乙烯精馏过程软测量技术应用研究
|
摘要
随着现代石化企业对控制、计量、节能增效和运行可靠性等要求的不断提高,各种测量要求也日益增多。在实际生产过程中,存在着一大类变量,由于技术或经济等多重原因,目前尚难以或无法直接通过传感器进行检测,但同时又是需要加以严格控制的、与产品质量关系非常密切的过程参数。这将直接影响到产品的质量以及系统的稳定运行,给企业带来不可低估的经济损失。软测量技术随即得到快速发展。
本文研究了软测量技术的若干重要方面,以最小二乘法和神经网络的理论为基础,在深入了解乙烯精馏塔工艺流程及精馏原理的基础上,进行乙烯精馏塔出口乙烯纯度软测量系统的研究。针对在线分析仪表存在测量滞后和不稳定,采用了最小二乘法、人工神经网络以及支持向量机建模技术,建立了乙烯纯度软测量的三种模型：PLS、RBFPLS及LS-SVM,并利用MATLAB系统,分别对三种模型进行离线仿真,分析了三种模型的优劣,最终得出最小二乘支持向量机模型(LS-SVM)适合用于乙烯精馏塔出口乙烯纯度估计。模型实施后,效果较好,能够为生产操作提供指导。
As modern petrochemical enterprise's control, metering, energy efficiency and operational reliability requirements and continuous improvement, measuring demand is growing. In the actual production process, there is a large class of variables, due to technical or economic and other reasons, it is difficult or not directly through the sensor for detecting, but is also needed to be strictly controlled, and the quality of the product very closely related process parameters. This will directly affect the quality of the product, as well as the stable operation of the system, for business economic loss should not be underestimated. Soft-sensor technology is rapidly developing.
In this article systematic, thorough research soft survey technology certain important aspects, take the least squares method and the neural network theory as the foundation, in the thorough understanding ethylene rectifying tower technical process and in the selective evaporation principle foundation, carries on the ethylene rectifying tower to export the ethylene purity soft measurement system the research. And was unstable in view of the online analysis measuring appliance existence survey lag, has used the least squares method, the artificial neural networks as well as the support vector machines modeling technology, has established the ethylene purity soft survey three kind of models.
本文研究了软测量技术的若干重要方面,以最小二乘法和神经网络的理论为基础,在深入了解乙烯精馏塔工艺流程及精馏原理的基础上,进行乙烯精馏塔出口乙烯纯度软测量系统的研究。针对在线分析仪表存在测量滞后和不稳定,采用了最小二乘法、人工神经网络以及支持向量机建模技术,建立了乙烯纯度软测量的三种模型：PLS、RBFPLS及LS-SVM,并利用MATLAB系统,分别对三种模型进行离线仿真,分析了三种模型的优劣,最终得出最小二乘支持向量机模型(LS-SVM)适合用于乙烯精馏塔出口乙烯纯度估计。模型实施后,效果较好,能够为生产操作提供指导。
As modern petrochemical enterprise's control, metering, energy efficiency and operational reliability requirements and continuous improvement, measuring demand is growing. In the actual production process, there is a large class of variables, due to technical or economic and other reasons, it is difficult or not directly through the sensor for detecting, but is also needed to be strictly controlled, and the quality of the product very closely related process parameters. This will directly affect the quality of the product, as well as the stable operation of the system, for business economic loss should not be underestimated. Soft-sensor technology is rapidly developing.
In this article systematic, thorough research soft survey technology certain important aspects, take the least squares method and the neural network theory as the foundation, in the thorough understanding ethylene rectifying tower technical process and in the selective evaporation principle foundation, carries on the ethylene rectifying tower to export the ethylene purity soft measurement system the research. And was unstable in view of the online analysis measuring appliance existence survey lag, has used the least squares method, the artificial neural networks as well as the support vector machines modeling technology, has established the ethylene purity soft survey three kind of models.
引文
[1]俞金寿,刘爱伦,张克进.软测量技术及其在石油化工中的应用[M].北京：化学工业出版社,2000
[2]Brosillow C B. Inferential Control of Process. AICHE[J],1978,24(3):485-509
[3]T. J. Mcavoy. Contemplative Stance for Chemical Process Control. Automatic[J]. 1992,28(2):441-442
[4]王旭东,邵惠鹤.神经元网络建模与软测量技术[J].化工自动化及仪表,1996,23(2)：28-31
[5]Tham M T. etal. Soft-sensing a Solution to the Problem of Measurement Delays [J]. Chem. Eng. Res and Des,1989,67(6):547-554
[6]Yingxu You and Tianyou Cai. Soft sensing based On artificial neural network [C]. Proceedings of the American Control Conference, Albuquerque, New Mexico, June 1997
[7]Luo Jianxu,Shao Huihe. Soft sensing modeling using neurofuzzy system based on rough settheory [c] ,Proceedings of the American Control Conference,Anchorage,AK May 8-10,2002
[8]杨金城,胡春.裂解深度神经网络控制系统的研究..石油化工自动化,2005,4：24-27
[9]鲁春燕.氧化铝输送中氧化铝粉流量的软测量方法研究.兰州理工大学.学位论文.2009
[10]刘刚.复杂热力系统数据建模方法的研究.华北电力大学(北京).学位论文.2008
[11]SH INSKEY F G. Distillation Control [M]. New York:McGraw-Hill,1984
[12]薛美盛等.精馏塔控制与节能优化研究综述.化工自动化及仪表,2006,4：15-19
[13]于静江,周春晖.过程控制中的软测量技术.控制理论与应用,1996,13(2)：137-144
[14]王宽全,徐宁,张大鹏.自动掌纹诊病的神经网络方法.计算机科学,2001,28(5)：104-106
[15]杨金城,胡春.裂解深度神经网络控制系统的研究.石油化工自动化,2005,4：24-27
[16]闫晶.基于MEDO粒子群算法的流量软测量RBF网络优化研究.燕山大学.学位论文.2009
[17]苏洪雨.基于聚类分析和多模型的传感器建模及其应用研究.东北林业大学.学位论文.2010
[18]关伟.基于神经网络的混合建模在发酵过程中的研究与应用.北京工业大学.学位论文.2009
[19]薄翠梅,张浞.基于神经网络的软测量技术在精馏塔上的应用.过程工程学报,2003,3(4)：371-375
[20]金强,杨凌,宋军.新型El man网络在乙烯精馏塔软测量建模中的应用.石油化工自动化,2008,6：35-38
[21]孙欣,王金春,何声亮.基于神经网络的过程软测量.自动化仪表,1996,15(9)：7-10
[22]徐科军,区建军,苏敬安,董联.预测铁水硅含量的神经网络初探.工业仪表与自动化装置,1996,3：8-11
[23]吴晓汉等.基于神经网络的软测量技术及应用.自动化与仪器仪表,1998,2：16-19
[24]丛爽,赵何.反向传播网络的不足与改进.自动化博览,1999,1：25-26,47
[25]Jagdish C.Patra. An Artificial Neural Network-based Smart Capacitive Pressure Sensor. Measurment. Vol 22, No 3-4, Nov-Dec 1997. pp:113-121
[26]蓝照斌,陈晓祥,史贤林.基于神经网络的树脂浓度控制.计算机自动测量与控制,2000,8(2)：24-25,28.
[27]吴国良.先进控制在乙烯精馏塔中的应用分析.石油化工自动化,2002
[28]王旭东,邵惠鹤.RBF神经网络理论及其在控制中的应用.信息与控制,1997,26(4)：273-283
[39]Galvan I M, Zaldivar J M. Application of Recurrent Neural Networks in Batch Reactor Part I:NARMA Modeling of the Heat Transfer Fluid Temperature. Chemical Engineering and Processing。1997,36(6):505-518
[30]高峰,彭双伟,王海洋,王立忠.乙烯精馏塔先进控制.当代化工.2002
[31]黄聪明,李志坚.改进的递归神经网络的化工动态系统建模.北京理工大学学报.2004.24(7)：596-599
[32]章卫国,杨向忠.《模糊控制理论与应用》[M].西北工业大学出版社,1999.10
[33]王惠文.《偏最小二乘回归方法及其应用》[M].中国人民大学出版社,1989.10
[34]王惠文,吴载斌,孟洁.《偏最小二乘回归方法及其应用》[M].中国人民大学出版社,1989.10
[35]K. A. Kosanovich, k. S. Dahl, M. J. Piovoso. Improve Process Understanding Using Multiway Principal Component nalysis [J]. Ind. Eng. chem. Res.1996,35 (1):138-146
[36]Tham. M. Tetal. Soft-Sensors for Process Estimation and Inferential ContrOl[J].Process Control,1991.1(1):3-14
[37]Kreisselmeier. G. Adaptive Observers with Exponential Convergence. IEEE Trans, Automat. Contr,1977. AC-22(1):2-8
[38]Guilandoust. M. T etc Adaptive Inferential Control. IEEE Proceedings 1987, 184(3):171-179
[39]Quineo. Marmol. etc. Application of an Extended Luenberger Observer to the Control of Multicompenent Barch Detilletion. Ind. Eng, Chem Res,1991,30(8): 1870-1880
[40]叶楠,吕勇哉.模式识别在状态估计中的应用一类软测量技术.仪器仪表学报,1988,9(4)：368-374
[41]穆罕默德.阿塔,蒋蔚孙.精馏塔板效率的在线模式识别.信息与控制,1993,22(1)：47-49
[42]周丽君,侯跃岭,敖晨虹.催化裂化装置稳定汽油饱和蒸气压的在线软测量与实时控制.内蒙古石油化工.2001
[43]俞金寿,刘爱伦,张克进.软测量技术及其在石油化工中的应用[M].北京：化学工业出版社,2000
[44]Cortes C, Vapnik V N. Support vector machine. Machine Learning,1995, (20):273-297
[45]Deniz O, Castrillon M, Hemandez M. Face Recognition Using Independent Component Analysis and Support Vector Machines. In:Proc. of International Conference on Audio and Video Based Person Authentication. Sweden:Lecture Notes,2001:59-64
[46]董恩清,赵鹤鸣,周亚同,张晓娣.支持向量机在语音激活检测中的应用研究.通信学报,2003,24(3)：70-77
[47]王李东,侍宏博.基于SVR算法的短期快速预测研究.继电器.2005,33(9)：17-21
[48]杜树新,吴铁军.回归型加权支持向量机方法及其应用。浙江大学学报,2004,38(3)；302-306
[49]张绍武,潘泉,张洪才,等.基于支持向量机和贝叶斯方法的蛋白质四级结构分类研究.生物物理学报,2003,19(2)：171-175
[50]高鹏飞.锅炉过热汽温的GGAP-RBF神经网络控制.重庆大学.学位论文.2010
[51]李颖新,朱云华,阮晓刚.基于支持向量机的肿瘤亚型识别.计算机工程与应用,2004,(17)：1-3
[52]邵信光,杨慧中,石晨曦.占不敏感支持向量回归在化工数据建模中的应用,东南大学学报(自然科学版),2004,34(s1)：215-218
[53]常玉清,王福利,吕哲.基于支持向量机的生物发酵过程软测量建模.东北大学学报(自然科学版),2005,26(11)：1025-1028
[54]王听.软测量技术及其在工业聚丙烯生产过程中的应用.浙江大学.学位论文.2006
[55]宋贤民.基于SVM的污水处理过程软测量建模研究.南昌大学.学位论文.2007
[56]T Mitchell著,曾华军,张音奎译.机器学习.北京：机械工业出版社,2003
[57]Suykens J A K Vandewalle J. Least squares support vector machine classifiers. Neural Processing Letters(S1370-4621),1999,9(3):293-300
[58]Suykens J A K Lukas L, Van Dooren P, De Moor B, Vandewalle J. Least square support vector machine dassifiers:a large scale algorithm. Proceedings of the European Conference on Circuit Theory and Design (ECCTD'99), Stresa, Italy,1999, (9):839-842
[59]Vapnik V N著,张学工译.统计学习理论的本质.北京：清华大学出版社,2000[60]Vapnik V N, Levin E, Le Cun Y. Measuring the VC dimension of a learning machine. Neural Computation,1994, (6):851-876
[61]李学迁.一种基于RS-SVM的工程项目风险评价方法.南昌大学.学位论文.2007
[62]Kok Seng Chua. Efficient computations for large least squares support vector machine classifiers. Pattern recognition,2003, (24):75-80
[63]Hameta B, Suykens J A K De Moor B. A comparison of iterative methods for least squares support vector machine classifiers. Technical Report 01-110, ESAT-SISTA, K. U. Leuven(Leuvcn, Belgium),2001
[64]Suykens J A K, Lukas L, Vandewalle J. Sparse least squares support vector machines, ESANN'2000, Bruges, Belgium,2000:37-42
[65]Zeng Xiangyan, Chen Xuewen, SMO—based pruning methods for sparse least squares support vector machines. IEEE transactions on Neural Networks, 2005,16(6):1541-1546
[66]Hoegaerts L, Suykesn J A K Vandewalle J, De Moor B. A. comparison of pruning algorithms for sparse lest squares support vector machines. ICONIP 2004, Caloutta, India,2004,(6):37-42
[67]姚崇龄.基于神经模糊系统的多模型建模方法及在软测量中的应用.厦门大学.学位论文.2007
[68]董纯.软测量技术及其在聚醋酸乙烯聚合度的在线检测中的应用.上海交通大学.学位论文.2008
[69]刘进军.RBF神经网络的改进及其应用研究.兰州大学.学位论文.2008
[70]宋军.基于递归神经网络的石化生产软测量建模方法研究.兰州大学.学位论文.2008
[71]刘松青.基于LS-SVM的软测量建模方法研究.南京理工大学.学位论文.2008
[2]Brosillow C B. Inferential Control of Process. AICHE[J],1978,24(3):485-509
[3]T. J. Mcavoy. Contemplative Stance for Chemical Process Control. Automatic[J]. 1992,28(2):441-442
[4]王旭东,邵惠鹤.神经元网络建模与软测量技术[J].化工自动化及仪表,1996,23(2)：28-31
[5]Tham M T. etal. Soft-sensing a Solution to the Problem of Measurement Delays [J]. Chem. Eng. Res and Des,1989,67(6):547-554
[6]Yingxu You and Tianyou Cai. Soft sensing based On artificial neural network [C]. Proceedings of the American Control Conference, Albuquerque, New Mexico, June 1997
[7]Luo Jianxu,Shao Huihe. Soft sensing modeling using neurofuzzy system based on rough settheory [c] ,Proceedings of the American Control Conference,Anchorage,AK May 8-10,2002
[8]杨金城,胡春.裂解深度神经网络控制系统的研究..石油化工自动化,2005,4：24-27
[9]鲁春燕.氧化铝输送中氧化铝粉流量的软测量方法研究.兰州理工大学.学位论文.2009
[10]刘刚.复杂热力系统数据建模方法的研究.华北电力大学(北京).学位论文.2008
[11]SH INSKEY F G. Distillation Control [M]. New York:McGraw-Hill,1984
[12]薛美盛等.精馏塔控制与节能优化研究综述.化工自动化及仪表,2006,4：15-19
[13]于静江,周春晖.过程控制中的软测量技术.控制理论与应用,1996,13(2)：137-144
[14]王宽全,徐宁,张大鹏.自动掌纹诊病的神经网络方法.计算机科学,2001,28(5)：104-106
[15]杨金城,胡春.裂解深度神经网络控制系统的研究.石油化工自动化,2005,4：24-27
[16]闫晶.基于MEDO粒子群算法的流量软测量RBF网络优化研究.燕山大学.学位论文.2009
[17]苏洪雨.基于聚类分析和多模型的传感器建模及其应用研究.东北林业大学.学位论文.2010
[18]关伟.基于神经网络的混合建模在发酵过程中的研究与应用.北京工业大学.学位论文.2009
[19]薄翠梅,张浞.基于神经网络的软测量技术在精馏塔上的应用.过程工程学报,2003,3(4)：371-375
[20]金强,杨凌,宋军.新型El man网络在乙烯精馏塔软测量建模中的应用.石油化工自动化,2008,6：35-38
[21]孙欣,王金春,何声亮.基于神经网络的过程软测量.自动化仪表,1996,15(9)：7-10
[22]徐科军,区建军,苏敬安,董联.预测铁水硅含量的神经网络初探.工业仪表与自动化装置,1996,3：8-11
[23]吴晓汉等.基于神经网络的软测量技术及应用.自动化与仪器仪表,1998,2：16-19
[24]丛爽,赵何.反向传播网络的不足与改进.自动化博览,1999,1：25-26,47
[25]Jagdish C.Patra. An Artificial Neural Network-based Smart Capacitive Pressure Sensor. Measurment. Vol 22, No 3-4, Nov-Dec 1997. pp:113-121
[26]蓝照斌,陈晓祥,史贤林.基于神经网络的树脂浓度控制.计算机自动测量与控制,2000,8(2)：24-25,28.
[27]吴国良.先进控制在乙烯精馏塔中的应用分析.石油化工自动化,2002
[28]王旭东,邵惠鹤.RBF神经网络理论及其在控制中的应用.信息与控制,1997,26(4)：273-283
[39]Galvan I M, Zaldivar J M. Application of Recurrent Neural Networks in Batch Reactor Part I:NARMA Modeling of the Heat Transfer Fluid Temperature. Chemical Engineering and Processing。1997,36(6):505-518
[30]高峰,彭双伟,王海洋,王立忠.乙烯精馏塔先进控制.当代化工.2002
[31]黄聪明,李志坚.改进的递归神经网络的化工动态系统建模.北京理工大学学报.2004.24(7)：596-599
[32]章卫国,杨向忠.《模糊控制理论与应用》[M].西北工业大学出版社,1999.10
[33]王惠文.《偏最小二乘回归方法及其应用》[M].中国人民大学出版社,1989.10
[34]王惠文,吴载斌,孟洁.《偏最小二乘回归方法及其应用》[M].中国人民大学出版社,1989.10
[35]K. A. Kosanovich, k. S. Dahl, M. J. Piovoso. Improve Process Understanding Using Multiway Principal Component nalysis [J]. Ind. Eng. chem. Res.1996,35 (1):138-146
[36]Tham. M. Tetal. Soft-Sensors for Process Estimation and Inferential ContrOl[J].Process Control,1991.1(1):3-14
[37]Kreisselmeier. G. Adaptive Observers with Exponential Convergence. IEEE Trans, Automat. Contr,1977. AC-22(1):2-8
[38]Guilandoust. M. T etc Adaptive Inferential Control. IEEE Proceedings 1987, 184(3):171-179
[39]Quineo. Marmol. etc. Application of an Extended Luenberger Observer to the Control of Multicompenent Barch Detilletion. Ind. Eng, Chem Res,1991,30(8): 1870-1880
[40]叶楠,吕勇哉.模式识别在状态估计中的应用一类软测量技术.仪器仪表学报,1988,9(4)：368-374
[41]穆罕默德.阿塔,蒋蔚孙.精馏塔板效率的在线模式识别.信息与控制,1993,22(1)：47-49
[42]周丽君,侯跃岭,敖晨虹.催化裂化装置稳定汽油饱和蒸气压的在线软测量与实时控制.内蒙古石油化工.2001
[43]俞金寿,刘爱伦,张克进.软测量技术及其在石油化工中的应用[M].北京：化学工业出版社,2000
[44]Cortes C, Vapnik V N. Support vector machine. Machine Learning,1995, (20):273-297
[45]Deniz O, Castrillon M, Hemandez M. Face Recognition Using Independent Component Analysis and Support Vector Machines. In:Proc. of International Conference on Audio and Video Based Person Authentication. Sweden:Lecture Notes,2001:59-64
[46]董恩清,赵鹤鸣,周亚同,张晓娣.支持向量机在语音激活检测中的应用研究.通信学报,2003,24(3)：70-77
[47]王李东,侍宏博.基于SVR算法的短期快速预测研究.继电器.2005,33(9)：17-21
[48]杜树新,吴铁军.回归型加权支持向量机方法及其应用。浙江大学学报,2004,38(3)；302-306
[49]张绍武,潘泉,张洪才,等.基于支持向量机和贝叶斯方法的蛋白质四级结构分类研究.生物物理学报,2003,19(2)：171-175
[50]高鹏飞.锅炉过热汽温的GGAP-RBF神经网络控制.重庆大学.学位论文.2010
[51]李颖新,朱云华,阮晓刚.基于支持向量机的肿瘤亚型识别.计算机工程与应用,2004,(17)：1-3
[52]邵信光,杨慧中,石晨曦.占不敏感支持向量回归在化工数据建模中的应用,东南大学学报(自然科学版),2004,34(s1)：215-218
[53]常玉清,王福利,吕哲.基于支持向量机的生物发酵过程软测量建模.东北大学学报(自然科学版),2005,26(11)：1025-1028
[54]王听.软测量技术及其在工业聚丙烯生产过程中的应用.浙江大学.学位论文.2006
[55]宋贤民.基于SVM的污水处理过程软测量建模研究.南昌大学.学位论文.2007
[56]T Mitchell著,曾华军,张音奎译.机器学习.北京：机械工业出版社,2003
[57]Suykens J A K Vandewalle J. Least squares support vector machine classifiers. Neural Processing Letters(S1370-4621),1999,9(3):293-300
[58]Suykens J A K Lukas L, Van Dooren P, De Moor B, Vandewalle J. Least square support vector machine dassifiers:a large scale algorithm. Proceedings of the European Conference on Circuit Theory and Design (ECCTD'99), Stresa, Italy,1999, (9):839-842
[59]Vapnik V N著,张学工译.统计学习理论的本质.北京：清华大学出版社,2000[60]Vapnik V N, Levin E, Le Cun Y. Measuring the VC dimension of a learning machine. Neural Computation,1994, (6):851-876
[61]李学迁.一种基于RS-SVM的工程项目风险评价方法.南昌大学.学位论文.2007
[62]Kok Seng Chua. Efficient computations for large least squares support vector machine classifiers. Pattern recognition,2003, (24):75-80
[63]Hameta B, Suykens J A K De Moor B. A comparison of iterative methods for least squares support vector machine classifiers. Technical Report 01-110, ESAT-SISTA, K. U. Leuven(Leuvcn, Belgium),2001
[64]Suykens J A K, Lukas L, Vandewalle J. Sparse least squares support vector machines, ESANN'2000, Bruges, Belgium,2000:37-42
[65]Zeng Xiangyan, Chen Xuewen, SMO—based pruning methods for sparse least squares support vector machines. IEEE transactions on Neural Networks, 2005,16(6):1541-1546
[66]Hoegaerts L, Suykesn J A K Vandewalle J, De Moor B. A. comparison of pruning algorithms for sparse lest squares support vector machines. ICONIP 2004, Caloutta, India,2004,(6):37-42
[67]姚崇龄.基于神经模糊系统的多模型建模方法及在软测量中的应用.厦门大学.学位论文.2007
[68]董纯.软测量技术及其在聚醋酸乙烯聚合度的在线检测中的应用.上海交通大学.学位论文.2008
[69]刘进军.RBF神经网络的改进及其应用研究.兰州大学.学位论文.2008
[70]宋军.基于递归神经网络的石化生产软测量建模方法研究.兰州大学.学位论文.2008
[71]刘松青.基于LS-SVM的软测量建模方法研究.南京理工大学.学位论文.2008