping kinetic model for heavy oil catalytic cracking(Ⅱ):Industrial validation of the model[J]. Journal of Chemical Industry and Engineering(China), 1995, 46(6):669-674.
[6]	熊凯, 卢春喜. 催化裂化(裂解)集总反应动力学模型研究进展[J]. 石油学报(石油加工), 2015, 31(2):293-306. XIONG K, LU C X. Research progresses of lump kinetic model of FCC and catalytic pyrolysis[J]. Acta Petrolei Sinica(Petroleum Processing Section), 2015, 31(2):293-306.
[7]	JACOB S M, BENJAMIN G, VOLTZ S E, et al. A lumping and reaction scheme for catalytic cracking[J]. AIChE Journal, 1976, 22(4):701-713.
[8]	欧阳福生, 刘永吉. 集总动力学模型结合神经网络预测催化裂化产物收率[J]. 石油化工, 2017, (1):9-16. OUYANG F S, LIU Y J. Prediction of the product yield from catalytic cracking process by lumped kinetic model combined with neural network[J]. Petrochemical Technology, 2017, (1):9-16.
[9]	MICHALOPOULOS J, PAPADOKONSTADAKIS S, ARAMPATZIS G, et al. Modelling of an industrial fluid catalytic cracking unit using neural networks[J]. Chemical Engineering Research & Design, 2001, 79(A2):137-142.
[10]	LIGURAS D K, ALLEN D T. Structural models for catalytic cracking(1):Model compound reactions[J]. Industrial & Engineering Chemistry Research, 1989, 28(6):665-673.
[11]	欧阳福生, 王磊, 王胜, 等. 催化裂解过程分子尺度反应动力学模型研究[J]. 高校化学工程学报, 2008, 22(6):927-934. OUYANG F S, WANG L, WANG S, et al. Molecular reaction kinetics model for deep catalytic cracking[J]. Journal of Chemical Engineering of Chinese Universities, 2008, 22(6):927-934.
[12]	郝鑫. 广义回归神经网络和遗传算法研究及其在化工过程建模中的应用[D]. 杭州:浙江大学, 2004. HAO X. The study of general regression neural network and genetic lgorithms and their application for chemical engineering[D]. Hangzhou:Zhejiang University, 2004
[13]	GONZAGA J, MELEIRO L A C, KIANG C, et al. ANN-based soft-sensor for real-time process monitoring and control of an industrial polymerization process[J]. Computers & Chemical Engineering, 2009, 33(1):43-49.
[14]	苏鑫, 吴迎亚, 裴华健, 等. 大数据技术在过程工业中的应用研究进展[J]. 化工进展, 2016, 35(6):1652-1659. SU X, WU Y Y, PEI H J, et al. Recent development of the application of big data technology in process industries[J]. Chemical Industry and Engineering Progress, 2016, 35(6):1652-1659.
[15]	李鹏, 郑晓军, 明梁, 等. 大数据技术在催化裂化装置运行分析中的应用[J]. 化工进展, 2016, 35(3):665-670. LI P, ZHENG X J, MING L, et al. Application of big data technology in operation analysis of catalytic cracking[J]. Chemical Industry and Engineering Progress, 2016, 35(3):665-670.
[16]	GUYON I, ELISSEEFF A. An Introduction to Variable and Feature Selection[M]. JMLR.org, 2003.
[17]	BERMEJO P, GÁMEZ J A, PUERTA J M. A GRASP algorithm for fast hybrid (Filter-Wrapper) feature subset selection in high-dimensional datasets[J]. Pattern Recognition Letters, 2011, 32(5):701-711.
[18]	CERVANTE L, XUE B, ZHANG M, et al. Binary particle swarm optimisation for feature selection:a Filter based approach[C]//Proceedings of the Evolutionary Computation, 2012.
[19]	MALDONADO S, WEBER R. A Wrapper method for feature selection using support vector machines[J]. Information Sciences, 2008, 179(13):2208-2217.
[20]	SEBBAN M, NOCK R. A hybrid Filter/Wrapper approach of feature selection using information theory[J]. Pattern Recognition, 2002, 35(4):835-846.
[21]	HSU H H, HSIEH C W, LU M D. Hybrid feature selection by combining Filters and Wrappers[J]. Expert Systems with Applications, 2011, 38(7):8144-8150.
[22]	ROBNIK ŠIKONJA M, KONONENKO I. An adaptation of Relief for attribute estimation in regression[C]//Proceedings of the Fourteenth International Conference on Machine Learning, 1997.
[23]	ROBNIK ŠIKONJA M, KONONENKO I. Theoretical and empirical analysis of ReliefF and RReliefF[J]. Machine Learning, 2003, 53(1):23-69.
[24]	张亚乐, 徐博文, 方崇智, 等. 一种改进的遗传算法在原油蒸馏过程优化中的应用[J]. 化工自动化及仪表, 1997, (3):12-17. ZHANG Y L, XU B W, FANG C Z, et al. Application of an improved genetic algorithm in optimization of crude distillation process[J]. Control and Instruments in Chemical Industry, 1997, (3):12-17.
[25]	何险峰, 周家驹. 遗传算法及其在化学化工中的应用[J]. 化学进展, 1998, 10(3):312-318. HE X F, ZHOU J J. Genetic algorithms and their applications in chemistry and chemical engineering[J]. Progress in Chemistry, 1998, 10(3):312-318.
[26]	王举, 陈中州, 袁希钢. 用于间歇化工过程最优设计的遗传算法[J]. 计算机与应用化学, 1999, (1):24-28. WANG J, CHEN Z Z, YUAN X G. A genetic algorithm for optimal design of batch processes[J]. Computers and Applied Chemistry, 1999, (1):24-28.
[27]	王定成, 方廷健, 高理富, 等. 支持向量机回归在线建模及应用[J]. 控制与决策, 2003, 18(1):89-91. WANG D C, FANG Y J, GAO F L, et al. Support vector machines regression on-line modelling and its application[J]. Control and Decision, 2003, 18(1):89-91.
[28]	朱树先, 张仁杰. 支持向量机核函数选择的研究[J]. 科学技术与工程, 2008, 8(16):4513-4517. ZHU S X, ZHANG R J. Study on selection of SVM kernel function[J]. Science Technology and Engineering, 2008, 8(16):4513-4517.
[29]	胡局新, 张功杰. 基于K折交叉验证的选择性集成分类算法[J]. 科技通报, 2013, (12):115-117. HU J X, ZHANG G J. K-fold cross-validation based selected ensemble classification algorithm[J]. Bulletin of Science and Technology, 2013, (12):115-117.
[30]	许友好, 龚剑洪, 张久顺, 等. 降低干气和焦炭产率的MIP工艺研究[J]. 石油炼制与化工, 2007, 38(10):7-11. XU Y H, GONG J H, ZHANG J S, et al. Study on the MIP process for reducing dry gas and coke yields[J]. Petroleum Processing and Petrochemicals, 2007, 38(10):7-11.
[31]	张丽新. 高维数据的特征变量选择及基于特征变量选择的集成学习研究[D]. 北京:清华大学, 2004. ZHANG L X. Study on feature selection and ensemble learning based on feature selection for high-dimensional datasets[D]. Beijing:Tsinghua University, 2004.
[32]	苏鑫, 裴华健, 吴迎亚, 等. 应用经遗传算法优化的BP神经网络预测催化裂化装置焦炭产率[J]. 化工进展, 2016, 35(2):389-396. SU X, PEI H J, WU Y Y, et al. Predicting coke yield of FCC unit using genetic algorithm optimized BP neural network[J]. Chemical Industry and Engineering Progress, 2016, 35(2):389-396.
[33]	欧阳福生, 方伟刚, 唐嘉瑞, 等. 以BP神经网络为基础的MIP工艺过程产品分布优化[J]. 石油炼制与化工, 2016, 47(5):95-100. OUYANG F S, FANG W G, TANG J R, et al. Optimizing product distribution of MIP process using BP neural network[J]. Petroleum Processing and Petrochemicals, 2016, 47(5):95-100.